1
Effects of Harmful Environmental Events on Reputations of
Firms*
Kari Jones
Department of Economics
Emory University
Atlanta, GA 30322-2240
email: kjone01@emory.edu
Paul H. Rubin
Department of Economics
Emory University
Atlanta, GA 30322-2240
email: prubin@emory.edu
Abstract: Many previous event studies have found unexpectedly large losses to firms involved in
negative incidents. Many of these studies’authors explain such losses as “goodwill losses” or
"reputation effects." To test this hypothesis, we search for residual losses (in excess of direct
costs) to firms involved in events which produce ill will, but do not affect the quality of their
final products nor break implicit labor or supply contracts. We find an overall insignificant
capital market response to a sample of 98 negative environmental events (representing all such
incidents reported in the Wall Street Journal between 1970 and 1992 in which electric power
companies or oil firms with listed stocks were involved). Although others have found similar
outcomes for more limited samples, our results enhance previous research by extending similar
findings to a broader range of environmental incidents over a longer time period. Further, our
findings suggest that the large residual losses of other studies may be due to reputation (and not
measurement errors or event study idiosyncrasies), but only when the notion of "reputation
effects" is limited to punishment solely by those who are directly harmed by the firms' conduct.
*
We would like to thank Owen Beelders, Robert Carpenter, Joel Schrag, and Susan Griffin for providing helpful
suggestions and comments. Additionally, we thank Xiaolan Wang of the Emory University Goizueta Business
School for carefully, quickly, and patiently fulfilling our requests for CRSP data. Errors and omissions are
attributable solely to the authors.
2
INTRODUCTION
Empirically, researchers have found large unexplained capital market losses to firms involved
in “negative” incidents. Implicitly and explicitly many authors attribute these residual losses to
attrition of reputational capital, often for lack of another explanation. Theoretical models of
reputation describe retaliation or punishment by a firm’s contractual partners when the firm
deviates from: an implicit agreement on quality of its product (by consumers); a profitmaximizing strategy (by shareholders); an implicit labor contract (by employees); or an implicit
purchasing agreement (by suppliers). However, some of the authors of the original event studies
(and those commenting on them) also imply that reputation may include punishment by the
firm’s contractual partners for harm done to others. Mainstream theoretical models of
reputational mechanisms do not predict that one group will punish when another group is
harmed. (Unless the group doing the punishing expects a positive probability of being harmed if
the firm’s devious behavior continues, punishment requires that the harmed group’s well being
enter the punishing group’s utility functions.)
The factors driving the unexpected results of previous event studies are yet to be explained.
Policy decisions and academic questions depend on correctly identifying the situations in which
reputational mechanisms are present. This paper provides evidence that reputational
mechanisms are a plausible explanation of the unexplained residual losses, when reputation is
defined traditionally – that is, only those who are (potentially) harmed incur the costs of
punishment.
3
The paper is organized as follows: Section I contains an overview of previous event studies
that relate to the question of reputation effects. Section II includes an explanation of these past
results within the traditional models of reputation and reviews examples from the academic
literature and popular press of the widespread assumption that a firm’s social reputation affects
its capital market value. We outline an empirical test of this assertion. Section IV includes a
discussion of the empirical procedure and the data. Finally, Section V presents the empirical
results and conclusions.
I.
UNEXPLAINED PAST EMPIRICAL FINDINGS
Unexplained losses have been found in a variety of previous event studies1 designed to
measure the effect (and effectiveness) of regulation on a firm or industry2, and more recently, the
effect of an even wider variety of nonregulatory events. Among the regulatory event studies,
significant capital market losses are associated with firms’involvement in Federal Trade
Commission censures for false and deceptive advertising (Peltzman (1981) and Mathios and
Plummer (1989)), government-ordered drug, automobile, and other product recalls (Jarrell and
Peltzman (1985), Hoffer, et al. (1988), Rubin, et al. (1988), and Bosch and Lee (1994)), other
product-safety-related regulatory (and private) actions (Viscusi and Hersch (1990)), Equal
Employment Opportunity (EEO) violations (Hersch (1991)), Occupational Safety and Health
Administration (OSHA) violations (Davidson, et al. (1989) and Fry and Lee (1989)), and
corporate crimes such as fraud and price fixing (Cloninger, et al. (1988), Karpoff and Lott
(1993), and Reichert, et al. (1996)).
1
2
Section IV contains an overview of the methodology for readers unfamiliar with the procedure.
See Schwert (1981) for an overview
4
Nonregulatory events also lead to large and often unexplained losses. These include the
changing of a product’s formula (Benjamin and Mitchell, undated manuscript), the Tylenol
poisonings (Mitchell (1989) and Dowdell, et al. (1992)), and airline crashes (Mitchell and
Maloney (1989) and Chalk (1987)). While these studies show the negative effects of a bad
reputation, one study provides evidence that good reputations can increase a firm’s value.
Chauvin and Guthrie (1994) found that firms experienced a statistically significant average gain
in market value from appearing on Working Mother magazine’s list of “best” employers.
In most of these studies the monetary losses to stockholders are shown to significantly
outweigh the direct and estimated indirect costs of the incidents. These results are surprising to
many authors. Peltzman (1981) characterized his findings as “amazing” and “… a mystery” (at
418), while Rubin, et al. (1988) label their extremely significant findings “surprisingly large” (at
37). For lack of a better explanation many authors characterize the residual losses (above and
beyond explainable costs) as losses of reputation or goodwill. Dowdell, et al. (1992), Jarrell and
Peltzman (1985), and Rubin, et al (1988) characterized the excess losses in their studies as losses
of a firm’s goodwill. Mitchell and Maloney (1989) dubbed their residual losses a “brand name
effect”.
Previous event studies involving environmental events show mixed results. Muoghalu, et al.
(1990) found that in hazardous waste lawsuits that allege damages from improper hazardous
waste disposal, defendant firms suffered significant losses. However, Harper and Adams (1996)
found that the average market reaction of a firm’s stock upon being named a potentially liable
party in a Superfund cleanup effort was not significantly different from zero. Laplante and
Lanoie (1994) found that for negative environmental events reported in the media, Canadianowned firms did not experience significant declines in stock market value either when an
5
environmental violation was announced or when a lawsuit was filed. This is consistent with the
small average penalty paid. The authors found that significant market adjustment occurred only
after a suit settlement was announced. It is not known if the firms experienced residual losses.
Karpoff, Lott, and Rankine (1998) found a statistically significant average loss of 0.85 percent to
firms involved in negative environmental incidents. The average loss was greater for events
where initial press reports of the incident occurred either at the allegation date or the date charges
were filed. However, when comparing these figures with the direct costs of the incidents, the
authors found no evidence that any part of this loss could be attributed to reputation effects.
Hamilton (1995) investigated the stock market effects of information releases concerning a
firm’s pollution activities. Manufacturing facilities must report annual releases of chemicals to
the EPA. This information is relayed to the public in a database called the Toxics Releases
Inventory (TRI). Hamilton studied both how the media treated TRI information and what affects
it had on stock prices of polluting firms. He found that on the day of the information release,
firms suffered, on average, a statistically significant drop in capital market value. This loss was
higher the greater the number of chemicals per facility. Capital market losses were also
positively correlated with number of Superfund sites. Hamilton’s explanatory regressions
suggested that potential liability and other direct monetary exposure issues, rather than consumer
forces, drove these losses.
Finally, Blacconiere and Patten (1994) reported that Union Carbide took a 27.9 percent, or
approximately $1 billion, hit from the Bhopal chemical leak, while its industry rivals suffered an
average 1.28 percent loss in capital market value. The authors attributed these losses to
investors’revisions of possible production-side risks and increased regulatory exposure. A
summary of these event studies is presented in Table 1.
6
II.
EXPLAINING PAST EMPIRICAL RESULTS
Traditional Theories of Reputation
Some of the studies reviewed in Section I attempt to explain their estimated capital market losses
by regressing them on study-specific potential explanatory factors. Others offer ad hoc
explanations of possible factors affecting the magnitude of losses. However, none attempts to
formally model the reputational mechanism at work when the authors refer to “reputation
effects”.
Traditional theories of reputational mechanisms have their roots in the concepts that are
articulated in Akerlof (1970), Klein, Crawford, and Alchian (1978), Klein and Leffler (1981),
Nelson (1970), and Nelson (1974), and are modeled formally in Shapiro (1983). Akerlof (1970)
notes that in some situations of asymmetric information between buyers and sellers, mutually
beneficial trades may be precluded by the prospects of cheating. Subsequently, economists
began describing and modeling the methods that have evolved in such markets to mitigate3 this
problem. In particular, firms often use reputation to guarantee product quality.
Klein, et al. (1978) first suggested that potential cheaters might offer a forfeitable hostage to
guarantee performance in interfirm contracting. Klein and Leffler (1981) applied this concept to
the consumer-producer relationship. High prices signal high quality, but consumers pay these
prices only if they receive some guarantee of high quality. Producers of high quality goods make
firm-specific investments that are forfeited if consumers discontinue purchases of the firm’s
output. Consumers realize that firms are unlikely to deviate from high quality because continued
high quality production allows them to recoup their investments in “hostages”4.
3
As Shapiro (1982) notes, full-information quality levels are unattainable even with a reputation mechanism.
Other investigations into the nature of the outcomes of various reputation mechanisms are found in the signaling
models of game theorists. See, for example, Allen (1984), Kreps and Wilson (1982), Milgrom and Roberts (1986),
and Kihlstrom and Riordan (1984).
