Economic fluctuations.

• Author: Hall, Robert E.
• Position: Program Report

Economic Fluctuations

In the three years since my last Program Report, the NBER's Program in Economic Fluctuations has grown from 38 to 69 participants and has produced more than 200 Working papers. This article summarizes some of the major topics that we have addressed.

Research on Overall Macroeconomic Performance

One important area of research in the economic fluctuations program involves tracking down the sources of booms and recessions in the United States and other countries and understanding the role of financial and other factors in propagating these fluctuations. To that end, Ben S. Bernanke and Alan S. Blinder have studied the interaction between the real and financial sectors of the U.S. macroeconomy.(1) Bernanke and Mark Gertler have developed theoretical models to show how the net worth of borrowers can affect the stability and cyclical characteristics of the economy.(2) In related empirical work, Bernanke and John Y. Campbell have studied the phenomenon of increasing leverage of U.S. corporations.(3)

Motivated by the recent European experience, Olivier J. Blanchard has explored channels through which supply and demand shocks have long-lasting effects on the economy's equilibrium. In particular, with Danny Quah, he has studied the roles of bargaining in the labor market and imperfect competition in the goods market.(4)

Lawrence H. Summers has urged greater attention to relative wage theories in attempts to understand unemployment.(5) The importance of relative wages in different jobs, emphasized by Keynes, is related closely to recent efficiency wage theories. Empirically, Summers and Blanchard have written about the persistence of European unemployment, asking why unemployment remains so high near the so-called "full" employment level. In another area of research, Summers and Chris Carroll have compared private saving behavior in the United States and Canada.(6)

Bennett T. McCallum has produced a number of integrated overviews of several major areas of macroeconomic analysis.(7)

Christina D. Romer has focused on the causes of business cycles in the pre-World War II period and the changes in the nature of business cycles between the prewar and the postwar eras. In one study, she examined the role of the Great Crash of the stock market in October 1929 in initiating the Great Depression.(8) Another study found that beneficial supply shocks related to an agricultural boom may explain both the rapid deflation in 1921 and the fact that total production declined only slightly in that year, despite large declines in aggregate demand.

Mark W. Watson has worked on theoretical and empirical issues associated with the analysis of long-run relationships in aggregate economic time-series data. He has studied properties of univariate detrending methods, developed tests for cointegration (with James H. Stock), studied asymptotic properties of estimators in autoregressive models in the presence of unit roots (with Stock and Christopher Sims), and investigated the relationship between permanent and cyclical shocks to the economy (with Robert G. King, Charles I. Plosser, Stock, and Matthew D. Shapiro). In addition, Watson and Stock have served as codirectors of an NBER project to develop a new set of leading and coincident economic indicators.(9)

David Romer has been conducting research on whether--and if so, how--nominal disturbances have real effects. With Christina Romer he has investigated this question empirically by using nonstatistical information to isolate monetary disturbances.(10) With Laurence Ball, Romer has derived and tested an implication of endogenous models of price rigidity that differed from the predictions both of New Classical models and of the traditional Keynesian models that simply postulate the existence of price rigidity.(11)

John H. Cochrane has studied the random properties of aggregate output in the United States. His work shows that GNP probably has a tendency to return to a long-run growth path after a disturbance, but that the rate of return is slow. His research on consumption concludes that rejections of the permanent-income hypothesis and consumption-based asset pricing models can be caused by behavior that costs consumers a few cents per month.(12)

Campbell has studied the persistence of economic fluctuations. The traditional view is that fluctuations have little long-run effect because output returns rapidly to trend. In work with N. Gregory Mankiw, however, Campbell has argued that there is no empirical evidence for this return to trend. In fact, postwar U.S. data are consistent with the view that a 1 percent shock to the level of GNP should lead one to revise one's long-run forecast of GNP by more than 1 percent.(13) In work with Angus Deaton, Campbell has developed the implications of persistent fluctuations for the behavior of consumption.(14)

Steven N. Durlauf has worked on econometric issues raised by the research of Campbell-Mankiw, Cochrane, and others, which show that GNP and other macroeconomic variables have important random movements at frequencies well below the usual business cycle. Durlauf also has collaborated with me on the detection of specification errors or noise in models with expectations. The methods we have developed have applications in the stock market, inventory, investment, consumption, and other branches of macroeconomic research.

Thomas J. Sargent has been studying learning in dynamic economic environments. Specifically, how do various adaptive ways of forming views converge to rational expectations? Sargent's research considers two approaches to learning: recursive least squares and classifier systems from the artificial intelligence literature. His work on least squares is with Albert Marcet; his work on artificial intelligence systems is with Ramon Marimon. Sargent is also working with Rodolfo Manuelli on monetary theory in the context of models with missing markets.

Victor Zarnowitz has carried out a comparative study of major features of past and recent U.S. macroeconomic fluctuations. His model of lead-lag interactions among measures of real growth, inflation, interest rates, monetary and fiscal changes, and early stages of investment and production processes (selected leading indicators) ranks the effects of these variables similarly for the...