MSC 515 Econometrics

--Empirical Issues in Finance Research

1
 Introduction and back ground

 The event study literature: basic facts

 Characterizing event study methods

2
 Event studies examine the behavior of firms’ stock
around corporate events
◦ The usefulness of an event study provides a measure of the
(unanticipated) impact of this type of event on the wealth
of the firms’ claimholders

◦ Event studies focusing on announcement effects for a

short-horizon around an event provide evidence relevant
for understanding corporate policy decisions

3
 Event studies serve an important purpose in capital
market research as a way of testing market
efficiency
◦ Systematically nonzero abnormal security returns that
persist after a particular type of corporate event are
inconsistent with market efficiency

◦ Event studies focusing on long-horizon following an event

can provide key evidence on market efficiency

4
5
 The use of daily rather than monthly security return
data has become prevalent
◦ More precise measurement of abnormal returns
◦ More informative studies of announcement effects

 The methods used to estimate abnormal returns and

calibrate their statistical significance have become
more sophisticated
◦ This change is of particular importance for long-horizon
event studies

6
 Serious limitations of long-horizon methods have
been brought to light and still remain
◦ Inferences from long-horizon tests require extreme caution

◦ The developments underscore and dramatically strengthen

early warnings about the reliability or lack of reliability of
long-horizon methods

7
 The model
◦ An event study tries to examine return behavior for a
sample of firms experiencing a common type of event
 Stock split
 Stock repurchase
 M&A
 Strategic alliance
 …

◦ The event might take place at different points in calendar

time or it might be clustered at a particular date

8
 Estimation period Event windows

Event day

Expected return The difference

Estimate parameters is abnormal
return
Actual return

9
 Let t=0 represent the time of the event

 For each sample security i, the return on the

security for time period t relative to the event
◦ 𝑅𝑖𝑡 = 𝐾𝑖𝑡 + 𝑒𝑖𝑡
 Kit is the normal (expected or predicted return given a
particular model of expected returns)
 eit is the component of returns which is abnormal or
unexpected
 𝑒𝑖𝑡 = 𝑅𝑖𝑡 − 𝐾𝑖𝑡

10
 eit is the difference between the return conditional on
the event and the expected return unconditional on the
event
◦ Thus, the abnormal return is a direct measure of the
(unexpected) change in security holder wealth associated with
the event
 The security is typically a common stock

◦ A model of normal returns must be specified before an abnormal

return can be defined
 Market model
 Constant expected returns model
 Capital asset pricing model

11
 Cross-sectional aggregation
◦ An event study seeks to establish whether the cross-
sectional distribution of returns at the time of an event is
abnormal

◦ In event study literature, the focus almost always is on the

mean of the distribution of abnormal returns

◦ Typically, the specific null hypothesis to be tested is

whether the mean abnormal returns at time t is equal to
zero

12
 The focus on mean effects (i.e., the first moment of
the return distribution)
◦ The event on average associated with a change in security
holder wealth, and if one is testing economic models and
alternative hypotheses that predict the sign of the average
effect

◦ For a sample of N securities, the cross-section mean

abnormal return for any period t is
1 𝑁
 𝐴𝑅𝑡 = 𝑖=1 𝑒𝑖𝑡
𝑁

13
 To examine whether mean abnormal returns for
periods around the event are equal to zero
◦ If the event is partially anticipated, some of the abnormal
return behavior related to the event should show up in the
pre-event period
 Information leakage

◦ The speed of adjustment to the information revealed at the

time of the event is an empirical question
𝑡2
 CAR: 𝐶𝐴𝑅 𝑡1 , 𝑡2 = 𝑡=𝑡1 𝐴𝑅𝑡
 Buy-and-hold method

14
 For a given performance measure, such as CAR, a test
statistic is typically computed and compared to its
assumed distribution under the null hypothesis
◦ Null hypothesis: Mean abnormal performance equals zero
◦ The null hypothesis is rejected if the test statistic exceeds a
critical value

◦ The test statistic is a random variable because abnormal returns

are measured with error
 Prediction about securities’ unconditional expected returns are
imprecise
 Individual firms’ realized returns at the time of an event are affected
for reasons unrelated to the event
 This component of the abnormal returns does not average to zero in
the cross-section

