Over the past few years, considerable progress has been made in the area of macroeconomic and financial forecasting. This special issue presents recent theoretical and empirical contributions related to forecast uncertainty. The assessment of forecasts and their uncertainty is of particular interest for policymakers confronted with the task of exploiting all the available information and evaluating the relative accuracy and relevance of forecasts from different sources. The special issue provides a collection of papers that contribute to the literature on forecast uncertainty by investigating different aspects of forecast uncertainty, including information uncertainty, parameter and model uncertainty (e.g. forecast and model combination, forecast model selection, forecast evaluation) and measurement uncertainty (in particular, real-time data analysis).
This special issue on 'Forecast Uncertainty in Macroeconomics and Finance' resulted from a refereeing process of papers submitted responding to the call for papers following the 5th Workshop on Forecasting Techniques on the same topic organized at the European Central Bank in December 2007. The majority of the papers published in this special issue were presented at that conference.
The papers in this special issue can be grouped into three broad topics: (i) forecast and information uncertainty, forecast evaluation and potentially unstable environments; (ii) density forecasts and forecast combination; and (iii) measurement uncertainty, real-time data analysis and indicators. In the following, we summarize the contributions of the papers included in this special issue to the different topics.
1. FORECAST AND INFORMATION UNCERTAINTY, FORECAST EVALUATION AND POTENTIALLY UNSTABLE ENVIRONMENTS
In the paper 'Measuring forecast uncertainty by disagreement: the missing link', Kajal Lahiri and Xuguang Sheng introduce a new measure of forecast uncertainty based on forecast disagreement. Using a standard decomposition of forecast errors into common and idiosyncratic components, and under fairly reasonable assumptions on the cumulated shocks and the idiosyncratic component, the authors decompose the forecast uncertainty into the disagreement and the variance of the accumulated aggregate shocks over the forecast horizon. The reliability of disagreement as a proxy for aggregate uncertainty, therefore, depends on the (perception of) variability of aggregate