stock price reaction of airlines and aircraft ...stock price reaction of airlines and aircraft...

57
STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel A Thesis Submitted to the University of North Carolina Wilmington in Partial Fulfillment of the Requirements for the Degree of Master of Business Administration Cameron School of Business University of North Carolina Wilmington 2010 Approved by Advisory Committee Vince Howe Joseph Farinella Co-Chair Nivine Richie Co-Chair Accepted by _____________________________ Dean, Graduate School

Upload: others

Post on 10-May-2020

7 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES

Pavel Šepel

A Thesis Submitted to the University of North Carolina Wilmington in Partial Fulfillment

of the Requirements for the Degree of Master of Business Administration

Cameron School of Business

University of North Carolina Wilmington

2010

Approved by

Advisory Committee

Vince Howe Joseph Farinella Co-Chair

Nivine Richie Co-Chair

Accepted by

_____________________________ Dean, Graduate School

Page 2: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

ii

TABLE OF CONTENTS ABSTRACT .............................................................................................................................. iii

LIST OF TABLES .................................................................................................................... iv

LIST OF FIGURES .................................................................................................................... v

LITERATURE REVIEW ........................................................................................................... 6

Impact on Airline Companies .............................................................................................................. 6

Impact on Aircraft Manufacturer Companies ...................................................................................... 9

HYPOTHESIS .......................................................................................................................... 11

RESEARCH DESIGN ............................................................................................................. 14

Data ................................................................................................................................................... 14

Methodology ..................................................................................................................................... 22

EMPIRICAL RESULTS .......................................................................................................... 27

CONCLUSIONS ...................................................................................................................... 35

REFERENCES ......................................................................................................................... 36

BIBLIOGRAPHY .................................................................................................................... 37

APPENDIX .............................................................................................................................. 39

Page 3: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

iii

ABSTRACT

The impact that aviation catastrophes have on airlines’ and aircraft manufacturers’ stock

price is examined in this paper. A priori, the stock price is expected to decrease after an

aviation catastrophe. To test this hypothesis, a sample of 150 air crashes involving civil

commercial aircraft is examined. An event study methodology is used to measure the

abnormal returns of airlines and aircraft manufacturers. In addition, cross-sectional analysis is

used to determine the factors that drive these abnormal returns. The results show that airlines

experience an average statistically significant stock price decline of -2.05% within one trading

day after the air crash. Aircraft manufacturers experience a stock price decline of -0.17%

during that time period. This decline is not statistically significant. The magnitude of the price

decline does not appear to be driven by the particular characteristics of the company or the

aviation catastrophe. This study contributes to the existing literature in two ways; it examines

both U.S. and foreign (non-U.S.) companies and updates the data set.

Page 4: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

iv

LIST OF TABLES

Table Page

1. Aviation Catastrophe Record for Airlines ............................................................................ 17 2. Aviation Catastrophe Record for Aircraft Manufacturers .................................................... 18 3. Aviation Catastrophe Characteristics (Panel A) and Classification of Air Crash Causes (Panel B) ............................................................................................................................... 21 4. Mean Abnormal Return and Mean Cumulative Abnormal Return for the Airline Sample Using the Market Model ........................................................................................................... 28 5. Mean Abnormal Return and Mean Cumulative Abnormal Return for the Airline Sample Using the Market-Adjusted Returns ......................................................................................... 29 6. Mean Abnormal Return and Mean Cumulative Abnormal Return for the Airline Sample Using the Comparison Period Mean-Adjusted Returns ........................................................... 30 7. Mean Abnormal Return and Mean Cumulative Abnormal Return for the Aircraft Manufacturers Sample Using the Market Model ..................................................................... 32 8. Mean Abnormal Return and Mean Cumulative Abnormal Return for the Aircraft Manufacturers Sample Using the Market-Adjusted Returns ................................................... 33 9. Mean Abnormal Return and Mean Cumulative Abnormal Return for the Aircraft Manufacturers Sample Using the Comparison Period Mean-Adjusted Returns ...................... 34

(A) = Airlines (M) = Manufacturers

Page 5: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

v

LIST OF FIGURES

Figure Page

1. Global passenger fatalities per 100 million passenger miles, scheduled commercial transport operations, excluding acts of unlawful interference................................................ 3 2. Global rate of accidents involving passenger fatalities per 10 million flights, scheduled commercial air transport operations, excluding acts of unlawful interference ...................... 4

Page 6: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

INTRODUCTION

Aviation business is highly responsive to global political and economical impacts.

Whether political climate, economical conditions or ecological concerns are under debate –

these impacts may be volatile and reactive, but the good news is that aviation shows high

adaptability to the external influences and is maintaining growth. As reported in the Federal

Aviation Administration1 Aerospace Forecast for 2010-2030, the present challenges will

remain for no less than twenty more years, but aviation will continue to grow over the long

term, despite current global economic conditions. Commercial aviation has suffered from the

turbulences in the economy during recession; bankruptcy of several major players in the

market, blown up fuel prices and notably shrunk investments – these challenges have slowed

down the business activity, however have not sharply decreased the number of flying

passengers and the growing demand for commercial aviation. The idea behind this trend is

that the economic growth will return along with the demand, and vice versa. As stated in the

forecast, “There has been a slowdown in air travel growth, and the FAA now calls for one

billion passengers to be flown in 2023” (p. 5).

The demand for commercial aviation weakened during recession in 2008 by reason of the

downturn of global economy. After four straight quarters of slowdown, the U.S. economy

continued growth in the fourth quarter of 2009. The growing U.S. and world economy will

also drive growth in business aviation demand over the long term. ”As the fleet grows, the

1 The Federal Aviation Administration (FAA) is an agency of the United States Department of Transportation, which is responsible for ensuring the safety of civil aviation (National Airworthiness Authority). FAA functions as an agency within the US Department of Transportation. The Federal Aviation Act of 1958 created the “Federal Aviation Agency” and gave it broad authority to combat aviation hazards as well broad rulemaking power. In addition, the FAA has sole responsibility for developing and maintaining a common civil-military system of air navigation and air traffic control. In 1968, Congress gave the FAA the power to set aircraft noise standards. Today, the agency's mission is "to provide the safest, most efficient aerospace system in the world."

Page 7: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

2

number of general aviation hours flown is projected to increase an average of 2.5 percent a

year through 2030” (p. 6), as reported by FAA.

The scenario of the FAA forecast for commercial aviation confirms that the world is

becoming more and more global, looking from both private and corporate perspective.

Moreover, globalization sets accordingly metamorphosed requirements for the level of

mobility and flexibility, in other words – the new level of dynamics, and this dynamics is very

physical. Businessmen and privates would like to be in the desired place on time, which often

means “as soon as possible”, and that – worldwide. For this reason, people fly and will

continue to fly more and more in the long term, which is obviously confirmed by the

mentioned forecasts. But what we are exploiting in order to fly for our private or business

motives, from service providers to machines, is produced and managed by human beings. And

unfortunately, we make mistakes, which occasionally are ruthless and irreversible.

However, as reported by the European Advertising Standards Alliance2 Annual Safety

Review 2009, ”The safety of aviation has improved from 1945 onwards. Based on the

measure of passenger fatalities per 100 million miles flown, it took some 20 years (1948 to

1968) to achieve the first 10-fold improvement from 5 to 0.5. Another 10-fold improvement

was reached in 1997, almost 30 years later, when the rate had dropped below 0.05. For the

year 2009 this rate is estimated to have stayed at 0.01 fatalities per 100 million miles flown.

The accident rate in this figure appears to be flat for recent years. This is the result of the scale

used to reflect the high rates in the late 1940s” (p. 11) (see Figure 1).

2 The European Advertising Standards Alliance (EASA) is the single authoritative voice on advertising self-regulation issues and promotes high ethical standards in commercial communications by means of effective self-regulation, while being mindful of national differences of culture, legal and commercial practice. These standards are promoted for example via EASA's Advertising Self-Regulatory Charter and EASA's Best Practice Recommendations. As a non-profit organization based in Brussels, it brings together national advertising self-regulatory organizations (SROs) and organizations representing the advertising industry in Europe and beyond.

Page 8: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

3

Figure 1: Global passenger fatalities per 100 million passenger miles, scheduled commercial transport operations, excluding acts of unlawful interference

Dark blue graph stands for passenger fatalities rate, light blue stands for 5 year moving average. Source: http://www.easa.europa.eu/

As also declared by EASA, ”The rate of accidents involving passenger fatalities in

scheduled operations (excluding acts of unlawful interference) per 10 million flights ranged

from 16 (1990) to 21 (1993) and showed no improvement from 1990 to 1993. From that year,

the rate dropped continuously until 2003, where it reached its lowest value, three. After

increases in 2004 and 2005, in line with the decreasing number of fatal accidents the rate

dropped in 2007 to four, increased to 5 in 2008 to drop back to 4 (estimate) in 2009. The 5

year moving average rate has remained almost constant since 2004” (p. 12) (see Figure 2).

Page 9: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

4

Figure 2: Global rate of accidents involving passenger fatalities per 10 million flights, scheduled commercial air transport operations, excluding acts of unlawful interference

Dark blue stands for fatal accident rate, light blue stands for 5 year moving average. Source: http://www.easa.europa.eu/

Although air travel statistically represents one of the safest modes of traveling large

distances, aviation catastrophes unfortunately do occur. When an air crash takes place,

whether provoked by human error, weather conditions, aircraft technical integrity, or

deliberate terrorism, it is always a tragedy and involves loss of many human lives. As brought

up by Chance and Ferris (1987), ”The immediate and long-range human loss cannot be fully

measured, much less adequately compensated. The ramifications on the companies involved,

their shareholders and the air transport industry as a whole are of concern to economists,

investors, regulators and the public” (p. 151).

What would be the financial consequences of an aviation catastrophe to the airlines and

aircraft manufacturers involved? And what is the reasonable way to delineate these “financial

consequences”? What would be the adequate measure for that determination?

Page 10: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

5

The evidence provided in this paper measures the goodwill loss. Thus, the change in

stock price captures the financial consequences of the aviation catastrophe. The null

hypothesis to be tested is that the mean excess return during the event windows subsequent to

an air crash is equal to zero. One test looks at the stock price reaction of the airlines involved

in the crash. Another test looks at the stock price reaction of aircraft manufacturers involved

in the crash. Both stock price reactions are expected to be negative.

The paper is organized as follows. The next section is dedicated to the literature review.

The third section looks at the hypothesis to be tested. The fourth section provides the

description of the data. The fifth section explains the methodology used to test the hypothesis.

The sixth section presents the empirical results of the analysis. The conclusions and the

importance of the results are discussed in the sixth section.

Page 11: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

LITERATURE REVIEW

The literature review is grouped in two subsegments – one group of articles examines the

stock price reaction of airlines, the second group looks at the performance of aircraft

manufacturers following the tragedy. Also, the uniqueness of this paper weighted against the

reviewed literature is to be mentioned.

Impact on Airline Companies

Beginning with the preceding efforts to examine the stock price reaction of involved

airlines following the aviation catastrophe, one of the tests of Chance and Ferris (1987) looks

at the stock prices of airlines involved in the crash and shows a loss of shareholder’s wealth of

1.2%. Their test also observes the airlines not involved in the crash; no reaction is observed

for those companies. The results suggest that the phenomenon is not industry-wide. An

immediate reaction or abnormal return for involved airlines one trading day after the crash,

which does not continue beyond the date of the crash, is what the results of their study state.

