Should we break up the large banks?

by on February 12, 2012 at 12:26 pm in Uncategorized | Permalink

In my column today, I say no.  Here is part of my argument:

…the logic of cutting down huge institutions could mean splitting the largest ones into several pieces. Yet banks do not always come in easily divisible parts. Such a move could amount to eradicating the largest banks rather than splitting them up — and eradication is both politically unlikely and potentially disastrous for the economy. In short, if the resulting parts of a divided bank cannot turn a profit, the split-up may prompt the very bailout it was trying to avoid.

Another fear is that American money market operations would move to larger foreign banks, which would have a newly found competitive advantage. If a financial problem arose, we would either bail out the foreign banks or rely on a foreign central bank to protect our own interests. Neither option seems appealing.

Even if a breakup went well, the incentives for the new, smaller banks would be unhealthy. Those banks could make mistakes or take on bad risks without being punished very much in terms of capitalization or revenue, because of their legally capped size. Even if they made big mistakes, these banks would probably be pushing on the frontier of maximum allowed growth. Eventually, the competitive process would cease to make these banks tougher or smarter or leaner, and we would just be cultivating another kind of banking system where bad or irresponsible decisions don’t lead to financial failure.

Most important bank failures spring from correlated risks, like the bursting of a real estate bubble, that affect many banks at roughly the same time. Bailing out a large number of smaller failing banks may be easier than bailing out a smaller number of large ones, since it is easier to apply bankruptcy and the procedures of the Federal Deposit Insurance Corporation to the smaller institutions. But that outcome hardly gets rid of bailouts.

There is still another problem. The more a bank is legally limited in terms of easily measurable size, the more it may resort to off-balance-sheet activities to make up the difference. “Breaking up big banks” may really mean making these less-transparent bank activities much more important to a bank’s fate.

The rest of the piece considers non-limited liability as an alternative for banking reform, and I thank Scott Sumner for drawing my attention to a recent piece by Eugene White (pdf) on this topic.   Stephen Williamson comments, see also his references.  I’ll respond to Arnold Kling’s remarks in a separate post, soon enough probably tomorrow.

Ted Craig February 12, 2012 at 12:44 pm
Willitts February 12, 2012 at 1:18 pm

Using food, cars, and houses in your opener as metaphors for financial institution regulation is never a good sign, especially from someone whose house, car, and diet is probably larger than it should be.

If shareholders find added value in divestiture, that’s their business.

Jamie_NYC February 12, 2012 at 12:45 pm

Tyler, you are right in your conclusion, but you do not mention two facts that render the whole issue of “too big to fail” irrelevant:
- S&L crisis: the biggest financial disaster up to 2008 was caused my myriad of tiny (some consisting of a single branch) financial institutions all following pretty much the same strategy, largely for regulatory reasons;
- Drexel Burnham: why was Drexel not too big to fail, butLehman was? Simply, because Drexel was a unique case in it’s time, while all other iBanks had portfolios similar to Lehman. The problem is not the size, it is the uniformity, a monoculture if you like, which is, as I mentioned, largely an unintended consequence of regulation.

Willitts February 12, 2012 at 1:22 pm

Great point. Regulation unwittingly creates identical incentives across institutions, leading to a coordination of systematic risk.

Ricardo February 13, 2012 at 1:58 am

What regulations made ibanks take on highly leveraged positions in subprime-related assets in the early 2000s? Investment banks spent the vast majority of their history not investing on their own account and not functioning as limited liability corporations. These two innovations — going public and investing shareholder’s capital in the financial markets — are relatively new but it is hard to see how they are the product of regulations.

GiT February 13, 2012 at 2:38 am

Why is regulation unique in producing a monoculture?

NeedleFactory February 12, 2012 at 12:59 pm

It is not the size of banks I fear, it is the “correlated risks” you mention. Philip Maymin puts it well in “Why Financial Regulation is Doomed to Fail”, to which you linked on August 10 2011.

Dale February 12, 2012 at 7:47 pm

I don’t find “correlated risks” very convincing as the cause of the problems. All financial institutions must deal with risk – some risks are more correlated than others. They failed to adequately gauge the correlation – either due to incompetence, poor incentives, or both. So, the source of the problem is not with the correlation of the risks. If bank failures “spring from correlated risks,” then this must be an argument for some kind of structural change to the sector. Perhaps the size of banks does have something to do with the failures (it is easier for a few large banks to have correlated risks than a great number of smaller ones).

mgj February 12, 2012 at 1:10 pm

Also address the gaming that will apply to the suggestion. E.g a shareholder will hold the bank shares in …..a coporation with limited liability.

Willitts February 12, 2012 at 1:11 pm

Before responding to a monumentally stupid proposal, it’s usually a good idea to recognize the chance it will come to fruition. Sure, bad ideas unchecked can become law over time, but when it comes to that point, all the good arguments in the world won’t stop it.