4
7
Thus, if a firm deviates from a quality commitment (or even if an incident occurs that signals
without one-hundred-percent certainty that the firm has deviated), customers punish the firm by
lowering their willingness to pay. The net present value of the future profit stream declines and
the capital market value of the firm falls.5 These notions of stochastic and nonstochastic signals
of deviations from prior commitments can be generalized to include the firm’s reputation with its
other contractual partners, namely its employees and suppliers. If a firm cheats in some way on
a commitment it has to a supplier, the firm is likely to incur higher input costs in the future.
Firms that cheat their employees may face wage premiums to attract future workers in a
competitive labor market.
Finally, firms also have reputations with investors. A firm’s involvement in a negative
incident may lead investors to reevaluate their faith that the firm’s management will steer clear
of such costly events in the future. Investors may also revise their subjective probabilities of
tighter future regulatory scrutiny or additional regulations. The higher the perceived risk of such
consequences, the lower the firm’s expected future profit stream, and the lower the share price.
In summary, the comments concerning the findings of previous event studies, framed in the
theory of reputation, suggest that “reputation effects” are transmitted several ways.
(1) The incident may result in a downgrading of the firm’s reputation with its customers,
employees, or suppliers. Depending on the nature of the implicit commitment, this loss of
reputation may result from
(a) a deviation from expected behavior (in a nonstochastic setting), or
(b) consumers’, employees’, or suppliers’revision of their estimates of the probability that
the firm has cheated (in a stochastic setting).
5
Ippolito (1992) reported in a study of the mutual fund market that consumers rationally react in this manner, thus
preventing a “lemons” market.
8
(2) The firm also may suffer capital market losses if the event leads to decreased faith in the
firm’s management on the part of investors. This may be the result of
(a) perceived increased risk of future incidents, and/or
perceived increased regulatory exposure.
While these reputational factors are reasonable partial explanations of observed residual
losses, they are incomplete interpretations for at least two reasons. First, no empirical evidence
exists to show that a combination of (1) direct costs, (2) consumers’response to deviations from
expected product quality, (3) employees’response to deviations from implicit employment
contracts, (4) suppliers’reactions to deviations from implicit purchasing agreements, and (5)
investors’decreased faith in management explain all of the losses reported in empirical studies.
In fact, both Peltzman’s work, (1981) and Peltzman’s and others’comments suggest otherwise.
Second, suggestions of the importance of social reputation in explaining these residual losses are
pervasive in the literature. That is, it is widely suggested that the firms’contractual partners
punish firms for harm done to others. Without a unified model of reputation (consistent with the
stylized facts), the social reputation hypotheses, while not compatible with economic theory,
continues to carry as much weight as the other ad hoc explanations of residual losses.
The (Conjectured) Social Component of Reputation
The foundations of the case for social reputation are found in suggestions and anecdotes in
the literature and also in the results of experimental economics. Many of the original event
studies’authors suggest that a firm’s social reputation can affect sales6. Some quotes are more
direct. Hersch (1991) suggests that in the aftermath of an EEO violation “costs include …
6
In contrast, Karpoff and Lott (1993) find that firms that are penalized for committing frauds that do not affect
consumers, suppliers, or stockholders (such as paperwork errors) suffer no unexplainable stock market losses.
9
adverse publicity that might result in the loss of sales … ” (at 140). Muoghalu, et al. (1990), in
reference to residual losses suffered by firms involved in Superfund cases, state that “stockholder
losses … include … public ill-will resulting from the lawsuit or the dumping” (at 358, note 5).
Hanka (1992) noted that “image-conscious firms fear the reputation consequences of pollution”
(at 26). Davidson, et al. (1994) attribute some of the significant losses subsequent to OSHA
violations to “negative publicity for the firm”.
These comments suggest that consumers, employees, or suppliers7 punish firms for engaging
in practices that are “socially irresponsible”. The losses suffered from this type of retaliation
would be in addition to other direct and indirect costs of the incident, including the reputational
losses for failing to honor implicit contracts with consumers, employees, or suppliers, and any
losses from decreased faith in management. Thus, such punishments would create the
unexplained “goodwill” or “reputation” effects.
These types of comments cover not only the negative effects of a bad reputation, but also the
positive effects of a good social reputation. Chauvin and Guthrie (1994) state “… if investors
believe that customers will prefer purchasing goods and services from ‘good’employers, [the
positive returns to firms on Working Mother’s “best employer” list] may also reflect estimates of
the effect that labor market reputation may have on sales” (at 551). Schwartz (1968) states that
“(g)ifts which enhance the public image of a corporation can advantageously shift the demand
curve for the corporation’s product” (at 480). Navarro (1988) concluded that giving to charity
increases demand or decreases demand elasticity for a firm’s product(s). Analogously, the
7
While socially conscious investors may divest themselves of the stock of firms they consider irresponsible, this
phenomenon is unlikely to affect the observed stock market loss. Assuming an efficient market, the investors
interested only in risk and return will bid against each other for the divested shares until the price reflects the value
of the firm.
10
tarnishing of a firm’s image could decrease demand or increase demand elasticity for a firm’s
product(s).
Making the “social reputation” case even stronger is the anecdotal evidence that suggests that
customers gain utility and disutility from characteristics of firms’production and financing
processes. Consumers readily purchase recycled products (such as notebook paper and paper
towels) which function no better than nonrecycled paper products, yet command price premiums.
Investors put money in socially conscious mutual funds that pay lower returns for the risk than
their socially-disinterested counterparts. Rothchild (1996) states that “[o]ver the past 12 months,
the 39 ethical funds tracked by Lipper [Analytical] have returned 18.2% vs. 27.2% for the S&P
500” (at 197). Rogers (1996) translates this into a $57.5 billion price investors were willing to
pay to avoid investing in undesirable firms over the year.
Extra-legal social enforcement mechanisms exist to enforce a wide variety of desired
behaviors – both economic and social. One example is boycotts. Consumer groups boycott
producers, and often successfully change producer actions. A November 8, 1990 Wall Street
Journal article reports that at least 300 boycotts of producer policies occurred in 1991 alone. The
authors report that “facing embarrassing publicity, many companies have changed their policies
to pacify the boycotters”8. The article further reports that, according to a consumer poll in June
1991, 27% of consumers boycotted a product because of a manufacturer’s record on the
environment.
Talk is cheap, but firms’actions, based on such beliefs, are not. Causal observation of
business practices such as voluntary divestiture from South Africa, advertisements touting a
firm’s expenditures on the environment or “fairness” in hiring practices, appointment of
environmentalists to boards of directors, and emphasis in stockholder reports on politically
11
correct firm policies suggest that firms reap some reward from consumer or investor knowledge
of these practices. Arguably, such practices do not improve the efficiency of the actual
production process, but, because they are common practice, firms must expect that they add to
profits. A spokesman for Reebok International Ltd. claims “[m]ore and more in the marketplace,
… who you are and what you stand for is as important as the quality of the product you sell”
(Hayes and Pereira, 1990, at B1).
Adding to the fervor of these suggestions of social reputation are experimental economists’
reports of evidence of widespread principle-based behavior in laboratory tests9 and behavioral
models in the literature which are generated to be compatible with these experimental results10.
III.
TESTING FOR EVIDENCE OF A SOCIAL REPUTATION EFFECT
Both the formal definitions of reputation in the literature (see, for example Shapiro (1982))
and the formal discussions in the above event studies assert that “reputation effects” refer only to
the reputation of a firm’s output in the goods market, its reputation for its dealings with its
employees or suppliers, and its management’s reputation in the capital market11. However, the
informal discussion suggests that social reputation affects profit. These assertions have
significant consequences. Voters, regulators, and taxpayers make assumptions concerning
retaliatory behavior trends when voting for (or otherwise affecting) government involvement in
markets.
8
For example, Karpoff, et al. (1998) note that the U.S. Sentencing Commission is
Hayes and Pereira (1990)
See, for example, Camerer and Thaler (1995) and Kahneman, et al. (1986) in which players are willing to sacrifice
money to reward players who are kind to others in previous rounds and punish those who were previously unfair to
their opponents.
10
See, for example, Bolton (1991) in which relative payoffs matter and Rabin (1993) which allows for agents to
reward those who are kind and punish those who are unkind.
11
Chauvin and Guthrie (1994) note that “… in all theoretical work on reputations, reputations have economic value
because they improve the efficiency of markets… ” (at 546). Schwartz (1968) notes that “(h)istorically, economists
have tended to ignore private philanthropic behavior and to regard it as economically irrational” (at 479).
9
12
explicitly considering the existence of reputational effects from environmental incidents in
setting its sentencing guidelines for corporate environmental crimes.
The goal of the empirical section of this paper is to test whether, as copious comments
suggest, social reputations are a factor in the unexplained losses of previous event studies. Our
procedure is to test for residual losses to firms involved in negative events that do not harm the
firm’s contractual partners (except through direct losses from the incident), but do affect their
social reputation by harming third-parties. Negative environmental incidents represent such
events.
We employ standard event study methodology to determine the effect that a negative
environmental event has on the market value of a firm. Under an assumption of efficient
markets, capital market participants produce an unbiased estimate of the value of a firm
reflecting all available information. A firm’s abnormal return12 immediately following the
release of news concerning the firm is the market’s unbiased estimate of the costs (or benefits) of
the news.
