THE ECONOMIC BENEFITS OF WAR
an analysis of Western arms companies during modern times
Author Thijs van Vulpen
s2380811 Date January 20, 2017 Supervisor Prof. dr. Elhorst JEL-code G10, N4 ABSTRACT
While Western citizens enjoy peace, safety and freedom, the Western countries seem to profit from the misery somewhere else in the world. In this paper the impact of individual periods of military conflict on abnormal stock returns of Western arms companies is investigated for the
21st century. The data shows that Western countries are large net arms exporters who transfer
their arms mainly to Asian recipients. Unfortunately, the hypothesis that Western arms companies profit from individual military conflicts could not be empirically verified.
2 ONLY THE DEAD HAVE SEEN THE END OF WAR
3
1 INTRODUCTION
It’s January 11, 2014, millions of people march in a peace demonstration against the three-day lasting terrorist attacks that started with the attack on Charlie Hebdo in Paris a few days earlier. Among the crowd are relatives of the victims, but also more than 40 world leaders who lock their arms to show unity. Although the walk from Place de la République to Place de la Nation is meant to protest against the violence, it feels odd that leaders of Western nations, which are known for their military strength, their large arms industries and their top notch military technology, have joined this march.
Terrorist threats are a great concern for Western countries due to the rise of terrorist
cells such as Al Qaida or Islamic State. However, the world has become less violent in terms of fewer armed conflicts and reduced crime rates, among others, which leads to the media
praising the 21st century as the safest era in history1. The Global Peace Index (GPI) measures
and then ranks all countries in the world on their relative peacefulness by creating an index number of three categories: (i) the safety and security in society, (ii) domestic and
international conflict and (iii) the level of militarization. In 2016, the index reports that most
peaceful countries lie on the Western hemisphere2. The European countries, Australia and
Canada are in the top 40 of most peaceful countries in the world, the United States has a modest state of peacefulness and Russia has a very low state of peacefulness.
As peaceful as circumstances in these countries might be, it also seems that many of
these countries profit from conflicts in the more violent countries. Research by Brauer (2007) shows that the United States, Russia, France, Germany and the United Kingdom are the leading arms suppliers in the world. Not surprisingly, most of the largest arms producing companies are located in those Western countries. While Western citizens are enjoying peace, freedom and safety in their home country, they appear to benefit from the danger, war and misery somewhere else in the world.
The objective of this research is twofold. First, the aim is to reveal the world’s largest
arms suppliers and recipients in the 21st century. A deeper analysis of the arms export, arms
import and the respective recipients of the relevant Western countries - which are selected by using a ranking of the top arms companies of the Stockholm International Peace Research
1 Kenny, C., 2015. 2015: The best year in history for the average human being. Retrieved from:
http://www.theatlantic.com/international/archive/2015/12/good-news-in-2015/421200/
Raphael, T. J., 2014. The world is actually safer than ever and here's the data to prove that. Retrieved from: http://www.pri.org/stories/2014-10-23/world-actually-safer-ever-and-heres-data-prove
2 Institute For Economics and Peace, 2016. Global Peace Index 2016. Retrieved from:
4 Institute (SIPRI) - will be presented. The companies with the largest arms sales will be
discussed as well. This will provide a first glimpse into the dark arms industry.
Second, a panel regression is used to investigate whether abnormal stock price returns
of Western arms companies are related to the individual international military conflicts their domestic country is engaged in. When a Western country engages in a new military conflict, the demand for arms is expected to increase which should result in increased arms sales for these companies and thus to higher stock returns. Firm and country factors are added to the regression in order to filter out factors that also explain the fluctuation in stock price returns.
The analysis concludes that the included Western countries are all large net exporters
controlling more than half - and when adding Russia 75 percent - of the arms market. Their main recipients are located in Asia, but also in Africa and South America. Furthermore, the largest arms producers are located in the United States who have individual arms sales that are about four times the arms export of the United States, suggesting that it is a large and
influential domestic industry. Unfortunately the regression analysis provides no empirical evidence that the stock price returns of Western arms companies respond positive to individual military conflicts.
