I've been asking for some time now what amounts to a simple question, "What is the minimum efficient scale for financial institutions?" Unfortunately, the answer is much harder to formulate than the question, and there hasn't been anything convincing (at least to me) on this topic. Here's the latest on this topic from Neil Irwin:
The case for the too-big-to-fail banks, by Neil Irwin: ...[I]n the last few months, there's been a new wave of calls to break up the "too big to fail" banks that were at the center of the crisis — and the beneficiaries of a massive wave of bailouts.
So, is splitting those banks up the answer? ... The move ... has a growing list of powerful allies. ... But what is the counterargument? ... A new paper from Patrick Sims of Hamilton Place Strategies, a policy and communications firm led by Bush administration White House and Treasury official Tony Fratto, amounts to a case for the big banks. (Hamilton Place counts major banks and their trade associations among its clients)..., here are some of the arguments that I believe have the most merit.
The first argument is also the simplest. This is a huge and complex global economy. With trillions of dollars in global trade and companies with hundreds of billions in assets, it takes giant banks with a global reach to supply them with the financial products they need to do business. ...
This is related to a second line of argument: That without huge, regulated banks, companies will turn to the unregulated "shadow banking" sector to meet their financial needs... The collapse of the shadow banking system was a crucial factor in the 2008 crisis... If a bank break-up sparked a vast expansion of shadow banking, it would make the financial system more vulnerable.
I find less persuasive Sims's argument that we shouldn't worry too much about concentration in the banking system because it's not quite as severe in the United States as in other countries. One of the paper's tables notes that the largest U.S. bank, JPMorgan Chase, has assets that amount to 15 percent of U.S. GDP...
As the too-big-to-fail debate heats up, these are the questions... What real value do megabanks create for the world? Do they need to be this huge to fulfill their roles? Can regulation be enough to tame their inherent risks, or do they need to be split up? ...
Here are some past posts on this topic. This was a response to a speech by William Dudley on the too big to fail problem (November, 2012):
Solving the Too Big to Fail Problem: ...One thing we really need to understand better is the minimum efficient scale for various financial activities. Whenever the topic of breaking banks into smaller pieces is raised, we hear that a "much reduced size threshold could sacrifice socially useful economies of scale and scope benefits." The key word is "could." As far as I can tell, the evidence on this point is very shaky -- we just don't know for sure what size is necessary to exploit efficiencies. My own view is that the minimum scale is likely smaller than many firms today, and hence there would be no harm in reducing firms size. This may not help much with stability, but there are still perhaps many benefits (e.g. reduced political and economic power) from reducing firm size and increasing the number of institutions engaged in important financial activities.
A response to Tyler Cowen's alternative to breaking up big banks (February, 2012):
Break Up the Banks? Here’s an Alternative: ... Some notes: First, there's an implicit assumption in this article about the minimum efficient scale for a bank. Tyler worries that breaking up banks will result in less efficient banking operations (i.e. higher cost) causing the smaller banks to fail altogether, or be less competitive with foreign banks.
However, I have not seen convincing evidence that banks need to be as large as they are for efficiency reasons (here's some evidence, but as I noted, I am not convinced by it). I am not advocating a per se rule here -- we shouldn't break them up just because. But if there's evidence that the size leads to undue political or economic power that is being exploited in the banks' favor, and if mega-mega-size is not necessary for efficiency, then there is definitely a reason to break them into smaller pieces. From my perspective, there is quite a bit of evidence that these banks have far too much political influence, and I think a case can also be made that it's unhealthy for the economy to have firms with such a large market share in particular segments of financial markets.
So I think that, absent of strong evidence that there actually are economic efficiencies associated with size (in which case they ought to be treated more like a regulated monopoly than a competitive marketplace), and the evidence that these banks are highly influential politically -- to the point where regulatory capture is more than a passing worry -- we should break these banks into pieces that are closer to the "minimum efficient scale" for financial institutions. ...
