Cyclicality of Credit Supply: Firm Level Evidence
Executive Summary — Bank lending falls in economic recessions. In particular, it shrank considerably during the recent economic downturn. Does such cyclicality of bank lending reflect a decline in banks' willingness to lend (the "loan supply" effect) or reduced demand for loans from firms (the "loan demand" effect)? The considerable attention that is given to banks' financial health by the Federal Reserve, Congress, and other branches of government is only warranted if the answer is supply. Focusing on U.S. firms that raised new debt financing between 1990 and 2009, HBS professors Bo Becker and Victoria Ivashina demonstrate that many large U.S. firms turn to the bond market when banks are in poor financial health. When times are better, the same firms get bank loans. Becker and Ivashina argue that the substitution between bonds and loans at the firm-level is a good economic proxy for the bank credit supply. Key concepts include:
- The incidence of bank loans, as compared to bonds, is very cyclical. Firms getting a bank loan are likely to stay with that form of debt in the near future. The pattern is similar in most years. But when banks are doing poorly, this pattern is reversed, and many large firms who would typically turn to banks for debt financing, instead issue bonds.
- The loan-and-bond mix for large firms is a strong predictor of a likelihood of firms without access to bond markets to raise bank debt.
We study the effect of bank loan supply through the business cycle using firm level data from 1990 to 2009. The contribution of our paper is to address two of the main empirical challenges in identifying the effects of bank credit supply. First, we focus on firms' choice between two close forms of external financing: bank debt and public bonds. By conditioning the sample of firms raising new debt, we can rule out a demand explanation for the drop in bank borrowing. Second, by doing the analysis at the firm level, we can directly address how the composition of firms raising finance varies through time. We find strong evidence of substitution from bank loans to bonds at times characterized by tight lending standards, high levels of non-performing loans to bank equity, low bank share prices and tight monetary policy. To illustrate our point, in the last half of 2007, 36% of all debt issues were bank loans. However, relative loan issuance fell to 8% by the first half of 2009, the lowest level in the period from 1990 to 2009. Although the bank-to-bond substitution can only be measured for larger firms (which have access to bond markets), we confirm that this substitution has strong predictive power for lending volume by small and unrated firms. 37 pages