Risk Management Exam

Risk Management
Fourth Edition
Anthony C Valsamakis
Eikos Risk Capital Limited, London
Robert W Vivian
School of Economics and Business Sciences University of the
Witwatersrand
Gawie S du Toit
College of Economic and Management Sciences
University of South Africa
Published by Heinemann Publishers (Pty) Ltd Every effort has been made to trace the copyright holders. In the event of unintentional omissions or
Heinemann House errors, any information which would enable the publisher to make the proper arrangements will be
Grayston Office Park appreciated.
Building No. 5
128 Foreword
Peter
Road
Sandton
e in text: the authors

@ in published edition: Heinemann Publishers (Pty) Ltd, 2010 Among the many duties of directors is the duty to take risk for reward. That, after all, is the essence of
business. Pension funds, on behalf of their ultimate beneficiaries, buy equities in companies because of
All rights reserved. No part of this publication may be reproduced, stored in a retrievaE system or transmitted in any
form oc by any means, electronic, mechanical, photocopying or otherwise without the prior written permission of the the expectation that their beneficiaries will receive a greater return than investments in the money market.
publisher.
In taking risk for reward, directors are dealing with foresight. By definition, this means dealing with
First edition published 1991 uncertain future events, which are shrouded in risk. Hence the wisdom ofthe legal principle that a court
Second edition published 2000
Third edition published 2005
should be slow to substitute its hindsight for the foresight of directors. There is much comfort in the chair
Fourth edition published 2010 of hindsight, compared with the heat of foresight in the boardroom.
16 15 14 While the board is responsible for the governance of risk and should create the framework for the
876
risks pertinent to the business of the company, the implementation of those processes is a management
ISBN 9780796231 185
function. The board must ensure that the framework or policy of risk that it lays down will support the
Typeset in 1 1/ T 3 Minion Pro by Boss Repro and Design Studio company's long-term strategic plan for its business.
Printed by Creda
The board should determine the level of risk tolerance. It should set limits for the levels of risk which
the company is prepared to pursue. Consideration must be given to both external and internal factors
impacting on the business of the company. Likewise, the board should set limits in regard to the company's adopt a recognised methodology, identifying risks and opportunities and measuring the impact on the
risk appetite. business of the company.
A risk committee, and if one cannot be established, the audit committee, should assist in the discharge The risk management plan should have the attributes of insight, information, incentives, instinct
of the duties and responsibilities in respect of risk management. A risk committee should consist of both (avoid following the herd), independence and interconnectivity.
the executive and non-executive directors, but essentially it is a management function. Consideration has to be given to avoiding risk, treating, reducing or mitigating risk, transferring the
The board should also ensure that a company has and maintains an effective on-going risk assessment risk to a third party, tolerating or accepting risk, exploiting the risk where the risk exposure represents
process, consisting of risk identification, risk quantification and risk evaluation. Such assessment should potential opportunity, terminating the risk or integrating some or all of the risk responses.
Risk can be seen in the categories of compliance, financial, operationalAcknowledgements:
strategic, human resource and sustainability. The latter two contain huge risksAs we embark on this, the fourth edition, it is opportune to reflect on the for the modern company. Head hunters search a borderless,
electronic worldcommercial practice of risk management and acknowledge the valuable input for skills. The loss of certain skills to a company entails risk. Governance,of professional colleagues and clients who have
all contributed over a period strategy and sustainability have become inseparable. For example, a beveragethat spans more than 20 years, in first creating and then converting theory manufacturer cannot plan long-term
without ensuring accessibility to potableinto practical reality. Thanks are due to colleagues of the Eikos Group, which, water. The combination of these three elements makes it essential to identifyas an independent
consultancy practising in this arena, is testament to this. and manage the risks involved. The scarcity of water on planet earth is an obvious sustainability risk to the beverage manufacturer.
The Global Reporting Initiative's G3 Guidelines have some 80 indicators. A survey has shown that
companies which have studied those indicators improved their risk management, because invariably an issue
which they had not thought of either directly or tangentially arises to improve their risk management.
The book Risk Management covers the concept of risk, risk identification, risk evaluation, the
insurance of risk and strategies involving risk. The book is essential reading for the modern corporate
executive.
Prof Mervyn E King SC

Contents Part I
Fundamental principles
Part l: Fundamental principles
Chapter 1 Principles of management applied to managing risk 2

Chapter 2 Concept of risk 26
Chapter 3 Decision-making under conditions of risk and uncertainty 55
Chapter 4 Corporate governance and enterprise risk management 79
105
Part Il: Managing risk
Chapter 5 Risk identification 106
Chapter 6 Risk evaluation 123
Chapter 7 Operational risk management 1 34
Chapter 8 Risk response: Risk control 152
Part Ill: Risk financing 220

Chapter 9 Risk financing 221
Chapter 10 Risk retention 238
Chapter 1 1 Captive insurance companies 259
Chapter 12 Insurance 312
Chapter 13 Finite insurance 336
Chapter Capital market instruments 348
Chapter 15 Composite financing strategies 361
Bibliography 387
Index 395
PRINCIPLES OF MANAGEMENT APPLIED TO MANAGING RISK
2 index representative of the entire market. A company that is closely related to the market will,
as far as the systematic component of its risk is concerned, follow the fortunes of the market
very closely. The factors that cause the market to increase and decrease will have a similar effect
on this type of company. Other companies are less sensitive to market volatility. Their share
price (and therefore their return to investors) will move only fractionally in comparison to
Principles of market movements. The factors that affect the entire market and therefore the systematic part
of a company's risk are general economic factors such as interest rates, inflation and economic
management applied to cycles.
The remainder of a company's risk is the unsystematic part. This is that part of the total risk
managing risk that is specific to a particular company. It is not caused by general factors that affect the entire
market, but by factors that are specific to a particular company or industry and its type of
business. For example, the fortune ofa company selling fashion goods is closely linked to
demand changes in that particular market.
While most of the technical aspects of risk management can be bought, rented or borrowed,
it is the management of these elements that cannot be outsourced. The purpose of this chapter
and the point of departure ofthis book are therefore to explain how the unsystematic or specific
risks of a company are managed. The following sections in this chapter therefore focus on the
case for risk management, the functions of management and how it applies to the management
of risk in an organisation. The chapter finally gives an overview of the development of risk
management in the United States (US), the United Kingdom (UK) and South Africa.
1 .1 Introduction 1.2 Case for risk management

Noted management guru Peter Drucker observed at a conference of risk managers that risk
Risk management in a loose sense is the art and science of managing risks. Since people are by
management may be as important as entrepreneurship and business acumen in propelling the
nature risk averse, they have always found ways, at least in part, to manage risks. In this sense,
economic grovnh of the Western world. He argues that society's ability to manage fortuity the
then, risk management is as old as humankind itself. As a systematic and holistic business
unexpected — is probably one of the characteristics that most distinguishes the developed from
discipline, however, it is a modern development.
the developing nations. Disasters befall rich and poor nations alike, but a society that is able to
Entrepreneurs and shareholders accept the risk associated with a certain type of business control and cushion itself against such events is better able to deploy its resources toward
in order to earn the reward associated with that particular type of business or industry. For economic and social advancement.
example, a person investing in the construction industry will be aware that this industry is In Against the Gods: The Remarkable Story ofRisk, Peter Bernstein makes a similar point
subjected to cyclical swings in demand but that the return is generally higher than that of stable when he says: 'What is it that distinguishes the thousands of years of history from what we
industries such as electricity supply or food. think of as modern times? Ihe revolutionary idea that defines the boundary between modern
Investors are exposed to two broad categories of risk, namely systematic risk and times and the past is the mastery of risk'.
5
4
unsystematic (or specific risk). Systematic risk is a market-related risk. It gives the relationship
between company performance and that of an
3
PRINCIPLES OF MANAGEMENT APPLIED TO
RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES MANAGING RISK
To place the importance of risk management into perspective, we only Consequence
need to consider table I. i, with its examples of the causes and consequences of human-made 1959—61 6 000 malformed
risks. 1985:
Without further elaboration, these incidents provide suffcient evidence of the devastating King's Cross underground station firet London, UK
real-life effects of the consequences of event risks and of the need to manage these risks
effectively.
1986:
Table 1.1 Causes and consequences of risks
Use of medical drug widely prescribed babies were born in 20

Thalidomide sedative drug that countries. The producer Kinross Gold Mine fire, South Africa
caused genetic damage of the drug paid 1988: Piper AEpha off-shore rig explosion, North Sea
in the babies of women who compensation.

took the drug.
1960: A coal-mining disaster 435 miners were

Coalbrook mining occurred when part of trapped underground 994: Merriespruit slimes dam disaster, Free State, South Africa
disaster, South Africa the mine collapsed. and alf attempts to rescue
them failed.
For unknown reasons, a 736 people were
1976:
Release of dioxin (dioxin safety valve in a factory evacuated and a
1995:
is 10 000 times more allowed the discharge ban was placed on
A locomotive fails down
toxic than cyanide), into the atmosphere the consumption of
Sevesor Italy of trichlorphenoi food produced in the a mine shaft, Vaai Reefs mine, South Africa 11 September 2001:
derivatives, including area. Compensation
dioxin. payments of
approximately R480m
were made. 250 000 m3 of
contaminated soil had to
be buried.
High rates of loss of life 1 500 people were killed

1984:
Leak of methyliso- and injuries occurred and 34 000 people
cyanate, Bhopal, India partly due to the fact sustained eye injuries,
that many people lived bums, and damage to
in a shanty town that respiratory tracts and
had been built around nervous systems.
the factory. 200 000 people had to leave

the area.
Accumulated grease 31 people died
on the running tracks (including one fire
of wooden escalators offcer) and many more
was ignited (perhaps RISK by MANAGEMENT:
a were injured.
FUNDAMENTAL PRINCIPLES
match) causing a quickly MANAGING RISK
spreading fire.
Welding operations 177 miners died in the
ignited polyurethane gas„filled mine.

foam that had been
installed in the mine. The
resulting fire released a
gas into the mine.
An explosion, resulting 165 men were killed, the

in fire and followed by platform was destroyed
further explosions, was and a total loss of
caused by an operator approximately RI 6bn
error during restart after a was sustained.
safety valve had been
removed.
The property of the 17 people were killed

Harmony Gold Mining and an estimated
Company, a mining R250m in damages was
slimes dam burst its sustained.
banks, causing sludge to
flood the mining town.
A locomotive feti down a 104 people were killed.

mine shaft at speed.
A sophisticated terrorist attack was launched in the US
when four aircraft were hijacked, two of which crashed into
the World Trade Center. This caused the twin towers to
collapse and the deaths of thousands of people.
In addition to the above tragic incidents are the huge corporate losses that occurred because of
fraud, mismanagement or other causes such as the well-
The case for an integrated and enterprise-wide approach to managing risk becomes more
prominent when the following trends are considered.
known 1 incidents at Barings Bank, Enron, HIH and, in South Africa, AA Mutual, MacMed,
LeisureNet and Saambou.
1
Dickson (1989: 20}.
Increasing sophistication of risk There has been a corresponding growth in captive insurance companies of more than 300% over
the period 1980-96. This is more than double the growth of the insurance market for the same
Risk has become more sophisticated recently through increased automation; extensive use of
period. Captive insurance companies increased by 17% from 1996 to 1997, and according to
robotics; and the discovery of new chemicals, treatments and processes, all of which spell
figures published by Tillinghast-Towers Perrin, there were 3 966 captives world-wide at the end
potential new risks. The danger exists Dicksonthat thel effective management of risk is sacrificed
of 1996 — a net gain of 171% over the previous year — generating approximately $18 billion
in the quest for progress. proposes the prevalence of a 'leapfrog' mentality concerning risk
of written premiums. By the year 2007, there were 5 119 captive insurers.
progress
This trend in itself dictates that the risk control and alternative riskfinancing aspects be
and change are left unchecked until a major incident occurs, prompting risk management to integrated and that trade-offs between risk retained and self-funded, on the one hand, and risk
leapfrog into the forefront of attention. insured, on the other, be understood and managed optimally.
Increasing concentration of risk 1.3 Reasons to manage risk

This concentration can take place in a number of ways although one of the most obvious is that
of complex functions inherent in the momentous advances in silicon-chip technology. This Introduction
concentration entails added risk, as does that resulting from the large size of buildings, Before describing the fundamental principles of risk management, the reasons for managing
manufacturing plants and organisations generally. Individual values and potential liabilities have risks in general are considered, followed by criteria to be considered when making specific
increased enormously over recent years. decisions.
In general, the reasons are linked to the corporate objectives of survival, growth and
maximising shareowner wealth. Corporate policy and good citizenship are factors that
Increasing awareness of risk influence decisions.4
Ihe public is more aware of risk and safety today. The impact of major disasters has tended to Risk and return are interrelated. any reduction in the risk profile of an organisation (against
accelerate the move toward what could be termed 'risk consumerism'. Particularly in the US, a given expected return) following a deliberate risk management programme will result in a
more efficient risk-return trade-off.5Thus, adopting a risk management programme that reduces
social and political forces have moulded what Kloman23 describes as an 'entitlement society' in
risk is of itself consistent with the general reasons for the existence of a firm. It is thus not
which citizens suffering misfortune refuse to accept personal accountability, turning instead to surprising that the adoption of a risk management programme features in most codes of
others and to the state for reimbursement or reward. corporate governance.
This, then, can be seen as the fundamental reasoning supporting the existence of a risk
management philosophy or programme within a firm. In theory, because investment decisions
Decline of insurance as a risk-financing technique entail the achievement of effcient riskto-return trade-offs, this dictates that risk and its
There has been a steady decline in the use of commercial insurance. Expenditure on insurance reduction per se should be considered as equally fundamental and significant.
relative to expenditure on the risk control and risk retention has 8
7
reduced. In 1977, for some of the Fortune 500 US corporations, the cost of property insurance
premiums comprised 26,3% of the 'cost of risk'. By 1996, this figure had stabilised to 14, 1%.
2
(1984: 34). 5
for a discussion of the effect of pure rlsk on the value oi the share price firmr see Main { 1 982- 237—47)
3
Captives are discussed at length in chapter 1 1, A optive is an insurance company w.med by a parent company to underwrlte (he parents risk
4
Mehr and &dges (1974:29-37).
Justification of specific decisions 9

If we accept that the general theory of the firm translates into making specific decisions to
reduce risk, the next step is thus to identify criteria to be applied by the risk manager when that, notwithstanding possible ineffciencies of, for exampl% purchasing insurance,
considering specific proposals aimed at reducing risk. Four such criteria will be discussed. organisations tend towards the purchase of insurance to 'transfer risk" Such behaviour
reinforces the view that risk aversion indeed influences the risk manager, an influence that
Steps taken are cost beneficial extends beyond that of risk 'transfer* to the insurance market or the implementation of a
The most common reason given for taking any decision is that steps that are taken to reduce deliberate retention funding to risk control steps.
risks are cost beneficial.6 This criterion supports the risk-toreturn premise outlined above and Risk management activities generally involve expenditure of funds rather than the creation
of income. Even excellent performance by the risk manager is often unidentified, but failure to
finds favour among risk managers who are finance orientated. Simplistically, this decision rule
recognise a source of loss or to take steps to prevent such loss and failure to arrange adequate
hinges upon whether the return or benefit derived from risk management measures outweighs
risk funding becomes readily apparent to top management and shareholders after the event
the expenditure incurred to reduce the risk. occurs. Individuals' risk aversion, therefore, also has a bearing on how and why risks are
Since the return is usually dependent on the probability ofa future outcome, probability managed.
theory discussed below, is relevant. This approach requires that a risk-return calculation be
carried out. However, an important qualification needs to be made. Although the cost-to-benefit Policy-based decisions
principle is sound, the emphasis should be on maxi-mising or achieving risk-return trade-off In practice many decisions to implement risk management programmes are simply based on
effciency in the longer term. Emphasis should not necessarily be placed on short-term cost- organisational policy formulated for reasons other than risk. Quite often, therefore, the decision
tobenefit relationships; the error in this practice is the disregard of the principle and its that reduces risk may not be motivated by risk management concerns at all. For instance, a
substitution for a short-term accept-or-reject criterion. company may be motivated to design and implement a quality assurance program-me because
Nevertheless, considering that expenditure on and benefits derived from risk management it strives for excellence rather than because it is concerned about the liability exposure of its
decisions cannot always be evaluated in the shorter term, the aim of minimising the long-term products. Similarly, the decision to introduce an employee safety programme and follow a
National Occupational Safety Association programme may well be a policy-based decision
cost of risk and thereby attaining an efficient long-term risk-to-return relationship remains the
arising from the organisation's concern about occupational injuries, despite the fact that section
fundamental principle of risk management.
35 of the Compensation for Occupational Injuries and Diseases Act has no liability exposure. It
is not suggested that this attitude is wrong or ineffectual, but rather that deliberate company
Risk aversion policy can have an unintentional impact on the risk management function. The risk manager
The concept of diminishing marginal utility states that people value more greatly what they should support, encourage and monitor the implementation of any corporate programme that
have than what they may gain in the future. People are thus prepared to sacrifice part of their reduces risk.s This results in the integration of company strategy and risk management.
existing wealth to safeguard the remaining portion of it. In short, people are risk averse and this
drives them to take steps to reduce risk, even if these steps are not strictly cost beneficial.
Hence, even under given cost-to-benefit conditions, an aversion to risk per se often dictates
aspects of the risk management programme. One finds
6 Most literature re&rs to cost-to-benefit being the main reason. or at the least Implies this. 'Lidging from the structure of the work For example
7 The influence of risk aversion •n the decision-making mocess of pu rchasing insuraru:e has been researcPed by numerous writers.The sernlnal on the atiitude
tahkrty (1985: 365) of (he firm; StaRes.- loss 'The reduction criteria for decisions makir•g can loss be reduction nude acccwding decisions to are chelf the impact same towards risk appears to be by Grayson ( 1 960). See also WI" lams (1986'577-9, Gordon Dickson (1982: 89), Smith and (1987-225). and sources quoted by Kihlstrom and
asthose on value used for of loss the firm's finarnr-ua existing The equity: common Greene yardstickand 6 the value Roth {1982 361}. 8 See. for example Vernon (1988).
Serbe.in (1 983: 02) recognise two types of Ulteria for the ratonale of loss amtrol. (i) economic and non-economtc. Mehr and Hedges (1974: 439 recognise cost-to-
benefit and non-I-rmetary criteda. Within the legislative requirements are considered.
MANAGEMENT: FUNDAMENTAL PRINCiPLES PRINCIPLES OF MANAGEMENT APPLIED TO MANAGING RISK
10 resources, it stands to reason that each one should have an organisational structure that will
accommodate its particular needs. Management must match the organisation's structure to
its resources, goals and strategies. This process is called organisational design.
Authoritative reasons
Finally, with regard to certain risks, e.g. fire, occupational safety or products, there are Leading refers to directing the human resources of the organisation and motivating them in
such a way that their actions accord with previously formulated goals and plans. Managers
authoritative reasons why risk management measures are implemented. Commonly, these are
do not act in isolation and do not only give orders — they collaborate with their superiors,
imposed through legislation* and accepted codes of practice.
equals and subordinates, and with individuals and groups to attain the goals of the
organisation. Leading the organisation means making use of influence and power to
motivate employees to achieve organisational goals. Leading means communicating goals
1.4 Functional approach to management throughout the organisation and motivating departments and individuals to perform as well
as they possibly can.
Functions of management
Controlling means that managers should constantly make sure that the organisation is on the
Risk management, as the term implies, is the management of risk. Seen in this light, it is the
right course to attain its goals. 'Ihe aim of control is therefore to check that performance and
application of well-established principles of management to the management of risk. It is thus
action conform to plans to attain the predetermined goals. Control also enables management
logical to briefly discuss these principles of management. Management is defined as the process to identify and rectify any deviations from the plans, and to take into account factors that
of planning, organising, leading and controlling the resources of the organisation to achieve might oblige them to revise their goals and plans.
ll
stated organisational goals as efficiently as possible.
The term 'manager' refers to anybody who carries out the four fundamental functions of management
A concise description of each of the fundamental management functions will further
described above.
explain the concept of management and the nature of the management process.
Planning is the management function that determines the organisation's mission and goals. Levels of management
It entails determining the future position of the organisation and the strategies needed to
Managers perform these functions at different levels in organisations.
reach that position. Hence the activities of the organisation cannot be performed in a
random fashion, but should follow a specific, logical, scientific method or plan.
At the strategic level, managers identify the organisation's mission, its goals and objectives, its
strategic plan and the evaluative processes used to measure the organisation's progress towards
Organising is the second step in the management process. Once the goals and plans have
its goals. nese activities are the responsibility Of top management. This is usually a relatively
been determined, management has to allocate the organisation's human and physical
small group of managers with whom the final responsibility and authority for executing the
resources to relevant departments or individuals. Duties have to be defined and procedures
management process rest. Top management is mainly concerned with long-term planning,
established to achieve the objectives. 'Thus, organising involves developing a framework
designing the organisation's broad structure, leading the organisation and controlling it.
or organisational structure to indicate how people and materials should be deployed to
achieve the goals. The success of an organisation lies in directing the different resources
The operations management function comprises those activities that move the organisation
towards the achievement of a common goal. The better the resources are coordinated and
towards its mission. Operations management concerns itself with the process of providing goods
organised, the more successful the organisation will be. Because organisations have
and services. Two levels of management are responsible for these activities. Middle management
different goals and
is responsible for
12 RISK
9 See chapter 8 for a list of laws that have an impact on risk management issues.
10 Health and safety In South Africa. for example, are governed by the Occupational Health and Safety Act and regulation* kn the US, safev„• and
specific departments and is primarily concerned with implementing policies and strategies
occupational health is governed by the Occupational Safety and Health Adminttratm and safety standards 11 Smit and De cronje
formulated by top management. It performs medium- and short-term planning, organising
11
functional areas, leading by means of departmental heads and controlling the activities of the
middle managers' own departments. Lower management or first-line management is management level, this requires a full understanding of risk management principles and the
responsible for smaller segments of the organisation, namely subsections. The management embedding of a risk management culture in the organisation. It also means that risk management
activities should be integrated with other business functions and risk control integrated with the
activities of lower management focus on daily activities. Its primary function is to implement
risk-financing activities, as well as self-funding with other ways of financing risks.
the plans of middle management, to apply procedures and rules, to provide technical assistance,
The interaction of strategic activities with operational activities is also essential. Failure to
to motivate subordinates and to accomplish day-to-day objectives. involve the operational level can result in a fragmented or contradictory approach in which a
business may face unnecessary risks. Reward systems must be designed to encourage risk
awareness and reward behaviour that is consistent with business objectives.
1.5 Definition of risk management
Global competition and turbulent markets with associated high levels of volatility have raised Proactive
risk management to the forefront of business thinking. The substantial costs of failure and the
Risks must be anticipated in advance and be catered for properly through both risk control and
equally large benefits that accrue from managing the ratio of reward to risk highlight the financing arrangements. In this way, risk management becomes an integral part ofgeneral
importance of risk management for the organisation. management, as opposed to a set of isolated functions comprising risk control and risk financing.
'This volatile environment requires an enterprise-wide approach to risk management that In order to reflect the managerial nature of the function and the need for an integrated
is comprehensive, inclusive and proactive. Each one of these requirements will be discussed approach to risk management, the following definition of risk management is adopted in this
book:
briefly.
Definition
Comprehensive Risk management is a managerial function aimed at protecting the orga nisation and its
To be comprehensive, an integrated approach to risk management requires that three key people, assets and profits against the physical and financial consequences of risk. It involves
planning, coordinating and directing the risk-control and the risk-financing activities in the
aspects of business organisations are considered, namely:
organisation.
(i) its strategy (ii)
the processes This definition firstly emphasises the managerial nature of the function by explicitly stating that
(iii) its people. risk management involves the planning, coordinating and directing of the risk-control and risk-
financing activities. Implicit to the definition is that, since risk management is a management
function, the risk manager will be involved in strategic decision-making. It also meets the
Inclusive requirement that people and processes should be involved by specifically referring to the
To be inclusive, risk management must involve all the levels of the organisation. At the protection of people, assets (and processes) and profits in the definition.
strategic level, it requires that the risk-to-reward ratio for all types of risk be considered, where Although one could claim that the field of risk management is rather definite in scope, the
the company's board of directors must play a leading role in setting a clear risk framework. activities it involves are both broad and complex. Successful risk management implies the
The framework should provide management with a guide that relates investors' expectations to employment of qualified personnel,
14 RISK
the risk-to-reward ratio. This framework should also provide guidelines as to what level of risk
will be accepted, which risks will be transferred by contract or negotiation, and which risk
should be insured. usually specialists in the varied techniques of risk management (control and financing) and
13 across the different areas of risk (see the following chapter on the nature of risk).
We can now proceed to consider the model for the management of risk that this definition
prescribes.
Once these guidelines have been set, they have to be fully integrated into the processes of
the organisation so that management and day-to-day activities support this vision. At the
frB*1.6 Risk management model Determining the objectives of the risk management function
The alignment of risk management goals and objectives with the mission of the organisation is a
Although the actual practice of risk management varies among organisations, certain elements
fundamental task of the risk manager. Tlis part of the risk management process identifies the
are common to all risk management programmes. Together, these elements comprise the risk
relationship between risk management and the mission of the organisation. The establishment of
management function, which is also referred to as the risk management model or the risk
risk management goals and objectives is critical, for they serve as the foundation for all risk
management process. This is depicted in figure 1. l . Figure 1.1
management activities. The goals and objectives provide the yardsticks against which the success
or failure of the programme is measured, and also determine the philosophy underlying risk
management activities. The objectives are formalised in a risk management policy that states the
aims and describes the policy measures for attaining these goals.
Risk identification
Logically, before any attention can be focused on the management of risk, the source of risk must
be identified. In many instances, identification is selfe evident, while in others it requires more
insight. Often, consultation with professional personnel/organisations is required to identify all
risk sources that may give rise to multiple risk situations.
Risk identification consists of two related activities. Firstly, risks that affect the organisation
must be identified. Identification of risks usually is accompanied by both hazard identification
and exposure identification. Hazards (or 'risk factors' in the case of speculative risks) are
activities or conditions that create or increase the likelihood of loss (gain) or the loss (gain)
amount. An improperly maintained piece of heavy machinery is an example of a hazard. An
exposure to loss (or gain) would be the object, individual or situation subject to loss (gain), e.g.
the worker who could be injured operating the improperly maintained equipment.
Identification is followed by analysis. It is not sufficient to know that hazards, risk factors
and exposures to loss or gain exist. risk manager must understand the nature of those hazards,
risk factors and exposures; how they come to exist; and how they interact to produce a loss or
gain. Perceptions of risk, as well as uncertainty, are also analysed, since they may be of profound
importance. For instances the hazards associated with the handling of a dangerous material may
be well understood by the organisation, but if employees perceive the risk to be extraordinary,
that assessment of the risk may become a de facto management reality for the organisation.
16
The Risk evaluation and assessment

15 This step is most important in the overall process as, together with identification ofpossible
sources of loss, it represents the foundation for planning, organising and managing the risk to
reduce the impact of possible losses.
The elements of a risk management model are discussed below.
Risk evaluation entails quantifying the risk and determining its possible impact on an
organisation. Head12 refers to the process of risk evaluation as the analysis of loss exposures,
where attention is focused on 'how frequent and how severe accidents are likely to be and how
they may interfere with the organisation's success'.
More particularly, risk evaluation and assessment concerns the following: the evaluation • Acceptance: Certain risks are inherent to a specific business. Trying to avoid them would
of both loss frequency and loss severity, which will provide, among other things, the two mean disinvesting from a particular industry. If the risk-return properties are acceptable,
significant measures of expected average loss and maximum possible loss. Since the the risk associated with that industry or type of business needs to be accepted.
characteristics of risk and the frequency and severity of losses are all constantly changings
• Mitigation: Mitigation follows on from risk acceptance and is aimed at lessening the impact
evaluation becomes a continuous process.
of the risk.
• an analysis of the financial strength of the organisation, which basically entails the assessment
of the firm's risk-retention capacity. The objective is to ascertain what the impact of a given Risk control activities are those that focus on avoiding, preventing, reducing, or otherwise
risk might be relative to the financial strength of the firm. controlling risks and uncertainties. Risk control activities can take simple forms, such as making
sure that a kitchen has functioning fire extinguishers. They also can be complex, such as
The evaluation process requires expertise in several disciplines and the use of various
developing a catastrophe contingency plan to use in a nuclear power plant emergency.
techniques that by their nature necessitate interacting with several organisational units. For
example, in evaluating liabilities emanating from contractual agreements or in evaluating
possible losses from a liabilityviewpoint of a new product, input from both the legal and Risk financing
research and development departments is required.
The final step in the process of risk management entails the financial provision for losses that
may occur. Risk-financing activities provide the means of reimbursing losses that occur and for
Risk control funding other programmes to reduce uncertainty and risk* or to enhance positive outcomes.
Normally, some losses will occur, despite risk control efforts. The financing of these losses can
The next step is to minimise the risk practically through the design and implementation of a
include measures such as the purchase of insurance coverage, the establishment of a captive
physical risk management programme.
insurance subsidiary or the use of letters of credit. The funding of a highway safety programme
Such programmes would aim to achieve the following goals:
through earmarked taxes would be a less obvious, but valid, illustration of risk financing.
• reduction of the magnitude of the exposure reduction of the frequency of the
loss-producing events dealing (physically) with loss-producing events recovering
In the selection of the most effcient method of financially providing for the consequences of risk, the
(physically) from loss-producing events.
following choices become evident:
the retention of risk under a deliberate self-funding plan
Risk control programmes may be referred to as practical in the sense that they are conducted 18 RISK
at the source ofthe risk While to a large degree the responsibility
the combination of risks (diversifying or hedging) to obtain the benefit of greater certainty in
Head (1982)
17 predicting the loss occurrences through the use of the law of large numbers. This method may
be used by business firms, individuals or, indeed} insurers. It has its limitations, however, since
the scope of combination or diversification may be limited. 14 the transfer of risk cost to other
for risk management in an organisation can be viewed as a staff function, the practical third parties through techniques such as insurance.15
implementation and monitoring of loss control program-mes are conducted by line
management. 6 Risk retention and risk transfer to third parties are discussed in detail in the following chapters.
Nevertheless, it should be pointed out here that the evaluation of the advantages and costs of
Three different responses to risk are possible: alternative methods and the selection of the most efficient method or technique are difficult
• Avoidance: Some risks can be avoided by not carrying out specific activities. The risk of tasks. Prudent risk retention is underpinned by relatively complex quantitative analyses aimed
injury can, for example, be avoided by not taking part in dangerous sports. Avoiding a risk at determining probable quantums of loss that are to be self-financed. Other aspects to consider
is, unfortunately, rarely possible in the business environment. include, for example, whether a given risk should be 'transferred' by insurance, or self-retained;
and, if the risk is retained as selfinsurance, a limit beyond which the risk should be 'transferred'
by purchase of excess insurance needs to be set.
6
Che of (he more significant aspects of successful risk management, however, is the cornmltment of top management to any such programmes,
Risk•control and risk-financing activities are not always mutually exclusive. Banks, for however, also involve the functions of leading, planning and decision1.7 Historical
example, are actively involved in credit and interest rate risk management activities (such as
development of risk management
interest rate swaps) that might reasonably be seen as both risk-financing and risk-control
activities. Early developments
control measures are those practical measures that people can take to minimise risk. An
Programme monitoring and administration example of risk control would be when primitive humans made a fire to keep wild beasts at
bay. It can be argued that when Pharaoh ordered Joseph to store up grain in case there was a
This element will reveal the need to revise decisions when a significant change in conditions
famine, this was a risk control measure. 16
causes the programme to fall short of its goals. It also establishes procedures followed in the
day-to-day operations of the risk management programme. For example, procedures for Risk financing can also be traced long into the past. The origin ofinsurance, particularly
17
communicating programme efforts or the structuring of a programme review and evaluation marine insurance, is veiled in antiquity and lost in obscurity. Some writers argue that
fall into programme monitoring and administration. insurance can be traced as far back as 3000 BC, when the Babylonians established bottomry18
as a means of handling the risks associated with international trade.19 Bottomry was a type
ofloan: a grantee issued a bond to the value of goods that were shipped between countries.
Repayment of the
14 Apart from the fact that the ways in which the risk manager can 'spread' ortllverstfy rrisk across several exposures are limited, It is pertinent also to
consider h19tær attitude toward indwldual riSk_ For shareholders, the effect of pure rusk may be lost within (heir invesmnent portfolios. If sha reholders holdefficlent\y
diverslbed investrnent portfolios, the effect of pure rlsk on a pa rt.cular stock in the portfolio Will have an Insignificant effect on the riskiness of the whole portfoliö_
However, rtsk managers cannot diversify away their em pbyment risk so easily. Conseqently, they may adopt more risk-averse postu:es in making corporate decislons. out that this event cannot be seen as an example Of insurance risk hnance. It can, however, be seen as an
See Elliott (197& 309—20),
15 Reference to the concept ot risk transfer to third partles can be twaced back to the old writers on Roman-Dutch jaw. Grotaus (1926), for example Gibbon (1776) In TheDeclineand Fan oftheRoman Empire that nautical insurance was practised by that empire has now been that the Roman govern
dehoes insurance as •an agreement whereby one person takes upon himself the risk of uncertain danger apprehended byanother, and the latter In his turn binds ment requ.red supplies of military stores to accept all rlsks, mainly because ot fraudinsurance start v,ith a discussion on this historical
himself to pay the former a premlurn'- The concept Of risk transfer therefore is one embedded in legal literature and referred to in most contemporary Wiitings. duelopment ot insurance.
r
Ho.'r eveq, it should be noted that conceptually risk cannot be transferred per se- As Channel! points out in Prudential msurare Com-pony Ltd. v iniond Revenue (1904)
2 65B. 'Where you insure a ship ar a house. you can not insure that the ship shall not be Jost the house burnt, but what. you do Insure is that a sum Of money snall be
pad upon the happening ofa cestain event."See Getz and Davis {lgo.- 77—81WIIIiams and Reins ('981 : 187—200) and Vivian {1990.26}.
19
Despite the fact that the listing of these elements implies a sequence, risk management is
not a sequential process. Certain elements logically appear to precede others (e.g. without risk
assessment, it is diffcult to consider the selection of risk-control or risk-financing methods).
The issues confronting an established risk management programme, however, can involve any
one of the five elements. Also, some overlap among the five elements is almost inevitable. For
example, a risk manager's review of insurance coverages proposed by a broker may reveal a
type ofrisk that the risk manager failed to recognise during risk assessment. Normally, the
review ofcoverages would be considered part of the programme administration, not of risk
assessment.
The management functions of planning, organising and controlling apply to all of the
elements of the risk management process, although a particular function may be more prevalent
in one element than the other. Programme evaluation will, for example, mainly involve the
control function, while risk control will mainly involve the organising function of allocating
resources to a particular course of action that has been decided upon. Both elements will,
MANAGEMENT: FUNDAMENTAL PR'NCIPLES PRINCIPLES OF MANAGEMENT APPLIED TO MANAGING RISK
20 accounting activities such as stocktaking» compiling financial statements, determining costs
and compiling statistics managerial activities such as planning, control, organisation,
command and coordination.
loan was based on the safe arrival of the goods — the seller had to repay the grantee of the bond
21
if the goods arrived safely; if not, the loan did not have to be repaid.
In approximately 700 BC, the Greeks and Phoenicians extended the bottomry principle and
developed bonds with special provisions that were aimed at reducing the risk to the grantee of the Although Fayol)s perception of the security function was limited in scope, if compared to the
bond. The bonds provided for specified loan terms, the placement of agreements in escrow (the present-day risk management function, it is found to have given early recognition to the need
for a security and loss-control function in an organisation focused on protecting the resources
custody of a third party), specified loan periods and specified interest rates, and included severe
of the company. It was not, however, until the early 1960s that risk management became
penalties for breach of contract. Some bonds also contained provisions that prohibited a ship from
established as a recognised management function, primarily through the impetus it gained in
deviating from a mutually agreed upon, prespecified route considered to be the least risky by the the US.
parties involved.
The above developments eventually led to the establishment of the insurance industry.
Development of risk management
Insurance contracts were compiled that indemnified an insured against losses caused by specific
perils. As the industry grew, these contracts extended beyond marine risks to cover fire, personal United States
accidents, burglary explosions, breakage of glass, loss of profits and legal liability claims. The modern concept of risk management originated in the US. 22 Traditionally the most
Organisations increasingly relied on insurance to provide indemnification against losses, and comprehensive method used by an organisation to deal with risk was to purchase insurance,
insurance buyers were appointed in many large companies. Risk management services were The first indication of the importance of the insurancebuying function came in 1931 when the
provided by insurers. 7 These included surveys that were conducted on the client's premises, American Management Association held a conference in Boston that led to the formation of
leading to risk control recommendations. The financial consequences of the risk were transferred an insurance division of that association. 23 This in turn led to the recognition of the insurance
to the insurance industry. manager instead of the insurance buyer. The modern origins of risk management can be traced
During the early part of the twentieth century, as large companies were being formed, to a speech delivered by Professor Wayne Snider on 24 November 1955 in which he suggested
theoreticians discussed the various management functions that should exist within these that the 'professional insurance manager should be a risk manager'.24 This was followed in 1956
companies. The recognition of risk management as a separate management function can be by an article by Russell Gallagher in the Harvard Business Review entitled 'Risk management:
attributed to Fayol,21 whose seminal work in 1916 on the functional approach to management New phase in risk controll Gallagher was the insurance manager for the Philco Corporation.
identified the Risk management thus has its origins in the insurance management function of the large firms
following six basic functions: or megafirm.
technical activities such as production, manufacture and adaptation This soon led to the adoption of the term 'risk management' as a general management
commercial activities such as buying, selling and exchange financial activities such as the function, but Snider25 pointed to the complete lack of literature and suitable courses on this
search for an optimum use of capital subject. In 1964, the Insurance Journal changed its name to the Journal of Risk and Insurance.
security activities such as the protection of property and people The first textbook on the subject, Risk Management in the Business Enterprise, appeared in
1963 and was written by Mehr and Hedges 26 This was followed in 1964 by Risk Management
7
In the nineteenth century. various fire insurers had their 0'.'.rn ire mark displayed ouside the Insured premises In the event of a me. the hre• fighting teams of the Life authontles Insures, established donated In 1331 fire-fighting , had a ire mark equipment This company to municipalities. was acquired It is thought •n 1894 by London
insurers were diS*tcbed to fight The system never really developed in South Africa. that at since least such one a fire servi& Insure-r. was the provldedSA FHe by and locat and Lancashire, which one
Assurance Company, ancestor or Mutual and Federal* %uth Africak, largest short-term insurer. For a history of this cornpany. seeVIvian
0996). 21 (1916),
and Insurance, authored by Williams and Heins. 27 In 1965, the Insurance Institute of America Risk management is a special sub-division of the discipline of general management - a
pioneered professional education in risk management and awarded sub-division having the basic goa's of preventing accidental losses and financing the
restoration of those losses which cannot be prevented, to enable an organisation to
fulfil its mission despite actual or potential accidental losses.
This approach emphasises the distinction between insurance buying and risk management as a
22 For the hlstor ical deu&pment of risk management. see Crockford (1982) and Klornan 1992). function on a general management level. Gahin33 takes a similar view in the following statement:
23 zadil,c (1%6- 17)_
'In the real world it is inconceivable
24 Tre speech was repue-d in the Natiomt U,rrdemnter.
25 snider (1956),
26 Mehr Hedges (1963)-
27 and Hein; 0964)
22 RISK MANAGEMENT'. FUNDAMENTAL PRINCIPLES 28 (1982: 180-6)
29 Kloman (1971} was the first to COIn the chrase tut the umbilcal cord to insurance, but the call was taken up by others: Strutt t 1991) makes san*
pant
30 Farthing (1988).
the designation ofAssociate ofRisk Management on the successful completion of a three- 31 vauohan (1982:36).
32 Hod (1 9S2).
semester course.28 Much of the debate about the role of risk management has centred around its 33 Gatlin (1984; 62).
relationship to insurance, and in particular to the need perceived by some to cut the 'umbilical 23
cord to insurance'.29
Risk management developed from the insurance management activities of companies, and risk exists in a vacuums independent of other activities. Treatment of therefore, be
should,
an integral part of the decision concerning the pure risk.'34 risk management is now seen as a
many felt that the close association with insurance buying restricted the development of the risk
that creates is far removed from the insurance-buying approach that stages of its
fully fledged management
management function. Firstly, risk managers restricted their activities to the handling of insurable summary,
development. Some authors like Robinson 36 and Gahin37 are of the opinion that the risk manager
risks; ands secondly, they tended to turn to insurance when faced with a risk problem instead of a view that
have to deal with speculative risks, particularly foreign exchange risks. Greene and Serbein 38
in the early
using more creative methods. Capacity problems in the insurance industry in the mid-1970s were express the view35that activities such as quality control, market research, business forecasting and
Kunath, Baglini,
resolved by establishing captive insurance companies, while increasing liability claims were met investments are within the scope of the risk management department. These views show risk
future will also
by obtaining additional management developing into a management function much wider in scope than envisaged by
Henri Fayol in his visionary work on management, which now has become a reality with the
liability cover.30
development of enterprise-wide risk management, which will be discussed in chapter 4.
Other developments confirmed the close relationship between insurance and
riskmanagement. In 1932, the insurance buyers ofNew York formed the Risk Research Institute,
and in 1950 the National Insurance Buyers' Association was launched. This later became the
American Society of Insurance Management, which eventually became the Risk and Insurance risk management in the UK was also closely linked to Although
Management Society.31 original attitudes towards this new function 1970s were cynical," brokers and underwriters
began to employ management consultants, while the financing aspects were being actively
From the late 1960s onwards, however, proponents of risk management submitted that the Kingdom
by merchant bankers, who were introducing new areas of expertise. risk management in
nature of the risk management function extended beyond the management of insurable risks and development of
the UK appeared in 1969, entitled Risk, and Insurance and written by Horrigam Another
insurance buying. Head,3Z for instance, points out that 'risk managers increasingly are learning to activities.
milestone Kluwer's 1973 publication, The Handbook ofInsurance.
evaluate and present their programmes in terms of their effects on the amount and timing of the
organisation's cash flow and therefore are subject to increasing integration with general On the educational side, the City University in London and the University been presenting
management', and makes the following statement regarding risk management: courses in risk management and insurance 1970s. In 1982, the Glasgow College
ofTechnology
book on started the program-me in risk management. In the same year, the
Chartered included a paper on risk management in its fellowship
Management
ofNottingham have
early
Institute
RISK MANAGEMENT: FUNDAMENTAL PR'NC'PLES
Association of insurance Managers and Insurance Consultants was formed in
December 1963, when 12 major insurance and industrial companies forum for a discussion of
problems encountered in the insurance organisation later changed its name to the Association
of
risk'has been accepted to describe risk situations that display dcmnside risk only. Th.; text uses the terms event rSWand 'pure alth0L&h Increasingly
the termevent rlsk'is gaining favour. (1981).
24
RISK MANAGEMENT: FUNDAMENTAL PR'NC'PLES
Insurance and Risk Managers in Industry and Commerce. In 1989, the Institute of Risk
Management introduced a three-year, non-degree programme in risk management.
South Africa
Risk management started to develop in South Africa in the 1970s. Most of the broking houses,
including Alexander Forbes, First Bowring and Glen Rand MIB, were involved in this
development, as were major industrial companies. The first known risk management statement
to be signed by a company was in 1975, when Mike Rosholt, chairman of the then Barlow Rand
Group, committed the group to a risk management programme.
The people involved in the early development included H Spooner and B Edmundson.
Spooner was one of the earlier authors to publish on the developments in South Africa. He
reported on risk management locally in 1976 and indicated that most corporations were familiar
with its fundamental principles.
In 1975, Edmundson formed the South African Risk Management Association, but it never
gained momentum and was dissolved in 1979. Various views as to why the association failed
exist, but Edmundson attributed its failure largely to the soft insurance market that prevailed
during that time: the focus of insurance brokers and risk management consultants on loss control
and insurance, the lack of education, and the lack of a sound theoretical base on which risk
management principles could be developed.
In 1986, a number ofthe major South African companies formed the South African Risk and
Insurance Management Association (SARIMA), membership of which is on a corporate basis.8
The Society of Risk Managers was formed in 1990, with Reg Lambert of Barlow Rand being
elected the first president. Membership is on an individual basis and the organisation's major aim
is the promotion of the professional status of risk managers in South Africa.
SARIMA and the Society of Risk Managers amalgamated in 2003 to form the Institute of
Risk Managers. The subject Risk Management is taught at formal and non-formal levels by
various educational institutions in South Africa.
PRINCIPLES OF MANAGEMENT APPLIED TO MANAGING RISK 25
1.8 Summary
From its insurance-buying origins, risk management has developed into a fully fledged
management function where the planning organising, leading and controlling functions of
management apply directly to the management of risk. The focus of risk management has
shifted from preventing and funding losses to that of managing all the risks in the organisation.
This is known as enterprisewide risk management, which is discussed in chapter 4.
8
Business Ch:y. 2 October 1986.
26
2.2 Elements of risk
Risk management is concerned with managing possible future consequences or outcomes. This
suggests that the notion of risk has a number of elements.
Introduction
Elements of risk
Definitions of risk
Basic risk classifications
Strategic risks — sustainability
Managerial risk classifications
Psychokogical influences on risk
Market failure
Economic viability of insurance companies
Economic viability of insurance markets
Summary
2.1 Introduction
In the previous chapter we developed a definition of and model for risk management. This
chapter describes the context of risk and provides a definition so that a sense of what managing
risk means is gained. It begins by describing the elements of risk, the condition of uncertainty
and its relationship with risk. Risk is then defined. The classifications of risk and the concepts
of peril and hazard are also considered.
RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES CONCEPT OF RISK
CONCEPT OF RISK 27
Outcomes
The first element is outcomes, which can be either positive (favourable) or negative (unfavourable).
Outcomes have at least five different forms. Firstly, negative outcomes can be damage to
property. Secondly they can be consequential losses that arise because of damage to property.
Thus, if a factory is destroyed) the owner will not only face the cost of repairing or replacing the
factory, but will also face a loss of the income that he/she would have earned had the factocy not
been destroyed. Thirdly, there could be death or injury to people, as in the case of employees
killed or injured in a fire. Fourthly, there could be legal liability claims against the owner. In the
case of the factory fire, for example, a fireman sent to fight the fire could be injured and decide
to sue the owner. Finally, the losses can be purely financial in nature, i.e. losses unrelated to
damage to property. "Ihus, employees may go on strike, resulting in a loss. All forms of losses
are usually expressed in financial (monetary) terms. The extent of the damage to the factory can
be expressed as the amount of money required to repair the factory, injuries to people as the
amount of compensation they receive, and so on. outcomes are, however, not all monetary in
nature. Thus, the owner of the factory may face criminal charges arising out of the fire, and the
possibility of criminal prosecutions will be regarded as much a risk as the financial loss. Non-
monetary outcomes are diffcult to quantify, since they are not expressed in a common currency,
and thus the tendency is to concentrate on outcomes expressed in financial terms.
Positive outcomes are usually expressed in monetary values, e.g. an investor will purchase
shares with the view to earning a profit. The profit is a positive outcome.
Outcomes can occur anywhere within an enterprise, and hence risk management is not
confined to any particular part of the organisation, but should extend throughout the
organisation, hence the term enterprise-wide risk management. Traditionally, risk management
was fragmented, concentrating more on certain types of outcomes, i.e. insurable outcomes.
Since risk management is concerned with the future, outcomes may be anticipatory rather
than actual For example, buildings may be designed to resist an earthquake, not because an
earthquake as taken place, but because it may take place. Past outcomes can be recorded and
measured, e.g. the cost of individual motor vehicle losses can be recorded. This suggests that
risk management requires data to be collected and analysed.
Events
The outcomes are often be traced to a specific time and place. This suggests that risk involves a
second element, an event. 'lhus a fire is an event that can be 28
determined with reference to a time and place, Positive outcomes, on the other hand, may not be
traced to a specific event. Company profits are earned over a period oftime. The discovery of a
overly
MANAGEMENT: FUNDAMENTAL PRINCIPLES
valuable idea by a research and development department can be regarded as a valuable event. Contemporary finance theory makes extensive use of the notion of risk, but concerned with
However, the profits from this discovery may only take place years later. The profits will accrue semantics, and the technical usage of the term 'uncertainty' has, to a degree, been limited. It is
not surprisingl to find that varying definitions of risk exist.
over a period of time and are not confined to an event per se.
probability.
Events can also be recorded. Thus, the number of motor accidents or industrial accidents
for example, defines risk as a combination of hazards measured by
can be recorded. Agains since events can be recorded, this suggests that the data can be subjected
to statistical analysis. Denenberg et al.2 define risk as uncertainty of loss, where the term 'risk' is implicitly
understood as uncertainty of financial loss and where the definition denies that the degree
of uncertainty needs be measurable or the probability of loss determinable.
Sources Greene and Serbein3 qualify the existence of many usages of the term risk, and there is
The cause of outcomes can often be traced to specific sources. This suggests that risk involves a therefore no single definition that is universally employed. Nevertheless, it is stated that the
third element, the source of the loss. Ihus, a fire may be the cause of damage to a building, term is understood to mean mainly the uncertainty of the occurrence of economic loss.
resulting in a financial loss. source of the loss is also often called the peril. Thus, it can be said a Athearn and Pritchett4 define risk simply as a condition in which loss or losses are possible.
Risk (pure), they state, involves only the possibilities of loss or no loss.
factory is exposed to the peril of fire.
As the result of the work of Frank Knight in economic theory, risk has a specific meaning.
It describes those situations where the probability and outcomes can be determined. If either
Environmental factors the outcome or probability or both cannot be determined, uncertainty is involved.
It may be anticipated that two projects could produce similar outcomes. This does not mean that,
specifying
from a risk perspective, these are the same. For example, two factories, each representing the the definitions cited above, authors have been careful to qualify depends to
interpretation
same capital investment, would both be exposed to a possible loss due to fire. If one manufactures some extent on the particular orientation of the Notwithstanding such qualifications, however,
of risk.
evidence exists of non-uniformity rather than disagreement concerning the fundamental tenets
steel products and the other manufactures wood products, the probability of a fire is different,
of risk when defining risk in the context of pure risk management and insurance.
however, with a fire being more likely in a factory that works with wood, a material that burns
easily. The probability and standard deviation of the outcomes of these two may differ. It will be With the emergence and growing importance of risk management as a systematised
discipline, it becomes necessary to provide a more rigorous definition of risk by enlarging on
said that seen in terms of the peril of fire, the one factory is more hazardous than the other.
the early concepts and, by so doing, adding clarity to the more contemporary definitions and
interpretations. The following definition is proposed:
2.3 Definitions of risk

Having gained an idea of the elements involved in risk, the next point is to discuss definitions of Risk is defined as the variation of the actual outcome from the expected outcome. Ifthjs
risk. Risk often has a contextual meaning and thus no single definition covers all possible definition is accepted, then the standard deviation is an a ppropriate measu re therefore
meanings. Furthermore, different disciplines assign different meanings to the word 'risk'. The implies the presence of uncertainty.
context in which risk can be viewed is so diverse that no single definition is sumcient to cover
all possible risks. This gives rise to interpretations and definitions suited only to specific areas 0974:4-3).
0983:24).
of study or disciplines. Hence, in an actuarial context, risk has a statistical interpretation; while Pritchett (1984'4-5).
30 RISK
in the world of insurance, the term 'risk' may be used to describe the subject of the policy (the
property or liability that is insured) or the peril insured under the contract.
29 nie definition of risk as the deviation of an actual outcome from the expected result or outcome
implies the following:
Uncertainty surrounds the outcome of the event. *Ihe decision maker is uncertain about the The act of insuring has the effect of pooling the exposures, and as the number of exposures
outcome, but predicts an expected outcome. The actual outcome may deviate from the pooled increases, so the variability in the aggregate outcome diminishes.
expected outcome. If the outcome were certain, there would be no uncertainty, there would It is contended that the degree of risk is solely dependent upon the variability rather than
be no deviation from the expected result and therefore no risk. upon the probability value surrounding an event or its outcome. Such an interpretation refutes
The extent of the uncertainty between the actual outcome and the expected outcome the viewpoint of those who consider risk in terms ofprobability ofoccurrence and who often
measure risk on a scale with certainty ofoccurrence at one end and certainty of non-occurrence
determines the level of risk. The greater the possible deviation
at the other end. The assumption that risk exists in varying degrees between these extremes,
and that risk is thus greatest where the probability of occurrence or non-occurrence is equal,
between the expected and actual outcomes, the greater the risk
does not fit with the basic tenet that risk refers to the degree of uncertainty or variability about
the occurrence, and not to the degree ofprobability that it will occur.
In a sense, the above definition links risk and uncertainty. Uncertainty prevails because outcomes
From a risk management perspective, uncertainty exists concerning:
of situations are not known in advance. Consequently, such
situations display risk To the extent that associated probabilities are assigned (objectively or
• whether the event or occurrence will take place; and if it does, what the outcome
subjectively) to possible outcomes, risk can be mathematically described. Where situations
(financial) of the event will be.6
dictate that associated probabilities cannot be assigned (objectively or subjectively), risk cannot
be quantified, and thus, from a risk management point ofview, whether one regards these The definition, for example, that risk is uncertainty about loss, is indicative of the orientation
situations as uncertain as opposed to risky becomes immaterial.
towards insurance, rather than towards risk management: i.e. it is one that concerns itself more with
the financial treatment of the
The degree of uncertainty surrounding the event determines the extent of the risk.
consequences of the event than with the business of managing risk. Managing risk implies not
Ibis distinction seems to be somewhat ignored in the definitions given earlier. The tendency in only the financial provision for the consequences of an event, but the effort to:
these definitions is towards the negative financial consequences associated with the event, which,
although such consequences must be seen as an important consideration, somewhat blurs the reduce or minimise the likelihood of the loss-producing event occurring
question of whether uncertainty, and hence risk, is present. For example, it is maintained that
insurers carry risk, since they are involved in funding the economic loss that is the consequence • reduce or minimise the adverse effects (mostly financial) once the event has
when an insured event does occur. In the extreme, it could be argued that, given a sufficiently occurred
large number of exposure units, the insurer's aggregate loss is known with some degree of
Taking cognisance ofthe points made above, the following expanded definition is proposed:
certainty and thus, notwithstanding the economic loss implied, there exists little or no risk to the
insurance market. The variability surrounding the expected aggregate loss is small under these
conditions and therefore the risk is limited — the insurer theoretically carries little risk and is
merely a mechanism for funding the financial consequences of the insured events.S
6 Simp:istlcally. the Situation Is analogous to that of throwirE a de The uncert*inty not only concerns the value of thrown b\.n also
5 Under these conditions, the insurer thecgetically faces only the liquidity solvency risk. Hovever, systematic risk may be inherent. Forin example in the life insurance
industry, the effect of acqulred immune deficiency syndrome (AIDS) has been to increase insures•risk, (possibly the sh0fte€term).The overall uncertalnty surrcm.lßding to whethep it will occtr. instance. atoll. There in the are case instances, Of a motor fieett where the event. one is defined concerned as an accident is not uncertaln,
the event What Will Is occur ure«ain as opposedis how many vehicks wilt be Gmaged by lithe nat,onal annual acc&nt rate Is say, i and if the fleet composes 203 vehicles,
the event ofdeath (in the books of the insurer} hasincreased. one could 25 vehicles to be damaoed each year. The concern is not the probability ofa vehde being involved in an accident — it is almost certain, that an accideru Will
take place. The concern •n this instance is about the number of accidents that Witl occur and about outcome
CONCEPT OF RISK 31
32
Obviously, from the viewpoints of the individual insureds, uncertainty, and hence risk,
does exist; each exposure viewed independently displays considerable variability in outcome.
Definition
Risk is defined as a deviation from the expected value. It implies the presence of uncertainty.
There may be uncertainty as to the occurrence of an event producing a loss, and uncertainty
as regards the outcome of the event. The degree of risk is interpreted with reference to the
degree of variability and not with reference to the probability that it Will display a particular
outcome? The standard deviation becomes a good measure of risk.
2.4 Basic risk classifications

Risks can be classified in a number of ways. Two divisions are discussed, basic classifications
and managerial classifications. Basic classifications examine the more popular ways in which
risks have been conceptually classified. Managerial classifications are those classifications made
in practice to enable these particular risks to be managed.
Managing risk requires both the ability to quantify risk in order to facilitate decision-making
under conditions of uncertainty and the recognition of the qualitative and often psychological
aspects of risk. In this chapter, therefore, probability theory and its application to risk
measurement is introduced, and for illustrative purposes, a relatively simple model for decision-
making under risk conditions is proposed.
A central theme of this book is the integration of the activities that collectively represent
the risk management process and discipline. The fundamental principles governing risk
management, which are translated into a simple integrated model, were discussed in the previous
chapter. This model identifies the two underlying facets in the risk management process,
namely:
the practical or physical management or control of risk the financing of risk, where the
insurance market is an important mechanism for financing the consequences of risk
In line with this integrated approach, this chapter also considers the definition of risk from a
financing perspective, particularly the question of risk as applicable to an insurer. Finally, in
the more general context, particular aspects relevant to the formation of a market for treating
risk are described.
7 coa:orne in Itselfcar. assume various forms. eq crirr.lnal from an accident. ernorLonaI stress due to the boss OF e and manyothers in this
study, however. the concern is primarily about the hnanti&l outcomes of risk.
33
Risk and uncertainty

Uncertainty
Uncertainty is the opposite of certainty. Certainty is the lack of doubt. It exists when a particular
consequence can be unambiguously predicted to follow a given event. Examples of certainty are
predictions from physical laws, such as the law of gravity or laws of motion in physics. The
predictions of these laws correspond to the actual outcomes.
Uncertainty arises from a person's imperfect knowledge concerning future events. It exists
in decision situations where the decision maker lacks complete knowledge, information or
understanding about the decision and its possible consequences. The perceived level of
uncertainty depends on information that an individual can use to evaluate the likelihood
ofoutcomes and the individual's ability to evaluate this information. In other words, uncertainty
is present in levels or degrees, as illustrated in table 2.1.
Uncertainty is concerned with the following two elements:
(i) uncertainty whether an event will occur; and

(ii) if the event does occur, what the outcome of the event will be.
Example
Uncertainty exists whether interest rates will change (decrease or increase) and also, if they do
change, what the direction and the extent of the change will be.
Table 2.1 The certainty-uncertainty continuum
Level of uncertainty Characteristics Examples
None (certainty) Outcomes can be predicted Physical laws, with
precision natural sciences
Level 1 : Objective Outcomes are identified and Games ofchance, uncertainty probabilities are known
cards, dice
Level 2: Subjective Outcomes are identified, but Fire, car accidents, uncertainty
probabilities are unknown investments Level 3: Total uncertainty Outcomes
are not fully
Space exploration, identified and probabilities genetic

research are unknown
Source: Adapted from Williams, Smith and Young (T 998)

MANAGEMENT: FUNDAMENTAL PRINCIPLES CONCEPT OF RISK
34 RISK My only reason for mentioning it here is to show why it seems necessary to define risk
with reference to the degree of uncertainty about the occurrence of a loss, and not
A high degree of uncertainty, as at level 3, reflects a significant lack of understanding and with reference to the degree of probability that it will occur. Risk in this sense is the
knowledge of the situation, resulting in a low level of confidence and assurance. Where there is objective correlative of the subjective uncertainty. It is the uncertainty considered as
complete uncertainty, the prediction of possible outcomes is impossible. embodied in the course of events in the external world, of which subjective
Uncertainty, which is a condition that results from an inability to foresee future events, has uncertainty is a more or less faithful interpretation.
been recognised as affecting all walks of life. The finite nature of our minds denies us the ability
to foresee, and hence control, the many happenings that affect our lives and those of others. In The distinction between risk and uncertainty then, is in the objective and
interpreting the notion of uncertainty, Shaclde9 offers the following:
of variability in the outcome of events.
aspects offered by Knight li may also be seen to hinge on a similar objective-
The interpretation
The word uncertainty can convey either the mind's consciousness of ignorance and its subjective distinction. Knight distinguished between measurable and immeasurable risk. More
consequent willingness to entertain an array of diverse hypotheses, or alternatively particularly, in the cases where possible outcomes and associated probabilities are known and
the mind's hesitant contact with any hypothesis which has not a complete case in its therefore the spread of values can mathematically measured, Knight defines such variability as
risk. When the distribution of possible values and probabilities is not known, the situation is
favour or has some partial case against it.
described as uncertain.
While many statisticians and philosophers believe that there are laws that precipitate change* be 'Ihus, where so many uncertainties (technical, political and economic) surrounding a value
it in attitudes, perceptions or the business environment, such laws and governing principles are exist* implying that the number be viewed as a random value, uncertainty is deemed to be
too vast to be regarded as charted knowledge. In their efforts to understand or minimise present. However, this interpretation implies that in situations where objective probabilities
uncertainty, people have attempted to determine causation, unfold patterns and 10give meaning surrounding outcomes cannot
to unexplained events, possibly in terms of a controlling power. where subjectively one may impose a range on the possible magnitude of outcomes,
assigned, but
Despite such attempts and the aversion to yield to domination, humankind's situation will then risk is present, but it cannot be measured very
always be characterised by the presence of uncertainty.
In distinguishing between uncertainty and risk in the definitions offered by Willett and
Risk Knight, attention should be focused more on the important similarities rather than on any
In a literal sense, the concepts ofriskand uncertainty are regarded as interrelated. The perception differences that may be identified. The modern tendency is to interpret risk as the absence of
is that uncertainty gives rise to risk. This is because where the outcomes of events are surrounded certainty, where certainty represents the situation where there is only one possible outcome.
by uncertainty, risk will be present. Reference to uncertainty is made when considering events Hence, a risky outcome is one that can assume a number of values, but the particular value is
whose outcomes are predictable, even though such outcomes may be assigned objective not known in advance.
associated probabilities. In such cases, the assigning of objective probabilities maybe seen as
being incongruent with the concept of uncertainty.
35 and hazards
The terms 'risk and 'peril' are often confused and used as ifthey mean the same thing. A peril has been
defined as the cause of a loss; 12 it has also been defined 36
Willett 10 draws the following distinction between risk and uncertainty:
9
Shackle 0961.09).
10
Denenberget al. (1974: 15-6).
RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES CONCEPT OF
contended, is also evidenced in the definitions to the ofpossibleexistence the risk-related concepts of a dishonest of carelessness Of peril ter&ncy. and or
as the source of loss. }3 For reasons that will become evident when considering the definition
hazardindifference Reference is to also loss. madetomoole Objectively. the hazard should {Mehr. 1 refer9S6; 16 In the 'itenl sense. moral hazard refers term
of hazard, it is preferable to consider peri_l as the source of loss.
28), #lere the behavioural intention characteristic is to refer to bloader human characteristics devoted (o loss-reducing activity This asgpct Is
broadbi to a (not necessarity unethical) the level of effort fu;ther debated in chapter 14-
Definition 37
Peril is defined as the source of loss,
The preceding sections aimed at providing a definition and classification of risk and to
Peril, therefore, is quite distinct from risk, which has been defined as the absence of certainty
relative to both the occurrence of a loss-producing event and its outcome. Typical perils are distinguish between the risk-related concepts of peril and hazard. Any discussion of risk
fires, explosions, storms and earthquakes. These perils give rise to risk, but are not defined as should, however, include both qualitative and quantitative aspects. The concern is to place
risks themselves. risk and particularly the management of risk in an overall framework within which it can be
Illus, for example, a fire in a warehouse is the peril against which insurance may be understood and managed. Managing risk implies not only the qualitative understanding and
purchased, while the oil drums stored on the warehouse floor represent a fundamental loss- perception of risk, but the application of developed quantitative techniques to arrive at some
causing circumstance or situation. This definition coincides reasonably with that provided by dimensional expression of the concept. Hence, the sections that follow aim to give an
Mehr,14 who states that behind the ostensible cause ofloss (peril) is hazard — except that it
understanding of:
claims to differentiate between risk as the source of loss and hazard as the fundamental loss-
causing circumstance or agent.15
the quantification of at least some aspects of risk through the use of quantitative techniques
the psychology of risk and decision-making under pure risk conditions
Definition
• market considerations for the financial treatment of risk.
Hazard relates to the circumstances surrounding the cause of loss.
Intuitively then, one situation in relation to another may be described as more hazardous if that Pure (or event) and speculative risks
situation is fundamentally considered more loss-causing; the term 'hazardous' should not be A classification often used is that of pure and speculative risk. Pure risks are those risks that
interpreted as describing the degree of risk, for the latter refers to the degree of uncertainty
only have the possibility of a loss, e.g. the risk of destruction of a building due to fire.
(variability).
Speculative risks are those that have the possibility of either the profit or loss. 'Ihus, entering
into a profit-making venture entails a speculative risk
A distinction can be drawn between physical and moral hazards.
Definition Insurable and non-insurable risks

The material or physical aspects in the circumstances surrounding the cause of loss are Another important classification to the risk manager is the distinction between insurable
termed physical hazards; the personal aspects or characteristics are termed moral and non-insurable risks. No completely satisfactory test has yet been formulated to
hazards.i6 determine if a particular risk is insurable or not. In theory, most pure risks should be
insurable. But some risks can be pure and non-insurable. For example, the possible
destruction ofa building is a pure risk, and generally
et al- (1974•-8).
this risk is insurable, but the destruction of the building due to war is generally not. Usually,
S 't has been stated (hat hazard represents definition condition ot fisk and that in particular incre3ses the the likelihood interpretation or chanceot of the degree loss
(Oenenberg Of risk proposed et in 1974.8; the previousMehrr I B": 29.
damage to property, consequential losses arising out of this property, injury to people, and
r
Paradoxically, when considering ) that untenainty the is cduced AS and rnenEianed hence nsk earher. is reduced- a literal It is sense unlikely has that, developed in liability claims from sudden and accidental events resulting in the abovementioned losses
definlng hazard in such of a risk, way, and the this. authors'it IS
are insurable.
hazard may imply {stricth, to this ambivalence. in the dehnitton
intenrlons have been to point
In reality, very few risks are insurable, which is surprising. prevented, particularly by an individual. Such losses are frequently catastrophic and therefore
social insurance may need to be established in order to mitigate their effects, instead of
commercial insurance.
Fundamental and particular risks
Fundamental risks arise from losses that are impersonal in origin and in consequence, and
originate in the economic, political or social interdependency of society, although they may
also arise from purely physical occurrences. 1 7 Most eclusiöns from Insurance pollcks referring to lard-based riSk5 rdate to fundamental risks, e. war and nuckar eycluslons.
RISK 39
38
Changes in attitudes, knowledge, technology and social conditions, and other factors,
Particular risks are losses that have their origin in discrete events that are essentially
sometimes political in nature, make it diffcult to determine whether the risk is fundamental or
personal in cause. Such risks would, for example, be fire damage to a building or the explosion
particular, and thus whether the cost of such risks should be borne by society or remain the
of a pressure tank.
responsibility of the individual. However, the relevance of such a classification or categorisation
One may further categorise operational risks by distinguishing between fundamental and
is real. 18
particular risks.
Definition Systematic risk

Fundamental risks arise when many losses can be traced to a single source in origin and in This classification is used extensively in finance theory. Systematic risk is a market-related risk.
consequence.
Thus, take for example the case of changes in the value of the rand against the dollar. As a result,
there is a market-wide impact and most of the values ofshares change. This would be regarded
Fundamental risks affect large parts of society or even the world, rather than individuals. Thus,
as a market-related risk. On the other hand, events could take place within a particular company
the risk of damage to assets due to wear and tear or corrosion is fundamental. 17
Most fundamental risks such as, for examples war and recession originate in the economic, that could have an impact on the value of its shares. For example, assume that the authorities
political or social interdependency of society, although they may also arise from purely decide not to register a drug that a pharmaceutical firm has spent huge amounts on developing.
physical occurrences, such as, for example, pollution, or even the Irish famine experienced The value of the shares of the firm may decline substantia}ly once the announcement is made.
during the nineteenth century, which was caused by the destruction of the potato crop. A further
example is damage due to the nuclear accident that happened at Chernobyl. Fundamental risks
are regarded as commercially uninsurable.
Systemic risk
Particular risks give rise to losses that have their origin in discrete events, which are In banking, in particular, there is theory that an event may cause the entire banking system to
essentially personal in cause. Such risks would include the explosion of a boiler or fire damage collapse; in other words, the entire banking system is exposed to systemic risk. Systemic risk is
to a building. the risk of the system collapsing, Some writers express doubt as to whether this risk is indeed
The distinction between fundamental and particular risks is not, however, definitive. The
possible. Recent events in South Africa, however, seem to suggest that systemic risk is indeed
classification of risk as either fundamental or particular depends on judgement. For example,
possible. It will be recalled that Regal Bank got into diffculties and the Reserve Bank, as the
while it was previously readily accepted that unemployment was a particular risk (caused by a
person's lack of ability or work ethic) , it is more widely accepted today that in most cases lender of last resort, refused to intervene to save the bank as it had previously done with Nedcor,
unemployment is a result ofthe economic system — a responsibility of society rather than of African Bank and Absa. Shortly after Regal Bank collapsed, a run took place on Saambou Bank.
the individual, and therefore a fundamental risk as opposed to a particular one. Again, the Reserve Bank did not act. Shortly thereafter, Absa got into diffculties with the Unifer
One reason for determining whether a given risk is particular or fundamental is to debacle. While this was still unfolding, a run took place against the BOE Bank. It certainly
establish whether commercial insurance may be appropriate, or in fact available, as a means
began to look as if South Africa's banking system was in danger. The Reserve Bank then
of financing the consequences of such risk. Losses arising from fundamental risks cannot be
announced, as it had previously done with Nedcor, that it would guarantee all deposits. Once
the Reserve Bank took this course of action, the run on BOE stopped and Absa was able to shore
up its balance sheet with the announcement that it would float a RI billion corporate bond. These
developments clearly point to the possibility of systemic risk.
18 can be argued that the legal liability üisis (tie irKtease in iiabi lib* I itigation and consequent sggralling ofthe price of insurance and even
•rgal . re-med* fi5ks of V.ewed capacity) As a log•cal from has a consequence dQent arsen because perspective. of risks classifylng such judiclaf as a personal risk systems as fundamental.
injury, have tried which to the used neglrgence private to be seen financmg system as particular of systems deljctual nsks. Onsurance) liability are now is to in inappropriate fund fact
treated what theasas
ærcelve as social hence the liability crisis,

See
40 their issues economr_. thrm.gh a environmental reporting and social performance. The Latest vers•on the GRI i'. called G3- http•flwww.globalreport.ng.orq (accessed
23 October 4 James
(2009). 6 stern
(2006).
2.5 Strategic risks: Sustainability 8 South Africa was ranked as 13th highest CO? emitter in the world by the United Nations' Carbon Dioxide Ana*sis Centre (CDIAD) in 2007 due to the burning of fuels.
Data availabk at httpY/data (accessed 26 October
Integral to a company's risk management is the identification of strategic risk including

41
sustainability risk. Business' sustainability is dependent upon ensuring that the four pillars of
strategic planning, financial, environmental, social and human, are appropriately addressed.)
Mervyn King2 asserts that 'governance, strategy and sustainability' are inseparable. 'lhe incorporates solutions to climate change. Joseph Stiglitz, who is presiding over the UN General
provisions of King Ill will require companies to report on their sustainability practices and Assembly discussions on reform of the international monetary and financial system, has called
performance, making the issue of sustainability a governance criterion as well. One of the best on leaders to ensure that goals such as meeting the Millennium Development Goals and
practice formats identified by the South African Institute of Chartered Accountants and the protecting the world against the threat of climate change must remain priorities and that 'the
International Federation of Accountants for sustainability reporting is the global reporting immediate steps taken in response to the crisis ... should provide an opportunity to accelerate
initiative (GRI).3 Sustainability is no longer an optional extra, but an 'essential component of progress towards these goals'.9
doing business', according to Ben Marx of the University of Johannesburg. 4 One of the key This presents a critical challenge to business as key drivers of development
consequences of incorporating sustainability reporting in King Ill is that doing so integrates
and contributors to emissions especially in developing countries not to go about their business
sustainability with financial reporting, making sustainability a key risk management concern.
as usual Businesses in developing countries need to explore innovative solutions to
The
sustainability risks and not simply pursue the same destructive paths towards industrial
2006 Companies Act in the UK introduced the requirement for businesses to report on their
development that has resulted in the world we have today. The UN Climate Summit in
social and environmental issues.
Copenhagen in December 2009 will present global targets for reducing emissions, which should
It is not difficult to understand why sustainability issues are of significant importance in
serve as a guideline for business in responding to its sustainability responsibilities.
risk management. Note the findings of Sir Nicholas Stern in the 2006 Stern Review: The
Economics of Climate Change6 that climate change will increase the frequency and severity The World Economic Forum 2009 Global Risks reporti0 has pointed to three main risks
ofweather patterns such as hurricanes and floods, with severe consequences both for humans facing the world in post-financial crisis. The first is that governments need to plan appropriate
and the economy. Costs
exit strategies for the massive bail-outs that have been the focus of addressing short-term crisis
associated with this trend have been estimated by the Global Humanitarian Forum over the past
management. Closely linked to this risk is the second element, which is the need for long-term
five years at around $230 billion in a single year, placing industries, businesses, infrastructure
strategic thinking around issues that may seem low priorities today, but in the big picture remain
and properties at risk/ The present threat to sustainability presented by climate change is likely
to be exacerbated by the trend towards increasing harmful emissions by large developing crucial to long-term stability: issues such as sustainability relating to climate change and
countries such as China, India, Brazil and South Africa.8 environmental degradation are specifically highlighted. Finally, the report calls for a global
It is easy to lose sight of the importance of sustainability alongside the global financial response in the form of global governance relating to issues of sustainability, including climate
tsunami that in 2008/09 engulfed the world, but business and political leaders around the world change, water, biodiversity, poverty, terrorism and human rights.
are calling for a solution to the financial crisis that In his call to action, King reports numerous initiatives being undertaken by forward-
thinking corporations both globally and in South Africa to address the problems caused by
pollution and to reduce the amount of environmentally damaging greenhouse gases they
produce. He tells of 'green' offce buildings in Manhattan and Johannesburg, Sasol's clean
Committee development mechanisms and African Explosives and Highveld Steel & Vanadium
*fich FK0duced (he King Il' code on goverrmce prlncipbes. which comes into effect
participation in carbon offsetting. ll South Africa's JSE Securities Exchange has begun offering
3 The GRI is a voluntary nerwork framework. of business that CA'tlines clvjl soaetyr princjpks labchJt for and organlsations professional to institutions use as a guid'ng around
mechanism pro.moting for transparency Treasurlng and on sustainabilityreportirva on shares in carbon credit investments. King's book highlights the tremendous importance of large
corporations as agents for change and identifies good corporate citizens as role models for how
Committee
companies should be addressing their sustainability whkh produced the King Ili code on governance principles, which comes into effect
3 The GRI is a voluntary network framewort. or business. cwll society, Labour and professlonal to institutions use as around promoting tansparency measur.ng and on
sustainabilityreporting on
issues through a environmental reporting and social that performance. outlines pnnclples The larefl orgarnsat'ons ve--rsU:n of (he CRI is called a guiding G3. See
mechanism http•j/vovw.globalreporting.org for (accessed their economic. 23 October 2009). 4 lames {2009 6 stern (2006).
to (2009%
Africa as 13th highest C02 ani«er in the world by the United Nations' Carbon Doxide Analysis Centre (CIAO) in 2007 g South was ranked due to the burning of
fossil fuels. Data available at (accessed 26 October 2009).
'2909:3-4).
RISK 41
40
incorporates solutions to climate change. Joseph Stiglitz, who is presiding over the UN General
Assembly discussions on reform of the international monetary and financial system, has called
2.5 Strategic risks: Sustainability on leaders to ensure that goals such as meeting the Millennium Development Goals and
protecting the world against the threat of climate change must remain priorities and that 'the
Integral to a company's risk management is the identification of strategic risk, including
immediate steps taken in response to the crisis ... should provide an opportunity to accelerate
sustainability risk. sustainability is dependent upon ensuring that the four pillars of strategic
progress towards these goals'. 9
planning, financial, environmental, social and human, are appropriately addressed. l Mervyn
King2 asserts that 'governance, strategy and sustainability' are inseparable. The provisions of This presents a critical challenge to business as key drivers of development
King Ill will require companies to report on their sustainability practices and performance, and contributors to emissions especially in developing countries not to go about their business
making the issue of sustainability a governance criterion as well. One of the best practice formats as usual. Businesses in developing countries need to explore innovative solutions to
identified by the South African Institute of Chartered Accountants and the International sustainability risks and not simply pursue the same destructive paths towards industrial
Federation of Accountants for sustainability reporting is the global reporting initiative (GRI).3 development that has resulted in the world we have today. The UN Climate Summit in
Sustainability is no longer an optional extra, but an 'essential component of doing business', Copenhagen in December 2009 will present global targets for reducing emissionsy which
according to Ben Marx of the University of Johannesburg.4 One of the key consequences of should serve as a guideline for business in responding to its sustainability responsibilities.
incorporating sustainability reporting in King Ill is that doing so integrates sustainability with i0
The World Economic Forum 2009 Global Risks report has pointed to three main risks
financial reporting, making sustainability a key risk management concern. The 2006 Companies
facing the world in post-financial crisis. The first is that governments need to plan appropriate
Act in the UK introduced the requirement for businesses to report on their social and
environmental issues. exit strategies for the massive bail-outs that have been the focus of addressing short-term crisis
It is not difficult to understand why sustainability issues are of significant importance in management. Closely linked to this risk is the second element, which is the need for long-term
risk management. Note the findings of Sir Nicholas Stern in the 2006 Stern Review: The strategic thinking around issues that may seem low priorities today, but in the big picture remain
Economics of Climate Changeo that climate change will increase the frequency and severity crucial to long-term stability: issues such as sustainability relating to climate change and
ofweather patterns such as hurricanes and floods, with severe consequences both for humans environmental degradation are specifically highlighted. Finally, the report calls for a global
and the economy. Costs associated with this trend have been estimated by the Global
response in the form of global governance relating to issues of sustainability, including climate
Humanitarian Forum over the past five years at around $230 billion in a single year, placing
change, water, biodiversity, poverty, terrorism and human rights.
industries, businesses, infrastructure and properties at risk,7 The present threat to sustainability
presented by climate change is likely to be exacerbated by the trend towards increasing harmful In his call to action, King reports numerous initiatives being undertaken by forward-
emissions by large developing countries such as China, India, Brazil and South Africa. 8 thinking corporations both globally and in South Africa to address the problems caused by
pollution and to reduce the amount of environmentally damaging greenhouse gases they
It is easy to lose sight of the importance of sustainability alongside the global financial
produce. He tells of 'green' offce buildings in Manhattan and Johannesburg, Sasolk clean
tsunami that in 2008/09 engulfed the world, but business and political leaders around the world
development mechanisms and African Explosives and Highveld Steel & Vanadium
are calling for a solution to the financial crisis that participation in carbon offsetting. ll South Africa's JSE Securities Exchange has begun offering
shares in carbon credit investments. King's book highlights the tremendous importance of large
RISK MANAGEMENT'. FUNDAMENTAL PRINCIPLES CONCEPT OF RISK
{2009. 186)
corporations as agents for change and identifies good corporate citizens as role models for how 13 King (2009'.183v
19 Doherty (l 2}.
companies should be addressing their sustainability 20 Geene and Serbeln (1983: 20).
43
finance personnel and

production
10
42 • environment.
An organisation, in seeking to earn a profit for its owners, can be affected by the collective
risks and responsibilities. King makes the case that addressing sustainability is not simply a effect of these various risks. But the organisation is also exposed to, for example, the
matter of social responsibility; rather, the business case sees a link between sustainability and possibility of fire loss to its plant and equipment, to liability claims due to defective
opportunity where sustainability is the basis for strategic and long-term thinking. 'Sustainable products, and to possible theft of assets and fraud by its personnel.
2
development makes a company more competitive, not less competitive? The costs of doing For the purposes of this discussion, risk in a corporate environment is subdivided into
business are increasing as energy sources are depleted, and the higher costs of electricity and three categories on the macro level, namely: inherent business risks (end-economic risks)
diesel will make production prohibitively expensive in time, forcing companies to rethink their and incidental risks pure21 and speculative risks operational risks.
strategic risks and ability to continue doing business as usual. The lack of key resources such
as water will require strategic responses from industries reliant on water as a factor of 'These categories are discussed below.
production, such as Coca-Cola. The challenge for businesses in responding to sustainability risk
is 'how to produce more with less'. 13
Inherent business risks and incidental risks
2.6 Manageriat risk classifications Inherent risks

