Changes in equity returns and volatility across different
Australian industries following the recent terrorist attacks

Keywords:
Terrorism
Equity market
Abnormal returns
Non-parametric test
Parametric test
Systematic risk
Australia

We investigate the impact of five recent terrorist attacks on equities

澳洲留學(xué)生dissertation

listed on the Australian Stock Exchange. Following the Global Industry
Classification Standard, we analyse how these events affect the
different sectors in Australia. Using parametric and non-parametric
tests, we investigate the relationship between stock returns for
equities listed in these sectors and terrorist attacks. We report
significant short term negative abnormal returns around the
September 11 attacks and to a lesser extent, the Madrid and London
bombings. Our evidence shows a weak positive equity response to the
Bali bombing, and no response from the Mumbai attack in the
Australian market. We also document negative industry abnormal
returns as high as 37.30% on the day in the Utilities sector. Our findings
show that systematic risk of certain sectors increased after the events
of September 11 but remained unchanged for the other attacks.
Crown Copyright © 2009 Published by Elsevier B.V. All rights reserved.

1. Introduction
Chan andWei (1996) postulates that political risks affect the risk and return of capital markets and in this
paper we investigate whether political events like terrorism activities affect foreign capital markets. Cam
(2006) provides a detailed analysis of the impact of the September 11, Bali and Madrid bombings on 135
industry equity indexes in the United States. The empirical evidence shows that September 11 had the most
influence on the US market with airline, hotel and leisure industries recording strong negative abnormal
returns while water, defence and telecommunications industries showing strong positive abnormal returns.
More recently, Nikkinen et al. (2008) showthat the response of international markets to September 11, 2001
terrorist attack differs according to the degree of integration of each region with the international market. In
our paper we adopt and augment the approach used by Cam (2006), and look at the impact of five recent
terrorist attacks on the Australian Equity Market.
Following Cam (2006), we do not assume that investors necessarily react negatively to terrorist attacks.
Equity holders tend to respond negatively to such events onlywhen they perceive an increase in the expected
costs of terrorist activities. We argue that market players may well not react if they do not perceive that the
attack has an impact on expected returns. It is possible that stock markets do not react negatively on days
surrounding a major terrorist attack.We believe that markets can respond differently to the different attacks#p#分頁(yè)標(biāo)題#e#
and that the variability in risk and returns differs significantly across different sectorswithin an economy. Our
industry analysis on howterrorismaffects returns of industrial portfolios contributes to the debate of Yao, Gao
and Alles (2005) who only limited their model to financial and economic factors.
Kim and In (2002) demonstrated that the Australian market is sensitive to international events. As such,
theAustralian Stock Exchange provides an ideal testing groundfor our arguments.Onthe one hand,Australia's
geographic isolationmay project the image of an investment haven. Yet, Australia's strong tieswith theUnited
States and the ‘war on terror’ may attract terrorist activity. Furthermore the Australian Stock Exchange was
among the first markets to open immediately following 9/11. Chen and Siems (2004), Ito and Lee (2005),
Richman et al. (2005) and Worthington and Valadkhani (2005) showed that the Australian market reacted
negatively to the September 11 terrorist attack. Using a long term regression analysis and assessing the
industry effects, Worthington and Valadkhani (2005) argue that only the Financial sector was negatively
affected. Richman et al. (2005) showed a negative long term effect on the overall Australian market. On the
short term analysis of the impact of September 11, Chen and Siems (2004) and Richman et al. (2005) argue
that the entire equity market fell. Our results support these two studies in that we do observe a negative
impact on the Australian market following 9/11.http://www.400190.com/dissertation_writing/Finance/
Our contributions are as follows. Firstly, we identify precisely which industries in Australia were affected.
Secondly, we look at how subsequent attacks impacted on these industries. Thirdly we modified the
methodologies used in the existing literature by excluding firm specific information and using regression
analysis to reinforce our findings.Most of the existing literatures fail to exclude firmspecific information and
thus report resultswhich contain both the impact of terrorist attacks and other non terrorist components. For
instance, Ito and Lee (2005) studied the impact of 9/11 and Bali bombings on both the domestic and
international airline demand. They do not observe any immediate downward spike but instead an ongoing
downshift in the Australian domestic and international airline demand, following September 11. They
explained that their results contain both the impact of 9/11 and the collapse of Ansett Australia. As for Bali
bombings, they document a decrease in the international demand only. Drakos (2004) argue that systematic
risk of major airline1 companies increased post 9/11 but fails to demonstrate the same increment on the
Australian airline company (Qantas). Our conclusions support Drakos (2004) as themajority of the Australian
industries studied did not result in an increase in their systematic risk. However,we identified certain sectors#p#分頁(yè)標(biāo)題#e#
with an increase in their systematic risk.Wethus argue that onemust be careful in generalising the findings of
Drakos (2004) as there are variations in systematic risk changes across industries. To the best of our
knowledge there is no current study that looks at the short termimpact of 9/11 on other Australian industries.
