Real-time forecasting US GDP from small-scale factor models

This paper proposes two refinements to the single-index dynamic factor model developed by Aruoba and Diebold (AD, 2010) to monitor US economic activity in real time. First, we adapt the model to include survey data and financial indicators. Second, we examine the predictive performance of the model when the goal is to forecast GDP growth. We find that our model is unequivocally the preferred alternative to compute backcasts. In nowcasting and forecasting, our model is able to forecast growth as well as AD and much better than several baseline alternatives. In addition, we find that our model could be used to predict more accurately the US business cycles.