The Liquidity Trap and U.S. Interest Rates in the 1930s
Hanes, Christopher, Journal of Money, Credit & Banking
RECENT EXPERIENCE OF low inflation has revived interest in the liquidity trap, "a situation in which conventional monetary policies have become impotent, because nominal interest rates are at or near zero: injecting monetary base into the economy has no effect" (Krugman 1998, p. 141). Many discussions of the liquidity trap argue (or assume) that it constrains a central bank's ability to control interest rates at any maturity as soon as interest rates have been pushed to zero at the shortest maturity--in modern financial markets, the overnight maturity. Monetary policy may still be able to influence real activity through exchange rates (Orphanides and Wieland, 2000, Svensson, 2000, McCallum, 2001) or effects of money balances on demand for assets in general, including durable goods (Brunner and Meltzer, 1968a, Meltzer, 2001). But the "interest rate channel" of monetary policy--special influence over yields on liquid assets, and hence on required returns for less liquid financial assets such as bank loans--is blocked. In the common view, a central bank cannot affect term premiums unless it can "twist" the yield curve by changing the maturity structure of outstanding government debt, which is doubtful (Johnson, Small, and Tryon, 1999, King, 1999, Eggertsson and Woodford, 2003, pp. 20-23). Thus, a central bank influences longer-term rates only through expectations of future overnight rates. Once overnight rates have been driven to their floor, a central bank has no reliable mechanism to validate, or contradict, these expectations. Perhaps the central bank can make announcements or adopt rules to boost the public's confidence that overnight rates will remain zero further into the future. But if expectations do not respond as desired, the central bank cannot push down longer-term rates merely by increasing reserve supply (Fuhrer and Madigan, 1997, Krugman, 1998, 2000, McCallum, 2001, Meyer, 2001, Woodford, 1999).
On this definition, the United States was in a liquidity trap through most of the 1930s. According to Krugman (1998), U.S. interest rates were "hard up against the zero constraint" (p. 137). From 1934 through the outbreak of the Second World War in August 1939, overnight rates were effectively zero. Treasury bill rates were usually less than a quarter of a percent. (1) At the same time, however, long-term rates remained well above zero: Treasury bond yields never fell below 2%. Meanwhile, monetary authorities followed policies that resulted in large shocks to the supply of high-powered money. Some of these money supply shocks were associated with news about monetary policy, but others were the result of transient factors that had no obvious implications for future money supply or overnight rates.
In this paper, I present evidence from the 1930s U.S. on the relation between reserve supply and longer-term interest rates when overnight rates are zero. To begin, I argue that recent discussions of liquidity traps overlook an idea from early literature on money demand such as Tobin (1958): the potential role of cash as an asset free of interest-rate risk, which implies that high-powered money demand is inversely related to the level of longer-term rates when short-term rates are zero. I present a model that applies this idea to banks' demand for reserves. The model matches conventional propositions about monetary policy when overnight rates are positive, but it also implies that, when the overnight rate is zero, an increase in reserve supply reduces long-term rates holding fixed expectations of future overnight rates. Finally, I examine the relation between reserve quantities and bond yields over 1934-39. I find that changes in bond yields were related to reserve quantities in a manner consistent with the model.
1. RESERVE DEMAND AND THE LIQUIDITY TRAP: ISSUES
Most literature on reserve demand and short-term interest-rate determination in the United States (for example, Poole, 1968, Strongin, 1995, Hamilton, 1996, 1997, Borio, 1997, Bernanke and Mihov, 1998, Woodford, 2000, Bartolini, Bertola, and Prati, 2002) is framed in terms of a common model that reflects institutions in place since the early 1920s. …