How big a bank is too big to fail?

Published by the R Street Institute.

The notion of “too big to fail”—an idea that would play a starring role in banking debates from then to now—was introduced by then-Comptroller of the Currency Todd Conover in testimony before Congress in 1984. Conover was defending the bailout of Continental Illinois National Bank. Actually, since the stockholders lost all their money, the top management was replaced and most of the board was forced out, it was more precisely a bailout of the bank’s creditors.

Continental was the largest crisis at an individual bank in U.S. history up to that time. It has since been surpassed, of course.

Conover told the House Banking Committee that “the federal government won’t currently allow any of the nation’s 11 largest banks to fail,” as reported by The Wall Street Journal. Continental was No. 7, with total assets of $41 billion. The reason for protecting the creditors from losses, Conover said, was that if Continental had “been treated in a way in which depositors and creditors were not made whole, we could very well have seen a national, if not an international crisis the dimensions of which were difficult to imagine.” This is the possibility that no one in authority ever wants to risk have happen on their watch; therefore, it triggers bailouts.

Rep. Stewart McKinney, R-Conn., responded during the hearing that Conover had created a new kind of bank, one “too big to fail,” and the phrase thus entered the lexicon of banking politics.

It is still not clear why Conover picked the largest 11, as opposed to some other number, although he presumably because he needed to make Continental appear somewhere toward the middle of the pack. In any case, here were the 11 banks said to be too big to fail in 1984, with their year-end 1983 total assets – which to current banking eyes, look medium-sized:

If you are young enough, you may not remember some of the names of these once prominent banks that were pronounced too big to fail. Only two of the 11 still exist as independent companies: Chemical Bank (which changed its name to Chase in 1996 and then merged with J.P. Morgan & Co. in 2000 to become JPMorgan Chase) and Citibank (now Citigroup), which has since been bailed out, as well. All the others have disappeared into mergers, although the acquiring bank adopted the name of the acquired bank in the cases of Bank of America, Morgan and Wells Fargo.

The Dodd-Frank Act is claimed by some to have ended too big to fail, but the relevant Dodd-Frank provisions are actually about how to bail out creditors, just as was the goal with Continental. Thus in the opposing view, it has simply reinforced too big to fail. I believe this latter view is correct, and the question of who is too big to fail is very much alive, controversial, relevant and unclear.

Just how big is too big to fail?

Would Continental’s $41 billion make the cut today? That size now would make it the 46th biggest bank.

If we correct Continental’s size for more than three decades of constant inflation, and express it in 2016 dollars, it would have $97 billion in inflation-adjusted total assets, ranking it 36th as of the end of 2016. Is 36th biggest big enough to be too big to fail, assuming its failure would still, as in 1984, have imposed losses on hundreds of smaller banks and large amounts of uninsured deposits?

If a bank is a “systemically important financial institution” at $50 billion in assets, as Dodd-Frank stipulates, does that mean it is too big to fail?  Is it logically possible to be one and not the other?

Let us shift to Conover’s original cutoff, the 11th biggest bank. In 2016, that was Bank of New York Mellon, with assets of $333 billion. Conover would without question have considered that—could he have imagined it in 1984—too big to fail. But now, is the test still the top 11?  Is it some other number?

Is $100 billion in assets a reasonable round number to serve as a cutoff? That would give us 35 too big to fail banks. At $250 billion, it would be 12. That’s close to 11. At $500 billion, it would be six. We should throw in Fannie Mae and Freddie Mac, which have been demonstrated beyond doubt to be too big to fail, and call it eight.

A venerable theory of central banking is always to maintain ambiguity. A more recent theory is to have clear communication of plans. Which approach is right when it comes to too big to fail?

My guess is that regulators and central bankers would oppose anything that offers as bright a line as “the 11 biggest”; claim to reject too big to fail as a doctrine; strive to continue ambiguity; and all the while be ready to bail out whichever banks turn out to be perceived as too big to fail whenever the next crisis comes.