Abstract
This study examines the effect of firm-level investor sentiment derived from news articles and Twitter media content on analyst herding. The results indicate improvements (deterioration) in investor sentiment derived from news and Twitter media content lead to an increase (decrease) in analyst herding. This effect is primarily driven by media content with positive sentiment, and the effect size is magnified when the news and Twitter media content share the same sentiment polarity. Finally, the effect of firm-level investor sentiment on analyst herding is most pronounced in firms with low valuation uncertainty. By establishing a link between firm-level investor sentiment derived from news and Twitter content and analyst herding, this paper shows that analyst herding is amplified by firm-level investor sentiment, and the effect is more pronounced for firms with low valuation uncertainty.
This is a preview of subscription content, access via your institution.
Notes
The quarterly EPS dispersion measure is log transformed to make the data approximately normal. Since some observations have a zero standard deviation, I add a small constant (0.00268) to all the dispersion values before taking the logarithm. Consistent with Liu and Nararajan (2012) and Berry (1987), I add a constant that minimizes the sum of the absolute values of skewness and excess kurtosis. The resulting distribution of log (DISP + constant) is approximately normal. The sum of the absolute values of skewness and excess kurtosis is 1.45.
See Cui et al. (2016) for a summary of Bloomberg’s sentiment measures and how they are computed. In summary, machine-learning techniques are used to construct news & social sentiment measures, which use classification engines trained to mimic a human expert in processing textual information. The classification is based on the question, “If an investor having a long position in the security mentioned were to read this news or tweet, is he/she bullish, bearish or neutral on his/her holdings?” Bloomberg then compiles and publishes company-level average daily sentiment scores, which are the confidence-weighted average of the past day’s story-level sentiments for both news and Twitter.
The absolute value of the prior quarter return is log transformed to make the data approximately normal. Since some observations have a zero return, I added a small constant (0.6185) to all the values before taking the logarithm. I add a constant that minimizes the sum of the absolute values of skewness and excess kurtosis. The resulting distribution is approximately normal. The sum of the absolute values of skewness and excess kurtosis is 0.45.
The quarterly debt-to-capital ratio and average bid-ask spread is log transformed to make the data approximately normal.
The debt-to-capital ratio is log transformed to make the data approximately normal. Since some observations have zero debt, one is added to all debt-to-capital ratios before taking the logarithm.
The absolute value of the mean EPS forecast is log-transformed to make the data approximately normal. Since some observations have a zero earnings estimate, I added a small constant (0.0717) to all the dispersion values before taking the logarithm. I add a constant that minimizes the sum of the absolute values of skewness and excess kurtosis. The resulting distribution is approximately normal. The sum of the absolute values of skewness and excess kurtosis is 0.34.
Bloomberg provides estimates of current analyst ratings on a scale from 1 and 5, with 5 indicating the strongest ranking (buy or similar) and 1 indicating the weakest (sell or similar). The first quartile is 3.4, the mean is 3.9, the median is 4.0, and the third quartile is 4.4. Given the positively skewed distribution, the ratings groups are grouped into three groups: analyst rating of 3 or less (group 1), rating from 3 to 4 (group 2), and ratings higher than 4 (group 3).
Using the PE ratio relative to the PE ratio of the firm’s relevant benchmark index as designated by Bloomberg (e.g., Standard and Poor’s 500, Russell 1000, etc.), firms above the 3rd quartile (1.575) are coded to 1.Firms with no earnings or negative earnings are also coded to 1; all other observations are coded to 0.
To measure economic policy uncertainty, I used the Economic Policy Uncertainty Index as derived by Baker et al. (2016). The index measures were downloaded from https://www.policyuncertainty.com/.
To mitigate the effect of extreme outliers, the quarterly return (prior to taking the absolute value), book-to-market ratio, and 90-day share price volatility measures are winsorized at the 1st and 99th percentiles and the debt-to-capital ratio and EPS dispersion measures are winsorized at the 99th percentile.
Following Newey and West (1994), where the optimal lags L = floor (4*(N/100)2/9), and N is the firm-level sample size, and given the maximum number of quarterly observations per firm is 20, two lags are used.
