Economies around the world are marked by major interventions in credit markets. Institutions ranging from central banks to the Grameen Bank operate under the assumptions that credit markets are imperfect, that these imperfections can be ameliorated, and that doing so increases output. There is surprisingly little empirical support for these propositions. This paper develops evidence on related questions by exploiting changes to a major intervention in U.S. credit markets, the Community Reinvestment Act (CRA). Using data on both banks and potential commercial borrowers, I find evidence that CRA does increase credit to small businesses as intended. I then exploit these CRA-induced supply shocks to identify the impact of credit increases on county-level payroll and bankruptcies. There is some evidence of real benefits at plausible implied rates of return on CRA borrowing, and little suggestion of crowd-out or adverse effects on bank performance. The findings therefore appear consistent with a model where targeted credit market interventions can improve efficiency. Ongoing work seeks to identify whether CRA does in fact ameliorate any particular type of credit market failure.

Economies around the world are characterized by major interventions in credit markets. Institutions ranging from central banks to the Grameen Bank operate under the assumptions that credit markets are imperfect, that these imperfections can be ameliorated, and that doing so increases output. There is surprisingly little empirical support for these propositions. The existence of important credit market failures is uncertain. A substantial body of work on investment-cash flow sensitivity concludes that many firms are liquidity constrained (Hubbard, 1998; Fazzari, et. al. 2000). Yet whether the observed liquidity sensitivity actually implies financing (e.g., credit) constraints has been questioned on both theoretical and empirical grounds (Kaplan and Zingales, 1997 and 2000).More direct tests of theoretical models of credit constraints (e.g., Stiglitz and Weiss 1981, Hart and Moore 1994) are rare, and they have produced little evidence of empirically important imperfections (e.g., Berger and Udell, 1992). Finding “real” (as opposed to merely “financial”) effects of finance might offer indirect evidence of underlying market failures and motivate interventions. But there is little to suggest that increasing credit (as many interventions seek to do) would increase output in steady-state; on the contrary, the finance literature suggests that banks may be the second-best solution to credit market frictions. A growing body of evidence does suggest that aggregate output increases with the quality of financial intermediation (Jayaratne and Strahan 1996; Rajan and Zingales 1998), but little is known about the effects of changes to credit supply; e.g., the existence of a bank lending channel for monetary policy remains relatively controversial.