The Impact of Segment Definition on the Accuracy of Analysts' Earnings Forecasts

Abstract

This study investigates forecast error determinants for a set of forecasts of annual corporate earnings, generated by UK analysts 22 months prior to the announcement dates. This study is particularly concerned with the impact of segmental data on forecast errors; the hypothesis under test is whether finer segment definitions provide market participants with improved insight. If segments are too broad or vague (e.g. rest of the world) it is unlikely that data for such segments will provide analysts with any additional information regarding the current corporate position or future prospects. The results of this study provide evidence of predictive gains to both line- of-business data and geographic data, although these gains appear to be concentrated within a sub-sample of firms for which analysts appear to have specific difficulty in forecasting earnings, i.e. those experiencing negative changes in earnings. The results also indicate that forecast errors are: negatively related to firm size; positively related to the magnitude of the change in earnings which the analyst must predict; and not significantly affected by the number of reported segments.