The controls include; log of mean analyst’ EPS, log of share price, log of the 90-day share price volatility, log of market capitalization, beta, log of total analysts covering the firm, log of the quarterly average bid-ask spread, the log of the debt-to-capital ratio, the absolute value of the quarterly return, the book-to-market ratio, the change in EPS, economic uncertainty index, dummy variables for whether the firm reported a loss, experienced a negative quarterly return, the firm incurred R&D expense, the firm has a high relative PE ratio, valuation uncertainty (where “low” is the reference), and analyst rating groups.
The covariance matrix of the regression coefficients are corrected for heteroskedasticity and autocorrelation. All subsequent Wald tests also corrected for hteroskedasticity and autocorrelation.
The correlation between news and Twitter sentiment is only 0.23, with 72% of the observations sharing the same sentiment valence.
Due to the perfect collinearity between the polarity indicator variables for news and Twitter sentiment when the subsample of observations with the same polarity is selected, I re-run the regressions individually for Twitter sentiment (columns 2 and 3) and news sentiment (columns 4 and 5).
The median rating is 4.0 (a buy rating) on a 1- 5 scale, and the 1st quartile is 3.41 (between a hold and buy rating).
The Baker and Wurgler market sentiment index was available through 2018 and was obtained from Dr. Wurgler’s website: http://people.stern.nyu.edu/jwurgler/. To arrive at quarterly measures, the quarterly average is derived from the monthly measures.
The Pearson correlation between the quarterly Baker & Wurgler index measure and the quarterly news and Twitter sentiment measures are only −0.0204, and 0.0051, respectively.
On a daily basis, Bloomberg compiles analyst ratings. A rating scale between 1 and 5 is used, with 5 being the strongest rating (buy or similar) and 1 being the weakest (sell or similar).
References
Aboody, D., Even-Tov, O., Lehavy, R., Trueman, B.: Overnight returns and firm-specific investor sentiment. J. Financ. Quantitative Anal. 53(2), 485–505 (2018). https://doi.org/10.1017/S0022109017000989
Agrawal, S., Azar, P.D., Lo, A.W., Singh, T.: Social media: Evidence from StockTwits and Twitter. J. Portf. Manage. 44(7), 85–95 (2018)
Antweiler, W., Frank, M.Z.: Is all that talk just noise ? The information content of Internet stock message boards. J. Financ. 59(3), 1259–1294 (2004). https://doi.org/10.1111/j.1540-6261.2004.00662.x
Avramov, D., Chordia, T., Jostova, G., Philipov, A.: Dispersion in analysts’ earnings forecasts and credit rating. J. Financ. Econ. 91(1), 83–101 (2009). https://doi.org/10.1016/j.jfineco.2008.02.005
Baird, P.L.: Do investors recognize biases in securities analysts’ forecasts? Rev. Financ. Econ. (2019). https://doi.org/10.1002/rfe.1094
Barber, B.M., Odean, T.: The courage of misguided convictions. Financ. Anal. J. 55(6), 41–55 (1999). https://doi.org/10.2469/faj.v55.n6.2313
Baker, M., Wurgler, J.: Investor sentiment and the cross-section of stock returns. J. Finance 61(4), 1645–1680 (2006). https://doi.org/10.1111/j.1540-6261.2006.00885.x
Baker, M., Wurgler, J.: Investor sentiment in the stock market. J. Econ. Perspect. 21(2), 129–151 (2007). https://doi.org/10.1257/jep.21.2.129
Barker, R.G.: The role of dividends in valuation models used by analysts and fund managers. Int. J. Phytorem. 21(1), 195–218 (1999). https://doi.org/10.1080/096381899335998
Baker, S.R., Bloom, N., Davis, S.J.: Uncertainty, measuring economic policy. Q. J. Econ. 131(4), 0–52 (2016). https://doi.org/10.1093/qje/qjw024
