A dynamic index model for large cross sections

This paper shows how standard methods can be used to formulate and estimate a dynamic index model for random fields—stochastic processes indexed by time and cross section where the time-series and cross-section dimensions are comparable in magnitude. We use these to study dynamic comovements of sectoral employment in the U.S. economy. The dynamics of employment in sixty sectors is well explained using only two unobservable factors; those factors are also strongly correlated with GNP growth.