Can Accounting-Information Based Strategies Generate Abnormal Return?
Efficient Market Hypothesis is one of the most important theories in investigating the relationship between abnormal stock return and accounting information. As Fama (1991) stated, if semi-strong EMH holds, any fundamental strategies could earn a zero expected abnormal return as stock price has already incorporate all publicly available information including the accounting announcement. Reinganum (1981, pp.24) uses data from 566 firms spanning 1975-1977 and illustrated that “abnormal returns cannot be earned over the period studied by constructing portfolios on the basis of firms standardized unexpected earnings.”However, several scholars hold various theoretical and empirical researches to doubt Fama's (1991) statement.
Rendleman et al. (1982), Joy and Jones (1979) regards the post-announcement drift as a critical anomaly contrasting EMH. Moreover, SUE strategies based on accounting information is proved to be able to capture abnormal return.
Numerous literatures are carried on to explain the drift anomaly and the profitability of accounting information based strategy. Meanwhile, academics formulated and develop amounts of models including simulating trading strategy analysis, earning-based model along with balance sheet associated models to test the impact of accounting information on stock price.
This essay is an attempt to analysis the linkage between accounting information and equity return by the means of currently theoretical and empirical literatures in this area. As the prelude to this essay, several hypotheses as long as their empirical evidence concerning the usefulness of accounting information in generating excess return are reviewed. Then this essay turns to the discussion of the development of methodology in testing this relationship.
Review Of The Evidence
Efficient Market Hypothesis
Weak-form EMH indicated that accounting information is useless since stock price adjusts quickly to publicly available information has already been used to exploit excess return and reflected in equity price.(Fama, 1991) EMH stems from the arbitrage activities of considerable profit-maximizing investors. Rational investors would compete to construct trading strategies based on all publicly relevant information to exploit abnormal return. The highly competitive arbitrages account for the immediate reaction of stock price to accounting announcement. Furthermore, on account of the stochastic characteristic of new released accounting data, stock return at announcement day is perceived as a random walk, (Malkial, 2003) therefore, any techniques based on accounting information would be useless.
Fama (1998) adopted a series of short-run “event studies”, and suggested that stock price responds quickly to the “events” containing accounting changes and earnings announcement. The same evidence is provided by Ball and Kothari (1991), who also employed “event study” and delved into the cumulative abnormal returns (CAR) around the announcement day in an effort to test the market efficiency. They assert that substantial abnormal return only occurs at the announcement. Similarly, Beaver (1968, pp.74) postulated that investors respond to the earnings announcement rapidly and thus there is no delay in reflection earning information in stock price. He found that almost all the abnormal trading volume occurred in the announcement week.
However, the prolonged post announcement drift challenges EMH, Jones et al.(1985) found “31% of price change occurs before announcement, 18% on the announcement day, and as much as 51% after the announcement.” Simultaneously, Foster et al. (1984) documented the existence of post announcement drift as the CAR of stocks with positive surprise exhibit an upward trend after the announcements. Meanwhile, considering the manipulation of earnings, Jegadeesh and Livnat (2006) focus on the impact of revenues announcement on stock price and get the conclusion that revenue surprises and abnormal returns on the calendar day and several days after the announcement co-integrated a positive relationship. These evidences suggest that accounting announcement carried information which is not already available to the market. And thus trading techniques such as Standardized Unexpected Earnings (SUE) strategy, which is long stock with positive SUE and short stocks with negative SUE, could earn abnormal return. (Bernard and Thomas ,1990)
Nevertheless, the excess earnings from SUE strategies are thought to be a compensation for bearing high risk. (Fama, 1991)
The test of EMH is joint test of risk-adjusted return models. The anomaly is due to weakness of CAPM rather than market inefficiency. The abnormal return available from post-announcement drift is a result of flawed proxy of “earnings and dividends for omitted variables” (Ball, 1978, pp.111) such as firm size (Banz, 1981). Fama and French (1996) documented that drift anomaly can be explained by the three-factor APT model as manifestations of risk premiums.
Uncertainty Resolution Hypothesis
Ball and Brown (1968, pp.316) employed the Hypothesis of Uncertainty Resolution to explicit the linkage between stock return and accounting announcements. They demonstrated that the disclosure of corporate reporting reduces the uncertainty about the firm's information such as future cash flow which is the determinant of stock price in Discounted Cash-flow Model. Thus, the resolution of uncertainty would result in the increase in equity return variance. As accounting announcement would not influence the variance in market portfolio return, covariance of the company's stock return and market portfolio tends to go up in event time, which trigger the upward trend in CAPM beta and thus enhance the expected value of the stock. Similar explanations are presented by subsequent studies of Salaman and Choi (1989), Subrahmanyam and Stapleton (1979). This hypothesis is consistent with Ball and Kothari's (1991, pp. 722) empirical evidence that significant rise in beta is detected.
