How Should Banking Be Regulated?

With the economy appearing to begin to stabilize, more attention is being given to possible measures for preventing a recurrence of an economic crash as calamitous as this one has been and continues to be. Advocacy of specific measures seems to me premature, however. This is not only because proposals for overhauling the regulation of banking adds additional uncertainty to the economic environment. It is also because the reform of banking regulation is an immensely complex undertaking that should be attempted only after careful deliberation. There is no urgency about strengthening bank regulation at present or in the immediate future. Memories are short, but there isn't going to be another housing or credit bubble in the immediate future even if no regulatory changes are made.

Congress seems about to pass a law that would create an investigatory commission, modeled on the 9/11 commission, to explore the causes of the current depressioin. The 9/11 commission issued, in the summer of 2004, an exhaustive narrative and analysis of the events leading up to the terrorist attacks of September 11, 2001, and of the failure of our intelligence agencies to discover the terrorists' plot in advance of its execution. Almost as an afterthought the commission made a number of ill-conceived recommendations for reform of the intelligence system, many of which were adopted with, I have argued in several books, meager results. One hopes the commission to investigate the causes of the current depression will devote more attention to formulating intelligent, constructive recommendations for reform than the 9/11 commission did. (It would be nice if we could occasionally learn from our mistakes.) But the completion of the commission's report will doubtless take a long time, so one can hope that the rush to reform banking regulation will be halted until the report is published and its recommendations digested.

Many reforms are being discussed, and in this entry I will offer my reactions to just a few of them. The first is to have two tiers of banking regulation, one for the vast majority of banks, the other for the handful (about 20) banks that among them account for about two-thirds of the revenues of the U.S. banking industry. The latter group, often described as banks "too big to fail" (more precisely, too big to be allowed to fail, that is, to go broke), would, in the suggested reform that I am discussing, be carefully regulated by a "systemic regulator" charged with responsibility for preventing the banks from taking risks that, if the risks materialized, could precipitate a general economic crash and require massive federal bailouts, as in the current crisis. A radical alternative that is unlikely to receive serious consideration is to break up the banks or shear away the parts of their business that do not fit the traditional understanding of commercial banking.

The concern with the big banks is understandable, but the idea of subjecting them to a "systemic regulator" doesn't make a lot of sense. To begin with, size per se is not a good indication of the macroeconomic significance of a financial institution. Lehman Brothers, which the government allowed to collapse last September, was not very large. But it occupied a strategic position in several important financial markets, such as commercial paper, letters of credit, and credit default swaps, with the result that its collapse, against the background of the severe weakening of the banking industry as a whole by the collapse of the housing and mortgage markets, caused a substantial disruption of the global finance industry.

Furthermore, regulations triggered by the regulated firms' size or (what is more difficult to ascertain and evaluate) strategic significance will induce many institutions to avoid becoming large enough or strategic enough to be subject to the regulations, and the result may a less efficient banking industry, if scale and position in financial markets confer substantial benefits.

Then too, the idea of the systemic regulator doesn't match up with the concern that underlies the expression "firms too big to be allowed to fail." The concern is valid: a firm that has reason to think that it will be bailed out by the government if it gets into deep financial trouble will be able to borrow at lower interest rates than other firms, and this will give it a competitive advantage unrelated to superior efficiency. The managers and shareholders of the banks that have had to be bailed out by the government have taken a shellacking, but not the bondholders; and that is the basis for the fear that banks too big to be allowed to fail can borrow at lower rates--their bondholders don't have to worry that the banks will go broke and default on the bonds.

But even if the 20 biggest banks were split five ways, so that 100 banks instead of 20 banks accounted for the bulk of bank lending, the "too big to fail" problem would not be solved. It is not any individual bank that is too big to be allowed to fail, but the banking industry as a whole. Any single bank, even the largest, could have been allowed to fail without catastrophic consequences. The problem was that the bursting of the housing bubble impaired the solvency of a very large fraction of the banking industry (weighted by size of bank). Had that very large fraction consisted of 100 banks rather than 20 banks, the need for bailouts would have been no less urgent and the resulting windfalls for the creditors of those banks no smaller.

We also need to worry about the identity and incentives of a systemic regulator. The regulation of the American banking industry, broadly defined as it must be to include all financial intermediaries (that is, lenders of borrowed capital) rather than just commercial banks, because of the substitutability of the financial products sold by the different types of financial intermediary, is divided among a large number of federal and state agencies. Is the systemic regulator to be layered over them? Presumably so, and that would be a recipe for bureaucratic turf wars, delay, and loss of information and control.

Moreover, any agency that has a single mission is likely to perform it out without sufficient attention to the costs. An agency charged with preventing the collapse of the banking industry is likely to place an exaggerated emphasis on safety, for it will be blamed if there is a collapse, but it will not be praised for the economic benefits of a less safe, but more dynamic and profitable, banking industry. There is an analogy to the security officers who conduct background checks on applicants for jobs in the CIA and other intelligence agencies. They are too conservative in their recommendations because from a career standpoint it is better to err on the side of advising against hiring an applicant who might turn out to be a security risk; if he does, they will be blamed, but if instead he turns out to do a terrific job, his supervisors will get the credit.

A further objection to the idea of a systemic regulator is that because banking is thoroughly international, tighter regulation of U.S. banking, by reducing the profitability of U.S. banks, will tend to divert banking business to foreign banks. I would not give too much weight to that objection, however, because it is quite possible that the U.S. banking industry grew too large during recent years--to the point where its private value exceeded its social value. An efficient financial system creates wealth, but some unknown but sizable fraction of the profits that it generates are merely transfer payments from losers to winners in trades. Such transfers may merely rearrange rather than augment the wealth of society, for example by sucking talented people into the finance industry who might have a larger social product in some other field; the example I give in the book is Harvard Ph.Ds in physics who go to work for financial firms, which highly value physicists' mathematical skills.

