What Should We Expect When Testing for Price Response to News in Securities Litigation? — No. 1:21-md-02989 (S.D. Fla., Dkt. 582-7, entered 06/21/2023)
Summary
Document 582-7 in Case 1:21-md-02989-CMA, entered on the FLSD docket 06/21/2023, is a 17-page August 2016 paper, What Should We Expect When Testing for Price Response to News in Securities Litigation?, by Dr. David I. Tabak, Senior Vice President of NERA Economic Consulting. The paper treats price response to news as a key consideration in testing market efficiency under Basic Inc. v. Levinson and Cammer v. Bloom, and explains why the Proof-by-Example approach, which presents a handful of news days followed by stock-price movements, was rebutted in In re PolyMedica Corp. Securities Litigation, 453 F. Supp. 2d 260 (D. Mass. 2006). Later pages discuss the comparison test endorsed by courts, which compares the percentage of days with news that have a statistically significant price movement with the percentage of days without news that have one.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
Cite this document
“What Should We Expect When Testing for Price Response to News in Securities Litigation? — No. 1:21-md-02989 (S.D. Fla., Dkt. 582-7, entered 06/21/2023).”
Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/gov-uscourts-flsd-590042-582-7-77920033e1/
, SHA-256 dfb542ba496f9151d4e700718c891f9a89227aec8adc63701cc78d8c02944810.Full text
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 1 of 17
August 2016
What Should We Expect When Testing for
Price Response to News in Securities Litigation?
By Dr. David I. Tabak* Introduction
The frequency with which a stock price is expected to respond to news is an important issue
in securities litigation. As one example, price response to news is an important factor in the
consideration of market efficiency. If a stock trades in an efficient market, there is generally
a presumption, under Basic Inc. v. Levinson, 485 U.S. 224, 108 S. Ct. 978, 99 L. Ed. 2d 194
(1988), that investors can rely on the market price of the stock to reflect the effects of any
material misrepresentation or omission.
Among the various tests employed to assess market efficiency, whether a stock, or other
security, responds to news has emerged as the key consideration. In part, this relates back to
the seminal Cammer v. Bloom case that set forth five factors that many courts have adopted
as part of their consideration of market efficiency. The Cammer Court noted, “Finally, it would
be helpful to a plaintiff seeking to allege an efficient market to allege empirical facts showing a
cause and effect relationship between unexpected corporate events or financial releases and an
immediate response in the stock price. This, after all, is the essence of an efficient market and
the foundation for the fraud on the market theory.”1
Electroniccopy
Electronic copy available
available at:
at:https://ssrn.com/abstract=2824565
http://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 2 of 17
The Proof-By-Example Approach
Following Cammer, there has been an evolving history of how experts and lawyers presented
empirical facts that might have been deemed to show a cause-and-effect relationship between
news and stock-price movements. An initially popular approach might be termed “Proof by
Example.” In this method, an expert or counsel for plaintiffs would present a handful of days
with news followed by a stock-price movement.2 A serious, and in fact, fatal, problem with
this approach is that one would expect to see such results if stock-price movements were
completely random and had no average correlation with news events. One can, for example,
look at a calendar and provide examples to “demonstrate” that Mondays are correlated with
even-numbered days of the month, or that Mondays are correlated with odd-numbered days
of the month. Similarly, in most securities-litigation cases, there are likely to be days with news
(“news days”) associated with large stock-price movements and news days associated with
typical or even small stock-price movements. Similarly, we would expect to see days without
news (“non-news days”) associated with both large and small stock-price movements.
The Proof-by-Example approach was forcefully rebutted in In re PolyMedica Corp. Securities
Litigation, 453 F. Supp. 2d 260 (D. Mass. 2006). In this case, the expert for plaintiffs
presented five days where there was news and what he claimed was an associated stock-
price response. The expert for defendants countered with an explanation of the reasoning
described above. The court agreed with the defense expert, and found that the plaintiffs’
expert’s “mere listing of five days on which news was released and which exhibited large
price fluctuations proves nothing.”3
The FDT Test and Its Successes
The failures of the Proof-by-Example approach, at least by practitioners and courts that
recognize those failures, raise the question of what a proper alternative should be. In 2004,
a law review article (co-authored by the author of this paper) and a submission by an expert
for defendants in a securities litigation advocated very similar, if not functionally identical,
approaches. “The ‘Less Than’ Efficient Capital Markets Hypothesis: Requiring More Proof from
Plaintiffs in Fraud-on-the-Market Cases” proposed a test that has since been referred to as
the “FDT Test.”4 The FDT Test can be conceptualized similarly to a medical study in which a
researcher wishes to determine if a drug is effective. To do so, the medical researcher runs
a study with two groups. One group, the control group, is either given no drug or is given a
placebo, while the other group, the treatment group, is given the drug in question. Results
between the two groups are then examined to see if they are different. In the FDT Test, stock-
price movements on news days (the treatment group) are compared to stock-price movements
on non-news days (the control group) to see if they are different. In particular, the FDT Test runs
an event study for each day in the period being examined and then uses the results of those
event studies to compare the fraction of news days with a statistically significant stock-price
movement to the fraction of non-news days with a statistically significant stock-price movement
to see if there is a difference between the two fractions. It further asks whether the difference
between the two fractions is itself statistically significant and is higher on news days.5
www.nera.com 2
Electroniccopy
Electronic copy available
available at:
at:https://ssrn.com/abstract=2824565
http://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 3 of 17
The 2004 case alluded to above is Lehocky v. Tidel Tech., Inc., 220 F.R.D. 491 (S.D. Tex. 2004).
In this matter, the court found as follows (emphasis added):
Plaintiffs’ expert[] conducted an event study using Tidel’s trading data. He
identified two-day periods in which information pertaining to Tidel was
released to the public and separated those two day[] [periods] from other
two day periods in which there was no public information pertaining to Tidel.
The periods were classified into “information” versus “non-information days”
…. [Plaintiffs’ expert] concluded that the price changes on information
days versus non-information days was statistically significant, meaning
there was a related cause and effect relationship between [a] the
release of information pertaining to Tidel and [b] Tidel’s stock price.
Simply put, [Plaintiffs’ expert’s] tests indicated that Tidel’s stock price
reacted within a two day window to news releases concerning Tidel,
which indicates market efficiency.