But we shouldn't fool ourselves into thinking that breaking up big banks into smaller pieces will necessarily make the financial system more stable. We had bank runs and financial meltdowns in eras where there were predominately small banks, think of the Great Depression for example. That's because it's the interconnectedness of banks that causes the problems, and smaller banks can be just as interconnected and hence just as vulnerable to a systemic shock as large banks. Perhaps there's a bit more diversification in larger banks that offers some protection, but the evidence is not strong on this point and big banks appear to be just as vulnerable as small banks to systemic troubles (and vice versa). ...
But whatever we do, we need to get on with it. Despite the Dodd-Frank financial reform bill and its directive to address this issue, the problem of bank runs in the shadow system -- a key factor in the financial sector collapse -- has not yet been solved. Work on this is underway, and new regulations are in the works, but for now the problem has not yet been resolved.
And a response to Bernanke's "smaller banks are not necessarily the answer" (October, 2009):
Bernanke: Smaller Banks Not Necessarily the Answer: ... Ben Bernanke does not want to lose "the economic benefit of multi-function, international (financial) firms," so he is hesitant to break large banks into smaller sized institutions. I don't have much problem with the economics, if there are efficiencies that come with bank size we should exploit them, especially if breaking up banks into smaller entities does little to reduce systemic risk but instead simply fragments the problem into many more pieces (though I'd still like to know where the minimum efficient scale is, anything larger than that is unnecessary). Obtaining resolution authority for banks in the shadow system is also very important, so I don't disagree with the emphasis on this in Bernanke's remarks.
But there seems to be the view that if they have resolution authority, higher capital requirements, etc., that will make the probability of a major breakdown small enough so that the expected benefits of size outweigh the expected costs. While I agree that obtaining resolution authority and other regulatory change is extremely important, I wouldn't bet my house, or housing and asset markets more generally, that this will eliminate the chance of a major breakdown, or make the chance small enough to justify huge, powerful, market-dominating institutions.
I would like to see more effort to measure and regulate connectedness within the system (which can be very high even with banks broken into smaller pieces) since that would add another layer of protection, the degree of leverage should come under scrutiny as well, and I would also like to see more attention to the political risks (e.g. capture of legislators and hence regulation) posed by large financial firms
Finally, A response to David Altig (this is the entire post from April, 2009):
Breaking Up Big Banks: As Usual, Benefits Come with a Side of Costs: David Altig warns that breaking up big banks could reduce efficiency:
Breaking up big banks: As usual, benefits come with a side of costs, by David Altig. macroblog: Probably the least controversial proposition among an otherwise very controversial set of propositions on which financial reform proposals are based is that institutions deemed "too big to fail" (TBTF) are a real problem. As Fed Chairman Bernanke declared not too long ago:
As the crisis has shown, one of the greatest threats to the diversity and efficiency of our financial system is the pernicious problem of financial institutions that are deemed "too big to fail."
The next question, of course, is how to deal with that threat. At this point the debate gets contentious. One popular suggestion for dealing with the TBTF problem is to just make sure that no bank is "too big." Two scholars leading that charge are Simon Johnson and
James Kwak (who are among other things the proprietors at The Baseline Scenario blog). They make their case in the New York Times' Economix feature:
Since last fall, many leading central bankers including Mervyn King, Paul Volcker, Richard Fisher and Thomas Hoenig have come out in favor of either breaking up large banks or constraining their activities in ways that reduce taxpayers' exposure to potential failures. Senators Bernard Sanders and Ted Kaufman have also called for cutting large banks down to a size where they no longer pose a systemic threat to the financial system and the economy.
…We think that increased capital requirements are an important and valuable step toward ensuring a safer financial system. We just don't think they are enough. Nor are they the central issue…
We think the better solution is the "dumber" one: avoid having banks that are too big (or too complex) to fail in the first place.
