A study of Credit Scoring Methods used by banks

SUBMITTED BY

Krutika Bajaj

UNDER THE GUIDANCE OF

Assist. Prof. Trupti Naik

A PROJECT SUBMITTED IN PARTIAL FULFILMENT OF

MMS TO

VIDYALANKAR INSTITUTE OF TECHNOLOGY

Wadala (East), Mumbai 400 037

April 2020
 A study of Credit Scoring Methods used by banks

Submitted By

Krutika Bajaj

UNDER THE GUIDANCE OF

Assist. Prof. Trupti Naik

A PROJECT SUBMITTED IN PARTIAL FULFILMENT OF

MMS TO

VIDYALANKAR INSTITUTE OF TECHNOLOGY

Wadala (East), Mumbai 400 037

April 2020

Signature of Faculty Guide Head of Department

 Declaration
I hereby declare that the study presented by me to Vidyalankar Institute of Technology School
of Management on the topic “A study of Credit Scoring Methods used by banks.” is
completed and submitted under the guidance and supervision of Assist. Prof. Trupti Naik.

Krutika Bajaj
 Acknowledgement
I take this opportunity to express my profound and sincere gratitude to Vidyalankar
Institute of Technology for providing me with avenues to explore the corridors of the
corporate world and gather invaluable knowledge and practical experience via Final
Project.

I take the privilege of offering a deep sense of gratitude to Assist. Prof. Trupti Naik,
Department of Management, VIT for providing me their guidance and inspiration to complete
the Final Project Report.

Krutika Bajaj
 Table of Content
Particulars Page No.
Sr. No.
Abstract
1.
1.1 Objective 1
Introduction
2.
2.1. Country wise credit scoring 4
2.2. credit scoring 15
2.3. Background of credit scoring 16
2.4. Difference between credit scoring and credit rating 16
2.5 How scoring model works? 17
2.6 Benefits of credit scoring 18
2.7 What influences credit scoring? 20
2.8 How are credit scoring models calculated? 21

Literature Review and Methodology

3&4
3.1. About credit scoring 21
3.2. Credit scoring procedures and techniques 23
Methodology
4.1 Statement of problem 25
4.2 what are the types of credit scoring 25
4.3 Rules based credit scoring methodology 27

Findings, Conclusion, Example & Bibliography

5. 39
5.1.Conclusion
40
5.2. Current credit scoring procedures
41
5.3. Example
44
5.4. Final note
45- 48
5.5. Bibliography
 1. Abstract
This study describes credit scoring techniques used for classifying risks and granting credit to
applicants in India. Loan products offered by commercial banks are from several kinds, since
housing loans, personal loans and business loans until education loans or vehicle loans, among
many other types. All loan products are categorized either as secured or unsecured loans. Credit
scoring techniques used for both secured and unsecured loans are broadly divided into two
categories: Advanced Statistical Methods and Traditional Statistical Methods. Some
methodologies are presented to discuss Indian situation and the different credit scoring methods
for personal finance used by commercial banks in India.

1.1 Objectives
The main objectives of the present study are the following ones:
To know the different credit scoring methods for personal finance available and used by
commercial banks.

1. Introduction
Initial Considerations
The human History allows well to understand the risks of lending money. Borrowing and
lending have long history related to human behavior. A credit risk component may be
associated with lending transactions, in order to reduce credit risk. Credit scoring methods
may be used to assess the credit worthiness of borrower. The information filled in the
personal finance application form is used to develop a numerical score for each applicant and
these scores may be used to discriminate on bad and good loans. Credit scores are broadly
classified into two categories based on the method used to obtain scores, i.e. deductive or
judgmental credit scoring and empirical or statistical credit scoring

Credit Scoring
Credit evaluation is one of the most crucial processes in banks credit management decisions.
This process includes collecting, analyzing and classifying different credit elements and
variables to assess the credit decisions. The quality of bank loans is the key determinant of
competition, survival and profitability. One of the most important kits, to classify bank
customers, as a part of the credit evaluation process to reduce the current and the expected risk

1
of a customer being bad credit, is credit scoring. Hand & Jacka stated that “the process of
modeling creditworthiness by financial institutions is referred to as credit scoring”. It is also
useful to provide further definitions of credit scoring. Anderson (2007) suggested that to define
credit scoring, the term should be broken down into two components, credit and scoring.
Firstly, simply the word credit means “buy now, pay later”. It is derived from the Latin word
“credo”, which means “I believe” or “I trust in”. Secondly, the word “scoring” refers to “the use
of a numerical tool to rank order cases according to some real or perceived quality in order to
discriminate between them, and ensure objective and consistent decisions”. Therefore, scores
might be presented as “numbers” to represent a single quality, or “grades” which may be
presented as “letters” or “labels” to represent one or more qualities. Consequently, credit scoring
can be simply defined as “the use of statistical models to transform relevant data into numerical
measures that guide credit decisions. It is the industrialization of trust; a logical future
development of the subjective credit ratings first provided by nineteenth century credit bureau,
which has been driven by a need for objectives, fast and consistent decisions, and made possible
by advances in technology”. Furthermore, “credit scoring is the use of statistical models to
determine the likelihood that a prospective borrower will default on a loan. Credit scoring
models are widely used to evaluate business, real estate, and consumer loans” . Also, “credit
scoring is the set of decision models and their underlying techniques that aid lenders in the
granting of consumer credit. These techniques decide who will get credit, how much credit they
should get and what operational strategies will enhance the profitability of the borrowers to the
lenders”

A credit score is a numerical expression based on a level analysis of a person's credit files, to
represent the creditworthiness of an individual. A credit score is primarily based on a credit
report, information typically sourced from credit bureaus.

Lenders, such as banks and credit card companies, use credit scores to evaluate the potential risk
posed by lending money to consumers and to mitigate losses due to bad debt. Lenders use credit
scores to determine who qualifies for a loan, at what interest rate, and what credit limits. Lenders
also use credit scores to determine which customers are likely to bring in the most revenue. The
use of credit or identity scoring prior to authorizing access or granting credit is an
implementation of a trusted system.

2
 Credit scoring is not limited to banks. Other organizations, such as mobile phone companies,
insurance companies, landlords, and government departments employ the same techniques.
Digital finance companies such as online lenders also use alternative data sources to calculate the
creditworthiness of borrowers.

Contents

 2.1. By country
o 2.1.1Australia
o 2.1.2Austria
o 2.1.3Brazil

o 2.1.4Canada
o 2.1.5China
o 2.1.6Denmark
o 2.1.7Germany
o 2.1.8India
o 2.1.9Norway
o 2.1.10South Africa
o 2.1.11Sri Lanka
o 2.1.12Sweden
o 2.1.13United Kingdom
o 2.1.14United States
o 2.1.15United Arab Emirates

By country

2.1.1 Australia

In Australia, credit scoring is widely accepted as the primary method of assessing

creditworthiness. Credit scoring is used not only to determine whether credit should be approved
to an applicant, but for credit scoring in the setting of credit limits on credit or store cards, in
behavioral modelling such as collections scoring, and also in the pre-approval of additional credit
to a company's existing client base.

3
Although logistic (or non-linear) probability modelling is still the most popular means by which
to develop scorecards, various other methods offer powerful alternatives,
including MARS, CART, CHAID, and random forests.

Prior to 12 March 2014 Veda Advantage, the main provider of credit file data, provided only a
negative credit reporting system containing information on applications for credit and adverse
listings indicating a default under a credit contract. Veda was acquired by Equifax in Feb
2016, making Equifax the largest credit agency in Australia.

With the subsequent introduction of positive reporting, lending companies have begun an uptake
of its usage with some implementing risk based pricing to set lending rates.

2.1.2 Austria

In Austria, credit scoring is done as a blacklist. Consumers who did not pay bills end up on the
blacklists that are held by different credit bureaus. Having an entry on the black list may result in
the denial of contracts. Certain enterprises including telecom carriers use the list on a regular
basis. Banks also use these lists, but rather inquire about security and income when considering
loans. Beside these lists several agencies and credit bureaus provide credit scoring of consumers.

According to the Austrian Data Protection Act, consumers must opt-in for the use of their private
data for any purpose. Consumers can also withhold permission to use the data later, making
illegal any further distribution or use of the collected data. Consumers also have the right to
receive a free copy of all data held by credit bureaus once a year. Wrong or unlawfully collected
data must be deleted or corrected.

2.1.3 Brazil

Credit scoring is relatively new in Brazil. Previously, credit reporting was done as a blacklist and
each lender used to assess potential borrowers on their own criteria. Nowadays, the system of
credit reports and scores in Brazil is very similar to that in the United States.

A credit score is a number based on a statistical analysis of a person's credit information, which
represents the creditworthiness of that person. It is the most important tool used by financial
institutions during a credit analysis that aims to assist the decision-making process of granting
credit and conducting business, in order to verify the likelihood that people will pay their bills. A
credit score is primarily based on credit report information, typically from one of the three major

4
credit bureaus: Serasa Experian, Boa Vista (previously Equifax do Brasil) and SPC Brasil.

There are different methods of calculating credit scores in Brazil. In general, scores range from 0
to 1000 indicating what is the chance of a certain profile of consumers paying their bills on time
in the next 12 months. The score is calculated from several factors, but practically it analyzes a
person's trajectory as a consumer, what includes up to date payments of bills, history of negative
debts, financial relationships with companies and updated personal data on credit protection
agencies, such as Serasa Experian, Boa Vista, SPC, Quod and Foregon.

2.1.4 Canada

The system of credit reports and scores in Canada is very similar to that in the United States and
India, with two of the same reporting agencies active in the country: Equifax and TransUnion.
(Experian, which entered the Canadian market with the purchase of Northern Credit Bureaus in
2008, announced the closing of its Canadian operations as of 18 April 2009).