15
 For the CAR, a standard test statistic is the CAR
divided by an estimated of its standard deviation
𝐶𝐴𝑅(𝑡1 ,𝑡2 )
◦
[𝜎 2 (𝑡1 ,𝑡2 )]1/2
 where 𝜎 2 𝑡1 , 𝑡2 = 𝐿𝜎 2 (𝐴𝑅𝑡 ) and 𝜎 2 (𝐴𝑅𝑡 ) is the variance of
the one-period mean abnormal return
 𝜎 2 𝑡1 , 𝑡2 = 𝐿𝜎 2 (𝐴𝑅𝑡 ) implies that the CAR has a higher variance
the longer is L, and assumes time-series independence of the one-
period mean abnormal return

 Event-time clustering renders the independence assumption for

the abnormal returns in the cross-section incorrect
 This would bias the estimated standard deviation downward and
𝐶𝐴𝑅(𝑡 ,𝑡 )
the test statistic given in 2 1 21/2 upward
[𝜎 (𝑡1 ,𝑡2 )]

16
 To address the bias induced by event-time clustering
◦ The significance of the event-period average abnormal
return can be gauged using the variability of the time series
of event portfolio returns in the period preceding or after
the event date
 For example, we can construct a portfolio of event firms and
obtain a time series of daily abnormal returns on the portfolio
for a number of days (e.g., 180 days) around the event date

 The standard deviation of the portfolio returns can be used to

assess the significance of the event-window average abnormal
return

17
 The cross-sectional dependence is accounted for
because the variability of the portfolio returns
through time incorporates whatever cross-
dependence that exists among the returns on
individual event securities

18
 The portfolio approach has a drawback
◦ To the extent the event period is associated with increased
uncertainty, the use of historical or post-event time-series
variability might understate the true variability of the
event-period abnormal performance
 The statistical significance of the event window abnormal
performance would be overstated if it is evaluated on the basis
of historical variability of the event-firm portfolio returns

 Estimate the likely increase in the variability of event and non-

event periods
 The ratio of the variances during the event period and non-event
periods might serve as an estimate of the degree of increase in the
variability of returns during the event period

19
 The join-test problem
◦ While the specification and power of a test can be
statistically determined, economic interpretation is not
straightforward because all tests are joint tests
 Event studies are well-specified only to the extent that the
assumption underlying their estimation are correct

 The joint test imposes challenge

 Whether the abnormal returns are zero
 Whether the assumed model of expected return is correct

20
 Additional assumptions concerning the statistical
properties of the abnormal return measures must
also be correct
◦ For example
 A standard t-test for mean abnormal return assumes that the
mean abnormal performance for the cross-section of securities
is normally distributed

 Depending on the specific t-test, there may be additional

assumptions that the abnormal return data are independent in
time-series or cross-section

21
 Much of what is known about general properties of
event study tests comes from large-scale simulation
(Brown and Warner, 1980)

◦ Intuition
 Different event study methods are simulated by repeated application
of each method to samples that have been constructed through a
random selection (or stratified random) of securities and random
selection of an event date to each

 If performance is measured correctly, these samples should show no

abnormal performance

 This make it possible to study test statistic specification, that is, the
probability of rejecting the null hypothesis when it is known to be
true

22
 Qualitative properties
◦ Specification

◦ Power against specific types of alternative hypotheses

◦ Sensitivity of specification to assumptions about the return

generating process

23
24
 Horizontal length has a big impact on the event
study test properties
◦ Short-horizon event study methods are generally well-
specified

◦ Long-horizon event study are sometimes very poorly

specified
 No procedure has been developed to achieve complete
confidence so far

25
 Short-horizon methods are quite powerful if the
abnormal performance is concentrated in the event
window

 Long-horizon event studies (even when they are

well-specified) generally have low power to detect
abnormal performance
 Both when it is concentrated in the event window and
when it is not

26
 With short-horizon methods the test statistic
specification is not highly sensitive to the
benchmark model of normal returns or assumptions
about the cross-sectional or time-series dependence
of abnormal returns

