Unit BSA 4123 – Risk Models

Week 4-6 – Ruin Theory

www.strathmore.edu
Introduction
Here, we consider the aggregate claims 𝑆(𝑡) arising up to time t.
Let:
𝑁(𝑡) be the number of claims arising by time t
𝑋𝑖 is the amount of the i-th claim
Then:
𝑆 𝑡 = 𝑋1 + 𝑋2 + ⋯ + 𝑋𝑁(𝑡)
𝑁(𝑡) is called a Poisson process
𝑆(𝑡) is a compound Poisson process
We will also consider the concepts of:
1. a premium security loading, which is an additional amount charged on an insurance premium to reduce the likelihood of
an insurance company becoming ruined
2. the adjustment coefficient, a parameter associated with risk
3. the Lundberg’s inequality
Finally, we will consider the effect of changing parameter values on the probability of ruin for an insurance company before
considering the impact of introducing reinsurance.

www.strathmore.edu
Basic concepts(1)
1. Notation
A function 𝑓(𝑥) is described as being 𝑜(𝑥) as 𝑥 goes to zero, if:
𝑓(𝑥)
lim =0
𝑥→0 𝑥
For example, if we consider the function: 𝑔 𝑥 = 3𝑥 + 0.5𝑥 2 + 0.004𝑥 3
0.5𝑥 2 +0.004𝑥 3
This can be written as: 𝑔 𝑥 = 3𝑥 + 𝑜(𝑥) since → 0 as 𝑥 → 0
𝑥
Also note that 𝑜(𝑥) does not represent an actual number and so the following are all equivalent:
• 𝑐 ∗ 𝑜(𝑥) where c is a constant
• −𝑜(𝑥)
• 𝑜(𝑥)

www.strathmore.edu
Basic concepts(2)
Example:
Which of the following functions are 𝑜(𝑥) as x → 0
i. 𝑥2
ii. 𝑒𝑥
iii. 𝑒 −𝑥 − 1 + 𝑥

Solution:
i. Yes
ii. No
𝑥2 𝑥3
iii. Yes, if we expand 𝑒 −𝑥 as a power series and simplify we get 𝑒 −𝑥 = 1 − 𝑥 + − + ⋯ and therefore:
2! 3!
𝑒 −𝑥 −1+𝑥 𝑥 𝑥2
= − + ⋯ which solves to 𝑜(𝑥) as x → 0
𝑥 2! 3!

www.strathmore.edu
Basic concepts(3)
In the actuarial literature, the word ‘risk’ is often used instead of the phrase ‘portfolio of policies’.
Here, both terms will be used, so that ‘risk’ will mean either a single policy or a collection of policies.
We will focus on claims generated by a portfolio over successive time periods.
Some notation is needed:
𝑁(𝑡) is the number of claims generated by the portfolio in the time interval [0, 𝑡] for all 𝑡 ≥ 0
𝑋𝑖 is the amount of the i-th claim, 𝑖 = 1, 2, 3, …
𝑆(𝑡) is the aggregate claims in the time interval [0, 𝑡] for all 𝑡 ≥ 0
Therefore:
∞
𝑋𝑖 𝑖=1 is a sequence of random variables
𝑁(𝑡) 𝑡≥0 and 𝑆(𝑡) 𝑡≥0 are both families of random variables, one for each time 𝑡 ≥ 0 i.e. they are stochastic processes
It can be seen that:
𝑁(𝑡)
𝑆 𝑡 = σ𝑖=1 𝑋𝑖
where 𝑆 𝑡 = 0 if 𝑁 𝑡 = 0

www.strathmore.edu
Basic concepts(4)
The stochastic process 𝑆(𝑡) 𝑡≥0 as defined above is known as the aggregate claims process for the risk.
The random variables 𝑁(1) and 𝑆(1) represent the number of claims and the aggregate claims respectively from the
portfolio in the first unit of time.
The insurer of this portfolio will receive premiums from the policyholders. It is convenient at this stage to assume, as will be
assumed throughout this topic, that the premium income is received continuously and at a constant rate.
Define the following:
𝑐 = 𝑡ℎ𝑒 𝑟𝑎𝑡𝑒 𝑜𝑓 𝑝𝑟𝑒𝑚𝑖𝑢𝑚 𝑖𝑛𝑐𝑜𝑚𝑒 𝑝𝑒𝑟 𝑢𝑛𝑖𝑡 𝑡𝑖𝑚𝑒
so that the total premium income received in the time interval [0, 𝑡] is 𝑐𝑡 .
It will also be assumed that 𝑐 is strictly positive.

www.strathmore.edu
The surplus process(1)
Suppose that at time 0 the insurer has an amount of money set aside for this portfolio.
This amount of money is called the initial surplus and is denoted by 𝑈.
It will always be assumed that 𝑈 ≥ 0.
The insurer needs this initial surplus because the future premium income on its own may not be sufficient to cover the future
claims (here we are ignoring expenses).
The insurer’s surplus at any future time 𝑡 > 0 is a random variable since its value depends on the claims experience up to
time 𝑡.
The insurer’s surplus at time 𝑡 is denoted by 𝑈(𝑡).
We can therefore define the following formula for 𝑈 𝑡 :
𝑈 𝑡 = 𝑈 + 𝑐𝑡 − 𝑆(𝑡)
This means that the insurer’s surplus at time 𝑡 is the initial surplus plus the premium income up to time 𝑡 minus the
aggregate claims up to time 𝑡.
Notice that the initial surplus and the premium income are not random variables since they are determined before the risk
process starts.
The above formula is valid 𝑡 ≥ 0 with the understanding that 𝑈(0) is equal to 𝑈.
For a given value of 𝑡, 𝑈(𝑡) is a random variable because 𝑆(𝑡) is a random variable.
Hence 𝑈(𝑡) 𝑡≥0 is a stochastic process, which is known as the cash flow process or surplus process.

www.strathmore.edu
The surplus process(2)

The figure above shows one possible outcome of the surplus process.
Claims occur at times 𝑇1 , 𝑇2 , 𝑇3 , 𝑇4 𝑎𝑛𝑑𝑇5 . At these times, the surplus immediately falls by the amount of the claim.
Between claims, the surplus increases at constant rate 𝑐 per unit time.
The model being used for the insurer’s surplus incorporates many simplifications, as will any model of a complex real-life operation. Some
important simplifications are that it is assumed that:
• claims are settled as soon as they occur
• no interest is earned on the insurer’s surplus
• There are no expenses associated with the process
• The insurer cannot vary the premium rate 𝑐
Despite its simplicity, this model can give an interesting insight into the mathematics of an insurance operation.

www.strathmore.edu
The probability of ruin in continuous time(1)
It can be seen from the previous figure that the insurer’s surplus falls below zero as a result of the claim at time 𝑇3 .
When the surplus falls below zero, the insurer has run out of money and it is said that ruin has occurred.
In this simplified model, the insurer will want to keep the probability of this event, that is, the probability of ruin, as small as
possible, or at least below a predetermined value.
Ruin can be thought of as meaning insolvency; although determining whether or not an insurance company is insolvent is, in
practice, a more complex problem.
Another way of looking at the probability of ruin is to think of it as the probability that, at some future time, the insurance
company may need to provide more capital to finance this particular portfolio.
Now to be more precise. The following two probabilities are defined:
𝜓 𝑈 = 𝑃[𝑈 𝑡 < 0, 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 𝑡, 0 < 𝑡 < ∞]
𝜓 𝑈, 𝑡 = 𝑃[𝑈 𝜏 < 0, 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 𝜏, 0 < 𝜏 ≤ 𝑡]
Where:
𝜓 𝑈 is the probability of ultimate ruin (given initial surplus 𝑈)
𝜓 𝑈, 𝑡 is the probability of ruin within time 𝑡 (given initial surplus 𝑈).
These probabilities are sometimes referred to as the probability of ruin in infinite time and the probability of ruin in finite
time.

www.strathmore.edu
The probability of ruin in continuous time(2)
Here are some important logical relationships between these probabilities for 0 < 𝑡1 ≤ 𝑡2 < ∞ and for 0 ≤ 𝑈1 ≤ 𝑈2 :
𝜓 𝑈2 , 𝑡 ≤ 𝜓 𝑈1 , 𝑡 - the larger the initial surplus, the less likely it is that ruin will occur in a finite time period (1.1)
𝜓 𝑈2 ≤ 𝜓 𝑈1 - the larger the initial surplus, the less likely it is that ruin will occur in an unlimited time period (1.2)
𝜓 𝑈, 𝑡1 ≤ 𝜓 𝑈, 𝑡2 ≤ 𝜓 𝑈 - for a given initial surplus 𝑈, the longer the period considered when checking for ruin, the
more likely it is that ruin will occur (1.3)
lim 𝜓 𝑈, 𝑡 = 𝜓 𝑈 - the probability of ultimate ruin can be approximated by the probability of ruin within finite
𝑡→∞
time 𝑡 provided 𝑡 is sufficiently large (1.4)

Example:
What is lim 𝜓 𝑈, 𝑡 ?
𝑈→∞

Solution:
As the amount of initial surplus increases, ruin becomes less and less likely. So the limit is zero.

www.strathmore.edu
The probability of ruin in discrete time(1)
The two probabilities of ruin considered so far have been continuous time probabilities of ruin, so-called because they check
for ruin in continuous time.
In practice, it may be possible (or even desirable) to check for ruin only at discrete intervals of time.
For a given interval of time, denoted ℎ, the following two discrete time probabilities of ruin are defined:
𝜓ℎ 𝑈 = 𝑃[𝑈 𝑡 < 0, 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 𝑡, 𝑡 = ℎ, 2ℎ, 3ℎ, … ]
𝜓ℎ 𝑈, 𝑡 = 𝑃[𝑈 𝜏 < 0, 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 𝜏, 𝜏 = ℎ, 2ℎ, 3ℎ, … , 𝑡 − ℎ, 𝑡]
It is assumed for convenience in the definition of 𝜓ℎ 𝑈, 𝑡 that 𝑡 is an integer multiple of ℎ.
The figure below shows the same realisation of the surplus process as given for continuous time, but assuming now that the
process is checked only at discrete time intervals.
The black markers show the values of the surplus process at integer time intervals (i.e. ℎ = 1)
The black markers together with the white ones show the values of the surplus process at time intervals of length ½.

www.strathmore.edu
The probability of ruin in discrete time(2)

