Identifying RIsk differences...

[Les Poissons]

There was that question asking if the plan was good at identifying risk differences. For part A, is it true that large insureds were better at identifying risk differences, and thatsmall insureds were inferior because there really isn't an increasing trend in the manual loss ratios (.8 drops to .79, in fact)? And for part B of the problem, I think I agreed that too much weight was given to small insureds, because there was somewhat of a decreasing trend in the standard LRs. For C, a drop from 1.07 to 1.05 in the offbalance factor indicates that rate inadequacy has improved from last year to the current year, but that aggregate rate levels are still inadequate. Anyone disagree?

[ramanujan]

I agree with parts a and c.

For part B, I wrote that too little weight was being given to small insureds because the LRs are decreasing with increasing mod.

If we want to correct the difference identified by the mod, ie have same modified loss ratios, we need to move closer to the manual loss ratios (which are increasing with increase in mod) and therefore we need to increase the weight given to the individual experience.

[Les Poissons]

isn't that what too much weight means? if there's too much weight, it's almost "over-correcting." .71 goes to .84 for credit risks and .87 goes down to .76. Shouldn't the standard loss ratios be randomly distributed around a certain number (say 1.00) for it to have neither too much nor too little weight?

[ramanujan]

Now that I think about it, you are right. The mods are over-correcting so we need to decrease credit/debit by reducing Z. Damn! Another point gone.