Estimating frequency distributions

[jack123]

Hi, I'm looking for some clarification on the frequency-severity rating approach in Chapter 14.

The notes state that first the claim frequencies should be grouped by policy year and trended. Then you need to fit a distribution to the frequencies. How do you do this? If you have 5 years of data you will only have 5 data points with which to fit your distribution.

 

[Katherine Young]

Yes, tricky isn't it, and this could easily be an exam question.

I'm a little concerned for you abc1, because of the number of questions you are asking on this forum. It's ok to ask questions of course, but it's important that you develop your own problem-solving skills too.

I recommend that you practise generating ideas, even for questions which, like this, you haven't considered before. After all, 'idea generation' is an important technique in the exam.

Therefore, I'm going to ask you to suggest your own answer to your question. How might you go about this in practice, if you had to do this in your own work?