Many investors look askance at insurance companies, because the insurance business is somewhat obscure and because management estimates play a huge role in determining results. So I thought it'd be a great exercise to talk to an insurance-industry executive about getting to understand the process of pricing and setting reserves. The following is a phone conversation I had with an industry executive, who chose to speak anonymously.

Q: What are some signs that outside investors should look for in determining whether an insurer conservatively or aggressively states its loss reserves?

A: In terms of reserving -- if you look at the past 10 years of [reserve] development [in the cumulative redundancy tables in an insurer's 10-K], that's a good starting point. To dig in deeper, look at the NAIC [National Association of Insurance Commissioners] statutory financials. Schedule P in the annual statutory filings contains a good deal of information by line.

But track record of management is also a key factor in assessing a company's reserve strength, too. A management team that had conservative reserves five to 10 years ago is probably conservative now. Management's track record is crucial because of the conundrum with reserves -- companies doing very well are going to want to hold higher reserves, and companies doing poorly will be inclined to hold lower reserves, so when you look at the current year's results of various companies, they will tend to look the same. But if you look at their various track records over time, you may see a different picture.

Q: As investors, we all have access, for the most part, to the same data -- SEC filings and so on -- so data quality isn't a big competitive advantage. For an insurer, is there a discrepancy between how good one insurer's data is versus another?

A: There is a tremendous difference in data quality, and that difference can be critical. Any model is as only as good as the data that goes into it. A number of companies may be struggling with 20-year-old legacy systems, or have multiple systems to integrate. It will be tougher for them to have robust data. I think many observers of the industry have this idea that the magic is in the models, but in reality, 80% or more of the battle is getting the good, accurate, timely data.

Q: How can an outside investor judge an insurer's data quality?

A: That's difficult, but you could look at what kind of information that insurer can produce. Do [specific insurers] generally produce a lot of specific and detailed information for the investor, or do they tend to produce relatively little and more anecdotal information? This is something I suspect investors could press more on by asking more specific questions about what sorts of information the company is able to produce.

Q: I remember listening to a Progressive (NYSE:PGR) conference call, and I was impressed with the way the people on that call could slice and dice their data in very specific ways. Is that what you're talking about?

A: Progressive is a good example of someone who seems to have a good hold on their data. The volume of information they produce ... should give investors comfort that they pay a lot of attention to data quality.

Q: Please describe your thoughts on risk modeling. What went wrong [in the insurance industry] with Hurricanes Katrina, Rita, and Wilma? Is this due to the risk premiums cycle?

A: There will always be an insurance cycle, but Katrina and Wilma were more about risk management. Rates were still relatively high for catastrophe business back then. Going back to 2001 [after the 9/11 attacks], there was a massive hardening of property-insurance rates. Rates at the time of Katrina were good, but the issue was that not everyone had correctly assessed their true downside from property catastrophe risk. In many cases, you had insurers and reinsurers who thought they were protected to a one-in-250-year event or a one-in-100-year event, based on a model that turned out to be wrong.

There are always problems in predicting the tail of any distribution [in terms of infrequent events], because you're trying to predict something that, by definition, doesn't happen a lot. If you're trying to predict the worst event in 250 years, there's not 250 years' worth of data to look at, and your odds of getting the estimate wrong are high.

Q: Which of your competitors do you admire and why?

A: Shand Morahan, part of Markel (NYSE:MKL), and Admiral Insurance Company, part of WR Berkley (NYSE:BER), have reputations for excellent underwriting. Based on what I see, those two companies appear to be disciplined competitors who pay a lot of attention to coverage.

While we don't compete with them, Geico [Berkshire Hathaway's (NYSE:BRK-A) auto-insurance subsidiary] is easy to admire, because of their expense advantage. And I think Arch Capital Group (NASDAQ:ACGL) had good market timing and good execution. Starting from 2001, they put together a really well-run, diversified insurance group with excellent results.

Q: How does an insurer manage its operating expenses?

A: The biggest expense, aside from losses, for many insurers is acquisition cost, so that's the most obvious place to gain an expense advantage. A company with low acquisition costs will almost certainly have an expense-ratio advantage. Beyond that, the biggest expense for most insurers is salaries.

In a competitive labor market, you have to pay competitive wages, so keeping your overall salary costs low is going to be a function of having highly productive workers. That, in turn, is a function of using technology to drive efficiency, and having a relatively flat, simple hierarchy. But ultimately, expense control has to be ingrained in the corporate culture. Every employee needs to be watching the expenses as though they were his or her own.

Q: Do you think pricing for 2007 is softening too much?

A: It's hard to generalize, because even in the hardest markets there will be pockets of underpriced business, and even in the softest markets there are pockets of opportunity. But overall, I think the prices are OK right now. I think there are some segments that were so distressed during the height of the hard market that they reached price levels that are difficult to sustain. Coming down off those very high rates is not too terribly troubling, and even inevitable.

The cycle is unavoidable; that's just human nature. It's hard for me to judge how long the current favorable pricing environment might last -- right now and for the next 12 to 18 months, I don't see rates being a problem. After that, who knows? There's definitely going to come a time when rates are going to become a problem. Companies that are disciplined will have to stop growing or even pull back a bit. One thing to keep an eye on is which companies are willing to pull back when pricing reaches their pain threshold. For those companies, there will be some short-run pain, but in the long run, those disciplined insurers will be better positioned to flourish in the next hard market.

The wild card in assessing the adequacy of pricing levels is the legal environment. The courts and the legislatures can have a tremendous impact on losses, and it is difficult if not impossible to predict how they might act in the future. For a number of years, the insurance industry had benefited from some favorable trends -- tort reform in particular. But it is impossible to say if those trends will continue, or if they will reverse. If the industry suffers major setbacks in the courts or the in legislative environment, rates that currently appear adequate may prove to be insufficient.

Related Foolishness:

Berkshire Hathaway is a Motley Fool Inside Value recommendation. Try our value-oriented newsletter service free for 30 days.

Fool contributor Emil Lee is an analyst and a disciple of value investing. He doesn't own shares in any of the companies mentioned above. Emil appreciates your comments, concerns, and complaints. The Motley Fool has a disclosure policy.