Experiments on the Post-Earnings-Announcement Drift

Public companies periodically release earnings announcements to inform investors and the public about their profitability. Classical finance theory predicts that investors in the market react immediately, causing stock prices to instantly adjust to reflect the information. However, stock prices in practice exhibit post-earnings-announcement drift (PEAD), meaning that firms stock prices seem to adjust only slowly to the news contained in earnings announcements. Stocks with positive news remain under- and stocks with bad news overvalued for as long as 6 months. While finance researchers have proposed a number of different explanations over the years, no clear consensus has yet emerged on what causes the PEAD. Many other market anomalies disappeared after their discovery, since sophisticated investors traded to profit from them, yet the PEAD persists since the 1970s and investors can exploit it to achieve excess profits. This makes it one of the most important anomalies to study. Previous research into PEAD relied on statistical analyses of real-market data to measure mispricing after earnings announcements. In contrast, this study investigates the PEAD using economic experiments, where subjects trade in an artificial market on a computer, imitating the process of trading in real life. An experimental approach is ideally suited for the exploration of PEAD, because the researcher controls the trading environment. Consequently, the researcher can examine different hypotheses, derived from potential explanations for the PEAD in the literature. An experimental approach furthermore allows for separating the PEAD from other occurrences in the market, which may also affect prices. While experimental finance research has grown strongly in recent years (with 2017s Nobel prize for Richard Thaler being only the most visible signal), few experiments touch upon the PEAD. Furthermore, none of them directly tests the explanatory factors proposed in the literature. This project will first establish more insights into the appearance and characteristics of the PEAD in a baseline setting. In study 1, it will then introduce earnings correlation and transaction costs as well as short-selling restrictions to probe their influence on the PEAD. In study 2, the project will explore whether a high frequency trader, a robot trader who corrects certain mispricings, is able to eliminate the PEAD. In study 3, the project will consider the impact of the timing of the announcement. Overall, the aim of the project is to provide a better understanding of investors behavior after earnings announcements and of the causes for the PEAD, by transferring it into an experimental setting. The described advantages of the proposed experiment will help to evaluate how well correlation in earnings can explain, how trading constraints affect, and whether changes in the timing of earnings announcements moderate the PEAD.

Corporations regularly report about their firm's performance in so-called "earnings announcements". Since these announcements typically contain new information that is not yet known to investors in capital markets, the firms' stock prices should quickly adjust to these news. Previous research finds, however, that prices tend to adjust slowly. After a positive earnings announcement, stock prices tend to drift upward (and after a negative announcement downward) for up to one year. This empirical phenomenon is referred to as the post-earnings-announcement drift (PEAD). While many researchers have studied the PEAD, no consensus has emerged so far. The PEAD is still considered a puzzle, despite its great relevance to both economic theory and investment practice. A first important issues in this context is that the earnings that firms announce are not independent from one announcement to the next. In other words, if firms announce higher-than-expected earnings in one quarter, they are more likely than not to again beat expectations in the next quarter (and vice versa). Some studies conjecture that the PEAD may be caused by investors not properly accounting for this pattern. A second important issue is the question of announcement timing. Firms can choose when to make their earnings announcements. Specifically, they can choose whether to announce before, during or after exchange opening hours. This timing decision may also be tied to the extent of the observed PEAD. This project used an entirely new methodological approach to disentangle potential causes for the PEAD that had been proposed in the literature. The researchers invited volunteers to computer labs to study their trading behavior in simulated markets. This method allows researchers to study phenomena under controlled conditions, and to clearly document causal relationships. Several pioneers of experimental research in finance and economics have received the Nobel Memorial Prize in Economics for their contributions to economic research. The results of the project show that the PEAD is significantly strengthened when earnings news are dependent over time, but that this dependence is not the original cause of the drift. PEAD is furthermore greater in the presence of trading frictions and certain algorithms that are designed to help make the market work better do not succeed in reducing PEAD. Finally, the project for the first time provides detailed evidence on earnings announcement timing decisions in the equity markets of France, Germany, Japan and the UK. The project's findings are of interest to investors and investment professionals, who already exploit the PEAD to earn extra profits. They are also relevant for regulators, who receive new insights into earnings announcements, particularly in the German market.