Further Insight into ‘What Should We Expect When Testing for Price Response to News in Securities Litigation?’

Event studies have been used in litigation to examine issues such as how a company’s stock price responds to news events and to assess the efficiency of the market in which a stock trades. Unfortunately, there are few benchmarks for the likelihood that news will in fact cause a statistically significant price movement. My paper explains why there is no theoretical answer to this question, and also provides some empirical data. When we examine large companies and events that would likely be material to investors, such as earnings surprises, only about half of those news announcements are associated with a statistically significant stock-price movement, though there is considerable variance in that proportion across companies and industries. Notably, this holds even though we allow for two chances (the day of the news announcement and the following trading day) to find a statistically significant price movement. 

An easy, though informal, proof of why there is no theoretical basis for an argument that a specific share of news days should be associated with a statistically significant price movement can be made by pointing out that statistical significance is determined at various levels, with the 5% significance level being the most common in financial economics. Assume that under the 5% significance level, some fraction of news days is associated with statistically significant price movements. Under another significance level, say the stricter 1% significance level, a different fraction of news events would be associated with statistically significant price movements. Thus, in the abstract, there cannot be some constant expected fraction of news days that should be associated with statistically significant price movements.

In my paper, I address the question of what percent of ‘news days’ is associated with a statistically significant stock-price movement through a number of analyses. One analysis examines quarterly earnings announcements for members of the S&P 500 Index as of December 31, 2015 over calendar years 2010 through 2015. Two results of this analysis are: (1) for the median company, only 54.2% of earnings announcements are associated with a statistically significant stock-price movement; and (2) only 2.2% of companies have statistically significant stock-price movements associated with more than 80% of their earnings announcements. 

Thus, if one is testing for market efficiency, the proper empirical benchmark for the percent of earnings announcements that should lead to a statistically significant price response is not 50% (unless we are willing to have half of the members of the S&P 500 Index fail) or 80% (unless we are willing to have 98% of the members of the S&P 500 Index fail). Instead, it may be worthwhile to examine the frequency of observing a statistically significant price response to the particular types of announcements in question for companies in whose stock we expect trades in an efficient market.

Additional results not presented in the paper demonstrate that there may be benefits to developing a proper peer group. For example, while the median company shows a statistically significant response to 54.2% of earnings announcements, that figure ranges from 37.5% for the median company in the utilities sector to 62.5% for the median companies in three sectors including technology. Both of these results make sense intuitively. Utilities are often regulated and even if earnings in one quarter are unusually high or low, regulators may be expected to bring the profit margin back into line in the long run, thereby lowering the importance of a single earnings announcement. In contrast, companies in the technology sector are viewed as among the most likely to change their long-run operations and profitability, meaning that each earnings announcement may signal more about the future in that sector than in the economy overall.

The overall import of the theoretical and empirical analyses is that we should not assume benchmarks for the likelihood that a news announcement should result in a statistically significant price response by a company’s stock; we should rather examine the relevant literature or conduct relevant studies of similar situations.

David Tabak is the Senior Vice President at NERA Economic Consulting and a guest contributor to the Oxford Business Law Blog.