Evolving macroeconomic dynamics and structural change: applications and policy implications

Abstract

The present thesis is a collection of three separate essays, each corresponding to a chapter.Each essay represents an application of structural dynamic analysis within TVC-VARmodels. In the first chapter, coauthored with Fabio Canova, we investigate the relationshipbetween changes in output and inflation and monetary policy in the US. There are variationsin the structural coefficients and in the variance of the structural shocks but only the latterare synchronized across equations. The policy rules in the 1970s and 1990s are similar asis the transmission of policy disturbances. Changes in inflation persistence are only partlyexplained by monetary policy changes. Variations in the systematic component of policyhave limited effects on the dynamics of the system. Results are robust to alterations in theauxiliary assumptions.In the second chapter, coauthored with Fabio Canova and Evi Pappa, we examine thedynamics of US output and inflation. We show that there are changes in the volatility ofboth variables and in the persistence of inflation. Technology shocks explain changes inoutput volatility, while a combination of technology, demand and monetary shocks explainvariations in the persistence and volatility of inflation. We detect changes over time in thetransmission of technology shocks and in the variance of technology and of monetary policyshocks.In the third chapter we study whether hours worked rise or fall after a positive technologyshock. According to the existing evidence it depends on whether they enter the VAR in levels(hours rise) or growth rates (hours fall). We argue that conflicting results may ultimatelyarise because important structural time variations in the US economy are a priori ruledout by empirical models. We identify technology shocks as the only shocks driving long-runlabor productivity using postwar US quarterly data. We find that, under both specificationsfor hours (levels and growth rates), (i) hours fall, and (ii) technology shocks explain about11-23% of total aggregate fluctuations giving rise to positive but small correlations betweenoutput and hours. Differences with respect to fixed coefficients VAR are due to instabilitiesin the relationship between labor productivity and levels of hours.To conclude, the main contribution of the present research work as a whole is twofold.First, from an empirical perspective, we provide new evidence on some important macroeconomicissues of the US economy. In all the applications new interesting results, absentin standard VAR, emerge. Second, from a theoretical perspective, we contribute to developnew tools useful for policy analysis within the class of TVC-VAR models.