Você está na página 1de 12

108

How A Specific Market Announcement May Impact The Stock Price Value Of A
Particular Firm An Event Empirical Study

Anastasia Vardavaki
16 Epidavrou str., Halandri
152 33, Athens, Greece
E-mail: anastasia_vardavaki@yahoo.gr

Dr. J ohn Mylonakis
10 Nikiforou str., Glyfada
166 75, Athens, Greece
E-mail: imylonakis@panafonet.gr

Abstract
This paper focused on the examination of the effect of a specific event on the stock price of a
particular firm. The firm of interest is a biotechnology company traded on NASDAQ. The main
event constituted the announcement of the discovery of a new cure for cancer by ENMD.
Furthermore, there were two other significant events that influenced ENMD price. Applying an
event study analysis, the paper tests whether these events had an effect on the stock price of the
firm, that is, whether the abnormal returns achieved due to these events are statistically
significant. Results found that the reaction to the positive event was strong and significant. On
the other hand, the negative reaction to the next event was weaker and almost insignificant.
This considerable difference can be explained by taking into account the type of the journal in
each case. One can conclude that what is relevant for the market is not the event itself but the
common knowledge of that event among investors. Furthermore it can be mentioned that in order
to examine the impact of an event on the market prices, one has to take into account the timing
and the extent to which this event has been revealed to investors.

Key words: Event Study, Stock Prices, Efficient Market Hypothesis

1. Introduction
An event study, in economics/finance/accounting research, is an analysis of whether there was a
statistically significant reaction in financial markets to past occurrence of a given type of event
that was hypothesized to affect public firms' market values. According to fundamental-based
approach to stock pricing, an asset should trade at the risk-adjusted present value of its expected
future cash flows. In frictionless market, an equilibrium price does not involve disagreement
among the market participants. Although the efficient-market hypothesis suggests that price
changes are unpredictable, it connects them with changes in investors beliefs about future cash
flows. Since expectations change with the arrival of new information, prices should change as
well. Thus, a specific event that takes place either inside or outside of a firm can influence
significantly its stock price.

Gur Huberman and Tomar Regev (2001) present a very instructive case by describing the
chronology of events that affected the stock price of EntreMed during the period 1997-1998.
They document three specific public announcements that provoked considerable influence on its
price. They point out that this is a very interesting case of how a non-event affected a firms
108

109

stock price and the price of other firms as well, and they try to explain why the real event did
not affect the stock price.

This paper aims to study the effect of the event in May 4, 1998 on the stock price of the firm.
This is conducted by an event study analysis, using the market model and estimating the
abnormal returns around the event. The NASDAQ Biotechnology Index and the NASDAQ
Composite are used as market benchmarks. The paper tests whether the abnormal returns
achieved due to this event, were significantly different from zero. Moreover, another event study
is conducted for the third event (November 12, 1998).

2. Event Study
Firstly, the meaning of an event has to be determined. An event is the public announcement
that affects a firm's market value. The event can be within the firm's control, such as an
announcement of a stock split, a merger, a security issue, an earnings announcement, a new
investment policy, a new product launch. Furthermore, the event can be outside the firm's
control, such as the event of a legislative act being passed, or a regulatory ruling being
announced, that will affect the firm's future operations in some way. Some events, such as a
regulatory change or an economic shock can affect many securities contemporaneously; other
events, particularly these that take place within the firm, are specific to individual securities.

An event study is an econometric procedure that isolates the stock price impact of the event or its
impact on firm value. Particularly, the event-study methodology is used to examine the reaction
of investors to positive and negative news. The methodology is based on the assumption that
capital markets are sufficiently efficient to evaluate the impact of new information (events) on
expected future profits of the firms. The main purpose of an event study is to test whether the
efficient market hypothesis holds. If abnormal returns disseminate rapidly after the event occurs,
event study proves that market is efficient. On the other hand, if abnormal returns persist after
the occurrence of an event or they are associated with an anticipated event, then it is obvious that
these facts contradict to the efficient market hypothesis. Furthermore, the event study analysis
measures the magnitude of an events impact.

