Nowcasting Finnish GDP growth using financial variables: a MIDAS approach

Abstract

We analyse the performance of financial market variables in nowcasting Finnish quarterly GDP growth. In particular, we assess if prediction accuracy is affected by the sampling frequency of the financial variables. Therefore, we apply MIDAS models that allow us to nowcast quarterly GDP growth using monthly or daily data without temporal aggregation in a parsimonious way. We find that financial market data nowcasts Finnish GDP growth relatively well: nowcasting performance is similar to industrial production, but financial market data is available much earlier. Our results suggest that the sampling frequency of financial market variables is not crucial: nowcasting accuracy of daily, monthly and quarterly data is similar.