ABSTRACT __________________________________________________________________________ This paper reviews the role of micro non-convexities in the study of business cycles. One important nonconvexity arises because an individual can work only one workweek length in a given week. The implication of this non-convexity is that the aggregate intertemporal elasticity of labor supply is large and the principal margin of adjustment is in the number employed—not in the hours per person employed—as observed. The paper also reviews a business cycle model with an occasionally binding capacity constraint. This model better mimics business cycle fluctuations than the standard real business cycle model. Aggregation in the presence of micro non-convexities is key in the model.

c environment is preferable to a less stable one. But if shocks cannot be eliminated, it may be in the long-run interests of society to adapt to the shocks. To the extent that countercyclical policies interfere with the necessary adaptations, they may do more harm than good. 1 Not surprisingly, the shift of professional opinion toward the shock-based view of business cycles has been accompanied by increasing debate about the sources of cyclical volatility. Few macroeconomists doubt that random shocks underlie business cycles, but they have been unable to agree on which random shocks, historically, have been the main causes of cyclical volatility. To a person not versed in business-cycle theory (including economists who are not macroeconomists) , this situation must seem somewhat paradoxical: How can macroeconomists be certain that shocks cause cycles, yet not agree on which shocks are responsible for cyclical volatility ? Moreover, if a person is told

There is no spiritual copyright in scientific discoveries, unless they should happen to be quite mistaken. Only in making a blunder does a scientist do something which, conceivably, no one else might ever do again. —Peter Medawar, The Strange Case of the Spotted Mice It takes an economist to read an economist.

Introduction In Klamer (1984, p. 138) Robert Solow 1 says that "the intellectual problem of new classical economics is to determine how one can make the size of observed fluctuations compatible with the belief in perpetual Walrasian equilibrium." Lucas (1977) writes about the problem considered by Adelman and Adelman (1959). "The Adelmans posed, in a precise way, the question of whether an observer armed with [time-series] methods: : : could distinguish between a collection of economic series generated artificially by a computer programmed to follow the: : : [IS/LM] equations and the analogous series generated by an actual economy." Later he mentions: "To date, however, no equilibrium model has been developed which meets these standards and which, at the same time, could pass the test posed by the Adelmans (1959). My own guess would be that success in this sense is five, but not twenty-five years off." He was exactly right. An

This paper outlines a bootstrapping approach to the estimation and analysis of macroeconometric models. It integrates for dynamic, nonlinear, simultaneous equation models the bootstrapping approach to evaluating estimators initiated by Efron (1979) and the stochastic simulation approach to evaluating models' properties initiated by Adelman and Adelman (1959). It also estimates for a particular model the gain in coverage accuracy from using bootstrap confidence intervals over asymptotic confidence intervals.

This paper applies a new Bayesian framework laid out in Faust and Gupta (2009) for evaluating the suitability of dynamic stochastic general equilibrium (DSGE) models for the task of monetary policy analysis. The paper characterizes practical monetary policy analysis as determining how intended policy should be revised in light of the structural interpretation of incoming news. The news is defined as the one-step ahead forecast errors and the first and second moments of this news sufficiently summarize its structure for the purposes of monetary policy analysis. To shed light on the structural implications of this news, the paper estimates the implied structural shocks from fitting the DSGE model of Smets and Wouters (2007) to data. These fitted shocks then provide a structural interpretation for the observed news. The paper then evaluates the first and second moments of both the forecast errors and structural shocks, finding strengths of the model and important shortcomings. The paper finds that the posterior distribution for the realized value of the mean and cross-correlations of the structural shocks are typically nonzero. This result is surprising because structural shocks by construction are supposed to be zero-mean and uncorrelated. The results, therefore, identify specific frictions and structural aspects of the model that cause misspecification and suggest areas of improvement for model building.

This paper outlines a bootstrapping approach to the estimation and analysis of macroeconometric models. It integrates for dynamic, nonlinear, simultaneous equation models the bootstrapping approach to evaluating estimators initiated by Efron (1979) and the stochastic simulation approach to evaluating models ’ properties initiated by Adelman and Adelman (1959). It also estimates for a particular model the gain in coverage accuracy from using bootstrap confidence intervals over asymptotic confidence intervals. 1

The difference between answering a ‘why’-question and answering a ‘how much’-question

Econometric busines cycle research (EBCR) combines economic theory and measurement in the study of business cycles, i.e., upand downs in overall economic activity. EBCR has four goals: description, forecasting, policy evaluation and implementation/testing of economic theories. In this paper I assess four EBCR methods: business cycle indicators, simultaneous equations models, vector autoregressive systems and real business cycle models. My conclusion will be that simultaneous equations models are—despite its shortcomings—the best vehicle for EBCR is all its goals are to be attained simultaneously.

This paper deals with the size of the random walk property of Colombia´s output in two periods 1925-1994 and 1950-1994. GDP and GDPPC were both found to be integrated of order one a result which is very well known. The sequences are highly persistent, specially in the period 1950-1994. The forecast error when an innovation of 1 percent enters into the economy is about 1.5 percent in the very long run, when GDP is considered. The response is about 1.3 percent in the case of GDPPC, which seems to give support to the idea that population growth is a source of nonstationarity in some macroeconomic aggregates. For the larger sample (1925- 1994) persistence is less. This result could cast some doubt on the method of estimation of GDP for the period 1925-1950. Finally, evidence of nonlinearity is found only in Hodrick-Prescott filtered variables dated between 1925 and 1994. This leaves open the question about whether the HP filter introduces nonlinearity in the high frequency variable that it generates.