Economic fluctuations.

Hall, Robert E.

The U.S. business cycle continues to be the main concern of the NBER's Program on Economic Fluctuations. We have examined the sources of the business cycle, propagation mechanisms, and policy responses. The program is organized around a number of specialized research groups, each of which studies some aspect of the business cycle with a combination of analytical and empirical methods. Their results are presented at research meetings three times a year.

The program also maintains the semi-official chronology of the U.S. business cycle. This function is the responsibility of the Bureau's Business Cycle Dating Committee.

The 1991 Business Cycle Trough

On December 22, 1992, the Business Cycle Dating Committee announced its determination that the U.S. economy reached a trough of activity in March 1991. Previously, the committee had determined that the economy reached a peak of activity in July 1990. The eight-month period between July 1990 and March 1991 thus entered the NBER's chronology as a recession. The recession ended in March 1991 and an expansion began; the expansion is still underway at this writing.

The committee waited nearly two years to identify the trough. Although this period was unusually long and the committee was criticized for the delay, the sluggish performance of the economy during 1991 and 1992 compelled the long wait. The committee followed its normal policy of waiting to determine the trough date until there was no doubt that any future downturn in the economy would be considered a new recession and not a continuation of the recession that began in July 1990. An important fact to the committee was that the broadest measure of economic activity--gross domestic product in constant dollars--did not surpass its previous peak, according to data available at the time, until the third quarter of 1992. Not until December did the overall pattern of economic activity appear to be strong enough to warrant the determination of the trough date.

The behavior of the economy in 1991 made deciding on a trough date particularly challenging. Two important monthly indicators related to the production and sales of goods--industrial production, and manufacturing-trade sales in constant dollars--had unambiguous troughs in early 1991 (in March and January, respectively). Two other monthly indicators had declined to close to their minimum values by early 1991, but continued to decline slightly for the rest of that year. Real personal income reached its trough in November 1991, at a minuscule 0.07 percent below its level in April. Total hours of all nonagricultural employees reached its trough in April. The choice of March 1991 as the trough date was based primarily on the fact that various averages of the monthly indicators reached clear troughs in that month.

The Business Cycle Dating Committee meets as needed to keep the chronology up to date. The next meeting will occur when it appears appropriate to determine the date of the next peak in activity, or if new data become available that suggest that an existing date should be adjusted.

Overview of Modern Research in Economic Fluctuations

The unifying theme of research in the program is movements over time in national output, employment, productivity, interest rates, and inflation. Interestingly, none of the research tackles all of these issues jointly. Members of the program are not engaged in a collective effort to build a complete empirical model of the U.S. or world economies. Instead, research generally isolates particular parts of the economy, such as the behavior of the consumer, or considers issues of the overall performance of the economy in terms of relatively simple and admittedly stripped-down models of the determination of output and employment.

Most of the research within the program looks at data on the U.S. economy within a rather fully specified analytical framework. The research is oriented toward scientific findings about technology, preferences, and other fundamentals, rather than toward descriptions and correlations. Most of the papers presented to the program's research groups and at its research meetings are quite technical, even when the basic factual results ultimately can be explained in a nontechnical way. A number of program members use methods from the frontier of econometrics, and are seen by the profession as econometricians as much as macroeconomists.

Where research considers issues of fluctuations by looking at particular sectors of the economy, the actual research closely resembles work by economists in other branches of applied microeconomics. Many of the papers produced within the program are of great interest to those in labor economics, industrial organization, and other fields of microeconomics. And many of the economists who attend group meetings or program meetings come from microeconomic specialties. Although the long-run goal of many researchers of providing a microeconomic theory of macroeconomic fluctuations has proven elusive, the assimilation of microeconomic thinking into macroeconomics, as it is practiced in the fluctuations program, has been complete.

Macroeconomic Complementarities

For a number of years, the EF program has had an active research group on macroeconomic complementarities under the direction of Russell W. Cooper. Complementarities--that is, favorable nonmarket interactions between firms, or between workers and firms--could play an important role in explaining the vulnerability of the national economy to relatively unimportant shocks.

In an economy without complementarities, markets have a strong role as shock absorbers. For example, a downward shift in consumption, such as apparently occurred around 1990 in the United States, should have relatively little effect on output and employment. Instead of triggering a recession, the shock should bring responses in credit markets that offset it. If consumers spend less, interest rates should fall, raising the demand for investment goods and exports.