While the terminology is slightly different, and the focus of plaintiffs’ expert in Tidel appears
to be on price changes, perhaps indicating the magnitude of changes and not merely
whether there was a change, the basic concept of employing a statistical comparison of
price movements on news days to those on non-news days is similar to the concept behind
the FDT Test.
More recently, courts have accepted a number of tests of market efficiency that directly follow
the FDT Test procedures, if not citing the FDT paper directly. See the following examples (with
all emphases added):
• Lumen v. Anderson, No. 08-0514-CV-W-HFS (W.D. Mo. Feb. 10, 2012).
The expert also concluded FCStone’s stock was approximately one-and-
a-half to two times as likely to experience a change in stock price on
days when company-specific news was announced than on days when
it was not. However, for legal purposes, the critical question is whether the
price quickly absorbs and reflects news about the company because this
is what allows the legal presumption that (1) the price incorporates public
information and (2) the investor relied on the price as the repository of that
public information. [The defendants’ expert’s] report establishes that
when news about the company is made public, the company’s stock
price immediately incorporates the information. This is the essence of an
efficient market — perhaps not for an economist’s purposes, but for purposes
of Basic.
• Smilovits v. First Solar, Inc., No. CV12-00555-PHX-DGC (D. Ariz. Oct. 8, 2013).
Plaintiffs submit that, overall, statistically significant price moves
occurred three times more often on days with relevant news than on
days without news.
www.nera.com 3
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 4 of 17
Furthermore, the court rejected defendants’ expert’s conclusion based on a discussion of days
when the stock apparently had no significant price change following news because, citing
Lumen, “there is no comparison to the number of times there were significant price changes
following important news.”
• McIntire v. China MediaExpress Holdings, Inc., 38 F. Supp. 2d 415 (S.D.N.Y. 2014).
For that reason, courts have instead endorsed the comparison test that
[Plaintiffs’ expert] used. See, e.g., In re Alstom SA Sec. Litig., 253 F.R.D. 266,
280 (S.D.N.Y.2008). This test “involves comparing the percentage of days with
news that have a statistically significant price movement to the percentage of
days without news that have a statistically significant price movement.” Paul A.
Ferrillo et al., The “Less Than” Efficient Capital Markets Hypothesis: Requiring
More Proof from Plaintiffs in Fraud-on-the-Market Cases, 78 St. John’s L. Rev.
81, 120 (2004). If the stock price is significantly more likely to change on News
Days than on Non-News Days, that suggests a causal relationship between
material news and the stock price. Id.
• In re Petrobras Securities Litigation, No. 14-cv-9662 (JSR) (S.D.N.Y. Feb. 2, 2016).
Both sides refer to [plaintiffs’ expert’s] methodology as an “FDT” test
because use of z-test to evaluate market efficiency was first proposed in a
law review article by three well-known securities econometric experts, whose
combined initials were “FDT.” See Paul A. Ferrillo, Frederick C. Dunbar, and
David Tabak, The “Less Than” Efficient Capital Markets Hypothesis: Requiring
More Proof from Plaintiffs in Fraud-on-the-Market Cases, 78 St. John’s L. Rev.
81, 119-22 (2004). [Defendants’ expert] contends that, because the article
was not peer-reviewed, a z-test cannot be used to show market efficiency.
Were [plaintiffs’ expert] using a novel or questionable statistical technique,
the Court would place more weight on the absence of peer review. But it is
not necessary for every application of a commonly used statistical technique
to be peer-reviewed. Indeed, the elegance of statistical methods is that they
can be applied to data sets of varying substantive significance, from rates
of emphysema to transactions on modern securities markets. Because the
Court is convinced that the z-test is a well-established and sound statistical
technique, the lack of peer review does not seriously undermine [plaintiffs’
expert’s] application of the z-test.
www.nera.com 4
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 5 of 17
Challenges to the FDT Test
Not all cases in which a test to examine the fractions of news and non-news days associated
with statistically significant returns was employed by plaintiffs have resulted in a finding of
market efficiency. Defendants do have at least two wins with the argument that all or nearly all
news dates should be associated with statistically significant returns:
• In re Federal Home Loan Mortgage Corp., No. 09 Civ. 832 (MGC) (S.D.N.Y. Mar. 27, 2012).6
A plaintiff must show that the market price responds to most new, material
news.
…
At his first deposition, [plaintiffs’ expert] testified that if new information
had been released into the market on each of his initial 28 event days, one
could expect to see statistically significant abnormal returns on each of the 28
days in an efficient market. In contrast, [defendants’ expert] testified that an
economist may conclude that a market is efficient if it reacts to news 80 to
90% of the time, depending on the number of news dates at issue.
• George v. China Automotive Systems, Inc., No. 11 Civ. 7533 (KBF) (S.D.N.Y. July 3, 2013).
Even assuming that the methodology was proper, showing that only seven
out of sixteen days resulted in a market reaction is an insufficient foundation
upon which to pronounce market efficiency. (Id. at 100.) (To state the obvious,
seven out of sixteen is less than 50%.)
In fact, defendants have frequently argued that some large number of dates with news should
be associated with statistically significant stock-price movements.7 Yet, it is not clear if there
is any case in which defendants have cited some empirical or even sound theoretical basis
for an argument that there should be some minimum fraction of news days associated with
statistically significant returns.
How Often Should Stocks Respond to News in a Statistically
Significant Manner?
Theoretical Issues
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.8 Assume that under
the 5% significance level, some fraction of news days is associated with statistically significant
price movements when the series of event studies that form the FDT Test is undertaken. Under
another significance level, say the stricter 1% significance level, a different, and smaller, fraction
would be associated with statistically significant price movements. In fact, a different fraction
would exist for each level of statistical significance (though with a discrete count of days, there
could be cases where the fraction would remain constant over ranges of levels of statistical
significance until the significance level is increased (reduced) to the point where another day
www.nera.com 5
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 6 of 17
fails (succeeds) in becoming statistically significant). Thus, there cannot be some constant
expected fraction of news days that should be associated with statistically significant price
movements in an efficient market, not even 50%, as that fraction would depend on the level of
statistical significance being employed.
