Business survey data: Do they help in forecasting GDP growth?

In this paper we examine whether data from business tendency surveys are useful for forecasting GDP growth in the short run. The starting point is a so-called dynamic factor model (DFM), which is used both as a framework for dimension reduction in forecasting and as a procedure for filtering out unimportant idiosyncratic noise in the underlying survey data. In this way, it is possible to model a rather large number of noisy survey variables in a parsimoniously parameterised vector autoregression (VAR). To assess the forecasting performance of the procedure, comparisons are made with VARs that either use the survey variables directly, use macro variables only, or use other popular summary indices of economic activity. Our DFM-based procedure turns out to outperform the competing alternatives in most cases.