New EOM6 Discussion

[uclatommy]

Working on my EOM6 and thought I'd check in with AO to see if I'm getting results consistent with what everyone else is getting.

1. Low reserve for retrospective calculation
   I'm showing a reserve in the neighborhood of $700. Looks like the claims development has mostly run its course after 6 months. Did anyone put back the manual reserve adjustments for Nov-12 and Dec-12 or did you leave them at $0?
2. Reserve deficiency in the summary of retrospective calculation
   It seems our reserves as of Dec-12 were deficient by $4K given what we know happened in the six months since then. Is that an accurate description of what's going on here? This seems like a pretty bad result (big loss) when reserves are in the neighborhood of $20K. The size of loss exceeds the margin for adverse deviation. Did anyone feel the need to change assumptions?
3. Loss ratio assumption seems to be low
   Anyone else seeing this as well?

Overall, it seems like the triangle method is developing completion factors that are a bit low. Anyone know why this is happening and how to improve it? Is claims inflation the culprit here?

[alex123]

I got similar answers.

1) I did adjust Nov 12, I am thinking of adjusting Dec 12 with a negative adjustment.

2) Yea, I completely agree. I thought that I was missing something when I got those numbers, too. I am not sure why this is happening though.

Other questions -

4) on the loss ratio estimate on the input tab, can I change those? I am not clear on how to come up with good inputs for the loss ratios.

5) On tab 13, I was looking at the trend rate displayed for June 13. It seems a lot lower than the last few estimates. I got between 2-5%, which is less than half as much as the trend comparison for retro Dec 12.

[aedunn05]

Is anyone coming out with with a big difference using the two methods for the 6/30/2013 reserve? I've been through it over and over and can't figure out why.

[uclatommy]

The premiums and membership increased significantly. This highlights one of the pitfalls of using the loss ratio method as mentioned in the module.

[shoang86]

I also got similar answers.

I got a reserve of about 780 and ending reserve with adjustment of 840.

I also go the redundancy/deficiency of $4.1k

I'm not sure about the manual adjustment, there's no documentation as to why that manual adjustment was made.



@uclatommy, i also went to ucla!

[undegaussable]

The Nov/Dec 2012 adjustments were for flu... they are pretty egregious!

[uclatommy]

It seems like with the development method, you would always have to make manual adjustments to the most recent 2 months because there just isn't enough development data to produce a reliable estimate of the final claims costs associated with claims in those months. You just kinda have to wave a magic want and come up with some adjustment and bs your way to a good reason.

[no driver]

I am not a health actuary, but in my understanding the whole point of the development method is to say "after 1 month we have typically paid out x% of the final claims, after 2 months..." etc. Then you scale the reported claims by the appropriate factors to determine the expected total claims that will be reported. I don't think that the most recent months necessarily need manual adjustments unless there's a justifiable reason to make them. The method should take care of inflating the current experience up to what would be typically experienced.

[lovmath]

Hi Guys,

I still don't understand how you determine the best approach to using either 3 month, 6 month, 9 month or 12 month for projection... what data charactersitics or calculation results do you look at to determine what to use.

[bp0001]

uclatommy,

The development method isn't fully presented in the model. When we prepare IBNR, we also use a credibility calculation. If expected payments are a low %, then we use theoretical claims(PMPM from 12 months prior inflated by trend) - paid in place of the completed claims.

This is how we do it in practice, not in FAP.