Thus, we calculate and analyze abnormal returns to firms involved in negative environmental
incidents. Absence of residual losses suggests that agents do not punish firms for harm done to
others, while presence of significant residual losses suggests that measurable reputation effects
may result from deterioration of social reputation. If residual losses to such events are found, it
will remain to be shown that these losses represent punishment for social reputation, and not, for
example, a measurement problem common to all event studies. As such, we conjecture that
consumers’propensities to punish decrease as the cost of punishment increases, but a
measurement error or other extraneous factor should not vary with costs of punishment.
Therefore, we prepare to test any residual losses for social reputation effects by collecting a
13
sample of environmental mishaps caused by electric power companies and oil companies. The
utilities represent an industry with few substitutes (where punishment would be expensive and,
therefore, less likely), while the oil companies represent an industry with many close substitutes
(where punishment would be relatively cheaper, and, thus, more likely). Finally, we perform a
cross-sectional analysis of the observed abnormal returns.
IV.
THE EMPIRICAL MODEL AND PROCEDURE
The Event Study Data and Methodology13
A list of potential events was drawn from all negative environmental events suffered by oil
concerns and electric utilities between 197014 and 199215 as reported in The Wall Street Journal
Index. Potential events had to meet three criteria. First, the event must have had a negative
environmental impact as the result of the actions of an oil division or electric power producing
division of the firm. This is because, if consumers were distressed upon hearing of pollution by
Acme Chemical, but were unaware that Acme was a subsidiary of ABC Oil Co., they would be
unable to punish, if they were so inclined. Second, the event must not have affected the quality
of the firm’s physical product. For example, some oil firms have been charged with switching
leaded and unleaded fuel. Using the wrong type of fuel not only causes pollution, but also
inflicts costly damage to a car’s engine and exhaust system. In such cases, consumers’material
self-interest would lead to decreased demand; retaliation for any resulting pollution could not be
separated16. Third, because all of our stock returns data are from the CRSP tapes17, the target
12
defined as the difference between the firm’s observed return and its expected return
For a “how-to” guide of event study particulars, see Armitage (1995) and Peterson (1989).
14
The environmental movement was gathering steam and the EPA was created this year.
15
Environmental fervor seems to be related to political climate and this year ended a presidential administration.
16
Most event studies also exclude any events that had other news about the firm reported in the event window,
because the effects of such news are inextricably summed with the effects of the event of interest. Because large oil
13
14
firm’s returns must be available on the CRSP tapes for the entire estimation period and event
window.
We use as our event date, the day that news of the event appears in the Wall Street Journal
(WSJ), unless the report suggests a more appropriate date. When picking our event dates, we
carefully considered the fact that if WSJ announcement dates are used as event dates when they
do not correspond to the date of the precipitating incident, the market may have adjusted to news
of the incident before the event date, consequently biasing announcement date losses downward.
The information to which market participants are most likely to react is not always the
precipitating incident. The market reaction to an announcement of an investigation that has a 5%
chance of resulting in a $1M fine will be much different that the reaction to the commencement
of an investigation that has a 95% probability of resulting in a $1M fine. The best option is to use
the date most likely to contain the bulk of the market adjustment to the “event”. If the event
were of the former type, this would be at the time of the precipitating incident. If the event were
of the latter type, this would be at the time that news that an actual loss will occur (or, possibly,
is more likely) reaches the market.
The WSJ announcement date is the appropriate event date for our study for three reasons.
First, there is precedent for using this methodology. Many event studies have used the WSJ
announcement date as the event date. Secondly, most of the other studies with which we are
comparing our results employ this methodology. To allow meaningful comparisons of our
results with past studies’results, we follow comparable procedures. The third reason is that we
feel the WSJ gets it right most of the time. That is, the Journal’s report of news that affects the
firms are mentioned in the financial news on an almost-daily basis, we left all events with concurrent news in the
sample to get as much information as possible. However, we calculated the results for various subsets of events,
depending on the nature of the concurrent news.
15
market is reasonably accurate and timely. At a minimum, there is enough uncertainty left when
the WSJ publishes a news story that the information is still novel.
Helmuth, et al. (1994) found significant market reaction to corrections printed in the WSJ.
Firms realized significantly positive returns on the day that corrections representing “good news”
were reported, and suffered significant losses when corrections representing “bad news” were
printed. Further, there was no post-event rebound of value. As the authors note, these results
imply that the “corrections are at least partially unanticipated by the market” suggesting “that a
subset of market participants depend on the WSJ for financial information.”
The results of previous event studies (particularly previous environmental event studies), and
the seemingly-random actions of the EPA and environmental groups, suggest that being
implicated in an environmental incident often is associated with a low probability of further
action or concern. Karpoff, Lott, and Rankine (1998), at 25, conclude that penalties stemming
from environmental incidents are highly variable and not easy to predict. Thus, news of many
potential events is unlikely to be printed until some official action is taken. However, some
events (e.g. Valdez and Three-Mile Island) seem highly likely to generate further costs (and,
thus, were reported very close to the precipitating incident date)18.
The final sample of 98 events is summarized in Table A-1 of the appendix. Included are the
name of the firm, a description of the event, the date that news of the event first appeared in the
Wall Street Journal, and any other firm-specific news reported in the Journal on the days
surrounding the event.
17
The CRSP tapes are a database of stock prices and other securities information administered by the Center for
Research in Stock Prices at the University of Chicago.
18
Additionally, to test for upward bias of abnormal returns resulting from late reports of important stories,
we control for the timeliness of the report in our explanatory regression of these returns.
16
Our goal is to identify the firms’abnormal returns from these negative environmental
incidents. The abnormal return to a firm from the event is the difference between the firm’s
actual return and the return we would expect for the firm in the absence of the incident. We
employ the market model to obtain parameters for generating expected returns. Specifically, for
each event, using returns for days just prior to the event date, we obtain an OLS estimate of:
RETit = α i + β i MRETt + eit
where
RETit ≡ the percentage return to the target firm of event i at date t,
MRETt ≡ the percentage return at time t to the NYSE/AMEX value-weighted portfolio,
and
eit is random disturbance to the event i target firm’s return at date t. For each event, the et are
distributed with a mean of zero and standard deviation of σi2.
Accordingly, α is a constant and β is the systematic risk of firm i’s stock over the estimation
period (beta in the CAPM model). Time is indexed t, such that t=0 on the event date.
The timespan used to estimate α and β is referred to as the estimation period. We calculate
expected returns using the 199 days immediately prior to the event date. The resulting estimates
of αi and βi, denoted ai and bi respectively, are used to calculate expected returns to the target
firm of event i around the event date. The difference between the actual return on any day and
the expected return calculated for that day (the prediction error) is the day’s abnormal return:
ARit = RETit − (ai + bi MRETt )
where
ARit is the abnormal return to the target firm of event i on day t.
17
The average abnormal return to the entire sample on any day t, denoted AARt, is the average
of the abnormal returns on day t for each event in the sample:
n
AARt =
∑
i =1
ARit
n
,
where n is the number of firms in the sample.
The time span over which the market is assumed to fully adjust to news of the event is called
the event window. The abnormal return to the firm accumulated over this adjustment period is
our best estimate of losses (or gains) to the firm from the event19. This measure is called the
cumulative abnormal return. For a given event window [e.g. the event window consisting of
days t=-1 and t=0, denoted (-1,0)], the cumulative abnormal return for event i, denoted CARi, is
defined as the sum of the abnormal returns for each day in the event window:
b
CARi = ∑ ARit ,
t =a
where
a ≡ the first day in i’s event window, and
b ≡ the last day in i’s event window.
For any given event window [e.g. the (-1,0) window], the average cumulative abnormal return to
the full sample, ACAR, is the average over the sample of each CARi:
19
We assume that the market reacts immediately to news of the event. However, because it was unclear for many of
our events whether this news reached the market before 4:00 p.m. on the day before it appeared in the Journal, we
use the (-1,0) event window. For comparison, we also calculate results for the day t=-1, day t=0, day t=1, and the
(-5,0), and (-1,9) windows.
18
n
ACAR =
∑ CAR
i
i =1
n
.
For a given event window, the total monetary loss to the firm’s shareholders (in dollars) from
the event is simply the product of the firm’s CAR and the value of the firm’s shares outstanding
on the day t=-2.
Cross-Sectional Analysis of the Event Study Results
After we calculate the cumulative abnormal returns to each event, we investigate the factors
contributing to these observed abnormal returns. The (-1,0) window CARs for all events are
regressed on event-specific explanatory variables in an attempt to identify other causal factors
that may explain the event study results. The first explanatory variable is report type. This is a
proxy for how novel the news reported on the event date is to market participants. The second
explanatory variable, action type, accounts for the possibility that different types of accusers
impose different costs on polluters.
The third explanatory variable is concurrent news. Events are categorized by type of
concurrent news in the (-1,0,1) window. The concurrent news on day t=1 is included since this
news may have reached the market on day t=0. Inclusion of the concurrent news variable will
net out any biasing effects of other firm-specific news in the event window. The fourth
explanatory variable is time. It may proxy for environmental consciousness and regulatory
and/or prosecutorial fervor in the environmental sector. In an alternative specification we
dummy for presidential administration.
19
The fifth explanatory variable is industry. Dummying by industry controls for the possibility
that oil firms’and electric firms’reactions to new information are systematically different, either
because the utilities are rate-of-return regulated, and/or because their stocks are traded more
thinly, on average, than oil stocks. The sixth and final explanatory variable, firm size, controls
for the possibility that extremely large firms’abnormal returns test insignificant because losses
from the incident are dwarfed by the sheer magnitude of the firms’market values.
