INSTITUTE AND FACULTY OF ACTUARIES

EXAMINATION

24 April 2023 (am)

Subject CS1 – Actuarial Practice

Core Principles
Paper B
Time allowed: One hour and fifty minutes

In addition to this paper you should have available the 2002 edition of
the Formulae and Tables and your own electronic calculator.

If you encounter any issues during the examination please contact the Assessment Team on
T. 0044 (0) 1865 268 873.

CS1B A2023 © Institute and Faculty of Actuaries

1 (i) Simulate 10,000 values from a N(0,1) distribution using an appropriate R
command. You should save the generated values in R for later use. [1]

(ii) Simulate 10,000 values from a χ24 distribution using an appropriate R

command. You should save the generated values in R for later use. [2]

(iii) Discuss the main disadvantage of using the inverse transform method for
simulating values from a chi-square distribution. [3]

(iv) Simulate 10,000 values from a t4 distribution using your answers to parts (i)
and (ii), explaining your answer.

You should save the generated values in R for later use. [5]

(v) Simulate 10,000 values from a t20 distribution by modifying your R code from
parts (i), (ii) and (iv) where necessary.

You should save the generated values in R for later use. [5]

(vi) Demonstrate an important property of the t distribution by presenting and

comparing two appropriate plots of the values simulated in parts (iv) and (v).
[6]
[Total 22]

CS1B A2023–2
2 Consider a sample of 1,000 insurance policies. For each policy, there is a record of
whether the policyholder has submitted at least one claim during the last 6-month
period, the age of the policyholder and the duration of the policy. The number of
policies, X, with one or more submitted claims is modelled as a random variable with
a Binomial distribution X ~ Binomial(1000, p).

(i) State the underlying assumptions that lead to the choice of a Binomial
distribution for X. [2]

Use the command load("CS1policies.Rdata") to load the observed data into

R. This will create a data frame called CS1policies with three entries in your R
workspace. You can check the names of the entries with names(CS1policies):

 age: age of the policyholder in years

 duration: duration of the policy in months
 claimed: takes the values 0 (no claim in 6-month period) or 1 (at least one claim
in 6-month period).

An Actuary wishes to fit different Generalised Linear Models (GLMs) to the data,
assuming that the number of policies with one or more submitted claims has a
Binomial distribution and the link function of the GLM is the logit function.

(ii) Fit a GLM to the data such that p depends on the age of the policyholder and
report the estimated parameter values. [5]

(iii) Fit a GLM to the data such that p depends on the duration of the policy and
report the estimated parameter values. [5]

(iv) Compare the fit of the models in parts (ii) and (iii) using the deviance and
the Akaike’s Information Criterion (AIC). [3]

(v) Fit a GLM to the data, such that p depends on both the age of the policyholder
and the duration of the policy, and report the estimated parameter values. [5]

(vi) Comment on which of the three models in parts (ii), (iii) and (v) is most
appropriate, justifying your answer. [2]
[Total 22]

CS1B A2023–3
3 The file heights.RData contains an R data frame (heights) that consists of a
random sample of heights of mothers (mother.height), fathers
(father.height) and their daughters (daughter.height) in inches.

(i) Draw a scatterplot for each pair of data. [3]

(ii) Comment on the relationships between the pairs of data by referring to the
plots from part (i). [4]

(iii) Fit a multiple linear regression model with the daughters’ height as the
response variable and mothers’ and fathers’ height as the explanatory
variables. Your answer should show a summary of the output and the fitted
equation. [5]

(iv) Assess the effect of each explanatory variable on the dependent variable by
referring to the R output from part (iii). [4]

(v) Plot the residuals of the model fitted in part (iii) using a graph that is
appropriate for checking if the estimated errors are independent of the
explanatory variables. [5]

(vi) Comment on the fit of the model by referring to the plot from part (v). [2]

A suggestion is made that when the mother’s height is 61 inches and the father’s
height is 63 inches, their daughter’s height should be 67.5 inches.

(vii) Comment on whether this claim is likely by using an appropriate 95%

predictive interval of the daughter’s height based on the fitted model. [4]
[Total 27]

CS1B A2023–4
4 The table below shows the total claim amounts (£1,000s) per year, Xij , over a 5-year
period for five insurance companies’ critical illness book of business.

Year, j
2016 2017 2018 2019 2020
A 5,000 2,720 3,170 2,950 6,300
B 3,680 3,360 3,900 2,910 8,020
Insurance
C 880 800 550 620 1,890
company, i
D 6,150 3,880 5,780 5,220 7,100
E 1,100 970 1,900 2,300 3,430

The data can be entered into R in matrix form using the following code:

claims <-
matrix(c(5000,3680,880,6150,1100,2720,3360,800,3880,970,
3170,3900,550,5780,1900,2950,2910,620,5220,2300,6300,
8020,1890,7100,3430),nrow=5,ncol=5)

(i) Calculate, using Empirical Bayes Creditability Theory (EBCT) Model 1, the
following:

(a) Em ϑ [2]

(b) Var m ϑ [3]

(c) E s2 ϑ [2]

(ii) Calculate the expected claim amount for Company E, using your answers from
part (i). [3]

The table below shows additional information that has been provided showing the
number of claims, Pij , over a 5-year period for the five insurance companies.

Year, j
2016 2017 2018 2019 2020
A 580 540 490 500 650
B 440 460 360 390 510
Insurance
C 180 160 170 120 220
company, i
D 990 1,000 880 930 1,100
E 210 200 175 170 130

The data can be entered into R in matrix form using the following code:

volumes <-
matrix(c(580,440,180,990,210,540,460,160,1000,200,490,
360,170,880,175,500,390,120,930,170,650,510,220,
1100,130),nrow=5,ncol=5)

CS1B A2023–5
 (iii) Calculate, using EBCT Model 2, the following:

(a) Em ϑ [7]

(b) Var m ϑ [4]

(c) E s2 ϑ [3]

(iv) Calculate the expected claim amount for Company E, using your answers from
part (iii), assuming that the number of claims for Company E in the following
year remains at 130. [3]

(v) Comment on the difference between models EBCT Model 1 and EBCT
Model 2 applied in earlier parts, by comparing your answers in parts (ii)
and (iv). [2]
[Total 29]

END OF PAPER

CS1B A2023–6