Tom West February 12, 2012 at 1:20 pm

Too much efficiency.

Ultimate efficiency has very little robustness, as everything is calibrated to absolutely maximal results under optimal conditions (and tends to breed a single “best-practices” business strategy). In a culture where efficiency and return is prized over everything else, any organizations pursuing robustness that is not constantly required (i.e. preparing for rare events) will have lower returns and are soon subject to liquidation by their more efficient peers.

Which culture is king among all others in pursuit of maximizing efficiency without being slowed by cultural traditions, fear of change, and general complacency?

The conditions that breed the highest standard of living in the world also guarantee highest vulnerability to correlated risk.

improbable February 12, 2012 at 2:02 pm

Nicely said. Also Jamie_NYC’s point about S&L, above.

(From someone watching this dogshow from India, which has been fairly isolated.)

steve February 12, 2012 at 1:25 pm

1) I think your suggestion still has timing issues. It will take months or years to determine what shareholders owe. (I suspect banks will still make use of off balance sheet accounting. Shareholders will sue claiming lack of knowledge about toxic assets.) Meanwhile, what happens to credit markets?

2) Why not just align incentives? Get rid of LLCs. Turn these banks into partnerships. Have full liability up to and including loss of personal assets. They can have the chance to make billions, or have the chance to lose it all. It only runs one direction now.


msgkings February 12, 2012 at 6:02 pm

#2: I’m a fan of this idea but I can’t see how to force that genie back in the bottle. Also, partnerships aren’t scalable the way corporations are, and the increasingly globalized world needs scale.

anon February 12, 2012 at 1:37 pm

Aren’t high capital requirements basically equivalent to increased shareholder liability (assuming no credit constraints)? Unlimited liability is quite hard to manage in practice, as other commenters point out.

Willitts February 12, 2012 at 2:39 pm

Of course. Capital consists mostly of tangible common equity. More capital means more capital at risk. This is fine if the capital requirements accurately reflect the inherent risk, but higher capital means lower return on equity.

With earnings under pressure from low interest rates and caps on fees, it might be difficult to raise capital. In absolute terms, banks make a lot of money, but relative to their inputs the returns are quite small, particularly when accounting for risk.

Bair is an advocate for higher capital – not controversial in the current regime. But its closing the barn door after the horses have run out unless we see another shock to the system in the near future.

We never should have had the financial crisis to begin with, and regulators are at least partly to blame. On the other hand, if regulators gained perfect foresight in, say, 2004, their demands for higher capital buffers would have faced stiff political opposition from both parties, banks, real estate interests, developers and their employees, and a lot of existing and prospective home owners.

Limited liability is hard to swallow.

Companies like AIG which took on the liability for supposed tail risk were also too big to fail. I’m not sure we will ever get past the government as the lender of last resort. That’s why it’s imperative to keep government out of the process of mandating, encouraging, guaranteeing or subsidizing private sector loans. No way this financial crisis would have happened without government help.

mw February 12, 2012 at 1:52 pm

Could you outline how and why under this scheme things would have been different during the last bubble? When they were raking in the cash during the run-up, why would shareholders have chosen a different board advocating less risk-rewarding, opaque short-termist compensation schemes?

Willitts February 12, 2012 at 2:45 pm

The ones who did, wouldn’t. Not all banks, though, drank the kool aid. I don’t know what goes on in board rooms that make one set of people risk takers and another set of people risk averse. Even a risk averse bank had gains to be made from sipping the kool aid. My guess is that the good banks had somebody who had the guts to say this looked like a previous financial crisis happening again. I would like to identify those banks and read their management discussion and analysis from several years ago.

mw February 12, 2012 at 2:02 pm

Also, could you explain why (or if?) you think the status quo for bondholders is ok, compared to for shareholders?

Zach February 12, 2012 at 2:51 pm

“Even if a breakup went well, the incentives for the new, smaller banks would be unhealthy. Those banks could make mistakes or take on bad risks without being punished very much in terms of capitalization or revenue, because of their legally capped size.”

Aren’t the consequences of a bad bet, relative to the size of the bank, essentially identical? And bets that pay off and push the bank value over the cap aren’t unrewarded; those gains are sold off and the bank distributes dividends, right? It’s not as if big banks don’t have plenty of incentive to bend or break the rules today. I don’t understand how there’s any change incentive that compares to, “If your bank fails, your investors will be mostly made whole to prevent an economic crisis.”

Some of your concern is essentially, “We’d probably screw this up.” I don’t follow the assumption that the government will both tell banks how to split up and do so in such a dumb way that some of the resulting companies fail. Another concern is that less-regulated, foreign banks will gain market share… in what industry is this not the case? How is this not the case in banking now with low-regulation tax havens?