There are, however, some key differences. One is that, unlike in the United States, where a
consumer is allowed only one free copy of their credit report a year, in Canada, the consumer
may order a free copy of their credit report any number of times in a year, as long as the request
is made in writing, and as long as the consumer asks for a printed copy to be delivered by
mail. Borrowell and CreditKarma offers free credit report and credit check and this request by
the consumer is noted in the credit report as a 'soft inquiry', so it has no effect on their credit
score. According to Equifax's ScorePower Report, Equifax Beacon scores range from 300 to
900. Trans Union Emperica scores also range from 300 and 900.

The Government of Canada offers a free publication called Understanding Your Credit Report
and Credit Score. This publication provides sample credit report and credit score documents,
with explanations of the notations and codes that are used. It also contains general information
on how to build or improve credit history, and how to check for signs that identity theft has
occurred. The publication is available online at the Financial Consumer Agency of Canada.
Paper copies can also be ordered at no charge for residents of Canada.

2.1.5 China

Private companies have developed credit score systems, these systems include Sesame Credit,
which is provided by Alibaba affiliate Ant Financial, and Tencent Credit. Part of the

5
government's Social Credit System uses credit information about citizens.

2.1.6 Denmark

The credit scoring is widely used in Denmark by the banks and a number of private companies
within telco and others. The credit scoring is split in two:

 Private: The probability of defaulting

 Businesses: The probability of bankruptcy

For privates, the credit scoring is always made by the creditor. For businesses it is either made by
the creditor or by a third party.

There are a few companies who have specialized in developing credit scorecards in Denmark:

 Experian (generic rating for business)

 Bisnode (generic rating for business)

The credit scorecards in Denmark are mainly based on information provided by the applicant and
publicly available data. It is very restricted by legislation compared to its neighbouring countries.

2.1.7 Germany

In Germany, credit scoring is widely accepted as the primary method of assessing

creditworthiness. Credit scoring is used not only to determine whether credit should be approved
to an applicant, but for credit scoring in the setting of credit limits on credit or store cards, in
behavioral modelling such as collections scoring, and also in the pre-approval of additional credit
to a company's existing client base.

Consumers have the right to receive a free copy of all data held by credit bureaus once a year. At
present Schufa, the main provider of credit file data, provides scores for about three-quarters of
the German population.

2.1.8 India

In India, there are four credit information companies licensed by Reserve Bank of India. The
Credit Information Bureau (India) Limited (CIBIL) has functioned as a Credit Information
Company from January 2018. Subsequently, in 2010, Experian, Equifax and CRIF High
Mark were given licenses by Reserve Bank of India to operate as Credit Information Companies

6
in India.

Although all the four credit information companies have developed their individual credit scores,
the most popular is CIBIL credit score. The CIBIL credit score is a three digit number that
represents a summary of individuals' credit history and credit rating. This score ranges from 300
to 900, with 900 being the best score. Individuals with no credit history will have a score of -1. If
the credit history is less than six months, the score will be 0. CIBIL credit score takes time to
build up and usually it takes between 18 and 36 months or more of credit usage to obtain a
satisfactory credit score.

2.1.9 Norway

In Norway, credit scoring services are provided by three credit scoring agencies: Dun &
Bradstreet, Experian and Lindorff Decision. Credit scoring is based on publicly available
information such as demographic data, tax returns, taxable income and
any Betalingsanmerkning (non-payment records) that might be registered on the credit-scored
individual. Upon being scored, an individual will receive a notice (written or by e-mail) from the
scoring agency stating who performed the credit score as well as any information provided in the
score. In addition, many credit institutions use custom scorecards based on any number of
parameters. Credit scores range between 300 and 999.

2.1.10 South Africa

Credit scoring is used throughout the credit industry in South Africa, with the likes of banks,
micro-lenders, clothing retailers, furniture retailers, specialized lenders and insurers all using
credit scores. Currently all four retail credit bureau offer credit bureau scores. The data stored by
the credit bureaus include both positive and negative data, increasing the predictive power of the
individual scores. TransUnion (formerly ITC) offer the Empirica Score which is, as of mid-2010,
in its 4th generation. The Empirica score is segmented into two suites: the account origination
(AO) and account management (AM). Experian South Africa likewise has a Delphi credit score
with their fourth generation about to be released (late 2010). In 2011, Compuscan released
Compuscore ABC, a scoring suite which predicts the probability of customer default throughout
the credit life cycle. Six years later, Compuscan introduced Compuscore PSY, a 3-digit
psychometric-based credit bureau score used by lenders to make informed lending decisions on
thin files or marginal declines.

7
2.1.11 Sri Lanka

According to the provisions of Credit Information Bureau Act No 18 of 1990 (as amended by
Act No 42 of 2008), CRIB has been delegated with power to issue credit reports to any subject to
whom that information is related to. The Bureau commenced to issue Self Inquiry Credit Reports
in December 2009.

2.1.12 Sweden

Sweden has a system for credit scoring that aims to find people with a history of neglect to pay
bills or, most commonly, taxes. Anyone who does not pay their debts on time, and fails to make
payments after a reminder, will have their case forwarded to the Swedish Enforcement
Authority which is a national authority for collecting debts. The mere appearance of a company,
or government office, as a debtor to this authority will result in a record among private credit
bureaus; however, this does not apply to individuals as debtors. This record is called a Betalnings
an märkning (non-payment record) and by law can be stored for three years for an individual and
five years for a company. This kind of nonpayment record will make it very difficult to get a
loan, rent an apartment, get telephone subscriptions, rent a car or get a job where you handle
cash. The banks, also use income and asset figures in connection with loan assessments.

If a person gets an injunction to pay issued by the Enforcement Authority, it is possible to

dispute it. Then the party requesting the payment must show its correctness in district court.
Failure to dispute is seen as admitting the debt. If the debtor loses the court trial, costs for the
trial are added to the debt. Taxes and authority fees must always be paid on demand unless
payment has already been made.

Every person with a Swedish national identification number must register a valid address, even if
living abroad, since sent letters are considered to have been delivered to that person once they
reach the registered address. As an example, Swedish astronaut Christer Fuglesang got
a Betalningsanmärkning since a car he had ordered, and therefore owned, passed a toll station for
the Stockholm congestion tax. At the time, he was living in the USA training for his first Space
Shuttle mission and had an old invalid address registered for the car. Letters with payment
requests did not reach him on time. The case was appealed and retracted, but the non-payment
record remained for three years since it could not be retracted according to the law.

8
2.1.13 United Kingdom

Credit scoring in the United Kingdom is very different to that of the United States and other
nations. There is no such thing as a universal credit score or credit rating in the UK. Each lender
will assess potential borrowers on their own criteria, and these algorithms are effectively trade
secrets.

"Credit scores" which are available for individuals to see and provided from Credit Reference
Agencies such as Call Credit, Equifax, Experian and TransUnion are marketed to consumers and
are not usually used by lenders. Most lenders instead use their own internal scoring mechanism.

The most popular statistical technique used is logistic regression to predict a binary outcome: bad
debt or no bad debt. Some banks also build regression models that predict the amount of bad debt
a customer may incur. Typically this is much harder to predict, and most banks focus only on the
binary outcome.

Credit scoring is closely regulated only by the Financial Conduct Authority when used for the
purposes of the Advanced approach to Capital Adequacy under Basel II regulations.

Credit scoring is closely regulated in the UK, with the industry regulator being the Information
Commissioner's Office (ICO). Consumers can also send complaints to the Financial Ombudsman
Service if they experience problems with any Credit Reference Agency.

It is very difficult for a consumer to know in advance whether they have a high enough credit
score to be accepted for credit with a given lender. This situation is due to the complexity and
structure of credit scoring, which differs from one lender to another.

Lenders need not reveal their credit score head, nor need they reveal the minimum credit score
required for the applicant to be accepted, because there may not be such a minimum score.

If the applicant is declined for credit, the lender is not obliged to reveal the exact reason why.
However industry associations including the Finance and Leasing Association oblige their
members to provide a satisfactory reason. Credit-bureau data sharing agreements also require
that an applicant declined based on credit-bureau data is told that this is the reason and the
address of the credit bureau must be provided.

2.1.14 United States

9
In the United States, a credit score is a number based on a statistical analysis of a person's credit
files, that in theory represents the creditworthiness of that person, which is the likelihood that
people will pay their bills. A credit score is primarily based on credit report information,
typically from one of the three major credit bureaus: Experian, TransUnion, and Equifax. Income
and employment history (or lack thereof) are not considered by the major credit bureaus when
calculating credit scores.

There are different methods of calculating credit scores. FICO scores, the most widely used type
of credit score, is a credit score developed by FICO, previously known as Fair Isaac Corporation.
As of 2018, there are currently 29 different versions of FICO scores in use in the United States.
Some of these versions are "industry specific" scores, that is, scores produced for particular
market segments, including automotive lending and bankcard (credit card) lending. Industry-
specific FICO scores produced for automotive lending are formulated differently than FICO
scores produced for bankcard lending. Nearly every consumer will have different FICO scores
depending upon which type of FICO score is ordered by a lender; for example, a consumer with
several paid-in-full car loans but no reported credit card payment history will generally score
better on a FICO automotive-enhanced score than on a FICO bankcard-enhanced score. FICO
also produces several "general purpose" scores which are not tailored to any particular industry.
Industry-specific FICO scores range from 250 to 900, whereas general purpose scores range
from 300 to 850.