One can also formally show that the expected fraction of news days that should be associated
with statistically significant stock-price movements will depend on (1) the distribution of market-
adjusted stock-price movements on news days and (2) the cutoff for determining what size
market-adjusted stock-price movement is deemed statistically significant. It is notable that under
the standard academic practice of not filtering news days from the period used to estimate the
“market model” that provides the parameters for determining whether days are associated with
statistically significant stock-price movements, the share of all days that are news days will affect
that cutoff. Again, there is no reason for the fraction of news days that should be associated
with statistically significant stock-price movements to necessarily be above or below 50%.
Intuitively, one can see at least part of the reason why the fraction of news days associated
with statistically significant price movements is indeterminate theoretically by considering
that the definition of a news day can differ. So, if one defines a news day as any day with
an earnings announcement, we might expect a certain fraction of those news days to be
associated with a statistically significant stock-price movement. However, if we define a news
day to be one with an earnings surprise (i.e., an earnings announcement where the reported
earnings are different from the consensus estimate, often proxied for by the average of
earnings estimates provided by analysts), then we might expect the fraction of news days
associated with statistically significant returns to be greater. Alternatively, if we have a broader
definition of news, such as mentions of the company in any news story, we might expect the
fraction of those news days associated with a statistically significant stock-price movement to
be lower. There is no theoretical or mathematical/statistical reason to believe that the fraction
of news days associated with statistically significant returns should exceed 50% as a general
matter in an efficient market.
Academic Literature
Given the lack of a theoretical standard for the fraction of news days that should be associated
with statistically significant returns, we might ask if there is any guidance from the academic
literature on this subject. There is in fact a long literature on the general inability of news to
substantially explain stock-price movements. One seminal article is Richard Roll, “R2,” Journal
of Finance 43, (1988). The abstract for this paper begins, “Even with hindsight, the ability
to explain stock price changes is modest. R2s were calculated for the returns of large stocks
as explained by systematic economic influences, by the returns on other stocks in the same
industry, and by public firm-specific news events. The average adjusted R2 is only about .35 with
monthly data and .20 with daily data.”
Following Roll’s paper, the literature typically focused on how much stock prices move even
after accounting for news (e.g., if stock prices tended to move by 10% on news days and
8% on non-news days, one could say that news explained 20% of stock-price movements).
However, the typically more relevant question brought forth by plaintiffs and defendants in
securities litigations is how often news should elicit a stock-price movement, the measure
examined in the FDT Test. There are a few articles that do address this issue directly. One such
www.nera.com 6
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 7 of 17
paper is Jacob Boudoukh et al., Which News Moves Stock Prices? A Textual Analysis, Working
Paper No. 18725, National Bureau of Economic Research, October 14, 2013. The authors find
that “identified news days are 32.5% more likely to coincide with the bottom and top 10% of
return days.”9 That is, while one would expect 20% of news days to be in the bottom and top
10% of returns if price movements were unaffected by news, instead 26.5% (i.e., 20% plus an
additional 32.5% of that 20%) of news days are in that group. Recall that the 5% significance
level represents the bottom and top 2.5% of the sample. As 26.5% of news days are in the
bottom and top 10% of the sample, 26.5% is an upper bound of the percentage of news
days that could be in the bottom and top 2.5% of the sample and thus be associated with
statistically significant returns at the 5% level.
Next, we turn to John M. Griffin, Nicholas H. Hirschey, Patrick J. Kelly, “How Important Is the
Financial Media in Global Markets?” Review of Financial Studies, 2011. Chart A on page 3971
of the article shows that if one sorts days by the magnitude of their price movements, with 100
representing the largest movement in absolute magnitude and 0 the smallest, the mean rank
for news days in the U.S. is about 57 or 58.10 That is, rather than all or nearly all of the news
days falling in the upper 5% of the distribution (i.e., ranks 95-100), which by definition is the set
of statistically significant price movements at the 5% level, the price movement for a “typical”
news day is above the midpoint (i.e., a rank of 50) but is not statistically significant (i.e., in ranks
95-100). That is, only news that is well above average in importance to investors should be
expected to give rise to a statistically significant price movement.
Both of these papers show that the assumption that half of all news days should be associated
with statistically significant returns does not hold in practice, at least for the news, stocks, and
time periods covered by these articles.
New Empirical Analyses
We now turn to furthering those analyses. In particular, in this paper, we address the
question of what percent of “news days” is associated with a statistically significant stock-
price movement.
As a first analysis, we begin with a methodology designed to give a high estimate of the
number of events that will be associated with statistically significant returns when the event
studies underlying the FDT Test are performed. Our sample is all members of the S&P 500 Index
as of December 31, 2015, excluding those that began trading after April 2014. The examination
period covers 2010 through 2015 and a news day is defined as an earnings announcement.
For each stock, a market model is calculated over each calendar year from 2009 through 2014
using the S&P 500 Index as the explanatory variable, and the parameters and standard error of
each market model are used to determine the statistical significance of the stock’s daily price
movements during the following calendar year.
The events examined are the earnings announcements by each company in the time period
examined. To account for the possibility that an earnings announcement came out either
before or after the close of trading, and to increase the probability of finding a statistically
significant return associated with each earnings announcement, we characterize an earnings
announcement as being associated with a statistically significant return if either the day
of the earnings announcement or the following trading day has a statistically significant
www.nera.com 7
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 8 of 17
return, or both.11 As a result, we would not expect merely 5% of news days to be associated
with statistically significant returns merely by chance, but (if markets are efficient and price
movements on consecutive days are independent), we would expect roughly twice that (in
fact, 9.75%) of news days to be associated with a statistically significant price movement.12 The
results are presented in column (1) of Table 1 below, which shows the number of companies in
different percentages of statistically significant returns across its earnings announcements.
