INTRODUCTION TO INSURANCE

WEEK 10
Recap
•Insurance companies operations
➢ Underwriting
➢ Production
➢ Claim settlement
➢ Investments

2
Discussions
•Felix is a property claims adjustor for a large property
insurer. Janet is a policyholder who recently notified the
company that the roof of her home incurred substantial
damage because of a recent rainstorm and as a results
affected some items in her home. Janet owns her home and
is insured under a standard homeowners policy with no
special endorsements.
•What questions should Felix ask before the claim is
approved for payment by his company?

3
Focus
•Risk measurement
•Probability distribution
•Expected value
•Variance and standard deviation
•Skewness
•Coefficient of Variation
•Pooling of Losses
Risk Measurement
•Random variables
➢ A random variable is a variable whose outcome is uncertain.
➢For example, suppose a coin is to be flipped and the variable
X is defined to be equal to GH¢1 if heads appears and GH¢0 if
tails appear.
➢Then prior to the coin flip, the value of X is unknown; X is a
random variable.
➢Once the coin has been flipped and the outcome revealed,
the uncertainty about X is resolved, because the value of X is
then known
Probability Distributions
•Information about a random variable is summarized by the
random variable’s probability distribution
•A probability distribution identifies all the possible
outcomes for the random variable and the probability of the
outcomes
Probability Distributions
•For example
Possible outcomes for X Probability
GH¢ 1 0.5
GH¢ 0 0.5
•We can also describe probability distributions graphically
•On the horizontal axis, we graph the possible outcomes and on
the vertical axis, we graph the probability of a particular
outcomes
Expected Value (Mean)
•The expected value of a probability distribution provides
information about where the outcomes tend to occur, on
the average
•The mean or expected value is found by multiplying each
outcome by the probability of occurrence, and then
summing the resulting products
•Mathematically, expected value can be defined as
EV = ƩXiPi
Expected value
•For example, assume that an actuary estimates the
following probabilities of various losses for a certain risk:
Amount of Loss (Xi) Probability of Loss (Pi) Xi Pi
¢0 0.30 ¢0 * 0.30 = 0
¢360 0.50 ¢360 * 0.50 = 180
¢600 0.20 ¢600 * 0.20 = 120
ƩXi Pi = ¢300
Expected value
Expected value
•Assume the probability distribution for the cedi amount of
damages to your car during the coming year is given as:
Amount of Loss (Xi) Probability of Loss (Pi)
GH¢ 0 0.5
GH¢ 500 0.3
GH¢ 1,000 0.1
GH¢ 5,000 0.06
GH¢ 10,000 0.04
•Find the expected value of damages.
Application of Expected Value
•𝐸(𝑥) = Σ(𝑃𝑟𝑜𝑓𝑖𝑡)(𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑃𝑟𝑜𝑓𝑖𝑡) −
Σ(𝐶𝑜𝑠𝑡)(𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝐶𝑜𝑠𝑡)
• Examples in Gambling
➢In the game where a pair of dice are rolled, you can bet
$1 that the sum of the dice will be 2. The probability of
winning is 1/36. If you win, the profit is $30. What is the
expected value of your profit?
Application of Expected Value
•E(x) = ($30-1)(1/36) – ($1)(35/36) = - $0.17 (Loss)

•Example 2
➢In a state lottery, you pay $1 and pick a number from 000
to 999. If your number comes up, you win $600. The
probability of winning is 0.001. What is the expected value
of your profit? Is this a gain or loss?
Application of Expected Value
•𝐸(𝑥) = ($600 − 1)(0.001) − ($1)(0.999) = − $0.40 (Loss)
Application of Expected Value
•Example in Business
➢You are considering investing $10,000 into a start-up
company. You estimate that you have a 0.25 probability of
a $20,000 loss, a 0.20 probability of a $10,000 profit, a
0.15 probability of a $50,000 profit, and a 0.40 probability
of breaking even (no profit). What is the expected value of
the profit? Should you follow through on this investment?
Application of Expected Value
•The probability distribution will be

OUTCOME PROBABILITY

-$20,000 0.25

$10,000 0.20

$50,000 0.15

$0 0.40
Application of Expected Value
➢E(x) = ($10,000)(0.20) + ($50,000)(0.15) +($0)(0.40) –
($20,000)(0.25)
➢ = $4500 (Yes)
Application of Expected Value
•Your company is considering developing one of two cell phones.
Your development and market research teams provide you with
the following projections.
Cell phone A:
Cost of development: $2,500,000
Projected sales: 50% chance of net sales of $5,000,000
30% chance of net sales of $3,000,000
20% chance of net sales of $1,500,000
.
Application of Expected Value
•Cell phone B:
Cost of development: $1,500,000
Projected sales: 30% chance of net sales of $4,000,000
60% chance of net sales of $2,000,000
10% chance of net sales of $500,000