The explanatory regression results are obtained by an OLS estimate of
CARi = α + β1 REPORTC i + β 2 REPORTRi + δ1 ACTIONFi + δ2 ACTIONPi +
δ3 ACTIONM i + τ1 NEWSG i + τ 2 NEWSBi + τ 3 NEWSN i + ηTIMEi +
ϕ INDi + γ
SIZEi + ei ,
where
REPORTCi = 1 if first news of event i is reported at the time of the filing of the first suit, the
first official allegation, the first official warning of an impending suit or charge,
or a settlement concurrent with official charges,
=0 otherwise.
REPORTRi = 1 if first news of event i is an update on a present court case, a judge or jury ruling,
a penalty ruling, or a settlement (not concurrent with the accusation),
=0 otherwise.
ACTIONFi =1 if event i involves only a federal action,
=0 otherwise.
ACTIONPi =1 if event i involves only a private action,
=0 otherwise.
ACTIONMi =1 if event i involves a multi-party action,
=0 otherwise.
NEWSGi =1 if other firm-specific news reported in event i’s (-1,0,1) window is expected to
increase the firm’s value,
=0 otherwise.
NEWSBi =1 if other firm-specific news reported in event i’s (-1,0,1) window is expected to
decrease the firm’s value,
=0 otherwise.
NEWSNi =1 if other firm-specific neutral news or news with an uncertain affect on the firm’s
value is reported in event i’s (-1,0,1) window,
=0 otherwise.
TIMEi = the last two digits of the year of event i,
INDi =1 if event i involves an oil firm,
=0 if event i involves an electric utility, and
SIZEi = the total dollar value of the firm i’s outstanding shares at t=-2, measured in thousands of
1987 dollars.
20
V.
RESULTS AND CONCLUSIONS
Abnormal and Cumulative Abnormal Returns 20
The CARs for the (-1,0) window for each of the 98 events are presented in Table A-2 of the
appendix, along with information concerning the costs of each event. A summary of the average
abnormal returns (AARs) and the average cumulative abnormal returns (ACARs) for various
windows is presented in Table 2. The split of negative and positive returns is not significantly
different from what may occur randomly, and relatively few of the firms’stocks reacted
significantly to news of the incident.
When all events are considered, the ACAR to the full sample for the (-1,0) window is
positive. Additionally, there are eight significantly positive CARs. These results are
counterintuitive, because, even if reputation effects are not present, there are direct costs to the
incidents. Four events for which the target firm had significantly positive abnormal returns on
either day t=-1, day t=0, or in the (-1,0) window also had “good” firm-specific news in the event
window (events 39, 59, 72, and 82). In addition, four more events (#18, 19, 20, and 21) shared
an event date on which Mideast tensions threatened to delay a possible loosening of oil
importing restrictions by President Nixon. Three of these CARs are significantly positive. Only
one event with bad news in the window (#13) produced a significantly negative abnormal return
on t=-1or t=0 or in the (-1,0) window. Removal of these nine (possibly biasing) events from the
sample yields an insignificant 0.25 percent, or $882,000, average loss to the remainder of the
sample. When the 34 events with any concurrent firm-specific news in the event window (good,
20
A subset of fourteen events was used to test the effect of different combinations of estimation period (150-day or
200-day) and market return calculation (equally-weighted or value-weighted) specifications on the predictive power
of the market model. No one measure or class of measures produced a significantly higher correlation between
targets’returns and market returns. The subsample was also used to determine whether either a procedure to net out
21
bad, or neutral) are excluded, the average cumulative abnormal return is an insignificant 0.016
percent gain21. These figures represent gross losses. Residual losses were not calculated since
the ACAR is insignificant. That is, netting out the losses due to direct costs can only make an
individual event’s return a larger (less negative) number. Thus the ACAR, net of direct costs,
cannot be significantly lees than zero.
In the full sample results, the t=-1 and t=0 windows show a significantly negative and a
significantly positive AAR, respectively. However, this result appears to be driven by one
unusual incident. Event 9 involves a radiation leak scare at a Rochester Gas and Electric plant
only a few years after Three Mile Island. At first news of the incident (t=-1), a mass sell-off
resulted in a 13.5 percent drop in the firm’s value. However, when more information was
released the following day (t=0), the stock gained back almost nine percent (and the stock’s price
was back to its pre-event level within five trading days). When event 9 is removed from the full
sample (and each of our subsamples), the average returns in the t=-1 and t=0 windows are no
longer significantly different from zero. With or without event 9, the ACAR for the (-1,0)
window is insignificant, because the two days' reactions tend to cancel each other out.
Insignificant average cumulative abnormal returns in the (-5,0) window and (-1,9) window
suggest that the stock market’s adjustment to the event did not occur outside of the (-1,0) event
window. The insignificant (-5,0) window suggests that news of the event did not reach the
market just prior to the report date, while the insignificant (-1,9) window suggests that
adjustment to the new news did not occur with a lag.
the effects of industry-wide increased regulatory scrutiny or a procedure to net out the residual losses due to
decreased faith in management was warranted for the full sample. Neither was found to be of significant benefit.
21
Additionally, when events 18 and 20 are removed, the (-1,0) ACAR is an insignificant 0.232 percent loss.
22
Explanatory Regression Results
Our event study results suggest that abnormal returns to firms involved in negative
environmental events are random and average approximately zero. Compared to event studies of
other negative events that employ a similar methodology, we find a much smaller effect from
negative environmental incidents. To test whether some causal variable(s) could explain the
variation of CARs over the different firms and incidents, we regress the observed cumulative
abnormal returns for the (-1,0) window on explanatory variables for report type, action type,
concurrent news type, time, industry, and firm size.
Consistent with our event study results, our explanatory regression results also suggest that
abnormal returns to firms are random and can be explained as white noise. The adjusted R2 is
0.0605 and none of the regressors has explanatory power at a significance level of 0.20 or lower.
The Durbin-Watson statistic is 1.8962 and a plot of the regression residuals appears normal.
Two alternate specifications were tested. First, because categorizing type of news in the event
window is a somewhat arbitrary choice on the part of the researcher, we test if the results are
robust to classifying events as having either no news in the window or any news (good, bad, or
neutral) in the window. In the second alternative specification, we dummy for presidential
administration instead of employing the time variable. The alternative specifications also have
very little explanatory power and yield intercept and slope coefficients that are not significant.
The insignificance of the report type variables suggests that use of event dates postdating the
initial incident date does not significantly bias the sample average cumulative abnormal return
(ACAR) upward. This is accentuated by the fact that 70 events with first news at the time of the
charge are being compared to only 18 events (including both the Valdez spill and Three Mile
Island) with first news at the time of the precipitating incident. The insignificance of the
23
coefficient on the dummy for industry suggests that oil and electric firms do not have
significantly different market reactions due to regulatory or thin-trading issues. Estimated
intercepts and slope coefficients for each specification, along with corresponding t-statistics, are
reported in Table 3.
If, as some authors have suggested, there were a reputational penalty to negative
environmental events, we would expect to find significantly negative residual losses in our study.
Instead, the cumulative abnormal returns appear to be random. In keeping with the standard
models of reputation, we find no evidence of a negative reputation effect.
Our results are consistent with Harper and Adams’s (1996) finding that firms experienced
insignificant changes in value upon being named a potentially responsible party in a Superfund
cleanup. They are also consistent with Laplante and Lanoie’s (1994) finding that Canadian firms
did not experience significant losses from announcements of environmental violations or
lawsuits. Our results, however, extend this finding of insignificant22 market losses in
environmental event studies to a broader class of regulatory as well as nonregulatory events.
Our results are also consistent with the assertions and findings of Karpoff, et al. (1998) that
formal penalties for committing environmental crimes are random and that stockholders realize
this. Because our sample covers a period twice as long as Karpoff, et al.’s, (1970 through 1992
versus 1980 through 1991) we can be confident that their findings are not merely a figment of,
for example, the “80’s mentality”, but span several decades and presidential administrations.
22
Muoghalu, et al. (1990) unsurprisingly found significant gross losses because they considered only a specific type
of environmental incident that had (potentially large) direct costs. The authors did not report residual losses (net of
direct costs), so reputation effects are not known. Hamilton (1995) found that firms suffered significant losses when
information about chemical releases at their facilities was made public. However, in the author’s explanatory
regressions, losses increased with number of chemicals at a facility but not with level of emissions, indicating
potential future liability was more important than extent of current harm. Also, number of Superfund sites was
24
Conclusions
We find an overall insignificant stock market response to a sample consisting of all negative
environmental incidents (regulatory and nonregulatory) reported in the Wall Street Journal over
the 1970 to 1992 period involving oil firms and electric power companies with listed stocks.
Thus, our results suggest that other researchers’findings of insignificant reputation effects from
regulatory environmental events extend to a wider class of regulatory and nonregulatory
incidents than previously has been considered in the literature. These events affect firms' social
reputations, but not the quality of their output or their reputations with employees or suppliers.
Thus, our findings contradict the widely-asserted hypothesis that when firms develop negative
social reputations (that is, negative reputations from harming third parties), their unaffected
contractual partners will incur the costs of punishment. Rather, our findings affirm the
traditional models of reputational mechanisms, which are effective when harms and punishments
are limited to contractual partners. Moreover, our findings provide evidence that the large losses
from harmful events are in fact due to stock market expectations that contractual partners will
punish firms for imposing direct harms.
associated with increased loss, but extent of media coverage had an insignificant effect, again indicating that
response to potential risk exposure outweighed reactions to the firms’social reputation and goodwill.