FICO scores are used by many mortgage lenders that use a risk-based system to determine the
possibility that the borrower may default on financial obligations to the mortgage lender. For
most mortgages originated in the United States, three credit scores are obtained on a consumer: a
Beacon 5.0 score (Beacon is a trademark of FICO) which is calculated from the consumer's
Equifax credit history, a FICO Model II score, which is calculated from the consumer's Experian
credit history, and a Classic04 score, which is calculated from the consumer's Trans Union
history.

Credit bureaus also often re-sell FICO scores directly to consumers, often a general-purpose
FICO 8 score. Previously, the credit bureaus also sold their own credit scores which they
developed themselves, and which did not require payment to FICO to utilize: Equifax's RISK

10
score and Experian's PLUS score. However, as of 2018, these scores are no longer sold by the
credit bureaus. Trans Union offers a Vantage 3.0 score for sale to consumers, which is a version
of the VantageScore credit score. In addition, many large lenders, including the major credit card
issuers, have developed their own proprietary scoring models.

Studies have shown scores to be predictive of risk in the underwriting of both credit and
insurance. Some studies even suggest that most consumers are the beneficiaries of lower credit
costs and insurance premiums due to the use of credit scores.

New credit scores have been developed in the last decade by companies such as
Scorelogix, PRBC, L2C, Innovis etc. which do not use bureau data to predict creditworthiness.
Scorelogix's JSS Credit Score uses a different set of risk factors, such as the borrower's job
stability, income, income sufficiency, and impact of economy, in predicting credit risk, and the
use of such alternative credit scores is on the rise. These new types of credit scores are often
combined with FICO or bureau scores to improve the accuracy of predictions. Most lenders
today use some combination of bureau scores and alternative credit scores to develop better
understanding of a borrower's ability to pay.

It is widely recognized that FICO is a measure of past ability to pay. New credit scores that focus
more on future ability to pay are being deployed to enhance credit risk models. L2C offers an
alternative credit score that uses utility payment histories to determine creditworthiness, and
many lenders use this score in addition to bureau scores to make lending decisions. Many lenders
use Scorelogix's JSS score in addition to bureau scores, given that the JSS score incorporates job
and income stability to determine whether the borrower will have the ability to repay debt in the
future. It is thought that the FICO score will remain the dominant score, but it will likely be used
in conjunction with other alternative credit scores that offer other pictures of risk.

Usage of credit histories in employment screenings has increased from 19% in 1996 to 42% in
2006. However, credit reports for employment screening purposes do not include credit scores.

Americans are entitled to one free credit report in every 12-month period from each of the three
credit bureaus, but are not entitled to receive a free credit score. The three credit bureaus
run Annualcreditreport.com, where users can get their free credit reports. Credit scores are
available as an add-on feature of the report for a fee. If the consumer disputes an item on a credit
report obtained using the free system, under the Fair Credit Reporting Act (FCRA), the credit

11
bureaus have 45 days to investigate, rather than 30 days for reports obtained otherwise.

Alternatively, consumers wishing to obtain their credit scores can in some cases purchase them
separately from the credit bureaus or can purchase their FICO score directly from FICO. Credit
scores (including FICO scores) are also made available free by subscription to one of the
many credit report monitoring services available from the credit bureaus or other third parties,
although to actually get the scores free from most such services, one must use a credit card to
sign up for a free trial subscription of the service and then cancel before the first monthly charge.
Websites like WalletHub, Credit Sesame and Credit Karma provide free credit scores with no
credit card required, using the TransUnion VantageScore 3.0 model. Credit.com uses
the Experian VantageScore 3.0 model. Until March 2009, holders of credit cards issued
by Washington Mutual were offered a free FICO score each month through the bank's Web site.
(Chase, which took over Washington Mutual in 2008, discontinued this practice in March
2009.) Chase resumed the practice of offering a free FICO score in March 2010 of select card
members to the exclusion of the majority of former WAMU card holders.

Under the Fair Credit Reporting Act, a consumer is entitled to a free credit report (but not a free
credit score) within 60 days of any adverse action (e.g., being denied credit, or receiving
substandard credit terms from a lender) taken as a result of their credit score. Under the Wall
Street reform bill passed on 22 July 2010, a consumer is entitled to receive a free credit score if
they are denied a loan or insurance due to their credit score.

The generic or classic FICO credit score ranges between 300 and 850. The VantageScore 3.0
score ranges from 300-850. The old VantageScore was between 501 and 990.

The first step to interpreting a score is to identify the source of the credit score and its use. There
are numerous scores based on various scoring models sold to lenders and other users. The most
common was created by FICO and is called FICO score. FICO is a publicly traded corporation
(under the ticker symbol FICO) that created the best-known and most widely used credit score
model in the United States. FICO produces scoring models which are installed at and distributed
by the three largest national credit repositories in the U.S (TransUnion, Equifax and Experian)
and the two national credit repositories in Canada (TransUnion Canada and Equifax Canada).
FICO controls the vast majority of the credit score market in the United States and Canada
although there are several other competing players that collectively share a very small percentage

12
of the market.

In the United States, the median generic FICO score was 723 in 2006 and 711 in 2011. The
performance definition of the FICO risk score (its stated design objective) is to predict the
likelihood that a consumer will go 90 days past due or worse in the subsequent 24 months after
the score has been calculated. The higher the consumer's score, the less likely he or she will go
90 days past due in the subsequent 24 months after the score has been calculated. Because
different lending uses (mortgage, automobile, credit card) have different parameters, FICO
algorithms are adjusted according to the predictability of that use. For this reason, a person might
have a higher credit score for a revolving credit card debt when compared to a mortgage credit
score taken at the same point in time.

The interpretation of a credit score will vary by lender, industry, and the economy as a whole.
While 640 has been a divider between "prime" and "subprime", all considerations about score
revolve around the strength of the economy in general and investors' appetites for risk in
providing the funding for borrowers in particular when the score is evaluated. In 2010, the
Federal Housing Administration (FHA) tightened its guidelines regarding credit scores to a small
degree, but lenders who have to service and sell the securities packaged for sale into the
secondary market largely raised their minimum score to 640 in the absence of strong
compensating factors in the borrower's loan profile. In another housing example, Fannie Mae and
Freddie Mac began charging extra for loans over 75% of the value that have scores below 740.
Furthermore, private mortgage insurance companies will not even provide mortgage insurance
for borrowers with scores below 660. Therefore, "prime" is a product of the lender's appetite for
the risk profile of the borrower at the time that the borrower is asking for the loan.

Several factors affect individual's credit scores. One factor is the amount an individual borrowed
as compared to the amount of credit available to the individual. As an individual borrows, or
leverages, more money, the individual's credit score decreases.

2.1.15 United Arab Emirates

Credit Scores in UAE are issued by AECB, a Federal Government-held company. Individual
Credit Scores are three-digit numbers used to predict the likelihood of an individual making their
loan and credit card payments on time, based on previous credit and payment behavior.

13
The Credit Score number in UAE ranges from 300 to 900. A low score indicates a higher risk,
whereas a higher score indicates a lower risk.

Issuer of Credit Score is Al Etihad Credit Bureau. its a Public Joint Stock Company wholly
owned by the UAE Federal Government. As per UAE Federal Law No. (6) of 2010 concerning
Credit Information, the company is mandated to regularly collect credit information from
financial and non-financial institutions in the UAE.

Al Etihad Credit Bureau aggregates and analyzes this data to calculate Credit Scores and produce
Credit Reports that are made available to individuals and companies in the UAE.

Credit Scoring is relatively new in UAE, But All major banks started using the same which
shows its high importance.

2.2 What is Credit Scoring?

1. Credit scoring is a statistical technique that combines several financial characteristics to form
a single score for assessing a borrower’s credit worthiness. The score does not predict a
company's ability to pay, but rather its willingness to pay in a timely fashion. The probabilities of
delinquency, as estimated by the model, are based on the analysis of previous applicants with
similar characteristics. Credit scorecards are “tools used to predict the behavior of new applicants
based on the performance of previous applicants” (U.S. Comptroller of the Currency, 1998).
Scorecards can also be used to predict the performance of existing accounts, based on past
experience of accounts with similar characteristics.

2. Credit scoring is a model applied by banks in their assessment and approval or decline of the
loan requests by SMEs. As there is a strong link between the payment behavior of the business
owner and that of the business, SME credit scores usually include financial characteristics from
both the business and the business owner. Credit scoring is based upon information like how the
repayment of the previous loans has gone, what is the current income level of the enterprise,
what are the outstanding debts, if any? It focuses on the credit history of the enterprise. As part
of the process, the lenders see whether the enterprise/business owner has the reliability and
honesty to repay the loan. It also examines how the enterprise has used credit before, its record
for repayment of bills, including utility bills, how long the enterprise has been in existence,

14
assets possessed by the enterprise and sustainability and viability of the activities that the unit is
engaged in. Credit scoring model draws inputs from historical information on the performance of
loans with similar characteristics.

2.3 Background of Credit Scoring

Credit scores have been widely used for many years in consumer credit markets e.g., mortgages,
credit cards, and auto loans. The model has facilitated lending procedures leading to availability
of low cost credits and faster decision making yielding significant growth in consumer credit
availability. Owing to the similarities of the small business in terms of being individual centric,
the principles utilized for making lending decisions to individual borrowers under the credit
scoring model were extrapolated to meet the credit requirements of small businesses. In the mid-
1990s, Fair Isaac and Company introduced one of the first credit scoring models developed
exclusively for SMEs, the Small Business Scoring Service (SBSS). Since then, many SME banks
in the U.S., as well as in Canada, the U.K., and Japan, have implemented some type of credit
scoring for SME borrowers.