Hence the first objective of this paper is to bridge the gap between the current short term literatures on the
effects of September 11 on the Australian sectors. Furthermore, investors can use this as a guide to make their
investment decision in Australia in the event of another terrorist attack. Such analysis will be beneficial to
portfolio managers that use the top–down investment process. The second stage of this process is to deal with
the factors influencing the industry and we contribute to this debate by adding the terrorist impact on the
different industries.We also observe that the Water sector is very sensitive to international terrorist attacks
and this may have some serious implications for Australian security.
Furthermore,most of the above literaturemay lead one to believe that terrorist attacks result in an increase
in terrorist risk, and therefore reflect a negative sentiment. We argue that such conclusions should not be
drawn until one considers the industry effects of terrorist attacks post September 11. To support our
hypothesis,we study the impact of the subsequent four terrorist attacks that occurred in Bali,Madrid, London
and Mumbai on the Australian Stock Exchange. By observing the industry effects in Australia, we can
determine how Australian investors' reacted to the recent major terrorist attacks. This study is unique in the
sense that it is the first study that looks at the short term effects of the five recent attacks on the different
Australian industries. Most of the current literature attempt to study the impact of one attack on the world
capital markets where we study how the major international terrorist attacks had an impact on one single
country. Our results are consistentwith the prior literature, in that September 11 did, indeed, have a negative
impact on the AustralianMarket. Furthermore, we find that, the market as a whole is fairly insensitive to the
major terrorist attacks post September 2001. Our contribution to this debate is that while we show that the
major terrorist attacks following September 11 did not affect the Australian equity market as a whole, certain
industries were more severely affected. In Section 2, we present the data and methods used in this paper.
Section 3 presents the empirical findings and Section 4 provides some concluding remarks.
2. Data and methods
2.1. Data
We use daily stock returns indexes, returns calculated from the All Ordinaries share price index, and the
10-year bond rate for the period, August 1999 to August 2006, obtained from Datastream. We have a total#p#分頁(yè)標(biāo)題#e#
of 1191 stocks in our sample. We construct industry portfolios based on the Global Industry Classification
Standards (GICS). One of the practical issues that we face in this process is the small number of firms within
some industry sectors. To overcome this issue, we study 13 out of the 14 industries described by GICS. The
number of firms in each of these industry sectors is shown in Table 1. Table 1 reports the descriptive
statistics for each of the different industries. The average daily return for the Computer, Health, Defence,
Water, and Retail sectors are negative. The banking sector shows a positive return while the remaining
sectors exhibit close to zero returns for the period. Further, we fail to reject the null hypothesis that the
returns for the Capital Goods, Insurance, Defence, Water, Banks and Utilities industries are normally
distributed. Table 1 also includes the standard deviation, skewness, excess kurtosis, range of returns, and
number of shares for each of the industry classifications. Details of the five terrorist attacks that occurred in
the United States, Bali, Madrid, London and Mumbai, are summarised in Table 2.
2.2. Methodology
We define daily return as:
DRi = ln
SRIit
SRIit−1
 ð1Þ
Table 1
Descriptive statistics of daily returns, for sectors in Australia from August 1999 to August 2006.
Mean Stdev Skewness Excess Kurt Range Count T-test statistic? JB-statistic
Return
Materials 0.053% 0.011 16.158 272.702 0.185 300 0.879 942631
Diversified Financials 0.007% 0.001 −0.488 15.040 0.015 106 0.534 1003
Energy 0.016% 0.003 2.859 22.627 0.032 156 0.619 3540
Real Estate 0.003% 0.002 −3.574 15.349 0.001 108 0.190 1290
Capital Goods 0.019% 0.001 0.247 0.819 0.006 55 1.119 2.097
Computers −0.101% 0.002 −0.894 7.790 0.015 112 −5.927 298
Pharmaceuticals −0.076% 0.004 2.517 13.123 0.026 50 −1.436 412
Health −0.156% 0.004 −3.860 19.241 0.023 39 −2.614 698
Insurance 0.058% 0.001 −0.455 −1.615 0.002 5 1.495 .716
Defence −0.117% 0.001 −1.016 1.007 0.005 15 −3.580 3.217
Water −0.384% 0.004 −0.667 −2.475 0.008 5 −2.306 1.647
Retail −0.048% 0.002 −1.484 18.649 0.030 221 −2.963 3284
Banks 0.074% 0.001 −0.785 2.140 0.003 8 2.539 2.348
Utilities 0.015% 0.003 0.606 3.643 0.014 11 0.152 6.754
All −0.012% 0.006 25.568 796.013 0.199 1191 −0.700 31574016

where
DRi is the daily return for stock i,
SRIit is the stock return index for stock i at time t.
SRIit−1 is the stock return index for stock i at time t−1.
Ex-post abnormal returns are estimated following Cam (2006) and Brown and Warner (1985). Ex-post
abnormal returns for each firm (ARit) are calculated as the difference between observed returns of firm i at#p#分頁(yè)標(biāo)題#e#
event day t and the expected return, E(Rit).
