Did Risk-Based Capital Allocate Bank Credit and Cause a `Credit Crunch' in the U.S.?

This paper examines the reallocation of bank credit from loans to securities in the early 1990s using data on virtually all U.S. banks from 1979 to 1992. The spectacular increase in bank and thrift failures in the 1980s raised concerns about depository institution risk and spurred interest in public policy prescriptions to reduce this risk. One of these pre-scriptions was the Basle Accord on risk-based capital, which mandates that international banks operating in the major industrialized nations hold capital in proportion to their perceived credit risks. Because capital is more expensive to raise than insured deposits, risk-based capital (RBC) may be viewed as a regulatory tax that is higher on assets in categories that are assigned higher risk weights. Therefore, it would be expected that implementation of RBC would encourage substitution out of assets in the 100% risk category, such as commercial loans, and into assets in the 0% risk category, such as Treasury securities. Thus, the allocation of credit away from commercial loans may have caused a "credit crunch," which the authors define as a significant reduction in the supply of credit available to commercial borrowers. Consistent with these expectations, U.S. banks did reduce their commercial loans and increase their holdings of Treasuries in the early 1990s. A number of alternative explanations for this change in bank behavior have been offered. The authors suggest several other hypotheses including the leverage credit crunch hypothesis reflecting banks' interest in reducing their required leverage capital ratio; the loan examination credit crunch hypothesis reflecting the more rigorous examination process which encouraged substitution into safe assets; the voluntary risk retrenchment credit crunch hypothesis reflecting management's voluntary substitution of safer assets to lower the cost of funding and reduce the risk of bankruptcy; and the macro/regional demand-side hypo-thesis reflecting the reduction in overall loan demand because of the downturn in the economy and the steep slope of the term structure. An additional hypothesis is that the decline in commercial lending reflects continuation of longer term trends in the declining demand for bank intermediation services. Unfortunately, all of these hypotheses are roughly consistent with the aggregate data, leaving unknown whether risk-based capital played a major part in the reallocation of bank credit or whether a supply side "credit crunch" even existed. It is possible that all of the theories were correct simultaneously. The paper takes a close look at the data at the micro bank level to try to distinguish among the alternative hypotheses, with emphasis on RBC. The method used by the authors is to examine how bank portfolios changed in the early 1990s from the 1980s, and to see how these changes are related to the risk-based capital ratios and other key variables. The authors' tests relate the growth rates of bank asset categories to several measures of perceived bank risk, including the Tier 1 and Total RBC ratios. The findings suggest that the RBC credit crunch hypothesis fares the worst of all the alternative explanations of the bank credit reallocation of the 1990s. They find that the effects of the RBC ratios on lending did not get consistently stronger in the early 1990s, and that the Tier 1 and Total RBC ratios generally acted to counteract each other in their effect on credit allocation. The other credit crunch theories examined are somewhat more consistent with the data, given that the relations to the leverage ratio and the "problem" loan categories generally have the predicted signs. However, the quantitative effects are not substantial. The only other evidence that is roughly consistent with the credit crunch hypotheses is that large banks, banks with weaker capital ratios, and banks supervised by the OCC have much more substantial credit allocation effects and greater lending reactions to perceived risk than do other banks. While it is difficult to disentangle these groups, since large banks tend to have weaker capital ratios and national charters, in none of these groups are most of the decreases attributable to the credit crunch hypothesis directly. The authors note that the findings do not rule out non-risk related credit crunch expla-nations. The authors state that such theories cannot be easily identified econometrically because they are not associated with observ-able variables on which to base a test. They conclude that the demand-side effects on lending are relatively strong, but exact attribution to the different hypotheses cannot be determined from their model.