2.4 Difference between credit scoring and credit rating

1. Quite often, I have seen bankers talking about credit scoring and credit rating in the same
breath. Therefore, before I move any further, let me attempt to clarify the difference between
credit scoring and credit rating. I would like to emphasize that these are two entirely distinct
concepts and to be employed in distinctly different scenarios. So what are these differences?
Credit scoring is a statistical technique that combines several pre determined characteristics to
form a single score to assess a borrower’s credit worthiness. The score allocated to any
application is the sum of the appropriate weights given by the values that the included
characteristics take for that application.

Any two identical applications will always receive the same score. Credit rating, on the other
hand, is based more on the experience and judgment of the credit officer and uses financial
indicators as key. The objective of scoring is to replicate manual analysis and approval of loans
at a lower cost, with greater speed, while the use of credit rating is reliant on manual analysis by
credit officers to supplement the rating provided by the tool.

15
2. To put it simply, credit scoring uses a retail lending approach to credit screening/decision
making and is recommended for smaller ticket size loans, where adequate reliable financial data
about the borrower is not available. Credit Rating, on the other hand, is a more appropriate tool
for larger, mid-segment or corporate loans which have relevant financial data/ business plans that
provide the basis for further credit analysis and information. Again, credit scoring is usually a
100 per cent automated mechanism for screening loan applicants on the basis of a pre-defined set
of parameters while credit rating is a more subjective and manual method for determining a
borrower’s ability to repay, based on a pre-determined set of criteria. Further, credit scoring
enables clear decision matrix like immediate approvals, declines or gray area loans, which can be
reviewed in greater depth. Declined loans are “knocked out” quickly while strong proposals get
quick approval. Additional assessment, if any, would only need to be made in respect of gray
area loans or proposals which are neither an unqualified ‘yes’ or an unqualified ‘no’. On the
contrary, credit rating requires a similar level of credit analysis for all loan applications and is
more comprehensive in nature, hence increasing the transaction cost per loan.

2.5 How Scoring Model works?

The objective of a scoring model is to use a relevant and representative sample to measure loan
performance and then develop characteristics that predict performance. The resulting model is
integrated into the decision making process. A scoring model applies different weights to the
characteristics used to predict the performance. The weights, or values, measure the influence of
that characteristic on the outcome. The weights and the level of influence are determined by
statistical analysis. Only those characteristics that exert a ‘significant’ influence are included in
the final model. The score allocated to any application is the sum of the appropriate weights
given by the values that the included characteristics take for that application. Under this model,
any two identical applications will always receive the same score, unlike in the manual
underwriting where the decision could change from one lender to the other or also by the same
lender depending on the interpretation on a given day.

2.6 Benefits of Credit Scoring

16
1. When used appropriately, credit scoring can benefit multiple stakeholders, including lenders,
borrowers, and the overall economy. For the lender, scoring leads to process automation which
facilitates process improvements leading to many by-products such as improved management
information, control and consistency. It also increases the profitability of SME lending by
reducing the time and cost required to approve loans and increasing revenues by expanding
lending opportunities as lenders can safely approve marginal applicants that an individual
underwriter might reject. International evidence has shown that credit scoring can assist in
overcoming the inherent benefit/cost trade-off that banks face when deciding whether or not to
invest in obtaining information regarding a potential borrower. A study that was meant to test the
credit scoring situation in US estimated that the cost of evaluating micro loan applications in the
US using credit scoring was reduced to around USRs.100 compared to a range of USRs.500-
USRs.1,800 prior to the introduction of credit scoring. The time saving involved meant that
banks could focus more time on marginal applications, existing loans that are showing signs of
distress and processing more loan applications. Use of credit scoring has also meant that the
marginal benefits of taking and maintaining collateral are not justified for small loans.

2. The Bank of England has also acknowledged that there is some evidence of banks being more
willing to lend on an unsecured basis when using credit scoring, which potentially improves the
access to bank finance for very small and start-up Small and Medium Enterprises.

3. For the borrower, the benefits from credit scoring include increased access to credit and, in
some cases, lower borrowing costs. In its study of SME credit scoring’s impact on access to
credit, the Federal Reserve Bank (FRB) of Atlanta found that, in general, the use of credit
scoring increased the amount of credit banks extended to the Small and Medium Enterprises. It
found that banks using scoring were more likely to lend to the Small and Medium Enterprises
that lacked sufficient financial information for approval through traditional underwriting
methods. Presumably, this is due to the inclusion of the business owner’s personal information
in the scoring process. The study also found that banks using scoring were more likely to lend
in low-income areas, a fact it attributed to greater objectivity in the underwriting process.

Specific Benefits of credit scoring:

 Speed - When used with an automated software system, each customer is evaluated or
scored in seconds. Customer orders are approved without delay. Also automated data

17
 feeds can speed up data entry of financials, ratings, account status and exposure, and
other key data points.

 Consistency and Accuracy - All factors involved in the credit score are considered and
used in the same way. Since credit scoring analyzes each customer using a similar set of
rules, there is consistency in the evaluation process by the entire staff at all
locations. Plus human error is eliminated.

 Reduced Bad Debt Losses - When approving new customers, all of the necessary factors
involved in the credit decision process are received and scored. High-risk customers are
identified as exceptions and reviewed by a credit analyst. Customers on the verge of
bankruptcy are not approved because of missed information or incorrect analysis. Orders
are not approved simply because the dollar amount is less than the automatic credit
approval amount. By quickly identifying existing customers that require immediate
attention you can hold orders that otherwise may have been shipped, and take legal action
while assets are still available to be attached.

 Reduced Personnel Costs - Particularly with firms having many thousands of customers,
the impact of credit scoring, combined with the use of an automated software system, can
significantly reduce personnel costs. Each activity in a credit department has personnel
cost associated with that activity. With automated credit scoring, fewer people are needed
to research customers, check references and make decisions.

 Collection Activities are Prioritized - Credit scores enable the credit executive to have
different collection strategies for low risk, medium risk and high risk customers. When
credit risk scores are coupled with amounts owed, collection activity is prioritized.

 Decision support and planning tools - Credit scoring enables the credit executive to
prepare reports that accurately reflect the quality of the total accounts receivable
portfolio, and other reports that will reveal if certain groups of customers carry more risk.
Additional reports will focus attention on the amount of bad debts for each risk category,

18
 pricing strategies, overly rigid or liberal applications of credit policy, credit department
staffing and expenses, and other issues vital to a firm’s future.

 Comply with audit mandates – Due to Sarbanes-Oxley Act, auditors are more likely to
require supporting materials on the collectability of receivables and class of
receivable. Expert reviews can reveal inadequate bad debt reserve and overlooked
significant write-offs from prior periods. This failure could result in material errors in
past and current financial statements. Credit scoring enables you to support decisions by
tracking data evaluated and how decisions were made. Furthermore, you can score to
regularly re-evaluate your portfolio in a timely manner. In this way, your decisions will
be more reliable and supportable for financial reporting.

2.7 What influences credit scoring?

Many factors are taken into account in the process of granting a loan. These include the
characteristics of the borrower (who they are), their economic situation, the amount of the loan
applied for, its purpose (i.e. what is to be financed by the loan) and the type of collateral. The variety
of these factors means that the risk is estimated using elements of quantitative and qualitative
analysis.

The quantitative analysis includes, first of all, an assessment of the financial standing of the customer
based on their income and monthly expenses. It may also include cash flow analysis of the
customer's accounts and credit history. While the qualitative assessment takes into account, among
others, marital status, education or employment form - for natural persons, and for enterprises - legal
form, industry in which they operate or the way of keeping accounts.

Equally important are past customer behaviors that adversely affects credit scoring:

 late payment of installments and other liabilities,

 exceeding credit card limits,
 large number of commitments entered into,

19
  no credit history of any kind.

2.8 Credit scoring model. How is it calculated?

Banks usually grant loans based on a credit scoring model which

combines qualitative and quantitative analysis. Credit scoring is based on statistical methods,
thanks to which it is possible to predict the probability of a certain event occurring in the future – in
this case a loan default.

The scoring process uses information about the customer collected at the application stage - mainly
data characterizing the customer, but also information about their past behavior. Each credit
institution considers a different set of features and assigns different point values to them. For
example, a highly educated person will usually score higher than a college dropout, but the exact
point value and its impact on the final score may vary from bank to bank. The sum of points from
particular characteristics is usually the final score. The range of the possible score is determined by
each bank / institution - the most popular credit scoring in the USA (FICO) gives results between 300
and 850, while the scoring provided by the Polish credit information bureau (BIK) can reach a
maximum of 100 points.

2. Literature Review
There are many studies developing this subject. Some studies are particularly interesting in order
to frame the analysis and to create some references for the developments that are intended to
develop in this study. Some modelling cases follow around studies on this area.

Hand and Crowde (2005), for example, used latent-variable technique for measuring underlying
aspects of credit customer behavior. The latent- variable model separates the observed variables
into primary characteristics (x) and behavioral characteristics (y). Then the study summarizes
them into overall measure of credit consumer scores.

Samreen et al. (2013) summarized the development of a credit scoring model known as Credit

20
Scoring Model for Corporations (CSMC), which could be used to evaluate the creditworthiness
of corporate borrowers before granting loan. Type I and type II errors of proposed model
(CSMC) have more accuracy rate with no errors as compared to LR and DA.

Li and Zhong (2012) introduced ensemble learning model for credit scoring. This model points
out moving from static credit scoring to dynamic behavioral scoring and maximizing revenue by
decreasing Type I and Type II errors. The challenges faced in building credit scoring models are
half- baked applicant’s information, missing values and inaccurate information.

