Can financial statement analysis beat consensus analysts’ recommendations?

Abstract

We examine whether investors can exploit financial statement information to identify companies with a greater likelihood of future earnings increases and whether stocks of those companies generate 1-year abnormal returns that exceed the abnormal returns from following analysts’ consensus recommendations. Our approach summarizes financial statement information into a “predicted earnings increase score,” which captures the likelihood of 1-year-ahead earnings increases. We find that, within our sample of consensus recommendations, stocks with high scores are much more likely to experience future earnings increases than stocks with low scores. A hedge portfolio strategy that utilizes our approach within each consensus recommendation level generates average annual abnormal returns of 10.9 percent over our 12-year sample period, after controlling for previously identified risk factors. These abnormal returns exceed those available from following analysts’ consensus recommendations. Our results show that share prices and consensus recommendations fail to impound financial statement information that helps predict future earnings changes.

This is a preview of subscription content, access via your institution.

Notes

  1. 1.

    Investors can interpret consensus hold recommendations to be (1) neutral, (2) ambiguous, or (3) veiled sells. Investors can interpret a consensus hold as neutral, implying the average analyst believes the firm’s shares are fairly priced and will perform in line with some benchmark portfolio index over some future holding period. Alternately, a consensus hold may be ambiguous if analysts are roughly equally divided on whether to buy or sell at the current share price. Investors can also interpret some consensus holds as veiled sells, because analysts rarely issue explicit underperform or sell recommendations. Some analysts issue a hold when they wish to disguise their negative beliefs to retain favorable access to the firm’s managers and avoid negative repercussions (McNichols and O’Brien 1997; Francis and Soffer 1997; Bradshaw 2004). Because they are open to different interpretations, consensus holds are not clear directional predictions of expected future share price performance.

  2. 2.

    We designed the predicted earnings increase score (PEIS) using the prior literature’s findings on the information in financial statement ratios but without peeking at the association between PEIS and future earnings changes or past or future stock returns data.

  3. 3.

    When we split the sample by recommendation level, we find earnings increases generated by 62.1, 60.1, 55.2, and 54.8 percent of firm-years associated with strong buy, buy, hold, and sell recommendations, respectively.

  4. 4.

    Our study adds to an already large stream of literature related to analysts’ stock recommendations. Research has examined individual analysts’ recommendation ability (Mikhail et al. 2004; Li 2005), analysts’ use of information in formulating signals for recommendations (Jegadeesh et al. 2004), the performance of recommendations (Barber et al. 2001, 2003, 2006; Ivkovic and Jegadeesh 2004), the information content of recommendations (Womack 1996; Asquith et al. 2005), and the relation between analysts’ earnings forecasts and recommendations (Francis and Soffer 1997; Eames et al. 2002; Bradshaw 2004).

  5. 5.

    Li (2005) states the I/B/E/S database contains recommendations from some of the larger brokerage houses that other databases do not contain. Our sample period extends to 2005, which is important because it contains observations after the adoption of new analyst and financial reporting regulations (e.g. NASD Rule 2711 and SEC Regulation FD).

  6. 6.

    Abarbanell and Lehavy (2003) use a similar classification system to form buy, hold, and sell portfolios.

  7. 7.

    Because only 1.6 percent of our sample observations have consensus recommendations equivalent to either “under-perform” or “sell,” we treat them as one combined “sell” category in our empirical analysis.

  8. 8.

    Prior research provides many insights into individual analysts’ earnings forecasts and recommendations, showing that individuals do not fully reflect the information available from financial statement analysis. We test consensus analysts’ recommendations because they diversify variation in individual analysts’ expectations and represent a proxy for the market’s expectations. We therefore seek to test whether analysts’ consensus recommendations and share prices fully reflect financial statement information.

  9. 9.