Barron, O.E., Byard, D., Kim, O.: Changes in analysts’ information around earnings announcements. Account. Rev. 77(4), 821–846 (2002). https://doi.org/10.2308/accr.2002.77.4.821
Barron, O. E., Kim, O., Lim, S. C., & Stevens, D. E.: Using analysts’ forecasts to measure properties of analysts’ information environment. Accounting Review, 73(4), 421–433. (1998) https://www.jstor.org/stable/248184
Barron, O.E., Stuerke, P.S.: Dispersion in analysts’ earnings forecasts as a measure of uncertainty. J. Account, Audit. Finance 13(3), 245–270 (1998). https://doi.org/10.1177/0148558X9801300305
Behrendt, S., Schmidt, A.: The Twitter myth revisited: Intraday investor sentiment, Twitter activity and individual-level stock return volatility. J. Bank. Finance 96, 355–367 (2018). https://doi.org/10.1016/j.jbankfin.2018.09.016
Berry, D. A. Logarithmic transformations in ANOVA. Biometrics, 43(2), 439–456. (1987) https://www.jstor.org/stable/2531826
Bhushan, R.: Firm characteristics and analyst following. J. Account. Econ. 11(2–3), 255–274 (1989). https://doi.org/10.1016/0165-4101(89)90008-6
Blankespoor, E., DeHaan, E., Zhu, C.: Capital market effects of media synthesis and dissemination: Evidence from robo-journalism. Rev. Account. Stud. 23(1), 1–36 (2018). https://doi.org/10.1007/s11142-017-9422-2
Blasco, N., Corredor, P., Ferrer, E.: Analysts herding: When does sentiment matter? Appl. Econ. 50(51), 5495–5509 (2018). https://doi.org/10.1080/00036846.2018.1486999
Bollen, J., Mao, H., Zeng, X.: Twitter mood predicts the stock market. J. Comput. Sci. 2(1), 1–8 (2011). https://doi.org/10.1016/j.jocs.2010.12.007
Bowen, R.M., Davis, A.K., Matsumoto, D.A.: Do conference calls affect analysts’ forecasts? Account. Rev. 77(2), 285–316 (2002). https://doi.org/10.2139/ssrn.216810
Bradshaw, M.T.: How do analysts use their earnings forecasts in generating stock recommendations? Accoun. Rev. 79(1), 25–50 (2004). https://doi.org/10.2308/accr.2004.79.1.25
Bradshaw, M.T.: Analysts’ forecasts: What do we know after decades of work? SSRN Electron. J. (2011). https://doi.org/10.2139/ssrn.1880339
Brenan, M.: Americans’ trust in mass media edges down to 41%. Gallup.com. (2019) https://news.gallup.com/poll/267047/americans-trust-mass-media-edges-down.aspx
Brown, L.D.: Earnings forecasting research: Its implications for capital markets research. Int. J. Forecast. 9, 295–320 (1993). https://doi.org/10.1016/0169-2070(93)90026-J
Brown, G.W., Cliff, M.T.: Investor sentiment and the near-term stock market. J. Empir. Finance 11(1), 1–27 (2004). https://doi.org/10.1016/j.jempfin.2002.12.001
Chen, H., De, P., Hu, Y., Hwang, B.H.: Wisdom of crowds: The value of stock opinions transmitted through social media. Rev. Financ. Stud. 27(5), 1367–1403 (2014). https://doi.org/10.1093/rfs/hhu001
Chiang, M.T., Lin, M.C.: Market sentiment and herding in analysts’ stock recommendations. North Am. J. Econ. Finance 48(1), 48–64 (2019). https://doi.org/10.1016/j.najef.2019.01.007
Cohen, D.A., Lys, T.Z.: A note on analysts’ earnings forecast errors distribution. J. Account. Econ. 36(1–3), 147–164 (2003). https://doi.org/10.1016/j.jacceco.2003.11.002
Colla, P., Mele, A.: Information linkages and correlated trading. Rev. Financ. Stud. 23(1), 203–246 (2010). https://doi.org/10.1093/rfs/hhp021
Conrad, J., Cornell, B., Landsman, W.R., Rountree, B.R.: How do analyst recommendations respond to major news? J. Financ. Quantitative Anal. 41(1), 25–49 (2006). https://doi.org/10.1017/S0022109000002416
Cui, X., Lam, D., Verma, A.: Embedded value in Bloomberg news & social sentiment data. Bloomberg L.P, New York (2016)