Davidson, Chandy and Cross (1987) examine the effects of large losses on firm value

in commercial aviation. In particular, they analyze the impact of large losses on the returns of

common stock. Prior to that, Sprecher and Pertl (1983) looked at large losses for a cross

section of industries. Davidson, Chandy and Cross extend their preceding study by focusing

on a specific industry with unique characteristics – the airline industry. The airline industry is

chosen for study due to its high vulnerability for large catastrophic losses – airplane crashes,

and the high degree of regulation imposed on the industry. This regulation results in the

mandatory liability insurance coverage. The impact of these losses is negative and statistically

significant on the day of the crash in a number of cases. These negative reactions are only

over a short term and reverse in the days immediately following the crash. This is contrary to

Page 12: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

7

the findings in the study of Sprecher and Pertl, in which large losses have a negative impact

on stock returns with no reversal in the first few days following the losses.

Borenstein and B. Zimmerman (1988) examine deviations from expected demand

following aviation disasters and determine little or no effect prior to the so called Airline

Deregulation Act3 and weak indication of a response to recent crashes. Since most crashes in

the sample involve total destruction of the aircraft and an average of more than 40 fatalities,

the average firm loss appears to be below the total social costs of the accident. According to

their paper, crashes are associated with a statistically significant loss in equity value of 1%

percent. However, reaction regarding demand or consumer response is weakly significant

before and after deregulation. The investigation of consumer response also finds that before

deregulation, travelers did not respond to crashes to an extent that is statistically discernible.

In the period since deregulation, the consumer response appears to have increased.

Mitchell and Maloney (1989) examine the "brand-name effect" of air crashes for

involved airlines. Their study focuses on the potential loss of consumer goodwill. To test the

theory of declining goodwill value caused by fears and changing consumption patterns, they

examine the abnormal stock market performance of airlines immediately after a crash. They

partition the sample into airline "at-fault" crashes, where a pilot error is at issue, and all

others. Mitchell and Maloney find a statistically significant negative market reaction for the

crash airline in the at-fault subsample only, concluding that "airline crashes cause consumers

3 The 1978 Airline Deregulation Act partially shifted control over air travel from the political to the market sphere. The Civil Aeronautics Board (CAB), which had previously controlled entry, exit, and the pricing of airline services, as well as intercarrier agreements, mergers, and consumer issues, was phased out under the CAB Sunset Act and expired officially on December 31, 1984. The economic liberalization of air travel was part of a series of “deregulation” moves based on the growing realization that a politically controlled economy served no continuing public interest. U.S. deregulation has been part of a greater global airline liberalization trend, especially in Asia, Latin America, and the European Union.

Page 13: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

8

to reduce their demand for the services provided by negligent carriers,"' contrary to Borenstein

and Zimmerman. They do not examine the impact of crashes on non-crash airlines.

Bosch, Eckard and Singal (1998) examine stock market reactions to commercial air

crashes to test the hypothesis that consumers respond by switching to rival airlines and flying

less. They focus on the stock price reactions of airlines not involved in the crash. The

underlying assumption is that if switching occurs, non-crash airlines should benefit to the

extent that they are direct competitors of the crash airline. The evidence supports both a

switching effect – the greater the overlap between airlines, the greater consumer switch, and a

spillover – also non-crash airlines with little overlap lose value. Thus, the crash airline suffers

significant financial losses from a crash, which appear to be related to consumer switching.

While this suggests a traditional market incentive to "supply" safety, it can only apply to

safety related factors under each airline's control. The evidence of a negative spillover

suggests that consumers and insurers may be concerned about other elements of the

commercial air travel system that are involved in the joint production of air safety.

Walker, Thiengtham and Lin (2005) examine the impact of aviation disasters on the short-

and long-term performance of airlines in one of their tests. They measure the abnormal

performance of airlines and airplane manufacturers subsequent to these disasters. Also, they

employ a series of univariate tests and regression analysis to determine the factors that drive

the abnormal returns for the observed firms. The paper concludes that airlines experience an

average stock price drop of 2.8% within one trading day after the corresponding news

announcement. The magnitude of the initial price decline appears to be driven by various

characteristics of both the firm and the accident itself. They observe that airlines' abnormal

performance is negatively related to firm size and the number of fatalities resulting from the

Page 14: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

9

accident. In addition, they observe that disasters that occurred in the U.S. and disasters caused

by criminal activity (in particular the 9/11 terrorist attacks) cause significantly larger stock

price drops in the days following the event.

Impact on Aircraft Manufacturer Companies

Continuing with the prior studies on the attempts to observe the performance of the

involved aircraft manufacturers subsequent to an air crash, Chalk (1987) investigates the

wealth impact on aircraft manufacturers of crashes involving their aircraft. He attempts to

separate out legal and endogenous market components of these costs using stock market data.

The results of the market model and regressions indicate that accidents, in which the

manufacturers are implicated and under condition that these aircrafts are still in production

result in a decrease in the market value of equity of -3.774%. In contrast, crashes that did not

implicate the manufacturer did not result in any stock price effect; neither did the involved

airplanes that are no longer in production. These results suggest that endogenous responses in

the product market are responsible for the negative effect on manufacturers.

Chance and Ferris (1987) examine the financial consequences of an aviation disaster to the

carriers; they also test the reaction of stock prices to aviation crashes. One of their tests looks

at the manufacturers involved in the crash and indicates no stock price reaction. On the

strength of this result, manufacturers not involved in the crash are not tested, as the not

involved airlines.

Walker, Thiengtham and Lin (2005) examine the impact of aviation disasters on the short-

and long-term performance of airplane manufacturers in one of their tests. Similarly to the

airlines, they measure abnormal performance of airplane manufacturers subsequent to these

Page 15: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

10

disasters. Also, they employ a series of univariate tests and regression analysis to determine

the factors that drive the abnormal returns for the observed firms. The paper concludes that

aircraft manufacturers experience an average stock price drop of 0.8% within one trading day

after the corresponding news announcement. No statistically significant dependencies are

observed in the conducted cross-sectional analysis testing the effect of the crash particularities

and 9/11 period.

Page 16: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

HYPOTHESIS

There is no doubt that both airlines and aircraft manufacturers are the “victims” of

physical losses and liability claims subsequent to an aviation catastrophe. However, the

financial impact of an air crash should not be measured exclusively by the examination of the

income statement. Companies can use accounting principles to "bury" negative information,

while emphasizing more sympathetic information, such as better advertising, higher service

quality, extended route maps and increased passenger miles, combining all that with the right

branding. Moreover, insurance payments to the company may compensate for much of the

loss, and subsequent are difficult to detect in the income statements.

Thus, related legal costs (increased insurance rates, victims’ claims compensations,

advanced safety principles), as well as physical (fleet loss or maintenance) costs for the

crashed aircraft, accordingly growing expenses on insurance quality, and loss of consumer

confidence and trust accompanied by flying less might not represent the de facto losses

suffered by involved companies. To test this hypothesis, a more accurate measure of the

financial impact of an aviation catastrophe will be applied by testing the price reaction of the

company's common stock subsequent to an air crash. The stock price reaction will capture the

markets’ long term impact on the companies and their reputation – the goodwill loss. In an

efficient stock market, the prices of securities react quickly to new information and fully

reflect its economic content. According to the works on this issue, the stock market is quite

efficient in responding to events, which are relevant to the wealth of shareholders (Fama,

1970, 1976).

The rationale of examining the response of stock prices as a measure of the financial

impact of an event is well known in the economic theory. Referring to Chance and Ferris

(1987), ”The price of a stock represents investors’ assessments of the discounted future cash

Page 17: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

12

flows of a company. That is, it reflects the company’s profitability, as defined by its future

cash generating capacity. The stock price, thus, represents the shareholders’ wealth, and

reflects the willingness of investors to commit capital to the firm. If the stock price reacts

negatively to an event, the implication is that investors foresee lower or riskier cash flows in

the future. If the stock price does not react, it can be assumed that the event conveys no

significant information about the expected value of the firm’s future cash flows” (p.152).

Thus, this paper examines the stock price reaction of involved companies subsequent

to the aviation catastrophes. The null hypothesis to be tested is that the mean excess return

during the event windows subsequent to the aviation catastrophe is equal to zero. The return is

expected to be abnormal and negative. One test looks at the stock price reaction of the airlines

involved in the air crash. Those stocks are expected to react negatively, but the duration and

magnitude of the reaction cannot be known a priori. However, if investors believe that the

fault lies with the aircraft manufacturer, then no reaction is to be expected. Another test looks

at the stock price reaction of aircraft manufacturers involved in the air crash. Also for these

companies a negative stock price reaction is expected. However, if investors believe that the

responsibility lies with the airline, then no reaction will be found.

On the strength of these expectations, the cross-sectional analysis is conducted in order to

test, if such variables as the amount of victims suffered in the air or on the ground, the

business location of the company (U.S. or non-U.S.), and the reason of the air crash affect the

magnitude of abnormal returns. In addition, the supposition, if the reaction materialized in

negative abnormal returns tends to vary before and after the terrorist attacks on 9/11 is to be

tested. Having a look at the tested variables, it is reasonable to assume that the amount of

victims injured or killed in the accident might amplify the negative stock price reaction

Page 18: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

13

because of immediate mass media resonance subsequent to a tragedy. Also, it is logical to

expect that the location of the company’s business might affect the stock price reaction,

depending on the U.S. and non-U.S. investors’ particular behavior, such as risk averseness.

Moreover, it is likely that the cause of the air crash might have a similar effect on the stock

price reaction, depending on the mass media immediate news announcement and preliminary

assumptions regarding the probable reasons of an accident. To comment this presupposition to

mention is that, if the pilots are in charge of the tragedy – the publicity and investors might

conclude that the involved airline is not reliable, hiring untrained pilots. Likewise, if the

technical condition of an aircraft is in charge of the accident – the publicity and investors

might conclude that the involved aircraft manufacturer is not acceptable for the market,

producing low quality products that do not fulfill the required safety principles. All this might

have a negative impact on the financial performance of the involved company. Lastly, the

9/11 terrorist attacks might affect the investor’s responsiveness, from the view of risk

averseness and behavior regarding their investment strategy subsequent to the aviation

catastrophes.

Page 19: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

RESEARCH DESIGN

Data

The data on aviation catastrophes is published by the U.S. National Transportation

Safety Board4. The NTSB data includes reports providing the accident or incident

information, such as the date and time, location and injury severity, the category and model of

an aircraft, as well as the type of operation and involved airline. Also, NTSB supplies a

probable (preliminary version) and/or factual (expertise conclusion subsequent to

investigation) cause of an accident. The database of NTSB allows selecting and hence,

restricting or limiting the screening criteria. The sample of examined aviation catastrophes in

this study includes approximately 150 crashes occurred within the last thirty years, and

involving 25 or more victims. The additional criteria applied to restrict the examined sample

are selected as described further on in this paragraph.