    The descriptive statistics in Table 2 reveal that the future returns data contain some extreme positive and negative observations. We do not delete or winsorize these extreme returns observations because they reflect ex post realizations of returns, which we do not know ex ante. When we form portfolios, we do not know whether realized returns will turn out to be positive or negative, extreme or not extreme. Thus, to delete or winsorize involves peeking ahead at the returns results. Our empirical results are not sensitive to extreme returns realizations.

  10. 10.

    Our finding of positive lagged abnormal returns associated with strong buy and buy recommendations is similar to the finding in Jegadeesh et al. (2004) that analysts’ buy recommendations include “glamour” stocks (for example, stocks with high momentum, growth, volume, and market multiples). Our data reveal an inverse relation between recommendation levels and future returns, consistent with evidence in Barber et al. (2003) and Drake et al. (2009).

  11. 11.

    The scoring model is similar to models used in Lev and Thiagarajan (1993), Abarbanell and Bushee (1997), Piotroski (2000), and Wieland (2006). We do not expect that the PEIS includes the optimal set of financial statement measures that help predict the future change in earnings, but it should distinguish those firms that are more or less likely to experience an earnings increase.

  12. 12.

    We chose the six signals included in PEIS only based on findings from research on the information in financial statement ratios and not based on any attempt to “fit” PEIS to future data on earnings changes or past or future stock returns.

  13. 13.

    We compute ATO using total assets in the denominator rather than net operating assets. Net operating assets can cause some unrealistically large values of ATO.

  14. 14.

    The descriptive statistics in Table 5 reveal that the variables that we use to construct the PEIS include some extreme positive and negative observations. We do not delete or winsorize these extreme observations in constructing PEIS because we do not use the values of the variables per se to compute PEIS and instead just use the cross-sectional ranks of these variables each year. Thus, extreme or winsorized values will not affect a firm’s PEIS.

  15. 15.

    For completeness, we note that, when we delete from our computations of PEIS the outlying top and bottom percentile of the observations for each variable each year, then PEIS prediction accuracy for this reduced sample changes marginally (64.4 percent of the firm-years in the top PEIS quintile and 43.3 percent of the firm-years in the bottom PEIS quintile generate earnings increases).

  16. 16.

    We base the inclusion of the control variables on evidence in (1) Fama and French (1992) that beta and the book-to-market ratio explain future returns, (2) Jegadeesh (1990) and Jegadeesh and Titman (1993) that short-run returns tend to persist in the subsequent year, (3) Haugen and Baker (1996) that low P/E ratio firms outperform high P/E ratio firms on a risk-adjusted basis, and (4) Sloan (1996) that trading strategies based on the extreme deciles of accruals generate abnormal returns.

  17. 17.

    We also estimated Eq. (1) including size to control for the possibility that nonlinear size effects influence our size-adjusted abnormal returns. As we report in the next section, our results remain essentially the same.

  18. 18.

    When we winsorize or delete the top and bottom percentile of the distribution of realized returns each year, we obtain results that are quantitatively and qualitatively similar to those in the paper. We do not believe winsorizing or deleting realized returns observations is appropriate because it requires peeking ahead to the returns results, which are not available at the time of portfolio formation.

  19. 19.

    We also estimate the models in pooled cross-sectional regressions and obtain similar results.

  20. 20.

    When we form portfolios using a less restrictive approach with long/short positions in the top/bottom quintile of PEIS each year (rather than the top/bottom quintile within each recommendation level each year), the abnormal returns results generally improve by roughly a full percentage point in both the buy-and-hold results in Table 7 and the controlled regression results in Table 8.

  21. 21.

    We also test a portfolio strategy including all sample hold recommendation stocks using relative weighting based on PEIS score rank. In our study, the higher the PEIS score the more likely the firm will experience an earnings increase so we scale our hedge variable from zero to one based on the quintile in which each falls each year. Specifically, we code our hedge variable as 4/4, 3/4, 2/4, 1/4, and 0/4, for the top to bottom quintile PEIS scores each year, respectively. The approach generated average annual abnormal returns of 18.6 percent and positive abnormal returns in 12 out of 12 years.

