- Earnings Announcements: Major, Recurring Events
- Market Reactions to Earnings Announcements
- Earnings Surprises
- Market Reactions to Earnings Surprises Are Full of Surprises
This chapter is from the book
This chapter is from the book
This chapter is from the book
Market Reactions to Earnings Announcements
How large an impact do earnings announcements have on share prices? We have studied this question by analyzing all the quarterly earnings announcements of the largest 1,000 stocks (measured by market capitalization) during the 1984–2009 period. We measured earnings announcement returns over a three-day window that spanned from the day before earnings announcement to the day after. We examined 110,495 market-adjusted (excess) returns to earnings announcements made by Russell 1000 firms during the 1984–2009 period. Table 1.1 shows the distribution of (market-adjusted) stock market reactions to these earnings announcements.1 Overall, for all the earnings announcement returns spanning the full 26 years, the median announcement return was 10 basis points for the three-day window—about zero. This suggests that there were about as many positive earnings announcement returns as there were negative ones. Next, we looked at the average (mean) announcement return (presented in the bottom row). Coming in at 20 basis points for the three-day window returns, we concluded that, on average, earnings announcement returns were not that large at all. However, it would be a huge mistake to conclude from these numbers that the market reactions to earnings announcements were always small. After all, the average is just that: an average. It doesn’t tell us much about the variation in the returns.
Table 1.1. Distribution of Earnings-Announcement Excess Returns of Russell 1000 Stocks from 1984 to 2009
Distribution |
3-Day Window Return |
Maximum |
246.24% |
90th Percentile |
17.41% |
75th Percentile |
2.74% |
50th Percentile (Median) |
0.10% |
25th Percentile |
–2.34% |
10th Percentile |
–5.59% |
Minimum |
–84.24% |
Mean |
0.20% |
To examine the variation in returns, consider the quartile cutoffs. Specifically, for the three-day window, the 75th percentile excess return was 2.74%, and the 25th percentile excess return was –2.34%. That means that 25% of the earnings announcements had excess returns greater than 2.74%, and similarly, 25% of the earnings announcements had excess returns more negative than –2.34%. These returns are quite large when compared to the expected returns during a typical three-day period where no earnings were announced. For instance, assume that over the past 50 years, the average return of a random stock was about 7% a year and its annual volatility was about 20%. There are roughly 250 trading days per year, so the mean of daily market returns is about 2.8 basis points (=0.07/250),2 which means that a typical three-day return is about 8.4 basis points. The volatility over any three-day window is 2.20%. So 50% of the three-day earnings announcement returns (those above the 75th percentile and below the 25th percentile) are more than one standard deviation away from the mean. Assuming a normal distribution, this should happen about 32% of the time, but during earnings announcements it happens more than 50% of the time.3 If we examine the announcement returns in the 90th and 10th percentiles, the size of the returns is obviously even more significant. The basic message is that market movements during earnings announcement periods are quite large when compared to returns during nonevent periods, and hence offer great opportunities for trading.
When we dig deeper and consider the returns from each individual year between 1984 and 2009, we find an increasing trend in the percentile returns over the period. For instance, the 90th percentile announcement-period returns started out at a hair less than 5.0% in 1984, but steadily increased to approximately 12% in 2009. Similarly, the 25th percentile started in 1984 at approximately –2.5%, but ended at about –4.5% in 2009. Even if we ignore years with elevated market volatility (e.g., 2000, 2001, 2008, and 2009), the conclusion is the same: Market reactions to earnings announcements have increased in size and intensity over time. This suggests that the opportunities to profit from options-based trades around earnings announcements have not diminished over the years.