(h/t modeled behavior) Here is David Henderson (a full article):
Opinion: Making Banks Too Small to Succeed?
…His plan would limit both the size of banks and the kinds of assets they are allowed to invest in. In other words, the government would take one the most regulated industries in the country – and regulate it more.
Unfortunately, this plan, if implemented, would go against much of what economists know about the value of size. Moreover, it would do nothing to address the real problem: “moral hazard.”
As Columbia University financial economist Charles Calomiris pointed out last fall, when it comes to banks, size has its advantages. A bank that wants to deal in a large international market is better off if it’s big. Those who deal with it are also better off because, to the extent that there are economies of scale, some of the cost savings make their way to customers. Also, all other things equal, the larger the bank, the more diversified it is. The more diversified it is, the less likely it is to fail….
Because banks during the Great Depression were so small, they were undiversified. So when the agriculture sector went under, in part because of the Smoot-Hawley Act that attacked free trade, many rural banks failed. Call it “too small, so we failed.”….
Had they been allowed to be big, many fewer would have failed. It’s worth noting that in Canada, which also had a downturn in its farm sector, the banks were larger and not one failed during the Great Depression….
We can. And the best way to do that is to give those closest to banks a strong incentive to make sure the banks aren’t taking undue risks.
One way to do that would be to get rid of deposit insurance….
A bunch of things:
1) SAndrew, one of the better finance bloggers/tumblr-ers who needs to be writing more, said in comments that he didn’t believe that someone could blame FDIC for the current financial problems. There you go.
Stockholders, CEOs, board members, regulators, ratings agencies and sophisticated financial players assessing counterparty risk couldn’t make heads-or-tails of what was going on in the financial markets – looks like it’s up to you and your $12,000 savings account to discipline the financial markets grandma!
I’ll come back to this line of argument later because it is becoming more popular, but a Rothbardian friend of mine (everyone should have one) told me a while ago, which discussing the idea of a CFPA, “Mike, the best form of consumer financial protection is a good old-fashioned bank run.” wtf? It was like the exact opposite of any argument I’ve normally consider.
2) “Too small, so we failed”, ha! What idiots back then. Could you imagine a financial sector so dangerous that a single agriculture sector failure caused it to fail? It’s ridiculous – like imagining a financial sector that collapsed because 20% of the mortgage “crop” planted between 2004-2006 failed, crops used as collateral to leverage into the repo markets. Oh wait….
3) James took apart that paper that Calomiris quoted for an example of returns to scale and found it lacking. You can read his critique there.
4) I rarely see studies that propose a return to scale past the 25 billion dollar mark. See this survey of the literature. That’s way less than anyone is proposing as a size limit. Nobody would argue that a bank should be less than $100 million in assets; there’s clear evidence that there is scaling effects up to that point. Recent studies dig pretty deep into the statistical toolbox to pull up returns to scale past $100 billion, but those usually end at 2006, at the peak of a credit cycle.
5) The argument for scale usually involves saying that only with huge scale can a firm credibly “warehouse” huge portfolios of derivatives, thus making market-making more efficient since a firm can internally hedge, etc. I talked about that argument here.
6) I created this chart here, using data from the “Stress Test” results that were issued by the Federal Reserve. This is the Fed asking the 20 largest bank-like companies to estimate their losses over the next year. So this is the bank’s word, not mine. I graph the expected percent of losses against size:
I would love to see more sophisticated research done, but looking at that is there any reason to believe the optimal size of a bank is past the $1trillion dollar mark? Is this the magic “diversification” everyone is talking about? There is well-documented regression to the mean for skill in finance in terms of size (see all the mutual fund literature, for one), and the only way to keep returns higher than normal past a certain size is to take on more leverage, and that’s exactly what has gotten us in the current situation.
Also note when those four largest banks fail, what the hell do we do? The next guy down who could credible take it over is going to be 1/4th the size of that institution or much, much less. We can force it into one of the other largest firms, punting on the issue….
Rortybomb 
Original: “nobody’s seriously attacking the FDIC, are they?”
Edit: “nobody serious is attacking the FDIC, are they?”
I kid. I’m sure Henderson is a more serious economist than I’ll ever be. Still, I’m surprised that there’s a cadre of (libertarian?) economists still out there who would like to see an end to deposit insurance. I’ve had Russ Robert’s Econtalk with Calomiris on my iPhone for a few weeks. Guess I should give it a listen.
Oh, and thanks for the shout-out, Mike.
The law of diminishing returns becomes the law of decreasing efficiency as human systems grow beyond a certain size. More and more resources need to be committed to management and controls, and this is seldom done because it greatly increases costs. Therefore, these inefficiencies result in untoward occurrences that eventually threaten system breakdown.
there’s a myriad ways of arriving at the same conclusion. You did it your way, I simply saw the words “Charles Calomiris” and reasoned by induction “all previous comments from Charles Calomiris have been wrong; therefore I assume this is also wrong”. We both got to the same place in the end.
Russ Roberts is definitely someone who falls in that zone of the world. I listened to a podcast of his recently (I can’t recall which one) where they talked about this type of moral hazard. (Odd to me – they focused on how bondholders being protected meant that they would engage in less oversight. They ignored the fact that bondholders being unprotected in the status quo ante mean that *managers* could be less cautious because they knew someone else would be performing some kind of monitoring role. A very partial equilibrium view of the world, I thought, but consistent with a lot of the “FDIC is bad” group.)
I understand the economics, certainly, but I guess I just completely disagree in terms of its universal application. I mean, the fact that I’ll take my kid to the emergency room if he falls off the roof increases the likelihood that he’ll do it, I suppose. And I guess in some infinitesimally small way I may drive more recklessly because I have insurance. And I certainly don’t investigate the financial condition of the firms I buy insurance from, knowing that they are state-regulated.
But in no way do I think that medical care, auto insurance, or insurance regulation are bad things to have. Quite the opposite.
FWIW, I don’t think we should abolish the FDIC, and I don’t think big banks are inherently great, although I haven’t read Charles Calomiris on this, so I’m open to persuasion in either direction. What I thought was the important point that David made was that small banks are no panacea. Small banks being blamed for causing panics goes back well into the 1800s.
Krugman mentions Canada today- He says their relatively better performance is due to regulation.
http://krugman.blogs.nytimes.com/2010/01/30/worthwhile-canadian-example/
Canada’s lack of bank failures probably also has to do with branch banking. Isn’t there any way to compromise, to shrink the size of the biggest banks while keeping their national reach? I think it can be done. Split each of Wells Fargo, Citibank, Chase, and BofA in half, but not in half east-west. Instead, split the branches such that the branches of Citibank A are close to Citibank B, and vica versa, so they are spread out as much as possible.