MAT00061M Statistics for Insurance 2021/22
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
MAT00061M
MMath and MSc Examinations 2021/22
Statistics for Insurance
1 (of 4) . (a) In an insurance company, two separate Poisson surplus processes are
being managed, where the respective parameters are given in the following table . In the table, U denotes the initial reserve, A denotes the Poisson parameter, 9 is the security loading and x is the typical claim size variable .
Table: Parameters for the two Poisson surplus processes
Process U A 9 Distribution of x |
||
A 500 20 0 .25 Exponential with mean 100 |
||
B 600 20 0 .30 |
Γ(2, 0 |
02) |
(i) For Process A, write down the adjustment equation and calculate the adjustment coefficient . [5]
(ii) For Process B, write down the adjustment equation and calculate the adjustment coefficient . [5]
(iii) Use the Lundberg upper bound as an approximation for the probability
of ruin, and rank the two processes A and B . [4]
(b) The aggregate claim process in a company is modelled by a compound Poisson surplus process, where the Poisson parameter is A = 20, initial reserves are U = 1200, a typical claim is exponentially distributed with mean 50 . The insurance company is considering buying proportional rein- surance where a portion a of every claim is retained . Let 9 = 0.2 be the insurer’s loading factor and 5 = 0.3 be the reinsurer’s loading factor .
(i) What is the minimum value of a that the insurance company should consider? [2]
(ii) Derive the adjustment coefficient for this process and write it as a
function of a . [6]
(a) Let s = x1 +× × ×+xN have a compound Poisson distribution, where A = 5
(b) A portfolio of 200 one-year fire insurance policies is summarised in the following table .
Table: Number of policies by incidence and claim distribution
|
Claim size |
distribution |
Claim probability |
Uniform [0, 48] |
Uniform [48, 96] |
0.02 |
80 |
60 |
0.04 |
40 |
20 |
For example, there are 140 policies where the chance of a claim being made is 0.02, and for 80 of these 140 policies if a claim is made, it is equally likely to be anywhere in the interval [0, 48] .
(i) Let s denote total claims from this portfolio during the year . Using a compound Poisson distribution to model s, determine the security loading factor 9 which is necessary to be 95% sure premiums exceed claims . [14]
(ii) If the number of policies were to triple in each of the four categories, what would be the necessary security loading factor? [8]
3 (of 4) .
(a) A random sample of 120 claims was observed from a portfolio, where
12d 12d
zi = 9, 000; zi(2) = 420, 000,
i91 i91
zi denotes the i-th claim observation .
(i) Model the data by an exponential distribution with density 9e-a_ (z > 0) and estimate the positive parameter 9 by the method of moments . [3]
(ii) Model the data by a distribution with density f (z) = 9ze-a_2 /2 (z > 0) and find the maximum likelihood estimate of the positive
parameter 9 . [5]
(b) A binomial generalised linear model with the logistic link function is used by a bank to predict the probability of success of new personal investment packages . The linear predictor is defined by n = ai + 8i z, where ai and 8i are the parameters with the index i denoting types of products (i = 1 for “property”, i = 2 for “equity”, and i = 3 for “bond”), the covariate z (in million pounds) represents the marketing budget for the product. Maximum likelihood estimates for these parameters have been obtained:
i = 1 (property): 1 = .0.045 , 1 = 0.028;
(ii) What is the predicted success probability of a new “equity” product
with a 2 million pounds budget? [4]
(iii) At which budget level will a “property” and an “equity”-based prod- uct have an equal probability of success? [2]
4 (of 4) . (a) The following table gives cumulative paid claims in a motor insurance
scheme over a four-year period, together with annual premium income and estimated loss ratios determined by an underwriter . Assume that the amounts have been adjusted for inflation and all claims are settled by the end of development year 3 .
Table: Motor insurance scheme: paid claims, premiums and loss ratios
Origin Development year |
||||||
year i |
0 |
1 |
2 |
3 |
Premium |
Loss ratio |
1998 |
31,766 |
48,708 |
64,551 |
69,003 |
76,725 |
92% |
1999 |
30,043 |
45,720 |
59,883 |
|
77,005 |
92% |
2000 |
35,819 |
54,790 |
|
|
78,520 |
90% |
2001 |
40,108 |
|
|
|
83,400 |
90% |
Use the Bornhuetter-Ferguson method and the weighted development factors to estimate the amount of reserves which should be set aside for the future claims at the end of 2001 . [12]
(b) A claim-size random variable is modelled by a Pareto distribution, i .e ., x ~ Pareto(a, 入) with a = 4 and 入 = 900 . A reinsurance arrangement has been made for future years whereby the excess of any claim over 600 is paid by the reinsurer .
(i) Determine the mean reduction in claim size for the baseline insurance company which is achieved by this arrangement . [8]
(ii) If claim inflation of 10% is expected for the next year and the excess level remains at 600, what will be the expected cost of a claim to the baseline insurance company? [6]
2022-08-11