Inherent business risks include all the activities, decisions and events that impact on the operating
Managerial classifications are made to assist in the managing of risks. A board of directors may
profit of an organisation. These risks are also inherent
decide to manage the risks facing the company, but how does it do this? One way would be to
break the organisation into more manageable sections. So it could divide risks into two to the main business of the organisation as reflected by the mission statement. They cause
categories, future risks (strategic risks) and day-to-day risks (operational risks). The day-to-day fluctuations in the operating profit of the company and eventually also in the earnings
risks could then be subdivided into risks facing each department of the company, such as the ofthe ordinary shareholder. Inherent business risks consist of two different types of risks.
marketing department, engineering (technical), financial and treasury, and so on. It is clear that The first results from variations affecting an individual company and is uncorrelated
managerial classifications are arbitrary in nature and those discussed here are by no means all with the rest of the economy. This is referred to as specific risk or unsystematic risk (as
that are possible. opposed to systematic risk; see above). The second type of risk stems from occurrences
From a risk management perspective the terms 'risk' and 'management' imply, in a sense, that affect the economy as a whole. This is referred to as systematic risk or market risk
that only some risks inherent in corporate activity will concern this process. This class of risks (see above).
has distinguishing characteristics that differentiate it from other forms of risks. Specific business risk, or the volatility of operating income, can be segregated further into the
following:
Doherty19 identifies four types affecting an organisation. These are: marketing
risk financial risk resource management risk environmental risk. Sales variability
Greene and Serbein20 classify risks along quite similar lines, namely property and personnel
• marketing
Sales variability is measured by the standard deviation of sales over time and
is dependent on consumer demand for the product. It is caused by market
the djst•ncuon has been made between pure and speculative rtsE bur as companies increasingly have adopted risk
•nagement programmes and the various risks have been classihed, thrs distinction, although conceptually sound, no longer adequate.
•speculative risks can be referred to as operational risks, and many ofthese are insurable.
MANAGEMENT: FUNDAMENTAL PRINCIPLES CONCEPT OF RISK
RISK The risks involved in transactions in financial assets and those that may result from fluctuating
financial claims (deposit liabilities) are referred to as financial risks. Businesses exposed to this type
ofrisk must, therefore, manage not only their non-financial (operating risks) and physical assets, but
factors that affect the demand for a company's products such as product design, promotion, must also manage their financial assets.
general income levels, price, and prices of competing and complementary products. Financial risks can generally be subdivided into the following main headings:
Operating leverage
Interest rate risk
Operating leverage is dependent on the production function and specifically the mix of fixed Interest rate risk refers to the changes to the net interest income that could arise owing to adverse
and variable cost inputs that go into producing goods. It is measured by the percentage
variations in interest rates.
change in operating earnings divided by the percentage change in sales over a specific
period of time. Liquidity risk
Liquidity risk is the risk that operations cannot be funded and financial
Fixed costs are costs that remain the same regardless of the level of sales volumes. commitments cannot be met timeously and cost effectively. Liquidity risk results from both the
difference between the size of assets and liability (the funding need) and the disproportion in their
Variable costs vary directly in relation to sales volumes. sizes.
Operating cost volatility is dependent on the proportion of fixed costs to total costs. Fixed Investment (capital) risk
costs add to the volatility of total costs by not declining when sales are dropping, but also Capital risk refers to the possibility that investments may be adversely affected from risks to which they
by not increasing when sales are picking up. are exposed.
stemming
Resource risks
Another factor that causes fluctuations in the operating profits of companies is resource risks. Credit risk is the risk that a financial contract will not be concluded according to the original set of
In the production process, the firm brings together a number of specific resources such as terms. It is the risk that a party to the contract will default.
labour, material, capital and technology. Changes in the productivity of these resources bring
about changes in the profits and therefore cause risk to shareholders.
Currency (or foreign exchange) risk concerns the possible impact that changes in exchange rates
Profit margin and turnover may have on the foreign exchange holdings or the commitments payable in foreign currencies by
The profit margin and turnover of companies also affect the operating profit. Increases in business organisations.
competition that may result in lower profit margins and/or a smaller turnover, therefore
causing risk to shareholders. lhese risks exist and arise indirectly from the business activities of the business.
The management of systematic risk entails a repositioning on the riskreturn curve. A

specific company or an investor may position him-/herselfon a risk-return curve to
determine the level of systematic risk he/she is prepared to accept given the level of return. risks
lhe required return increases with increasing levels of systematic or market risk. Operational risks refer to risks of a non-speculative nature that have no potential for showing a profit.
Traditionally, many of these can be insured.
A typical example would be the destruction of an asset by fire. If a fire Occurs, a loss is
Incidental risks
incurred; if the fire does not occur, then no loss occurs. Pure risks are usually referred to as
Incidental risks are those risks that arise naturally from the activities of a business, but are
insurable losses, since the financial consequences of these losses may be transferred to an
incidental in the sense that they do not form part of the main business of the organisation,
insurance company by insuring against 46
yet are necessary to ensure the continuation Of the main business of the entity.
The principle subcategory of incidental risks is financial risk.
45
CONCEPT OF RISK
accepting speculative risks — a decision that appears to be contradictory, since both situations may have
2.7 Psychological influences on risk
the same objective probability of occurring.24
Attempting to understand human behaviour has for a long time preoccupied the minds of
The importance ofpsychological influences on risk behaviour in economic decisions has
many in the discipline of Psychology. This has led to the development ofa recognised area
increasingly been understood and is studied in the field of behavioural economics and finance.
ofstudy and subdisciplines or specialisations such as Social Psychology, Clinical
Psychology, Psychology of Education and others.
Interest has been directed for some considerable time to the influence of risk on decision- 2.8 Market failure
making — more particularly on determining substitutes for risk and incorporating these in
decision-making techniques, usually quantitative ones, so as to reflect the effects of
uncertainty more accurately. Individuals' perception of risk and their capacity to absorb it are Public, merit and private goods
critical to the study of risk management and insurance. Situations of risk are perceived Market failure is often discussed in the context of insurance. Market failure has a number of
differently by each person, and such perceptions often depend on an individual's reaction to interrelated meanings. Its most common meaning is the failure of a market to provide some specific
risk or are even dictated by a particular entity's attitude towards risk treatment. Put simply, goods or services in general, or, more specifically, goods or services to a specific type of people.
one person's reaction towards risk, which could be described as aggressive, might be Insurance often features in this debate. Thus, some may allege that the aged cannot afford medical
contrasted with another's, whose attitude is one ofaversion to risk and who will purchase treatment, and in this case it is argued that the market has failed. In other words, there are consumers
insurance merely not to tempt fate. who require medical treatment, but the private market will not provide the medical service. A
An individual's perception of and reaction to risk are influenced by many factors. Wealth consequence of allegations of market failure is that the government must intervene in the private
(i.e. risk-bearing capacity), family background, previous experience, position or status, and market and provide the goods and services that the market has failed to provide. The discussion of
geographical location are just a few of what seem to be an endless list of factors that influence market failure in an insurance context is thus commonplace and important. A central issue is therefore
risk-taking decisions. Not only have psychologists not identified all such factors, but the whether it is legitimate for the government to supply a particular good or service or whether this
effects of those already identified on the kinds of decisions that are made under conditions should be left to the private sector. It can be argued that the state should provide goods and services
of risk and uncertainty have not been satisfactorily quantified. called public goods, the private sector private goods and in the greyarea in between are merit
Research has shown that the attitudes towards risk are unstable, in that an individual's
goods. The debate is then directed at defining these three classes of goods.25
attitudes can change when he/she is being subjected or exposed to the attitudes of a group.11
The implication that group decisions are less risky than individual ones may be inaccurate. The oldest example ofa public good is the provision of a national defence. It has been accepted
This influence of group attitudes is particularly relevant when placed in the context of for a long time that the duty of the state is to provide protection for all. Thus, national protection is
organisational behaviour towards risk and its overall profile. This often dictates what is an said to be a public good. The task then becomes one of defining the characteristic ofa public good.
acceptable level of risk to be borne and hence what level of risk should be insured. It is likely Stiglitz26 suggests that a public good has two characteristics, i.e. zero marginal costs and the
that the perceptions and attitudes of those individuals responsible for insurance are swayed impossibility ofexcluding some people. In the case ofnational defence, the marginal cost of
to coincide with those portrayed by the group as a whole. Organisational culture or ethos, protecting every additional individual is very small and it is oot possible to protect only some
therefore, affects the specific area of risk perception and management. people. If people cannot be excluded from receiving a public good, the provision of public goods
Personal or group characteristics are not the only influences on reaction to risk. The leads to the free rider problem. A private good can only exist where a public good is not provided.
A market for private goods cannot exist if some consumers are not excluded. It is thus diffcult to
type of risk (pure as opposed risk to situations speculative) are frequently also exerts
see a market for a private national defence developing where
12
desirable an influence.and Williams argues that speculative pure risk situations
undesirable. Again, it is not uncommon for individuals or organisations to purchase
A for this by of utility theory. see Friedman and (1948-279-304).
insurance and thereby avoid risk while simultaneously 47 74).
11 12
Rim (1964; 70-9). (1966; 579).
RISK MANAGEMENT'. FUNDAMENTAL PRINCIPLES CONCEPT OF RISK
4-8 RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES
2.9 Economic viability of insurance companies
the state is already providing that good. Thus, Musgrave and Musgrave 13 point out 'the Law of large numbers
market can function only in situations where the exclusion principle applies'. Ifpeople can We will now examine the question of the failure of insurance companies. Insurance companies do
be excluded, this gives rise to competition and rival goods as other suppliers offer goods occasionally 'fail', e.g. in South Africa, companies 49
and services to those who are excluded.
There are some goods that it is generally believed should be consumer goods, but not
all people will participate in the consumption. This class of good is called a merit good. such as Parity, AA Mutual and IGI. In reality, the last two did not fail as such. When the AA Mutual
Thus, for example, it is generally believed that all children should receive an education.
and IGI were finally wound up, having paid all their liabilities, the companies ended up with excess
Education can be a private good, i.e. it can be rationed to those who can pay for it. Those
assets, so it can be argued that neither was ever legally insolvent. They were wound up by regulatory
who do not pay can be excluded. The state may, however, compel children to consume
education (and usually does). As a private good, not all children will be provided 28 with an action, not legal insolvency.
education, and so the government provides it as a merit good. An insurer faces many types of risk. The most significant of these is risk of technical
Where the private market fails to provide a good, this is often seen as a reason for the insolvency, i.e. the risk that its liabilities exceed its assets. Since most insurance industries are
state to intervene and provide the good. Market failure thus leads to government regulated, the risk is that a company's assets do not exceed its liabilities by the amount prescribed in
intervention. the regulations. Should this happen, the regulator will take steps to have the company wound up. 29
An insurer's assets can fall below the prescribed level for a number of reasons. One is that the
claims exceed the premium income. In fact, in most cases, because of the competitive nature of
Market failure and insurance
insurance markets, claims always exceed premium income, and the difference is covered by
Market failure can thus be said to exist in a number of contexts. Where, for example, people
investment income.30
want a good, but the market does not provide it, it can be said the market has failed; in this
Clearly, there is a point where persistent underwriting losses can cause solvency problems. In
case, the market has failed to form. Where aged people need medical treatment* but medical
suppliers will not supply it, it can be said the market has failed. A second more limited highly competitive markets, the insurer will simply not be able to recover.
example is where a limited exclusion exists. Thus it may be that an 18-year-old may not be The insurer in setting its premium must first determine the premiums required to cover the
able to purchase motor vehicle insurance unless the insurance forms part of his/ her parents' claims; this is the pure premium. The pure premium is therefore the premium that, if calculated
insurance policy. In this case, it can be said the market has failed to provide this class of correctly, would be just adequate to cover the claims. This premium must be increased in order to
person with insurance. The market in general has not failed, since it is generally possible for
cover the expenses ofthe business and payments to the shareholders. This final premium, i.e. the one
people to purchase insurance.
that is charged to insureds, can be referred to as the commercial premium. The difference between
Yet another example is when an insurance company fails. It is argued that governments
should intervene, this time by way of regulation to ensure that insurers do not fail. Some the pure premium and the commercial premium is usually defined in terms of ratios, i.e. 60:40. Sixty
aspects of market failure and insurance are now examined. per cent of the commercial premium is used to pay for claims (pure premium) and 40% is needed
for expenses and profits. The claims, and hence the pure premium, are more completely described
in terms of probability distributions (rand losses per exposure unit), which is itself reflective of
frequency and severity distributions (as described above) for that particular class of insurance.
The insurer's risk is not that a single loss will occur, but that the aggregate cost of claims
incurred during the financial period will exceed the expected aggregate income from premiums.
This risk, as explained above, can be translated into a measure of the dispersion around the expected
13
Mugrave and Musgrave (1989, 42). 28
stigiitz (1986 g l).
CONCEPT OF RISK
mean. It is clear that although an expected aggregate loss is assumed to incur long-term it
is possible that losses may take on values ranging from zero to very high amounts. Hence,
any underlying pure premium probability distribution with a mean pure premium per
exposure unit of say X has a dispersion around x, calculated as its standard deviation o ,
which describes this aspect of the
iötventy takes •nto consideration minimum statutory surpluses
09AC,€nsated uR. as examples, for a number the markets of factors, seldom such as reflect inflation.underwriting profits.
MANAGEMENT: FUNDAMENTAL PRINCIPLES RISK
50 RISK 33 The discussion foElows closeb,r the law dehnltion numbers of risk to an the insurer central as llrnit grated theorem, by Houston which (1%4; States 51 that:1-
38).•The distribltion ot the mean value ofa set 34 This isa corntxnation of of large and mean and variance0 normal
distribution with of n independent variance 2/n and as Identically n tends towards dlstributed Infinity-randocn r See Freund. variaUeseach Wil%ams havTHJ and
2
approachesa dealt with further in chapter 15. g and ct Perles (1988" 299—301 This aspect is
insurer's risk, namely, the variation of pure premium from the average pure premium for that
CONCEPT OF 51
particular line of insurance business."
In addition to the variation within the pure premium distribution, the insurer faces an far has been to define and characterise risk — to provide some framework within which
additional risk, the sampling risk. No insurer insures the entire population of exposures (of a uncertainty may be conceptualised in an attempt to deal with it. Both quantitative and qualitative
line of business), but only a sample of that population. Should an underwriter charge a premium considerations have been discussed. The problem in the economic and management sense is to
X, derived as the average from the pure premium probability distribution, there is a risk that the develop a rationale — a model of choice — for uncertain situations that will help outline the
relevant considerations for the creation of a market (insurance).
underwriter's sample of insureds (exposures) is not identical to the population of exposures from
The model that economists have developed revolves around the concept of utility, as
which X was estimated. Simply stated, the risk is that the sample's actual average pure premium,
introduced in the preceding section. It characterises the preferences of individuals for goods and
say Xs, may differ from X by more than an expected amount. This is influenced by two factors,
services, i.e. the model explains the workings of a market for insurance from the insureds'
i.e. the quality of business and the number of exposure units. perspective. Importantly, it allows one to conclude that in an uncertain world an insurance market
Having recognised that the insurer faces this additional sampling risk, one can now will arise and hence will mitigate the effects of risk on society as a whole, which is an important
describe the risk in a mathematical form that combines both elements of risk, i.e. the variation consideration for economists.
The premise underlying the theory of utility is that generally more goods, income, etc. are
in the pure premium distribution (o ) and the measure of sampling error. 33
preferred to less; an individual therefore receives a level of utility or satisfaction in relation to
It is significant to note that such a mathematical expression illustrates the principle ofrisk the extent of his 'package: Secondly, the more the individual already has, the less satisfaction
reduction through pooling. The insurer's units insured risk will (sample vary inverselysize). 34 that individual receives from an extra unit (goods, service, income and even wealth).
To illustrate the economist's basic model of decision-making under conditions of
with the square root of the number of exposure uncertainty, the assumption is made that the utility curve in figure 2.1 applies to a consumer,
Therefore, it is the combination ofa large number ofexposures that enables an insurer to act as whose preferences for income are represented by OU.35
underwriter and reduce risk by increasing the predictability of average losses. The pooling
effect, or as can be stated more technically, the law of large numbers, reduces the sampling risk Figure 2.1 Hypothetical utiJjty curve (OU)
of insurers. Utility
In order that an insurer be viable it must charge a commercial premium which is in excess
of the insured's own losses.
31
30
Insureds' rationale for purchasing insurance

20
In order that an insurer be viable, it must charge a commercial premium that is in excess of the
insureds' own losses. On a cost-to-benefit basis, the cost on insurance produces negative
10
benefits. 'Ihis requires an answer to the question: Why do insureds pay amounts in excess of
their own losses? The concern thus
28 ;
Wealth
10 20 30 40 50
Rands ( '000)
32 me standard devlatioo may be mathematically defined as follows:
Let us assume further that the following risk situation prevails:
where:
o of the pure premium dlstributlon pp all possibk? consumers will display a diminishing marginal utility of and display an income. :ndividuals
pure premium values = mean pure premium n the number Ofexposure units. neither gamble insure have a straight utility curve (called a constant marginaf utility
which is assumed to be fairly large.
RISK
as-&2— %ere is, of course, a limit to what consumers will pay. This limit is determined by the
52 RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES
indifference premium. In practice, competition also
CONCEPT OF
Value of the house
Probability of the house being totally destroyed by fire 0,20

Definition
R6 000
Pure premium
A consumer's indifference premium is that premium at which a consumer is 1indifferent to buying
of RIO 000
or not buying insurance.
R30 000
2.10 Economic viability of insurance markets

Commercial prem•um (loss ratio 60%)
Clearly, if the commercial premium were limited to the pure premium, the market could not be
viable and hence insurance would not exist. The application of utility theory indicates that
Clearly, since the commercial premium exceeds the cost to the insured, from an economic point
of view the insured should not purchase insurance. Once utility is introduced, however, a consumers will pay in excess of the pure premium, indicating that a viable insurance market
different picture emerges.36 Using the utility curve shown in figure 2.1, utility values are can exist. Experience indicates the correctness of this conclusion. The fact that consumers do
substituted for monetary values. indeed purchase insurance gives rise to other problems. If consumers are risk averse and
demand insurance, why then is insurance limited to such a restricted range of risks - mainly
The following conclusions may be drawn:
damage to property (including consequential losses), and death and injury to people?
No insurance Wealth (income) Utiles per OU Economists have for some time tried to answer this question." A suggested solution is the
Event: No fire R40 000 31 problem of moral hazard, which could threaten the economic viability of insurance. Consumers
Event: Fire RIO 000 10 generally pay a small premium to receive a large payout. This creates an incentive to make the
Expected utiles + 26,8 insured event happen. The mere existence of insurance gives rise to a moral hazard problem,
Insurance purchased and insurance can only exist under circumstances where moral hazard can be managed. Moral
Event: No fire R30 000 28 hazard gives rise to a number of issues.
Event: Fire R30 000 28
28
Expected Asymmetry of information
utility
Insureds have an incentive not to reveal adverse information about their risks. The logic is that
The question posed is whether the individual will purchase insurance. The answer is dependent if the insurer knew the truth, the insurer would either decl_i_ne to offer insurance or, if it did,
upon whether the cover is offered at an acceptable premium. it would do so at a higher premium. 'lhus, asymmetry of information becomes a matter of
If the consumer does not purchase insurance, two states of nature exist: a fire and no fire concern in insurance matters. This implies heterogeneity of insureds concerning their utility
state, each associated with probabilities. From the above table, it will be seen that the expected curves and, thus, indifference premiums. The lack of information that would allow insurers to
utility value of not purchasing insurance (i.e. the consumer's wealth will increase by the cost of determine the correct probabilities of loss may mean that insurance is offered at premium levels
the commercial premium) is 26,8 utiles. On the other hand, if the consumer purchases insurance, that cannot provide any profit to underwriters. e This being the case, premium levels are then
his wealth remains constant at the value of the house, which has a utility value of 28 utiles. This adjusted upwards, but, because of the diversity of the groups insured, certain consumers'
consumer has more satisfaction in purchasing insurance, even though it costs more than the indifference premiums •nay be exceeded and therefore total revenue to insurers declines,
consumer's cost of losses. Utility theory thus appears to explain why consumers purchase despite the ncrease in the price of cover.
insurance, or, stating it differently, why a market for insurance exists.
53
determines the premium that insurers can charge.
36The of Mllity is discussed 10 greater detail in chapter 3.

54 RISK into account the financial mechanism of an insurance market to fund the
possible financial outcomes resulting from risk
Having reviewed what may be regarded as the key features of the area
Adverse selection
A second related problem is that ofadverse selection by insureds, which occurs when insureds decide to
purchase insurance when they believe losses will arise, but do not pass this information on to the
insurers. An example of this would be where someone discovers that he has cancer and goes to purchase
medical or life insurance, but, of course, withholds the medical information. Adverse selection is an
example of asymmetry of information. The process of adverse selection, therefore, entails the purchase
of cover by those who are increasingly certain that they will need insurance.
of uncertainty or risk, the chapter that follows focuses on decision-making under conditions of risk and
uncertainty.
Indemnity and co-insurance 55
One method of dealing with the moral hazard problem is to ensure that the
insured never receives more than the loss suffered, and preferably carries some
of the loss. What the insured recovers is thus limited to the indemnification of
his loss, and preferably the insured should be a co-insurer for a portion of the Chapter 3 Decision-making under
loss. This is usually in the form of an excess.
Clearly, dealing with the moral hazard problem restricts the scope of conditions of risk and
insurance and may thus explain why insurance is not much more widespread
in its application. uncertainty cneoo E*Estes
2.1 1 Summary
The objective of this chapter has been to provide a general background to Introduction
risk management by examining the concepts of risk and uncertainty, and the
Expected monetary value (EMV) criterion
definition and classification of risk, With regard to the latter, a degree of Probability theory
rigour was imposed, while explaining, from a risk management point of 3.4 Probability distributions
view, the two distinct elements of event and outcome, which are inherent Expected utility criterion
In uncertainty. Maximin decision criterion
The discussion has, of necessity, considered the question of risk
3.7 Minimax regret decision criterion
quantification and decision-making under uncertain conditions, taking Maximax decision criterion
Summary
RISK
3.1 Introduction
Attempts have been made over a long period of time to develop a theory for decision-making in general, and
in particular in those situations that involve conditions of risk uncertainty. Some of these theories and their
application to risk management are discussed in this chapter.
3.2 Expected monetary value (EMV) criterion

The earliest theory suggested that decisions be made in terms of the EMV decision criterion. The origin of
this theory is usually ascribed to Pierre de Fermat (1601-65) and Blaise Pascal (1623—62). According to this
criterion, when facing a decision where the outcomes can be in monetary terms and where the probabilities of
these outcomes are known, the decision maker should choose the path that has the greatest EMV. The EMV
is also represented by symbols such E{X} and g. Mathematically it is defined as follows:
DECISCON-MAKING UNDER CONDFTIONS OF RISK AND UNCERTAINTY
57
MANAGEMENT: FUNDAMENTAL PRINCIPLES DECISION-MAKING UNDER CONDITIONS OF RISK AND UNCERTAINTY
56
Definition
EMV=
Also
Also
where:
EMV
where
diagrammatically
Figure
Thus, if the EMV of, say, a profit-generating project exceeds the stake (S), then the decision maker should undertake the project. This theory requires that the probabilities and outcomes be determined. Hence, a discussion on
probability theory and its relationship to risk and risk management is necessary.
Example
Assume that a building has a capital value of R50 million and that the probability of a fire causing a total loss (which for the purposes of the example is the only loss that is of concern) is 0,0001. What premium should an insurer charge?
Use the EMV decision criterion to answer the question.
These facts can be set out in a states of nature matrix.
States of nature
Totals
Fire No fire
Outcomes (R50 000 000) 0
Probability otoool 059999 1

Contribution to EMV (R5 000) {R5 000)
If a fire takes place, the insurerfaces a R50 million loss.The EMV ofthe loss is R5 000. The pure premium Of R5 000 is thus required to cover the cost of the loss. This amount would, of course, only be
suffcient to pay for the losses. For insurance to be economically viable, an additional amount to pay for the administration costs and to return a profit to shareowners is required.
The successful operation of insurance is thus based on statistical theory, which is needed to refine our understanding and management of an insurer's risk. The more fundamental aspects in particular have been examined; other risk aspects
do, of course, exist, so that the risk borne by the insurer can assume a multiplicity of meanings. What is relevant is that while the theories of probability and sampling have greatly contributed towards approaching risk from a more rigorous
viewpoints their inherent assumptions present major
I mitations, so that it becomes necessary to adjust one's approximations of expected outcomes by factors that are non-quantitative in nature.
ulere are instances where EMV = Ep.A is inadequate, since this equation does not indicate the degree of confidence that the EMV will be achieved. For example, assume you have a car valued at R200 000. You check with some friends
who have a fleet ofsuch vehicles and they advise you that the probability Of this type of vehicle suffering a total loss is 031. You decide not to insure this vehicle, but to set aside the premium to pay for the possible loss of the vehicle.
58 RISK
3.3 Probability theory noeJ
The pure premium is O, I x R200 000 = R20 000. The formula indicates what the pure premium The risk elements included the concern whether or not an event will occur and, if so, the extent
should be, but this information is not particularly useful. In fact, in this case, an outcome of R20 of the outcome. Where the possible range of outcomes can be specified together with their
000 will never be achieved, since the actual outcomes are either RO or R200 000. On the other probabilities, probability theory is of value.
hand, if you have a fleet of50 vehicles and set aside an amount of50 x R20 000 = RI 000 000 Where outcomes and probabilities are knowns risk is involved in terms of economists' meaning
p.a. in a fund (P), this amount is adequate to pay for five new vehicles a year. So now the question of risk
becomes, How confident can you be that the fund of RI 000 000 will be adequate? An amount There are various types of probabilities:
ofRI 000 000 + RX is needed, where RX is chosen to give the desired degree of confidence. To • objective or a priori probabilities
determine RX, a more detailed understanding of probabilities and probability distributions is • subjective probabilities
required. The mere expected value is inadequate. • relative frequencies leading to empirical probabilities.
A priori probabilities reliability of these estimates based on a number of underlying assumptions: firstly, a suffciently
large number of past events have been observed and, secondly, casual influences temain the
A priori probabilities - also referred to as objective probabilities — are those that are intuitively same as those in the past. Naturally, the future is seldom a perfect tepeat ofthe past, and hence
known. For example, the probability of a head appearing when a coin is flipped is 1 /2. A head margins oferror in estimates must be anticipated. Statistical techniques have been developed to
is one of only two equally likely outcomes (head or tail), Intuitively, since there is no logical measure the probable range of deviation ofactual from expected outcomes. Empirical
reason why a head should be any more or less likely than a tail, the probability of a head probabilities are derived from frequent events, such as death in a very large population or car
appearing is 1 /2. accidents arnong a large fleet. Empirical probabilities can be misleading where infrequent vents
A priori probabilities are determined without observing the actual frequency of occurrence; are involved. Infrequent events are sometimes called Black Swan events.
the probability values are calculated on the basis of deductive reasoning alone and assuming
that the occurrence of the event is due to chance alone, i.e. subject to randomness alone. A priori
probabilities of outcomes are usually determined for problems associated
with games of chance such as dice rolling and coin spinning. ' based in known guesses.
However, it is not possible to determine most probabilities intuitively. They must either be
60 RISK
estimated or determined by experimentation (i.e. empirically) or observation.
59
Since they are by definition infrequent, a database does not include these and probabilities from
such a database can be misleading. Empirical probabilities are also prone to problems of
Subjective probabilities survivor bias, since the database usually only includes data about survivors.
Subjective probabilities are those that the decision maker must simply estimate. Usually, the It can be seen, therefore, that statistical probabilities can alter either on account ofchanges
estimate is made by contrast to a known probability. Subjective probabilities are resorted to in our condition ofknowledge ofcauses (affecting deductive reasoning) or because of changes in
where insuffcient information is available to determine the probability in a more objective underlying forces (affecting inductive probabilities). For the purposes of this book, the empirical
fashion. Thus, if one is asked to determine the probability of dying of anthrax, the approach will probability is regarded as the relative frequency with which an event occurs in the long
be to know that the probability of dying in a car accident is 0,0001, then estimating the run.
probability of dying of anthrax. Experiments have indicated that people tend to underestimate It is interesting to note that in the short-term insurance industry, traditionally probability
well-known events and overestimate rare events. Thus, the probability of dying of a heart attack theory has not been used to determine premiums. Loss ratios are used. The relationship between
will be underestimated and dying of probability theory and loss ratios is now discussed. In short-term insurance, the total income P
from a class ofbusiness is known and as the losses (L) are recorded, the losses become known.
Assume thus that the total premium is known, i.e. P = N.p, where P is the total premium. It is the
fund available to pay for the losses (L) and N is the number of insured units upon which identical
Relative frequencies leading to empirical probabilities premiums (p) are levied. The losses (L) are distributed in different class intervals C or L = inc;
A third method ofdetermining probabilities is by observation, experimentation or measurement. the losses equal the sum of the losses in each class interval. Since the pure premium P is equal
In practice, this is the most common method. One cannot, for example, intuitively determine the to the loss L = Enc, by definition it is the amount required to pay for the losses, i.e. the ratio L/P
a priori probability of death. This probability is obtained from data. equals 1. In this event:
One can generalise an estimate of past outcomes and extrapolate these to a future event by
observing its relative frequency of occurrence. ms, if the probability of death over a given period P.N = En.c or
is required, a sample ofpeople alive at the beginning of the period can be obtained and the
number of deaths during the period recorded. The ratio of the number of deaths to the total
number of people in the sample then can be determined. This ratio is the relative frequency. This is the formula that would be used to determine the premium if probability theory were to
Where there is confidence that the relative frequency represents the probability, this be applied. The loss ratio method is thus a practical way of arriving at the premium. Of course,
probability is referred to as the empirical. This confidence usually comes with increasing in order to cater for expenses and returns to shareholders, the loss ratio is usually less than l, say
numbers. Such probabilities are also referred to as statistical or inductive probabilities. l The 0,6.
The loss ratio method is particularly useful, since it can determine the premium where a Example
range of losses occur, i.e. not only a single total loss, and where more than one loss per insured A firm that transports cars to customers faces the possibility of damage to the cats while in
takes place. This method is limited to cases where previous events have taken place. transit. If a damaged car is considered to be a total loss and the probability of loss of a single
car is O, I, the binomial distribution can be used to determine the distribution of the number
of damaged cars. When a group of two cars is considered, N = 2 and p = O, 1. The
probabilities of O, 1, or 2 losses can be calculated using the following formula:
3.4 Probability distributions
Probability distributions are useful tools for evaluating risk. If the risk manager knows the
probability distribution of the outcomes of an event, it is straightforward to calculate the most
probable outcome or the riskiness of the event. Three theoretical distributions are useful in risk
management, 61
the binomial distribution, normal distribution and Poisson distribution. "Ihe usefulness of each
distribution depends on the type of problem the risk manager is addressing.
The binomial distribution

Ihe binomial distribution is particularly useful to the risk manager, since it deals with two states
of outcome, i.e. a loss or no loss. Take our above example ofa car fleet of 50 vehicles with a
probability of0,I that a vehicle will be written off. Each vehicle can either suffer a loss or no loss.
In fact, it is possible for any vehicle to suffer more than one loss during the year, but for the
purposes of this illustration, this will not be taken into consideration. 'Ibe fleet can suffer exactly
O losses, 1 loss, 2 losses ... 50 losses. Thus, for a fleet of 50 vehicles, a total of 51 possible losses
is possible. In general, if the fleet has N vehicles, N + 1 losses are possible.
The binomial distribution is a two-parameter distribution described by the following two parameters:
the total number of events (N) the probability (p) that the
event will take place.
The binomial distribution can be used to determine the probability distribution of the number
of accidents that can take place in a fleet of N vehicles that are independently exposed to the
risk of accident.
formula for the binomial distribution is:
r!(N— r)! V(I —p)N-r where:
p(r) = probability that exactly r accidents will take place
N! is called the factorial 0! is

equal to l.
RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES DECISION-MAKING UNDER CONDITIONS OF RiSK AND UNCERTAINTY
62 63
0 losses 21 x = 0181.
The normal distribution
Probability of exactly The normal distribution is an example of a two-parameter distribution. The distribution is
2! = o, 18. determined completely by two parameters. The only information required is the expected value
Probability of exactly 1 toss — x (or mean — g) and the standard deviation (o). Once the expected value and the standard deviation
are known, the probability ofany desired value can be determined. The normal distribution is
2 tosses = 2! x (0,1 = 0,01.
Probability of exactly bell-shaped. A usefulness of the normal distribution is that, under certain circumstances, other
distributions such as the binomial distribution can be approximated by the normal distribution.
For the binomial distribution, the expected number of losses (g) is given by the formula Np, A typical normal distribution is shown in figure 3.3.
while the standard deviation (o) is given by UNE(TZF).
Figure 3.3 A normal distribution
Take another example: assume that there are 50 items each exposed to a loss with equal
probability. This means there are 51 possible outcomes ranging from 0 losses to all 50 items
suffering a loss. 'Ihe probability distribution is shown in figure 3.2.
Figure 3.2 A binomial distribution
50 exposed items
Or 1200
0.1000
orogoo
0.0600
illustrate the use of the normal distribution, take a problem where the outcomes can be
0,0400
described by the binomial distribution, which approximates to the normal distribution. Assume
0,0200 that there are 500 units with a mean value oflosses ofR5 000 and a standard deviation of R500.
The standardised normal distribution table from which the probability that an outcome will
0.0000 exceed e, certain selected value for a normal distribution can be obtained from any textbook on
Exact number of accidents
uExact number of accidents
statistics.
The expected value is = NP = 50 x 1/2 = 25, with a probability of this exact number of accidents Ef the pure premium is set at the expected value of R5 000 (or at the beginning of the year
being 0, 1123 (derived from the above formula). a fund of R2,5 million is created), the issue to be determined is how likely is it that the cost of
We are now in a position to determine a degree of confidence. We can be 66941%
claims will exceed the size of the fund. Since the distribution is symmetrical, there is a 50%
confident that there will be 26 or fewer accidents, 76,01% that there will be 27 or fewer, and
83,89% confident that there will be 28 or fewer. probability that the fund will be exhausted. The risk manager may, on the other hand, set the
We are thus now beginning to be in a position to determine A in the equation EMV = premium at a greater value. This greater value can be set as a multiple of the standard deviation.
Ep.A ± A. The results are indicated in table 3.2.
64
Table 3.2 Selected threshold points for a binomial distribution that approximates to the
6,7
normal distribution any risk typically addressed by managers is accurately described by any known theoretical
distribution. Despite this reservation, the principles used to measure risk using these distributions
size = expect
apply generally to risk management, and the results may apply approximately to risk management
mu%tipteof
standard deviation problems.
Expected value (W) + O o R2 500 000
Expected value (p) + 0 R2 512 500 The Poisson distribution

Expected value (p) + 1 0 R2 525 000 The Poisson distribution is named after a French mathematician who observed that 'trouble comes
Expected value (p) + 1 0 R2 537 500 in bunches'. Like the binomial distribution, the Poisson distribution is a discrete distribution,
making it useful for describing the possible number of accidents. Unlike the binomial distribution,
Expected value (g) + 2 a R2 550 000
however, the Poisson is a single-parameter distribution. The single parameter is the distribution's
Expected value (y) + 0 R2 562 500
expected value. The formula for the Poisson distribution is:
Expected value (p) + 3 R2 575 000
Expected value (p) + 0 R2 587 500 r!
where:
It now can be seen that by using the normal distribution it is possible to select A to give the
g = average or expected number of accidents e = the base of natural logarithms
degree of confidence required. Thus, if A = o, the confidence level is 84, 13%.
(approximately 2,718) r! = r(r— l) (r — 2) ... with 0! = I p(r) = probability of
each value r.
Example
The probability of claims exceeding the selected fund size of R2 537 500 is 6,7% (according The Poisson distribution approximates the binomial distribution where N is very large and p is
to table 3.2), which is the probability of claims exceeding By symmetry, the probability of
very small, holding the expected value NP constant g. The Poisson distribution could apply to a
claims being less than p - rand equivalent value is:
very large number of units exposed to risk, each facing a very small chance of an accident. For
example, a vehiclekilometre may be considered the unit of exposure for vehicle accidents, with
R2 500 000 (145 x R25 000) = RI 462 500.
a very small probability of an accident in a single kilometre. This approach allows for the
The probability of claims being greater than R2 537 500 or less than R2 462 500 is the sum possibility of multiple accidents to the same vehicle, which the binomial distribution does not
of the two probabilities, i.e. + allow.
The value of claims must lie either outside these bounds or within them, hence the Another possible application is to use a worker-day (one worker for one day) as the unit of
probability of an outcome lying between R2 462 500 and R2 537 500 is: exposure to work-related injury.
Table 3.3 shows a Poisson distribution with an expected value (g) of 2,0. In this example, g = 2 and e-2
Usually, a risk manager's concerns are focused on cost exceeding a given fund size.
is approximately 0,1353.
It is important to note that the distribution and the resulting estimates in table The mean p of a Poisson distribution is also its variance. Consequently, the standard deviation
3.2 apply only approximately to observed occurrences. It is doubtful whether of a Poisson distribution is the square root of its expected value:
65
The single-parameter feature makes the Poisson distribution convenient to use, as only one threshold(s) = required accuracy.
parameter estimate is required. This same feature is also a shortcoming - the Poisson distribution
is not appropriate as a model unless the assumed relationship between the expected value and of this concept can be illustrated by the following example.
the standard deviation is reasonable.
Table 3.3 Poisson distribution with expected value of

wants to be 95% certain that costs will not exceed RI,5 million. risk manager is
0,1353 probability of O accidents willing to tolerate at most a 5% probability exceed RI,5 million. Thus, if costs follow
0,2706 probability of 1 accident a normal distribution value of RI .2 milllon and a standard deviation of Æ2 371 , the
0,2706
Probability of 2 accidents meets the requirement to the threshold points for a normal distribution presented
0.1804 probability of 3 accidents chance is present that normally distributed costs wilt tying 11645 standard
0,0902 deviations above the expected value.
Probability of 4 accidents 0,0361
Probability of 5 accidents
For the assumed distribution of costs the threshold value [5:
Probability of 6 accidents0,012
0,0034
371) = RIS million.
Probability of 7 accidents
0.009
Probability of 8 accidents Normally distributed costs with an expected value of RI 10 million and standard deviation of R400
Probability of 9 accidents0.0002 000 do not meet this requi rement, as
1,0078
000) = RI 658 000.
Total
than the RI million level to which the risk manager wishes to

As for the binomial distribution, the calculation becomes a diffcult task when the possible
number of accidents becomes larger than, say, 20. Fortunately, the central limit theory (the
normal distribution) applies when the number of accidents (r) becomes large, which allows us
to approximate a Poisson distribution using a normal distribution with expected value N and Expected utility criterion
standard deviation The approximation is very close if p > 50, in which case the
Utility theory was adopted in an attempt to explain why people make decisions differently than
Poisson distribution becomes virtually indistinguishable from a normal distribution with a mean
of 50 and a standard deviation offi = 7,07. suggested by the EMV criterion. Take the example suggested by Bernoulli (1738), where a
poor man finds a lottery ticket that will pay R20 000 if a head appears. The expected value of
Accuracy of estimates this game is RIO 000. According to the EMV theory, the poor man should not accept an amount
a wealthy
less than RIO 000. If man, who is going to purchase a ticket in any event, sees the
If the probability distribution of costs is known or can be reasonably approximated, it is possible poor fellow picking up the ticket and offers the poor fellow say R8 000 for the ticket, e general
to estimate the maximum probable cost (MPC) or to determine conditions for predicting consensus is that the poor man will accept the offer rather than play the game. He would prefer
outcomes with a specified degree of accuracy. Ihe procedure is based on two requirements: a certain R8 000 rather than face the risk of getting nothing. Here we have an example of a
(i) a known probability distribution of costs, or a reasonable approximation, and parameters person taking an amount less
that can be estimated
(ii) a degree of risk tolerance or tolerance for the chance of an incorrect prediction.
67
these requirements are met, the problem can be formally stated equation:
68 RISK
than predicted by the EMV calculation. Bernoulli agued that people do not value money in a DECISION-MAKING UNDER CONDITIONS OF RISK AND UNCERTAINTY 69
linear fashion, but that the marginal utility of wealth (U(W)) diminishes as wealth (W) increases.
If the poor man values R8 000 as 1 utile and money is valued linearly, then R20 000 is valued at
Figure 3.4 Hypothetical utility curve
2,5 utiles. If he plays the game, then the EMV of the game is 1,25 utiles, which is greater than
the I utile represented by the R8 000. Therefore, he should play the game rather than accept the Utility
R8 000. On the other hand, if he values R20 000 as only 2 utiles, then the EMV is l, which
represents the R8 000. Under these circumstances, he would accept any amount in excess Of R8
000 and not take a chance to play the game.
This suggests expected utility value (EUV) methodology for utility calculations similar to 30
the EMV criterion:

Utiles
EUV = Ep.U(W). 20
In the above example, the utilityvalue ofR20000 was established, i.e. t.J(20 000)
= 2 EUV = I, and this was compared with the utility value of R8 0009 i.e. 10
000) = 1.
It is important to note that this problem could be solved by applying mathematics to utility.
This accepts that utility is cardinal in nature. Income
10
Utility has a long history in economic theory. From 1776 onwards, after Adam Smith 20
30 40 50
published his Wealth of Nations, economists were taken by the concept ofutility, but had Rands ('000)
diffculties in measuring it. In the end, it was accepted that utility-based preferences could be at
least be ranked, so this branch of inquiry developed into the concept of ordinal utility. Preference The concept of diminishing marginal utility implies that losses will result in greater
proportional loss of utility than identical gains, i.e. increasing marginal disutility. This explains
theory then developed out of utility theory.
why mathematical probabilities alone do not explain the purchase of insurance. Plainly, the
Modern cardinal utility was ushered in by the publication ofVon Neumann and
disutility accompanying a premium payment will be less than that associated with suffering an
Morgernstern's Theory of Games and Economic Behaviour. This work produced the five axioms
uninsured loss, even after making allowances for the assigned probabilities of loss.
of cardinal utility:
Utility is, of course, personalised in nature and consequently each person has his/her own
• comparability (sometimes called completeness) utility function, based on a myriad of factors such as, for example, wealth, age (or life cycle)
• transitivity (sometimes called consistency) strong independence . measurability and income (or earnings). A utility function could nevertheless be conceptualised, and
• ranking. disutilities to potential monetary loss subjectively assigned in order to examine the resultant
decision rules. Although the utility function would necessarily be only hypothetical, the
The concept of utility14 purports to characterise an individual's preference for goods and services.
exercise or methods of arriving at an appropriate course of action would not be dissimilar to
Generally, more goods and services are preferred to less and generally the more one has, the less those discussed earlier, using monetary values.
satisfaction (measured in utiles) one derives from an extra unit of income (denominator for the
acquisition of goods and services referred to). Hence, the concept displays diminishing marginal
utility.
14
The concept of utility has become ba5iC for the current microeconomiic the authors theory attempted of individual to reconcile econornÉ risk seeking behaviour. and
Reference risk averslon can to be cardjgmadel
to the article by Friedman and Savaoe 0948}. where utllity.

RISK MANAGEMENT. FUNDAMENTAL PRINCIPLES In
71 the
states of
3.6 Maximin decision criterion nature matrix show in table 3.4, the decision would be to purchase insurance without a deductible,
since if a fire occurs, this is the best option.
States of nature Pessimistic state of
nature (fire)
Alternatives Fire No Fire 3.7 Minimax regret decision criterion
(RIOm) (RIOm)
Regret may be defined and quantified as the difference between the monetary loss of a course of
No insurance
(R8 000) (R8 000) (R8000) action selected and the value of the final outcome. It is a practical parameter for consideration,
Purchase insurance (pay
Yes since often, after the action has been selected and an outcome materialises, some regret may be
premium)
(RI 05 000) (R5 000) (RI 05 000)
experienced for having selected the 'incorrect' course of action. It can also be seen as a
performance yardstick in the short term — it is not inconceivable that decision makers'
Table 3.4 Maximin decision criterion performance is evaluated either consciously or unconsciously by the application ofthis measure
ofregret. Indeed, it can be argued that the measure ofregret (and concomitant consequences)
exerts influences, often irrational in nature, on the decisions of organisations concerning pure
risk situations.
The same course ofaction ofpurchasing insurance cover would most likely be chosen if it
were done to satisfy the objective ofminimising expected regret. Again, this would be because
Purchase insurance with R 100 000 deductible
the regret (or, as a further extension, the anxiety factor) associated with the expenditure of
premium when no loss occurs will most likely be outweighed by the regret (or anxiety)
In most cases, players aim to win when playing games. Sometimes, however, the consequences
experienced when a loss occurs and when no insurance had been purchased.
of losing may be untenable. fius, take the risk of a fire. A fire may or may not occur. If it does,
It is significant to note that this latter comparison ofregrets becomes more meaningful
and the company does not have insurance, the consequences for the company may be fatal. Under
when the decision maker is the professional risk manager whose performance may well be
these circumstances, the decision maker should not look at the favourable outcomes (no fire), but
measured in the light of actual events. Thus, in a practical context, the anxiety factor, although
the unfavourable outcomes (a fire) and select the best option if the worst-case situation happened.
diffcult to quantify, becomes an important influence in the decision-making process.
This is known as the maximin strategy. Under this strategy, the decision would be the selection of
the same course of action as with the use of monetary values, given the fact that normally a person
would not be adverse to having additional 'rands'. When the maximin strategy is adopted,
probabilities are ignored. On the other hand, if probabilities of occurrence are assigned to the 3.8 Maximax decision criterion
various outcomes (expected monetary loss/disutility) the course of action selected may indeed be
Table 3.5 Maximax decision criterion
different to that where the objective is to minimise monetary loss. This is almost certain to be the
case if applied to the question of purchasing insurance. The cost of premiums is almost certain to
States of Optimistic state nature of nature
exceed the expected value of losses. This could result in the decision rules pointing toward the
purchasing of insurance rather than to the retention (to varying degrees) of risk (no fire)
Alternatives Fire No fire
No insurance (R 10m) (RIOm)
Yes
Purchase insurance (R8 000) (pay premium) (R8 000) (R8 000)
Purchase insurance with (RI 05 000)

RI 00 000 deductible (R5 000) (R5 000)
Conversely to the maximin, the maximax objective is characterised by extreme
outlook. optimism, In where the states the course of nature of action matrix to in be table
adopted 3.5, is if based the maximax on an optimisticdecision criterion is adopted, the decision
maker will work on the assumption that a fire will not take place and, if so, then not to purchase
insurance is the best Outcome. It should be clear that this decision criterion is not normally
applied to the decision to purchase insurance.

72 RISK
73
Examples Figure 3.5 Loss severity distribution curve4

Now that the theory has been set out, two worked examples are given to demonstrate the theory.
Com ined
00200 Esorn te
Example 1 : Determining the premium for a vehicle fleet

010160
The following example, taken from data relating to losses3 over the period of four years,
illustrates expected costs. The purpose of the exercise is to determine the premium that
should be charged given the claims history shown in table
3.6. An 84% confidence level is required and assumes that a 60:40 split exists o,0D80 between cost of claims and other expenses.
Table 3.6 Frequency and loss costs per exposure for sample vehicle fleet per o,oooo severity interval
€2 500,00
1 750,SO 17 -gooso 3(Ü000.oo
Loss range Year 1 Year 2 Year 3 Year 4 37 500,so 87 soo,ooj 50 oooso
2753 2 753 2 999 2 837 each year can be determined from the application of the
The expected cost for
51 69 80 52 standard formula, EMV = EPC
1001 2500 68 62 65 56
2 501 5 000 39 50 34 50 Table 3.7 Expected cost for each year
5001 10000 42 40 34 47
Year
10001 25 000 35 40 45 36 Mid point (O n ErÜN.c Year En/N.C2 n Year inm.C3 Year in/N.c4
25 001 50 11 14 18 21 0.00 2753 0,00 2 753 o,oo 2 999 o,oo 2837 c,oo
50001 75000 7 4 6 4 1 000 500,50 51 8,47 69 11,34 80 1417 52 834

001
75 001 100 000 10 3 4 2501 5 000 1 750.50 3968 48.5539,51 5062 3564 65 34,57 56 31,43 100001 200000 4 2 5 5001 10000 ? 500,50 42 104,55 40 98,5361,58 3434 38,757749 47so 60,12
200 000 400 000 2 2 7 10001 25 mo 17 500.50 35 20329 40 22939 45 239.30 36 201,99
3013 3 045 3 291 3119 5C000 37 50050 11 136,91 14 172,42 18 20s, 11 25269
5000t 75000 62 500,50 7 145,21 4 82,10 6 113,93 4 80,15

100000
The loss severity distribution is depicted graphically in figure 3.5. A frequency/ +20001 200000 t 50 87 000,50500.50 4 29,0414 102 287,3698>2 53 227,9079,76 54 11222 severity graph can be
determined for each year, or a combined graph for the 400000 300 00000 199,14 98.52 2 i62,32 7 67329 four years. From the figures, an estimate can be made of the graph. Figure 3.5 3013 113,80
3045 1 175,91 329! 1211.30 3119 77353 indicates the combined graph and estimates of the curve.
3 These bosses reøesent octual loss experience {Inflation adjusted) ora leading South Afncan hauber The bosses relate to accidental damage, 4 '1 n that even when no losses occur, the insurer incurs administrative cost* fire and theft and exclude losses ofa ilabiiity nature
RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES DECISION-MAKING UNDER CONDITIONS OF RISK AND VNCERTAINTY
74
RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES
75
From table 3.7, it can be seen that the EMV for each year is RI 113,80, 3.9.
RI 175,91, RI 211,30 and RI 773.53. The mean of these four values is

Table 3.8 Hypothetical situation of risk: Data for decision rules
RI 318.64. This would represent the pure premium at a 50%
Table 3.9 Assumed financial impact resulting from courses of action under loss and no-toss outcomes
confidence level. If a normal distribution is assumed by adding one
standard deviation (a), an 84% confidence level is assumed. The
standard deviation for the four years is R264,95 or a pure premium States of nature
of RI 583,59 should be charged at an 84% confidence level. Provision Decision Loss No loss
should also be made for administration costs. Thus, if a 60:40 split is (R)
assumed between costs of losses and administration costst the 1. Retain risk entirely
commercial premium would be R2 639. If the fleet size remains at 3 Insurable loss 150000
119 vehicles, then the size of the fund would beR8231 041. Uninsured loss 60000
Tota' 210000
2. Purchase insurance
Example 2: Considering alternative courses of
action Premium 12 000 12 000
Uninsured loss 60000
It is assumed that a company is exposed to total loss as indicated in
Total 000
table 3.8 and is considering one of four alternative courses of action
3. Purchase insurance with deductible
against the risk of such a loss:
Premium 9000 9000
(i) retain the risk entirely
Cost of deductible
(ii) purchase insurance
(iii) purchase insurance with a R5 000 deductible (i.e. an element Uninsured loss 60 000
of the risk is retained with a 'commensurate' premium Total 24 000
reduction) 4. Retain risk, but implement loss prevention measures Insurable loss
(iv) retain the risk, but implement certain loss prevention 2 000
Uninsured toss 60 000
measures - ata capital cost of R20 000 and depreciated over a
10-year period. These measures will reduce the severity of Cost of prevention measures 20 000 Total 000
any possible loss, as indicated in table

Value (R)
A more appropriate method than assuming either loss or no loss is to assign
Assets 150 000some weight to possible outcomes. Such weighting reflects somewhat on the Uninsured loss 60 000chance of a Joss occurring or not occurring. The resultant loss then becomes
insurance premium 1 2 000the expected monetary loss and represents the average period loss over a large Insurance premium with a R5 000 deductible 9000number of periods.
Probability of lossUsing the probabilities of losses assigned in table 3.8 and applying these to without prevention devices 0,05the monetary outcomes as listed in table 3.9, the expected losses
are computed with prevention devices 0103in table 3.10 in respect to the four courses of action shown in table 3.9.
Table 3.10 Expected losses associated with each course of action Table 3.11 Possible monetary regrets associated with each course of actio
EMV
States of nature Regret (R)
0,05 0,95 Action Assuming a loss outcome Assuming a n040ss o
Probability
Fire No fire 1 210 000-72 000 = -138 000 No regret
Decision
(RIO 500) 2 72 000-72 000 = No regret -12 000
(R210000)
000) 3 74 000-72 000 = -2 000 -9 000
2 (R72 000) (R12000) (RI 5
4 190 000-72 000 --118000 -2 000
3 (R74 000) (R9 000) (RI 2 250)
Probability 0,03 0,97 1
4 (R230 000) (R2 000) (R8 840)

76 77
The above results indicate that in order to minimise the expected Possible regrets associated with each of the four courses of action
monetary loss, the appropriate course of action to adopt would be to according to the illustration are set out in table 3.11.6
retain the risk, but to institute loss prevention measures. If the

S The study of subjecdve probability is stifl in its infancy. Nevenheless, studies have indicated that th15 divergence between
probabilities correctly reflect the possibility of loss and if insurers use probabi'ities 6 'normarand. fvaherrnore. predictable within limits See Cohen (1960: 14) and Edwards (1961 : 478).
the best possible course of action undel eatb outcome and the various courses actiC0
these probabilities, then, theoretically, insurance would never be
The minimax and minimin objective applied in the context of regret analysis would result in the purchasing of
purchased. Insurers' costs include administrative and other costs,
insurance (action 2) and the retention of the entire risk (action respectively.
resulting in premiums charged to the public being higher than the
As noted previously, this set of decision rules take no cognisance of the probability of loss and therefore
expected value of loss. The probabilities assigned by individuals
once applied, the expected monetary regrets associated with each course of action become those indicated in
seldom coincide with those assigned by insurers. Access to relevant table 3.12.
and sufficient data is a major consideration, andr in the final analysis,
it is more likely that individuals would assign probabilities that are Table 3.12 Expected monetary regret associated with each course of action
arrived at subjectively rather than objectively. It is argued that such

probability evaluations, based on belief, judgment and the like, may Action Expected monetary regret (R)
vary widely from the underlying mathematical probabilities and -138 900
2 400
hence may be considered to be irrational?
-2
Table 3.10 sets out the assumed financial impact resulting from
3 -8 650 -118
4
the four courses of action (given in table 3.9) under loss and no-loss -5 480
outcomes. Although it is recognised that there may in fact be
In the longer term, the course of retention of risk with loss prevention measures (action 4) will yield
numerous objectives to be satisfied, this illustration refers to four
specified objectives around which decision rules may be applied: the least average expected regret.
• to minimise the maximum potential monetary loss (minimax)

As mentioned earlier, the objective has been to define and develop a set of decision criteria and thereby to offer a
• to minimise the minimum potential monetary loss (minimin)
framework within which one may rationalise decision-making as far as risk taking is concerned. Resultant monetary
• to minimise the expected monetary loss to minimise the
values as measures of potential loss have their shortcomings, nevertheless, for they may have varying effects for
expected monetary regret.
different individuals.
The expression of expected loss or regret in absolute terms may not necessarily sumce, so that it becomes
necessary to consider the decision in the context of utility.
6 The regret •s based upon the difference between available-
Chapter 4
Corporate governance and
enterprise risk
management
7879
3.9 Summary
Decision theory under conditions of risk and uncertainty lends itself to a number of mathematical
techniques, as indicated in this chapter. The most commonly used is the EMV decision criterion.
However, this did not seem to explain actual decisions made by people, and the expected utifity
criterion was therefore developed, which appeared to be better able to explain some decisions. 't
became clear that in some instances, particularly where catastrophic losses can occur,
probabilities may not be appropriate at all, and other decision criteria that emphasise the
magnitudes of tosses rather than probabilities were developed.
4.1 Background
Traditional risk management practices tended to focus on the management of insurable risks,
such as losses resulting from fires, thefts and liability claims. These practices also included the
responsibility for the buying of insurance and occasionally for occupational health and safety
programmes. Over the past few years, there has been a marked trend towards the expansion of
risk management to include the management of other risks in the organisation, a trend that seems
to be driven by the following factors:
The restructuring oforganisations has tended to broaden the responsibilities of all managers.
• Increasing competition has forced organisations to scrutinise cost structures.

Certain practices such as total quality management (TQM) and just-in-
time (JIT) all stress the need to control risk and to do so in an integrated fashion.
The consolidation of financial services has resulted in the integration of insurance, banking
and other financial services, which in turn has led to a broader thinking about risk
financing.
RISK MANAGEMENT: FUNDAMENTAL PRINC'PLES CORPORATE GOVERNANCE AND ENTERPRISE RISK MANAGEMENT
80 Guidelines and rules on good corporate governance spell out the principles that directors
and boards must apply in governing and managing the business to fulfil its responsibility
fie absence of coordinated risk management practices has been a feature of many sensational towards shareholders and other stakeholders. Several of these guidelines and rules have
been formulated for corporations operating in different parts of the world. One of the best
and highly publicised disasters and events.
known is the Committee on 81
But perhaps the most significant driving force behind the development of enterprise risk
management (ERM) was the emergence ofcorporate governance and its emphasis on the Financial Aspects of Corporate Governance (the Cadbury Committee) in the UK, which
boardroom risk management. The following section will therefore briefly review corporate made a report available in 1992 that offered guidelines to large companies as to how they
should conduct their affairs. At the core of the report was a Code of Best Practice (the Code)
governance and particularly its relation to risk management.
suggesting specific procedures for companies to follow. Although these procedures are not
mandatory, the London Stock Exchange requires every listed company to include a statement
in its annual report confirming that it is complying with the Code or giving reasons for non -
4.2 Corporate governance eencl ca
compliance.
Introduction
Corporate governance refers to the relationships among the management of a corporation,
its board, its shareholders and other relevant stakeholders and also to the specific The Code resulted in the publication of a similar document for South Africa by former Judge,
Mervyn King and Geoffrey Bowes' This report sets out the following 15 principles ofcorporate
responsibilities of boards of directors and management to ensure and maintain these
governance for the boards ofdirectors and people responsible for the direction of a business
relationships.
enterprise:
The board of a company and its management are firstly accountable to its shareholders
as the owners and the suppliers of risk capital. The overriding objective of any business is the
preservation and greatest practicable enhancement over time of the wealth of its
1: exercise leadership, enterprise, integrity and judgment in directing
shareholders. This means that managers and directors must focus on maximising the long-
so as to achieve continuing prosperity and to act in the best interest of the
term benefits to shareholders in terms of profits and cash flows and on minimising the risks.
business enterprise in a manner based on transparency, responsibility and fairness
Wealth creation for the shareholder means maintaining and expanding the net earnings
capacity of the company over the long term instead of trying to make short-term profits on Principle 2: ensure that through a managed and effective process board appointments are
risky projects that may jeopardise the continued existence of the company. made that provide a mix of proficient directors, each of whom is able to add value and to bring
independent judgment to bear on the decision-making process
The board and management, secondly, have a responsibility to other stakeholders. To
attain the objective of maximising wealth and to ensure the sustained prosperity of the
Principle 3: determine the corporation's purpose and values, determine the strategy to
business and its shareholders, management therefore has the responsibility to establish good achieve its purpose and to implement its values in order to ensure that it survives and thrives,
long-term relationships with other stakeholders such as employees, customers, suppliers and and ensure that procedures and practices are in place that protect the corporation's assets
the local community by also accommodating and promoting their interests. Future cash flows and reputation
can only be sustained and maximised if good relationships with these stakeholders are
Principle 4: monitor and evaluate the implementation of strategies, policies, performance criteria and
maintained. All boards have this responsibility, and their policies, structure, composition and
business plans
governing processes should reflect this.
RISK MANAGEMENT'. FUNDAMENTAL PRINCIPLES CORPORATE GOVERNANCE AND ENTERPRISE RISK MANAGEMENT
must take risk management Into conslderation, see Valsamakis (1999).
5: ensure that the corporation complies with all relevant laws, and codes of best
83
business practice
82
King Il
It was felt that the earlier codes, while sound in principle, lacked detail. Later codes contained
Principle 6: ensure that the corporation communicates with shareholders and other stakeholders
mere detail on how to implement these principles. The King Il report was published in March
effectively
2002. This report has the following six sections: board and directors
Principle 7: serve the legitimate interests ofthe shareholders of the corporation and account to • risk management internal audit
them fully • integrated sustainability reporting
• accounting and auditing
Principle 8: identify the corporation's internal and external stakeholders and agree on a policy, • compliance and enforcement.
or policies, determining how the corporation should relate to them
Unlike the King I report, which covered only one element ofrisk management, namely internal
Principle 9: ensure that no one person or a block of persons has unfettered power and that there control, the King Il report addresses risk management as a core element of corporate governance.
is an appropriate balance of power and authority on the board, which is, among other things,
The King Commission points out that corporate governance can, in part, be viewed as a
usually reflected by separating the roles of the chief executive cffcer (CEO) and chairperson,
company's strategic response to the need to assume prudent risks, appropriately mitigated, in
and by having a balance between executive and non-executive directors exchange for measurable rewards. The following are the most important recommendations in
the section on risk management:
Principle 10: regularly review processes and procedures to ensure the effectiveness of its
internal systems of control, so that its decision-making capability and the accuracy of its
• Ihe responsibility for risk management resides in the board of directors. Various people and
reporting and financial results are maintained at a high level at all times
functions within companies can assist, but the ultimate responsibility rests with the board.
Principle 11: regularly assess its performance and effectiveness as a whole, and that of the • Directors need to oversee the total process and at the end form their own opinion on its
individual directors, including the CEO effectiveness.
• The Code acknowledges the accountability of management towards the board for designing,
Principle 12: appoint the CEO and at least participate in the appointment of senior management; implementing and monitoring the process of risk management. In practice, the board, in liaison with
ensure the motivation and protection of intellectual capital intrinsic to the corporation; and senior management, will set risk strategy policies.
ensure that there is adequate training in the corporation for management and employees, and a
• The risk management process will be effective only if it is integrated with the day-to-day activities
succession plan for senior management
of the company, and if the risk strategy is incorporated into the language and culture of the company.
Companies should develop a system of risk management and internal control that builds more
Principle 13: ensure that all technology and systems used in the corporation are adequate to
robust business operations and delivers: a demonstrable system of risk identification a
properly run the business and for it to remain a meaningful competitor
commitment by management to the process a demonstrable system of risk mitigation activities
Principle 14: identify key risk areas and key performance indicators of the business enterprise a system of documented risk communications a system of documenting the cost of non-
and monitor these factore compliance and losses a documented system of internal control and risk management an
alignment of assurance of efforts to the risk profile a register of key risks that could affect
shareowner and relevant stakeholder interests.
Principle 15: ensure annually that the corporation will continue as a going concern for its next
fiscal year.
2 events can the share price of the firm; see. for enrnple. Pretty (19m It self-evident that governance
85 1. Risk
management is
Internal auditors need move 'from the back of boardroom' and need inseparable from
• Identifying key risks will entail a systematic, documented assessment of the processes and outcomes the company& strategic and business processes. Some risks must be taken in pursuing opportunity,
surrounding these risks and should address the company's exposure to at least the following:
but a company should mitigate its exposure to losses by responsible risk taking and welldefined risk
physical and operational risks
strategies. Risk management is intrusive and should not be viewed only as a reporting process to
• human resource risks technology risks business continuity and disaster recovery credit and
satisfy governance expectations and it is also essential in small companies with their traditionally
market risks compliance risks.
higher rate of business failure or missed opportunities.
• The Code points out that risk should not only be viewed from a negative perspective. The principle
2. The management should be responsible for the risk management process. Management is
of 'the higher the risk, the higher the return' will always apply. The 'risk appetite' of the company
will affect the management of risks within the company. accountable to the board for designing, implementing and monitoring the process ofrisk management
• Regular reporting from management to the board will assist the board in fulfilling its responsibilities and integrating into the day-to day activities of the company.
in terms of managing the risks. These reports should provide a balanced assessment of significant 3. All staff should practise risk management in their day-to-day activities.
risks and the effectiveness of the system of internal control. Significant control weaknesses, 4. The board should be responsible for the process of risk management. The development of the process
including the impact they have had, or may have had, as well as the actions taken to rectify the
of risk management and monitoring should be the responsibility of the board. Its members should
weaknesses, should be reported.
decide the company's appetite for risk and calculate its risk-bearing capacity, as well as the tolerance
• In the company's annual report, directors must at a minimum disclose that they are accountable for
the process of risk management, and that an effective process, which is regularly reviewed, has limits for key risks.
been instituted. They should provide specific assurances with regard to procedures that will allow 5. The board should approve the company's chosen risk philosophy.
the company to continue its critical business in the event of a disaster. The effectiveness of the
6. The board should adopt a documented risk management plan.
company's overall system of internal control needs to be disclosed.
7. The board may delegate the responsibility of risk management to a dedicated risk committee.
The compliance with corporate governance principles therefore places greater emphasis on an 8. Risk assessments should be performed on an ongoing basis.
integrated risk management function that covers all the risks in an organisation (i.e. ERM) The board should approve key risk indicators for each risk, as well as 9.
tolerance levels.
than ever before.
10. Risk identification should be directed in the context of the company's purpose. Risk
identification should not adopt a conceptual view or limit itself to a fixed list of risks, but
King Ill should adopt an all-embracing approach. It should not be limited to strategic risks.
A draft of the King Ill report was released in February 2009. At time of writing it was out for Operational risk management must form part of the risk management plan.
comment and will be introduced on I September 2009 and become effective from March 2010. Il. ne board should ensure that key risks are quantified and are responded to appropriately.
The main difference between King Il and King Ill is that King Ill applies to all entities 12- Internal audit should provide independent assurance on the risk management process.
and not business corporations only. It also moves away from the principle of 'comply' to
13. The board should report on the effectiveness of risk management.
'apply and explain'. For example, the board should be led by an independent non-executive
86
director. Where this is not the case, an explanation needs to be given.
to the to be strategically positioned and remain independent.
The report describes risk management 'as the practice of identifying and analysing the risks 14. rlhe board should ensure that the company's reputational risk is protected.
associated with the business and, where appropriate, taking adequate steps to 15. The board should determine the extent to which risks relating to sustainability are addressed
manage these risks'. and reported on.
The following principles are addressed in the chapter on risk 16. The board should ensure that information technology (IT) is aligned with business
management: objectives and sustainability.
17. The board should consider the risk ofthe unknown as part ofthe qualitative and quantitative Corporate governance codes have an expectation that an organisation-wide approach will be
risk assessment process. adapted to the management of risk. Because of the prevalence ofhigh-profile corporate failures,
18. Compliance should form part of the risk management process. investors and stakeholders do not want to be caught unawares. There is an expectation that
internal controls will be based on a thorough and properly structured process of risk
management.
4.3 Enterprise risk management 3 x$lntroduction Technology

New IT makes information quickly and readily available. It supports the globalisation drive by
providing instant information about new products, international suppliers and alternative
As stated in the previous section the King Il and King Ill reports require that risk be treated on
markets. Technology also allows growth
an integrated basis in the organisation. It also acknowledges that risk is inherent in any
opportunity. ERM is an approach to the management of risk that addresses risk in an integrated in the scale of operations and provides the means to manage operations at a distance.
fashion.
Intangible assets
The purpose of this section is to explain what ERM is and to show how it helps companies
to manage risk. It also sets out some ways in which ERM can be implemented. First, though, In the late 1970s, the book value of financial and physical assets on average equalled some 95%
we need to look at the trends that are changing the way businesses create value, forging the of market value. Today it is 20% or less. Ihe other 80% derives from intangible assets such as
compelling case for a more proactive and integrated approach to risk management. knowledge, brands, and relationships with employees, customers, suppliers, business partners
and investors. Increasing reliance on assets that are not easily locked up or insured has made
Until recently, risk management approaches were fragmented, addressing risk in a
business value more volatile than ever before and transformed the demands placed on risk
compartmentalised way. These approaches cannot deal with a company's continuallyevolving
management. A company's greatest assets can walk out the door if they are dissatisfied or can
risks and opportunities created by globalisation, advances in technology and a greater reliance
find a better deal elsewhere.
on intangibles assets such as the knowledge of its staff members. These changes create a lot of
uncertainty and plenty of new opportunities and require established businesses to adopt new
These trends force businesses to search for new ways to create value. Excellent products, effciency and
business models and adapt to changing risk profiles.
great service have become the norm and no longer
nie key drivers of these changes that require an integrated approach to the management
of risk are as follows. distinguish a company from its competitors. New sources of value have to be identified and
managed properly. Fundamentally, this means taking a new look at the assets or sources of
Globalisation value that make up a business. These are no longer just physical or financial. New asset
categories are needed to recognise the value inherent in the many intangible contributors to
Investors, customers, suppliers and even employees may disappear overnight to competitors
business, like knowledge, brands, people and relationships.
based anywhere on the planet. The erosion of national
In this new economy, level, every companies decision have a company to decide makes whether
involves to respond takingto
on risks. At a strategic changes in their environment by adapting or reinventing their business
to risk Enterprise management is preferlEd In this but terms will be

used
model. 88
87
boundaries and local market barriers means that businesses are more exposed to global There are also new risks in the assets and processes they need to execute their business model.
competition. But globalisation also means that businesses with a real competitive advantage can Put simply, risk and uncertainty are unavoidable when creating value in the new economy.
expand their markets internationally and exploit opportunities globally as never before in history. When less about the future is known and assets are harder to hold on to, more can potentially go
wrong. With risk playing such a central role, the imperative is to put risk management at the
Corporate governance centre of business strategy and planning.
What is enterprise risk management? Firstly, ERM is a process, which means that it is a set of continuous actions designed to
meet the risk management and corporate objectives. These actions are pervasive and
ERM is essentially a risk management perspective on management, as explained in chapter 1. It
inherent in the way management runs the business.
is a structured and systematic process that is interwoven into existing management
Secondly, ERM should be applied across the enterprise, at every level and unit. To
responsibilities. It provides a framework based on analysing risks and opportunities, with an
successfully do this, an organisation must take all its activities into account. ERM considers
ultimate objective of creating value for the shareholders. ERM entails the alignment of an
activities at all levels of the organisation, from strategic planning and resource allocation to
organisation's strategy, processes, people, technology and knowledge to meet its risk
marketing, human resources, production and credit review.
management purpose and offers a systematic and integrated way ofidentifying and responding
Thirdly, it is designed primarily to manage downside and exploit upside risks in direct
to all the sources of risk. It aims to provide a coherent framework to deal with all the risks that
relation to an organisation's strategy and risk appetite. ERM integrates risk management
result from operating in the new economy. It puts managers in the best position to see the risks
activities with strategic management and business planning processes so that the
they face, to assess their implications relative to the entire organisation and to plan an integrated
organisation:
response. Il-ne enterprisewide approach aims to provide one consistent framework for
responding to risk in an organisation. identifies the opportunities for creating value that present the most attractive risk/reward
trade-offs designs a business model that is responsive to these opportunities obtains a
Although many definitions of ERM exist, the following two capture its essential
holistic, enterprise-wide understanding of the risks inherent in the firm's assets, processes
characteristics. The first one is adapted from the one provided by the Casualty Actuarial Society
and the information used in decisionmaking
(CAS)4 and defines ERM as follows:
acquires the capabilities to manage the risk collects and analyses the data to produce
timely business risk information selects and implements the best strategy for exploiting
Definition desirable risks while concurrently mitigating undesirable risks supports units to achieve
Enterprise risk management is the process by which an organisation identifies, assesses, their goals in a controlled environment.
controls, exploits, finances and monitors risks from all sources for the purpose of increasing
Fourthly, it includes risks from all sources (financial, operational, strategic) and also the risks
the orgarusation's short- and long-term value to its shareholders.
created by 'natural hedges' and 'portfolio effects' from treating these in the collective.
'Ihe second definition is the one proposed by the Institute of Internal Auditors Research Fifthly, it involves the coordination of risk management activities. This spans:
Foundation. s risk assessment (including identification, analysis, measurement and prioritisation) risk
mitigation (including control processes)
90
4 The CAS presents a process for ERM, The primary purpose of the organ•sation's report is to consider the role of actuaries in risk but It gues a useful general ovenjievv
risk financing (including internal funding and external transfer such as insurance and
too. 11 includes a risk classification meant for general use. rather than limited to a specific industry. The report is available fl om the CAS wehslte at hedging) risk monitoring (including internal and external reporting and feedback into risk
http•//casact_org.%esearch/erm/.
5 Miccolis, Hively and Merkky {2001 : xxii). The Institute of Internal Auditors Research Foundation sponsored and supported this research.
assessment).
ERM differs considerably from 'traditional' risk management, which focused narrowly on the
89 scope of the risks, the treatment methods, and the impact and nature of risks. This approach
Definition was confined to pure risks or hazards and to property/liability risks, and the treatment methods
Enterprise risk management is a rigorous and coordinated approach to assessing and relied to a large extent on purchasing insurance. As a result, the traditional approach focused
responding to all risks that affect the achievement of an organisation's strategic and financial
primarily on the risks for which an insurance solution was available without regard for the
objectives. This includes both upside and downside risks.
materiality of the risk or its impact on the bottom line.
'Ihe main differences between the conventional approach to risk management and the
Most of the definitions reflect certain fundamental concepts:
enterprise-wide approach are summarised in table 4. L
Table 4.1 The differences between the conventional and enterprise-wide approaches to risk
management
Conventional approach Enterprise-wide approach

Integrated Source:
Fragmented
Deloach
Negative Positive
(2000: 16)
Reactive Proactive
Ad hoc Continuous
Cost based Value based
Narrowly focused Broadly focused 2. Shape risk
Functiona liy driven Process d riven
Approaches to implementing ERM 15

Two general approaches to the implementation of ERM have emerged.
(i) The measurement-driven approach: This approach focuses on identifying the key risk
factors facing an organisation and understanding their materiality and probability of
occurrence. Risk mitigation activities are focused on the most material risks, with
appropriate mitigation strategies.
91
Figure 4.1 The measurement-driven approach to ERM
Source: Adapted from Tillinghast-Towers Perrin (n.d.)
(ii) The process-control approach: This approach focuses on key business processes and
accompanying uncertainties that arise in the execution of the business plan. The emphasis
Figure 4.1 depicts this approach to ERM.
is on linking the process steps, reporting relationships, methodologies and data collection.
1- Assess risk Prioritise risk The goal is to manage risk events by achieving consistency of application across the
factors
business spectrum, thereby limiting the possibility of risk occurring. This approach to ERM
Identlty risk
is depicted in figure 4.2.
factors
92
15
MiccdiS. Hively and Merkley (2001.
Figure 4.2 The process-control approach to ERM The steps to implementing ERM
Business people Manage- Methodo-
Business
strategies and risk logies
The steps to implementing ERM are as follows.
System
Example of
sourcing risk
process-control approach -
(procurement) and data
1. Develop a common framework for risk
and policies management processes reports A common framework for risk enables managers to identify and prioritise all of the risks in th
organisation. It also helps to develop a common understanding of what risk is. commo
93 framework must be wide enough so that it captures all sources and classifications of risk. As a
example of such a wide framework, Deloach7 breaks risk into three broad categories, as follows
Data
Supplier Supplier Suppiier Supplier
strategy Supplier alliances status rating
Supplier certification Crosse reports system warehousing Procurement system
selection Integrated functional Commodity Transac-
policy contracting commodity teams benchmark tion
Ethical analysis
RISK MANAGEMENT'. FUNDAMENTAL PRINCIPLES CORPORATE
GOVERNANCE AND ENTERPRISE RISK MANAGEMENT
Environment risk arises when external forces affect a firm's performance, or make its
choices regarding its strategies, operations, customer and supplier relationships,
organisational structure, or financing obsolete or ineffective. These forces include the
actions of competitors and regulators, shifts in market prices, technological innovations,
changes in industry fundamentals, the availability of capital, or other factors outside the
companys direct ability to control.
Process risk arises when business processes do not achieve the objectives they were
designed to achieve in supporting the firm's business model. For example, characteristics
of poorly performing processes, or 'process risks', include:
poor alignment with enterprise-wide business objectives and strategies
ineffectiveness in satisfying customers inemcient operations diluting (instead of
creating or preserving) value failing to protect significant financial, physical,
customer, employee/ supplier knowledge, and information assets from unacceptable
losses, risk-taking misappropriation or misuse.
Information for decision-making risk arises when the information used to support business
decisions is incomplete, out of date, inaccurate, late or simply irrelevant to the decision-
making process.
These three components of business risk are interrelated. The environment risks and process
risks that the firm faces are driven by the external and internal realities of the business.
Information for decision-making risk is directly affected by the effectiveness and reliability of
information-processing systems and informal 'intelligence-gathering' processes for capturing
relevant data, converting that data to information, and providing that information to the
RISK MANAGEMENT: FUNDAMENTAL PRINC'PLES CORPORATE
GOVERNANCE AND ENTERPRISE RISK MANAGEMENT
standards
Source: Adapted from Deloach (2000)

Each approach incorporates the risk management steps outlined in chapter l, namely:
risk identification
• risk evaluation risk mitigation (which includes risk control, risk hedging and risk financing)
• risk monitoring.
In implementing ERM by adopting any one of the two approaches given above, the following
should be considered:
The range of organisation operations: The range and size of organisational activities may
prescribe that ERM be first piloted in one or a small number of business units or locations
with eventual roll-out to the entire operation. 'Ille sources of risk: Some organisations
confine the initial scope of ERM to a selected subset of their risks, e.g. property catastrophic
and currency risks. Eventually, all the risks are layered in, in sequence.
The types of risk management activities: Some organisations confine their initial vision to
the identification and evaluation (or assessment) of risk. Others begin by designing an
integrated programme around a subset of risk sources. Still others begin by measuring and
model)ing virtually all sources of risk, regardless of their priority.
CORPORATE GOVERNANCE ANO ENTERPRISE FISK MANAGEMENT
94 RISK MANAGEMENn FUNDAMENTAL PRINCIPLES
appropriate managers in the form of written reports and oral communications.

Process risk is sometimes virtually indistinguishable from information for
decision-making risk, because information is needed to make informed
decisions.
nie three components of business risk provide a broad foundation on
which more specific categories of risk can be identified and detailed. 'Ihey are
depicted in table 4.2.
Table 4.2 The business risk framework
Environment risk
Process risk
Sources of uncertainty
Information for
decision-making risk
Source: Deioach (2000)

95
MANAGEMENT. FUNDAMENTAL PRINCIPLES CORPORATE GOVERNANCE AND ENTERPRISE RISK MANAGEMENT
An example of an ERM-orientated vision statement is the following:

Business risk management is a continuous process, and an element of corporate governance. It promotes (ii) Define overall management goals and objectives
effcient and effective assessment of risk, increases risk awareness and improves the management of risk
th roughout the group. This includes anticipating and avoiding threats and losses as well as identifying
and realising opportu nities.
Affects the viability of the

business Affects the
execution of the business A broad statement of goals and objectives articulates the value proposition of risk
Affects the relevance and management, e.g. to establish sustainable competitive advantage optimise the cost of
reliability of information
managing business risk, make informed and conscious risk management choices on an
that supports value-creation
enterprise-wide basis, and improve overall business performance.
decisions
The above framework can be used to identify risks across the enterprise.
(iii)Develop strategies to achieve overall goals and objectives; align risk strategies and business strategies
A sound strategy for the organisation as a whole and its units and divisions provides a
2. Establish overall risk management goals
framework for accepting and rejecting risk as the firm seeks new sources of value in the
A broad statement of objectives articulates the value proposition of risk management, e.g. to marketplace. The risk strategy must be consistent with other company strategies, regardless
establish sustainable competitive advantage, optimise the cost of managing business risk, make of whether it is formulated separately or integrated with existing business strategies.
informed and conscious risk management choices on an enterprise basis, and improve overall Business and risk strategies should be developed concurrently — at least at a high
business performance. level — so they are in sync with one another. They can then be defined in greater detail
and subsequently realigned.
3. Develop risk management strategies When aligning risk management strategy with business strategy, there are many issues
The development of risk management strategies consists of the following five steps: to consider, for example:
(i) Articulate the risk management vision Is risk management a continuous process or a periodic activity?
The vision provides a sense of purpose and focuses subsequent development Are certain risks more vital to our value creation strategies than others?
of more specific risk management goals and strategies.
Are certain risks more common enterprise-wide than others?
96
Do our primary risk-taking activities capitalise on our core competencies?
How strong are our risk management capabilities relative to those of our competitors,
(iv) Develop a formal risk management policy
particularly for mission-critical risks?
The risk management policy assigns responsibility for performing key tasks, establishes Who is responsible for developing risk strategies and assigning risk authorities? Who
accountability with the appropriate managers, defines boundaries and limits, and formalises is responsible for executing those strategies and authorities?
reporting channels. These elements provide the baseline for a sound policy statement.
While the actual format and details of a baseline policy structure vary from one firm to
another, the following issues should be explicitly addressed:
the objectives of assessing, managing and monitoring business risk the responsibilities
of risk owners and risk oversight personnel the roles and responsibilities of operating
managers in managing business risk at the business unit and divisional levels the
overall enterprise risk tolerances linked to established business objectives and strategies
(e.g. if management's goals are to earn R2,50 per share and retire R50 million in debt
during the coming year, how much exposure to earnings and cash flow variability can
the business withstand?) the boundaries and limit structures linked to enterprise goals
that clearly specify management's 'risk tolerance' (i.e. how much risk is the firm willing
to accept?). This includes prescribed exposure limits for authorised business activities
that are potentially risky if not well managed.
the risk authorities, i.e. who is authorised to commit the firm's resources in conjunction (v) Obtain approvals, communicate widely and evaluate periodically The risk
with volatile and high-risk activities and execute specific business risk management management vision, goals and objectives, strategy and policy, and any changes to them are
strategies? (a) approved by the board of directors; (b) implemented under the direction and
the required risk-reporting and approved methodologies for measuring risk. coordination of a senior executive and/or a chartered risk management executive committee
reporting directly to the CEO; and (c) widely communicated throughout the firm.
Once the specific strategies have been selected, the policy structure is expanded to address: Executive management should periodically evaluate the firm's risk management
a description of the business risks that management has determined are 'mission-critical' goals and objectives in cooperation with the board* key executive and operating
strategies for managing different types of risks, including acceptable or preferred risk managers, and process/activity and risk owners. A feedback and re-evaluation process
management techniques and prescribed tools, products and practices the value of can be executed through facilitated meetings, task forces and focus groups, and through
evaluating business risk according to uniform processes to enable enterprise-wide the ongoing activities of existing executive and risk management committees. Overall,
aggregation, analysis and knowledge sharing. management should do the following:
design and implement risk management capabilities to execute the strategies build on
The policy structure:
this foundation to continuously improve strategies, processes and measures for
summarises the overall process for building and improving risk management
individual risk
capabilities to be applied by the enterprise and each of its business units
• aggregate multiple risk measures link aggregate measures to enterprise
97
performance build on the above capabilities to formulate an ERM
strategy.
institutionalises the measures of success by which risk management will be evaluated
defines the value proposition of risk management in the organisation.
Principles of enterprise-wide risk management
All told, the policy in an ERM environment is a broad one that is consistent with the organisation's
The Institute of Risk Management SA identified the following principles applicable
structure and culture.
to the implementation of ERM. These principles follow directly from the King Il
Policies should be well documented to avoid ambiguity. Manuals should be as brief as report:
possible and 'user-friendly: The nature and level of detail of the policy documentation is a
function of management's operating (i) Establish formal framework board of accountability ERM for the
organisation.for risk management.
philosophy, the complexity of the business and the nature of its risks. (ii) Develop a
(iii)Establish organisational structures for ERWI.