Event study involves the following steps:

1) Identification of the event of interest and definition of the event window. This means that
the timing of the events occurrence has to be identified, which is not necessary the period during
the event but it can also be the investment period immediately preceding the announcement of
the event.

2) Determination of the firm selection criteria. In this stage one has to arrange the security
performance data relative to the timing of the event. The day of the announcement is determined
as period zero and the number of preceding and following periods with 1,-2 .. and +1,+2
respectively is selected.

3) Prediction of the normal return during the event window in the absence of the event. This
step involves the estimation of the normal return based on data from the pre-event measurement
period. Several methods may be used to estimate normal returns:
109

110


* Constant mean return model:
t t
r c + = , where is the mean of returns and c is a white
noise;

* Market model which assumes a linear relationship between the return of any security to the
return of the market portfolio:


R
it
= o
i
+ |
i
R
mt
+ c
it
(1)
E(c
it
) = 0, Var (c
it
) = o
c


where t is the time index, i =1,2, N stands for security, R
i,
R
mt
are the returns on security
and the market portfolio respectively during the period, and c
it
is the error term for the security.
The market model differs from the first model because the market return is used as an
independent variable affecting the securitys return.
Equation (1) is generally estimated over a period which runs between 120 and 10 days prior to
the event. The event window is defined as the period from 10 days prior to the event to 10 days
after the event. With the estimates from equation (1), one can predict the normal return during
the days covered by the event window.

4) Estimation of the abnormal return within the event window, where the abnormal return is
defined as the difference between the actual and predicted return. This step aims to separate the
security-specific component of the return from the securitys total return. Therefore, to estimate
the security-specific return (abnormal return) one should subtract from total return the securitys
alpha and beta times the markets return on that day. The alphas and betas are the same as those
estimated form the pre-event regression. Thus, the abnormal return is calculated as:

t m t
E
t t t
R R R R AR
,

| o = = (2)

5) Testing procedures. In this stage one must test the null hypothesis that the abnormal returns
are near zero, i.e. the observed returns are near the expected (normal) returns. Thus, if the event
is unanticipated and the t-statistic is significant on the day of the event but insignificant on the
days following the event, a reasonable conclusion is that the event does affect security returns
but that it does not contradict the efficient market hypothesis. In contrast, if the t-statistic is
significant on the post-event days as well, one might conclude that the market is inefficient
because it does not rapidly absorb new information. Market is also inefficient if t-statistic is
significant on the day of the event but the event was partly anticipated. It must be pointed out
that the less anticipated the event is, the more powerful the event study analysis is.

Mackinlay states that under the null hypothesis, the abnormal returns are normally distributed
with mean zero and variance equal to the variance of the residuals from regression (1) plus a
component of adjustment due to sampling error in the estimated o and | .

( ) window event t AR N AR
t t
e ~ ) ( , 0
2
o
110

111

where,
(
(


+ + =
2
2
, 2 2

) (
1
1
) (
m
m t m
t
r
L
AR
o

o o
c
(3)
c
o =standard deviation of the residual from regression (1);
t m
r
,
=market return relative to the considered day;
2
,
m m
o =sample mean and variance of market returns calculated on the estimation window;
L =length of the estimation window.

The variance of the abnormal returns from (3) can be approximated by the first term of (3) only,
(by the variance of the regressions residuals), if the estimation window L is wide enough.
Therefore, the abnormal returns are asymptotically distributed as normal . ) , 0 (
2
c
o N

Each single return in the event window can be tested. The most relevant test is the one on the
event date (day 0). Moreover it is interesting to test the cumulative abnormal returns on the
appropriate event sub windows, typically (-1,0) and (-1,+1). The cumulative abnormal returns are
defined just as the sum of the abnormal returns on the chosen sub window ( : ) ,
2 1
T T

=
=
2
1
) , (
2 1
T
T t
t
AR T T CAR (4)
Following the asymptotic approximation, the cumulative abnormal returns under the null
hypothesis, are normally distributed with mean zero and variance defined as:

2
1 2 2 1
2
) 1 ( ) , (
c
o o + = T T T T (5)

Thus, one can test under the Ho. )) , ( , 0 ( ) , (
2 1
2
2 1
T T N T T CAR
t
o ~

6) Empirical results. In this stage one can test if there are any violations in the assumptions of
the model or if we have to change the sample size. Furthermore, one can make some diagnostics
tests.

