Information Uncertainty and the Post–Earnings Announcement Drift in Europe
Summary (4 min read)
1. Introduction
- It has recently been noted that some important non-earnings information is also released at the time of the announcement.
- Last but not least, the authors investigate the role played by information uncertainty in explaining the pay-offs associated with the abnormal volume and the abnormal return effects.
- Therefore, as with the abnormal return effect, the authors find that firms which surprise the market in one quarter continue to do so one year later.
- The next section discusses the relevant literature and motivates their analysis.
2. Literature Review and Motivations
- It is also obvious that some important non-earnings related news are released at the time of an announcement.
- Finally, a number of recent empirical findings suggest that market participants incorrectly value non-financial information.
- In line with earlier findings, Brandt et al (2006) report a strong postearnings announcement drift when ranking stocks on the basis of their abnormal return at the time of the earnings announcement.
- Moreover, Zhang (2006) demonstrates that investors under-reaction to public information is even more pronounced in cases of greater information uncertainty.
3.1 Sample Description
- The universe for their study consists of exchange listed firms belonging to the FTSE All-World Developed Europe Index from 1997 to 2010.
- Interim report dates comprise semi-annual announcements as well as announcements for quarter one to quarter three.
- Moreover, the authors use widely available 9 market data to compute their indicators of market surprise, so that the main constraint that they face in terms of data coverage is the availability of interim and fiscal year-end report dates.
- The authors also note some seasonal patterns in the relative and absolute coverage of announcement dates.
- The authors can see that the proportion of interim report dates is large since the beginning of their sample period.
3.2 Variable Definitions
- The authors first indicator of market surprise is the abnormal return recorded over a three-day window centred on the announcement date.
- It is the average daily share volume over a three-day window centred on the announcement date divided by the average daily share volume estimated over days -8 through -63.
TABLE 1 ABOUT HERE]
- The market-adjusted quintile spread return is equal to 0.57% in the five days that follow the event and increases to 1.66% when considering a 60-day holding period.
- Interestingly, the authors find that the premiums of the abnormal return and abnormal volume strategies are driven to a large extent by the outperformance of the quintile five of earnings announcement abnormal returns (i.e. those stocks with a positive surprise) and abnormal volume, respectively.
- These findings are worth pointing out because they raise the possibility that the earnings announcement abnormal volume effect exists despite the medium-term price momentum anomaly.
13 [FIGURE 5 ABOUT HERE]
- As with the abnormal return effect, the authors find, looking at Figure 5 and Table 3 (Panel B), that the premium earned by the abnormal volume strategy in the days following the announcement is not short lived.
- The cumulative difference in the adjusted long-term performance of stocks that fall in respectively quintile five and quintile one of earnings announcement abnormal volume experiences an upward drift until 90 days after the event.
- This period of underperformance is followed by a significant pick-up in the strategy returns, starting 250 days after the event.
- Stocks that surprise the market in one quarter seem to continue to do so in the same quarter one year later.
- Investors appear to be systematically surprised by these announcements.
4.2 The Role of Information Uncertainty
- One of the theoretical explanations for the post-earnings announcement drift is that investors under-react to public information.
- The authors use the specific volatility of stock returns as a proxy for information uncertainty and the dispersion of beliefs about valuations.
- Bekaert et al (2010) show, for a large sample of developed countries, that aggregate idiosyncratic volatility is well described by a stationary, mean reverting process with occasional shifts to a higher mean, higher variance regime.
- In Table 4 , the authors see, when considering a 20-day and a 60-day holding period, that a strategy based on earnings announcement abnormal return (Panel A) works significantly better within highly volatile stocks than within stocks with low volatility.
- When considering a short horizon, however, no significant difference can be found.
4.3 The Earnings Announcement Abnormal Return & Abnormal Volume Effects:
- In Table 6 , the authors intersect independent rankings based on earnings announcement abnormal return and abnormal volume, and measure, over different holding periods, the abnormal performance of stocks that fall in each intersection.
- The evidence for the marginal impact of the abnormal volume is, however, more statistically significant when considering longer holding periods.
- The quintile spread returns earned by this strategy in each intersection of abnormal volume are typically positive, and in several instances also statistically significant.
- Therefore, their empirical evidence seems to confirm the idea that these two empirical patterns are the results of distinct behavioural biases.
4.4 Trading Strategy with Monthly Rebalancing
- Given their independent information contents the authors investigate in this section the combined predictive power of the earnings announcement abnormal return and abnormal volume effects.
- Indeed, under this specification, a score is used for the first time on average two weeks after the earnings announcement.
- For their trading strategy to be easily implementable, it is important that it works within large-capitalisation names.
- The authors also distinguish between periods of high and low aggregate specific risk on the basis of idiosyncratic risk rankings at the end of every months.
- Equation ( 5) is ran using the monthly returns from the trading strategy of Figure 8 , where the top quintile of earnings surprises is bought and the bottom quintile sold.
[TABLE 7 ABOUT HERE]
- The earnings surprise strategy generates a large raw quintile spread return of 0.70% per month.
- The raw quintile spread return earned in the largest capitalisation subset of their universe is equal to 0.58% (0.56%).
- Similarly, to the extent that investors are already aware of large firms, it is at first unclear how an earnings announcement could significantly increase their visibility.
- As expected, the authors find, in Tables 7 and 8 , that the trading strategy earns a larger premium within those stocks that have higher levels of idiosyncratic risk.
- The positive premium for the low volatility subset of the universe loses, however, its statistical significance once the authors control for the Fama-French / Carhart factors.