Table 1. Excess Returns on Earnings Announcements1 for 504 Companies in the S&P 500 Index,
as of December 31, 2015,2 over the Period January 4, 2010 through December 31, 2015
Daily Returns Predicted Using the S&P 500 Index 3
All Earnings Small Earnings Medium Earnings Large Earnings
Announcements 4 Announcements 4 Announcements 4 Announcements 4
(1) (2) (3) (4)
Assuming that there is a significant return if the excess return on the announcement day or the following day is
statistically significant at the 5% level:
Total Number of Companies5,6 496 487 478 448
Average Number of Earnings 23.3 10.8 5.0 8.6
Announcements7
Percent of companies with the percent of earnings announcements with statistically significant returns in the range:
< = 20% 2.0% 8.2% 15.5% 12.9%
> 20% and < = 40% 19.2% 18.5% 21.8% 18.3%
> 40% and < = 60% 45.8% 37.4% 24.9% 33.5%
> 60% and < = 80% 30.8% 25.5% 21.8% 19.6%
> 80% and < = 100% 2.2% 10.5% 16.1% 15.6%
Percent of earnings announcements with statistically significant returns for the median company:
54.2% 53.8% 50.0% 50.0%
Notes and Sources:
Price data and dividend data obtained from FactSet Research Systems, Inc.
Returns were adjusted for dividends and the effective trading date of the earnings dates were used.
1 Earnings announcement dates obtained from FactSet Research Systems, Inc.
2 S&P 500 Index constituents, as of December 31, 2015, obtained from Bloomberg, L.P.
3 Company stock price returns for each calendar year are predicted by regressing the daily logarithmic returns of the company’s stock on the daily logarithmic returns
of the S&P 500 Index over the previous calendar year. Stock price returns on January 4, 2016, are predicted by regressing the daily logarithmic returns of the
company’s stock on the daily logarithmic returns of the S&P 500 Index, over the period: January 2, 2014, through December 31, 2014.
4 Earnings announcements are classified as “Small,” “Medium,” or “Large” based on the absolute value of the I/B/E/S amount earnings-per-share (EPS)
surprise reported divided by the company stock price on the report date. “Small” events have an absolute EPS surprise of less than or equal to 0.075% of the
closing price on the report date, “Medium” events have an absolute EPS surprise greater than 0.075% of the closing price on the report date but less than or
equal to 0.15% of the closing price on the report date, and “Large” events have an absolute EPS surprise greater than 0.15% of the closing price on the report
date. EPS surprises are calculated from the mean I/B/E/S EPS estimate for that earnings report.
5 The following companies were excluded from the analyses in columns (1) to (4) because their shares began trading publicly after April 2014: CSRA, Inc., Hewlett
Packard Enterprise Co., Synchrony Financial, Alphabet, Inc. (class C shares), Navient Corp., PayPal Holdings, Inc., Columbia Pipeline Group, Inc., and Baxalta, Inc. In
addition, News Corp. was excluded from the analyses in columns (2) to (4) because of an inability to locate I/B/E/S EPS surprise data for 2010-2015.
6 Values in this row represent the total number of companies with at least one earnings day of the type indicated in columns (1) to (4).
7 Calculated for only companies with at least one earnings day of the given type.
www.nera.com 8
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 9 of 17
A number of results are apparent from Table 1 and the data underlying that table. First,
it is rarely the case that all or nearly all of the earnings announcements are associated
with statistically significant returns. Even allowing for two opportunities to find a
statistically significant return (on either the day of or the trading day following the earnings
announcement, as discussed above) for each earnings announcement, only 2.2% of S&P 500
companies have more than 80% of their earnings announcements associated with statistically
significant returns.13
For the median S&P 500 company, 54.2% of its earnings announcements were associated
with statistically significant returns, again allowing for two opportunities to find a statistically
significant return. Thus, the argument that even a majority of news announcements should
be associated with statistically significant returns would not hold for approximately half of the
members of the S&P 500 under this set of parameters.
Next, we refine this analysis by dividing the earnings announcements, or surprise, into “Small,”
“Medium,” and “Large” categories by comparing the actual earnings figure to the consensus
estimate from I/B/E/S.14 If the earnings surprise divided by the stock’s closing price on the
earnings report date yields a figure of at most 0.075%, the earnings announcement is defined
to be Small; if the earnings surprise comes in more than 0.075%, but within 0.15%, of the
closing price on the report date, the earnings announcement is defined to be Medium; and
if earnings surprise come in more than 0.15% from the closing price on the report date, the
earnings announcement is defined to be Large.15 Columns (2) through (4) of Table 1 show
the percent of Small, Medium, and Large earnings announcements that are associated with
statistically significant returns.
As seen in Table 1, when we define news days to include only one of the categories of earnings
announcements, we still find fewer than 17% of S&P 500 companies have a proportion of
news days to be associated with statistically significant returns that exceeds 80%, with both the
Medium and Large categories coming in between 15% and 17%.16
In Table 2, we replicate Table 1 but add an industry index for each market model. In particular,
we add the SPDR Sector ETF corresponding to each company’s GICS Sector Code. As seen in
the table, the results do not change substantially from what we observe in Table 1, though
more earnings-announcement dates tend to be associated with statistically significant returns
(e.g., if we examine Medium and Large earnings surprises, 18.6% and 16.5% of companies for
the two categories, respectively, have at least 80% of the announcements dates associated with
a statistically significant return.)
For our last set of analyses of earnings announcements for S&P 500 companies, we examine
whether the price movements for the Small, Medium, and Large earnings announcements are
in the direction one would expect based solely on the direction of the earnings surprise.17 This
expands upon the results from the two academic papers discussed above, neither of which
imposes a directionality requirement on the returns. The results are shown in Table 3.
www.nera.com 9
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 10 of
17
Table 2. Excess Returns on Earnings Announcements1 for 504 Companies in the S&P 500 Index,
as of December 31, 2015,2 over the Period January 4, 2010 through December 31, 2015
Daily Returns Predicted Using the S&P 500 Index and SPDR Sector ETFs3
All Earnings Small Earnings Medium Earnings Large Earnings
Announcements 4 Announcements 4 Announcements 4 Announcements 4
(1) (2) (3) (4)
Assuming that there is a significant return if the excess return on the announcement day or the following day is
statistically significant at the 5% level:
Total Number of Companies5,6 496 487 478 448
Average Number of Earnings 23.3 10.8 5.0 8.6
Announcements7
Percent of companies with the percent of earnings announcements with statistically significant returns in the range:
< = 20% 1.6% 6.8% 14.2% 10.9%
> 20% and < = 40% 16.9% 18.5% 19.5% 19.4%
> 40% and < = 60% 46.4% 36.8% 24.3% 29.7%
> 60% and < = 80% 32.3% 27.1% 23.4% 23.4%
> 80% and < = 100% 2.8% 10.9% 18.6% 16.5%
Percent of earnings announcements with statistically significant returns for the median company:
54.2% 55.0% 52.3% 53.8%
Notes and Sources:
Price data and dividend data obtained from FactSet Research Systems, Inc.
