Exam MAS-I Practice Question 46
Study guide: Exam MAS-I
An actuary will fit a two-parameter Pareto distribution to 40 liability claims by maximum likelihood, using a numerical search. She first needs starting values for both parameters. A few very large claims dominate the data, and prior studies of this line put \alpha near 1.5. The method that supplies the starting values must:
- estimate both parameters from the data;
- not rely on a moment of the Pareto that is infinite, or that the sample estimates erratically, when
\alphais near 1.5; - not be driven by the one or two largest or smallest claims.
Which method best meets all three requirements?
- A) Method of moments, matching the sample mean and sample variance
- B) Matching the sample mean and the smoothed empirical median
- C) Percentile matching at the smoothed 5th and 95th percentiles
- D) Percentile matching at the smoothed 25th and 75th percentiles
- E) Percentile matching at the smoothed median, with
\alphaset to 1.5
Answer this question and see the worked solution