ARit = Rit − EðRitÞ ð2Þ
The daily expected return, E(Rit), is estimated using an excess return CAPM over the last 260 observed
daily returns:
EðRitÞ= β0 + β1
~r
mt −~r  ft ð3Þ
The abnormal return for industry I, ARIt at time t is then obtained by averaging the abnormal return of
each firm within the industry.
ARIt =
1 澳洲留學(xué)生dissertation
NX N
i=1
ARit ð4Þ
We exclude all firms with firm specific information 15 days on either side of the event day from the
industry portfolio, where firm specific information is defined as any announcement made on the Australian
Stock Exchange. This enables the model to capture only the impact of the terrorist attacks and to be free
from other firm specific information.
2.3. Parametric tests
The parametric tests used in this study rely on the important assumption that the industry abnormal
returns and cumulative abnormal returns are normally distributed. The standard t-statistic for the abnormal
return is:
tARIt =
ARIt
SDðARItÞ ð5Þ
where ARIt is defined as above and SD(ARIt) is an estimate of the standard deviation of the abnormal returns.
By cumulating the periodic abnormal return for each industry over five days,we obtain thefive day cumulative
abnormal return, CAR5It.
CAR5It =X
5
t=1
ARIt ð6ÞThe t-statistic for the five day cumulative abnormal return (CAR5) is obtained by dividing CAR5It by the
standard deviation of the five day cumulative abnormal return, SD(CAR5It).
tCAR5It =
CAR5It
SDðCAR5ItÞ ð7Þ
2.4. Non-parametric tests
The literature dealingwith abnormal returns showthat they are not normally distributed. In particular, the
distribution of the abnormal returns tends to exhibit fat tails and positive skewness. Under these circumstances,
parametric tests tend to reject the null too often when testing for positive abnormal performance and
too seldomwhen testing for negative abnormal returns. Corrado and Truong (2008) highlighted the need for a
robustness test in event studies in Asia-Pacific financial markets and as a result we turn to an alternative
ranking test developed by Corrado (1989). This non-parametric test is more powerful at detecting the false
null hypothesis of no abnormal returns.
We transform each firm's abnormal returns, ARit into ranks, Ki over the combined period, Ti of 260 days
and are denoted as:
Ki = rankðARitÞ ð8Þ
Following Cam (2006), the period is broken up into the 244 days prior to the event, the event day and
15 days after the event. The ranks in the event period for each firm are then compared with the expected#p#分頁(yè)標(biāo)題#e#
average rank (K?i
) under the null hypothesis of no abnormal returns. The expected average rank (K?i) is givenby
Ki = 0:5 +
Ti
2 ð9Þ
As such, the non-parametric t-statistic, tnp, for the null hypothesis of no abnormal returns for each
industry is therefore given by:
tnp =
1N
P N
i=1
Ki − Ki
SD K   ð10Þ
where SD(K?) is the standard deviation of the average rank and is denoted by:
SD K   = ffi1ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffififfiffiffiffiffiffiffi
TX T
t=1
1
N2XKit−Ki2 vuut
ð11Þ
2.5. Regression analysis
Using the CAPM, we then test if terrorist attacks have had an impact on the systematic risk of Australian
industries on the days of the attack. We include a multiplicative dummy variable in the standard CAPM to
test this possibility:
~rIt −~rft = /I + β1I
hr~mt −r~fti+ β2I
hr~mt −r~ftiTD +~
eit ð12Þ
where
r?It is the industry i's return at time t
r?ft is the risk free return at time t
r?mt is the return on the market at time t
D is a dummy variable that takes the value of 1 on the day of the event, and 0 otherwise. This
variable is meantto capture the effect of terrorist attacks on the systematic risk.
ε?it is the error term
?i is the intercept of the regression equation (E(?i)=0)
βI
1 is the CAPM beta
βI
2 is the coefficient of the dummy variable.
The inclusion of an additive dummy variable in the above Eq. (12) results in a near singular, variance–
covariance matrix. As a result we estimate a separate equation to test if the intercept was affected by the
attacks;
~r
It −~rft = uI + α1I
~r
mt −~r h fti+ α2I
D +~eit ð13Þ
We gathered the returns for each industry 244 days prior to the event, and 15 days after the event.
Standard tests and residual diagnostics revealed no major concerns with the above two econometric
models. We also test if these dummy variables were redundant in the above equations using a Wald test for
restrictions.
Using the terrorist event date as the breakpoint, we apply the Chow test to test if the CAPM has changed
after each terrorist attack. Further, we consider how specifically the terrorist attacks impacted on the long
term systematic risk. The test determines whether the level of risk, particularly captured by structural
changes, is altered after the event day:
~r
it = uI + δ1I
~r
mt −~r h fti+ δ2I
~r
mt −~r h ftiTSD + δ3I
SD +~eit ð14Þ
where SD is a structural dummy variable that takes the value of 0 prior to the event, and 1 after the day of#p#分頁(yè)標(biāo)題#e#
the event. This variable captures the structural changes and influence of terrorist attacks on the systematic
risk over a long term horizon.