The leading screening criteria for the data collection are the aviation type (civil,

commercial), the operation (civil, commercial, U.S., non-U.S. or foreign) involved and the

time window of the catastrophes. Thus, air crashes involving national (U.S.) and foreign (non-

U.S.) operations using both national and foreign aircraft fleet, belonging to the civil

commercial aviation, occurred worldwide (location) from January 1981 until May 2010 are

observed. To mention here is that NTSB does not provide with the accident data prior to the

beginning year of the selected time period. The second and constrained criterion for the

selection of aviation catastrophes is the data availability. As a source, Bloomberg is used for

the data collection. As both U.S. and non-U.S. catastrophes are examined, numerous data on

4 The National Transportation Safety Board (NTSB) is an independent U.S. Government investigative agency responsible for civil transportation accident investigation. In this role, the NTSB investigates and reports on aviation accidents and incidents, certain types of highway crashes, ship and marine accidents, pipeline incidents and railroad accidents. When requested, the NTSB will assist the military with accident investigation. The NTSB is also in charge of investigating cases of hazardous waste releases that occur during transportation.

Page 20: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

15

the publicly owned companies (i.e. Arabic, Chinese or Russian) are unlisted, or have never

been published. Privately owned companies are also common for the airline industry. Data for

those companies is also unlisted and not available on Bloomberg. Many companies, even if

they used to be European or U.S. are bankrupt for different reasons, or else acquired or

merged; also for those companies the correct stock price information is delisted, or no more

available on Bloomberg. For supplementary explanations regarding data availability on

Bloomberg, please find the table of airlines and aircraft manufacturers with the data

availability comments in the appendix of this paper. The third criterion is the number of

fatalities involved in the air crash. Aviation catastrophes involving 25 or more human losses

are examined. Reason for this limitation is to pay attention to crashes having been reflected in

the mass media and having more public resonance, rather than looking at less reactive crashes,

which might appear to be outliers for the observation.

Having applied the determined screening criteria, a sample of 30 records for airlines and

86 records for aircraft manufacturers is examined, involving civil commercial aviation

catastrophes with more than 25 fatalities and dating from January 1982 to the May 2010, or

factually – the present time. The reason of separating the data into the two different samples is

also the data availability; daily stock price information is available for most of the aircraft

manufacturers, which usually are EADS and Boeing. The accordant data for airlines is much

more internationally spread and thus is less obtainable. The idea of separating the samples is

therefore to embrace as much available information as possible for each of the autonomous

samples, instead of being stuck by the least informative sample. For this reason, the sample of

catastrophic events for aircraft manufacturers is almost three times larger than the one for

airlines.

Page 21: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

16

To visualize the data, Table 1 represents summary statistics for the examined airlines’

sample. The sample covers the crashes dated from February 1985 to January 2009, operated

by North and South American, as well as French, Kenyan, Russian, Brazilian, Chinese, Swiss,

Korean, Swedish and Singaporean airlines, involving 25 or more fatalities. Out of the total 30

crashes – 17 were operated by national (U.S.) airlines, and 13 – by foreign (non-U.S.) airlines.

The number of fatalities fluctuates from the minimum 25 to the maximum 265 victims, as a

result an Airbus Industrie A300-605R operated by of American Airlines flight 587 that

crashed into a residential area of Belle Harbor, New York in December, 2001. Most of these

crashes (12) are reasoned by the flight crew members’ fault, 7 – by the technical integrity of

the involved aircraft, 3 – by air traffic control failure, 5 (or 3?) – by criminal or terror activity,

and 3 – by different reasons, such as wing icing, cargo explosion on board or a combination of

two or more unlucky coincidences (i.e. pilot error and weather conditions). Two crashes are

still under investigation. Lastly, 10 of the observed crashes occurred before 9/11, and 18 –

after the 9/11 terrorist attacks.

Page 22: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

17

Table 1: Aviation Catastrophe Record for Airlines

Date Involved Airline Victims U.S. or non-

U.S. Airline

Crash Reason

06.01.09 Air France 228 non-U.S. Under investigation 02.12.09 Colgan Air 50 U.S. Flight Crew Members 09.14.08 Aeroflot-Nord 88 non-U.S. Flight Crew Members 07.17.07 TAM Linhas Aéreas 199 non-U.S. Flight Crew Members 05.05.07 Kenya Airways 114 non-U.S. Flight Crew Members 09.29.06 Gol Air 154 non-U.S. Air Traffic Control 08.27.06 Comair Inc 49 U.S. Flight Crew Members 11.21.04 China Eastern Airlines 55 non-U.S. Wing Icing 11.12.01 American Airlines 265 U.S. Flight Crew Members 10.08.01 Scandinavian Airlines System 118 non-U.S. Air Traffic Control 09.11.01 American Airlines 92+64 U.S. Crime or Terror 09.11.01 United Airlines 65+44 U.S. Crime or Terror 10.31.00 Singapore Airlines 83 non-U.S. Flight Crew Members 07.25.00 Air France 113 non-U.S. Unlucky Combination 01.31.00 Alaska Airlines 88 U.S. Technical Integrity 01.30.00 Kenya Airways 169 non-U.S. Technical Integrity 09.02.98 Swissair 229 non-U.S. Technical Integrity 12.19.97 Silkair 97 non-U.S. Flight Crew Members 08.06.97 Korean Airlines 228 non-U.S. Flight Crew Members 07.17.96 Trans World Airlines 230 U.S. Technical Integrity 12.20.95 American Airlines 160 U.S. Flight Crew Members 10.31.94 American Eagle Airlines 68 U.S. Wing Icing 09.08.94 USAir 132 U.S. Technical Integrity 07.02.94 USAir 37 U.S. Flight Crew Members 03.22.92 USAir 27 U.S. Wing Icing 03.03.91 United Airlines 25 U.S. Technical Integrity 02.01.91 USAir 34 U.S. Air Traffic Control 07.19.89 United Airlines 111 U.S. Technical Integrity 12.07.87 Pacific Southwest Airlines 43 U.S. Crime or Terror 08.02.85 Delta Airlines 135 U.S. Flight Crew Members

Table 2 represents summary statistics for the examined aircraft manufacturers’ sample.

The sample covers the crashes dated from January 1982 to May 2010, including the majority

of the U.S. (Boeing, Lockheed, McDonnell Douglas, Bombardier) and the minority of the

European (Airbus) aircraft manufacturers, involving 25 or more fatalities. Out of the total 86

Page 23: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

18

crashes – 77 involved national (U.S.) aircraft models, and 9 – foreign (non-U.S.) aircraft

models. The number of fatalities involved in this sample varies from the minimum 25 to the

maximum 349 victims, as a result of a mid air collision about 40 miles southwest of New

Delhi, India, between a departing Saudi Arabian Airlines B747-168 and an arriving Kazakh

Airways TU-154 in December, 1996. Most of the crashes (31) are reasoned by the flight crew

members’ fault, 16 – by the technical integrity of the involved aircraft, 9 – by air traffic

control failure, 6 (or 2?) – by criminal or terror activity, 17 – by different reasons, such as

wing icing, cargo explosion on board, or a combination of two or more unlucky coincidences

(i.e. mechanical failure and pilot error). Nine crashes are still under investigation. Lastly, 48

of the observed crashes occurred before 9/11, and 38 – after the 9/11 terrorist attacks.

Table 2: Aviation Catastrophe Record for Aircraft Manufacturers

Date Involved Aircraft Victims U.S. or non-U.S.

Aircraft Crash Reason

05.22.10 Boeing 158 U.S. Under Investigation 05.12.10 Airbus 103 non-U.S. Under Investigation 01.25.10 Boeing 90 U.S. Under Investigation 06.29.09 Airbus 152 non-U.S. Under Investigation 06.01.09 Airbus 228 non-U.S. Under Investigation 02.12.09 Bombardier 50 U.S. Flight Crew Members 09.14.08 Boeing 88 U.S. Flight Crew Members 08.24.08 Boeing 65 U.S. Technical Integrity 08.20.08 Boeing 154 U.S. Flight Crew Members 04.15.08 McDonnell Douglas 40 U.S. Technical Integrity 11.30.07 McDonnell Douglas 57 U.S. Flight Crew Members 09.16.07 Boeing 89 non-U.S. Flight Crew Members 07.17.07 Airbus 199 non-U.S. Flight Crew Members 05.05.07 Boeing 114 U.S. Flight Crew Members 01.01.07 Boeing 102 U.S. Technical + Pilot 10.29.06 Boeing 96 U.S. Air Traffic Control 09.29.06 Boeing 154 U.S. Air Traffic Control 08.27.06 Bombardier 49 U.S. Flight Crew Members 07.08.06 Airbus 124 non-U.S. Flight Crew Members 12.10.05 McDonnell Douglas 107 U.S. Weather Conditions 10.22.05 Boeing 117 U.S. Weather Conditions

Page 24: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

19

Continued 09.05.05 Boeing 145 U.S. Flight Crew Members 08.23.05 Boeing 45 U.S. Weather Conditions 08.16.05 Boeing 160 U.S. Icing 08.14.05 Boeing 121 U.S. Technical Integrity 02.03.05 Boeing 104 U.S. Under Investigation 11.30.04 McDonnell Douglas 25 U.S. Flight Crew Members 11.21.04 Bombardier 55 U.S. Icing 12.25.03 Boeing 141 U.S. Flight Crew Members 07.08.03 Boeing 115 U.S. Technical Integrity 03.06.03 Boeing 102 U.S. Technical Integrity 07.01.02 Boeing 71 U.S. Air Traffic Control 05.25.02 Boeing 206 U.S. Technical Integrity 04.15.02 Boeing 138 U.S. Air Traffic Control 01.28.02 Boeing 92 U.S. Flight Crew Members 11.12.01 Airbus 265 non-U.S. Flight Crew Members 10.08.01 McDonnell Douglas 118 U.S. Air Traffic Control 09.11.01 Boeings 92+65+64+44 U.S. Crime or Terror 10.31.00 Boeing 83 U.S. Flight Crew Members 08.23.00 Airbus 143 non-U.S. Flight Crew Members 07.17.00 Boeing 60 U.S. Flight Crew Members 04.19.00 Boeing 131 U.S. Air Traffic Control 01.31.00 McDonnell Douglas 88 U.S. Technical Integrity 10.31.99 Boeing 217 U.S. Suicide or Technical 08.31.99 Boeing 80 U.S. Flight Crew Members 09.02.98 McDonnell Douglas 229 U.S. Technical Integrity 04.20.98 Boeing 52 U.S. Weather + Pilot 12.19.97 Boeing 97 U.S. Flight Crew Members 08.06.97 Boeing 228 U.S. Flight Crew Members 05.08.97 Boeing 35 U.S. Pilot + Weather 11.23.96 Boeing 125 U.S. Crime or Terror 11.12.96 Boeing 349 U.S. Air Traffic Control 10.22.96 Boeing 34 U.S. Under Investigation 10.02.96 Boeing 70 U.S. Technical Integrity 07.17.96 Boeing 230 U.S. Technical Integrity 05.11.96 McDonnell Douglas 110 U.S. Cargo Explosion 02.29.96 Boeing 123 U.S. Weather Conditions 02.06.96 Boeing 189 U.S. Technical Integrity 12.20.95 Boeing 160 U.S. Flight Crew Members 12.03.95 Boeing 72 U.S. Under Investigation 08.09.95 Boeing 65 U.S. Flight Crew Members 01.11.95 McDonnell Douglas 52 U.S. Under Investigation

