Underwriting Small Employer Groups

[Polly Nomial]

My state is guaranteed issue for all small group products, but allows the use of a +/- 30% rateband. For new groups we do health underwriting and a risk load is applied. I have no problem with this.

I would like to know what to do in the subsequent years. My company has been using loss ratios to increase or decrease the risk loads. This makes no sense to me from a credability standpoint. Also, a health condition that was used to assess a risk load may still be there even though the claims remain low.

Any thoughts? Maintain the initial risk load? Increase/decrease it for bad/good loss ratios? Periodically re-underwrite the groups via diagnosis? Increase/decrease rates towards the average rate over a period of several years?

[PinkFloydActuary]

My state is guaranteed issue for all small group products, but allows the use of a +/- 30% rateband. For new groups we do health underwriting and a risk load is applied. I have no problem with this.

I would like to know what to do in the subsequent years. My company has been using loss ratios to increase or decrease the risk loads. This makes no sense to me from a credability standpoint. Also, a health condition that was used to assess a risk load may still be there even though the claims remain low.

Any thoughts? Maintain the initial risk load? Increase/decrease it for bad/good loss ratios? Periodically re-underwrite the groups via diagnosis? Increase/decrease rates towards the average rate over a period of several years?

At renewal, underwriting should verify what the load was for and whether it needs to continue as is. This can be done via diagnosis codes/drugs used over the past year. Since the drug data is usually much more up to date, that needs to be looked at carefully in case something new on the group has popped up.

[Dr T Non-Fan]

Loss ratio is so wrong.
Risk adjusters would be slightly better.

[mr.c]

Can't you use a credibility function that maxes out at 5% or so? Then a group can feel like they're bringing down their rates without being unreasonable. Most of the rest could just be poolwide trend/age adjustment on their old rates. We tried implementing a book rate fairly recently for some percentage of experience "missing" for newer employees. Healthy groups just loved that one.

Do you think that it would be cost effective to rexamine health status of so many individual groups? Can you get a big enough health change increase approved on top of a large trend increase?

[Dr T Non-Fan]

It's probably better to quantify the credibility (maxing at 5%) than to dismiss it outright.
It would be more effective to prove it somehow, as the underwriters always seem to know better. ( = "This is what we were taught, by someone who learned it 20 years ago.")

[mr.c]

Oh, yeah. All of those proofs of the credibilty formulas in the exam materials. I forgot about them.

[Goldenhands]

I work at a small ins. company under similar local laws. We have 35% bands instead of 30%, but close enough. LR is used as a factor, but medical management also gives projections on anybody they believe will be high dollar, i.e. over say $2000.

A problem with LR, outside of credibility, is that you may have someone that has a serious illness that will incur high claims next year, but claims haven't been showing up yet. Premiums will always lag the data too much if you don't take this into account.

One of the issues not asked is how to convey to agents the increases you develop. The pro to LR is that agents understand it. If you're giving a rate increase higher than trend/census changes you can justify to an agent by showing a high LR. However, agents also understand if there's a high dollar case on the books just diagnosed then they don't have much bargaining (read whining) room.

[Polly Nomial]

...One of the issues not asked is how to convey to agents the increases you develop. The pro to LR is that agents understand it. If you're giving a rate increase higher than trend/census changes you can justify to an agent by showing a high LR. However, agents also understand if there's a high dollar case on the books just diagnosed then they don't have much bargaining (read whining) room.

True enough about the simplicity of the rational, but it's so hit and miss. For example a difficult pregnancy can raise an LR, but once it's over, it's over...

[Dr T Non-Fan]

That's why risk adjusters are better. If only the methodology were complete...