Dominant Sectors in the US: A Factor Model Analysis of Sectoral Industrial Production
This paper considers the role of sector-specific production growth shocks as major sources of cross-sector comovement in the business cycle. The dominant sectors are identified according to the number of factors detected in the residuals after treating the production growth in some candidate sectors as observed factors. We build on the properties of the principal component method and analytically show this approach can consistently identify the dominant units (sectors) in a large panel data set even if the assumption of strongly influential factors is moderately relaxed. We provide empirical evidence that growth in a few industrial sectors in the US provide suitable approximations for an unknown common factor. Using data on the intersectoral material input-outputs and the cross-sector capital flows, we find that the dominant sectors have an important role as suppliers of capital products to other sectors.