(iv) Establish a structured process of risk assessment.
Develop a risk-based control environment.

98 RISK MANAGEMENT: FUNDAMENTAL PRINCIPLES
The management of insurable risk in a corporate context
(vi) Establish a system of risk monitoring. The scope ofthe insurable risk silo in an organisation
(vii) Establish a process of risk reporting. The job title crisk manager' is now accepted by large corporations.8 It is not uncommon, however,
(viü) Embed the process of ERM into the organisation. to find risk managers carrying additional titles. The management of operational risks is also
assigned to executives under whose control and responsibility falls the corporate treasury function,
(ix) Establish assurance processes for key risks and for the risk management process.
where a risk manager should be reporting. With the formalisation of risk management as a
(x) Incorporate the risk-related aspects of integrated sustainability reporting into the ERM
management function, there is a proliferation of specialist consultancy firms providing a service
framework.
to organisations in return for a fee.9
When discussing the functions of the risk manager, there is a debate
4.4 Organisational considerations
Risk silos
It is now accepted that the board of directors is responsible for ER.Ä(. The problem that now
arises is how to implement an ERM programme. For example, international companies face risk
over the types of risks to be managed. Should the risk manager's function be confined to
from every part of the globe. There is no single answer as to how the board responsibility is
insurable risks only, or should it extend to all types ofrisks? It has been suggested that the
translated into a practical programme covering every part of the enterprise. In practice, different
boundaries of risk management should be extended. 10 It still remains unclear as to how such a
techniques are used, such as the 'top ten' technique whereby each unit in the enterprise lists and
progressive course for risk management will be followed. It is generally believed that the
ranks the risks facing that unit. These risks are then passed upwards in such a way that all the
management of all operational risks should be included in the ambit of the risk manager and that
risks facing the enterprise are then examined and re-ranked until the top ten risks are identified.
risk management should be regarded as a specialist function, as should the treatment of
At the corporate level, management can then concentrate on managing these top ten risks. This
speculative risk within the organisation. li In efficiently carrying out such a function, a wider
technique may not cover future risks, and the board must still deal with these future or strategic
definition of risk may well have to be adopted, but the starting point must be the implementation
risks.
of a risk management process along the lines described by the model.
Although senior management will concentrate on the 'top ten' risks, this does not mean
There is, of course, no single method of applying risk management within an organisation,
that the other risks that have been identified should not be managed. The risks identified in
as each is different, and thus the method of implementation must suit the needs of the particular
each unit must receive due attention by that unit. In fact, it can be argued that these units are
entity. Nevertheless, the purpose or object of the risk management department simply stated
best placed to deal with these risks. Each unit will have the specialised knowledge of the risks
could be to ensure that the functions of risk management are implemented and that the risks
concerned and the technical skills needed to deal with them. For example, if the electrical
facing the company are being managed.
engineers have identified that their factory is exposed to risk ofloss ofelectricity due to a
transformer failure, it is only electrical engineers who will understand techniques such as
'circulating current protection' and 'Bucholtz gas-operated relays' that could deal with these
a
extent. the æsponsibility for managing persona' risk has shifted from the individual or famuy to the organisation. such a oe evidenced by the development of employee
risks. Clearly, each unit has specialised risk problems and specialised solutions to those
benefit plans (retirement me-dial did, group and personal The such a shih of is further supported by the success and growth of management consultancy firms
problems. It would be incorrect to believe that a generalist has specialised knowledge of offering
everything. This leads to the conclusion that there is a place for the so-called 'silo approach' to becoming widely used by south African the relevant in the US, Serbe;n can
The extent to Which the risk management practice is formally corporate management act,vity in south Africa can be from the membership south African Risk and
risk management, despite the criticism of this approach. Insurance People elected as representatives
One ofthese areas that lend themselves to management at a corporate level is the be heads of RI or risk management departments have least to five become in insurance associationor risk activity. Organ.satms that spend million or more
on insurance annually ace eligible ottitles Of managers see Parrott (1969). Kiornan (1987: Robinson and Kunath : 42).
management of insurable (event) risks. This area can in and of itself be regarded as a silo
requiring specialised knowledge to manage the risk. The
99
functional operation of this area will now be examined in greater detail, since these details are known
from cost-of-risk surveys.

100 101
Within the framework of what has been deduced in the preceding chapter, The risk control responsibility the organisation must ensure that the risk management model is effectively There is a significant
organisational difference between the financing and implemented, namely: risk control responsibilities entailed in risk management. Whereas even in a that steps are taken to control risk physically multifaceted
organisation a centralised risk finance department can ensure that that adequate provision, where appropriate, is made to finance losses that appropriate risk-financing programmes are implemented, this is not
possible arise. with risk control programmes. These programmes, which include, for instance, detailed risk inspections, demand that far more personnel are involved, so In order to achieve the risk management
objectives, organisations deploy day-to-day risk control remains a line management responsibility. The risk the requisite resources and establish risk management departments for this management department
could be used in a consultative role (fulfilling a staff purpose. The complexity of the risk management departments varies widely, function) to assist and give guidance to line management.
depending on the organisation's size and the scope of its activities, but these There are various methods of implementing risk control programmes, departments are mostly divided into risk finance and control
divisions.l depending on the type of organisation. Both the fundamentals and the level While there is an obvious interrelationship between the activities 13 of these two of sophistication will differ when one
compares risk control programmes divisions, there is also an apparent difference in emphasis. implemented for a multiple site company to companies that have, for example, one or two major sites (with
concomitantly higher relative asset values). The The risk-financing responsibility latter may employ a highly qualified risk control staffng complement of
Traditionally, the purchase of insurance cover and the implementation of other engineers, safety officers and even medical practitioners who direct attention risk-financing strategies have been regarded as the main
responsibilities of the to the psychological or sociological aspects of risk control.
risk manager. This originated primarily from the interpretation of the function of risk management as serving to protect the organisation from the negative External consultants financial consequences of
operational risks. The risk management department is also charged with the responsibility Tre activity entails the structuring, implementing and administration of procuring the expertise and services of
14
external risk management of the appropriate risk-financing programme, including insurance, for the consultants. The extent of the use of external specialist consultancies is organisation. Thus, among
other things, the risk manager needs to be familiar evidenced by examining the number of employees typically employed in US with the scope of the cover afforded by each respective policy covering the
risk management departments. Usually risk management departments tend major perils as identified through the macro identification process. to be relatively small in size, and they contract with external
consultants to The responsibility of risk financing in the organisation is far reaching. In provide more specialised services, especially risk control services. This is not major organisations, these financing
programmes can be complex and hence surprising when one considers the multitude and diversity of risks facing an this function is specialised. It is not uncommon for the risk manager to devote organisation.
most of his/her time and energy to the risk-financing responsibility. Risk financing is the line responsibility of the risk manager.
While there is obvious need to examine the two elements separately, the The reporting function of insurable risk managers in model developed in this study stresses the integration ofthese two
programmes organisations so as to achieve an encompassing risk management rationale or holistic The development of a formal programme of risk management in an approach. organisation is of relatively
recent origin. In the US, it was only during the late 1920s and early 1930s that industry began to formalise the responsibility for treating operational risks. 1S Today, risk management departments are
the common, understanding particularly of their in the position US, 16 and and, a indeed, survey of their these limitations.proves interesting in
14Grene and Serbeir, (1983:3-9), Head 50); Richards (1 985; S4-6)15 et el.
(1974 72)12 Robinson and 1931,-42)1 3 Greene and Serbein {1933: 80}
CORPORATE GOVERNANCE AND ENTERPRISE RISK MANAGEMENT
103
Table 4.4 Risk manager responsibilities
Responsibility level
General Shared Reco- Not
Functional area authority authority
mmend respononly
Sible
Property and liability insurance purchase 82,3

10,8 Liabi\ity claims
63,0
management 27,2
Worker' compensation insurance purchase 70,7 12,3 7,7

RISKMANAGEMENT. FUNDAMENTAL PRINCIPLES
102 RISK
Table 4.3 Risk manager reporting relationship
Function to which risk manager reports
Finance
Treasury
CEO/presjdent
Administration
Legal
Other
Human resources
Source: RIMS, Cost-of-Risk Survey, various
The various cost-of-risk surveys examine a number

the risk manager relative to the organisation. These
in tables 4.3 and 4.4. Table 4.3 specifically shows
relationships, while table 4.4 indicates the
responsibilities within an organisation.
It is evident that the majority of risk managers
r
ofthe organisation. nlis is in line with prior investigations
18
and Cerveny, who found that approximately
to the head of finance, although they had access
This is not surprising, since the major functional
corporate risk manager is risk financing. Naturally,
the operations and general survival of the business,
—
insurance cover especially to protect against
be underestimated. However, this narrow functional
insurance matters does encourage the risk management
biased towards a cost-to-benefit and cost-to-risk
full contribution that the discipline can make to
organisation.
Workers' compensation claim management 46,3 26,3 20,3
17 (1981:12).
Property loss prevention 43,6 36,6
Employee/public safety 23,8 37,7 227 156
Environmental affairs 11,3 35,3 26,7 26,7
Employee benefits — welfare plans 12,3 18,6 10,1 59,0 Pension/profit-
sharing
deferred compensation 72 13,1 72,0 plans
Security 18,1 24,2 49,9
Selection of brokers/ agents 78,2 13,5
Source: RIMS (T 997)
Part I l
RISK MANAGEMENT. FUNDAMENTAL PRINCIPLES
Managing risk
This part of the text focuses on the management of risk. As explained in chapter 2 (part I), this
entails identifying the risk, determining its impact on the organisation and controlling it. The
management of business risks (both inherent and incidental business risks) and event risks is
discussed, and, where necessary, the text draws a distinction between each type of risk.
RISK MANAGEMENT. FUNDAMENTAL PRINCIPLES
RISK IDENTIFICATION 107
106 Risk evaluation
Chapter 5
Risk control Risk financing
Risk identification Risk management is an ongoing process, not a single event. The model in figure 5.1 dictates that the
process starts with risk identification and evaluation as a prelude to risk control (for both business and
event risks), and finally provides financially for the consequences of event risk.
Although identification is considered the most important step in the risk management process, it has
generally been neglected in the relevant literature. A number of factors may have contributed to this. One
factor is the preoccupation of risk managers with risk financing and, in particular, insurance. This results in
risk identification and evaluation being viewed in the context of insurance provision and coverage limits,
which has an emphasis on macro-risk identification. The evaluation of risk is thus dealt with through the
description of risk exposures made known to the insurance market by insurance underwriting surveys, the
organisations' risk managers and their insurance intermediaries.
Introduction
The second factor, possibly the more significant, is that the techniques and general methodology
Risk identification
underlying risk identification can be diverse and complex, requiring specialised skills related to specific
Summary
disciplines, particularly those of engineering. It is not unexpected, therefore, that particular techniques are
developed by specialist consultants in, for example, the chemical engineering profession, from whom risk
management departments obtain advice. aspect can also be said to apply to the question of risk evaluation,
5.1 Introduction which may require specialised input by consultants with the necessary mathematical/ statistical or
actuarial background.
In part I, the fundamental principles of risk management were addressed, A third factor is that, in practice, an overlap exists between risk identification/evaluation and risk
providing an underlying framework for managing risks. The model suggests a control.] Risk identification and evaluation deserve separate study because they are diverse and complex
process that is indicated in figure 5.1. subjects in their own right. The intention in this chapter is not to provide a comprehensive study of risk
identification and evaluation methods, but rather to provide a more allencompassing insight so as to
Figure 5.1 The risk management process achieve a holistic approach to risk management that practitioners can consider.
Risk Identification
5.2 Risk identification
Every risk management programme must necessarily be put in motion by the process of risk identification, because,
obviously, a risk cannot be managed
A
examples may Illustrate the point eg. a risk manager nauces welding takirya place in a warehouse containing flammabk goods immediatety Identify this as a maJCK hazard In mak ing this assessment he realises
that a fire will ca use substantlar inancial ross Par t of the is to estimate the possible mjgnitu& of loss. Similarly. a "Sk manager for a chain saye may be concerned about damage taused motivated people.
One possibi chat she will Investigate isatson, and she wilt design a rlsk control pograrnme for this possibility
A possibillty is a mortar attack. intuitive" however. without assessment. (he risk manager may regard this possibillty as being so remote
•equire any further action. Her assessment of the probability of a mor tar attack would cause her to abandon any further Identification ot control procedures
RISK MANAGEMENT: MANAGING RISKS
108
if it is not first identified. Hence, risk identification must be viewed as the single most
important function of the risk management 2 process and should be approached in as
systematic a way as possible.
A systematised approach to risk identification may be achieved by first considering what
risks the organisation faces on the macro and micro levels.
Definition
Macro-risk identification is the identification of major risks that may have a significant impact
(financial and otherwise) on the organisation.
Micro-risk identification is the Identification of sub-risks within the major risk class and is key
to the risk control objectives.
Macro-risk identification
This concerns the identification of the major sources and types of risks facing the organisation.
For business, this involves an analysis, at the macro level, of external environmental factors by
using techniques such as industry analysis, competitor analysis, market analysis and country
analysis.
RISK MANAGEMENT. MANAGING RISKS
with the assistance ofa reputable insurance intermediary, can greatly assist the
Insurance classifications of risks or perils and their consequences are therefore useful. Table
5.1 gives a typical classification list.3
Table 5.7 Usual insurance covers and related perils
Insurance classifications Perils
Assets, all risks (perils) Damage to assets (unless excluded) from any perils (unless
excluded)
Fire, explosion Loss or damage to property due to fire or explosion Natural perils and others
Earthquake, storms, floods, hail. snow
Crime Theft, burglary, fraud, fidelity, arson, sabotage, terrorism and others
Engineering Machinery breakdown; loss ofor damage to machinery and

utilities
Transit Loss of or damage to goods in transit
Motor Loss of or damage to the vehicle, loss or damage to third parties, injuries to occupants
Legal liability Environmental impairment, products liability, employers' liability, death or injury to
persons
Consequential losses Business interruption losses due to the loss of assets due to Insured
perils
Insurance reviews may cover uninsured perils, as well as perils for which no insurance cover
is available, or is available only at extremely high premiums due to a prevailing restricted
insurance market. Uninsured risks are usually identified and discussed during the insurance
review.
Organisational charts and flow charts

Organisational charts assist in gaining a macro view or overview of the organisation. To the
list of perils identified by the typical insurance review, the risk manager can add other risks
facing the organisation. Such identification can be achieved by drawing an organisational chart
that illustrates the different aspects of the organisation's activities and structure. For example,
the chart
of nature offer limited benefit those risk who bear the of loss may not be by means of that assists In evaluating
risks and the adoption of preventativemeasures. SeeWi11iams and Hans (1981 : 45-9).
Jn the case of event risks, the principal concern following macro identification is to
insulate the firm as far as possible from the major negative financial consequences of loss
arising from the major risk sources. In most instances, financial provision for such
consequences can be made by the purchase of insurance cover. Certainly, in the extreme, the
insurance market represents an effcient financing mechanism, facilitated by the pooling effect
that is evident where such risks are concerned. Hence, macro-risk identification and
classification should logically follow accepted insurance classifications. These classifications
are thus useful for macro-risk identification purposes.
'Ihe process also involves identifying the impact of environmental (or dynamic) changes
that can affect event risks. These changes are outside the control of the company, but impact
upon it.
Insurance reviews
Generally, macro-risk identification is conducted by carrying out an insurance review. The
insurance industry has for centuries been a mechanism for handling pure risks and has thus
developed comprehensive risk reviews that,
2 0'Connell (1976; 34).
110
shown in figure 5.2 may indicate the main aspects of the organisation's activities under such
headings as contracts, retail and specialised subsidiaries.
Figure 5.2 depicts the structure of the organisation as a whole. It is also possible to
construct a chart that could, for example, show the management and administration set-up,
as shown in figure 5.3. Organisational charts are important in identifying consequential losses.
Figure 5.2 Organisational chart: ABC Manufacturing Company
ABC Manufacturing
RISK MANAGEMENT. MANAGING RISKS RISK IDENTIFICATION
Figure 5.3 Organisational chart: Management and 111
Manufacturing Company
administration of the ABC

Managing director
Research/ Finance
Marketing develop-
ment
Purchasing Production
Accounting
Genera I General manager manager

contracts DEF Trading division
Company
General General manager
manager ABC ABC
Civil Eiectronics
Engineering
Research/
development
Works
Marketing &
the organisational of
Production
sales
the company's to identify
There are numerous ways in which but the
concentrations' particularly to
key point is that all areas identify contrast to structure
The organisational chart attempts may be drawn, be
duplications, dependencies and risk Flow chart shown. to
charts are important risks (consequential activity should highlight
losses). In are not restricted to the 'areas' of risk and
organisational interruption
and evaluate flow charts
organisational charts, of can include
the firm and
Company
Own retail Subsidiaries

Contracts
ABC Civil ABC

Motor Contracts Aviation Engineering Electronics
Own Customer premises
premises
DEF Trading Company

112 A flow chart such as this therefore facilitates, for example, the identification and
interdependency of the various processes and stages of production, and it is useful to the risk
manager in identifying risks and interpreting potential risk exposures. Although this will, in
suppliers. These charts can be used to describe any form of 'flow' within the firm, e.g. the
all probability, introduce an element of business risk to the risks identified, for which no cover
production flow, by which raw materials are converted into the finished product. A simplified
is available or sought, it does provide a thorough analysis and should produce a systematic
flow chart is shown in figure 5.4, which identifies broad areas of risk. With this as a base, a
strategy. Among other things, this process can accomplish the following: the identification of
more detailed chart can be drawn up that considers each stage of the production process and
risks that are extraordinary to the organisation the identification of a broad set of risks, e.g.
describes likely lossproducing events causes and consequences.
general liability, and the focus on a sub-set for which cover may be sought, e.g. public liability
the opportunity to discover any interrelationship between event (pure) and business risk where
Figure 5.4 A production flow chart it occurs.
RISK IDENTIFICATION 1B
Analysis of financial statements

Another approach that has been applied more recently (and which is relevant event and business
risk interrelationships) is to identify risk by analysing a firm's financial statements? All financial
losses have an adverse effect on either balance sheet or income (profit and loss) statement, and
thus an analysis of organisation's financial statements can prove most useful as a starting point
risk identification.
Values that are at risk can, for example, be established by examining the balance sheet, after
which it can be established which perils could occur may destroy this value. The income statement
may show rental expense payments that may indicate the possibility ofa contractual liability
Moulding plant exposure; or indicate a hidden asset if, for instance, long-term rental obligations under current
lease agreement are markedly lower than current market rates. The income statement will also
disclose sources of income and thereby processes which losses may arise, such as liability to third
parties in the event of malfunctioning products.
Thus, a flow-chart method of identification coupled with an analysis of organisation's
financial statements provide a systematic procedure for reviewing, in proper order, the company's
operations. However, an aspect that should always be borne in mind is that a company is dynamic
and constant changes give rise to new and added risks.
Finished goods store
Finally, it must be stressed that the risk management process should follow continuous 'lhe effectiveness of any inspection is related to the competence and expertise of the person
strategy, whereby the risk is identified, evaluated, physically controlled where possible, and then carrying out the survey. The expertise that is needed, especially for multifaceted operations,
provided for from a financial standpoint. order to achieve this, it is necessary to structure the is extensive and could cover many disciplines, e.g. mechanical, electrical and chemical
various processes in complementary way. If a list of the various perils that pose a threat to the engineering, fire prevention and security.9
organisation were compiled without taking into account the way in which risk is to be financed • Inspections follow the law of diminishing returns. If the same person is sent to the same site a
or without reference to the existing insurance cover, risk identification process will be of little second time, he/she is unlikely to identify much more in addition to that which was detected
use in considering additional insurance cover. in the first visit.
Micro-risk identification
the term implies, micro-risk identification refers to the process ofidentifying micro risks that
can usually be prevented by introducing effective risk measures. 6
8 'Inspectionsr have Greene been recognised and Serbein by 0983; rnosr 70) and In Collecta the held end of Mayer tiSk management (1984.28}. 35 fom•nng an Integ ral partof the risk
managernent process
See, for example,
be argued that this approach lepresents a micro-risk Identification tecbnlque. Without wishing to refute this poinrot view. the alm S5ing accounting statement 9 As mentioned Sef«e, companies engage the sewites of professiona' risk managemem organi%tions who empby experts in these By virtue of thelr specialisati0ft, they are in a position
analysis as part of the process is to allow a more easy of the discussion so far and to indicate that che anatßls of accounrlng statements is only one wav among to draw comparsans between one company and another and thereby achieve compliance standards. It should also be noted ttEt For those compantes whose land-based asset values
several other- of identifying macro-risk facing an organisation. require world-whde insurance covert often direr-I insurance and veine.ugance ma frets WIB require surveys to be conducted by their in-house rlsk surveyors.
or the instal the lation danger water of a n inoperative gauges.water supply is Identified. then risk control steps can be implemented such as procuring a permit
114
Inspections usually focus on physical observations, whereas often the problems are
systems related. For example, an inspection will reveal neither poor design controls nor
Risk managers use several methods to identify the various risks. Dickson 7 points out that unsatisfactory product markings that could cause product liability claims. Techniques
identification methods have their origins in solving problems within an organisation or even an such as hazard and operability studies (HAZOP), failure mode and effect analysis
industry, and that the identification process requires creativity. The task of matching the method (FMEA)Y fault-tree analysis (FTA), hazard indices and safety audits have been
to the risks is important, although it must be recognised that it is not possible to lay down firm developed to overcome this drawback. These are now discussed.
guidelines as to how this is to be achieved, as the circumstances are different within each individual
firm. Nevertheless, having classified these broadly under macroand micro-identification methods, Hazard and operability studies
the following micr&identification methods can be considered. A HAZOP study has been described by Ozog10 as 'a systematic technique for identifying
hazards or operability problems throughout an entire facility'. The study is a qualitative
Risk inspections approach to risk identification that can be employed at the planning stages of projects. These
The obvious and most commonly used method to identify risks is to conduct a physical inspection. studies are used most often in the chemical industry.
8 The main advantage of this method is that the sites/plants are seen at first hand and, importantly, In essence, the HAZOP study is an important enquiry into the operations of a plant in
such a visit brings one into contact with those people whom one relies on for much of the terms of the hazards that it is to. The study follows the basic principle that extremely complex
information concerning risks and hazards on the shop floor. problems should be broken down into manageable parts, which are then examined carefully
This method has a number of drawbacks, however. so as to identify all associated hazards. Dicksonll states that the HAZOP study is concerned
Apart from the distance that may have to be travelled to reach the site, it is a time-consuming with four main questions:
exercise. Before carrying out an inspection, therefore, it is essential that as much preparatory (i) What is the part that is being examined intended for?
work as possible is done. This preparatory work relates primarily to pre-planning or (ii) What are the deviations from the declared intention?
programming to ensure that the visit will coincide with other tasks to be performed. A logical (iii)What are the causes of the deviations?
approach to risk identification incorporating, for instance, the design and use of a report that
(iv) What could be the consequences of the deviations?
is to be completed while the site is being inspected will minimise the chance of overlooking
important aspects.
Put simply, each part of a process is studied to ascertain its (intentiont Lists are then Loss of life and damage to property would be incurred.
compiled showing all possible deviations from the normal operating conditions, as well as
how these deviations might arise and what their consequences might be. Failure mode and effect analysis
HAZOP studies have become popular in practice, especially in the US, where it is
Ozog13 describes the primary function of an FMEA study as being 'to evaluate the frequency
estimated that at least half of the chemical industry has used the technique for new facilities
and consequences of component failures on the process'. This method, while identifying
or processes. With existing facilities, a typical interval between reviews is from one to five
risk sources, concentrates more on the effect of failure with the objective of minimising its
years. nese studies have been found to be effective in micro-risk identification. 12 14 The
consequences. FMEA is therefore usually applied to specific equipment. analysis is
thus primarily concerned with the evaluation of consequences and the frequency of failures.
The aim is to attain a fail-safe failure mode.
14 FMEA can be applied. forinstance, when considering the Installation overhead conducors, Should anyotthese detach andfall the they pose a serious to (he
overcome this one might position catch the overhead in an anempt to achi2•. re a fail-safeWIure
161). referring to Lövvley (?974).
116
'lhe major disadvantage of this technique is that the analysis does not extend to operational
procedures. Because FMEA concentrates on component failures and not on errors in operating
The main advantage of the HAZOP study is that each part of a complicated system is
and other procedures, it has a limited application. The method is also tedious to apply to complex
examined in detail. Furthermore, in view of the many different skills required to conduct the
installations.
study correctly, it is necessary to involve a team of people, and this in itself is advantageous.
Such a team approach encourages communication and is synergetic, which is a proven
advantage in any risk management department. Fault-tree analysis
The main disadvantage is the time involved in carrying out the HAZOP study. It calls FTA is not a means of identifying risk per se, but primarily a technique for analysing risk. It is
for a considerable investment of time and resources. Probably the most famous example of consequently used only-after some other type of technique or risk inspection has been carried
where a HAZOP study could have been used effectively is the Flixborough disaster, which out and a major risk or hazard has been identified.
occurred on I June 1974. Some pipework in a chemical plant was undergoing alterations The fault tree is a diagrammatical representation ofall events that may give rise to some
when its supports gave way, causing the pipework to fail and release 36 tons of cyclohexane major event. FTA concentrates on ascertaining all the possible combinations of individual
into the atmosphere. The resulting vapour cloud ignited and an explosion occurred, killing failures that can lead to what is termed the 'top event' or 'ultimate' risk. The diagram is made up
28 people and causing millions of rands worth of damage. The possible consequences of of a series of AND or OR gates. The final failure rate, frequency or 'probability' is then
these alterations could have been determined beforehand and the appropriate risk control ascertained by the application of probability theory.
steps taken to avoid the disaster. For example, if the risk of a major toxic release (seen as the top event) has been identified
The HAZOP study is applied to each item of equipment and is carried out by addressing by some other means, FTA can be applied to analyse the factors or combinations that could
standard questions or considerations to that equipment. considerations include:
result in that top event occurring.
What deviations could arise?
These combinations are shown in a sequential arrangement on a fault tree that indicates
How can these arise?
the frequency or 'probability' of failure of each event. By estimating the individual failure
What are the implications?
probabilities and then applying the appropriate probability theory to the expressions, the
Are there any further implications?
probability or frequency of the top event can be determined. Sensitivity analysis can then be
performed by altering the various constituent factors to gauge the effect of alterations to the system,
Applied to the Flixborough situation, the answers would be: or to bring the top event frequency to a desired design level
The pipe could break if the supports were not adequate. Cyclohexane would
Figure 5.5 shows a fault tree of a simple event: a tank exploding. The top event is an explosion
escape.
with a design frequency of explosion of 0,0002 per year (or I explosion every 5 000 years). 16 The
A massive explosion would occur. top event is contingent on both pressure rising and the relief valve failing to operate, which is an
AND gate. Pressure will rise if either the pump fails or there is an excessive input of the raw one major problem (not apparent in the other methods described), i.e. the derivation of
material, which is an OR gate. The relative frequency or probability of fåilure of each individual probabilities, particularly where these probabilities are very low. Should the relevant
event is also shown, which collectively ascertains the probability of the top event occurring. probabilities not be accurate, then the resulting calculation, particularly the measurement of
The probability of the pressure risk is + = 2. The probability of the tank exploding is 2 x I x 10- the likelihood of the top event, will be suspect. The method is also subject to the criticism
4 = which equals the design frequency. that the company defines an acceptable level for a failures which often includes the death of
people.
*.1985: 161 J. The view taken in this study — and the one adopted by Dickson (1 98?: 27) - is that the debate on whether identlfication of from analysis of risk is
irrelevant. Hence, FTA. which is a risk-anahßß technique, Is Included in the fisk-ldentihtation methods. Th.; • %motes the integration ofthe activities Of identification and
as opposed to presenting them as separate and posslbly unrelated. seem to be very low. but In fact it is not. Il means 1 tank/5 000 tanks wilt explode every
year. Since trære are hundreds of thousands c:' in the world, many tanks a year Will explode. Hazard indices
118
Dickson]$ describes the hazard index as a technique that attempts to express the degree of hazard
by using a number. In doing so, indices can be used to identify process areas with significant loss
If the calculated top event frequency is two low, then additional steps would be taken and the potential. The process involves the measurement of the likelihood of loss and the expression of the
new top event frequency determined. The procedure will continue until the top event result as a number to which others can be compared and annual changes monitored.
frequency is brought within the design frequency.
The most common method of expressing the degree of hazard is the Dow Fire and Explosion
Figure 5.5 Fault tree illustrating individual failures leading to a top event occurring Index. This is used primarily for insurance or underwriting purposes and has been expanded to
include business interruption losses. While various attempts have been made to further expand and
Top event improve it, they have met with limited success. 19
TANK EXPLODES
00002/y
Safety audits
A common practice used to identify safety risks is the general safety audit. In the course of the
AND
year, most industrial plants are inspected and audited by a number of people. In South Africa, it is
mandatory for the safety representative of a company to inspect the workplace.
RELIEF An extremely important aspect in the field of risk management is personnel safety
PRESSURE VALVE FAILS
RISES 104/y (particularly in the US and UK). Various systems to ensure personnel safety have been devised,
2/y including the use of specially designed safety audit sheets used to facilitate risk identification. Il-
OR
me safety audit sheets usually include some statutory requirements and defined objectives. One of
the bestknown audit systems was that developed and implemented by the National Occupational
EXCESSIVE
Safety Association as a management by objectives system.
PUMP
INPUT
FAILS
or5/y
Legislation and codes ofpractice
The implementation of established legislation and codes of practice achieves both compliance and
risk control. What is not obvious is that it also serves as a means of identifying risks.
In the manufacturing process, for example, risk identification techniques may not identify all
Source: Dickson (1987:72)
the sources that could give rise to a product liability claim. Often the risk source is only identified
-me main advantages of FTA stem from the fact that this method is a structured approach to when the defective product is returned and a claim arising out of the defective product is made.
risk identification. It simplifies analysis of complex systems and traces causes and impacts. Systematic compliance with legislative and other authoritative codes of practice dealing with
Apart from the relatively long time it takes to correctly carry out the study, it suffers from quality assurance will clearly mitigate the risk of defective products.
IAkkson 0987: 77).
Cickson (1987. 77-83), where, using the Dow Index. an Alustration is provided of how a hazard index Is constructed.
20 example. it asbestos cause occupational the disease. (asbestosis). s,mply observing the product not identify risk. and so specialist tesearch is needed to (ink substance
120 to
21 •n oche•' the risk may not be present initially, but will develop eg metal fatigue or stress cocosjon.
such systematic compliance covers a host of possible sources, and, through the compliance
process, these sources are therefore identified and risk control measures implemented. It is clear
Key features of the risk-identification process
that the same outcome is achieved in the field of occupational safety when safety legislation and
In concluding this section on risk identification, some key points can be made.
codes are implemented.
It is unlikely that one particular method or technique of identification will be enough to
Research identify all the risk exposures and address all the associated problems. It is better to use a
combination of methods in order to ensure that identification is as complete as possible.
Naturally, situations exist where risk cannot be easily identified. The intricacy of the product,
Since the various methods themselves have developed in response to solving problems
for example, may mean that observation, however systematised, will fail to identify the
within particular industries certain methods are found more useful in some industries than
associated risk.20
others.16
Under circumstances such as these, specialist research, often on an ongoing basis, is required
The process ofidentification is greatly assisted and improved byconsultation with as many
to identify and evaluate the effects of associated risk. Moreover, the complexity of situations of
people outside the risk management department as possible, e.g. various line managers
this type should not be underestimated, as there may be a time relationship between the
and others in the workforce who know the organisation and who possibly have their own
circumstance and the risk'] and even the situation where using the product causes risk in some
views on the risks that exist, particularly in their specific areas of work.
conditions or environments and not in others. Unfortunately, these aspects frequently indicate that
Risk identification is an ongoing process, and should not be regarded as an isolated or
research is conducted after the loss has occurred, which makes research not an identifying process,
once-off exercise. It is essential that risks that have been identified are monitored and new
but one concerned with understanding the risk and perhaps mitigating its effects. A particular type
of research involves examining loss events (accidents, insurance claims, near misses and so on). risks highlighted.
If an event occurs, lessons can be learnt from these. Events are thus a source of risk identification. Finally, risk identification involves a certain degree of creativity. While past routines used
in risk identification provide some system and rigour to the activity, no limit should be
Risk sourcing imposed on any lateral thinking about risk identification.
Risk sourcing is an approach where possible sources of risks are identified and evaluated rather
than the risks themselves. The purpose is to try to understand the underlying causes of risk. For
example, a manufacturing company that is concerned about the time it takes to get a product to 5.3 Summary
the market may have to re-evaluate its business processes and suppliers, and even its distribution
processes. The model for risk management developed in part I identified the various steps in the risk
Risk sourcing has the additional benefit that it helps managers to understand the type and management process. Risk identification and risk evaluation were shown as the preparatory
availability of data that will influence (a) how the risk can be measured, and (b) the selection of steps to risk control and the financing of losses. The aim of this chapter was to provide
insight into the identification and evaluation of risk.
risk control measures to control or eliminate the risk.
In so doing, the intention was also to address the relative neglect, in the general risk
management texts, ofthese aspects ofthe discipline. Since risk identification and evaluation
represent integral parts of the total risk management process, any holistic approach to the
subject requires that these steps be understood and their rote appreciated. The aim is to
16 f16
W for example, is an appropriate method of Identifying risk in an industrial pr«ess that involves goods or materials the diffetent stages, However. whese now i',
not the main activrty, another method identifrcatlon vauld be more suita ble.
122 RISK MANAGEMENT: MANAGING RISKS 6 evaluation
viewed in a reasonably rigorous way by first describing a macro-identification process, and then a micro-identification
process. Chapter
The various methodologies used in these two processes were described. Important features of risk
identification were also listed. It was stressed that the various risk identification methods must not be seen in
isolation. Risk identification is a dynamic process in which it is essential that risks that have been identified are
Risk
evaluated and monitored, and new risks highlighted.
es€(o« Introduction
Loss dimensions
develop a structured and integrated process of risk management. Hence, the question of risk Loss
identification was itself 123 frequencies and probabilities
Risk profiles
Measuring risk control parameters
Statutory risk assessment and
evaluation
Summary
6.1 Introduction
124
The purpose of explaining risk evaluation in this section is to develop a methodology for risk management in
a logical series of sequential steps. This section introduces further evaluation parameters as an extension of those
discussed in chapter 3, which provided a definition of risk, and where the quantification of risk was considered.
The aim of this chapter is to provide insight into the following:
• loss dimensions (the size of a loss or losses) loss frequencies and probabilities
• loss/frequency distributions variability as a measure of risk.
6.2 Loss dimensions

From a risk management perspective, one ofthe most important considerations is determining the size or quantum of a possible loss. Viewed 'internally', this provides a backdrop against which efforts to
manage the risk can be measured. And viewed 'externally', it becomes a departure point from which the insurance market (or some other financing mechanism) gauges the risk.
EVALUATION 125

While the largest loss may occur in location A, this location itself may comprise
By way of illustration, consider an organisation with three operational locations as indicated
separate asset elements, and it is therefore unlikely that the insurerS probable exposure
in table 6.1.
from fire is location A's full asset value of R260 million. An inspection of location A
is needed to determine with greater accuracy the exposure of the insurer. In an effort
Table 6.1 Asset values per location
to describe the possible loss more accurately, the following measures are suggested: 3
Asset value • full asset value (FAV)
Location • r maximum foreseeable loss (MFL)4
(R millions)
• zestimated maximum loss (EML)
260
• normal loss expectancy (NLE) q expected aggregate annual loss (EAAL).
160 B
c 180
600 Maximum foreseeable loss (MFL)
Total
Definition
It is clear that an insurer underwriting the risk of fire is exposed to a possible claim equal to the
The maximum foreseeable loss is defined as the value of the largest possible loss from a
full asset value (FAV) ofR600 million. It is, ofcourse, unlikely that all locations will be totally
single event that is reasonably foreseeable under the most adverse conditions.
destroyed by a fire during the same period of insurance. In all probability, the largest loss would
occur in location A, which has the largest value of assets.2
In the case ofa single asset (e.g. a building or a vessel), the MFL would possibly be the value
of the asset, since it is conceivable that the entire asset could be destroyed, e.g. as a result
offire. In this case, however; the MFL is not necessarily the combined asset value, since even
in the case of a single asset, the protective systems could be 'designed' to contain the peril
2 All the information depicted in tabk is nevertheless important, because [he insurer may, instance. need to arrange reinsurance protection for itself should the company not from spreading. Nevertheless, it should be emphasised that the MFL as a measure is the
be in a position to carry the entire exposure.
pessimistic forecast ofloss size, i.e. a worst-case scenario.
It should be noted that different EML and MFL values exist for the different perils and
classes of insurance, and that the determination of these values is largely subjective.
has t>en written concernlng these measur% and consequently the various Mlters do not always agree on tfk terminology used. This adopts [he terrnnology
and definition prov»ded by Frledfander {1977. 26 onwards}, whKh hå5 found wide acceptance in British and South Insurance practice. See also Berliner (1
986; 314—21).
4S
}nonym for. MEL is MPL (maxrmum posstble loss}, it Is suggested that, since a gceat deal Ofconfusron exists concerning the of and Other terms
describing loss the mam measures used EO describe severityofloss are MFL and WL.
126 a Estimated maximum loss (EML) 17

Definition
The estimated maximum loss is defined as the reasonably estimated extent of loss resulting from a single event, given that risk control
measures instituted may not entirety perform the task of loss containment or loss reduction attributed to them.' It is generally less than a
total loss and the maximum foreseeable loss.
17
The ANHican literature usually refers to PML (probabte nuximum loss}, i.e. the manmum amount of boss foreseeaue under ordinary circumstary:es. 6 Friedlander (1977: 27).
RISK EVALVATfON 127
RISK MANAGEMENT: MANAGING RISKS RISK

The EML, therefore, as a measure of loss, is usually smaller than the MFL. Insurers frequently will measures. Fortunately, however, the vast majority of losses do not fall into catastrophe
adopt the EML figure as the reasonable limit of maximum exposure faced, while accepting that a categorisation; these relatively minor-type loss quantums make up the NLE measure. 7
higher maximum loss may indeed occur. Usually, MFL and EML measures relate to the material
damage resulting from fire and explosion perils, although such measures may also incorporate
resultant losses due to business interruption. In determining the fire and explosion MFL/EML, for
The normal loss expectancy is the average loss that could result from a single all risk control
example, it is necessary to do the following. Divide the risk into fire areas formed by one or more
measures operate as expected.
buildings/ compartments not separated structurally or spatially. given that
Determine the fire area with the highest potential loss, taking into consideration the fire load Thus, if a fire breaks out during normal working hours, it is likely that it will be offce and will
(i.e. the quantity of combustible materials within the fire area) and the likelihood of explosion,
damage only part of that offce.
and the susceptibility of plant/goods to smoke, water, corrosion damage and the like. one
Estimate the MFL/EML that could be sustained. The MFL would be the maximum monetary
loss that could be sustained on a risk exposure as a result of a single fire or explosion when the annual aggregate loss (EAAL)
most unfavourable One of the most important estimates of losses is the expected annual aggregate of individual
circumstances combine and when, as a consequence, the fire is not at all or not satisfactorily losses. Thus, a company may have a motor fleet of 16 000 vehicles, from which a number of
combated, and is therefore only stopped by impassable obstacles or by the lack of anything else individual losses may result. The aggregate of these losses must be estimated if these losses are
to burn. The EML would be an estimate of the monetary loss that could be sustained as a result to be funded. Insurers therefore always keep records of the annual aggregate cost of losses and,
of a single fire or explosion considered to be within the bounds of probability. The estimate using the loss ratio, determine the premium, as explained in previous chapters.
ignores such remote coincidental events and catastrophes that may be possibilities, but that still
remain unlikely.
6.3 Loss frequencies and probabilities
Normal loss expectancy (NLE) The process of risk evaluation, as stated, requires the numeric assessment of risk exposure. The
section on loss dimensions concentrated on providing various definitions of loss quantums used
The measures ofloss defined so far give a pessimistic estimate to possible losses, expected or
in the field of risk management.
probable under varying worst-case conditions and assumptions. There is therefore a 'catastrophe
Associated with loss dimension is the question oflikelihood ofloss, which, when expressed
loss-type' connotation associated with these
in numerical terms, implies the frequency, relative frequency and empirical probability of losses
or events. This concept has been referred to previously: in defining normal loss expectancy, for
example, the qualification was expressed that, ordinarily, the majority of losses experienced do
not fall into the catastrophe-type category. Put simply, this states that ordinarily a low frequency
of catastrophe losses (i.e. of a high quantum or severity) occur. The added dimension of
frequency or probability of loss becomes the remaining component in the evaluation process of
assessing the risk. 8
7 Filaj[ander (1977:
8 sectitn on the theory of probablhry and its anlication co risk measurement in chapter 3 showed that 'n computing the pure premium, loss frequency was comblned
uvlth average loss severity to a rlive at the expected loss_
EVALUATION 129
128
Loss frequency
Definition 1-2 000

The loss frequency is the measure of the number of events9 that occur in a 11
defined time or consequence interval. 2 001-4 000
20
4 001-6 000
The frequency may relate, for example, to the number of motor accidents 6 001-8 000 8
8001-10 000 6
per year, the number of accidents per year within a cost interval, and the
10001-12000 4
frequency of industrial accidents per thousand worker hours worked.
12 001-14 000 3
Since risk management's concern is also with the 'consequences' of
14 001-16000 2
events, one could, by way of illustration, observe the frequency of motor
loss costs arranged in cost intervals, as depicted in table 6.2. 16 001-18000 18 1
Table 6.2 shows that the motor fleet under review experienced 62 001-20 000
accidents during the period defined, and that there is a relatively low 20
frequency of large losses and, conversely, a relatively high frequency of Total 62
smaller losses. The tabulation of the frequency of events against respective There is a close association, therefore, between frequency and the concept of probability; both represent
intervals of loss costs, as depicted in the table, is useful when transforming expressions of, or surrogates for, likelihood. It is not the intention in this section to further discuss the concept
the results into a graphical representation (i.e. distribution curves). Such
of probability. Rather, we will reinforce its application in the process of loss predictability, and hence risk
distributions have properties that are used for predictive purposes in risk
management. evaluation.
However, it is important to point to an area of confusion among risk management practitioners. Inductive
probability theory suggests the estimation of an outcome of a future event by observing its relative frequency
Loss probability
of occurrence in the past. To avoid confusion s requires proper understanding
The notion of probability and its applicability to the discipline of risk
K
management were discussed in chapter 3, where the definition given was
of the following terms: event', 'outcome, 'frequency of occurrence; 'relative frequency' and 'probability'. The
that empirical probability is the relative frequency with which an event
contention is not that such terms are generally not clearly understood, but that ill-considered reference is made
occurs as the number of observations tend towards infinity. More
specifically, it was stated that relative frequency is the proportion of time to them, so that probabilities are sometimes 'assigned' to events that do not display a probability of outcome,
in which an event takes place; while the relative frequency with which it but rather frequency of occurrence.
takes place in the long run is its probability. For example, if it is said that there have been four major earthquakes in South Africa over a 400-year
period, the tendency is then to say that the probability of an earthquake occurring in South Africa is 0>01,
Table 6.2 The frequency of motor loss costs in cost intervals
simply because this fraction is much less than one. This, however, is clearly incorrect, since this is the annual
Cost interval Number of frequency of earthquakes and not the probability of earthquakes. It could be possible to have 100 earthquakes
times the in one year, in which event the annual frequency is 100 and it becomes clear that the ratio is not a probability.
toss falls
Consequently, reference is made, for instance, to a probability of an accident occurring when there is certainty
into the
that an accident (event) will occur.
interval
Actually, the intention is to refer to the probability of an accident or occurrence having an outcome of,
(R)
6 say, X loss value, where such probability is estimated by observing the outcome's relative frequency of
occurrence — associated with an outcome. For example, if electrical transformers fail at the rate of 1 in 200
years, it is not very useful to discuss the possibility of a transformer failure for a municipality that has 20 000
installed transformers. It
9 pr. event usual ly elatesto an outcome. Thus, if it is said the there væTeX fatal Kies during theyear, impliclt in th IS isa double statementwere X accidents, and (b) in X
cases Rhere were X fatalities. If the total populaticn if N and if a rela(lve frequency is to be determined, it muld be
X.AI for fatalities and (N — XEN for non-fatalitæ.
10 Cummins and Frelfeldeq (i979a• 20-6, 32-6).
130 EVALUATION 131
is a virtual certainty that a transformer will fail. In all likelihood, what needs to be determined is
the probable number of transformer failures.
6.4 Risk profiles
Once the various probabilities have been estimated, the impact of risk control measures Once a macro-risk identification exercise has been carried out and the MFL, EML and
on the probabilities and the effectiveness of risk control programmes can also be estimated. probabilities have been determined, these can be depicted in a table and the risk profile ofthe
organisation calculated. Table 6.3 indicates the format of a typical risk profile.
Loss/frequency distributions Table 6.3 Format of a typical risk profile

Table 6.2 summarised a loss experience pertaining to a motor fleet: it showed the number of Consequential
accidents against costs (expressed in intervals). It is useful to depict such findings graphically. Property Bodily Liability
The data in table 6.2 plots what is referred to as a 'frequency distribution', which is illustrated bsses injuries exposure
i
in figure 6.1.
EML Proba- Proba- ProbaNumber Proba-
EML
Figure 6.1 Loss/frequency distribution bility bility bitity EMI

bility
Fire RXXX Y RXXX Y Y RXXX y
25 Explosion RXXX RXXX Y
y Rxxx Y
Earth-
20
Rxxx Y RXXX
c 15 quake Y Rxxx Y
10 Crime y RXXX
Y Rxxx Y
5 6.5 Measuring risk control parameters

In practice, measuring the performance - or effectiveness — of risk control programmes requires the
consideration of:
2 000 6000 10000 14 000 IS 000 22 000
fatality or accident or loss frequency rates
Loss size • the degree ofcompliance with standards.
The loss tendency is clearly discernible from the figure. The larger quantum losses are
Fatality or accident or loss frequency
experienced less frequently than the relatively lower quantum losses. Various curves exist that
One measure of the effectiveness of a risk control programme is the extent to which the
describe similar loss tendencies (and other tendencies) that have been studied and whose
ofpeopleexposed changes over a period of time. The purpose of a risk control
properties, for example, find predictive use in the field of risk management. 10 Suffce it to say
programme, as indicated previously* is to reduce the frequency ofloss-producing events or the
here that distribution curves are significant as instruments in the valuation process in the
severity of these events. If the programme is successful, then it can be expected that the number
management of risk, as set out in chapter 3.
of loss events will decrease.
Figure 6.2 indicates the fatality and accident rates for south African gold mines over a frequency cannot be measured or the measurement is quite meaningless. It therefore does not
period of 100 years. It is clear that in both cases a meaningful decrease has taken place. Of help to plot the earthquake frequency/severity curves at any particular location, since
course, statistical measurements do not prove 132
11 (1991: 1997b; 2009)-
EVALUATION 133
causation, and therefore these measurements often lead to acrimonious debates. ll It is, in any
event, obvious that the fatality rate in the mining industry has decreased considerably over the the frequency is so low that this curve cannot be used to indicate the effect of improved
years, but was this due to improved safety systems or not? A second question is: Can the rate standards against time, and the number of earthquakes cannot be controlled by specific risk
be improved still further? It has often been pointed out that an improvement in these measures control programmes. Thus, the frequency of earthquakes will not indicate whether an
must follow the law of diminishing returns. The dramatic decrease in the fatality rate realised earthquake risk control programme is effective or not. On the other hand, engineering or
during the earlier part of the previous century cannot be repeated or sustained indefinitely. building standards can be defined and an audit conducted to determine if buildings have been
erected according to these standards. In practice, in the safety field, both measures are used,
Figure 6.2 Fatality and accident rates in South African gold mines, 1904-2004 i.e. the accident/fatality rates and the degree to which the safety standards are being
complied with.
7 70
60
6
6.6 Statutory risk assessment and evaluation
50
5 It can happen that a law imposes the obligation to carry out risk assessment, as in case of the
mining industry where the obligation exists in terms of section Il of the Mine Safety Act.
40
30
6.7 Summary
2 20
Risk can be expressed in numerical terms. The insurer is primarily interested in the
magnitude ofthe exposure that he/she has agreed to insure. This magnitude is expressed
jn concepts such as MFL, EML and NLE. probability theory is applied to individual losses, it
0 quickly becomes apparent that, as the exposure units increase, the risk that the risk
1904 1914 1924 1934 1944 1954 ! 964 1974 1984-- •1994 2004 manager faces is not whether or not an individual event will take place, but rather what
Year
the variation around the 1 expected aggregate loss will be.
Fatality rate Accident rate
Source: Vivian (2009: 11)
Degree of compliance with standards

A further measure ofrisk control is the degree ofimplementation of risk control standards. These
standards or objectives can be set and an audit carried out to determine the degree to which
they have been adhered to. This method is particularly useful, since in many instances loss
OPERATIONAL RISK MANAGEMENT
RISK MANAGEMENT: MANAGING RISKS OPERATIONAL RISK MANAGEMENT
136
7.3 Operational risk factors

Defining the scope of operational risk allows a more focused approach, especially when qualifying and quantifying the risk
exposures. qualifying and quantifying of risk exposures are imperative when these exposures threaten the well-being of the
organisation and must therefore be managed. It is, furthermore, important to note that these operational risk factors could cause
a risk that could negatively influence the business operations. As such, these factors are also known as causal factors. According
to Dowd,' the focus on causes is imperative for the management of operational risk. However, to be able to measure the
operational risks and assess their potential impact on the organisation in terms of their frequency and severity, it is also necessary
to determine the loss events and their potential effect.
An example of how operational risk can be analysed in terms of the causal factors, risk events and effects is illustrated in
figure 7.1.
OPERATIONAL RISK MANAGEMENT 137
for business operations and could have a •or impact on the organisation if they are negatively
influenced by risk events. s such, it is crucial that processes, as an operational risk factor, are
clearly identified and understood.
Youngs states that process risks are associated with day-to-day business processing tbat
could have a negative effect on the operations of the organisation. A key component of
operational risk is the risk of failure in the processing of data, which could, in generals result in
processing failures such as transactional errors relating to failures as a result2f incorxgct data
and reconciliation failures, which relate to failures of reconciliation procedures or controls
designed to check for errors or verify operational performance.
Process risk also stems from business processes being insuffcient and causing unexpected
losses. However, most of the time, processes are linked to systems, which could also be a
separate operational risk factor.
Systems (technology)
The modern world's dependency on technology and systems causes system risk to be a
significant part of operational risk exposures and it should thus be controlled on a daily basis.
Furthermore, the rapid pace of technology development could also have a major influence on a
business, and effective
risk management should ensure that the latest technology is identified and implemented. By
doing this, the organisation will be able to conduct its business according to the latest
technological developments, which could result in a competitive advantage. However, it is also
imperative that the organisation be prepared for any shortcomings relating to systems that could
have a negative effect on its operations. For examples system outages could result in a loss of
business, or old technology could negatively influence the effciency of crucial business
processes.
As a result of this, it is crucial that systems form a core part of operational risk
management. However, the correct functioning of systems could also be dependent on the
human factor. As such, it is also important to understand people risk as a factor of operational
risk.
5 Young (2006. 15}.

Figure 7.1 Analysis of operational risk in terms of causes, events and effects
Each causal factor will be discussed in more detail in the next sections.
Processes
ve or anisation is de endent on rocesses to im lement its strate and achieve its 0b' ectives. As such, rocesses can also be re
arded asthe latform
4 Dowd (2003'.36).
staff should be aware of the communication policy throughout the organisation related
to risk reporting and decision-making.
The ethics of risk management, which should include a code of conduct, should be
communicated to all staff members.
The levels of authority of various staff to carry out their responsibilities relating to risk
management should be communicated to all staff members.
Performance incentives for risk management should be implemented. Performance should

be monitored and those responsible for performance against set targets should be
ll
adequately rewarded and those acting unfavourably should be disciplined. Haubenstock
states that the design of
performance incentives should not provide incentives to people to operate contrary to the
desired risk management values.
e r%ubenstock {2003. 257)-

1).gussain
COCO 93b
'i Saubenstock {2003: 2581
RISK MANAGEMENT: MANAGING RISKS 139
Risk management culture
management culture consists of the principles and related values of

managing operational risk.
bottom line for a successful risk

management culture is enclosed in a
risk language for the organisation. Haubenstock 8 defines a risk
management culture as the set of shared attitudes, values, goals and
practices
characterise how an organisation considers risk in its daily activities.
states that a risk management culture is all about beliefs» attitude,
judgement, approach and outlook, which manifest themselves through
behaviour.
Sometimes the risk management culture is taken for granted; however,
it is
to explicitly identify and embed a risk management culture for the
organisation. The risk management culture should address the following risk
management principles:
members of staff at the different
management levels should be
involved. According to Hussain, 10 an approach
to risk management needs
developed that involves staff and the
processes at all levels of the
organisation. The management and the workforce should challenge the old
paradigm (i.e. way of doing things) and develop a more flexible
approach
with the ever more complex and changing
environments that
businesses have to face.
Definition
140
Staff must be adequately skilled and trained in risk management to ensure that operational risk
exposures can be identified and effectively addressed.
The external environment in which the business operates should be monitored and assessed to ensure that internal
processes are aligned to meet or exceed expectations.
The management of operational risk should take place on a daily basis. eme dynamic nature of operational risk should
ensure that it forms part of every employee's daily activities.
Adherence to legal and regulatory requirements must be promoted throughout the organisation by means ofpolicies and
procedures. The business's appetite for risk will vary in accordance with its objectives and culture, and evolving conditions
in the overall business environment.
A number of questions can be used to establish whether an organisation has a sufficient and effective risk management culture
embedded in it. These questions are illustrated in table 7. l .
'Ihe next important aspect ofoperational risk management is the governance structures, which
will be discussed in the following section.
7.5 Governance structures

Definition
Governancestructuresaretheformallyapprovedorganisationalrisk management structures and committees that
govern operational risk management.
In light of this definition, the governance structure for operational risk consists of two components,
namely:
• an organisational structure
• a control structure.
Organisational structure
To ensure a successful operational risk management function, it is imperative that a formal
operational risk management structure exists in the organisation. lhis structure should be based
on the management strategy and approach to operational risk management. Ibere are two ways
to approach the implementation of an operational risk management structure. Firstly, it can be
structured according to the functional operational risk management activities, and secondly,
according to the business organisational structure.
The primary functions of operational risk management can be categorised according to the
components of an operational risk management process, namely:
risk identification
• risk evaluation
• risk control and monitoring
• risk financing.
A functional structure for operational risk management is illustrated in figure 7.2.

Table 7.1 Questions to evaluate the status of an organisation's risk management culture
Question
Is an approved risk management policy in place?
Do risk management responsibilities reflect on all employees'job
2.
descriptions?
Does the organisation have a formally approved code of conduct that is

3.
effectively communicated throughout the organisation?
Is the risk appetite reviewed regularly in accordance with changing

4.
strategies and business objectives?
5. Are policies and procedures aligned with regulatory requirements?
Are risk management successes monitored and are employees rewarded 6.

accordingly?
Many more questions could be formulated to determine the status of a risk management culture, but they would depend on
the degree of development of risk management within the organisation. However, it is clear that an embedded risk
management culture within the organisations plays an important role in the success of operational risk management.
RISK MANAGEMENT: MANAGING
142 successes of the planned responses, but also on changes in the business that might signal new
emerging risks. Therefore, there is a close link between risk monitoring and business processes.
The risk monitoring process should be able to serve as early warning to business management
Figure 7.2 Functional structure for operational risk management regarding new risk exposures or threats. ms will allow management to be proactive in managing
the new exposures in terms of control measures. The second aspect, which is related to risk
monitoring, is risk reporting. Risk reporting can be regarded as one of the more important
activities of risk management.
Risk reporting involves the collating of relative information and producing risk reports for top
management. A number of important issues should be considered during the process of risk
reporting, e.g.:
Risk data must be accurate and timely.
Risk data must be analysed to ensure that only relevant information is included in a risk report.
Risk reports must include information that will allow management to make effective
decisions regarding the implementation ofcontrol measures and the disclosure of
information to stakeholders (including regulatory bodies).
further important part ofrisk rep Orting is the governance of risk management.
According to the Institute of Directors in Southern Africa's King Ill report, 13 board may appoint a
On the other hand, risk management could be structured according to the business structure.
dedicated risk committee to assist it in carrying out its
An important aspect in this regard is the reporting lines, which are illustrated by figure 7.3.
Figure 7.3 Operational risk management structure according to the business structure
Chapman (2008: TX).
143 !nstitute of in Southern Africa (2009; 79).
RISKS
144
Each business should have a dedicated risk management function. Although these risk oficers
report to the respective business unit managers, they also have a reporting line directly to the
responsibilities in relation to risk management. The board should specifically consider the risks
group risk department or chief risk offcer (CRO). The CRO has a reporting line to the CEO, as
that may affect the sustainability of the organisation. As such, it may be appropriate to mandate
well as the board risk committee.
a risk committee — the board risk committee — to ensure oversight of sustainability, which
includes considering and recommending the sustainability report to the board for approval. It
Control structure is clear

that the board risk committee plays an important 14 role in the risk management
An important part of a risk management process is control. Two critical components of risk control are
process. The board risk committee should:
risk monitoring and risk reporting.
• oversee insurance arrangements, including the acceptance of risks that will not be insured
Risk monitoring is the process of observing the progress of risk controls in terms of the action
• oversee IT strategy, governance and risk management
plans. As such, it is imperative that it is embedded in the business processes and the risk
12 • review the risk management maturity of the organisation, the status of risk management
management culture. According to Chapman, risk monitoring must focus not only on the
activities and the significant risks facing the organisation. (The organisation's risk
management arrangements should incorporate risk reporting processes, including risk The components of a typical risk management process can be described as follows:
trends, risk materialisation, forecasting and emerging risks.)
)ßisk identification is regarded as the first step of the process and aims to analyse the strategy,
• consider the risk management strategy and policy and monitor the risk management process objectives and processes of the business in terms of risk exposures and threats. These are
(effective and continuous monitoring is an essential part of the risk management process) also known as the inherent risks, which make up the initial risk exposure without taking into
consider and evaluate, among others the following: a register of key risks; estimated costs account any control measures.
of significant losses; whether risk management costs are consistent with the risk profile of
*Risk evaluation involves the measurement (quantifying) and assessment (qualifying) of the
the business; material losses; reduction in earnings or cash flows caused by unforeseen
inherent risks. The inherent risks are evaluated to determine the potential severity and
incidents; material changes to the risk profile; details of risk finance arrangements that
likelihood of risk events, as well as the adequacy of the risk controls. The result of the
could expose the company; the risk-bearing capacity ofthe business; due diligence
evaluation process is the residual risks, which are the risks that remain after taking control
activities; and information technology risks
measures into account.
• ensure that risk assessments, risk reports and assurance on risks overseen by other board *Risk control is the next component, which aims to address the residual risks that remain a
committees are referred to those committees for their consideration. threat to the business. Current control measures need to be improved and additional control
measures should be developed. In this regard, it is imperative that actions plans be developed
Although the abovementioned activities are delegated to the board risk committee, certain that will ensure that the effects of risk events, should they occur, are minimised or
activities cannot be delegated to a committee. Examples of eliminated.
these activities are: *Risk financing is a risk management component that is often not considered, although it forms
the approval of the organisation's risk appetite the approval of the risk management
a crucial part ofthe process. This component forms a close link with risk control, as it also
strategy for the organisation the approval of the disclosure reports to external parties, such
strives to ensure that the cost of risk management does not exceed the benefits. The various
as regulators.
elements of risk financing can be summarised as follows:
Provision should be made for operational losses during the budget process, where small
From the above discussion, it can be deduced that governance structures form an important
losses are expected during the normal operations of the business and subsequently
part of an operational risk management framework and should
budgeted for to eventually be written off against the budget.
146 RISK MANAGEMENT: MANAGING RISKS
14 1 nsdtute of DI recus in Southern Africa (2009.80}.
145
The cost of specific control measures implemented to prevent loss events from
occurring should be budgeted for.
ensure that the roles and responsibilities are clear and embedded throughout the organisation.
Insurance policies should be arranged for those risk events for which a loss will be
After an approved risk management structure is in place and the posts adequately filled, the next
step is to develop and implement a formal operational risk management process. shared with a third party should the event occur. Thirdparty insurance and self-
funding are popular insurance methods.
Capital should be allocated for operational risk, where the organisation determines a
.6 Operational risk management process percentage of capital to be allocated to operational risks. These types of exposures are
usually catastrophic and could have a major influence on the existence of the
A clearly defined operational risk management process could be regarded as component of a
organisation.
typical operational risk management framework. A risk management process aims to provide a
\-Risk monitoring entails a continuous evaluation ofthe business operations to ensure the
structured approach to risk management so that the risk exposures and threats are identified and
effectively managed in a proactive way. It also aims to ensure the formulation of a realistic adequacy of control measures and the identification of new risk exposures and threats to
operational risk appetite. the organisation. As such, risk monitoring can be regarded as an early warning system for
RISK MANAGEMENT: MANAGING
management, in order to be proactive in identifying new risks and implementing control Figure 7.4 An operational risk management approach
measures to prevent a risk event or to minimise the effect should the risk event occur.
Risk Identification
Various methods and tools are available for operational risk managers to ensure an effective
risk management process. However, currently four primary are being used. These are discussed
Loss management Key 115k
below.
Indicators
Loss management: Actual losses are recorded and analysed determine the risk exposures Risk
appence
Risk profile
in terms of the event, values and causes. Corrective measures are then implemented to
Scenanos Risk & control se/frassessments
Risk control & monttoq ina
prevent similar events from occurring. As such, preventative actions are taken based on
what has happened in the past.
•k Key risk indicators are monitored on a regular and continuous basis to determine trends and
measured against a pre-set threshold. As soon as the threshold is breached, it is an risk profile is determined from various reports resulting from the implementation of
the operational risk methodologies. It could, for example, be presented in the format of pie
indicator to management to take action to prevent a negative impact on the business.
charts, histograms and line charts. The risk profile indicates the past, current and future
Corrective measures are thus taken, based on current information. risks, which can be used as an input to determine the operational risk appetite.
• {Risk and control self-assessments are performed by analysing business objectives and
processes. Inherent risks are identified based on potential risk events that could occur Definition
some time in the future. Operational risk appetite is the amount of risk an organisation is prepared to
4Scenario management is another method used to confirm potential future risks, based on
the views and experience of experts.
For the unacceptable risks, management will, firstly, try to avoid the risks by terminating
the riskactions or
The results of the abovementioned tools are various risk reports, which can • ensuring a balanced approach between being risk seeking and risk averse be used to determine an operational
profile for the organisation. The ensuring a better view ofrisk expenditure, which will ensure that the cost- processes involved,
147 or, secondly, try to transfer the effect of the risk event should it occur by means of, for
example, third-party insurance for expected risk events or capital allocation for
unexpected risk events. It is important that the organisation should determine the
operational risk appetite by using the available information and methodologies. Defining
an operational risk appetite benefits the organisation by:
assisting strategic planning by aligning strategic objectives and operational
risk profile, in turn, serves as an input to determine a realistic operational
risk appetite. approach to operational risk management is illustrated in
figure 7.4.
of-risk does not exceed the benefits inculcating a culture of risk awareness throughout the organisation
MANAGEMENT: MANAGING RISKS
148 RISK MANAGEMENT: MANAGING Once the impact and likelihood of operational losses are derived, it is possible to determine the
operational value at risk, which can be defined as the value of expected loss at a chosen confidence
level for a particular time period. This can be calculated by the aggregated loss distribution that
• ensuring sound decision-making by top management improving the corporate governance 15
allows a prediction for operational losses. Figure 7.5 illustrates the aggregated distribution,
of the organisation improving the reputation of the organisation improving effective
which can be derived by means of the Monte Carlo simulation model, which gives the probability
and embedded risk management in the organisation.
for the loss amount and is used to derive both the expected and unexpected losses, given a
particular confidence level.
Notwithstanding the above, there is one issue that still requires attention to ensure the effective
management of operational risk. This is the issue of the quantification of operational risk. Figure 7.5 Aggregated loss distribution
Expected fosses Unexpected losses
Except for actual losses, which are a quantifiable component, there is no simple objective
method available to quantify operational risk. As mentioned at the start of this chapter, it is
imperative to be able to measure the risks in order to manage and control them. In this regard,
operational risk modelling is being researched in order to objectively quantify operational risk.
Catastrophic
This will be introduced in the next section.
7,7 Operational risk modelling

Impact
The aim of operational risk modelling is to determine an objective value for the risks. As
mentioned earlier, the quantification of operational risk is one of the more diffcult aspects of
the management discipline. It is especially difficult to identify future unexpected risk events, Adapted from Young (2006:84)
let alone quantify them. As such, one method of quantifying future operational risks is by
Flhis
stochastic modelling. distribution curve can also be used as an input to determine the risk appetite of the
organisation, i.e. the risk an organisation is prepared to accept or tolerate. Once an aggregated
Stochastic modelling aims to calculate the operational value at risk, which is generated
loss distribution has been determined, it can be used to serve as an input to finalise a realistic risk
through the process of determining the severity and likelihood of operational losses. This appetite. For example, the expected losses can be used as an input to determine what management
modelling approach takes into account the loss history of the organisation in terms of a is prepared to tolerate or accept should such a risk incident occur. These types
statistical distribution analysis. The process involves collating loss data, which is used to
determine loss distributions for the impact and likelihood of losses. Once the distributions are
calculated, the expected and unexpected losses are derived from aggregated distributions.
(2002: 137)
The loss distribution calculation involves two types of models: 150 RISK
• Impact models aim to illustrate the impact of risk events by means of exponential
distributions such as Weibull distributions and Pareto distributions. Another technique
of risk incidents will usually fall in the category of low impact/low likelihood, indicated by A in figure
to deal with the impact of operational risk events is extreme value theory (EVT), which
7.5, and would involve implementing additional control measures, which could mean an increase
provides a way of getting around the problem of a shortage of operational risk data. This in the cost of controls. Secondly, in the category of medium-impact/medium-to-high-likelihood
theory is an area ofstatistics concerned with modelling the limiting behaviour ofsample incidents (B in figure 7.5), these can serve as an indication to top management to involve a third
extremes, which indicates that, for a large class of distributions, losses in excess of a party (insurance) to share the risk consequences should the risk incident occur. Finally, in the high-
high-enough threshold all follow the same distribution. impact/low-likelihood category (C in figure 7.5), management can decide to allocate a capital
OPERATIONAL RISK MANAGEMENT 149 amount for a potential catastrophic incident. These types of incidents are disastrous and could
cause the organisation to stop operating. However, should capital be allocated for this type of risk,
Likelihood models aim to perform an estimation of the frequency of operational risk events it could ensure that the organisation will have enough capital reserves to continue its business.
in the future. Typical operational risk frequency distributions for example, the Poisson Banks are typical organisations that would fall into this category and are also compelled by law to
distribution and the binomial distribution. allocate capital to operational risk management.
7.8 Summary
Stochastic modelling for operational risk is only one example ofmodelling. Other modelling
methods are being developed and implemented, such as causal modelling scenarios and
This chapter aims to introduce operational risk management as a separate risk
simulations, which are used to predict the potential behaviour of processes and estimate potential
management discipline. As such, a definition of operational risk was identified and the
losses. They entail creating a simulation model of operational processes. The parameters for factors involved were discussed. The embedding of a risk management culture forms an
modelling are derived from historical loss data sources supplemented by scenarios. According to important part of an operational risk framework, together with suffcient governance
Cruz,16 the simulation is used to derive distributions of operational losses, from which expected and structures. The risk management process is most probably the most important part of risk
unexpected loss statistics are calculated. In order to achieve this, various linear and non-linear management, as it provides the methodologies and tools to identi&, evaluate, control,
finance and monitor the risks.
models are available. Examples of linear methods are multivariate factor analysis, multifactor
Finally, risk modelling was discussed to illustrate how operational risk can be quantified in
modelling and non-linear models, which include the Bayesian belief networks, which can be used order to add value during the determining of a realistic risk
to provide a more accurate picture of the processes.
The modelling of operational risk is still at a very basic stage and is being tested by various Operational risk management ts fast becoming one of the more important management
disciplines, and it is therefore imperative for all organisations to pursue a positive approach
industries and institutions. However, it is clear that to be able to quantify operational risk,
to this discipline to ensure the effective continuation
additional methods will have to be used to add to the only objective, quantifiable current measure,
which is the loss history. The models used to quantify operational risk may differ from organisation
to organisation, depending on the business strategy and developments relevant to a particular
business. It is, furthermore, important to use modelling to serve as an objective input to determine
a realistic risk appetite.
152
Chapter 8
Risk response.• Risk
control
Introduction
Meaning of risk control
MANAGEMENT: MANAGING RISKS
Development of risk control Definition

General principles and objectives of risk control The physical control of risk, or the prevention and control of losses — usually referred to as risk control
Approaches to loss prevention l — is mainly used for the control of event risk, whereas hedging is mainly used for mitigating the
Risk control legislation consequences of incidental risk such as foreign exchange risks and other financial risks.
Specific risk control programmes and codes of practice
Hedging of incidental risks The larger part of this chapter will focus on risk control, while hedging will be described later in the chapter.
Summary We describe the meaning ofrisk control and, by reference to its development, set out its principles and
objectives. We also discuss different approaches to loss prevention and control, and describe specific
programmes and legislative requirements designed to control risk. Note that the professional disciplines
encompassed within the risk control activity are both broad and specialised.2 The intention is not, therefore,
8.1 Introduction to provide an authoritative discussion of such risk control disciplines or programmes, but to describe the
risk control activity within the context of the overall risk management discipline so as to achieve the
As the title suggests, this chapter deals with the response to risk. The preceding
necessary integration of the various elements constituting risk management.
chapters set the background against which this and the following chapters
should be viewed. Having identified and evaluated risk, we are now at the stage
where we can respond to the risk. There are three possible responses to
1 Various terms refer to tte aspect of prevention or conuol of losses, and Ibere is no consensus as to its exact meaning. Earlier South African risk management ITOde'5 1-rsed
business risks (including event risks). These are to: accept the risks transfer the the term •physlcal rlsk management'(Vivian 1985; 109}. grltßh waters also use the term 'phys•cal 115k conuol' (Chacon and Carcar 1992) Historically. the emphasis on managing risks has been
on the purchase of insurance, I.e. a financial technque. In the 'ate 1970s and early 1980s, writers in SOLIth Afrr-a started applying an integrated or model, Which includes Pleventatrve measures.
risks mitigate the risks. In order to distinguish prevencatwe fro-rn financial measures. the former were referred to as pbyscal risk management steps, and were deplcted as such in various articles published at Ihe
lime: seeVMian (1 985: I -9; 1987: 10—3}.The term 'physical risk managenent' also appears in English literature (Carter 1 982; 5). The term aEso appears, for exampde, in the EB4 BA curriculum
of the Glasgow College of Technology and the prospectus of the Institute of Risk Mana9ement- The American literatue refelS ma inly to loss control' sometimes 'rlsk control!
2 In th'5 regard. it iS appropriate to consr-del t6e observations reported by Crockford (1982. 1 71); 'The theories Of ('Sk management which have emerged from the insurance
marld have very largely ignored. and have been ignored by, the forms of risk management. which have been •n the engineering wodrJ. It would be necessary to decide whethei or not the
In the case ofbusiness risks, the decision to accept a risk should not be evaluated literature from [Ills area forms parr of the literature of management to be studied. A possible solution would be to treat this literature as that cfa specie list fi-eld of loss control and to subjecr
Il to the ame criteria for admis90n as the literature say. fire pæventaon.'
in isolation, but on an aggregate basis. The aggregate effect on the risk-return
properties should be evaluated in a risk-return space. Only if the acceptance of
the risk results in improving the risk-return relationship should the risk be
accepted. In the case of event risks (where risk has a negative impact only), the
RISK RESPONSE: RISK CONTROL 153
decision to retain the risk should only be considered in the case of risks with a
small impact on the organisation.
The decision to transfer the risk* e.g. in the case of the factoring ofcredit risk, should be carefully
evaluated in terms of the impact of the decision on the riskreturn properties of the organisation. As we have
seen, the decision to insure event risk or a combination of event and business risks amounts to transferring
the financial consequences of these risks and not the risks themselves.
Risk mitigation is an extremely important risk response alternative. In the case of incidental economic
losses, where the risk is incidental to the economic activity of the company, and particularly event risks,
where the risk results in losses only, it is important that the severity and occurrence of these risks be
restricted to a minimum. There are two main mitigation processes, namely: risk control hedging.
154 RISK 3 See chapter figure 1.1. which depicts die Integrated risk management poeess, having given consideralion to risk-financing elementand to {he process of risk
management in the context ot the corporate objectmæ of shareholders"ueaith maximisaton.
155
Event risks can be differentiated broadly as those being related to assets (property) and
those that give rise to liabilities. In view ofcertain fundamental differences between these Definition
two categories, this chapter will include a separate discussion on the legal basis of the latter Risk control is therefore defined as a method of countering risk. In the broader sense (as intended),
risks.
it includes alf activities conducted for the purpose of:
• eliminating or reducing the factors that may cause loss to a person or organisation; and
minimising the actual loss that occurs when preventative methods have not been fully
8.2 Meaning of risk control effective.
It is not difficult to define what is meant by risk control; the literature abounds with examples
categorising actions by individuals losses.and 3 organisations aimed at preventing and It was stated in chapter 3 that risk implies uncertainty surrounding: the
controlling risks and resultant magnitude or severity of losses the frequency of losses.
Depending on the depth of analysis, however, risk control is publicly perceived either as
In order to achieve consistency, risk control should be directed at controlling the two
any process that simply prevents losses or curtails their effects, or as a complex philosophy
uncertainties of severity and frequency-4 However, since risk management is concerned with
that incorporates a number of subtle definitions and disciplines. It may, of course, be
the management of risk, risk control is not only directed at the elements of uncertainty but also
suggested that complexity of definition is often the result of an attempt to highlight a concept
encompasses planning before an event occurs, during the event and after it occurs. In the
or phenomenon that is not normally easily recognised. This applies to the definition of risk
broader sense, contingency, emergency and disaster planning all fall within the ambit of risk
control.
control.
In risk management circles, debates are held regarding the differences — if indeed there
are any - between loss control and risk control, and whether these concepts constitute what
Most risk control activities are therefore follows: directed firstly, towards those
is also termed physical risk management. The reaction is not altogether unexpected; minimisingactivities
traditionally, non-financial aspects of risk management have been fragmented within an
future losses, and these can be categorised as pointing toward controlling the (previously
organisation and treated as separate activities. Thus, in an effort to integrate all relevant
identified or detected) possible adverse occurrence and then endeavouring to eliminate it; and,
areas, subtle differences in definition are bound to arise.
secondly, those directed towards minimising the loss after the loss has occurred (loss control
For the purpose of this discussion, the following course is adopted. Risk control entails activities)'
any activity aimed at preventing losses or minimising the consequences of losses that may Risk control should be measured in economic and humanitarian terms. The former is
arise from all operational risks facing the organisation, even relatively unpredictable risks that generally applied when considering the prevention of property losses, while the latter is
pose a threat to the organisation's existence. The aim is to treat risk control as pervasive in relevant to the loss of human life values. In some cases, both measures coincide, e.g. in the
the business environment and not to limit it to the activities of prevention oflosses that are prevention of industrial accidents.
considered to be inevitable, i.e. those activities that can be described as loss control activities. By implication, risk control is mostly possible and effective in relation to particular risks.
Clearly, we emphasise the concept that risk management is a holistic discipline: loss In chapter 2, section 2.4, such risks are defined as those of which the causes can be controlled
control must be viewed within the context of risk control; and risk control must be integrated to a greater extent by the exposed person or organisation. Control activities are effective in
with the discipline of financial risk management, which then constitutes the science of risk handling the risks of fire, burglary, boiler explosion and the like. Their effectiveness is directly
management. proportional to the degree of control that can be exercised. Conversely, such activities are not
as effective with fundamental risks, which are essentially economic, social, political or natural
in their origin.
MANAGEMENT: MANAGING RISKS RISK RESPONSE: RISK CONTROL
This emphasis is qu ite srgnficant in the risk evaluation stage. 6 to Discussion which these perl!s on account In the of the activities fact that the eau attention focused wider 1900 area. on them 49.2 specifically, notably theas
5F r
o example. risk and loss control can be focused on designing contingency plans to deaf with the eventuality of a fire devebping: fire detection fire extinguishing they lepresenred the peds
techniques dealing with the fire whlle it is bumjng; and. once the fire has ceased to burn. endeavouring to as GPidly as possible as much as possible through salvager and
minimiS1ng the disruption caused by the fire- most people were exposed. Obviously. US. for example. encompass average lifetime in years {Spiegel 1966:34). currently. accomplishments.
figure years (Denenberq 1974-. 81). provide account
156 this detailed histo„cal is development 70 of et al. see Allen The (1943). intention Heinrich Is not (1950) to and Blake a detailed (1963).chronological For
more
7 oenenberg et al. (1974-82-4).

