Post Earnings Announcement Drift: A Summary and Critical Discussion.
1.0 Introduction
The Post Earnings Announcement Drift (PEAD) anomaly is an anomaly resulting from the ability of market participants to predict future abnormal returns by using information that is contained in past earnings announcements (Brown and Pope,1996). The long existence and the great robustness of the PEAD anomaly violates the semi-strong EMH which states that all public information is fully and instantaneously reflected in the share price (Soffer and Lys,1999). Hence, the PEAD anomaly provides investors with opportunities to earn abnormal profits (Shleifer,2000).
In this paper, an overview of the PEAD phenomenon and a discussion of its length are first given. Possible explanations for the PEAD, particularly, the delay in price response and misspecification of CAPM, are then investigated by examining the supporting and opposing arguments.
2.0 PEAD Anomaly: an Overview
Ball and Brown(1968) first established the existence of a PEAD anomaly through recognizing that following earnings announcements, the estimated cumulative abnormal returns persistently drifted up in relation to firms announcing good news and persistently drifted down in relation to firms announcing bad news (Bernard and Thomas,1989). This is supported by Foster, Olsen and Shevlin(1984) (FOS thereafter), Freeman and Tse(1989), and Bernard and Thomas(1989) who recognized that for a time period of up to four years, a positive correlation exists between the estimated post announcement abnormal returns and the estimated unexpected earnings.
As illustrated by Figure 1, by plotting the cumulative abnormal returns for 10 portfolios with diverse earnings news over 120 days[1] where each portfolio consists of firms that were assigned on the basis of their standing level in relation to the unexpected earnings level (scaled by the standard deviation of previous forecast errors) and the cross sectional distribution of the scaled unanticipated earnings over the previous quarter, one can easily detect the presence of the PEAD anomaly (Bernard and Thomas,1989).
Figure 1 shows that the estimated post earnings announcement abnormal returns diverge monotonically in accordance to the Standardized Unexpected Earnings (SUE) deciles. The specific study conducted by Bernard and Thomas(1989) reveals that over the 60 trading days following the earnings announcement, undertaking a position that is long in the 10th portfolio[2] and short in the 1st portfolio[3] yields approximately 6.31% in abnormal returns.
The studies conducted by Ball and Brown (1968), FOS (1984), Freeman and Tse (1989), McWilliams (1966), Joy, Litzenberger, and McEnally (1977), Latane and Jones (1979), Abarbanell and Bernard (1992) and Collins and Hribar (2000) all conclude that the PEAD anomaly is existent and market participants can earn abnormal returns on consistent basis by exploiting it. Nevertheless, although the anomaly was generally significant in all the studies, the extent of its significance in each study differs.
3.0 PEAD: For How Long?
According to Bernard and Thomas(1989), the majority of the drift takes place in the first 60 days following the announcement of earnings. The drift lasts for approximately 9 months for small firms and around 6 months for large firms Bernard and Thomas(1989). This is supported by Watts(1978) whose study established that the drift lasts for 6 months for large firms. Bernard and Thomas(1989) found that a large amount of the drift that takes place in the first 60 days following the announcement of earnings occurs within the first 5 days following the announcement of earnings. In fact, Bernard and Thomas(1989) established that the percentage of the drift that occurs within the first 5 days following the announcement of earnings in relation to the drift that takes place in the first 60 days following the announcement of earnings is 13% for small firms, 18% for medium firms, and 20% for large firms.
4.0 Possible Explanations: Delay in Price Response and Misspecification of CAPM
4.1 .1 Arguments Supporting the Delay in Price Response Explanation
It is clear that there is a delay in the price response which can be due failure to instantaneously incorporate information to the share price, or due to the costly information processing (Bernard and Thomas,1989). This delay causes the semi-strong form of EMH to be violated. This is because of the fact that despite the large number of well-documented research conducted explaining the PEAD anomaly, the PEAD anomaly has continued to persist for well over 33 years in which prices continued to adjust long after earnings were announced (Asthana,2003)(Lev and Ohlson,1982)(Bartov,1992). If the semi strong EMH was to hold, then all the abnormal profits arising from exploiting such anomaly would have been negated long time ago, drawing an end to the persistence of such anomaly (Bidwell and Riddle, 1981)(Bartov,1992).
Despite the fact that explaining the reason behind the market’s failure to respond instantaneously to earnings announcement has proven to be of great difficulty, amongst the possible reasons are the transaction costs[4] and the influence of investors’ action in not fully inferring the implications of the earnings information announced (Bernard and Thomas,1989). The incomplete market response is also supported by (Kormendi and Lipe,1987).
The transaction cost argument is confirmed by (Bernard and Thomas,1989) study which established that the PEAD did not exceed an upper bound that is roughly equal to the transactions costs incurred by an individual investor. This holds even when adjusting for the firm’s size.
4.1.2 Arguments Questioning the Delay in Price Response Explanation
Although there is a general consensus on the fact that the delay in price response explains the PEAD, Bernard and Thomas(1989) raise some issues that question the extent of the ability of the delay in price response to solely explain the PEAD. These issues include raising inquiries in relation to the reason behind the large concentration of the PEAD around the following quarter’s earnings announcement when transactions costs are the reason for the PEAD. Also, other unresolved issues include the non-elimination of the PEAD by traders with insignificant transaction costs and the non-movement of the price by market makers to the theoretical level from the first trade following the announcement of earnings (Bernard and Thomas,1989).
4.2.1 Arguments Supporting the Misspecification of CAPM Explanation
Ball, Kothari, and Watts(1988) state that many PEAD studies assumed a stationary beta, which results in an upward bias in the extent of the abnormal returns generated by exploiting the PEAD anomaly. However, allowing for betas to shift annually results in an insignificant PEAD (Ball, Kothari, and Watts,1988).
