Nowcasting Belgium

This paper proposes a method that takes into account the calendar of European and Belgian intraquarterly data releases to automatically update GDP growth expectations or nowcasts in realtime. The role of surveys is well known in the nowcasting literature, but this is the first paper that has attempted to isolate quality from timeliness as independent properties that can be expressed in function of the model parameters. The modeling framework allows for the incorporation of different kinds of survey data directly in levels and features a parsimonious specification of the GDP revision process which does not impose strict assumptions regarding the rationality of the statistical agency. The results in the empirical section emphasize the quality of survey data, which allows the model to produce accurate real GDP growth nowcasts for Belgium three months prior to the publication of the official flash estimate.