Nowcasting of Investment. A real-time estimation with high-frequency indicators.
Working papers | 2018 | N 78
Keywords:
Nowcasting, Dynamic factor models, Real-time forecasts, Forecast poolingAbstract
In the present study, a real-time estimation of Investment evolution was performed using a wide array of high-frequency economic indicators, a methodology known in the literature as Nowcasting. To conduct the Nowcast, three groups of monthly frequency indicators were considered, and dynamic factor models were employed to forecast quarterly Investment growth. Additionally, a pooling exercise or forecast combination was conducted. Based on the Giacomini and White test, it was concluded that factor models and their combinations exhibit better predictive ability compared to an AR(1) model used as a benchmark, and that the inclusion of a greater number of indicators does not necessarily improve forecast performance.
JEL classification: C22, C53, E37