7) Interpretation and Conclusions. This stage is very important because it provides a solid
economic framework and useful interpretation of the results. It gives considerable information on
the market efficiency hypothesis and it explains how a specific event can affect the firms market
value.

3. Background on EntreMeds case
Huberman and Regev (2001) conducted an analysis on EntreMed case in order to present the
significant effect of an enthusiastic public announcement, which was not a real event, on the
firms stock price. They intended to emphasize the fact that enthusiastic public attention can
induce a permanent rise in share prices, even though no genuinely new information is presented.

EntreMed is a biotechnology company traded on the NASDAQ. It is a small company with rights
to commercialize a potentially cancer-curing process. Major movements in its stock price
occurred on November 28, 1997, May 4, 1998, and November 11, 1998. In the scientific journal
111

112

Nature on November 27, 1997, was announced that EntreMed was engaged in finding a cure for
cancer, a resistance-free cancer therapy-. Inside the article a breakthrough in cancer res
Dr. Folkman scientist was reported. In the same issue, an article
earch by
stated: the results are
nanticipated and could herald a new era of cancer treatment.
of
stock
en 27
NMDs stock traded between 9.875 and 15.25, with
nnualized volatility equal to 81%.
s

particularly ENMD received tremendous attention in the
ational media in subsequent weeks.
ng May 4,
ing the same period the
ASDAQ Combined Biotechnology Index lost about 24% of its value.
24.875. However, that price was still twice the closing price
rior to the Times article of May 4.
to the
s
deed,
rategic partnership with ENMD, had a significant benefit from the announcement of May 4.
u

Reports of this discovery appeared also in the popular press, such as the New York Times and
Newsday on November 27, 1997, as well as in the electronic media, such as CNNs MoneyLine
and CNBCs Street Signs. It seems that an effort was made to bring the news to the attention
circles wider than the scientific community. On November 28, ENMD itself issued a press
release that covered the news and the companys licensing rights. The closing price of its
was 11.875 on November 26, and 15.25 on November 28; thus, the news caused a price
appreciation of 28.4%. The unusually high trading volume on November 28 and December 1
(figure 2) indicates that the market paid attention to the specific news. In the period betwe
November 1997, and 3 May 1998, E
a

Five months later, on Sunday, May 3, 1998, the New York Times article presented virtually the
same information more prominently, with a label A special report. It had favorable comment
from various experts, including a Nobel Prize winner. As a result, ENMDs stock, which had
closed at 12.063 on the Friday before the article appeared, opened at 85 and closed at 51.81 on
Monday, May 4. The Friday-close-to Monday-close return of 330% was extremely unusual.
Not surprisingly, the Times article and
n

At this point it has to point out that, although some of the May 4-price increase was temporary, a
substantial portion of it was permanent. EntreMeds stock price fell in the days followi
to close the week at 33.25, still almost three times higher than its price a week earlier.
Furthermore, it did not fall below $20 till late August 1998, and by late fall it had not closed
below $16.94, which was 40% higher than its May 1 price. However, dur
N

On November 12, 1998, a front page article in the Wall Street Journal reported that other
laboratories had failed to replicate Dr. Folkmans results. ENMDs stock price decreased on that
day considerably by 24% to close at
p

