QUESTION 1 a) The following information are given: + 80% of claims are of Type M and the other 20% are of Type N. + Type M claim sizes follow a normal distribution with mean 5,500 and variance 1,000,000. + Type N claim sizes follow a normal distribution with mean 6,300 and variance 1,000,000. A claim file exists for each of the claims, and one of them is randomly selected. The claim file selected incomplete and indicates only that its associated claim size is greater than 8,000. Determine the posterior probability that a Type M claim was selected. By Bayes’ Theorem, Pr(X > 8,000 | M)Pr (M) Pr(M LX > 8.000) = 5S eaoOTM) Pr(M) + Pr > BOO0 IN) Pre) 100 — 5,500) 7000 Pr(x > 8,000 |) 0062 10 — 6,300 7,000 (1.7) = Pr(x > 8.000 |.) = 1- | 0146 (0.0062)(0.8) [W.0062) (0.8) + (0.0446) (oz ~ 99579 Pr(M |X > 8,000) =b) Based on survey conducted by Rustagi and Wright (1995), Bayesian methods were litte used in current statistical practice at that time compared to other statistical techniques. In your opinion, what are the possible reasons for this? « It does not tell how to select a prior. There is no correct way to choose a prior. Bayesian inferences require skills to translate subjective prior beliefs into a mathematically formulated prior. If you do not proceed with caution, you can generate misleading results. « Itcan produce posterior distributions that are heavily influenced by the priors. From a practical point of view, it might sometimes be difficult to convince subject matter experts who do not agree with the validity of the chosen prior. + Itoften comes with a high computational cost, especially in models with a large number of parameters. In addition, simulations provide slightly different answers unless the same random seed is used. Note that slight variations in simulation results do not contradict the early claim that Bayesian inferences are exact. The posterior distribution of a parameter is exact, given the likelihood function and the priors, while simulation-based estimates of posterior quantities can vary due to the random number generator used in the procedures.QUESTION 2 The table shows the agaregate claim amounts (in RM million) for an international insurer's fire portfolio for a 5-yéar period, together with some summary statistics. ‘Aggregate claim amount, Year] 1 2 3 4 5 1 53 56 64 55. 37 Country, 2 Cd 6 59 2 36 3 81 4 7 80 79 4 50 46 52 56 et The volumes of business for each country for the insurer are as follows: Volume of business, Year] 1 2 3 4p 5 € 7 21 19 24 20 18 23 aT 30 29 33 25 12 9 5 7 "1 10 28 25 33 38 a 36 ‘Year 6 using EBCT Model 2.QUESTION 2 varegate claim amountYear I te Ie chime of bustuessitear J Total claim amount per unit of risk volume, Yearj 1 2 3 4 5 1 3.1176, 2.7619 3.2105 2.2917 2.85 Country, 2 2.8696 2.3333, 1.9667, 2.1379, 1.7576 f 3 675 9.333 5.1333 4.7059 7.1818 4 1.7857 1.84 1.5758 1.4737 1.4878 Step 2:P,, Pand P* Country./ P; Ba 1 101 79.39 2 142 99.28 3 64 55.32 4 165 107.32 P=472 P= 17.96 Step 3: County. x Yuey-%* Pyy —2F 4 2.81 11.18 13.05 2 2.17 18.86 242.98 3 6.27 157.32 984.72 4 1.61 3.54 154.39 E|m(0)] = 2.67 alst(@) = 218+ 18 ses ts725 5 ses & Var{m(o)] = 13.05 + 242.98 + 984.72 + 154. 39 11.93] “ae a ‘Country Estimated credibility | Risk premium per unit EBCT Premium factor volume 4 0.9666 2.805324 50.495832_ 2 0.9760. 2.182 54.55 3 0.9483 6.08388. 60.8388 4 0.9793 1.631942 58.749912QUESTION 3 IfX ~ Gamma (a, ) , show that the MGF of X is: M(t) = (1-3) ° My(t) = E(e*) =f eit te Ped = [pert eax ‘ 4 aay [perpen onar wos (ga) = (EGY =(0-9)" rece