 For long-horizon methods, the specification is quite

sensitive to assumptions about the return
generating process

27
 In calculating the test statistic in an event study, a key
input required here is the individual security return
variance
◦ Standard deviation of daily returns on individual securities using
all CRSP commonstock securities from 1990-2002
◦ For each year, firms are ranked by their estimated daily standard
deviation
 Firms with missing observations are excluded
◦ The numbers under mean and median columns represent the
average of the annual mean and median values for the firms in
each decile and for all firms
◦ The number of firms in each decile ranges from 504 in 2002 to
673 in 1997

28
29
 Cross-sectional tests examine how the stock price
effects of an event are related firm characteristics
◦ For a cross-section of firms, abnormal returns are
compared to firm characteristics
 To provide evidence to discriminate among various economic
hypotheses

 Cross-sectional tests are a standard part of almost

every event study
◦ They are relevant even when the mean stock price effect of
an event is zero

30
 Abnormal returns vary cross-sectionally is that
economic effect of the event differs by firm
◦ Events are endogenous, reflecting a firm’s self selection to
choose the event, which in turn reflects insiders’
information
 Standard estimates of cross-sectional coefficients can be biased
 Appropriate procedures for treating self-selection and partial
anticipation issues is the subject of an entire section
 Will discuss later

31
 Additional important issues
◦ Some studies not on the stock price effect of an event, but
on predicting a corporate event (using past stock prices as
one explanatory variable)

◦ Other issues
 Whether the event was partially anticipated by market
participants
 Whether the partial anticipation is expected to vary cross-
sectionally in a predictable fashion
 For example, market participants might anticipate that managers
of firms experiencing high price run-up are likely to make value-
destroying stock acquisitions

32
 All event studies have to deal with
◦ Risk adjustment and expected/abnormal return modeling

◦ The aggregation of security-specific abnormal returns

◦ Calibration of the statistical significance of abnormal

returns

 These issues are critically important with long

horizons

33
 Long-horizon event studies have a long history,
including the original stock split event
◦ Fama et al. (1969)

 Many long-horizon studies document apparent

abnormal returns spread over long horizons
◦ Whether the apparent abnormal returns are due to mispricing, or
simply the result of measurement problems, is an unresolved
issue among financial economists

◦ The methodological research in the area is important because it

demonstrated how easy it is to conclude there is abnormal
performance when none exists

34
 In long-horizon tests, appropriate adjustment for
risk is critical in calculating abnormal price
performance
◦ In short-horizon tests in which risk adjustment is
straightforward and typically unimportant
◦ The error in calculating abnormal performance due to
errors in adjusting for risk in a short-horizon test is likely
to be small
◦ Daily expected returns are about 0.05%
 Even if the firm portfolio’s beta risk is misestimated by 50%
(1.5 instead of 1 [rue beta]), the error in the estimated
abnormal errors is small relative to the abnormal return of 1%

35
 In multi-year long-horizon tests, risk-adjusted
return measurement is the fatal weakness
◦ Even a small error in risk adjustment can make an
economically large difference when calculating abnormal
returns over horizon of one year or longer
 Make little difference for short horizons

◦ It is unclear which expected return model is correct, and

therefore estimates of abnormal returns over long horizons
are highly sensitive to model choice

36
 Which model of expected return is appropriate remains
an unresolved issue
◦ CAPM as a model of expected returns being thoroughly
discredited as a result of the voluminous anomalies evidence

◦ Fama-French three-factor model

 It is not clear about the economic underpinning of including size,
book-to-market, and momentum factors

 From the standard point of event study analysis, it is essential to

include them
 To isolate the incremental impact of an event on security price
performance

 These three factors are applicable to all stocks sharing those

characteristics, not just the sample of firms experiencing the event

37
 Two main methods for assessing and calibrating
post-event risk-adjusted performance
◦ Characteristics-based matching approach

◦ Jensen’s alpha approach

 Calendar-time portfolio approach

 There is still no clear winner in a horse race

◦ Both have low power against economically interesting null
hypotheses
◦ Neither is immune to misspecification