3. Empirical findings
This section reports the results of five different terrorist attacks on the Australian Stock Exchange. Using
parametric tests and a non-parametric test we test whether the returns and systematic risk of 13 Australian
industries were affected by these five events. We confirm that there is a strong negative impact on returns
for most of the industries and a general increase in systematic risk of some industries during the US attack.
Interestingly, we do not find similar evidence for the subsequent attacks. Surprisingly, the attack on a
neighbouring country, Indonesia, had a positive impact in some Australian sectors.

Table 4
Cumulative abnormal returns on Australian industry indices following five terrorist attacks.
Industry Sep-11 Bali Madrid London Mumbai
CAR5 T-stat CAR5 T-stat CAR5 T-stat CAR5 T-stat CAR5 T-stat
Banks −5.92% −1.23 −2.96% −1.32 −0.94% −0.66 0.24% 0.68 0.12% 0.17
Pharmaceutical −7.23% −2.76 2.21% 1.12 −2.21% −1.41 0.05% 0.01 −2.85% −1.31
Materials −10.71% −3.84 −0.76% −0.03 −0.91% −0.32 1.97% 1.24 −2.16% −0.68
Water 14.33% 0.89 13.24% 0.94 −9.54% −1.69 −6.12% −1.43 0.46% 0.23
Defence −9.47% −1.84 0.79% 0.10 −2.38% −0.79 −3.21% −1.12 −0.06% −0.01
Insurance 0.18% 0.08 −1.17% −0.06 3.35% 1.34 2.96% 1.02 −1.10% −0.54
Health −3.95% 0.07 −0.93% −0.06 −0.49% −0.10 3.01% 1.43 2.87% 1.22
Capital Goods −5.46% −2.43 0.24% 0.45 −3.91% −1.54 3.48% 2.42 −1.88% −0.99
Real Estate −3.94% −2.72 0.25% 0.05 1.95% 1.44 0.33% 0.09 −0.88% −0.82
Group Retailing −7.99% −3.92 2.13% 1.52 0.43% 0.06 0.59% 0.45 −0.34% −0.09
Utilities −23.76% −1.62 13.10% 1.00 −3.43% −1.13 −23.55% −1.15 0.19% 0.04
Energy −1.32% −0.99 −1.24% −0.68 −2.98% −1.22 5.75% 1.84 0.19% 0.03
Diversified Financials −1.29% −0.79 0.98% 0.17 −2.44% −1.32 −1.13% −0.61 −1.05% −1.13
This table presents five day cumulative abnormal returns and the parametric t-test results for 13 Australian industries after
September 11, Bali, Madrid, London and Mumbai terrorist attacks.

3.1. United States — September 11
Tables 3 and 4 summarise the parametric empirical results for September 11 for the different sectors.
Following Cam (2006), we report the abnormal return on the day, and the five day cumulative abnormal
return as well as their respective t-statistics for the 13 different industries. It should be noted that, unlike#p#分頁(yè)標(biāo)題#e#
the US market that opened 6 days after the attack, the Australian market opened the day after the attack. In
other words, we are assessing the performance of the Australian stock market on the 12th of September of
2001. The results reported in Tables 3 and 4 show a consistent negative effect on equities listed in the
Australian Stock Exchange following the September 11 attack. Fig. 1 supports this hypothesis, except for
the Water and the Insurance sectors; all the other industries illustrate both a negative abnormal return and
a negative five day cumulative abnormal return.
Columns 2 and 3 of Table 3 report the abnormal returns and the parametric t-statistics for the various
sectors. Table 3 shows that the returns in theMaterials sector fell by 4.35% after the September 11 attack, and
the t-statistic shows that this value is statistically different fromzero. With the exception2 of Banks, Insurance,
Energy and Diversified Financials, all the other industries exhibited a negative and significant abnormal return.
In other words nine out of thirteen sectors were affected by the event. The sector that was affected the most
was the Utilities sector, which fell by 37.30%. Such a large percentage fall is not unusual, given Cam (2006),
reports a 35% fall in the returns of Airline and Airport industry and after the September 11 attacks in the US.
While Cam(2006) US industry classification differs fromthe GICS classification thatwe use, some similarities
can be observed in theMaterials and Real Estate industries. These two industries suffered on September 11 in
both the US and Australia though the magnitude of the impact is moderately higher in the United States.
Surprisingly, we observe a decrease in the Defence and Water industries in Australia. Such result is
inconsistent with Cam (2006) who shows a positive return in these two sectors in the US.