It can be seen from the figure above that in discrete time with ℎ = 1, ruin does not occur for this realisation of the surplus
process before time 5, but ruin does occur (at time 2½) in discrete time with ℎ = ½

www.strathmore.edu
The probability of ruin in discrete time(3)
Listed below are 7 relationships between different discrete time probabilities of ruin for 0 < 𝑡1 ≤ 𝑡2 < ∞ and 0 ≤ 𝑈1 ≤ 𝑈2 :
𝜓ℎ 𝑈2 , 𝑡 ≤ 𝜓ℎ 𝑈1 , 𝑡 - the larger the initial surplus, the less likely it is that ruin will occur in a finite discrete time period
(1.5)
𝜓ℎ 𝑈2 ≤ 𝜓ℎ 𝑈1 - the larger the initial surplus, the less likely it is that ruin will occur in an unlimited discrete time period
(1.6)
𝜓ℎ 𝑈, 𝑡1 ≤ 𝜓ℎ 𝑈, 𝑡2 ≤ 𝜓ℎ 𝑈 - for a given initial surplus 𝑈, the longer the period considered when checking for ruin, the
more likely it is that ruin will occur (1.7)
lim 𝜓ℎ 𝑈, 𝑡 = 𝜓ℎ 𝑈 - the probability of ultimate ruin can be approximated by the probability of ruin within finite
𝑡→∞
time 𝑡 provided 𝑡 is sufficiently large (1.8)
𝜓ℎ 𝑈, 𝑡 ≤ 𝜓 𝑈, 𝑡 - the more often we check for ruin, the more likely we are to find it; therefore since 𝜓 𝑈, 𝑡 involves
checking for ruin at all possible times, we would expect that it would be greater than 𝜓ℎ 𝑈, 𝑡 (1.9)
Intuitively, it is expected that the following two relationships are true since the probability of ruin in continuous time could be
approximated by the probability of ruin in discrete time, with the same initial surplus (𝑈), and time horizon (𝑡), provided that
ruin is checked for sufficiently often, ie provided ℎ is sufficiently small.
lim 𝜓ℎ 𝑈, 𝑡 = 𝜓 𝑈, 𝑡 - (1.10)
ℎ→0+
lim 𝜓ℎ 𝑈 = 𝜓 𝑈 - (1.11)
ℎ→0+

www.strathmore.edu
The Poisson and compound Poisson process(1)
∞
We now make some assumptions about the claim number process, 𝑁(𝑡) 𝑡≥0 , and the claim amounts, 𝑋𝑖 𝑖=1 .
The claim number process will be assumed to be a Poisson process, leading to a compound Poisson process 𝑆(𝑡) 𝑡≥0 for
aggregate claims. The assumptions made in this section will hold for the remainder of this topic.
The Poisson Process:
The term “Poisson process” describes the number of claims arising from a time period of length 𝑡.
If the number of claims 𝑁 arising from a single time period has a Poisson distribution with parameter 𝜆 then the number of
claims 𝑁(𝑡) which arise over a time period of length 𝑡 is a Poisson process, ie 𝑁(𝑡) has a Poisson distribution with parameter
𝜆𝑡
The Poisson process is an example of a counting process. Here, the number of claims arising from a risk is of interest.
Since the number of claims is being counted over time, the claim number process 𝑁(𝑡) 𝑡≥0 must satisfy the following
conditions:
i. 𝑁 0 = 0 , ie there are no claims at time 0
ii. for any 𝑡 > 0 , 𝑁(𝑡) must be integer valued
iii. when 𝑠 < 𝑡, 𝑁 𝑠 ≤ 𝑁(𝑡), ie the number of claims over time is non-decreasing
iv. when 𝑠 < 𝑡, 𝑁 𝑡 − 𝑁(𝑠) represents the number of claims occurring in the time interval (𝑠, 𝑡]

www.strathmore.edu
The Poisson and compound Poisson process(2)
The claim number process 𝑁(𝑡) 𝑡≥0 is defined to be a Poisson process with parameter 𝜆 if the following conditions are
satisfied:
I. 𝑁 0 = 0, and 𝑁 𝑠 ≤ 𝑁(𝑡) when 𝑠 < 𝑡
II. 𝑃 𝑁 𝑡 + ℎ = 𝑟 𝑁 𝑡 = 𝑟 = 1 − 𝜆ℎ + 𝑜(ℎ)
𝑃 𝑁 𝑡 + ℎ = 𝑟 + 1 𝑁 𝑡 = 𝑟 = 𝜆ℎ + 𝑜(ℎ)
𝑃 𝑁 𝑡 + ℎ > 𝑟 + 1 𝑁 𝑡 = 𝑟 = 𝑜(ℎ)
III. When 𝑠 < 𝑡, the number of claims in the time interval (𝑠, 𝑡] is independent of the number of claims up to time 𝑠
Condition (II) implies that there can be a maximum of one claim in a very short time interval ℎ. It also implies that the
number of claims in a time interval of length ℎ does not depend on when that time interval starts.
Example:
Explain how motor insurance claims could be represented by a Poisson process.
Solution:
The events in this case are occurrences of claim events (ie accidents, fires, thefts etc) or claims reported to the insurer.
The parameter 𝜆 represents the average rate of occurrence of claims (eg 50 per day), which we are assuming remains
constant throughout the year and at different times of day.
The assumption that, in a sufficiently short time interval, there can be at most one claim is satisfied if we assume that claim
events cannot lead to multiple claims (ie no motorway pile-ups etc).

www.strathmore.edu
The Poisson and compound Poisson process(3)
Time to the first claim
We now derive the distribution of the time to the first claim.
Let the random variable 𝑇1 denote the time of the first claim.
Then, for a fixed value of 𝑡, if no claims have occurred by time 𝑡, 𝑇1 > 𝑡 . Hence:
𝑃 𝑇1 > 𝑡 = 𝑃(𝑁 𝑡 = 0)
We know that:
𝑁(𝑡)~𝑃𝑜𝑖(𝜆𝑡)
Thus:
𝑒 −𝜆𝑡 (𝜆𝑡)0
𝑃 𝑁 𝑡 =0 = = 𝑒 −𝜆𝑡
0!
Therefore:
𝑃 𝑇1 ≤ 𝑡 = 1 − 𝑃 𝑇1 > 𝑡 = 1 − 𝑒 −𝜆𝑡
This is the distribution function of an exponential distribution
So that:
𝑇1 has an exponential distribution with parameter 𝜆
The time to the first claim in a Poisson process has an exponential distribution with parameter 𝜆

www.strathmore.edu
The Poisson and compound Poisson process(4)
Time between claims
For 𝑖 = 1, 2, 3, … let the random variable 𝑇𝑖 denote the time between the (𝑖 − 1)𝑡ℎ and the 𝑖𝑡ℎ claim
Then:
𝑃 𝑇𝑛+1 > 𝑡| σ𝑛𝑖=1 𝑇𝑖 = 𝑟 = 𝑃 σ𝑛+1 𝑛
𝑖=1 𝑇𝑖 > 𝑡 + 𝑟| σ𝑖=1 𝑇𝑖 = 𝑟

= 𝑃(𝑁 𝑡 + 𝑟 = 𝑛|𝑁 𝑟 = 𝑛)
= 𝑃(𝑁 𝑡 + 𝑟 − 𝑁(𝑟) = 0|𝑁 𝑟 = 𝑛)
We now use the independence of claim numbers in different time periods to remove the conditioning:
𝑃 𝑁 𝑡 + 𝑟 − 𝑁 𝑟 = 0 𝑁 𝑟 = 𝑛 = 𝑃(𝑁 𝑡 + 𝑟 − 𝑁(𝑟) = 0)
𝑃 𝑁 𝑡 + 𝑟 − 𝑁 𝑟 = 0 = 𝑃 𝑁 𝑡 = 0 = 𝑒 −𝜆𝑡
since the number of claims in a time interval of length 𝑟 does not depend on when that time interval starts
Thus, the inter-event times also have an exponential distribution with parameter 𝜆
The time between claims in a Poisson process has an exponential distribution with parameter 𝜆
Note that the inter-event time is independent of the absolute time i.e. the time until the next event has the same
distribution, irrespective of the time since the last event or the number of events that have already occurred.
This is referred to as the memoryless property of the exponential distribution.

www.strathmore.edu
The Poisson and compound Poisson process(5)
Example:
If reported claims follow a Poisson process with rate 5 per day (and the insurer has a 24-hour hotline), calculate:
i. the probability that there will be fewer than 2 claims reported on a given day
ii. the probability that another claim will be reported during the next hour

Solution:
(i) The expected number of claims reported on a given day is 5.
So the number of claims reported on a given day has a 𝑃𝑜𝑖𝑠𝑠𝑜𝑛(5) distribution and the probability that there will be fewer
than 2 claims is:
𝑒 −5 (5)0 𝑒 −5 (5)1
𝑃 𝑁 <2 = 𝑃 𝑁 =0 +𝑃 𝑁 =1 = + = 𝑒 −5 + 5𝑒 −5 = 0.04043
0! 1!
(ii) The waiting time until the next event has an 𝐸𝑥𝑝(5) distribution.
We need to find a probability using the exponential distribution. To do this, we can use the cumulative distribution function:
𝑃 𝑇 < 𝑡 = 1 − 𝑒 −𝜆𝑡
So the probability that there will be a claim (or several claims) during the next hour ( 1/24 of a day) is:
5
1 −
𝑃 𝑇< =1−𝑒 24 = 0.18806
24

www.strathmore.edu
The Poisson and compound Poisson process(6)
The Compound Poisson Process
We now combine the Poisson process for the number of claims with a claim amount distribution to give a compound Poisson
process for the aggregate claims.
The following three important assumptions are made:
• the random variables 𝑋𝑖 ∞ 𝑖=1 are independent and identically distributed
• the random variables 𝑋𝑖 ∞ 𝑖=1 are independent of 𝑁(𝑡) for all 𝑡 ≥ 0
• the stochastic process 𝑁(𝑡) 𝑡≥0 is a Poisson process whose parameter is denoted 𝜆
This last assumption means that for any 𝑡 ≥ 0, the random variable 𝑁(𝑡) has a Poisson distribution with parameter 𝜆𝑡, so
that:
𝑒 −𝜆𝑡 (𝜆𝑡)𝑘
𝑃 𝑁 𝑡 =𝑘 =
𝐾!

With these assumptions, the aggregate claims process 𝑆(𝑡) 𝑡≥0 is called a compound Poisson process with Poisson
parameter 𝜆.
By comparing the assumptions above with the assumptions in the Introduction, it can be seen that the connection between
the two is that if 𝑆(𝑡) 𝑡≥0 is a compound Poisson process with Poisson parameter 𝜆, then, for a fixed value of 𝑡 ≥ 0, 𝑆(𝑡) has
a compound Poisson distribution with Poisson parameter 𝜆𝑡.
Note the slight change in terminology here: ‘Poisson parameter 𝜆’ becomes ‘Poisson parameter 𝜆𝑡’ when a change is made
from the process to the distribution.

www.strathmore.edu
The Poisson and compound Poisson process(7)
The common distribution function of the 𝑋𝑖 ′𝑠will be denoted 𝐹(𝑥) and it will be assumed for the remainder of this topic that
𝐹 0 = 0 so that all claims are for positive amounts.
𝐹(𝑥) is defined to be 𝑃(𝑋 ≤ 𝑥).
In the continuous case we would find 𝐹(𝑥) by integrating the probability density function (PDF):
𝑥
𝐹 𝑥 = ‫׬‬−∞ 𝑓 𝑡 𝑑𝑡
The probability density function of the 𝑋𝑖 ′𝑠, if it exists, will be denoted 𝑓(𝑥) and the 𝑘𝑡ℎ moment about zero of the 𝑋𝑖 ′𝑠, if
it exists, will be denoted 𝑚𝑘 , so that:
𝑚𝑘 = 𝐸 𝑋𝑖𝑘 for 𝑘 = 1, 2, 3, …
Whenever the common moment generating function of the 𝑋𝑖 ′𝑠 exists, its value at the point 𝑟 will be denoted by 𝑀𝑋 𝑟 .
Where:
𝑀𝑋 𝑟 = 𝐸[𝑒 𝑟𝑋 ]
Since, for a fixed value of 𝑡, 𝑆(𝑡) has a compound Poisson distribution, it follows from Risk Theory that the process
𝑆(𝑡) 𝑡≥0 has:
Mean - 𝜆𝑡𝑚1
Variance - 𝜆𝑡𝑚2
𝑙𝑛𝑀𝑋 𝑟
moment generating function 𝑀𝑆(𝑡) 𝑟 , where: 𝑀𝑆(𝑡) 𝑟 = 𝑀𝑁(𝑡) 𝑙𝑛𝑀𝑋 𝑟 = 𝑒 𝜆𝑡(𝑒 −1)
= 𝑒 𝜆𝑡(𝑀𝑋 𝑟 −1)

www.strathmore.edu
The Poisson and compound Poisson process(8)
For the remainder of this topic the following (intuitively reasonable) assumption will be made concerning the rate of
premium income:
𝑐 > 𝜆𝑚1
so that the insurer’s premium income (per unit time) is greater than the expected claims outgo (per unit time).
Otherwise the insurer would be charging premiums that were less than the amount it expected to pay out in claims.
Please note that, in the real world this assumption may not always be true, especially during periods of competitive pressure
when premium rates are soft.