Which model should your company develop? Explain

Application of Expected Value
•Although cell phone A has twice the risk of losing $1
million, it has the greater expected value. So, using
expected value as a decision guideline, your company
should develop cell phone A.
Expected Value
•You take out a fire insurance policy on your home. The
annual premium is $300. In case of fire, the insurance
company will pay you $200,000. The probability of a house
fire in your area is 0.0002.
➢a. What is the expected value?
➢b. What is the insurance company’s expected value?
➢c. Suppose the insurance company sells 100,000 of these
policies. What can the company expect to earn?
Expected Value
Expected value = (0.0002)(199,700) + (0.9998)(−300) = −
$260.00
•a. The expected value over many years is −$260 per year. Of
course, your hope is that you will never have to collect on fi
re insurance for your home.
•b. The expected value for the insurance company is the
same, except the perspective is switched. Instead of − $260
per year, it is +$260 per year. Of this, the company must pay
a large percent for salaries and overhead cost.
Expected Value
•c. The insurance company can expect to gross $30,000,000
in premiums on 100,000 such policies. With a probability of
0.0002 for fire, the company can expect to pay on about 20
fires. This leaves a gross profit of $26,000,000.
Expected Value
•An insurance company sells a one-year term life insurance
policy to a 75-year old woman. The woman pays a premium of
$750. If she dies within 1 year, the insurance company will pay
$25,000 to her beneficiary. According to the 2014 Social Security
Actuarial Life Table, the probability that a 75-year old woman
will be alive in 1 year is 0.9746.
➢a. Find the expected value
➢b. Find the expected value of the insurance company’s profit
on the policy. Interpret the result.
➢Suppose the insurance company sells 200,000 of these
policies. What can the company expect to earn?
Expected Value
• a. 𝐸(𝑥) = ($25,000 − $750)(1 − 0.9746) − ($750)(0.9746) = −
$115
• b. 𝐸(𝑥) 𝑜𝑓 𝑃𝑟𝑜𝑓𝑖𝑡 = ($750)(0.9746) − ($25,000 − $750)(1 −
0.9746) = $115
➢This indicates that if the insurance company sells many
policies like this one, the company can expect on average to
earn a $115 profit per policy. (Gain)
Expected Value
•The insurance company can expect to gross $150,000,000 in
premiums on 200,000 such policies. With a probability of
0.0254 for death, the company can expect to pay on about
5,080 people. The cost will be $127,000,000
•This leaves a gross profit of $23,000,000.
Variance and Standard Deviation
•The variance of a probability distribution provides information about
the likelihood and magnitude by which a particular outcome from
the distribution will differ from the expected value
•In other words, variance measures the probable variation in
outcomes around the expected value.
•If a distribution has low variance, then the actual outcome is likely to
be close to the expected value.
•Conversely, if the distribution has a high variance, then it is more
likely that the actual outcome from the distribution will be far from
the expected value
Variance
•A high variance implies that outcomes are difficult to
predict
•Mathematically, the variance is denoted by

µ= expected value
Xi= possible outcomes
Pi= probability of outcomes
Standard Deviation
•A number that measures the concentration of the values
about their mean. The smaller the standard deviation
relative to the mean, the less the dispersion and the more
uniform the values. It is the square root of the variance.
•The standard deviation is mostly used to measure risk
•Denoted by
Variance and Standard deviation
•Find the variance and standard deviation
Amount of Loss Probability
0 0.30
360 0.50
600 0.20
Variance and standard deviation
Var (𝛿 2 ) = 0.30 (0 − 300)2 + 0.50 (360 − 300)2 + 0.20
(600 − 300)2
= 27,000 + 1,800 + 18,000
= 46,800

SD (δ) = 46800 = 216.33

Variance and standard deviation
Distribution 1 Distribution 2 Distribution 3

Loss outcome Prob Loss outcome Prob Loss outcome Prob

250 0.33 0 0.33 0 0.4

500 0.34 500 0.34 500 0.2

750 0.33 1000 0.33 1000 0.4

Variance and standard deviation
•Higher standard deviations, relative to the mean, are
associated with greater uncertainty of loss; therefore, risk is
higher.
Skewness
•Skewness measures the symmetry of the distribution. Many
loss exposures have skewed probability distributions.
•Recognizing this characteristics of skewed distributions is
important when assessing the likelihood (high or low
probability) of either low losses or large losses
Skewness
•The coefficient of skewness is given by;
Skewness
•For symmetric distribution, 𝛾 = 0
•The mean, median and the mode are all equal for
symmetric distributions.
•For positively skewed distributions, 𝛾 > 0
•For negatively skewed distributions, 𝛾 < 0
Symmetric distribution (𝛾=0)
Positively skewed distribution (𝛾>0)
Negatively skewed distribution (𝛾<0)
Coefficient of Variation (CV)
•CV is good for comparing variation between two groups. A high
CV indicates high variability in the data set.
•For same units data with different means, one is interested in
the size of the standard deviation relative to the mean. i.e

𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
➢CV =
𝑀𝑒𝑎𝑛
•For data with different units, standard deviations cannot be
directly compared.
Coefficient of Variation: same units data,
same means
•Comparing coefficient of variation for two discrete distributions
Distribution 1 Distribution 2
Outcome Prob Outcome Prob
GH¢250 0.33 GH¢0 0.33

GH¢500 0.34 GH¢500 0.34

GH¢750 0.33 GH¢1000 0.33

Standard deviations can be compared! Reason: Means coincide!

Coefficient of Variation: Same units data,
different means
•Comparing coefficient of variation for two discrete distributions
Distribution 1 Distribution 2
Outcome Prob Outcome Prob
GH¢50 0.33 GH¢0 0.33

GH¢150 0.34 GH¢500 0.34

GH¢750 0.33 GH¢1000 0.33

Standard deviations CANNOT be compared!

Reason: Means are different!
Coefficient of Variation: Different units
data
•Comparing coefficient of variation for two discrete distributions
Distribution 1 Distribution 2
Outcome Prob Outcome Prob
$ 50 0.33 GH¢0 0.33

$ 250 0.34 GH¢500 0.34

$ 750 0.33 GH¢1000 0.33

Standard deviations CANNOT be compared!

Reason: Data have different units!
Pooling of Losses
•Pooling is the spreading of losses incurred by the few over the
entire group, so that in the process, average loss is substituted
for actual loss
•It involves the grouping of a large number of exposure units so
that the law of large numbers can operate to provide a
substantially accurate prediction of future losses.
•Pooling implies
➢the sharing of losses by the entire group and
➢ prediction of future losses with some accuracy based on the
law of large numbers.
Pooling of Losses
•The primary purpose of pooling, or the sharing of losses, is to
reduce the variation in possible outcomes as measured by the
standard deviation or some other measure of dispersion, which
reduces risk.
•For example, two business owners each own an identical
storage building valued at $50,000. Assume there is a 10%
chance in any year that each building will be destroyed by a
peril, and that a loss to either building is an independent event.
➢Calculate the expected loss for each owner and the standard
deviation
Pooling of Losses
•Expected loss = $0 (0.90) + ($50,000)(0.10) = $5,000
•SD = 0.90(0 − $5000)2 + 0.10($50,000 − $5000)2
=$15,000
Pooling of Losses
•Suppose instead of bearing the risk of loss individually, the two
owners decide to pool (combine) their loss exposures, and each
agrees to pay an equal share of any loss that might occur. Under this
scenario, there are four possible outcomes:
Possible Outcome Prob
Neither building is destroyed 0.90 * 0.90 = 0.81
First building destroyed, second 0.10 * 0.90 = 0.09
building no loss
First building no loss, second building 0.90 * 0.10 = 0.09
destroyed
Both buildings are destroyed 0.10 * 0.10 = 0.01
Pooling of Losses
•So now the probability distribution will be
Loss Outcome Prob
$0 0.81
$25000 0.09
$25000 0.09
$50000 0.01

•Calculate the expected loss and standard deviation.

Pooling of Losses
Expected loss = 0.81 * $0 + 0.09 * $25,000 + 0.90 * $25,000 + 0.01 * $50,000 =
$5,000

SD =
0.81(0 − $5000)2 + 0.09 ($25000 − $5000) 2 +0.09 ($25000 − $5000)2 + 0.10($50,000 − $5000)2

SD = $10,607
Pooling of Losses
Thus, as additional individuals are added to the pooling
arrangement, the standard deviation continues to decline
while the expected value of the loss remains unchanged.
Pooling of Losses
•Suppose that Edmund and Irene enter a pooling arrangement. Each
of them has a loss distribution of

Loss Outcome Prob

$0 0.80
$20000 0.20

➢Calculate the expected loss and standard deviation before pooling

➢Calculate the expected loss and standard deviation after pooling.
➢Is pooling of losses beneficial? Why?
Costs of Pooling Arrangements
•Pooling arrangements reduce risk, but they involve costs:
➢Adding Participants
➢Marketing
➢Underwriting
➢Verifying Losses
➢Collecting Assessments