THE RELATIVE IMPORTANCE OF AGGREGATE AND SECTOR-SPECIFIC SHOCKS AT EXPLAINING AGGREGATE AND SECTORAL FLUCTUATIONS

In this note, we present a simple exercise aimed at disentangling the sectoral and aggregate effects of both aggregate and sector-specific shocks. We construct a structural multi-sectoral model of the economy and show that, after several hypotheses, estimation of such structural model by Vector Auto-Regressions (VAR) is possible. The most important assumption that we use is that sector-specific shocks have no contemporaneous effects on sectors other than the sector where they originate, although we do not exclude that they have some effects on the former sectors after some lag. We estimate the model using monthly U.S. data (the industrial production index) so that the latter assumption is not too strong. Our results suggest that the aggregate effects of sector-specific shocks are not large but, on the other hand, their sectoral effects dominate the sectoral effects of aggregate shocks.