www.strathmore.edu
Probability of ruin in the short term(1)
If we know the distribution of the aggregate claims 𝑆(𝑡), we can often determine the probability of ruin for the discrete
model over a finite time horizon directly (without reference to the models), by looking at the cashflows involved
Example:
The claims arising during each year from a particular type of annual insurance policy are assumed to follow a normal
distribution with mean 0.7𝑃 and standard deviation 2.0𝑃, where 𝑃 is the annual premium.Claims are assumed to arise
independently. Insurers assess their solvency position at the end of each year.
A small insurer with an initial surplus of KES 100,000 expects to sell 100 policies at the beginning of the coming year in
respect of identical risks for an annual premium of KES 5,000. The insurer incurs expenses of 0.2𝑃 at the time of writing each
policy. Calculate the probability that the insurer will prove to be insolvent at the end of the coming year. Ignore interest.
Solution:
Using the information given, the insurer’s surplus at the end of the coming year will be:
𝑈 1 = 𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑠𝑢𝑟𝑝𝑙𝑢𝑠 + 𝑝𝑟𝑒𝑚𝑖𝑢𝑚𝑠 − 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑠 − 𝑐𝑙𝑎𝑖𝑚𝑠
= 100000 + 100 ∗ 5000 − 100 ∗ 0.2 ∗ 5000 − 𝑆(1)
= 500,000 − 𝑆(1)
The distribution of 𝑆(1) is:
2
𝑆 1 ~𝑁 100 ∗ 0.7 ∗ 5000, 100 ∗ 2.0 ∗ 5000 = 𝑁[350000, 1000002 ]

www.strathmore.edu
Probability of ruin in the short term(2)
So the probability that the surplus will be negative is:
500000−350000
𝑃 𝑈 1 < 0 = 𝑃 𝑆 1 > 500000 = 𝑃 𝑁 350000, 1000002 > 500000 = 1 − Φ = 1 − 0.93319 =
100000
0.067
Example:
If the insurer expects to sell 200 policies during the second year for the same premium and expects to incur expenses at the
same rate, calculate the probability that the insurer will prove to be insolvent at the end of the second year.
Solution:
Using the information given, the insurer’s surplus at the end of the coming year will be:
𝑈 2 = 𝑖𝑛𝑖𝑡𝑖𝑎𝑙 𝑠𝑢𝑟𝑝𝑙𝑢𝑠 + 𝑝𝑟𝑒𝑚𝑖𝑢𝑚𝑠 − 𝑒𝑥𝑝𝑒𝑛𝑠𝑒𝑠 − 𝑐𝑙𝑎𝑖𝑚𝑠
= 100000 + 100 ∗ 5000 + 200 ∗ 5000 − 100 ∗ 0.2 ∗ 5000 − (200 ∗ 0.2 ∗ 5000) − 𝑆(2)
= 1,300,000 − 𝑆(2)
The distribution of 𝑆(2) is:
2
𝑆 2 ~𝑁 300 ∗ 0.7 ∗ 5000, 300 ∗ 2.0 ∗ 5000 = 𝑁[1050000, 1732052 ]
So the probability that the surplus will be negative is:
1300000−1050000
𝑃 𝑈 2 < 0 = 𝑃 𝑆 2 > 1300000 = 𝑃 𝑁 1050000, 1732052 > 1300000 = 1 − Φ =1−
173205
Φ(1.443) = 0.074

www.strathmore.edu
Probability of ruin in the short term(3)
Example:
The number of claims from a portfolio of policies has a Poisson distribution with parameter 30 per year. The individual claim
amount distribution is lognormal with parameters 𝜇 = 3 and 𝜎 2 = 1.1.
The rate of premium income from the portfolio is 1,200 per year. If the insurer has an initial surplus of 1,000, estimate the
probability that the insurer’s surplus at time 2 will be negative, by assuming that the aggregate claims distribution is
approximately normal.
Solution:
First, we need the mean and variance of the aggregate claims in a two-year period. The expected number of claims will be 60.
So, the mean and variance are (using the formulae for the first two moments of a lognormal distribution):
𝐸𝑆 2 = 60𝑒 3+1.1/2 = 2,088.80
𝑣𝑎𝑟 𝑆 2 = 60𝑒 2(3+1.1) = 218,457
Ruin will occur if 𝑆(2) is greater than the initial surplus plus premiums received. So we want:
3400−2088.8
𝑃 𝑆 2 > 1000 + 2 ∗ 1200 ≈ 𝑃 𝑁 0, 1 > = 1 − Φ 2.8053 = 0.0025
218457

The probability of ruin is approximately 0.25%.

www.strathmore.edu
Premium security loadings(1)
So far, we have used 𝑐 to denote the rate of premium income per unit time, independent of the claims outgo.
In some circumstances it is more useful to think of the rate of premium income as being related to the rate of claims outgo.
For the insurer to survive, the rate at which premium income comes in needs to be greater than the rate at which claims are
paid out.
If this is not true, the insurer is certain to be ruined at some point.
Sometimes 𝑐 will be written as:
𝑐 = (1 + 𝜃)𝜆𝑚1
where 𝜃(> 0) is the premium loading factor
The security loading is the percentage by which the rate of premium income exceeds the rate of claims outgo. So, for the
Poisson process outlined above, we have:
𝑐 = 1 + 𝜃 𝐸 𝑆 = (1 + 𝜃)𝜆𝑚1
where 𝜃 is the security loading.
𝜃 is also sometimes called the ‘relative security loading’. It might typically be a figure such as 0.2, ie 20%.
The insurer will need to adopt a positive security loading when pricing policies, in order to cover expenses, profit,
contingency margins and so on.
Note that this does not mean that ruin is impossible. It is quite possible for the actual claims outgo to exceed substantially its
expected value. So even in this situation the insurer’s probability of ruin is non-zero.

www.strathmore.edu
Mean, variance and MGF of the total claim amount(1)
For a compound Poisson process 𝑆(𝑡) the mean and variance of the total claim amount are given by:
Mean:
𝐸𝑆 𝑡 = 𝜆𝑡𝐸(𝑋)

Variance:
𝑣𝑎𝑟 𝑆 𝑡 = 𝜆𝑡𝐸(𝑋 2 )

MGF:
𝑙𝑛𝑀𝑋 𝑟
𝑀𝑆(𝑡) 𝑟 = 𝑀𝑁(𝑡) 𝑙𝑛𝑀𝑋 𝑟 = 𝑒 𝜆𝑡(𝑒 −1)
= 𝑒 𝜆𝑡(𝑀𝑋 𝑟 −1)

www.strathmore.edu
A Technicality(1)
In the next section a technical result will be needed concerning 𝑀𝑋 𝑟 (the moment generating function of the individual claim
amount distribution), which, for convenience, will be presented here.
It will be assumed throughout the remainder of this topic that there is some number 𝛾 (0 < 𝛾 ≤ ∞) such that 𝑀𝑋 𝑟 is finite for all
𝑟 < 𝛾 and:
lim 𝑀𝑋 𝑟 = ∞ ……………………. (1)
𝑟→𝛾 −
For example, if the 𝑋𝑖′ 𝑠 have a range bounded above by some finite number, then 𝛾 will be ∞; if the 𝑋𝑖′ 𝑠 have an exponential
distribution with parameter 𝛼, then 𝛾 will be equal to 𝛼
Suppose for example that claim amounts have a continuous uniform distribution on the interval (0,10) so that they are bounded
above by 10. Then the moment generating function of the claim distribution is (from the Tables):
𝑒 10𝑟 − 1
𝑀𝑋 𝑟 =
10𝑟
This is defined for all positive values of 𝑟, and so in this case 𝛾 = ∞. We can see that as 𝑟 → ∞, the limit of the MGF is infinite. If the
claim distribution is 𝐸𝑥𝑝(𝛼), the MGF (as stated in the Tables) is:
𝑟 𝛼
𝑀𝑋 𝑟 = (1 − )−1 =
𝛼 𝛼−𝑟
this tends to infinity as 𝑟 tends to 𝛼 from below
In the next section the following result will be needed:
lim (𝜆𝑀𝑋 𝑟 − 𝑐𝑟) = ∞ ……………………….. (2)
𝑟→𝛾 −
If 𝛾 is finite, (2) follows immediately from (1)
This is because 𝜆, 𝑐 and 𝑟 would all have finite values in the limit

www.strathmore.edu
A Technicality(2)
Now it will be shown that (2) holds when 𝛾 is infinite.
First note that there is a positive number, 𝜀 say, such that:
𝑃 𝑋𝑖 > 𝜀 > 0
The reason for this is that all claim amounts are positive.
So, if we pick a small enough number ( 𝜀 = 0.01 maybe), we’re bound to get a proportion of claims whose amount exceeds
this.
This probability will be denoted by 𝜋. Then:
𝑀𝑋 𝑟 ≥ 𝑒 𝑟𝜀 𝜋
This follows by considering the claims below and above 𝜀:
𝑀𝑋 𝑟 = 𝐸 𝑒 𝑟𝑋 = 𝐸 𝑒 𝑟𝑋 |𝑋 ≤ 𝜀 𝑃 𝑋 ≤ 𝜀 + 𝐸 𝑒 𝑟𝑋 |𝑋 > 𝜀 𝑃 𝑋 > 𝜀 ≥ 0 + 𝑒 𝑟𝜀 𝜋
Hence:
lim (𝜆𝑀𝑋 𝑟 − 𝑐𝑟) ≥ lim 𝜆𝑒 𝑟𝜀 𝜋 − 𝑐𝑟 = ∞
𝑟→∞ 𝑟→∞
Here the 𝑒 𝑟𝜀 term is tending to +∞, while the −𝑐𝑟 term is tending to −∞. Remember that in such cases the exponential
term always ‘wins’. So the limit is +∞.

www.strathmore.edu
The adjustment coefficient and Lundberg’s inequality(1)
We will now look at the probability of ruin and introduce the adjustment coefficient, a parameter associated with risk. The
letters 𝑅 and 𝑟 will be used interchangeably for the adjustment coefficient.
Lundberg’s inequality
Lundberg’s inequality states that:
𝜓 𝑈 ≤ 𝑒 −𝑅𝑈
where 𝑈 is the insurer’s initial surplus and 𝜓 𝑈 is the probability of ultimate ruin.
𝑅 is a parameter associated with a surplus process known as the adjustment coefficient. Its value depends upon the
distribution of aggregate claims and on the rate of premium income.
Before defining 𝑅 the importance of the result and some features of the adjustment coefficient will be illustrated.
Note that if we can find a value for 𝑅, then Lundberg’s inequality tells us that we can find an upper bound for the probability
of ruin. This is a very useful result.
The figure below shows a graph of both 𝑒 −𝑅𝑈 and 𝜓 𝑈 against 𝑈 when claim amounts are exponentially distributed with
mean 1, and when the premium loading factor is 10%.

www.strathmore.edu
The adjustment coefficient and Lundberg’s inequality(2)

It can be seen that, for large values of 𝑈, 𝜓(𝑈) is very close to the upper bound, so that 𝜓(𝑈) ≈ 𝑒 −𝑅𝑈
In actuarial literature, 𝑒 −𝑅𝑈 is often used as an approximation to 𝜓(𝑈) .
𝑅 can be interpreted as measuring risk. The larger the value of 𝑅, the smaller the upper bound for 𝜓(𝑈) will be.
Hence, 𝜓(𝑈) would be expected to decrease as 𝑅 increases.
𝑅 is a function of the parameters that affect the probability of ruin, and 𝑅 ’𝑠 behaviour as a function of these parameters can
be observed.
Note that 𝑅 is an inverse measure of risk. Larger values of 𝑅 imply smaller ruin probabilities, and vice versa.

www.strathmore.edu
The adjustment coefficient and Lundberg’s inequality(3)
The figure below shows a graph of 𝑅 as a function of the loading factor, 𝜃 , when:
(i) the claim amount distribution is exponential with mean 10, and
(ii) all claims are of amount 10

Note that in both cases, 𝑅 is an increasing function of 𝜃.