25
BIBLIOGRAPHY
Akerlof, George A., “The Market for ‘Lemons’: Quality Uncertainty and the Market
Mechanism,” Quarterly Journal of Economics, 1970, volume 84, pp. 488-500.
Allen, Franklin, “Reputation and Product Quality,” Rand Journal of Economics, 1984, volume
15, pp. 311-327.
Armitage, Seth, “Event Study Methods and Evidence on their Performance,” Journal of
Economic Surveys, 1995, volume 8, pp. 25-52.
Benjamin, Daniel K. and Mark L. Mitchell, “Commitment and Consumer Sovereignty: Classic
Evidence from the Real Thing,” undated manuscript.
Blacconiere, Walter G. and Dennis M. Patten, “Environmental Disclosures, Regulatory Costs,
and Changes in Firm Value,” Journal of Accounting and Economics, 1994, volume 18, pp.
357-77.
Bolton, Gary E., “A Comparative Model of Bargaining: Theory and Evidence,” American
Economic Review, 1991, volume 81, pp. 1096-1136.
Bosch, Jean-Claude and Insup Lee, “Wealth Effects of Food and Drug Administration (FDA)
Decisions,” Managerial and Decision Economics, 1994, volume 15, pp. 589-599.
Camerer, Colin and Richard H. Thaler, “Anomalies: Ultimatums, Dictators and Manners,” 1995,
Journal of Economic Perspectives, volume 9, pp. 209-219.
Chalk, Andrew, “Market Forces and Aircraft Safety: The Case of the DC-10,” Economic
Inquiry, 1986, volume 24, pp. 43-60.
Chalk, Andrew J., “Market Forces and Commercial Aircraft Safety,” Journal of Industrial
Economics, 1987, volume 36, pp. 61-81.
Chauvin, Keith W. and James P. Guthrie, “Labor Market Reputation and the Value of the Firm,”
Managerial and Decision Economics, 1994, volume 15, pp. 543-552.
Cloninger, Dale O., Terrance R. Skantz, and Thomas H. Strickland, “Price Fixing and Legal
Sanctions: The Stockholder-Enrichment Motive,” Antitrust Law and Economics Review,
1988, pp. 17-24.
Davidson, Wallace N. III, Dan Worrell, and Louis T. W. Cheng, “The Effectiveness of OSHA
Penalties: A Stock-Market-Based Test,” Industrial Relations, 1994, volume 33, pp. 283-296.
Dowdell, Thomas D., Suresh Govindaraj, and Prem C. Jain, “The Tylenol Incident, Ensuing
Regulation, and Stock Prices,” Journal of Financial and Quantitative Analysis, 1992, volume
27, pp. 283-301.
26
Fry, Clifford L. and Insup Lee, “OSHA Sanctions and the Value of the Firm,” The Financial
Review, 1989, volume 24, pp. 599-610.
Hamilton, James T., “Pollution as News: Media and Stock Market Reactions to the Toxics
Release Inventory Data,” Journal of Environmental Economics and Management, 1995,
volume 28, pp. 98-113.
Hanka, Gordon, “Does Wall Street Want Firms to Harm Their Neighbors and Employees?”
manuscript, November 25, 1992.
Harper, Richard K. and Stephen C. Adams, “CERCLA and Deep Pockets: Market Response to
the Superfund Program,” Contemporary Economic Policy, 1996, volume 14, pp. 107-115.
Hayes, Arthur S. and Joseph Pereira, “Facing a Boycott, Many Companies Bend,” The Wall
Street Journal, November 8, 1990, pp. B1 & B7.
Helmuth, John A., Ashok J. Robin, and John S. Zdanowicz, “The Adjustment of Stock Prices to
Wall Street Journal Corrections,” Review of Financial Economics, 1994, volume 4, pp. 6977.
Hersch, Joni, “Equal Employment Opportunity Law and Firm Profitability,” The Journal of
Human Resources, 1991, volume 26, pp. 139-53.
Hoffer, George E., Stephen W. Pruitt, and Robert J. Reilly, “The Impact of Product Recalls on
the Wealth of Sellers: A Reexamination,” Journal of Political Economy, 1988, volume 96, pp.
663-670.
Ippolito, Richard A., “Consumer Reaction to Measures of Poor Quality: Evidence from the
Mutual Fund Industry,” Journal of Law and Economics, 1992, volume 35, pp. 45-70.
Jarrell, Gregg and Sam Peltzman, “The Impact of Product Recalls on the Wealth of Sellers,”
Journal of Political Economy, 1985, volume 93, pp. 512-536.
Kahneman, Daniel, Jack L. Knetsch, and Richard H. Thaler, “Fairness and the Assumptions of
Economics,” Journal of Business, 1986, volume 59, pp. S285-S300.
Karpoff, Jonathan and John R. Lott, Jr., “The Reputational Penalty Firms Bear from Committing
Criminal Fraud,” Journal of Law and Economics, 1993, volume 36, pp. 757-802.
Karpoff, Jonathan M., John R. Lott, Jr., and Graeme Rankine, “Environmental Violations, Legal
Penalties, and Reputation Costs,” October 23, 1998, working paper, Social Science Research
Network.
Kihlstrom, Richard E. and Michael H. Riordan, “Advertising as a Signal,” Journal of Political
Economy, 1984, volume 92, pp. 427-450.
27
Klein, Benjamin and Keith B. Leffler, “The Role of Market Forces in Assuring Contractual
Performance,” Journal of Political Economy, 1981, volume 89, pp. 615-641.
Klein, Benjamin, Robert G. Crawford, and Armen A. Alchian, “Vertical Integration,
Appropriable Rents, and the Competitive Contracting Process,” Journal of Law and
Economics, 1978, volume 21, pp. 297-326.
Kreps, David M. and Robert Wilson, “Reputation and Imperfect Information,” Journal of
Economic Theory, 1982, volume 27, pp. 253-279.
Laplante, Benoit and Paul Lanoie, “The Market Response to Environmental Incidents in Canada:
A Theoretical Analysis,” Southern Economic Journal, 1994, volume 60, pp. 657-672.
Mathios, Alan and Mark Plummer, “The Regulation of Advertising by the Federal Trade
Commission: Capital Market Effects,” from Research in Law and Economics, 1989, volume
12, edited by Richard O. Zerbe, Jr., JAI Press, Greenwich, Connecticut, pp. 77-93.
Milgrom, Paul and John Roberts, “Price and Advertising Signals of Product Quality,” Journal of
Political Economy, 1986, volume 94, pp. 796-821.
Mitchell, Mark L. and Michael T. Maloney, “Crisis in the Cockpit? The Role of Market Forces
in Promoting Air Travel Safety,” Journal of Law and Economics, 1989, volume 32, pp. 329356.
Mitchell, Mark L., “The Impact of External Parties on Brand-Name Capital: The 1982 Tylenol
Poisonings and Subsequent Cases,” Economic Inquiry, 1989, volume 27, pp. 601-618.
Muoghalu, Michael I., H. David Robison, and John L. Glascock, “Hazardous Waste Lawsuits,
Stockholder Returns, and Deterrence,” Southern Economic Journal, 1990, volume 57, pp.
357-370.
Navarro, Peter, “Why Do Corporations Give to Charity?” Journal of Business, 1988, volume 61,
pp. 65-93.
Nelson, Phillip, “Advertising as Information,” Journal of Political Economy, 1974, volume 82,
pp. 729-754.
Nelson, Phillip, “Information and Consumer Behavior,” Journal of Political Economy, 1970,
volume 78, pp. 311-329.
Peltzman, Sam, “The Effects of FTC Advertising Regulation,” Journal of Law and Economics,
1981, volume 24, pp. 403-448.
Peterson, Pamela P., “Event Studies: A Review of Issues and Methodology,” Quarterly Journal
of Business and Economics, 1989 volume 28, pp. 36-66.
28
Rabin, Matthew, “Incorporating Fairness into Game Theory and Economics,” American
Economic Review, 1993, volume 83, pp. 1281-1302.
Reichert, Alan K., Michael Lockett, and Ramesh P. Rao, “The Impact of Illegal Business
Practice on Shareholder Returns,” The Financial Review, 1996, volume 31, pp. 67-85.
Rogers, T. J., “Profits vs. PC,” Reason, October 1996, pp. 36-43.
Rothchild, John, “Why I Invest with Sinners,” Fortune, May 13, 1996, p. 197.
Rubin, Paul H., R. Dennis Murphy, and Gregg Jarrell, “Risky Products, Risky Stocks,”
Regulation, 1988, pp. 35-39.
Schwartz, R. A., “Corporate Philanthropic Contributions,” The Journal of Finance, 1968, volume
23, pp. 479-497.
Schwert, G. William, “Using Financial Data to Measure Effects of Regulation,” Journal of Law
and Economics, 1981, volume 24, pp. 121-158.
Shapiro, Carl, “Consumer Information, Product Quality, and Seller Reputation,” Bell Journal of
Economics, 1982, volume 13, pp. 20-35.
Shapiro, Carl, “Premiums for High Quality Products as Returns to Reputations,” The Quarterly
Journal of Economics, 1983, volume 98, pp. 659-679.
Viscusi, W. Kip and Joni Hersch, “The Market Response to Product Safety Litigation,” Journal
of Regulatory Economics, 1990, volume 2, pp. 215-230.
