In the August 2005 draft of their paper entitled “Analyst Conflicts and Research Quality”, Anup Agrawal and Mark Chen examine whether the forecasts quality of stock analysts relates to conflicts of interest from the investment banking and brokerage businesses of their employers. They define forecast quality in terms of: (1) accuracy; (2) bias; (3) frequency of quarterly earnings per share (EPS) forecast revisions; and, (4) relative optimism of long-term earnings growth forecasts. By cross-referencing the forecasts of 3,000 analysts with line-of-business revenue breakdowns for their respective employers (163 different firms) over the period 1994-2003, they find that:
- The bias and accuracy of quarterly EPS forecasts do not relate systematically to business conflicts.
- Brokerage business conflicts systematically elevate the frequency of quarterly EPS forecast revisions. Trade generation appears to be one motive for this bias.
- Brokerage business conflicts systematically elevate long-term earnings forecasts. Analyst reputations may be less at stake for long-term forecasts than short-term, and an optimistic long-term earnings forecast supports higher stock prices [thereby preserving relationships between analysts and the management of companies they cover].
In summary, conflicts from the brokerage business (not investment banking) play an important role in shaping analyst forecasting behavior.
