You are on page 1of 2

CS2B: Chapter 1 Poisson processes - Questions Page 1

Questions
1.1 A small general insurance company sells three different types of policy, motor insurance, house
insurance and travel insurance. It received 709 claims over the 10-year period from 2008 to 2017,
which are tabulated in chronological order in a matrix. You should run this code to read the data:

claims = read.csv("Claims Data.csv",header=T,sep=",")


claims$claim.date <- as.Date(claims$claim.date, format("%d/%m/%Y"))
head(claims)

[Note that we have specified the date format for you.]

(i) Create three new separate tables for the motor insurance claims, house insurance claims
and travel insurance claims, showing the date and size of each claim. You should show
the first few rows of each table and specify the number of claims of each type of policy. [5]

Hint: you may wish to use the functions head and nrow.

The company is investigating modelling its future claim numbers as a Poisson process.

(ii) State the distribution of the waiting times between consecutive events of a Poisson
process Nt ~ Poi  t  . [2]

(iii) (a) Expand your tables to include a column for the observed waiting times (in days)
for all the claims. You should again show the first few rows of each table.

(b) Repeat this separately for the motor, house and travel claims. [10]

Hint: you may wish to use the diff and as.numeric functions.

(iv) Test whether the observed waiting times for the combined claims are indicative of a
Poisson process.

(There were no days with more than two claims and you may assume that if two claims
were on the same day, the waiting time between them was between 0 and 0.5 days.) [10]

Hint: you may wish to use the qchisq function to obtain the critical value for the test.

(v) Calculate the Poisson parameter to be used for the motor insurance claims by:

(a) examining the actual waiting times of the motor insurance claims

(b) thinning the Poisson parameter used for all claims.

Comment on your answers. [6]

(vi) Repeat part (v) for both the house insurance claims and the travel insurance claims. [8]

The Actuarial Education Company © IFE: 2020 Examinations


Page 2 CS2B: Chapter 1 Poisson processes - Questions

(vii) Test whether the observed waiting times are indicative of a Poisson process for the:

(a) motor insurance claims

(b) house insurance claims

(c) travel insurance claims. [12]

(viii) Use the results of parts (iv) to (vii) to discuss whether the company should model its
future claim numbers as a Poisson process. [6]

The claim amounts of each individual claim are assumed to be independent of each other and
follow a gamma distribution. The company decides to model the total future claim amount for
each type of policy as a compound Poisson process, and uses the method of moments to fit the
parameters of the claim size distribution:

Nt
St   X j
j 0

where X j ~ Ga  ,   .

(ix) Code a function to calculate the sample mean, standard deviation, and the estimated
values for the parameters  and  for:

(a) motor insurance claims

(b) house insurance claims

(c) travel insurance claims [7]

Hint: you may wish to use the mean, var and sqrt functions.

(x) Calculate the mean and standard deviation of the total annual claim amounts for:

(a) motor insurance

(b) house insurance

(c) travel insurance [5]

(xi) Besides those already discussed in part (viii), discuss any other issues the company should
consider before proceeding with this model for total future claim amounts. [6]
[Total 77]

© IFE: 2020 Examinations The Actuarial Education Company

You might also like