An alternative to creating a systemic regulator is simply to reimpose some of the old rules that made banking, which as I have stressed is inherently risky, less risky than an unregulated banking industry would be. They include such regulations as forbidding payment of interest on demand deposits (which gave banks a very cheap source of capital, enabling them to make money by lending that capital at the low interest rates that safe borrowers pay) and limiting branch banking--that is, limiting competition among banks and thus making it easier for them to make money without taking many risks (risk and return being positively correlated).

There are two objections, apart from the reduction in the availability of credit that more conservative lending practices would bring about. The first is the difficulty of deciding which financial intermediaries should be subject to such regulations. Should hedge funds be forbidden to pay interest to their investors? To open branch offices? Those would be ridiculous regulations. Hedge funds would attract no lenders at zero interest, and they have no branch offices. But if left unregulated, which is essentially their condition today, they would be in a position to eat the lunch of the commercial banks if the latter were subjected to tight regulation and they were not. Yet any regulation of hedge funds would have to be tailor made to the unique character of such enterprises, rather than being a carbon copy of the regulation of commercial banks; and the same goes for the regulation of any other nonbank financial intermediary. Designing a mosaic of regulations that would create a "level playing field" for all financial intermediaries would probably be impossible.

Second, trying to limit risk taking by financial intermediaries is like trying to compress a balloon by squeezing it without bursting it. Suppose a bank were forbidden to pay interest on demand deposits. The amount of those deposits would plummet as depositors sought alternative investments. If banks were forbidden to have branches, competing financial intermediaries would gain a leg up because banks would no longer be as convenient for customers as they had once been. Moreover, if risk taking is profit maximizing, banks will find ways of imparting risk to their operations. Suppose banks are somehow constrained not to make risky loans. They can make loans risky nevertheless just by increasing the ratio of debt to equity in their capital structures.

To illustrate this important point, let us suppose that a bank has a debt-equity ratio of 30 to 1, and that it makes a $100 loan, so that the bank's equity stake is only $3 and the rest is borrowed, and the interest rate on the loan is 5 percent. The loan will yield the bank $5 in annual interest (if there is no default), and that will generate a large percentage return on equity even after subtracting the interest rate on the borrowed capital used to make the loan. Suppose the bank borrows at 3 percent. Then its spread is 2 percent (5 percent, the interest rate on the loan it makes, minus 3 percent), which yields a net gain of $2 ($5 minus $3) on its $100 loan--2/3 of its equity stake ($2 net gain divided by $3 equity stake). That is a very high rate of return. Of course, there is a lot of downside risk (think of the loss to the bank's shareholders if the value of the loan declines by 10 percent). The regulators could put a ceiling on the bank's debt-equity ratio to limit the downside risk. But how would they determine the ratio? And would they impose a ceiling on the debt-equity ratio of all potential lenders?

There is much more to be said about the difficulties of re-regulating banks. But to keep this entry to a manageable length, I shall discuss just one more proposal, and that is to eliminate the various subsidies to home ownership, such as the deductibility of mortgage and home-equity interest from income tax. That is not going to happen, but what is more within the limits of the politically feasible (if only barely) is to eliminate Fanny Mae and Freddie Mac, the huge mortgage companies that financed much of the housing bubble, and to repeal laws such as the Community Reinvestment Act of 1987 (along with measures by the Clinton and Bush administrations) that encouraged risky mortgage lending, for example to members of minority groups who might not be able afford to make a normal down payment on a house.

Fannie Mae and Freddie Mac are what are called "Government-Sponsored Enterprises," but until they collapsed last summer and were placed under federal control they were private corporations. Their pertinacious and well-informed critic, Peter Wallison of the American Enterprise Institute, contends that these GSEs, along with the laws that encouraged them (and indeed encouraged the fully private mortgage sector as well) to make risky mortgage loans, were single-handedly responsible for the housing bubble.

I have no truck with the GSEs and see no great social interest in promoting a nation of homeowners rather than a nation of renters (one advantage of a nation of renters is that relocation to a different city or state to pursue new job opportunities is easier when one rents rather than owns one's home, and job mobility is one of the great strengths of the American economy). But I think Wallison exaggerates the effect of the GSEs and the "ownership society" propaganda on the risky mortgage lending that has gotten the banking industry in so much trouble.

He contends that the GSEs and the government's encouragement of risky mortgage lending caused loose lending practices to spread to the prime loan market, vastly increasing the availability of credit for mortgages and thereby leading to speculation in houses and ultimately to the housing bubble.There are three objections to this diagnosis.

The first is that to the extent that the GSEs provided a market for risky mortgages, the loss when the default rate on those mortgages skyrocketed fell on the GSEs rather than on the banking industry as a whole. The second objection is that much of the risky lending occurred completely outside the chain of distribution that included the GSEs. Banks that originated mortgage-backed securities that were sold to other banks, hedge funds, foreign investors, and other investors had no connection to the GSEs; they made their own decisions about making, buying, and packaging mortgages. Third, I can't think of any economic process by which the GSEs' providing a market for substandard mortgage loans would have induced other financial intermediaries to do likewise. If they had thought that the mortgage loans were excessively risky, they would have ceded this part of the market entirely to the GSEs.