About 1 1/2 years ago, I wrote a seven-part series on investing in insurance stocks. ?It is still a good series, and worthy of your time, because there aren’t *that* many writers freely available on the topic.

This particular article deals expands on part 4 of that series, which deals with insurance reserving. ?I wanted to do this at the time, but I was short on time, and wrote out the general theory there, while not actually doing the time-consuming job of ranking the conservativeness of P&C insurers reserving practices.

Let me quote the two most important sections from part 4:

 

When an insurance policy is written, the insurer does not know the true cost of the liability that it has incurred; that will only be known over time.

Now the actuaries inside the firm most of the time have a better idea than outsiders as to where reserve should be set to pay future claims from existing business, but even they don?t know for sure.? Some lines of insurance do not have a strong method of calculating reserves.? This was/is true of most financial insurance, title insurance, etc., and as such, many such insurers got wiped out in the collapse of the housing bubble, because they did not realize that they were taking one big nondiversifiable risk.? The law of large numbers did not apply, because the results were highly correlated with housing prices, financial asset prices, etc.

Even with a long-tailed P&C insurance coverage, setting the reserves can be more of an art than science.? That is why I try to underwrite insurance management teams to understand whether they are conservative or not.? I would rather get a string of positive surprises than negative surprises, and you tend to one or the other.

and

What is the company?s attitude on reserving?? How often do they report significant additional claims incurred from business written more than a year ago?? Good companies establish strong reserves on current year business, which depress current year profits, but gain reserve releases from prior year strongly set reserves.

So get out the 10K, and look for ?Increase (decrease) in net losses and loss expenses incurred in respect of losses occurring in: prior years.?? That value should be consistently negative.? That is a sign that he management team does not care about maximizing current period profits but is conservative in its reserving practices.

One final note: point 2 does not work with life insurers.? They don?t have to give that disclosure.? My concern with life insurers is different at present because I don?t trust the reserving of secondary guarantees, which are promises made where the liability cannot easily be calculated, and where the regulators are behind the curve.

As such, I am leery of life insurers that write a lot of variable business, among other hard-to-value practices.? Simplicity of product design is a plus to investors.

Today’s post analyzes Property & Casualty Insurers, and looks at their history of whether they consistently reserve conservatively each year. ? Repeating from above, management teams that reserve conservatively?establish strong reserves on current year business, which depress current year profits, but gain reserve releases from prior year strongly set reserves. ?This should give greater confidence that the accounting is fair, if not conservative.

So, I went and got the figures for “Increase (decrease) in net losses and loss expenses incurred in respect of losses occurring in: prior years,” for?67 companies over the past 12 years from the EDGAR database. ?Today I share that with you.

When you look at the column “Reserving by Year,” that tells you how the reserving for business in prior years went over time. ?A company that was consistently conservative of the past twelve?years would have “12N’ written there for twelve negative adjustments to reserves. ?Using Allstate as an example, the text is “5N, 1P. 3N, 3P” which means for the last 5 years [2013-2009], Allstate had negative adjustments to prior year reserves. ?In 2008, it had to strengthen prior year reserves. ?2007-2005, negative adjustments. ?2004-3,?it had to strengthen prior year reserves.

Now, in reserving, current results are more important than results in the past. ?Thus, in order to come up with a score, I discounted each successive year by 25%. ?That is, 2013 was worth 100 points, 2012 was worth 75, 2011 was worth 56, 2010 was worth 42 points, etc. ?Since not all of the companies were around for the full 12 years, I normalized their scores by dividing by the score of a hypothetical company that was around as long as they were that had a perfect score.

Now, is this the only measure for evaluating an insurance company? ?Of course not. ?All this measures?in a rough way is the willingness of a management team to reduce income in the short-run in order to be more certain about the accounting. ?Consult my 7-part series for more ways to analyze insurance companies.

As an example, imagine an insurance company that consistently writes insurance business at an 80% combined ratio. ?[I.e. 20% of the premium emerges as profit.] ?I wouldn’t care much about minor reserve understatement. ?Trouble is, few companies are regularly that profitable, and companies that understate reserves tend to get into trouble more frequently.

Comments and Surprises

1) Now, it is possible for a company to game this measure in the short run, where the management aims to always release some reserves from prior year business whether it is warranted or not. ?That may have happened with Tower Group. ?Very aggressive in growth, after their initial periods, they consistently released reserves for eight years, before delivering huge reserve increases for two years.

Now, someone watching carefully might have noticed a reserve strengthening for their non-reciprocal business in 2011, and then strengthenings in mid-2012, before the whole world realized the trouble they were in.

2) Notice in the red zone (scores of 40% and lower) the number of companies that did subprime auto insurance — Infinity, Kingsway, and Affirmative. ?That business is very hard to underwrite. ?In the short run, it is hard to not want to be aggressive with reserves.

3) Also notice the red zone is loaded with companies with much recent strengthening of reserves. ?Many of these companies are smaller, with a few exceptions — the law of large numbers doesn’t apply so well with smaller companies, so they mis-estimate more frequently. ?I won’t put companies with less than $1 billion of market cap into the Hall of Shame. ?It’s hard to get reserving right as a smaller company.

4) As for larger companies, they can be admitted to the Hall of Shame, and here they are:

Hall of Shame

• AIG
• The Hartford
• AmTrust Financial Services
• Mercury General, and?
• National General Holdings

AIG is no surprise. ?I am a little surprised at the Hartford and Mercury General. ?National General Holdings and Amtrust are controlled by the Karfunkels, who are aggressive in managing their companies. ?Maiden Holdings, another of their companies is in the yellow zone.

Final Notes

I would encourage insurance investors to stick to the green zone for their investing, and maybe the yellow zone if the company has compensating strengths. ?Stay out of the red zone.

This analysis could be improved by using prior year reserve releases as a fraction of beginning of year reserves, and then discounting by 25% each year. ?Next time I run the analysis, that is how I will update it. ?Until then!

Full disclosure: long TRV, ENH, BRK/B, ALL