Regulate risk-taking rather than bank size

Oct 21, 2013

James Barth, Ross Levine, Apanard (Penny) Prabha

Publisher: MarketWatch.com

The government tells us that never again will any bank be too big to fail.

The basis for this claim is that big banks now hold more capital to cover potential losses, undergo increased regulatory scrutiny to detect and correct emerging problems before they become serious, and face stringent restrictions on activities deemed too risky. And if these preemptive actions do not work, regulators can now liquidate a deeply troubled big bank in an orderly manner without using government funds.

Some prominent bank regulators and banking experts remain skeptical that the new regulatory approach will put an end to too-big-to-fail and government bailouts.They recommend a more drastic solution: break up the biggest banks.

While sharing the concern that reform efforts may fall short, we believe that too little attention has been devoted to a cost-benefit analysis of breaking up the biggest banks and the details of such an unprecedented action. Meanwhile, too much attention has been focused on blaming big banks for the 2008 financial crisis.

When is a bank too big? Of the 28 "globally systemically important" banks identified by the Financial Stability Board-- which promotes the implementation of effective regulatory, supervisory and other financial sector policies--eight are U.S. banks with combined assets equal to 90% of GDP as of June 2013. Other countries have fewer such banks but with higher asset-to-GDP ratios. For example, Switzerland has two, the U.K. and France have four each, but with combined assets-to-GDP ratios of 333%, 310 %, and 297%, respectively. Germany has only one such bank with a ratio of 69%, which is not that far below the total combined assets of all eight U.S. banks relative to GDP. The Netherlands, Sweden, Spain and Japan also have such banks whose assets-to-GDP ratios are larger than the comparable ratio for the U.S. banks.

The biggest U.S. banks therefore are relatively small players in their own country, and would become even smaller if broken up when compared to rivals elsewhere in the world. Is there any convincing evidence that the ratios for big banks are already too high when operating in a globally competitive environment?

Drawing a bright line in identifying big banks that are or will become a systemic risk is also extremely difficult. Consider the following: Three big banks will be in violation of the Safe, Accountable, Fair & Efficient Banking Act of 2012--limiting a bank's non-deposit liabilities to 2% of GDP--should it become law. Furthermore, a proposal from MIT's Simon Johnson limiting the asset size of banks to less than 2% of GDP would put seven banks over the line.

At the same time, two banks that are identified as globally systemically important banks would be allowed to grow bigger since their asset size is substantially less than this ratio. One might also note that big banks already are prohibited under current law from growing bigger through acquisitions when the result would lead to the new bank's total deposits exceeding 10% of all U.S. bank deposits. The arbitrariness and inconsistencies of these cutoffs is obvious.

One must also confront the seemingly simple question of "What is a bank?" before breaking up big banks. This becomes more complex when one begins to consider how big banks should be broken up. Those eight globally systemically important U.S. banks, which are holding companies, have total assets of $10.2 trillion and 23 bank subsidiaries with assets of $7 trillion. The remainder of their assets consists of different mixtures of securities, insurance policies and other products. Most of the big banks also have assets in other countries. Citigroup C -0.18%, for instance, has the largest share of foreign revenue at 57%.

Indeed, most of the world's big banks have a mix of businesses with different kinds of assets and liabilities, which make cross-border comparisons difficult. The question remains: how to break up big U.S. banks such that they are not put at a competitive disadvantage to the big, universal banks in other countries that don't impose such limits?

Also, to the extent that the demarcation line is adjusted for individual banks of different degrees of bigness, will the end result once again lead back to a too-big-to-fail problem? Moreover, what do the break-up proponents do about big companies that own banks, such as General Electric, Toyota and Target? And what do they propose to do with the big money-market mutual funds that function like banks, financial firms such as Fannie Mae and Freddie Mac, and big insurance firms like AIG that were bailed out during the financial crisis?

Trades and trade-offs

We also note that the latest crisis was primarily triggered by regulatory policies that encouraged excessive risk-taking by financial institutions of a variety of types and sizes. While many big banks promoted and exploited these policies for substantial short-run profits, they were not the only institutions to engage in such behavior. And, in the end, it was the elected representatives and the regulatory authorities who passed the laws and the implementing regulations in ways that created a dangerously fragile financial system which eventually imploded.

Admittedly, bigness might be bad. Apart from the too-big-to-fail issue, big banks possess considerable lobbying power that may be used to influence regulators. One way to illustrate the potential dangers that big banks pose is to answer the question, why do banks become big? There is insufficient evidence that big banks provide better services than smaller banks or lower the overall costs of banking. From this perspective, banks might become big for the purposes of becoming too-big-to-fail, too-big-to-discipline, and big enough to shape the regulatory rules in their favor.

Yet, despite these legitimate concerns, there is far too little evidence on the costs and benefits of breaking up big banks. After all, the government bailed out thousands of small savings and loans in the early 1980s through various accounting gimmicks when there was no too-big-to-fail problem. Instead, it was a too-many-to-fail problem.

Thus, in the absence of compelling evidence that the size of a bank per se is a problem rather than the riskiness of a bank relative to its owner-contributed equity capital, we recommend both more research and improvements in the regulatory environment to limit the systemic risks of banks. (For example, we favor a tangible equity leverage ratio of at least 9%.) In addition, given the poor past performance of the regulatory authorities in reducing the likelihood and severity of crises, it is also prudent to establish better procedures to hold regulators more accountable for achieving stability