29
TABLE 1
Summary of Previous Event Studies
Event Type
Average Loss in
$
Average Loss
as Percentage
of Value
Cite(s)
FTC cease-and-desist order for
false advertising
Automobile recalls
as high as $250M for
larger firms
n.a.
3.2 – 6.4
FDA-mandated drug recalls
FDA disciplinary actions
Drug packaging regulations
(subsequent to Tylenol
poisonings)
CPSC-mandated recalls
EEO violations;
EEO class-action suits
OSHA sanctions
$1.8M (residual)
n.a.
$310M
5.6 (residual)
2.22 – 3.42
11.83
$146M
$18.5M
$16M (residual)
$3.9 – 22.1M
5.4 – 6.9
0.29 – 0.48;
15.6
0.53 - 2.1
Corporate indictments
$9.15M (residual; not
sig.)
n.a.
1.38 – 3.1
Fry & Lee, 1989;
Davidson, et al., 1994
Reichert, et al., 1996
17
Cloninger, et al., 1988
$33.3M
1.2
Muoghalu, et al., 1990
n.a.
0.85 net loss to full
sample;
no residual
insignificant
Karpoff, et al., 1998
Laplante and Lanoie, 1994
Charges and pleas in pricefixing cases
Hazardous waste
mismanagement suits
Environmental violations
insignificant
Peltzman, 1981;
Mathios and Plummer, 1989
Jarrell & Peltzman, 1985;
Hoffer, et al. 1988
Jarrell & Peltzman, 1985
Bosch & Lee, 1994
Dowdell, et al., 1992
Rubin, et al., 1988
Hersch, 1991
Being named a potentially
responsible party at a
Superfund site
Violation of Canadian
environmental regulation
n.a.
Coca-Cola adopts New Coke
Criminal fraud against
stakeholders and government
Tylenol poisonings (first
wave)
Bhopal, India chemical leak
$500M
$40 - 60.8M
insig. for
announcement;
2.0 for settlement
8.1
1.34 – 5.05
$1.24B (residual)
14.3 (residual)
Mitchell, 1989
~$1B (Union
Carbide)
n.a. (Union Carbide’s
rivals)
$4.1M
$11 – 18.3M
(residual)
$21.32M
27.9 (Union
Carbide)
1.28 (Union
Carbide’s rivals)
0.28
1.34 – 1.55
(residual)
3.8
Blacconiere and Patten
(1994)
n.a.
0.28 – 2.2
Chauvin & Guthrie, 1994
Chemical releases reported
Airline crashes before and
after deregulation
Aircraft-manufacturer-at-fault
crashes
Included in Working Mother’s
“best” employers list
n.a.
Harper & Adams, 1996
Benjamin & Mitchell
Karpoff & Lott, 1993
Hamilton, 1995
Mitchell & Maloney, 1989
Chalk, 1987
30
TABLE 2
Summary of Results for Various Windows
All events:
Significantly-positive-tosignificantly-negative ratio
Total-positive-to-total-negativeratio
(Probability of getting the actual
split or even more skewed,
assuming a .5 probability of a
positive return)
AAR or ACAR - % return
(z-statistic)
Average change in $ value ($K)
t=-1
6:7
t=0
7:7
(-1,0)
8:10
(-5,0)a
7:6
(-1,9)b
4:7
45:53
(.48)
58:40
(.09)
52:46
(.61)
49:49
(1.00)
55:43
(.27)
-0.2171
(-2.1041)‡
-4239
0.3907
(2.1815)‡
8254
0.1736
(0.0538)
4018
0.2461
(0.0676)
-19506
0.3353
(0.1531)
61656
Without events with (1) concurrent good news and significantly positive returns or (2) concurrent bad news and
significantly negative returns:
t=-1
t=0
(-1,0)
(-5,0)a
(-1,9)b
5:7
2:7
2:9
4:5
3:7
Significantly-positive-tosignificantly-negative ratio
38:51
50:39
44:45
42:47
48:41
Total-positive-to-total-negative(.20)
(.29)
(.99)
(.67)
(.53)
ratio
(Probability of getting the actual
split or even more skewed,
assuming a .5 probability of a
positive return)
-0.327
0.0768
-0.25
-0.14
-0.142
AAR or ACAR - % return
(-2.6334)*
(0.3366)
(-1.6189)
(-0.7754) (-0.6279)
(z-statistic)
-3532
2520
-1010
-12803
42582
Average change in $ value ($K)
Without events with conflicting news and event 9:
t=-1
5:6
Significantly-positive-tosignificantly-negative ratio
38:50
Total-positive-to-total-negative(.24)
ratio
(Probability of getting the actual
split or even more skewed,
assuming a .5 probability of a
positive return)
-0.177
AAR or ACAR - % return
(-1.5991)
(z-statistic)
-3169
Average change in $ value ($K)
t=0
1:7
(-1,0)
2:8
(-5,0)a
4:4
(-1,9)b
3:7
49:39
(.34)
44:44
(1.00)
42:46
(.75)
47:41
(.59)
-0.023
(-0.3475)
2285
-0.200
(-1.3718)
-882
-0.056
(-0.5445)
-12724
-0.167
(-0.6806)
43005
* significant at 1% level
‡
significant at 5% level
a
These are estimates, because the estimation period and event window overlap for these calculations.
b
Because of data restrictions event 5 = (-1,5), event 7 = (-1,8), and event 9 = (-1,8).
31
TABLE 3
Explanatory Regression Results
Intercept term
or Coefficient on:
Intercept term
REPORTC
REPORTR
ACTIONF
ACTIONP
ACTIONM
NEWSG
NEWSB
NEWSN
Estimated Value
(t-statistic - 86
d.f.)
0.01761
(0.5490)
0.00660
(0.7247)
-0.00121
(-0.1065)
0.00710
(0.9664)
0.00212
(0.1813)
-0.01320
(-1.124)
0.00437
(0.6848)
-0.00439
(-0.3952)
0.00504
(0.4746)
NEWSC
TIME
-0.00039
(-1.056)
Estimated Value
(t-statistic - 88 d.f.)
0.01964
(0.6344)
0.00752
(0.8479)
-0.00006
(-0.0054)
0.00616
(0.8689)
0.00209
(0.1838)
-0.01355
(-1.192)
CARTER
REAGAN
BUSH
SIZE
0.00755
(1.112)
-2.2000E-11
(-0.1307)
-0.00632
(-0.5164)
0.00543
(0.5746)
-0.00138
(-0.1198)
0.00753
(0.9973)
0.00161
(0.1344)
-0.01374
(-1.143)
0.00514
(0.7947)
-0.00366
(-0.3249)
0.00671
(0.6216)
0.00303
(0.5642)
-0.00042
(-1.160)
FORD
IND
Estimated Value
(t-statistic - 83 d.f.)
0.00744
(1.113)
-3.2563E-11
(-0.1975)
0.00502
(0.4013)
-0.01106
(-1.127)
-0.01030
(-1.398)
-0.00387
(-0.4721)
0.00484
(0.6598)
-5.9286E-11
(-0.3450)
32
APPENDIX
TABLE A-1
Summary of Events
No.
1
2
Firm
Event
Description
Event
Date
Shell Oil
Standard Oil of
California (Chevron)
Detroit Edison
Platform explosion
Tanker collision
12/02/70
01/19/71
Soot release
08/28/73
Standard Oil of
California (Chevron)
Philadelphia Electric
Co.
Standard Oil of
Indiana (Amoco)
General Public
Utilities
Philadelphia Electric
Co. (Peco)
Rochester Gas &
Electric Co.
Long Island Lighting
Co. (Lilco)
Tank puncture
10/13/75
Radioactive release
10/10/77
Tanker spill
03/17/78
Radioactive release
03/29/79
Radioactive release
06/22/79
Radioactive release
01/26/82
Storage tank leak
02/12/85
11
12
13
Ashland Oil
Exxon
Exxon
Storage tank collapse
Tanker spill
Pipeline spill
01/05/88
03/28/89
01/03/90
14
British Petroleum
(BP)
Olin Corp.
Pipeline ruptured by tanker
02/08/90
Charged by Justice Dept.
with water pollution
02/10/70
3
4
5
6
7
8
9
10
15
Other firm-specific
news in event
window
On t-0, high temperatures
led to record electrical
demand; extra power
bought from Canada
On t=1, more
developments in case of
political football
Shoreham plant
On t=-1, seven plaintiffs’
lawyers appointed by
state and federal court
judges in AK to
committee to coordinate
litigation against
company over Valdez
spill
On t=-1, Judge reduced
three felony counts
against Valdez captain to
one
33
Table A-1 – continued
16
17
Public Service
Electric and Gas Co.
Texaco
Charged by Justice Dept.
with water pollution
Charged with by Justice
Dept. water pollution
Charged by Justice Dept.
with water pollution
02/10/70
18
Mobil Oil
19
Atlantic Richfield
Co. (ARCO)
Charged by Justice Dept.
with water pollution
02/19/70
20
Phillips Petroleum
02/19/70
21
Standard Oil of
Indiana (Amoco) &
its sub. American Oil
Co.
Charged by Justice Dept.
with water pollution
Charged by Justice Dept.
with water pollution
22
Florida Power &
Light
02/25/70
23
Standard Oil of
California (Chevron)
24
Consolidated Edison
of NY (Con Ed)
25
Standard Oil of
California (Chevron)
Nixon administration
threatened to file suit (and
later did) to stop construction
of a hot water canal
(potential thermal pollution)
FTC & California Air
Resources Board dispute
company’s claims
concerning the
environmental friendliness of
gasoline additive
Alleged thermal and
chemical pollution, first
news mentions dead fish
Oil platform explosion
leading to subsequent oil
slick
26
Standard Oil of New
Jersey’s sub. Humble Oil &
Refining Co.