Moreover, the authors of the paper point out that the news about a breakthrough in cancer
research affected not only the stock of the firm that had direct commercialization rights
development but it also caused spillover effects, so other firms could benefit from this
innovation. They mention that the optimism was contagious since members of the NASDAQ
Combined Biotechnological Index reacted similarly in direction, if not in magnitude. In
seven members of the Index had returns that exceeding 25%, after the event of May 4.
Furthermore, Bristol-Myers Squibb (BMY), a major pharmaceutical firm that had forged a
st

112

113

The authors of the paper believe that this market behavior can probably be explained by
analyzing the investors features. If they were not expert in the sector, they most likely read the
second article, rather than the first in Nature. This confirms that the announcement of an
more significant than the event itself. Eventually, the authors conclu
event is
de that prices probably move
n no new news, and the movements may be concentrated on stocks that have common
It shows clearly the peak in ENMD price on the 4 of May 1998. After
at day, the price decreases again but it stabilizes at a significantly higher level than that in the
igure1: EntreMed Price


o
characteristics, but not necessarily similar economic fundamentals.

The following figure indicates the excessive changes in the price of the EntreMed from J anuary
1997 to December 2000.
th
th
period before the event.

F

0
10
20
30
40
50
60
1/01/97 5/21/97 10/08/97 2/25/98 7/15/98 12/02/98
ENMD_PRICE
113

114

The next figure shows the significant change in the volume of EntreMed. Indeed, the relevance
of the market reaction is even more evident, considering the traded volumes during the examined
period. The volume of ENMD security had a sudden huge increase after the announcement.
Although the price changes are excessive after the date of the event, the changes in traded
volumes are insignificant. It can be claimed that the path of volume follows more closely the
returns path.

Figure 2: EntreMed Traded Volume
0
4000
8000
12000
16000
20000
24000
1997:01 1997:07 1998:01 1998:07 1999:01
ENMD_VOL




















4. Empirical Results
4.1 Data
In this part an event study analysis will be conducted in order to check if the events on May 4,
1998 and on November 28, 1998 had an effect on the stock price of EntreMed. Daily data on the
stock price and volume of ENMD and the prices of the NASDAQ Biotechnology Index and of
the broader NASDAQ Composite has been obtained from Datastream. The data set covers the
period from 1 J anuary 1997 to 1 April 1999. The total sample includes 250 observations. Returns
are calculated as ln-returns where R
t
=ln(p
t
/p
t-1
). The NASDAQ Biotechnology Index consists of
firms engaged in bio-medical research to develop new treatments and cures for diseases. To enter
the Index, a firm must have a minimal market capitalization of $50 million. The timing of the
data set needs to be rescheduled in order to consider the event date as the reference point in time.
The event date becomes time zero, and all the other dates are rearranged by adding or subtracting
one for any day after or before the event date.

4.2 Results from Event Study
4.2.1 Main Event May 4, 1998 Event
One consider an event window of 25 days (26 including the event date itself), from 13/04/1998 (-
15) to 18/05/1998 (+10). The estimation window that is commonly used is around 3 months. In
114

115

this case the estimation period of 90 days is from 17/11/1997 (time -120) to 13/04/1998 (time
16). In this period the returns seem to be quite stationary.

The returns of the two indices (the generic NASDAQ Composite Index (NASCOMP) and the
more specific of the sector NASDAQ Biotechnology Index (NASBIOT)) are used as market
returns. While their behavior is in general similar, on the event date (day 0) NASBIOT presents
higher return than NASCOMP. This is obvious since the impact of the event on the components
of the Biotechnology Index is more precise and higher than on the NASCOMP.

In fact there was a significant optimism reflected in the prices movements, as a spillover effect
of the ENMD event. This means that using the NASBIOT in equation (2) higher abnormal
returns are taken than using the NASCOMP. If the purpose is to analyze the ENMD story
compared to the other companies of the sector, the more appropriate index would be the
NASBIOT, as it represents the general biotechnology sector returns. On the other hand, if we
aim at measuring the total effect of the event, we have to use the NASCOMP, since it is
representative of the whole securities market.

4.2.2 Empirical results
From the regression of ENMD returns using both the NASBIOT and the NASCOMP on the
period from 120 to 16, the following estimated parameters derived.

