Do financial analysts get intangibles?

Abstract

It is widely agreed that corporate financial reports provide deficient information about intangible assets. However, investors are exposed to substantial information beyond financial reports, such as managers' direct communications to capital markets and analysts' reports. We ask: To what extent do these non-financial report sources compensate for the intangibles-related deficiencies of financial statements?

To address this question we assume that analysts' forecasts of earnings reflect, among other things, the beyond-financial-report information we seek, and we use simultaneous equations to estimate the incremental information contribution of earnings forecasts over the information contained in financial reports, thereby isolating value-relevant information not available in financial reports. We focus particularly on intangibles-related information, by comparing analysts' contribution for firms with and without R&D. We find that, to some extent, analysts do compensate for the intangibles-related information deficiencies of financial reports, but definitely not for all the deficiencies. Accordingly, we identify the ‘weakest links’–industries in which analysts do not get intangibles.

ACKNOWLEDGEMENTS

We thank I/B/E/S for providing analysts' earnings forecasts for this project. We also thank seminar participants at City University of New York (Baruch College), the University of Cyprus (Nicosia), Kent State University, London Business School, University of California-Berkeley and the 1999 AAA annual meetings (San Diego) for helpful comments. Eli Amir is grateful to the Israel Institute of Business Research for financial support.

Notes

¹This is often referred to as ‘price chasing’ by analysts.

²The validity of these signals was confirmed by various researchers, such as Abarbanell and Bushee (1997) and Francis and Schipper (1999).

³Capital market researchers use various model specifications, some different from the one we present in equation (1). To examine the robustness of our findings to model specification, we re-estimated our regressions with the Easton–Harris (Easton and Harris, 1991) specification, where returns are regressed on the levels and changes of earnings (along with the levels and changes of the present value earnings forecasts). Estimates from these Easton–Harris regressions are very close (in terms of R ² values–the statistic of interest to us) to the estimates generated by the model in equation (1) and reported in Tables 2 and 3.

⁴The Wall Street Journal's ‘Heard on the Street’ column (12 April 2000) argued that analysts adjust their forecasts and recommendations to justify price increases, rather than conduct an independent analysis. An article in the New York Times (8 August 2000), ‘A Bull Retreats in Downgrade of Web Shares’, reports on Morgan Stanley's analyst Mary Meeker, and Merrill Lynch's Henry Blodget, raising loss-per-share estimates for Internet stocks and downgrading recommendations in response to a sharp market decline. There was no market reaction to these downgrades, which is consistent with the argument that analysts reacted to the market, rather than the reverse.

⁵Conrad et al. (2002) found that analysts are equally likely to upgrade or downgrade a company's stock following a large stock price increase. However, analysts are more likely to downgrade a company's stock following a large stock price decline. They also found that upgrades are more likely if there is an investment banking relation with the company for which the recommendation is issued.

⁶The historical (from the 1920s to present) risk premium in the United States is approximately 7%. However, many experts believe that the risk premium has declined significantly in the decades since 1980, to levels in the range 3–5%. Claus and Thomas (2001) provide support to this level of risk premium.

⁷We interpret the significant coefficient on the present value of analysts' forecasts as indicating the information contribution of analysts to investors. However, we cannot rule out the possibility that investors are ‘fixated’ on analysts' earnings forecasts, and their reaction to the forecasts leads to mispricing.

⁸In the original Lev and Thiagarajan (1993) study, the AR coefficient was negative and significant only during periods of high inflation. However, we have no explanation for the positive coefficient in Table 2.

⁹While the theme of this study is intangibles, we focus in our empirical analysis on R&D, since R&D is the only intangible that is systematically reported by US companies.