Monetary Policy and Credit Conditions: Evidence from the Composition of External Finance: Comment

In recent years, considerable research has they take as evidence for a bank lending explored the role played by bank lending in channel. the transmission of monetary shocks. In conWe reexamine KSW's analysis. In contrast trast to the traditional Keynesian transmission to KSW's use of aggregate data, we analyze mechanism that operates strictly through inthe mix of bank and nonbank debt separately terest rates, a bank lending channel allows for small and large firms. As is well known, central bank actions to affect the supply of loans financing patterns differ sharply across these from depository institutions ("banks") and, in two groups. Only the very largest corporations turn, the real spending of bank borrowers. issue significant amounts of commercial paEmpirical work on the existence of a bank per; conversely, small firms issue essentially lending channel (for example, Stephen King, no commercial paper, depending instead on 1986) generally has focused on the correlabanks as their primary source of finance. With tions among aggregate output, bank debt, and heterogeneous firms, a given movement in the indicators of monetary policy. This work, aggregate debt mix can reflect any number of however, is plagued by the problem of idendevelopments at the firm level. tifying shifts in loan demand from shifts in Moving to disaggregated data forces us to loan supply. Evidence that both output and modify KSW's measure of the debt mix. Bebank loans fall after a monetary tightening cause small firms issue so little commercial does not identify whether the decline in loan paper, KSW's measure-the ratio of bank volume reflects a constriction of loan supply debt to the sum of bank debt and commercial or a dampening of loan demand through the paper-is essentially pegged at unity for these traditional interest rate mechanism. firms. Thus, KSW's mix variable cannot posAnil K. Kashyap, Jeremy C. Stein, and sibly capture shifts in the relative importance David W. Wilcox (1993), henceforth KSW, of bank and nonbank finance for small firms. cut through this identification problem by exThis is a serious shortcoming because small amining relative movements in bank loans and firms are often presumed to bear the brunt of commercial paper after monetary shocks. Their a bank lending channel. Therefore, the mix intuition is straightforward: a monetary shock variable in this comment includes all forms of that operates through the usual interest rate short-term nonbank debt, not merely commerchannel lowers the demand for all types of ficial paper. This allows for meaningful substinance, while a monetary shock that operates tution between bank and nonbank debt for through a bank lending channel affects the small firms yet remains consistent with the supply of only bank debt. KSW find that bank spirit of KSW. The range of potential subloans outstanding decline relative to commerstitutes for bank debt is a crucial issue for cia1 paper after a monetary contraction, which KSW's analysis and for our own. In Oliner and Rudebusch (1995), we show that the results presented here remain valid when we broaden the mix variable even further by including * Oliner: Board of Governors of the Federal Reserve trade credit and long-term debt. System, Washington, DC 20551; Rudebusch: Economic Our results cast serious doubt on KSW's Research Department, Federal Reserve Bank of San Franstory about the transmission of monetary polcisco, San Francisco, CA 94105. We thank Tom Brennan icy. Using data for the U.S. manufacturing secfor excellent research assistance. The views expressed here are ours alone and do not necessarily represent those tor, we find little evidence that a monetary of the institutions with which we are affiliated. shock changes the mix of bank and nonbank