Image by Herve Boinay

There's a common refrain among bank analysts that investors should buy bank stocks when they're priced at half of book value and then sell them at two times the same measure -- that is, buy them when they're cheap and sell them when they're dear.

It's something I've quoted before. And it's something that's wrong. If anything, in fact, the opposite is true; you should only buy bank stocks that trade at a healthy premium to book value relative to others in the industry.

The siren song of value investing
The main problem with using valuations to inform a bank investment is that it presupposes a finite holding period.

Banking is a cyclical industry based in large part on interest rate fluctuations. When interest rates are high, banks make great money on the spread between the cost of deposits, many of which are often noninterest bearing, and the yield on their loans and securities portfolio.

Because banks earn higher profits in these periods, their stock prices typically respond in kind by trading at a higher multiple to book value. By contrast, when interest rates are low, as they are now, banks make less money and correspondingly trade for a lower multiple to book value.

These fluctuations are an inherent and not infrequent aspect of the market for bank stocks. The experiences of Wells Fargo (NYSE: WFC) and JPMorgan Chase (NYSE: JPM), two of the most stable banks over the last two decades, serve as cases in point.

In the mid-1990s, both were trading at fairly reasonable levels given the respective qualities of the operations. Shares of Wells Fargo were down around two times book value while JPMorgan's bottomed out around one times book.

Within five years, however, both shot up considerably. Wells Fargo's more than doubled while JPMorgan's roughly tripled. Fast forward another five years and they were back down in the neighborhood of their former levels.

The three problems with low multiples
Now, I know what you're thinking. Far from proving my point that one should ignore valuations, this actually refutes it. Had you correctly timed the cycle around the turn of the century, you would have made a pretty penny with either of these stocks.

But herein lies the issue -- or, rather, issues. In the first case, it's impossible to time the market on a systematic basis. I concede that everybody gets lucky once in a while, but relying on luck as a strategy is imprudent to put it mildly.

In the second case, moving in and out of the market causes you to sacrifice the power of compounding returns. I can't emphasize enough how important this mathematical phenomenon is. In my opinion, it's the only guaranteed way to achieve extraordinary results over the long run.

Finally, at least when it comes to bank stocks, the absolute last thing you should be interested in is a substandard company -- which, while the market isn't perfect, is precisely what a comparatively low valuation multiple implies.

Take Citigroup (NYSE: C) as an example. Presently, its shares trade for a 28% discount to book value. That's the largest discount among top banks. The only one that comes even close is Bank of America (NYSE: BAC), which trades for a 20% discount to book.

Why is Citigroup so, as some might see it, cheap? The answer is that, you get what you pay for.

Since at least the Great Depression, Citigroup has been one of the least consistent operators in the financial space -- coming within a hair's breadth of failing on multiple occasions. During the most recent crisis, its shareholders were diluted so egregiously that it could take upwards of a century before they're made whole.

By comparison, a bank like New York Community Bancorp, which didn't see its valuation drop to even close to the same levels as Citigroup's, was strong enough to reject TARP assistance and avoid a dilutive share offering. How did it do this? Simply by running a great bank.

So, if valuations don't matter, what does?
The point here is simple. When you pick bank stocks, pick them for the right reasons. And being "cheap" isn't one of them.

Buy the best. Hold onto them for a long time. And let your investments do the work for you.

That's how people get rich.