Nowcasting macroeconomic aggregates in Argentina: comparing the predictive ability of different models

Abstract

Monetary policy making requires a correct and timely assessment of current macroeconomic conditions. While the main source of macroeconomic data is quarterly National Accounts, often published with a significant lag, higher frequency business cycle indicators are increasingly available. Taking this into account, central banks have adopted nowcasting as a useful tool for having an immediate and more accurate perception of economic conditions. In this paper, we extend the use of nowcasting tools to produce early indicators of the evolution of two components of aggregate domestic demand: consumption and investment. The exercise uses a broad and restricted set of indicators to construct different dynamic factor models, as well as a pooling of models in the case of investment. Finally, we compare different approaches in a pseudo-real time out-of-sample exercise and evaluate their predictive performance.