Page 25: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

20

Continued 12.29.94 Boeing 54 U.S. Pilot + Weather 09.08.94 Boeing 132 U.S. Technical Integrity 07.02.94 McDonnell Douglas 37 U.S. Flight Crew Members 12.21.92 McDonnell Douglas 56 U.S. Weather + Pilot 06.06.92 Boeing 47 U.S. Technical Integrity 03.03.91 Boeing 25 U.S. Technical Integrity 02.01.91 Boeing 34 U.S. Air Traffic Control 05.05.90 McDonnell Douglas 27 U.S. Technical Integrity 01.25.90 Boeing 73 U.S. Flight Crew Members 10.21.89 Boeing 131 U.S. Flight Crew Members 07.19.89 McDonnell Douglas 111 U.S. Technical Integrity 06.07.89 McDonnell Douglas 174 U.S. Flight Crew Members 02.08.89 Boeing 144 U.S. Flight Crew Members 12.21.88 Boeing 270 U.S. Crime or Terror 11.15.87 McDonnell Douglas 28 U.S. Pilot + Icing 08.16.87 McDonnell Douglas 156 U.S. Flight Crew Members 08.31.86 McDonnell Douglas 82 U.S. Air Traffic Control 12.12.85 McDonnell Douglas 256 U.S. Icing 09.06.85 McDonnell Douglas 31 U.S. Flight Crew Members 08.02.85 Lockheed 135 U.S. Flight Crew Members 01.21.85 Lockheed 70 U.S. Flight Crew Members 01.01.85 Boeing 29 U.S. Under Investigation 07.09.82 Boeing 153 U.S. Weather Conditions 01.13.82 Boeing 78 U.S. Pilot + Icing

For data processing, daily stock price returns are retrieved for all publicly traded

national (U.S.) and foreign (non-U.S.) airlines and aircraft manufacturers. Daily stock price

information is exported from Bloomberg. The time window of one year around the crash for

each company is covered in order to calculate the abnormal returns, using the expected and

actual returns on the stocks. Where necessary, the daily stock prices denoted in local foreign

currencies are converted into the U.S. dollars to homogenize the output. In addition, return

information for value-weighted world market index are retrieved in order to measure the daily

market returns. The MXSI5 world index – a free float-adjusted market capitalization weighted

5 The MSCI World Index is a free float-adjusted market capitalization weighted index that is designed to measure the equity market performance of both developed and developing markets. As of February 2008 the

Page 26: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

21

index that is designed to measure the equity market performance of both developed and

developing markets is applied. The purpose of using the MSCI world market index is to

homogenize the market indices instead of employing the local national indices for

international companies.

To illuminate the cross-sectional analysis on the strength of the expectation of the

negative abnormal returns discussed beforehand in this paper, Table 3 represents definitions

of the variables used in the subsequent examination. Panel A provides definitions for variables

that characterize the aviation catastrophe. Panel B provides the classification of the air crash

causes that were immediately announced in the mass media.

Table 3: Aviation Catastrophe Characteristics (Panel A) and Classification of Air Crash Causes (Panel B)

Panel A

Victims People killed in the crash in the air and on the ground

U.S. or non-U.S.* U.S. or non-U.S. (airline or aircraft manufacturer)

Before or After 9/11* Crash dated before or after 9/11 terrorist attacks

Panel B (*)

Flight Crew Fault* Pilot error; oversight; lapse in judgment or failure to exercise

Technical Integrity* Engine failure; instrument failure; design failure

Air Traffic Control* Air traffic control error; incorrect commands by dispatcher or incorrect command interpretation by the crew

Crime or Terror* Hijacking ; suicide; 9/11

Under Investigation* Unknown

None of these* Wing icing; cargo explosion; weather conditions or a combination of unlucky coincidences

*Dummy variable

MSCI World Index consisted of 23 developed market countries with 1945 companies and combined market capitalization of USD 26,065 billion.

Page 27: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

22

Methodology

An event study methodology is applied in this paper, in particular the benchmark

models using a separate estimation period, including the market model (MM), the market-

adjusted return model (MAR) and the comparison period mean-adjusted return model to

measure the abnormal stock price performance of airlines and aircraft manufacturers

subsequent to aviation catastrophes. The application of each benchmark model requires two

time series of return data for each security-event: an estimation period for estimating the

benchmark parameters (including standard deviation), and an event period for calculating and

testing the abnormal returns.

Event study methodology measures the abnormal return of a stock as the difference

between the actual return and the expected return, around the time of an event. Event studies

draw on the efficient market hypothesis of Fama (1969), which states that capital markets are

efficient in processing information by establishing correct new stock price equilibrium as soon

as new information about a firm becomes available. The logics underlying the hypothesis is

the belief that investors in capital markets process publicly available information on firm

activities and external events influencing a company, and that they consider not just the

impact on current performance but also on the performance of the company in future periods.

When additional information becomes available, the company's stock price should change

rapidly and should reflect investors' revised consensus of the firm's future profitability.

The strength of the method lies in the fact that it captures the overall assessment by a

large number of investors of the discounted value of current and future firm performance

attributable to individual events, which are reflected in the stock price and the market value of

the firm. Changes in investors' beliefs regarding the future profitability of a firm are reflected

Page 28: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

23

in abnormal returns – risk adjusted returns in excess of the firm's expected return – after the

catastrophe. Abnormal returns thus provide a unique means of associating the impact of a

crash announcement on the firm's expected profitability in future periods.

To calculate the effect of an event, it is necessary to estimate what the price of the

stock would have been had the event not occurred. To do this, and to control for overall

market effects, the price of the stock is regressed against a market index. The estimated

coefficients from that regression are used to calculate the predicted value of the stock over the

time window in which the stock price is adjusted. This yields the regression:

(1) tstmts RR ,,10, εββ ++=

where tsR , is the return of stock s at time t: tsR , = ( tsprice , - 1, −tsprice )/ 1, −tsprice . The

subscript t indicates time, the subscript s indicates a specific stock, and the subscript m

indicates the market. The ts ,ε is a random error term for stock s at time t, and the β 's are

coefficients to be estimated. As mentioned before, the MXSI value-weighted world market

index is used to proxy for the market. The date of the event is denoted as t = 0. To estimate the

expected return, the data from t = (-255, -21) is used, which is 255 trading days (less than one

year) of pre-event data.

The coefficient estimates from the regression shown in Equation 1 are used to predict

the expected return over various post-event windows. The model provides a good fit, that is,

when examining the quality of the model for other dates on which no accident occurred the

cumulative abnormal returns (AR_C) are very close to zero and statistically insignificant.

Page 29: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

24

To estimate the abnormal return (market model) of a stock on day t, Brown and

Warner (1985) are followed and the expected return on the stock is subtracted from its actual

return on that day:

(2) )( ,10,, tmtsts RRAR ββ +−=

The coefficients 0β and 1β are estimates of the true parameters obtained via ordinary

least squares (OLS) regression. The abnormal returns are simply the prediction errors of the

model over the event window. Notice here that AR are abnormal returns, that is they are

returns over and above the return predicted by general market trends on a given day. The

assumption of the methodology is that the abnormal returns are the result of the

announcement and not some other random event occurring on the same day. The strength of

the method is linked to the improbability of random events across different companies on

different days coinciding with the announcement of an aviation disaster. The standard errors

are calculated by the formula defined by Judge, Kill, Griffiths, Luetkepohl, and Lee (1988).

(3) ))(

)(11()var(

1

2,

2,2

,

−++=∑ =

T

t Mtm

Mtmss

RR

RR

TSAR τ

where 2sS is the variance of the error from the estimation model, MR is the mean market

return over the prediction interval, and T represents the number of days in the estimation

interval. The τ indicates observations within the event window, while the t indicates

observations in the estimation interval. Notice, then, that the standard error on any given day

τ of the prediction interval is a function of how much the actual market return on that day

Page 30: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

25

deviates from the mean market return during the estimation interval. Thus, on days on which

the market return is very different from the expected market return, the standard errors of

abnormal returns are greater. Notice also that the standard error depends on the length of the

estimation interval, such that longer estimation intervals lead to lower standard errors.

Under the assumption that the returns on each day are independent, the standard errors

are cumulative, so the proper standard error is the cumulative standard error. This is due to the

fact that adding independent normal variables requires adding the standard errors. Thus, we

have the following equations to describe a company's AR_C, and the variance of the

cumulative abnormal returns, var (AR_C):

(4) AR_Cs,t = ARs,ii=0

τ∑

(5) var(AR _Cs,τ ) = var(ARs,i)i=0

τ∑

From these equations the average AR_C across all firms and the variance of the AR_Cs

can be calculated. The resulting equations are:

(6) AR _Cτ =1N

AR _Cs,τs=1

N

(7) var(AR _Cτ ) =1

N 2 var(AR _Cs,τ )s=1

N

To test the hypothesis that the mean AR_C is different from zero on any given day,

then, one would use a t-test, which under the hypothesis of zero returns, is of the form:

Page 31: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

26

(8) t =AR _Cτ

var(AR _Cτ )≅ t(α ,df =N −1)

As mentioned before, the two benchmark models alternative to the market model are

used to calculate the abnormal returns. The market adjusted returns are calculated by

subtracting the observed return on the market index for day t, tmR , , from the rate of return of

the common stock of the s firm on day t:

(9) tmtsts RRAR ,,, −=

The comparison period mean adjusted returns are calculated by subtracting the

arithmetic mean return of the common stock of the firm s calculated over the estimation

period, sR , from its return on day t:

(10) ststs RRAR −= ,,

The definitions of the average abnormal return, cumulative average abnormal return

and average compounded abnormal return are analogous to those for market model abnormal

returns.

As mentioned before, the null hypothesis to be tested using the methodology discussed in

this section is that the mean excess return during the event windows subsequent to the aviation

catastrophe is equal to zero. The return is expected to be negative.

Page 32: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

EMPIRICAL RESULTS

The results of this study show that there is a statistically significant negative stock price

reaction stated by corresponding negative abnormal returns of the involved airlines from one

to six days subsequent to an aviation catastrophe; however, the results show no evidence of a

statistically significant stock price reaction for the involved aircraft manufacturers.

The stock price reaction of involved airlines using the market model show negative

abnormal returns of -2.12% one day, -1.86% two days and -6.40% six days subsequent to the

aviation catastrophe statistically significant at 99% level. Furthermore, negative mean

cumulative abnormal returns of -2.16% one day, -3.53% two days, -2.31% five days and -

4.24% ten days subsequent to the air crash statistically significant at 99%, 99%, 90% and 95%

level respectively are observed (see Table 4). According to the market adjusted model,

negative abnormal returns of -2.09% one day, -1.94% two days and -6.35% six days are

observed subsequent to the aviation catastrophe significant at 99% level. In addition, negative

mean cumulative abnormal returns of -2.19% one day, -3.62% two days, -2.69% five days and

-4.76% ten days subsequent to the aviation catastrophe statistically significant at 99%, 99%,

95% and 99% level respectively are recognized (see Table 5). The comparison-period mean

adjusted model confirms negative abnormal returns of -1.94% one day, -1.82% two days and -

6.45% six days subsequent to the aviation catastrophe significant at 99% level. Also, negative

mean cumulative abnormal returns of -2.00% one day, -3.34% two days, -2.08% five days and

-5.17% ten days subsequent to the air crash statistically significant at 99%, 99%, 90% and

99% level respectively are detected (see Table 6).

