Tom Brown knows bank stocks. For the past 17 years, he's been analyzing and investing in them at his hedge fund, Second Curve Capital. How does someone like Brown pick banks to invest in? I asked him that in a recent conversation.

The founding of Second Curve Capital

Brown spent years cultivating a reputation as a top analyst of financial stocks. He worked at a number of prominent Wall Street firms throughout the 1980s and 1990s, before joining Julian Robertson's hedge fund, Tiger Management, in 1998 to lead its financial-services group. Two years later, Brown founded Second Curve Capital, a hedge fund that focuses on banks and other companies in the financial services sector. He's run the fund ever since.

The last decade and a half have been a minefield for bank investors, with the financial crisis causing hundreds of banks to fail and forcing many others to dilute shareholder value by raising new capital at the nadir of the downturn. In December 2009, for instance, in the largest public equity offering in history at the time, Citigroup (NYSE:C) issued 5.4 billion shares of common stock at around $3 each. Just two years earlier, its book value per share had been as high as $25. The impact on Citigroup's shareholders from this and other crisis-era moves will take decades to reverse.

C Chart

Citigroup data by YCharts.

Brown's fund has persevered through all of this, netting a 20% average annual return, according to Insider Monkey, which tracks hedge fund performance. Today, Second Curve Capital's latest 13F lists positions in 24 companies, including Citigroup, Bank of America (NYSE:BAC), Capital One Financial (NYSE:COF), and Zions Bancorporation (NASDAQ:ZION).

How Brown analyzes bank stocks

Brown's approach to analyzing bank stocks is relatively straightforward:

We evaluate banks based primarily on two factors as a screening tool. First, what do we think they can make two years from now? And second, what's a fair multiple for them to trade at? Then we rank them all and do more work on the ones that look the most undervalued and we try to understand why the other end of the spectrum are most overvalued.

To determine how much a bank is likely to make in two years' time, Brown uses an estimate of a bank's normalized earnings. This neutralizes for factors -- such as elevated credit costs in a recession or expenses following a merger or acquisition -- that temporarily weigh on a bank's performance.

"So, in 2009 when a company wasn't making any money, we'd say what do we think they can earn when they've rebuilt their capital and their credit problems are gone," said Brown.

Staying the course

This in and of itself isn't unique, as other investors and analysts use similar tactics. What sets Brown's approach apart, however, is that he maintains his focus on normalized earnings through all stages of the credit cycle. Other investors and analysts tend to alternate between valuation metrics depending on the state of the economy.

As Brown explained:

At the beginning of the 1980s, price-to-book value was the preferred valuation metric, particularly for smaller banks. Then we went to an earnings multiples in the second half of the '80s and '90s. Then in the Great Recession, everyone switched back to price-to-book. And then in the last 12 months, we've seen most but not all go to an earnings-driven valuation model.

It was only recently that analysts at Keefe, Bruyette & Woods, or KBW, an investment bank that serves the financial services sector, switched back to using an earnings multiple to assess bank valuations for the first time since the crisis. "We switched last year around the time of Brexit," KBW's Brian Kleinhanzl told me. Other prominent analysts and investors have said the same thing.

Stock traders at work.

Investors can learn a lot from studying how hedge funds analyze stocks. Image source: Getty Images.

The one aspect of Brown's approach that does change depending on the economy is the period over which Second Curve projects a bank's normalized earnings. Ordinarily, as already noted, Brown forecasts a bank's earnings two years out. But in a recession, while he uses the same methodology, he employs a longer period of time to get to normalized earnings, and then discounts that value back.

A lesson for retail investors

For the typical retail investor, both approaches present challenges. Switching between earnings and book value multiples presupposes a familiarity with both metrics, as well as the ability to gauge when to alternate between the two. Meanwhile, Brown's approach presupposes a detailed enough familiarity with specific banks, and banking business models more generally, to be able to assess a bank's normalized earnings power.

These complications aside, an important lesson to take away from Brown's battle-tested method of analyzing bank stocks is to compare a bank's current valuation to what that bank seems likely to earn in the future. This may seem obvious, but most retail investors probably don't take the time to actually do so.

John Maxfield owns shares of Bank of America. His cousin is a managing director at Second Curve Capital. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.