This is not surprising as 𝜓(𝑈) would be expected to be a decreasing function of 𝜃, and since 𝜓(𝑈) ≈ 𝑒 −𝑅𝑈 any factor causing
a decrease in 𝜓(𝑈) would cause 𝑅 to increase.

www.strathmore.edu
The adjustment coefficient and Lundberg’s inequality(4)
Question:
Why would 𝜓(𝑈) be expected to be a decreasing function of 𝜃?
Solution:
𝜃 is the security loading, and 𝜓(𝑈) is the probability of ruin for a fixed level of surplus 𝑈.
If 𝜃 increases the premiums we charge will increase, and we should become more secure, ie the probability of ruin should
fall.
Note also that:
The value of 𝑅 when claim amounts are exponentially distributed is less than the value of 𝑅 when all claim amounts are 10.
This result should not be surprising.
Both claim amount distributions have the same mean, but the exponential distribution has greater variability. Greater
variability is associated with greater risk, and hence a larger value 𝜓(𝑈) would be expected for the exponential distribution,
and a lower value of 𝑅.
This example illustrates that 𝑅 is affected by the premium loading factor and by the characteristics of the individual claim
amount distribution. 𝑅 is now defined and shown, in general, to encapsulate all the factors affecting a surplus process.

www.strathmore.edu
The adjustment coefficient – compound Poisson process(1)
The surplus process depends on:
• the initial surplus
• the aggregate claims process, and
• the rate of premium income.
The adjustment coefficient is a parameter associated with a surplus process which takes account of two of these factors:
• aggregate claims, and
• premium income.
The adjustment coefficient gives a measure of risk for a surplus process.
When aggregate claims are a compound Poisson process, the adjustment coefficient is defined in terms of:
• the Poisson parameter
• the moment generating function of individual claim amounts, and
• the premium income per unit time

www.strathmore.edu
The adjustment coefficient – compound Poisson process(2)
The adjustment coefficient, denoted 𝑅, is defined to be the unique positive root of:
𝜆𝑀𝑋 𝑟 − 𝜆 − 𝑐𝑟 = 0 ……………………………………………(1)
So that:
𝜆𝑀𝑋 𝑅 = 𝜆 + 𝑐𝑅 ………………………………………………….(2)
Note that equation (1) implies that the value of the adjustment coefficient depends on the Poisson parameter, the individual
claim amount distribution and the rate of premium income.
However, writing 𝑐 = (1 + 𝜃)𝜆𝑚1 gives:
𝜆𝑀𝑋 𝑟 = 𝜆 + 𝑟(1 + 𝜃)𝜆𝑚1
𝑀𝑋 𝑟 = 1 + 𝑟(1 + 𝜃)𝑚1
𝑀𝑋 𝑟 = 1 + (1 + 𝜃)𝑚1 𝑟
so that 𝑅 is independent of the Poisson parameter and simply depends on the loading factor, 𝜃 , and the individual claim
amount distribution.

www.strathmore.edu
The adjustment coefficient – compound Poisson process(3)
Example:
An insurer knows from past experience that the number of claims received per month has a Poisson distribution with mean 15, and that claim
amounts have an exponential distribution with mean 500. The insurer uses a security loading of 30%. Calculate the insurer’s adjustment
coefficient and give an upper bound for the insurer’s probability of ruin, if the insurer sets aside an initial surplus of 1,000.
Solution:
The equation for the adjustment coefficient is:
𝑀𝑋 𝑟 = 1 + (1 + 𝜃)𝑚1 𝑟
We have that:
1
𝑋~ exp , 𝑠𝑜 𝑡ℎ𝑎𝑡 𝑀𝑋 𝑟 = (1 − 500𝑟)−1
500

𝜃 = 0.3
𝑚1 = 𝐸 𝑋 = 500
Substituting these into the equation:
(1 − 500𝑟)−1 = 1 + 1.3 500 𝑟 = 1 + 650𝑟
1 = (1 − 500𝑟)(1 + 650𝑟)
1 = 1 + 650𝑟 − 500𝑟 − 325000𝑟 2
150
0 = 150 − 325000𝑟 → 𝑟 = = 0.000461538
325000

From Lundberg’s inequality: 𝜓 𝑈 ≤ 𝑒 −𝑟𝑈 ≤ 𝑒 −0.000461538∗1000 = 0.63031

www.strathmore.edu
The adjustment coefficient – compound Poisson process(4)
Example:
Write down the equation for the adjustment coefficient for personal accident claims if 90% of claims are for KES 10,000 and 10% of claims are for KES
25,000, assuming a proportional security loading of 20%. Show that this equation has a solution in the range 0.00002599 < 𝑅 < 0.00002601
Solution:
The adjustment coefficient satisfies:
𝑀𝑋 𝑟 = 1 + (1 + 𝜃)𝑚1 𝑟
The distribution of the individual claim sizes 𝑋 is:
10,000 𝑤𝑖𝑡ℎ 𝑝𝑟𝑜𝑏 0.9
𝑋=ቊ
25,000 𝑤𝑖𝑡ℎ 𝑝𝑟𝑜𝑏 0.1
So:
𝐸 𝑋 = 0.9 10000 + 0.1 25000 = 11500
𝑀𝑋 𝑟 = 𝐸 𝑒 𝑅𝑋 = σ 𝑒 𝑅𝑋 𝑃 𝑋 = 𝑥 = 0.9𝑒 10000𝑅 + 0.1𝑒 25000𝑅
The security loading is 𝜃 = 0.2
So the equation for the adjustment coefficient is:
0.9𝑒 10000𝑅 + 0.1𝑒 25000𝑅 = 1 + 1 + 0.2 ∗ 11500𝑅
0.9𝑒 10000𝑅 + 0.1𝑒 25000𝑅 = 1 + 13800𝑅
At R = 0.00002599: 1 + 13800𝑅 − 0.9𝑒 10000𝑅 + 0.1𝑒 25000𝑅 = 0.000035
At R = 0.00002601: 1 + 13800𝑅 − 0.9𝑒 10000𝑅 + 0.1𝑒 25000𝑅 = −0.000018
Since there is a reversal of signs (and we are dealing with a continuous function), the difference must be zero at some point between these two values,
i.e. there is a solution of the equation in the range 0.00002599 < 𝑅 < 0.00002601

www.strathmore.edu
The adjustment coefficient – compound Poisson process(5)
If the equation for 𝑅 has to be solved numerically, it is useful to have a rough idea of 𝑅 ’𝑠 value. The previous equation (2) can be used to find a simple
upper bound for 𝑅 as follows:
𝜆 + 𝑐𝑅 = 𝜆𝑀𝑋 𝑅
∞
= 𝜆 ‫׬‬0 𝑒 𝑅𝑥 𝑓 𝑥 𝑑𝑥
∞ 𝑅𝑥 2
> 𝜆 ‫׬‬0 (1 + 𝑅𝑥 + )𝑓 𝑥 𝑑𝑥
2!
∞ 1
> 𝜆 ‫׬‬0 (1 + 𝑅𝑥 + 𝑅2 𝑥 2 )𝑓 𝑥 𝑑𝑥
2
1 2
= 𝜆(1 + 𝑅𝑚1 + 𝑅 𝑚2 )
2
This inequality is true because all the terms in the series 𝑒 𝑅𝑥 are positive. So 𝑒 𝑅𝑥 must always be greater than the total of the first few terms.
So we have:
1
𝜆 + 𝑐𝑅 > 𝜆(1 + 𝑅𝑚1 + 2 𝑅2 𝑚2 )
1
𝑐 − 𝜆𝑚1 𝑅 > 2 𝑅2 𝜆𝑚2
1
𝑐 − 𝜆𝑚1 > 2 𝑅𝜆𝑚2
2 𝑐−𝜆𝑚1
𝑅< 𝜆𝑚2

If we assume that 𝑐 = (1 + 𝜃)𝜆𝑚1 we have that:

2 1+𝜃)𝜆𝑚1 −𝜆𝑚1 2𝜃𝑚1
𝑅< →𝑅<
𝜆𝑚2 𝑚2

Therefore, if the value of 𝑅 is small, then it should be very close to this upper bound since the approximation to 𝑒 𝑅𝑥 should be sufficient

www.strathmore.edu
The adjustment coefficient – compound Poisson process(6)
Example:
Find an upper limit for the adjustment coefficient in the previous example, and comment on your answer
Solution:
We have that:
𝐸 𝑋 = 11500
𝐸 𝑋 2 = σ 𝑥 2 𝑃 𝑋 = 𝑥 = 0.9 100002 + 0.1 250002 = 152,500,000
So:
2𝜃𝑚1 2∗0.2∗11500
𝑅< = = 0.0000302
𝑚2 152500000

This is a reasonable initial estimate, compared with the correct value of approximately 0.000026.

www.strathmore.edu
The adjustment coefficient – a lower bound for R(1)
A lower bound for 𝑅 can be derived when there is an upper limit, say 𝑀, to the amount of an individual claim.
For example, if individual claim amounts are uniformly distributed on (0,100), then 𝑀 = 100.
The lower bound is found by applying the inequality:
𝑥 𝑅𝑀 𝑥
𝑒 𝑅𝑥 ≤ 𝑒 +1− 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 𝑀 ………………………(3)
𝑀 𝑀
This inequality is proved through the series expansion of 𝑒 𝑅𝑀 :
𝑥 𝑅𝑀 𝑥 𝑥 ∞ (𝑅𝑀)𝑗 𝑥
𝑒 +1− = σ +1−
𝑀 𝑀 𝑀 𝑗=0 𝑗! 𝑀
𝑗 𝑗−1 𝑥
∞ 𝑅 𝑀
=1+ σ𝑗=1
𝑗!

∞ 𝑅𝑥 𝑗
≥1+ σ𝑗=1 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 𝑀 since 𝑥 𝑗 ≤ 𝑀𝑗−1 𝑥 if 0 ≤ 𝑥 ≤ 𝑀
𝑗!

