Nowcasting economic activity in Argentina with many predictors

Abstract

We pool a large data set of business cycle indicators to produce Nowcast of contemporaneous GDP growth. We also conduct Nowcast using factors for a restricted subset of the indicators. Using an AR(1) benchmark to compare the forecasting performance of both Nowcasts, we conclude that only the Nowcast with pooling outperforms this univariate model. The Giacomini and White (2004) test is employed to evaluate the out of sample forecasting performance of the pooling compared to the AR(1). In general, results indicate that a rich data set approach can provide valuable predictions about GDP behavior for the immediate future.