8 SeeVa'samakis (1983-20-7) fix a summary af the developrnent offil? insurarce in England.
8.3 Development of risk control 9 Denenbero et al. 0974: 84)
RISK RESPONSE: RISK CONTROL 157
In understanding current risk control activities, it is appropriate to briefly describe their historical
development, with specific attention to the development of loss control in the areas of fire,
service and research organisations are constantly developing equipment and techniques aimed at
explosion and industrial accidents.6 Early risk control programmes were directed at specific perils
preventing fire, explosions and similar loss-producing
and did not comprise a holistic method of handling risk by management techniques. 10
Flhe present pace of technology has resulted in a higher concentration of assets in

businesses, which create new fire hazards, and prevention specialists are continually required
Fire risk control
to research and combat these. Specialised management techniques are required to ensure that
The first organised loss-prevention activities in which the insurance industry played a dominant modern fire technology bears and minimises the risk from fire.
part were directed toward the prevention of property losses resulting from fire. 7 Fire perils have
the following characteristics:
Everyone is exposed to this risk. Engineering risk control
• Losses occur relatively frequently. Parallel with fire prevention, methods to prevent boiler explosion were developed. For
A considerable number of these losses are severe, even catastrophic.
example, during the early 1980s, state and local governments in the US started paying
Losses may be diffcult to prevent, since in many instances they are due to the actions or particular attention to losses caused by boiler explosions. 11 Insurance companies specialising
negligence of others. in flywheel, engine breakdown, turbine and electrical insurance - which were ultimately
combined under the general classification of machinery insurance — appointed engine and
These factors naturally suggest that fire and its prevention have an extremely important bearing on boiler inspectors. emphasis was on the prevention of loss rather than providing an indemnity
the economy, particularly on the insurance sector, which has traditionally been the medium to meet subsequent to the loss — an objective evidenced by the fact that loss payments generally were,
the financial consequences of this risk It is therefore not surprising that insurers originated early and continue to be, less than the amounts paid out for prevention.
fire prevention developments, appointed surveyors to inspect and report on all properties to be
insured, and stipulated that certain requirements and standards should be met before insurance was
offered.8 Occupational accidents
Thus, from early times, the practice of the careful inspection of property and the insistence Although, like fire perils, industrial accidents have a widespread effect on people, the
on the periodic follow-up of these inspections were present. As fire insurance developed, it became development of industrial accident prevention historically was somewhat different.
apparent that understanding the causes of loss and their elimination were the most effective The growth of industry naturally drew attention to the increasing number and severity of
prevention techniques. Consequently, insurance companies started keeping records of the industrial accidents. However, in those early times, these accidents were considered a result
circumstances of losses. This in turn prompted research into, 9for instance, the causes of ofthe socio-economic conditions ofworkers. It was assumed that the number of accidents
spontaneous combustion and the effects of friction. could be reduced by considering contributing factors such as working hours, pay, lighting,
The preoccupation with risk control, coupled with increasing sophistication, meant that the heat, sanitation and other working conditions. It was commonly believed that workers' wages
industry began employing people with varied, but specialised, knowledge of engineering. Today, were too low and that not only the working conditions, but also the non-working environment
the scientists of numerous manufacturing, needed improvement. 12
Early efforts in industrial accident prevention were made in Russia, one of the first nations
to pass comprehensive occupational safety legislation. 13
1
0As •n example, rnsurance companies such as Munich Re, A'liönz and the Arnerican mutual insurer Factory Mutual have special•sed risk expe«ise covering a wtde S The history or workers' compensator, legislator, in south Africa has been researched by gudlender (1979); see also Budlender (1983; 1984). It appears that the emphasis
range Of losses.These insurers gwoduce autboritative publications in the held of risk control. distilhng the considerable on irrvoved safety standards and a compensation system originated employer groups and government and not ernpbyees, until recentlyr organised labour has not taken
much in these matters. More recently. general term. compensation. has been in of compensation. In south Africa, the Compensation for Occupational and Diseases Act
P€acticaJ 11 Oenenberg experience et al. 0974 gained 8S).by the industry. See, for example, the Alikmz Hardb@k oftoss (Allianz. 1987). NO. 130 of 1993 the term'employee• This cont,nue to use more internationalterm - •worker's (or workers') section 35 of the Compensauon for Occupational Injuries and
Diseases Act, in particular. which limits the employee'and his/her dependant'
12 Blake (1963: 34-6)
17 (1901: 268-79) covered the position of employer: liability and compensation in the dunng nineteenth century. At this time, south African low was Influenced by the
13
I
s interesting to note (hat this legislation. although passed in 1835, was not In fact enforced See Pascoe (1981.486). position In UK. and most writers commenting on the of south *frit-an law rebycm UK law- See, for example. 8uöIender (1979 137)-
158 159
common law, which, as a study of UK cases clearly indicates, was not kindly disposed to this
Accident control in South Africa
kind of action. In the well-known and often criticised case of Priestly v Fowler, a precedent was
South African industry developed a standardised approach to accident control via the National set when the court ruled that an employer was not liable for the negligence of a fellow
Occupational Safety Association (NOSA) programme, discussed below. Similarly, the mining employee.)9 Illis established the so-called doctrine of common employment. The doctrine of
industry developed a similar system, the Mine Safety Management System (MSMS). These assumption of risk also applied to occupational accidents, as did the doctrine of contributory
systems utilise the concept of management by objectives (MBO) and proved to be successful. negligence. The employer was free to contract out of the liability. All these combined to ensure
During the earlier stages of development in South Africa, the emphasis of safety that the common law provided little assistance to the injured employee. It is clear that,
management was on legislation and supervision by government officials. Safety and, to a whatever the theory may have been, in reality the rights ofinjured employees to recover
limited extent, occupational health in industry were governed by the Factories, Machinery and damages by common law during the nineteenth century were very limited indeed.
Building Works Act No. 22 of 1941, usually referred to as the 'Factories Act',) 4 and in the mining Development of workers' compensation in South Africa. At the turn of the twentieth
industry by the Mines and Works Act No. 27 of 1956. Specialised industries, e.g. the explosives century, South Africa's legal system depended heavily on developments in the UK and, as was
industry, have their own legislation, i.e. the Explosives Act No. 15 of 2003. the case there, only a limited right of recovery for injury existed. Large companies in the mining
Even the legislative approach emphasised management techniques and the management industry, however, were not satisfied with this state of affairs and a scheme to compensate
of risk. For instance, the Factories Act required the appointment of a responsible engineer to injured workers was introduced. As early as 1894, the mining industry established the Rand
oversee machinery. The management approach became evident in the Machinery and Mutual Assurance Company.20 Since the mining industry had established its own fund, when
Occupational Safety Act No. 6 of 1983, which repealed the Factories Act. This Act, which, it was general legislation was introduced, the existing mining system had to be taken into
felt, did not give due regard to health issues, was in turn replaced by the Occupational Health consideration.
and Safety Act No. 85 of 1993, while the safety aspect of the 1956 Mines and Works Act was In 1914, a consolidating act, based on the UK Act of 1906 and the New Zealand Workmen's
replaced by the Mine Health and Safety Act No. 29 of 1996. Compensation Act, was adopted that altered the common law.21 The new Act provided for
compensation in the case of all accidents 'arising out of, or in the course of, employment' and
Workers' compensation in South Africa 'S where the accident was not due to the 'serious and wilful misconduct of the employee'. In
Employees' rights in the nineteenth and earlier twentieth centuries. Closely allied to the terms ofthis Act, the injured worker could choose between claiming under the common law or
question of compensation for occupational injuries is the right of the employee to sue the under the new statutory provisions. The worker did not, however, have recourse to both. In
employer and hence the exposure of the employer to the so- called employers' liability risk16 1917, the Act was extended to provide compensation for certain industrial diseases.
covered by the employers' liability policy and the workers' compensation schemes. The position In the early Act, the UK example was followed where the employer was obliged to pay
regarding employers' liability must be viewed against the historical development of these compensation. However, there was no guarantee that the employer would be able to do so.
actions. The next step was to make the purchase of insurance cover against this risk compulsory. The
In the nineteenth century and the early part of the twentieth century, an employee injured Workmen's Compensation Act No. 25 of 191422 was repealed by the Workmen's
in the course of employment had little chance of receiving compensation for injuries sustained Compensation Act No. 38 of 1934>23 which in turn was repealed and replaced by the
at work.)7 Remedies were confined to the Workmen's Compensation Act No. 30 of 1941. Finally, it was also repealed by the
Compensation for Occupational Injuries and Diseases Act No. 130 of 1993.
14 The use ofthe shotened title unföttun.tely resulted n the erroneous view IS pointed that out the in Act only of applied the to factor.es In fact. 3,three separate issues
18 1837 3M and WI.
were factories. machinery and building
2019This case 's often used as the point of departure In legal theory. e.g. Dvorkin {1986: 2),
21 Act No. 25 of 1914, "tilch consolidated a-Cts promulgated in the pre- 1994 Cape, Transvaal and Netal provinces. For a
comrnentary on this cmsult Barry (1914).
22 8arvy (1914} published a commentary on the 1914 Acr23 Nathan (1935) published a commentary on this Act.
RISK MANAGEMENT: MANAGING RISKS RESPONSE: RISK
RISK CONTROL 161
160
modification and expansion of the common law rights and duties of employers Insofar as the
risk manager is concerned, workers' compensation is and employees in the event of industrial accidents followed. Under these important.24 The Compensation for Occupational Injuries and Diseases Act
circumstances, the employer had little incentive, other than a moral one, to specifies that an injured employee cannot sue his/her employer,25 but must pursue preventative practices.
seek compensation in terms of the legal mechanism set up by the Act. As was the case in the UK, three principal defences were pursued. 30 'Ihe Workers' compensation in other countries her contributory own
negligence, negligence he/she defence contributed denied to a the worker occurrence compensation of the accident.if, by his/
The UK. The UK has a long history of employment legislation.% In 1802,27 the second, the fellow servant defence, allowed the employer to show that a fellow first Factory Act No. 42 Geo Ill c 73 was passed,
followed by a number of other employee contributed to the accident; and the third defence, the assumption Acts, which were consolidated in 1878. "lhe Act of 1844 allowed the inspector of risk, enabled
an employer to claim that when an employee accepted to bring an action for damages on behalf of a worker injured by factory employment, he/she also accepted the associated risk of injury.31 machinery.
This was the first step in granting compensation to the injured Clearly, the legal position of employers defending themselves against worker and was the precursor to the Employers' Liability Act No. 43 and
No. claims brought against them by employees was extremely strong. The laws 44 Vict c 42 of 1880. The Workmen's Compensation Act was passed in 1897 narrowing and defining more clearly their liability
— did not achieve much in based on the German system and a further act was passed in 1900, but did not encouraging loss prevention. It was costly for employees to pursue recourse displace the doctrine
of common employment. In 1877 a Select Committee through the legal system, and this encouraged employers to resist claims, of the House of Commons was appointed to consider the subject, but did
rather than to attempt to reduce them through loss prevention. 32 not go so far as to recommend that the doctrine of common employment be In tracing the development of industrial accident prevention
in the US, abolished. Rather, it advised that the employer be liable for the negligence of dissatisfaction with the negligence system acted as a catalyst for the passing of certain people to whom he/she had
delegated his/her rights as 'master' over workers' compensation laws.18 These provided specifically for compensation the workers. These people were called 'vice-masters' in the report. Although to the
injured worker and his/her dependants when injury arose out of or in the Act imposed liability, it was established soon after that the employer may connection with his/her employment. These laws
prompted an increase in the contract out of this liability.28 number of claims and in the amounts claimed, which in turn precipitated the By 1900, the legislation passed by the Conservative Party government
of production of more data regarding accidents and their causes. As a result, the the day had established not only that the employer would be liable for injury development ofindustrial accident prevention
was greatly stimulated.
caused by fellow workers, but also that compensation should be paid for every Today, largely because of the great improvement in plant facilities and the accident arising out of the work, whether caused
by negligence or not. Other fact that the more important causes of industrial accidents have been identified acts were also applicable to accidents, e.g. the Fatal Accidents Act No. 9 and as being human
actions rather than physical circumstances, a greater emphasis No. 10 Vict c 93 of 1845. is placed on safety and remedies that are non-physical in nature. Significantly, Other countries. Germanywas the
first countryto pass workers' compensation there is a tendency to include the aspect of rehabilitation within the scope legislation. In the US, where workers' compensation is a state matter, industrial of
preventative efforts. 'Ihese efforts are regarded as attempts to minimise accident prevention reflected the earlier UK and German efforts. loss through the maximum recovery of physical, emotional and
occupational position in each state must be studied in a depth beyond the scope of potential following injury or occupational disease.
this text and only general remarks can be made here. Accident control efforts It is interesting to note that in the US, rehabilitation centres receive began by passing laws requiring factory inspections
and setting up central support from insurance companies, particularly those involved with workers' bureaux of labour statistics to assist in determining the nature and causes of compensation.34 Ihere
is also evidence that some companies are in the process industrial accidents." Subsequent legislation followed, requiring, for example, of establishing rehabilitation centres for claimants injured in traffic
accidents. the guarding of dangerous moving machinery. More importantly, in time, the
24Vivian (1986-441. 29 et al. (1974.86}.
30 For a detailed discusston on these defencesr consult any textbook on delict or tort In Scoth Africa. the leading works McKerron (1 97)), 80berg (1984}.
26The history that follows is taken from Act Ruegg mentioned Van der Walt (1979). and Neethlng. Potgreter and Vlsser (2006) 31 See 1837 3M &WI.
(1901-241by Pascoe (1974:474) is the 1825 Act. 27 It is 32 et
irueresting to note that the first 357.
163
18
The first of these laws was passed by the US Cmgress in 1903 and applied to ceatain hazardous types of fedaal ernpbyment. By 1921, 42
states had enacted such laws; today, all US states have them. 34 Denenberg et (1974-91)-
RISK MANAGEMENT: MANAGING RISK RESPONSE: RISK CONTROL
Reduction ofthe magnitude ofthe losses

This entails the careful assessment of the magnitude of exposure per se with a view to identifying
and implementing action plans that will reduce the size or quantum of the exposure. From a
practical viewpoint, such action may mean, for example, the purchase of back-up equipment that
would result in diminishing the vulnerability following a breakdown. Another example is the
installation of party-perfect fire walls to divide buildings into separate compartments.
Reduction ofthe frequency ofloss-producing events

The possibility of reducing the frequency of loss-producing events exists independently from
reducing the magnitude of an exposure or loss. This entails a critical analysis of the causes of a
loss, followed by the implementation of plans and procedures to eliminate these causes. For
instance, the frequent cause of transformer failure may be identified as the result of connections
that overheat. Periodic infrared inspections that should detect overheating may reduce the
frequency of losses from this source.
Planning for and recovery from losses

Risk control, in meeting its objectives, is also concerned with planning in order to deal with
loss situations and to recover from losses.
Recovering from loss-producing events (contingency planning)

Contingency planning can be compared closely with emergency planning in that it also adopts
a preconceived plan in the event of a particular type of loss. The objective is to have a contingent
plan of action in place that will ensure recovery from the loss in the shortest time and the most
economical way. Contingency planning thus encompasses both financial and non-financial
aspects of recovery.
An example of contingency planning is identifying alternative suppliers of materials to
deal with the loss of a critical supplier.
Dealing with loss-producing events (emergency planning)

These are preconceived, specific designs or emergency plans implemented when a particular
type ofloss occurs, with a view to reducing the effect of such a loss. Such plans may entail
communicating the potential risk exposure to all those who could be affected. Instruction and
information concerning the potential risk, coupled with an emergency plan of action, will
generally prompt a greater understanding by customers of an organisation and will reduce the
negative impact in the event of a loss. This also achieves the desired integration of the
RISK MANAGEMENT: MANAGING RISK RESPONSE: RISK CONTROL
164 Attention is thus directed at such physical factors as construction design,

materials, apparatus, the supply of amenities, protective devices, machine
maintenance and even the physical factors of adjacent locations (risk exposures).
risk programme with the other areas of operations and with other participants
(often outside the organisation) in the economic equation.
39 Holisrlc "Industrial plans ate formulated as mutual aid plans.
An example of an emergency plan is the drafting of evacuation plans for high- at The acual approach adopted by individual organisations and socieæs differs from time time. early twentieth the beter part of nineteenth of Inquiry, a
the emphas 00 enforcement. As result (he regulatory approach was adopted in the UK. south Africa, more enlightened selfapproach. has
rise buildings, or plans to deal with a major toxic spill or the release of nuclear with the National safety ET-regulatory. always
materials. It is usually necessaryto draft emergency plans in conjunction with RESPONSE: RISK CONTROL 165
other bodies such as civil defence organisations and local authorities. In some
cases, as a result of major industrial disasters (e.g. in Bhopal, India), emergency
Generally, the engineering approach to loss prevention is the first to be considered and
planning is becoming mandatory.
implemented. In most risk situations, many inanimate factors that may cause losses can be eliminated
relatively inexpensively (or their effects reduced). Moreover, these efforts remain effective over a
considerable period of time. The complexity of the loss-prevention approach is obviously a function of the
8.5 Approaches to toss prevention risk that may, at one extreme, require a simple remedy, but, at the other extreme, may require intricate
knowledge of physics, chemistry, electricity, mathematics and so forth. The trend is, nevertheless, toward
Loss prevention generally requires, firstly, the identification of the source (or
more complex and technical prevention techniques in trying to keep abreast of new technology.
cause) of loss and the impact of environmental factors; and, secondly, the
Under the broader definition of the engineering approach, one can also incorporate a relatively
elimination or reduction ofthese effects. In summarising the various approaches
recent approach to prevention, labelled the systems safety approach." Systems safety was primarily
to loss prevention, one should recognise that causes of loss are either animate or
developed and is still used mainly by the US Department of Defense, although its application has also
inanimate in nature, and that it is this distinction that determines, generally, the
been used in attempts to solve social problems.43
approach taken to prevent and control losses.
40
The systems safety approach is used to identify and correct hazards from the early conceptual stage
Denenberg et al. refer to two customary approaches: the engineering
of a product through to its design and eventual operation. Therefore, it envelops the complete ambit of
approach, and the human or personal approach. fie former takes cognisance of the
operations and testing procedures, inspections, even reviews by top-level management, and the
physical or mechanical (inanimate) properties of the risk situation, while the latter
motivation of workers so as to ensure a systematised and integrated approach to safety.
considers the personal (or animate) aspects.
The objective of this particular approach cannot be said to differ from that ofconventional
Approaches such as the educational approach, the statistical approach and
approaches, but it does emphasise the identification ofpossible causes of losses before they occur.
the enforcement approach are also referred to.41 Although different in some ways,
Conventional loss prevention often analyses losses or 'accidents' to identify causes, but understandably,
these approaches may in fact be seen as complementary to the basic engineering
in the sectors in which systems safety application is needed, the cost in property and, indeed, lives lost is
and human approaches to loss prevention.
inestimable. Hence, this approach requires the use of the expensive and most advanced management
skills, mathematics, computer science (simulation) etc. in order to attain the set objectives. It follows,
therefore, that because of the costs involved, systems safety is usually possible only in state or government
Engineering approach
projects; its success has, however, prompted the formation of a number of private enterprises offering
The objective of the engineering approach is to control the physical aspects of the systems safety services to companies in the private sector.
risk situation, i.e. the risk is treated from an engineering perspective. It involves
the use of the physical sciences to prevent loss-producing events from occurring
and to minimise the loss should they occur.
Human or personal approach

The human or personal approach is based on the fact that although losses are caused in part by physical
factors, the most important causes ofloss are human errors, or at least can be traced to the actions of
humans." Consequently,
42
Some authors, e.g. Denenberg et al (1974: l), refer to systerns safetyasa novel approach and creat it as supplementary totheengineering approach. Systems safety practices encompass
most comprehensively the engineering a hu man s&nces — hence the reference to the approach as being 'systernadsed'and integrated.
44 Oeraenb«g et •t al. (1974 (1974:93}.100).

166
prevention of losses entails the study and remedying of people's habits, faults and maladjustments to their personal and work environments.
'Ihe nature of the approach also indicates that it is most practically applicable in the field of industrial accident prevention. It was principally after World War I that in the industrial accident field it became apparent
that physical causes of loss to property were in most instances easily corrected, provided that attention was paid to the actions of people.
Human causes of accidents include complacency, irresponsibility, ignorance, boredom, defiance, timidity, fatigue, misunderstanding, alcoholism, communication barriers and emotional stress. Some of these causes
often do not appear on loss reports and, as has been pointed out,45 remedial steps aimed at avoiding repetition might therefore not go far enough.
The focus of attention on the human element was largely attributed to the research of Heinrich, who developed what has been described as the personnel administration approach.19 Heinrich's research led to three
basic premises:20
(i) Eighty-eight per cent of all accidents were primarily the result of people's unsafe actions.
19
20
Heinrich (1950). heinrkh's research still toms the cornerstone of modern accldent prevention. See Baker and Coetzee (1 go: 168}.
RISK MANAGEMENT: MANAGING RISKS RISK RESPONSE:
R'SK RESPONSE: RISKRISK CONTROL
CONTROL 167
Educational, statistical and enforcement approaches

The educational, statistical and enforcement approaches to risk control were mentioned
earlier. It was stated that the view adopted was that these could logically be seen as
complementary rather than separate strategies.
The increasing importance ofriskmanagement as a science and its applicability in

practice needs no further emphasis — evidence of this claim is plentiful. 218 This trend
obviously demands that a similar advance in education is necessary.
This education should extend beyond the established training and awareness
programmes and general prevention activities. It is recognised that for risk management to
be successfully implemented, there must be an appreciation of the concept, philosophy and
policy that originate from the top level of management. In addition, theoretical work is
presently gaining ground as the subject is gaining recognition as a discipline for study and
research at university level.
The use of statistics, when developed properly, includes the formulation of statistical
techniques for the collection and analysis of facts relating to losses. Although considerable
progress has been made in this respect, unfortunately the effective application of statistics is
often absent, particularly in the private sector. 49
This state of affairs can be traced to a lack of education and, as yet, a less-than-fully
developed policy toward risk management. Again, in many instances, because of the
mechanism of insurance and the misconception that risk transfer is thus effectively achieved,
there is insufficient emphasis on statistics (at the risk source). Rather, the statistics collected
and collated by insurers and their intermediaries are unduly relied on. so
When considering enforcement, one may reason that human nature is fallible and prone
to procrastination, forgetfulness and other such weaknesses. Thus, in order to make
prevention measures effective, it is argued that these should be used carefully and, moreover,
consistently. Naturally, this means strict enforcement through statutes and safety codes.
In some situations where, for example, an employer—employee relationship exists,
such enforcement is achievable by pressure on the employee. Many risk control programmes
incorporate a penalty and reward system that may induce the proper and consistent use of
prevention methods.51 However, in the wider sense of the term, there may be varying degrees
21
S Denenberg et al. {1974- 105).
MANAGEMENT: MANAGING RISKS RISK RESPONSE: CONTROL
(ii) The causes of major injuries were the same as those of minor injuries. Since it was rationalised
that the severity of the injury was in fact a matter of chance or providence, efforts to prevent
accidents should be directed toward the prevention of all causes of loss. of enforcement. At one end of the continuum there could be a deliberate policy laid down by
(iii) Indirect losses from industrial accidents amounted to four times the direct losses, and this top management
phenomenon drew attention to the managerial type of loss control.
48 See for Pretty ( 1999.
49See Oenenberg etal. (1974.533) and Hurley (1 962: 186-8).
The principles subscribed to by the human approach to loss prevention have been significantly 50 te u
s h. significant number of organjsations 'subcontract' responsibility managlng statistics to the insurance market and its particlparl& AS a
extended. Accident cause analysis has been further developed and modified so that efforts are
their managements that have not insuhcient subscribe appreoation delrberate of the direct or Indirect costs resulting from losses. Mostlyr to the this attitude is
channelled to identifring underlying causes (heredity and environment) that could appear in the prevalent among or*nisatlons do to any risk management philosophy or programme, except extent of purchasing *Isurance.
guise of workers' faults. The use of the human sciences (psychology and sociology) has become 51 F
oreyarnple. achievement.'n the NOSA system, resuJt5 are measured against dehned criteria and orga nisations are awarded star [ati ngs dependlag on the degree or
increasingly popular.
Today, in a risk management discourse, we also find references to the humanitarian approach
to prevention. Again, further discussion is not necessary, as it can be seen as an extension of the
basic human approach. In general, the humanitarian approach considers the preservation of life
values as the key objective. The term 'humanics' has been adopted to describe the principal theme
of this approach, which is concerned not only with the prevention of accidents, but also with the
effective reduction of the suffering and expense caused by injury through the use of medical and
rehabilitation services.
168 169
RISK MANAGEMENT: MANAGING RISKS RISK RESPONSE: RISK CONTROL
legal regulation may be necessary. This concern may well not be
intended
constituting its enforcement of risk control, and, at the other as criticism of the enforcement approach to risk control, but
extreme, there means of enforcement. rather that the
could be legislation as its problem is not the interpretation of the various requirements,
but the question
It is therefore evident that the degree ofenforcement increases of how an organisation will ensure that the law is complied
progressively with.
from a motivation to control risk due to a natural aversion to it The view that ought to be followed is that, notwithstanding any
(which is governing
evident in most people), through to subscription to a well-developed legislation, the most effective risk control is achieved when a
and deliberate and
specific risk control programme, insistence on the application of integrated risk management philosophy is entrenched in the
standard codes of practice and risk control requirements prescribed overall policy of
by law.
Turning to the South African context, we will now consider specific
risk control programmes, codes of practice and legal requirements.
8.7 Specific risk control programmes and codes
8.6 Risk control legislation of practice
As mentioned earlier, legislation plays an important part in risk
control. In
South Africa, a great deal of legislation exists prescribing risk
control measures. Safety
The most legislation is indicated in table 8, 1.
Introduction
Table 8.2 indicates the number ofoccupational accidents reported
Table 8.1 Some South African legislation impacting on risk control measures
each year.
Short title of the Act Number Year

Table 8.2 Reported occupational accidents in South Africa, 1996-2007
Occupational Health and Safety Act 85 1993
26 1956
Explosives Act 15 2003 Year Number of reported accidents
Explosives Act T 996/97 159
266
Mines and Works Act (largely repealed) 27 1956 1997/98
Mine Health and Safety Act 29 1996 289 952
46 1999 1998/99 2% 259
Nuclear Energy Act
National Environmental Management Act 107 1998 1999/00
242 126
Environmental Conservation Act 73 1989 2000/01
223 615
AtmOspheric Pollution Act 45 1965 2001/02 280 631
1997
Water Services Act 108
36 1998 2002/03 230 274
National Water Act
Hazardous Substances Act 15 1973 2003/04

237 533
National Building Regulations and Building 103 1977 2004/05
218873
Standards Act 2005/06
237 980
2006/07 213 226
In addition to stipulated risk control measures, each Act also contains Source: Compensation Commissioner's Annual Reports, various years
numerous regulations, all with the purpose of providing the legal
framework
within which risk is to be managed and controlled. Against a background of rapidly changing industrial relations and
the
With so many laws, it is not surprising that some risk management increasing participation by employees and employee organisations
in what
commentators" are concerned when the opinion is expressed that may be termed the 'new' industrial society, greater awareness and
further planning in the area of occupational safety and health are
required.
52 Friedman and Fnedman (1980 22749).
170
Statutory requirements
Apart from humanitarian, moral and social responsibilities on the part of management to provide An accident is an undesired event that results in physical harm to a person and/ or damage
a safe working environment for employees, there are onerous statutory duties in this regard. In
to property. It is usually the result of a fortuitous contact with a source of energy above the
South Africa, statutory intervention in working conditions is regulated mainly by the legislation
threshold limit of the body or structure.
indicated in table 8.1. Among other things these Acts provide for the registration and control of
factories and mines. also make provision for the supervision of the safe use of machinery and
Measurement
precautions to be taken against accidents and occupational diseases.
Disabling injury frequency rate (DIFR). The disabling injury frequency rate relates the injuries
Should a company fail to comply with statutory requirements, severe penalties can be
to the hours worked during a specific period and expresses them in terms of a million-hour unit
imposed. It is therefore essential that a risk manager, or whoever is responsible for safety (e.g.
by means of the following formula:
the factory engineer), should keep abreast of the changing legislation in safety matters, familiarise
him-/herself with legislation covering the organisation's activities and adopt a systematic
DIFR = Number of disablin in•uries x 1 000 000
approach to ensure that legislative requirements are complied with.
Total number ofperson-hours worked
The above formula can be modified to show the approximate percentage of the workforce injured annually,
Responsibility
the so-called incidence rate (IR).
Responsibility for safe working conditions and practices extends from top management down to
the shop floor. It is not suffcient for management merely to supply safety equipment and clothing; IR = Disablin in•uries x 200 000
it has to ensure that employees use this equipment. Similarly, safe working practices must be Total number of person-hours wor ed
enforced and employees must be clearly warned of the penalties that will be imposed if they fail
to comply with safety instructions. IR DIFR
5
Safety programmes and organisations Disabling injury severity rate (DISR). The disabling injury severity rate relates the days charged
Three main safety programmes were developed to manage safety and occupational health risks to the hours worked during the period and expresses them in terms of a million-hour unit by
and, as such, supplement the legislation: means of the following formula:
(i) the Mine Safety Management System (MSMS) of the Chamber of Mines
(ii) the International Mine Safety Rating (ISR) DISR = Total da s char ed x 1 000 000
Total number o person-hours worke
(iii)the National Occupational Safety Association (NOSA) programme. The basic principles of accident prevention
The unsafe acts ofpeople are responsible for the majority ofaccidents. Research has found that out
While the MSMS was designed specifically for the mining industry, the NOSA programme of every 100 accidents:
found wide application in commerce, industry and the mining industry in general.
54 {1937; 7).
National Occupational Safety Association (NOSA)53

NOSA was established in 1951. It was modelled on the pattern of the Industrial Accident
Prevention Association of Ontario, Canada. Historically, the bulk Of the funds for NOSA came
from the Accident Fund, but increasingly it had
53 For a discuss•on the NOSA system, consult Matttysen (1983: 1 64} and the commemorative edition Of SafetyManagement (June 1991}-
RISK 171
to find its own funds. For various reasons, it was unsuccessful in this and, unfortunately, increasingly it
got into financial diffculties.
South Africa achieved a high degree of success in preventing occupational injuries. The
programme also resulted in a reduction in damage to property and disruption of business
activities.54 The modus operandi for achieving this success is the management-by-objectives
accident prevention programme training courses, publicity material and safety promotional
activities.
Accident prevention: Definitions and concepts
Definition
172 to employed persons.56 Its operation is excluded where the Mine Health and Safety Act No. 29 of 1996
and the Explosives Act No. 15 of 2003 apply.

88% are caused by unsafe human acts The Act mainly allows the relevant minister to set safety standards through the
10% are caused by unsafe environmental and mechanical conditions promulgation of regulations. The Act also identifies certain specific duties of employers, users
• 2% are caused by acts of providence.
of machinery, safety representatives and inspectors.
From the above, the conclusion can be drawn that 98% of all industrial accidents can be
• Safety standards
prevented.
An innovation of the OHSACT is the ability to use legislation to promulgate safety standards."
The four basic reasons for the occurrence of unsafe acts provide a guide to the selection of
Regulations promulgated under the Factories, Machinery and Building Work Act of 1941 were
appropriate corrective measures. Briefly, the reasons for the unsafe acts of people may be: (i)
initially deemed to have been promulgated under the OHSACT and remain in force until they
the wrong attitude
are repealed.5*
(ii) lack of knowledge or skill
(iii)physical unsuitability
Exemptions
(iv) an improper mechanical or work environment. The relevant minister may grant exemptions from all or any of the provisions of the Act and
regulations or from any prescribed safety standard.

"Ihe following are the basic methods available for preventing accidents: e
engineering revision education, training, persuasion and appeal personnel selection,
Safety organisation
placement and adjustment
In terms ofthe Act, all operations falling under the jurisdiction ofthe OHSACT are required to
• enforcement (discipline).
have certain people responsible for various aspects of safety at the workplace.
The most valuable methods of accident prevention are similar to the methods required for the
• People responsible for safety
control of the quality, cost and quantity of production. In many cases, the same faulty practice The OHSACT imposes substantial responsibilities on employers and users of machinery:
is involved, and the reason for the existence of the fault is similar for both accident occurrence
The employer has the overall duty to ensure that much of the safety legislation is complied
and unsatisfactory production.
The employer and user ofmachinery, usually a company, is legally responsible for with. Where the employer is a corporate body, the chief executive oiücer (CEO) of the
supplying a safe environment, tools and safe methods of working, as laid down in Occupational body will have overall responsibility.22Corporate CEOs and all other employers who are
Health and Safety Act No. 85 of 1993 and the Mine Health and Safety Act No. 29 of 1996. not personally involved in the day-to-day running of the operation shall designate in writing
management of the company is legally obliged to carry out this responsibility. the senior full-time manager or person at the operation as having that responsibility. In turn,
senior managers may delegate some or all of their safety duties to employees under their
Occupational Health and Safety Act (OHSACT)
authority. Ihe delegation of duties does not relieve an employer of any responsibility or any
Of greater importance to risk control than the Act itself are the specific regulations promulgated
liability under the OHSACT. In brief, the requirement of the Department of Manpower is
or deemedS5 to be promulgated in terms of the OHSACT.
that the managing director of a company shall be designated by a written resolution of the
Application of the OHSACT board of directors in terms of regulation 4(1) GAR. He/
This is the principal act for the protection of employees in the workplace, other than in mines.
The Machinery and Occupational Safety Act No. 6 of 1983 replaced the long-standing
Factories, Machinery and Building Works Act of No. 22 of 1941, and in turn was replaced by
the Occupational Health and Safety Act NO. 85 of 1993. Many of its provisions are universal 57 SS See. 10 particular. for example, the machinelY Incorporation provisions, of Safety which Standards are not — Electrlcal confined (GN2271to the occupational '85)
Situation.
in application and are not confined
58 Extensive use has been made of (he power to incorporate safety
35 Regulations are in terms oi an exlsdng When an Act is replaced, The then existing regulations are usually deemed be promulgatedin termsOf new Act
173
22
oromutgatedSec. I Of the now repealed Machinery and Occupational safety Amendment Act NO. 40 of 1989. The date of commencernent Is still to be
RISK RESPONSE: RISK
RISKS RISK CONTROL
174 The figures for occupational diseases include items such as hearing loss, which in the ordinary
sense of the word would not be regarded as a disease.
she may then, in terms of regulation 4(4) GAR, delegate these duties to the manager(s) in It is not always possible to maintain a close distinction between occupational health and
charge of operating sites" occupational safety. Some authors have expressed the view that there should be no such
A user of machinery is defined in the OHSACT, and it should be noted the definition does distinction.61
not restrict the scope to the employment situation. Occupational health can be subdivided into the following disciplines:62
With the promulgation of the OHSACT, the emphasis moved to selfregulation. One of the • occupational safety
measures of self-regulation is the appointment of safety representatives. The safety • occupational hygiene
representative is absolved of any civil liability for failing to do anything he/she may be • occupational medicine.
required to do in terms of the OHSACT.
Historical background63
A further innovation is the appointment of safety committees. Members of the committee are
absolved of civil liability. Occupational health, in its narrow sense of occupational safety, was initially concerned with
the prevention of accidents, particularly personal injury arising from mechanisation and the
Risk control organisation for safety
use of new and powerful forms of energy such as steam and electricity. As the industrial safety
Where the size of the operation warrants a risk control organisations it is recommended that the
movement expanded, it became evident that certain job-related environmental conditions such
safety organisation be included in the risk control structure. This department should
as dust, fumes and chemicals have a detrimental effect on the health and wellbeing of the
systematically implement and audit the safety systems.
people exposed to them. It therefore became increasingly clear that it was necessary to monitor
and treat the environment and the people in the environment.
Occupational health
The legislative emphasis was also put on the prevention of accidents, although many of
the regulations promulgated in terms of the Factories Act dealt with occupational health
Definition issues.
Occupational health is concerned with any substance, process or circumstance that may affect Concern about the position of occupational health issues resulted in the appointment of
a person's health or safety in the workplace.
the Erasmus Commission of Enquiry on Occupational Health (RP55/1976). The commission
considered its terms of reference to relate specifically to occupational diseases only, but clearly
Table 8.3 Reported cases of occupational diseases
its findings could have an indirect effect on the field ofindustrial accidents.64 The commission
Year N umber of cases of diseases noted that the gold mining industry had little to be ashamed of. The commission made a
2000/01 3 361 number of recommendations, including the establishment of a single body to control industrial
2001/02 4 689 health, which should be vested in the Department of Health, and that one principal industrial
2002/03 5018 health act should be promulgated.
2003/04 5 358 The recommendations were not immediately implemented because of, amongother
2004/05 3 822
things, the appointment ofthe Wiehahn Com mission .65 Eventually, the Machinery and
2005/06 4 564
Occupational Safety Act No. 6 of 1983 was promulgated, the ambit ofwhich is broad enough
2006/07 3720
to include occupational health and safety, and the emphasis shifted away from the Erasmus
Source: Compensation Commissioner's Annual Reports, various years Commission."
175
61 See, example. Benatarr Metz and Elmes (1 gag: 952-3).

63 62 Piek For a {1986, discussion 1 1 9) states on some that hlstorica) patlanöl issues medicine see Baker is (1986.5)-one oftwo legs of occupational health, the other bang occupa
hygiene'.
To determine the environmental factors or stresses that influence health and well-
64 Para- 2.38 of the reporr_
being, one has to be familiar with the processes and work procedures involved. The main
65 Para, 24 part 4 of the WehGhn Report environmental factors are as follows:
60 For an opanion on regulation 4. see Swart (1990},
66 Conslderabk discusslon did hovveverr contlnue. See. for example, Metz and Benater (1987; 72) and Baker (1986: T 03)-
176 MANAGEMENT'. MANAGING
chemical: fluid, dust, smoke, mist, fumes, fog, vapour, or gas and hazardous
substances68
67 Piek (tm. 119}.

A further commission, the Niewenhuizen Commission of Inquiry into Compensation for 68 Commonsubstances include lead and mercury.
Occupational Diseases in the Republic of South Africa (RP100/1981), was appointed. A 177
number of bills were promulgated, including an Occupational Medicine Bill (Notice 20/1984;
GG9029 1984) and a Draft Compensation for Occupational Diseases Bill.
physical: electromagnetic and ionising radiation, noise,69 vibration and extreme
temperatures and pressures, and non-ionising radiation such as that produced by welding
Occupational safety
and video display units70 biological: insects, mites, fungi, moulds, bacteria and viruses
(resulting from bad handling of wastes sewage and food, and lack of personal hygiene)
Definition
ergonomical:71 body attitude and work position, continuous repetitious movement, high work
Occupational safety covers physical safety from an accident prevention perspective.
demands and mental fatigue.
This constitutes the major thrust of the well-known NOSA programme, and broad subjects
The occupational hygienist recognises the above factors as threats to health and as the cause of
include housekeeping; safe working practices; mechanical, electrical and personal
stress and significant discomfort.
safeguarding; safety organisation; and administrative and legal requirements. Note that in the
Machinery and Occupational Safety Act 'safe' means free from any threat that may cause
Occupational medicine
bodily injury, illness or death, and 'safety' has a corresponding meaning (sec. I(l)(xxiii)).
Occupational hygiene Definition
Occupational hygiene is sometimes called industrial hygiene or environmental health, and Whereas occupational hygiene concerns the monitoring and treatment of the workplace
according to Piek,67 deals with the identification, evaluation and control of potential hazards environment, occupational medicine, in general terms, concerns the monitoring and
treatment of people in that environment.
in the workplace.
Definition In this case, knowledge of a person, ofsubstances as well as and methods work processes to eliminate
Occupational hygiene may be defined as a science devoted to the identification, that could or reduce adverselysuch
evaluation and control of those environmental factors arising in or from the workplace affect the health hazards, is
that cause disease, damage health, or impair the faculties of workers and other members required.
of the community.
The cause/source of an accident is usually clear, but this is not the case with occupational
Products liability
diseases. Environmental issues are therefore ofgreater importance where occupational Another risk control programme is the products liability risk control programme. It should be integrated
health is concerned. with quality assurance programmes.
RISK RESPONSE: RISK
RISKS RISK CONTROL
Products can be the source of extensive liability claims. One of the wellknown examples The approach adopted in this following section is to illustrate the various forms of products
is the Thalidomide disaster of the 1960s, where the drug caused babies to be born with Liability by referring to practical examples.
deformities. In the UK, 452 children suffered brain damage as a result of the drug. It took a
long time to resolve the issue of compensation and the Thalidomide Trust was eventually 69 A measure is aud,omerric sqeen,ng. For a comment on its limitations. see Metz {1987; 35}.
issue ofdangers associated video d,splay units is Even though billions have been spent on research. there
established in the UK. The manufacturer and the government contributed towards the funding
71 Bridges and Taros (1 986:39. 72
of this Trust." Following this disaster, there has been a dramatic escalation in product liability Wallace (1982).
claims in various parts ofthe world.

RISK RESPONSE: RISK CONTROL
180 RISK MANAGEMENT: MANAGING RISKS 79 time Product or liability surveys carried out of numerous Incorrect companies despatch by is a common the etrcghave indicated that Virtually every company
has anothef, Incorrectly rkspatc.hed modUCtS.
81 The of liability have been researched in s,wth Africa. see De (1979: 81: 1980, i v, 83, 234).
18T
Grant v Australian Knitting Mills Ltd77

In this case, a large number of garments had been made without mishap, with the exception
The following are examples:
of a pair of underpants that had not been properly rinsed. This caused the plaintiff to
A surgeon sued the car- and tyre-manufacturing firms for RI million in
experience a serious skin disorder. The manufacturer produced evidence of high-quality
control standards employed at the factory. The court rejected that this proved lack of loss of earnings arising from an accident that occurred when one of the rear tyres of his vehicle
negligence, ruling that the mere fact that the garment was put on the market place was
burst.82
evidence of negligence. The court evoked the so-called res ipsa loquitur doctrine.'8
An animal feed manufacturer was successfully sued when a number of racehorses died because
their feed contained a poisonous substance'

This case suggests that if a product is admittedly defectives the manufacturer has little
defence. In addition to court cases, other claims are reported in the media. There have been numerous
incidents that could have given rise to product liability claims, and also incidents that indicated
Liability for incorrect despatch
the concern of commerce and industry for possible product liability claims. For example, a major
It is possible for the despatch department of a company to despatch the wrong product.79 The supermarket chain withdrew a teat used on babies' bottles and pacifiers from the market
receiver does not detect the error and uses the product, with detrimental effect. because of the danger that the teat might disintegrate in a baby's mouth. Rooibos tea was also
withdrawn from the market for a period for alleged health reasons. Furthermore, motor
Marais v Commercial General Agency Lte
manufacturing companies abroad have been sued for defective design after their vehicles were
A seed merchant inadvertently supplied a farmer with seeds of a character different from
involved in accidents and subsequently burst into flames, killing the occupants."
those supposedly purchased and was held liable for the resultant loss of the value of the
crop. The decision turned on breach of contract.
Risk control steps
'Ille court, however, said that the case may also have succeeded on other grounds, such as Existing quality assurance or control programmes
negligence. The objective ofa product liability risk control programme is the prevention of product liability
claims. If the product is carefully designed, manufactured and distributed, the frequency and
Further examples of products liability claims severity of losses should decrease dramatically. A product liability risk control programme should
Most people are under the impression that South Africa has not had many products liability be integrated into existing programmes that are designed to ensure that product quality is of a
claims.81 This is not entirely correct. South African law has historically treated products high standard. These quality assurance and control programmes are already well established.
liability claims as normal delictual or contractual claims and dealt with them accordingly. Product design, manufacture and distribution affect virtually every department in a
The majority of these claims are dealt with as insurance claims and have been settled company, and therefore one person should be responsible for product quality or the
out of court. It is only in exceptional instances that the cases do go to court. Notwithstanding product risk control programme. For example, if the raw materials are defective, the final
this, in recent years a number of product liability cases have been mentioned in the media product could well be defective; this is a supplier problem. So, too, if the manufacturing dye
and some have gone to court. used in the manufacturing process is defective, the product could be defective; this is a
manufacturing problem. If the advertising material incorrectly describes the product, a
products liability claim could arise; this is a marketing problem.
In the one case, the head of stores is involved, in the other the manufacturing department, and in the
third the marketing department. When designing
78The application of this doctrine has long been criticised as imposing negl igence where none in fact exists.
MANAGEMENT'. MANAGING RISKS RtSK RESPONSE,'
CONTROL
82 The Stars July 24 July 1987. REK 183
83TheCilizen, 15 February 1985
84 Tne most famous of which is the VS caseGrimshaw v FordMot* Ca

182 RISK
Contract review. When a new product is placed on the market, the product liability
aspects must be taken into consideration. Requirements and planning for these aspects
a product's liability risk control programme, it must be ensured that such a programme covers should not be neglected, as the oversight could prove to be costly indeed.
all the departments involved, that the system is easily understood and that all employees in the
organisation are familiar with the programme. A manufacturer produced a very inexpensive item. After making tens of thousands of
these items, the die used in the manufacturing process became worn and had to be
Product liability programmes and quality assurance codes of practice The point of departure replaced. However, the size of the new die was incorrect and, as a result, the products
being manufactured were slightly too big. The product was put into operation and the
with a products liability programme is the implementation of recognised quality assurance codes
loosely fitting part resulted jn a multimillion rand consequential loss claim. The claim could
of practice. When designing a product liability risk control programme, the risk manager should have been prevented simply by using testing equipment (which would have cost R25 at
work closely with the quality assurance manager and the quality assurance programme. In most) that would have detected that the new die was faulty. The onfy reason the
accordance with the general philosophy of risk management, the risk manager plays a necessary testing instrument was not purchased was that when the company received the
order and designed the product, nobody determined that this quality test was needed to
consultative role and can therefore never be responsible for the implementation of the quality
establish whether the product quality was suitable.
assurance programme. In addition to implementing the elements of these codes of practice,
further aspects may not be covered in suffcient detail, even though they are important from a
A detailed quality plan must therefore be drawn up for every product manufactured.
product liability point of view and should therefore be implemented, namely a product recall
Design control. The most critical aspect of any consumer product from a liability point
programme and a programme to deal with the issue of liability for mere words. It is necessary to
of view is its design. If the design is incorrect, all products manufactured from the design will
ensure that the product liability programme covers these additional elements.
be incorrect. In fact, the quality assurance process will ensure that all products conform to
the faulty design.
Elements of quality assurance codes
Management responsibility. To minimise the risk of product liability claims, a holistic quality
The case of the BC disposable lighter can illustrate this problem. Approximately 250
assurance programme must be implemented. Senior management therefore has to ensure that million lighters per annum are sold in the US alone. If the design is defective, it can result
such a programme exists, and that responsibility has been assigned to one person in the in millions of claims. tn the US, a woman was paid R65 million by BIC as a result of a lighter
organisation for the programme and to various other people for individual sections of the that ignited in her pocket, causing burns to her body.85 One small design error can
therefore result in a company's incurring sufficient claims to make it insolvent.
programme
The quality system. It is essential that a properly defined quality assurance organisation be
The design control programme should include a design review, field testing, reliability testing,
established within the company and that various persons be assigned the responsibility for the
incorporation of safety codes and standards, and so forth.
prevention of products liability claims. One person, appointed by the general manager of the
Document control. The quality assurance programme is essentially a programme
site, must be given the overall responsibility for the programme. This person must ensure that
defined by means of documents. It is accordingly very important that these documents be
the programme is kept up to date. Furthermore, no documentation pertaining to the
kept up to date, and that no changes are made to the documents, except by those duly
programme should be altered without his/her specific consent.
authorised and clearly designated to perform the task. The system should also be designed
It is not suffcient simply to implement a system; provision should be made for the system to be in such a way that it is obvious from the document itself whether it is up to date.
reviewed regularly by internal and external auditors. Two aspects in particular should be
reviewed, i.e. whether the regulations of the system are being complied with and whether the 14 April 1982
system itself is adequate. In practice, it is found that while compliance is audited, the system 184
itself is often not revised.
RISK MANAGEMENT: MANAGING RISKS RISK RESPONSE.' RISK CONTROL
One ofthe largest industrial disasters was caused at Flixborough when alterations were day, line and batch were involved. Since the so-called defective feed was also supplied to
carried out to the piping arrangements of a chemical plant. The temporary supporting other customers on the same day who did not suffer any ill effects, this was a good
structure collapsed, resulting in the release of an explosive vapour. This ignited, causing an indication that the feed was not defective. Indeed, when the relevant batching instructions
unconfined vapour cloud explosion. were examined, it became clear that the product was indeed not defective.
Installation, manufacturing, repairs and alterations should all be carried out in accordance with Process control. The steps required during the manufacturing process can be defined using work
the standards set out in the approved documentation. Strict control over the documentation and instructions.
over changes to the documentation is required. However, documentation can also be a two- The product liability programme functions in terms of work instructions, which indicate
edged 86 sword and could be used against the person who prepared the document. the steps required to manufacture the product according to the design specification. Therefore, a
Purchasing. If the inputs, materials and services are defective, the final product could be section dealing with manufacturing control should be established. In practice, the manufacturing
defective. All incoming materials and services must therefore be carefully inspected. In order controls and procedures are probably the most widely used and understood aspect of any system.
to carry out such an inspection, specifications must be drawn up against which the incoming Although many companies do not have a defined quality system, they do operate according to
materials can be checked. specified manufacturing instructions.
The importance of inspecting raw materials cannot be overemphasised. All raw materials Once the work instructions have been determined, the manufacturing process must be
must be inspected by an approved statistical sampling programme. The integrity and identity of programmed for implementation. Normally this will consist of defining the steps one should take
the materials should be confirmed and recorded, to ensure that the product is manufactured correctly. This plan is put into action by drawing up
specific instructions for various employees on the procedure of manufacturing the product. It
The Kinross mining disaster is an example of what can happen if the incorrect material is therefore concerns the issuing of work instructions. The work instructions should be
supplied. In the Kinross case, the mine contracted for the supply and installation of a comprehensive and cover all aspects relating not only to the product, but also to safety.
substance called polyisocyanurate.v Although the mine had ordered polyisocyanurate, it For example, assume that a work instruction has been issued for the inspection of raw
was supplied with polyurethane, with disastrous results (see table 1 .1 materials. The work instruction should include the method required to safely inspect the raw
materials. If this is a dangerous chemical, the work instruction should also include safety
It is therefore essential that at the very least the identity of products is confirmed and that all instructions, such as that protective clothing must be worn.
harmful and dangerous products, as well as all safety critical products, are carefully inspected. The specific work instructions should integrate the requirements of other systems and the
Purchaser-supplied products. In practice, the purchaser may often supply some of the statutory requirements. Thus the NOSA MBO requirement for protective clothing should be
materials for the manufacture of a product him-/herself. In this case, the quality assurance incorporated into the document, as should the requirements in terms of the Occupational Health
programme should take this factor into consideration when determining the final standards and and Safety Act. In this way, all the different requirements can be integrated into one document,
integrity of the product. Purchasersupplied materials should be subject to the same controls as i.e. the work instructions.
any other material. However, liability for defects in the final product as a result ofpurchaser- Inspection and testing. A well-designed quality system incorporates inspection and testing
supplied materials should be determined and agreed upon beforehand. requirements and should define which inspections and tests must be carried out to ensure
Product identification and traceability. If a product is defective, remedial action must be product integrity. These tests may be expensive to carry out and add to the cost of the product.
taken as soon as possible to minimise the extent of the damage. identification and traceability Unfortunately, in recent years, cases have come to light where the defined tests were not
may be necessary to prove whether a defect is involved. Therefore, it should be possible to
identify and trace products to the correct manufacturer.
87 86 Vivian Unpublished {1990; 10).judgment 1045 Of the judgment handed down by MaggtrorevR PJeruson.
185
In one case, it was alteged that a certain product was defective and had caused the death
of certain pets. Fortunately, the bag that contained the alleged defective feed was found
and returned to the manufacturer. From the label, the manufacturer could establish which
186 187
Once non -conforming products/material have been identified, the programme should specify
carried out and the test documents were falsified.88 Not only is this obviously fraudulent, but also
which person has authority to deal with the material and also record what action was taken.
dangerous.
The decision maker should not be a member of the production department.
Inspecting, measuring and testing equipment. In most cases, the product is subjected to a
The decision to declare material as non-conforming could be very expensive. It is
number of inspections and tests during the manufacturing process. These testing stages are
therefore not inconceivable that a production manager may decide to "take a chance' and
conducted with the aid of testing equipment. The equipment could be very simple, ranging from a
release non-conforming material onto the market in the hope that there will be no returns.
simple spring balance or a go-no-go gauge to multimillion rand electronic equipment. If the
For example, it was reported that drugs nearly at the end of their shelf life were being sold to
equipment is defective or incorrectly calibrated, the tests will be misleading. Therefore, a clearly
the public. Rather than the drugs being destroyed, they were sold at a discount. 89
defined programme is needed to ensure that all testing and measuring instruments are accurate. The
Corrective action. lhe programme should specify that corrective action be taken and
programme must also define the right type of instrument and correct margins of error for each
which person has the authority to take such corrective action should the product be defective
instrument.
or something go wrong with the manufacturing process.
It appears that a faulty instrument caused one of the great nuclear disasters, that at Three Mile The person deciding on the corrective action must not be a member of the
Island. The importance of correct instrumentation and checking that the instrumentation is accurate manufacturing staff, but should fall under the control of the quality assurance staff. Thus, if
cannot be overemphasised. In some instances, the specifications are laid down by law. a product is defective, somebody must decide whether the degree of defectiveness is so
Inspection and test status. As has already been indicated, a significant portion of the quality severe that the product should be abandoned, which could cause extensive losses, or whether
assurance programme is devoted to inspections and tests of one kind or another. 'Ihese start with the production process should be continued.
the receipt of the raw materials and continue through the different stages of the manufacturing Handling, storage, packaging and delivery. Once a product has been manufactured, the
process. There must be written evidence showing that the material has been inspected and whether code should specify procedures to ensure that the integrity of the product remains intact until
or not it passed the inspection. The test status should be indicated. it is in the hands of the consumer.
If a product is found to be defective, this should be indicated and the product isolated. Once This is particularly important in industries such as the pharmaceutical industry, where
identified as defective, the suspect product should be dealt with as defined in the control of non- the expiry date of the product must be shown on the product itself, so when the consumer
conforming material specification of the code of practice. buys it, he/she can ensure that it is still safe and effective. If the expiry date is not indicated,
the product may be manufactured correctly, but because its shelf-life has expired, it may have
In one case, defective products were withdrawn from the market, but since the test status was lost its effectiveness. This type of claim does not arise from the manufacturing process itself,
but from preservation aspects of the product.
not indicated and the defective products were stored near the production line, the products were
Labelling, too, is an important aspect of product liability. Labels must be accurate and
accidentally released back onto the market.
disclose suffcient information to warn against any dangers associated with the product.
Quality records. A number of documents are generated by the quality assurance system.
Control of non-conforming products. Specific provision should be made in the quality assurance The different records should be defined and controlled in terms of the requirements relating
programme to control materials that do not comply with specifications. to product records. The records should be comprehensive, starting at the raw material stage
and continuing to include work instructions, all quality inspections and tests, and, finally, the
A defective product was detected, but there were no adequate procedures for its control. The despatch and sales documents. It should therefore be possible to trace the total history of the
unmarked defective product was stored next to the production line and placed with the product from its records.
marketable products. Any quality assurance programme should therefore specify how non-
89 5hestar.14 1987.
conforming products are to be controlled.
188
88 New Scientist, 20 September 1 984; The Srar. 12 January 1989.

RISK MANAGEMENT: MANAGING RISKS RISK RESPONSE.' RISK CONTROL
The quality assurance programme should also indicate the period for which records on legal cases that gave rise to product liability claims, a product liability claim could be made
should be kept and the methods available for locating the various records. even though the product was not defective. A liability claim could arise, for example, if the
Internal quality audits. The system must make provision for an audit of compliance with consumer was misled by the company concerning the purpose and effectiveness of the product
the requirements of the system. These internal audits should be conducted on a regular basis. Therefore, the product liability risk control programme should also include a programme aimed
Training. Because a quality assurance programme is comprehensive and complicated, it at training employees (as well as other techniques) to ensure that liability does not attach as a
normally takes three to four years to implement such a programme from the day the decision result of words.
is made to its final implementation. Thus, an adequate training programme must form an
integral part of the programme, so that employees are adequately trained to operate the system. In a case, a company representative was asked If the product would be suitable for purpose
A. He assured the buyer that it was. After the product was manufactured, it was found to be
Services
unsuitable for this purpose. The company was held liable.
A large section of commerce and industry is devoted to the provision of services, not products.
Special procedures should be drawn up to minimise liability claims arising from the provision
of services. These procedures will obviously differ, depending on the type of service provided. Sales representatives should therefore be familiar with the consequences of using incorrect words
and promises concerning the product. Sales and technical literature should also be thoroughly
In one case, a contractor supplied transportation services. He delivered products to checked as part of the quality assurance programme to ensure that it contains the correct
various manufacturers, which were pumped into storage tanks. One day he discharged
product A into the storage tank containing product B, instead of discharging it into tank information.
A.
Liability for gradually developing environmental impairment

Statistical techniques. Not every incoming item needs to be inspected — particularly if these
are not safety critical - as this can be expensive. In such a case, a sampling plan can be Codes of practice are also of assistance in managing environmental risks.
developed using statistical techniques. Sampling then takes place in accordance with the plan
and deviations from the sample can be used as an indication of the integrity of the entire batch Environmental concerns
of raw material. The quality assurance programme should therefore define the sampling plan One of the recent exceptions to a general liability insurance policy is liability attributable to the
in terms of which product will be inspected. gradual or ongoing release of pollutants. 90 It is therefore advisable to treat gradual pollution
separately from pollution caused by sudden and accidental events.
Product recall programmes
The types of cost involved in environmental impairment are also unusual. Generally, it
The product recall programme generally does not form part of the quality assurance
concerns not the damage caused* but the rehabilitation of the environment. lhese costs could,
programme, although it does form part of a programme such as the good manufacturing
however, run into billions of rand.
practices of the pharmaceutical industry. This is an area of particular concern and interest to
the risk manager, and in product liability literature this is often the main or only element ofthe In the US, the consumer movement, which gave rise to the products liability crisis, was
quality assurance programme dealt with in the product recall programme. ushered in by Ralph Nader's Unsafe at Any Speed. The other well-known consumer movement
Since this aspect is normally part of the quality assurance programme, the risk manager should is the environmental movement, which was ushered in with Rachel Carson's book Silent Spring
ensure that there is adequate provision for the recall of defective products. (1962). Since the 1960s,
189
(1989. 14}.
190 RISK
Liability for words
The quality assurance programme is generally aimed at ensuring that the product is manufactured there has been increasing concern for the environment world-wide, and South Africa is no
according to the specifications. If it meets the specifications, the quality assurance manager will exception. The public is quick to speak out and turn its opinion against those companies who
be satisfied that his/her programme is functioning adequately. As was seen from the discussion are perceived to pollute and damage the environment.
South African case law The Rylands v Fletcher (1868) rule. Many industrial sites store a wide range ofdangerous
Liability for impairment recognised in principle substances, and even water could under certain circumstances be classified as a dangerous
In South Africa, very few actions for damages arising from gradually developing substance.
environmental impairment have been instituted. In fact, until fairly recently, very little had It is quite possible that a dangerous substance, including water, escaping from a premise
even been written on the subject.91
can cause damage. In such cases, the South African courts could possibly invoke the sc»called
However, our courts would not approach the claims for environmental impairment as
'Rylands v Fletcher rule'. This rule imposes strict liability upon occupiers of dangerous property
something different or unique, simply because it was an action involving the environment.
Ihey would simply apply the general principles of delict, which were discussed earlier. and states that:
a person who for his own purposes brings on his lands and collects and keeps there
An early case in which the generat principles were applied to an environmental problem
anything likely to do mischief if it escapes, must keep it in at his peril, and, if he does
was Regal vAfrican Superstate (Pty) Ltd. In this leading case, in which liability for an
omission was recognised, a previous owner (a company) of a piece of la nd had been not do so, is prima facie answerable for all the damage which is the natural
liquidated. The company had used the land for quarrying purposes and left slate waste in consequence of its escape.94
a position where the floodwaters of the Elands River could reach it. Under these
circumstances, the floodwaters would wash the slate downstream and cause harm to a If this rule is applied, negligence is not required and it need only be shown that the
neighbour's land. The neighbour sought a court interdict against the previous owner, but dangerous substance that escaped caused the damage or loss. It is generally accepted that the
since the owner had been liquidated, the action came to nothing. When a new owner
Rylands v Fletcher rule does not form part of our law, but in view of the changes to our
bought the property, the neighbour instituted an action against this new owner. Clearly,
the new owner had not created the nuisance or danger that the slate would wash common law in recent years, the court may, given the opportunity, indeed decide to make this
downstream, as it was already in existence when he bought the property. lhe action failed rule applicable Liability may also attach because of: solid waste disposal groundwater
on these facts, but the court held that:92 'if it was reasonably practicable to avert the still contamination.
threatened damage then the failure to do so would be unlawfful, and then the appellant
would have a basis for a petition for an interdict and possibly also for a claim for
compensation for damage which he might suffer' and 'the only acceptable basis of liability • A number of vegetable farmers in Tala Valley and in other parts of Natal issued summons to obtain
was a failure to take reasonably practicable steps to prevent the situation complained oft a court interdict against manufacturers of agricultural chemicals, on the basis that the chemicals
(1963 (1) SA 102 (A)). that they were using contained a substance known as 2.4D herbicide, which could conceivably
damage crops.% The Witbank Town Council pumped water into an underground mine a number of
The court thus recognised that a claim for compensation could succeed93 if the actions of years ago, apparently to put out an underground fire. As a result, it is alleged that the properties in
one person caused damage to another as a result of environmental impairment. the area are beginning to crumble.% An action for R 100 million damages was instituted.
A Johannesburg-based property developer was to sue the Palaborwa Mining Company
91 Rabie {1 and Rabi* 983},

after the developer's stud cattle died from copper poisoning. The developer alleged that
copper dust settled on the grass or fodder in the area and that the cattle died as a result
93 The court also recogni*d that liability could attach for omissms. This paint was developed in later Judg«nents and culmlnated in the EweKC*
191 of eating this. The farm was situated some 15 kilometres downwind from the mine,"
illustrating how wide the affected area of danger could be.
It should be clear from these quotations and the developments that have taken place in 94 (1971:246}.
% 14 December 1987% 22 June 1989.

South African common law since 1963 that if a person suffers damage that arises from 97 14 August 1989.
gradually developing environmental impairment, he/ she may very well succeed with an action
at law.
RISK MANAGEMENT.• MANAGING RISKS RISK RESPONSE: RISK CONTROL
192 People on neighbouring land may be affected by the activities of an industrial company
and by various impairment agents. These need not be dangerous or hazardous, but may simply
be a nuisance or cause discomfort, as in the case of noise and odours. While neither of these
A municipal dump has been in existence for a number of years outside Linbro Park and two agents poses any particular health hazard or health threat, they are causes for concern and
water from the refuse dump has seeped into the water table, contaminating the nearby so an audit should determine whether any problems are known to have arisen from these
water boreholes? This is an example of groundwater pollution and illustrates the question particular sources.
of environmental harmony. The dump was established long before the residential area, Another way in which the surrounding areas can be affected is by the disharmony caused
but as people moved into the area, they became dissatisfied with the presence of the
by the activities of an industrial site. It often happens that people move into the area long after
refuse dump and petitioned to have it moved. The reason for this action was that a
an industrial site has been developed and sooner or later pressure groups develop that aim at
dumpsite and a residential site could not exist in harmony.
• A Vereeniging schoolboy was walking across an ash dump when he fell and was severely having the industrial site removed. A record should thus be kept of developments in the area,
burnt. As a result, he instituted a claim against the owners and controllers of the d ump a and township developments in particular should be carefully monitored.
nd turned down an offer of R2000009g for settlement. This is an example of someone being Air. Air can be contaminated by activities carried out at an industrial site, and pollution
injured as a result of the location of a solid-waste disposal site. agents can be dangerous to health and property. Atmospheric pollution is controlled in terms
of legislation and the audit should ensure that the necessary permits and monitoring
procedures are in order. The audit should then concentrate on identifying the sources or agents
Risk control steps: Structuring the risk identification audit that pollute the atmosphere and risk control steps should be directed at those particular
Again, the point of departure for an environmental impairment risk control programme is the sources.
identification of the risk. A large part of the literature100 on the subject of environmental Water. Water is controlled in terms of the Water Services Act No. 108 of 1997 and the
impairment in the risk control field concerns the structuring of a risk identification audit. This National Water Act No. 36 of 1998, which have very strict provisions and stringent
audit is somewhat more diffcult to structure and carry out than the audit of, say, the identification monitoring requirements. Again, some sort of statutory risk control programme should be
of an event such as a possible accident, since the causes of the environmental impairment are not implemented to ensure that the requirements of legislation on water have been met.
that obvious to the person carrying out the audit. The audit should concentrate on identifying the source of the water, its discharge point,
For example, in the Witbank case mentioned above, the dangers of pumping water into an and whether and where any contaminants are let into the water. Once it has been determined
underground mine were not self-evident. Even if they were noted, the person conducting the audit that the water is or can be contaminated, risk control steps aimed at those specific aspects of
may not realise that this could be the source of a RIOO million lawsuit. It is therefore so much contamination should be implemented.
more difficult to identify the possible sources and consequences of actions in the case of risks of It should also be borne in mind that contamination need not be ofa chemical nature. It
this type. could be something as innocuous as heat. Often factories draw water for cooling purposes and
While an audit may be structured in a number of ways, a logical way of going about it would then return the heated water to the streams. Such heated water could be the cause of changes
be to examine the areas that could be affected by environmental impairment agents and then to or damage to the ecology.
identify the agents that could affect that particular aspect of the environment. The areas normally Groundwater. Groundwater pollution is much more serious in that the pollution can take
affected by environmental impairment are the surrounding land, air, running water and place without anyone being aware of it. Groundwater is also controlled by legislation and the
groundwater. requirements of the legislation, particularly the regulations promulgated in terms of the Water
Surrounding land. The most obvious people to be affected by environmental issues are those Services Act, should be met.
on surrounding properties. A description and profile of surrounding properties should be drawn On a large industrial site, boreholes should be sunk around the site and samples taken at
up. regular intervals to determine whether the groundwater has been contaminated. In addition,
the identification audit should be aimed at
194 RISK MANAGEMENT: MANAGING RISKS
98 26Oecernber 1986.
99 TheStan 28 July 1988.
100 On environmental aud1Klng, see Russell (1985: 27 onwards)
193
101 1981 36 to.
monitoring the groundwater condition if contamination is detected. The next step of the
195
audit should be to concentrate on determining which aspects do, or could, cause the
contamination of the groundwater. Dams. Storage dams can be the source of environmental impairment, particularly to
Sources ofcontamination should be identified. %ese include items such as dams, groundwater. If storage dams are used and have been identified in accordance with the
particularly cooling or holding dams, underground storage tanks, etc. In the case of dams, abovementioned risk identification programme, steps should be taken to ensure that they are
contaminated water and chemicals are often pumped into the dam, and as the water not the source of contamination.
evaporates, the concentration of the chemicals rises and the chemicals seep into the Monitoring systems should be installed on the dam to detect or monitor any leaks. The
groundwater. Eventually the contaminated groundwater could surface several kilometres condition of the groundwater in the vicinity of the dam should also be monitored. If a dam
away from the site. develops a leak, the detection system should pick this up and steps should be taken to seal the
Another source of groundwater contamination is underground tanks. If there are leak in such a way that further leaks will not occur. Special linings may have to be installed for
this purpose.
underground tanks on site that contain chemicals, these should be identified and the type of
Underground tanks. These are a further possible source of groundwater pollution and
chemicals, as well as their possible effects, should be
hold the potential for economic disaster. 102 In the US, Congress amended the Resource
listed. Conservation and Recovery Act in 1984 and gave the US Environmental Protection Agency
Closely allied to underground tanks are underground pipes. Solid waste is another source
broad new areas to regulate, together with unusually specific legislative directives. Among the
of groundwater contamination. Often waste buried years earlier decomposes and seeps into
new items that Congress targeted for regulation are underground storage tanks, in effect
the groundwater. For this reason, solid waste dumps and tips should also contain leaching changing the way in which such tanks may be managed.
points or boreholes from which water flowing from the dumps can be monitored. The new legislation in the US specifies how tanks are to be installed and monitored. These
regulations, which include the following, give a good indication of the risk control steps that
Franschoekse Wynkelder (Ko-operatief) Beperk v SAR & one should consider with regard to underground tanks:
In this case the South African Railways & Harbours (SAR & H) employed a contractor, Union Weedkiller The usage ofsingle-wall tanks is discouraged and the construction of tanks with spill
Services (Pty) Ltd, to spray the undergrowth alongside the railway line with weed killer. The weed
containment systems encouraged.
killer apparent}y seeped into the water table and caused damage to vines on nearby farms. Because
the harvest of grapes could not be sold to the cooperative society, the society suffered a financial Tanks and piping should be designed so that theycan be visuallyinspected on a frequent
loss of R975 664, which it sought to recover from the SAR & H. basis. Tanks that cannot be visually inspected require secondary containment and alarm
However, because the action was brought by the cooperative society and not the systems capable of rapidly detecting a leak. The main emphasis of the programme should
individual farmers, the exception by the SAR & H was upheld, on the basis that the damage be the prevention of leaks from tanks and pipelines.
the cooperative suffered was too remote. The basis of the action was negligence and, had Regular tank inspection and testing should be undertaken. The inspection frequency
the individual farmers sued, the result may have been different. should not be less than once per year. Testing programmes should form part of the plant
maintenance system.
Transportation. Environmental impairment need not only originate on the site itself, but can Underground or groundwater monitoring is a prerequisite for the underground storage tank
come from activities conducted off-site. For this reason, it is necessary to carefully programme.103
investigate the transportation activities of a company. It should be determined whether the All new underground tanks must be installed to a recognised, approved standard of construction.
company transports any substance that could contarninate or impair the environment. If so, • A detailed tank management plan that addresses operator training, emergency response
and if an accident occurs, the environment could be contaminated off-site. This has indeed procedures, and operation and inspection procedures must be drawn up and kept up to
occurred in South Africa. On a number of occasions, hazardous substances that were being date.
transported spilled from trucks, polluting rivers.

Underground tanks and pipes must have corrosion protection and make use ofnon- South Africa are very strict, but also cooperative, and if the programme is drawn up in consultation
corrosive backfill materials that are installed in accordance with the requirements of a with these bodies, the legislative requirements are, as a general rule, complied with.
properly qualified professional engineer.
Russeo and Ward (19871

104 Details of ur&rground PI ping can be found in Chemical fryineefing, 7 Ckcernba 1987. p 22.
103 DetaiE Of an underground monitoring system can be fCNjnd in Russell (1987). 1934 SA
196 197
It is clear from this list that the US takes the possibility of environmental impairment from Fire risk control
underground tanks very seriously indeed. Many of those relatively closely involved in this aspect of risk control do not always realise
Underground pipes. %ese can also be the source of pollution, and detailed requirements must the nature and extent of the damage caused by accidental fire and explosion.
be drawn up for the installation of these pipes. 104 This programme is similar to the programme for Direct property damage losses in South Africa run into hundreds of millions of rand
underground tanks. annually. For instance, in 1997 the direct loss was estimated at RI,2 billion. This figure does
Solid waste. Contaminated solid waste has become a world-wide problem. In some well- not include consequential losses, which probably total three times the direct loss amount.
publicised cases, ships loaded with such waste have travelled around the world in search of a In addition to the monetary loss of property or assets, the loss of life resulting from
country to accept the cargo. Before dumping, the type of waste should be investigated and all waste fires and explosions must also be taken into consideration. In 1997, the number of deaths
sites carefully recorded on a plan. reported as directly attributable to fire was 160. However, this number actually only reflects
Where possible, a properly qualified hazardous waste contractor should be appointed to dispose deaths that occurred in urban areas, and figures compiled by the Department of Statistics
of this type of waste. The contractor should then take the waste to a specially designated disposal reflect an annual average in excess of 1 000.
site. When such disposal takes place, it is important to ensure that all liability passes to the Fire also causes unemployment and loss of business opportunities. Notwithstanding
contractor. The conditions of the contract should thus be examined very carefully. The location of insurance cover, it is estimated that more than 50% of all businesses that suffer a major fire
any buried waste should be carefully recorded on site plans. never recover sufficiently to resume their business.
Regulations promulgated for solid waste sites can be used to develop a statutory risk control
programme.
Legal requirements, codes ofpractice and standards
Many aspects of fire and explosion prevention and protection are prescribed by law. The
Smit v Suid-Afrikaanse Vervoerdienstew most important of these are to be found in the National Building Regulations and Building
An employee of the South African Transport Services dumped some carbide in a rubbish bin at Standards Act No. 103 of 1977, the Fire Brigade Services Act No. 99 of 1987, the Mine
a station, a place that was readily accessible to the public. Willy Smit, a boy of 1 2, found the Health and Safety Act No. 29 of 1996 and the Occupational Health and Safety Act No. 85
carbide and used it to make a bomb by placing it in a bottle with water. The buried bomb of 1993.
exploded, injuring the boy.
Other legislation also applies to a varying extent and includes, among other things, the
Hazardous Substances Act No. 15 of 1973, the Explosives Act No. 15 of 2003 and local
Atmospheric pollution. This is controlled in terms of the Atmospheric Pollution Prevention
authority fire by-laws. In addition to the above, the numerous regulations pertaining to these
Act No. 15 of 1985, and a statutory risk control program-me should be drawn up to comply with
laws also apply. Codes of practice dealing with fire risks are extremely important for risk
the requirements of this Act. The necessary permits should also be obtained and permit conditions
control purposes.
strictly adhered to.
Another valuable source of information is the National Fire Code published by the
Surface water. In many instances, pollution from a site lands on the ground and is washed into
National Fire Protection Association of America. This is a comprehensive guide and it is
a river during a rainstorm. Where this is identified as a problem, the storm water system should be
updated annually.
set up in such a way that substances are treated before leaving the site.
In South Africa, the Automatic Sprinkler Inspection Bureau (ASIB) is the body with
Water. Water pollution is also controlled by legislation and the necessary statutory risk
the responsibility not only to monitor sprinklers installed in buildings on behalf of both the
control programmes should be drawn up in accordance with legislation. The water authorities in
insured and the insurer, but also to publish and update the rules for automatic sprinkler
installations. This rulebook is a fully comprehensive document that has been developed over
more than a century of fire engineering experience and is an indispensable tool for those
engaged in fire risk control.
In recent times, the need for effective risk control has prompted some businesses,
mainly corporate groups, to draw up tailor-made risk control standards of their own.
198 Depending again on the size of the business, an essential part of any fire procedure is the
nomination of fire teams, team leaders, and fire wardens or marshals. In these people are vested
the responsibility of initially taking charge in the event of an emergency until the arrival of the
Requirements ofinsurance companies public fire service or the person holding overall responsibility.
Depending upon the wording of the policy, the insurance company may impose certain
requirements on the insured. These requirements usually apply to firefighting equipment and Fire-safe building design
Among the many prerequisites of any building is fire safety. This can be broken down into the following
systems such as automatic sprinkler systems. It is usually a prerequisite for such equipment
aspects.
and systems to comply with ASIB rules.
In some cases, especially where hazardous materials such as liquefied petroleum gas or Life safety. Every building should be designed so that in the event of fire, people inhabiting that
flammable liquids are being used, handled or stored, the insured will be required to comply building may have free and unimpeded access to safe escape routes.
with the codes of practice. Building protection. Buildings should be designed and constructed in such a way that in the event
of fire, the building affords the maximum protection to itself and its contents until fire-fighting
insured will, in all cases, be expected to comply with statutory requirements as they
operations can begin, and also during such operations.
apply to the particular premises or business being conducted. Ihese include, among others, the
Building survival. High-value structures require that the building skeleton be capable of suffering
OHSACT and the National Building Regulations.
a major fire without collapse or serious structural damage. Rehabilitation of such a building is
In those rare cases where legislation is unclear or where no provision as such is made for
far less expensive than demolition and reconstruction, while business interruption costs are also
a particular process or activity, the insurance company may determine the requirements it
minimised.
deems necessary. These requirements are generally not made conditions of the policy.
It is of particular importance to note that once the terms of a policy have been agreed Site planning and layout
upon, any loss that may occur and that can be proved to be attributable to the non-compliance It is vital that fire protection be perceived as one of the priority items to be dealt with by the
by the insured with these terms may result in the repudiation of a claim by the insurer. design team. Possible future expansion must be kept in mind. In all cases, the National Building
Regulations and codes of practice must be applied.
Risk control
• Inspection programmes Construction methods
By far the most important aspect of fire and explosion risk control is an adequate preventative While it is possible to construct an entirely fire-resistant building, such a structure would have no
inspection programme. No business undertaking is too small to benefit from this, and in most practical use. Buildings must be habitable and useful, and the resultant fire risk must be dealt
cases the results are extremely cost effective in the long term. Better insurance rates can be with.
obtained and business interruption costs are eliminated or at least minimised.
Non-combustible materials should be used as far as is practicable. The term (non-combustible'
does not necessarily mean that they are fire resistant. In this way, steel requires a protective
Emergency procedures
The absence of a predetermined and formalised fire or emergency procedure often directly covering if it is to reliably carry a load through a fire, while glass has very little or no resistance
causes a fire to develop into a major loss. An almost universal necessity is that every business to fire.
must have an effective plan of action in the event of a fire, explosion or other emergency. One The structure itself should ideally be designed in such a way that it adds as little as possible to the
of the prerequisites of such an action plan is that it should be simple, effective and ultimate fire load.
understandable to all employees.
In all businesses, an authoritative person should be charged with the overall responsibility for Fire detection
risk control and also for the formulation of a fire procedure. In small operations, one person One of the most important aspects of loss prevention as a result of fire is an early warning that a
would, no doubt, carry out this task on his/ her own, but in larger organisations he/she would fire has started. Modern fire-detection systems, if properly engineered, installed and maintained,
require the assistance of a nominated or elected committee. can be extremely valuable in raising the alarm at an early stage,
T 99
200
In order to determine the type of system best suited to a particular building, it is essential to In this type ofpolicy, natural perils are covered in the property damage section. The actual perils
have a thorough knowledge of the nature of the risk against which it is to be protected. covered are generally specified, as are those specifically excluded from the policy.
Although it is impossible to prevent a natural peril from happening, there are tried and tested
Other fire-risk control measures
risk control techniques and measures that can reduce the loss of life and damage to property that
These include alarm systems, portable and fixed fire extinguishers, the identification of so often characterise these perils.
hazardous materials, and fire investigations.
Risk control techniques and measures

Natural perils Risk avoidance
One needs only a cursory knowledge of natural perils to realise that many of these perils are not
evenly distributed around the world or in a particular country. While some perils, such as fire and
Natural perils are those hazards or pure risks generally arising from natural causes, as lightning, occur widely, others, such as earthquakes and severe storms are limited to particular parts
opposed to losses brought about by human actions,
of the world. In South Africa, for instance, damaging earthquakes are rare events for the country as
a whole, and even then they are generally restricted to specific areas. This characteristic of natural
This definition explains why damage caused by natural perils is often referred to as being the
result of an act of God. Down through the ages, natural perils have been responsible for many perils leads directly to the risk-avoidance technique of risk control.
of the worst losses in history, and well-known examples include fire, floods, storms and In accordance with this technique, areas prone to the occurrence of one or more perils will be
earthquakes. Because of the largescale destructive power of natural perils, the term avoided where possible. For example, given reasonable alternatives it is better to avoid sites that
'catastrophe' has come to be associated with natural perils, especially from an insurance point
are subject to regular river flooding or tornadoes. This is easier to accomplish for certain perils
of view.
than for others. In many cases, the location of the premises depends on factors beyond management
Virtually all natural perils are insurable;)06 in fact, modern short-term insurance practice
has its origin in the need to spread the risk of losses brought about, in the main, by natural control, e.g. thermal power stations near sources of coal, or sugar mills near cane farms.
perils in the maritime industry.