= 𝑒 𝑅𝑥
Inequality (3) can be used to show that:
1 𝑐
𝑅> log( )
𝑀 𝜆𝑚1

when individual claim amounts have a continuous distribution on (0, 𝑀)

This is the lower bound for 𝑅 that we are trying to find.

www.strathmore.edu
The adjustment coefficient – a lower bound for R(2)
The starting point is the equation defining 𝑅:
𝜆 + 𝑐𝑅 = 𝜆𝑀𝑋 𝑅
𝑀
𝜆 + 𝑐𝑅 = 𝜆 ‫׬‬0 𝑒 𝑅𝑥 𝑓 𝑥 𝑑𝑥
𝑀 𝑥 𝑥
≤ 𝜆 ‫׬‬0 ( 𝑒 𝑅𝑀 + 1 − )𝑓 𝑥 𝑑𝑥
𝑀 𝑀
𝜆 𝑅𝑀 𝜆
= 𝑒 𝑚1 + 𝜆 − 𝑚
𝑀 𝑀 1
We can then rearrange this to obtain:
𝜆 𝑅𝑀 𝜆
𝜆 + 𝑐𝑅 ≤ 𝑒 𝑚1 + 𝜆 − 𝑚1
𝑀 𝑀
𝜆 𝑅𝑀 𝜆
𝑐𝑅 ≤ 𝑒 𝑚1 − 𝑚1
𝑀 𝑀
𝑐 1 1 𝑅𝑀 𝑅𝑀 2 𝑅𝑀 3 𝑅𝑀 𝑅𝑀 2
≤ [𝑒 𝑅𝑀 −1] = [1 + + + + ⋯ − 1] = 1 + + +⋯
𝜆𝑚1 𝑅𝑀 𝑅𝑀 1 2! 3! 2 3!
𝑅𝑀 𝑅𝑀 2
<1+ + + ⋯ =𝑒 𝑅𝑀
1! 2!
Therefore:
𝑐 𝑐
< 𝑒 𝑅𝑀 → 𝑙𝑜𝑔 < 𝑅𝑀
𝜆𝑚1 𝜆𝑚1
So:
1 𝑐
𝑅> 𝑙𝑜𝑔
𝑀 𝜆𝑚1
1 (1+𝜃)𝜆𝑚1
and for 𝑐 = (1 + 𝜃)𝜆𝑚1 we have that: 𝑅 > 𝑙𝑜𝑔
𝑀 𝜆𝑚1
1
𝑅> log(1 + 𝜃)
𝑀

www.strathmore.edu
The effect of changing parameter values on ruin probabilities(1)
Recall the following definitions for ruin probabilities:
𝜓 𝑈 = 𝑃[𝑈 𝜏 < 0, 𝜏 > 0] 𝑖. 𝑒. 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑢𝑙𝑡𝑖𝑚𝑎𝑡𝑒 𝑟𝑢𝑖𝑛
𝜓 𝑈, 𝑡 = 𝑃[𝑈 𝜏 < 0, 0 < 𝜏 < 𝑡] 𝑖. 𝑒. 𝑝𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑟𝑢𝑖𝑛 𝑖𝑛 𝑓𝑖𝑛𝑖𝑡𝑒 𝑡𝑖𝑚𝑒 𝑡
We will now discuss the effect of changing parameter values on 𝜓 𝑈 and 𝜓 𝑈, 𝑡 .
Some features of 𝜓 𝑈 and 𝜓 𝑈, 𝑡 will be illustrated by a series of numerical examples.
In these examples the same basic assumptions will be made as in previous sections. In particular, it will be assumed that the
aggregate claims process is a compound Poisson process.
In addition, it will be assumed throughout that:
• the Poisson parameter for the number of claims is 1 (a)
• the expected value of an individual claim is 1 (b)
• individual claims have an exponential distribution (c)
The implication of Assumption (a) is that the unit of time has been chosen to be such that the expected number of claims in a unit of
time is 1. Hence 𝜓 𝑈, 500 is the probability of ruin (given initial surplus U ) over a time period in which 500 claims are expected. The
actual number of claims over this time period has a Poisson distribution (with parameter 500) and could take any non-negative
integer value.
The implication of Assumption (b) is that the monetary unit has been chosen to be equal to the expected amount of a single claim.
Hence 𝜓 20,500 is the probability of ruin (over a time period in which 500 claims are expected) given an initial surplus equal to 20
times the expected amount of a single claim.
The advantage of using an exponential distribution for individual claims (Assumption (c)) is that both 𝑒 −𝑅𝑈 and 𝜓 𝑈 can be
calculated for these examples.

www.strathmore.edu
A formula for 𝝍(𝑼) when 𝑭(𝒙) is the exponential distribution(1)
The formula for 𝜓(𝑈) when individual claims amounts are exponentially distributed with mean 1, and when the premium
loading factor is 𝜃, is given by the following result:
𝐹 𝑥 = 1 − 𝑒 −𝑥
−𝜃𝑈
1
𝜓 𝑈 = 𝑒 1+𝜃
1+𝜃
You are not required to derive or memorise this result for this unit
This result has only been presented in order to illustrate how, for this particular distribution, the ultimate ruin probability is
affected by changes in parameter values.
Question:
Is 𝜓(𝑈, 𝑡) an increasing or decreasing function of 𝑡?
Solution:
𝜓(𝑈, 𝑡) is the probability of ruin at some point between times 0 and 𝑡.
This should increase with time since the longer the time period, the more chance there is of ruin.
It should be intuitively obvious that 𝜓 𝑈, 𝑡1 < 𝜓 𝑈, 𝑡2 for 𝑡1 < 𝑡2 since if a scenario produces ruin before time 𝑡1 , then ruin
has also occurred before time 𝑡2 .

www.strathmore.edu
A formula for 𝝍(𝑼) when 𝑭(𝒙) is the exponential distribution(2)
The figure below shows a graph of 𝜓(15, 𝑡) for 0 ≤ 𝑡 ≤ 500.
The premium loading factor 𝜃 = 0.1 so that the premium income per unit time is 1.1.
Also shown in the figure are 𝜓(15) (dotted line) and 𝑒 −15𝑅 (solid line) for this portfolio. These last two values are shown as
lines parallel to the time axis since their values are independent of time.
Here:
• 𝜓(15, 𝑡) has been worked out using a numerical method not described here
−𝜃𝑈
1
• 𝜓(15) has been calculated using the equation 𝜓 𝑈 = 𝑒 1+𝜃 introduced earlier
1+𝜃
• 𝑒 −15𝑅 is the upper bound given by Lundberg’s inequality.

www.strathmore.edu
A formula for 𝝍(𝑼) when 𝑭(𝒙) is the exponential distribution(3)
Example:
Calculate the value of 𝜓(15)
Solution:
−𝜃𝑈
1
𝜓 𝑈 = 𝑒 1+𝜃
1+𝜃
−0.1∗15
1
𝜓 15 = 𝑒 1.1 = 0.23248
1.1

Example:
Calculate the value of 𝑒 −15𝑅
Solution:
𝛼𝜃 1∗0.1
𝑅 is worked out from 𝑅 = =
1+𝜃 1.1
0.1
−15∗
𝑒 −15𝑅 = 𝑒 1.1 = 0.2557
The features of note in the previous figure are:
i. 𝜓(15, 𝑡) is an increasing function of 𝑡
ii. for small values of 𝑡, 𝜓(15, 𝑡) increases very quickly (its value doubles as t increases from 25 to 50 and doubles again as t increases from
50 to 100)
iii. for larger values of 𝑡, 𝜓(15, 𝑡) increases less quickly and approaches asymptotically the value of 𝜓(15)

www.strathmore.edu
A formula for 𝝍(𝑼) when 𝑭(𝒙) is the exponential distribution(4)
General reasoning should help you to understand (ii) and (iii).
You would expect a much higher probability of ruin before time 50 than before time 25 since the overall performance of the
fund could easily change in such a short time period.
However, if premium rates are expected to be profitable in the long term, then at time 400, say, a significant surplus will have
built up in most cases and so the probability of ruin at time 425 won’t be that much higher than at time 400.
We are assuming here that accumulated surpluses stay in the fund and are not, for example, distributed to shareholders.

www.strathmore.edu
Ruin probability as a function of initial surplus(1)
Question:
Is 𝜓(𝑈, 𝑡) an increasing or decreasing function of 𝑈?
Solution:
Decreasing. The bigger the initial surplus, the less chance there should be of ruin.
The figure below shows values of 𝜓(𝑈, 𝑡) for 𝛼 and for three values of the initial surplus, 𝑈 = 15, 20, 𝑎𝑛𝑑 25. The premium
loading factor is 𝜃 = 0.1 as in the previous figure. For 𝑈 = 15 the graph of 𝛼 is as in the previous figure.

www.strathmore.edu
Ruin probability as a function of initial surplus(2)
Example:
Calculate the value of 𝜓(20)
Solution:
−𝜃𝑈
1
𝜓 𝑈 = 𝑒 1+𝜃
1+𝜃
1 −0.1∗20
𝜓 20 = 𝑒 1.1 = 0.14756
1.1

The features of note in this figure are:

i. The graphs all have the same general shape
ii. Increasing the value of 𝑈 decreases the value of 𝜓(U, 𝑡) for any value of 𝑡
iii. each of the three graphs approaches an asymptotic limit as 𝑡 increases. Note that 𝜓 20 = 0.14756 and 𝜓 25 =
0.0937
𝜓 20 is a non-increasing function of 𝑈
An increase in 𝑈 represents an increase in the insurer’s surplus without any corresponding increase in claim amounts. Thus,
an increase in 𝑈 represents an increase in the insurer’s security and hence will reduce the probability of ruin.

www.strathmore.edu
Ruin probability as a function of premium loading(1)
Question:
Is 𝜓(𝑈, 𝑡) an increasing or decreasing function of 𝜃?
Solution:
Decreasing. If everything else remains unchanged, then increasing the premium income will reduce the probability of ruin.
The figure below shows values of 𝜓(15, 𝑡) for 0 ≤ 𝑡 ≤ 500 and for three values of the premium loading factor, 𝜃 =
0.1, 0.2 𝑎𝑛𝑑 0.3.
The features of note in this figure are:
i. the graphs of 𝜓(15, 𝑡) all have the same general shape
ii. increasing the value of 𝜃 decreases the value of 𝜓(15, 𝑡) for any given value of 𝑡; this is in fact true for any value of 𝑈,
and is an obvious result since an increase in 𝜃 is equivalent to an increase in the rate of premium income with no change
in the aggregate claims process
iii. it can be seen that when 𝜃 = 0.1, 0.2 𝑎𝑛𝑑 0.3, 𝜓(15, 𝑡) is more or less constant for 𝑡 greater than about 150.
For 𝑡1 < 𝑡2 , the difference 𝜓 15, 𝑡1 − 𝜓(15, 𝑡2 )represents the probability that ruin occurs between times 𝑡1 and 𝑡2 . This
point will be discussed further in the next Section.
It is clear by general reasoning that 𝜓(𝑈) must be a non-increasing function of 𝜃. In the case of exponential individual claim
amounts, 𝜓(𝑈) is a decreasing function of 𝜃.
𝑑 −1 −2
𝜓 𝑈 =− 1+𝜃 𝜓 𝑈 −𝑈 1+𝜃 𝜓(𝑈)
𝑑𝜃

www.strathmore.edu
Ruin probability as a function of premium loading(2)

www.strathmore.edu
Ruin probability as a function of the Poisson parameter(1)

The figure above shows 𝜓(15, 10) as a function of 𝜆 for three values of the premium loading factor, 𝜃 = 0.1, 0.2 𝑎𝑛𝑑 0.3.
We considering the following two risks:
Risk 1: aggregate claims are a compound Poisson process with Poisson parameter 1 and 𝐹 𝑥 = 1 − 𝑒 −𝑥 . The premium
income per unit time to cover this risk is (1 + 𝜃).
Risk 2: aggregate claims are a compound Poisson process with Poisson parameter 0.5 and 𝐹 𝑥 = 1 − 𝑒 −𝑥 . The premium
income per unit time to cover this risk is 0.5(1 + 𝜃).
The unit of time is taken to be one year. The only difference between these risks is that twice as many claims are expected
each year under Risk 1. This is reflected in the two premiums.

