Using the Flow of High Frequency Information for Short Term Forecasting of Economic Activity in Argentina

We exploit the richness of a large data set of daily and monthly business cycle indicators by pooling them to produce Nowcast of contemporaneous real GDP growth. We conduct predictions based on a pooling of bivariate forecasts which uses these indicators as predictors of GDP (Nowcast with pooling). We also conduct a Nowcast exercise with factors for a restricted subset of business cycle indicators. When comparing the predictive accuracy of Nowcast with pooling and with factors with that of an AR(1) model, only the Nowcast with pooling outperforms the AR(1), indicating that the use of information released within the quarter helps to improve GDP growth prediction. The methodology then offers an encouraging and valuable approach to provide timely information for policy decision making.