Considering the transaction cost, the SUE strategies cannot bring investors excess return. Watt (1978) uses data from 73 firms covering the period 1962-1968 and indicated that SUE trading techniques cannot exploit extra profit for the large transaction cost.
Behaviour Finance Hypothesis
Nevertheless, the received view in Behavior finance is that accounting information is useful in exploit abnormal return on account of irrational investors and limits to arbitrage.
Naïve expectation Hypothesis
Rendleman et al. (1987, pp.142) firstly introduced the “Naïve expectation hypothesis” in which stock price cannot adjust immediately to its intrinsic value on the announcement day as some “naïve” investors fixate on the variance of earnings and ignore the serial correlation in equity earnings. Bartov and Ball's (1996) study shows that the extent to which investor's expectation deviates from the serial correlation of share returns is amounts to 50%. Rendleman et al. (1987, pp.142) also documented that strategies with sophisticate models can generate abnormal return. Corroborating evidence is presented by Bernard and Thomas (1990, pp.307), who implemented an empirical test using 85，000 quarterly return announcements data spanning 12 years. Their result evidenced that the abnormal return exists and lasts for three quarters after the announcements and the excess return dovetails with the prediction of serial correlation. Furthermore, they proved the profitability of SUE strategy.
Edwards (1968), Bodie et al. (2005) advocated that the market is inefficient in reflecting all publicly accounting information and traced the announcement drift to conservatism. According to Bodie et al. (2005), investors prone to revise their initial beliefs slowly and make insufficient adjustment. Barberis et al (1998) constructed a model to explain this hypothesis; they indicated that investors are conservative to belief in mean reversion and underreact to recent earnings surprises, then after a chain of surprise, investors would believe in the trend, and overreact to the accounting information, leading to the short-run momentum. A similar explanation is provided by Kahneman et al. (1982). They attributes the anomaly associated with accounting information to anchoring heuristic, in which people anchor too heavily on initial beliefs and underweighted new information. Daniel (1998) pointed out that investors are overconfident in their analysis and overreact to public signals such as corporate reporting. Moreover, self-biased attribution gives rise to further overreaction. As a result, stock return on announcement day would not adjust rapidly to its value and exhibit a post-earning-announcement drift.
Limits To Arbitrage
Ball (1992) considers the drift anomaly as a result of arbitrage costs. The high implementation costs of SUE technique discourage arbitrageurs from fully employing this trading strategy. Johnson and Schwartz (2000) delved into the impact of implementation costs on the post-1990 announcement drifts and documented that stocks with low information acquiring costs exhibits weak earnings drift. The empirical result is consistent with Ball's (1992) explanation. In De Long et al.(1990)'s model, knowledgeable investors tends to reluctant to exploit the arbitrage opportunity such as anomalous earnings at the announcement on account of the post announcement drift caused by Noise traders. The agency problem is deemed as another plausible reason for limits to arbitrage. (Shleifer and Vishny, 1997) Since investors focused their attention on the earnings and ignore the accrual components (Stem, 1974), a zero-investment strategy with a long position in low accruals stocks and short position in high accrual stocks could generate abnormal returns, however, the unattractive characteristics of these firms and high information-processing and strategy-executing costs impeded the implementation of accrual strategy. (Lev and Nissim, 2006)
Another alternative explanation of drift anomaly in Behavior Finance is information obscurity. (Merton, 1987) Brav and Heaton (2002) endorse the view that investors underestimate the stocks of firms with uncertainty information and modify previous valuation at the announcement day, giving rise to the abnormal return on the release of accounting information data. If so, anomalous variance in stock price is much more pronounced among obscure-information firms, this hypothesis is empirically documented by Zhang (2006).
Another alternative hypothesis in support of the usefulness of accounting information in acquiring excess return is information Hypothesis. According to Regan and Niederhoffer (1972), directors are prone to spread good news ahead of the release of corporate reporting which include negative information. Under such hypothesis, pre-announcement excess return is predicted to be positive while at-announcement abnormal return is supposed to be negative. Ball and Kothari (1991, pp. 722) employed CAPM to examine information hypothesis. However, the insignificant positive abnormal return contradicts this hypothesis.