Hussein and Pointon (2011) reviewed 214 articles/books of credit scoring applications. The
important and key determinants of credit scoring models have been investigated. The matrix
(ACC rate criterion) measures the proportion of correctly classified cases. ACC rate is a
significant criterion in evaluating the classification capacity of proposed scoring model. ROC
curves also known as Lorentz curves is a two dimensional graph that represents the proposition
of sensitivity (1-type II error) on y-axis and specificity (1-type I error) on x-axis. The maximum
distance between ROC and diagonal is equal to constant times K-S statistics.

Bellotti and Crook (2009) developed a credit score model with inclusion of time varying macro-
variables like interest rate and unemployment rate using Survival Analysis. Survival Analysis is
competitive in comparison to LR as a credit scoring method for prediction. The inclusion of
macro- variables gave a statistically significant improvement in predictive performance.

Kessy (2011) examined the link between the loan processing and monitoring in banks, and asset
growth and empowerment of individual customers. Findings revealed that better allocation and
utilization of financial institution’s economic capital not only facilitates outreach to more under
banked and unbanked productive poor people but also empowering them by stimulating
investments and increasing productivity in a cost-effective way for poverty reduction.

Azam et al. (2012) evaluated the significance of loan applicant socioeconomic attributes on
personal loan decision in banks using descriptive statistics and logistic regression. The model
identified that out of six independent variables only three variables (region, residence status and
year with the current organization) have significant impact on personal loan decision.

Matthew and Sarah (2013) investigated credit risk and default among Ghanaian banks. It was
suggested that banks should tighten their credit assessment tool, i.e. CAMPARI (Character,
Ability, Model, Purpose, Amount, Repayment and Insurance) model. It was recommended that

21
the Central Bank should facilitate the establishment of a vibrant credit- referencing bureau in
order to provide the credit history for banks customers.

Koh et al. (2006) illustrated the use of data mining techniques to construct credit scoring
models. The construction of this model has five steps: defining the objective, selecting
variables, selecting sample and collecting data, selecting modelling tools and constructing
models, validating and assessing models.

Nancy et al. (2013) studied the Credit Risk Assessment Model of SBI Bank. SBI loan norms are
flexible and differ from case to case. Loans applicant information will be checked from RBI
willful defaulters lists.

Credit scoring procedures and techniques

Credit scoring was primarily dedicated to assessing individuals who were granted loans, both
existing and new customers. Credit analysts, based on pre-determined scores, reviewed
customers’ credit history and creditworthiness to minimize the probability of delinquency and
default.

Basically, credit scoring is a method which can be used to classify or quantify the risk factors
relevant for a borrower's ability and willingness to repay the loan. Credit scoring allows lenders
to predict likely loan outcomes based on the use of statistical techniques, which allow objective
predictions as to whether a loan will produce a good or bad outcome. Credit scoring can be used
on a standalone basis or as a part of the credit evaluation process. When used on a standalone
basis, credit scoring assists in classifying applicants into accept/reject groups or good/bad
credits; when used as part of the credit evaluation process, credit scoring can help to measure the
credit risk of the applicants (Thomas et al, 2019; Bhatia, 2006).

Durand (1941) was the first to recognise that one could use the same techniques to discriminate
between good and bad loans. “Credit scoring is essentially a way of recognizing the different
groups in a population when one cannot see the characteristic that separates the groups”
(Thomas, 2017). Commercially, the credit scoring was first developed in the 1950s by Bill Fair
and Earl Isaac, but has only come into increasing use in the last two decades (Thomas, 2017).
The main aim of the credit scoring model is to build a single aggregate risk indicator for a set of
risk factors from analysis of data representative of the lender's own previous lending experience

22
(Thomas, 2017; Bhatia, 2006).

As per the information collected from the bankers, the credit scoring for personal loans is done
in line with the RBI guidelines. Almost all banks are following BASEL II and III guidelines. As
per the experience of bankers, credit scoring process includes collecting, analyzing and
classifying different credit elements and variables to assess the credit decisions. The quality of
bank loans is the key determinant of competition, survival and profitability. It is one of the most
important kits, to classify a bank’s customers, as a part of the credit evaluation process to reduce
the current and the expected risk of a customer as being a bad credit. The objective of credit
scoring models is to assign loan customers to either good credit or bad credit or predict the bad
creditors. Therefore, scoring problems are related to the classification analysis.

Probably the earliest use of statistical scoring to distinguish between “good” and “bad”
applicants was made by Durand. Bankers will assign some weight age to the loan application
form filled by borrower. Every bank has its own credit score cut points, which will cluster the
customer into different risk groups.

A wide range of statistical techniques are used in building scoring models. Most of these models
are statistical, being some of them non-linear; models are applicable to build an efficient and
effective credit scoring system that is effectively used for predictive purposes. Techniques, such
as weight of evidence measure, regression analysis, discriminant analysis, probit analysis,
logistic regression, linear programming, Cox’s proportional hazard model, support vector
machines, decision trees, neural networks, k-nearest-neighbour, genetic algorithms and genetic
programming, are all widely used techniques in building credit scoring models by credit analysts,
bankers, lenders and computer software developers and providers

Advanced statistical methods vs. traditional statistical methods

Advanced statistical techniques, such neural networks and genetic programming, provide an
alternative to conventional statistical techniques, such as discriminant analysis, Probit analysis
and logistic regression. The point of using sophisticated techniques, such as neural nets, is their
capability of modeling extremely complex functions, and, of course, this stands in contrast to
traditional linear techniques, such as linear regression and linear discriminant analysis.
Probabilistic neural nets usually trains presented cases faster than multi-layer feed-forward nets,
and classifies them in the same way or better than multi-layer feed-forward nets, even through

23
multi-layer feed-forward nets have been shown to be excellent classifiers (Palisade, 2005; Irwin,
et al., 1995). However, a range of sophisticated algorithms for neural nets training - making them
an attractive alternative to the more conventional techniques - has become available (Masters,
1995; Palisade, 2005). Also, genetic programming is one of the most successful alternatives to
traditional techniques recently used in this field. Genetic programming is utilized to
automatically determine the sufficient discriminant functions and the applicable features
simultaneously. Dissimilar neural networks may only suit large datasets, but genetic
programming can positively perform well even with small data-sets (Nath et al, 1997). Different
credit scoring tools and techniques used by banks are discussed in existing literature.

3. Methodology

4.1 Statement of the Problem

Considering the exposed above, it can be said that the main income for retail banking is the
interest generated for the loans and advances. If this interest or loans are not paid regularly it
becomes a big problem for the bank. When a borrower fails to meet the legal obligations (or
conditions) of a loan, he is said to have defaulted on his/her loan. These defaults increase the
level of non-performing assets. In order to decrease the level of non-performing assets, the bank
has to develop loan application screening methods, which distinguishes applicants as bad and
good applicants, considering the credit scoring methods. Almost all banks have credit scoring
methods, but still they have non-performing assets. This study is carried out to know the credit
scoring methods used by banks.
4.2What are the types of credit scoring?

Scoring models can be classified according to different criteria. Thus, we can talk about a scoring of
individuals or companies (division based on the assessed entity) or credit card, cash or mortgage
scoring (depending on the type of product applied for by the client). Taking into account who created
and managed the scoring model, we can talk about internal scoring (prepared by banks for their own
needs) or external scoring (created and made available by specialized institutions, e.g. credit
information offices).

24
There is a very clear dividing line between application and behavioral scoring. The first one is
designed to evaluate new customers on the basis of data provided by the customer in the credit
application. Behavioral scoring, on the other hand, is determined on the basis of the history of the
customer's behavior concerning the service of financial products. Therefore, it is calculated for
regular customers, mainly in order to resell new products or change the terms and conditions of
existing products (e.g. increase the credit card limit).

For the most part, the goal of the scoring models is to determine the risk of debt default. However,
more and more emphasis has been placed recently on using this method for other purposes:

 profit maximization (what credit terms should be offered to the customer to be accepted, i.e.
risk-based pricing)
 increasing the effectiveness of marketing campaigns by investigating whether the customer
will be interested in a given product,
 fraud scoring,
 attrition scoring,
 improving debt management by determining whether the customer will be able to repay the
loan in the event of financial problems.

Regardless of the type, the scoring models enable an objective assessment of credit risk, which is a
key element of the credit granting process. In order to make the credit calculation as accurate,
transparent and low-risk as possible, banks increasingly automate it and use ready-made systems that
allow for performing a credit assessment model in a point system. The use of such tools reduces the
probability of granting doubtful loans and allows for accelerating the entire credit process while
reducing the risk of human error.

Individual customer scoring

The risk assessment of natural persons is based on their personal characteristics. The most commonly
used are: age, marital status, education, number of dependents, seniority, form of employment,
occupation, etc. The risk assessment of natural persons is based on their personal
characteristics. Financial characteristics also play a very important role, such as monthly income,
possible additional income or information about expenses (repayment of other loans, living costs,

25
bills paid, etc.). If possible, data on the financial history of the customer are also taken into account,
such as the number of previous commitments made and the history of their repayment or information
about possible overdrafts.

Depending on the type of credit applied for, the list of characteristics to be taken into account in the
risk assessment may vary. The scoring model for a mortgage loan is usually more comprehensive
than that for a credit card. When buying a house on a loan, the scoring also includes a number of
information about the property itself (including its value, LTV level, etc.). It also puts more emphasis
on the income data.

Enterprise Scoring (SME)

In the case of small and medium-sized enterprises, the risk of bankruptcy and insolvency is examined
by assigning points. Their capital, debt and development strategy are taken into account. In the case
of enterprises, the credit calculation is made on the basis of: the characteristics of the industry, the
characteristics of the company, its previous financial results. Credit risk analysis is also performed
taking into account the company's structure, source of financing, competition, and even the
qualifications of employees, mainly those at high levels, are examined. The credit assessment model
in the case of a company also takes into account the financial condition of the company,
its planned projects, liquidity, financial liabilities and industry risk assessment.

