When Ben Bernanke was asked about the "too big to fail" problem not too long ago, the WSJ Economics blog reports:
Federal Reserve Chairman Ben Bernanke voiced skepticism that breaking-up big banks is the way to solve the so-called too big to fail problem...
Asked for his thoughts on Bank of England Gov. Mervyn King’s recent speech that advocated breaking up banks that were so large that their failure would represent a risk to the broader financial system, Bernanke said that making banks smaller would not necessarily be the solution to the problem. Smaller banks can also play important roles in financial systems, he said. He noted that during the 1930s, the U.S. didn’t have too many large bank failures, but the country suffered thousands of failures of smaller banks that added to the woes of the Great Depression. “I don’t think simply making banks smaller is the way to do it,” he said.
Still, more than once during his comments to the Economic Club of New York, Bernanke emphasized that it is crucial that large financial firms be allowed to fail in order to return market discipline to the financial system.
It is not at all clear to me that breaking large banks into smaller pieces addresses the connectedness issue. Smaller banks can be just as interconnected as larger banks, and hence simply breaking banks up without examining the effect it has on the underlying financial network connections may not reduce systemic risk.
Joseph Stiglitz says break them up whenever possible, regulate them heavily when it's not possible:
Too Big to Live, by Joseph E. Stiglitz, Commentary, Project Syndicate: A global controversy is raging... Mervyn King, the governor of the Bank of England, has called for restrictions on the kinds of activities in which mega-banks can engage. ... King is right to demand that banks that are too big to fail be reined in. In the United States, the United Kingdom, and elsewhere, large banks have been responsible for the bulk of the cost to taxpayers. ...
The crisis is a result of at least eight distinct but related failures:
... There are, of course, costs to regulations, but the costs of having an inadequate regulatory structure are enormous. We have not done nearly enough to prevent another crisis... King is right: banks that are too big to fail are too big to exist. If they continue to exist, they must exist in what is sometimes called a “utility” model, meaning that they are heavily regulated.
- Too-big-to-fail banks have perverse incentives; if they gamble and win, they walk off with the proceeds; if they fail, taxpayers pick up the tab.
- Financial institutions are too intertwined to fail...
- Even if individual banks are small, if they engage in correlated behavior – using the same models – their behavior can fuel systemic risk;
- Incentive structures within banks are designed to encourage short-sighted behavior and excessive risk taking.
- In assessing their own risk, banks do not look at the externalities that they (or their failure) would impose on others, which is one reason why we need regulation in the first place.
- · Banks have done a bad job in risk assessment – the models they were using were deeply flawed.
- · Investors, seemingly even less informed about the risk of excessive leverage than banks, put enormous pressure on banks to undertake excessive risk.
- · Regulators, who are supposed to understand all of this and prevent actions that spur systemic risk, failed. They, too, used flawed models and had flawed incentives; too many didn’t understand the role of regulation; and too many became “captured” by those they were supposed to be regulating.
In particular, allowing such banks to continue engaging in proprietary trading distorts financial markets. Why should they be allowed to gamble, with taxpayers underwriting their losses? What are the “synergies”? Can they possibly outweigh the costs? Some large banks are now involved in a sufficiently large share of trading ... that they have, in effect, gained the same unfair advantage that any inside trader has.
This may generate higher profits for them, but at the expense of others. It is a skewed playing field – and one increasingly skewed against smaller players. Who wouldn’t prefer a credit default swap underwritten by the US or UK government; no wonder that too-big-to-fail institutions dominate this market.
The one thing nowadays that economists agree upon is that incentives matter. ... Given the lack of understanding of risk by investors, and deficiencies in corporate governance, bankers had an incentive not to design good incentive structures. It is vital to correct such flaws – at the level of the organization and of the individual manager.
That means breaking up too-important-to fail (or too-complex-to-fix) institutions. Where this is not possible, it means stringently restricting what they can do and imposing higher taxes and capital-adequacy requirements, thereby helping level the playing field. ...
Even if we fix bank incentive structures perfectly ... the banks will still represent a big risk. The bigger the bank, and the more risk-taking in which big banks are allowed to engage, the greater the threat to our economies and our societies. ... What is required is a multi-prong approach, including special taxes, increased capital requirements, tighter supervision, and limits on size and risk-taking activities.Such an approach won’t prevent another crisis, but it would make one less likely – and less costly if it did occur.
I think limiting connectedness and limiting leverage ratios are both essential elements of reform. There will always be vulnerabilities, even in a system that has only small financial institutions, and we may not be able to identify the vulnerabilities in time. Shocks are going to happen. Limiting connectedness and leverage ratios for both big and small firms (along with regulation on what types of activities they can engage in, which addresses an aspect of connectedness) will reduce the magnitude of the damage to the financial system and the broader economy that those inevitable shocks are able to bring about.