In Discussion Paper No. 1244, the objective of Marie Forni and Research Fellow Lucrezia Reichlin is to learn about both the sources and the manifestation of the business cycle. They develop a method for analysing large cross-sections with non-trivial time dimensions and apply it to output and productivity for 450 US industries from 1958-86. They use a dynamic factor analytic model to decompose output and productivity into common and idiosyncratic dynamics. The common component describes both the (possibly heterogeneous) direct effect of shocks common to all sectors and the asymptotic indirect effects produced by inter-sectoral autoregressive linkages (either input-output or demand). Their framework is an extension of the dynamic factor model of Sargent and Sims (1977) for large cross-sections.

The authors introduce the following methodological innovations. First, they develop a method based on dynamic principal components to identify the number of common shocks to the 450 sectoral output and productivity growth rates. Second, they estimate the unobserved common component through sectoral averages. Third, they use a new method for identifying the technology shock based on the innocuous idea that technological innovations are mainly positive. Lastly, they propose and construct a measure of the degree of substitution between complementary effects of shocks. Their empirical results suggest that: first, there are only two common shocks driving output and productivity in the 450 sectors; second, technological innovations produce positive co-movements at business cycle frequencies, but they are not sufficient to explain all cyclical fluctuations; third, technological shocks are strongly correlated with the growth rates of the investment in machinery and equipment sectors and their inputs.

Let's Get Real: A Dynamic Factor Analytical Approach to Disaggregated Business Cycle
Marie Forni and Lucrezia Reichlin

Discussion Paper No. 1244, September 1995 (IM)