In the case of the smallest companies, however, very often, in addition to taking into account the
parameters characterizing the company, the owner of the examined company is also the owner. It
turns out that the profile of the owner and their personal credit history is more important than the
numbers describing their business. This is particularly often used in the case of the smallest entities
operating on the market for a short period of time.

4.3 Rules Based Credit Scoring Methodology

CRF thanks Credit & Management Systems, Inc. http://www.icmsglobal.com

Today’s credit professionals must make accurate, on-the-spot decisions. Of course, you have
high value and/or high volume of credit decisions. Departments are consolidating and slimming
down as well as under more scrutiny from auditors. Under these conditions, how do you perform

26
the detailed and consistent analysis needed to avoid unneccessary credit risk? Credit scoring can
help. Credit scoring, by definition, is a method of evaluating the credit worthiness of your
customers by using a formula or set of rules. Depending on the make up of your customer base,
credit scoring can produce considerable benefits to some firms and somewhat lesser benefits to
others.
The Different Types of Credit Scoring
Credit scoring is based on the assumption that past experience can be used as a guide in
predicting credit worthiness. There are two types of credit scoring models. Both can be
statistically validated.

Judgmental Scoring Model

A judgmental scoring model is based on traditional standards of credit analysis. Factors such as
payment history, bank and trade references, credit agency ratings and financial statement ratios
are scored and weighted to produce an overall credit score.

The determination of which factors to use, and how each will be scored and weighted, is
generally based on the credit executive's past experience with their company, the products or
services they sell, and the industry they are in. Judgmental models are enhanced by comparing
industry financial profiles using peer groups from (RMA) Risk Management Associates
Statement Studies. Including scoring factors that reflect the individual characteristics and
policies of their own firm further enhances the judgmental model.

Judgemental scoring is the most straightforward to implement because it uses your credit policies
and decision process, the number of rules are easily set, and the grading scale can be simple or
complex. Therefore, it is easier to understand and augment.

Statistical Scoring Model

Statistical models function in much the same way as judgmental models. However, in choosing
the factors to be scored and weighted they rely on statistical methods rather than the experience
and judgment of a credit executive.

27
Statistical models consider many factors simultaneously, a process that calculates and analyzes
multivariate correlation to identify the relevant tradeoffs among factors, and assigns statistically
derived weights used in the model. The key factors are generally captured from credit agency
reports and the credit files of the client.

Statistical models are often described as a scorecard, a pooled scorecard, and a custom scorecard.
A scorecard uses data from one firm. A pooled scorecard uses data from many firms. A custom
scorecard blends a statistical model with some of the factors used in a judgmental model.
An Example Judgmental Scoring Model
A judgmental scoring model is used in this example because it is an effective scoring model, and
the simplest scoring model to implement. It uses your credit policies and decision processes, the
number of rules are easily set, and the grading scale can be simple or complex. And it does not
require the services of a costly third party to create and maintain.

In this scoring model, the goal is to calculate an overall risk score based on the principle that the
risk or credit worthiness of a customer can be evaluated on:
1. Traditional credit information - This is credit information that is non-financial data that
you would normally use when making a credit decision. Items such as control years, trade
references and pay history fall into this category.
2. Credit Agency information - The outside ratings that you normally consider in your credit
decision process. The D&B rating, Paydex score and Experian Intelliscore fall into this
category.
3. Financial Statement scores – Ratios scored comparing peer companies within the same
financial year by industry by size (sales or assets).

Each piece of credit information, from trade references to financial information, will be scored
on a 1-6 scale in which 1 is the best score and 6 is the worst.

You establish the level of emphasis for each item by giving it a weighting. This is a subjective
element that “customizes” the scoring model to your specifications. Sample values for each
category are shown below.

28
 Traditional Credit Information Scoring:

Information Weight Score Result

Pay history to others 0.15 3 0.45
Pay history to others 0.15 3 0.45
Bank rating 0.05 3 0.15
Trade reference #1 0.04 2 0.08
Trade reference #2 0.03 3 0.09
Trade reference #3 0.03 3 0.09
Industry credit group 0.05 3 0.15
Control years 0.10 5 0.50
NSF checks reported 0.05 6 0.30
Collection claims reported 0.05 1 0.05
Suits 0.10 1 0.10
Judgments/tax liens 0.20 1 0.20

Traditional Score 2.61

As you can see in the sample provided, traditional items typically consist of non-financial
information that you would normally use when making a credit decision. Feel free to add the
items that you use in your particular industry.

29
These items might include some of the following:

Plant Capacity
Profit Margin
Country Risk
Annual Purchases
Number of Beds in a Hospital
Collateral
Occupancy Rate of Beds
Competitive Index
Credit Agency Information Scoring:

Information Rating Weight Score Result

D&B Rating 4A2 0.20 2 0.40
D&B Paydex 75 0.20 3 0.60
NACM Score 85 0.20 2 0.40
Experian days beyond 16 0.20 3 0.40
terms
Experian Intelliscore 0.20 3 0.00
Credit Agency Score 2.61

Sample Scoring
Parameters and
their
corresponding
translation
tables

30
The Credit Agency category will contain the third party ratings that you consider most
important when making a credit decision. As in the traditional category, you will assign
weights, in percentage form, to emphasize the importance of each rating. Also like the
traditional category, items which are not scored are not used in the evaluation process--the
items that are scored are simply re-weighted.

Also note since the scoring uses a common grade scale, understanding the multiple credit agency
scores is easier to understand by eliminating the proprietary agencies alphabet soup of rating
scores.

Financial Statement Scoring:

One ratio is not enough to analyze a financial statement. The best way to determine a firm's
financial quality is to assess by comparing ratios of peers in three categories:
· Liquidity
· Profitability
· Leverage
The 12 ratios listed below will provide a good measurement of the financial strength or weakness
of a firm.

Liquidity Ratios
 Current Ratio
 Quick Ratio
 Sales / Receivables
 Cost of Sales / Inventory
 Cost of Sales / Payables
 Sales / Working Capital

Profitability & Operating Ratios

 % Profit Before Taxes/Tangible Net Worth
 % Profit Before Taxes/Total Assets
 Sales/Total Assets

31
 Leverage & Coverage Ratios
 EBIT / lnterest
 Fixed Assets/Tangible Net Worth
 Total Liabilities/Tangible Net Worth

In scoring a financial statement, the above 12 ratios will be computed and then scored through
comparison to a published peer group such as the industry ratios published by Risk Management
Associates in their Annual Statement Studies. Risk Management Associates (RMA), through
their member banks, will receive financial statements throughout the year and then separate them
into groups that are similar in the following ways:
1. Same industry - as defined by the SIC code.
2. Same financial statement year.
3. Same size - as defined by size of sales or assets

RMA will then calculate and publish the upper quartile, the median and lower quartile ratios for
each of 12 ratios. How they do this is reflected in the following hypothetical sample. The small
sample size of 17 firms is used for illustration purposes only:
You will need to expand these three ratio levels in order to score on a scale of 1 to 6.

Current Ratio Score

Above 2.2 1
Between 2.1 & 1.8 2
Between 1.8 & 1.4 3
Between 1.4 & 1.1 4
Between 1.1 & 0.8 5
Below 0.8 6

Each ratio is now scored on the 1 to 6 scale. The average score is determined for the liquidity,

32
profitability, and leverage categories, and then multiplied by the weight assigned to each
category, to arrive at the financial statement score.

Information Weight Score Result

by category
Liquidity Ratios
Current Ratio 3
Quick Ratio 4
Sales / Receivables 2
Cost of Sales / Inventory 3
Cost of Sales / Payables 4
Sales/Working Capital 2
30% 18 / 6 = 3.00 0.90
Profitability Ratios
% Profit Before Taxes / Tangible Net 3
Worth
% Profit Before Taxes / Total Assets 2
Sales / Total Assets 2
40% 7 / 3 = 2.33 0.93
Leverage Ratios
EBIT / Interest 1
Fixed Assets / Tangible Net Worth 3
Total Liabilities / Tangible Net Worth 2
30% 6 / 3 = 2.00 0.60
Financial Statement Score 2.43

Overall Risk Score

The overall risk score is developed by multiplying the traditional score, credit agency score, and
financial statement score by the weight assigned. When a category score is not available the

33
remaining scores are automatically re-weighted. In the example below, if a financial statement
had not been available, the traditional score would be weighted at 30/40's, or 75%, and the credit
agency score would be weighted 10/40's, or 25%.

Category Weight Score Result

Information Weight Score Result
Traditional score 0.30 2.61 0.78
Credit Agency 0.10 2.49 0.25
score
Financial Statement 0.60 2.43 1.46
score
Overall Risk Score 2.49

Overall Risk Score Scale

Highest Quality 1.00 to 1.83
Good Quality
1.84 to
2.66
Average 2.67 to 3.50
Below Average 3.51 to 4.34
Poor Risk 4.35 to 5.17
High Risk 5.18 to 6.00

As you can see in this example, the account is of “good quality” in terms of creditworthiness
with a 2.49 Overall Risk Score.

34
4.4 Management Reports Made Possible By Credit Scoring
I. Portfolio Report
Historically, most credit executives used the Days Sales Outstanding (DSO) figure to measure
the quality of their accounts receivable. This assumes, of course, that customer payment trends
are related to risk. Actually, how a customer pays your firm is often a poor indicator of risk.
Many high-risk customers pay promptly or within acceptable terms. Conversely, low risk
customers often are given longer terms to accommodate special inventory programs.