Table 1: Regression of ENMD returns on NASBIOT and on NASCOMP
Market Index o
|


R
2

NASBIOT 0.002021 0.683455 0.02254
NASCOMP 0.001607 0.463722 0.00992

One can notice from the table that the R
2
is very low in both cases, particularly in the
NASCOMP. This implies that the model has a low explanatory power and so there are other
relevant regressors that are important to explain ENMD returns. Moreover it is found that the
estimated coefficientso and are not significant. From this evidence one can conclude that the
expected returns calculated with (1) are lower than the true normal returns, and therefore the
abnormal returns will be overestimated.
|


Using the equation (2) one can calculate the abnormal returns on the period from 15 to +10. The
following table presents the mean and the standard deviation of the abnormal returns estimated
with the two regressors.

Table 2: Descriptive Statistics of Abnormal Returns
Market Index Mean of AR Standard Deviation of AR
NASBIOT 0.03862 0.29785
NASCOMP 0.03923 0.29984

Although the difference between the two cases is negligible, the abnormal returns estimated
using NASBIOT have lower mean. Particularly, on the date of the event, the abnormal return is
1.4536926 in the NASBIOT case and 1.454741 in the NASCOMP case.
115

116

At this point the null hypothesis of no impact of the event on the prices on the event date is
tested. Thus, the Ho is that the abnormal return on the event date should be zero, rather than
1.4536926.

Table 3 presents the variance of the abnormal returns and of the residuals, obtained from the
equation (3) and the asymptotic approximation. Furthermore, the table presents the results from
testing the null hypothesis on cumulative abnormal returns. The cumulative return for the sub
window (-1,0) is 1.4524 and for the (-1,+1).2639. For this test equation (5) is used.

The first test gives a strong rejection of Ho in the case of NASBIOT, confirming the importance
of the market reaction to the event. It is obvious that the Ho will be rejected for NASCOMP case
too, since the abnormal returns are higher on average. This implies that the market reaction to the
event is highly significant. Particularly, the announcement of the event on May 4, 1998, had a
considerable effect on the ENMD prices. As far as the sub windows are concerned, the
cumulative abnormal returns around the event are significantly different from zero, that is,
different from the expected cumulative returns.

Table 3: Test for the abnormal returns
Returns
2
(AR
t
)
2

5% normal value
AR on event day 8.7372 29.0311 1.645
CAR (-1,0) 14.5804 1.645
CAR on (-1,+1) 8.4591 1.645


4.2.3 November 12, 1998 Event
Conducting the same procedure, 105 days estimation window (from 28/05/1198 to 21/10/1998)
and 25 days event window (22/10/1998 to 26/11/1998) are chosen.

On the event date the ENMD return is 0.27118 while the return from the main event (4 May) is
1.45769. Only by comparing these two values, one can infer that the market did not react to that
event to the same degree. By making the tests it is found that the effect is significant, but much
more restricted.

4.2.4 Empirical Results
Running the same regression for this particular estimation period the following estimated
parameters are found.

Table 4: Regression of ENMD returns on NASBIOT and on NASCOMP
Market Index o
|


R
2

NASBIOT -0.002331 1.22112 0.221526
NASCOMP -0.001418 1.29197 0.218114

One can notice that, in this case, the explanatory power of the model is greater than before, since
the values of R
2
are higher. This means that, during this period, the market index explained more
adequately the returns of ENMD.

116

117

Using the 25-days-event window (15, +10) the abnormal returns of ENMD are calculated. The
following table presents the main descriptive statistics.

Table 5: Descriptive Statistics of Abnormal Returns
Market Index Mean of AR Standard Deviation of AR
NASBIOT -0.001468 0.080811
NASCOMP -0.005389 0.085461

As before, the average abnormal returns, in absolute value, are higher in the NASCOMP case.
However, on the event date, the abnormal return is 0.26899 using the NASBIOT as market
index and 0.26206 using the NASCOMP. The surprising thing is that, although on 12
November the NASBIOT return is positive, the ENMD returns and NASCOMP returns are
negative. This implies that this event did not affect the ENMD price and the whole
biotechnology sector in the sane way. It can also be claimed that perhaps the event is irrelevant
in explaining this difference. Thus, since the two indices present different behavior both of them
are tested.