Simulating Trading Strategy Analysis
Ou and Penman (1989) used 68 accounting ratio as a summary value measure to construct an investment strategy to predict the one year ahead earning changes from 1973 to 1983, and as the summary value measure is extracted from financial statement which is exposed to the public, this empirical analysis can be used to test whether accounting information is reflected in the stock price.They found that the strategy generated market adjusted excess return by 8.3% for 12 months and 14.5% for 24 months (Ou and Penman, 1989, pp.328)
Nevertheless, Holthausen and Larker (1992) replicated the model put forward by Ou and Pennman (1989) and found that their strategy worked well in the first subperiod from 1978 to 1982, meanwhile, it performed poorly subsequent to 1983, and the excess returns from the hedge portfolio for both 12 months and 24 months are negative. Holthausen and Larker pointed out that Ou and Penman's (1989) strategy is fragile for three reasons: firstly, the excess earning predicted by their strategy has weakly association with returns (Larcker, 1998). Secondly, Ou and Penman's model excluded accounting entries which “provide significant marginal explanatory power in the presence of other variables”. Thirdly, 8 ratios of the 68 accounting ratios contained a considerable number of missing data.
Considering these problems, Holthausen and Larker (1992) developed a new LOGIT model by adding over-the-counter firms to the sample and dropping 8 accounting ratios, they implemented an investment strategy based on the prediction of stock returns directly and used three excess returns metrics to test whether their strategy based on accounting information can yield 12-month abnormal returns over the 1978-1988 period.
Stock analysts devoted considerable attention to the earnings announcement analysis. But empirically examining the relationship between accounting statements and stock price is complicated, since it is difficult to form a complete equity valuation model with accounting data. (George,1967,pp.3)
The release of accounting statements can influence stock price is due to investors' activities. After receiving annual report, investors will adjust their expectation about the stock and decided to buy or sell the equity. The expectation variance can be measured by the simultaneous change in stock price. Thus, the impact of financial statements on stock price can be measured by the variance in the stock price. (George,1967,pp.3)
In the regression model, the elasticizes of most accounting data measuring net income is only around 0.02. Thus, George (1967,pp.22) concluded the earning is not informative in valuating equity price.
However, it is argued that the tested weak linkage between earning announcement and variance in stock price is due to the improper model George (1967) constructed. Beaver (1968,pp.70) utilized the George's (1967) methodology to formulate a model in order to the magnitude of stock price and volume change around the earnings announcement day of 143 firms' stock over the period 1961-1965. The result showed that both the trading volume and stock price changed dramatically on the release day of earning announcement. ( George,1967,pp.74)
Balance Sheet Associated Model
However, earning-based model failed to consider the accounting information in balance sheet, which explains how the earnings are generated. (Chen and Zhang, 2007, p.221) As demonstrated in Statement of Financial Accounting Concept (SFAC), “income statement can be interpreted most meaningfully….only if it is used in conjunction with a statement of financial position'' (FASB, 1978a, b). For example, book value showed on the balance sheet is an important index to judge whether the stock is mispricing.
Under the assumption of the persistence in the abnormal return, expected future earnings can be calculated through current earnings and anticipated abnormal return, which is defined as the difference between earnings and book value times risk-free interest rate.( Feleham and Ohlson,1995,pp.691-pp.692) As the stock price of a firm is closely related to the present value of expected cash flow and current assets value, Feleham and Ohlson (1995) derived the return model as a function of four variables form income statement such as operating earnings, and three variables including book value from balance sheet. (Feleham and Ohlson,1995, pp.691)
Through theoretical analysis, they identified stock price movements as a linear function of earnings yield and book value. (Feleham and Ohlson,1995, pp.721)
However, in Feleham and Ohlson's model (1995), all current investment activities is assumed as a “predetermined linear stochastic process” (Zhang, 2000, p.272) and have a NPV equal to 0, however, in reality, firms will choose projects with positive NPV and determine to contract or expand the scale of operation by making investment decision. Due to the convexity property of option, the equity price cannot be viewed as a linear function of stock price and earning yields.
Therefore, Zhang (2000) made a modification by adding “endogenous investment decision” variable, which is the value of option to decide whether invest on a project, in Ohlson's (1995) model. Moreover, he assorted sample firms in terms of operating efficiency and potential growth, and analysed the relationship between accounting variables and stock price based on this category. (Zhang, 2000, pp.272) Utilizing the model above, Zhang (2000, pp.292-pp.293) stated that the stock return with endogenous investment is a convex function of earnings and book value.
However, previous studies have not formed a model using accounting variables to explain the stock price movements with high R2.