Worthington and Valadkhani (2005) also use time series analysis of Australian equity markets and
concluded that the Financial sector was the only sector affected negatively by September 11. While
industry classification is different, our analysis shows no evidence of statistical fall in the Diversified
Financials (see Tables 3 and 4). Thus, our findings are inconsistent with Worthington and Valadkhani
(2005). A direct comparison of their study, however, is not possible given that they use a different model. In
their model specification, they include other catastrophic and natural events, like Sydney hail storm,
Canberra bushfire, Victorian gas supply crisis, HIH collapse, September 11 and Bali bombings. In addition their analysis is of a long term nature while ours is on a short term basis. Chen and Siems (2004) assess the
short term effect of September 11 on the global capital market. Using a major market index, they showed
that the Australian equity market fell by 4.19%. Using an international capital asset pricing model, Richman#p#分頁(yè)標(biāo)題#e#
et al. (2005) document a negative impact of about 4.6% on the All Ordinaries index. Our findings are thus
consistent with Chen and Siems (2004) and Richman et al. (2005) as we show a clear and consistent fall in
various industries in Australia. Whilst our analysis does not specifically look at the airline industry, we are
consistent with Ito and Lee (2005) in terms of negative sentiment surrounding the event. Fig. 1 shows the
ranking of the abnormal returns. From the Fig. 1, we can observe that Health, Energy, Diversified Financials,
and Insurance sectors are the least affected by the September http://www.400190.com/dissertation_writing/Finance/ 11 terrorist attack.
Except for the Water and Insurance sectors, all other sectors exhibit a negative cumulative abnormal
return over the following five days. Note that our approach is consistent with most studies as this
methodology supports the hypothesis of negative sentiment after the September 11 attack. The second
column of Table 4 shows that the Utilities sector was the worst performing sector with −23.76% as CAR
over the next five days (see Fig. 1) though the t-statistic implies that this is not statistically different from
zero. The sectors that recorded statistically significant drop were Pharmaceutical (−7.23%), Materials
(−10.71%), Capital Goods (−5.46%), Real Estate (−3.94%) and Group Retailing (−7.99%). Note that all
these sectors also exhibit a negative abnormal return on the day following the attack. From Fig. 1, we
observe a positive five day CAR for Water but a quick look at the t-statistic in Table 4 (column 3) reveals
that this is not statistically significant. It is noticeable from Fig. 1 that the CAR5 is marginally higher than the
event day AR for most industries, implying that the market continued to plummet over the following five
days. Our findings are consistent with Chen and Siems (2004) who showed that cumulative abnormal
return is around −6.81% six days after the event and −8.60% eleven days after the attack. This result is
inconsistent with the Cam (2006) whofound that the CAR over the following six days is lower than the
abnormal return for US firms.
As a robustness test, we consider the non-parametric3 results in Table 5 in our discussion. The negative
impact of the 9/11 event on Australian industries was also detected by the non-parametric tests. The
results in Table 5 show that except for Water and Group Retailing, all the other industries have a negative
non-parametric t-statistic. For instance, column 2 of Table 5 shows that the non-parametric t-statistic is
−3.09 for the Material industry. This reflects the negative abnormal returns identified earlier in the
parametric tests. Generally speaking the results of the non-parametric tests supports the results observed#p#分頁(yè)標(biāo)題#e#
in the parametric analysis.

Table 5
The impact of five terrorist attacks on Australian industry indices — non-parametric results.
Industry Sep-11 Bali Madrid London Mumbai
Banks −3.02 1.40 −0.94 0.93 0.35
Pharmaceutical −0.22 2.10 −0.90 0.18 0.88
Materials −3.09 1.25 −1.42 −0.44 −0.87
Water 1.83 20.33 −2.22 −5.24 0.23
Defence −2.64 1.34 −0.08 −1.67 0.57
Insurance −0.39 1.00 −1.12 0.11 −0.45
Health −0.54 −0.35 1.13 1.44 0.07
Capital Goods −3.35 −0.27 −1.64 −1.28 −1.29
Real Estate −0.42 1.42 −1.33 −0.95 −1.10
Group Retailing 0.96 0.11 −1.69 −0.65 −1.63
Utilities −2.84 0.63 −0.47 0.33 0.13
Energy −0.28 0.23 −2.49 0.48 −1.78
Diversified Financials −0.22 0.27 −0.99 0.43 −0.86
This table presents the non-parametric t-test results for 13 Australian Industries after September 11, Bali, Madrid, London and Mumbai
terrorist attacks.

Based on the above discussion, we can conclude that Pharmaceutical, Material, Water, Defence, Health,
Capital Goods, Real Estate, Group Retailing and Utilities industries were strongly negatively affected on the
day following the September 11 attack and these findings are consistent with Chan and Wei (1996) who
reported that unfavourable political news are associated with negative returns. It is generally assumed that
following a terrorist attack, returns of equities fall as a result of an increase in systematic risk. Our next
objective will be to test if the industries negatively affected by 9/11 experienced a general increase in their
systematic risk. The multiplicative regression analysis (see Eq. (12)) attempts to test this hypothesis
immediately after the attack. Columns 2 to 4 of Table 6 report the results of the multiplicative dummy
variable model (Eq. (12)). A positive (negative) coefficient of the multiplicative dummy variable (βI
2) reflects
an increase (decrease) in systematic risk. The sign of the coefficient (βI
2) appears to be positive in seven out of
the nine industries discussed above. When the coefficient of the multiplicative dummy variable is statistically
different from zero, it implies a significant statistical change in the systematic risk of the industry. The
t-statistics results from column 4 of Table 6 show that systematic risk statistically increased in only four
sectors namely Capital Goods, Defence, Health and Water sectors out of the nine sectors that recorded a
statistical dropped in their abnormal return. For example the systematic risk of Capital Goods was 0.130932
(see column 3 of Table 6) prior to the attack and increased by 0.780447 (see column 4 of Table 6) after the
attack. The systematic risk increased from 0.130932 to 0.911379 after the attack. TheWald test4 reveals that#p#分頁(yè)標(biāo)題#e#
for this industry (and for Defence, Health and Water sectors), that the dummy variable is not a redundant
variable. The general increase in systematic risk after political news was also documented by Chan and Wei
(1996). On the other hand, there is no statistical evidence of an increase in systematic risk in the remaining
five industries. Another key finding of this study is that terrorist attacks do not always lead to an increase in
systematic risk and that terrorist risk varies significantly across industries. These results are consistent with
Drakos (2004) who finds no evidence of an increase in the systematic risk of a major Australian airline
company (Qantas). The general observation of Drakos (2004) on the other hand is that systematic risk
generally increased for all the major international airline companies except for Qantas and KLM.