Table 6: Test for the abnormal returns
Returns
2
(AR
t
)
2

5% normal value
NASBIOT
AR on event day -2.42068 -5.06182 1.645
CAR (-1,0) -2.81754 1.645
CAR on (-1,+1) -2.38944 1.645
NASBIOT
AR on event day -2.330396 -4.92166 1.645
CAR (-1,0) -2.81089 1.645
CAR on (-1,+1) -2.34349 1.645

Since the absolute values of the calculated numbers are greater than the normal values with 5%
significance level, the null hypothesis of no impact of the event is rejected. However, in this
case, the rejection is not so strong. Particularly, in the case of abnormal returns the rejection is
very weak. It has to be mentioned that using the correct form of the variance, rather than the
asymptotic approximation, the null in the case of cumulative abnormal returns is most likely
accepted, like in the case of abnormal returns. Therefore, one can conclude that although the
reaction of the market to the first event was absolutely significant, the results in the second event
are ambiguous.

5. Conclusions
This paper focused on the examination of the effect of a specific event on the stock price of a
particular firm. The firm of interest is EntreMed, a biotechnology company traded on NASDAQ.
The main event (4 May 1998) constituted the announcement of the discovery of a new cure for
cancer by ENMD. Furthermore, there were two other significant events that influenced ENMD
price. Prior to the main event, there was a publication five months earlier (November 1997)
mentioning the same fact. The third event, about 6 months after the main event, denied, in some
way, the initial announcement, by stating that the new cure could not be replicated by other
research.
117

118


Conducting an event study on the main and third event, some significant conclusions were
reached. It was found that the reaction to the positive event was strong and significant. On the
other hand, the negative reaction to the next event was weaker and almost insignificant. This
considerable difference can be explained by taking into account the type of the journal in each
case. It could be argued that an article published on Sunday NY Times is undoubtedly more
popular than the last article.

Finally, it should be mentioned one very surprising fact. Although the first publication on 12
November 1997 was the true event, it did not provoke significant impact on the market. Instead,
the event of May 4, which was actually a non-event, was treated as the main event and
therefore it caused outstanding effect on the firms price and generally on the whole sector. This
is the reason that Huberman and Regev state in their article (2001): an enthusiastic public
attention may induce a permanent rise in share prices even when no genuinely new information
is presented. Therefore, one can conclude that what is relevant for the market is not the event
itself but the common knowledge of that event among investors. In conclusion, in order to
examine the impact of an event on the market prices, one has to take into account the timing and
the extent to which this event has been revealed to investors.

References
Campbell, J ohn Y., Lo, Andrew and MacKinlay, A Craig (1997), The Econometrics of
Financial Markets, Princeton University Press
Copeland, Thomas E. and Weston Fred J . (1988), Financial Theory and Corporate Policy
Addison Wisley, 3
rd
Edition
Elton, Edwin J . and Gruber, Martin J . (1995), Modern Portfolio Theory and Investment
Analysis, 5
th
edition, J ohn Wiley & Sons, Inc
Grinblatt, M. and Titman S. (1998), Financial Markets and Corporate Strategy, McGraw Hill
Huberman, Gur and Tomar Regev (2001), Contagious Speculation and a Cure for Cancer: A
Non-Event that Made Stock Prices Soar, J ournal of Finance, 61(1), pp. 387-396
Mackinlay, A Craig (1997), Event Studies in Economics and Finance, J ournal of Economic
Literature, 25, pp. 13-39
118

Copyright of Culture & Religion Review Journal is the property of Franklin Publishing Company and its
content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's
express written permission. However, users may print, download, or email articles for individual use.

Você também pode gostar