age/sex factors

[Polly Nomial]

When my company computes age sex factors, they completely remove from the calculation anyone whose claims excede 3 standard deviations for any age/gender band. This seems odd to me. Wouldn't it make more sense to simply cap the claims at $50,000 or so? Or does it?

[Patience]

Using the 3 x SD is much less arbitrary. It also moves the top range based upon the calculated average.

If a young male has 1% of the claims over $50,000 then there would be very little adjustment. If 8% of 60 year old women have claims over $50,000 then your taking a much bigger adjustment and lowering the average for the older person much more than the younger person.

By taking the 3xSD you are trimming a constant % of outlier claims off of all the samples equally. This would keep the comparable ratios constant, where a constant cap would flatten the aging curve.

[balmact]

Although, capping at a certain dollar amount could make sense if they were running a med stop loss starting at that dollar amount. However, then you would still be incorporating the price of the risk exceeding the dollar amount. But it should be cheaper than if the front carrier takes on 3SD of risk. That is about the only time that I could foresee setting a dollar amount cap instead of a ratio or SD.