  22. 22.

    We also estimated Eq. (1) including a size variable, to control for the possibility that nonlinear size effects could influence our size-adjusted abnormal returns. Our results remain essentially the same. For example, the conditional hold strategy returns change from 19.3 percent without the size variable to 18.9 percent with the size variable, and they remain positive in 12 out of 12 years.

References

  1. Abarbanell, J., & Bushee, B. (1997). Fundamental analysis, future earnings, and stock prices. Journal of Accounting Research, 35, 1–24.

    Article  Google Scholar 

  2. Abarbanell, J., & Bushee, B. (1998). Abnormal returns to a fundamental analysis strategy. The Accounting Review, 73, 19–45.

    Google Scholar 

  3. Abarbanell, J., & Lehavy, B. (2003). Can stock recommendations predict earnings management and analysts’ earnings forecast errors? Journal of Accounting Research, 41, 1–31.

    Article  Google Scholar 

  4. Anderson, M., Banker, R., & Janakiraman, S. (2003). Are selling, general, and administrative costs “sticky”? Journal of Accounting Research, 41, 47–63.

    Article  Google Scholar 

  5. Asquith, P., Mikhail, M., & Au, A. (2005). Information content of equity analyst reports. Journal of Financial Economics, 75, 245–282.

    Article  Google Scholar 

  6. Barber, B., Lehavy, R., McNichols, M., & Trueman, B. (2001). Can investors profit from the prophets? Security analyst recommendations and stock returns. Journal of Finance, 56, 531–563.

    Article  Google Scholar 

  7. Barber, B., Lehavy, R., McNichols, M., & Trueman, B. (2003). Reassessing the returns to analysts’ stock recommendations. Financial Analysts Journal, 59(2), 88–96.

    Article  Google Scholar 

  8. Barber, B., Lehavy, R., McNichols, M., & Trueman, B. (2006). Buys, holds, and sells: The distribution of investment banks’ stock ratings and the implications for the profitability of analysts’ recommendations. Journal of Accounting and Economics, 41, 87–117.

    Article  Google Scholar 

  9. Beneish, M. D., & Vargus, M. E. (2002). Insider trading, earnings quality, and accrual mispricing. The Accounting Review, 77, 755–791.

    Article  Google Scholar 

  10. Bernard, V., & Thomas, J. (1990). Evidence that stock prices do not fully reflect the implications of current earnings for future earnings. Journal of Accounting and Economics, 13, 305–340.

    Article  Google Scholar 

  11. Bernard, V., Thomas, J., & Wahlen, J. (1997). Accounting-based stock price anomalies: Separating market inefficiencies from risk. Contemporary Accounting Research, 14(2), 89–136.

    Article  Google Scholar 

  12. Bradshaw, M. T. (2004). How do analysts use their earnings forecasts in generating stock recommendations? The Accounting Review, 79, 25–50.

    Article  Google Scholar 

  13. Drake, M., Rees, L., & Swanson, E. (2009). Should investors follow the prophets or the bears? Evidence on the use of public information by analysts and short sellers Working paper, Texas A&M University. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1269427.

  14. Eames, M., Glover, S., & Kennedy, J. (2002). The association between trading recommendations and broker-analysts’ earnings forecasts. Journal of Accounting Research, 40, 85–104.

    Article  Google Scholar 

  15. Fairfield, P., & Yohn, T. (2001). Using asset turnover and profit margin to forecast changes in profitability. Review of Accounting Studies, 6(4), 371–385.

    Google Scholar 

  16. Fama, E., & French, K. (1992). The cross-section of expected stock returns. Journal of Finance, 47, 427–465.

    Article  Google Scholar 

  17. Fama, E., & MacBeth, J. (1973). Risk, return and equilibrium: Empirical tests. Journal of Political Economy, 81, 607–636.