Page 33: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

28

Table 4: Mean Abnormal Return and Mean Cumulative Abnormal Return for the Airline Sample Using the Market Model

Page 34: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

29

Table 5: Mean Abnormal Return and Mean Cumulative Abnormal Return for the Airline Sample Using the Market-Adjusted Returns

Page 35: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

30

Table 6: Mean Abnormal Return and Mean Cumulative Abnormal Return for the Airline Sample Using the Comparison Period Mean-Adjusted Returns

The stock price reaction of involved aircraft manufacturers using the three different

benchmark models does not show statistically significant negative abnormal returns.

According to the market model, negative abnormal returns of -0.12% one day, -0.19% three

days and -0.17% six days subsequent to the aviation catastrophe are reported. Thus, relatively

minor mean cumulative abnormal returns of -0.17% one day, 0.06% two days, 0.05% five

days and -0.31% ten days subsequent to the air crash are statistically not significant (see Table

Page 36: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

31

7). Using the market adjusted model, negative abnormal returns of -0.09% one day, -0.22%

three days and -0.13% six days subsequent to the aviation catastrophe are observed. Hence,

the market adjusted returns show mean cumulative abnormal returns of -0.12% one day,

0.17% two days, 0.19% five days and 0.10% ten days subsequent to the air crash (see Table

8). Looking at the comparison-period mean adjusted model, no significant stock price reaction

is not recognized either with the minor negative abnormal returns of -0.14% one day, -0.36%

three days and 0.23%% six days subsequent to the tragedy. The model delivers negative mean

cumulative abnormal returns of -0.24% one day, -0.01% two days, -0.16% five days and -

0.82% ten days subsequent to the event with no statistical significance (see Table 9).

Page 37: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

32

Table 7: Mean Abnormal Return and Mean Cumulative Abnormal Return for the Aircraft Manufacturers Sample Using the Market Model

Page 38: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

33

Table 8: Mean Abnormal Return and Mean Cumulative Abnormal Return for the Aircraft Manufacturers Sample Using the Market-Adjusted Returns

Page 39: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

34

Table 9: Mean Abnormal Return and Mean Cumulative Abnormal Return for the Aircraft Manufacturers Sample Using the Comparison Period Mean-Adjusted Returns

The results of cross-sectional analysis do not show statistically significant evidence of the

stock price reaction effect explained by the tested variables (see results of the regression

analysis in the appendix of this paper). As no abnormal returns are observed for aircraft

manufacturers, the cross-sectional analysis for these companies is not performed.

Page 40: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

CONCLUSIONS

The impact of aviation catastrophes on stock price reaction of airlines and aircraft

manufacturers is examined in this paper. The null hypothesis to be tested is that the mean

excess return during the event windows subsequent to an air crash is equal to zero. The results

show that involved airlines experience an average stock price decline of -2.05% within one

trading day after the aviation catastrophe, which is statistically significant. However, involved

aircraft manufacturers experience a stock price decline of only -0.1% during that time period,

which is not statistically significant and can be interpreted as no reaction. These results are

consistent with those of Walker, Thiengtham and Lin. According to the conducted cross-

sectional analysis, the magnitude of the price decline does not appear to be driven by the

characteristics of the company or the aviation catastrophe particularities. This result is not

consistent with the study mentioned before. From the perspective of an investor who holds the

shares of the respective airlines and aircraft manufacturers (to less extent) in their portfolios,

an aviation catastrophe obviously has a wealth diminishing effect that persists from one to six

days subsequent to the tragedy. Under the efficient market hypothesis, such prolonged

downward trends should not be observed as they could be exploited by short-selling the stock

of the respective airlines immediately after the news announcement and then repurchasing the

stock at a lower price one week later.

Page 41: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

REFERENCES

FAA http://www.faa.gov/data_research/aviation/aerospace_forecasts/2010-2030/media/2010%20Forecast%20Doc.pdf

EASA http://easa.europa.eu/communications/docs/annual-safety-review/2009/RLY_EASA_Annual_100722.pdf

The Effect of Aviation Disasters on the Air Transport Industry: A Financial Market Perspective; Author(s): Don M. Chance and Stephen P. Ferris; Source: Journal of Transport Economics and Policy, Vol. 21, No. 2 (May, 1987), pp. 151-165; Published by: University of Bath and The London School of Economics and Political Science

On the Performance of Airlines and Airplane Manufacturers Following Aviation Disasters; Authors(s): Thomas John Walker, Dolruedee, Jum Thiengtham, Michael Yi Lin; Source: Canadian Journal of Administrative Sciences, Vol. 22, No. 1(What? 2004), pp. 21-34; Published by: Concordia University

Page 42: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

BIBLIOGRAPHY

Market Forces and Commercial Aircraft Safety. Author(s): Andrew J. Chalk; Source: The Journal of Industrial Economics, Vol. 36, No. 1 (Sep., 1987), pp. 61-81; Published by: Blackwell Publishing Market Forces and Aircraft Safety: The Case of the DC-10. Author(s): Andrew J. Chalk; Source: Economic Inquiry, Vol. 24, (Jan., 1986), pp. 43-60; Published by: Blackwell Publishing The Effect of Aviation Disasters on the Air Transport Industry: A Financial Market Perspective. Author(s): Don M. Chance and Stephen P. Ferris; Source: Journal of Transport Economics and Policy, Vol. 21, No. 2 (May, 1987), pp. 151-165; Published by: University of Bath and The London School of Economics and Political Science The Competitive Impact of Air Crashes: Stock Market Evidence. Author(s): Jean-Claude Bosch, E. Woodrow Eckard, Vijay Singal; Source: Journal of Law and Economics, Vol. 41, No. 2 (Oct., 1998), pp. 503-519; Published by: The University of Chicago Press Market Incentives for Safe Commercial Airline Operation. Author(s): Severin Borenstein and Martin B. Zimmerman; Source: The American Economic Review, Vol. 78, No. 5 (Dec., 1988), pp. 913-935; Published by: American Economic Association Crisis in the Cockpit? The Role of Market Forces in Promoting Air Travel Safety. Author(s): Mark L. Mitchell and Michael T. Maloney; Source: Journal of Law and Economics, Vol. 32, No. 2 (Oct., 1989), pp. 329-355; Published by: The University of Chicago Press Large Losses, Risk Management and Stock Prices. Author(s): C. Ronald Sprecher and Mars A. Pertl; Source: The Journal of Risk and Insurance, Vol. 50, No. 1 (Mar., 1983), pp. 107-117; Published by: American Risk and Insurance Association Large Losses, Risk Management and Stock Returns in the Airline Industry. Author(s): Wallace N. Davidson, III, P. R. Chandy, Mark Cross; Source: The Journal of Risk and Insurance, Vol. 54, No. 1 (Mar.,1987), pp. 162-172; Published by: American Risk and Insurance Association On the Performance of Airlines and Airplane Manufacturers Following Aviation Disasters. Author(s): Thomas John Walker, Dolruedee Jum Thiengtham, Yi Lin Michael; Source: Canadian Journal of Administrative Sciences, (Mar., 2005), Vol. 22 Issue 1, pp. 21-34 Fear of Flying? Economic Analyses of Airline Safety. Author(s): Nancy L. Rose; Source: The Journal of Economic Perspectives, Vol. 6, No. 2 (Spring, 1992), pp. 75-94; Published by: American Economic Association The Adjustment of Stock Prices to New Information. Author(s): Fama, E., Fisher. L. & Jensen, M.; (1969); International Economic Review. / a ( l ) . 1-21

Cody, Ronald R. & Smith, Jeffrey K. (1997). Applied Statistics and the SAS Programming Language. Prentice Hall, Upper Saddle River, New Jersey 07458 4th Edition

Page 43: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

38

Studenmund, A.H. (2005). Using Econometrics: A Practical Guide, Fifth Edition, Pearson Addison Wesley Publishers. Stock, J. H. & Watson, M.W. (2006). Introduction to Econometrics, Second Edition. Pearson Addison Wesley Publishers.

Page 44: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

APPENDIX

ABNORMAL RETURNS FOR AIRLINES

1 ESTIMATION PERIOD: Ends 46 days before the event date; 255 days in length. TOTAL SECURITY-EVENTS IN REQUEST FILE: 30

SECURITY-EVENTS DROPPED: 0 SECURITY-EVENTS WITH USEABLE RETURNS: 30 Minimum days of return data required for parameter estimation: 3 Any non-trading date has been converted to the next trading date.

NOTE: Useable returns means all nonmissing returns except the first day after a missing estimation period return.

2 Results of Daily Security Return Data Input

Esti- mation Event Period Period Returns Returns Ticker Event Date <=255 <=21 Reason if no useable returns AF 07/25/2000 142 15 AFLT 09/14/2008 137 15 AFLYY 06/01/2009 137 14 ALK 01/31/2000 142 15 AMR 10/31/1994 140 15 AMR 12/20/1995 141 14 AMR 09/11/2001 137 10 AMR 11/12/2001 137 14 CEA 11/21/2004 130 14 DALRQ 08/02/1985 135 15

DALRQ 08/27/2006 137 14 GOL 09/29/2006 136 15 KNAL 01/30/2000 133 14 KNAL 05/05/2007 130 14 KP 08/06/1997 164 17 PNCL 02/12/2009 138 13 SAS 10/08/2001 136 15 SIA 12/19/1997 136 14 SIA 10/31/2000 138 14 SRN 09/02/1998 136 15 TAM 07/17/2007 138 15 TWAIQ 07/17/1996 139 15 UALAQ 07/19/1989 139 15 UALAQ 03/03/1991 137 15 UALAQ 09/11/2001 137 10 UAWGQ 12/07/1987 142 15 UAWGQ 02/01/1991 137 15

UAWGQ 03/22/1992 137 15 UAWGQ 07/02/1994 137 14 UAWGQ 09/08/1994 136 13

Page 45: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

40

3 Parameter Estimates and Estimation Period Statistics

Market Index=Value Mean % of Raw Market Total Residual Event Total Returns Model Res- Return Standard Autocor- Ticker Date Return >0 Alpha Beta iduals>0 Variance Deviation relation* DALRQ 02AUG1985 0.00177 43.70% 0.00073 1.29 48.89% 0.00031 0.01628 0.1332 UAWGQ 07DEC1987 0.00111 46.48% 0.00018 0.69 49.30% 0.00056 0.02232 0.0141

UALAQ 19JUL1989 0.00208 46.76% 0.00057 0.93 44.60% 0.00018 0.01251 0.0081 UAWGQ 01FEB1991 -0.00489 35.04% -0.00438 1.02 49.64% 0.00095 0.02918 -0.0094 UALAQ 03MAR1991 0.00002 48.18% 0.00057 0.64 43.07% 0.00085 0.02858 -0.0906 UAWGQ 22MAR1992 0.00136 35.77% 0.00095 1.41 44.53% 0.00117 0.03315 0.0422 UAWGQ 02JUL1994 -0.00527 29.20% -0.00554 0.71 49.64% 0.00082 0.02854 -0.0561 UAWGQ 08SEP1994 -0.00285 32.35% -0.00344 1.06 46.32% 0.00096 0.03062 -0.0245 AMR 31OCT1994 -0.00099 40.71% -0.00135 1.33 48.57% 0.00023 0.01417 -0.0112 AMR 20DEC1995 0.00130 46.81% 0.00097 0.43 48.94% 0.00026 0.01607 0.0269 TWAIQ 17JUL1996 0.00733 44.60% 0.00622 1.54 41.01% 0.00219 0.04654 0.1293 KP 06AUG1997 0.00101 45.12% 0.00052 0.31 42.47% 0.00060 0.02522 0.0945 SIA 19DEC1997 -0.00050 44.12% -0.00053 0.06 48.53% 0.00025 0.01586 -0.0479 SRN 02SEP1998 0.00057 48.53% 0.00001 0.57 48.53% 0.00030 0.01677 0.0864 KNAL 30JAN2000 0.00155 50.38% 0.00150 0.10 47.37% 0.00078 0.02817 -0.1450 ALK 31JAN2000 -0.00081 43.66% -0.00132 0.66 42.25% 0.00053 0.02254 -0.0868 AF 25JUL2000 0.00017 47.18% 0.00012 0.05 47.18% 0.00059 0.02433 0.0077 SIA 31OCT2000 -0.00011 50.00% 0.00006 0.41 51.45% 0.00051 0.02227 -0.0016 AMR 11SEP2001 0.00086 41.61% 0.00213 1.33 48.91% 0.00080 0.02460 -0.0781