Risk elimination
In the context of conventional short-term insurance, there are two main forms of natural
perils coverage. This technique has the same final result as risk avoidance except that in the case of elimination,
the particular risk already exists and elimination will generally require a relocation to an area that
Fire and allied perils is free from the natural peril that is to be eliminated.
Fire and 'allied perils' is a popular form of insurance cover. In this context, allied perils are
those historically associated with fire insurance, and against which cover is effected by Legislation
adding them to a fire policy. nie following are the usual allied perils: earthquake explosion It is obviously in the interests of government and the local authority to avoid or eliminate, as far as
riot and strike malicious damage special perils spontaneous combustion sprinkler leakage. possible, natural perils affecting the people for whom 202
106 See chapter '4. 146 for a discursion of insurance agalnst the consequences of natural perils.
201 these authorities are responsible. For this reason, a number of statutes have been designed with the
objective of avoiding unnecessary exposure to natural perils. An example of this legislation is the
Water Services Act No. 108 of 1997, which states in section 169(a) that no township may be
Special perils generally constitute an extension to a fire or profits policy. These special perils usually
established or extended without the flood levels expected at prescribed flood recurrence intervals
include the following:
storm, wind, water, hail or snow aircraft, aerial devices or articles being indicated on the layout plan. This alerts the authorities who are empowered to approve the
dropped from them impact by animals or road vehicles. establishment or extension concerned to the fact that certain areas are vulnerable to flooding.
Assets all risks Risk reduction

This technique is aimed at the reduction of loss or damage brought about by one or more natural Levels ofsecurity risks
perils. Risk reduction starts at government level with certain legislation such as the National The range of security risks as outlined above is so wide that it is useful to consider them at three
Building Regulations, where the minimum standards laid down are aimed at protecting the safety levels.
and well-being of society by regulating the way in which buildings are constructed. Many of the (i) Risks to personal security affect the individual in his/her private capacity.
features required serve to protect these buildings against water, storm, hail and earthquakes. (ii) Security risks to business organisations threaten industrial, commercial and other private-sector
Earthquake-resistant buildings are being developed in Japan and California, and similar guidelines undertakings.
apply in South Africa.
(iii)National security risks directly or indirectly threaten national assets or the nation as a whole.
• Emergency planning
Security in the risk control context
Natural perils such as floods, storms and earthquakes will often constitute emergency situations
Security as a risk control discipline cannot be isolated into a totally separate field, as it has
where and when they strike. Emergency planning is an important risk-control measure aimed at
implications for a number of other risk control disciplines such as safety and fire. For
reducing the severity of the impact. example, certain measures that may be desirable from a security point of view may not be
acceptable in a fire protection context, such as the locking of exits and other means of
escape from buildings.
Crime perils
Definition Motor vehicle risk control

Security, as a business management risk control discipline, is concerned with the provision of
Nature and extent of the problem
appropriate standards of protection for people, property and other assets against the security
Motor vehicles represent a significant part of the capital and operating expenditure of any
risks to which they are exposed.
organisation, not only because of high running costs, but also because of costs associated
with fortuitous losses such as accidents. Such losses can be attributed mainly to a lack of
This definition correctly implies that security risks must first be identified and evaluated in the effective risk control measures.
organisation concerned before appropriate standards of protection can be considered.
Security risks in a business setting include not only criminal acts, but also so-called Consequential losses
The opinion that accident costs are adequately covered by insurance and that the company
mischievous acts, negligence, carelessness and certain public liability risks, all of which result in
would not suffer any financial loss is grossly misleading. Apart from the cost of repairs or
incidents that threaten the business's security.
replacement and the payment of third-party claims other factors must also be taken into
203 consideration. These include the policy excess, loss of use of the vehicle, management
time in investigation, the retrieval of the damaged vehicle, the hire of a replacement
Security risks are classified into the following four main categories: vehicle and loss of productivity.
(i) criminal acts Insurance premiums are linked to claims payments, and the higher the number of claims
(ii) negligence, carelessness and mischief (iii)public liability losses in a given year, the higher the premiums will be the following year. Properly implemented
risk control measures may therefore favourably affect insurance premiums.
(iv) unconventional threats and mischief.
Many ofthese threats, particularly conventional criminal acts, are commonplace and have Scope ofvehicle risk control
been around for centuries. The protective measures relating to these traditional or Motor vehicle risk control measures can be applied in every operating location, whether the
conventional threats are generally in place and form part of existing security programmes. organisation has only a scooter or a large fleet of motor vehicles. The elements of a risk
More recently, attention is being directed at some unconventional criminal threats and those control programme are given below.
arising in the socio-economic arena.
Regardless of size, every organisation should have a defined motor vehicle risk control (i) commercial and industrial vehicles: goods-carrying and passenger vehicles (ii) allocated and
policy, and a policy statement signed by the CEO - should be issued to all members of staff allowance vehicles: motor vehicles assigned on a permanent basis to a company or vehicles
concerned with the use and operation of company or allowance scheme vehicles. owned by drivers who receive an allowance from the company
(iii)pool vehicles: vehicles not allocated to designated people, but which are available to authorised
employees
Responsibility for the coordination, decision-making and daily control of the
(iv) off-road vehicles: vehicles used exclusively on-site by appointed or authorised employees.
motor vehicle risk control programme should be delegated to a member of management either
on a full- or part-time basis.
Freight
The nature of the freight being transported influences the type of risk control measures that need to be
applied. The following may be considered.
Guidelines should be compiled for the selection of drivers detailing the minimum standards
required.
• Explosives and hazardous substances
Driver induction and training Ifexplosives are transported, the relevant provisions of the Explosives Act No. 15 of 2003 and
A continuing programme of education, evaluation, training and retraining of drivers should chapter 6 of the regulations must be complied with.
be carried out. 'The regulations governing the conveyance ofhazardous substances promulgated in terms of the
Hazardous Substances Act No. 15 of 1973 must be complied with.
A programme to encourage discipline and motivation and to provide opportunities for Passengers
The provisions relating to passenger-carrying vehicles of the Road Transportation Act No. 74 of 1977 must
advancement should be introduced.
be studied if employees are to be transported by company transport.
A programme to ensure that the vehicles are maintained in a roadworthy condition is • Extra-heavy loads
The vehicle must be custom designed for the load being transported and the pre-trip checklist
essential.
should include checking the stability of the load. Codes of practice are a useful reference guide
Accident reporting and investigation in this regard.
A procedure for reporting and investigating accidents in order to determine causes and to 206
reduce the frequency of their recurrence should be established. 205
Effective cost control Consequential loss (business interruption)

A method for ensuring that effective cost control is achieved should be developed and maintained.
One of the major exposures facing any company is the exposure to loss of profits as a result
of an incident that interrupts or interferes with normal business activities.
The annual accident frequency and vehicle accident loss rate should be determined on a
It will be recalled that one of the classifications of risk was according to the financial
continuous basis. Separate figures should be recorded for the different types of vehicles owned
consequences of the risk. When an analysis was done on this basis, it was indicated that the
by the organisation. Ihese rates should also be calculated in respect of each driver for control
and supervision purposes. following five classes of financial consequences exist: (i) the costs of the damage to assets
(ii) the consequential losses that result from the damage to these assets
Fleet types (iii)pure financial losses (iv)
In most organisations, the various types of vehicles used will invariably be categorised legal liability claims
according to the different operating conditions, procedures and instructions that will apply. The (v) employee benefit costs.
following four classifications would be applicable to the majority of organisations.
The consequential loss programme deals mainly with the second category, namely financial
consequences associated with damage to assets.
It is important to keep this classification in mind, because otherwise confusion will result Production can also be lost if other items of equipment, such as testing and quality assurance equipment,
between classes (ii) and (iii), above. The main difference is that generally, pure financial
are damaged.
losses are not insurable.
There are many instances where a company can incur pure financial losses, e.g. in the
• Damage to internal services
case of a strike that may bring production to a halt. This results in loss of sales or income, and Consequential losses can occur if services such as the electrical, mechanical, steam, compressed air or
these are generally not an insurable loss. Measures to prevent and deal with strikes are, computer services are damaged.
however, the responsibility ofthe industrial relations manager and not the risk manager.
If one designs a consequential loss programme without taking into consideration the Utilities
clear distinction between categories (ii) and (iii), above, it is possible that the programme will Consequential losses are not only caused by damage to owned assets. A business that depends
encroach upon speculative risks. Indeed, if the distinction is not borne in mind at all times, on third parties, e.g. suppliers, may also be affected if there is damage to such third-patty assets.
the risk control programme will encroach on all the activities of the firm. It follows that a business may be affected by losses occurring on suppliers' premises.
Production can also be interrupted as a result of damage to utilities such as electricity, waters
gas, roads and steam. Another source ofconsequential loss is the loss of a distribution facility.
The nature of the interruption risk control programme
From the above, it is clear that interruption losses can arise from a wide range
A risk control programme is in many ways similar to the legal liability risk control
programme in that both are aimed at managing the consequences of an event. fie control of
the event itself will often be dealt with according to another risk control programme. Thus, Time-related aspects ofconsequential losses
extensive consequential losses can be suffered ifa factory is destroyed by fire, but the
prevention of fire is the function of a fire risk control programme. The nature of interruption losses
Consequential losses are time related because the longer production is halted for any reason,
A consequential risk control programme has its own objectives. For example, the
the longer the period ofloss ofincome will exist, or the longer additional working costs will be
purchase of a spare transformer with the specific objective of limiting consequential losses
incurred to keep the products or services in the marketplace.
cannot be viewed as part of a fire risk control programme.
207
Sources ofconsequential losses

Loss of aggregate assets
It is important to identify those loss situations that could result in consequential losses. It is
obvious that the destruction of assets can be a major source of consequential losses, because ifan
asset that is normally used in the production ofincome is destroyed, then consequential losses
can follow.
Current assets
Loss of current assets such as raw materials and finished goods can also result in consequential
losses. If raw materials are damaged, say in a fire, production time may be lost. If finished goods
are destroyed, loss of sales can result, which could also result in a loss ofmarket share.
Production assets
Consequential losses can arise if production equipment is damaged.
Testing equipment
208 R'SK 107 McDonald 221).10 this vwrk. It is poinred out that a group of affiliared insurance companies found that in IS of 21 cases of damage to steel plants, the
consequential bosses were oreater than the mateqai damage losses.
108 109 The Atkins way (1981:69.in which the sum insured is calculated (and What this tn fact n-pans) is Important If a risk is to beevaluated.
• Indemnity period 209

Since consequential losses are time related, the insurance industry imposes a time limit to such
losses. The period for which losses are covered is called the indemnity period, and the result is that Risk identification
long-term losses (i.e. those extending beyond the indemnity period) will not be covered by
insurance. • Intangible nature ofinterruption loss
When one deals with losses arising from damage to assets, risk identification is both
Danger of loss of market share important and diffcult. Regarding damage to assets, all that is needed is first to identify the
assets that can be damaged and then the perils to which these assets are exposed. Once it
One of the most serious consequences of a long halt in production is the possibility of the company
has been determined which assets are at risk, the cost of the assets can be calculated and
losing market share. Tüs is further aggravated by the fact that because this is a long-term
an indication of the exposure due to loss or damage to the assets determined.
consequence, it is not insurable. In cases where long-term losses may occur, it is important that an
adequate risk control programme, with the specific aim of preventing long-term losses, be It is much more diffcult, however, to determine the exposurefor interruption losses.
installed. Insurance is not the solution to loss of market share. While consequential losses have a financial implication, it is not possible simply to
examine the asset and then to determine the consequential losses. Table 8.4 lists some of
Accounting and insurance aspects of interruption losses Some forms of interruption losses are the assets that are normally involved in an analysis to determine what the consequences
of the loss of these assets will be. Clearly, the unique perils and consequences of damage
insurable by means of specialised related to each asset will have
insurance.
In some instances, consequential losses exceed the cost of material damage. 107 Interruption
Table 8.4 Assets from which consequential losses can arise
insurance cover will generally only respond to claims for damage to assets from insurable perils.
Consequential loss insurance is thus not a free-standing insurance against consequential loss; it is
peril Consequence
intimately tied to physical damage to assets as a result of insured perils. Atkins illustrates this point
Composite assets buildings
clearly:tos 1 Administration
2 Production
By whatever name it is known (and for the present purpose it will be called 'profits' 3 Storage
insurance), its basic function is to pay for the financial loss that may be sustained by a (a) Raw materials
business undertaking should its operation be interrupted or interfered with as a result of (b) Finished goods
4 Services
'insured damage' to the physical assets of the undertaking occurring ....
5 Transport
Usually, the insured damage is that which would be covered by a 'fire and allied
Equipment
perils'policy, but there are other forms of policy, such as 'engineering' and 'marine
I Production machinery
policies', to which profits insurance, in one form or another may be related. 2 Services
Therefore, the interruption-or-profits policy is activated by damage to physical assets that 3 Testing
4 Transport
are normally covered by another policy such as fire and engineering policies. If this distinction is
Assets belonging to utilities
not borne in mind, risk control programmes will tend to deal with management and engineering 1 Water
problems that might often be quite unrelated to event risks, but rather to increased business 2 Electricity
3 Steam
effciency.109
Assets belonging to suppliers and the final projected impact on the gross profit estimated with a reasonable degree of
T Materials certainty.
2 Services 211
210
Risk evaluation
From table 8.4, it is clear that the analysis includes assets that do not belong to the business. Time periods used in the evaluation
The point is again emphasised that it is possible to insure against the loss of profits that can Interruption loss analysis (including flow diagrams) lends itself to the concept of risk evaluation
occur as a consequence of a business interruption caused by a supplier due to damage to his/her by means of various forms of probability analysis. This is particularly true of items such as
assets. The insurance is not actually against the loss of the supplier's materials, but against loss electrical distribution works, which can be extensively analysed and which lead to the field of
of the supplier's material due to damage to the supplier's assets. reliability studies and value engineering.
As indicated in table 8.4, the impact ofvarious perils on the assets must be assessed. The To avoid unnecessary complexity in the evaluation, one should remember that the
perils include fire, impact, internal breakdown and theft. In order to complete table 8.4, the exposure is time related and that the evaluation has mainly two limits of concern. In the first
repair and replacement costs have to be estimated. The downtime will also be related to the instance, the evaluation is not concerned with small, short-duration breakdowns that can
spares that are kept or are available. normally be funded by a maintenance budget. One could, therefore, comfortably work on a
Each item used in the production process must be identified and carefully considered. At lower limit of up to a seven-day breakdown. The other important limit is the insurance period
the same time, any special features such as PCB insulation in transformers or unique features, of indemnity. This period can range from about three months to one or two
which makes a machine diffcult to replace, will be discovered and the impact on downtime
In order to simplify the calculations, as a first approximation of the exposure, one should
estimated.
simply work on the basis that a certain item of machinery will be out of operation for a period
Ifit becomes clear that the downtime is unacceptable, e.g. ifthe replacement time is more
of, say, six months and determine the exposure on that basis. If one is reasonably sure that it is
than a few weeks, contingency plans should be developed to deal with this eventuality.
improbable that an out-of-operation period will last longer than six months, as in the case of a
normal electric motor, then it is not necessary to be concerned with losses oflong-term items
Flow diagrams
such as a
Since consequential losses are intangible, it helps considerably if the potential loss sources are
expressed in some tangible form such as a diagram.
If the procedure for selecting a fixed period of outage is not adoptecL loss-distribution
It is for this reason that flow diagrams are used extensively in determining the interruption
curves must be developed for each asset. The analysis then becomes much more complicated.
loss exposure. These diagrams indicate the flow of materials from the supply of raw materials
through to the delivery of the finished goods to the customer. The diagrams indicate the By selecting a fixed period ofoutage, the MFL and EMI} can be calculated, since the cost of
various items of equipment required for production purposes and other inputs, including the outage is time related.
distribution system. The flow diagrams can then be analysed to determine the effects of an
interruption from any single item, such as a machine, on the total flow of goods. Failure frequencies
Once probabilities and marginal contributions are included on the diagram, the flow A number of organisations have devoted considerable effort to establishing failure frequencies
diagram lends itself to a systematic analysis. for a large range of assets. can be used in the place of distribution curves as a first approximation
in the evaluation of losses.
Use of financial and management accounts
The most useful tool for the analysis of consequential loss values is the company's financial
and management accounts. By using the income and expenditure statements, a 'what if' Engineering risk control
analysis can be undertaken. If the financial statements are available on a computer
Loss statistics ofmachinery breakdowns
spreadsheet, the analysis becomes that much easier. So, for example, the impact of the loss
Since we are trying here to determine the consequences of a loss or damage to assets, and particularly
of an item of equipment can be estimated and changes made to the spreadsheet. The income
machinery, we will use loss statistics that have been
and expenditure figures affected by the loss of or damage to the equipment can be adjusted
110 See 6, section 62-
213
212
each type ofmachine must be examined to plan the desired risk control steps. In other words,
collected over the years. Various sources can be consulted, but probably the best known is the specific steps can be taken to prevent the breakdown of a transformer and to prevent the
course that is run by ICI in London. breakdown of a piston machine, for example.
Munich Re, one of the world's largest reinsurance companies, has done extensive work
in the field of insurance claims from machinery breakdowns. Altogether, 16 005 cases of Risk controlsteps: Machinery breakdown
damage that occurred during the period 1969-74 were evaluated. These cases were distributed Upon examination ofthe breakdown statistics, the various ratios ofthe different types of faults
among the various types of machinery as follows: can be determined. %erefores note that for electrical machines, product faults comprise some
9 135 cases of damage to electrical machines 41% of the faults, whereas operational faults
1 045 cases of damage to steam generators
1 215 cases of damage to fluid-flow machines In its study, Munich Re came to some general conclusions regarding loss control steps for
the different types offaults. The different faults and the relevant risk control steps that could be
1 000 cases of damage to piston machines
taken are summarised below.
3 610 cases of damage to mechanical handling and lifting machines.
Product faults. In terms of numbers and costs, the proportion of damage resulting from
Munich Re found that as the proportions of the groups of causes of damage included vary product faults is most pronounced in the case of fluidflow machines and steam boilers. It is
widely with the type of machines considered, the key measures necessary for successful loss heaviest in terms of cost in the case of electrical machines and installations, but it is relatively
prevention also vary. low in terms of numbers, indicating high repair costs per individual case of damage.
With these three types of equipment, the trend of technical developments towards larger
Once a machine has broken down, the cause of the breakdown must be identified. To a
and more powerful items ofequipment is particularly noticeable. For example, 50 years ago, a
large extent, the classification system chosen is arbitrary and the causes can be classified in
50 MW electrical generator was regarded as a large generator, but today, 960 MW units are
various ways. Munich Re has classified causes into three categories:
routinely installed.
(i) Product faults. These include all causes of damage that are the result of manufacture,
Newly developed components that have not been sufficiently tested in a practical operation
which include faults in planning and design such as deficient layout calculation, incorrect
are simultaneously installed in a large number ofmachines with increasing frequency.
choice of materials and unsuitable geometry; faults in processing, such as incorrect heat
Munich Re suggests the following: 112
treatment, machining errors and assembly faults; and faulty materials.
(ii) Operational faults. These include causes of damage that originate during the operation Methods of preventing damage as a result of product faults as primary cause, involve
of the installation, which include the loosening of components; the failure or non-
utilisation of knowledge from toss events regarding calculation, design and
response of protective devices; servicing faults; damage arising from corrosion, ageing,
manufacture by way of:
etc.; as well as causes that can be traced back to external influences, e.g. natural forces,
close cooperation with manufacturers in special cases early publication of
foreign bodies, over-voltages from the grid, and so on.
experience jn speciafised journals
(iii) Handling faults or attendance faults. These can be regarded as a subsection of operation
• open exchange ofexperience at conferences and seminars.
faults. In about 95% of the damage cases, external influences did not play any part.
In this way, an important contribution can be made towards ensuring that, with different
From the analysis, it is evident that the vast majority of machinery failures have been as a
manufacturers, damage arising from identical causes will be prevented from happening again.
result of factors that could have been prevented.
Further support for the prevention of damage is offered by purposeful inspection and
In designing risk control steps to prevent machinery breakdowns, generalised steps such
replacement at the appropriate time of components that have led to damage to similar machines,
as a relevant management system — applicable to all types of machinery — could be
as well as a result of basic deficiencies.
specified. However, once this has been done,
11 1 Munich Ret 1978)

i 12 Munich
214 215
Withelectricalmachines, the key risk control points are: the application of damage
Operational faults. Munich Re notes that the proportion of damage from operational
experience for calculation, design and manufacture; revisions at regular intervals, if
causes, in terms of numbers and costs, is highest in the case of piston machines, followed
possible; and the training of personnel.
by electrical machines and installations. In the case of the latter, the high number of cases
With fluid-flow machines and steam generators it is vitally important to apply damage
of damage with relatively low costs is conspicuous, indicating that there are many causes
experience in manufacturing machines and also inspection and overhauls. It is also essential
and instances of fairly inexpensive types of damage.
to consider the distribution of the findings ofdamage causes that were the result ofinspections
MunichRe suggests the following operational
carried out on turbines by Allianz and on steam generators by TUV.
Withpiston machines, thekeytoloss preventionliesintheimprovement of servicing and
• supervision of the installation, when not only sudden, but also gradual changes in
maintenance, as well as in regular inspections.
operational data (pressure and temperature, deficiencies, power input, etc.) must be
With mechanical handling and lifting equipment, damage arising from handling faults
noted and their causes established
occurs most often, as should be expected. In this case, therefore, loss prevention lies
primarily in training personnel by demonstrating examples ofdamage, as well as by
• constant control and supervision of measuring, control and protective pamphlets and instruction plates. In addition, everything possible must be done to reduce the
proportion ofhandling faults by improving the way in which staff operate equipment.
• constant control of the condition of operational and auxiliary materials (lubricating,
cooling, refrigerating media, etc.) regular servicing, maintenance and overhauls Planned maintenance systems
• overhauls and inspections at regular intervals so that the components that have reached We have already seen that the key factor in the prevention of loss or damage to equipment, and
the end of their lives through wear, corrosion, erosion or temperature influences can be in particular production machinery, is an adequate inspection and maintenance programme.
replaced before failure occurs. These inspection programmes are best formalised in terms ofwhat is known as a 'planned
maintenance system'. In such a system, the maintenance requirements of machinery are defined
On the basis of systematic evaluations of damage and inspections, oroptimumreliable beforehand and then carried out in accordance with the programme.
We will now examine some of the elements required for a successful planned maintenance
supervisory and control systems have been introduced for the determination ofthe most programme (which can be used to prevent the damage to an asset that gives rise to consequential
favourable points in time for carrying out inspections. losses).
Handling faults. The largest portion of damage from faults arising from mishandling is
Responsibility for the planned maintenance programme. If the prograrnme is to function
found in a group of machines concerned with mechanical handling and lifting. However, in
properly, responsibility for it must essentially be assigned to one person or one department. It is
other groups of machines dealt with here, the cause of damage is also noteworthy.
equally essential that this person or department should not form part of the production department,
For loss prevention, Munich Re suggests the following: as planned maintenance and production have different aims. In large organisations, a separate
the selection of suitably trained personnel for operating, and servicing machines planned maintenance department should be established and responsibility for the planned
in actions continual training of operating personnel — especially the required maintenance programme be assigned to that department. This department should not fall under
when machines break down - through courses, lectures, literature and pamphlets. Their the responsibility of production management.
work should also be inspected regularly. the installation of easily operated fitting and Work instructions. At the heart of a successful planned maintenance programme is what can
control elements, warning notices, and instruction plates the provision of easily be referred to as work instructions or planned maintenance cards.
understood, detailed operating instructions. There should be a separate planned maintenance card for each and every item Of equipment or
machinery. This should not be a general card, but must relate to the specific item that is being
General risk control measures for different classes of machines inspected.
In the study conducted by Munich Re, the machinery was classified into different
216
categories. Munich Re then summarised the general steps that could be taken for each class
of machinery.
This planned maintenance card should indicate the specific maintenance work to be carried • investigate alternative suppliers and markets assess the vtdnerability of the supplier and
out on that particular item of equipment. To determine such work requires that a detailed market to loss of assets
investigation be undertaken of the types of failures and faults generally associated with that • assess the ability of the key supplier and market to recover from loss or damage to their
particular item. In this ways if a planned maintenance programme — which would include the
assets.
maintenance of an electric machine - is to be instituted, the specific inspections that need to
be carried out for that type of machine must be determined. Once this has been determined,
the maintenance card should deal with this specific item.
In determining the maintenance work required on the machine, one should take
8.8 Hedging of incidental risks
cognisance ofthe existing literature on the subject and particularly the insurance literature Incidental risks arise from the mainstream activities ofa company, as explained in chapter 2. For
regarding the causes of failures. Tie major causes of failures should be investigated and the example, a car manufacturer that imports certain components is exposed to foreign exchange rate
steps required to avoid any further failures should be incorporated into the planned changes. These rate changes may influence the profit position of the manufacturer negatively or
maintenance card. positively, even though foreign exchange dealings are not its main line of business. Investors
It is not sufficient to simply define the project or work to be done in terms of the planned associate the company with its main business risk (the end economic risk), i.e. the manufacturing
maintenance project; it is also necessary to ensure that only competent and qualified people of motor vehicles. Any additional risks, like foreign exchange risks, should therefore be This can
carry out that work. be achieved by forecasting foreign exchange movements that would eliminate the risk. Since
capital markets are efficient, it is virtually impossible to out-predict the market. Alternatively,
On the card, there should be space for comments and the signature of the person doing
these risks can be managed by rather inflexible on-balance sheet transactions, such as borrowing
the work to indicate that it has been done. Where applicable, the statutory requirements must
in the foreign currency or by moving production abroad. Hedging using derivative instruments is
be incorporated into the planned maintenance costs.
much more flexible and cost effective, and can be used to eliminate the impact of incidental risks
Inspection frequencies. The various frequencies of inspection and other maintenance
such as changes in foreign exchange rates, interest rates, and even price changes of commodities
work should be determined by consulting the relevant literature, and the manuals and
and financial instruments.
specifications. These requirements should be incorporated into the work instructions and into
The principle of hedging is illustrated in figures 8.1 and 8.2. Figure 8.1 is the risk profile
the control system for the planned maintenance system.
ofan importer. As the exchange rate increases (and the price of the imported goods therefore
increases), the profits and therefore the value of 218
Spares policies
As we have already seen, a consequential loss is time dependent. Therefore, the more quickly
machinery can be repaired to get it back in operation, the lower the loss will be. It stands to
Figure 8.1 The risk profile of an importer
reason that the programme should not only concentrate on preventing a machine from
breaking down, but should also include contingency plans for locating spares. Chanqe in
profit tvalue)
A spares inventory should therefore be drawn up and kept up to date. Where high
Change in
vulnerability is detected, spare equipment should be purchased. hanqe rate
(price of
imported
goods)
Risk control: Suppliers
As a general rule, it is not possible for a customer to tell a supplier how he/she should manage
his/her own business or assets. Therefore, it is not generally acceptable for a client to tell a
supplier that his/her planned maintenance 217
Risk profile
programme is not up to standard; although it is becoming acceptable in terms of quality assurance codes
of practice for a client to inspect and evaluate
The following risk control steps can, however, be implemented in this regard: identify key Figure 8.2 Pay-off profile for a forward contract
suppliers of goods, services identify key markets
pay-off
profile of
should be identified and applied to risk control the approach(es) to risk control should be
forwatd
contract integrated with the other activities within the organisation so as to achieve optimal results.
Resulting The objectives of risk control programmes were stated as endeavouring to reduce the frequency and
exposure
severity of loss costs resulting from risk exposures and, through emergency and contingency planning, to
deal with and recover from loss-producing events.
Approaches to loss prevention were then described. These were categorised as engineering, human*
Risk of educational and enforcement approaches. Risk control programmes and codes of practice relative to South
impor
ter Africa were given and, under the heading of assets programmes, general assets perils were listed around
which specific risk control programmes are designed. Finally, the more important acts prescribing regal
risk control measures were listed.
Early in the chapter, it was mentioned that risks can be broadly differentiated as being assets (property)
Figure 8.2 gives the pay-off diagram of one of the financial derivative instruments, i.e. a
related and Itability related, and that these types of risks are fundamentally different. Thist together with
forward contract. A forward contract obligates its owner to buy a given asset on a specified
the fact that a 'liability crisis' is currently in evidence, dictates that with regard to risk treatment, a separate
date at a price specified at the origination of the contract. If at maturity the price is higher,
examination of this subject is called for.
the holder of the contract makes a profit; if the price is lower, the owner suffers a loss. Also,
Hedging, as a response to incidental risks, was also discussed and illustrated, showing how a derivative
superimposed in figure 8.2 is the risk profile of the importer. As can be seen, the decline in
instrument such as a forward contract can be used to hedge profits and the value of the company against
profits (or the value of the company) is exactly offset by an increase in the value of the
movements in, for example, exchange rates.
forward contract. For this example, the forward contract therefore provides a perfect hedge
against movements in the exchange rate and the value of the company.
Other derivative instruments such as futures and options can be used in a similar way
to hedge against risk. Ihese instruments will be discussed in more detail in chapter 14.
219
8.9 Summary
The aim of this chapter was to describe risk control responses to various identified risks.
Against the background of risk identification and evaluation, risk control is concerned with the
physical management of risk. An important aspect considered was the meaning of risk control
and its position within the overall discipline of risk management.
Risk control was established as a subdiscipline of physical risk management, which encompasses the
overall process of risk identification and evaluation. In turn, physical risk management, when integrated
with the activities related to financial risk management, constitutes the entire discipline of risk
management Risk control was simply defined as a method of meeting risk. This definition recognised the
broad aims of:
• eliminating or reducing the factors that may cause loss (toss prevention/ control)
minimising the actual loss that occurs when preventative measures have not been fully effective.
The historical development of risk control, particularly in the areas of fire, explosion and
industrial accidents, was sketched, followed by the general principles and objectives of risk
control. These principles included the important premises that: • management principles
Chapter 9
Risk financing
MANAGEMENT: FINANCING FINANCING
Introduction
Interrelationship between risk control and financial elements in the risk
management process Risk-financing mechanisms
Risk management within a financial decision-making structure
Summary
9.1 Introduction
The earlier chapters in this text concentrated on the nature and definition of risk and the position
of event (pure) risk within the framework of risk management. The introduction ofa model
provided a logic that can be applied to the management of risk. The model dictated that after
identifying and evaluating the risk, an 'action' point was reached when the practical aspects of
risk control had to be considered.
Viewing this text as a progression of the logic provided by the risk management model
means that we have reached a further turning point. The model now prescribes that attention be
focused on the question of the financial provision for fortuitous losses that may occur from risk
exposures that were identified, evaluated and controlled to the extent possible.
The path that was followed and explored now requires some further clarification. By
differentiating the risk management elements and arranging these sequentially for discussion,
their significance in the overall process is highlighted. More importantly, the sequence allows
the easy integration of these aspects into one all-encompassing risk management model,
221
FISK MANAGEMENT: RISK FINANCING
FINANCING
222
Retention
Tie purpose of this chapter is threefold. Firstly, it is necessary to link the elements of the basic
model to achieve an overall process of risk management, which includes the considerations of loss RISK FINANCING
External financing
financing. It will be seen that not only is the integration necessary across the model, but that
effective risk management implies integration within the model.
Secondly, the chapter will provide a brief overview of risk-financing mechanisms, the aim being
Pre-loss
to distinguish between pre-loss and post-loss financing. The intention is not to examine such
Cash flow Captive
mechanisms in detail, but rather to provide a framework within which the resulting discussions of Provisions Commercial Capital
insurance Financing
Reserves insurance Market
the financing techniques of insurance and self-funding can be placed. companies State risk
Products
facilities
Equity financing
Thirdly, we investigate the underlying rationale for risk management by viewing the risk Post-loss
management function in relation to the corporate objective of value maximisation. 'Ihis criterion is
finally used as a base from which to critically assess the risk management model and ultimately The risk-financing function becomes important when considering the pursuit of minimisi_ng
adapt it to incorporate decisions concerning investment and finance — so achieving the desired the 'total cost-of-risk' to an organisation.
holistic approach to the management of event risk.
VCost-of-risk
9.2 Interrelationship between risk control and financial Risk management professionals have for some time been concerned with the problem of
measuring their effectiveness in controlling risk costs. l The 'costof-risk* concept was thus
elements in the risk management process
developed by Barlow2 as a useful measurement device for reporting the results of the risk
and insurance management function. As defined, the cost-of-risk is the sum of:
The risk-financing element within the basic risk management model
(O insurance costs unreimbursed losses (self-insured, self-
In chapter 8 it was stated that risk control entailed any activity aimed at reducing the severity and retained) risk control and loss prevention expenses @
frequency of losses that may arise from the pure risks an organisation faces. Through systematic administrative costs.
identification, evaluation and control, this objective is achieved by:
• eliminating or reducing the factors that may cause loss events Insurance costs
• minimising the extent of the actual loss that occurs when preventative methods are not fully Insurance costs comprise the following two components:
effective. (i) Direct insurance cost: This is directly related to the total of all insurance premiums. A number of
additional aspects must be taken into account
The risk management process is incomplete, however, until one considers the element of
financing (diagrammatically shown in figure 9.1).
RISK 223
I o„ncan 23-33).
Doughs Barlow. a former president of the Risk and Insurance Management society and risk manager for Canada:
d
eveloped the concept of as a measurement for leportjng the results of the risk and insurance management
Vigure 9.1 The risk-financing element within the basic risk management model 224 RISK RISK
RISK MANAGEMENT
FINANCING
to determine the actual cost of insurance. rlhese include the opportunity cost of the funds Administrative expenses
expended as insurance premiums, the tax implications of premiums and claims, and the These expenses include the following:
cost of capital relative to the value of the expected claims. one must include costs of
clerical costs in the handling of insurance matters the cost
statutory insurance, such as workers' compensation and unemployment insurance, but
ofhandling self-insured losses the cost of reporting and
reduce the cost
investigating loss occurrences the cost of an in-house risk
(ii) Opportunity cost: This is the rand value spent on insurance premiums that could, for management department the cost of a risk management broker
example, have been spent on improving the production line or invested in the production (if applicable).
process, thereby generating more profit for the enterprise. The 'opportunity' of those
additional profits is lost due to having to spend the money on insurance premiums. When considering these costs, any business should endeavour to achieve these three risk
management objectives:
Unreimbursed losses
Ni) financial: to reduce the ratio of total cost-of-risk (relative to some base, such as revenue)
Unreimbursed losses can arise from the following:
and to maintain the improvement over a sustained period resources: to conserve assets
• excesses provided for in insurance policies
and preserve the physical health ofemployees through safe working habits and a safe
• inadequate sums insured breach of policy conditions or warranties contained in policies
and the consequent forfeiture of insurance protection work environment people it serves and through the public: the to reduction improve the
• losses resulting from risks against which there is no available insurance cover, although the enterprise's image conditions among thatthe
risks are known of accident-causing
• losses that are uninsured either intentionally or unintentionally uninsurable losses affect the product it makes and the services it delivers.
• the insolvency of the insurer.
Figure 9.2, below, shows how the total cost-of-risk, against a yardstick of revenue, is used to
Risk control and loss prevention expenditure ascertain the effectiveness with which the pure risk area is controlled. Initially, the costs will
Ihis includes the following: increase as capital is expended on risk control. Then, as risk control begins to affect losses, a
depreciation on major capital costs to reduce risk, e.g. depreciation following the reduction in self-insured losses and insured losses results with a consequent reduction in self-
installation of a sprinkler system (note that the depreciation is the operational cost, and insurance costs and insurance premiums.
this forms part of the risk control expenditure, The central issue becomes one of optimal financing (pre-loss)3 so that, while the risk
control process is being addressed, the financing of risk and losses as they occur does not
the cost of time consumed in the identification and evaluation of risk, including the cost introduce any inefflciencies. Immediately, the preoccupation with overall cost effciency
of outside consultants, if appointed ongoing operational and procedural expenses for risk becomes evident. When one considers the definition offered by Williams and Heins 4 that risk
control measures such as security guards, maintenance of fire equipment and cash- management is 'the minimisation of the adverse effects of risk at minimum cost through its
carrying identification, measurement, and control', one detects the objective of balance between the
services risk control training costs/seminars/courses subscriptions to safety, security and benefit from the removal or reduction of risk and the costs of such action. Definitions afforded
risk management societies, journals, by other writers, e.g. Greene and Serbein,5 also appear to reflect similar cost-to-benefit trade-
offs.
management time taken up by risk-control and risk-financing (insurance) meetings. trade-off referred to is the direct result of the interrelationship between risk control, on
RISK 225
the one side ofthe risk management model, and risk financing,
A
fitlnction exists between what is termed •pre-loss' and •post-loss' financing. Ibis distinction is made latel in the chapter However, it is iP>CSite to stare here that 9 Naturally, this is not to say that the "deal situation is one in which cor achieve zero or close to zero insurance or third-party financePre-loss hnancial provision for
insurarxe and detlberate self-funding are examples of pæ-bss financing. The concern. when refe•ence is made to 4%lharns "*King optimisation. and Heins (1981 is catastrophe contingenoes through Insurance must be more by virtue ofthe diversihcation and pooling achieved by tip Insurance sector,
already It been dealt is with by (he holding of a diverslfied portfolio. can In this case, that the the purchase of insurance merlts the closer investigation effect of insurance
with : 16).regard to pre-loss hnanctngS&eene and Seibein (1983:3). (Oe Vilit3hasS 10 the company owned by drvasified shareholders, it be argued specific rlsk dealt With by pcoling
and Vivian. I Is that most insurance wrlters in the field ignore the matter of investroent. v,hkhr
226 Il This consideration most imponatü. is interesting included in the oveqall risk management as process, facilitates a rmre all-eryompassjng management model
model {or rlsk developed management earlier in In re13tIon this chapter to the incorporatoperation; es
of the business organisatlon a whole. See Doherty (1985: 23-34).Tie risk the question of
investment.
on the other. What becomes clear is that, in the shorter term, possible economies RISK 227
evident in risk financing, e.g. a reduction in the premium to insurers, should determine effort or
expenditure on risk control activities — and that an optimal balance should be sought. The relevance of this distinction between investment and financing decisions becomes
Practical considerations, however, dictate that longer-term horizons are more significant, clearer once the company's objective is defined as the maximising of the wealth of
shareholders in the long run, i.e. the value derived from the firm's projected earnings stream.
and this often means that 'levels' of risk control and insurance expenditures may not be that
12
closely related to shorter-term cost-to-benefit trade-offs.6 Nevertheless, this does not imply that
The above factors lead to the conclusion that in pursuing the practical risk management
the concept of balance or optimisation is unimportant. On the contrary, once one takes long-
objective of minimising the cost of risk, the following qualifications are necessary:
term horizons into account, optimisation again becomes vitally important to risk managers.
A longer-term view should be adopted when cost-to-benefit trade-offs relating to levels
It can be argued that risk control endeavours should not necessarily depend on short-term of risk control expenditure are considered. Risk control should be undertaken within the
cost-to-benefit trade-offs per se;7 risk control should be undertaken within the limits imposed limits dictated by broad corporate financial objectives and constraints.
by broader (often strategic) corporate financial objectives and constraints, and once Risk-financing decisions should be taken with due reference to the investment decision.
determined, the risk finance function should then be optimised.
ne implication for financing through insurance following such desired Simplifying the model may be appropriate so as to illustrate the integrative path in the risk
optimisation is interesting. When insurance is regarded as a transfer of risk cost8 and an insurer management process. If 'cost-of-risk' is defined as comprising three basic elements: 13 (i)
as a financial intermediary, then, in the long run, being more reliant on this type of intermediary expenditure on risk control,
financing must imply a higher cost than being less reliant.9 (ii) expenditure relating to retained risk/self-funding (deliberate and nondeliberate), and
Even with long-term considerations in minds the close interrelationship remains between (iii)expenditure on insurance cover, then, simplistically, one objective of risk management
risk control and the financial elements that constitute the risk management process. In assessing can be stated to be the long-run minimisation of the cost-of-risk to an organisation.
the interrelationship between risk control and financial elements, there is, however} a most
significant aspect that needs to be examined. Risk is identified, measured and handled on the Diagrammatically, the cost-of-risk is shown in figure 9.2.
implicit assumption that decisions relating to it will affect both investment decisions (the
commitment of resources to productive use) and financing decisions (those concerning the 12 See Pretty (1999) This important study examlnes various dynamics of corporate risk behavlout- Orr part of the wok concerns itself with eauating the impact of on
shareholder value and 'recoverioywith •non-recovering'firrns Conclusions following the regression analysis conducted whae the regression variables BÆre Size. impact,
funding of such investments). Industry. insurance, fatalitles, country, responsibility, class and
i
. There is ammo non-recoveærs initial negative response cfover of marketcapita'lsation,
Often, the investment decision is ignored and emphasis is rather placed on 2. 'n tip 61st two or three months. the magnitude ofthe estimated financial loss is Significant among non-recoverers.
3. There are targe numbers of fatalities from the non-recovering portfoJH' ofdisasters. and thls seems to govern morket 'eaction in the first
the insurance purchase alone, being a source of finance, without due attention being paid to
4. Thereaftä, the issue of management's responsibility for accident or safety lapses appears to explain shareFH)Jder value response. by contrast. whether the losses lave'e fully
whether such financing is indeed necessary at all, 10 or necessary and prudent to support post- covered insurance does not appear to have slgnfrant influence.
the research presents evidence of the substantial impact of a catastrophe on a fum's stock returns and trading volume- The impact on share Dice is derived from two sets of
factors. The hrst is the djrect financial consequence of the catastrophe, particularly regard to cash flows
loss reinvestment. ll
will form a opinion of how the catastrophe affect cash flows and will adjust the share accordingly. The second set of föctors js 'Mirecr.These have
an impact on returns from what the catastrophe reveals about management skilb not hitherto .refÉcte•d in the valoe of the stock. Management is pbced in the spotlight
and faces rhe oppcvtunity to demonstrate ski" or otherwise in an extreme situation the market will evaluate management'sabi!ityt whit-h will result In reassessmentot
the fum's future cash flow position. both
6 See chapter 4. in %hiCh the point is made that often, ifu ndue emphasls Is placed on fort-term cost-to-benefit relationships, dangerexEts of produclng correct short-term Tik effect coukf therefore be elther positive or negative positive where the benefits of what is revealed management outweigh the financial loss followrng the
answers that may result in disasters In the long Ion. cerastrophe; unfavourable If the revelation is negative since this will amplify the negative Impact of the financial loss What is interesting is that the results of the study
suggest that it is the indirect factors that dominate the I m pact on stock Insurance as a financial will not be to protect the firm the share price effects ofcatastrophes,
7 Recent corpora(e governance princlp'es endorse this view. In the LIK management Ot listed companies is accountable risks beyond tradlttonal financlal threats. and the firm's insurance *rategy should therefoæ not be viewed as substrute fol high-quality risk management and contingency plann.ng and systems. see also
8 It is emphasised that insurance transfec cost, not risk (Vivian, 1990).

should be noted. however. that by simply defintng the cast-of-ask as the sum of three constituent elements, the assoc-latec costs should not be ianor«d. For exam re. in
cons&ing expenditure relating to setf•fundjng and Insurance any administrative or procuæment costs must
FINANCING
for- There has been debate among risk managers around the question of risk cost apportionment, regard to expenditure on "5k conti01 measuæs. | 4 Depicting the cost-of-IA ekmentswithln a shape is, ofcourse. intention, however, is to the on whi,-h the
The question is often asked as to whether ail such expendtcure should be regarded as risk as opposed to some element being regarded as expenditure. The of the cost-of.fisk a deliberate risk control. of the the of
problem is jn fact intemea to reduce the frequencyor severity of loss should be included in tost-of*i5k'equation. Is housed. intends to signify the catastrophe element in risk situations, 'middle'part of the triangle refers to the element of bosses are reasonably and where
228 RISK self-retentim/se'f-finncinq is a more financing technique. regard to incidental such as of and agency the debate see tbe
footnote and the section dealing wlth cost-of-risk-
RISK 229
Figure 9.2 The cost-of-risk!4

While recognising the very cyclical nature of the insurance market (one parameter against which
short-term cost-to-benefit analyses are conducted)
and accounting for costs other than risk financing (e.g. costs associated with loss of productivity
and morale), it can be argued that risk control expenditure in the longer term is optimised once
Insurance
The 'feasible technological saturation' has
expenditure costofrisk
Expenditure/exposure through been achieved. That is to say, one should constantly direct attention to and make expenditure on
retention risk control within the feasible limits imposed by the organisation's operations, shareholders and
so on until, technologically, the effects of risk can be minimised no further.
Expenditure on risk control While it is recognised that such a policy may be criticised for being normative and possibly
idealistic, in the long term its effects on the selfretention and transfer of risk costs must be seen
as beneficial.
It is important to describe the cost-of-risk in such a way, for it emphasises that effective risk
management is based, firstly, on managing the physical risk situation through risk control
and, secondly, on directing attention to the most efficient method of financing. Given that Micro integration
the 'boundaries' that separate the three elements are not absolute or well defined, the
challenge faced is to combine or integrate these to achieve a long-run cost efficiency. Definition
The model also proposes a 'macro-integration' and 'micro-integration' process. Micro integration is the integration of the risk*financing options of self-funding and third-party
funding (insurance).
Macro integration
In organisations where size, resources and philosophy dictate that significant self-retention of risk
is indeed necessary and cost effcient, the question of determining the point-of-risk cost transfer
Definition
to the insurance market is a key issue for risk managers and risk consultants in practice.
Macro integration is the optimisation of expenditure on risk control within the context of
long-run benefit from the removal or reduction of risk. It is also not surprising that most academic thinking has centred around
this problem. lS When one considers the insurance market as a mechanism or market that
In this context, optimising becomes somewhat of a misnomer: in the short term it may be reallocates financial resource (supply and demand) with all its associated notions, such as capacity
possible to balance or optimise expenditure against benefit as measured against, say, a and transfer costs, the question of€optimal mix' becomes an interesting economic problem. 23
discount in premium, but longer-term considerations (including strategic and/or
The intention thus far has been to link the elements of the risk management model.
governance-related factors), as mentioned earlier, introduce added difficulties.
Reference to macro and micro integration aims to illustrate the earlier point that a holistic
23
A comprehensive discussion of composite hnanclag strateges can be fimjnd in Äerty (1985.327-61).
approach to risk management means integration across the model (risk control and financial) and
integration within it (insurance and self-funding).
9.3 Risk-financing mechanisms

This section is concerned with issues relative to financing risk. It is important to stress that when
risk-financing mechanisms are considered, it should not be
that S See rlsk chapter control 1 has 5, which dscusses Ilttk attention.and refers to insuonce purchasing optimisation rno&ts developed. A surveyciacademic Mitangs Indicates
MANAGEMENT: RISK FINANCING FINANCING
230 RISK In adapting the basic risk management model to take account of risk financings we must
consider in which setting to place risk financing and risk management
thought that these refer only to insurance or self-funding. We can categorise financing
mechanisms broadly into two strategies:
(i) pre-loss financing (through insurance, retention funding, and hence some form of composite
funding or 'mix') Chapter 3 deals with finite and spread-loss insurance as alternate ask (ART) techniques to these of risk-financing mechanisms are also covered. Often these do not
qualify as insurance as elements, e.g. sumclent Sisk transfer to the Insurer, may not be in evidence. 19Doherty (198%272).
RISK 231
(ii) post-loss financing (through, for example, cash resources, debt or equity finance).
as a whole. This setting entails the understanding and agreement of corporate objectives or
While we refer to two broad strategiess note that, in practices depending on the class of risk
criteria to be satisfied when risk management decisions are taken.
being financed, a combination of pre- and post-loss financing is used. The burning-cost insurance
Not only is this necessary for adding direction, it also represents a point of criticism of
and retro-rating plans are examples of a structured combination where a deposit insurance
the conventional thinking on the subject of risk management, which traditionally has been
premium is payable, the balance being calculated according to a formula dependent on the loss
17
made from an insurance point of view rather than from that of the wider base of financial
experience. Spread-loss insurance also represents this form of financing. Here, the insurer
management.20
finances the loss, and the pay-back burden 18 (in the form of an is spread over a period of time.
additional premium) The development of financial management has prompted a definite corporate
objective. The firm is assumed to be a profit-seeking organisation owned by the holders ofits
It is true to state that in the context of risk management, however, consideration is mostly
equity. It is furthermore assumed that owners' welfare is measured by the value of
given to pre-loss financing and particularly to insurance. Although there is often a logical case
ownership claims thus making the appropriate corporate objective the maximisation of the
for adopting a post-loss financing strategy, particularly when it can be shown to be better suited
total value of the firm.2i
to satisfying the company's objectives (shareholders' wealth maximisation), practical
considerations that may have no cost foundation make pre-loss financing (principally through The use of value maximisation as the corporate objective is consistent with the agency-
insurance) the most chosen alternative in practice. I type relationship that exists between the shareholders and managers of the firm. Paid
managers therefore have to satisfy the aspirations of shareholders, but may indeed also have
In the following chapters, we focus attention on issues related to preloss financing. Overall,
other personal objectives that exert forces in other directions.
insurance, self-insurance and composite funding (incorporating the question of self-funding—
insurance integration) require investigation. This aspect is an important one when viewed in the practical context. When one
considers invested human capital or the 'employment risk' faced by such managers (a risk
that is not easily diversified away), managers understandably adopt a more risk-averse
9.4 Risk management within a financial decision-making attitude when taking corporate decisions. 24 This factor adds a further dimension to the
reasons why insurance is purchased and could also explain the more conventional
structure orientation towards insurance as a financing technique.
Nevertheless, the shareholders' wealth maximisation objective does provide a measure
Corporate objective of maximising shareholders' wealth of performance. lhe share price of the firm represents the capital market's assessment of its
expected future earnings, given available information. Thus, the wealth maximisation
24
indicate that risk managers adapt more risk-averse postures in making corporate decisrons [see chapter 2).
objective implies that decisions are taken that will seek to maximise the size ofthe firm's be termed 'post-loss financing', since the funding arrangement is set in place after the loss has
future earnings stream, or its corollary: to minimise its level of pure or specific risk been experienced.
Conversely, the purchase of insurance, or the establishment of an internal reserve fund
from which potential future losses can be met, is termed 'pre-loss financing'.
Interestingly, apart from 'undeliberate' post-loss funding resulting really from
20 The practl%l application Ot risk management entails the integration Of varöus subject disciplines. Some may argue that risk runagernent stoutd fatl within the
disclptine Of financial manayrnent, others who adopt mole behavmra' view Stress che 'management' jn risk management. An arternauve view is that rlsk management 'undeliberate' self-retention of risk, most risk management financing strategies are pre-loss in
is technological in nature. losses being srde products of the technological O tocesses carned out by the firm. The view takeo in this text is that. regardless orientation,
projects are assessed agatnst corpoate obJectives.finarw management simply offers a decision frameworkWith a perspective that does not compete With alternative
subject areas In its claims risk management. Rather, it should be perceived as cornpkrneatary. Most risk managerrkr-tt thinkrng and writing has envged from an
nature.25 It is doubtful whether there is meaningful discussion in practice as to the optimal
background, where little attention has been dlrected toward the motivating objectiveofSisk management activity to rnaxmse share
source of finance. It is suggested that this is partly due to the non-integration of the risk
v21 alue. It Should A few be Wiiters pointed have out introduced that the value this thinklng; maximlsation see o*cüve ty ( 1985) implies and 3 Olson risk•adjusted and SirnklSS
management objective with that of the firm — and partly because of inherent objectives and
k•nger-terrn (1 982), view.This IS necessary •n order to reconcile the value maximisation objective With the miccæconomic behaviour of the firm, where typicaJY the firm Is assun-
»edto be a wealth maximiser.

security considerations evident in risk managers (a point that was made earlier).
At mni rofi
p t misatiön maxjrmsation and the with Is lattes a rnoxirti$ing appear muttl-penod is to broadly maximise to be profit coma risk.djusted contljct equity pat•ble over of a value with sångle
Objectives analogue sharehowers'Ho»vever, period, betæen of Single li and one wearch generally the peraod considers theory maxiprofit rnisation.free of that, the maxirnisatiori{Dohetty, oi
for consideotions firm the and most financial Of rnrcroeconomic management. risk. 1985. 17). then "It Thus, may The analysis be tte forrner coxivded ecoromic has seeks been concept that 23 The crlterla fix mak.rq investment iSions are the class oftecbniques known as discou cash flow (DCF). CCF cuteria are not speclfic 10 any class of investment decislons
rnaxrrnisestatic, valueI.eof a nd may be said to have equal validity for risk management deasions- Under OCE tie cash flows (both poe.iti•.e and negative) resuhing from a are capital Ised at an
appropnate discount rate, tfthe capitalised value is posulve, accepance offbe poject increase the vajue of the firm. The project should not be accepted if the capitalised
232 valve is negative, since this decreases the value of the
ownership and revenues clalms. can be accepted Is not that the EMV of slnce the Insurance in most Instarres purchase perfect declsion infon-nation must be negative,
is noq available. Application It becomes of DCF even requires That dtffarecostsr-ul[ are quantified. This an easy task* are pobeb•llsdc in should be that such dficulties
with ask decislons, because natLlTe. Neverthekss,the polnt

Rationalising along these lines, and assuming that risk management is to be encompassed
must take into account any risk effects and spillover effects on (he costs of other funds, as b e 24 Naturally, the cost
within a financial management decision-making structure, the same objective can be adopted of capital calculation ev•dent. Hence. debt financing may be return nominally cheaper insurance for the exarngE. cost ofequity.but debt may Introduce added financial
isk
leverage r that in turn increases required rates of to sharehdders and therefoæ
for risk management.
25 Self-retentlon of risk is the result the of risks not formally Identified or evaluated finance Unckr rhese corditions. hence (he firms termtjndeliberate' accept the risks
postfor obSwn
account. perhaps non-deliberately. In event of losses arising fro-n qnh risk sources. is sought.
financing.
Financing and investment considerations RSK 233
Defining the risk management objective therefore differentiates the financing decision from
the investment decision. Decisions concerning the sources of funds (e.g. insurance and self- Turning now to the investment decision, as it pertains to risk management, let us consider
insurance funds) are referred to as financing decisions, while those involving the use of funds three scenarios.
are referred to as investment decisions. The first concerns investment in the replacement of destroyed assets. To illustrate this, a
As with an investment appraisal, well-developed criteria exist for comparing competing firm working at surplus capacity would view the reinvestment decision in a different light to a
sources offinance (an evaluation that risk managers should logically effect in accordance with firm that, for example, was working at full capacity. The latter firm would be more adversely
the finance decision introduced). 23 If we remember that the principal objective is to maximise affected by a loss of output (hence revenue and profits) if a resource were destroyed.
the firm's share value, 24the source of finance should be the one that leads to the lowest cost Parallel to the finance decision, the reinvestment decision is often ignored. As Mehr and
of capital. Hedges26 point out, the omission may be explained under the assumption of survival and
continuity of operations being corporate objectives.25 Nevertheless,
Consequently, this criterion introduces to the risk manager added sources of finance
other than insurance or deliberately retained funds to finance the self-insured exposures. The
even when continuity is not assumed explicitly, a high emphasis on insurance to
use of fund sources such as new issues of debt or equity and term loans or lines ofcredit can
finance losses has a similar effect ... the goals of survival and continuity are not
25
exßanation could be [he of Indemnity- The very intention of insurance to restore the asset to its pre-loss value The ir•etment decision is möde almost by default as part of the 29 in some instances. e.g. when leglslation Is taken into account, tte Investment may be compulsory.
decjsion to purchase insurance.
28 Ooherty (1 985: 28-9).
MANAGEMENT: RISK FINANCING REK FINANCING
necessarily compatible with value maximisatjon, and the decision to reinvest A third investment consideration refers to decisions regarding contingent liabilities. Liability
js not necessarily a rational economic one.28 losses do not relate to the destruction of assets of the firm, and it is the third parties, rather, that
suffer damage. The firm, however, acquires an obligation to pay compensation.
Yet the decision to invest is not inconsequential if the firm is to pursue a strategy based
Nevertheless, such contingent liabilities require investment decisions similar to those
on value maximisation. involving the destruction of corporate resources. Investments may be made with the purpose of
The second scenario concerning investment decisions involves the reducing the expected accountability and burden of such losses. "Ihe benefit of such investments
evaluation of risk control expenditure. It can be argued that such risk reduction lies in the reduction of financing costs of the liability payments.
investments have identifiable cash flows and that therefore the investment Ihe more difficult problem is whether a liability loss threatens the continuation of the firm's
decision can be assessed along the same lines as one would assess any investment operations. In this regard, it is not dissimilar to the reinvestment decision that arises from the
loss or destruction of the firm's resources.
project or proposal.
Although a liability claim does not directly affect the ability to generate future earnings, it
With risk control investment, the pay-off takes the form of a reduction in the
can affect the firm's ability (and indeed willingness) to continue with its operations or part of its
cost of contingent losses: if the capitalised values exceed the capitalised costs,
operations.27 Depending on the source of finance, the value of ownership claims may reduce
the expenditure is justified.
substantially; should retained earnings be used, for example, the value for shareholders would
Here it is necessary to make certain important qualifications, however. reduce by the amount of the liability claim.
Firstly, an investment decision assessed under the wealth maximisation criterion
However, there are also indirect costs. Product defect and community resistance, for example, may cause
assumes implicitly that a long-term horizon is being contemplated. Il-ms, for the
the demand for the firm's product to decrease; while employee consciousness of work hazards and morale
assessment to be of value, with such a time horizon, all benefits and costs -
may cause production costs to increase. Such effects may prompt the firm to cease operations after a loss,
including indirect costs (and savings) arising from such factors as, for example, although production was profitable before the event.28
employee consciousness of work hazards and worker morale — must be
accounted for. This in itself becomes a diffcult task.
Secondly, if this type of valuation criterion is accepted, the investment The risk management model within a financial management decision-
should also logically be evaluated in the light of other competing29 and, most making structure
often, mutually exclusive investments. This is particularly relevant to risk control
Risk financing has been evaluated in the light of corporate objectives. The discussion focused
investment proposals, where the capital outlay is quite significant and hence
on the issues of investment and finance decision-making.
comparable with alternative possible investments.
26 Nehr and Hedges Q 9'4: 2).

RISK 26
235
Often, this evaluation renders risk control investment sub-optimal. Yet in the long run, The primary reason was to avoid the danger of isolating risk management, or risk financing, from
such investments could safeguard the firm from disaster. Hence the statement made earlier the general finance and investment disciplines.
in this chapter that firms should ordinarily achieve an 'operating effciency' level through It is worthwhile summarising the warnings drawn from this evaluation, since discussions
investment in risk control; such a level should be within feasible bounds dictated by overall in later chapters must be seen in relation to such evaluation and subject (where applicable) to
organisational and financial constraints. these warnings. Moreover, they form the basis for adapting the basic risk management model
given in figure 9.3.33
26
If. for e,arnpie. liability insurance is in place. the benefit may rake the form of reduced prem.ums given the of to being jeopardised. questioning by firms of continuation of operations along the lines as the destruction of physical mav
indeed be the catalyst needed to effect appropnate
change in the legal system and in the litigious nature of some societies.
27
A number oi US have withdrawn perfectly products from the market as a result of but unsuccessful. liability clang.
28
This argument comes significantly Into focus whenviewed against the 8, which discusses the l,ability crisis. Even clans may be financed by trough insurance. which
may have a of capital lower than, for tera,ned earnings. insurers' results have forced such large increases memiums thu Rhe continuation of operations still nevertheless
• Any risk management structure should not support an inappropriate measure of the costs of
loss events. The cost should not be assumed to be replacement value necessarily, but
considered against the impact of any loss of earnings stream.
lhe productivity of resources employed by the firm is central to the reinvestment decision.
Emphasis on corporate risk aversion and the supposed desire for continuity of production
should be evaluated in this light.
There is a distinction between those decisions that involve investment and the funding of
such investments, i.e. financing decisions. Insurance is a source of finance, and failure to
identify the nature of financing and investment decisions leads to alternative sources of
finance not being considered. The approach to the handling of risk should be an integrative
one. Considerations about risk control and its management bear on risk financing.
Furthermore, risk should not be considered simply as an 'additive'; possibilities for
diversification exist, with risks being spread over various risk management exposure units,
for example, and even by the firmis shareholders through portfolio investment
diversification. 34
Given the warnings drawn from the evaluation of risk management against the corporate
objective of value maximisation, figure 9.3 represents the final adaptation of the risk
management model first given in figure 9.1. The differentiation between investment and
financing decisions further clarifies the process of integration:
On the macro level, investment in risk control activities is considered with due reference
to the (long-term) financial investment decision, i.e. whether post-loss reinvestment or
abandonment should take place.
Should investment in risk control be decided upon, the model then directs attention to the
financing decision (post-loss reinvestment option) and to the question of the micro
integration of the pre-loss financing means of insurance and self-funding.
å3 these warnings follow closely che Oltiqtre by Doherty (1985:35—8) of the w-called tradltionai itsk management structure. The model is a
34 w?ht. the adaptation warnings of tmpiy that the proposed existence by Doherty of nonefficlency (T 985:32).should the suggested decision structuse not oe adopted.
To the extent that these exrst, then. theoretically. such onortunlty costs should be inccwposated into the total cost of risk equation. The rack of discussion costs in risk
management texts and aty:ks on •cost-of-rise Is suggestive of the non-finance orientation of traditional rbk arogemertr-
RISK MANAGEMENT: RISK FINANCING
RISK FINANCING 237
236
9.5 Summary
Figure 9.3 Risk management model within a financial management decision-
making structure the physical and financial elements of risk management and to investigate the
underlying rationale (financial) for risk management, with the view to modifying
the risk management model by placing it within a financial framework of decision-
making.
By introducing the concept of 'cost-of-risk', attention was focused on the
aspects of integration and optimisation. It was held that these could be
approached on a macro and micro basis. Relevant conclusions reached with regard
to cost-ofrrisk optimisation were as follows:
A longer-term view is necessary when determining levels of loss control expenditure.
Risk control should be undertaken within the broader corporate financial objectives and
constraints.
Decisions should be taken with due reference to investment considerations.
The broad categorisation of risk-financing mechanisms into pre-toss and posttoss

financing was discussed. It was concluded that in a practical context, micro
integration entails the optimum use and mix of pre-toss financing sources/
€0
strategies, i.e. insurance and self-insurance.
0 co o The placement of risk management within a financial decision-making structure
required the adoption of the corporate objective of maximising
shareholders'weatth. An important difference was drawn between the financing
and investment decision.These discussions led toa summary of risk management
warnings that formed the basis for adapting the risk management model to one
O that now reflects risi«financing considerations.
238
Chapter 10
Risk retention
Source: Adapted from Doherty {1985: 32)

This chapter dealt with the element of
risk financing in the risk management
process. The major objectives were to
describe the interrelationship between
RISK MANAGEMENT: RISK FINANCING RISK
10.1 Introduction chapter Il for a more detailed discussion of captive insurers). Along this continuum, other
10.2 Unfunded retalned risk
programmes have developed such as structured insurance programmes (SIPS) and cell-
10.3 Funded retained risk
10.4 captives or rent-a-captive, which are captive facilities offered by insurance companies.
Implementing a retention
10,5 programme
10.6 Evaluation of retention funding
10.7 using the value of the firm criterion 10.2 Unfunded retained risk
10.8 Case for retention programmes
10.9 In insurance terminology, it is usual to refer to potential losses that are not insured as
Other practical considerations and
retention program mes 'retained losses'. Retention takes place when a risk cost is ignored, or when the risk
Deductibles situation is identified and evaluated, but a decision is made not to fund it externally. Seen
Summary
from the position ofthe risk manager, a retained risk may be funded or unfunded. A risk is
retained and unfunded when no specific provision is made for the financial consequences
of a loss.2
10.1 Introduction Based on this, and bearing in mind the array of risks facing a business, not funding for
risk is the most common risk-financing method. Economic or speculative risk is included in
Ihe risk management model shown in figure
the array of risks facing the organisation. If one accepts the descriptive parameters of end
1.1 makes reference to risk retention. If the
and intermediate economic risk given in chapter 12, then of course the provision for end
cost of a loss is not funded externally, it is
economic risk would serve no purpose. If such risk could be provided for, there would exist
retained. This chapter discusses retention
no justification for profit, as profit presupposes the assumption of risk by the risk bearer.
funding in greater detail. To do this, we
examine the meaning of risk retention? The Unfunded retained risks can be divided into two broad categories: (i) risks
efficacy of retention funding is then that are insurable, but retained (ii) risks that are uninsurable.
appraised
Generally, the first category is characterised by high-frequency, lowseverity losses. An
Practical facilities for retention range example of this would be each and every excess on motor losses. An insurer may impose a
from a retention fund (at one end of the
continuum) in the form of an accounting small excess of, say, RI 500 on each and every claim. Risk managers may decide not to make
provision or reserve in the books of a firm to any provision for these excesses, in which event these losses are funded out of the
a captive insurance company (as the
continuum moves towards external operational expenses of the firm. The second category would include speculative risks that
financing). A captive insurance company is a are usually purely financial in nature.
capitalised insurance company that
underwrites the risks of its parent Ihere are numerous reasons why certain risks are not funded. Often, the cost of
organisation (refer to
making provision to fund risk through insurance, for example, may far exceed the
1 It is more correct to refer to retention. but. toavoid unnecessarily cluttering the text, we use risk The risk of
damage as such cannot be The reader is also referred to section 10.3 This text the but subject to perceived benefit to the insured. In such cases, although insurance is available, the cost-
made. The used because used in the reinsurance to the risk retained by the direct
239 to-perceived benefit prompts the risk manager to retain

R'SK RETENTION
2 Derenberg et al. (1974:122).
The direct Insurer rewas risk it does not reinsure.
240 fund into which contributions are paid and from which withdrawals can be made following
a loss. Retention is sometimes referred to as self-insurance; it is preferable to refer to
funded risk retention, which is any plan of risk retention in which a programme or
the risk. In south Africa, the risk of an earthquake,
procedure has been set up to fund losses when they occur.
for example, is small, and accordingly many risk
The use of the term 'self-insurance' gives rise to a number of controversial issues. The
managers may decide not to insure the risk or
definition of insurance determines what is meant by self-insurance, since the latter implies
make any other provision for it.3
that it is a form of insurance. Insurance is defined by some as 'pooling'? by others as
Another reason why a risk may not be
'transfer'6 of risk, and still others in terms of both pooling and transfer.' Thus, for those
funded is because it may be impractical to do so.
who define insurance as pooling, the term self-insurance is an accurate description of
The risk of loss or damage of property from war
deliberate pre-loss retention funding. uhose who say that risk 'transfer' is a prerequisite
on land — being fundamental by nature - is not,
for insurance hold that self-insurance does not meet the requirement of risk transfer, since
for example, insurable and an organisation would
not ordinarily fund for losses from this source.
3 It events with be
On many occasions, however, risks are not
An example where event not insured materialised the destruction of both Twin Towers of the world
funded due to a lack of foresight, knowledge, Center in New York on I I September 2001. The likelihood of both being estimated not insuring joint possxbllity,
4 The way In which the fund is retalned is dif-used below.
care, judgement or a market. Since there is no 5 Kulp and Hall 10).
6 MO't&ray, glanchard and Williarns (1969:62).
ready source of information regarding 7 Bickelhaupt and Magee (1970:62)
RETENTION 241
uninsurable risks, organisations should establish
which risks are not covered or are excluded from
insurance contracts and compile a list of conceptually at least, the risk is kept within the same economic family. %ese arguments
uninsured or uninsurable risks based on this confuse the concept of insurance and the mechanism of insurance. Insurance is an
information. Insurance brokers usually have a list enforceable contract, while pooling is the mechanism that explains how and why
of the most common uninsured risks insurance can succeed in replacing uncertainty with certainty.
The existence of a funded retention programme might also lead to the illusion
that the process operates similarly to insurance and 'reduces' or 'eliminates' risk. 8 As
10.3 Funded retained risk Doherty9 points out, this misconception arises from the idea that the quantum or size
of the loss will be offset exactly by a payment from the fund. This creates the illusion
A firm may find that a particular type of risk
that the funded firm is immune to risk, where the losses are less than the fund.
consists of a large number of independent and
However, contrary to this illusion, with retention funding, the firm will suffer a decrease
homogeneous exposure units, and that the
in value following a loss. Before the loss, the firm had the asset and cash in the fund.
historical loss patterns display a well-defined loss
After the loss, the firm has only the repaired or replaced asset and no cash. The cash
distribution. In such cases, aggregate annual
flow caused by the loss is simply given a different accounting label. Cash as an asset is
losses can be fairly accurately predicted and
paid out to repair or replace the damaged or lost asset, cash decreases, and the value
hence the annual costs can be estimated. This
of the relevant asset increases by the same amount.
amount can be prepaid into a fund and in this way
The practice of retention funding should not be regarded as insurance,
predicted losses can be economically retained. 4
notwithstanding that pooling may indeed exist — and particularly where the balance
This, then, represents another form ofpre-loss
of the retention fund is considered an asset in a firm's accounting books.
financing, i.e. the establislment of an internal
The position should also be considered
where facilities offered by commercial
insurers are used. In this instance there is
(transfer' and a limited risk transfer to the
firm- IO These facilities involve the payment of
a premium to the commercial insurer. While
the nature of the premium may be argued,
the firm does not regard the fund or any
residual value that may accrue to the firm as
an asset. In this case, the value of the firm is
not affected by individual insured losses.
It is diffcult to estimate the extent
ofretention funding. A survey conducted
some time ago by Goshayll revealed that the
majority of firms did not fund retained risks.
However, with the passage of time and the
changes brought about by the development
of risk management, it is doubtful that the
results of a similar survey would be the same
today, L2 and it can be accepted that the
retention levels by firms have significantly
increased. It is estimated that there are in
excess of 4 000 captives active in the world,
of which about 2 000 are single-owner
captives and the remainder are joint
interest/multi-owner/
a
rtuality. mere can be ro real reduction or eliminaton Of Cisk Of damage to
assets through the medium Of insurance, Refer to discussions thrs
issue in chapter 2- Insulance does not reduce or elimhate risk and
there is no risk transfer to insurer as such. but rather the to pay for
the financial consequences ansing from losses incured.
are
briefly
discussed in
section 10 7,
which
examines and
insurer, 0964:
126}.
'2 Doherty (1985:292)-
MANAGEMENT: RISK FINANCING RISK RETENTION
accessible to finance losses as they occur, then some be revealed. that have resaves or to such should
242 conduct an evaluation of the
14 Although be saving In costs. the is greater. so that the of bad Loss the may Indeed witness higher cost than would have applied insurance. It is customary
against to combine levela being surpassed In insurance be excess loss and aggregate
association-type captives. Some 1 500 captives are domiciled in Bermuda, the original and 243
probably the best-serviced captive domicile. other popular domiciles are Barbados, Bahamas, the
Cayman Islands, Guernsey, the Isle of Man, Luxembourg and various states in the US, with
smaller numbers in Hong Kong and Singapore. number of new captives registered in the five Determining the opportunity for and feasibility of retention
largest domiciles appears in table 11.1. In money terms, it has been variously estimated that funding
world-wide gross premiums to captives are nearly $10 billion and gross investable assets are of The first step in developing a retention programme is to determine the riskfinancing techniques
the order of $15 billion (1987). available to the risk bearer. Although in the wider sense this would involve an analysis of post-
There does not appear to be any agreement in the literature as to the purpose of risk funding. loss financing techniques as well, it is generally implied that the risk should be evaluated from
In general, funding is associated with financial stability and plaming. The motivation for retention an insurance availability standpoint. The risks faced by the firm should therefore be classified
funding is similar to that applying to unfunded retention. There are two mdamental reasons for as insurable or uninsurable risk. This will naturally reveal the feasibility and opportunity of
293 funded risk retention in comparison to insurance, and will facilitate any evaluation of the
retention funding. Firstly, it may not be possible to fully 'transfer' the risk through insurance
or any other device, and the situation may be that there is no market for a particular risk. The retention strategy.
second is that it may be less expensive than conventional insurance 'risk transfer' techniques. This When embarking on a course of funded risk retention, the firm accepts for its own account
reasoning applies particularly to what has been termed high-frequency/low-severity risks where the responsibility for the financial consequences of (pure) risk. As a second step in the process,
conventional transfer' devices such as insurance can prove to be relatively expensive, given the therefore, an evaluation of the firm's risk retention capacity is essential. The firm's risk
cost structures of commercial insurance companies. In these circumstances, retention funding retention capacity, or the limit of loss-bearing capacity, should be gauged by the degree to
permits substantial savings in insurance premiums, with an opportunity to reduce loss costs and which the firn* cash flow can be diverted from its usual uses to meet losses without
the motivation to prevent losses by implementing sound risk control programmes. 14 significantly disrupting normal operations.
10.4 Implementing a retention programme Analysing the statistical characteristics of the retained risks
The characteristics of risk retention are similar to those of insurance, so that many of the risk-
Practical steps in implementing a retention programme financing choices facing the firm are similar to choices facing an insurer. Among these are the
Several practical issues must be considered when implementing a retention programme. These edent of retention funding for the firm and the size of reserves. In order to make decisions on
include the following: these matters, an analysis of historical loss patterns is necessary in order to estimate possible
future loss costs. The statistical analysis involves determining loss distributions and using these
determining the opportunity and feasibility of retention funding
to predict the likely future costs of losses.
• analysing the statistical properties of the retained risk
As a general rule, risks displaying high-loss frequency with relatively low severity permit
• evaluating the appropriate retention strategy. i6
predictability, and hence are amenable to retention. It is important for the risk manager to
establish how a retention fund will stand up to claims. He/she must, therefore, assess the
probability that the fund will be insufficient to meet these demands. This involves determining
13 ARhough reference is mode to basic incentives, there are, of Othel incentives to retain risk It may be vortiuvhik to mention that on occasion.
especially with strategic industries. bea reluctance to to an Insurer the necessary to setting. one finds the ruin probability of the retention fund.)' Similar to protecting insurance reserves, the risk
government and little incentive to finance losses other than through Ifthe motivation for such an Is risk-financing such costs
with frequency lossesthen the has obvious merit. However, if the is based the that substantial is available manager must protect the retention fund against ruin.
29
pr.ctice. onefinds various risk retention capacity fomt'lae in use. one view is that risk retent\on capacity should be measured largeiy in Qi the creation and which theo determines deglee ofpredictability the absorute also risk mea retentk'nns that capacity in effect level the (Gibson. advantage 1983: i—S).Otherwise
availability of liquid reserve funds — malnta•ning that prudent poky requtres assets co be Immedl&tely available to loss costs. Others assign weights tofinancial evidenc with Insurance becomes'redundant'tQttveby IOW rlsk allows me firm to retain aod internally finance these losses without incurång the higher edmifitstratlon
ratios (l•quidity. profitability and even let.eage) and argveata welghtedaveragecomputation,
costs Of ttE insurer
7 'Oin probability of a rerentvon fund may be Jehned as the that the .ggregate retained losses a given pecod exceed the the fund in that period (Doherty.
1985:306).
RISK MANAGEMENT'. RISK FINANCING RISK RETENTION
244 of the continuum a simple internal retention fund, and, at the other, the setting up of a captive
insurance company. Various factors influence firms' decisions regarding the level of
sophistication required of their retention programme — this aspect is examined in section
Evaluating the retention strategy in comparison to alternative 10.7.
financing techniques 245
frie third step is to compare the funded risk retention strategy in comparison to the alternative
risk-financing techniques of insurance. Basically, this process 10.5 Evaluation of retention funding using the value of the
begins at one extreme by comparing the total cost of risk under a retention strategy with
firm criterion
conventional insurance cover. The intention is to determine the trade-off between risk retained
Retention funding is simply a means ofproviding finance from current income or reserves
and risk transferred.
created out of current income to pay for losses. The opportunity costs of retention funding are
It must be recognised that, notwithstanding a high degree of loss incurred before a loss takes place. Contributions to a retention fund are akin to insurance
predictability, it is obvious that the retention fund will not cater for catastrophic losses. In this premium payments. As with other forms of financing, the method of evaluating the superiority
of retention funding is based on the effect that funding may have on the value of a firm. This
area of risk, the characteristics (severe, but infrequent) are by nature more amenable to
criterion dictates that, should funding have a more positive effect on the value of the firm than
insurance where the large exposure base will ensure that the pooling will facilitate financing.
other forms of financing, it is to be preferred over these other forms. In order to evaluate the role
Hence, a retention strategy does not usually imply non-insurance, but rather selective of funding, we must first examine how the value of the firm can be determined.
insurance, probably on the basis of excess of loss or aggregate stop loss protection.
Encompassed in the objective of total risk cost minimisation, therefore, lies the question of the
Value of the firm
optimum integration of retention with insurance following a risk-return analysis.
The value of the firm at any point of time is:
Determining the optimal trade-off between insurance and retention faces
the sum of the value of operations, plus the value of
certain diffculties. Firstly, the cost of retention can be predicted, but cannot be known in
investments.
advance. Only the passage of time can determine the exact cost
and hence the possible benefit of retention. By then, the decision regarding retention will already The value of operations is determined by discounting the net cash flow from operations as the
have been made. Secondly, for many risks, it may be difference between the weighted average cost of capital and the expected growth rate in net cash
difficult to determine insurance costs under varying assumptions of retention and, indeed, flow. The value of an investment is usually taken at market value, since the return on an
more diffcult to determine future insurance costs. investment is assumed to be equal to the cost of capital, and is referred to as a zero-net-present-
Nevertheless, it may be possible to analyse insurer costs broadly in order to value investment. This implies that the discounted value of the cash flow equals its market value.
estimate possible opportunities for savings. Given that retention is considered for those risks,
or sections of risk, where loss predictability is reasonably accurate, insurer expenses become Role of funding
a major factor and might be accurately assessed; under assumptions of required profit to
It is necessary to clarify certain issues that are central to the understanding of the role of funding.
30
insurers, it is possible to construct 'what if' scenarios and direct a risk-return trade-off
analysis.
Size ofthe fund
Finally, having decided to retain the risk, the last consideration is to determine the
Firstly, if the fund is too small to cover the losses, any loss will have to be borne by the firm
appropriate fund retention facility. Earlier it was indicated that one could consider at one end
as an unbudgeted item. This results in shifting the burden towards other forms of funding,
30
Very sophisticated costs Of models nee have arwd expected been deveboped values Of to pure asslsr prenalurn in trade-off li\ely analyss. to be charged One such by This is discussed in chapter 15.
a
insucers. technique assuming uses simulation varying levels theocy of "Skto expected
20 Doherty 300)-
including external financing (e.g. debt) with an associated increase in the cost of such finance
247
and transaction costs. Secondly, accepting that the intention of establishing a fund is to optimise
the risk management activity and not to make a profit, the retention fund can be regarded as a
of the fund. Value will only be added in the event of the risk-adjusted return on the fund
zero-net-present-value investment. If the fund is too large, it 246
surpassing the risk-adjusted cost, i.e. if the fund is not a zero-netpresent-value investment.
The significant conclusion2t is that 'the case for establishing a fund is entirely
reduces the value of the firm, since excess funds could have been invested in operations and independent of the firm's loss exposure — hence, there is no risk management case for
earned a return in excess of the cost of capital. This implies that ideally there should be an establishing a retention fund under the assumption of no transaction costs'.22
optimum fund size. Doherty's analysis gives rise to the following modifications to the considerations for
establishing a retention fund, outlined in the previous section.
The first step of determining the opportunity and feasibility of retention funding assumes
Secondly, there is the issue of transaction costs. This is arguably the most significant aspect to be a more significant profile. The investment of resources in fund assets transcends the mere
considered when evaluating the case for retention funding. If a large loss occurs, this entails comparison with insurance; the feasibility of funding is dependent on the fund's expected
raising new external debt or equity, which gives rise to issue and underwriting costs, and may also return in relation to the firm's hurdle rate or cost of capital.
be associated with time delays and hence interruption costs. These delay costs should be regarded The statistical analysis, shown as the second step, retains its significance only insofar as
as part of the transaction costs. it claims to provide an estimate of expected loss costs to be borne by the firm. Since the
case for establishing a retention fund is not influenced by the value of losses, the
statistical estimate will not be a factor in determining the value ofthe fund. Should the
The effect that funding may have on the firm9s cost of capital is the third issue to consider. criteria dictate that investment in fund assets is favourable, other factors would determine
Funding merely shifts the cost to another accounting label, and the firm is exposed to the same the fund's assets value.
riskiness of the retained losses. Since the investor will discount the cash flows at the rate
process ofevaluating the retention funding strategy remains, in concept, an important
determined by the systematic component of the risk, the loss beta value19 is important in
consideration. However, rather than evaluating the funding strategy, the process depends on
determining the cost of equity.
evaluating risk retention as a risk management strategy in comparison to alternative risk
In the following sections, we present Doherty's20 views on the role of retention funding where: financing mechanisms (usually insurance).
the firm has no debt and alternative sources of finance do not involve transaction costs Finally, under the assumption ofno transaction costs, the results show that retention
funding has no risk management case for it. Hence, in a strict sense, it should not be formally
• alternative sources of finance involve transaction costs.
identified or defined as an alternative pre-loss vehicle or means of financing risk.
It is prudent at this point, however, to put Doherty's analysis in perspective. Paying
Evaluation of retention funding under the assumption of no insurance premiums also decreases the value of the firm. In addition, it can be assumed that,
transaction costs over the long term, insurance premiums will at least equal losses, which means that the value
of the firm will be decreased by the loss experience even when insurance is purchased. The
Under the assumption of no transaction costs, the value of the firm can be depicted as an equation
absence of 'frictional' costs, the effect of a possible favourable loss experience, the timing of
where the firm's value appears as the sum of: the value of the commercial non-risk management
cash OUtflows and the possibility of earning more than the cost of capital on the retention
activities of the firm and the value of the retention fund, reduced by
fund make retention funding more appropriate than insurance from a value-of-the-firm
• the value of the firm's loss exposure.
perspective, even in the absence of transaction costs.
The case for establishing a fund then exclusively on the relationship between the risk-adjusted 300}.
22 fir
m may choose to hold finar-xbl assets In the Erm of@fund, but it will be rational to do so only on rhe basls Of Investment return those assets
return on the fund assets and the risk-adjusted cost
248
19The beta is a measure ofthe systematE risk of a company.

Evaluation of retention funding under the assumption of specifically earmarked for reinvestment may facilitate recovery. In addition, the existence of a
fund following a sizeable loss may avoid any market nervousness and hence unfavourable issue
transaction costs commanded by alternative sources of finance
terms that might have prevailed had the firm been required to raise new external finance.
As mentioned earlier, post-loss sources of finance may have associated costs such as
Transaction costs arise when the loss exceeds the assets of a retention fund and the firm is
underwriting and issue fees and/or cost of delays.
forced to consider alternative sources of finance for the purpose of reinvestment. It follows that
Introducing transaction costs to the question, the value of the firm emerges as the sum of:
the larger the retention fund, the smaller the probability that an unfunded loss will occur and thus
the value from commercial (non-risk management) operations and the value of the
the smaller the probability of incurring a transaction cost. However, this benefit needs to be
retention fund (the value of establishing the fund and investing its assets net of the cost
balanced against the opportunity costs of maintaining the fund. Probably, since the investment of
of raising capital to finance the fund)» reduced by the loss exposure contribution and
these funds is subject to certain limitations (e.g. they may demand a high liquidity condition), the
the contribution arising from the prospect of incurring transaction costs for unfunded
return is less than the cost of capital. Consequently, a retention fund must be viewed in a similar
losses.
light to the management of the firm's general liquidity, where an optimal level implies a trade-off
between the cost of maintaining liquid resources and the benefit of avoiding cash embarrassment
It is evident that, with the likelihood of transaction costs, the value of the fund is dependent (or transaction costs in the case of pure riskretention funding).
upon the loss exposure — thus, there is a risk management case for the establishment of a
Maintaining an optimal fund level introduces two further important qualifications. The first
retention fund. concerns the fact that, generally, retention funds are regarded as being separately earmarked for
An incorrect argument put forward for the establishment of a fund is that it represents an the purpose ofproviding for losses. Indeed, this would facilitate recovery from losses, particularly
orderly method for financing the capital demands induced by a loss, and that therefore the firm if interruption losses are a feature. In a sense, the immediate availability of corporate liquidity
is not exposed to any sudden financial demands. The argument, however, ignores the economic implies a form of de facto risk management.
costs or opportunities facing the firm. Should the fund opportunities induce investment, the However, ifretention funds are to be seen as separate from the firm's general funds and if no
investment would ordinarily be made. The investment decision would be taken independently transfers are permitted between the funds, then the funding requirement to maintain a given
of any risk management motive. The fund itself does not protect a firm from the random probability that either or both funds will be ruined is generally higher than the requirement to
incidence of loss. When such a loss occurs, the firm will lose value, the result being identical maintain a composite fund at the same probability of ruin.24 The contention now is that any
to that where the loss was unfunded. demand for liquidity should be met jointly with other demands for corporate liquidity, and the
The position can also be seen in the following light. The establishment ofa retention fund risk manager should cooperate with the other managers, each recognising the composite nature
is no more than bringing forward the date on which financing is raised for the purpose of of demands likely to be made on those liquid resources.
reinvestment following a loss. The case for bringing such a date forward rests purely on The second qualification concerning the optimisation of the fund level involves the
whether the investment of the fund assets yields a return in excess of the cost of capital. integration of the fund with the purchase of insurance. The dynamic nature of funding must be
Again, the firm's risk exposure and risk management motive have no bearing on the
recognised.
investment decision.
23 Delays can occur when insurers rely on purely defences to claims such as a lack of Insurable interest or the non-din-losu,e of
250
10.6 Case for retention programmes
A case can be made for risk retention on risk management grounds if retention avoids the
• Fund level optimisation criteria dictate that the level should be reviewed (usually together
costs imposed by other forms of financing. Included in these costs 249
with the insurance covers that relate to a particular period) and not merely be allowed to
accumulate asset balances that are more than required as estimated by any statistical
is the delay factor in securing alternative forms of risk finance,313 an aspect that may become
significant when one considers the effect of interruption loss. The establishment of a liquid fund
31
; Consida th.s statement in general. If separate funds were establshed with values X and whæh wee then cocnbioed into one fund * the ruin cannot rise,
Since each opportunity that av311abfe before the funds wefe combined still ums. The combining funds 'n fact achieves a •pool,ng effect' - the combinauon of funds
has ony g,ven additional fol Intert,ansfers. Hence, the
methodology. The latter strategy would, probably, introduce inefficiencies from the point interpreted in the narrower sense. Aspects of security, and job evaluation and human capital
ofview of the opportunity cost of holding the fund. considerations effectively focus risk managers' attention on the specifics of funding and
Earlier, mention was made of the fact that the fund's objective is not to provide for insurance. Ihis means that the motive to accumulate fund assets is strong, as is the tendency to
catastrophic risks, and for this purpose selective insurance would ordinarily be purchased. design retention funds so that the fund is not labelled or seen as an asset.
Since the insurance market is itselfdynamic in nature, the optimal fund level would also be Any separation of the risk management activity from the general financial management
influenced by the riskreturn trade-off between retention and transfer. Liquidity activity of the firm is disappointing. A central viewpoint expressed in this text is that the function
optimisation considerations must therefore not be isolated from those concerning the should be 'elevated' and integrated with the other activities, which not only include finance
integration of retention and insurance. The composite nature of the fund in comparison to activities, but all those that by their nature influence the risk profile of the firm. In the final
the other demands on liquid resources should be recognised, as should the composite nature analysis, it is exactly this aspect with which the risk manager is concerned. A general appreciation
of the fund in comparison to any insurance coverage. of the risk management function and the process of integration (as suggested by the risk
management model) must have the effect of reducing the tendency to see risk management in
isolation and reduce any associated ineffciencies.
The question of taxation raises further interesting debate. Establishing a retention fund
10.7 Other practical considerations and retention
should not be inspired by any favourable tax effects as such, since this motive has no underlying
programmes foundation in risk management. But this is not the same as saying that risk managers should take
no account of tax-effcient means of retention funding that may be available. This question
prompts a discussion of retention facilities/plans that have been developed to facilitate risk
Practical considerations with regard to retention funding
retention through the medium of an insurance company.
An evaluation of risk retention would be incomplete without a discussion of certain practical
considerations that, as will be seen, influence the way funding is dealt with in practice. Mostly,
these considerations revolve around the issue of taxation and the personal objectives of risk Retention and insurers
managers.
The origins of these programmes may be found in retention facilities defined as rent-a-captives.
Amounts raised to establish a provision in the books of account of the firm for the financial
The terminology is somewhat unfortunate, as these facilities cannot be regarded as captives (or
consequences of pure risk are generally not deductible for company tax purposes. On the other
mutual captives), since there is no requirement for the firm to capitalise, as is the case when a
hand, a premium paid to an insurance company is regarded as a tax-deductible expense. It can firm establishes its own captive insurance company. The rent-a-captive is not a subsidiary of its
be argued that where the retention fund is used to pay for a loss, the net after-tax position is parent. The intention, rather, is to use the facilities of an established insurance company to
identical, since the losses would be claimed as tax deductions.26 The situation therefore affects
achieve the objectives of a captive insurance company.
the firm only insofar as it implies an accumulation of fund reserves in a non-tax-efficient
The rationale is simply to pay as a premium to an insurer the amounts required to satisfy the level of
manner. The counter-argument, of course, is that the accumulation of 'excess' fund assets is in
risk retention.32 Instead of the firm's internally 252
itself undesirable from an optimisation standpoint, as debated earlier.
25 There to be an agency-type relationship between the shareholders and managers of a firm. Paid need to satisfy of shareholders,
but may indeed have other personal objectives chat exert forces in other directions
26 holds on the assumption that, following a [he ,einvesu-oent or expense relates to a revenue-producing i.e. an expense actually Incurred will be asa tax deduction if held retention fund financing the losses as they occur, the insurance company would settle the
the of income For a of retention, see Divaris (1988 131)
251 losses as insurance claims when these are presented to it.