A credit score that uses many types of credit information to evaluate a customer's risk is a better
measurement than one single factor such as how that customer pays your firm. If this is true
when evaluating a single customer, it is also true when evaluating all of your customers.

It follows then that a better measurement of the quality of your accounts receivable portfolio is to
use the credit score for each customer along with the amount owed by that customer at the end of
each month. By combining the total amount owed by customers in each risk category the credit
executive will have a picture of how much of their portfolio is high quality, how much is high
risk, and everything in between. The following report is an example of portfolio analysis using
credit scores:

Portfolio Analysis
Year 2017 2018 2019
A/R Balance Rs.20,087,185 Rs.22,174,030 Rs.25,108,033
High Quality Rs.2,812,001 Rs.1,980,004 Rs.1,081,003
(1.00 - 1.83) 14.00% 8.93% 4.31%
Good Quality Rs.5,429,002 Rs.5,588,005 Rs.4,750,004
(1.84 - 2.67) 27.03% 25.02% 18.92%
Average Rs.5,631,003 Rs.5,080,006 Rs.5,250,005
(2.68 - 3.50) 28.03% 22.91% 20.91%
Below Average Rs.6,215,179 Rs.7,326,007 Rs.9,010,006
(3.51 - 4.33) 30.94% 22.04% 35.88%

35
Poor Quality Rs.2,200,008 Rs.3,759,007
(4.34 - 5.17) 9.92% 14.97%
High Risk Rs.1,258,008
(5.18 - 6.00) 5.01%
Percent 100.00% 100.00% 100.00%

This Portfolio Analysis illustrates a negative trend in the quality of the portfolio over a three-year
period. In 2017 there were no Poor or High Risk customers, but in 2019 almost 20% of the
customer base were considered Poor or High Risk. While the Portfolio Analysis uses an entire
customer base, the credit executive may also want to segment their portfolio by product line,
customer type or sales region to determine if certain groups of customers carry more risk than
others.

II. Bad Debt Reserve Report

Using the portfolio analysis reflected above, and past bad debt experience for each risk category,
the credit executive can forecast an amount for the Bad Debt Reserve that is more accurate than
basing it only on a percentage of projected sales.

Suppose that bad debts over the past several years had averaged 0.001% of annual sales. Using
just that single variable, the Bad Debt Reserve for 2003 would be forecast at Rs.270,000 or
0.001% of projected sales of Rs.270 million.

However, if the credit executive forecasted the amount of bad debts by using projected accounts
receivable and the past bad debt experience for each risk category, a much higher amount would
be set aside.

The calculation for Bad Debt Reserve using this method would look as follows:

Bad Debt Reserve Forecast

36
Year 2019 2003 Bad Debt Bad Debt Reserve
% in the past
A/R Balance Rs.25,108,033 Rs.27,000,000
High Quality Rs.1,081,003 Rs.1,163,700 0.0%
(1.00 - 1.83) 4.31% 4.31%
Good Quality Rs.4,750,004 Rs.5,108,400 0.0%
(1.84 - 2.67) 18.92% 18.92%
Average Rs.5,250,005 Rs.5,645,700 1.0% Rs.56,457
(2.68 - 3.50) 20.91% 20.91%
Below Rs.9,010,006 Rs.9,687,600 2.0% Rs.193,752
Average 35.88% 35.88%
(3.51 - 4.33)
Poor Quality Rs.3,759,007 Rs.4,041,900 3.0% Rs.121,257
(4.34 - 5.17) 14.97% 14.97%
High Risk Rs.1,258,008 Rs.1,352,700 4.0% Rs.54,108
(5.18 - 6.00) 5.01% 5.01%
Bad Debt Reserve Forecast Rs.425,574

The Bad Debt Reserve of Rs.425,574 is probably more accurate than the Rs.270,000 amount
because it considers that the 2003 portfolio has a much higher percentage of Poor And High Risk
customers than in 2017 and 2018.
Basing the bad debt reserve on past bad debt experience by category is somewhat similar to the
computations by life insurance companies. While life insurance companies do not know which
individuals in the 75 - 80 year old category will die next year, they can compute the percentage
of the category very accurately.

5. Findings and conclusion

Who Uses Credit Scoring And What Do They Have In Common?
The first firms to use credit scoring were credit card companies and the consumer lending

37
divisions of commercial banks. The huge number of transactions involved in consumer credit
necessitated a computer generated score to approve and service their customers in a cost
effective and timely fashion.
Fortunately, the credit information they needed for their statistical scorecards was readily
available, much of it free. Their credit application provided data concerning employment, annual
salary, home value, mortgage and other obligations. Additional data was available in consumer
credit reports that were usually very comprehensive.
The information required to conduct a credit appraisal for a consumer is far less than that of a
business. A salary of Rs.50,000 can be measured easier than a bank reference or a financial
statement.
Finally, the dollar amount of each consumer credit transaction is usually low, so that a single
scorecard rejection of a sale will have little impact on overall sales revenue.
All of the above factors contributed to making credit scoring quite successful for those firms in
the consumer credit business.
The second group to use credit scoring was banks, leasing companies and finance companies
lending to small business firms. Similar to the credit card business, the numbers of transactions
were high and the dollar amount on each relatively low. Since they could not afford to spend
hours gathering and analyzing credit information on each transaction they turned to statistical
scorecards.
Now, credit scoring is beginning to be used by business credit providers of all types.

38
5.1 Conclusion
Whether you decide to use credit scoring or not is an individual decision based on your customer
composition and other circumstances of your own company. Which of the credit scoring methods
you use, be it judgmental or statistical, is also an individual decision based on the perceived
benefits to your company and the cost of implementation. Rule-based scoring is straightforward
to implement because it uses your credit policies and decision processes, the number of rules are
easily set, and the grading scale can be simple or complex.
However, if you decide to implement credit scoring you will need computer processes that work
with the scores. If credit scoring is being used manually, you will only have limited benefits. The
advantages of speed, accuracy, audit tracking, and reduced personnel costs can only come if you
add credit scoring logic and functionality into your existing computer systems or obtain them
from a third party system.

5.2 Current Credit Scoring Procedure

The different types of loans offered by Indian banks are home loans, auto loans, personal loans,
business loans, loans against property, gold loans and credit cards loans. The process of loans
sanctioning involves personal loans product promotions, customers contacting the bank, filling
up the application form. Physical evidences for the support of a loan process are application
forms, agreement, loan balance statement, and acknowledgement of repayment.

The interface (people) responsible for carrying the whole loan process is a bank branch, a bank
loan executive, a bank manager, a credit scoring executive, a CIBIL and a credit manager.
Customer requirements are: availability for different loan products, low interest rates, low EMI,
maximum repayment time periods, diminishing rates of interests, easy documentation, higher
LTV and the credit scoring. Once banks receive the filled application form of applicant for
personal finance, it undergoes through various stages. The application form will be sent to the
credit risk department where credit scores are calculated.

There are two types of credit scoring done by banks in India i.e. internal and external scoring.
The ranges as well as groups vary from bank to bank, while as the external CIBIL scores are the
same for all banks. The CIBIL scores vary from 300 to 900 or NA (Not applicable) or NH (No
History), 300 being the lowest and 900 the highest. Banks prefer applicants with a score higher
than 700. The external scoring is carried out to check the applicant’s banking with other banks

39
and other financial institutions. The components of the CIBIL credit score and credit report are
payment history (35%), Amount Owed (30%), Length of Credit History (15%), New Credit
(10%) and types of Credit used (10%) for scoring the applicant. CIBIL has all the transaction
details of all borrowers. Normally the whole process will take place within 10 working days.

The commonly assessed customer details are: bank history, income, banking (Annual Quarterly
Balance AQB), stability, etc. The whole processing fee will be paid by banks. The internal credit
scoring will be done within the bank, while as for the external scoring bank has to pay towards
CIBIL. Bank will not charge the processing fee to the applicant. The credit score can be
improved considering the following principles:

1. Always pay your dues on time.

2. Always keep your credit balance low on your credit card.

3. Maintain a healthy mix of credit

4. Monitor your and your guarantor accounts balance frequently.

There are 4 main factors which mainly affect the score:

1. payment history,

2. high utilization of credit limit,

3. higher percentage of credit cards or personal finance and

4. many new accounts opened recently.

5.3 Example

Credit scoring methods used by various banks are somewhere similar. Only few parameters are
different.

As this information is internal and cannot be given publically, below are the common methods

40
used by almost all the banks.

Common methods used by banks:

1) Ratio analysis

2) Judgemental scoring Model

3) Statistical scoring Model

and many more.

Below are few models used for different types of loans:

Credit scoring model for Home loan

Parameters Score weights

Employment Type 10

Work experience/years in business or practice 5

Marital status 5

Age 5

Relationship with our bank 5

Residence Type 3

No. of years residing in current address 2

Monthly Average Disposable Income 5

EMI/NMI Ratio 15

Number of applicants 5

Parameters Score weights

Repayment mode 10

Purpose of Home Loan 5

LTV Ratio 15

41
CIBIL Score 10

Total 100

Key terms:

1.What Is the Loan-to-Value (LTV) Ratio?

The loan-to-value (LTV) ratio is an assessment of lending risk that financial institutions and
other lenders examine before approving a mortgage. Typically, assessments with high LTV
ratios are higher risk and, therefore, if the mortgage is approved, the loan costs the borrower
more.

Additionally, a loan with a high LTV ratio may require the borrower to purchase mortgage
insurance to offset the risk to the lender.