Returns were adjusted for dividends and the effective trading date of the earnings dates were used.
1 Earnings announcement dates obtained from FactSet Research Systems, Inc.
2 S&P 500 Index constituents, as of December 31, 2015, obtained from Bloomberg, L.P.
3 Company stock price returns for each calendar year are predicted by regressing the daily logarithmic returns of the company’s stock on the daily logarithmic returns
of the S&P 500 Index and the SPDR Sector ETF corresponding to the company’s GICS Sector Code, over the previous calendar year. Stock price returns on January
4, 2016, are predicted by regressing the daily logarithmic returns of the company’s stock on the daily logarithmic returns of the S&P 500 Index and the SPDR Sector
ETF corresponding to the company’s GICS Sector Code, over the period: January 2, 2014, through December 31, 2014. The SPDR Technology Sector ETF (ticker:
“XLK”) was used for companies with the GICS codes 45 (Information Technology) and 50 (Telecommunication Services).
4 Earnings announcements are classified as “Small,” “Medium,” or “Large” based on the absolute value of the I/B/E/S amount earnings-per-share (EPS) surprise
reported divided by the company stock price on the report date. “Small” events have an absolute EPS surprise of less than or equal to 0.075% of the closing price
on the report date, “Medium” events have an absolute EPS surprise greater than 0.075% of the closing price on the report date but less than or equal to 0.15% of
the closing price on the report date, and “Large” events have an absolute EPS surprise greater than 0.15% of the closing price on the report date. EPS surprises are
calculated from the mean I/B/E/S EPS estimate for that earnings report.
5 The following companies were excluded from the analyses in columns (1) to (4) because their shares began trading publicly after April 2014: CSRA, Inc., Hewlett
Packard Enterprise Co., Synchrony Financial, Alphabet, Inc. (class C shares), Navient Corp., PayPal Holdings, Inc., Columbia Pipeline Group, Inc., and Baxalta, Inc. In
addition, News Corp. was excluded from the analyses in columns (2) to (4) because of an inability to locate I/B/E/S EPS surprise data for 2010-2015.
6 Values in this row represent the total number of companies with at least one earnings day of the type indicated in columns (1) to (4).
7 Calculated for only companies with at least one earnings day of the given type.
www.nera.com 10
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 11 of
17
Table 3. Excess Returns on Earnings Announcements1 for 504 Companies in the S&P 500 Index,
as of December 31, 2015,2 over the Period January 4, 2010 through December 31, 2015
Daily Returns Predicted Using the S&P 500 Index and SPDR Sector ETFs3
All Non-Zero Small Non-Zero Medium Large
Earnings Earnings Earnings Earnings
Announcements 4 Announcements 4 Announcements 4 Announcements 4
(1) (2) (3) (4)
Assuming that there is a significant return if the excess return on the announcement day or the following day is statistically
significant at the 5% level and is in the same direction as the earnings news:
Total Number of Companies5,6 495 487 478 448
Average Number of Earnings 23.1 10.7 5.0 8.6
Announcements7
Percent of companies with the percent of earnings announcements with statistically significant returns in the range:
< = 20% 9.1% 30.2% 27.6% 19.4%
> 20% and < = 40% 48.9% 40.0% 32.0% 30.4%
> 40% and < = 60% 38.4% 23.0% 22.2% 27.2%
> 60% and < = 80% 3.4% 3.9% 11.1% 11.8%
> 80% and < = 100% 0.2% 2.9% 7.1% 11.2%
Percent of earnings announcements with statistically significant returns for the median company:
37.5% 31.3% 33.3% 41.0%
Notes and Sources:
Price data and dividend data obtained from FactSet Research Systems, Inc.
Returns were adjusted for dividends and the effective trading date of the earnings dates were used.
1 Earnings announcement dates obtained from FactSet Research Systems, Inc.
2 S&P 500 Index constituents, as of December 31, 2015, obtained from Bloomberg, L.P.
3 Company stock price returns for each calendar year are predicted by regressing the daily logarithmic returns of the company’s stock on the daily logarithmic returns
of the S&P 500 Index and the SPDR Sector ETF corresponding to the company’s GICS Sector Code over the previous calendar year. Stock price returns on January 4,
2016, are predicted by regressing the daily logarithmic returns of the company’s stock on the daily logarithmic returns of the S&P 500 Index and the SPDR Sector
ETF corresponding to the company’s GICS Sector Code, over the period: January 2, 2014, through December 31, 2014. The SPDR Technology Sector ETF (ticker:
“XLK”) was used for companies with the GICS codes 45 (Information Technology) and 50 (Telecommunication Services).
4 Earnings announcements where the EPS reported was exactly the same as the mean EPS estimate are excluded. Earnings announcements are classified as “Small,”
“Medium,” or “Large” based on the absolute value of the I/B/E/S amount earnings-per-share (EPS) surprise reported divided by the company stock price on the
report date. “Small” events have an absolute EPS surprise of less than or equal to 0.075% of the closing price of the stock on the report date, “Medium” events have
an absolute EPS surprise greater than 0.075% of the closing stock price on the report date but less than or equal to 0.15% of the company price on the report date,
and “Large” events have an absolute EPS surprise greater than 0.15% of the closing price on the report date. EPS surprises are calculated from the mean I/B/E/S EPS
estimate for that earnings report.
5 The following companies were excluded from the analyses in columns (1) to (4) because their shares began trading publicly after April 2014: CSRA, Inc., Hewlett
Packard Enterprise Co., Synchrony Financial, Alphabet, Inc. (class C shares), Navient Corp., PayPal Holdings, Inc., Columbia Pipeline Group, Inc., and Baxalta, Inc. In
addition, News Corp. was excluded from the analyses in columns (2) to (4) because of an inability to locate I/B/E/S EPS surprise data for 2010-2015.
6 Values in this row represent the total number of companies with at least one earnings day of the type indicated in columns (1) to (4).