Before describing the features of such self-insurance programmes, it is
28
The objectives of risk managers, however, are often perceived as being personalised. It desirable to examine the rationale and motivation for insurers to offer such facilities. The best
should be recalled that other factors ultimately modify the firmS objective. However, more starting point is the realisation that with the development of risk financing (coupled with other
importantly from a risk management perspective, the risk manger's function is most often factors), firms have begun to retain higher levels of risk for their own account. All the reasoning as
32
is easy to see why such a facility has been termed a rent-a•captive. In essence [he firm rents the insurance companywcaptivet 'Icence to Insurance business. There is have to capitatise its own captive; the '•nre:ated CM1pany does not have r.0 90 th rough the formalities associated with the setting upof a captive; the rent-a-captive
natually a facility fee pay&ble to the ensurer, which. theoretically. compensates the insurer for rhe stralfl Sdvenc,' 'esvlting from its writing the self-insurance prerniurni (telated items unrelated Insulance business (which the US context js good for its tax defence position); and the rent-a-captive {related company)
Kiornan and Rosenbaum 0982: 135) desorbe such rent-a,captive f"'litres -o yowned captives The advantages of such a scheme are that the unrelated company dees rot
r r
'e-ceives fees tor {he service. and usually some the investment income as well.
described in earlier sections of this chapter has encouraged a more active involvement by insureds 253
and, particularly in larger organisations, this has resulted in the restriction of insurance purchase to
catastrophe-type covers. on the contractual agreement between the parties, a portion of any excess of premium over actual
As a general statement> then, this has caused premium incomes to conventional insurers to losses may be refundable to the insured, or used to fund the following period's risk financing
drop as firms more emciently finance those risks or sections of risk where, because of their nature requirements. Thus, the hybrid nature of these programmes becomes apparent. 'Transfer' and
(i.e. the extent of their predictability and severity), the advantages of pooling offered by insurers pooling exist, and risk is assumed by the insurer.
becomes more 'redundant'. Interpreted in a slightly different way, insurers' inherent cost structures Apart from cell-captive or rent-a-captive facilities, a facility has recently developed where
disqualify them from providing an efficient pre-loss means for financing risk under these underwriters assist the insured to spread a particular loss over a period of time, thus smoothing
31
conditions. the impact ofa large loss on the financial statements of the insured. These facilities or
This realisation has prompted certain insurer organisations to differentiate risks according to programmes have been termed structured insurance programmes (SIPS) and have materialised
similar risk-retention criteria and to regard such highfrequency/low-severity risk situations as primarily as a response to the sophisticated risk-financing needs of today's risk managers who
'niche market' opportunities. Tailoring their risk-financing involvement in these situations, are seeking to achieve effective asset and balance sheet protection.
SIPs 32 enable the insured to manage losses in a structured manner through the following
insurers are able to offer such facilities at an economic price or premium.
Most self-insurance facilities are composite programmes. Ihe optimal integration of retention characteristics.
with insurance implies the 'transfer of risk' at some point. "Ihusi where the insurer facilitates the
retention programme and writes the residual risk, the policy may be seen as a composite k• Risk transfer
mechanism encompassing both the relatively predictable low-severity part and the relatively In an SIP, risk transfer can take the form of:
unpredictable high-severity exposure. 29 an underwriting risk (actual aggregate loss incurred exceeds the projection) a timing risk
(actual claims payments occur faster than projected) interest rate risks (earnings on invested
The facility as a whole may alternatively be regarded as 'layered?0 The first layer caters for
assets fall short of the projection) credit risks (uncertainty surrounding the collection of future
the retained portion of the risk, while the second provides excess loss cover for the more
premiums).
catastrophic element. This does not imply that the insurer bears no risk whatsoever in the retention
of the first layer. Generally, in competing for such premiums (or retention funds), the insurer
grants cover in excess of what has been provided for in terms of the total premium. Depending Risk spread/risk combination
Risk spread/combination can take various forms. For example, one could spread the impact
of large losses over time through a multi-year policy or combine several exposure years into
28 is controversy surrcunding the issue af self-insurance facilities. Fo« instance, the tax implications ot selßnsurance premiums are not a single policy term. It is also possible to provide coverage for different lines of business on
all clear: some &T9ue that effectively there -IS real 'transfer' of risk to the irsurerr and that therefore the 'veneer' of insurance is inadequate allow payments intothe facl'ity a combined basis.
to be regarded as insurance premiums. In (befnal analysis, the insurer may bear some risk, tH_Jt the extent bearing that would alkyweligitil ity for Insurance Is a debat3b'e
point. From an economic standpoint. one cou Id argue that insurers have adapæd their cost structures to mai ntaln their competitiveness this niche market and have
diversified their portfolto by writing relatively low-rlsk.dvwreturn busiress. To the utent of other that the insurances Inclusion at Of a this business improves overall results.
(self•insurance which in turn facilities) might lead is advantageousto the creation of InsuraX 2 • Aggregate limits
capacity and the provlsioq more competitive price. thß activity insurance structures (IRIS}. In order to limit the total exposure, an overall aggregate limit is normally applied for the
29 Such programmes are known as as±eamless' programmes or two Integrated or roore rusk aod an absolute monetary amount Of insurance
•kcontract period, and, if necessary, annual pay-out limitations are set that define the maximum
30 Layering can be desctibed an arrangement whereby insurers each provide
amount available in any one given year.33
coverage. or layer. The insurer Of the first layer is primaryJ because it responds first in covefir" a loss Insurers oi subsequent layers respond sequerwer as necessary. to
cover any large loss. See PJNtS (1985:42)-
32
31 The Readers interested ale referred reader to is chapter referred 13 to fcw Cheng an In-depth (1990). discussjon Ofsuch programmes.
33
The underwriter first dete'rnines the expected loss. Then, based on assumptions concerning the *tern of payments and interest rates individual risk(s), the discounted will be
established_ The diffeænce benæen the ex-pared Poss and the discounted loss the ane interest rate risk are assumed. The between the expected and the aggregate lirn,t
represents the
254 Cash flow protection
The capacity to retain a portion of losses is constrained by liquidity and time. The shorter the
time frame, the lower the tolerance for interruptions to cash flow. For example, the extent
• Commutation
of losses a company can fund in a year is greater than that it can fund in a month. The longer
This last feature is of particular importance in an SIP. Commutation is a system whereby any excess
time period allows funds to be built up and also reduces the variability of the loss experience
premium (surplus) plus interest is returned to the insured."
that is encountered on a monthly basis.
The use of a deductible places a ceiling on the amount that the company has to
The following section will examine deductibles and contingency policies as the first option in
establishing a retention programme. contribute to each loss when it is settled. reduces the disruption to cash flow. Large
deductible programmes are flexible in that they can be tailored to fit the amount ofrisk an
insured is comfortable retaining, as well as allowing for the unbundling of services if the
insured so desires.
10.8 Deductibtes 35
What is a deductible? Types of deductibles

Economic theory requires that to deal with the moral hazard problem, the insured should always The concept of the deductible requires that the insured bears a portion of the losses arising
bear a portion of the loss. The deductible achieves this purpose. A deductible is the part of the from pure risk exposures. This can refer to losses from a single event or over a specified
insured loss an insured retains for its own account. The deductible is subtracted from the amount period. Each arrangement has a different effect on the distribution of losses that are retained
with which the insurer reimburses the company in the event of a loss. rlhe greater the deductible, by the organisation. 'Ibe main forms of deductible are the following.
the greater is the extent of retention funding. Deductibles may range from small amounts to
millions of rands. Each and every/per loss
Otherwise known as the simple or straight deductible, this is one of the simplest and yet
A large deductible programme may offer many ofthe advantages ofa retention ftlnding
most effective deductibles in use. It applies to each loss and is subtracted before any loss
programme without the increased expenses that may accompany the establishment of, for
payment is made.
example, a captive insurance company. This is because the insurer provides all the necessary
administrative services and pays all the claims. The insurer then bills the customer for amounts
Aggregate deductibles
that fall within the policy's deductible. Also, for many companies, the full retention of losses may
Another type of deductible is the aggregate deductible, which applies for an entire year. With
be neither advisable nor desirable. More specifically, the use of a deductible offers the insured the
an aggregate deductible, the insured absorbs all losses until the deductible level is reached.
following benefits.
At that point, the insurer pays for all losses over the specified amount. Sometimes, the
aggregate and straight deductibles are used together.
Compared to the straight deductible, aggregate deductibles are not as successful in
With a deductible, the insurer offers a discount on the full premium cost to the insured. There is
eliminating the cost of processing small claims, because all losses will likely be reported to
a trade-off. In return for its contribution to the cost ofthe loss, the insurer reduces the premium
the insurer for credit toward meeting the deductible. In addition, because losses may be
charged to the insured. It is in the insurer's interest for the insured to retain losses at lower levels
fully paid after the deductible is met, the ability to reduce moral hazard is not as great as
because of the disproportionate administrative costs associated with these losses and to
with the straight deductible.
discourage the insured from claiming for very small losses that the insured would usually not
bother to repair if they were uninsured.36 The benefits to the insured are that it enjoys an
• Disappearing deductible
immediate reduction in expenditure in the form of lower premiums and an increased cash flow.
When a disappearing deductible is used, the size of the deductible decreases as the size of
34 Typically, commutation is in the fdlowing way on receiving the premium. a pregreed pe«entage is deducted to underwriting expenses, including a charge for Insuring the loss increases. Finally, at a given level of loss, the deductible completely disappears. The
the risk. Loss payments Of also be deducted men. to the amount interest earnings credited (which are usually calculated on the bas,s ofa predetermined interest rate
factor). reduction in the deductible results from the fact that losses are adjusted according to a
are products on the market these small losses that are more valet-type products, The are products still insurance products requi@to
RETENTION 255 formula such as:
256 400,00
2
290.00 100
3
240,50 250
where: 4
P = payment by insurer 191,75 500
L = loss 5
D = deductible 137,50 5 000
6
R = recapture factor, where R has a value between 0 and 1 and increases as
65,00 10 000
L increases.
To judge the value of the deductibles, additional data is required. ms is shown in table 10.2.
Franchise deductible
Table 10.2 Loss data for private cars, 2004—08 (adjusted for changes in fleet and inflation)
A franchise deductible is expressed either as a percentage of value or as a rand amount. Under a
franchise deductible, there is no liability on the part of the insurer, unless the loss exceeds the
amount stated. Once the loss exceeds this amount, however, the insurer must pay the entire claim. Year Number of Number of Total cost (R) vehicles incidents
In insurance for ships and their cargoes, it is common to use a franchise deductible expressed as 2004
a percentage of the amount insured. Thus, the policy might provide that no loss is payable unless 2 000 320 820 000
2005
the loss equals or exceeds 3% of the total value. But once the loss reaches the 3% level, the insurer 1 700 200 700 000
is responsible for 100% of the claim. 2006
2 100 480 1 104000
2007
2 200 640 1 638 500
Determining deductible levels 2008
360 875
Various deductible selection rules exist in the literature. We will limit our discussion to the least
Total 000 2000 5 137 500
cost rule, which does not take financial capacity explicitly into account.
The analysis of this information reveals the following:
Ihe least cost rule is based on the proposition that the cost of pure risk to the organisation is equal Average number of incidents per annum over the five-year period is:
to the insurance premium plus the cost of losses retained under the deductible. The initial 2 000/5 = 400.
formulation ofthe rule assumes that losses that occur under the deductible are equal to the full Average annual cost is:
amount of the deductible.
5 13 500/5 = RI 027 500.
The rule states that the level of deductible selected should be that which gives the lowest To apply the TEC rule, the value of q (the average number of incidents per vehicle) is calculated as
total expected cost (TEC). Formally, it is expressed as:
follows:
TEC = P + qD where:
P insurance premium D = deductible level q = average annual frequency of occurrence - 2 000/10 000
of loss per exposure unit.
RETENTION 257
or
Table 10.1 Insurance rates (per vehicle) for a commercial fleet of private cars
- 400/2 000 = 0,2.
We can
Option Option premium (R) Deductible (R) now examine expected cost of the various deductible levels.
RISK MANAGEMENT: RISK
Chapter 1 1 Captive
insurance companies
258 ftNANClNG
259
Table 10.3 Calculations for optimum deductible
400,0
310,0
Definition
How do captives work?
290,5
The captive insurance industry
291,8
Captive domiciles
1 137,5
Types of captives
Features of and motivations for forming captives
Captive ownership structures
According to the TEC rule, option 3 with the R250 deductible should be
selected, as it yields the lowest TEC.
Deductible funding policies (contingency risk policies)33

1 1.1 Definition
Deductible funding policies — also known as contingency risk policies - are similar to large deductible
programmes, except that a customer's premium is placed in a fund that is used to reimburse the
Definition
insurer for losses paid within the deductible layer. 'Ihe fund earns interest at a predetermined rate
A captive l insurance company is defined as an insurance company that is owned by a
and, at the end of the contract, any excess is returned to the insured. These programmes are
parent company (or companies) and that writes the insurance business (in sum or in part)
designed as another mechanism for the insured to keep more cash flow benefits, while still allowing
of its parent(s)?
the insurer a reasonable rate of return.
Bawcutt3 notes the following about captives:
10.9 Summary I the key factor is that the insurance company (captive) is owned by the parent and therefore
has its own captive business. Defining captives

The influence of practical considerations on the implementation ofa risk retention strategy is not diffcult to see. The introduction
of transaction costs results in a risk management case for retention funding, while the practical considerations of tax add a further Credit the namercaptive'is glven to Fred Rerss. who coined the name nearly 70 years ago (Phelamet Comm Isslon. 1990: ST). 2 Some capoves,
specifically US captlves, write non-parent-related risks. This has la'gely occurred because oftax conslderations.
dimension to the risk manager's evaluation criteria and provides further choice regarding the retention method used.
Depending on the size of the organisation and the intended sophistication surrounding its risk retention, the firm may as a first step consider larger
deductibles and contingency policies and may even consider the incorporation of its own captive insurance
company.
33
These are sm•larto rent-a-captive far-Il Ides.
RISK MANAGEMENT: RISK FINANCING CAPTIVE INSURANCE COMPANIES
260 8 shown that the pure premium to insurers p or the expected monetary value of the claims. Each insured theoretica l* its own annual claims the pooling effect the to
spread own expected claims costs
9 Clearly. this refus to short-term insurance contracts and to long-te€rn Oife) assurance.
CAPTIVE INSURANCE COMPANIES 261
becomes difficult however, for there have developed a number of captive types,4 and
as one moves away from the pure captive that insures only the parent's business, one
about the captive industry, and examines the features of captive insurance and the
is attracted to the suggestion that it would probably be more helpful to accept the
motivations for forming a captive. It includes discussion of certain issues relevant to South
origins of captives and move towards a situation where the word 'captive' disappears
African-owned captives.
and where ... [it] is referred
to as an insurance affliate of a corporation.5 There is considerable debate and
unnecessary confusion over the nature of captive insurers, basically resulting from the
11.2 How do captives work?
American concept that insurance requires risk shifting, risk distribution or risk transfer.
Captives function in a similar way to commercial insurers, i.e. they insure and reinsure and
Although use of the term 'transfer' of risk is substantiated by the fact that captives adopt rating invest their funds. The primary difference between captives and commercial/professional
and reserving procedures, it was initially argued that the risk does not in fact leave the corporate insurers is that captives concentrate on their parents' risks. The parent would most likely
group and that therefore there is no real 'transfer' of risk to a third party, i.e. the captive may be have good risk control programmes and procedures in place long before it would consider
seen to be a member of the same 'economic familf.6 The economic family concept, however, cuts self-insurance to a degree that requires the use of a captive.
across the well-established principle ofthe corporate entity, or juristic persona, and it is not There are two common arrangements whereby captives provide cover to their parent
surprising that the American courts have begun to reject this concept. (and subsidiaries and affiliates), namely the direct method and the indirect method.
It is submitted that the correct test to determine whether a particular transaction or scheme Direct method: Premiums are remitted directly to the captive, which retains a portion
is one of insurance or not is not to enquire whether risk transfer7 or loss spreading8 is involved, according to the liability it wishes to carry and cedes or reinsures the remainder.
but to determine whether an enforceable insurance contract is present, obliging one party (the Indirect method: Risks are insured directly with a conventional commercial insurer,
insurer), for a consideration (usually a premium), to indemnify another party (the insured) in the which, by agreement, reinsures a portion of the risk with the captive. This is known as
event of a defined loss-producing event materialising. 'fronting' (for which a fee is generally payable) and is an arrangement whereby a captive
Thus, the test is whether, when an individual loss occurs, the claimant, in exchange for a contracts with an insurer licensed to do business in a jurisdiction (where the captive is
consideration, has an enforceable right to indemnification, say from a captive insurer. If such a not able to be licensed or is excluded for some reason) to provide the required covers
right exists, it is argued that the party obliged to provide such indemnification is indeed an and then reinsure some (or all) ofthem with the captive.
insurer. Moreover, it is necessary that an insurance contract/policy exists before any reinsurance
can be arranged, and hence it is only when a captive is a legitimate insurer that reinsurance of Often, parent companies have more than one captive. A company may, for example, have
the captive is possible. one captive for property and one for casualty; one for domestic and one for foreign
As concerns the circumstances in which a captive insurer can be used, it would appear that business; one for deductibles and one for excess reinsurance, and so on.
the earlier conclusions relating to retention funding would equally apply to captives. This
chapter provides certain facts and estimates Use of captives
The primary motivation for the formation of captives these days seems to have moved away
4 and (1982-. divide a Into the categories pure waiting ot extanal only pædominantty risks). centrethe of the parent); captive pöffollo indudes although not overwhelming
from the original more simple motives such as reduction in costs, access to reinsurers, etc.
po,tiön
(where the parent's represents small the total risks and group captive owned by two o, more
Today, the motives for captive formation are management and control issues such as
5 It suggested that the term considered to describe a insurer in place the improved risk control; improved ability to manage total cost-of-risk, possibly across a
family'concept was specifically rejected in a number of v united states F.2d act {198.9 and see insurance, 7 August a discussion of the issues. In no dispersed
circumstances do we adopt the economE family argument advanced by thegovernment! multifaceted and geographically group; access to reinsurance markets;
investments; and, due to targeted risk control, the provision of covers that would not traditional insurance structures analyse loss trends on an annual basis at renewal, the introduction
otherwise be available and the extension of covers at reasonable cost. of a captive can provide the ability for loss trends to be identified at more frequent intervals
262
during the insurance year.
Captives are used to build reserves for specific covers required of th e organisation that are
Perhaps the most important driver in the use of captives is the ability to manage the either unavailable from or too expensive in the market, which is driven by the overall market's
organisation's overall total cost-of-risk. Forty-seven per cent of UK-owned captives regard the perception of risk. The captive can also be used to extend covers where market offerings are not
reduction of total cost-of-risk as the reason for using a captive. Whereas the trend for the formation adequate and to provide cover for the specific contractual requirements of the organisation. 263
of new captives does not follow the insurance cycles which is driven by external factors, captives,
which only have to react to the insured's risk profile, can be and are used to manage the effect on
Where specialist covers are purchased through markets where there is a concern over the
total cost-of-risk through successive periods of increasing insurance rates in the general market. financial strength or credit quality of the insurer or reinsurer, the captive, with a clearer vision
The focus on risk retention is also driven by the need to achieve control over the total cost-of-risk of the real picture, can be used to mitigate these concerns. Captives are also used to access
in order to smooth this cost and to achieve predictability of costs in line with predictions of actual legislated markets and markets that do not accept direct placements, including reinsurance
losses. markets.
In the European Union (EU), for example, captives are used to eliminate the
The use of a captive allows for the aggregation of risks from across the organisation and the
disadvantages of 'passporting' requirements. Passporting refers, for example, to a case where
ability to present a far larger, unified and diverse portfolio ofrisks to the market. The economies
a captive located in the EU or the European Economic Area (EEA) is able to write business
introduced through this scale and diversity enable price reductions that contribute towards the through the EU/EEA without the need for separate insurance licences in each member state.
reduction in total cost-ofrisk. In some cases, analysis of the diversification effect points to This benefit is particularly useful for employee benefit-type covers and is being used by US
reductions of cost-of-risk of up to 30%. Further, the aggregation of risks (premium and claims) and other non-EU multinationals that operate in the El-J. Current moves by US regulators may
into one or two risk vehicles introduces a level of control over the costof-risk across the entire affect such cases.
Insurers rely on a combination of premiums and investment yield to cover claims.
organisation, a level of control and discipline that was not easily achieved without the use of a
Captives, of necessity, have to follow the same practice; thus, a reasonable return on capital is
consolidation vehicle such as a captive or insurance subsidiary. 'Ihis centralised insurance-buying
generally expected. Return on capital is now often one of the key performance indicators of
function, together with the need to take control of the risk.financing strategies of subsidiaries in the effective management of a captive. There is, however, a need to be able to balance this
order to realise economies of scale and scope across the organisation, is suggested as one of the requirement for investment yield against the organisation's overall requirement to reduce the
key reasons why multinationals form captives. total cost-of-risk. Investment yields and effective financial management of a captive cannot
be achieved simply by manipulating premium rates, which have to be properly related to the
Captives are often used to harness the group profile to ensure reasonable deductibles for
risks covered, but must be achieved through risk control and sound investment practices.
subsidiaries. In this way, the smaller subsidiaries benefit from the diversification and greater
buying power achieved by the larger group.
Lines of business underwritten by captives
The ability to identify organisational loss trends is a key benefit of consolidation of risk
Historically, captives were used to provide for protection on exposure lines that may have
information through a single risk vehicle. The closer monitoring of claims through the captive
been diffcult to obtain (or too costly) in the insurance market. Captives were used to house
structure is often accompanied by a greater focus on claims across the organisation, compared protection for long-tail liability exposures of the organisation, and their formation was often
with the monitoring that would take place in conventional placements. Because the captive often triggered at a time of significant organisational change (such as the need on privatisation to
provide for previous potential liabilities) or through some significant loss event (such as the
uses overall savings in premiums to provide more comprehensive cover (covering some part of
9/11 World Trade Center attacks and the need to provide for terrorism exposures).
conventional insurance deductibles), the ability to assimilate more complete loss information
Recently, however, captives have been formed to underwrite the more traditional
across the group in a single vehicle is a clear advantage provided to the organisation. Whereas casualty and property lines (often with reinsurance arrangements in place) in order to take
advantage of the various benefits (risk management, management of total cost-of-risk, etc.) The traditional lines of business written by captives include property exposures, motor,
offered by the insurance subsidiary to its parent organisation. construction, marine, aviation and liability (primarily general liability, products liability,
The long-tailed casualty line seems to be the largest business risk Underwritten by workers' compensation, group life and disability, motor liability, directors' and offcers'
captives. In the survey of UK-owned captives, 41 % of captives U nderwrite casualty lines. liability and professional liability). Equipment and motor vehicle warranty risks, credit life
Included in casualty are employee benefit-type risks. This trend is following that of captive and credit risk (debtors' insurance) are being housed in captives, as are post-retirement
owners in the US, where permanent 264
medical benefits, mortgage default, and pollution and environmental impairment exposures.
total disablement and medical expense exposures are being included in their captives.
European organisations are also expanding their use of captives to include US employee 'Ihe introduction of the Terrorism Risk Insurance Act (TRIA) in the US has given rise to the
benefits. Of the formations of new captives in Vermont during 2005, the most popular line of requirement of all insurers, including domestic based captives, to provide terrorism cover. All
coverage was general liability, followed by professional liability, motor liability, workers' US-domestic captives and risk retention groups (except those writing medical malpractice and
compensation and then property, with a growing number of captives writing medical other lines specifically excluded by the Act) are required to offer terrorism coverage and are
malpractice insurance. subject to the 3% surcharge. The introduction of TRIA (and its recent extension) has encouraged
a rise in the number of captives that are issuing terrorism insurance policies.
Figure 11.1 Business underwritten by UK-owned captives UK-based research has provided a breakdown of lines of business underwritten by captives
in various sectors that are owned by UK companies (see figure 1 1.1). Assets comprise property
damage and business interruption risks. The 'other' segment is a reflection of more specialised
Assets
risks that are directly relevant to the particular owner of the captive, e.g. personal lines covers
25%
and credit insurance.
Casualty Other
41%
11.3 The captive insurance industry
Whereas there are examples going back to the late 1700s and 1800s ofcompanies (or groups of
Motor Marine
12% companies) forming entities to insure their own risks, Bermuda has been credited with
4% developing the modern captive insurance concept in the early 1960s.
There was significant interest from US companies in Bermuda following the relaxation of
exchange control in the early 1970s. In the 1970s, brokerages and insurance managers formed
in Bermuda to take advantage of the growing captive market. In 1978, Bermuda introduced
Source: Marsh (2007) legislation to regulate the insurance industry, and this legislation, together with its various
amendments and regulation, as well as other benefits offered, e.g. good management facilities,
Underwriting long-tailed casualty business in a captive is advantageous to an organisation, as
excellent banking and financial facilities, geographic proximity to the US, good living conditions
the higher premium levels increase investment opportunities in the captive, as opposed to
a stable economy (and currency), a tax structure that mitigates any double taxation, etc.,
these funds passing into the general market.
resulted in Bermuda becoming the leading captive domicile and also a leading global financial
In the US, a small, but growing number of organisations are funding longterm employee
centre.
benefit covers through their captives, following the granting of specific permission to do so
A convenient starting point is to offer some facts and estimates about captives that will
by the US Department of Labour. In addition to the significant cash flow benefits this
show how pervasive this form of risk financing has become. The majority of captive insurance
introduces, employers are able to obtain more flexibility in the design of their benefit plans
companies in the world have been formed by US corporations. The first reliable survey by Risk
when they are funded through a captive.
Management Reports in 1974 reported the total number of captives as 316. According to the
Insurance Information Institute, this number increased to 5 119 captives registered world-wide
at the end of 2007.
MANAGEMENT: RISK FINANCING CAPTIVE INSURANCE COMPANIES
266 267
Captive companies are domiciled in various domiciles in the world. Hawaii n.a. n.a. n.a. 160 163
Table 11 lists the top captive domiciles for the period 1990—2007.
Other US
domiciles/ states 245 123 123 447 317 436
Table 11.1 Captive domiciles, 1990—2007
Location Number (estimate) Other non-US 197 244 285 295 268 308
1990 19% 1997 2000 2006 2007 Total 2 876 3189 3 361 4 204 4951 5119
Table 11.2 Captive ownership by country of parent company, 2007
Bermuda T 225 1 050 1 564 989 958
1 038
Cayman 449 517 740 765
350 418 Country of parent company Ownership (96)
Islands
Barbados 175 167 195 199 235 256 United States
57
Turks and n.a. n.a. 169 173
n.a. n.a. United Kingdom 10
Caicos islands
Guernsey 339 381
Sweden
190 324 375 368
6
240 254 264 208 210
Luxembourg 140 France
Isle of Man 159 167 168 161 155 5

Luxembourg 2
26 142 163 154
Ireland
134 El Canada 2
51 Belgium 2
42 49 52 60 62
Singapore Australia
Vermont 189 305 361 563 567 2
293
Other 14
British Virgin n.a. n.a. n.a. 400 409
Total
Islands n.a.
100
South n.a. n.a. n.a. 146 158
n.a. Source: Marsh (2008)
Carolina
Property
Captives are used to underwrite a variety of risks. Table 11.3 lists the major
Sources: Offshore Finance, 1990; Business Insurance, 27 April 1998; Business Insurance, 3 March 2008. damage 20
The figures may not be strictly comparable.
General or third-party liability 18
Employers' liability and workers' compensation 12
Captive ownership by the country of the parent company at the end of 2007 is depicted in table
Auto liability 9
11.2.
Professional indemnity
types of business written by captives. 8
Financial products
6
Table T 1.3 Types of business underwritten, 2007 Health and medical 4
Product liability 4
Type of business Marine
4
Other
15
Total
100
The 10 Gwen the di"culty in obtaining reports on ceil companies' individual cells, this numbes is mosc probably an underestimate Of the number of cell/PCC structJles
premium throughput and assets of captives world-wide continue to grow and seem largely
in the market
unaffected by the vagaries of the insurance market cycles. 268
Table T 1.4 Use of captives by industry

The total premium volume is now believed to approach 20% of all worldwide commercial
insurance premiums. These figures make captives important entities in the insurance Industry
industry and a number of individual captives compare in size with many conventional Take-up factor Utifisation factor
insurance companies. Agriculture, forestry and fishing 50% 1.50
Captive formation usually follows a deliberate risk retention strategy by organisations. Mining
It is interesting to note that, according to the Insurance Information Institute, from 1988 to 61% 1.95
2006 risk retention group premiums rose by a staggering 1 008%; the increase is around Construction
257% for the period 2000-06. This leads to the conclusion that the number of captives
46% 1.56
(vehicles representing deliberate retention by organisations), as quoted above, is an
Manufacturing 45% 1.46
underestimate — especially when one includes the formation of cell captives (governed by
Transportation, communications,
protected cell company or incorporated cell company legislation). The cell captive industry
electricity, gas and sanitary services 40% 1.54
has shown significant growth (and in our opinion, as consultants operating in this arena, is
Wholesale trade 28%
becoming the favoured structure of ail but very large insured organisations). AM Best
T .27
estimates that by 2003, 50% of US commercial market had migrated to alternative risk
Retail trade 42% 1.38
transfer (ART)-type structures, and those in the front line will argue that this trend is as result
Finance, insurance and real estate 55% 1.54
of cell structures that offer retention features to clients in a less-costly and lesstime-
Services 49% I .53
consuming way. Today, 27 domiciles offer some form of cell legislature, and it is estimated
Public administration 40% 1.00
that there are at least 265 cell companies with around 1 200 individual cells. 10 Of the top 1
500 global companies, only 47% have captives at the parent level. Of the top 1 500 global companies, 61% of the companies falling within the mining' sector
have at least one captive, with approximately two captives per company. Illis was found to
Ihe following industry sectors lead in the use of captives: mining, be the highest utilisation of captives both from a take-up within the particular industry sector
including oil and gas extraction finance, insurance and real as well as from a per-company utilisation perspective.
estate manufacturing services agriculture, forestry and fishing When further analysed, these trends continue into the specific oil and gas extraction sub-
construction. sector of mining: 85% of the companies in the mining sector fell within the oil and gas
extraction sub-sector. Of these, 61% owned at least one captive, with approximately two
captives owned per company.
The high take-up factor in some industries is in part historic. Many financial institutions
established captives during the 1990s to address mortgage indemnity guarantee risks. In the
UK, on privatisation, utility companies formed captives to manage historic liability exposures.
Other industries with large exposure profiles (e.g. energy) have made significant use of
captives to deal with the lack of capacity available in the insurance market.
CAPTIVE INSURANCE COMPANIES
RISK MANAGEMENT;
When analysing the types of risk covered by captives in industry sectors, the results of US and Canada
UK research regarding the industry groupings of manufacturing, service suppliers and
The size of the US and its relatively free insurance environment have resulted in the majority
financial services are indicated in figure I I .2.
of captives being owned by US parent corporations. Most of these captives are located in
RISK FINANCING
270 Bermuda, which is geographically close to the US.
are close to | 000 captives in Bermuda, and approximately 80% are owned by US parents.
Most of these companies were set up in Bermuda in the 1960s and 1970s. Formations
Figure | 1.2 Business underwritten by UK-owned captives dropped in the early 1980s owing to the soft insurance 271
Other Service suppliers Assets
Manufacturing
37%
Assets
market, but a resurgence of formation was seen in 1986. In 1978, the Bermuda authorities
28%
Motor looked at the possibility of introducing insurance legislation, and in 1981 the Bermuda
29% Insurance Act was introduced. Following the decision by a Harvard-based medical
Other
Casualty
5% malpractice captive to form its insurance company in the Cayman Islands rather than in
37%
Bermuda, the Cayman Islands grew significantly in the late 1970s, particularly for group and
Casualty Marine Motor
association captives. As a natural consequence of this grovnh, the Caymans also introduced
Marine
19% 2% 19% its own insurance legislation.
Financial services Assets In recent years, the most significant development for US companies has been the
19% questioning by the US Internal Revenue Service (IRS) of the tax status and, in particular, the
tax deductibility, of premiums paid to captives. When this development was first mooted,
many captives felt that one way to avoid the problem was to enter into unrelated or non-
Casualty parent business, on the basis that at a certain level the tax authorities would regard the
Other captive as a genuine insurance vehicle and grant tax deductibility. Unfortunately, this hope
35% has not been realised and tax deductibility is an important issue for US companies. In
Motor addition, the tax position and the unavailability ofproduct liability and medical malpractice
3% capacity in the 1970s saw the growth of association and group captives with shared
ownership, rather than the single ownership common to most captive companies.
Source: Marsh (2007)
The development of captives in the US and, in particular, growth among those formed
within those states specialising in providing captive facilities have continued, as has the
The even distribution of coverage lines in the manufacturing sector is indicative of the nature
grov,th throughout the 1980s of association and group captives. Vermont now handles close
of the business and exposures experienced in that sector. The picture is very different in the
to 600 captives, and many other states are now considering introducing legislation to
financial services sector, where long-term casualty-type exposures (including professional
facilitate onshore captives. British Columbia in Canada has also introduced legislation
indemnity and financial institution risk) dominate the types of cover written in captives, with
attractive to Canadianowned companies to allow for the formation of onshore captives. This
'other' also being significant. 'other' segment reflects the personal lines insurances that
may signal an increasing trend to locate within the same country as the captive's parent.
financial institutions offer and in which their captives actively participate.
Europe
Captive development within Europe is dominated by UK companies, which own 10% of
captives world-wide. Research within the FTSE 100 companies indicates that 74% of these
companies make use of a captive/insurance subsidiary. This high take-up within the FTSE Outside the EU, Switzerland has a number of captives domestically, and particularly
100 has been attributed to these companies having the critical mass of premium, the offshore in well-established captive locations such as Bermuda. There is also the opportunity
necessary risk profile and the amount of losses to be able to justify the need for a captive. to establish captives within Switzerland, and this has been done in Zurich.
The greater the si ze Of the company, the greater the probability that financial benefit will In Scandinavia, there are captives in Denmark, Norway, Sweden and Finland. In Sweden,
be derived through the use of a captive. there are a number of domestic captives, while a growing number of Swedish parent companies
have located captives offshore, particularly in Luxembourg, but also in Bermuda and the
Guernsey.These captives Ille imposition were originally of exchange located controls in Bermuda limited 12
Cayman Islands. Sweden has a larger number of captive insurance companies per head of
and the establishment subsequently inof
population than anywhere else in the world. The last count of the number of Swedish-owned
captives amounted to over eighty, with the majority in Luxembourg.
Apart from the captive locations already mentioned, other possible growth areas include
but more recently it is because of insurance market attitudes. However, the situation is
Cyprus and Malta. Cyprus may be of interest to ship owners and companies with Middle East
changing, with many large French companies having formed captive insurers, particularly in
connections. Malta, a relative newcomer to the captive scene, has introduced attractive
Luxembourg. In addition, many large French insurance and reinsurance companies, as well as
insurance legislation and provides a concessionary low tax rate of 5% for captive insurance
insurance brokers, have acknowledged the role that captives can play in corporate insurance
companies. It has a developing captive management infrastructure.
and have established their own captive insurance management operations and, in some
There is no doubt that Europe provides an important market for the growth of captive
instances, their own overseas reinsurance entities. There is, therefore, growing interest and
insurance companies. It is diffcult to obtain full details of the number of European-owned
greater capability available to the French insurance buyer.
captives, as European multinationals tend to favour confidentiality in their operations, in
On a more limited scale, Italy, and to a greater extent Spain, are also following the French
contrast to the more open environment that exists in the US. However, it has been estimated
example. As exchange control regulations ease and the freedom ofservices legislation begins
that there are more than 300 captives, and this is expected to grow to close to 1 000 over the
to have an impact, one can expect a more common approach to the use of captives in risk-
next decade. The wider European market transforms the potential for
financing programmes. This will be accompanied by the disappearance of the individual country
postures that have for so long inhibited the availability of the captive concept to many large
continental European insurance buyers.
offshore the and KLM and Swissair.
272
new captives in the territories of the Channel Islands, Gibraltar, the Isle of Man and the
Republic of Ireland. After a long period of gestation, Guernsey introduced full insurance
legislation in September 1986 and is highly regarded as an important area for captives and
offshore reinsurance companies, with around 400 insurance companies licensed at the end of
2007. In February 1997, the Protected Cell Companies Ordinance was introduced.
Another very popular location for UK-owned companies has been the Isle of Man. ne Isle
of Man introduced specific legislation in 1986 that eliminated the normal corporate tax rate
of 20% where insurance companies writing nonIsle of Man business were established, and put
into place special insurance legislation designed to ensure proper supervision and control.
One of the most important developments in recent years has been the establishment of
Luxembourg as a captive reinsurance centre. Luxembourg provides attractions through the
ability to defer tax. Dublin in Ireland is also of particular significance being able to write direct
CAPTIVE INSURANCE COMPANIES
RISK MANAGEMENT;
business (unlike Luxembourg), and, as in the case of Gibraltar, captives can write direct policies
covering risks of organisations within the El-J.
EU captive development is at an early stage, but promises to be an exciting process and
may involve important changes in the way in which EU captives are structured. The process
will also be influenced by freedom of services legislation and the EU tax position on companies
established outside the EU, in particular those established in low- or nil-tax areas. Offshore
captives owned by Dutch parents have usually been located in the Netherlands Antilles. This
has advantages for captives and important specific benefits for the Dutch parent, because
dividends distributed by the captive can be returned to the parent in the Netherlands without
any payment of additional Dutch corporation tax.
Belgium is also a very competitive insurance market from an insurance perspective, and
2% ofthe captives world-wide are owned by Belgian companies. However, in view of the
relatively small number of major companies with Belgian parentage, the potential is limited.
Germany is a growing and very important market for the formation of new captives.
Historically, most captives have been established within Germany. Captive establishment has
been frustrated by the influence of the German insurance market on captive owners including
equity interest and board representation. This has also been exacerbated by the perhaps
unique concept of the captive broker. ll
Historically, France and Italy have been difflcult areas for captive formation,
In the early years, this was due to a restrictive exchange control environment,
This involves companies forming insurance broking that earn Insurance con-•rnissiöns on premiums p.a•d to In sure'S with they place their business,
274 reported on them. The government then appointed a further commission to investigate the
operation of captive insurers, again under the chairmanship of Justice Melamet (Mehmet 11).
14
captive formation, and, combined with freedom of services in the insurance sector and the
Some of the terms of reference of the commission (with regard to captives) were to
possibility of more European countries joining the EU (or at least developing economic ties),
inquire into and report and make recom-mendations on: the operations of captive insurers and
provides a market as large as that in the US.
the necessity for South African insureds to use their services inside or outside the Republic
the extent to which the services ofcaptive insurers are currently being used by South African
insureds and the effect on among other things, the: insurance industry outflow of funds loss
Asia
of taxation
The country in this region with the most potential for captive formation is obviously Japan.
Although a number of Japanese captives have been established, many in Bermuda, the investment of the funds of captive insurers the advantage gained from captive
peculiarities of the Japanese insurance market have precluded significant growth of captives. insurers as genuine risk carriers for the South African insurance industry as to, among
However, as Japanese corporations expand throughout the world and establish their own other things: an awareness of risk management by the local personnel, and their training
operations in Europe and elsewhere, there is a growing tendency for these subsidiary companies therein the promotion of entry into international reinsurance markets the practice
to establish a captive insurance subsidiary. In very recent times, Japaneseowned captives have whereby local insurers act as a front for foreign captive
been established in Luxembourg and Dublin, and this must be the beginning of an expansion of
the captive concept for the overseas subsidiaries of Japanese corporations.
It is not the intention here to offer commentary on the wider issues and effects ofcaptives as
In Asia, there are two good captive locations: Hong Kong and Singapore. Hong Kong does
per the terms of reference given to the commission, nor to debate the recommendations
not have specific captive insurance legislation and its high capitalisation requirement means that
proposed by the commission's reports (1988/1990). Rather, this section provides a summary
it would only be of interest to a substantial company with major insurance programmes.
of the Melamet Commission's recommendations with regard to offshore captives, and an
Singapore has specifically developed both insurance regulations and a low tax rate for captive
insight into the specific aspects of taxation and exchange control as they pertain to South
insurance companies writing non-Singapore business, and has become the favoured location
African-owned captives. These recommendations were as follows (with the relevant section
among Australian-owned captives.
number of the commission's report
A committee consisting of representatives from the exchange control authorities, the
Registrar of Financial Institutions and the Commissioner for Inland Revenue should be
South Africa
formed to control offshore captives. (5.8.1) The committee should recommend what
South Africa has a well-developed captive market. There is record of more than 30 South African-
funds, if any, and in what form, accumulated in the captive be returned to the parent
owned offshore captives (and more than five onshore captives) at the offices of the Financial
(5.8.2. l) The Commissioner for Inland Revenue should investigate any unjustified
Services Board and the exchange control department of the South African Reserve Bank It has
deductions of premiums for income tax purposes. (5.8.2.2) Possible contravention of
been estimated that in 1988 premiums to captives totalled over R300 million. 13
sections 60(2) and section 75 of the Insurance Act should be investigated. (5.8.2.3)
Following the collapse of the AA Mutual insurance company, a large shortterm insurer, the
South African government appointed a judicial commission of enquiry under the chairmanship
of Justice DA Melamet (Melamet I). While carrying out its task, this commission came across 14 April 1989. The Melamet Cornrntssion of Inquiry {1988, which appointed the collapse of the Insurance Ltd, that the and for insurers,
be the of the Ccu-nrnißicner for Inland Revenue and the Exchange Contrai
captive insurers and
276
Annual returns should be submitted to the exchange control authorities and the Registrar
13 Mekmet (l "O; 52). Since 1988, there has been a significant increase in the number of South African captlves formed. Hence, the as quoted in 1986 are an of Financial Institutions, as prescribed. (5.8.3)
underestimate ofthe number of South African Mfsbore and onstu:ue captives today. Also, in the past six years, has been conslderable activity the Sc.uth African cell-
captive and Melamet Commission can be because it only looked at the outward claims Bovv than of premlums outflow of and premiums.did not report on [he inward
flow of claims. The period under Imestioation• charm-tensed by a greater inflow of
The Commissioner for Inland Revenue should investigate premium deductions to rent-a-
275 captives. (5.8.4)
MANAGEMENT: RISK FINANCING CAPTIVE tNSURANCE COMPANIES
Deduction for premiums remitted to offshore captives should not be allowed for income
1 1.4 Captive domiciles
tax purposes, except amounts spent on reinsurance, less commissions received thereon.
(5.8.5) Potential captive owners often found that the insurance legislation in their own countries did
not provide the level ofdesired flexibility. Legislation and capital requirements in most
countries were focused on controls of an existing openmarket insurance market, and these
In the main, the recommendations of the commission were not implemented and the
controls restricted the captive's ability to achieve its objectives and add value to its parent
subsequent growth of captives owned by South African corporations has been significant.
company. As a consequence, potential captive owners sought jurisdictions where the
Over thirty captives have been formed in recent years, with most of these being established
environment was more suitable for captive formation. At this time, there was no or little
in either the Isle of Man or Guernsey. However, a number of South African captives have been
supervision ofinsurance within these jurisdictions and, although not necessarily served by a
located elsewhere, including Bermuda, the Cayman Islands and Jersey.
network of double taxation treaties many had low or no tax arrangements, which were a
significant attraction to captive owners.
Other domiciles with significant captive histories include Guernsey, Jersey, the Cayman
Other parts of the world
Islands and the Isle of Man. During the 1980s, many of these domiciles followed Bermuda's
Australia is probably the most competitive insurance market in the world, and this has a
example and introduced legislation to encourage the establishment of captives and other
significant impact on the relatively small formations of captive insurance companies. There
insurance vehicles. Once a small number of jurisdictions had been identified as offering the
are, however, in the region of thirty captives owned by Australian companies, and most of
appropriate economic and regulatory framework, captive owners were further attracted to
these are located in Singapore, although a number exist in Bermuda, Cyprus and the Cayman
these jurisdictions ifspecialist expertise was available in the supporting service
Islands. Captives established offshore are being particularly affected by the Australian accruals
tax legislation. 'This has the effect of taxing the profits of a captive as though it were resident
within Australia. The future benefits of establishing a captive in a low-tax area must therefore
be questioned, particularly given the Australian insurance market situation. US-based captives
New Zealand has also introduced controlled foreign company legislation that severely While current moves by the US regulators are a cause for concern, the introduction in 1981
inhibits the benefits of establishing captives offshore in territories other than a list of seven of the Product Liability Risk Retention Act in the US facilitated the development of new
developed countries. The potential for significant captive formation from New Zealand is also vehicles for transferring risk. This Act provided for the formation ofrisk retention groups and
limited due to the relatively small number of large companies in the country. risk-purchasing groups in the areas ofproduct and completed operations liability and was
In the Caribbean, locations come and go as politicians seek the development of an extended in 1986 to include all areas of commercial liability except workers' compensation.
offshore insurance sector as an important diversification from tourism and local industry. lhis Act exempts both risk retention and risk-purchasing groups from many ofthe state laws
Barbados has been particularly successful over the last twenty years, mainly as a result of an that would normally govern insurance organisations. In 1981, the state of Vermont was the
attractive double tax treaty arrangement with Canada and, early on, a treaty with the US that first off the mark to establish captive-specific
enabled captive parents to avoid federal excise tax on insurance and reinsurance premiums.
Approximately thirty US states are captive insurance company domiciles, and
After much debate, Bermuda eventually negotiated a similar treaty, but both treaties were
ended at the beginning of this decade and this advantage for Barbados and then Bermuda considering the aggregate number ofcaptives in all ofits states, in 2006 the Us was the
disappeared. The Bahamas has been a captive location since the early 1960s and has leading domicile country with 1 251 captives (of which 45% were located in Vermont), or
continually tried to regain the position that it lost during 25% of the world-wide total.
278 RISK
CAPTiVE tNSURANCE COMPAN'ES 277
this decade, when the government antagonised the captive community by threatening the introduction The US is followed by Bermuda and the Cayman Islands in terms of the number of
of a profits tax. licensed captives. In 2006, Bermuda had 20% of the world's captives (989 captives) and the
Cayman Islands had (740 captives).
MANAGEMENT: RISK FINANCING
Why the growth in US-based captives? Why Vermont? I. Bermuda has long-standing and progressive insurance legislation.
According to the Aon Global Risk ConsultingL6 2007 survey of the top 1 500 global (Gl 500) 2. The co-operation and flexibility of the Bermudian regulators and the extent to which
companies, the US-based Gl 500 companies clearly favour the establishment of captives consultation between the insurance regulators and the industry takes place are significant
onshore in the US. In 2005 and 2006, 80% of Gl 500 US-parented companies set up their factors. There is a regular review of the legislation and the practice of insurance.
captives in US onshore domiciles. This growth has possibly been triggered by recent US
3. Bermuda is a significant and leading financial centre that provides effcient access to
Department of Labour decisions to allow the underwriting of employee benefit risks and the
banking and reinsurance markets and expertise. Ibe presence of significant commercial
reinsurance of group life benefits by US-based captives.
insurance and reinsurance companies on the island with large amounts of capacity allows
As is reflected in the numbers, Vermont is seen as the leading US-based domicile. The
captives access to open-market underwriting capacity that is not found in other captive
view is that this is attributable to Vermont's being first-to-market in the captive space.
domiciles.
Vermont's regulatory environment is seen as its key draw card. Numerous reports mention the
4. Withits long-established association ofprofessional insurance management companies,
excellent relationship captives and their managers have with the Vermont regulators.
there is access to a wide range of regulated captive insurance managers.
Bermuda as a dominant captive domicile

Rise of Asia
Bermuda first developed the concepts of captive insurance in the early 1960s. According to
Only 27% of the leading companies in Asia had captives in 2006. A significant growth in the
the Bermuda Insurance Development Council, the bulk of Bermuda's captives are US-owned
number of Asian captives is expected as the captive concept gains recognition in this region.
entities that are often used to insure and reinsure retentions on general liability, auto liability,
Asian corporates are now in a phase of rapid regionalisation and globalisation, with gross
workers' compensation, property and marine programmes and to access the reinsurance
domestic product in China, India and the Association of South*East Asian Nations states
markets.
projected by the International Monetary Fund to grow by more than 8% in 2008 and 2009. *Ihis
The fastest-growing category of captives in Bermuda since the early 1990s has been the
growth is accompanied by new risk exposures, and companies are recognising the importance
healthcare or medical malpractice sector out of the US.
of ensuring that these risk exposures are controlled and managed in an effcient and cost-
The growth in the number of new captives licensed within Bermuda has slowed over the
effective way. In addition, with the increase in emphasis on corporate governance in Asia,
last five years and, in fact, there has been a decline of 14% in the overall number of captives
enterprise risks, such as the risk to business interruption posed by a potential pandemic flu, are
licensed between 2003 and 2006.
being focused on. Captives are being seen as vehicles to fill the gaps between the requirements
The US is the biggest single source of captive business for Bermuda, accounting for more
ofcorporates and the covers that are available in the market With this shift in the culture of risk,
than 60% of the island's insurance formations. However, Bermuda is seeing a change in this
Asia is seen to be at the initial stages of more sophisticated risk management solutions, and a
trend, with new sources of business emanating from Australia and emerging markets in
significant growth in the number of captives
Africa, the Far East, the Pacific Rim and Latin America as these markets become more
progressive in terms of the self-insurance retention strategies.
Singapore is the largest and most established captive domicile in the AsiaPacific region,
Despite the decline in new captives licensed, Bermuda is still seen as a leading captive
with over 60 captives registered. Other competitive domiciles are being established in the
domicile. Whereas it is no longer able to dominate other
region (e.g. Labuan; see below) to meet the growing demand for captive insurance vehicles
from the Far East.
16 Aon Global Risk Consulting (2007),
279
domiciles in terms of some financial factors, e.g. administration, financial services, cost of
capital, etc. (see above)) Bermuda is still attractive due to the following factors, in the main:
17 Global Risk Consutting (2007),
280 CAPTIVE INSURANCE COMPANIES 281
requirements for setting up captive insurance companies. Since then, the number
Choice of domicile by industry sector sometimes dictated by
ofinsurance licences has grown. Business is flowing into Labuan from Malaysia, Australia,
legislation Hong Kong, Indonesia, Japan and %ailand.
Aonl' survey of the top Gl 500 companies showed that Bermuda and Vermont retail
trade sector (64% are domiciled in these domiciles), transportation sector (60%), communication
sector (56%), oil and gas extraction sector (54%), and pharmaceutical and chemical Domicile selection
manufacturing sector (49%). With over 100 domiciles vying for business, there is a need to be able to compare their
relative advantages and disadvantages. The issues to take into account when making a
Health services companies are domiciled mainly in the Cayman Islands (57%). Over 30% of
selection ofa domicile can be broadly categorised into the following areas.
the Cayman Islands captives are healthcare captives.
Companies in the finance/insurance sector do not seem to favour any particular domicile
for their captives, with 28 domiciles across the world hosting captives in this sector of the Gl 500 Stability ofthe political environment
companies. Bermuda and Vermont account for 31% of the Gl 500 finance and insurance sector An analysis ofthe support for captive enabling legislation by all political parties vying for
companies, followed by Luxembourg at 15%, Guernsey at and Ireland at 9%. control of the government is crucial when comparing domiciles. Should control of
government or state change, the views on captives, their regulations and the benefits
enjoyed under the previous government may change.
Emergence of new domiciles An example ofthis is the Bahamas, where a change in government brought with it a
change in legislation that was not taken well by captives and that resulted in their flight to
other domiciles.
Malta came onto the European captive domicile scene in May 2004 when it acceded to the EU. A further example is the introduction of TRIA in the US, which forced captives to write
Marsh, which has over 80% of the captive market in Malta, observed a high level of interest terrorism risks as a back-stop to federal terrorism insurance. There is a view that this
during 2006 and 2007 in Malta, principally due to its capital requirements, operational cost requirement and the uncertainty of the nature of its filture under successive governments
effectiveness and approachable regulator. Malta's EU membership now provides the right of may force companies to consider locating their captives outside of the US, where such
Maltese captives to passport their licences into all EU and EEA states. Malta's redomiciliation requirements are not present.
regulations allow captives that are operating in other jurisdictions the opportunity to transfer
their domicile to Malta. Malta has an extensive double tax treaty network and protected cell Legislation and regulatory issues
legislation. The domicile should have regulation that adequately addresses the issues ofthe character,
capacity and capital of the owners and managers of captives in order to perform their
responsibilities and to competently manage their business. The presence of competent and
Anguilla, in the Caribbean, passed its new insurance legislation in 2004 and Protected cell hands-on regulators in a domicile is important to ensure the adequate management and
Company Regulations in 2005. This legislation is based on best practice legislation from the control of captives and other insurance vehicles in the domicile. Not only does this protect
leading international insurance domiciles. the beneficiaries of the captive, but it also protects the captive's reinsurers.
The ability of the legislators to adapt legislation in a responsible manner is an
important domicile selection consideration. Bermuda and Vermont are generally regarded
Labuan is a federal territory of Malaysia. In 1997, the Labuan government amended the 1990 Offshore as the leaders in this, with both having regulators who are very accessible to captive
Insurance Act, effectively lowering the capital managers and owners and who are prepared to be flexible when flexibility is justified.
282
Ability to accept and write business The ability of the domicile to supply these essential support services for the captive at the
required level of cost and professionalism was in the past a key consideration in domicile
The ability of an insurance subsidiary to accept business from different states or countries based
selection. However, the key domiciles have now attracted the required network of support
on its domicile is an important - often overriding consideration. Some domiciles prevent the
services needed to support their captive industry, and most international brokers, reinsurers and
writing of certain lines of business. Some jurisdictions, especially developing countries who wish
captive managers are represented in all major domiciles.
to protect their local insurance markets, prevent the placement of insurance business into
unapproved offshore structures. This can be particularly onerous for multinational companies
Cost
wishing to make use of a single captive. Solutions often involve the fronting oflocal business by
local insurers and the reinsurance of risk into the offshore captive. The most obvious cost is the minimum amount required to capitalise the captive vehicle. These
With the introduction of 'passporting' in Europe, captives domiciled in a particular EU or costs vary among domiciles and differ depending on the nature of the business being
EEA state are now able to accept business from all EU/EAA states without having to be underwritten by the captive. Other costs will include annual fees paid to the regulator, the costs
separately licensed in each state, as was the case in the past. of annual audits and reports by service providers as required by the regulators, the costs of the
As previously mentioned, within the US, there has been a restriction on the ability of captives meetings of the captive board, and taxation.
to write employee benefit-type business. Recent rulings by the Department of has seen this
change, but the process of applying for such a ruling is seen as onerous and it is therefore easier
to operate captives elsewhere for this type of exposure. Current moves by US legislators will also 11.5 Types of captives
have an impact. Captive insurance companies have, over the years and in various guises, developed into the
key mechanism of risk financing for large corporations. Owing to their flexibility and self-
Specialisation within domiciles
management (even if indirectly through appointed managers), they are the only real solution
Certain domiciles specialise in supporting captives that write certain lines of cover. The Caymans, to the most important problems that face risk managers when conventional insurance markets
for example, is the preferred domicile for captives writing hospital and medical malpractice are not able to satisfy or keep up with their needs.
coverage. Nevertheless, because of the entrepreneurial, actuarial and technical skills and experience
Further, due to geographic localities, domiciles tend to become familiar with the regulations of professional insurers and reinsurers, and their insurance business spread and capacity,
ofneighbouring or close jurisdictions. US-based domiciles are, naturally, very familiar with US captives are not likely to replace these markets. Indeed, a realistic viewpoint of captive
federal tax legislation, as is Bermuda, for very different reasons. managers towards the professional insurance markets should be one of cooperation and trade
whenever terms suitable to all parties can be negotiated. This is especially the case when
Accessibility dealing with catastrophe-level risks.
Captives can be classified on size into the following categories.
As most captives require an annual meeting in the domicile, access to that domicile is a
284
consideration. Flying times and local conditions are often an important factor in domicile
selection. Furthermore, the ability to combine business trips on captive business with other
business that can be conducted in the domicile due to the presence ofreinsurance and other
Paper captives
financial markets can be a factor.
A paper captive is established on a confidential low-key basis to provide funding facilities for
283
the risk exposure of the group. These companies are usually established on an exempt basis
and located in areas where little or no legislation exists and where the taxation rates are
For US-based companies, having the ability to form captives within the US itself is a key
minimal. They are often located in one tax haven and managed from another. They may be
advantage, compared to having the captive formed in Guernsey, Jersey or even Bermuda.
used for contingent risk, as well as conventional insurable exposures.
such captives are cheap to run and would most probably be managed outside the company
Support services
by an accountant or lawyer. "Ihey remain confidential to outsiders, but are susceptible to
Support services include local captive managers who understand local legislation and
legislation of the domicile in which they are located and to revenue authorities in the country
regulations and who have the ability to interface with the regulators; actuarial support; legal
where the parent domiciles.
support; and financial support in the form of accountants, investment managers and bankers.
MANAGEMENT: RISK FINANCING CAPTIVE tNSURANCE COMPANIES
Small-scale captives its own captive. The normal captive management services such as statutory reporting, claims
management and investment reporting are provided by the owners of the rent-a-captive and are
The small-scale captive is normally a captive that has recently been established, retaining only
not the responsibility of the participants.
a small risk, but with the view of eventually becoming a fullscale captive.
Compared to a wholly owned captive, the rent-a-captive programme has the following
Small-scale captives may be located domestically or offshore. They are normally managed
advantages:
outside the company. Since management is shared with other captives, the management fees
The rent-a-captive may offer more flexibility to an insured that only has a short-term need
are relatively low.
for specific insurance coverage. A wholly owned captive company remains an ongoing
They are reliant on the insurance market for support in two respects. Firstly, they usually
business regardless of whether the need for the specific insurance still exists.
need fronting facilities for policy issue and technical services, and secondly, they are generally
The costs associated with a rent-a-captive are usually lower than those of a wholly owned
dependent on the market for reinsurance, which would play a major role in their overall
captive. There are no establishment costs, and the management costs are shared by the
financial programme.
various participants. A captive company has ongoing operating expenses even when there
Small-scale captives make up approximately 80% of captives throughout the world.
are no active insurance programmes.
Rent-a-captive participants with smaller premium volumes may receive the benefits
Full-scale captives ofincreased investment returns because of the pooling of all of the participants' funds. The
increased size of the investment portfolio allows the investment manager more flexibility
A full-scale captive is a captive that has been established for many years and is considered by
to improve the performance of the fund.
its owners to be a major financial asset. It will be big enough to have its own management and
The use of a rent-a-captive may not require an initial capital investment from the insured
will be largely independent from the market because of its high level of risk retention. It can
company.
be located either domestically or offshore, depending on the needs of the parent.
Captives can also be broadly classified into 'ownership' and 'business philosophy' types.
than fcwty-frve ofthe total nurnbef of captives are clasg6ed as captives or rent-a-captives. majority c/ these ate reglst«ed in
'9rtetermtent-ecaptive'ls See the 1997 CoptivonsuranceCommny a mlsnomer-The arrangement Directory.is not one ofa captive whi:h irnplies so-ilk ownership or
investment. Rent-a-captlve are insurance contract-orientated; te,thecontractofinsurancepovides mechanism to return any surplus to the rnsured. As a
Non-owned captives or 'rent-a-captive' the term tontingency policy' has been IntrodKRed and Is belng used inaeasingiy too describe such retention arrangernents.
286
of
over the past years, there has been an increase in the use by organisations formal insurer-
related retention vehicles, such as rent-a-captives and core-cell 285
Participation in a rent-a-captive programme can usually be achieved in a significantly
18 shorter time than that required to establish a wholly owned captive company.
captives. Illis has been prompted by the need of insurers to get involved in retention-type
'products' and the examination by insureds Of costs relative to captive vehicles. Arguably, the
obscurity offered to insureds through, say, renta-captive arrangements has also led to their
Core-cell company
increasing popularity.
Description: Unlike regular captive companies, rent-a-captives are not formed to insure the The core-cell company takes the rent-a-captive one step further. A cell captive is a niche
risks oftheir owners, but those of unrelated organisations. These organisations wish to avail insurance company owned and controlled by the sponsoring or core capital. Generally, this
themselves of a retention vehicle such as a captive, but do not wish to form their own captive capital is suffcient to provide the minimum required solvency, as well as some additional
insurance company, so they become lessees ofa rent-a-captive. 19 The rent-a-captive is therefore capacity to serve as a "guarantee fund'. The main distinction is that each insured takes up a
an insurance company established to insure the risks of unrelated organisations (lessees) with the shareholding in the company through a particular class of shares to form its own 'cell'. Such
express purpose of returning underwriting profits and investment income to the lessees. In share capital is then used to support the particular cell's solvency margin and risk exposure.
contrast to traditional insurers, rent-a-captive owners are generally not in business to make a In the event of the capital of any one cell being depleted, the assets of only the core cell
profit by taking underwriting risks. Their income is derived from management fees and rentals are at risk, but those of the other leased cells are not. Ihis segregation of cell liability is achieved
paid by the participants. through legislation (for protected cell companies — PCCs) or contract. PCCs actually
Advantages: The principal advantage ofa rent-a-captive is an organisation's ability to guarantee that each cell within the company will be shielded not only from sharing capital and
recapture underwriting profits and investment income without having to establish and manage surplus with other cell owners, but also from any legal action brought against any other cell in
the company. Guernsey was the first domicile to enact specific legislation for PCCs. 20 The
Cayman Islands recently introduced similar PCC legislation, and Mauritius is currently Investment returns: Better rates of investment returns are achieved through the pooling of
replicating the Guernsey legislation. Gibraltar allows for a contractual structure. Isle of Man is the cell captive's total funds when compared with the returns achieved on individual (small
friendly toward the contractual approach, but has indicated opposition to statutory PCC captive) investment funds.
legislation, as has Jersey.21 In Bermuda, where most cells are registered, each new registration
is through an individual act of parliament, but PCC legislation is being
introduced. Single-parent captive
The advantages of core-cell companies include the following:
This is a wholly owned captive.
Ease of establishment: The core capital provider has already undergone the regulatory
and formation aspects of establishment; therefore a new insured wanting to join a cell
captive can do so by simply taking up a block of shares in the company and begin writing Joint interest or association captive
its insurance programme through the facility. This would be established by a number ofpersons in the same trade, profession or business who have a
Entry level lowered: The entry level for a business wanting to use a captivetype facility common insurance need that can best be met by
has been lowered and companies ofa small to medium size can
20 The Protected Cell Companaes Ordinance. 1997. 11-•15 is a stand-ahne law.- howeverr it is eons•dered an adjurEt to che Companies Law (Guernsey). I the
Insurance Butsiness Law (Guernsey}, 1986. and the Protection of each of cell.Investors Law (Bailiwick of Guernsey). 1987. lawallow: a PCC to create one more cells
fCK the purpose segregating the assets
Multi-parent captive
21 Many v,ithin (be ind LIStry are not fully convinced of the effectiveness of PCC legislation-There are questions as to wheKher courts outsidet?k jurisdict.on lhis will generally arise from a mutual need or common objective ofa number of insurance
would a-ccept the (0 arise PCC wher, concept in and any whether dispute assets held outside an the action jurisdk'tion was instituted could against be fully a pmcected
Of
PC-C in a In foreign an insoavenc.YjurisdKtiOClThe first question is likely between part\esr case. some bekeve that it would not be possible to that the li ofliabillty
buyers. The most common motivating factors are the advantages pooling capital and
provisions in the Guernsey model legislatlon management resources and obtaining a spread of risks. T•rere are instances where an insurer is
upheld Report Offshore 'strands expected
be whether PCCs In all under circumstances. [he legi dative The Tegime Edwards would survive on rhe legal Ilenge •n Review the courts outside to
included among the shareholders — usually a common lead insurer, which then contributes the
Guernseyacknowledge Rhat there is no way of
major portion
287
start making use of such retention facilities. Some offshore cell facilities will accept business
with annual premium levels as low as EIOO 000 (about
RI .2 million at 2009 exchange rates).
Ease of exit: It is relatively easy to exit a cell captive by selling the run-off risk, declaring the
distributable profits back by way of dividend and then redeeming the shares.
Flexibility on capital: With some facilities, only a relatively small amount of capital needs
to be invested in the cell, since the remaining required solvency is provided by the core
capital.
Accounting treatment/non-consolidation: From an accounting perspective, the cell shares
need not be consolidated into the parent company, and as a general rule, they will be
accounted for in the books of the parent at the original cost price.
Lower operating costs: Byvirtue ofthe po oling ofthe various administrative functions under
one licence, the operating costs should be lower than those of a wholly owned captive.
Taxation aspects: An important advantage of a cell captive relates to controlled foreign
corporation (CFC) legislation. When properly structured, a cell company may be classified
as a norFCFC, with all the obvious advantages.
RISK MANAGEMENT'. RISK FINANCING
CAPTIVE
RISK FINANCING 24
290 1.00
China 30
22
27
Sweden 1.76
1.42 26 16
Australia 27
1.45 21
1.69 25
Switzerland 2
RISK MANAGEMENT:
Spain 4 insurance vehicles could potentially lead to inefficient allocation of capital due to minimum levels
16
of prescribed regulatory capital and the stepped levels of regulatory capital required in a number
Brazil 2.00
11 of the major international jurisdictions. In addition, each additional jurisdiction would also
12
220 necessitate a minimum fixed annual operating cost. Furthermore, if the entities were using an
Belgium economic determination ofcapital levels as opposed to a regulatory measure, spreading the
12
Taiwan n.a. 9 various programmes over a variety of insurance vehicles would lead to the benefit of the
10
41 diversification of risk being lost ory at the very least, watered down.
Finland 1.50
Current work by the International Association of Insurance Supervisors (IAIS) on so-called
1.21
Other 'Solvency 11>23 could ultimately lead to more pressure on solvency margins and 'capitalisation'
issues, and increased costs could become a serious issue for businesses involved in multiple
insurance vehicles.
of the ten Gl 500 companies with five or more captives, seven are US parented.
ptimising investment strategy

lnåustry utilisation factors
In addition to the benefits to be derived from optimising the use of capital, additional benefits
Table 11.6 Industry utilisation factors could also be obtained by a consolidated investment strategy involving either the pooling of the
investments ofthe various insurance entities or rationalising into fewer entities or a single entity
Utilisation factor Take-up factor (or even the conversion of existing captives into cost-emcient cells in a segregated account
Industry 50%
company). The benefits could be derived by way of economies of scale and cost reduction (the
Agriculture, forestry, fishing elimination of duplicated costs, the negotiation of lower commissions based on higher volumes,
61%
1.95
Mining 46% etc.) and by the lowering of risk by improved diversification of investment due to greater capital
1.56
Construction 1.46 45% availability. Entities could possibly even afford to adopt a higher risk profile (and thus seek higher
Manufacturing returns) in respect ofsome of their investments should the amalgamation lead to the freeing up
Transport, comms., elec., gas, 1.54 of capital reserves beyond those of the immediate solvency and liquidity requirements. Issues
arising on rationalisation
sanitary 28%
1.27
Should the decision be made to rationalise a multi-captive scenario to take advantage of capital
Wholesale trade 42%
1.38 optimisation, cost reductions and improved investment returns, expert advice would need to be
Retail trade 55%
1.54 sought to ensure that the resultant processes are correctly managed. The processes of novation,
finance, insurance, real estate 49%
153 run off, reinsurance of long-term tails, etc. entail a number of legal issues that need to be
Services 40% addressed to ensure that all the parties to the contract are adequately covered. For example, in
1. the case of novation, differences between the various jufisdictions (i.e. conflict of laws) could
00 Public administration imal use of capital create problems. Both parties to the novation would need to be adequately covered in respect
When considering any situation where a parent has multiple captive/specia.l„ purpose ofany residual claims,
insurance subsidiaries, one needs to evaluate the trade-off between the relative cost of
the inefficient use of capital against the benefits received by operating in multiple
jurisdictions or by the separation or 'ring fencing' of potential liabilities. The spread of
premjse behind Solvency li is a simple one. common risK•based capitol framework for all European insurers and reinsurers-
insurance over multiple
292
and all liabilities forming part of the agreement should pass concurrently with the transfer of Captives may offer certain tax advantages over commercial insurance or
related assets, with adequate protection against third parties. In certain cases, access may need to
other forms of retention.
be provided to the old vehicle in the case of other assets being required to meet any claims arising
As in, for example, the US, commercial insurance premiums in South Africa are deductible
out of the novation. The legal liability ofthe directors and officers ofboth entities involved will
for purposes of calculating tax. If insurance is not purchased, the firm may deduct (most)
also need careful consideration.
uninsured losses as they are incurred. %ese provisions are not neutral in their impact on the
decision to purchase insurance if the time value of money is considered; the payment of
Practices relating to the consolidation ofcaptive results into parents' financials premiums precedes the occurrence of losses, and hence the tax benefit from premium payment
In terms of international accounting standards, the general principle is that where there is evidence arises earlier than the tax benefit from deduction of losses uninsured. This creates a bias in
of control, the parent body will be forced to consolidate the results of a subsidiary body. In favour of insurance over other methods of financing losses, but consideration needs to be given
jurisdictions where the international standards are not applied, the local accounting practice will to the recognition ofa captive as an insurer and the tax treatment of a captive's underwriting
dictate whether the results of the subsidiary entity are consolidated or not. The nature of the parent and investment income.25 These aspects will be dealt with briefly under the section that
- e.g. a trust or an unincorporated partnership, as opposed to incorporated entity; a listed entity, examines the operations of South African-owned captives.
as opposed to an unlisted entity; as well as the requirements of the 'controlling' entity — also
influences the accounting.
Genera! market conditions may result in particular types of cover
being difficult or expensive to obtain in the commercial insurance
1 1.6 Features of and motivations for forming captives market.
A popular argument is that captives were established because the commercial market has failed
Investment income earned on reserve funds accrues to the captive, as to satisfy the needs of insureds by failing to provide premiums that fairly reflect the risk
opposed to an unrelated insurer. exposure or the insured's loss record. The foundation of this argument lies in the imperfection,
or cross-subsidisation (over different types of insureds), that may arise in the commercial
Reserve funds and paid-up capital form a pool of investable funds; such investment income
market. A captive provides the opportunity for a firm to use its own data and set its own rates.26
serves to reduce the overall cost of operating a captive. In the case of purchasing commercial
This function is only of value if the firm is able to produce a more reliable estimate of the
insurance, the investment income accrues to the insurer, but in a competitive environment, this
probability distribution of future losses than that produced by the commercial market.
income might be reflected in premium rates. However, imperfections may restrain insurers from
fully passing on the investment income to the insureds in the form of lower premium rates. Also,
the investment regulations of the domiciles of captives may be more liberal, which, given that Underwriting costs are often lower than those evident in the
the captive has access to skilled and effective portfolio managers, may result in superior
commercial insurance market.
performance. Finally, the investment income from offshore locations is often treated more
Commercial insurance is expensive relative to the expected value of loss. Kloman and
favourably from a taxation viewpoint than is domestic investment income. 24 Naturally.. value
Rosenbaum27 claim that with commercial insurance, the expense
will be created if the after-tax yield exceeds the cost of capital.
be thatttEtaxation ofcaptive insurancetmsactionsis complex, variousreasonsTax impceclse in definjrtg insuance therefore. the circumsta'Ees under whlch
captive will be recEnlsed as an pnsurance company for tax purposes. Also. captive transactions •e Often internaqonal oature. thussubjecttotheJurisdctioosofmore
thanone tax authority. Itis not intended to provide an in-depth an&lysis ø•; pertarnjng to captive insurance. This would entail a separate intensive study in its right.
of well as it be far to state that the liability crisis (discussed in chapter has prompted a considejable number of hrrns (some non-pccflt ;anisatiöns) to establish their own captives
that pcovide liability-related insumce covers arnd that alsoact as expedient vehicles inrisk/claims/ see Otto {T989_ 1—9).
293 (ear. and Rosenbaum 0922; t36)
this section
RISK MANAGEMENT. FINANCING CAPTIVE INSURANCE COMPANIES
294 RISK the loss is transferred to a related third party, this moral hazard cost is nonexistent. With a
captive, therefore, incentives to formulate effective risk control programmes are sustained.
The features described in the preceding section provide the motives for establishing a
ratio is 0,3 and higher, as compared with expense ratios of captives that range from 0,05 to
captive. When observed in an overall sense, these can be condensed into the following two
0,1. Part of this saving arises from the exclusion of marketing costs (including commissions)
broad motives.
and insurer profits.
A captive may be formed with the object of obtaining insurance on

The captive has access to the reinsurance market, which, by virtue of its
'wholesale' nature, is thought to be more competitive than the direct more advantageous terms.
(retail) market. The intention here is to establish the captive as a conduit to the reinsurance market,
which is perceived to offer more-favourable terms and conditions.30 Thus, there is the motive
By forming a captive, a firm is not restricted to placing its insurance in the direct market, but
to obtain meaningful pre-loss financing from an unrelated reinsurer, although this does not
has access to the more competitive reinsurance market. The firm will enjoy even greater
imply that the captive will not retain some risk for itself.
advantages and more flexibility when negotiating deductibles with commensurate premium
discount. In the direct market, insurers are often reluctant to issue policies with high levels of perceived advantages of directing the risk premium to reinsurers through the
deductibles or to offer significant premium reductions. medium of the captive are also (qualitative' in nature. Since the cession of premium is in effect
a cession from the direct insurer (the captive) to the reinsurer, there may be the opportunity
Several reasons have been proposed to explain the difference in intensity of competition
to disguise the risk source. This is not to suggest that captive parents generally wish to maintain
between the reinsurance and the direct market. These may be briefly itemised as follows:
anonymity, but, rather, that the opportunity to do so exists. A major motivation for South
Direct writing is usually more regulated than reinsurance. Regulation is not aimed at
African organisations to form a captive has been that it could allow more concealed
keeping rates low, but rather at protecting solvency and protecting local markets from
accessibility to international markets. A captive could therefore facilitate the marketing of
international competition? Moreover, often such regulation is protectionist in nature,
certain risks in, say, politically sensitive or nervous markets.31
resulting in inefficiency
market for direct insurance is nation-wide in its dimensions, while reinsurance is

A captive may represent a genuine pre-loss financing alternative to insurance.
often transacted on an international basis, and this exposes firms to more intensive
Where there is the intention to retain risk within the organisation, the captive competes with
competition.
alternative methods of reinvestment financing such as internal retention funding, new debt
issues, new equity and commercial insurance. In such cases, the captive may or may not be
Typically, direct insurers incur higher costs ofproduction and marketing than do reinsurers.
heavily reinsured. It is probable that in the early years there will be substantial reinsurance of
the captive until its retention capacity grows to allow it to carry more ofthe parent's exposure.
Incentives for effective risk control are stronger than under There may also be the opportunity (and willingness) for the captive to write unrelated third-
commercial insurance. party risks, in which case the captive assumes a wider responsibility and acts as what is
termed a 'senior captive'.
It is thought that insurance leads to the moral hazard problem (in its correct definition) when
29The is Swaziland legislation that tobe placed
Royal and a offence in of sec. 3 ofthe control of order
b
ebefrasfikntioned earlier, arise from the differences in theIntensityofcornpetitlon in the directandreinsuance ma•kets. Although rhe eWhents put forward are
losses are transferred to an independent third party. With commercial insurance, the higher plausible. it should be that no sclentiöc stu$' exists to support the hypothesis.
ISSue was seen by South African insureds as being •31strategic Impottance. Captives were seen to faclhtate transfer to makets wrliing to bushess so long as the
total claims costs resulting from moral hazard factors will ultimately be borne by policyholders risk SOL111.-e was djsgu.sed. is not possu to app•ase the effectivenessofsuch a mechanism (exante}. as is part on the strength of econorntc motives Of
(re)insurers under Imposed conditions trade restrictlöns
as a group, but if 295
296
RISK MANAGEMENT. FINANCING CAPTIVE INSURANCE COMPANIES
(g) Other risk management and investment considerations: A group facility/ insurance vehicle
1 1.7 Captive ownership structures
acts as a central vehicle through which risk management objectives (rewards and penalties)
Any decision to invest in a special-purpose vehicle or a specially designed third-party insurer are achieved. They also have proven to facilitate and improve risk management data
to solve a risk-financing need is generally driven by the consideration of the following criteria: collection, collation and analysis. While these might be seen as 'soft' criteria, many
organisations have reported them as significant. Finally, the insurance vehicle can also
(a) Smoothing of losses: Insurance is essentially a smoothing mechanism providing provide a foreign currency hedge for those risks for which the conventional covers/ hedges
organisations with protection against the erratic nature of losses. One would typically look are inadequate. Over time, with the accumulation of risk reserves, the insurance facility
to determine a first-layer risk exposure to set the level at which one would wish to smooth can also offer further risk diversification through a strategy of international investment that
the volatility of the risk exposure. may prove complementary to the entitys overall investment activities.
(b) Reserving: Risk is often considered as a one-year phenomenon, the effects of which can
be gauged, say, at the financial year end and neatly processed for the sake of the entity's Contemporary solutions
financials. In practice, this is not the case and liabilities generally take a long time to Much ofthe above has been historicallyachieved by organisations incorporating their own
finalise. In some cases, this could even take as much as 10 years. Business entities today captive insurance companies, or cell captives. Today, these still operate and represent significant
attempt to avail themselves of facilities that effectively provide them with a reserving insurance conduits satisfying a variety of risk management needs. Ihese alternatives were termed
capability to better finance their strategic exposures and be more prepared for the downside 'alternative risk transfer' (ART) techniques and referred mainly to seff-insurance arrangements.
and volatility of these. Over time, the rules and accounting treatment ofsuch arrangements have developed, so that today
some of the criteria listed above are not satisfied through the
(c) Premium pricing management: In a shorter term, reserving allows the entity seeking risk
The industry, however, together with consulting participants, has been developing
finance more flexibility when negotiating renewal terms with insurers in a hard-market
insurance solutions, given that the need for these still remain with organisations that have
cycle. The pricing and capacity vagaries of the insurance market can therefore be more
complex and diverse risk exposures. These are quite contemporary solutions that seek to be
effectively managed.
more insurance related and are termed insurance and risk integration strategies' (IRIS). In
essence, specialist or niche insurers offer their licence and capacity to insure first-layer and
(d) Transfer of risk: A well-structured facility allows the transfer of risk from the entity's
strategic risks on a basis that is affordable and where the policy terms are flexible, but robust,
balance sheet covering both shorter-term and longer-term exposures. This criterion would
thereby offering their insureds some control over the financing of their risk exposures and
be considered in tandem with the ones listed above.
allowing these to be smoothed over time. Such strategies are seen to go hand in hand with an
organisation's enterprise risk management initiatives, where having identified exposures and
(e) Retention variations and buy-downs of exposures: This is a very practical consideration.
evaluated these in terms of balance sheet threat, organisations seek financial solutions aimed at
Often a group first-layer exposure is negotiated that provides the entity with the most
providing for some of this volatility over time.
efficient results with insurers, but from a divisional perspective, this level is unaffordable
298
and not in 'balance' with its scale of operation. A well-structured facility/insurer can offer
buy-down cover at a scientifically derived premium. One can typically achieve a practi cal
architecture within the main insurance structure. The mechanics of the policies offered by such specialist insurers vary, but typically a
number of specific key issues need to be addressed, as follows: The insurance contracts need
(f) Access to reinsurers: A group facility in respect of a major corporate ent ify• allows access
to be entirely at arm's length.
to the reinsurance/wholesale market. Cost savings may well be achievable by accessing
There needs to be significant risk transfer.
reinsurers in this way, instead of, say, paying
There must be no asset right to the insured.
297
Reserves are accumulated and accessible through other means.
higher prices for local capacity, which then itself reinsures to the wholesale
RISK MANAGEMENT: F-INANCING CAPTIVE INSURANCE COMPANIES
Premium reserve/capacity needs to be secure. An exit for all future eventualities, the trend in accounting practice has tended towards minimising
strategy needs to be available. fie policy needs to be reserving and seeking to account for all possible/potential assets.
cost effective. While each independent jurisdiction has its own specific tax rules, taxation is often driven
There should be no negative taxation implications. by accounting treatment either because of the nature of the relevant tax legislation or simply
because of practical issues of disclosure, and therefore conflict between insurance requirements
Historically, the following ownership structures have developed. and accounting treatment often leads to negative tax treatment of insurance-related issues.
Single parent: This would be a wholly owned captive, with 100% of the equity owned directly
by the parent. Specific international accounting issues
lhe current international insurance accounting environment is governed by various statutes and
Rent-a-captive: This would be a so-called segregated cell-type company, the individual cells
guidelines; more recently, International Financial Reporting Standards (IFRS) 4 and
of which are capitalised by the third-party owner and the capacity provided
is rented to clients of the rent-a-captive. International Accounting Standards Board (IAS) 39 have had a major impact, and there could be
far-reaching implications with the development of so-called 'Solvency Il'.
PCC: A protected cell company is a segregated cell company created in terms
oflegislation in the form of a PCC act. IFRS 4: An overview
IFRS 4, which deals with insurance contracts issued by any entity and reinsurance contracts
ICC: The incorporated cell company is similar to a PCC, with the exception that the individual
issued or held by an entity, is a concise but comprehensive statement as to the required
cells have the ability to contract with one another and can at a later stage
accounting treatment of insurance contracts by insurers. It is written in language that enables
be spun off as separate individual legal corporate entities.
firms to structure insurance policies with substantial certainty as to the accounting treatment
RRG: A risk retention group would be established by a number of people in the same trade, that will flow from the contract terms adopted. This does not mean that IFRS 4 provides easy
profession or industry who have common insurance needs that can best be solutions for specially structured insurance policies, but certainty, even in respect of aspects
met by combining resources. regarded as negative, is always a virtue in planning.