Cumming, G.: Inference by eye: Reading the overlap of independent confidence intervals. Stat. Med. 28(2), 205–220 (2009). https://doi.org/10.1002/sim
Daniel, K., Hirshleifer, D., Subrahmanyam, A.: Investor psychology and security market under- and overreactions. J. Finance 53(6), 1839–1885 (1998). https://doi.org/10.1111/0022-1082.00077
Davis, A.K., Tama-Sweet, I.: Managers’ use of language across alternative disclosure outlets: Earnings press releases versus MD&A. Contemp. Account. Res. 29(3), 804–837 (2012). https://doi.org/10.1111/j.1911-3846.2011.01125.x
De Long, J. B., Shleifer, A., Summers, L. H., & Waldmann, R. J.: Noise trader risk in financial markets. Journal of Political Economy, 98(4), 703–738. (1990) https://www.jstor.org/stable/2937765
Demirakos, E.G., Strong, N.C., Walker, M.: What valuation models do analysts use? Account. Horizons 18(4), 221–240 (2004). https://doi.org/10.2308/acch.2004.18.4.221
Diether, K.B., Malloy, C.J., Scherbina, A.: Differences of opinion and the cross-section of stock returns. J. Finance 57(5), 2113–2141 (2002). https://doi.org/10.1111/0022-1082.00490
Dugar, A., Nathan, S.: The effect of investment banking relationships on financial analysts’ earnings forecasts and investment recommendations. Contemp. Account. Res. 12(1), 131–160 (1995). https://doi.org/10.1111/j.1911-3846.1995.tb00484.x
Dunham, L., Garcia, J.: Measuring the effect of investor sentiment on liquidity. Managerial Finance 47(1), 59–85 (2021). https://doi.org/10.1108/MF-06-2019-0265/full/html
Duru, A., Reeb, D.M.: International diversification and analysts’ forecast accuracy and bias. Account. Rev. 77(2), 415–433 (2002). https://doi.org/10.2308/accr.2002.77.2.415
Easterwood, J.C., Nutt, S.R.: Inefficiency in analysts’ earnings forecasts: Systematic misreaction or systematic optimism? J. Finance 54(5), 1777–1797 (1999). https://doi.org/10.1111/0022-1082.00166
Engle, R.F., Ng, V.K.: Measuring and testing the impact of news on volatility. J. Finance 48(5), 1749–1778 (1993). https://doi.org/10.1111/j.1540-6261.1993.tb05127.x
Gandhi, P., Loughran, T., McDonald, B.: Using annual report sentiment as a proxy for financial distress in U.S. banks. J. Behav. Finance 20(4), 424–436 (2019). https://doi.org/10.2139/ssrn.2905225
Garcia, D.: Sentiment during recessions. J. Finance 68(3), 1267–1300 (2013). https://doi.org/10.1111/jofi.12027
Garfinkel, J.A.: Measuring investors’ opinion divergence. J. Account. Res. 47(5), 1317–1348 (2009). https://doi.org/10.1111/j.1475-679X.2009.00344.x
Greenwich Associates. (2019). Investing in the digital age: Media’s role in the institutional investor engagement journey. pp 1–16. https://www.greenwich.com/market-structure-technology/investing-digital-age
Han, B., Yang, L.: Social networks, information acquisition, and asset prices. Manage. Sci. 59(6), 1444–1457 (2013). https://doi.org/10.1287/mnsc.1120.1678
Henry, E.: Are investors influenced by how earnings press releases are written? J. Bus. Commun. 45(4), 363–407 (2008). https://doi.org/10.1177/0021943608319388
Hong, H., Kubik, J.D., Stein, J.C.: Social Interaction and Stock-Market Participation. J. Finance 59(1), 137–163 (2004). https://doi.org/10.1111/j.1540-6261.2004.00629.x
Hribar, P., & McInnis, J.: Investor sentiment and analysts’ earnings forecast errors. Management Science, 58(2), 293–307. (2012) https://www.jstor.org/stable/41406389
Kaplanski, G., & Levy, H.: Sentiment effect on analysts ’ recommendations : Time-series and cross-section analyses. Working paper. Bar-IIan University and The Hebrew University oJerusalem. (2010)