Oil spill; tanker chartered,
but not owned or operated,
by company; subsequent
State suit
02/16/70
27
Shell Oil Co.
03/30/70
28
Standard Oil of New
Jersey’s sub. Humble Oil &
Refining Co.
Interior Secretary ordered
three drilling platforms
closed down for alleged
violations of safety
regulations
Private suit alleging
pollution caused physical
and mental harm to humans
and livestock
02/10/70
02/19/70
02/19/70
03/02/70
On t=-1, one of five firms
to relinquish rights to
dawsonite oil-shale tracts
On t=-1, one of five firms
to relinquish rights to
dawsonite oil-shale tracts
On t=0, elected a director
and a president of a sub.
On t=0, received a $3.2M
DSA contract for fuel oil
and gas
03/11/70
02/11/70
08/26/70
On t=0, one of three
refiners saying that they
will market unleaded fuel
when car makers produce
compatible engines
On t=-1, named a
president of a subsidiary
On t=+1, rights offering
for common shares
valued at $387M became
effective
On t=+1, Humble plans to
raise $50M by offering
abroad $30M of 5-year
notes and $20M of 15year debenture
34
Table A-1 – continued
29
Mobil Oil Corp.
Fined by NY state judge for
failing to halt discharge of
industrial waste
10/06/70
30
Mobil Oil Corp.
10/14/70
31
Standard Oil of New
Jersey’s sub. Humble Oil &
Refining Co.
32
Continental Oil
33
Union Oil of
California
34
Tenneco Oil Co.
35
Gulf Oil Corp.
36
Kerr-McGee Corp.
37
Mobil
38
Cleveland Electric
Illuminating Co.
Olin Corp.
U.S. attorney filed criminal
informations for water
pollution
Justice Dept. charged
company with knowingly
failing to provide subsurface
safety devices on offshore
wells
Charged by Justice Dept.
with failing to install and
maintain subsurface safety
devices on offshore wells
Charged by Justice Dept.
with failing to install and
maintain subsurface safety
devices on offshore wells
Charged by Justice Dept.
with failing to install and
maintain subsurface safety
devices on offshore wells
Charged by Justice Dept.
with failing to install and
maintain subsurface safety
devices on offshore wells
Charged by Justice Dept.
with failing to install and
maintain subsurface safety
devices on offshore wells
Charged by Justice Dept.
with failing to install and
maintain subsurface safety
devices on offshore wells
Charged by federal grand
jury with water pollution
Charged by federal grand
jury with water pollution
39
11/16/70
On t=+1, announced
plans to increase output
of some fuels to ease
anticipated winter
shortage
On t=-1, Libyan affiliate
agreed to raise posted
price of Libyan crude
On t=0, lifted wholesale
prices of some oils and
kerosene in East
11/23/70
11/23/70
12/23/70
12/23/70
On t=0, offered 8%
increase in 1971 wages,
6% in 1972, topping other
refineries
12/23/70
12/23/70
04/30/71
04/30/71
On t=0, second quarter
earnings to be
substantially ahead of
first quarter
On t=0, reported a
quarterly dividend was to
be paid on 06/09
On t=+1, received $4.3M
contract to operate Army
ammunition plant
35
Table A-1 – continued
40
Shell Oil
Charged by federal grand
jury with water pollution
04/30/71
41
Crown Central
Petroleum Corp.
05/04/71
42
44
Atlantic Richfield
Co. (ARCO)
Union Oil Co. of
California
Signal Oil Co.
45
Getty Oil Co.
46
Atlantic Richfield
Co. (ARCO)
Amerada Hess Corp.
Charged by FTC with
making false antipollution
claims about a gasoline
additive
Charged by Justice Dept.
with water pollution
Charged by Justice Dept.
with water pollution
Charged by Justice Dept.
with water pollution
Charged by Justice Dept.
with water pollution
Charged by Justice Dept.
with water pollution
Charged by Justice Dept.
with water pollution
Charged by Justice Dept.
with water pollution
Sued by Justice Dept. to
clean up an earlier oil spill
Sued by car dealer for
damage to his cars from air
pollution
Charged by Justice Dept.
with water pollution
Sued by environmentalists
for violating election laws by
setting up a “front group” to
oppose a proposed clean
environment statute to avoid
publicity
Oil spill
Pennsylvania Insurance
Commissioner issued
“Investor’s Guide to
Polluters” listing firms and
government units the state is
taking action against for
pollution so that insurance
companies could “use their
investment program to end
activities of polluters”
43
47
48
49
Coastal States Gas
Producing Co.
Texaco
50
Consolidated Edison
Co. (Con Ed)
51
Olin Corp.
52
Standard Oil of
California (Chevron)
53
54
Texaco
Sun Oil Co.
On t=-1, announced
$25M expansion of oil
refinery to produce
components for lead-free
gas
05/19/71
05/19/71
06/01/71
06/01/71
06/01/71
On t=0, ordered $13.2M
storage vessel
06/22/71
06/22/71
06/28/71
06/29/71
09/21/71
03/08/72
07/25/72
09/29/72
On t=+1, exec. VP retires
in ill health
36
Table A-1 – continued
55
Gulf Oil Corp.
56
Standard Oil of
California (Chevron)
57
Potomac Electric
Power Co.
58
Central Illinois
Lighting Co.
59
Union Electric Co.
60
Exxon Corp.
61
Texaco Inc.
62
63
Consolidated Edison
Co. (Con Ed)
Exxon Corp.
64
Tampa Electric Co.
65
Chevron USA
66
Gulf Oil Corp.
Pennsylvania Insurance
Commissioner issued
“Investor’s Guide to
Polluters” listing firms and
government units the state is
taking action against for
pollution so that insurance
companies could “use their
investment program to end
activities of polluters”
Issued notice of violation of
Clean Air Act by EPA for
constructing new storage
tanks without EPA approval
Notified by EPA that plants
were in violation of state
emissions regulations
Notified by EPA that plants
were in violation of state
emissions regulations
Notified by EPA that plants
were in violation of state
emissions regulations
Charged by Justice with
water pollution
09/29/72
Charged by Justice Dept.
with water pollution
Received notice of emissions
violation from EPA
Agreed to consent order with
EPA for dumping drilling
waste into sea
Florida Department of
Environmental Regulation
filed suit alleging company
is constructing a generator
without a permit
Ordered by Bay Area
Pollution Control District to
close a refinery because of
an earlier agreement to limit
pollution (later ruled by a
judge to be misinterpretation
of the agreement)
Agreed to proposed consent
judgement with a Justice
Dept./EPA regarding water
pollution at a refinery
12/02/74
02/06/74
06/10/74
06/10/74
06/10/74
On t=0, received rate
boost of $39.9M
12/02/74
On t=-1, one of 3
companies to tell AEC
they plan pilot plants to
make nuclear fuels
On t=-1, plans to deter
project to convert coal to
natural gas
04/02/75
On t=+1, asked for record
increase in electric rates
09/03/76
03/18/77
11/25/77
On t=-1, consortium
headed by co. placed
contract for outside firm
to operate CO2 pipeline in
SW Texas
01/23/78
On t=-1, a joint venture
involving Gulf hired a
contractor in development
of oil shale tract
37
Table A-1 – continued
67
Ohio Edison Co.
68
Philadelphia Electric
Company (PECO)
69
Carolina Power and
Light Co. (Carolina
P&L)
Ethyl Corp.
70
71
72
Cleveland Electric
Illuminating Co.
Atlantic Richfield
Co. (ARCO)
73
Southern California
Edison
74
Mobil Oil Corp.
75
Exxon Corp.
76
Commonwealth
Edison Co. (Com Ed)
77
Texaco
Sued by EPA for violation of
state’s air pollution
regulations
Plants found in violation of
air pollution control
standards by EPA
Fined by NRC for
improperly disposing of
radioactive materials
Charged in administrative
complaint by EPA with
using leaded gasoline in
company-owned vehicles
designed for unleaded fuel
Sued by EPA for violating
emissions standards at plants
EPA plans fine for using
leaded gasoline in company
cars designed for unleaded
fuel
08/03/78
Reached an out-of-court
settlement with the state air
resources board and local air
quality officials on a dispute
over emissions
Suit filed and settled with
Alabama over water
pollution charges
NY state environmental
officials asked NY state
attorney general to take
action for alleged water
pollution violations
Sued by Justice Dept. on
behalf of EPA to force an
expanded cleanup of toxic
spillage from electrical
equipment
Fined by EPA for water
pollution; no finding of
liability
03/11/82
11/24/78
08/08/80
On t=+1, placed contract
for power plant
simulator/training center
03/19/81
03/19/81
04/30/81
On t=-1, net income
figures out
On t=0, separately, but in
the same article ARCO
completed a 3-year
shipping agreement to
build an 80-mile pipeline
across Panama leading to
about $100M in shipping
tariffs over 3 years
09/28/82
10/17/83
On t=+1, company
mentioned in “Heard on
the Street” column; news
unclear
02/22/84
On t=+1, placed large
order with Swiss firm
08/22/85
On t=-1, another
company to buy Texaco’s
50% stake in $1.4B
copper-mining project in
Chile
38
Table A-1 – continued
78
Atlantic Richfield
Co. (ARCO)
Fined by EPA for water
pollution; consent decree
08/22/85
79
Chevron USA
Penalty judgement won by
EPA in civil suit for air
pollution
10/07/85
80
Ashland Oil Inc.