Whereas it was proposed that IFRS 4 should be applied for financial periods beginning on
Mutual: "Ihis would arise from the mutual need or common obj ective Of a number of buyers. or after 1 January 2005, entities were permitted — and, indeed, encouraged — to apply the
The most common motivating factors are the advantages of pooling standard as soon as possible. Many international
i
capital and management resources and obtaining a spread of risks.
nsurance companies opted for early adoption, and IFRS 4 can still be seen as the current
Essential features here are that the management standards of the benchmark.
participating organisations are known to one another and risk spread is a
reality. the discussion Carnation Oftax relating a 10 the US see Pine, of the Sta ngerand Wright (1978: 1 1), which IS Peter found was Theodore published Commissioner after the IRS rssued 3B its
tubng, but
299
'79. l). and in the following cose and court gives cases clear and explanatlon subngs; Hefvering IRS v Le position. Giese 312 Furthel US 531 reference {1941 j; in Pine and Wright (1979:37)
TC 101 Mid 1 Wavaht(1 962};
Ruling 77-316 Carnation ComponyvCommjssioner 71 TC 400 (1987). affrmed 640 E 2d 1010: Cen Denied 454 VS 965 {1981b, Humana F. 2d WL (1989).
11.8 Accounting, tax and regulatory environmental trends

The establishment and operation of captives raises a number of tax issues that have been subject
32
to much debate, particularly in the US. Insurance needs and accounting practice have tended
to diverge over the past decade - while sound insurance principles would call for the provision
MANAGEMENT. RISK FINANCING
INSURANCE COMPANIES
300 that the revenue authorities may well see some of the disclosure now being 'imposed' by
accountants on their clients by the interpretation of accounting practice as supporting the
disallowance of premium expenditure for tax purposes. This can, of course, create real problems,
Appendix A to IFRS 4 contains a crisp definition of an insurance contnct as as, to some extent, there is an element of subjectivity in the tests for the deductibility of expenses.
This introduction of an element of subjectivity in the tests may well lead a court to look to the
a contract under which one party accepts significant insurance risk (defined as 'risk,
treatment of the insurance premium and/or the relevant policy of insurance in the financial
other than financial risk, transferred from the holder of a contract to the issued') from
another party . . . by agreeing to compensate the policy holder if a specified uncertain statements of the taxpayer as indicating the true intention of the taxpayer. It follows, then, that it
future event . . . adversely affects the policy holder. would be important for any insurance policy or plan to be organised in such a way as to lead to
accounting treatment that accords with the taxation treatment one seeks to achieve.
Appendix B of IFRS 4 expands upon and amplifies this definition by way of examples. Complication in the revenue authorities' interpretation of transactions surrounding insurance
The question as to whether or not the insurance risk can be regarded as significant is contracts could lead not only to the disallowance of premiums and the bringing 'to book' of assets
discussed in IFRS 4, Appendix B, paragraphs 22—28. It is clear that risk is significant 'only to the parent's balance sheet, but also to an attack on the reserving at the level of the insurance
if, an insured event could cause an insurer to pay significant additional benefits in any vehicle.
scenario'. It is not clear as to what benefits are to be measured against in order to determine
that there are 'additional* benefits, but presumably it will be a measure against premiums paid. Application of IAS 39
The word 'significant' is not quantified, but on reading Appendix B, paragraphs 22-28, one is Some contracts previously classified as insurance contracts may not be insurance contracts as
left with the impression that, while there must be a measurable 'addition'* the test is not to be defined in IFRS 4. If these contracts create financial assets and liabilities (deposits), IAS 39
a harsh one. Various attempts have been made to benchmark the passing of significant risk applies.
and a range of between 10% (as an absolute minimum) and 20% (a safety zone) has been Financial assets will be measured in accordance with IAS 39, often at fair value. To avoid
applied in various circles. Appendix B, paragraphs 22—28 would imply this sort of range. an accounting mismatch, entities often change their accounting policies for insurance liabilities
While it is clear that accounting treatment will not be definitive as to the treatment for so that both assets and liabilities reflect changes in market conditions (particularly interest rates).
tax purposes, the adoption of the definition of 'insurance contract' in IFRS 4 to be applied Solvency Il has increased the level of debate in this area.
internationally must be regarded as bringing to bear significant influence on the tests to be The following points have triggered the trend.
applied by revenue authorities and income tax courts. This will apply to the treatment of both
the insured and the insurer. Development of derivatives: "Ihe current rapid increase in the creation of and trade in
An important aspect of IFRS 4 is the requirement in paragraph 34 that a 'discretionary derivatives, allowing for geared or leveraged investments and hedging, has also impacted on
participation feature' (defined in Appendix A) in an insurance contract be identified and insurance programmes. Derivatives have much in common with insurance contracts, and any
separately accounted for by the issuer of the contract. While the entire standard, and thus its concern over the way that derivative instruments are reported/handled can thus spill over i nto
paragraph 34, refers to the issuer, it has obvious implications for the insured, who may well the way that captives are monitored. It appears that much of the current international accounting
find that the deductibility of premiums paid in respect ofpolicies so accounted for by the issuer practice development, e.g. as in IAS 39, has its origins in the development of derivatives.
is queried by the revenue authorities. 302
While reference has been made above to IFRS 4 as a point of departure one must
understand that, unless so provided in legislation, there is no direct connection between the
accounting and taxation treatment of insurance premiums and policies. This statement is Focused legislation, e.g. tax treatment in the US: A topical issue in the US relating to the IRS's
supported by the findings of various courts internationally in the course of deciding on appeals treatment of insurance transactions with captive entities has its origins back in 197% when the
in taxation cases, These are only a few instances where income tax acts refer to accounting IRS issued Revenue Ruling 77-316, which disallowed insurance status to any captive whose only
treatment or practice, insurance premiums, and/or policies. business was derived from its parent. The ruling laid down the principles of the economic family
CAPTIVE 301 theory, in terms of which since the captive was a wholly owned subsidiary of the parent and it
wrote no unrelated third-party insurance business, it did not meet the two tests of risk transfer and
risk distribution required to qualify as an 'insurance company'. The view was that since both the
In practice, the revenue authorities will often base a disallowance of expenditure on the form
parent and the captive subsidiary were members ofthe same 'economic family, the risks in the
of disclosure in the relevant financial statements. In the case of insurance, there has been a concern
RISK MANAGEMENT'. RISK FINANCING CAPTIVE INSURANCE COMPANIES
captive had not been actually transferred off the balance sheet and certainly not distributed over Standard model with no further modification
a pool of unrelated entities. 27%
Standard model with further modification
After numerous tax and circuit court challenges, culminating in a ruling issued in 2001 15%
(Revenue Ruling 2001—31), the IRS acknowledged that the economic family theory was Stress test/scenario analysis 71%
Stochastic internal models
insufficient evidence on its own merit to judge a captive's status as an insurance company.
Combination of different models 56%
In a complete reversal of the abovementioned process, the current regulation proposed by
the IRS seeks to defer the tax deductibility of incurred losses - i.e. a captive's loss reserves and Single-period model 2%
the parent's premium, obtained from related party business (the captive's parent) - until the loss Multi- period model 26%
is actually 35%
This means that for tax purposes, the once-acceptable insurance transaction between parent and Use ofeconomic versus regulatory capital levels: The KPMG survey indicated that the market
captive would be disallowed. has tended to lead the regulatory environment in this regard with a significant number of the
Despite the obvious intentions of the IRS to attempt to tighten the tax net and improve the entities surveyed placing less significance on economic capital indicators as opposed to
timing of its revenue collection, it is thought that this proposed regulation may well have the regulatory capital requirements, to set levels ofoperational capital required. In the event of the
opposite effect of what it intends and cause a loss of tax revenues at both the federal and state full implementation of Solvency Il, which will force companies to employ a capital level that
levels resulting from an exodus of US-domiciled captives to offshore domicilium and a better matches the true risks they face, there should be a convergence between economic and
reduction in the numbers of newly formed US domestic captives. This process would be regulatory capital in the future.
facilitated by the fact that while the proposed regulation will disallow a captive from
Issues paper on the regulation and supervision ofcaptive insurance companies: This was adopted
restructuring itself specifically to avoid application of the proposed legislation, there is no law
at the IAIS annual general meeting in Beijing in October 2006. It was expected that the IAIS
that forbids a captive's parent from forming an offshore company into which the domestic
captive guidance paper (discussed at the 14th annual working meeting in Guernsey in 2007)
captive would move 100% of its business, thus winding down the domestic captives or engaging
would be finalised in 2008.
in any number of other legal strategies that would protect its insurance accounting status.
other taxation considerations: Income taxes are only one of the aspects that need to be
South African regulation
considered in structuring an insurance programme. In addition. other taxes such as premium tax
Four etchange
aspects ofregulation are important to captive insurers taxation, insurance, control and
(see references to 'passporting'), stamp duties. etc. will often drive the decision-making as to the
accounting standards. Each is considered in turn.
jurisdiction and location of insurance vehicles, programmes and policies. Many insurance
304
jurisdiction$ provide for a zero tax situation, which is important for the purpose complications
in respect of double taxation, as opposed to the
303
Taxation ofcaptive insurers
misconception of being driven purely by a general desire for tax avoidance. Stamp duty and similar Taxation is transaction specific and the final tax liability is determined by the
exemptions may also be provided. treatment of a number of transactions. A number of these transactions as applied to captive insurers
are considered.
Risk and capital management: The general consensus of the industry surveys reviewed indicates Tax deductibility of premiums paid to captives: It has been recognised and accepted for
that the insurance industry lags behind the banking industry in managing investor capital several centuries that taxation must be levied on income and not on capital. 33 To determine the
effciently and has generally not embraced the need to calculate economic capital requirements, income earned by a company, all expenses incurred in the production of revenue must be
as opposed to basic regulatory capital. subtracted from the revenue. This principle is encapsulated in south African tax legislation and is
It appears that there is inadequate use ofinternal capital adequacy models — only 47% ofthe to be found in the so-called general deduction formula, derived from reading sections 11 (a) and
companies surveyed in the KPMG Risk and Capital Management Survey conducted in 2004 23(g) ofthe Income Tax Act No. 58 of 1962, as amended. In terms of these two sections, read
used an internal capital adequacy model. Germany, Bermuda and South Africa were the together, a company that derives an income (revenue) from the carrying on of trade, business or a
territories that were most advanced in this regard. 'lhe survey further revealed the following
venture may claim a deduction from this income (revenue) if that company has:
breakdown of usage among the 47% that responded positively:
• actually incurred expenditure or loss in the Republic June 1992 that the Income Tax Act would be amended to exempt non-residents which was the
• in the production of income that • is not of a capital nature during

original intention. This exemption was achieved when section 10(1)(hA) was passed. The
amendment that grants the exemption (as subsequently amended) reads:36
the period of assessment.
interest received by or accrued to a person (other than a company) who is ordinarily

The application of these principles allows bona fide risk-related premiums paid to captive resident outside the Republic or a company which is managed and controlled outside
insurers to be tax deductible. It is not necessary to discuss the application of these principles to the Republic: Provided that -
captive insurers, since it has been the practice, quite correctly, to allow as a deduction insurance (i) the exemption under this section shall not apply to any natural person who was at
premiums paid to captive insurance companies. There does not appear to be any reason why any time ordinarily resident in the Republic if such person has during the year of
deductibility in respect of genuine risk-related premiums paid to captives should not be allowed assessment carried on business in the Republic; (ji) for the purposes of this paragraph
for tax purposes. It has been the practice of companies to claim such deduction, as noted by the the expression 'Republic' shall include any country which has for the purposes of
applying any regulation made under Section 9 of the Currency and Exchanges Act,
Melamet Commission, which accepted that companies were in fact making deductions for
1933 (Act No. 9 of 1933), been included in the common monetary area;
premiums paid to captive insurance companies.
Taxation of interest income earned in South Africa

Repeal of non-resident withholding taxation on interest: A captive writing business inside
South Africa, but registered and located outside South Africa, may be liable in south Africa for
tax on interest income accruing on deposits held in south Africa. This is because the source u Yavo•ær, February 1993, 28
ofAct No. 90 oi 1938. held in R v Buchanan 1914 AD S09:'For business to be carried on In the Republic, the premiums must be received within It and tere *'$1
be manager, secretary or agent and DavlS. 1983.26).
principle recognises that this income is from a source in south Africa. is complex
history behind this issue. 306
(iii) for the purposes of this paragraph, so much of any dividend as has been distributed
by any unit portfolio constituting a company in terms of paragraph (e) of the
33 See in paitkular the writings of David Riff-ardo {1817) and JS Mill (1848 Bk 5, par.. 7)
305 definition of 'company' in section 1 out of interest derived by such unit portfolio
which is exempt from tax in the hands of such unit portfolio under the provisions of
rlhe offshore captive company can for tax purposes be a non-resident person. From 1967 to paragraph (iA), shall be deemed to be interest;37
1988 a non-resident withholding tax was levied on interest accrued in South Africa. In 1988, the
(iv) the exemption under this paragraph shall not apply to any natural person unless
then minister of finance announced that non-residents would no longer be subject to South African
such person was physically absent from the Republic for a period or periods of at
taxation.34 This statement culminated in Parliament abolishing section 64A—F of the Income Tax
Act, i.e. the withholding tax provisions. 3S For a while there was confusion regarding the meaning least 1 83 days in aggregate during the year of assessment in which such interest
of this statement. The confusion arose because of the source principle. Source is notoriously was received or accrued;3B and
diffcult to determine, as it is determined by factual circumstances. In terms of case law, the one (v) the exemption under this paragraph shall not apply to any interest received by or
test for source ofinterest on a loan (or deposit) is the place where the provision of credit was made.
accrued to a companywhich is managed and controlled outside the Republic, if such
If the supply and use of that credit takes place in the Republic, the source of interest earned is the
Republic. This interest is thus included in the taxpayer's gross income and, prior to the repeal, a interest is effectively connected with the business carried on by that company in the
withholding tax was paid. Republic."
The minister had announced that interest income on non-resident accounts would no longer
be subject to this tax. Eventually, the Receiver of Revenue indicated that because of the source Hence, from 1992, interest income earned by offshore captive insurance companies became
principles despite the repeal of the withholding tax, the income was still subject to taxation. This technically exempt from tax. However, the Receiver of Revenue was clearly not happy with this
caused some confusion and the announcement caused considerable volatility in foreign exchange position, and a proviso was introduced in 1996, namely proviso which appeared to disqualify
transactions. The matter was then clarified by the then minister of finance, who announced on 3 captives from the exemption. The position is now governed by the source and residence principle.
Taxation of captive profits: Since the repeal of the strict application of the source principle, non- Insurance regulation: The Insurance Act
resident companies could be subject to taxation in South Africa. If the offshore captive is a domestic A number of provisions of the Short-term Insurance Act No. 53 of 1998 are now considered.
company for tax purposes, i.e. one that is managed and controlled in South Africa, 40 then it would Three sections in particular are relevant. These are sections 7, 8(2) and 56-63.
also be subject to tax in South Africa. Offshore captives are not usually managed and/or controlled Registration as an insurer: Section 7 (previously section 5)" prohibits the captive
from within the Republic. insurer, as an insurer, from carrying on insurance business in South Africa unless it is duly
registered. The relevant portion of section 7 reads as
Captive insurance company: The question now of whether the captive income represents South
African-sourced income for purposes of tax needs to be examined. Different tax regimes exist -
the US, for example, has a world-wide tax base. South Africa has changed from the source to the
residence principle.
Section 28(2) of the Income Tax Act deals with the determination of taxable income of a
taxpayer carrying on short-term insurance business in the Republic of South Africa.41 An offshore
captive would be liable in terms ofthis section for tax in South Africa if it were carrying on short-
term insurance (or reinsurance) business in South Africa.42 This section and others can only be
need be met so to sec. 7 Of Act. which those on
applicable if the taxpayer carries on short-term business in the Republic.

"QCisxaed as an msorer under rot this Act on insurance the purpose business. . of carrylng no person on tt)öt shall carry of on insurance any class buslness:of insurance business in the Republic
37 Para. ('Il) added by sec. ofAct No. 1 13 of 1993. unless
38 Para. (tv) added by sec. 10(1) {d} of Act No. 21 of 1 gos
39 Para. (PIA} inserted bye,ec, i 011 and ) (n) ofAct No. of1 41 non-resldents of 1992; para_ (v} added by set 5 as of amendedAct No. 36 of t 996 510<1} (HA) was deleted in of
40 51411 NC) of Act 32 of 2004 defined the as position a Sou(h African company is or governed 3 company by that is managed and controlkd In In the Republic.
41 AdornestiC Getz am Oavis company' is on R 1 91 4 AD 509, state_'FCK business to be carried on the Republic. tre p«emilJÉM (1 26).
rely•ng rece•ved väthin it and there be a manager, secretary or agent here: 42 The bass of taxation being as is set out in this section.
307
Dividends: Where captives are established mainly for the purpose ofproviding a more effcient
risk management system, there is no real intention to pay dividends to the parent company.
In this case, thus, the issue ofdividends does not arise. If it does, it should be treated in the
same way as dividends to parent
Captives and tax avoidance: It has been suggested that a strong motive for the formation of
captives is tax related, but in view of the extensive use of captives where tax deductibility does
not exist, the correctness of this view cannot be sustained. Should a captive insurance company
be formed solely for the purposes of avoiding, postponing or reducing tax liability, then it
would run the risk of falling foul of tax avoidance provisions in the legislation.
Summary: Taxation
In summary, South African-owned offshore captives may be liable for tax on interest income
earned on deposits made in South Africa. In order not to attract South African tax in terms of
various other sections, the following principles should apply: 43
Management and control must be outside of South Africa. The offshore captive must be
operated on the principles and practices of a normal insurance company, with transactions
being conducted at arm's length.
Premium to risk relationships should be normal and in accordance with the market place.
RISK MANAGEMENT: F-INANCING CAPTIVE INSURANCE COMPANIES
308 Sections 56-63: Placing insurance business outside of the Republic of South Africa:
Sections 56—63 replace section 60 of the previous Act and are the closest that the Short-term
Insurance Act comes to prohibiting the placing of insurance business outside of the Republic.
7(1) No person shall carry on any kind of short-term insurance business, unless, that person
- These sections are not as comprehensive as the repealed sections in that there are no
(a) is registered or deemed to beas a short-term insurer, and is authorised to carry on the kind of equivalents to the confusing section 60(2), which was applicable to placing business outside
short-term business concerned under this Act; or of South Africa and hence to captive insurers.
(b) is authorised under section 56 (the Lloyd's provisions) to do so, and carries on that A general observation is that the new Act is less clear than the repealed Act, and this is
busi ness in accordance with this Act. problematic.
The wording in the new Act is broader than in the old one. The old Act specifically referred to Exchange control
carrying on business in the Republic of South Africa, but the wording of the new Act is not limited For a long time, the detailed rules governing exchange control were not widely publicised or
to the Republic. The previous Act referred to class of insurance, which was defined;45 the new Act accessible to the public. The system was governed by circulars from the Reserve Bank to the
refers to any kind of insurance, which is not defined. It is not clear what the effect of the broader commercial banks, which it has appointed as authorised dealers. It is no surprise, therefore, that
wording is, if indeed there is any. it was in an exchange control case that the court ruled that ignorance of the law can indeed be
Assuming the Act is confined to South Africa, in order not to contravene this section, the an excuse.48 Offshore captives are concerned with exchange control issues and ostensibly the
risk manager should ensure that the captive is either registered in its own right or as a cell captive, authorities were concerned about the extent to which captives were responsible for the outflow
in which case the host insurer must be registered. of funds from South Africa.49 It is necessary, therefore, to note the requirements with regard to
If the captive is not registered in South Africa as an insurer, it is a foreigndomiciled captive offshore flows with respect to captives. This could be capital sums needed to capitalise the
and must then make sure that it does not carry on insurance business in the Republic. In this case, captive or premiums paid to captives. The placement of insurance outside the Republic is
the captive should not have an offce, secretary, manager or agent in South Africa and all governed by the Insurance Act and exchange control regulations.50 It is explained,51 among
negotiations, transactions, etc. should be conducted outside of the Republic. The new Act other things, that:
contains deeming provisions that should be noted."
no person may place his insurance outside the Republic unless if not placed in
accordance with the provisions of section 60(T) or 60(2) of the Insurance Act, the
Section 8(2) rendering services as an intermediary: The new Act does not appear to contain an
placement abroad has previously been approved by the Exchange Control
equivalent to section 75, which read:
Authorities'
Any person who induces or attempts to induce any person to enter into a contract of insurance
The Insurance Act provisions were repealed, as discussed above, but the principle
with a person who is not registered as an insurer under this Act, or to make an application to
enter into such a contract with such last mentioned person shall be guilty of an offence and liable remains. The permission ofthe South African Reserve Bank is required for the formation and
to a fine. capitalisation of a captive insurer and for the remittance of the establishment capital. It is also
required that, on an ongoing basis, the renewal programme of the captive, including the
Clearly, this provision only applies to a person who induces or attempts to induce; it does not
captive's reinsurance and expense requirements, be approved by the authorities from the
apply to the insured that decided to place its risk with a captive insurer. 47 The new Act has a
standpoint of
much broader provision in section 8(2)which prohibits persons from providing services as an
intermediary in relation to a short-term policy, subject to some exclusions.
1977 3SA513A.
estimated that between R20D COO million leaves the country as premiums to captrves. Unf0Jturotely. the did of the foreign exchange rulings couerlng
45 The dehnllion oft-lass' was inserted by sec. I (c) of Act No. 10 or 1965. insurance and are set in Commission {19m: paras. 7.1-7
46 Sec. 713}.
47 This is the i nte:pret3tion applied (o sec, 75 as denved an amendment from the wording by Parliament-of the section as corrct ly noted by Melamet Il (Melarnetcommia sec. 60(2) does not prevent the risk manager from Placing the risk the captive. which no is required
1990: 47 & 95}. Any broader meaning väll requlre 4.hi0i%iøn, 1990:47}.
309 310
RISK MANAGEMENT', RISK FINANCING
remittance of funds.53 In granting permission for the outflow of capital for the formation of the Captives have been found to be almost indispensable in the buifding of reserves for those covers
captive, the Reserve Bank will specify conditions governing the offshore captive. required by organisations if such covers are either unavailable or too expensive in the general
insurance markets.
Since the captive's liability would ordinarily be in the designation of the local currency, the
retained portion of the risk would ordinarily be held by the captive in a non-resident's bank
Captive are said to mitigate concerns over the financial strength or credit
account in South Africa. This gives rise to the tax issue discussed above.
Captives are stated to be useful jn allowing access to some legislated markets and creating entry
1 1 .9 Summary to markets that do not accept direct placements.
Captives are currently being used to underwrite the more traditional casualty
The salient features of this chapter can be summarised as follows:
• Since the late 1980s, the number of captives world-wide has increased at the rate of
approximately 4% per annum, with 7% growth having been experienced between 2003 Seventy-four per cent ofFTSE 100 companies make use of captives.
and 2006.
The leading captive insurance domiciles in 2006 were Bermuda, the Cayman Islands, Vermont, the
British Virgin Islands and Guernsey.
• Motives for captive formation were determined to be management and control issues,
led primarily by the need for organisations to produce effective
Decisions in recorded
risk financing provision while managing the total cost-of-risk, inclusive of both premiums as being primarily dnven by the need for smoothing of losses, reserving, premium
paid and payment of self-insured losses. pricing managemerL transfer of risk. retention variations and buydowns of exposures,
access to reinsurers, risk management, and investment
• Captives are said to be used to manage the effect on total cost-of-risk through successive
periods of increasing insurance rates in the general market,
Current work by IAS on so-called 'Solvency f!' could ultimately lead to more pressure
• Experience demonstrates that the use of a captive allows for the aggregation of risks from on solvency margins and 'capitalisation' issues, and increased costs could become a
across organisations and the ability to present a far larger, unified and diverse serious issue for businesses involved in multiple insurance
portfolio of risks to the market with the consequent more favourable market response.
Jn addition to the benefits to be derived from optimising the use ofcapital, additjonaf
Centratising the placing of insurance, together with the control of overall risk-financing benefits could also be obtained by a consolidated investment 312
strategies, is suggested as another of the key reasons why multinationals form captives.
The ability to more accurately track organisational loss trends is also identified as one of
the material benefits of consolidation of risk information through a captive risk-
financing vehicle or vehicles. Chapter 12 Insurance
53 Athough not a legal requirement, in practice Is also necessary to seek and obtain the approval of the financial Servtes form an offshore captive and to place
insurances with it. It was recommended by the Mdamet Cornmtssion that a committee Of representatives from the exchange control authorities, the Registrar
Of Financial Institutions aru:i the Cornmssioner for Inland formed to monitor offshore captives {1988; 80).
Introduction insurers' activities into areas of speculative or business risk, and we will therefore discuss the
Macro considerations of insurance treatment ofspeculative risk through the medium of insurance. A discussion of insurance pricing
12.3 Limits of insurability then follows. The chapter concludes with a brief overview of the various short-term insurance
12.4 Requirements of an insurable peril policies.
12.5 Financial treatment of speculative risk through insurance
12.6
Insurance pricing (rate-making) theory
12.7 Specific policies 12.2 Macro considerations of insurance
ff / had my way, would write the word 'Insure' over the door of every cottage and
upon the blotting book of every public man because I am convinced that, for sacrifices
12.1 Introduction which are inconceivably small, families be secured against catastrophes which
otherwise would smash them up.
Insurance is an important method of meeting the financial consequences of risk. It has been Winston Churchill
traditionally defined as the business of transferring event (insurable) risks by means of a two-
party contract. In a macroeconomic sense, the essential benefit accruing to society from the institution of
While the intention is not to dispute this definition, it is important to realise at the outset insurance is the reduction of the uncertainty of losses in the aggregate. In as much as the
that insurance provides a mechanism for the transfer of the cost of risk rather than the pooling effect, inherent in insurance, represents an efficient risk-financing or loss transfer
transfer of risk itself. This is a primary criterion for regarding insurance as a mechanism of risk
device, this mechanism reduces the impact of pure losses, aids economic progress and is thus
financing within the overall mou for risk management. The consequence of this is that there
can be a transfer of neither risk nor the responsibility for such risk. This fact places a greater a valuable contribution to the well- being of society.
emphasis on the need for proper and effective risk management, which in the
Insurance allows the prospective entrepreneur to more easily secure and commit risk
final analysis is the focus of this text.]
capital to the purchase of plant, equipment, and other factors of production and business
1 It should be noted although advanced with a of and facilities. Without insurance, the entrepreneur would be exposed to event risk in addition to
proposed inthe It the insurance function with
being oneof In
speculative risk. Anticipating the possibility of the destruction of his/her property, he/she
aggregate claim payments the premium
would about a reduction in the
the insurance suggest that. imply. the of independent
would necessarily reserve part of his/her funds for the purpose of financing such losses in
order remain in business. This accumulation of individual reserve funds could be seen as
the in
the risk in fact very - the be seen charged a would in the run pro.,e - with little or norisk to attached.cater
restricting output, and presumably also affecting other economic units, Guch as lending
INSURANCE 313
institutions, which would require a greater liquidity position.
Having made this distinction, it follows that the concern of this text is primarily with the to more The availability fully
commit of insurance, assets to on the the basic other operation hand, permits
analysis of the role that insurance plays in the model for risk management. Insurance as a of the the business, entrepreneurand
method for treating risk requires examination and evaluation. We recognise, however, that the
subject of insurance is a wide one, and would require a separate study if it is to be analysed in Premiums accumulated by the insurer would represent the reserve assets ofthe
detail. 314
The purpose ofthis chapter is to consider particular aspects of insurance in order to provide
an insight into insurance as a method of financing risk costs. As a starting point, we briefly many individuals or business units. In a sense, insurers could be regarded as a clearing house that
consider macro considerations of insurance, followed by the reasons why an insurer 'assumes alleviates the impact of event risk. Such action on a large scale suggests a more effcient utilisation
risk' and a discussion of the requirements of an insurable peril (i.e. what makes a peril insurable,
and allocation of resources.
as opposed to uninsurable). An interesting development has been the extension (limited) of
2 It follows. however, then that the if additional the presence bosses of insurance become invites 3 social moral cosK hazard of insurance. (in the literal Simiärly, sense)
It is reasonable to assume that the availability of insurance also permits a reduction in the
Of
it with be consequent argued that Ircreases if the existerce in severity of losses, the insu red loss do activities Ihrough the lure of prem iurn reduction, then iety
prices of consumer goods and services. Firstly, the expansion in output made possible through
benefits-
insurance should create economies of scale that should have a positive impact on prices. Secondly,
encourages to engage In preveauon
in the absence of insurance, the entrepreneur, when faced with greater risk, would presumably INSURANCE 315
require a greater return, recoverable through higher prices. The entrepreneur's loading for this
factor of risk would undoubtedly exceed the premium cost, since insurance economies of scale The statistical properties of the insurance pool have already been described in chapter 3.
It was stated that the insurer faces a risk, described as the variation within the pure premium
through pooling would not be possible.
distribution. Mathematically this is the standard deviation
Turning, finally, to the social costs of insurance, much of the premium payments represent
transfer payments among the insureds (with a zero gain in the long run) and, as such, are not a
12.1 where:
cost to society. The social cost of insurance is determined by the amount of productive factors
standard deviation of the pure premium distribution all possible pure
consumed in the process of providing the service, which relate mainly to insurers' general
premium values g mean pure premium
overhead costs and costs of operation.
N = number of pure premiums.
Losses themselves, although a loss to society, do not represent a social cost of insurance, as
they are not attributable to the existence of insurance. There is also the added risk resulting from the fact that insurers cannot insure the entire
population of exposures, but only a sample of it. The risk exists that the underwriter's sample
of exposures is not identical to the population of exposures from which the pure premium
was estimated.
12.3 Limits of insurability
Thus, if the insurer decides to charge g as its pure premium, it is faced with the risk that
Insurance, n. An ingenious modern game of chance in which the player is permitted to its sample of insureds will have an actual average pure premium, defined as X3, that differs
enjoy the comfortable conviction that he is beating the man who keeps the table. significantly from g. This element of risk to the insurer may be termed sampling error, and
Ambrose Bierce, The Devil's Dictionary (1881-1911) is a function of the size of the sample.
We therefore recognise that just as a pure premium distribution exists, so does a
The function of the insurer is to diversify the claims cost of individual insureds (exposures) over distribution ofsample means, each sample being composed of a number of exposures from the
underlying pure premium distribution.
a wide base of policyholders who are commonly exposed to the possibility of loss. To make this
The mean of the distribution of sample means (R) has the same value (U) as the
function feasible, from the insurerk perspective, several requirements need to be satisfied. The
underlying distribution from which the sample is selected. The central limit theorem justifies
question is why and how is the insurer able to accept the risks, (the costs of which) the insureds
approximating (when N is sufficiently large) the distribution of sample means (X) with a
are willing to transfer?
normal distribution whose mean is p and whose variance is: 3
The most relevant aspect is that the combination of many similar exposures
(the practice of pooling) brings into operation the which law then of large enhances numbers.0118
12.2
Through pooling, there is an increase in knowledge, ability to predict the range of probable losses Diagrammatically, the relationship can be depicted as shown in figure 12.1.
from a risk exposure.
limit states that the dist„butlon ofthe mean value ofa set of n independent and distributed random variables ng mean p and variance appc,aches a
distribution w,th mean u end va„ance as n tends towards infinity & Pedes, 1988: 299901).
Relation of pure premium distribution to
RISK MANAGEMENT'. RISK FINANCING INSURANCE
distribution of sample means
316 units increases, the probability that the average loss for the portfolio will exceed some critical value in excess
of the mean falls. As N approaches infinity, the ruin probability approaches zero (equations 12.1 and 12.3).
Pure premium distribution Thus, both risk measures, the standard deviation and the ruin probability, reveal that portfolio risk is
Distribution of sample means for
Figure 12.1 Relation of pure premium
a given sample size inversely related to the size or number of exposures in a portfolio: These properties therefore make
distribution to distribution of sample
means insurance a feasible riskfinancing means or mechanism.
There are also instances where insurers do not enjoythe statistical advantages derived from pooling,
$ losses
but where they nevertheless accept such unique risks. In these cases, it may be said that perceived
$ losses
differences in uncertainties exist surrounding the risk: the insurer's subjective valuation ofthe risk may be
'Ihus, there is a relationship between the standard deviation of the pure premium distributiona andthe standard lower than that of the insured.
deviation ofthe distribution ofthe sample means: It may be possible too that the insurer has greater knowledge of the outcome of the event, based on
factors other than the law of large numbers, or that the reward for assuming the risk is suffcient to overcome
opp 12.3
the uncertainty of the insurer. Naturally, the insurer's knowledge may indeed be inferior to that of the
insureds and there may be a moral hazard involved. Insurers may anticipate this situation, commonly referred
It will be seen that where the insurer has but one exposure or insured, i.e. when N I, the risk is equal to o and the to as adverse selection, and make adjustments to their premium calculations.6
insurer faces the same risk as the insured prior to transfer. only when the insurer combines more than one insured An insurer may also be willing to accept the risks of the insured on account of its greater financial
does the risk reduction process become evident, i.e. the advantage of pooling comes into play. Indeed, the insurer's capacity. Because of the decreasing marginal utility of money, the difference in the relative value ofrisk to
risk varies inversely 4 with the square root of the number of exposures insured (equation 12.3). the insurer and the insured may permit a mutually acceptable contract.
In sum, an insurance portfolio can be seen as a collection of insured Finally, an insurer may accept risk even when faced with concentrations of value in a particular
exposure units, each represented by a contract with an external party - and exposure, geographical location or line of insurance, because it may reduce its uncertainty by transferring
the insure/s interest lies in the aggregate claims to be paid under the part or all of the risk to another insurer. This process is termed reinsurance, and is a very important facet of
portfolio. The insurer's portfolio risk, as measured by the standard deviation the insurance business/
of the average loss, tends towards zero as the number of exposure units in
the portfolio tends towards infinity.
Furthermore, according to the central limit theorem, the distribution of the average loss tends towards a normal 12.4 Requirements of an insurable peril
distribution as the number of exposure units increases. 'Ibis latter tendency permits the employment of the normal
approximation method to estimate the probability (i.e. the probability of ruin) that the average loss for the portfolio In chapter 3, the distinction was drawn between the terms 'peril' and 'risk. Peril was defined as the source
will exceed some critical value, determined, for example, by the insureds reserves. As the number of exposure causing the financial loss, as opposed to risk,
be r
4 contributions of probability and sampling tbeo,y toward the of insurance well However. note
Si
of the many limitations when to situations such for t should be noted that these propertles relate to a portfdio offisbs 01 exposures that are but sirn•lar (homogeneous). broadly results are evident when the Insuuers portfolO as
composed of heterogeneous, but exposures. Risk tends to Olsappær as Number of policies becomes very large. but this is on (tk assumption that the relative co€nposition of
the portfolio is independent ci Its
that conditions in the the past. In the of insurable this seldom the When risks ere indepeMent - and therefore policies are correi&ted - the distobution of •voge loss does not tend t€mard ocwmal
n
umber policies Of polices tends tow&rds Increases Infinity. More Such tmpojtantJy. portföli:ys the are rrgkiness characterised of the by distrlbution a minimum of level the ofaverage risk that loss cannot
deviations In experience are not due to these theoretical concepts are used for does bediversified nol tend twvard away. zero The as risk theof
chaprei whN$i 2, section displays 2.8a hpgh degree of geographical rnterdependence. is an example of such a 'isk category
317
Pand IC•cess quota of reinsurance. shate) often theceding Impry that insurera relatively may bevewed small part as of themaiketing the risk is retairkd interrvædiaryto by the direct
thereinsurer Insurer. and Treatyreinsurarve therefoae the uttlmate arrangernentscarrier 'he Ek is in fact reinsurer. The pricing Interrelationship between insure and (Snsurer can be venj
cofflplex, but on the premise that the risk carrier is reinsuler. It is not surprising that chervs in market pgicing (the hard and soft market cycle) are led by reinsurance 1963:6-14).
J
of calculating first of the expected outcomes of future events. other adjustments. often of a nature
required
318 which was defined as the uncertainty surrounding the event and its possible financial outcome.
The reason for buying insurance is therefore to negate or reduce the financial consequences
arising from the uncertainty surrounding a given peril.
Perils that give rise to the risk of financial loss are, for example, fires, thefts and motor If the peril insured against produces a lossi this must be definite in time and quantum (size
collisions. Such perils create the types of risk that have been termed event (pure) risks, for or extent). requirement can be seen as a statistical and administrative one; in order to
which insurance provides a feasible risk-financing mechanism. Of course, not all perils that correctly adjust claims and accurately reflect liabilities, the insurer must be capable
give rise to pure risk situations are insurable. ofsubstantiating the extent of losses.
The requirements of an insurable peril (i.e. the characteristics that make a peril insurable) In practice, insurable perils meet this requirement only in varying degrees. The timing and
are governed by the statistical and administrative factors discussed earlier, which are
severity of a loss are often not easily determinable (e.g. in cases involving health and fidelity
associated with the practical operation of the insurance technique.
guarantee insurance). This indefiniteness ofloss in time and quantum introduces an element
The qualifications offered below are not to be seen as all that are possible, but are
presented rather with the aim of identifying the types of problems that must be considered in of moral hazard, which makes the underwriting of covers of this type more diffcult.
deciding upon the insurability of a particular peril.8 premium must be reasonable in relation to the potential financial loss. Ihis requirement has
implications for both the insurer and the insured.
There must be a large number of homogeneous exposures. In order to permit the
operation of the theory of probability, the requirement of large numbers is necessary. From an insurer's viewpoint, it must make economic sense to provide insurance; an economic
Homogeneous exposures are those that display similar expectations of loss — reliance premium-to-risk relationship implies that some minimum premium exists, which the insurer
can be placed on statistical probabilities only if the sample to which the probability is
requires to make insurance viable. Flhus, in the case of covers relating to individuals or
applied is similar to the one from which the probability was derived?
households, where the potential financial loss is small (relatively), the insurer would require an
The insured group (exposure units) must be independent and hence not exposed to a
aggregation of exposures, or simply stated, a suffciently large volume of business.
catastrophe whose consequences cannot be estimated, Statistically, the estimated
From an insured's perspective, it is also not sound financial management to insure against
probability of loss will only apply ifthe units that are exposed to loss are independent.
Exposures that are not independent, i.e. risks that are correlated, exhibit a minimum level perils that have a high probability of occurrence. The suggestion here is that where perils display
of risk that cannot be diversified away. Such risks are therefore difficult to underwrite. high probability of loss (and usually are accurately predictable), the insurance mechanism
represents a weak method of financing.
Examples of such risks include earthquake and unemployment.
What is entailed, in essence, is dollar swapping', and given the insurer's inherent costs of
The occurrence of the event insured against must be fortuitous. Strictly interpreted, this
means that the timing and severity of the loss should be entirely out of the control of the administration, these, added to the high frequency claim costs, result in a premium that
insured. Of course, this requirement is not met literally, since most perils are subject to approaches (and may even surpass) the value of the potential loss. This, in part, has prompted
some degree of control by the insured. Indeed, the fact that many losses are within the the growth of 'self-insurance' mechanisms as a means of providing for risks of this type. 12
control of the insured allows risk management and risk control activity to be carried out.
The implication is not that the advantages of pooling are not in evidence, but that
infrastructurally, insurers 'disqualify' themselves from providing a cost-effcient means of
financing. 13
g It recogn,sed that a peril may theoretically fail to meet one or of the requ,rements. that insurers m,ercome such
9 Reference should be made to section 12.3, which dlscusses the statistical properttes and results where lisps are heterogeneoj independent. and where the risks are
Whether regard tothe life insured rnsurance, Wii' the outcome but rather ofdeath when he/she is certain WI" and. die.indeed, known In advance by the insured. However,the
not independent. i.e. where they arecarelated.
risk ro beconsidered Ch&ter 2, section 24.
319
me questkn of risk retention is examined In chapter 10.
small insurers sornettrnesdo notappreciatethe cationale cost constraints offaedand are opponents as this category and critics of risk is concerned) and have developedtnSouth
a minority Insurers have recognised these (insofar insurance products.
The outcome ofthe event should not be known in advance by the insured and moral hazards 320
ll should not be the subject of insurance. Underwriters are reluctant to insure in situations
where there may be an incentive by the insured to exploit a loss, as is the case where amounts
of insurance in relation to, say, property values or income levels are excessive.
But further along the spectrum, there is evidence of insurers being asked to consider insuring
12.5 Financial treatment of speculative risk through
risks that are very different from the definition of pure risk with the aim of facilitating the operations
insurance of business units. An example of this is residual value insurance, which would guarantee some
minimum economic value attaching to an asset after a determined period of time.
As discussed in chapter 2, section 2.4, risk can be classified as pure or speculative. This
The answer to this question lies within two broad interpretations. Firstly, one should be
distinction is made because it is generally accepted that insurance applies to pure risk
cautious in defining economic or speculative risk. A wide range of economic risk exists where one
situations. 'Ihere are cases, however, where insurance is applied to speculative-type risks.
could, for example, put resultant economic loss from a pre-defined event (pure) risk at one end of
An example of such a case is the availability of credit guarantee insurance to indemnify a
this range and business risk proper (the speculative risk assumed because of the possibility of gain)
company for losses through default.
at
The following reasons are given for the non-insurability of speculative risk. Firstly,
limited adequate statistics are available to enable one to predict probability of loss. It The risk in question is often loosely defined. The point made earlier is that there is no
can also be argued that even if such statistics were available, events may not be regular economic reason for insuring the 'extreme' case of speculative risk. Put differently, if one added
enough to permit accurate prediction. further types of risk to economic risk — e.g. end risk and intermediate risk — then the extreme
Secondly, the probability of loss may be too high to make insurance a practical case ofevent (pure) business risk is an end economic risk for which risk elimination or
mechanism, given an inherent administration cost and required return for the risk diversification through insurance becomes futile or superfluous. 14
assumed. While this argument applies to speculative risk, it does also often apply to The second interpretation stems in fact from intermediate economic risk. One could define this
event risk. As discussed earlier, where the probability of loss is high (even in event risk as a series of risks that in the aggregate constitute or result in end economic risk. Such intermediate
situations), economic or cost constraints may make insurance unsuitable when compared risks are de facto speculative in nature, but do not represent the de jure economic risk taken by a
to other forms of financing risk, e.g. self-insurance mechanisms. company in its attempt to make profits.
nuirdly, the most important reason for not insuring speculative risk is that the process The choice of the word 'intermediate' in this context is deliberate. 'Ihe modern interpretation
would be self-defeating. The assumption of risk by individuals is based mostly on the of the business of an insurer is that it acts as a financial intermediary: in exchange for a pre-paid
expectation of gain; to insure against the possibility of loss would substantially reduce fee (the premium), the cost offinancing reinvestment is transferred to a financial intermediary (the
or eliminate the extent of gain, since the premium required would in theory be almost insurer). 15
the same as the expected return. Paradoxically, the insuring of speculative risk, e.g. With such a broad-based financial approach to the definition of insurance, it can be argued
business risk, would in this instance represent real transfer of risk from the individual that the class of intermediate economic risk may be added to the class of risks for which insurance
or business unit to the insurer. is a feasible risk elimination or risk reduction technique (given, of course, the required criteria for
the insurer to assume risk in the first place). Hence, this viewpoint suggests and encourages a
Thus, from an economic standpoint, insurance of speculative (business) risk is questionable. greater participation by insurers in the wider range of risks facing the business
The substantial premium required would lead to uneconomic output and the reduction of
profit. Also, it may be argued that the transfer of such risk is superfluous* for, ifit were the Given the necessary information, knowledge and spread of risk, insurers could consider
intention to transfer risk to the insurer, then this could otherwise be achieved by the insurer insuring (or part-insuring) intennediate economic risk (which would have the effect of reducing
buying the business unit's shares or stock. end economic risk to some extent) without
I-nis being the case with speculative risk, it still may be asked why, as
insurance contracts become more sophisticated, more and more economictype risks are
concepts or end and intermediate from attempt to define economic that be and the setting of fundamental upon which request for such Insurance may be based 'n
being insured. the final the that one js try,ng to The following; if the risk intermediate ,isk. rhen insurance is the criteria for risk through etc. that apply to Jfthe risk identified as
economic risk. exists for risk The would effect be by virtue of its
It is accepted, for example, that when insuring the value of an asset, n ot only the
322
material value is considered, but also the economic value attach ed to the asset. Loss-of-
profits insurance resulting from a pure risk event is commonplace.
321 going against the basic economic reason for a business accepting business risk. So long as the
economic risk is incidental, there may be opportunity for insurers to intermediate and therefore to
offer businesses another way of managing risk.t6
operation of the law of large numbers. Given the warnings mentioned earlier, it is possible to
12.6 Insurance pricing (rate-making) theory
estimate within reasonable accuracy the ratio of losses to total exposures.
A problem with premium rates is, of course, that they apply to a period of time subsequent
Scope of the discussion to the time they were calculated. Insofar as conditions of the past (reflected in the loss
This section discusses the theory behind the rate-making process - a 'rate' being defined as the experience and used to predict future losses) differ from those expected in the future,
price an insurer charges for each unit of risk accepted (or, from the standpoint ofthe insured, the adjustments to the statistics and hence to the rate must be made. nese adjustments will
price per unit of protection provided). Particular attention is given to: • the nature and principles necessarily introduce an element of subjectivity to an otherwise objective basis of predictability.
of rate-making problems in deriving rates general methods of calculating premium rates. But, as Hurley19 states: 'This is no denial of the value of statistics It is solely a reminder that
improper rates can result from statistics -- when the statistics are an imperfect representation of
future loss expectancies.' In addition to the determination of risk categories and the prediction
Nature and principles of rate-making of claims, the prediction and allocation of expenses are a constituent of rate-making.
In most lines of insurance, the expense component is smaller than the component used to
cover the cost of losses. Claims predictions of many insurer organisations are similar for
several lines of insurance or risk, and the competitiveness, therefore, of a particular rate most
The premium that an insured pays for coverage is derived by multiplying the I appropriate rate
by the number of homogeneous exposure units (units of risk). often depends on the element of costs and margin for surplus and profit. 20 One should also
note that expenses are often a proportionate addition to the pure premium, as many expense
Hence, the premium reflects both the likelihood of occurrence of loss and the size of the potential items to some extent vary directly with the size of the basic premium. There is naturally the
loss. Viewed in this way, it follows that a lower premium will result from both a reduction in fLxed cost element, which is not a function of the pure premium.
exposure value, and, say, from a superior quality of the risk insured. As concerns the margin element, this includes a contingency allowance and a provision
for profit (or for surplus in the case of the insurance mutual). The allowance is designed to
From an insurer's point of view, the rate-making function is seen as the process ofpredicting
provide funds to offset adverse claims experience
claims and expenses following the ofrisk, and then apportioning the estimated claims and
i7
expenses, including the required profit margins, to the various classes of policyholders. In
perception of thisaspect of loss predictabdlty may range from intrigue to disbelief. and the operations Ofinsure's have often been {o gambling. so it is
practice, however, this process is not so precise and refined as it first appears to be. necessuy to draw attentroa to the differences between insurance and gambling. 80th transactions payment ofa relativety small amount in exchangefor the return
of a much gseater arnount upon the occurrence ofa chanceevent With insurance. shoutd the event occur. the recovery under the contract does not (purport to)
This apportionment applies across various classes of policyholders, and. therefore, before enhance the participantS ons-lied) uealth positron. Gamblng. on the Other hand, speculative risk where the porticipynt has chance 93in or loss. Moreover. the risk
•s tm-existent untlt the gambler creates it An Insured though, IS exposed to rhe vis.k whether he/she chr»ses to insure or not; the 'nsurarxe tr*M.action reduces
calculating a rate, it is necessary to select and group the types the possible financial Impact tesulting from the utKertainty.
(1962: 188)
16 broadening lines stated here is a new and important for the future. presented of
"'Bt•ent required settling that to costs be an in insureT When the business would one considers need of insurance.to totaltypicallycosts consida in terms are Ot rrurketing the classic (acqulsition)
has been developed where the methods economic definition. costs,officethen adminrstrative included in these expenses IS the (tnf'&stsuctural}norrnal profit
period.
323
or classes of risks for which the rate will apply. Rate-making classifications are required both
in predicting claims and in allocating costs. To the extent possible, this process of risk
classification attempts to identify homogeneous risks where loss characteristics are similar,
which then makes possible the allocation of costs (including profit) to the risk as a function of
its profile and size.
Earlier, under the section that examined the circumstances under which an insurer will
accept risks, we reviewed the question of loss predictability. Obviously, this is the most
significant factor in the rate determination process.18 The procedure relies principally upon the
324 'NSURANCE 325
of fierce price competition (when underwriting principles are often ignored) to a position of
by the insurer. If these claims do not materialise, then these allowances are added to profits
'extreme' price regulation. When one uses the definition of an insurer as being a financial
or surpluses. Although the profit factor can be considered and accounted for in various ways,
intermediary facilitating transfer from supply to demand, this problem then becomes more
insurance companies tend to use realistic assumptions with respect to claims and expenses
apparent. One can only conclude that the price instability denotes an intermediary mechanism
and then add an additional margin to provide their profit.21
that at present is imperfect and undeveloped. It is interesting that price instability has been a
factor encouraging self-retention of risk, which is evidence of the market's inability as a
mechanism to effciently facilitate transfer.
Problems in rate derivation
In chapter 2 we discussed economic considerations for treating risk As such, the reasons
It may be said that insurance pricing involves problems not ordinarily encountered in the
and requirements necessary for a market to form were given. We also examined the possible
setting of prices for other services and products. As has been shown, the price setting in the
sources of market failure and directed particular attention to the question of adverse selection
insurance market involves a range of subjective and anticipatory aspects. Apart from these
resulting from asymmetrical information between supply and demand. 2 With respect to the
problems associated with the statistical and anticipatory nature of insurance, a few specific
question of rate derivation, asymmetrical information introduces the problem of moral hazard,
issues need to be examined, and these are now discussed.
which we will now examine.
The nature of insurance and particularly the social responsibility attached to insurers
providing protection to society in general has encouraged regulation to some extent, or at
Definition
least public inspection, which often requires detailed justification of the internal
Moral hazard is the behavioural effect that insurance might have on the level of effort or
composition of a rate structure. This is difficult in most businesses, even with developed cost
expenditure devoted to loss-reducing activitiesp
accounting techniques, and in insurance the problem is more pronounced.
Notwithstanding this diffculty, in a sense, it is necessary to reconcile, on the one hand, Given that the value of individual and corporate wealth (property value) is reduced by risky
the fiduciary responsibility of insurers (implying an adequacy in pricing/contingency events and by claims from employees, third parties, defective products, and the like any money
provision) with, on the other hand, reasonableness and equity (as relates to the fair spent on reducing the probability or severity of loss will benefit the owner and increase value.
treatment of individual insureds). There is thus an incentive to protect ownership rights and to undertake forms of protection that
A trade-offrepresents a problem in its own right, but it becomes exaggerated when yield
applied to the insurance market. While the strength of the underlying statistics of large
insurers usually makes their loss predictions very accurate, in many cases the experience of An insurance policy, however, which transfers the cost of losses from the owner to the
one company is not extensive enough to allow an accurate prediction to be made. As a insurer, reduces the effect of investing in loss reduction activities and removes the incentive for
result, one finds a form of price cooperation being practised, and even cartelisation, the owner to try and prevent losses. In order to restore such an incentive, the insurer may
especially in a hard market cycle when insurers attempt to recover from losses incurred and modify the pricing structure or the conditions ofcover. Hence, rate reductions may be allowed
restore their solvency margins. Ihis can be seen as being contrary to economic philosophy for the installation of preventive devices, and premiums may be increased if the chances of
and even against the law. losses
It is, of course, diffcult to measure, let alone ensure, acceptable levels Of
reasonableness and equity. The nature of the problem is apparent by the large fluctuations The problem, however, arises from the inequality in the information on reduction available
in insurance pricing and availability of capacity as the market moves through its cycle. This to the insured and insurer. Although the insurer may be able to distinguish some form of loss-
implies that the market swings from a position reducing activity in the insured's
21 Ore other significant element accounted for the rate is the Investment income element. Although, stricty it nay be (hat investrænt income accruing to
insurers fall outside the factors that determ•ræ rates, since such factors should retlæt underwriting consideratms fac•ng the insurer, in finds that the of Insureds a to they the insureds of and premiums may mean that insurance offe,ed at premium that cannot profits to underwriters.
investment climate plays an important role. Investc underwnring has been identified as having a contrabutlng Influence on insurance prici ng cycles.The point in the Information that differentia' the no buy a
ofview taken •n this sectm the nature and principles of rate-sealag Is a punst one, which divorces the influence of investment income from the underwritlng
insurers. However, che reader Is referreo toValsamakiS 0983: 6—8J 120—65}.
of poorer to share the costs of benefit payr-nenti seechapter 2. sec-tons 2.7 and 28,
The derivation of class rates involves all of the problems associated with selection of
the be pointed out that categorised hazards•are neithe, unethical that often risk categories and the prediction of losses and expenses; judgement too plays a role in class
It confuse illegal.the of the
rate calculations. As few risks fall into distinct categories or classes judgement is required to
326
set class boundaries and ensure homogeneity of risk.
operations, it may not be possible to observe everyday housekeeping and work practices that may Individual or experience risk rating
give rise to the perils that are insured against. These unobserved practices cannot be controlled or With the growing concern by organisations over the management of risk and the expenditure
rewarded by pricing adjustments or contract conditions. Consequently, insurers may not be able to on insurance, the class rating method has come under criticism, since it is seen to be
charge the correct rate or premium. discriminatory. Policyholders who have better than average loss experience argue that the
From the insured's or risk manager's point ofview, moral hazard is simply rational economic broad classes should be restructured into more homogeneous groups that further distinguish
behaviour. The rational risk manager would take account of any premium incentive in designing a better' risks from others.
risk control programme, since the premium saving offered to the insured may represent a major
part of his/her expected benefit? Definition
This argument supports the view that premiums, sensitive to changes in the expected value of Individual risk rating is based on the principle that the number of exposure units of some
loss, will provide the insured with information on the effectiveness of different forms of loss insureds is sufficiently large to conclude that tosses that are (or are not) incurred reflect
reduction. To the extent that asymmetry of information remains a feature in the market, then the to some degree the actual characteristics of the risk of the particular insured.
efficient dissemination of loss prevention information will be frustrated?
Hence, individual risk rates reflect more directly the actual loss experience of particular
policyholders. The nature of this approach has prompted the usage of the term Sexperience
General methods of calculating rates rating'.
Insurers may take two fundamental approaches to setting rates and premiums. Paradoxically, the The application ofindividual risk rating raises a broad question concerning the function
two approaches reflect conflicting philosophies. The two approaches are class rate-making and of insurance. In the extreme, every insured will pay his/her own losses and the pooling
individual risk rating. The latter approach takes more account of the individual's loss experience function of insurance is destroyed. It can be assumed, on the other hand, that some form of
and risk characteristics when it is used to set premium rates. pooling is involved, because those losses that exceed expectations (according to experience
rated assessments) are subsidised by those with fewer losses than anticipated.
Class rates
It follows that as predictability of loss becomes more certain, so the pooling effect is
minimised. Under these conditions, insurance becomes a non-efficient method of treating the
financial consequences of risk, which is a point made earlier in the chapter,
A class rate is obtained by apportioning the anticipated losses and expenses for a given class of From the point of view of equity, the technique is seen as appropriate, although some
peril or coverage across alt the exposure units.
argue that the general result of individual rate-making has probably been to increase class
rates.26 It is felt that insurers are not entirely happy with having to adopt this approach, but
This procedure incorporates, naturally, the pooling or averaging principle, which is the foundation
do so as a result of competitive necessity. All observers recognise, nevertheless, that
of insurance. It implies that the losses of the less fortunate are subsidised by the premium
individual risk-rating will continue to grow in use.
payments of the more fortunate. Hence, losses of a particular policyholder affect the rate only to
the extent that such losses generate a change in the class rate.
2EQenberg et al, (1974: S 1 9).

328
24 is that should to
Rather, the rat.onal risk manage, should account ofthe economic consequences of eoendlture on any control gogramme. 25 As points of market lend to view loss-
prevention Information. Competition tends to lead to premium that discriminate among expectancies. Insurers often pool their to and insuredsdatabases. Insurers.
Three methods ofindividual risk-rating be considered, namely:
and typically to
risks in order to provide Info'mation to underwriters (i) experience rating
327
(ii) retrospective rating (retro-rating) (iii) schedule rating. premium, however, does ensure that the insured's payments will not exceed a given amount
even in the light of very high losses.
Experience rating In essence, the technique provides cover only above the maximum premium, with the
insurer providing a claims service below that level, the cost of which is catered for in the
minimum premium paid by the insured.30
Experience rating utilises the past loss experience of a particular insured to determine, at least
in part, the rate to be paid in the future?8 Schedule rating
expenence rates are prospective, where the basic formula used is12.4 Definition
Schedule rating uses a schedule that lists supposedly average physical characteristics for
a given type of policyholder, and each insured is compared with the schedule for rating
where: purposes.
A = the actual losses losses expected to be
incurred. This comparison then determines the extent to which the appropriate class rate is modified.
The technique requires the actual inspection of the property being insured and tends
The experience rate derived as shown above would then be applied to the basic formula used. to place emphasis on the physical characteristics, although some non-physical aspects (e.g.
For example, if a firm incurred actual losses of R50 000 in a given period, when it was safety programmes) are also considered. The process does, however, rely on judgement,
expected to incur losses of RIOO 000, then the rate for the following period would be reduced particularly when a physical feature of the property varies from the average.
by 50%. To the resultant rate must be added the expense element and consideration must be Schedule rating was first developed for fire insurance, and is still used today in this field
given to the size of risk. Thus, the rate of an insured may not be determined entirely by his/her of insurance. Hurley3i rationalises its use as a technique on the following basis:
past experience.
It is generally accepted that over-alt rate levels and classification relativities must
Retro-rating take into consideration the actual loss statistics. However, there are many
students of the business who believe that final individual risk rates on an important
segment of the fire insurance business are most equitably distributed by the
schedule rating approach .... It may well be that fire loss expectancy for other than
A retrospective rate utilises the experience of the past period to determine the rate for that trivial losses may be of such a low order Of magnitude that a fire rating system
period? based solely on loss statistics may prove not feasible within the normal tolerances
for credibility standards.
retro-rating requires the payment of a deposit (often minimum) premium at the
inception of the policy protection period, and then the policyholder's experience for the
been very concerned the use of retro•plans. slnce the technique largely dismisses (heir rnatoc function of Cisk handl@ on insurer an intermediary
period determines whether he/she receives mechanism for finanonq risk this argument in a different Irght. in any event. the increased
•eQf seif-insuance has forced insurers to offer retro•satedcontracts.
187).
330
27 It be noted that generic definition However. of does not ixiudeschedule rating. rating. most
e»racteristics for given type of is listed under ttE form of individual rating. 28 Oenenberg al.
(1974; 520).
29 Derenberg et al.0974:520).
INSURANCE 329
12.7 Specific policies
a partial refund or pays an additional amount. There is also a maximum premium provision. Insurance and the risk manager
This method is the closest possible approach to self-rating, where the insurance would Insurance remains the single most important method of financing the consequences of event (i.e.
be merely a service contract that provided no real protection against the risk. The maximum pure) risk, and the risk manager needs to have a firm grasp of a number of insurance issues.
In particular, he/she needs to understand the law of contracts and how this law applies to A general average loss is a toss caused by or directly consequential on a general
the insurance contract, the law of insurance,32 specific insurance contracts, insurance practice, average act. It includes a general average expenditure as well as a general
insurance regulation," what cover is provided by the various statutory funds, 34 and the average sacrifice.
characteristics of various insurance markets.35 There is a general average act where any extraordinary sacrifice or expenditure
is voluntarily and reasonably made or incurred in time of peril for the purpose
It is not intended to deal with all of these here, but only to review the more common types
of preserving the property imperitled in the common adventure.
of insurance policies available.
Where there is a generat average loss, the party on whom it falls is entitled,
subject to the conditions imposed by maritime law, to a rateable contribution
from the other parties interested, and such contribution is called a general
Marine insurance average contribution.
Subject to any express provision in the policy, where the assured has incurred a
lhe oldest class of insurance is marine insurance.36 people involved in a marine venture are
generat average expenditure, he may recover from the insurer in respect of the
exposed to what is referred to as the 'perils of the sea',37 perils on the sea and extraneous risk.
proportion of the loss which falls upon him; and, in the case of a general average
These perils cause loss or damage. Ships are lost at sea; cargo is lost at sea. The cost of
sacrifice, he may recover from the insurer in respect of the whole loss without
transporting the cargo (freight) is paid, but since the cargo is lost, the expenditure is lost. Lives having enforced his right of contribution from the other parties liable to
are lost at sea. Liability claims can be instituted as a consequence of a marine venture. contribute.
The marine policy is in the first instance an assets policy. Subject to any express provision in the policy, where the assured has paid, or is
The most famous form of marine policy was the so-called Lloyds SG policy,38 which was liable to pay, a general average contribution in respect ofthe subject insured, he
phased out in 1982 with the introduction of the Lloyd's Marine Policy (the MAR Form) and the may recover therefor from the insurer.
Companies Marine Policy of the Institute of London Underwriters. It is, however, the Institute
In the absence of express stipulation, the insurer is not liable for any general
Clauses that are important. The first of these, the '1888 Time Clauses', were accepted by the average loss or contribution where the loss was not incurred for the purpose of
Institute of London Underwriters in 1888 and were subsequently issued annually. Each avoiding, or in connection with the avoidance of, a peril insured against.
complete set of Institute Clauses to be used with the Lloyd's policy constitutes a self-contained Where ship, freight, and cargo, or any two of those interests, are owned by the
insurance cover. same assured, the liability of the insurer in respect of general average losses or
contributions is to be determined as if those subjects were owned by different
UK Marine Insurance Law was codified at the turn of the century in terms of the Marine
persons.
Insurance Act (1906). This Act is very important, since
People engaged in a marine venture are exposed to legal liability claims. However, the usual
32 The leading includes south the textbooks Act, on the the Medical ofthe Schenes insurance Act contract and the Cooperatives are marine cover provided by the Institute Time Clauses (Hulls) (1 October 1983) does not provide
Getz (1993) Actand Lawn (2002) 33 Regulation
full protection against liability claims, but only limited cover in terms of the so-called 3/4
34 compensation, third-party motor insurance and insolvency legislation.
Collision Liability Clause.39
36 insurers produce easy-to-lead publications the of insurance. see, example. Mellert {1989) an
37 of theses' lacks precise definition, but has been in use for centuries. It has beentraced back to the
of the referred to terms of section 30 of the of Is annexed to the asthe iomewnters refer to thisas the unnitg•down clause (ROC)
absurdform and of incoherent document: to Nevertheless, the Stamp the of document was in use fCK 200 years.to and as 332
331
The liability of the insurer is limited to three-quarters of the amount paid by the insured as
essentially it is a codification of not only marine insurance law, but insurance law in general.
a consequence of liability claims. The obligation of the insurer is limited to three-quarters of
An unusual risk to which the person who uses marine transport is exposed is the risk of
the insured value of the insured vessels and the clause does not cover the costs of the
'general average' being declared. General average loss is described as follows in the Marine
insured for the removal of wrecks, loss of life, personal injury or pollution. It is clear that
Insurance Act:
very limited marine liability cover is provided.
Since the marine policy provides only limited liability cover, a need exists to provide a
more comprehensive protection. In the marine market this need is filled by the P&l Clubs'
Aviation insurance After centuries ofexperience, the modern fire policy has evolved to become very
This is a modern addition to the insurance industry. However, since both the marine and sophisticated and is designed to cater for a wide range of costs associated with fire and natural
aviation markets concern forms of transportation, the tendency in modern theory is to deal perils.
with marine and aviation together.

Aviation gives rise to a number of risks: loss or damage to the aircraft (hull); loss or
Accident market
damage to cargo; the cost of search for missing aircraft; extensive liabilities attached to
The development of rail transport and factories led to the desire to be compensated for an
people on the ground; and liability to crew and passengers. accident. This gave rise to a number of accident policies: railway accident, employers' liability
Some forms ofaviation liability are governed by legislation and international treaties. (which initially was not actually liability based, but accident based} and gave rise to the workmen's
compensation movement and employee benefits) and personal accident policies. It was
customary for large insurance companies to have accident departments.
Fire insurance
The next comprehensive short-term insurance market to develop was the fire market. This Example
policy provides indemnification against loss of or damage to assets from a specific peril, The following is a list of covers listed by the Liverpool & London & Globe in
namely fire. A fire can be caused by a number of factors. An earthquake can be preceded 1936:
Accident (Personal) and Disease *All Risks'
(sometimes by a mere few seconds) or followed by fires. Other perils, such as explosions, Aviation (Personal Accident) Boarding Housekeepers' Liability
can precede the fire or be caused by the fire. Boiler Burglary, Private House and Business
In such cases, confusion arises as to the cause of or the proximate cause of the loss Premises
Cash in Transit Contingency
incurred: fire, explosion or earthquake. The tendency was either to exclude fire as a result
Combined Domestic Cycle
of the other natural perils or to extend the fire policy to provide indemnification against a Commercial Motor Vehicles Domestic Servants
variety of natural perils. Court and Govern ment Bonds Electrical Plant
The fire-fighting process could entail a variety of expenses, as the fire department may Defects ofTitle Fidelity Guarantee
Driving Accidents Goods in Transit
charge for its services. After the fire, indirect costs such as the employment of accountants
Farmers' Public Liability Gun Risks
to quantify the cost of the loss can be incurred* In order to rebuild the damaged asset, the Gas Engines Hold-up
Golfers' indemnity Lifts, Hoists and Cranes
service of architects may be required If a factory has to be rebuilt, the municipal regulations
Householders' Comprehensive Lost Document Indemnity
may have changed and Innkeepers'Liability Missing Beneficiaries
Live Stock (Horses and Cattle) Money, Insurance of Motor Traders' Road and
Garage Risks
40 FO« a diSCuSSlDtl Of the P&l Clubs. consult Dillon and Van Niekerk (1983) Machinery Personal Leabjlity
INSURANCE 333
Missing Documents Plate Glass
334
increased costs may be incurred to meet the higher standards. It may be desirable not to replace
the building in exactly the same form as it was before the fire, but to make use of the rebuilding Motor Cycles Property Owners' Indemnity
process to carry out capital improvements. It may also take some time for the building to be Oil Cabinets Redeemable Securities
rebuilt, by which time inflation has reduced the value of the money. It may also be necessary to Personal Accident Schools' Public Liability
Petrol Pump Installations Sickness and Accident
make provision for infraction in the policy.
Private Cars Steam Engines
Public Liability Third Party (Schools)
Refrigerating Plant Tractor
Securities in Transit Time Loss (Power Plant Breakdown)
Liability policies
Sports Club Indemnity Travellers' Baggage
Liability developed as part of the accident market, but is today regarded as a separate form of
Third Party (Farmers) Traders* Collective
insurance. In addition to the general liability, a number of specific forms of liability cover exist. As
Tourists' and Traveliers' Wireless
Personal Accident Workmen's Compensation noted, the marine hull cover provides a measure of liability cover. Section B of the motor policy
Ihe modern motors liability and consequential loss policies developed from the accident provides liability cover and the aviation policy also provides liability cover.
markets that were created at the turn of the century. The purpose of the liability policy is to indemnify the insured against amounts that he/she
It is still usual today for companies to retain the 'accident' structure in reporting and becomes legally liable to pay, usually (except in professional liability policies) arising from
controlling results. To a degree, it is necessary because of reinsurance treaties. accidental loss or damage, or injury to people. The cover is essentially in line with the damage to
property or injury to people. Liability for pure financial loss is not covered, nor are claims for
violation of personal rights such as defamation (unless through the arrangement ofspecial
Motor policies and markets extensions).
Because of the size of business in the motor class, modern practice tends to treat motor policies
as a separate class.
The operators of motor vehicles are exposed to a number of risks. Firstly, they are
exposed to the loss or damage of the motor vehicle itself. "lhis is normally covered in terms
of section A of the motor policy, which is essentially an assets policy. Certain assets associated
with the vehicle or risks cannot be covered because of their nature. Damage to tyres and
springs not associated with an accident is not covered, nor is wear and tear on the vehicle.
Generally consequential losses are not covered, except possibly the cost of towing and
transportation of the damaged vehicle.
The operators are exposed to legal liability claims and provision to indemnify the insured
and other people is catered for in terms of section B of the policy. Finally, people may be
injured and limited provision is made to pay for medical costs associated with the motor
accident in terms of section C of the policy.
Consequential loss market

It has always been accepted that when people are injured or assets are damaged, consequential
losses of a financial nature arise. Thus, an injured person who 335
is unable to work loses his/her income. A building destroyed by fire cannot produce the rent income
for the landlord. Without a specific extension, these losses are not covered by an asset policy such
as the fire policy. Consequential loss policies were developed during the last century to cater for
losses that arise as a consequence of loss-producing events.
Consequential losses can be treated as part of the market that deals with the major perils. Thus,
consequential losses suffered as a result of fire can be treated as a part of the fire market, while
consequential losses suffered as a result of machinery breakdown form part of the engineering
market.
INSURANCE
336 FINITE 337
insurer, for example. l Nevertheless, it is opportune to examine this type of insurance and its rationale.
Chapter 13 Definition
Finite insurance Finite risk insurance is an alternative risk-financing technique where insurance and self-funding
are combined with emphasis on the time value of money.
Some people even regard finite insurance as a new approach to insurance where people do not
rely on the law of large numbers, but rather seek risk diversification by spreading losses over
time.
Finite insurance is a relatively new term for a variation offinancial insurance aimed at
offering to corporate insurance buyers a method of risk financing that combines the 'off-balance
sheet' advantages of conventional insurance with a minimum transfer of insurance premium.
Although finite insurance transactions have increased in popularity and have been used
13.1 Finite risk insurance
Problems with finite reinsurance
extensively by many insurance companies, use has been mainly restricted to the insurance
13.2
13.3 industry to help stabilise earnings by spreading the peaks and troughs of annual losses over a
Characteristics of finite insurance Functions of finite
13.4 insurance much longer period of time? While the traditional method of laying off exposures has been to
Why has finite reinsurance evolved? use the normal reinsurance market, the finite reinsurance product enables the exposure not to
13.6 Major types of finite cover be transferred, but deferred and spread over time, thereby minimising the loss of income if the
13.7 Summary anticipated or potential losses do not arise.
The use of finite insurance and reinsurance by the insurance industry is now recognised as
a possible mechanism for major corporate insurance buyers faced with uninsurable exposures
such as pollution; or where the pricing is high, capacity is limited, or the relationship between
13.1 Finite risk insurance
premium and the cover provided Cthe rate on line') seems high; and, lastly where, given an
effcient mechanism, self-insurance might be a more cost-effective approach.3
What is finite risk insurance?
Many corporate insurance buyers have addressed these insurance problems by
Total self-insurance, as explained in chapters 10 and Il, is not usual. More often, several
establishing their own captive insurance companies. However, in most cases, the need for high
methods of funding risk will be used at the same time, one of which is known as finite risk
cover at relatively low premium income has created difficulties due to the need for substantial
insurance (also referred to as financial insurance or financial reinsurance). This financing
capitalisation, questioning by tax authorities on the deductibility of premiums and the fact that
method has traditionally been seen as an alternative risk transfer (ART) technique. It must be
any loss within the captive (if consolidated with its parent) may not achieve the main objective
noted, however, that in recent years, ART and its role in the insurance sector has come under
Ofspreading the exposure over a longer time frame than an annual accounting period.
considerable scrutiny by regulators and other organisations, including accountants'
organisations. There have been significant changes in the way these transactions are viewed;
principally, they are gauged against the underlying principles of insurance to determine
whether indeed they constitute genuine insurance contracts where there is suffcient evidence F
inite reinsurance has especially tecentiy come undea considerable gruttny Insurance executives have faced charges relating to shorn aimed 3t avording detecrlon by auditors
and 'egulators.• see Green (2009).
of risk transfer to the
FINITE INSURANCE
2
insurers insurance or providing programmes some are cover also (cash ernerging flow provision) withjn mega for d16cuit corporate jnsurance composite risks-
poxyamrnes. These are providing further capacity to the teöder is referted to the sectlon in chapter i covering accountir-A tax and jegu[3t01yconsiderations.
a maximum gain of R30 million — although the range of results typically moves closer together
338 when allowance is made for investment income.
FINITE 339
The possibility of purchasing an insurance cover that provides the spreading mechanism,
The pricing of a finite (re)insurance contract is a combination of the assessment of the
enabling the risk to be handled 'offthe balance sheet' and achieving this at a fraction of the
underlying risk (its frequency and severity) and financing decisions (future probable interest
conventional insurance cost (if losses do not occur) is clearly an attraction and has resulted in
rates and the return required on capital).
many companies examining the finite risk insurance concept.
Such a contract has generated additional capacity, often while conventional capacity is
The essential idea of finite insurance is to eliminate volatility in earnings. The products are
not available R250 million (but only R250 million once in aggregate over five years, not five
usually composed of three characteristics:
annual limits of R250 million each).
The contract has an aggregate limit of cover (hence the term 'finite'), and may also
The wording ofcontracts is often more flexible, because the finite reinsurer will typically
contain annual sub-limits to prevent the full limit being drawn down too quickly.
write the contract 100% and retain the risk 100%, so that there are no co-insurers and
The policies are long term (typically five or ten years), with the total price based on a
reinsurance contract conditions to be accommodated.4 Financially, the contract has smoothed
combination of claim payments, instalment premiums and the investment income
cash flows; e.g. iftwo R75 million losses are incurred, one each in the second and third year of
earned on the 'fund' over the policy period.
the contract, they will in effect be repaid out of the premiums being paid in years one to five
The typical annual finite (re)insurance premiums are higher than conventional insurance
(i.e. preand post-loss funding). The insured's reported earnings will also be smoothed - instead
(where comparable cover is available), but to offset this, there is usually a commutation
of losses impacting the results of whichever year they fall in, profits will be reduced each year
option offering a substantial return of premium to the insured if loss experience under
by the same R50 million premium, which ideally will
the contract is low.
As such, finite (re)insurance can be viewed as cover priced at a high rate on line, but with a
significant profit commission provision. For instance, a finite (re)insurance policy may offer a
13.2 Problems with finite reinsurance
five-year aggregate limit ofR250 million (with a sub-limit of R75 million in any one year) for
five annual premiums of R30 million each and with an 80% return of the experience fund at There are concerns over whether this technique offers true risk transfer, and so is a contract
the end of the contract (the experience fund is formed by the build-up of premiums paid in, ofinsurance, or isjust a banking arrangement. Ifit is not insurance, then because of the
together with the accrual of investment income, less any claims payments made under the necessary accounting treatment of such contracts, they
contract).
This is a complex issue, but essentially it turns on two key points that relate to the definition of
Assume that cover for the comparable R50 million layer is not available on a conventional
insurance risk. For insurance to exist, there should be:
basis. Additional capacity has been generated by such a finite (re)insurance contract; in the
first year's RYO million of cover, in fact only R75 million is actually accessible, for a cost of R30 • uncertainty with regard to the amount of a loss, i.e. 'underwriting risk'
million. However, if claims are made, the substantial profit commissions available at the end • uncertainty with regard to the timing of a loss payment, i.e. 'timing risk'5
Of the contract are impaired, so that there are elements of pre-loss and post-loss funding in
However, major companies in the finite insurance marketplace are prepared to develop
such contracts. If we ignore investment income (purely for the sake of simplicity), the
structures (integrated risk and insurance strategies — IRIS) that involve a genuine transfer of
maximum that the assured can benefit over the fiveyear duration of the contract is RIOO
insurance risk combined with the more traditional finite insurance approach.6 Ibis produces a
million — i.e. claim recoveries of R250 million against a premium cost of R150 million —
so-called (blending' programme that should substantially increase the chances of premium
whereas the maximum cost to the assured over the five years is R30 million — i.e. R150 million
payments allowed as a business expense by the tax authorities and the product itself,
premium less R120 million profit commission. The financial outcomes for the fini te (re)insurer
assuming it comprises the necessary characteristics to make it an insurance contract.
are the exact reverse — lying somewhere between a maximum loss of R100 million, against
Many large and well-respected companies carefully consider the possibility ofusing finite The finite insurance contract has an aggregate limit of cover so that insurer's exposure is
insurance, now known as finite risk insurance. some ofthe largest limited to the amount ofaggregate cover offered (hence the term (finite insurance'). This
again differs from conventional insurance contracts, where the exposure of the insurer is
usually unlimited. Finiie
makes finite reinsurance an ideal mechanism for what we speculative again referred to the section chapter that IAS 39 'FRS 4 and guidelines
uncommon for insurers offer insurance which in essence combine under contract the 341
340
insurance contracts may also contain annual sub-limits. This prevents the early exhaustion of
the entire cover offered.
reinsurers in the world, who have set up specialist teams to work together with major corporate
premiums of typical finite insurance contracts are higher than those of
buyers and develop tailor-made risk-financing products, acknowledge this interest. These
conventional insurance where similar cover is provided. To offset this higher premium,
contracts are mainly aimed at combining traditional insurance exposures with a substantial level
there is usually a commutation option, which offers a substantial return of premium to
of risk funding for problem areas such as pollution liability. It is expected that these developments
the assured if the loss experienced under the contract is lower than the expected losses.
will continue as the issues of tax and accounting are resolved and the finite insurance market
offers more than the minimum, if any, risk transfer that was a characteristic of the early 'time and
Finite insurance can thus be viewed as limited cover being provided at higher rates, but with
distance policies.
substantial discounts being offered for better-than-expected claims experience.
Each product is tailor made and the major players include the largest reinsurers in the world.
It is important to bear in mind that, at the end of the contract period, the insured receives most of
the premium paid plus investment income (less any losses); this is, of course, very different from
13.4 Functions of finite insurance
conventional insurance.
The most important function offinite insurance is to smooth fluctuations in the insured's loss
experience over the course of a multi-year contract. It therefore meets the insured's need for
13.3 Characteristics of finite insurance stable, long-term cover at calculable prices.
Finite insurance programmes allow the insured to improve and control key balance sheet
The most important characteristic of finite risk insurance is the insureds limited (finite) risk
figures for captive insurance companies such as the solvency ratio and the reserve ratio.
assumption. In finite contracts, the insured firstly transfers the risk of an unexpectedly rapid
Finite products can also be used to increase the capacity to underwrite risk. By
settlement of losses from the loss reserve to the insurer, but the insurer's liability is limited to a
increasing a captive's solvency ratio, such products extend the captive's capability to
predetermined amount. The insured is responsible for losses above this predetermined amount.
underwrite additional risk.
Another distinguishing feature of finite contracts is that the contract period runs over several
years. This reflects the fact that risk managers' problems do not fit neatly into one-year intervals.
The insured benefits from long-term cover under favourable terms and finite insurers benefit
13.5 Why has finite reinsurance evolved? 7
from a continuous flow of premiums.
An additional feature is that profits accruing over the multi-year period are paid back to the We may ask ourselves why companies would enter into a contract where only a limited
insured to a substantial extent. The insured receives compensation for the limitation of risk amount of the financial consequences of risk is transferred to the insurance company. We
through proper risk control. would expect that companies would prefer to protect themselves in times of economic
Expected investment income is also explicitly considered in the calculation of the premium. volatility and hardships by purchasing as much insurance as possible. The following reasons
The time value of money is taken into account and the insured receives the benefits of the cause companies to purchase finite insurance.
investment income derived from the funds paid to the insurer.
Finite insurance or finite reinsurance differs from conventional insurance in terms of the
following: Increasing profitability
Finite insurance contracts are multi-year contracts. Conventional insurance is normally Many finite reinsurance contracts appear, at least initially, to have very favourable effects on
contracted for one year only. profits. This is because, under conventional accounting,
FINITE INSURANCE
a d
d ition to the jeasons cited if! this section. the gaowing acceptance of the Strategg: value Of "5k managet•nent has also given fin'te a boost- These contracts can oe useful
when constructing a financing plan takjng into account the strategic experiences that a
Boorarjon faces
FINITE INSURANCE
342
the premium is less than the benefit recognised from the policy. The premium set by the 343 Increasing solvency
reinsurance company takes into account the significant delays experienced in settling
claims and the investment income that will be generated on the funds set aside to meet The second main reason for using finite reinsurance is increasing solvency.
them. By taking this into account, the reinsurer is able to charge a much lower premium. Given that reinsurance is generally allowed in calculating a company's level of solvency, some
Such arrangements not only have the effect of increased profitability for the cedent, but companies choose to enter into such arrangements merely to achieve beneficial effects on solvency.
also, in passing, enhance its solvency position. This is typically achieved by financial quota share policies that reduce net premium and claim levels
Here is a very simple example to illustrate this type of time and distance - and thus solvency requirements — while minimising the transfer of the risk to the
arrangement.
1 2 3 4 5 Total
Year: This type ofpolicy is particularly effective where solvency requirements are set by reference to
—premium (100) (100) volumes of business and credit may be taken for reinsurance recoveries of up to a maximum of 50%.
Cash flow
—recoveries 30 30 30
Profit and loss (70) 30 30 30 30 50 Loss financing

