IFOCAST: Methods of the Ifo short-term forecast

The assessment and forecast of the economic situation in the current and coming quarter is one of the key tasks of economic activity forecasting. The Ifo Institute bases its short-term forecasts of GDP on the three-stage IFOCAST approach. In the first stage, available monthly indicators, for example the Ifo Business Climate, are extrapolated and aggregated to a quarterly level. Special attention is given to industry production, which is updated with the aid of disaggregated Ifo survey data. In a second step the gross value added of individual economic sectors is predicted with the help of bridge equations. Using a combination approach, a number of models are joined in order to compensate for model insecurity. In a third step the quarterly forecasting of individual economic sectors is aggregated by means of the economic weights for the forecasting of GDP. Experience in both the forecasting literature and in practice shows that this approach supplies reliable short-term forecasts and is flexible enough to cope also with extreme developments. In addition to the multi-phased standard procedures, this article discusses the mixed-frequency models and boosting algorithms that complement the standard approach in the test phase.