Kadous, K., Mercer, M., & Zhuo, Y.: Undue Influence? The Effect of Social Media Advice on Investment Decisions. SSRN. (2017) Retrieved from https://doi.org/10.2139/ssrn.2968407
Kaplanski, G., Levy, H.: Analysts and sentiment: A causality study. Quat Rev. Econ. Finance 63, 315–327 (2017). https://doi.org/10.1016/j.qref.2016.06.002
Ke, B., Yu, Y.: Why don’t analysts use their earnings forecasts in generating stock recommendations? China Account. Finance Rev. 22(1), 31–66 (2020). https://doi.org/10.2139/ssrn.1011449
Kothari, S.P.: Capital markets research in accounting. J. Account. Econ. 31(1–3), 105–231 (2001). https://doi.org/10.1016/S0165-4101(01)00030-1
Kumar, A.: Hard-to-value stocks, behavioral biases, and informed trading. J. Financ. Quantitative Anal. 44(6), 1375–1401 (2009). https://doi.org/10.1017/S0022109009990342
Lim, T.: Rationality and analysts’ forecast bias. J. Finance 56(1), 369–385 (2001). https://doi.org/10.1111/0022-1082.00329
Lin, H.W., McNichols, M.F.: Underwriting relationships, analysts’ earnings forecasts and investment recommendations. J. Account. Econ. 25(1), 101–127 (1998). https://doi.org/10.1016/S0165-4101(98)00016-0
Liu, X.G., Natarajan, R.: The effect of financial analysts’ strategic behavior on analysts’ forecast dispersion. Account. Rev. 87(6), 2123–2149 (2012). https://doi.org/10.2308/accr-50212
Loewenstein, G., Hsee, C.K., Weber, E.U., Welch, N.: Risk as feelings. Psychol. Bull. 127(2), 267–286 (2001). https://doi.org/10.1037/0033-2909.127.2.267
Loughran, T., McDonald, B.: When is a liability not a liability? Textual analysis, dictionaries, and 10-Ks. J. Finance 66(1), 35–65 (2011). https://doi.org/10.1111/j.1540-6261.2010.01625.x
Maat, H.P.: How promotional language in press releases is dealt with by journalists: Genre mixing or genre conflict? J. Bus. Commun. 44(1), 59–95 (2007). https://doi.org/10.1177/0021943606295780
Mayer, J., Hanson, E.: Mood-Congruent Judgment Over Time. Pers. Soc. Psychol. Bull. 21(3), 237–244 (1995)
Mayew, W.J., Sethuraman, M., Venkatachalam, M.: MD&A disclosure and the firm’s ability to continue as a going concern. Account. Rev. 90(4), 1621–1651 (2015). https://doi.org/10.2308/accr-50983
McGurk, Z., Nowak, A., Hall, J.C.: Stock returns and investor sentiment: textual analysis and social media. J. Econ. Finance 44(3), 458–485 (2020). https://doi.org/10.1007/s12197-019-09494-4
Morris, M., Counts, S., Roseway, A., Hoff, A., & Schwarz, J.: Tweeting is believing? Understanding microblog credibility perceptions. In Proceedings of the ACM 2012 Conference on Computer Supported Cooperative Work, 441–450 (2012). Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/14571036
Nagar, V., Schoenfeld, J., Wellman, L.: The effect of economic policy uncertainty on investor information asymmetry and management disclosures. J. Account. Econ. 67(1), 36–57 (2019). https://doi.org/10.1016/j.jacceco.2018.08.011
Newey, W. K., & West, K. D.: A simple, positive semi-definite, heteroskedasticity and autocorrelation consistent covariance matrix. Econometrica, 55(3), 703–708 (1987). https://www.jstor.org/stable/1913610
Newey, W. K., & West, K. D.: Automatic lag selection in covariance matrix estimation. The Review of Economic Studies, 61(4), 631–653 (1994). https://www.jstor.org/stable/2297912
Nikkinen, J., Peltomäki, J.: Crash fears and stock market effects: Evidence from web searches and printed news articles. J. Behav. Finance (2019). https://doi.org/10.1080/15427560.2019.1630125
Nofsinger, J.R.: Social mood and financial economics. J. Behav. Finance 6(3), 144–160 (2005). https://doi.org/10.1207/s15427579jpfm0603_4