On the other hand the additivedummy variable Eq. (13) shows the impact of September 11 on the intercept
of the CAPM. Once more we focus the industries stated in the previous paragraph. Columns 5 to 7 of Table 6
present the findings of the regression. As from Column 7, we can observe that the intercept was statistically
decreased only in case of theMaterials and Group Retailing sector and did not change for the remaining ones.
In estimating Eqs. (12) and (13), we only show the short term impact of the September 11 attacks on
the Australian industrial sectors. We applied the Chow breakpoint test to the standard CAPM to determine
whether the model has changed after the terrorist attack. The results are consistent with a change in the
model for most of the sectors but fail to explain whether the intercept or the slope (long term systematic
risk) of the model has changed. By applying Eq. (14), we can establish the direction of change in the long
term systematic risk. That is, we test if the increase in systematic risk observed on the first trading day after
Table 6
The impact of September 11 attack on Australian industry indices — regression analysis.
r?It−r?ft=?I+βI
1[r?mt−r?ft]+βI
2[r?mt−r?ft]?D+ε?it r?It−r?ft=φI+αI
1[r?mt−r?ft]+αI
2D+ε?it
Industry Intercept Coefficient Coefficient Intercept Coefficient Coefficient
?i βI
1 βI
2 φi αI
1 αI
2
Banks 0.00053 0.02305 −0.35243 0.00063 0.00760 −0.00283
T-statistics 1.07 0.36 −1.44 1.24 0.12 −1.38
Capital Goods −0.00131 0.13093 0.78045 −0.00116 0.18550 0.00024
T-statistics −1.78 1.36 2.13 −1.52 1.97 0.08
Defence −0.00484 0.21257 1.60708 −0.00497 0.30119 0.00752
T-statistics −3.06 1.03 2.05 −3.04 1.50 1.14
Diversified Fin. −0.00117 0.11324 −0.56717 −0.00079 0.10022 −0.00805#p#分頁(yè)標(biāo)題#e#
T-statistics −1.82 1.35 −1.78 −1.20 1.25 −3.05
Energy −0.00179 0.20441 0.75736 −0.00133 0.27409 −0.00471
T-statistics −2.54 2.23 2.17 −1.83 3.07 −1.61
Health −0.00239 0.10368 1.28423 −0.00202 0.19990 −0.00150
T-statistics −2.56 0.85 2.77 −2.08 1.67 −0.38
Insurance −0.00122 0.37069 −0.28128 −0.00084 0.37446 −0.00702
T-statistics −0.91 2.11 −0.42 −0.61 2.20 −1.26
Materials −0.00246 0.12699 −0.04050 −0.00213 0.14233 −0.00538
T-statistics −4.23 1.68 −0.14 −3.60 1.96 −2.25
Pharmaceutical −0.00321 0.17629 0.61371 −0.00272 0.23917 −0.00571
T-statistics −2.86 1.20 1.10 −2.35 1.68 −1.22
Real Estate −0.00074 0.17403 0.06821 −0.00070 0.17987 −0.00030
T-statistics −1.55 2.81 0.29 −1.44 2.99 −0.15
Group Retailing −0.00220 0.07097 −0.06104 −0.00182 0.08744 −0.00615
T-statistics −4.15 1.03 −0.23 −3.40 1.33 −2.84
Utilities −0.00084 0.08612 1.02482 −0.00032 0.17487 −0.00473
T-statistics −0.37 0.29 0.90 −0.14 0.60 −0.50
Water −0.00071 −0.34035 3.78276 −0.00019 −0.08769 0.00467
T-statistics −0.23 −0.84 2.45 −0.06 −0.22 0.36
This table presents the regression analysis results for 13 Australian industries after September 11 terrorist attack. The first
multiplicative dummy variable equation illustrates the impact on systematic risk and the second additive dummy variable equation
the event persists in the long term. The results, presented in Table 7 column 4, show that 30% of the
industries exhibit an increase in systematic risk in the long run. For example, the systematic of the Health
industry increased by 0.51 after the September attacks.