    Article  Google Scholar 

  18. Francis, J., & Soffer, L. (1997). The relative informativeness of analysts’ stock recommendations and earnings forecast revisions. Journal of Accounting Research, 35, 193–211.

    Article  Google Scholar 

  19. Haugen, R. A., & Baker, N. L. (1996). Commonality in the determinants of expected stock returns. Journal of Financial Economics, 41, 401–439.

    Article  Google Scholar 

  20. Ivkovic, Z., & Jegadeesh, N. (2004). The timing and value of forecast and recommendation revisions. Journal of Financial Economics, 73, 433–463.

    Article  Google Scholar 

  21. Jegadeesh, N. (1990). Evidence of predictable behavior of security returns. Journal of Finance, 47, 410–428.

    Google Scholar 

  22. Jegadeesh, N., Kim, J., Krische, S., & Lee, C. M. C. (2004). Analyzing the analysts: When do recommendations add value? Journal of Finance, 59, 1083–1124.

    Article  Google Scholar 

  23. Jegadeesh, N., & Titman, S. (1993). Returns to buying winners and selling losers. Journal of Finance, 48, 881–898.

    Article  Google Scholar 

  24. Lev, B., & Thiagarajan, S. R. (1993). Fundamental information analysis. Journal of Accounting Research, 31, 190–215.

    Article  Google Scholar 

  25. Li, X. (2005). The persistence of relative performance in stock recommendations of sell-side financial analysts. Journal of Accounting and Economics, 40, 129–152.

    Article  Google Scholar 

  26. McNichols, M., & O’Brien, P. (1997). Self-selection and analyst coverage. Journal of Accounting Research, 35, 167–199.

    Google Scholar 

  27. Mikhail, M., Walther, B., & Willis, R. (2004). Do security analysts exhibit persistent differences in stock picking ability? Journal of Financial Economics, 74, 67–91.

    Article  Google Scholar 

  28. Ou, J., & Penman, S. (1989). Financial statement analysis and the prediction of stock returns. Journal of Accounting and Economics, 11, 295–330.

    Article  Google Scholar 

  29. Penman, S., & Zhang, X. (2006). Modeling sustainable earnings and P/E ratios with financial statement analysis. Working paper. Columbia University and University of California, Berkeley. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=318967.

  30. Piotroski, J. (2000). Value investing: The use of historical financial statement information to separate winners from losers. Journal of Accounting Research, 38, 1–41.

    Article  Google Scholar 

  31. Sloan, R. (1996). Do stock prices fully reflect information in accruals and cash flows about future earnings? The Accounting Review, 71, 289–315.

    Google Scholar 

  32. Wieland, M. (2006). Identifying consensus analysts’ earnings change forecasts with incorrect signs. Working paper, University of Georgia.

  33. Womack, K. (1996). Do brokerage analysts’ recommendations have investment value? Journal of Finance, 51, 137–167.

    Article  Google Scholar 

Download references

Acknowledgments

We gratefully acknowledge helpful comments from S. Baginski, L. Bamber, M. Billings, M. Illueca, C. Nichols, the editor and reviewers, and workshop participants at the University of Georgia, Cornell University, the University of Toronto, and the BBVA Foundation—IVIE International Seminar on Accounting, Financial Institutions, and Capital Markets in Castellon-de-Plana, Spain. I/B/E/S provided analyst forecast data as part of a broad academic program to encourage earnings expectations research.

Author information

Affiliations

Authors

Corresponding author

Correspondence to James M. Wahlen.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Wahlen, J.M., Wieland, M.M. Can financial statement analysis beat consensus analysts’ recommendations?. Rev Account Stud 16, 89–115 (2011). https://doi.org/10.1007/s11142-010-9124-5

Download citation

Keywords

  • Earnings predictions
  • Financial statement analysis
  • Consensus recommendations
  • Abnormal returns
  • Sell side analysts

JEL Classification

  • G10
  • G11
  • G14
  • G17
  • M41