The sorts of shocks and dumb decisions you’re worried about could be avoided by phasing in a risk limit for financial institutions over time, with risk being a function of liquidity and leverage. If we’re worried about foreign banks putting us out of business, use diplomatic leverage as the largest importer in the world.

sammler February 12, 2012 at 2:55 pm

But Steve Randy Waldman (Interfluidity) has already solved this problem!

Absolute caps on interest-on-reserves, paid per institution, provide a tunable and marginal incentive for the transfer of assets out of the very largest banks — or for their breakup, if their size does not truly generate synergies.

NAME REDACTED February 12, 2012 at 5:41 pm

So they break up into a bunch of separate banks that have contracts between them that cause them to behave (and have the same systematic risk) as if they were a single bank?
Sorry dude, but technocratic solutions do not work. They are too easy for anyone with half a brain to arbitrage.

Master of None February 13, 2012 at 2:43 pm

I think that any analysis of this issue should consider the sentiments voiced at Interfluidity: The moral hazard and pervasive sense of unfairness of bailouts/TBTF is rotting the work ethic of an entire generation of Americans.

Ken February 12, 2012 at 2:58 pm

I agree in general that breaking up the big banks would be difficult and result in consequences, both foreseen and unforseene, that would likely be harmful.

My preference would be to limit the ability to become to big to fail. let any single bank grow as large as its talent and treasure permit, but do not permit it to grow by acquiring other financial institutions. And, possibly, considering limiting its reach. Any bank might be permited to operate in any five contiguious states (or X number) plus abroad, but no further.

I live in a mid-sized midwestern city with two major national banks and four regionial banks dominating the market. in each case they entered by acquiring a locally owned institution or one that had within recent years had acquired a local institution. in no case did they ‘start fresh’. in every case they gutted the local personnel and reduced their community relations (financial, personnel, jobs, lending) to a substantial degree. They prosper. The city wilts. surely this can not be either good public policy or in the long term economic interest of the nation.

fractalist February 12, 2012 at 3:23 pm

Break em up…

Saturation Macroeconomics: Gobbledy-Gook or the Real Deal?

Time for a new mathematical model, a new paradigm, for macroeconomics?

Is there a patterned science representing the time dependent evolution of macroeconomics?

The last paragraph of the Economic Fractalist main page ….

The ideal growth fractal time sequence is X, 2.5X, 2X and 1.5-1.6X. The first two cycles include a saturation transitional point and decay process in the terminal portion of the cycles. A sudden nonlinear drop in the last 0.5x time period of the 2.5X is the hallmark of a second cycle and characterizes this most recognizable cycle. After the nonlinear gap drop, the third cycle begins. This means that the second cycle can last anywhere in length from 2x to 2.5x. The third cycle 2X is primarily a growth cycle with a lower saturation point and decay process followed by a higher saturation point. The last 1.5-1.6X cycle is primarily a decay cycle interrupted with a mid area growth period. Near ideal fractal cycles can be seen in the trading valuations of many commodities and individual stocks. Most of the cycles are caricatures of the ideal and conform to Gompertz mathematical type saturation and decay curves.

For the Wilshire, the US composite equity index March 09 to October 2011 was a 4 phased Lammert growth and decay fractal series..

x/2.5x/2x/1.5x :: 5/13/10/7 months. That’s an empirical real system observation – available to all – of the time dependent workings of the macroeconomic system.

2005 was the description, the hypothesis – March 2009 to October 2011 was the empirical asset valuation evolution…

The flash crash on 6 May 2010 ….. does that not meet second fractal criteria?

“A sudden nonlinear drop in the last 0.5x time period of the 2.5X is the hallmark of a second cycle and characterizes this most recognizable cycle.”

Maybe this is all occurring by chance alone …. Likely…. Very very very likely ….not.

NAME REDACTED February 12, 2012 at 4:12 pm

The incentive for the bank is to split off all the bad assets in one and the good assets in the other. Then to allow the bad bank’s value to go to zero and the good bank’s value to quickly double.

Vagabundus February 12, 2012 at 4:39 pm

If a bank is “too big to fail” and causes a systemic risk, then there are uncompensated externalities, mainly free riding on bailout insurance. If there was a tax on banks with capital exceeding say 250 billion sufficient to cover these costs, then banks could choose to avoid this by breaking themselves up into stand alone divisions. This would be more efficient than nationalization, a court ordered, Fed or FDIC break up.
But what the hell do I know.

NAME REDACTED February 12, 2012 at 5:50 pm

Then the banks create a set of contracts between them that make them behave as one single larger bank.
Regulatory arbitrage FTW.

Vagabundus February 12, 2012 at 6:36 pm

But failure of A would not lead to contagion of B. The assets of B would not cover the liabilities of the A during audit. A would fail earlier, and with less systemic risk to others.