7 Calculated for only companies with at least one earnings day of the given type.
www.nera.com 11
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 12 of
17
The results are similar to those in Table 2, but with the additional restriction that to qualify
as a statistically significant result, the price movement has to be in the same direction of the
earnings surprise, reducing the number of announcements that are considered to be associated
with a statistically significant price movement. There are (at least) three theoretical reasons
for this. First, some of the difference may be due to noise, in which a day with, say, a positive
earnings announcement should engender a positive price movement, but due to a large
negative degree of noise, the price movement is actually negative and statistically significant.
However, this cannot reasonably be expected to explain the full difference in the tables. In
particular, since we are using the 5% significance level, we would expect to obtain a statistically
significant price movement in the “incorrect” direction either (1) if the movement is in the
“incorrect” direction on the first day but not in the “correct” direction on the second, or (2)
neither in the “correct” nor “incorrect” direction on the first day, but in the “incorrect” direction
on the second day. Consequently, there is a 4.8125% chance of finding a statistically significant
movement in the “incorrect” direction.18 Second, it could be the case that the direction of
an earnings surprise presents an incorrect view of the net effect of the announcement. For
example, suppose that a company announced that earnings unexpectedly fell 5% this quarter
because the company is spending on expanding production such that earnings will be up by
20% for several quarters. If one just looked at the current earnings figures, this would seem to
be a negative announcement. Yet, the market might easily view this news as positive on net.
Third, the consensus expectation for earnings may be incorrect because some analysts had not
updated their analysis prior to the earnings announcement or had highly idiosyncratic views that
skewed the average estimate.
Because earnings announcements are not days with no news, our expectations for price
movements in the “incorrect” direction are lower than those above. As a simple measure, we
remove those announcements with statistically significant movements in the “correct” direction
and assume that 4.8125% of the remaining earnings announcements should have a movement
in the “incorrect” direction merely by chance. Table 4 then shows the expected percentage
of earnings announcements with movements in the “incorrect” direction, using the market
model including the industry adjustment. In Table 4, we see that when looking at the S&P
500 Index, there is a notable fraction of earnings announcements with statistically significant
price movements in the “incorrect” direction, with 17.2% of all responses to non-zero earnings
announcements across all S&P 500 companies appearing to be in the “incorrect” direction
compared to 36.7% in the “correct” direction. Notably, the ratio of “incorrect” to “correct”
movements tends to fall as we move from Small to Medium to Large earnings announcements,
suggesting that the overall news content of announcements with larger earnings surprises
more closely aligns with the direction of the earnings surprise than is true for announcements
with smaller earnings surprises. This suggests that caution is required in interpreting the overall
direction of a news announcement relative to market expectations.19
www.nera.com 12
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 13 of
17
Table 4. Excess Returns on Earnings Announcements1, Considering Directionality for 504 Companies in the
S&P 500 Index, as of December 31, 2015,2 over the Period January 4, 2010 through December 31, 2015
Daily Returns Predicted Using the S&P 500 Index and SPDR Sector ETFs3
All Non-Zero Small Non-Zero Medium Large
Earnings Earnings Earnings Earnings
Announcements 4 Announcements 4 Announcements 4 Announcements 4
(1) (2) (3) (4)
Assuming that there is a significant return if the excess return on the announcement day or the following day is statistically
significant at the 5% level and is in the same direction as the earnings news:
Total Number of Companies5,6 495 487 478 448
Average Number of Earnings 23.1 10.7 5.0 8.6
Announcements7
Assuming that a “correct” movement is a statistically significant return on that day or the following day and in the same
direction as the earnings announcement, and assuming that an “incorrect” movement is a statistically significant return
in the opposite direction of the earnings news:
Percent of Earnings Announcements 36.7% 33.1% 38.4% 40.5%
with a Correct Movement
Expected Percentage of Earnings 3.0% 3.2% 3.0% 2.9%
Announcements with an Incorrect Movement8
Observed Percentage of Earnings 17.2% 22.1% 16.1% 11.2%
Announcements with an Incorrect Movement9
Excess Percentage of Earnings 14.1% 18.9% 13.1% 8.4%
Announcements with an Incorrect Movement
Excess percentage of earnings announcements with an “incorrect” movement for the median company:
13.7% 21.0% 15.7% 10.0%
Notes and Sources:
Price data and dividend data obtained from FactSet Research Systems, Inc.
Returns were adjusted for dividends and the effective trading date of the earnings dates were used.
1 Earnings announcement dates obtained from FactSet Research Systems, Inc.
2 S&P 500 Index constituents, as of December 31, 2015, obtained from Bloomberg, L.P.
3 Company stock price returns for each calendar year are predicted by regressing the daily logarithmic returns of the company’s stock on the daily logarithmic returns
of the S&P 500 Index and the SPDR Sector ETF corresponding to the company’s GICS Sector Code over the previous calendar year. Stock price returns on January 4,
2016, are predicted by regressing the daily logarithmic returns of the company’s stock on the daily logarithmic returns of the S&P 500 Index and the SPDR Sector
ETF corresponding to the company’s GICS Sector Code, over the period: January 2, 2014, through December 31, 2014. The SPDR Technology Sector ETF (ticker:
“XLK”) was used for companies with the GICS codes 45 (Information Technology) and 50 (Telecommunication Services).
4 Earnings announcements where the EPS reported was exactly the same as the mean EPS estimate are excluded. Earnings announcements are classified as “Small,”
“Medium,” or “Large” based on the absolute value of the I/B/E/S amount earnings-per-share (EPS) surprise reported divided by the company stock price on the
report date. “Small” events have an absolute EPS surprise of less than or equal to 0.075% of the closing price of the stock on the report date, “Medium” events have
an absolute EPS surprise greater than 0.075% of the closing stock price on the report date but less than or equal to 0.15% of the company price on the report date,
and “Large” events have an absolute EPS surprise greater than 0.15% of the closing price on the report date. EPS surprises are calculated from the mean I/B/E/S EPS
estimate for that earnings report.
