BOA chairman Ken Lewis has written an opinion for the WSJ ("Some Myths About Banks") containing the following "myth" and rebuttal:
The banks are insolvent. In the past 18 months, we've seen fewer than 50 bank failures. That compares to about 2,000 failures or closings of commercial banks or savings institutions between 1986 and 1991. There may be more to come, but the vast majority of banks will weather this economic storm.
Lewis is absolutely right - only 50 banks have failed this year. But the absolute count is not indicative of the inherent risk (and potential catastrophe) which could result from such failures. Here's a chart I put together with data from the FDIC of banks seized/supported in each month since 1980:
The blue bars show the number of banks (left axis). As Lewis indicates, the number of recent failures is dwarfed by the S&L crisis in the 80's and early 90's. However, the red bars show the asset values of the seized banks (right axis), and the tall bar at the right represents almost half the value of the combined assets of banks seized during entire S&L crisis! That bar corresponds, of course, to the $300 million in assets held by Washington Mutual. This chart does not even show the $1.2 billion in assets held by the subsidiaries of Citigroup, which received an explicit FDIC guaruntee in late 2008 (those figures were actually in the FDIC data but I manually excluded them from the chart). Nor does it include the billions of implicit guaruntees the FDIC has made on bank-issued debt.
Bank assets may not be the best direct proxy for the FDIC's liabilities, but they certainly speak to the responsibility the agency assumes in seizing a bank. Needless to say, it is a strange argument to make that the number of banks failed - rather than size or impact of failure - is a chief measure of the economy's health. Even the S&L crisis is better characterized by cost than by quantity (though the two are correlated, lending spurious truth to silly claims like this one).
But I have digressed from Lewis' actual point that the reduced number of bank failures is indicative of solvency. He's right - fewer banks are have been declared formally insolvent than in the 1980's. But again, however, is the number or size of greater concern? If three banks go under which are sufficiently large to take the FDIC with them (not to mention the rest of the world), then does it matter that they numbered only three?