Filtering German economic conditions from a large dataset: The new DIW economic barometer

This paper presents a revised version of the DIW Economic Barometer, the business cycle index of the German Institute for Economic Research (DIW Berlin). As in earlier versions, we put forward a factor model on a monthly frequency to filter the latent state of the aggregate economy. In the new version, the resulting business cycle factor is based on more than 300 variables. The main methodological changes relate to (i) the estimation procedure, (ii) treatment of publication lags and missings, and (iii) the decomposition of the index into contributions from different sectors of the economy. Alongside several practical advantages, we also document a better historical nowcasting performance of the new index.