5.1 The Trading Strategy Returns and Stock Illiquidity
- Chordia et al (2009) report that the post-earnings announcement drift in the US is prevalent mainly within relatively illiquid stocks.
- To the extent that market capitalisation does not perfectly control for stock illiquidity, it could be that the significant quintile spread return earned by the earnings surprise strategy within the largest capitalisation names disappears once stock illiquidity is taken into account explicitly.
- A stock is allocated in the high liquidity subset if its illiquidity score falls below the median value that month.
- Controlling for illiquidity only marginally reduces the quintile spread return of the strategy within the largest capitalisation subset of the universe.
- The raw and abnormal returns earned by the strategy continue to be both economically as well as statistically significant, hence possibly suggesting that their control for market capitalisation already captures most of the difference in firm liquidity.
5.2 Earnings Announcement Abnormal Volume and Absolute Returns
- Several authors have documented a contemporaneous correlation between the trading volume of a stock and its absolute return.
- Since absolute return closely proxy for stock volatility, it could be argued that the abnormal volume effect only reflects this risk premium.
- To assess whether this is the case, the authors first run the monthly strategy of Figure 8 , using the earnings announcement abnormal volume and the absolute announcement return on the threeday announcement window.
- Second, every months the authors run two cross-sectional regressions of the earnings announcement abnormal volume of each stock against their absolute announcement return.
[TABLE 10 ABOUT HERE]
- Table 10 shows that the absolute return strategy fails to earn a significant quintile spread return over the period of their study.
- In contrast, the volume strategy generates a significant raw quintile spread return of around 0.51% (0.38%) over the study 21 period.
- Moreover, controlling for the absolute return around the announcement reduces but does not eliminate the significance of the premium earned by the earnings announcement abnormal volume strategy.
6. Conclusion
- By studying these anomalies in a large sample of European firms that spans a period of 14 years.the authors.
- Finally, to avoid biasing their statistical tests, the authors consider the earnings surprise of a firm only if its post-event returns do not overlap with those of its previous surprise.
- 5 the authors look at the performance of the earnings announcement abnormal return (Panel A) and abnormal volume (Panel B) strategies following periods of high and low aggregate specific risk.
- To compute the characteristics benchmark, the authors match each firm to a portfolio of securities that fall in the same Size (market capitalization), Book-to-Market and Momentum terciles.
- The authors then use those rankings to create quintile spread portfolios that they hold until the end of the following month, when the ranking process is repeated.
Did you find this useful? Give us your feedback
Citations
12 citations
6 citations
Cites background from "Information Uncertainty and the Pos..."
...A more recent study in Europe was conducted by Gerard (2012)....
[...]
...More recently for example, Gerard (2012) has shown a more behavioural reason for the PEAD....
[...]
4 citations
4 citations
References
1,397 citations
"Information Uncertainty and the Pos..." refers background in this paper
...Moreover, Zhang (2006) demonstrated that investor underreaction to public information is even more pronounced in cases of greater information uncertainty....
[...]
1,213 citations
884 citations
"Information Uncertainty and the Pos..." refers background in this paper
...These findings echo those of Gervais, Kaniel, and Mingelgrin (2001), who showed that a positive relationship exists between short-term abnormal trading volume and subsequent returns, even without conditioning on earnings announcements....
[...]
750 citations
Additional excerpts
...54 www.cfapubs.org ©2012 CFA Institute and Thomas 1989; Bhushan 1994; Bartov, Radhakrishnan, and Krinsky 2000; Mendenhall 2004; Chordia, Goyal, Sadka, Sadka, and Shivakumar 2009)....
[...]
625 citations
"Information Uncertainty and the Pos..." refers background in this paper
...…offered an alternative interpretation of the causes of the post– earnings announcement drift with respect to the autocorrelation structure of forecast errors, and Livnat and Mendenhall (2006) put forward an explanation based on a more general hypothesis of underreaction to earnings information....
[...]
Related Papers (5)
Frequently Asked Questions (7)
Q2. What is the main constraint that the authors face in terms of data coverage?
the authors use widely availablemarket data to compute their indicators of market surprise, so that the main constraint that the authors face in terms of data coverage is the availability of interim and fiscal year-end report dates.
Q3. What is the strongest evidence of abnormal behaviour?
In other words, it is in within those stocks that suffer form larger degrees of information uncertainty and/or higher limits to arbitrage that the authors find the strongest evidence of abnormal behaviour.
Q4. Why do Bekaert et al (2010) show that idiosyncra?
Bekaert et al (2010) show, for a large sample of developed countries, that aggregate idiosyncratic volatility is well described by a stationary, mean reverting process with occasional shifts to a higher mean, higher variance regime.
Q5. What are the main reasons for the post-earnings announcement drift?
For instance, Liu and Thomas (2000) show that a significant portion of the market reaction around earnings announcement is due to nonearnings related information.
Q6. how do the authors determine whether the abnormal volume strategy continues to generate a significant return?
each month the authors run:i t i tt i ttt i t Return AbsoluteReturn AbsoluteVolume Abnormal εββα +⋅+⋅+= −−++ (8)The residuals of these regressions are then used in monthly strategies, to assess whether, controlling for absolute return, the abnormal volume strategy continues to generate a significant return.
Q7. What is the effect of controlling for stock illiquidity?
To the extent that market capitalisation does not perfectly control for stock illiquidity, it could be that the significant quintile spread return earned by the earnings surprise strategy within the largest capitalisation names disappears once stock illiquidity is taken into account explicitly.