CASEY B. MULLIGAN University of Chicago XAVIER SALA-I-MARTIN Yale University U.S. Money Demand: Surprising Cross-sectional Estimates THE SPECIFICATION of the money demand functionhas important impli- cations for a number of macroeconomic issues. First, if policymakers are to be responsible for achieving price stability, they need reliable quantitative estimatesof money demand. ' In particular, if the money de- mandfunctionis stable, the income elasticity yields the rate of money growththatis consistent with long-run price stability. Second, macroeconomic theoristsneed quantitative estimates of the money demand functionin orderto determine the exact predictions of their models. In Keynesian models, for instance, the relative ability of monetaryand fiscal policy to affect the real economy depends on the elasticities of the demandfor money. For a given interest elasticity, a largerincome elasticity implies a more vertical LM curve; as a result, monetary policy is relativelymorepotentthanfiscalpolicy. In fact, part of the debate between monetarists and fiscalists in the 1950sand 1960s was over the "slope"of the LM curve. Such issues are especially im- portant to manyeconomists of the 1990s,who are called upon to assess We havebenefited fromdiscussions withRobert Barro, PaulCashin, Lawrence Chris- tiano,Cagun Dena,Jordi Gali,Marvin Goodfriend, Robert Lucas, Gregory Mankiw, Julio Rotemberg, EtsuroShioji,Christopher Sims, Nancy Stokey, members of the September 1992 Brookings Panel,andespeciallyAndrew Atkeson.This material is baseduponwork supported by the National Science Foundation. 1. That economic research promotes "prosperity andpricestability" has alwaysbeen a primary goal of the Brookings Papers on Economic Activity. Readers of BPEA find a formal statement of this goal on the first pageof every volume. 285