Loan-to-Value Ratio Formula and Calculation

Home buyers can easily calculate the LTV ratio on their home.

The process involves dividing the total mortgage loan amount into the total purchase price of the
home. For instance, a home with a purchase price of Rs.200,000 and a total mortgage loan for
Rs.180,000 results in a LTV ratio of 90%. Conventional mortgage lenders often provide better
loan terms to borrowers who have LTV ratios no higher than 80%.

An LTV ratio is calculated by dividing the amount borrowed by the appraised value of the
property, expressed as a percentage. For example, if you buy a home appraised at Rs.100,000 for
its appraised value and make a Rs.10,000 down payment, you will borrow Rs.90,000 resulting in
an LTV ratio of 90% (i.e., 90,000/100,000).

2.CIBIL Score

CIBIL Score is a three-digit numeric summary of your credit history. The score is derived using
the credit history found in the CIBIL Report (also known as CIR i.e Credit Information Report).
A CIR is an individual’s credit payment history across loan types and credit institutions over a
period of time. A CIR does not contain details of your savings, investments or fixed deposits.

42
The score weight of 65 and above is considered to be good and should be moved
further for assessment. But if the score comes in between 55-65 then the higher
authority comes into existence. And then final decision is with the higher authority.

5.4 Final Notes

In a new or emerging market, the operational, technical, business and cultural issues should be
considered with the implementation of the credit scoring models for retail loan products. The
operational issues relate to the use of the model and it is imperative that the staff and the
management of the bank understand the purpose of the model. Application scoring models
should be used for making credit decisions on new applications and behavioral models for retail
loan products to supervise existing borrowers for limiting the expansion or for marketing new
products. The technical issues relate to the development of proper infrastructure, maintenance
of historical data and software needed to build a credit scoring model for retail loan products
within the bank. The business issues relate to whether the soundness and safety of banks could
be achieved through the adoption of quantitative credit decision models, which would send a
positive impact in the banking sector. The cultural issues relate to making credit irrespective of
race, colour, sex, religion, marital status, age or ethnic origin. Further, models have to be
validated so as to ensure that the model performance is compatible in meeting the business as
well as regulatory requirements.

Thus, the above issues have to be considered while developing and implementing credit scoring
models for retail loan products.

5.5 Bibliography
o "Equifax Completes Acquisition of Australia's Leading Credit Information Company,
Veda Group Limited, for Total Consideration of USD$1.9 Billion". Equifax Australia. 25
February 2016. Retrieved 6 March 2018.
o ^ Anonymous (17 August 2010). "Credit reporting agencies". Retrieved 6 March 2018.
o ^ McGeorge, Noah. "fico ratings" (Atlanta). Marcus Arnold. Barff. Retrieved 28
February2013.

43
o ^ "Fragen & Antworten". Archived from the original on 1 January 2011. Retrieved 31
March 2011.
o ^ § 8 Abs 1 Z2 DSG
o ^ § 26 DSG
o ^ §§ 27 and 28 DSG
o ^ Jacomini, Leonardo (9 November 2017). "Score de crédito: o que é e como ele
funciona".
o ^ "Tudo sobre score de crédito" [All about credit score] (in Spanish). Archived from the
original on 29 May 2019. Retrieved 29 May 2019.
o ^ "Entenda Como Funciona o Score de Crédito do Consumidor". QuickBooks. 24 April
2015.
o ^ Equifax form (Canada)
o ^ TransUnion form (Canada)
o ^ Understanding Your Credit Report and Credit Score (Canada)
o ^ Ming, Cheang (2017). "FICO with Chinese characteristics: Nice rewards, but
punishing penalties". CNBC.
o ^ "Lecture: Court in the Akten | Sunday | Schedule 35th Chaos Communication
Congress". fahrplan.events.ccc.de.
o ^ Jump up to:a b "Reserve Bank of India". rbi.org.in.
o ^ "Reserve Bank of India". rbi.org.in.
o ^ "Reserve Bank of India". rbi.org.in.
o ^ "Introducing our newest service offering: Compuscore PSY". Retrieved 11
December2017.
o ^ Crib.lk (Sri Lanks)
o ^ "Betalningsanmärkning". kronofogden.se. Enforcement Authority. Retrieved 16
August2015.
o ^ Liljemalm, Anna (3 February 2015). "Så minskar skulder chansen att få
lägenhet" [Debts reduce chances to get an apartment]. gp.se. Göteborgs-Posten.
Retrieved 16 August 2015.
o ^ Olsson, Caroline (17 March 2008). "Fuglesang hamnar hos kronofogden" [Fuglesang
at the Swedish Enforcement Administration]. aftonbladet.se. Aftonbladet. Retrieved 16

44
 August2015.
o ^ https://www.equifax.co.uk/resources/loans_and_credit/credit-myths-the-truth-about-
credit.html
o ^ Ross, Katherine (27 December 2018). "Here's What Your Credit Score Is According to
an Experian Executive". TheStreet.
o ^ Hulland, Louise. "Credit Reference Agencies". BBC.
o ^ Report to the Congress on credit scoring and its effects on the availability and
affordability of credit
o ^ Jump up to:a b Credit-based insurance scores: Impacts on consumers of automobile
insuranceArchived 11 May 2009 at the Wayback Machine
o ^ "No Evidence of Disparate Impact in Texas Due to Use of Credit Information by
Personal Lines Insurers - III". iii.org.
o ^ An overview of consumer data and credit reporting
o ^ Jump up to:a b Wernau, Julie (28 April 2010). "TransUnion battling attempts to ban
employment credit checks". Chicago Tribune. Retrieved 23 October 2011.
o ^ Equifax Credit Report Dispute FAQs retrieved on 18 December 2008
o ^ Chase drops WaMu secured card, FICO score retrieved on 17 January 2009
o ^ President Obama on Final Passage of Wall Street Reform: An End to Bailouts, a
Beginning for Accountability published 15 July 2010, retrieved on 26 July 2010.
o ^ "Score Distributions". FICO.
o ^ "Credit Score in UAE". moneymall.ae. 'MoneyMall' leading personal finance portal in
UAE. Retrieved 22 November 2018.
o Abdou, H. (2009c). Genetic programming for credit scoring: The case of Egyptian public
sector banks. Expert Systems with Applications 36 (9): 11402-11417.
o Abdou, H., Pointon, J. (2009). Credit scoring and decision-making in Egyptian public
sector banks. International Journal of Managerial Finance 5 (4): 391-406.
o Altman, E. I. (1968). Financial Ratios, Discriminant Analysis and the Prediction of
Corporate Bankruptcy. The Journal of Finance XXIII (4): 589-609.
o Anderson, R. (2007). The Credit Scoring Toolkit: Theory and Practice for Retail Credit
Risk Management and Decision Automation.
o New York: Oxford University Press.

45
o Arminger, G., Enache, D., Bonne, T. (1997). Analyzing Credit Risk Data: A
Comparison of Logistic Discriminant, Classification Tree Analysis, and Feedforward
Networks. Computational Statistics 12 (2): 293-310.
o Baesens, B., Gestel, T. V., Viaene, S., Stepanova, M., Suykens, J., Vanthienen, J. (2003).
Benchmarking State-of-the-Art Classification Algorithms for Credit Scoring. Journal of
the Operational Research Society 54 (6): 627-635.
o Bailey, M. (2001). Credit scoring: the principles and practicalities. Kingswood, Bristol:
White Box Publishing.
o Beynon, M. J. (2005). Optimizing object classification under ambiguity/ignorance:
application to the credit rating problem. Intelligent Systems in Accounting, Finance and
Management 13(2). pp. 113-130.
o Bhatia, M. (2006). Credit Risk Management and Basel II: An implementation guide.
London: Risk B
o Bishop, C. M. (1995). Neural Networks for Pattern Recognition. New York: Oxford
University Press Inc.
o Blochlinger, A., Leippold, M. (2006). Economic Benefit of Powerful Credit Scoring.
Journal of Banking & Finance 30(3): 851-873.
o Boyle, M., Crook, J. N., Hamilton, R., Thomas, L. C. (1992). Methods for credit scoring
applied to slow payers. In Credit Scoring and Credit Control, Thomas, L. C., Crook, J.
N., Edelman, D. B., eds., Oxford University Press, Oxford, 75-90.
o Breiman, L., Friedman, J. H., Olshen, R. A., Stone, C. J. (1984). Classification and
Regression Trees. Belmont: The Wadsworth.
o Caire, D. (2004) Building Credit Scorecards for Small Business Lending in Developing
Markets. London: Bannock Consulting,
o Caouette, J. B., Altman, E. I., Narayanan, P. (1998). Managing Credit Risk: The Next
Great Financial Challenge. New York: John Wiley & Sons Inc.
o Chen, M., Huang, S. (2003). Credit scoring and rejected instances reassigning through
evolutionary computation techniques. Expert Systems with Applications 24(4): 433-441.
o Chuang, C. and Lin, R. (2009). Constructing a reassigning credit scoring model. Expert
Systems with Applications. 36 (2/1). pp.1685- 1694.
o Coffman, J. Y. (1986). The proper role of tree analysis in forecasting the risk behaviour

46
 of borrowers. MDS Reports 3, 4, 7 and 9. Management Decision Systems, Atlanta.
Cramer.
o Desai, V. S., Conway, D. G., Crook, J. N., Overstreet, G. A. (1997). Credit scoring
models in the credit union environment using neural networks and genetic algorithms.
IMA Journal of Mathematics Applied in Business and Industry 8 (4): 323-3463.
o Durand, D. (1941). Risk Elements in Consumer Instalment Financing, Studies in
Consumer Instalment Financing. New York: National Bureau of Economic Research.

47