2013 Spring 11
Could you please explain why, in part (c), we need to credibility weight the normalized relativities. In my calculation, I did the credibility weighted average on the rebased relativities (the base class having relativity 1), and got similar results. Thanks in advance.
Comments
While using rebased relativities before credibility weighting can give similar results, starting with normalized relativities helps address any data issues early on. Your answer still may have received full credit though.
In general, credibility weighting the normalized relativities is slightly better because it makes your rates more reliable and stable. Normalizing first ensures fair comparison between different segments and minimizes bias.