Publication

This paper proposes an approach for estimating the uncertainty associated with model-based macroeconomic forecasts. The authors argue that estimated forecast intervals should account for the uncertainty arising from selecting the specification of a model from sample data. To this end, they formalize a model selection procedure that specifies the lag structure of a model and accounts for aberrant observations. They then apply the procedure to assess the risk of deflationary developments occurring in Germany over the four years after 2003.