Paul Krugman has noted one big potential problem with this line of attack:
As I argued in my last column, while the problem of "too big to fail" has gotten most of the attention—and while big banks deserve all the opprobrium they're getting—the core problem with our financial system isn't the size of the largest financial institutions. It is, instead, the fact that the current system doesn't limit risky behavior by "shadow banks," institutions—like Lehman Brothers—that carry out banking functions, that are perfectly capable of creating a banking crisis, but, because they issue debt rather than taking deposits, face minimal oversight.
In addition to that observation—which is the basis of calls for a systemic regulator that spans the financial system, and not just specific classes of financial institutions—there is another, very basic, economic question: Why are banks big?
To that question, there seems to be an answer: We have big banks because there are efficiencies associated with getting bigger—economies of scale. David Wheelock and Paul Wilson, of the Federal Reserve Bank of St. Louis and Clemson University, respectively, sum up what they and other economists know about economies of scale in banking:
…our findings are consistent with other recent studies that find evidence of significant scale economies for large bank holding companies, as well as with the view that industry consolidation has been driven, at least in part, by scale economies. Further, our results have implications for policies intended to limit the size of banks to ensure competitive markets, to reduce the number of banks deemed "too-big-to-fail," or for other purposes. Although there may be benefits to imposing limits on the size of banks, our research points out potential costs of such intervention.
Writing at the National Review Online, the Cato Institute's Arnold Kling acknowledges the efficiency angle, and then dismisses it:
There's a long debate to be had about the maximum size to which a bank should be allowed to grow, and about how to go about breaking up banks that become too large. But I want to focus instead on the general objections to large banks.
The question can be examined from three perspectives. First, how much economic efficiency would be sacrificed by limiting the size of financial institutions? Second, how would such a policy affect systemic risk? Third, what would be the political economy of limiting banks' size?
It is the political economy that most concerns me…
If we had a free market in banking, very large banks would constitute evidence that there are commensurate economies of scale in the industry. But the reality is that our present large financial institutions probably owe their scale more to government policy than to economic advantages associated with their vast size.
I added the emphasis to the "probably" qualifier.
The Wheelock-Wilson evidence does not disprove the Kling assertion, as the estimates of scale economies are obtained using banks' cost structures, which certainly are impacted by the nature of government policy. But if economies of scale are in some way intrinsic to at least some aspects of banking—and not just political economy artifacts—the costs of placing restrictions on bank size could introduce risks that go beyond reducing the efficiency of the targeted financial institutions. If some banks are large for good economic reasons, the forces that move them to become big would likely emerge with force in the shadow banking system, exacerbating the very problem noted by Krugman.
I think it bears noting that the argument for something like constraining the size of particular banks implicitly assumes that it is not possible, for reasons that are either technical or political, to actually let failing large institutions fail. Maybe it is so, as Robert Reich asserts in a Huffington Post item today. And maybe it is in fact the case that big is not beautiful when it comes to financial institutions. But in evaluating the benefits of busting up the big guys, we shouldn't lose sight of the possibility that this is also a strategy that could carry very real costs.
I'd like to know the source of the scale economies. The paper linked above estimates returns to scale, but not their source. As noted in the introduction (and also noted by David):
Our results indicate that as recently as 2006 banks faced increasing returns to scale, suggesting that scale economies are a plausible (though not necessarily only) reason for the growth in bank size...
Without knowing the source of the changes in costs as banks increase in size, the (non-parametric) results -- results that differ from most previous work -- are hard to evaluate. I've been hoping for good estimates of the size and nature of the economies of scale for banks, but I'm not fully convinced by this evidence.
One final note. Even if the scale economies are there, we may still want to restrict the size of banks. The benefit from larger scale must be balanced against the increased risk to the financial system and the increased risk of political/regulator capture that comes with size and power. If the costs of size outweigh the benefits, then size should be restricted below the minimum efficient scale (How much risk comes with size depends, in large part, upon your faith in resolution authority discussed above. I am not fully convinced it will work.)