02/16/2012 03:25 pm ET Updated Apr 17, 2012

Tyler Cowen Tackles the Big-bank Problem

Over the weekend, Tyler Cowen put up a column in The New York Times arguing a) that financial institutions should not be broken up and b) that shareholders should be liable for $1.50 of loss on every dollar of stock they buy. Presumably, since Cowen is sketchy on details, that extra half a buck would come in the form of collateral, as Stephen Williamson in a post about the idea suggests. (He likes the plan.)

How that would work escapes me, given that shareholders don't often deal with collateral like the fixed-income crowd. Arnold Kling in another post responding to Cowen's column (he doesn't like it) scratches his head at the practical implications: "This sounds good in principle, but I wonder how it would work in practice. Suppose that you sell your shares in Shakee Bank to me today, and tomorrow Shakee has to be taken over by the FDIC. Am I liable for losses, which probably were caused by decisions made before I bought my shares? Suppose that Shakee has accumulated $8 billion in losses, and all its shareholders of record obtained their shares for a collective $0.10. What happens then?" Kling clings to his idea that if you're going to try to shape good behavior, you should just throw bank executives in jail when they go bust.

In fact, under Cowen's plan one would expect that shareholders to stampede at the first sign of any trouble (as opposed to hanging around and trying to fix anything, which is a hassle), dumping their shares at a price that would attract a buyer aware of the $1.50 overhang. One would anticipate that this would create a permanently low bank share price, particularly given the opacity of most big-bank holdings.

Cowen's proposal, and its various responses, has attracted a feisty little mob of commenters, most of them (though not all) negative to the idea. Nearly everyone seems to focus on some aspect of the incentivization issue: Whether increasing the liability on shareholders resolves the agency cost problem that nearly everyone seems to believe lurks beneath bank failures. As several, including Williamson, note, Scottish and Canadian banks in the early 19th century had unlimited (even double!) liability banks -- shareholders were responsible for all the losses -- and a record of stability in relatively deregulated eras. The dream of getting rid of regulators, who are accused of everything from feckless behavior to creating a monoculture, continues to run strongly; everyone dreams of a structural answer, a mechanical solution to resolve everything from busts to bailouts.

But we do not live in the 19th century. Even limited liability, like Cowen's $1.50 (or more), would send some bank shareholders fleeing for the exits. If bankers then retreated down the risk curve, profits would undoubtedly fall, making banks a less desirable investment. From a regulatory perspective, this might be just fine. But what if the effect is a reduction in both liquidity and lending? Isn't Cowen's idea then simply another scheme to shrink the banks? The question no one seems to approach is what will be the macroeconomic effect of a smaller, less liquidity-driven big-bank sector. (Cowen also doesn't say whether his idea would apply to all banks, even already beleaguered smaller banks, or just systemically important financial institutions.) We are living in a financially driven post-industrial economy, not a hard-money, agriculture and industrial 19th century. You can't go back, even wishfully. What's next? Debtors' prisons?

Cowen spends the first part of his column arguing why breaking up the banks is such a bad idea. This part of the argument is also sketchy. He doesn't wrestle with any of the extant schemes for bank reform, from a return to Glass-Steagall to the British idea of ringfencing certain consumer functions to the notion of creating small consumer banks with 100 percent capital backing or installing hard caps on size (deposits, assets, capital). All of these might end up in the same hardscrabble neighborhood he takes us to: They would constrain credit and liquidity as a tradeoff for resolving too-big-to-fail. Cowen believes the breakup option would produce a major mess: "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." He also fears foreign banks would take over money market operations and pose bailout risks, and that "the new, smaller banks would be unhealthy." Cowen seems to believe that since size would be capped, they would gun for growth, creating greater risks.

One can accept the underlying thrust of Cowen's concerns: A major experiment in financial engineering might spawn a myriad of unintended consequences, not to say difficult politics. You can accept that, however, without embracing Cowen's scheme. Banks are not that indivisible. Breaking them up does not guarantee "eradication." You can agree that a return to Glass-Steagall is knee-jerk or ahistorical (like a return to unlimited liability) without accepting the notion that big banks are woven so tightly together that they can't be split up. Hell, they buy and sell assets regularly. Cowen is correct that many of these new banks post-breakup would face a share price and profitability problem. Many of banking's "new" businesses -- trading, M&A advisory, brokerage -- were added because of a sense that traditional commercial banking was not profitable enough for shareholders. Breaking up a big bank unwisely will create a variety of businesses, some with lousy profits and the incentive to take on risk; others with large, if volatile, profits and a steady state of high risk. The former might resemble the S&Ls; the latter could look an awful lot Drexel Burnham Lambert, Bear Stearns Cos. and Lehman Brothers.

And again, Cowen's idea would take us to that very same place if it proved to be effective. Pumping up liability for shareholders would -- in theory -- force bankers to reduce profits. Shares would be less expensive, certainly compared with other nonbank financial services organizations that would then possess more potent currency to grow and diversify. You will have created a bank ghetto.

Perhaps there is a simple structural remedy out there that will satisfy all of our desires for profits, liquidity and soaring shares while allaying all of our fears. But I doubt it. Regulators -- flawed, human, imperfect (just like the rest of us) -- remain an important component of the protections built into the system. It shouldn't be the only component. Perhaps we should toss a few bankers in jail, or at least toughen the rules -- though separating out business judgment from fraud will never be easy in banking. Perhaps we should think harder about liability, though without the nostalgia for eras long gone that few sane folks would want to return to. Perhaps we should ban certain dangerous businesses, or hive them off -- aware that old-fashioned banking needs profits from somewhere to survive. We wish for simple answers, but there are none as long as we want all the good things finance showers upon us, and none of the bad. -- Robert Teitelman