CAS Exam 8 Fall 2013, Question 3
Study guide: Exam 8
An actuary is creating a workers compensation classification rating plan. The actuary has access to frequency data from years 2002-2012 for fatal claims, which he believes may have low credibility.
After developing and testing a multivariate credibility procedure, the actuary finds the following results:
| Hazard Group | Sum of Squared Prediction Errors – Fatal Claims: Prediction based on Hazard Group | Prediction based on raw data less the holdout sample | Prediction based on credibility procedure |
|---|---|---|---|
| A | 100 | 90 | 80 |
| B | 110 | 70 | 105 |
| C | 90 | 115 | 75 |
Part a (0.25 points)
Briefly describe the purpose of a holdout sample.
Part b (0.5 points)
Justify an appropriate holdout sample from the available frequency data for the actuary to use in the classification analysis.
Part c (0.5 points)
Discuss what the actuary's predicted results imply about his credibility procedure.
Part d (0.5 points)
Suppose the actuary suspects that there may be an intrinsic downward trend in frequencies of fatal claims between 2002 and 2012 due to improved safety in his clients' workplaces. Propose a way for the actuary to test this theory.
Answer this question and see the worked solution