Ein Kurzfristindikatormodell für Prognosen der internationalen Konjunktur

When forecasting, all available information should be used. As the most recent indicators are mostly published in monthly or higher frequency while the GDP is released in quarterly frequency, this can pose a problem in an econometric framework. This paper presents the KOF short-term indicator for international forecasts, which uses both bridge equations as well as a MIDAS approach. The forecast performance of the models is then assessed by conducting a real-time experiment using a reduced dataset for the US. Even with this reduced set forecast performance can be increased by up to 25 percent in comparison to an autoregressive model.