By contrast, with complementarities there can be a magnification of small shocks. If the productivity of one firm depends on the level of activity of related firms, a negative shock can move the economy to a substantially lower level of output at its new equilibrium.

Another potential result of complementarities is multiple equilibriums. An economy can have a high-level equilibrium, in which firms and workers mutually reinforce each other's productivity. The same economy can have a low-level equilibrium, in which no single firm can expand profitably because of the low level of activity of other firms.

Currently, work in the complementarities group falls into three broad areas. Attacks on the tough theoretical issues raised by complementarities are being made by Douglas Gale, Alberto Trejos, Randall Wright, Andrew John, B. Ravikumar, Costas Azariadis, and Bruce Smith. The second type of work mixes theoretical and quantitative analysis. The common goal is to assess the quantitative plausibility of models with complementarities. Group members taking this approach are Satyajit Chatterjee, Russell W. Cooper, Paul Beaudry, Michael Devereux, and Jordi Gali. The last approach is strongly empirical. Suzanne Cooper and Steven N. Durlauf have developed a method for detecting the presence of the types of behavior associated with complementarities in aggregate data. Peter Klenow is studying complementarities that take the specific form of one firm learning from the experience of another firm making the same product. Klenow is using detailed data for computer chips.

Consumption Inequality and Its Aggregate Implications

A second important area of research in the program since its inception 15 years ago is the behavior of the consumer. Under the direction of Marjorie A. Flavin and Knut Mork, and more recently Orazio Attanasio, most of the research applies sophisticated models of household behavior to survey data on their purchasing decisions. Households gather and process information about their future well-being in order to make current decisions. Families do not just consume a certain fraction of their current income; they tend to put windfall income into the bank and to consume it slowly over future years. Our research also has rejected the opposite view--that families follow exactly the optimal information-processing strategy to distinguish temporary from permanent changes in income. There is some evidence that families let consumption track income too closely to be explained by models of optimal behavior. But a newly developed body of thought holds that refinement of the models can explain these departures.

Research by Attanasio and Steven J. Davis shows how the relative wage gains experienced by older and better-educated households during the 1980s were reflected in relative consumption gains. Angus Deaton and Christina H. Paxson study the effects of demographic changes in the United States, the United Kingdom, and Taiwan on the cross-sectional inequality in consumption. Wouter den Haan extends the method of parameterized expectations to solve and simulate several versions of an asset pricing model. Giuseppe Bertola is studying the effect on aggregate saving and growth of factor distribution. James Banks, Richard Blundell, and Agar Brugiavini are measuring consumption and income variance, and quantifying their effects on consumption growth. R. Glenn Hubbard, Jonathan S. Skinner, and Stephen P. Zeldes are using numerical simulations to assess the importance of the precautionary motive for saving. Several realistic features, such as liquidity constraints, failing health, and "means-tested" Social Security provisions are part of the model.

Micro and Macro Perspectives on the Aggregate Labor Market

A third area of research in recent years has been the functioning of labor markets. Under the direction of Wright and Richard Rogerson, the EF program has had a research group on labor supply and related issues; Davis led a separate group on employment dynamics. Recently these groups combined.

The elasticity of labor supply is of central importance in understanding overall economic fluctuations. When the economy booms, many people are working, and the typical worker is putting in unusually long hours. In a slump, both the number of people at work and their hours fall. Can these movements be interpreted as occurring along a labor supply curve?

The traditional macroeconomic answer was no. A recession was seen as a time of involuntary unemployment, a breakdown of the labor market. Modern thinking, which emphasizes market interpretations of fluctuations, has tried to provide a deeper analysis of the issue. One view is that the labor supply may be much more elastic in the short run than in the long run. Another looks at the time allocation of workers between working and looking for work. A further area of research in the macro-labor field is the sources of unemployment, job destruction; and the sources of reemployment, job creation. Very detailed micro data from individual plants provide important new insights into these processes.

The macro-labor group has been meeting for three years. The group's focus is on the behavior and performance of the aggregate labor market, including employment and productivity fluctuations, human capital accumulation and growth, and unemployment. Dale Mortenson and Christopher Pissarides have studied matching models of the labor market to analyze the relationship between job creation and destruction and cyclical fluctuations and growth. They build on the empirical work of Davis and John C. Haltiwanger on firm-level employment dynamics, which also has been represented in the group. Victor Rios-Rull shows how labor market regulations can emerge endogenously and create long wave growth cycles. Rogerson and Raquel Fernandez study the interaction between income distribution and education, and point to inefficiencies in a system of local provision of public education.