38
 An appealing feature of using BHAR is that buy-and-
hold returns better resemble investors’ actual
investment experience than periodic rebalancing
entailed in other approaches to measuring risk-
adjusted performance

39
 Once a matching firm or portfolio is identified,
BHAR calculation is straightforward
◦ A T-month BHAR for event firm i is defined as
 𝐵𝐻𝐴𝑅𝑖 𝑡, 𝑇 = 𝑡=1 𝑡𝑜 𝑇(1 + 𝑅𝑖,𝑡 ) − 𝑖=1 𝑡𝑜 𝑇(1 + 𝑅𝐵,𝑡 )
 Where RB is the return on either a non-event firm that is
matched to the event firm i, or it is the return on a matched
(benchmark) portfolio

◦ When matching the benchmark firms, a non-event firm that

is closest to an event firm on the basis of firm size, book-to-
market ratio, and past one-year return

40
 The distinguishing feature of the most recent
variants of the approach is to calculate calendar-
time portfolio returns for firms experiencing an
event, and calibrate whether they are abnormal in a
multifactor regression
◦ e.g., CAPM or Fama-French three factor

41
 Assume a sample of firms experiences a corporate event
◦ The event may spread over several years

 Assume that the research seeks to estimate price

performance over two years following the event for each
sample firm

 In each calendar month over the entire sample period, a

portfolio is constructed comprising all firms
experiencing the event within the previous T month

42
 Because the number of event firms is not uniformly
distributed over the sample period, the number of
firms included in a portfolio is not constant through
time
◦ As a result, some new firms are added each month and
some firms exit each month
◦ The portfolios are rebalanced each month and an equal or
value-weighted portfolio excess return is calculated

43
 The resulting time series of monthly excess returns is regressed on
 The CAPM market factor

 The three Fama-French factors

 The four Carhart (1997) factors as follows

 𝑅𝑝𝑡 − 𝑅 = 𝑎𝑝 + 𝑏𝑝 𝑅𝑚𝑡 − 𝑅 + 𝑆𝑝 𝑆𝑀𝐵𝑡 + ℎ𝑝 𝐻𝑀𝐿𝑡 + 𝑚𝑝 𝑈𝑀𝐷𝑡 + 𝑒𝑝𝑡

 Inferences about the abnormal performance are on the basis of the

estimated ap and its statistical significance. Since ap is the average
monthly abnormal performance over the T-month post-event period, it
can be used to calculate annualized post-event abnormal performance.

44
45
 Mitchell and Stafford (2000) and Brav and Gompers
(1997) favor the Jensen-alpha approach

 Loughran and Ritter (2000) argue against using the

Jensen-alpha approach because it might be biased
toward finding results consistent with market efficiency
◦ Their rationale is that corporate executives time the events to
exploit mispricing

◦ Jensen-alpha approach, by forming calendar-time portfolios,

under-weights managers’ timing decisions and out-weights other
observations

46
 Assessing the statistical significance of the event
portfolio’s BHAR has been particularly difficult because
◦ Long-horizon returns depart from the normality assumption that
underlies many statistical tests

◦ Long-horizon returns exhibit considerable cross-correlation

 Because the return horizons of many event firms overlap
 And also because many event firms are drawn from a few industries

◦ Volatility of the event firm returns exceeds that of matched firms

because of event-induced volatility

47
 Long-horizon buy-and-hold returns, event after
adjusting for the performance of a matched firm,
tend to be right skewed

◦ The right-skewness of the distribution of long-horizon

abnormal returns on event portfolios appears to be due
largely to the lack of independence arising from
overlapping long-horizon return observations in event
portfolios

48
 Cross-correlation
◦ Specification bias arising due to cross-correlation in
returns is a serious problem in long-horizon tests of price
performance
 Economy-wide and industry-specific factors would generate
contemporaneous co-movements in security returns is the
cornerstone of portfolio theory and is economically intuitive
and empirically compelling

◦ If the test statistic in an event study is calculated ignoring

cross-dependence in data, even a fairly small amount of
cross-correlation in data will lead to serious
misspecification of the test

49