Estimation, Evaluation, and Selection of Actuarial Models
Estimation, Evaluation, and Selection of Actuarial Models
Estimation, Evaluation, and Selection of Actuarial Models
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
22 CHAPTER 2. MODEL ESTIMATION<br />
Example 2.22 Use the empirical distribution in Example 2.6 to estimate the expected number <strong>of</strong><br />
accidents for one driver in one year.<br />
For this discrete distribution the expected value is P 5<br />
x=0<br />
xp(x) which can be estimated by<br />
5X<br />
x=0<br />
xp 94,935 (x) = 0 81,714<br />
94,935 +111,306 1,618 250 40<br />
+2 +3 +4<br />
94,935 94,935 94,935 94,935 +5 7<br />
94,935<br />
= 15,487<br />
94,935 =0.16313. ¤<br />
When all <strong>of</strong> the observations are present (no grouping, censoring, or truncation), one view <strong>of</strong> an<br />
empirical estimate is that it is calculated by doing to the data what you wish you could do to the<br />
population. In the previous example, we wanted to calculate the population mean–the empirical<br />
estimate calculates the sample mean.<br />
Example 2.23 For Data Set B use the empirical distribution to estimate the probability that a loss<br />
payment is greater than 150% <strong>of</strong> the mean payment.<br />
The mean <strong>of</strong> the empirical distribution is the sample mean or (27+82+···+15,743)/20 = 1,424.4.<br />
150% <strong>of</strong> that value is 2,136.6. In the sample, two <strong>of</strong> the observations exceed that value. The<br />
empirical estimate is 2/20 = 0.1. ¤<br />
Exercise 15 Consider Data Set B, but left truncated at 100 <strong>and</strong> right censored at 1000. Using the<br />
Kaplan-Meier estimate from Exercise 6 estimate the expected cost per payment with a deductible <strong>of</strong><br />
100 <strong>and</strong> a maximum payment <strong>of</strong> 900.<br />
Example 2.24 For Data Set C use the empirical distribution to estimate the expected premium<br />
increase when raising the policy limit from 125,000 to 300,000.<br />
In general, when the policy limit is u = c m (that is, the limit is a class boundary), the expected<br />
cost per payment is<br />
E(X ∧ u) =<br />
=<br />
Z u<br />
0<br />
mX<br />
j=1<br />
xf(x)dx + u[1 − F (u)]<br />
Z cj<br />
c j−1<br />
x<br />
n j<br />
n(c j − c j−1 ) dx + c m<br />
kX<br />
j=m+1<br />
mX n j c 2 j<br />
=<br />
− c2 kX<br />
j−1<br />
+ c m<br />
n(c j − c j−1 ) 2<br />
j=1<br />
j=m+1<br />
⎛<br />
⎞<br />
= 1 mX<br />
⎝<br />
c j + c j−1<br />
kX<br />
n j + c m n j<br />
⎠ .<br />
n 2<br />
j=1<br />
j=m+1<br />
Z cj<br />
c j−1<br />
n j<br />
n<br />
n j<br />
n(c j − c j−1 ) dx