The Continental Illinois Bank Building in Chicago. Commonly referred to as the first too-big-to-fail bank, Continental Illinois was seized by the FDIC in 1984 after a fear of bad loans ignited a bank run. Source: Wikimedia Commons.

In 2003, Howard Marks, one of the world's most successful investors, wrote a memo to his clients titled "The Most Important Thing" (link opens PDF). He then proceeded to list 18 such factors, ranging from the avoidance of leverage to the necessity of a contrarian approach to investing. But ascending above the rest was, and remains, the relationship between price and value. It truly is the most important thing.

I thought about this recently when I read through a presentation that the chief executive officer of New York Community Bank (NYCB 2.46%) gave at this year's Barclays Global Financial Services Conference. Amidst the requisite self-congratulatory introduction, CEO Joseph Ficalora made two comments that were particularly notable.

He observed that NYCB's compounded annual growth rate since its initial public in 1993 is nearly 28%, equating to a total return of more than 3,400%. That's on par with the greatest businesses and money managers of our era. On top of this, he noted that the bank has had 52 consecutive quarters, "wherein we've had [zero] losses on the assets that we created."

This is astounding. Those 52 quarters include both the more recent financial crisis and the brief recession following the bursting of the dot-com bubble. Over the last five years alone, nearly 500 banks have been closed by the FDIC, and countless others have suffered debilitating losses that forced them to dramatically dilute their existing shareholders.

Indeed, the more representative examples are those of KeyCorp (KEY -1.23%), Huntington Bancshares (HBAN -0.38%), and Citigroup (C -1.88%), which have lost 46%, 59%, and 69% of their book values, respectively, since the beginning of 2008.

How did NYCB avoid a similar fate? The answer is deceivingly simple. "When you think about what, in fact, distinguishes a bank as a lender," Ficalora said, "it's how much money it loses on the asset that it chooses to take risk with."

The corollary of this rule was articulated by Carl Webb, the second in command at Diamond A Ford, an exceptionally successful bank buyout firm founded by the billionaire Gerald J. Ford. "Banks get in trouble for one reason," Webb observed. "They make bad loans."

The most important thing for running a successful bank, in other words, is not originating bad loans.

That's it. It's not about hedging. It's not about creating a diversified business model with credit cards and investment banking units. It's not about making smart acquisitions. What matters is the quality of the assets that a bank originates and holds onto.

And lest there be any doubt, aside from NYCB, the most successful banks over the last few years have not been the universal banks like Citigroup and Bank of America (BAC 0.45%), both of which have seen their respective stock prices decline by 73% and 55% over the last five years as a result of bad loans and imprudent acquisitions.

Instead, they've been the stodgy, old-time lenders like U.S. Bancorp that never abandoned their focus on the principal functions of banking, gathering deposits, and writing good loans.