On March 1, Warren Buffett released his latest annual letter to Berkshire Hathaway shareholders. Many in the investment world consider it required reading. And over the last 49 years, Berkshire's book value per share growth (which Buffett considers "a rough tracking indicator" of intrinsic value growth) has been 19.7% compounded annually, so clearly Buffett and team have an exceptional ability to create shareholder value.

I encourage you to read the letter in its entirety, but here's one kernel of wisdom I consider highly applicable to the banking business:

"As I've long told you, Berkshire's intrinsic value far exceeds its book value. Moreover, the difference has widened considerably in recent years. That's why our 2012 decision to authorize the repurchase of shares at 120% of book value made sense. Purchases at that level benefit continuing shareholders because per-share intrinsic value exceeds that percentage of book value by a meaningful amount. We did not purchase shares during 2013, however, because the stock price did not descend to the 120% level. If it does, we will be aggressive."

Berkshire has grown book value per share at 20% per year, for decades. Of course, it pays no dividends, which complicates the comparison of it to a company with a 40-50% dividend-payout ratio, like a bank. Furthermore, price appreciation in Berkshire's securities portfolio contributes meaningfully to book value growth, and banks don't generally dabble in stocks. But 20% is still Berkshire's long-term ROE. I doubt that a 20% ROE bank (a rarity) could maintain such an ROE for long if it didn't pay out 40-50% of earnings as dividends, plus some more on repurchases. Lowering "E" helps banks increase "ROE".

Berkshire chooses not to repurchase shares at a price higher than 1.2x book value (or about 1.7x tangible book). What does that imply about all the banks and thrifts with ROE far below that of Berkshire that buy in shares at equivalent or higher book value multiples?

It implies that they're doing something reckless. No, it screams that they're doing something reckless.

Some banking institutions have actually shrunk tangible book value per share since 2000. This is sometimes due to post-2008 share issuance-related dilution, but it is just as often due to overpriced acquisitions and/or share repurchases. But Wells Fargo, a Buffett favorite, has actually grown tangible book value per share faster than Berkshire has since the end of 2000 (13.5% compounded annually versus 9.8%, per my calculation).

If a management team of a bank with low tangible book value per share growth were to respond that not all banks can be Wells Fargo, they'd be right. This bank can't replicate all the things that make Wells great, but that doesn't mean that it should close or sell, any more than the emergence of Tiger Woods should lead hordes of professional golfers to quit the game.

But the presence of a Tiger Woods motivates all pro golfers to try harder. When you compare some low-growth banks' strategic decisions to those of Wells, it's almost as if the low-growth banks think they're playing a completely different game — like a game of golf in which high scores are better than low ones. Otherwise, they'd presumably feel humiliated by falling so short.

Money managers pay homage to Buffett's investment record when his name is mentioned, without them necessarily choosing to employ his simple, almost folksy, yet time-tested investment philosophy. What Buffett insights are especially relevant to bank CEOs? I'd say there are three:

Care deeply about the price at which you make an investment, whether it's an investment in your own shares or in another company. Don't delude yourself about how great the future will be. And look for empirical evidence that you have an "edge", rather than simply assuming it.

What would the U.S. banking business look like in five years if every bank could get advice from Warren Buffett today, and they all listened to it?

Harvard Winters, a former investment banker, writes research on banks.