If you're going to invest in bank stocks, it's important to buy the best. Even though poorly run banks can produce spectacular short-term returns, over the long run they expose investors to severe dilution, if not abject failure.
Our history of panics
In the history of the United States, we've experienced somewhere around 17 banking panics. That's more than any other country in the world, perhaps with the exception of Argentina, and it means that on average, every dozen years or so we experience a crisis in which hundreds, sometimes thousands, of banks are rendered insolvent.
A bank's unique susceptibility to failure stems from two simple facts. First, banks are highly leveraged, generally borrowing $9 for every $1 of capital. This means a mere 10% decline in the value of a lender's assets, which consist largely of loans and fixed-income securities, can completely wipe out its capital base.
And second, banks are run by humans, who are prone to a variety of behavioral biases and emotional impulses that are particularly ill-suited for banking. In the midst of an asset bubble, for instance, it's natural for a banker's emotions to convince them to continue lending on over-inflated collateral, even though doing so will expose the bank to greater losses in the future.
Perhaps no one has captured these tendencies better than Fred Schwed, Jr. in his classic satire on Wall Street, Where Are the Customers' Yachts?:
Your truly conservative banker cannot be stampeded into unwary speculations by the hysteria of a boom. [...] He sits tight through 1926, 1927, and 1928. Unfortunately, he begins to come into the market in 1929. He begins cautiously enough, like an old maid trying out lipstick in the privacy of her room. (Watching these young whippersnappers make fortunes for three long years does something to the sturdiest character.) But he pulls out again, and, while a nice piece of money is lost, no one is ruined. He apologizes to himself for having had a human moment and resumes his thirty-year-old policy of listening attentively and saying "no."
How to identify good banks
If the goal is to avoid bad banks, then how does one go about identifying good ones? As I've discussed at length elsewhere, the answer to this is more straightforward than you might think, given how complex banks have become in recent decades.
The most important statistic to examine is the efficiency ratio, which is a bank's noninterest (e.g., operating) expenses divided by its net revenue. This tells you how much of a bank's revenue is set aside to cover future loan losses, pay taxes, and distribute or accrue to shareholders.
The objective is to have a ratio between 50% and 60% -- though, depending on a bank's net revenue relative to its assets, some banks can generate outsized returns with efficiency ratios as high as 65%. Of the nation's biggest banks, only Wells Fargo and U.S. Bancorp consistently fall within the ideal range.
This metric matters not only because it shows how efficient a bank is, but also because of its spillover effect on a bank's risk management: Less efficient banks must compensate for the drag on their earnings, and they tend to do so by originating higher-yielding, and thus riskier, loans.
It's for this reason Warren Buffett has said in the past that "only a very low-cost operator or someone operating in a protected, and usually small, niche can sustain high profitability levels" in highly competitive and commoditized industries like banking.
So take it from Buffett, not from me: If you want to buy the best bank stocks -- i.e., those that can outperform their peers without subjecting you to an outsized risk of failure -- then the best place to begin, and arguably even end, your search is with the efficiency ratio.