UALAQ 11SEP2001 0.00079 45.99% 0.00185 1.11 48.18% 0.00086 0.02681 -0.1278 SAS 08OCT2001 -0.00069 47.79% -0.00033 0.27 47.79% 0.00052 0.02264 -0.1016 AMR 12NOV2001 -0.00080 37.96% 0.00101 1.21 50.36% 0.00072 0.02355 -0.0720 CEA 21NOV2004 0.00294 53.08% 0.00283 1.15 46.15% 0.00061 0.02365 -0.0223 DALRQ 27AUG2006 -0.00434 41.61% -0.00410 0.33 50.36% 0.00248 0.04973 0.1237 GOL 29SEP2006 0.00388 48.53% 0.00172 2.65 41.91% 0.00150 0.03379 -0.0793 KNAL 05MAY2007 -0.00174 46.92% -0.00187 0.34 52.31% 0.00048 0.02199 0.0389 TAM 17JUL2007 0.00210 50.00% -0.00006 1.88 44.20% 0.00063 0.02310 0.0769 AFLT 14SEP2008 -0.00123 49.64% -0.00092 0.72 49.64% 0.00027 0.01484 0.0961 PNCL 12FEB2009 -0.00517 38.41% -0.00277 0.80 42.75% 0.00357 0.05757 0.1556 AFLYY 01JUN2009 -0.00146 48.18% 0.00039 1.26 51.09% 0.00206 0.03454 -0.1172 Mean -0.00007 44.28% -0.00014 0.88 47.20% 0.00088 0.02633 -0.0013 Median 0.00010 46.23% 0.00015 0.76 48.35% 0.00062 0.02399 -0.0055 * First order autocorrelation of market model abnormal returns

Page 46: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

41

4 Market Model, Value Weighted Index Mean Abnormal Positive: Patell Generalized Day N Return Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ -10 18 -0.49% 7:11 -1.375$ -0.706 -9 14 0.16% 8:6 -0.037 0.745 -8 9 -0.02% 2:7( 0.101 -1.501$ -7 16 -0.61% 7:9 -1.946* -0.276

-6 18 -0.26% 8:10 -0.864 -0.234 -5 21 0.46% 12:9 1.058 0.913 -4 18 -0.56% 7:11 -1.765* -0.706 -3 18 0.38% 9:9 1.459$ 0.238 -2 14 1.37% 9:5 3.343*** 1.281 -1 18 -0.17% 11:7 0.384 1.182 0 20 -0.48% 11:9 -1.248 0.699 +1 22 -2.12% 7:15( -4.067*** -1.445$ +2 21 -1.86% 9:12 -5.267*** -0.398 +3 21 0.18% 7:14 1.117 -1.273 +4 18 0.32% 7:11 0.352 -0.706 +5 18 1.17% 8:10 5.180*** -0.234 +6 17 -6.40% 8:9 -10.739*** -0.011 +7 22 1.83% 13:9 3.319*** 1.117 +8 23 -0.65% 7:16( -0.796 -1.610$ +9 23 0.81% 15:8> 1.188 1.731* +10 23 0.11% 11:12 0.306 0.060

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

5 Market Model, Value Weighted Index Mean Cumulative Precision Abnormal Weighted Positive: Patell Generalized Days N Return CAAR Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ (0,+1) 26 -2.16% -2.31% 7:19< -3.849*** -2.071* (0,+2) 27 -3.53% -4.59% 10:17 -6.183*** -1.058

(0,+5) 28 -2.31% -1.71% 10:18 -2.076* -1.217 (0,+10) 30 -4.24% -4.62% 11:19 -2.830** -1.155 ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

Page 47: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

42

6 Market Adjusted Returns, Value Weighted Index Mean Abnormal Positive: Patell Generalized Day N Return Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ -10 18 -0.62% 6:12 -1.568$ -1.145 -9 14 0.14% 7:7 0.036 0.240 -8 9 0.01% 3:6 0.415 -0.810 -7 16 -0.56% 8:8 -1.761* 0.257

-6 18 -0.12% 8:10 -0.480 -0.200 -5 21 0.59% 13:8) 1.419$ 1.387$ -4 18 -0.56% 7:11 -1.754* -0.673 -3 18 0.36% 9:9 1.340$ 0.272 -2 14 1.36% 9:5) 3.152*** 1.311$ -1 18 -0.14% 11:7 0.487 1.217 0 20 -0.56% 9:11 -1.420$ -0.161 +1 22 -2.09% 7:15( -4.039*** -1.408$ +2 21 -1.94% 8:13 -5.438*** -0.799 +3 21 0.01% 8:13 0.810 -0.799 +4 18 0.04% 6:12 -0.267 -1.145 +5 18 1.20% 8:10 5.108*** -0.200 +6 17 -6.35% 8:9 -10.727*** 0.021 +7 22 1.71% 12:10 3.062** 0.728 +8 23 -0.52% 7:16( -0.561 -1.573$ +9 23 0.55% 13:9 0.653 0.934 +10 23 0.10% 9:14 0.261 -0.737

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

7 Market Adjusted Returns, Value Weighted Index

Mean Cumulative Precision Abnormal Weighted Positive: Patell Generalized Days N Return CAAR Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ (0,+1) 26 -2.19% -2.68% 5:21<< -3.944*** -2.817** (0,+2) 27 -3.62% -5.36% 6:21<< -6.359*** -2.559**

(0,+5) 28 -2.69% -2.58% 9:19( -2.591** -1.554$ (0,+10) 30 -4.76% -6.18% 9:21< -3.389*** -1.844* ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

Page 48: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

43

8 Comparison-Period Mean Adjusted Returns Mean Abnormal Positive: Patell Generalized Day N Return Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ -10 18 -0.61% 8:10 -1.680* -0.206 -9 19 -0.07% 11:8 -0.476 0.964 -8 16 -0.01% 4:12< -0.057 -1.752* -7 21 -0.97% 9:12 -2.878** -0.368

-6 23 -0.27% 11:12 -1.032 0.093 -5 23 0.19% 13:10 0.401 0.928 -4 23 -0.79% 7:16( -2.146* -1.579$ -3 19 0.54% 12:7) 1.383$ 1.423$ -2 19 0.43% 10:9 1.818* 0.504 -1 18 -0.09% 11:7 0.692 1.211 0 20 -0.46% 10:10 -1.099 0.281 +1 22 -1.94% 7:15( -3.716*** -1.414$ +2 21 -1.82% 10:11 -4.782*** 0.070 +3 22 0.09% 7:15( 0.458 -1.414$ +4 18 0.36% 8:10 0.434 -0.206 +5 18 1.30% 8:10 5.527*** -0.206 +6 17 -6.45% 8:9 -9.556*** 0.016 +7 22 1.59% 13:9 2.702** 1.149 +8 23 -0.77% 5:18<< -0.886 -2.414** +9 23 0.25% 13:10 0.000 0.928 +10 23 -0.45% 9:14 -0.913 -0.743

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

9 Comparison-Period Mean Adjusted Returns Mean Cumulative Precision Abnormal Weighted Positive: Patell Generalized Days N Return CAAR Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ (0,+1) 26 -2.00% -2.46% 6:20<< -3.488*** -2.430** (0,+2) 27 -3.34% -4.91% 8:19< -5.608*** -1.794* (0,+5) 28 -2.08% -1.62% 8:20< -1.750* -1.940*

(0,+10) 30 -5.17% -6.05% 10:20( -3.330*** -1.485$ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

Page 49: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

44

ABNORMAL RETURNS FOR MANUFACTURERS

1 ESTIMATION PERIOD: Ends 46 days before the event date; 255 days in length. TOTAL SECURITY-EVENTS IN REQUEST FILE: 87 SECURITY-EVENTS DROPPED: 5 SECURITY-EVENTS WITH USEABLE RETURNS: 82 Minimum days of return data required for parameter estimation: 3 Any non-trading date has been converted to the next trading date.

NOTE: Useable returns means all nonmissing returns except the first day after a missing estimation period return.

2 Results of Daily Security Return Data Input Esti- mation Event Period Period Returns Returns Ticker Event Date <=255 <=21 Reason if no useable returns BA 01/13/1982 140 15 BA 09/07/1982 139 14 BA 01/01/1985 142 13 BA 12/21/1988 142 14

BA 08/02/1989 139 15 BA 10/21/1989 137 15 BA 01/25/1990 139 14 BA 01/02/1991 141 13 BA 03/03/1991 137 15 BA 06/06/1992 137 15 BA 08/09/1994 139 15 BA 12/29/1994 139 12 BA 03/12/1995 140 15 BA 09/08/1995 135 14 BA 12/20/1995 141 14 BA 02/10/1996 139 14 BA 02/29/1996 137 14 BA 06/02/1996 140 14 BA 07/17/1996 139 15 BA 10/22/1996 140 15 BA 11/23/1996 140 14

BA 12/11/1996 140 15 BA 06/08/1997 138 15 BA 08/05/1997 138 15 BA 12/19/1997 140 14 BA 02/09/1998 141 13 BA 04/20/1998 137 14 BA 08/31/1999 138 14 BA 10/31/1999 141 15 BA 01/31/2000 142 15 BA 04/19/2000 139 14 BA 07/17/2000 140 15 BA 10/31/2000 140 15 BA 08/10/2001 136 15 BA 11/09/2001 133 15 BA 01/07/2002 137 13 BA 01/28/2002 136 14

BA 04/15/2002 132 15 BA 05/25/2002 132 14 BA 06/03/2003 137 14 BA 08/07/2003 135 14 BA 12/25/2003 141 12 BA 11/30/2004 141 14 BA 03/02/2005 139 14 * No data or security-event eliminated before count.