Spread loss contracts help insurance companies to finance their losses. This type of arrangement
—premium (100)
may be viewed like a bank account: regular payments are made into the account with an automatic
—recoveries 150
loan facility in the event of a loss. Ihe cover obtained under such contracts is limited to the amount
—'profit' 50
Let us assume a premium of RIOO, and a five-year settlement arrangement with total of premium paid by the cedent, together with the notional interest earned. If the cedent suffers a
recoveries of R150. loss in excess of this cover, the reinsurer will pay the excess, but the cedent will be obliged to continue
The cash flow arising from this contract from the cedent's perspective is as shown paying renewal premiums until the reinsurer's net loss is made good. Such arrangements have the
in the example given above. effect of smoothing an insurer's results during periods of hardship in the market. Figure 13.1
Under conventional accounting practice, this contract gives rise to a net credit in illustrates this principle.
year one ofR50, being the net gain to the insurer ofthe RIOO premium paid, less the
Figure 13.1 Spread loss contracts
R150 claim to be recovered.
This type of finite reinsurance first came into being in the late 1970s and early
1980s, particularly in the Lloyd's market, where syndicates purchasing 'time and
Results
distance' policies generated an initial market. Flhe primary objective of these products
was to recognise the time value of money and to set a premium that took into account
A - Without spread loss
future investment income.
— Wi spread bss
Time and distance policies typically set out with certainty both the amount of
claim recoveries and the date of settlement, so neither underwriting nor timing risk is
present. Therefore, under current criteria, these policies would fail both tests of risk
transfer, since there is no timing or underwriting risk
A variation on this theme is the loss portfolio contract that guarantees the overall
Time
recovery but payments under the contract are dependent on when original losses are
actually settled. In this case, there is timing risk, but no underwriting risk This may not
constitute finite reinsurance under current thinking.
ADCs have the following advantages:

13.6 Major types of finite cover
They provide partial cover against IBNR and IBNER losses.
Owing to the diversity ofthe products that are available, it is virtually impossible to provide
They facilitate mergers and acquisitions. They increase
an overview of all the finite reinsurance products. The following is a representative sample
corporate value.
from the broad spectrum of possible products.
Finite quota shares (FQSs)

Loss portfolio transfers (LPTs)
In contrast to the previous two finite contracts, FQSs are designed mostly as prospective covers
Through an LPT, a risk or captive manager can cede future payment obligations resulting for current and future business. FQSs focus not only on improving solvency and increasing
from underwriting in the past. The insurer (or reinsurer) assumes the cedent's reserves for underwriting capacity; but they also contribute to smoothing a captive underwriting result. This
outstanding losses. The insurance premium approximately equals the net present value of can be achieved by, for example, an anti-cyclical quota share treaty Its basic principle is that the
the ceded loss reserves. 'lhe insurer also charges a profit and cost margin, as well as a risk insurer pays the cedent a commission that increases as the loss ratio increases and thus helps the
premium that reflects the timing and possible reserve risks accepted. cedent exactly when assistance is most urgently required. This type ofsliding-scale commission is
LPTs focus on timing risk. Along with the risk reserve, the insurer also accepts the risk exactly opposite to what is commonly found in the insurance market. The moral hazard that is
relating to the settlement of losses over time. The insurer runs the risk of loss in the event linked to this type of contract can be avoided by entering into a multi-year contract that allows
that the settlement of losses is unexpectedly rapid. LPTs provide the insured with the both parties to balance their results over time.
following benefits: The advantages of FQSs are the following:
The balance sheet figures of the cedent are markedly improved. The solvency ratio (the They increase and stabilise underwriting capacity for the captive. They smooth the
ratio of equity to premium volume) of a captive that enters into an LPT with an insurer underwriting results of the insured.
or reinsurer is, for example, improved, since the ceded liability exceeds the insurance
premium. This opens the possibility for the captive to write more of its parent's business.
LPTs enable captives to give up certain areas of business, to close the books on them Spread loss treaties (SLTs)
and to enter into new fields.
The problem many captives and large corporates face is that even though they can estimate with
LPTs are often indispensable in mergers and acquisitions by eliminating old risks and relatively high reliability the total losses that will be incurred during a future period, the
providing investors with more confidence that old liabilities will not get out of control. distribution of losses among individual years is uncertain. Tlis gave rise to the development
They avoid costly and lengthy run-off activities of losses. ofSLTs. The main purpose ofSLTs is to cope with timing risk and to smooth fluctuations in results
by spreading
Adverse development covers (ADCs) SLTs are characterised by two features:

(i) Premiums are accumulated over the entire term of the contract. (ii) Losses are
ADCs provide cedents with cover for losses resulting from the past. In contrast to LPTs,
distributed over a multi-year period.
however, no loss portfolios are ceded. The focus is instead on the insured's need for cover
in excess of the loss reserves already formed. essentially involves protection against losses
The insurer provides advance financing for any temporary negative balances on the account, but
that have already been incurred, but not reported (IBNR), as well as against losses that have
also runs some underwriting risk, because the insured does not have to balance the account
been incurred, but for which inadequate reserves have been made (IBNER), i.e. the
completely at the end of the period.
nightmare Of inadequate loss reserves. The insurer accepts part of these risks.
The advantages ofSLTs are as follows:
345
Owing to their design features, SLTs smooth the loss experience of the
They reduce the variability in the underwriting capacity of a captive. 347

346
13.7 Summary
They transfer the timing risk to the insurer.
They stabilise insurance costs by insulating the captive (cedent) from reinsurance market
Insureds are increasjng}y considering their insurance requirements in a multi* year time frame.
cycles. Most companies' business pfans extend over many years, and, under the correct circumstances,
finite insurance provides companies with the opportunity to do the same with their risk
management programmes by creating multi-year arrangements for at least part of their
Guaranteed cost approach insurance requirements.
Finite insurance also aligns the interests ofjnsureds and insurer and therefore results in more
The guaranteed cost approach is a well-known method where self-funding is combined with efficientty priced coverage, while simultaneously reducing the cost of arranging risk annuafly.
insurance. It is a popular approach that provides stability to the self-funding programme by Beyond the structural benefits, finite insurance programmes are also individuaffy designed to
partially transferring the financial consequences meet the client's needs.
In today's volatile environment, finite products, if correctly structured (satisfying the
of the risk to an insurer. It also provides elements such as risk control and claims-adjusting
underlying principles of insurance) provide a viable solution to funding needs by assisting risk
services so that the insured does not have to seek alternative sources for these services.
managers to take a long-term planning approach and to manage their exposures better.
The guaranteed cost programme is a prospective premium rated plan in which a fixed
premium is established prior to the policy's effective date.
The premium is developed from market rates and the insured's historical loss experience.
Although the premium is fixed, it may be adjusted after the policy expiration to reflect
significant differences between actual and estimated losses and/or dramatic changes in
exposure levels.
Retrospectively rated (retro-rated) programmes

Retro-rated programmes are widely used by companies in the US for their global casualty
risk programmes. They differ from ordinary insurance plans in that the premium varies with
the insured's loss experience, which allows the insured to directly influence its ultimate
insurance cost through risk control.
An incurred loss retro programme allows the insured to pay a premium that is adjusted
retrospectively, based on the insured's actual loss experience. The premium thus determined
is modified by a formula that contains prenegotiated minimum and maximum limits. A
deposit premium is charged at the inception of the policy, and this premium is adjusted after
the policy has expired to reflect the actual losses incurred.
A retrospectively rated programme is like a self-insured programme up to the maximum
premium. The lower the losses of the insured, the lower the premium; the higher the losses,
the higher the final premium up to the maximum will be.
CAPITAL MARKET INSTRUMENTS
349
$30 1996b)-
348
Table 14.1 Major tosses, 1989-2008, inflated to 2008 prices
Year Event Insured loss ($ billions)
Chapter 14 2008 Hurricane Ike

20.0
2001 Attack on World Trade Center 31.0
Capital market instruments 2005 Hurricane Katrina
115.0
Hurricane Fran
3.8
1995 Hurricane Opal 5.0
1995 Texas hailstorm

2.6
1994 Northridge earthquake 29.8
1993 Winter storm in Lothar 32.0
14.1 Introduction 1992 Hurricane Andrew 37,0

14.2 Catastrophe bonds (CAT bonds)
1992 Hurricane iniki 3.8
14.3 i Insurance derivatives
4.4 Catastrophe Risk Exchange (CATEX) 1991 Oakland fire 4.1
14.5
Double-trigger products
14.6 1989 Hurricane Hugo 10.0
Securitisation of catastrophe risks
14.7
Summary 1989 Loma Prieta earthquake 2.4
Total
296.5
14.1 Introduction Source: Constructed from various sources. A record of catastrophes can be found jn the
publications of the Property Claims Insurance Services and the Swiss Re Sigma series of
Sl 50 billion is not enough [insurance] to covet a $7 trillion economy and $12 to $15 trillion in publications.
property. The arithmetic just isn't there .... There's $15 to 20 trillion in the private capital markets. Add
that in and the numbers can work. Richard L Sandor, chairman and CEO, Centre Financial Products
Risk-financing instruments such as derivatives that access funds available in the capital
The need for risk transfer mechanisms has increased dramatically over the last decade. Changing market for funding catastrophe-type losses provide a solution to this limited capacity
demographics has resulted in a concentration of wealth in the high-risk catastrophe areas such as available in the insurance markets.
California, Florida and Texas in the US. This has resulted in large catastrophes such as Hurricane
Bernstein introduces the subject of derivatives under a most appropriate heading:
Andrew, which caused over $ 15 billion in insured damage in 1992, and Hurricane Hugo, which
'The fantastic system of side bets? He describes derivatives as the most sophisticated of
caused losses of approximately $4 billion in damage in 1989. The size of these catastrophes in
comparison to the net world-wide insurance capacity suggests that more capital is required to financial instruments, the most intricate, the most risky and very much in vogue. Yet, to
cover catastrophe-type risks.l many people, 'derivatives' is a dirty word.3
Despite the mystery surrounding these instruments, derivatives go far back in history.
Their use arose from the need to reduce uncertainty.4
I on the capacity available conservative maximum loss the gap is to be between
CAPITAL MARKET INSTRUMENTS
Today's derivatives differ from their predecessors only in some respects: they are valued Stock index options 3 286 242
mathematically and managed by computer, the risks they are designed to respond to are more Selected OTC instruments 500 3 450 7 777 11 200
complex, and they are put to further, more complex use. Interest rate swaps 400 2312 6 177 8815
Currency swaps 100 578 900 915
2 eernstein (t996).
Caps, collars, floors and swaptions 561 700 T 470
limi3
ted as 1994, toProetel the some even'S & of Gamble the themselves apparently and Gem-Ian but also Metatlgesellschaft linked dåivatrve to the arrangements øestige AG. aruj reputation skÆdenIy 1 083 5 742 16616 20 038
came of (he unstuck, firms suffer'og causing these esurnous losses These losses. Incl!ÆedThe surpose was not
Total
Source: Valsamaki$ (1999)
351
4 In the tuJapanese tvælfth lip bubble century, feudal of the lords sellersseven traded at tee-nth fikdleva[ in conuacts century trade involved that fai'S offered signed trodlryg conuactsprotection in
optlons called against gettres tulips the de rather (Cire than prorn@future of In bad the weatt*f tulips themselvesdelivery or warfare. of items Much they of the sold famousIn the
The next sections will briefly overview derivatives and other capital market instruments in the
350
US.
Derivatives are financial instruments that have no value of their own; they derive their value from
the value of some other asset, thereby enabling one to hedge the risk of unexpected fluctuation in 14.2 Catastrophe bonds (CAT bonds)
value/price. Derivatives cannot reduce the risks associated with the volatile assets, but they can
Catastrophe bonds provide returns to investors based on insurance events, in contrast to
determine who takes on the speculation and who avoids it.
normal bonds, where the return is determined by financial events. They are debt
Derivatives have use only in an environment of volatility. Their popularity is a function of the
instruments with relatively high yields that insurers issue. In the event of a catastrophe,
increasing volatility and uncertainty that is emerging even in areas long characterised by stability. For
the investor foregoes the interest payment, or part ofthe principal, or both.
example, until the early 1970s, exchange rates were fixed and the price ofoil varied over a narrow
Ihe relatively high interest rates and the diversification opportunities that these
range. The appearance of new risks in areas once considered stable has prompted the search for
instruments offer attract investors. The usual investment portfolio is affected by market
alternative and more effective tools of risk management. As Bernstein points out, derivatives are
risk, which is impossible to diversify because it affects the entire financial market. CAT
symptomatic of the state of the economy and of financial 5 markets, not the cause of the volatility that
bonds provide an additional investment alternative that is totally unaffected by variables
is the focus of so much concern'.
that normally affect the rate of return, since there is no correlation between insurance
Table 14.2 illustrates the explosion in the derivative markets. 'Ihe table shows the dollar value of
events/catastrophes and market fluctuations. Their patterns of gains and losses are
outstanding (existing) positions in derivative securities for which data is available. As can be seen,
different to those of other financial instruments and they therefore improve the return-
from 1986 to 1994 the market has grown from $1 trillion to $20 trillion. A survey in 1995 revealed
torisk ratios of investors who include them in their portfolios.
that when one includes the total notional amount of 'over-the-counter' (OTC) contracts, including
forwards and optional contracts, these markets add up to about $50 trillion (the gross domestic product
of the US is $7 trillion).
14.3 Insurance derivatives
Table 14.2 Global markets for selected derivatives outstanding contracts ($ bn)ö
1986 1990 1993 1994 Definition
Exchange-traded Instruments 583 2 292 7 839 8 838 A derivative is a financial instrument of which the return is based on that of an underlying
instrument,
interest rate futures 370 1 454 4 960 5 757
Interest rate options 146 600 2362 2 632
Currency futures 10 16 30 33
Currency options 39 56 81 55
Stock index futures 15 70 119 128

CAPITAL
MARKET INSTRUMENTS
Hedging Risk managers use derivative instruments to hedge their exposures to identified risk
exposures. The following are the two main derivative instruments used for hedging
purposes.
Definition
Hedging is defined as arranging for two different or opposing positions so that the potential
losses from one position tend to be more or Jess offset by profits from the other position.
S {1996: 305). nsk ofderwatives involves changes in contract

6 From a risk management perspective. these figures are a lit-tk misleading. The market be about SI 7 trillionvalues, rot notlonal amounts
The estimates of the replacement value ior all OTC contracts would
352 FINANCING the rise in the futures index. In other words, the purchaser creates a hedge by
purchasing catastrophe futures. If an unexpected catastrophe occurs, the
value of the index will rise and a rise in the price of the future will help to offset
Options
the unexpected cost of losses.
An option contract conveys from one party to another the right (not the
obligation) to buy or sell a specified asset at a specified price on (or before) a
specified date.
Catastrophe index futures
h should be noted that the principle underlying options contracts is that the buyer has the right, and not the
Ihe Chicago Board of Trade (CBOT) introduced index futures in December 1992.
obligation, to buy or sell the specified asset at a specified date. In other words, the buyer of an option has the right to
'Ihese are financial tools that allow the captive owner to offset its 353
'walk away' from the option contract or exercise it. There are two basic types of option contracts — the put and the
call.
The call option is the right to buy an asset at a specified price at a specified date. If at expiry the price of the catastrophic losses by using capital gains from the rise in the futures index. In other words, the insurer
asset is higher than the exercise price, the owner of the call option benefits at the expense of the seller. In the case creates a hedge by purchasing catastrophe insurance futures. Ifan unexpected catastrophe occurs, the value
Of the call, the buyer of the option will limit his/her possible loss to the cost of the premium. profit, however, is of the catastrophe index will rise and a rise in the price of the futures will help offset the unexpected cost
unlimited. The buyer of a call option will not exercise the option if the price of the asset in the market is lower than of losses. The hedge will only cover losses to the extent that the insurer's loss experience matches the loss
the exercise price. He/she can buy the asset in the market at a lower price than the option price. experience of the companies that comprise
A put option is the right to sell an asset at a specified price at a specified date. In this case, the buyer of the
Futures contracts are available for the eastern, mid-western and western regions of the US. The
option will 'walk away' and not exercise the option if the price of the asset is greater than the exercise price. The
contracts cover aggregate losses for the period in the following lines of insurance: private and commercial
holder of the put option can sell the asset in the market at a higher price than the option (strike) price.
car damage fire multiple peril for home owners
Futures contracts • commercial buildings and farm owners earthquake
A futures contract is a notional commitment to take delivery (purchase) or to make delivery (sell) of a given quantity • commercial inland insurance
of a specific instrument on a specified future date at a price determined at the time oftaking out the contract. A
The contract unit corresponds to the underwriting result on premiums earned
price is therefore set today for settlement in the future.
The basic principle of hedging with financial futures is that a futures In the property/casualty industry, catastrophe insurance futures are, for example, based on an index
position taken must be such that the gains caused by interest rate changes on that measures the level of catastrophic loss for ten of the hundred companies that report to the Insurance
the future compensates the holder of the future exactly for the losses incurred Services Offce. These ten companies are selected on the size of their business and the diversity of the risk
on assets or liabilities. they represent. The losses that occurred in the previous quarter are statistically weighted and introduced
Catastrophe index futures and options are insurance derivatives that into a formula to determine the index value. If the average loss development by the due date is less
allow the purchaser to offset its catastrophe losses by using capital gains from favourable than the market participants originally expected (when the contract was concluded), the pool
NSTRUMENTS
index drops and the sellers of the futures are credited with the profits. In the reverse situation, the net index
rises and the net value of the buyers of the futures becomes positive.
The price mechanism is illustrated in the pay-off diagram in figure 14.1 (overleaf). Assume, for
instance, that a captive wants to hedge itselfagainst a possible increase in losses and the index currently
stands at a loss level of 70%. corresponds to an index level of or $30 000 per contract. captive sells futures
on the index at this level (shorts the index) commensurate with its premiums. Ifthe loss ratio then declines
over the period of the future to a level of 75%, the index drops to a level of 0,25, which corresponds to a
spot price of$25 000. By buying in the index at the spot price of$25 000 and selling it at $30 000 (the value
ofthe futures contract), the captive makes $5 000 profit per contract. This profit on the future then offsets
the increase in the loss ratio on the captive's own portfolio.
RISK MANAGEMENT: RISK FINANCING CAPITAL MARKET INSTRUMENTS
354 Catastrophe options
The CBOT launched a complex of catastrophe insurance options in September 1995. The
contracts, called PCS catastrophe insurance options, track nine property claims services
Figure 14.1 The pay-off diagram for insurance index futures transactions
catastrophe loss indices covering US exposures nationally and regionally?
Value of long position With symmetrical hedging transactions using insurance index futures, the debits on an
(put buyer)
insurance company generally remain fixed, even if the resulting loss experience is more
10
favourable than expected. The hedging strategy surrenders any potential profit resulting from a
loss ratio that is less than expected. 'Ibis disadvantage vanishes with an asymmetric hedging
transaction, such as insurance index options.
Option premium
-10
index at contract date Figure 14.2 The pay-off diagram for an insurance index put option
-20
0/3 0.4 015
Pool index at maturity
To determine the price of a $25 000 catastrophe insurance futures contract, the CBOT uses the
following formula: o
$25 000 x (Incurred catastrophe losses/estimated property premium).
For example, assume that the ten insurance companies using the pool will experience $300
million in catastrophic losses during a quarter, that 75% of the losses will be reported by the In the case of the put option, the buyer of this option buys the contract (the premium) and obtains
end of the quarter, and that the insurance companies will collect $4 billion in property the right to sell the index at a given price. The put buyer makes a profit on the put if the loss
premiums. The contract price is then: experience is lower than expected. His/ her profit is determined by the value of the long position
in the put, less the premium paid.
$25 000 x {($300 million x 75%)/$4 billion)}
= $25 000 x 0,05625
= $1 406,25 per contract.
Catastrophic event-triggered equity put
Large companies have diffculties in arranging sufficient catastrophic cover or they may in
With the contract price, the captive can determine how many contracts it needs to hedge its certain instances not be able to arrange any cover at all because this risk is uninsurable.
catastrophe risks with the following formula:
Traditional forms offinance such as loans appear as a liability on the balance sheet, thus
producing a negative impact on a company's ability to attract capital and solvency ratio, and are
Number of contracts = (Earned premium(contract size) x (Hedged losses/ % reported).
often not available when most required (such as in the event of a major catastrophe hitting the
company).
For example, a company with $15 million in earned premiums using $25 000 contracts could
hedge 100% of its potential losses by buying 800 contracts. It we assume 75% reported losses,
'QOT(196).
this number is calculated as follows:
356
($15 million/$25 000) x

$600 x 1,3333 One way of solving this problem is for a large, highly rated financial institution to sell a put
800 contracts.
to a corporation, to become effective following certain specific events. The puts which entitles
355
the holder to sell equity instruments to the institution under certain circumstances, may then Based on its current financial position, Ostrich Meats feels it can assume a retention ofRI
be used to provide value after a catastrophic event. The equity that is made available under million per occurrence for product liability. Setting its retention at this level produces a
the put typically takes the form of non-voting preferred shares. After the catastrophic events considerable saving compared to expected losses. Ostrich Meats produces meat and feathers
the company must absorb the uninsured loss, and this is reflected in the income statement. for the fashion industry. Its raw materials are ostriches, which it processes into finished
However, by exercising its put option, the company can restore shareholders' equity and products. It maintains a strong financial position, but has experienced wide swings in earnings
financial leverage/solvency ratios. Essentially, the sale of such a put option ensures a supply of due to the changing price of its raw material. Ostrich Meats has little choice but to decrease
standby capital, helping to guarantee the corporation's survival in dire circumstances. its retention to a less optimal level to avoid the unacceptable results that might occur in a poor
year.
The double-trigger arrangement is ideally suited to the situation in which Ostrich Meats
14.4 Catastrophe Risk Exchange (CATEX) finds itself. In traditional arrangements, the retention is set at the beginning of a policy year —
This is an electronic system for trading risk that began operating in 1996. It improves the link when the financial results of the company are not yet known. However, if the retention can be
between capital markets and the insurance industry. Using an open 'bid and asked' market, the adjusted according to an index that simulates the financial results of the company, then a higher
CATEX enables licensed risk bearers to exchange (or swap) catastrophe exposures offered by retention can be selected. In the case of Ostrich Meats, we simply construct an index that varies
other subscribers. For example, a trader can swap California earthquake risk for Florida with the price for ostriches on the date of a loss. For example, if the business plan for Ostrich
hurricane risk. The exchange transactions are processed via an electronic mailbox in which the Meats is based on a price of, say, R20 per kilogram, then all losses will be paid using a retention
risk carriers post the precise descriptions of catastrophe risk they consider for exchange. At the of RI million. If the price of ostriches is R30 at the day of the loss, then the revised retention will
opening of daily trading, the CATEX system calculates a set of hypothetical exchange rates that be R666 666 (which is calculated as Rlm/(30/20). If the price goes down to R15, then the
are based on historical claims distributions and the latest dealing rates. Ihese can be regarded retention increases using the same formula, resulting in a retention of RI.3 million. The double-
as relative prices, which adjust to the actual, market situation automatically and within split trigger cover is likely to be most effective when the retention or premium variation is captured
seconds, so that supply and demand are brought into balance. within a somewhat narrow range rather than allowing for wider fluctuations.
This helps both insurers and reinsurers to diversify their portfolios and to employ capital While double triggers are getting the attention of a growing number of risk managers, few
more efficiently. of these programmes have been created. One obstacle is the difficulty of creating formulas
that achieve the floating relationship between risk and financial results.
14.5 Double-trigger products

14.6 Securitisation of catastrophe risks
Double-trigger cover is designed to provide cover if the financial environment in which the
There is a growing convergence between capital and insurance markets. This is driven by
risk manager operates changes. The risk manager's decision to retain a certain amount
demand for more capacity by insurers and reinsurers, on the one hand, and by investors
oflosses is based on a particular financial environment. But what happens if a significant loss
looking for innovative investment vehicles that are uncorrelated to traditional market risks
occurs at a time when another element in the organisation's financial environment is moving
and that provide higher yields and excess returns, on the other. In response to this need by
the wrong way? The 'double trigger' is a retention tool designed to address this problem.
both insurers and investors investment banks, reinsurers and insurers have created modules
Under the double trigger, the amount of risk assumed under the retention programme
for
is set as a variable based on an outside index that
CAPITAL MARKET INSTRUMENTS 357
example is adapted from Sanderson (1997).
358
captures a financial risk affecting an organisation's financial results. It is best explained by referring to
the following example.8

the securitisation of catastrophe risks in order to place such risk directly with investors in the Figure 14.4 The CEA insurance programme
form of securities.
10,5
The benefits of securitisation to both investors and insurers or reinsurers are summarised in Priority/cei!ing
table 14.2. (S billion)
Contingent excess cover by participating insurers:
$2 billion
Table 14.3 The benefits of securitisation to cedents and investors9
7
Earthquake risk bonds: $15 billion
Investors Series A Series B Series C
Cedents 6
Attractive valuation Layer financed by state genera! revenue bond: $1
Diversification of sources
billion
Uncorrelated diversification
New capacity 4
Sophisticated risk assessment ReinsurerS layer: $1 billion
Innovative structures
Competitive performance
No credit risk
Stable pricing Contingent excess cover by participating insurers:
$3 billion
The so-called initial contribution by participating insurers: Sl

insurance securitisation is defined as the transfer of underwriting risk from billion
'earthquake risk bond'
I insurers and reinsurers to investors by transforming underwriting cash flows into marketable securities.10
(ERB) is part of a state
insurance
fie concept of securitisation is based on the following method of operation. n in return for a
programme of the California Earthquake Authority (CEA). The project consists of multiple
specified capital payment, the investor purchases 'insurancelinked bonds' that are issued by a
layers and involves the financial markets, as well as the entire insurance and reinsurance
trustee on acceptable terms. The insurancelinked bond is a security with a conditional or
industries. The structure of the programme is depicted in figure 14.4.
unconditional entitlement to repayment and usually an entitlement to a variable interest rate.
The ERB's have a 10-year maturity and a cash-flow structure that targets the risk
The rate of interest is dependent on certain insured catastrophe risks. These risks are
appetite of an appropriate investor group. Their semi-annual interest coupons promise a
documented in the security in question, hence the term 'risk securitisation'.
contingent yield that is clearly above the risk-free return on government bonds. If an
Interest is provided from two sources: firstly, from the insurance premiums paid by the
earthquake with losses exceeding $7 billion occurs in California during the risk period (i.e.
insurance company to the trustee in order to transfer its catastrophe risks, and, secondly, from
within the first four years after the securities have been issued), the interest payments reduce
payments derived from investments in short-term government bonds made by the trustee and
to as little as 0% for the rest of the maturity period. In the extreme case where a catastrophe
financed mainly by investors' capital and insurance premiums. If no catastrophe claims arise
follows immediately after the bonds were issued, investors would receive a zero return.
during the term of the insurance-linked bond, the investor receives full repayment of capital,
However, the repayment of principal at par value is fully guaranteed at final maturity,
as well as the agreed rate of interest. However, if there are claims, the insurance company is
independent of the loss pattern.
initially reimbursed from the general fund, comprising insurance premiums, investors' capital
In order to raise the necessary capacity of $1,5 billion with this bond structure, an ERB
and investment returns already received. The balance is then available to investors.
volume of more than $3,5 billion must be issued. Approximately $2 billion of this amount
will have to be invested in government securities to ensure the repayment of principal at
9 Ct-,01f10ky 098}. forms
10 Nicholas (1998).
maturity. The remainder the actual risk capital for the fifth layer in the CEA
CAPITAL MARKET INSTRUMENTS 359
programme.
The principles of securitisation can be further illustrated by using an example of a *Ctroo paitia[lY been adapted from Durrer (19960}.
specially structured bond issued to cover earthquake risks in California.12

1 1 Wagner {1998),
360361
14.7 Summary
Chapter 15
Composite financing
strategies
Although there will always be a demand for traditional products, insurance derivatives and capital
market products present insurance buyers and investors with interesting and viable alternativefunding
mechanisms.There is good reason to be optimistic about the development of these products,
notwithstanding some initial impediments to their growth.
Tremendous investment is being made in improving and developing them and the growing
number of conferences and articles in journals will aid in developing understanding of how they can
be incorporated into a conventional funding programme to provide additional protection and a cost-
effective alternative to other funding options.
Introduction
Statistical analysis in risk management
Composite financing strategies: Transfer optimisation
Optimisation models based on expected utttity Summary
1 5.1 Introduction
Ihe two preceding chapters dealt with insurance and retention as pre-loss means of financing
the consequences of pure risk exposures. Evident in the discussions is the interrelationship
between these strategies. Both financing means were evaluated from the point ofview of a firm's
objective ofmaximising value, which entailed the consideration of alternative sources of finance
and their associated costs.
Although not ignoring the post-loss alternative sources of finance, the discussion has
centred more around the problem of a composite retention/ insurance strategy, given that certain
practical considerations modify the value maximisation and risk management objective. The
aspect of risk finance integration or optimisation now needs to be explored.
From the standpoint of risk retention, the firm acts as its own underwriter. In doing so, it
becomes imperative that the statistical properties of the risk (retained) should be understood. If
there is to be optimal transfer to the insurance market, for example, this cannot be achieved
without the process Of statistical analysis — the two aspects thus become interlinked in the risk
management discipline. Hence, this chapter combines both the area of Statistical analysis in risk
management and the question of optimal composite financing strategy.

RISK MANAGEMENT: RISK FINANCING COMPOSITE FINANCING STRATEGIES
362 of the information available to the risk manager is not fixed, yet the assembly of such
information into a form that can be used to make financial decisions is critical to the quality of
such decisions.
Statistical analysis in risk management entails the accurate estimation of loss costs and The objective is the estimation of a particular risk management statistic, termed the maximum
probabilities. To show its importance, the following should be considered:
probable yearly aggregate loss (MPY).
• Such estimates are essential for decision-making in risk retention and selffunding plans,
providing aspects such as estimates of maximum probable yearly aggregate loss.
Definition
By carrying out a risk-return exercise, the analysis enables a firm to identify economies of
risk retention. The maximum probable yearly aggregate loss is that value that wilt equal or exceed in a
stated proportion of afl cases the total foss amount during a oneyear period from a
The analysis is useful in designing and placing insurance packages, because of the need for
specified peril.2
accurate statistics when 'marketing the firm's risk to potential insurers. 1
Since it represents the largest total loss amount that an exposure or group of exposure units is
Chapter 3 introduced the theory of probability and its application to risk measurement. The
likely to suffer during the one-year period, the value requires an implicit or explicit choice ofa
intention was only to provide an initial outline to assist in the development and understanding of
probability or likelihood level. By implication, if the probability or likelihood level is, say, n,
such concepts as long-term expected costs and pure risk (premium) to insurers, which would then
then (I — n) represents the probability of ruin (i.e. the probability of losses exceeding the MP
be referred to in subsequent chapters. The chapter also introduced the concepts of regret and
Y).
utility to explain decision-making under conditions of risk.
To illustrate, a 90% MPY of, say, R50 million implies that aggregate losses will in 90%
The section in this chapter that covers statistical analysis must be seen, therefore, as an
of the cases be equal to or less than R50 million. Doherty 3 defines the probability of ruin as
extension of the statistical theory developed earlier and deals with the more sophisticated aspects
'the probability that the aggregate losses in a given period exceed the value of the fund in that
surrounding the prediction of loss costs and associated probabilities, The section that follows
period'. Hence, if a retention fund of R50 million is established at the beginning of the period,
covers composite financing strategies. The concern here is with the prospect of the optimal form
the probability of ruin, assuming a 90% MPY, is (l — 0>90), or 10%. lherefore, what is needed
of financing that requires a package drawing on two or more competing sources, e.g. the use of
is an estimate of the probability distributions for losses of various types that face the firm that
debt, internal liquid resources and insurance. The primary consideration, however, is the partial
can then be aggregated to estimate the distribution of total loss costs facing the firm.
use of insurance, i.e. the integration of insurance and retention funding as an optimal composite
Concepts such as MPY, estimated maximum loss (EML) and maximum possible loss
financing strategy. 'Ihe integration of insurance and retention funding requires a statistical
(MPL) are mentioned frequently in the risk management and insurance literature, and it is
analysis to be conducted. This provides the reasoning for combining, in one chapter, the subjects
necessary to briefly clarify the difference between the EML and the MP Y, after also defining
of statistical analysis as applied in risk management and transfer optimisation.
what is meant by EML.4
15.2 Statistical analysis in risk management Definition

The estimated maximum loss is that value that will equal or exceed, in a stated proportion of all cases,
the amount of toss from a specified peril or perils?
Maximum probable yearly aggregate loss and estimated
maximum loss In less technical language, EML is the largest loss that can occur under the worst conditions that are
In statistical analysis, the concern is primarily with data that describes prospective losses and likely to occur. EML is a severity concept associated
their associated probabilities. The quantity and quality

2 t:urnrrains and Frejfe/der a 978/79; S?)
1 This point is Increasingly significant. The role of statistics In the markeclng of risk to g7tentiaæ insurers es now regarded m serious Ilght, Attention is focused on of terms can be ite confusing, because analysts have not been to fully agree on standard definitions for and s. The definitionsadopted here may be as
having a as they are intended to of a nature.
data qualit-y and the numeric of risk managers. consultants and insurance brokers. It possibly state that considerable asymmetty in Information has arisen in past
because of unclear lines of responsibility concerning the collator: dissemi natlan of loss statistics and a general lack of appreciation M (he im portance of loss
experience sta tisEiC%The mic natLlteof coupled with the more deliberate and sclentrfic appmach to f isk management. has considerably altered perceptions 3bout 5 Cutnrntns and Freife'der 1977. 261 (1978/79: McGuinness SS).31-40), and and Heins (198] 67).
sr-atßti% 364
363
with individual losses, i.e. it does not consider the frequency/probability of the individual loss problems, however. The choice of MPY loss values and probabilities is somewhat arbitrary,
occurring to arrive at an expected loss for the given level of likelihood. Instead, it represents the there being little theoretical guidance to indicate how and what levels should be chosen.
maximum severity of an individual loss for the given level of likelihood. 'Ihe EML does not Consequentlyy a careful analysis of the firm's financial position, coupled with informed
apply to a group of exposure units, as it is valid only for individual exposure units, and, in fact, judgement, is also necessary before selecting the loss level.
for individual losses. The MPY, on the other hand, applies to one exposure unit or a group of Establishing the probability value is possibly more difficulti since management is generally
exposure units and is that value that will equal or exceed, in a stated proportion of all cases, the unaccustomed to thinking in terms of probabilities, especially small probabilities. There is also
sum of all losses sustained by the group of exposure units during a one-year period.6 the related problem that the MPY represents only one single point on the probability distribution
of total loss costs. Hence, in complex situations, the selection of the best alternative necessitates
a more thorough analysis of the distribution; management might react very differently to a 0,90
Applications of estimated maximum loss MPY of R500 000 and a 0,995 MPY of RI.5 million. Evaluation of the entire probability
EML is often used by insurers in deciding whether their capacity permits the acceptance of distributions or even multiple MPY points is, however, a very challenging task.
particular exposures? For example, in the field of commercial fire insurance, underwriters use
the concept of line limits, which indicate the maximum exposure to loss that the company will
Sources of data
accept on any given property. In risk management, although the EML could be used to determine
Before proceeding with the discussion ofestimating techniques, it is appropriate to look at the
and limit the amount of insurance to purchase, thereby reducing the premium outlay, such a
sources of data available to the risk manager. The most obvious source of data is the past loss
reduction should rather be achieved by increasing retention through increased deductibles, but
experience of the firm (and other firms engaged in a similar activity). The historical data may be
maintaining coverage limits equal to the total exposure.8 Insurance is most valuable for large
presented as a frequency distribution, and, given the fact that the underlying factors producing
infrequent losses, which, when they occur, are most likely to cause the firm severe financial
the losses have remained unchanged over a period of time, the frequency distribution could be
problems.
used to estimate the distribution of future losses a process known as statistical inference.
Unlike an insurance company, which usually has at its disposal a much more
Applications and limitations of maximum probable yearly comprehensive statistical base, the individual, non-insurance firm typically does not hold a
diversified or large portfolio of risk exposure units. Therefore, a point to consider is that its
aggregate loss
historical toss experience may be a limited and random selection of events that is not
While it is useful for a firm to know its EML values, the MPY has greater applicability,
representative of all possibilities. The quality ofdata will depend mostly on circumstance, but it
especially when an increase in retention limits is being considered. Naturally, the possible losses
is generally related to the size and quality of the sample from which the experience is drawn. ne
to individual exposures will influence the aggregate loss distribution, and different deductibles
quality of internal data depends on the number of exposure units and the length of experience. 9
will lead to different MPY values. It is, however, the MPY and not the EML that is of primary
The internal data may, of course, be supplemented by the use of external data, but its use
relevance to risk management decision-making; knowledge of the MPY can, for example, assist
requires, firstly, accessibility and, secondly, presupposes relevance. The relevance of
in decisions concerning the trade-off between higher risk following further retention and the
information depends on whether the firm is representative of the industry. Although statistical
resulting savings in premium.
testing may assist in
The principal function of the MPY in situations where the firm may have the opportunity
to choose between different levels of retention is to provide
Fc
10gether. years, it these would factors have determine i ODO vehicle-years the size oi of the expeoerv_e statistxal fram sample which from to which estjmate data probabilities is drawn.
Thus, and Ifwould a firm be has regarded owned 100 as having vehicles a sample for a periodSize
6 a stated probability value, fo, one exposure unit is less than the EML for (hat unit. This is because the MPY is based on the loss cost distribution that incorporates 000 comprising 100 exposure units. each covered for 10 years.
no-loss case, while the EML is obtained on the condition loss has occurred. probability to the no-loss case. any specified cumulative probability level will be reached 366
at smaller value for the MPY
8 Of the decision to Increase the deductible implies the support of sound analyses of costs. probabil,cies and other factorex
365 establishing whether samples of loss experience for two firms may be drawn from the same
probability distribution, if internal data is limited, whether external data is relevant becomes a
additional information about the alternative loss cost distributions, i.e. information to matter of subjective judgement.
supplement the usual summary statistics, such as the mean and variance. Its use is not without
A second source of information comes from subjective, non-statistical sources, which include, losses, that of Xi the denotes ith outcome the possible is denoted outcomes by Pr(Xi), (i = 1 (x 2
for example, technical, organisational and economic judgements made by the firm about the represents 3, ...y n), and the randomthat the
probability variable: total annual losses), then the following equations can be used to
probabilities surrounding an event happening and the economic consequences of the event. Often,
determine the mean, variance and skewness of the total loss distribution: L2
statistical and subjective data can be combined, so that statistical inference is tempered with
subjective judgement — a process that resembles the use of credibility formulae by insurers.
The mean or expected value:
E (X) = Pr(Xi)
Estimation of maximum probable yearly aggregate loss from loss
data "Ihe variance 02(x):
A number of different MPY estimation procedures exist; this section examines several of the more
prominent approaches. In estimating the MPY, we attempt to locate a particular point or points on The standard deviation is the square root of the variance: o(x) =
the probability distribution of total losses. One way to develop such information is to observe total
cost of losses (of the particular peril) over a time period and to develop an empirical distribution The skewness, (X) is developed by using the formula: 13
from the observed values. The observations can be divided into convenient ranges and the relative
frequency (the number oflosses within each range divided by the number of observations) and the
03(X)
cumulative frequency (indicating the proportion of losses lying within the range in question or
below that range) can be calculated. The frequency values may be used as probability estimates. In order to calculate a value for the MPY from one of the approximation formulae, three basic steps must
The MPY would then be selected from this empirical distribution.
be followed:
An alternative approach is to use the available data to develop two probability distributions,
one relating to the frequency of losses, the other to the severity of losses. The combination, (i) a probability level must be selected,
mathematically, of these distributions would derive the total loss distribution, a process known as (ii) the loss distribution statistics calculated,
convolution. The process is somewhat richer than direct estimation of the distribution of aggregate (iii)and these statistics must be inserted into the approximation formula of the method chosen.
loss since the additional information on frequency and severity is used, but the computation of an
MPY estimate is relatively diffcult. Consequently, other procedures have been developed that i 71 27heseformu Tne mean and vanance are gjven have in most been discussed 'n chapter see, 8for example, Freund,Widiarns and ( 1988}. It must be noted that the formulae
when the requ,red probability distributions are known. If the total toss distribution is unknown, its mean, variance and skewness must be estimated from the
available data. The equations be used. probability values are not known. sub}ect to certain jssurnptlons the estimates ofthe mean. •nriance and skewness can be obtatned
estimate the MPY without determining the total loss cost distribution, but rely instead on certain using thefolkwlng fo«nulae•
statistics (sample mean and variance) that can be computed from the data.
There are, thus, two groups of estimation methods for MPY — one contains the methods
that require the estimation of the form and shape of the underlying loss distributions, while the
n denotes the of years of data. the observation the and where the symbols and stand for the sample
other contains the procedures that base the MPY on statistics developed from the loss data.
denominator is obtained by taking the root ofthe variance (the standard the resulting The numerator
the third moment about the mean. For almost au risk loss that the loss distribution
10 A credibility formula can be defined simply as 3 weighted average of, say, a statistical estimate and a subjectlve judgemental 368
(Doherty, 1985; 79)
367
Three methods of approximation are discussed below.
Methods based on loss distribution statistics
The normal approximation
The MPY calculation methods based on loss distribution statistics utilise the mean, variance
and, in some cases, the skewness of the total loss cost distribution. u Skewness indicates the This approximation assumes that the distribution of total losses can be approximated by a
extent of the departure of the loss distribution from symmetry. normal probability distribution. Given that an MPY probability of a is chosen (ruin probability
Suppose that there are n number of possible outcomes for total annual is (l — the MPY estimate is the value on the total loss distribution that is equal to the mean
plus a multiple of the standard deviation. The number of standard deviations is indicated by
za, which is a point from a standard normal probability distribution (the value of Za is obtained distribution is so highly skewed that this causes the total loss distribution to be skewed too. By
from a table of the standard normal probability distribution) . 15 ignoring the skewness, this method will produce an underestimate of the true MPY.
The formula for obtaining the normal approximation of the MPY estimate is the
Ihe Chebyshev method
following:
The MPY calculation formula for the Chebyshev method is similar to that for the normal
MPY approximation, except that the standard deviation is multiplied by a different number. The
formula is as follows:
where:
MPYN = the maximum probable yearly aggregate loss estimator utilising the normal MPYc = E(X) + ko(X)
approximation where:
MPY = the maximum probable yearly aggregate loss estimator based on Chebyshev's
= the mean of the distribution of total losses
theorem k = the solution to the equation k = VI/(I — a), where a is the MPY probability
= the standard deviation of the distribution of total losses z = the
level.
standard normal random variable z = the value of the random variable z where
Pr (Z z ) = a.
This method is based on a result from statistics that is called Chebyshev*s theorem. 18 Since the
theorem is valid for all probability distributions, there is no need to be concerned with the
If the actual values of the parameters, mean and standard deviation of the total loss
actual form of the total loss distribution. The Chebyshev method produces an estimate of MP Y
distribution are not known and estimates are used instead, then the formula must be
that is larger than the true
adjusted for sampling error. There are two adjustments — one is to replace za with a
comparable value from a student t probability distribution (because the standard deviation
is not known); the other change adjusts for using an estimate of the mean rather than its The normal power method
The normal power method is similar to the normal approximation technique, but purports to
actual value. 16
correct the skewness of the distribution. The procedure requires the estimation of only one
The problem with this method is that the probability distribution of total losses may not more statistic from the observed loss data than either of the methods previously discussed.
closely approximate a normal distribution. Although the central The formula for the MPY calculation using this method is:
MPYNP E(x) + [za + (1/6)

14 A good discussion of MP"' estimation technjques using loss distribution staustlcs which can fotlows be found an in intuitrve Cummins estimation and Freifelder
81 O(X) where:
procedure-09709: This metho.::lb-27). Cummins and Freffelder also include in their d•scusslon Allen-Duvall method, is not reviewed I-ere, bur the above publication
and a publicatlon that a standard by Allen norrnal and random Duvall (1971 varoble ) can will be assume referred a to value •n thls that regard.is less than or to
MPY = the maximum probable yearly aggregate loss estimator utilising the normal power
15These tables usually give the cumulative is probability that a standard norrnal random varia"e will be less than or equal to Z. with a probab•lity Ota. TO a certain method
number. The Interpretation of z
Illustrate, suppose a 90% MPH value Is desired: then o formula 090 should and from be the used table to produce one can determine a normal approximation that z 8] (x) = the skewness of the total loss cost distribution and where the other
is 1,64 MPY that applies at the 16 In order 10 adjust for sampling errorr ttk followlng a level:
where:
MPY* — the normal approximat•on MPYest4mate adjusted for sampling error n = the
terms have the same interpretation as those provided earlier.
number of annual total loss observatlans utilsed to estimate x and 5(x)
t = student random varlable fon_, = the value of for a student random vanable n- I degrees of freedom
where- As an exjmple. suppose 0.99; the value Ofk is TO theorem 'The that a randorn take on a value within standard dev,atjons of the mean is at lust
As an the norrnal distribution caser tables of Ibe r dsttibution are available in most statlstics books. 370
The multiplier, recognises the fact that is being used In plate of E(K}, while tie substitution of for 20 recognlses the regAacefikT oi a" by
369
For skewed distributions, the normal power method is the most accurate for
limit theorem can be used to show that the distribution of total losses will approach normality
as the number of losses increases, for many type of perils the individual loss (severity)
estimating MPY, but one limitation is that it tends to overstate the true MPY if the loss probability distributions will yield a workable mathematical expression for the probability of
distribution is highly skewed. 19 aggregate losses - one such combination being a Poisson frequency and a gamma severity
distribution. Z3
Methods based on loss distribution estimation Once the mathematical expression for the total or aggregate loss distribution is obtained,
The objective of MPY analysis is to estimate the true MPY value accurately. The methods the MPY value can be computed. The cumulative total loss distribution can be tabulated using
described above utilise approximation formulae to accomplish this task, and in some cases this the mathematical formula that describes the distribution; this involves calculating the probability
task is accomplished adequately. However, since the value of the MPY depends on the form of different total loss amounts and then summing these numbers. The MPY value is the smallest
and shape of the total loss distribution, methods that estimate the actual distribution of total loss figure ofwhich the cumulative probability equals or exceeds the MPY probability
aggregate losses are the most accurate way to calculate MPY.20 There are two principal methods level.24
of estimating loss distributions: analytical tabulation and simulation.

The simulation method
The analytical tabulation approach In contrast to analytically tabulating the distribution of total losses, an alternative approach is to
'Ibis approach involves the use ofestimated frequ ency and severitydistributions. These perform a computer simulation of future loss experience.25 Ihe results of the simulation can then
distributions are then combined mathematically (convolution) to develop the estimated total, or be used to estimate the total loss distribution. This technique, like the analytical tabulation
aggregate loss distribution. The process of combining the frequency and severity distributions method, requires the development of separate frequency and severity distributions of the actual
can be performed in two ways. loss experience.
The first is to use empirical distributions, whereby the development of the total loss Simulation is a mathematical procedure that aims to provide synthetic loss experience.
distribution is essentially a counting problem. 'The actual process is to 'count' or list the various Each replication represents the simulated loss experience for a period of one year, with the
total loss amounts that can arise and then to determine the probability of each outcome. 21 occurrence of loss determined by a random process. A simulation has two principal steps.
Although this process is the least complex, it may be practical only when the frequency and Firstly; a random number is generated. Using the frequency of loss distribution, this number
severity distributions have a fairly small number of classes. Suppose the maximum frequency determines how many losses have occurred in the simulated year.26 Secondly, additional random
shown in the frequency distribution is p, and the number of severity classes is q, 22 then the numbers are generated, one for each of the simulated losses. These
number of possible loss sequences is:

23 The Poisson or negative to describe the frequency of distribution accurately. In addition to the gamma distribution. other distributions that might be used to describe the
seventy distribut,on are the and Pareto distiibutions. to choose among these, various statistical that measure the goodness available. for example, the Chl-square test. Risk
management consultants often ignore the goodness-of-fit problem by making an assumption about the proper theoretical distribution. The of the incorrect theoretical
distributions lead to substantial errors the calculated values of and other For of theoretical frequency and severity that do not Yield expressions for the aggregate loss
analysis methods be This approach the whereby the tabulated distribution then summed or cumulated to permit the MPY value to be development of such 24 is sometimes
to derive a for the cumulative of and in cases the
25 Risk management consulting companies have constn„-ted simulatlon models over the yea,s to assist them and the,r clients scenarios for the various they
Without much diffculty, one can gauge how quickly tabulation can become unmanageable: face ourlng past deode, the has been involved a risk s,muJation mode' within the Eikos A of the Is
given the next section and as computer screen snapshots. The author would to acknowledge the of professional assisted by the at
for instance, if a firm had, say, 20 losses in a single year, combined with, say, a severity diffaent stages of the process and into reality. Thanks extended van Warren Koch and Ron and stoop.
who continues to with the author refining and extending its capabilities Improving the enhance the model as a technique for unde,
distribution of 20 classes (to give some precision to defining loss values), then one is left with
random number selected and unity, which be compared with the cumulative shown for frequency to determine many are assumed. if, for the of no loss Is 05 and
sequences or outcomes in excess of x 10. the of loss occurring is 08 and the random number lies in the range then one loss is to have arisen
19 Accurate results have been repolted for See Beard. Pentlkalnen and Pesonen (1977),
20 With these approaches possible to approximate the true MPY value at least as accurately as with any of the statistical approxim at't" methods
21 A readable pvesentation Of this method. supported by an example, IS given by Cummins ard freifelder (1979a: 26-37). 22 la the practK-al Sill-jaliO'1, the empirical
severity of loss data is condensed into a set of jntervals_ By using the mldpointS (o a Il values in the intervaE. the severity distribution may be si m plihed to permst
the computation without the use of a computer. use intervals can produce inccuracies, and is usually better to obtain the precise results.
371
The other approach is to use theoretical probability distributions to describe the empirical
frequency and severity data. Obviously, the theoretical distributions must be chosen carefully to
ensure that the data is accurately described. one limitation is that only certain combinations of
27 This is another undertaken by a numbe procedure between similar zero to the and one unity described and comparing for the assumption that number
372
of with loss the frequency. cumulative The probabilities value of the for loss loss is chose'sever*' selecting (i randorn assurned two losses), a second loss be chosen b/
making another ran&wn numbefllütnn
numbers and the severity of loss distribution are used to determine the size of the simulated appropnate e if the simulation process the simulated for that repcom paring it with increases the cumulative with the number probabllity of iterations fot severlty (n)r This
although procedure reasonably will generate good resu\ts are usually loss experience obtainable with iterations. n
individual losses.27 The total loss figure for the simulated year is the sum of the values of the
28 Accuracy simulation one should pen to run the simulation at least 5 000 times to achieve a sufficient Probablyr when using the accuracy- Howeverr
simulated individual losses. computer and time costs are factors that must be taken into consideratlon with Lh.s tech nique.
COMPOSITE FtNANCING STRATEGIES 373
To develop a simulated total loss distribution, the process is repeated many times and the
results of these replications are tabulated to produce a probability distribution.28 The simulated
the annual aggregate for retained loses. Figure 15.2 provides an example output of an
cumulative probability distribution is then developed and the MPY estimate is determined from
allocation produced by a simulation model.
this distribution. Another development in insurance loss modelling is the equalisation of probabilities
ofruin. While maintaining a fixed likelihood of ruin ofthe group aggregate, this functionality
Practical application of the simulation method calculates the change in aggregate contribution of a specific division, given changes in its
The synthetic loss experience produced by the simulation method in the previous section can be deductible structure. With this technology, the firm can, for example, compute any
used to calculate the simulated yearly aggregate insured losses and the simulated yearly additional aggregate contribution due by a division for a given decrease in its divisional
aggregate retained losses. 'Ihese are obtained by applying the applicable deductibles and limits deductible while maintaining the likelihood of ruin of the group aggregate at its current
to the simulated individual losses and aggregating them.
Probability distributions can be compiled for the simulated aggregate insured losses and Figure 15.1 Group simulation: Divisional allocation
simulated aggregate retained losses using the techniques described in the previous section. The STR.]CTURE MOCU
probability distributions provide a means of calculating the maximum probable yearly aggregate
insured losses and maximum probable yearly aggregate retained losses for given levels of
likelihood. These provide the firm with respective guidelines for the size of the annual aggregate
for retained losses and for the annual insurance premium for insured losses. The latter can assist
the firm greatly when negotiating an annual insurance premium with an insurer. Figure 15.1
provides an example output of a simulation model used to price the aggregate retained losses
and aggregate insured losses. The average total loss plus a multiple of the standard deviation of
total loss is used as a price
estimate.
By applying the simulation method for each division in a firm, it is possible to simulate
Figure T 5.2 Simulation model: Group simulation
individual losses for each division. The divisional deductibles can be applied to simulated
divisional individual losses and the maximum probable yearly aggregate insured losses and the
maximum probable yearly aggregate retained losses computed for each division with
accompanying probability distributions for each of insured and retained losses per division. The
ratio of aggregate insured losses within each division provides an effective ratio for allocating
annual insurance premiums to divisions. In the same way, the ratio of retained losses within each
division provides an effective ratio for allocating Orns
374
Source of figures 15.1 and 15.2: Eikos Risk Group

RISK MANAGEMENT: RISK FINANCING COMPOSITE FINANCING
Standardisation of data P, = an index value for current prices
To conclude this section on statistical analysis in risk management, it is useful to include a Pi = an index value for prices in year i.
discussion ofwhat is termed the standardisation of data. The use of historical loss data implies
Adjustment of the loss frequency data for growth in exposure and adjustment of the loss
that the economic and hazard conditions prevailing in the period over which the data was
severity data for inflation often remove the most significant distortions that could arise from
generated will continue into the future. However, the firm may have grown in size, and even if
the use of historical data. But these adjustments do not address the effect of any underlying
its operations have not altered to an extent that might give it a somewhat different hazard
trends that may be evident. It could be, for example, that the risk has become more hazardous
profile, the effect of size implies that it will generate a higher frequency of losses? Also, the
over recent years on account of, say, poor risk control (maintenance) or increasing moral
effect of inflation will magnify the cost of the given losses.
hazard. Tests for these distortions can be undertaken by searching for trends in the data
These considerations mean that the basic historical data must be standardised so that the
series, and where separate loss frequency and loss severity data is available, each should be
probability estimates will relate more correctly to the firm's current level of operations and
tested and adjusted for trend, before proceeding with convolution or curve fitting. 3i
reflect the present price levels. This standardisation can be conducted separately on both the
Another problem that arises when using frequency and severity distributions is that
frequency and severity distributions.
events that have not arisen are ignored and events that have arisen are often given too much
adjustment to loss frequency can simply be achieved by multiplying historical values
importance. The question that arises is whether the distributions have 'a missing tail'. could
by the ratio of the real value currently at risk to the real value at risk in year i:
present severe problems, particularly if the loss data is drawn from a small number of
exposure units and/or covers a relatively few number of years. %ese issues can be addressed
= fri(V/Vi)
by considering the actual loss data, frequency and severity as constituting a sample that is
where: drawn from an underlying distribution with a definite form. If the form is known, then curve
fa i = the adjusted loss frequency fitting techniques can be used to estimate the parameters of the distribution from the sample
= raw frequency for year i the real value of the property data to 'smooth out' the distributions and fill in the missing tail. 32
currently at risk V the real value of property at risk in year i.
The adjustment will provide a basis for estimating the loss frequency distribution, given the 15.3 Composite financing strategies: Transfer optimisation
current size of the exposure.30 However, this type of adjustment may not always be
appropriate. Liability claims, for example, may not bear a close relationship to the value ofthe Introduction
firm's assets, so that methods for size adjustment depend to a large extent on the nature and Chapters 12, 13 and 14 examined the mechanism ofinsurance and the aspect of
properties of the individual risk. Nevertheless, some appropriate form of size standardisation riskretention as separate means ofpre-loss financing the financial consequences of risk. With
is usually necessary. retention funding, it was shown that a risk management case exists for funding under the
Adjustment of the loss severity values for the effect of inflation simply entails the condition of transaction costs associated with the
selection of a suitable price/cost index. The adjusted severity level for a loss arising in year i is:
Trend analysis nethods are widely used and discussed in rnost statistical texts. Sevecal statlsti€at methods are avarlable- For instance, there the Daniel-Spearman rank
corelation test (a non-parametrrc tenr which tests for trend, but does provide an appropriate ulJustment
s - srt(Pc/Pi) there are regression techniques that ft B trend line to the data and thus an adjustment factor to enable propct•on of value afuture yeac_ 'n interpreting
the results ofSimple regressions. the correlation coefficient and the student t-value determine thegoodness f:t'and whether the tiend detected 'S statistically
the dl'tobuuon.
29 the firm that is hazard conditions, significant
See the eariier section in this chapte€ on che use of theoretical probability distributions to describe empirical frequency and severity data.
30The may result in fractional values of per
f to be fitted. these fractional values should 376
STRATEGIES 375
where: alternative sources of finance. In separating the investment from the financing decision, both
pre-loss means of financing need to be evaluated against each other and in turn against the
Sti = the raw severity value for loss in year i
post-loss alternatives. Ellie concern now is with the question of integration, particularly with Full insurance does not necessarily imply, therefore, that shareholders are indifferent
regard to the integration of insurance and retention funding. as to whether a loss arises, and nor does insurance exactly compensate the firm's owners
For various reasons, the firm may not wish to purchase full insurance. We have already for loss in the value of their ownership claims.36 However, full insurance serves as a yardstick
mentioned the fact that expenses incurred in the settlement of relatively low severity claims for identifying the other strategies, or risk-sharing (partial insurance) options.
may be disproportionately high when settled by insurance. Because the regularity of such losses Various forms of risk sharing are encountered in insurance contracts; this examination
often permits the firm to budget accurately for their settlement, the insurance alternative would is limited to the most common forms, namely, deductibles, coinsurance and policy limits.
imply a 'trading of dollars' for these small losses. 'The prospect of larger losses, however, A deductible is a provision for reducing the insurance settlement on any loss by a
presents the firm with a substantive risk exposure and the attendant financing problems. specified sum. These may be arranged on a 'per loss' basis or may be related to the
The solution, it would seem, is to insure high-severity losses, but retain the predictable low- accumulated losses within a period of the policy (cumulative deductibles), which permit the
severity losses. Leading on from this argument is the related question of costs; rationally there firm to establish a clear upper limit of its retained liability within any accounting period.
should be a balance such that risk sharing can be optimal. Although optimisation has been viewed Deductibles are often used to avoid the disproportionately large settlement costs for
traditionally from the standpoint of expected utility, where the problem is defined as deriving small claims. Depending on the risk characteristics ofhigher retentions, the premium
the optimal level of risk retention (through co-insurance or deductibles) so as to maximise the discount for an increased deductible and the costs of competing forms of finance, a firm
expected utility of wealth, the optimisation question 33 may also be more simply regarded as one may even consider very substantial deductibles.37
of economic order quantities. Co-insurance implies that the insurer pays an agreed proportion of the ACV of the loss,
leaving the insured to bear the residual proportion. Ihe premium for proportional co-
insurance is usually scaled proportionately to the full insurance premium. Unlike the case of
Risk sharing through partial insurance deductibles, the firm faces a somewhat open-ended retention liability, but opportunities
Before considering composite financing strategies, however, it is important to first consider the may be explored by balancing the risk characteristics of the retained distribution, the
available forms of insurance and partial insurance. Full insurance provides compensation for the premium discount and the costs of alternative financing methods.
full value of loss as defined in the insurance contract or policy. rlhe compensation is for the cost Ihe imposition ofpolicy limits means that the basis of indemnity under the insurance
of replacement or the repair of destroyed or damaged property with similar materials and contract limits the liability of the insurer to a value that is less than the ACV of the property.
within a reasonable time.34 For interruption loss, the methods for calculating loss follow rules In the case of liability insurance, although the suits brought against the insured are open-
designed to measure the reduction in profitability from the level that might have prevailed had ended, the liability policy will invariably specify an upper limit to the settlement. Policy limits
the loss not occurred. The insurance settlement is thus led by the principle of indemnity: the are generally considered
insurer will restore to the insured the financial loss from the insured event and will usually limit
recovery to the 'actual cash value' (ACV) of the loss.
35 Equrty toss ps not determined by theactual cash value Ofthe propaty. but by its capacity for gene•ating wealth In productive use- Similarly.
The ACV is not, however, equivalent to the economic loss suffered by the owners of the thecase ofa liability claim, the equity value may not be related only to the cost of defending and satisfying litigation. Litigation may affect the abilrty to successfully market
its product, raise capital. etc
firm, for economic loss relates to the loss of the value of equity
36The imptlcauonofthis statement as regards the risk fiWügemenc activity of the firm is not difi:utr to perceive. It places further emphasls on the liskcontrol activtcy
and identifiesa furthe' ekrnent of cost to be•accountedfor in the total cost of risk equation. as discussed in chapter 9.The notion ofCost. resulting from the
poss'bledifference between actual cashvalue and economjc loss, be &dded to the cost of rnsurers have been 'eluctant tooffercontracrs Wilh substantial deductibles-
Tbls may be partly because of the diffcultv in pricing contracts tequirlng anest•rnatlon of thetail ofthelossdistribution) and because ofa reluctance to part •Mth a
substantiat portion of their OtefTIium incomethat would generate Investrnent in-_ome Deductibles are. on the other hand. imposed by rnsurevs, especially when
33 insurance models based on utility theory are discussed late, In this chapter. Certain considerations limit their use in management this there is the "Rention of restorng rncentivesfot the insured to undertake losspreventi0fVreduct;on rtkasures Recentw. however. there has been evidence firms
limitation has prompted the consideration of the from a different perspective - based
increasing therr letention levels {Doherty. T985: 331
34 legal In incurred of liability in defending theclaim, will be made good b,rthe insurer.ofthe settlement
378
STRATEGIES 377
unattractive for the business firm, since the retained loss is open-ended and diffcult to budget
as a result of the loss.35 Ownership claims of shareholders represent the rights to income
for, and the uninsured component of the loss lies at the tail of the loss distribution — which is
generated from corporate investment and the rights to share in the residual value of the
where insurance is most beneficial.
firm. The valuation ofthese rights is determined by capital market activity, and this fact is
Figure 15.3 graphically depicts the effects of the abovementioned forms of risk sharing. A
not considered in the design of an insurance policy.
specimen distribution of ACV is shown, which represents the distribution of ACV suffered by a
Deductible

firm if no insurance were in place. The distribution could then be segmented into the ACV
distribution of retained loss for the firm and the distribution of contingent payments made by D
the insurer with the various forms of risk sharing. (b)
Axes: pr — Probability
The effect of a deductible is illustrated in figure Losses below the deductible (D) are
ACV - Actual cash value
retained (line (i)). The dotted line (ii) shows the distribution of insurance payments for
pr
contingent losses; this distribution follows the vertical axis showing a probability close to zero
of insurance payments that arise if prospective losses turn out to be no greater than D.
Figure 15.3 (c) shows that co-insurance compresses the range of potential values of ACV
retained loss toward the vertical axis. The distribution of policy payments would also be
Proportionateco-insurance Policy limit
compressed toward the vertical axis in the same fashion. If a 50% proportionate co-insurance
were arranged, the distribution of policy payments would be identical to that of retained loss.
The impact of a policy limit (OL) is illustrated in figure 15.3(d). The limit considered in this
illustration is a cumulative limit on payments made by the policy in a specified period. The
distribution of retained loss shows: a high probability of no retention (shown by line (i) following
the vertical axis) the open-ended nature of prospective losses whereby the distribution
converges on but does not reach the horizontal axis.
Conversely, the distribution of policy payments (dotted line (ii)) shows the limited value of the
insurer's payments, OL. ACV
i)
In comparing the forms of retention, it is commonly held that deductibles are more effcient
devices for reducing risk than are proportionate devices such as co-insurance (or quota share o
reinsurance). Although this case has been motivated simplistically, evidence shows that a risk (c) o
averter would prefer a policy with a deductible to one with the same expected payout, but (0 Distribution Of retained loss
having the risk divided on a proportionate basis.38 (ii) Distribution Of policy payments Source: Doherty
(1985:330)
Evaluation of alternative financing sources

In considering a composite financing strategy, which could include retention (through
deductibles), insurance and outside finance, one would have to consider and appraise the
38 0985, 339-42k
STRATEGIES 379
associated relative costs. In respect of outside financing (debt and equity), the transaction
costs, which would include issue
Figure 15.3 Effect of deductible, proportionate co-insurance, and policy limit on hypothetical ACV
distribution
Loss distribution
RISK MANAGEMENT:
380 RISK F!NANCING
The respective costs of each form of financing are represented in figure 15.4. Each diagram shows the total actual pre
expenses and underwriting costs, must be allowed for. In respect of internal liquid funds, one would
the cheapest form of financing when losses are relatively low (avoiding transaction costs). However, it is also evident
have to consider the opportunity costs of maintaining these fiulds for risk management purposes.
within a cash management programme. Liquidity limitations would cause the firm to access either the insurance marke
The inclusion of such transaction costs brings a greater opportunity for satisfying the condition for insurance and retention. In
relatively high severity values.
order to explore the possibilities of a composite programme the competing sources are separated and their cost structures considered.
Table 15. I summarises the financing costs from different sources.
Figure T 5.4 Present value of financing costs from alternative sources
Table 15.1 Examples of financing costs from different sources39
Internal liquid resources
Source: Doherty (1985:348)

1. New insurance
2. New issue + k7)-l
The following can be noted: Insurance New issue

•
*450
Financing losses from internal sources avoids the issue costs of external funding and
the premium loadings associated with insurance. Assuming losses of size L, the cost
of financing such losses internally is L, and the present value is the expected financing
cost discounted at the appropriate risk-adjusted rate (1%).
• The insurance premium is represented by (1 + where E(L) is the expected value of the
Cins
loss and a represents the mark-up or transaction cost. This form of premium is implied
by rating methods that calculate the premium at a constant rate applied to some
measure of exposure. Since the premium is paid in advance, no discounting is
required.
Axes:
• For new debt and equity, the main transaction costs are incurred with issue (shown as
L - Losses
b), and since issue costs typically increase modestly with the size of the issue, the cost R — Cost of financing
is represented as a linear function of the loss, where c < a, revealing the realistic
situation where insurance transaction costs are higher than the variable costs of a new
issue. lhe present value of expected financing cost is obtained by
discounting at some rate ki that is appropriate to the riskiness ofthe contingent
financing costs. For the sake of simplicity, this rate could be the (opportunity) rate
used for internal financing, which is calculated in the usual way: Source: Doherty (1985:348)
39 0. b. and c are positive constants With a -pc, k' and are rsk-adjusted d.qount rates; and P is the mobatil•ry that a loss that is above threshold level ofA will anse-
382 Conclusion
A comparison ofthe approaches maybe undertaken by means ofrisk-adjusted present values,
Supposing that small losses, below some value L = d, can be financed from liquid sources where the decision-making process could adopt the following sequence: Select a first layer of
(i.e. the adoption of a cumulative deductible d), one could combine the sources into a deductible to be financed by internal resources.
COMPOSITE FINANCING STRATEGIES 383
composite financing scheme, where the financing is layered. Losses above d are funded by
insurance or by the new issue of debt or equity. If insurance is purchased, the premium
would be (I + a) times the expected value of the insured loss, which can be represented as Choose between insurance and new issues on the basis of the expected value of financing costs.
follows:
Naturally, this process presupposes that the aggregate loss distribution is known or accurately
estimated by statistical techniques discussed in the earlier section; and that the statistical
properties of the loss distribution, namely the expected value, the standard deviation and the
Figure 15.5 shows the total cost for different values of loss, as OADG, when small losses are
measure of skewness, are computed. With these parameters, and with knowledge of the
retained internally and losses in excess of d are insured — and OABF, when small losses are
respective transaction costs and risk-adjusted discount rates, the financing strategy can be
retained internally, but losses in excess of d are funded through new issues. Ihe total cost of
evaluated and a decision made on the form of the second layer of funding, given a first retention
financing through new issue is shown by the continuation along line BF; the difference in
layer in the form of a suitable deductible.
gradient between BF and 450 represents the variable transaction cost.
There remains, however, the selection of the retention portion or deductible level, which,
as the decision process indicates, should be given first consideration. The size ofthis layer will be
Figure 15.5 Composite two-layer financing programme
determined by its risk characteristics. More specifically, this layer should finance only losses or
450
portions of losses that arise with sufficient predictability so that they can be accommodated
G within the firm's liquidity or cash budget.

The method to be used may be developed by using the characteristics of the loss
distribution. The deductible d can be chosen at some level (say a probability level of0,25) where
the probability ofsurpassing d (ruin probability) is acceptably low. If the probability is fLxed too
low, implying that losses might well fall below this level (L < d), the firm will find itself with surplus
cash, which is costly in terms of investment earnings foregone. Finally, to gauge whether
payments financed in this way (i.e. under the deductible at the appropriate probability level) are
stable, the expected value and standard deviation of the prospective payments can be
determined.
"Ihe comparison of the risk-adjusted costs of financing will determine whether the optimal
strategy dictates the integration ofretention with insurance (OADG), or the integration of
Layer 1 Layer 2 retention with new issue financing (OABF).34
OABF: Total cost with deductible (d) financed by intemal liquid resources and residual of larger losses financed by new
issue
OADG: Total cost with deductible (d) financed by intemal liquid resources and residual of larger losses financed by
insurance policy subject to deductible (d)
15.4 Optimisation models based on expected utility
The above methodology, used to provide guidance for deductible selection (the risk retention
Source: Doherty (1985:349)
level), implied selection in a way that contains the riskiness of the retained risk at some
34
The methodology is well illustrated by Doherty ( 935: 35 -5) means Of a case study.
1 T
553-68). Doherty and ( 1983:555-65} and Cozotino {1978:449-71
See %Tlith (1968:69-77).

acceptable level. Several other types of insurancepurchasing models exist that may be found in (ii) the decision maker's loss outcomes utility function, according i.e. the to his/her rule that
preferences, assigns utility must valuesbe
the risk management literature." These models mainly use expected utility to derive the
to monetary available.
optimum level of retention through the adoption of co-insurance or deductibles.
384 42 Refeence can be made to 3. where the concept 01 utitity was introduced and a utility-b3*d decision for the purchaseof
431he propositionsconcaning the purchase ofinsurance derived from this type ofmodelassume aversionto risk riskmanagement decision
makets. so that, for example. If the premium than equal to or less than the of expected the Insured valueof loss the loss. rational (rationally) risk avettet insurance will will
retain be part purchased. the risk'f the: a
Utility theory has been theoretically recognised as a decision-making technique to facilitate
premium is constant multi* 1 cf tre expected deduct,ble will be preferred to If individual A is sore risk averse than Individual then A will more insurance tha a &
the evaluation of probabilistic alternatives. The method assigns a single value, the expected COMPOSITE FINANCING STRATEGIES 385
utility, to each probability outcome. The expected utilities are compared and the largest expected
utility is the preferred alternative.42
In the latter requirement lies a major limitation.
In order to compute the expected utility, each possible monetary outcome in the probability
Although most ofthe theoretical problems involving utility functions have been solved,
distribution needs to be converted to a utility value. These utility values are numbers that express utility theory has not been applied in many practical situations; theorists have been unable to
the preference of the decision maker for various amounts of money. Since the decision maker is convince practitioners that the method has merit: to them, the theory seems abstract and even
assumed to be risk averse, large losses would be assigned lower utilities than smaller losses; the abstruse"
utility curve therefore displays diminishing returns. In sum, the following problems limit the use of utility theory for risk management purposes:
In the case of risk management, utility values could be assigned to each possible outcome of The identification of precise risk management strategies means full identification of the
a loss distribution and would then be weighted by multiplying them by the probabilities decision maker's utility function. It is extremely doubtful that the firm (i.e. its board) can
ofoccurrence ofthe associated outcomes. The expected utility for the distribution is obtained by
expound a utility function or an attitude toward risk that satisfies the propositions of the
the summation of the weighted utility values. In considering an optimal level for insurance, the
utility decision models developed.
expected utility typically takes the following form:
Even if the utility function were availabley the appropriate utility is that of the firm's
Expected utility shareholders, who regard their equity holdings as financial assets traded in a capital market
according to market-determined risk-return relationships. Since the risk is clearly priced in
where:
the market (assuming a widely held firm), it is consistent to impose the market price of risk
W = the decision maker's wealth if a loss of size Li occurs (the wealth level is
determined by the level of insurance [INS)) P = the probability that a loss of size Li will on risk management decisions, i.e. risk preferences are measured by capital market activity.
occur U = an index of the decision maker's utility. Virtually all the utility-based models formulate the insurance decision as one in which
insurable risk is the only source of risk. Hence, possible portfolio effects within a firm are
The problem (optimisation) is therefore to maximise expected utility ofwealth, where wealth ignored. The risk-reducing effects of insurance may lose some value if shareholders are able
depends on loss experience (which is random) and the purchase of insurance (which is a
to diversify away this risk in their personal portfolio holdings.
decision)."
The expected utility approach can clearly be useful in helping to identify broad risk
management choices consistent with aversion to risk and may help 1 5.5 Summary
to eliminate specific choices from a decision. The approach presupposes that two fundamental Having discussed in chapters 1 2, 13 and 14 the risk-financing means of insurance and
requirements are satisfied: retention funding the objective of this chapter has been to describe composite financing
strategies. As this approach requires a statistical analysis to be performed, it was necessary
(i) the probability distribution of losses must be known, or at least accurately estimated, and
to begin the chapter by exploring further the use of statistical analysis in risk management.
The section covering statistical analysis concentrated particularly on the estimation of the
maximum probable yearly aggregate loss (MPY), but also drew a distinction between the MPY
and the probable maximum loss (PML) measure. Two broad methods of estimating the MPY
were discussed. One is based on loss distribution statistics, which incorporates the methods
of normal approximation, Chebyshev and the normal power. The other is based on toss
distribution Basel Committee on Banking Supervision
estimation, where the analytical tabulation approach and simulation method were examined.
Allen, PM 1943. The First Hundred Years Sound Practicesfor the Management
44 Cummlns and Freifelder (19709.44}.
1843-1943. Philadelphia: Mutual Fire and Supervision ofOperational Risk. Basel:
386 Insurance Company. Bank for International Settlements.

Allen, TC 1972. Quantitative risk Barry JT 1918. lie Workmen; Compensation
management methodology: A historical Act. Johannesburg: Law and General.
perspective.' Journal of Commerce, March. Bawcutt> PA 1982. Captive Insurance
The chapter proceeded by considering composite financing 5trategies. Firstly, we Allen, TC and Duvall, RM 1971. Companies: Establishment, Operation and
theoretical and practical approach to Management. Cambridge: Woodhead-
focused on risk sharing through partial insurance, graphically depicting the effects of
risk management.' American Society of Faulkner.
a deductible, proportionate co-insurance and policy limits on hypothetical actual cash Insurance Management. Bawcutt, PA 1997. Captive Insurance
value distribution. It was stated that a risk averter would commonly prefer a Allianz 1987. Allianz Handbook ofLoss Companies. London: Witherby.
deductible as a device for reducing risk to having the risk divided on a proportional Prevention. Berlin and Munich: Allianz Beard, RE, Pentikäinen, T and Pesonen, E
basis. Versicherungs AGO Risk Theory. New York John Wiley.
Secondly alternative external financing sources, i.e. new issues of debt and equity, Aon Global Risk Consulting 2007. Global Benatary SR, Metz, yr and Elmes, pc 1983.
were evaluated against insurance and internal liquid resources. The present values of 500: A Captive Insight. London: Aon. 'Occupational health and safety in South
Arrow, KJ 1965. 'Insurance, risk and resource Africa.' South African Medical Journal.
financing costs of these sources were separately depicted and then combined to
allocation.' In Aspects ofthe 'Iheoty of Berliner, B 1985. 'Large risks and limits of
show the total cost of financing under a composite twolayer financing programme
Risk-Bearing. Reproduced in Dionne, G insurability: Geneva Papers on Risk and
that entails financing losses though internal liquid resources and insurance, or and Harrington, SE 1992. Foundations Insurance, Vol. 10,
internal liquid resources and a new issue of debt or equity. ofInsurance Economics: Readings in Berlonghii A and Mattman, JW 1987. 'A new
It was concluded that a comparison of the risk-adjusted costs offinancing will Economics and Finance. Boston: Kluwer understanding of risk management: Risk
determine whether the optimal strategy dictates the integration of retention with Academic. Management, April.
insurance, or the integration of retention with new issue financing. Athearn, JL and Pritchett, ST 1984. Risk and Bernstein, PL 1996. Against the Gods: Ike
In determining an optimal risk-financing strategy involving risk retention and Insurance. St Paul: West Remarkable Story ofRisk. New York: John
Atkins, NG 1981. 'Loss ofprofits insurance: Wiley.
insurance, utility theory has been recognised as a decision-making technique to
Businessman's Law, Vol. I l, Bickelhaupt} PL and Magee, JH 1970. General
facilitate the evaluation of probabilistic alternatives. The chapter briefly describes
Insurance. Homewood: Irwin.
the problem of maximising the expected utility of wealth and the fundamental Biger, N and Kahane, Y 1978. 'Risk
requirements it presupposes, and enumerates the limitations of this theory. From a Baglini, NA 1976. Risk Management in the considerations in insurance rate-maläng.'
practical risk management decision-making point of view, it is concluded that it is International Firm. New York: Risk Studies Journal ofRisk and Insurance, Vol. 45.
extremely doubtful whether the firm (i.e. its board) can expound a utility function Foundation. Blake, RP 1963. Industrial Safety. Englewood
that satisfies the propositions relative to the decision models developed. Baker, M 1986. 'Legislation on occupational Cliffs: Prentice-Hall.
387 health in South Africa: South African Boberg, POR 1984. The Law ofDelicf. Cape
journal ofContinuing Medical Education , Town: Juta.
Bakeri M and Coetzee, AC 1983. An Bridges RS and Jacobs, GG 1986.
Introduction to Occupational Health 'Ergonomics: Ergo systems and
Nursing in South Africa. Johannesburg: occupational health.' South African Journal
Wits University Press. Of Continuing Medical Education.
Bibliography

Risk Management Exam

Uploaded by

Copyright:

Available Formats

You might also like

Risk Management Exam

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Risk Management Exam

Uploaded by

Copyright:

Available Formats

Risk Management

e in text: the authors

Prof Mervyn E King SC

Chapter 1 Principles of management applied to managing risk 2

Part Ill: Risk financing 220

1 .1 Introduction 1.2 Case for risk management

Use of medical drug widely prescribed babies were born in 20

in the babies of women who compensation.

1960: A coal-mining disaster 435 miners were

High rates of loss of life 1 500 people were killed

in a shanty town that respiratory tracts and

had been built around nervous systems.

the factory. 200 000 people had to leave

Welding operations 177 miners died in the

ignited polyurethane gas„filled mine.

An explosion, resulting 165 men were killed, the

The property of the 17 people were killed

A locomotive feti down a 104 people were killed.

Increasing concentration of risk 1.3 Reasons to manage risk

Justification of specific decisions 9

aspects of the risk management programme. One finds

The Risk evaluation and assessment

following six basic functions: or megafirm.

2.3 Definitions of risk

2.4 Basic risk classifications

Risk and uncertainty

(i) uncertainty whether an event will occur; and

Table 2.1 The certainty-uncertainty continuum

Level of uncertainty Characteristics Examples

None (certainty) Outcomes can be predicted Physical laws, with

precision natural sciences

probabilities are unknown investments Level 3: Total uncertainty Outcomes

are not fully

Space exploration, identified and probabilities genetic

Source: Adapted from Williams, Smith and Young (T 998)

Definition Insurable and non-insurable risks

Definition Systematic risk

ærcelve as social hence the liability crisis,

Data availabk at httpY/data (accessed 26 October

Integral to a company's risk management is the identification of strategic risk including

finance personnel and

2.6 Manageriat risk classifications Inherent risks

The management of systematic risk entails a repositioning on the riskreturn curve. A

iötventy takes •nto consideration minimum statutory surpluses

Insureds' rationale for purchasing insurance

Value of the house

Probability of the house being totally destroyed by fire 0,20

2.10 Economic viability of insurance markets

36The of Mllity is discussed 10 greater detail in chapter 3.

3.2 Expected monetary value (EMV) criterion

These facts can be set out in a states of nature matrix.

Outcomes (R50 000 000) 0

Probability otoool 059999 1

The binomial distribution

formula for the binomial distribution is:

r!(N— r)! V(I —p)N-r where:

p(r) = probability that exactly r accidents will take place

N! is called the factorial 0! is

Figure 3.2 A binomial distribution