As Paul mentioned in the comments on the last post, I answered a question at Stockpickr.com today. James Altucher, the bright guy who founded the site, asked me if I would answer the question, and so I did. Here is a reformatted version of my answer, complete with links that work:

Q: Any thoughts on how to value an insurance business? What are the best metrics to use? In particular I’m looking at small cap insurers (P&C) as potential acquisition targets. Does that change the methodology?

A: That’s not an easy question, partly because there are many different types of insurance companies, and each type (or subsector) gets valued differently due to the degree of growth and/or pricing power for the subsector as a whole.

Now, typically what I do as a first pass is graph Price/Book versus return on equity for the subsector as a whole, and fit a regression line through the points. Cheap companies trade below the line, or, are in the southeast corner of the graph.

But then I have to make subjective adjustments for reserve adequacy, excess/noncore assets, management quality, pricing power on the specific lines of business that write as compared to their peers, and any other factors that make the company different than its peers. When the industry is in a slump, I would have to analyze leverage and ability of the company to upstream cash from its operating subsidiaries up to the parent company.

Insurance is tough because we don’t know the cost of goods sold at the time of sale, which requires a host of arcane accounting rules. That’s what makes valuation so tough, because the actuarial assumptions are often not comparable even across two similar companies, and there is no simple way to adjust them to be comparable, unless one has nonpublic data.

My “simple” P/B-ROE method above works pretty well, but the ad hoc adjustments take a while to learn. One key point, focus on management quality. Do they deliver a lot of negative surprises? Avoid them, even if they are cheap. Do they deliver constant small earnings surprises? Avoid them too… insurance earnings should not be that predictable. If they become that predictable, someone is tinkering with the reserves.

Good insurance managements teams shoot straight, have occasional misses, and over time deliver high ROEs. Here are three links to help you. One is a summary article on how I view insurance companies. The second is my insurance portfolio at Stockpickr. The last is my major article list from RealMoney. Look at the section entitled, “Insurance & Financial Companies.”

Now, as for the small P&C company, it doesn’t change the answer much. The smaller the insurance firm, the more it is subject to the “Law of Small Numbers,” i.e., a tiny number of claims can make a big difference to the bottom line result. Analysis of management, and reserving (to the extent that you can get your arms around it) are crucial.

As for takeover targets, because insurers are regulated entities, they are difficult to LBO. Insurance brokers, nonstandard auto writers, and ancillary individual health coverage writers have been taken private, but not many other insurance entities. State insurance commissioners would block the takeover of a company if it felt that the lesser solvency of the holding company threatened the stability of the regulated operating companies. The regulators like strong parent companies; it lets them sleep at night.

One more note: insurance acquisitions get talked about more than done, because acquiring companies don’t always trust the reserves of target companies. Merger integration with insurance companies has a long history of integration failures, so many executives are wary of being too aggressive with purchases. That said, occasionally takeover waves hit the insurance industry, which often sets up the next round of underperformance, particularly of the acquirers.