The informativeness of analyst forecast revisions and the valuation of R&D-intensive firms

Prior studies (e.g., [McNichols and O'Brien, 1997] and [Diether et al., 2002]) find that analysts are less willing to disclose unfavorable earnings forecasts than to disclose favorable forecasts, and this tendency induces an optimistic bias in disclosed forecasts that increases with the degree of earnings uncertainty. Building on these findings, we predict that, in the context of R&D-intensive industries, there should be differential informativeness and asymmetric valuation roles for upward versus downward analyst forecast revisions. Consistent with our predictions, we find the following evidence: (i) analyst forecast revisions contain a downward bias, causing upward revisions to under-represent, whereas downward revisions to over-represent, changes in true earnings expectations, with the extent of over/under-representation greater for firms with higher R&D expenditures; (ii) upward revisions are associated with more rapid reductions in earnings uncertainties (proxied by forecast dispersions) than downward revisions, mainly for high R&D firms; and (iii) upward revisions are more effective in mitigating the return differentials between high and low R&D firms (as documented in Chan et al., 2001).