5 The following companies were excluded from the analyses in columns (1) to (4) because their shares began trading publicly after April 2014: CSRA, Inc., Hewlett
Packard Enterprise Co., Synchrony Financial, Alphabet, Inc. (class C shares), Navient Corp., PayPal Holdings, Inc., Columbia Pipeline Group, Inc., and Baxalta, Inc. In
addition, News Corp. was excluded from the analyses in columns (2) to (4) because of an inability to locate I/B/E/S EPS surprise data for 2010-2015.
6 Values in this row represent the total number of companies with at least one earnings day of the type indicated in columns (1) to (4).
7 Calculated for only companies with at least one earnings day of the given type.
8 Expected Percentage is calculated by multiplying the probability of a statistically significant movement in the wrong direction by chance (4.8125%, calculated as
0.025 * [0.95 + 0.975]) by the percent of announcements that did not have a “correct” movement.
9 Calculated as the difference in percent of earnings announcements with a statistically significant movement when the direction of the movement does not matter
and the percentage of earnings days with a statistically significant movement when the movement must be in the same direction as the earnings news.
www.nera.com 13
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 14 of
17
Conclusion
As has been well-documented in the academic literature, a large share of stock-price
movements cannot be substantially explained by identifiable news events. The analyses
in this paper examine a similar question, considering the fraction of news events that are
associated with statistically significant stock-price movements for individual companies
when the traditional event-study methodology is employed. The results show that when we
examine large companies and events that should be material to investors, such as relatively
large earnings surprises, only about half of those news announcements are associated with a
statistically significant stock-price movement. Notably, this result 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.
While these results are not surprising given the prior academic literature, they do suggest that
one should be careful in setting a priori expectations about how frequently one would expect
a stock price to respond to what may appear to be material news. If “material” were defined
with a very high bar, one would still presumably expect that most such material news would
be associated with a statistically significant price movement for stocks that trade in efficient
markets, though that bar might be reasonably high. Thus, assumed absolute standards about
how frequently stocks that trade in efficient markets should respond to news in a statistically
significant manner ignore both theoretical issues with such assumptions and empirical evidence.
Finally, we also see that one should be careful about assuming that the direction of a price
movement can always be easily inferred from a casual, or potentially even a careful, reading
of news. When examining earnings announcements, we see more statistically significant price
movements in the opposite direction of the earnings surprise than we would expect by chance
alone. This indicates that there may be additional information in earnings announcements
beyond just the overall surprise in the reported results, or that we have not properly understood
what market expectations were before the announcement. Again, the message is that one
should be careful in setting a priori expectations about the frequency and direction of how one
would expect a stock price to respond to what might appear to be material new information.
www.nera.com 14
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 15 of
17
References
* Senior Vice President, NERA Economic Consulting. I thank Marcia Kramer Mayer and Jordan Milev for comments
and Stefan Boettrich, Kyle Monk, John Shymansky, Shadman Torofder, and Ashutosh Venkatraman for their
contributions to the analysis. I also note that my colleagues at NERA and I have been involved in some of the cases
discussed in this paper.
1 Cammer v. Bloom, 711 F. Supp. 1264, 1287 (D.N.J. 1989).
2 See, for example, In re Accredo Health, Inc. Securities Litigation, 2006 WL 1716910 (W.D.Tenn.) (“As to the fifth
[Cammer] factor, plaintiffs offer the declaration of its expert, [ ], who opines that ‘material new information resulted
in an immediate response in Accredo’s stock price.’ [Plaintiff’s expert] Decl. ¶ 28. [Plaintiff’s expert] describes one
example[.]”) and In re The Mills Corp. Securities Litigation, 257 F.R.D. 101 (E.D. Va. 2009) [2009 BL 82258] (“A
finding of market efficiency for the preferred shares is further supported by the considerable price drop in Mills
preferred stock on August 11, 2006.”). In these cases, the reference to a single example of a stock-price movement
is the only basis given for satisfying the fifth Cammer factor (i.e., price response to new information).
3 In re PolyMedica Corp. Securities Litigation, 453 F. Supp. 2d 260, 270 (D. Mass. 2006).
4 Paul A. Ferrillo, Frederick C. Dunbar, and David Tabak, “The ‘Less Than’ Efficient Capital Markets Hypothesis:
Requiring More Proof from Plaintiffs in Fraud-on-the-Market Cases,” 78 St. John’s L. Rev. 81, 119-22 (2004).
5 This is the standard form of the FDT Test. One can also compare the magnitude of the stock-price movements
on the two types of days, for example. Different versions of the test might all point to the same conclusion
regarding price response to news for the same stock over the same time period, though that need not be the
case. The standard form of the FDT Test, looking at the fractions of news and non-news days with statistically
significant stock-price responses to news, is well-suited for determinations relevant to securities litigation because
it is more directly aimed at the binary question of whether a stock price responds to news. In particular, the FDT
Test examines whether any stock-price response is above the threshold for statistical significance, a result that has
often been taken in securities litigation to correspond to the binary question of whether investors should have been
able to rely on the price of the stock to reflect public information. An additional consideration is that examinations
of the magnitude of a stock-price response may be heavily driven by a few very large movements that raise the
average price response to news and may not be indicative of how the stock price would respond to typical news
or to news of lesser importance. This would be particularly problematic if, contrary to proper statistical technique,
the event ending the class period is included in the test and if the combination of a news event and large stock-
price movement following that event was part of the reason why the litigation was brought (i.e., there would be a
selection-bias issue in including such a day as a news event, as its inclusion would have been driven, at least in part,
by the fact that there was an associated stock-price movement). Still, a combination of various tests will typically be
helpful in either validating or challenging the broader applicability of a single test.
6 It is worth noting that after apparently making numerous errors in his initial report, the expert for plaintiffs first cited
to the FDT article in a second report. The court discounted this, stating, “Furthermore, I question [the expert]’s use
of these tests for the first time in a second and inconsistent event study.”
7 See, for example, Första AP-Fonden v. St. Jude Medical, Inc., Civil No. 12-3070 (JNE/HB) (D. Minn. Dec. 22, 2015).
(“[Defendants] point out, for example, that one iteration of [Plaintiffs’ expert’s] comparative analysis showed that
SJM stock responded to news in a statistically significant amount only 36.4% of the time.”). The court found that
the stock did trade in an efficient market.