The structure of the paper is as follows. The second section discusses previous
research. The third section explains the methodology. The fourth section discusses the data and the preliminary results. The fifth section shows the regression results. The final section concludes, discusses some limitations and provides ideas for future research.
2 LTERATURE REVIEW
The arms industry is a mysterious industry. At least that is what can be concluded from the limited data sources that are available about this sector (see e.g. Brauer, 2007; Smith et al., 1985). Smith et al. (1985) state that economists pay little attention to the trade in weapons, but that this is probably the consequence of the poor data availability. Much of the data is
incomplete, many arms transfers are illegal or concealed under different national accounts and even if the transfer is declared, the price as well as the costs of the transaction are often
unknown. The latter issue stems from the fact that the transaction price often does not only include the equipment, but also spare parts, training or access to technology, among others.
Research by Brauer (2007) finds that the share in the arms market of the top ten
former and current non-high income countries3, with the exception of Russia, is 8.8 percent
3 Ranking the countries from highest to lowest value of arms transfers: Ukraine, China, Israel, Belarus,
5 during the period 2000 until 2004 which is almost negligible. Moreover, when removing Spain, Israel and South Korea from this group, because they have reached the status of high-income countries in 2004, the market share of non-high high-income countries drops to 6.6 percent. According to Brauer (2007) explanations could be their poverty levels, their small size or a lack of competitiveness resulting, for example, from lower technology levels. Brauer (2007) thereby shows that Western countries such as the United States, Russia, France, Germany and the United Kingdom, lead the arms market. This finding is supported by Garcia-Alonso and Levine (2007) who find that the largest suppliers are from the Western world - except for China and Russia - and that the greater part of the arms deliveries are sold to the developing world.
Furthermore, Garcia-Alonso and Levine (2007) state that most arms agreements are
being made by the Western world and that developing countries are the most important recipients. Arms agreements are often concealed in treaties such as the Weapon Cooperation
Agreements (WCAs)4, which are investigated by Kinne (2016) who studies its effect on
weapon flows during 1995 till 2010. He finds that weapon flows show an increasing pattern in the world from 2000 until 2010. Apparently, the number of network ties positively influences the arms trade activity and Kinne (2016) concludes that countries that engage more in the global arms trade are more likely to be chosen as WCA partner.
According to Brozska (1987) many arms suppliers trade weapons for economic
reasons. Although there are certain benefits to arms control such as a decrease in the danger of military confrontation escalation, reducing economic development impediment in the Third World and improving the position of the civilian goods industry, these suppliers are unlikely to favour more arms transfer control. Moreover, according to Brozska (1987), the “growing commercial element” seems to weaken the control over conflict escalation, because private companies get more grip on influencing the arms trades, which is especially the case in the United States. Smith et al. (1985) state that weapon producers face large Research and
Development (R&D) costs and would not undertake the arms deals if they were not profitable. According to Smith et al. (1985) the arms market is risky and unattractive from a commercial perspective and therefore the government plays an important role in subsidizing their
domestic arms industry, which leads to their conclusion that government policy rather than market forces drives the expansion of arms supply. This government support has both a strategic and a political motive. The strategic motives are independence of foreign supply,
6 access to the latest technology and producing weapons that suit the domestic military’s
requirements (see e.g. Bauer, 2007; Kinne, 2016; Smith et al., 1985). Additionally, improving the domestic industry improves a country’s export demand and if the export price lies above the marginal costs of producing the weapons, the revenues can be used to contribute to the high R&D costs (see e.g. Garcia-Alonso and Levine, 2007; Smith et al., 1985). The political motives of weapon export are to assist allies and to influence their behaviour (Smith et al., 1985).
It is hoped that the arms industry leads to stimulating the domestic economy by
profiting from arms exports, reducing imports, improving employment and domestically benefiting from technology developments (see e.g. Brauer, 2007). However, according to Smith et al. (1985) the boost from exports can lead to economic momentum that undermines the former mentioned strategical and political motives. Strategically it can lead to weapons designed for the export market instead for domestic requirements and politically the need to maintain sales by enhancing the country’s reliability as an arms supplier can dominate foreign policy or security objectives. Furthermore, Smith et al. (1985) state that the arms industry is less labour intensive than the civilian goods industry and the scarce scientific and
technological resources that are devoted to the development of military equipment might better be utilised for improving productivity or competitiveness - and thus net exports - of the civilian goods industry.