www.strathmore.edu
Ruin probability as a function of the Poisson parameter(2)
Consider Risk 2 over a new time unit equivalent to two years. Then the distribution of aggregate claims and the premium income per unit time
are now identical to the corresponding quantities for Risk 1.
Hence, the probability of ruin over an infinite time span is the same for both risks. The solid line in the figure below shows an outcome of the
surplus process for Risk 1 when 𝜃 = 0.1. The dotted line shows the same surplus process when the unit of time is two years.
This illustrates that any outcome of the surplus process that causes ultimate ruin for Risk 1 will also cause ultimate ruin for Risk 2. There is thus
no difference in the probability of ultimate ruin for these two risks. It is only the time (in years) until ruin that will differ.
Measuring times in years, the probability of ruin by time 1 for Risk 1 is the same as the probability of ruin by time 2 for Risk 2. For example, the
value of 𝜓(15, 10) when 𝜆 = 50 is the same as the value of 𝜓(15, 500) when 𝜆 = 1

www.strathmore.edu
Reinsurance and ruin(1)
One of the options open to an insurer who wishes to reduce the variability of aggregate claims from a risk is to effect
reinsurance.
This is a form of insurance in which an insurance company obtains insurance cover from other insurance companies
(reinsurers) against the risk of losses.
A reduction in variability would be expected to increase an insurer’s security, and hence reduce the probability of ruin.
A reinsurance arrangement could be considered optimal if it minimises the probability of ruin.
As it is difficult to find explicit solutions for the probability of ruin, the effect of reinsurance on the adjustment coefficient will
be considered instead. If a reinsurance arrangement can be found that maximises the value of the adjustment coefficient, the
upper bound for the probability of ultimate ruin will be minimised. As the adjustment coefficient is a measure of risk, it
seems a reasonable objective to maximise its value.
We consider the effect on the adjustment coefficient of proportional and of excess of loss reinsurance arrangements.
Throughout this section we will use the notation:
𝑋 = 𝑖𝑛𝑑𝑖𝑣𝑖𝑑𝑢𝑎𝑙 𝑐𝑙𝑎𝑖𝑚 𝑎𝑚𝑜𝑢𝑛𝑡,
𝑌 = 𝑎𝑚𝑜𝑢𝑛𝑡 𝑝𝑎𝑖𝑑 𝑏𝑦 𝑡ℎ𝑒 𝑑𝑖𝑟𝑒𝑐𝑡 𝑖𝑛𝑠𝑢𝑟𝑒𝑟 and
𝑍 = 𝑎𝑚𝑜𝑢𝑛𝑡 𝑝𝑎𝑖𝑑 𝑏𝑦 𝑡ℎ𝑒 𝑟𝑒𝑖𝑛𝑠𝑢𝑟𝑒𝑟

www.strathmore.edu
Proportional reinsurance(1)
Under proportional reinsurance, the reinsurer covers an agreed proportion of each risk and the reinsurance premium is in
proportion to this risk ceded.
For example, the reinsurer might agree to pay 30% of each claim. The insurer would then pay 70% of the claim amount.
If the insurer retains a proportion 𝛼 then:
𝑌 = 𝛼𝑋
𝑍 = (1 − 𝛼)𝑋
𝐸 𝑌 = 𝐸 𝛼𝑋 = 𝛼𝐸(𝑋)
𝐸 𝑍 = 𝐸 1 − 𝛼 𝑋 = 1 − 𝛼 𝐸(𝑋)

Example:
Write down expressions for 𝑣𝑎𝑟(𝑌) and 𝑣𝑎𝑟(𝑍)
Solution:
𝑣𝑎𝑟 𝑌 = 𝑣𝑎𝑟 𝛼𝑋 = 𝛼 2 𝑣𝑎𝑟(𝑋)
𝑣𝑎𝑟 𝑍 = 𝑣𝑎𝑟 1 − 𝛼 𝑋 = 1 − 𝛼 2 𝑣𝑎𝑟(𝑋)

www.strathmore.edu
Proportional reinsurance(2)
Question:
Consider the following insurer:
The number of claims has a Poisson distribution with parameter 30 per year. The individual claim amount distribution is lognormal with parameters 𝜇 =
3 and 𝜎 2 = 1.1. The rate of premium income from the portfolio is 1,200 per year. The insurer has an initial surplus of 1,000.
This insurer is investigating the possibility of using proportional reinsurance. It has approached a reinsurer, who uses a security loading of 50% to
calculate its reinsurance premiums. If the insurer decides to reinsure 20% of each risk in the portfolio, estimate the effect the reinsurance will have on its
probability of ruin at Time 2. You can assume that the aggregate claim distribution is approximately normal.
Solution:
We first need to calculate the reinsurance premium. Since the reinsurer takes responsibility for 20% of each risk, and uses a loading factor of 50%, the
reinsurance premium (per annum) will be:
𝑅𝑃 = 1 + 𝜃𝑅 𝜆𝛼𝑚1 = 1.5 ∗ 30 ∗ 0.2 ∗ 𝑒 3.55 = 313.32
So over a two year period, the insurer will pay 626.64 for the reinsurance.
We now use 𝑆𝑛𝑒𝑡 (2) for the insurer’s aggregate payments (net of reinsurance). We need the mean and variance of 𝑆𝑛𝑒𝑡 (2), which are, using the formulae
for a compound Poisson distribution as before:
𝐸[𝑆𝑛𝑒𝑡 (2)] = 60 ∗ 0.8 ∗ 𝑒 3.55 = 1,671.04
𝑣𝑎𝑟[𝑆𝑛𝑒𝑡 (2)] = 60 ∗ 0.82 ∗ 𝑒 8.2 = 139,812.49
So, ruin will occur if:
𝑆𝑛𝑒𝑡 2 > 1000 + 2400 − 626.64 = 2,773.36
Using a normal distribution approach, we have:
2773.36−1671.04
𝑃[𝑆𝑛𝑒𝑡 2 > 2773.36] = 𝑃 𝑁 0,1 > = 1 − Φ 2.9481 = 0.0016
139812.49

www.strathmore.edu
Proportional reinsurance(3)
From the above question, we can infer as follows:
The reinsurance has reduced the probability of ruin to some extent, i.e. from about 0.25% (which was calculated earlier) to
about 0.16%.
However, this result is probably quite sensitive to the assumptions made (as we are near the tail of the normal distribution),
and slightly different assumptions might give us very different results.
We will also want to look at the effect of reinsurance on profitability. As we are paying a reinsurance premium, it is likely that
the overall effect on profitability is negative (although the effect on security is positive, as we have seen).
There is likely to be a trade-off between security and profitability here.

www.strathmore.edu
Excess of loss reinsurance(1)
Under excess of loss reinsurance, the cost to an insurer of a large claim, a cluster of claims arising from a single event (such as
an explosion) or claims over a given period (such as a catastrophe) is capped with the liability above a certain level being
passed to a reinsurer.
In most cases the reinsurer’s maximum liability is limited, but we will assume there is no such limit for now. We refer to the
capped amount as the retention of the insurer.
For example, the insurer might retain the first ten million of claims arising during the year, with the reinsurer paying any
excess above this amount. If the claim amount turns out to be less than ten million, then the insurer will pay this amount in
full and the reinsurer will pay nothing. However, if the claim amount is twelve million, then the insurer will pay ten million
and the reinsurer will pay the excess of two million.
We can apply the same type of logic as used for proportional reinsurance if the insurer decides to buy excess of loss
reinsurance. You might like to think about the effect on the probability of ruin if the insurer in the previous example
purchases excess of loss reinsurance with an individual retention of 2,000, say, and a security loading of 50% as before.
If we have a retention limit M , and no upper limit, then:
𝑋 𝑖𝑓 𝑋 < 𝑀
𝑌=ቊ
𝑀 𝑖𝑓 𝑋 ≥ 𝑀
0 𝑖𝑓 𝑋 < 𝑀
𝑍=ቊ
𝑋 − 𝑀 𝑖𝑓 𝑋 ≥ 𝑀
Or alternatively:
𝑌 = min(𝑋, 𝑀)
𝑍 = max(0, 𝑋 − 𝑀)

www.strathmore.edu
Excess of loss reinsurance(2)
Question:
Calculate 𝐸(𝑌) if 𝑋 has an exponential distribution with parameter 0.01, and the insurer has an excess of loss reinsurance
arrangement with retention limit 𝑀.
Solution:
The formula for an expectation is ‫𝑥𝑑 𝑥 𝑓𝑥 ׬‬
We have to calculate 𝐸(𝑌) by carrying out two integrals, to allow for the two different ranges of 𝑋:
𝑀 ∞
𝐸 𝑌 = ‫׬‬0 𝑥𝑓 𝑥 𝑑𝑥 + ‫𝑥𝑑 𝑥 𝑓𝑀 𝑀׬‬
𝑀 ∞
𝐸 𝑌 = ‫׬‬0 0.01𝑥𝑒 −0.01𝑥 𝑑𝑥 + ‫ 𝑀׬‬0.01𝑀𝑒 −0.01𝑥 𝑑𝑥
Using integration by parts:
0.01𝑥 −0.01𝑥 𝑀 𝑀 0.01𝑥 −0.01𝑥 0.01 −0.01𝑥 ∞
= − 𝑒 − ‫׬‬0 − 𝑒 𝑑𝑥 + −𝑀 𝑒
0.01 0 0.01 0.01 𝑀
1 𝑀
𝑀
= −𝑥𝑒 −0.01𝑥 0 − 𝑒 −0.01𝑥 + −𝑀𝑒 −0.01𝑥 ∞ 𝑀
0.01 0
1 1 1
= −𝑀𝑒 −0.01𝑀 − 𝑒 −0.01𝑀 + + 𝑀𝑒 −0.01𝑀
0.01 0.01 0.01
1
= (1 − 𝑒 −0.01𝑀 )
0.01

= 100(1 − 𝑒 −0.01𝑀 )

www.strathmore.edu
Excess of loss reinsurance(3)
Question:
Calculate 𝑣𝑎𝑟(𝑍) (in terms of 𝑀 ) if 𝑋~𝑈(0,100) where the insurer has an excess of loss reinsurance arrangement with retention
limit 𝑀, 0 < 𝑀 < 100
Solution:
To find 𝑣𝑎𝑟(𝑍), we need to find 𝐸[𝑍 2 ] since 𝑣𝑎𝑟 𝑍 = 𝐸 𝑍 2 − 𝐸(𝑍) 2

𝑀 100
𝐸 𝑍 2 = ‫׬‬0 02 𝑓 𝑥 𝑑𝑥 + ‫𝑀׬‬ 𝑥 − 𝑀 2 𝑓 𝑥 𝑑𝑥
1
From the Tables, the pdf of 𝑈(0,100) is
100
So:
100
2 100 𝑥−𝑀 2 𝑥−𝑀 3 100−𝑀 3
𝐸 𝑍 = ‫𝑀׬‬ 𝑑𝑥 = =
100 300 𝑀 300