Based on Zhang's research (2005), Chen and Zhang (2007) developed a more comprehensive model to explain stock price movements via accounting variables, the explanatory variables in their model comprise variance in discount rate and four factors influencing cash flow, as stock return associates with cash flow and discount rate. Considering accounting factors related with cash flow, they identified earning yields, a measurement of present return, as one of the independent variable. At the same time, expected future return depends on the operating efficiency and operating scale: the former can be represented by changes in profitability; the latter could be shown via capital investment, which measures the current operating scale, and changes in growth opportunities, a representative of expected future scale. (Zhang and Chen, 2007, pp.220)
Thus, Chen and Zhang's model (2007) included present earning level (reflected by earning yields), earning variance (reflected by change in profitability), accounting information in balance sheet (reflected by investment capital) and the surrounding financial environment (reflected by growth opportunity and discount rate).(Zhang, 2007,pp.222) Most accounting factors mentioned in previous research are incorporated in their model.
Also, they estimated the model with pooled cross sectional sample from Compustat over 1983-2001 periods. The R2 of the regression is 17.4%, much more significant than the R2 of less than 10% in earning based model, that is, Chen and Zhang's model (2007) holds more explanatory power in the observed stock price movements than other models.
This paper reviews the opinions of numerous theoretical and empirical literatures. According to Fama(1991), Semi-strong implies the uselessness of accounting information as stock price adjusts simultaneously with the arrival of new information. Empirical evidence provided by Ball and Kothari(1991), Bear(1968) is consistent with Fama's (1991) assertion. However, the pronounced post-announcement drift indicated the inefficiency of stock market. Moreover, Behaviour Financier traces the drift anomaly to the irrational characteristic of investors the inefficiency attributes of the security markets. If so, accounting information can be employed to generate abnormal return.
Meanwhile, this essay provides a clear scenario of the development of the methodologies in this area. The methodologies implemented in the studying the impact of accounting information on stock price could be categorized into three streams: Simulating strategy analysis, earnings-based model and balance sheet associated model.
Overall, it is empirically evidenced that accounting announcements carry information not available to the market and stock price adjust to its intrinsic value gradually. Therefore, accounting information equipped analyst with a useful weapon to exploit excess return. However, it is noteworthy that the comparability problems prevent the application of accounting information analysis for formulating trading strategies based on such information. The flexibility of General Accepted Accounting Principles (GAAP) guidelines in depreciation methods, valuation methods of R&D expenditures and adjusting methods in an inflation circumstance leads to the difficulty in both cross-sectional and over time comparison of financial statements. (Bodie et al, 2008,pp.597)
A valuable avenue for further study appears to be a theoretical or empirical research concerning the influence of incomparability of financial statement on the profitability of accounting-information based strategies.
1. Ball, R. “Anomalies in relationships between securities' yields and yield-surrogates.” Journal of Financial Economics 6 (1978): 103-126.
2. Ball, R. “The earnings-price anomaly.” Journal of Accounting and Economics 15 (1992), 319-346.
3. Ball, R. and E. Bartov. “How naïve is the stock market's use of earnings information?” Journal of Accounting and Economics 21 (1996), 319-337.
4. Ball, R. and P. Brown. “En empirical evaluation of accounting income numbers.” Journal of Accounting Research 6 (1968), 159-178.
5. Ball R. and S.P. Kothari, 1989, Nonstationary expected returns: Implications for tests of market efficiency and serial correlation in returns, Journal of Financial Economics 25, 51-74.
6. Ball, R., Kothari, S., 1991. Security returns around earnings announcements. The Accounting Review 66, 718-738.
7. Banz, R. W., "The Relationship Between Return and Market Value of Common Stocks," Journal of Financial Economics (March 1981), pp. 3-18.
8. Barberis, N., Shleifer, A., Vishny, R., 1998. A model of investor sentiment. Journal of Financial Economics 49, 307-343.
9. Beaver, W., 1968. The information content of annual earnings announcements. Journal of Accounting Research, pp. 67-92 .
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.
11. Bodie, A. Kane, and A. J. Marcus. Essentials of Investments. McGraw Hill Irwin, Boston, 5th edition, 2004.
12. Bodie, Z., A. Kane and A. J. Marcus (2008), Investments, Irwin, 567-610.
13. Brav & Heaton (2002) “Competing theories of financial anomalies”, Review of Financial Studies, 15, 575-606
14. Burton G. Malkiel, The Efficient Market Hypothesis and Its Critics , The Journal of Economic Perspectives, Vol. 17, No. 1 (Winter, 2003), pp. 59-82
15. Chen, P. and G. Zhang (2007), How do accounting variables explain stock price movements? Theory and evidence, Journal of Accounting and Economics 43, 219-244.
16. Choi, S., and G. Salamon. 1989. Accounting information and capital asset prices. Working paper, Indiana University, Bloomington.
17. Daniel, K., Hirshleifer, D., Subrahmanyam, A., 1998. Investor psychology and security market under- and overreactions. Journal of Finance 53, 1839-1885.
18. Delong, Shleifer, Summers, & Waldmann (1990) “Noise trader risk in financial markets”, Journal of Political Economy, pp.98.
19. Edwards, W. “Conservatism in human information processing.” In B. Kleinmuntz (Ed.) Formal Representation of Human Judgment (New York: Wile, 1968), 17-52.
20. Fama (1991) “Efficient capital markets: II”, Journal of Finance, 46, 1575-1617.
21. Fama (1998) ”Market efficiency, long-term returns, and behavioural finance”, Journal of Financial Economics, Volume 49, Issue 3, 1 September 1998, Pages 283-306.
22. Fama & French (1996) “Multifactor explanations of asset pricing anomalies”, Journal of Finance, pp. 1151-1168.
23. Feltham, G., Ohlson, J., 1995. Valuation and clean surplus accounting for operating and financing activities, Contemporary Accounting Research 11, 689-731.
24. Foster, G., C. Olsen and T. Shevlin, 1984, Earnings releases, anomalies, and the behavior of securities returns, The Accounting Review 59, 574-603.
25. George J. Benston, 1967, Published Corporate Accounting Data and Stock Prices, Journal of Accounting Research, Vol. 5, Empirical Research in Accounting: Selected Studies, pp. 1-56.
26. Holthausen, R. W and D. F. Larker (1992), The prediction of stock returns using financial statement analysis, Journal of Accounting and Economics 15, 373-411.
27. Jegadeesh, N. and J. Livnat (2006), Revenue surprises and stock returns, Journal of Accounting and Economics 41, 147-171.
28. Jones, C., R. Rendleman, H. Latané. "Earnings Announcements: Pre-and Post-Responses." Journal of Portfolio Management 11(Spring 1985):28-32.
29. Johnson, B. and W. Schwartz. 2000. Evidence that capital markets learn from academic research: earnings surprises and the persistence of post-earnings announcement drift. University of Iowa working paper.
30. Joy, O. M. and C. P. Jones, "Earnings Reports and Market Efficiencies: An Analysis of the Contrary Evi- dence," The Journal of Financial Research (Spring 1979), pp. 51-63.
31. Kahneman, D., P. Slovic and A. Tversky. Judgment Under Uncertainty: Heuristics and Biases (New York: Cambridge University Press, 1982).
32. Lev and Nissim (2006) “The persistence of accrual anomaly”, Contemporary Accounting Research, 23, 193-226
33. Merton (1987) “A simple model of capital market equilibrium with incomplete information”, Journal of Finance, 42, 483-510.
34. Niederhoffer, V., and P. Regan. 1972. Earnings changes, analysts' forecasts and trading volume. Financial Analysts Journal (May-June): 65-71. Penman, S. 1984. Abnormal returns to investment.
35. Ou, J. A. and S. H. Penman (1989), Financial statement analysis and the prediction of stock returns, Journal of Accounting and Economics 11, 295-329.
36. Reinganum, M. R., "Misspecification of Capital Asset Pricing: Empirical Anomalies Based on Earnings' Yields and Market Values," Journal of Financial Economics (March 1981), pp. 19-46.
37. Rendleman, R.J., C.P. Jones and H.A. Latané, 1987, Further insight into the standardized unexpected earnings anomaly: Size and serial correlation effects, Financial Review 22, 131-144.
38. Rendleman, R. J., C. P. Jones and H. A. Latane, "Empirical Anomalies Based on Unexpected Earnings and the Importance of Risk Adjustments," Journal of Financial Economics (November 1982), pp. 269- 287.
39. Shleifer and Vishny (1997) “The limits to arbitrage”, Journal of Finance, 52.
40. Stapleton, R., and M. Subrahmanyam. 1979. Multiperiod equilibrium: Some implications for capital budgeting. TIMS Studies in Management Sciences 11: 233-48.
41. Stern, J., “Earnings per Share Don't Count, Financial Analysts Journal, July-August, 1974, 39-43.
42. Watts, R. L., "Systematic 'Abnormal' Returns After Quarterly Earnings Announcements," Journal of Financial Economics (June/September 1978), pp. 127-150.
43. Zhang (2006) “Information uncertainty and stock returns”, Journal of Finance, 61, 105-136.
44. Zhang, G. (2000), Accounting information, capital investment decisions, and equity valuation: theory and empirical implications, Journal of Accounting Research 38, 271-295.