Does Soft Information Matter? An Application of Gravity Models to Financial Analysts' Forecasts
29 Pages Posted: 19 Mar 2011
Date Written: March 15, 2011
We study whether the financial analysts' concern to maintain friendly relationships with firms' managers in order to preserve their access to ‘soft’ qualitative information entice them to issue pessimistic (“earnings surprise management” hypothesis) or optimistic (“management access” hypothesis) forecasts.
One important innovation of our approach is to apply the methodology of gravity models to relationships between firms and analysts. The second interest of our paper is to propose a measure of soft information. Using a panel regression of analysts forecast error on observable firm-specific, analyst-specific and both firm-analyst characteristics, we decompose the fixed effect in order to extract a pair-specific effect. We compute the contribution of the pair effect to the total fixed effect for each observation. We interpret this contribution as a measure of soft information, as it describes the part of the unobservable variables which is due to every analyst-firm pair. It allows us to check whether soft information contributes to analysts' pessimism or to analysts' optimism.
We use a data set provided by ThomsonReuters that contains the one-year ahead Earnings Per Share forecasts issued by 4 648 analysts about the earnings of 241 French firms over the period 1997-2007.
We find that, for each analyst-firm pair, a low pair-effect is associated with a low forecast error, and that a high pair-effect is associated with a high forecast error, may the error be optimistic or pessimistic. This suggests that the need to preserve their relationship with firms' manager induce some analysts to issue pessimistic forecasts while inducing some others to issue optimistic forecasts about firms' EPS. Although they refer to opposite forecast patterns, pessimism and optimism both result from the same analyst's concern to preserve access to soft information released by managers.
Keywords: panel regression, financial analysts, forecasts, accuracy, optimism, pessimism, soft information
JEL Classification: D84, G17, G24
Suggested Citation: Suggested Citation