3 Results of Daily Security Return Data Input

Page 50: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

45

Esti- mation Event Period Period Returns Returns Ticker Event Date <=255 <=21 Reason if no useable returns BA 05/09/2005 141 15 BA 08/14/2005 140 15 BA 08/16/2005 140 15 BA 08/23/2005 139 15

BA 10/12/2005 140 15 BA 10/22/2005 140 15 BA 09/29/2006 136 15 BA 10/29/2006 139 15 BA 01/01/2007 142 12 BA 05/05/2007 138 15 BA 09/16/2007 136 15 BA 11/30/2007 138 14 BA 04/15/2008 136 15 BA 08/20/2008 135 15 BA 08/24/2008 136 14 BA 09/14/2008 136 15 BA 01/25/2010 141 14 BA 05/22/2010 138 14 BDRBF 11/21/2004 137 14 BDRBF 08/27/2006 137 14 BDRBF 12/02/2009 141 14

EAD 08/23/2000 * * No data available for estimation period. EAD 12/11/2001 * * EAD's data end before 20011211. EAD 08/07/2006 * * EAD's data end before 20060807. EADSF 07/17/2007 112 15 EADSY 12/05/2010 * * Database ends on 20100823. EADSY 01/30/2000 * * EADSY's data start after 20000130. EADSY 01/06/2009 88 14 EADSY 06/29/2009 138 13 LK 01/21/1985 140 15 LK 02/08/1985 137 14 MCDGL 06/09/1985 123 15 MCDGL 12/12/1985 138 14 MCDGL 08/31/1986 137 14 MCDGL 08/16/1987 138 15 MCDGL 11/15/1987 142 15 MCDGL 07/06/1989 46 13 MCDGL 07/19/1989 46 15

MCDGL 05/05/1990 140 15 MCDGL 12/21/1992 143 14 MCDGL 02/07/1994 142 15 MCDGL 11/01/1995 140 15 MCDGL 11/05/1996 141 15 * No data or security-event eliminated before count.

Page 51: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

46

4 Parameter Estimates and Estimation Period Statistics

Market Index=Value Mean % of Raw Market Total Residual Event Total Returns Model Res- Return Standard Autocor- Ticker Date Return >0 Alpha Beta iduals>0 Variance Deviation relation* BA 13JAN1982 -0.00137 38.57% -0.00123 1.09 50.00% 0.00035 0.01731 0.0536 BA 07SEP1982 -0.00081 38.13% -0.00074 1.01 50.36% 0.00038 0.01861 -0.0015

BA 01JAN1985 0.00208 44.37% 0.00147 1.23 47.18% 0.00030 0.01527 0.0573 LK 21JAN1985 0.00234 47.14% 0.00157 1.64 50.71% 0.00044 0.01816 -0.3024 LK 08FEB1985 0.00178 46.72% 0.00104 1.49 51.09% 0.00035 0.01591 -0.1527 MCDGL 09JUN1985 0.00504 53.66% 0.00466 0.38 39.02% 0.00178 0.04234 0.0222 MCDGL 12DEC1985 -0.00036 53.62% -0.00151 1.33 55.80% 0.00023 0.01367 0.1411 MCDGL 31AUG1986 0.00161 57.66% 0.00074 0.46 54.74% 0.00014 0.01133 -0.0011 MCDGL 16AUG1987 -0.00022 44.93% -0.00126 0.53 47.83% 0.00015 0.01173 0.0149 MCDGL 15NOV1987 0.00021 44.37% -0.00049 0.36 47.18% 0.00013 0.01129 -0.0058 BA 21DEC1988 0.00202 47.89% 0.00164 0.63 46.48% 0.00018 0.01293 -0.0838 MCDGL 06JUL1989 0.00401 65.22% 0.00348 0.24 45.65% 0.00010 0.01024 0.3567 MCDGL 19JUL1989 0.00401 63.04% 0.00390 0.04 41.30% 0.00010 0.01030 0.3523 BA 02AUG1989 0.00214 49.64% 0.00087 0.87 45.32% 0.00018 0.01243 0.1384 BA 21OCT1989 0.00296 55.47% 0.00175 1.00 46.72% 0.00022 0.01342 0.1283 BA 25JAN1990 0.00213 51.80% 0.00121 1.05 46.04% 0.00028 0.01541 0.0334 MCDGL 05MAY1990 -0.00121 37.86% -0.00138 0.20 55.00% 0.00016 0.01271 0.2433 BA 02JAN1991 0.00015 48.94% 0.00100 0.98 48.23% 0.00040 0.01751 0.0404 BA 03MAR1991 -0.00020 47.45% 0.00082 0.89 46.72% 0.00049 0.02035 -0.0385

BA 06JUN1992 -0.00090 40.88% -0.00078 0.94 47.45% 0.00033 0.01697 0.0208 MCDGL 21DEC1992 -0.00228 43.36% -0.00231 -0.13 47.55% 0.00078 0.02802 -0.0434 MCDGL 07FEB1994 0.00240 53.52% 0.00224 0.24 46.48% 0.00033 0.01823 -0.0531 BA 09AUG1994 0.00102 46.04% 0.00099 0.10 46.04% 0.00029 0.01714 -0.0479 BA 29DEC1994 -0.00051 43.88% -0.00055 0.52 51.08% 0.00014 0.01175 -0.0516 BA 12MAR1995 0.00032 45.00% 0.00036 0.40 48.57% 0.00014 0.01198 -0.0156 BA 08SEP1995 0.00225 48.15% 0.00208 0.28 45.93% 0.00020 0.01427 -0.0823 MCDGL 01NOV1995 0.00339 53.57% 0.00315 0.37 44.29% 0.00022 0.01462 -0.2038 BA 20DEC1995 0.00195 48.23% 0.00167 0.35 46.10% 0.00021 0.01459 -0.1219 BA 10FEB1996 0.00199 51.08% 0.00197 0.42 46.04% 0.00025 0.01554 -0.1170 BA 29FEB1996 0.00180 50.36% 0.00157 0.60 45.99% 0.00025 0.01570 -0.1339 BA 02JUN1996 0.00075 45.00% 0.00029 0.74 42.86% 0.00024 0.01535 -0.0834 BA 17JUL1996 0.00131 47.48% 0.00066 0.89 44.60% 0.00027 0.01599 -0.0812 BA 22OCT1996 0.00089 48.57% 0.00034 1.29 43.57% 0.00023 0.01411 -0.0172 MCDGL 05NOV1996 -0.00017 43.97% -0.00080 1.15 46.10% 0.00022 0.01426 -0.0928 BA 23NOV1996 0.00188 53.57% 0.00104 1.30 43.57% 0.00021 0.01350 -0.0094 BA 11DEC1996 0.00114 51.43% 0.00051 1.47 45.00% 0.00020 0.01294 0.0138

BA 08JUN1997 0.00045 49.28% 0.00004 1.06 50.72% 0.00020 0.01331 0.0111 BA 05AUG1997 0.00097 49.28% 0.00007 1.07 50.72% 0.00023 0.01411 0.0022 BA 19DEC1997 0.00025 48.57% -0.00050 1.30 50.00% 0.00044 0.01898 -0.0759 BA 09FEB1998 0.00032 47.52% -0.00035 1.06 47.52% 0.00042 0.01889 -0.0273 BA 20APR1998 -0.00050 44.53% -0.00052 1.05 47.45% 0.00051 0.02092 -0.0010 BA 31AUG1999 0.00145 54.35% 0.00094 0.56 50.00% 0.00076 0.02733 0.1359 BA 31OCT1999 0.00329 55.32% 0.00316 0.44 45.39% 0.00050 0.02223 0.0741 BA 31JAN2000 0.00353 57.75% 0.00328 0.33 49.30% 0.00048 0.02175 0.1022 BA 19APR2000 0.00131 52.52% 0.00103 0.33 51.08% 0.00046 0.02126 0.0228 BA 17JUL2000 0.00277 52.86% 0.00246 0.49 50.71% 0.00069 0.02584 0.0625 BA 31OCT2000 0.00410 54.29% 0.00416 0.29 52.14% 0.00064 0.02532 0.0664 BA 10AUG2001 -0.00125 45.59% -0.00076 0.73 49.26% 0.00055 0.02205 -0.0677 BA 09NOV2001 -0.00338 45.11% -0.00188 0.88 51.88% 0.00054 0.02119 0.0071 BA 07JAN2002 -0.00167 45.99% -0.00155 1.11 50.36% 0.00061 0.02150 0.0968 BA 28JAN2002 -0.00007 50.00% -0.00043 1.13 49.26% 0.00059 0.02137 0.0697 BA 15APR2002 0.00060 55.30% 0.00066 1.24 51.52% 0.00059 0.02142 0.0897

BA 25MAY2002 0.00112 55.30% 0.00119 1.38 46.97% 0.00058 0.02053 0.0378 BA 03JUN2003 -0.00082 48.18% -0.00144 0.97 45.26% 0.00058 0.02049 -0.0114 BA 07AUG2003 0.00294 53.33% 0.00179 0.90 47.41% 0.00043 0.01821 0.0522 BA 25DEC2003 0.00307 53.90% 0.00133 1.21 46.81% 0.00030 0.01461 -0.0526 BDRBF 21NOV2004 -0.00172 41.61% -0.00159 -0.60 45.99% 0.00235 0.04863 -0.1044 BA 30NOV2004 0.00077 51.06% 0.00069 1.07 52.48% 0.00017 0.01143 -0.0976 BA 02MAR2005 0.00117 53.24% 0.00042 1.16 50.36% 0.00016 0.01097 -0.0578 BA 09MAY2005 0.00088 49.65% 0.00032 1.25 51.77% 0.00018 0.01163 0.0011 BA 14AUG2005 0.00094 50.00% 0.00035 1.23 50.00% 0.00018 0.01174 -0.2023 BA 16AUG2005 0.00144 50.71% 0.00084 1.14 47.86% 0.00021 0.01341 -0.2613 BA 23AUG2005 0.00123 50.36% 0.00068 1.07 48.20% 0.00022 0.01361 -0.2726 BA 12OCT2005 0.00100 51.43% 0.00092 1.09 47.14% 0.00021 0.01345 -0.3229 BA 22OCT2005 0.00046 50.00% 0.00033 1.05 47.86% 0.00021 0.01376 -0.2710 BDRBF 27AUG2006 0.00450 52.55% 0.00321 1.41 39.42% 0.00093 0.02927 0.0670

Page 52: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

47

BA 29SEP2006 0.00144 57.35% 0.00078 0.97 50.00% 0.00023 0.01354 0.0771 BA 29OCT2006 0.00122 54.68% 0.00045 1.03 46.76% 0.00024 0.01384 0.0886 BA 01JAN2007 0.00162 53.52% 0.00075 1.03 47.89% 0.00026 0.01450 0.0956 BA 05MAY2007 0.00097 52.17% 0.00001 0.98 44.93% 0.00020 0.01310 0.0048 EADSF 17JUL2007 0.00168 54.46% 0.00058 0.98 52.68% 0.00041 0.01971 0.1446 BA 16SEP2007 0.00158 55.88% 0.00137 0.66 42.65% 0.00012 0.01015 -0.1670 BA 30NOV2007 0.00017 52.17% -0.00017 0.79 48.55% 0.00016 0.01107 -0.0244 BA 15APR2008 -0.00145 44.12% -0.00101 0.79 47.79% 0.00026 0.01397 -0.0095 BA 20AUG2008 -0.00251 42.22% -0.00201 0.71 46.67% 0.00036 0.01745 0.1084 BA 24AUG2008 -0.00240 41.91% -0.00188 0.71 46.32% 0.00036 0.01753 0.1096

BA 14SEP2008 -0.00216 44.85% -0.00167 0.78 49.26% 0.00042 0.01887 0.1138 EADSY 06JAN2009 -0.00109 45.45% 0.00601 1.31 50.00% 0.00223 0.03457 -0.0396 EADSY 29JUN2009 0.00056 47.83% 0.00180 1.23 44.20% 0.00191 0.03289 -0.0359 BDRBF 02DEC2009 0.00386 52.48% -0.00155 1.92 46.81% 0.00189 0.03578 -0.0805 BA 25JAN2010 0.00251 49.65% 0.00025 1.00 48.94% 0.00048 0.01897 -0.0029 BA 22MAY2010 0.00267 52.17% 0.00245 0.95 42.75% 0.00035 0.01664 -0.0016 Mean 0.00099 49.65% 0.00067 0.85 47.78% 0.00044 0.01778 -0.0082 Median 0.00107 49.82% 0.00067 0.97 47.48% 0.00028 0.01562 -0.0016 * First order autocorrelation of market model abnormal returns

