Use of Dynamic Factor Analysis in Macroeconomic Forecasts

This paper uses the dynamic factor analysis methodology developed by Stock and Watson (1998) in order to forecast inflation and the Imacec, an index of economic activity of common use for the Chilean economy. Our results indicate that using factors in the process of forecasting these macroeconomic variables improves significantly out of sample projections. Additionally, we find that factor augmented Phillips curve forecasts perform better than conventional Phillips curve forecasts based only on output gap measures.