8 See, for example, the Reference Guide on Statistics in the Reference Manual of Scientific Evidence, Third Edition,
Federal Judicial Center, 2011, p. 251 (“In practice, statistical analysts typically use levels of 5% and 1%. The 5%
level is the most common in social science, and an analyst who speaks of sig¬nificant results without specifying the
threshold probably is using this figure.” Internal footnote omitted.)
9 The authors “break down all news articles into either ‘unidentified news’ or an ‘identified news’ category based on
whether the article can be tied to a set of relevant events (broken down into 14 categories and 56 subcategories).
… Moreover, these 158 thousand stock-days are narrowed down further to stock-days defined as ‘complex’, either
by number of articles, number of events or disagreement in tone across articles.” (p.3) The finding of more than
just a chance number of news days associated with statistically significant returns “is much more pronounced for
identified news days associated with one of our three measures of complexity, i.e., multiple articles, multiple events
or disagreement across articles. Respectively, these days are 82.2%, 78.3% and 89.2% more likely to coincide
with extreme returns days.” That is, 36.44%, 35.66%, and 37.84% of the three types of “complex” news days,
respectively, are associated with the bottom and top 20% of stock-price returns.
News is identified as “all documents that pass through the Dow Jones Newswire from January 1, 2000 to December
31, 2009. For computational reasons, we limit ourselves to the S&P500 companies (with at least 40 days of
available data).” (p. 11)
10 The authors employ the Factiva database and “identify news for a firm using full-text searches for the firm’s name in
the headline or lead paragraph of articles.” (p. 3948)
www.nera.com 15
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 16 of
17
References
11 Due to the large dataset, we do not investigate whether news came out before or after the close of the market
for each event. However, if the news event occurred before the close of trading on the announcement day, then a
price movement solely on the following day would generally not be taken as an indication of market efficiency.
12 There would be a 95% chance of not having a statistically significant return on any day if we were considering
chance alone. Thus, the probability of not having a statistically significant return on two consecutive days is 95% ×
95%, or 90.25%, meaning that there is a 9.75% chance of having a statistically significant return on at least one of
two consecutive days due solely to chance.
13 Column (1) of Table 1 shows the fraction of S&P 500 companies that have different percentages of their earnings
announcements associated with statistically significant returns. For example, 2.0% of the companies have up to
20% of their earnings announcements associated with a statistically significant return while an additional 19.2% of
S&P 500 companies have between 20% and 40% of their earnings announcements associated with a statistically
significant return. The sum of the five entries in the middle block of the column will add to 100%, with any
apparent discrepancy due to rounding.
14 I/B/E/S, the Institutional Brokers’ Estimate System, is a commonly used data provider that collects, among other
information, analyst earnings estimates and calculates an average or consensus estimate that can be compared to
later-reported earnings. As earnings surprises are on a continuum, the Small, Medium, and Large categories are, of
course, set by arbitrary cutoffs solely for the purposes of exposition.
15 Here, we effectively assume that most earnings announcements are made after the close on the report date. An
alternative analysis that measures the earnings as a percent of the stock price the day before the announcement
yields similar results. In addition, alternative analyses that look at the absolute magnitude of the earnings surprise
in cents per share and others that use the S&P SmallCap 600 Index instead of the S&P 500 Index also yield similar
results to the main analyses in this paper. Results of alternative analyses are available from the author upon request.
16 One might be surprised that the percent of companies with more than 80% of news days having statistically
significant returns is higher for all three categories of earnings announcements (i.e., 10.5%, 16.1%, and 15.6%
for Small, Medium, and Large earnings announcements) than for earnings announcements overall (2.2%). A
similar result is seen with the percent of companies with no more than 20% of news days being associated
with statistically significant returns. These results are because there are fewer Small, Medium, or Large earnings
announcements than total earnings announcements, and with fewer announcements, there are more likely to be
companies with very few Small, Medium, or Large earnings announcements, making it easier for them to have, for
example, either 0% or 100% of those announcements be associated with statistically significant returns.
17 Note that here, we replace Small earnings announcements with Small Non-Zero earnings announcements.
18 This is calculated as the probability of (1) above (2.5% × 97.5%) plus the probability of (2) above (95% × 2.5%).
Note that if there is both a “correct” and an “incorrect” movement on the two days, the return is classified as being
in the “correct” direction overall.
19 This is even more likely to be relevant in situations where explicit expectations do not exist. For example, if a
company is sued for securities fraud several days after it made an announcement that caused its stock to fall, the
market may consider the filing of a securities-fraud lawsuit to be highly likely. In such a situation, the actual filing
could of course be negative news, but could also be positive news if the scope of the lawsuit (e.g., the time period
covered) is less than expected by the market.
www.nera.com 16
Electronic copy available at: https://ssrn.com/abstract=2824565
Case 1:21-md-02989-CMA Document 582-7 Entered on FLSD Docket 06/21/2023 Page 17 of
17
About NERA
NERA Economic Consulting (www.nera.com) is a global firm of experts dedicated to applying
economic, finance, and quantitative principles to complex business and legal challenges. For
over half a century, NERA’s economists have been creating strategies, studies, reports, expert
testimony, and policy recommendations for government authorities and the world’s leading
law firms and corporations. We bring academic rigor, objectivity, and real world industry
experience to bear on issues arising from competition, regulation, public policy, strategy,
finance, and litigation.
NERA’s clients value our ability to apply and communicate state-of-the-art approaches clearly
and convincingly, our commitment to deliver unbiased findings, and our reputation for quality
and independence. Our clients rely on the integrity and skills of our unparalleled team of
economists and other experts backed by the resources and reliability of one of the world’s
largest economic consultancies. With its main office in New York City, NERA serves clients
from more than 25 offices across North America, Europe, and Asia Pacific.
Contact
For further information and questions, please contact the author:
Dr. David I. Tabak
Senior Vice President
+1 212 345 2176
david.tabak@nera.com
The opinions expressed herein do not necessarily represent the views of NERA Economic Consulting or any other NERA
consultant. Please do not cite without explicit permission from the author.
Electronic copy available at: https://ssrn.com/abstract=2824565
File and source
- File
- gov.uscourts.flsd.590042.582.7.pdf
- Size
- 709,096 bytes
- SHA-256
- dfb542ba496f9151d4e700718c891f9a89227aec8adc63701cc78d8c02944810
- Original
- No public link identified.