Ozik, G., & Sadka, R.: Media and Investment Management (2012). https://ssrn.com/abstract=1633705 or https://doi.org/10.2139/ssrn.1633705
Ozsoylev, H.N., Walden, J.: Asset pricing in large information networks. J. Econ. Theory 146(6), 2252–2280 (2011). https://doi.org/10.1016/j.jet.2011.10.003
Pew Research Center.: News use across social media platforms 2018. Pew Research Center (2018). Retrieved from https://www.journalism.org/wp-content/uploads/sites/8/2018/09/PJ_2018.09.10_social-media-news_FINAL.pdf
Plumlee, M.A.: The effect of information complexity on analysts’ use of that information. Account. Rev. 78(1), 275–296 (2003). https://doi.org/10.2308/accr.2003.78.1.275
Qian, H.: Time variation in analyst optimism: An investor sentiment explanation. J. Behav. Finance 10(3), 182–193 (2009). https://doi.org/10.1080/15427560903167753
Ramnath, S., Rock, S., Shane, P.B.: Financial analysts’ forecasts and stock recommendations: A review of the research. Found. Trends Finance 2(4), 311–421 (2008). https://doi.org/10.1561/0500000023
Sadka, R., Scherbina, A.: Analyst disagreement, mispricing, and liquidity. J. Finance 62(5), 2367–2403 (2007). https://doi.org/10.1111/j.1540-6261.2007.01278.x
Schmierbach, M., Oeldorf-hirsch, A.: A little bird told me, so I didn’t believe It: Twitter, credibility, and issue perceptions. Commun. Quat. 60(3), 317–337 (2012). https://doi.org/10.1080/01463373.2012.688723
Seybert, N., & Yang, H. I. (2012). The party’s over : The role of earnings guidance in resolving sentiment-driven overvaluation. Management Science, 58(2), 308–319. https://www.jstor.org/stable/41406390
Shiller, R.J.: Narrative economics. Am. Econ. Rev. 107(4), 967–1004 (2017). https://doi.org/10.1257/aer.107.4.967
Subrahmanyam, A., Sorescu, S.M.: The cross-section of analyst recommendations. J. Financ. Quantitative Anal. 41(1), 139–168 (2006). https://doi.org/10.2139/ssrn.590123
Tetlock, P.C.: Giving Content to Investor Sentiment : The Role of Media in the Stock Market. J. Finance 62(3), 1139–1168 (2007)
Tetlock, P.C.: All the news that’s fit to reprint: Do investors react to stale information? Rev. Financ. Stud. 24(5), 1281–1512 (2011). https://doi.org/10.1093/rfs/hhq141
Tversky, A., Kahneman, D.: Judgment under uncertainty: Heuristics and biases. Science 185(4157), 1124–1131 (1974). https://doi.org/10.1126/science.185.4157.1124
Twitter Twitter announces first quarter 2019 Results. Twitter (2019a) Retrieved from https://s22.q4cdn.com/826641620/files/doc_financials/2019/q1/Q1-2019-Earnings-Release.pdf
Twitter Twitter for business. Twitter (2019b). Retrieved from https://business.twitter.com/
Vosoughi, S., Roy, D., Aral, S.: The spread of true and false news online. Science 359(6380), 1146–1151 (2018). https://doi.org/10.1126/science.aap9559
Welch, I.: Herding among security analysts. J. Financ. Econ. 58(3), 369–396 (2000). https://doi.org/10.1016/S0304-405X(00)00076-3
Xu, N., Jiang, X., Chan, K.C., Wu, S.: Analyst herding and stock price crash risk: Evidence from China. J. Int. Financ. Manage. Account. 28(3), 308–348 (2017). https://doi.org/10.1111/jifm.12062
Acknowledgments
I would like to thank the two anonymous referees and the editor for their helpful comments and suggestions.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix 1
Rights and permissions
About this article
Cite this article
Garcia, J. Analyst herding and firm-level investor sentiment. Financ Mark Portf Manag 35, 461–494 (2021). https://doi.org/10.1007/s11408-021-00382-8
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11408-021-00382-8