5 Market Model, Value Weighted Index Mean

Abnormal Positive: Patell Generalized Day N Return Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ -10 43 0.30% 24:19 2.033* 1.055 -9 38 -0.20% 18:20 -0.400 -0.050 -8 22 -0.80% 6:16< -5.524*** -1.925* -7 41 0.20% 22:19 1.550$ 0.754 -6 48 0.13% 22:26 0.330 -0.269 -5 57 0.19% 30:27 1.472$ 0.734 -4 55 -0.19% 25:30 -1.480$ -0.345 -3 46 0.59% 30:16>> 3.126*** 2.368** -2 41 -0.09% 19:22 -0.751 -0.184 -1 50 -0.09% 20:30 -0.953 -1.101 0 54 -0.08% 25:29 -1.418$ -0.218 +1 64 -0.12% 30:34 0.141 -0.144 +2 66 0.25% 38:28) 1.448$ 1.594$ +3 61 -0.19% 34:27 -0.646 1.245

+4 49 0.21% 28:21) 1.364$ 1.313$ +5 50 0.01% 23:27 0.162 -0.251 +6 50 -0.17% 22:28 -0.241 -0.535 +7 55 0.06% 25:30 0.209 -0.345 +8 64 0.13% 32:32 1.507$ 0.356 +9 66 -0.40% 28:38 -1.768* -0.870 +10 59 -0.11% 24:35 -1.845* -1.092 ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

6 Market Model, Value Weighted Index

Mean Cumulative Precision Abnormal Weighted Positive: Patell Generalized Days N Return CAAR Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ (0,+1) 71 -0.17% -0.23% 31:40 -0.807 -0.694 (0,+2) 81 0.06% 0.03% 39:42 0.262 0.067 (0,+5) 82 0.05% 0.19% 41:41 0.449 0.403 (0,+10) 82 -0.31% -0.20% 41:41 -0.342 0.403 ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

Page 53: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

48

7 Market Adjusted Returns, Value Weighted Index

Mean Abnormal Positive: Patell Generalized Day N Return Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ -10 43 0.29% 23:20 1.978* 0.544 -9 38 -0.20% 19:19 -0.182 0.081 -8 22 -0.68% 8:14 -4.673*** -1.218

-7 41 0.30% 24:17 1.991* 1.178 -6 48 0.27% 24:24 1.082 0.091 -5 57 0.35% 32:25 2.279* 1.027 -4 55 -0.15% 25:30 -1.228 -0.577 -3 46 0.58% 31:15>> 3.007** 2.449** -2 41 -0.07% 17:24 -0.796 -1.009 -1 50 0.05% 23:26 -0.381 -0.473 0 54 -0.05% 25:29 -1.369$ -0.448 +1 64 -0.09% 33:31 0.205 0.355 +2 66 0.34% 38:27) 1.882* 1.338$ +3 61 -0.22% 31:30 -0.837 0.231 +4 49 0.27% 26:23 1.624$ 0.521 +5 50 0.03% 24:26 0.474 -0.190 +6 50 -0.13% 21:28 -0.067 -1.038 +7 55 0.13% 25:30 0.540 -0.577 +8 64 0.21% 36:28 1.877* 1.105 +9 66 -0.29% 28:38 -1.284$ -1.124

+10 59 -0.04% 25:34 -1.468$ -1.071 ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

8 Market Adjusted Returns, Value Weighted Index Mean Cumulative Precision Abnormal Weighted Positive: Patell Generalized Days N Return CAAR Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ (0,+1) 71 -0.12% -0.23% 32:39 -0.726 -0.720 (0,+2) 81 0.17% 0.14% 40:40 0.594 0.007

(0,+5) 82 0.19% 0.40% 44:38 0.818 0.782 (0,+10) 82 0.10% 0.32% 44:38 0.469 0.782 ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

Page 54: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

49

9 Comparison-Period Mean Adjusted Returns Mean Abnormal Positive: Patell Generalized Day N Return Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ -10 43 0.24% 23:20 1.826* 0.618 -9 52 -0.39% 19:33< -2.027* -1.766* -8 49 -0.28% 22:27 -2.209* -0.544 -7 53 -0.04% 28:25 0.419 0.590

-6 63 0.04% 31:32 -0.035 0.067 -5 66 0.11% 32:34 1.185 -0.048 -4 60 -0.08% 26:34 -0.941 -0.844 -3 52 0.55% 33:19> 2.759** 2.118* -2 52 -0.36% 24:28 -2.190* -0.379 -1 50 -0.02% 20:30 -0.812 -1.242 0 54 -0.15% 23:31 -1.868* -0.910 +1 64 -0.14% 28:36 -0.040 -0.805 +2 66 0.24% 34:32 1.185 0.444 +3 61 -0.36% 31:30 -1.249 0.318 +4 49 0.07% 25:24 0.612 0.314 +5 50 0.13% 29:21) 0.717 1.304$ +6 50 -0.23% 21:29 -0.246 -0.959 +7 55 -0.05% 23:32 -0.387 -1.033 +8 64 0.11% 32:32 1.398$ 0.195 +9 66 -0.49% 25:41< -2.118* -1.772* +10 59 -0.25% 25:34 -2.397** -0.985

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

10 Comparison-Period Mean Adjusted Returns Mean Cumulative Precision Abnormal Weighted Positive: Patell Generalized Days N Return CAAR Negative Z Sign Z ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ (0,+1) 71 -0.24% -0.41% 27:44< -1.235 -1.813* (0,+2) 81 -0.01% -0.16% 35:46 -0.237 -1.003 (0,+5) 82 -0.16% -0.14% 40:42 -0.252 0.000

(0,+10) 82 -0.82% -0.95% 34:48( -1.348$ -1.326$ ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ The symbols $,*,**, and *** denote statistical significance at the 0.10, 0.05, 0.01 and 0.001 levels, respectively, using a generic one-tail test. The symbols (,< or ),> etc. correspond to $,* and show the direction and generic one-tail significance of the generalized sign test.

Page 55: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

50

Cross-sectional Analysis

1 Market model (MM)

Regression Statistics

Multiple R 0.581712443

R Square 0.338389366

Adjusted R Square 0.127876892

Standard Error 0.052840962

Observations 30

ANOVA

df SS MS F Significance F

Regression 7 0.031417986 0.004488284 1.607455175 0.185575944

Residual 22 0.06142768 0.002792167

Total 29 0.092845665

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 99.0% Upper 99.0%

Intercept 0.003211757 0.032334569 0.099328896 0.921776686 -0.063846035 0.070269549 -0.087931506 0.09435502

Death -9.75283E-05 0.000159315 -0.612172259 0.546699416 -0.000427928 0.000232871 -0.000546599 0.000351542

US or non-US 0.005808812 0.023040252 0.252115838 0.803291934 -0.041973745 0.05359137 -0.059136037 0.070753662

Terr. attacks -0.023112412 0.0236251 -0.978299027 0.338562513 -0.07210787 0.025883046 -0.089705805 0.043480981

Flight crew -0.025624079 0.028728483 -0.891939861 0.382077372 -0.085203305 0.033955147 -0.106602664 0.055354506

Air Traff. Control 0.041847061 0.039067479 1.071148215 0.295713706 -0.039173932 0.122868053 -0.068274633 0.151968754

Techn. Integr. -0.05648663 0.034328611 -1.645467979 0.114086915 -0.12767981 0.014706551 -0.153250609 0.04027735

Crime and terror 0.03158484 0.041885045 0.754083925 0.458794664 -0.055279428 0.118449107 -0.086478886 0.149648565

2 Market Adjusted Returns (MAR)

Page 56: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

51

Regression Statistics

Multiple R 0.592672792

R Square 0.351261038

Adjusted R Square 0.144844096

Standard Error 0.048733737

Observations 30

ANOVA

df SS MS F Significance F

Regression 7 0.028290597 0.004041514 1.701706429 0.160492214

Residual 22 0.052249498 0.002374977

Total 29 0.080540095

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 99.0% Upper 99.0%

Intercept 0.005427702 0.029821267 0.182007748 0.857243272 -0.05641782 0.067273223 -0.078631174 0.089486577

Deaths -2.7285E-05 0.000146932 -0.185698515 0.854382685 -0.000332003 0.000277433 -0.00044145 0.00038688

US or non-US 0.001974713 0.021249378 0.092930383 0.926799967 -0.0420938 0.046043226 -0.057922101 0.061871526

Terr. attacks -0.023943393 0.021788767 -1.098886988 0.283698156 -0.06913053 0.021243744 -0.085360612 0.037473827

Flight crew -0.038494614 0.026495474 -1.452875097 0.160373907 -0.093442864 0.016453635 -0.113178892 0.036189664

Air Traff. Control 0.023791642 0.036030841 0.660313258 0.515907586 -0.050931748 0.098515031 -0.077770508 0.125353792

Techn. Integr. -0.066504625 0.031660315 -2.10056738 0.047362972 -0.132164099 -0.000845151 -0.15574733 0.02273808

Crime and terror 0.023180643 0.038629403 0.600077705 0.554585841 -0.056931835 0.103293122 -0.08570622 0.132067507

3 Comparison-Period Mean Adjusted Returns (CP)

Regression Statistics

Page 57: STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT ...STOCK PRICE REACTION OF AIRLINES AND AIRCRAFT MANUFACTURERS FOLLOWING AVIATION CATASTROPHES Pavel Šepel ... An event study methodology

52

Multiple R 0.546368489

R Square 0.298518526

Adjusted R Square 0.075319875

Standard Error 0.057875665

Observations 30

ANOVA

df SS MS F Significance F

Regression 7 0.031359544 0.004479935 1.337456675 0.280059851

Residual 22 0.073691036 0.003349593

Total 29 0.105050581

Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Lower 99.0% Upper 99.0%

Intercept -0.005015098 0.035415417 -0.141607761 0.88867858 -0.078462176 0.068431981 -0.104842518 0.094812322

Deaths -7.74052E-05 0.000174495 -0.443595969 0.661666758 -0.000439285 0.000284475 -0.000569263 0.000414453

US or non-US 0.013811505 0.025235534 0.547303871 0.589679572 -0.038523789 0.0661468 -0.057321309 0.08494432

Terr. attacks -0.023298871 0.025876106 -0.900400959 0.37765801 -0.076962631 0.030364889 -0.096237303 0.049639561

Flight crew -0.025458444 0.031465741 -0.809084503 0.427130568 -0.090714397 0.039797509 -0.114152692 0.063235805

Air Traff. Control 0.044472349 0.04278984 1.03932029 0.309944187 -0.044268348 0.133213046 -0.076141772 0.165086471

Techn. Integr. -0.049535863 0.037599451 -1.317462412 0.201240325 -0.127512352 0.028440625 -0.155519544 0.056447817

Crime and terror 0.036357517 0.045875865 0.792519478 0.436521225 -0.058783204 0.131498238 -0.092955356 0.16567039