Empirical Methods in Macroeconomics

Durlauf and Francis Diebold recently organized a new group on empirical methods in macroeconomics, focusing on the interface between time-series econometrics and empirical macroeconomics. Their first meeting was in Palo Alto on February 6. Lucrezia Reichlin is working on the problem of invertibility in multivariate time-series models, and developing new diagnostic procedures for assessing the likelihood of neglected moving average components in VAR analyses. Robert F. Engle is working on modeling "common features" in economic and financial time series. Simon Potter is working with nonlinear impulse response functions and exploring their application in a number of macroeconomic contexts. Durlauf and Louis Maccini are using the Durlauf-Hall signal extraction approach to decompose inventory fluctuations into a model-consistent part and a "noise" part, thereby providing a natural metric for evaluating the fit of inventory models. Kenneth D. West and David W. Wilcox are exploring the properties of advanced estimation methods using Monte Carlo methods. Halbert White is continuing his work on neural networks, with particular focus on their strengths and weaknesses as approximators. Bent Sorenson is examining whether productivity is cointegrated with other determinants of long-run growth.

Impulses and Propagation Mechanisms

Since 1989, a group under the direction of Lawrence J. Christiano and Martin Eichenbaum has studied impulses and propagation mechanisms in the business cycle. Its winter 1992 meeting, organized jointly with the University of Arizona, focused on theoretical and empirical aspects of the interaction of monetary policy and business cycles. Four of the seven papers concerned international aspects of the problem. Alan C. Stockman presented a theoretical model of the impact of price rigidities on the trade-off between exchange rate targeting and monetary control. Don Schlagenhauf and Jeff Wrase analyzed the effects of monetary policy shocks on nominal and real exchange rates. Eichenbaum and Charles L. Evans showed empirical evidence of the effects of U.S. monetary policy shocks on exchange rates. Nobuhiro Kiyotaki discussed joint work with John Moore on the role of balance sheet constraints in propagating business cycle shocks. Christiano, Eichenbaum, and Evans gave a progress report on their ongoing work studying the impact of monetary policy shocks on the Flow of Funds data. Mark Gertler and Simon Gilchrist presented their work on the response of large and small firms to monetary policy shocks. Vittorio U. Grilli and Nouriel Roubini described international evidence on the exchange rate effects of monetary policy shocks.

At the NBER Summer Institute, the group discussed three papers on the interaction of monetary policy and the business cycle. Gertler and Gilchrist described the differential response to monetary policy shocks in the inventory investment behavior of small and large manufacturing firms. Stephen D. Oliner and Glenn Rudebusch looked at other differences in the response of large and small firms to monetary policy shocks. Christiano and Evans presented a paper with Eichenbaum on the effects of a monetary policy shock on the borrowing and lending activities of different sectors of the economy.

Two papers analyzed international estimates of productivity growth. Evans showed that measures of monetary policy shocks lead to a prolonged rise in standard measures of productivity. Patricia Reynolds presented improved estimates of productivity growth for different countries, and contrasted them with the existing literature. Two papers focused on fiscal policy in equilibrium business cycle models. Lee Ohanian presented a progress report on his work with Thomas Cooley that uses an equilibrium model to assess the effects of the policies pursued by the United Kingdom and the United States during World War II on their postwar economic performance. Ellen McGrattan discussed her ongoing study with Phillip A. Braun on the impact of fiscal policy on U.S. economic performance during and after World War II. David Levine analyzed the implications of income risk for the equity premium puzzle. Gali presented his research on models of spatial location; his theoretical discussion focused on the role of externalities and increasing returns in the propagation of sectoral shocks.

High-Frequency Economic Fluctuations

Until his recent move to the staff of the Council of Economic Advisers, Matthew D. Shapiro was the director of a research group on fluctuations that occur relatively rapidly over time. The group also is concerned with variations in economic activity over space and their relationship to similar variations over time. Participants include John Shea, Mark Bils, and James Kahn, who are working on spatial--temporal issues; Benjamin Eden, whose research in the group deals with inventory fluctuations; Valerie A. Ramey, who is continuing her research on the intertemporal elasticity of substitution; Ana Aizcorbe, who is studying variations over time in the productivity of auto assembly plants; and Susanto Basu and Argia Sbordone, whose research deals with labor hoarding and cyclical capital utilization.