We now need 𝐸 𝑍 :
100
100 (𝑥−𝑀) 𝑥−𝑀 2 100−𝑀 2
𝐸 𝑍 = ‫𝑀׬‬ 𝑑𝑥 = =
100 200 𝑀 200

Therefore:
2
100−𝑀 3 100−𝑀 2 100−𝑀 3 100−𝑀 4
𝑣𝑎𝑟 𝑍 = − = −
300 200 300 40,000

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(1)
Consider the effect of proportional reinsurance with retention 𝛼 on the insurer’s adjustment coefficient.
We will assume that the insurer’s premium income per unit time, before payment of the reinsurance premium, will be
written as (1 + 𝜃)𝜆𝑚1 , which represents the expected aggregate claims per unit time for the compound Poisson process
with a loading factor 𝜃.
We will also assume that the reinsurance premium is calculated as (1 + 𝜉)(1 − 𝛼)𝜆𝑚1 . Since the reinsurer pays proportion
1 − 𝛼 of each claim, (1 − 𝛼)𝜆𝑚1 represents the reinsurer’s expected claims per unit time.
Thus, 𝜉 is the premium loading factor used by the reinsurer. Hence, the insurer’s premium income, net of reinsurance, is:
[ 1 + 𝜃 − 1 + 𝜉 1 − 𝛼 ]𝜆𝑚1
The insurer will charge a premium of:
(1 + 𝜃)𝜆𝐸 𝑋 = (1 + 𝜃)𝜆𝑚1
The reinsurer will charge a premium of:
(1 + 𝜉)𝜆𝐸(𝑍)
But: 𝐸 𝑍 = 1 − 𝛼 𝐸 𝑋 = (1 − 𝛼)𝑚1
So the reinsurer’s premium is:
(1 + 𝜉)𝜆(1 − 𝛼)𝑚1
And therefore, the net premium received by the insurer is the difference:
[ 1 + 𝜃 − 1 + 𝜉 1 − 𝛼 ]𝜆𝑚1

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(2)
It will also be assumed that:
𝜉≥𝜃
If this were not true, it would be possible for the insurer to pass the entire risk on to the reinsurer and to make a certain
profit.
This of course ignores commission, expenses and other adjustments to the theoretical risk premium.
For the insurer’s premium income, net of reinsurance, to be positive:
1 + 𝜃 > (1 + 𝜉)(1 − 𝛼)
Thus:
(𝜉−𝜃)
𝛼>
(1+𝜉)

Example:
What range of values is possible for 𝛼 if 𝜃 = 0.2 and 𝜉 = 0.4?
Solution:
(𝜉−𝜃)
𝛼>
(1+𝜉)
(0.4−0.2)
𝛼> = 0.1429
(1.4)

So the possible range for 𝛼 is: 0.1429 < 𝛼 ≤ 1

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(3)
There is, however, a more important constraint on the insurer.
The insurer’s net of reinsurance premium income per unit time must exceed the expected aggregate claims per unit time.
Otherwise ultimate ruin is certain.
Net of reinsurance, the insurer’s expected aggregate claims per unit time are 𝛼𝜆𝑚1 .
Thus:
1+𝜃 − 1+𝜉 1−𝛼 >𝛼
Which solves to:
𝜃
𝛼 >1−
𝜉

Example:
So what is the range of possible values of 𝛼 now, given the figures in the previous question?
Solution:
𝜃
𝛼 >1−
𝜉
0.2
𝛼 >1− = 0.5
0.4
So:
0.5 < 𝛼 ≤ 1

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(4)
𝜃
The equation 𝛼 > 1 − specifies the insurer’s minimum retention level since:
𝜉
𝜃 𝜉−𝜃
1− ≥ 𝑤ℎ𝑒𝑛 𝜃 ≤ 𝜉
𝜉 1+𝜉

If the premium loading factors are equal, then inequality becomes 𝛼 > 0. In this case there exists a risk sharing arrangement
and any retention level is possible.
If, however, 𝜉 > 𝜃 then the insurer has to retain part of the risk.
Same Loadings:
First consider the case where both the insurer and the reinsurer use 𝜃 as the premium loading factor.
The adjustment coefficient will be found as a function of the retention level 𝛼 , when 𝐹 𝑥 = 1 − 𝑒 −0.1𝑥 .
The distribution of the insurer’s individual claims net of reinsurance is exponential with parameter 0.1/𝛼.
This can be seen by noting that if 𝑌 = 𝛼𝑋, then:
𝑦
𝑃 𝑌≤𝑦 =𝑃 𝑋< = 1 − 𝑒 −0.1𝑦/𝛼
𝛼
Note that the assumptions for the claims process and the adjustment coefficient equation apply equally well in the presence
of reinsurance, provided that we use the net premium and the net claim amounts in the adjustment coefficient equation.

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(5)
Question:
What will be the general equation for 𝑅, the direct insurer’s adjustment coefficient, when there is reinsurance?
Solution:
From the previous material, we know that the equation for R is: 𝜆 + 𝑐𝑟 = 𝜆𝑀𝑋 (𝑟)
With reinsurance this will become: 𝜆 + 𝑐𝑛𝑒𝑡 𝑟 = 𝜆𝑀𝑌 (𝑟) for the direct insurer
But we know that:
𝑐𝑛𝑒𝑡 = 1 + 𝜃 𝜆𝐸 𝑋 − (1 + 𝜉)𝜆𝐸(𝑍)
So, the equation for 𝑅 becomes:
𝜆 + [ 1 + 𝜃 𝜆𝐸 𝑋 − 1 + 𝜉 𝜆𝐸 𝑍 ]𝑟 = 𝜆𝑀𝑌 (𝑟)
Or
1 + [ 1 + 𝜃 𝐸 𝑋 − 1 + 𝜉 𝐸 𝑍 ]𝑟 = 𝑀𝑌 (𝑟)
In the case of proportional reinsurance this is:
1 + [ 1 + 𝜃 𝐸 𝑋 − 1 + 𝜉 (1 − 𝛼)𝐸 𝑋 ]𝑟 = 𝑀𝑌 (𝑟)

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(6)
Hence the equation defining 𝑅 is:
𝜆 + 𝑐𝑅 = 𝜆𝑀𝑋 (𝑅)
0.1𝑥
∞ 0.1 −
𝜆 + (1 + 𝜃)𝜆10𝛼𝑅 = 𝜆 ‫׬‬0 𝑒 𝑅𝑋 𝑒 𝛼 𝑑𝑥
𝛼
1
→ 1 + 1 + 𝜃 10𝛼𝑅 =
1−10𝛼𝑅
And therefore:
𝜃
𝑅= 𝑓𝑜𝑟 0 < 𝛼 ≤ 1
1+𝜃 10𝛼

It can be seen that 𝑅 is a decreasing function of 𝛼.

This is sensible as the larger the retention 𝛼, the larger the risk for the insurer and so 𝜓(𝑈) would be expected to increase,
and R to decrease, with 𝛼.

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(7)
Different Loadings:
Now consider what happens when 𝜉 > 𝜃.
We will assume that:
• The individual claim amount distribution is 𝐹 𝑥 = 1 − 𝑒 −0.1𝑥
• The insurer’s premium loading factor is 𝜃 = 0.1

Case A: 𝝃 = 𝟎. 𝟐
Suppose first that the reinsurer’s premium loading factor is 𝜉 = 0.2, so that the insurer’s (net) premium income per unit time
is (12𝛼 − 1)𝜆.
𝜃
The equation 𝛼 > 1 − shows that the insurer must retain at least 50% of each claim. Hence, a value of 𝛼 will be sought in
𝜉
the interval [0.5,1] that maximises the value of 𝑅. The equation defining 𝑅 is:
𝜆
𝜆 + (12𝛼 − 1)𝜆𝑅 =
1−10𝛼𝑅
Or:
1
1 + (12𝛼 − 1)𝑅 =
1−10𝛼𝑅

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(8)
And therefore:
1
1 + (12𝛼 − 1)𝑅 =
1−10𝛼𝑅
We have that:
1 + 12𝛼𝑅 − 𝑅 1 − 10𝛼𝑅 = 1
1 − 10𝛼𝑅 + 12𝛼𝑅 − 120𝛼 2 𝑅 2 − 𝑅 + 10𝛼𝑅 2 = 1
2𝛼𝑅 − 120𝛼 2 𝑅 2 − 𝑅 + 10𝛼𝑅 2 = 0
10𝛼 − 120𝛼 2 𝑅 2 + 2𝛼 − 1 𝑅 = 0
10𝛼 − 120𝛼 2 𝑅 + 2𝛼 − 1 = 0
(2𝛼−1)
𝑅= for 0.5 < 𝛼 ≤ 1
(120𝛼 2 −10𝛼)
If we differentiate 𝑅 with respect to 𝛼 (using the quotient rule for differentiation) and find the value of 𝛼 when the derivative
is zero, we end up with the following quadratic expression:
24𝛼 2 − 24𝛼 + 1 = 0
The roots of this quadratic are 0.9564 and 0.0436, and so the turning point which is of interest is 0.9564.
Consider the following values:
𝛼 = 0.5, 𝑅 = 0; 𝛼 = 0.9564, 𝑅 = 0.00911; 𝛼 = 1, 𝑅 = 0.00909
This shows that 𝑅 has a maximum in [0.5,1] at 0.9564.

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(9)
Case B: 𝝃 = 𝟎. 𝟑
The value of 𝛼 is now found that maximises 𝑅 when the reinsurer’s premium loading factor is 0.3.
The calculations are very similar to the previous case.
The insurer’s net premium income is (13𝛼 − 2)𝜆, so that the equation defining 𝑅 is:
𝜆
𝜆 + (13𝛼 − 2)𝜆𝑅 =
1−10𝛼𝑅
Or:
1
1 + (13𝛼 − 2)𝑅 =
1−10𝛼𝑅
Which leads to the solution:
(3𝛼−2)
𝑅= for 2/3 < 𝛼 ≤ 1
(130𝛼 2 −20𝛼)

If we differentiate 𝑅 with respect to 𝛼 (using the quotient rule for differentiation) and find the value of 𝛼 when the
derivative is zero, we end up with the following quadratic expression:
39𝛼 2 − 52𝛼 + 4 = 0
The roots of this quadratic are 0.0820 and 1.2514, and so there are no turning points in the interval [0.67, 1].
And 𝑅 as a function of 𝛼 in this interval increases from 0 at 𝛼 = 2/3 to 0.00909 at 𝛼 = 1. Thus, the value of 𝛼 which
maximises the adjustment coefficient is 1

www.strathmore.edu
Maximising the adjustment coefficient under proportional reinsurance(10)
It is not always possible to increase the value of the adjustment coefficient by effecting reinsurance.
Note that when an insurer effects reinsurance, this reduces the variability of the insurer’s aggregate claims.
A reduction in variability is associated with an increase in the value of the adjustment coefficient.
However, when 𝜉 > 𝜃, the insurer’s premium loading factor, net of reinsurance, decreases, and the value of the adjustment
coefficient is expected to decrease with the loading factor.
When the reinsurer’s premium loading factor was 0.3, the reduction in the insurer’s security caused by the reduction in the
loading factor has a greater effect on the adjustment coefficient than the increase resulting from reinsurance for all values of
𝛼.
Loading Factor (net of reinsurance):
The insurer’s premium loading factor, net of reinsurance, implied by the previous expression [ 1 + 𝜃 − 1 + 𝜉 1 − 𝛼 ]𝜆𝑚1
can now found, and shown to be an increasing function of 𝛼.
The loading factor is found by dividing the expected profit per unit time by the expected claims per unit time.
The expected profit per unit time is: 1 + 𝜃 − 1 + 𝜉 1 − 𝛼 𝜆𝑚1 − 𝛼𝜆𝑚1 i.e. net premiums less expected net claims
1+𝜃 − 1+𝜉 1−𝛼 −𝛼 (𝜉−𝜃)
So the loading factor is: 𝜃 ′ = =𝜉−
𝛼 𝛼
𝑑𝜃′ (𝜉−𝜃)
Now = 2 which is positive since 𝜉 > 𝜃, so that 𝜃′ is an increasing function of 𝛼. Thus, the net loading factor increases
𝑑𝛼 𝛼
as the retention level increases.