07/08/86
81
Texaco Inc.
82
Chevron USA
Civil complaint and consent
decree with Justice Dept.
over charges of persistent
water pollution at a refinery
Sued by EPA for air
pollution
Sued by US Attorney’s
office for water pollution
83
Atlantic Richfield
Co. (ARCO)
01/05/87
84
Atlantic Richfield
Co. (ARCO)
85
Chevron Corp.
86
Amerada Hess Corp.
87
Atlantic Richfield
Co. (ARCO)
Signed consent order with
NY State Department of
Environmental Conservation
to clean up pollution from a
defunct refinery (site is on
Superfund list)
Agreed to settle charges of
illegally releasing waste
water and sludge into LA
County water treatment
facility
Sued by US attorney on
behalf of EPA for violating
benzene rules
Sued by state of Alaska for
not preventing contamination
of shoreline from Valdez
spill
Sued by state of Alaska for
not preventing contamination
of shoreline from Valdez
spill
On t=-1, to sell one-third
interest in Stillwater
Mining Co. for $15M
On t=0, plans to offer up
to $200M of adjustablerate notes that mature in
9-48 months to raise
$1.2B
On t=-1, year-old
acquisition of company
used as a comparison to
current action of another
firm in “Heard on the
Street” column
On t=+1, SEC charged
the company and one of
its chairmen with foreign
bribery
07/16/86
08/27/86
On t=-1, mentioned in
“Heard on the Street” that
company’s stock would
be a good value if energy
prices turnaround
05/29/87
08/31/88
08/16/89
08/16/89
On t=0, company to
introduce low-emission
regular gas to So. Calif.
market on 09/01 that will
“significantly reduce air
pollution” from older cars
that don’t have catalytic
converters
39
Table A-1 – continued
88
British Petroleum Co.
89
Mobil Corp.
90
Phillips Petroleum
Co.
91
Unocal Corp.
92
BP America
93
Unocal Corp.
94
BP Oil
95
Exxon Corp.
96
BP Oil
97
Mobil
98
Exxon Co.
Sued by state of Alaska for
not preventing contamination
of shoreline from Valdez
spill
Sued by state of Alaska for
not preventing contamination
of shoreline from Valdez
spill
Sued by state of Alaska for
not preventing contamination
of shoreline from Valdez
spill
Sued by state of Alaska for
not preventing contamination
of shoreline from Valdez
spill
Sued by fishermen for spill
by another company’s tanker
because at-fault party
contesting liability
Settled suit brought by Sierra
Club over water pollution
Consent decree with US
Attorneys office and EPA for
water pollution from refinery
Cited by EPA for violations
of Clean Water Act for
unauthorized tank washing
transfers to the water
treatment facility (no actual
pollution; dispute over
meaning of permit)
Cited by EPA for violations
of Clean Water Act for
unauthorized tank washing
transfers to the water
treatment facility (no actual
pollution; dispute over
meaning of permit)
L.A. County officials filed
suit for an earlier oil spill
when a pipeline failed
Fined by EPA for violating
air and chemical reporting
regulations
08/16/89
On t=-1, BP Chemicals’
silicone surfactant
business bought by a
Union Carbide division
08/16/89
08/16/89
08/16/89
10/23/89
02/23/90
10/24/90
08/23/91
08/23/91
01/10/92
12/04/92
Dividend will be paid at
later date to shareholder
of record on t=0
40
TABLE A-2
Abnormal Returns, Gross Losses, & Costs
Event-by-event results using (1,0) window:
Event
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Information on Direct Costs Reported in WSJ
$47M, $3.3M of which was insured, + $210K/day for ~1 month and
$53K/day in lost production, + liability for people killed
Lawsuits filed for $3.5B; $6M in cleanup and private damages; $2500
fine
n.a.
n.a.
? fines on similar incidents $40K - $200K
$75M cleanup costs; $204M (14 years later), had $50M in insurance
>$60M in damages; >$1B in lawsuits
n.a.
Replacement power $280K/day for 4 months + repairs est. to be
> $1M deductible
Cleanup costs; ≤$5000 fine
~$18M cleanup, fines, claims (settled within 18 months)
$B’s
$18M in 2 suits (settled within 18 months) + cleanup costs?
$3.9M (5 years later) + cleanup costs?
≤$2500 fine, compliance?
≤$135,000 fine, compliance?
≤$135,000 fine, compliance?
≤$2500 fine
≤$2500 fine
≤$2500 fine
≤$2500 fine
$2500 fine + $35M over next 4 years on pollution control
Compliance?
State sued for $5M; compliance?
Fined $1M; $31.5M private suit filed
$250M suit filed
$340,000 fine
$1.2M suit filed
$10,000 fine + compliance?
$2500 fine
$300,000 fine
$242,000 fine
$24,000 fine
$32,000 fine
$250,000 fine
≤$20,000 fine
≤$150,000 fine
Change in Change in
value
value
($K)
(%÷ 100)
0.03394
102046
0.01804
78066
-0.00847
0.01784
-0.00184
-0.00285
-0.06263*
-0.01938
-0.04679‡
-6270
93534
-2597
-19967
-68258
-23371
-12284
0.00082
-0.0427§
-0.0181§
-0.0264§
-0.01887
-0.02351
0.00241
0.00407
0.0979*
0.0529§
0.0472§
0.02029
0.01225
‡
-0.0361
-0.00397
-0.02713
-0.0320*
-0.00745
-0.01976
0.02762
0.00267
-0.02308
0.02586
-0.02233
0.01766
0.02735
-0.00601
-0.01528
731
-76564
-1654658
-861123
-68324
-10620
1952
27842
368706
138621
70759
54852
12341
-147046
-4353
-105012
-359357
-20896
-296112
148563
14539
-354897
37325
-20850
22501
171042
-4648
-83942
41
TABLE A-2 – continued
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
Suspended fine of $10K, but must agree to spend $31.5M on pollution
control
≤$375,000 fine
≤$2500 fine
n.a.
≤$2500 fine
≤$7500 fine
≤$5000 fine
≤$2500 fine
≤$5000 fine
? ≤$2500 fine
? ≤$2500 fine
$65,000 suit filed
$2.3M suit filed
≤$7500 fine
n.a.
$15M in suits filed; cleanup costs?
n.a.
n.a.
Compliance costs = n.a.; possible fines if no compliance
2 years after incident fined $37,500 and ~$37M/each spent on 3
plants’pollution control equipment
Compliance costs = n.a.
Compliance costs = n.a.
≤$2000 fine
≤$30,000 fine
Compliance costs = n.a.
$100,000 fine; some compliance costs?
$6.8M suit filed; plus loss of use of facility?
Won in court, legal costs = n.a.
$15,000 penalty + compliance costs
Compliance costs ~$400M (uncertain timetable; order for equipment
2.5 years later); $1.55M penalties, $100M spent already
~$100M more in compliance costs
n.a.
≤$422,000 fine; compliance?
Potentially $M’s in fines; compliance?
2 years later sued for ≥$330,000; compliance?
Compliance?
$2M in penalties; $500,000 cleanup
$500,000 private settlement and $1.5M government settlement (within
1 year)
Compliance?
$600,000 fine
$340,000 fine
$6M penalty
$765,500 penalty plus compliance?; private suit settlement?
Suit seeks $M’s in fines and compliance
-0.02305
-12021
0.03861
-0.01139
0.00565
-0.00267
0.00637
0.04147
0.02695
0.01254
-0.0544‡
0.02028
-0.01840
0.01030
0.01032
-0.00003
0.00530
0.01218
0.02289
0.00616
-0.01805
22625
-37992
217
-8682
6943
15383
45335
38880
-43277
18409
-177979
12009
5017
-1246
45047
16502
111692
30185
-6049
0.00638
0.0640*
0.02471
0.02809
-0.00778
0.00087
0.00004
-0.01318
-0.01198
0.0358*
632
25513
340244
159552
-5387
20547
12
-89565
-58977
33547
-0.00927
-0.00076
-0.0579‡
0.01769
‡
0.0519
-0.0260§
-0.00991
0.00951
-11506
-689
-37156
12689
587068
-72746
-102963
317178
0.00597
0.01208
0.00324
-0.01481
-0.00752
-0.00726
22836
102401
41870
-195699
-13571
-51315
42
TABLE A-2 – continued
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
Suit seeks $8.8M in fines and compliance - $36M over past 3 years +
$50M over next 20
Agreed to $6M in cleanup costs over 2 years then possibly more
$581,000 settlement
$M’s in penalties sought + compliance (questionable if already in
compliance)
n.a.
n.a.
n.a.
n.a.
n.a.
n.a.
$104.8M suit filed (liability not clear)
$5.5M settlement; already in compliance
$2.3M fine
n.a.
n.a.
n.a.
$178,000 fine
n.a. = not available in WSJ
§
= return is significantly different from zero at 0.1 level of significance
‡
= return is significantly different from zero at 0.05 level of significance
*
= return is significantly different from zero at 0.01 level of significance
0.0425§
615436
-0.00816
-0.01891
-0.00223
-86401
-297069
-34045
-0.02291
0.00492
-0.00088
§
0.0246
-0.01139
0.00032
0.00020
-0.02807
0.01452
0.00570
0.02324
-0.01818
0.02065
-73927
86052
-2097
524338
-66149
1806
513
-202769
50443
412924
50091
-476997
1499452