3.2. Bali
Among all the terrorist attacks studied in this paper, the Bali bombing is geographically the closest to
Australian soil. The event occurred on Saturday 12th October 2002 and the first day that the Australian
market traded after the attack, was on the Monday 14th October 2002. The results of the parametric test on
sector returns for this day are shown in Table 3 (Columns 4 and 5). Only the Water sector was significantly
affected, and interestingly it was positively affected on the first day that the market traded. The robustness
test also support the claim of a positive effect in the Water sector on the first day of trading. The third
column of Table 5 shows the results on the non-parametric test on the various Australian industries. The
non-parametric t-statistic is positive and significant for the Water industry. Over the 5 day trading period,
there were no significant cumulative abnormal cumulative returns recorded (see Table 4) for Bali bombing.#p#分頁(yè)標(biāo)題#e#
We can therefore conclude that immediately after the Bali attack, only one sector, Water, was positively
affected while all other sectors were insensitive to the event. Just like the five day CAR analysis, the
regression analysis5 shows no significant results. Based on the empirical results, we can further conclude
that Bali bombings did not have a negative effect on the Australian market and on the contrary had a
positive influence. We may interpret this positive result as a substitution effect of terrorist attacks. Our
5 Note that we do not have any regression analysis as they show all the insignificant results.
Table 7
The long run impact of September 11 attacks on the systematic risk of Australian industry indices.
r?It−r?ft=δ0+δI
1[r?mt−r?ft]+δI
2[r?mt−r?ft]?SD+δI
3 (SD)
Industry δ0 δ1 δ2 δ3
Banks 0.00 0.02 −0.27 0.00
T-statistics 1.26 0.38 −1.04 −0.95
Capital Goods 0.00 0.13 0.48 0.00
T-statistics −1.57 1.37 2.24 −0.68
Defence −0.01 0.21 0.72 0.00
T-statistics −3.09 1.02 1.77 0.48
Diversified Fin. 0.00 0.12 −0.28 −0.01
T-statistics −1.19 1.41 −0.82 −2.59
Energy 0.00 0.21 0.38 −0.01
T-statistics −1.91 2.30 2.91 −2.51
Health 0.00 0.11 0.51 −0.01
T-statistics −2.17 0.88 3.09 −1.40
Insurance 0.00 0.37 0.00 −0.01
T-statistics −0.61 2.13 0.00 −1.18
Materials 0.00 0.13 0.20 −0.01
T-statistics −3.61 1.73 0.65 −2.34
Pharmaceutical 0.00 0.18 0.25 −0.01
T-statistics −2.39 1.24 1.61 −1.69
Real Estate 0.00 0.17 0.09 0.00
T-statistics −1.45 2.81 0.36 −0.26
Group Retailing 0.00 0.07 0.21 −0.01
T-statistics −3.42 1.10 0.76 −2.93
Utilities 0.00 0.09 0.32 −0.01
T-statistics −0.16 0.30 1.14 −0.85
Water 0.00 −0.34 0.66 −0.01
T-statistics −0.11 −0.83 2.46 −0.49
This table presents the regression analysis results for 13 Australian Industries after September 11 terrorist attack. The first
multiplicative dummy variable equation illustrates the impact on systematic risk and the second additive dummy variable equation
shows the impact on the intercept (see Eq. (14)).
hypothesis is that investors move their investments from countries directly under attack to the
neighbouring country in search of an investment haven. Unfortunately our findings show very weak
evidence of substitution effect as only the Water sector was affected. Worthington and Valadkhani (2005)6
showed that only the Australian Consumer Discretionary sector was negatively affected by the Bali
bombings. Ito and Lee (2005) also document a negative impact after this event. They recorded a 6% fall in#p#分頁(yè)標(biāo)題#e#
the international demand for airline in the Australian market. Our empirical findings about the Bali
bombings appear to conflict with the existing literature in this field.
3.3. Madrid
The bombings in Madrid occurred on Thursday 11th March 2004. We examine the Australian industry
reactions both immediately, and five days following the event. The results of the parametric test
immediately after the attacks and five days after the attacks are shown in columns 6 and 7 of Tables 3 and 4
respectively. Based on these two parametric tests, only the Insurance industry was significantly negatively
affected. This only incurred on the day following the attack, and this negative effect disappeared after five
days, i.e. CAR5 for the Insurance sector turned into a positive number. The non-parametric test also detects a
negative sentiment on the event day for the insurance sector but fails to show some statistical significance.
Of the five terrorist events that we examine, Madrid suffered the second highest injury and fatality rate, and
we conclude that this event had a negative impact in only one industry of the Australian economy while the
majority of the sectors were unaffected. This result can be regarded as another contribution to the literature
as at present there is no study that looks at the impact of Madrid bombings on the Australian market.

3.4. London
On Thursday 7th July 2005, London was subject to terrorist attacks. Due to our close ties with the
western world, one would have thought that it may have had quite an impact on Australian stock market.
Surprisingly enough, the Australian stock market's response to the attack was muted. The trading day
immediately after the attack saw the Water sector produce an abnormal return of .1.44% (see Table 3,
Column 8), and was the only industry to produce a significant result. The non-parametric t-statistic also
supports the negative movement in the Water sector. The Capital Goods sector showed an unusual
cumulative abnormal return of 3.48% over five days. However the majority of the industries were
immunized from the London bombings. Although the London terrorist attacks were a major global event, it
only affected one industry in the Australian equity market on the day of the impact.