Three Pipe Problem February 12, 2012 at 6:46 pm

I never thought I’d say it, but I do believe that we should consider intervening because of the systemic risk posed by these big banks. However, breaking up big banks seems like a bad idea for reasons Tyler covered. Having an arbitrary limit (e.g. your 250 billion) also seems like a bad idea.

I prefer the concept of a tax tied to size of the institution. The tax should be negligible for small institutions and very large for very large institutions. I’m thinking a power function of some kind…

Vagabundus February 13, 2012 at 1:28 am

Power function is a great idea. 250 billion is a WAG. Small banks should not have to pay anything since they don’t pose systemic risk. I would also suggest adding a haircut to deposit insurance.

Bob February 12, 2012 at 5:46 pm

Several issues I do not think you have fully considered: What you suggest is in substance a new class of bank capital. Unfortunately by its nature it is very “iffy.” How would one enforce the claim against shareholders — some of whom will be hard to locate, have limited assets, be shell entities or located in foreign countries? The enforcement of complex mortgage backed security arrangements (or the attempts to get modifications) suggest the complexities that would tax the legal system. If one required the obligation to be insured, one would be adding a tax on the system which would itself likely be an administrative mess. Also, the problem with the super large banks is not only the consequences to the broader economy if they fail. They are arguably too big to manage, too complex to regulate and very, very difficult to account for properly. You point to the Canadian system. But there are so many structural and regulatory differences that it provides at best very limited guidance. And, the chance that the major US banks would accept a similar system is likely not more likely than highly remote.

The Other Jim February 12, 2012 at 6:04 pm

Maybe we could stop forcing banks to loan money to unqualified people who live in the districts of powerful Congressmen.

Maybe that would be a good idea.

Bill February 12, 2012 at 6:32 pm

There is a different way of deconcentrating risk of large banking other than breaking up large banks.

Ask yourself: How did large banks grow larger during the 70s through the late nineties: They did it through acquisitions.

A simple policy of discouraging, or not subsidizing, acquisitions by or among the top 8 banks would lead to a decline of the larger firms.

There is support for this based on history: First, the Federal government has encouraged and subsidized large banks to acquire troubled banks and financial institutions. A policy of not subsidizing large banks from acquiring troubled banks, but instead having middle and small banks acquire these banks or assets would go a long way of growing the small and medium banks. Second, there is a history of this type of merger program: during the 60s and 70s the FTC had industry specific merger guidelines which blocked the largest firms in some industries from acquiring mid size firms, and encouraging mid size firms to grow, or small size firms to grow.

Finally, we should not ignore that banks may need to diversify risk. Today, we diversify by having larger banks. In the past, banks diversified by having underwriting pools with other banks. We could create institutional mechanisms that encourage joint venturing among small, large and medium size banks to underwrite larger risks that currently are underwritten only by large banks. Thinking about joint ventures and pools could be a way to diversify risk, effectively let banking systems handle larger risks, with growing and concentrating risks in few institutions.

Ted Craig February 12, 2012 at 7:38 pm

What about the banks that are already there?

Bill February 12, 2012 at 8:05 pm

Competition will make them relatively smaller if we stop subsidizing their acquisitions and look for other parties to pick up dying banks. Those mid range banks that would be acquiring dying banks will grow, increasing competition with larger banks in those segments where larger banks do have an advantage. Remember that we were going to subsidize Citi’s acquisition of a West Coast bank before another bank stepped in and paid more for it.

The PolyCapitalist February 12, 2012 at 6:42 pm

As YS and others have put it, ‘why do we keep indulging the fiction that banks are private enterprises?’

anonymous... February 12, 2012 at 6:44 pm

Why is breaking up big banks always assumed to be some kind of magic cure?

Spain and Italy are too big to fail too, shall we break them up into about a dozen pieces each?

Ted Craig February 12, 2012 at 7:39 pm

There are factions in both countries that support that idea.

NAME REDACTED February 12, 2012 at 7:47 pm

The Basques would be very happy to hear that!

anonymous... February 12, 2012 at 10:30 pm

Yes, but that’s a matter of politics, language and culture, not economics.

The point is simply this: perhaps we should cure “too big to fail” by addressing the “fail” part rather than the “big” part.

Turning “big” to “small” won’t magically solve anything by itself. The savings and loan crisis of the 1980s involved a bunch of institutions that were individually small enough to fail, but collectively a very big problem.

dead serious February 13, 2012 at 9:11 am

I’m pretty sure that the Basques would consider themselves better off economically if independent.

Ted Craig February 13, 2012 at 9:06 am

Actually, the bigger issue is the Catalonians and the Piedmontese.

dead serious February 12, 2012 at 10:42 pm

The argument that breaking up a bank is so difficult is not very compelling. It is *much* harder for a bank to subsume another entity or enter a new line of business and yet this doesn’t curtail this kind of activity.

Comments on this entry are closed.

Previous post:

Next post: