|
Article Excerpt
ABSTRACT
Post-earnings announcement drift is among the most persistent market anomalies. Two possible causes of the drift are transaction costs that dissuade investors from trading on earnings information immediately, and the market's inability to fully interpret the implications of earnings information. This study hypothesizes that information technology advances have significantly reduced the transaction costs of trading equity stocks and improved the informational efficiency of the capital market. These two changes should, therefore, reduce drift.
Using a sample of over 27,000 firm-quarter observations for more than 1,600 firms, the results support the hypothesis that drift has declined significantly with the growth of information technology. The results are robust after controlling for factors correlated with time, changes in the value-relevance of earnings, and previously identified variables that affect the drift, such as size, investor sophistication, and magnitude and sign of analysts' forecast errors. The study provides evidence that the information technology revolution may reduce trading friction and improve informational efficiency.
I. INTRODUCTION
Post-earnings announcement drift (henceforth, PAD) is the predictability of abnormal returns after the earnings have been announced. Ball and Brown (1968) initially observed the drift. Several subsequent studies have examined the phenomenon in great detail. (1) This anomaly has persisted for over 33 years despite extensive research to explain it. According to Kothari (2001), the PAD anomaly "poses a serious challenge to the efficient market hypothesis." Transaction costs and market inefficiency were two factors commonly identified as the causes for PAD (Bhushan 1994; Bernard and Thomas 1989; Freeman and Tse 1989, among many others). This study posits that the rapid growth that took place in the field of information technology (henceforth, IT) during the '90s reduced PAD as a result of decreased transaction costs and increased informational efficiency.
Advances in information and communication technologies have caused the cost of information search, production, storage, and dissemination to decline dramatically over the last decade. This has profound impact on both the institutional and informational structure of capital markets (Guldimann 2000). Thus, IT growth has improved the informational efficiency of the equity market. (2) During the last decade, competition from several alternate Internet sources (such as Ameritrade, E*Trade, Charles Schwab, Trading Direct, etc.) has also significantly driven down the transaction costs of trading. Based on a sample of over 27,000 observations from Compustat, CRSP, IIB/E/S, and COMPACT DISCLOSURE databases, the study reports evidence consistent with the hypothesis that the IT revolution has reduced PAD.
Prior research shows that the magnitude of the drift is directly related to the magnitude and sign of unexpected earnings, and the absolute magnitude of the drift is inversely related to firm size and investor sophistication (Foster et al. 1984; Bernard and Thomas 1989; Bartov et al. 2000). The research design in the current paper, therefore, controls for firm size; magnitude, and sign of the forecast error, and investor sophistication. In addition, recent research (Landsman and Maydew 2001; Chang 1999; Francis and Schipper 1999; Collins et al. 1997; Lev 1997; Ely and Waymire 1996; Ramesh and Thiagarajan 1995) finds that the value relevance/information content of earnings has changed with time. The research design in this study also controls for this factor, since the earnings quality can affect the market's efficiency in absorbing new information. To capture the effect of IT, the study develops a proxy based on the criteria used by the Statistical Abstracts of the United States (U.S. Census Bureau 2000). Uti lizing portfolio, correlation, and regression analyses, the study documents evidence consistent with PAD decreasing with the advancement of IT.
The findings in this study are important because they add to our understanding of PAD. The study presents evidence that the anomaly is weakening with the progress in IT. Indirect evidence is also presented that trading frictions are decreasing and the informational efficiency of the market is increasing with the improved technology. The results should be of particular interest to proponents of market efficiency.
The rest of the study is organized as follows. The next section discusses the literature and builds the hypothesis. The third and fourth sections present the research design and sample selection procedure, respectively. The fifth section explains the empirical results, followed by the conclusion in the sixth section.
II. EXISTING RESEARCH AND DEVELOPMENT OF HYPOTHESIS
Ball and Brown (1968) were the first to report the PAD phenomenon and they suggest that one of the explanations for the delayed adjustment to earnings information is that the value of the information is less than the transaction costs that would be incurred by the investor who wished to take advantage of the opportunity for abnormal gain.
Foster et al. (1984), who document PAD over 60 trading days after the announcement, examined PAD in detail. The drift is inversely related to size and positively related to the sign and magnitude of the unexpected earnings. The sign and magnitude of the unexpected earnings explain approximately 80 percent of the variation in cumulative abnormal returns and approximately 65 percent of the variation is explained by firm size. Subsequently, Bernard and Thomas (1989) document that most of the drift occurs during the first 60 trading days (about three months) subsequent to the earnings announcement. A disproportionately large amount of the 60-day drift occurs within five days of the earnings announcement. They also show that the magnitude of the drift is related to the magnitude of unexpected earnings and the absolute magnitude of the drift is inversely related to firm size. They find some evidence that transactions/opportunity costs exceed gains from immediate exploitation of information for a sufficiently large number of traders. Their results are also consistent with a market that fails to recognize the full implications of current earnings for future earnings. (3)
Prior research has used time-series properties of earnings to build the earnings expectations. Since earnings forecasts by security analysts are more accurate than time-series forecasts, Freeman and Tse (1989) examine PAD using analyst forecast errors. They find that PAD is actually 50 percent larger when analysts' forecasts are used to measure unexpected earnings in comparison to time-series forecasts. Bhushan (1994) presents evidence that the magnitude of the post-earnings announcement drift is positively related to the direct and indirect cost of trading. The relation subsumes the previously documented inverse relation between drift and firm size. More recently, Bartov et al. (2000) show that investor sophistication, proxied by institutional investor holdings, is negatively correlated with the observed post-announcement abnormal returns.
From a reading of the extant research discussed above, it is clear that two of the main reasons thought to be responsible for PAD are the constraints imposed by transaction costs and the market's failure to comprehend the implications of time-series properties of unexpected earnings. With the dawn of the Internet era in the '90s, several convenient sources of security trading (such as, Ameritrade, E*Trade, Charles Schwab, Trading Direct, etc.) have opened up. With the sudden expansion in the suppliers of electronic trade, the vast increase in the availability and accessibility of computers almost everywhere--in homes, trains, cars, cellular phones, even airplanes, the cost of trading securities has been greatly reduced. According to Guldimann (2000), brokers traditionally charged around $300 for a 5,000-share transaction on the exchange (that is, six cents per share). Today, an investor can trade for a mere fraction of this cost on the Internet. For example, E*Trade charges $14.95 for trading 5,000 shares of s tocks listed on major exchanges. The charge progressively decreases as the number of trades increases. These significant reductions in transaction costs should enable traders to cash in on earnings information faster, thereby reducing PAD.
The technology revolution of the '90s made significant strides in the fields of data storage, availability, and exchange. (4) While the Internet can trace its existence to ARPANET in 1969, it took off as a commercial tool with the development of the World Wide Web in 1991. The number of websites exploded with the development of Mosaic (predecessor of Netscape) from 600 in 1993 to over one million in 1997. Today, even very small companies can set up websites to provide information to present and potential investors. News wires and investment-related sites can be searched from home or office 24 hours a day to find information about companies. The number of users of these sites has increased dramatically over time. For example, the number of subscribers to America Online increased from 6.2 million as of June 30, 1996 to 23.2 million as of June 30, 2000. (5) Yahoo! reports that it provides services to more than 120 million users each month worldwide. (6) The number...
|