Out of the five attacks studied in this analysis, the Water sector is significantly affected by three of these
events. Water sector was negatively affected by September 11 and London and was positively affected in
Bali. Hence Water sector becomes the most sensitive industry around terrorist attacks.
3.5. Mumbai
Although Mumbai's terrorist attacks claimed 207 lives and injured 714 people, the response from
Australia equity market was immaterial. The empirical testing of this event produced no abnormal
performance results. The Mumbai evidence shows that it is wrong to assume that terrorist attacks will#p#分頁(yè)標(biāo)題#e#
impact negatively on stock markets. As such investment havens do exist even under terrorist attacks.
4. Conclusion
Studying the impacts of the recent terrorist attacks on the Australian industries, we are able to identify
various market effects. September 11 event had the most impact on the Australian market. The majority of
the industries were down on the day of the event, and just under 40% of the industries were still negatively
affected 5 days after the event. Approximately one third of the industries studied showed an increase in
systematic risk following the 9/11 attacks. Madrid and London bombings, the two European attacks, had
mild negative impacts on the Australian market. Surprisingly the lesson learnt from the Bali attacks was
positive for Australia. With only one sector positively affected, this can be interpreted as weak evidence of
substitution effect. Using the Bali bombing evidence, we argue that terrorist attacks do not always nurture
negative sentiment but can also be good for the neighbouring country. Another interesting finding is that
the Mumbai bombing had no effect on the Australian market. The Mumbai evidence can be used to
demonstrate that some capital markets can be insensitive to some terrorist attacks, and thus investment
havens may exist even just after an attack. Finally the industry that was most sensitive to the terrorist
attacks was the Water sector. Australia has not been drastically affected by terrorist attacks post 9/11 and
we can thus conclude that investment haven may exist after those events.
Acknowledgement
The authors wish to acknowledge the invaluable research assistance of Jason Lont, Rebecca Vassil, and
Christian Kypreos in gathering the data and completing some of the empirical analysis. We would also like
to thank George Tawadros, Robert Faff, Richard Heaney and the anonymous reviewers for their assistance
with the methodology, insights and comments that have greatly enhanced the quality of the paper. This
paper was presented at the 14th Annual Global Finance Conference and we thank all participants for their
comments. Any remaining errors, however, are our own.
References
Brown, S.J., Warner, J.B., 1985. Using daily stock returns: the case of event studies. Journal of Financial Economics 14 (1), 3–31.
Cam, M., 2006. The impact of terrorism on United States industry indexes. School of Economics, Finance and Marketing. InRoyal
Melbourne Institute of Technology, Melbourne.
Carter, D.A., Simkins, B.J., 2004. The market's reaction to unexpected, catastrophic events: the case of airline stock returns and the
September 11th attacks. Quarterly Review of Economics and Finance 44, 539–558.
Chan, Y., Wei, K.C.J., 1996. Political risk and stock price volatility: the case of Hong Kong. Pacific-Basin Finance Journal 4, 259–275.
Chen, A.H., Siems, T.F., 2004. The effects of terrorism on global capital markets. European Journal of Political Economy 20, 349–366.#p#分頁(yè)標(biāo)題#e#
Corrado, C.J., 1989. A nonparametric test for abnormal security price performance in event studies. Journal of Financial Economics 23,
385–395.
Corrado, C.J., Truong, C., 2008. Conducting event studies with Asia-Pacific security market data. Pacific-Basin Finance Journal 16,
493–521.http://www.400190.com/dissertation_writing/Finance/
Drakos, K., 2004. Terrorism-induced structural shifts in financial risk: airline stocks in the aftermath of the September 11th terror
attacks. European Journal of Political Economy 20, 435–446.
Ito, H., Lee, D., 2005. Comparing the impact of the September 11th terrorist attacks on international airline demand. International
Journal of the Economics of Business 12 (2), 225–249.
Kim, S., In, F., 2002. The influence of foreign stock markets and macroeconomic news announcements on Australian financial markets.
Pacific-Basin Finance Journal 10, 571–582.
Nikkinen, J., Omran, M.M., Sahltrom, P., Aijo, J., 2008. Stock returns and volatility following the September 11 attacks: evidence from
53 equity markets. International Review of Financial Analysis 17, 27–46.
Richman, V., Santos, M.R., Barkoulas, J.T., 2005. Short- and long-term effects of the 9/11 event: the international evidence.
International Journal of Theoretical and Applied Finance 8 (7), 947–958.
Worthington, A., Valadkhani, A., 2005. catastrophic shocks and capital markets: a comparative analysis by disaster and sector.
Economics Working Paper. InUniversity of Wollongong, Australia.
Yao, J., Gao, J., Alles, L., 2005. Dynamic investigation into the predictability of Australian industrial stock returns: Using financial and
economic information. Pacific-Basin Finance Journal 13, 225–245.