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(1)
We now consider the effect of excess of loss reinsurance on the adjustment coefficient.
The following assumptions are made:
• the insurer’s premium income (before reinsurance) per unit time is 1 + 𝜃 𝜆𝑚1
• the reinsurance premium per unit time is (1 + 𝜉)𝜆𝐸(𝑍), where 𝜉 ≥ 𝜃 is the reinsurer’s premium loading factor, and 𝑍 =
max(0, 𝑋 − 𝑀).
The insurer’s individual net claim payments are distributed as 𝑌 = min(𝑋, 𝑀), and the insurer’s premium income, net of reinsurance,
is:
𝑐𝑛𝑒𝑡 = 1 + 𝜃 𝜆𝑚1 − (1 + 𝜉)𝜆𝐸(𝑍)
which gives the equation defining R as:
𝜆 + 𝑐𝑛𝑒𝑡 𝑅 = 𝜆𝑀𝑌 (𝑅)
𝑋 𝑖𝑓 𝑋 < 𝑀
We know that: 𝑌 = ቊ
𝑀 𝑖𝑓 𝑋 ≥ 𝑀
Therefore:
𝑀 ∞ 𝑀 ∞
𝑀𝑌 𝑅 = 𝐸 𝑒 𝑅𝑌 = ‫׬‬0 𝑒 𝑅𝑥 𝑓 𝑥 𝑑𝑥 + ‫׬ = 𝑥𝑑 𝑥 𝑓 𝑀𝑅 𝑒 𝑀׬‬0 𝑒 𝑅𝑥 𝑓 𝑥 𝑑𝑥 + 𝑒 𝑅𝑀 ‫𝑥𝑑 𝑥 𝑓 𝑀׬‬
But the second integral is just integrating the PDF from M to ∞ . This is the same as 𝑃(𝑋 > 𝑀), which can be written as
1 − 𝐹(𝑀).
And therefore the equation for 𝑅 becomes:
𝑀
𝜆 + 𝑐𝑛𝑒𝑡 𝑅 = 𝜆{‫׬‬0 𝑒 𝑅𝑥 𝑓 𝑥 𝑑𝑥 + 𝑒 𝑅𝑀 (1 − 𝐹 𝑀 )}

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(2)
Illustration:
To illustrate ideas, look at the situation when:
𝑋~𝑈(0, 20)
1
𝑓 𝑥 = = 0.05 for 0 < 𝑥 < 20
20
Then for 0 < 𝑀 ≤ 20:
20
𝐸 𝑍 = ‫ 𝑥 𝑀׬‬− 𝑀 0.05 𝑑𝑥 = 10 − 𝑀 + 0.025𝑀2
And:
𝑀
𝑀𝑌 𝑅 = ‫׬‬0 𝑒 𝑅𝑥 0.05 𝑑𝑥 + 𝑒 𝑅𝑀 (1 − 0.05𝑀)
0.05
= 𝑒 𝑅𝑀 − 1 + 𝑒 𝑅𝑀 (1 − 0.05𝑀)
𝑅
The equation for 𝑅 must be solved numerically for given values of 𝜃 and 𝜉.
The figure below shows 𝑅 as a function of 𝑀 when 𝜃 = 𝜉 = 0.1.
As previously stated, any retention level is possible when the premium loading factors are equal.
𝑅 is a decreasing function of 𝑀.

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(3)

From the figure above, 𝑅 goes to ∞ as 𝑀 goes to 0.

When 𝜉 > 𝜃, there is a minimum retention level for the same reason as previously discussed.
𝜃
Recall the lower limit for 𝛼 given by the equation 𝛼 > 1 − applied when we were considering proportional reinsurance.
𝜉

For example, when 𝜃 = 0.1 and 𝜉 = 0.2 the insurer’s net premium income 𝑐𝑛𝑒𝑡 is:
𝑐𝑛𝑒𝑡 = 1 + 𝜃 𝜆𝑚1 − 1 + 𝜉 𝜆𝐸 𝑍 = 1.1𝜆 ∗ 10 − 1.2𝜆 10 − 𝑀 + 0.025𝑀2 = 11𝜆 − 1.2𝜆 10 − 𝑀 + 0.025𝑀2
and this must exceed the insurer’s expected claims, net of reinsurance.
The insurer’s expected net claims equal 𝜆𝐸 𝑋 − 𝜆𝐸(𝑍), which gives:
10𝜆 − 𝜆 10 − 𝑀 + 0.025𝑀2 = 𝜆[𝑀 − 0.025𝑀2 ]

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(4)
Thus we now have that:
11𝜆 − 1.2𝜆 10 − 𝑀 + 0.025𝑀2 > 𝜆 𝑀 − 0.025𝑀2
−1 + 1.2𝑀 − 0.03𝑀2 > 𝑀 − 0.025𝑀2
−0.005𝑀2 + 0.2𝑀 − 1 > 0
−𝑀2 + 40𝑀 − 200 > 0
𝑀2 − 40𝑀 + 200 < 0
5.8579 < 𝑀 < 34.1421
Hence, the minimum retention level is 5.8579.
Similarly, when 𝜉 = 0.4, the minimum retention level is 10.

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(5)

The figure above shows 𝑅 as a function of 𝑀 for the following combinations of 𝜃 and 𝜉:
• 𝜃 = 0.1 and 𝜉 = 0.2(solid line)
• 𝜃 = 0.1 and 𝜉 = 0.4(dotted line).
Without reinsurance, i.e. for 𝑀 = 20, the insurer’s adjustment coefficient is 0.014 (irrespective of the reinsurer’s loading
factor).
From the figure above, it can be seen that, for 𝜉 = 0.2:
• 𝑅 𝑀 > 𝑅 20 𝑓𝑜𝑟 9.6 < 𝑀 < 20
• 𝑅 𝑀 < 𝑅 20 𝑓𝑜𝑟 𝑀 < 9.6

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(6)
And for 𝜉 = 0.4
𝑅 𝑀 < 𝑅 20 𝑓𝑜𝑟 𝑅 < 20
Hence, for 𝜉 = 0.2 it is possible for the insurer to increase the value of the adjustment coefficient by effecting reinsurance,
provided that the retention level is above 9.6.
However, when 𝜉 = 0.4, the insurer should retain the entire risk in order to maximise the value of the adjustment coefficient.
As in the case of proportional reinsurance, the insurer’s expected profit per unit time is reduced if reinsurance is effected.
Question:
Claims occur as a Poisson process with rate 𝜆 and individual claim sizes 𝑋 follow an 𝐸𝑥𝑝(𝛽) distribution.
The office premium includes a security loading 𝜃1 .
An individual excess of loss arrangement operates under which the reinsurer pays the excess of individual claims above an
amount 𝑀 in return for a premium equal to the reinsurer’s risk premium increased by a proportionate security loading 𝜃2 .
Derive and simplify as far as possible an equation satisfied by the adjustment coefficient for the direct insurer.
Solution:
The adjustment coefficient equation is 𝜆 + 𝑐𝑅 = 𝜆𝑀𝑋 (𝑅)
The net rate of premium income for the direct insurer equals the rate of premiums charged to the policyholder minus the
rate of premiums paid to the reinsurer:

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(7)
1+𝜃1 𝜆 ∞
𝑐= − (1 + 𝜃2 )𝜆 ‫ 𝑥 𝑀׬‬− 𝑀 𝛽𝑒 −𝛽𝑥 𝑑𝑥
𝛽

The second term can be integrated using the substitution 𝑦 = 𝑥 − 𝑀, and identifying the integral as the mean of an
𝐸𝑥𝑝(𝛽) distribution. This gives:
𝜆
𝑐 = [ 1 + 𝜃1 − 1 + 𝜃2 𝑒 −𝛽𝑀 ]
𝛽

The individual net claims are the claims paid to policyholders minus the recoveries from the reinsurer. So the MGF
(which is valid for all values of 𝑅 < 𝛽 ) is:
𝑀 ∞ 1
𝑀𝑋 𝑅 = ‫׬‬0 𝑒 𝑅𝑥 𝛽𝑒 −𝛽𝑥 𝑑𝑥 + ‫ 𝑒𝛽 𝑀𝑅 𝑒 𝑀׬‬−𝛽𝑥 𝑑𝑥 = [𝛽 − 𝑅𝑒 − 𝛽−𝑅 𝑀
]
𝛽−𝑅

So the equation for the adjustment coefficient is:

𝜆 1
𝜆+ 1 + 𝜃1 − 1 + 𝜃2 𝑒 −𝛽𝑀 𝑅 = 𝜆 [𝛽 − 𝑅𝑒 − 𝛽−𝑅 𝑀
]
𝛽 𝛽−𝑅

Cancelling 𝜆 and multiplying through by 𝛽(𝛽 − 𝑅) gives:

𝛽 𝛽−𝑅 + 𝛽−𝑅 1 + 𝜃1 − 1 + 𝜃2 𝑒 −𝛽𝑀 𝑅 = 𝛽[𝛽 − 𝑅𝑒 − 𝛽−𝑅 𝑀
]
Cancelling the 𝛽 2 from both sides gives:
−𝛽𝑅 + 𝛽 − 𝑅 1 + 𝜃1 − 1 + 𝜃2 𝑒 −𝛽𝑀 𝑅 = −𝛽𝑅𝑒 − 𝛽−𝑅 𝑀

Cancelling 𝑅’𝑠 to exclude the trivial solution gives:

−𝛽 + 𝛽 − 𝑅 1 + 𝜃1 − 1 + 𝜃2 𝑒 −𝛽𝑀 = −𝛽𝑒 − 𝛽−𝑅 𝑀

www.strathmore.edu
Maximising the adjustment coefficient under excess of loss reinsurance(8)
Question:
𝑥2
Use the approximation 𝑒 𝑥 ≈ 1 + 𝑥 + to find an approximate numerical value for the adjustment coefficient for the previous
2
question in the case where:
𝛽 = 0.05
𝜃1 = 0.3
𝜃2 = 0.4
𝑀 = 10
Solution:
Using the values given, the adjustment coefficient equation becomes:
−𝛽 + 𝛽 − 𝑅 1 + 𝜃1 − 1 + 𝜃2 𝑒 −𝛽𝑀 = −𝛽𝑒 − 𝛽−𝑅 𝑀

−0.05 + 0.05 − 𝑅 1.3 − 1.4𝑒 −0.05∗10 = −0.05𝑒 − 0.05−𝑅 10

Multiplying by −20𝑒 0.5 to clear some of the fractions gives:

𝑒 0.5 − 1 − 20𝑅 1.3𝑒 0.5 − 1.4 = 𝑒 10𝑅
Expanding the LHS and applying the approximation to the RHS:
0.90538 + 14.8668𝑅 = 1 + 10𝑅 + 50𝑅 2
−0.09462 + 4.8668𝑅 − 50𝑅2 = 0
Solving this using the quadratic formula (taking the smallest positive root) gives:
−4.8668± 4.86682 −4(−50)(−0.09462)
𝑅= = 0.0268
2(−50)

www.strathmore.edu