FOS(1984) found that using the Security Return Model (SRM) method[5] instead of the Earnings Based Model (EBM) method does not suggest the existence of a PEAD. Bernard and Thomas(1989) show that the PEAD that is identified when the EBM is used reflects an unidentified risk premium. This is because one of the problems that the SRM aims to mitigate is the risk adjustment problem of the EBM (Bernard and Thomas,1989).
Other supporting arguments include CAPM’s failure to incorporate market imperfections in which one example is the non-incorporation of taxes.
4.2.2 Arguments against the Misspecification of the CAPM Explanation
Bernard and Thomas(1989) does not support the conclusion that misspecification of the CAPM explains the PEAD anomaly as the findings of their study is inconsistent with Ball, Kothari, and Watts(1988) finding in terms of the stationary betas causing a significant upward bias in abnormal returns. For instance, Bernard and Thomas(1989) established that the size of the estimated beta shifts in their study was only around 8% as great as would be required for a full explanation of the drift’s magnitude.
Additionally, Bernard and Thomas(1989) findings are inconsistent with the explanation that PEAD is a result of excluding risk factors apart from the systematic risk. For example, they found a disproportionate amount of drift based around the earnings announcement of the subsequent quarter.
The non incorporation of taxes argument is also rejected by Bernard and Thomas (1989) who found that incorporating the effect of taxes would only make an insignificant difference that does not largely explain the PEAD (Bernard and Thomas, 1989).
5.0 Conclusion
In conclusion, there is a remarkable consistency in the results of research that suggest the existence of the PEAD. It is well known that the PEAD anomaly has survived even the toughest of checks, to the extent that it was referred by Fama(1998) as being the sole recognized anomaly that is above suspicion.
Explaining the PEAD phenomenon is a controversial issue. However, there is a general agreement that the PEAD phenomenon can be explained by the delay in price response and the misspecification of CAPM. While there is a general consensus that the delay in price response is a good explanation, the misspecification of CAPM explanation is subject to rejection by some researchers.
References List
Abarbanell, J. and Bernard, V.(1992). Tests of analysts’ overreaction/underreaction to earnings information as an explanation for anomalous stock price behavior. The Journal of Finance 45:3, pp.1181-1207.
Asthana, S.(2003). Impact of information technology on post-earnings announcement drift. Journal of Information Systems 17:1, pp.1-17.
Ball, R. and Brown, P. (1968). An empirical evaluation of accounting income numbers. Journal of Accounting Research 6:2, pp.159-178.
Ball, R., Kothari, S., and Watts, R. (1988). “The Economics of the Relation Between Earnings Changes and Stock Returns.” Working paper, University of Rochester.
Bartov, E. (1992). Patterns in unexpected earnings as an explanation for post-announcement drift. The Accounting Review 67:3, pp.610-622.
Bernard, V. and Thomas, J. (1989). Post-earnings announcement drift: delayed price response or risk premium? Journal of Accounting Research 27 (Supplement), pp.1-36.
Bidwell, C. and Riddle, J. (1981). Market inefficiencies–opportunities for profits. Journal of Accounting Auditing & Finance 4:3 (Spring), pp.198-214.
Brown, S. And Pope, P. (1996). “Post-Earnings Announcement Drift?” Working Paper. New York University.
Collins, D. and Hribar, P. (2000), Earnings-based and accrual-based market anomalies: one effect or two? Journal of Accounting and Economics 29, pp.101-123.
Fama, E. (1998). Market efficiency, long-term returns, and behavioural finance. Journal of Financial Economics 49:3, pp.283-306.
Foster, G., Olsen, C. and Shevlin,T. (1984), Earnings releases, anomalies, and the behaviour of securities returns, The accounting review 59, pp.574-603.
Freeman, R. and Tse, S. (1989), The multi-period information content of accounting earnings: confirmations and contradictions of previous earnings reports, Journal of
Accounting Research 27 (Supplement), pp.49-79.
Joy, O., Litzenberger, R. and McEnally, R. (1977). The adjustment of stock prices to announcements of unanticipated changes in quarterly earnings. Journal of Accounting Research 15:2, pp. 207-225.
Kormendi, R. and Lipe, R. (1987). Earnings Innovations, Earnings Persistence, and Stock Returns. Journal of Business, pp.323-46
Latane, H. and Jones, C. (1979). Standardized unexpected earnings–1971-77. The Journal of Finance 34:3, pp.717-724.
Lev, B. and Ohlson, J. (1982). Market-based empirical research in accounting: a review, interpretation, and extension. Journal of Accounting Research 20 (Supplement), pp.249-322.
McWilliams, J. (1966). Prices, earnings, and P-E ratios. Financial Analysts Journal 22:3, pp.137-142.
Shleifer, A. (2000). “Inefficient markets: an introduction to behavioral finance”. Oxford University Press, US
Soffer, L. and Lys, T. (1999). Post-earnings announcement drift and the dissemination of predictable information. Contemporary Accounting Research 16:2, pp.305-331.
Watts, R. (1978). Systematic “abnormal” returns after quarterly earnings announcements. Journal of Financial Economics 6, pp.127-150.
[1] 60 days before the earnings announcement date and 60 days after the earnings announcement date
[2] Portfolio with the maximum unexpected earnings
[3] Portfolio with the minimum unexpected earnings
[4] For example, the bid-ask spread, short-selling costs and brokerage commissions.
[5] Under this method, firms are assigned to portfolios based on the firms’ projected abnormal returns over the period starting 60 days from the earnings announcement until the announcement day