Still, it seems that investors can earn abnormal returns if they exploit conflict by
systematically trading in assets. On average the stock markets’ reaction is positive to conflict and greater for international than for internal conflict (see e.g. Guidolin and La Ferrara, 2010). However, Schneider and Troeger (2006) find that if conflict happens unexpectedly the stock market reacts negatively, but when it concerns an unexpected cooperation then the market reacts often positively. The reaction of the stock market appears to be influenced by whether investors think the conflict to be economically costly and whether the event is better than the original scenario. The latter argument is supported by Rigobon and Sack (2005) who
7
3 METHODOLOGY
The aim of this paper is to investigate the dependency of Western arms companies on
international military conflicts during the 21st century. Stated by Brauer (2007), Western
countries lead the arms market and therefore their arms companies are expected to profit from the military conflicts they are engaged in. A new military conflict creates additional demand for arms or other military equipment which results in higher sales for these arms companies. Investors recognize the increase in firm performance which is reflected by positive stock returns. Therefore the hypothesis states that Western arms companies see their stock price returns improve during individual military conflicts.
In order to test the hypothesis empirically, one has to calculate the abnormal returns of
each firm and compare them to the periods of conflict. If the hypothesis is true, the abnormal returns are higher when their domestic country is in military conflict. Note that abnormal returns can be defined as in equation (1):
ARi,t = Ri,t – E(Ri,t) (1)
where ARi,t is abnormal return, Ri,t is actual return and E(Ri,t) is expected return for firm i at
day t. Daily stock returns can then be calculated using equation (2):
Ri,t = Si,tS – Si,t–1
i,t–1 (2)
where Ri,t is the return and Si,t is the stock price of firm i at day t and Si,t–1 is the stock price of
firm i at day t – 1. The conversion to returns means that the sample period is shortened by one period. Next, the stock returns are converted to annual values by taking the average of the daily observations for each year.
To isolate the abnormal returns that can be attributed to the independent military
8 and country factors. Equation (3) shows the first part of the regression model:
Ri,T = Σβv,iFFFFv,,i,T + Σβw,jCFCFw,,j,T + ei,T (3)
where Ri,T is actual return, FFv,i,T represents a firm factor v of firm i in year T, CFw,j,T
represents a country factor w for country j in year T, βv,FFi and βw,CFj are their respective
coefficients and ei,T are the residuals for firm i in year T. Note that the relevant country factors
for firm i depend on the country j it is located in. For example, airplane producer Boeing is located in the United States and therefore only country factors of the United States will be included in the regression for Boeing.
It is highly probable that there exists a significant time trend in stock prices which
influences the returns. Therefore the model is adjusted for period and cross-section effects. Instead of correcting for these effects by first regressing the returns on period and cross-sectional dummies, the model will correct for these effects by including cross-section and period fixed effects. The utilisation of random effects is excluded, because the number of variables that are included in the model is larger than the number of years, which is an essential condition in EViews for using random effects. Another way to capture a time trend is to use a singular linear time trend variable. The model below adjusts the model in equation (3) for both cross-section and period fixed effects:
Ri,T = αi + λT + Σβv,iFFFFv,i,T + Σβw,jCFCFw,j,T + ei,T (4)
where αi and λT represent cross-section and period fixed effects for firm i in year T,
respectively. Note that ei,T represents the part of the residuals that are not explained by the
model and can also be referred to as abnormal returns. The second part of the model is to regress the residuals of equation (4) on the periods the respective domestic country was in conflict:
ei,T = ΣβzC,jδz,j,TC + ui,T with δ ∈ [0,1] (5)
where ei,T are the abnormal returns of firm i in year T, δz,j,TC is a dummy variable for conflict z
9 is the error term. The conflict dummies are constructed as the number of months per year a country is engaged in military conflict. For example, if the United Kingdom engaged in a military conflict which lasted from June 2007 to the end of that year, the dummy for that
specific conflict will count seven-twelfth5 for all British companies for 2007. The conflict
dummy for conflict z is 0 if conflict z did not occur in a given year T, it is between 0 and 1 if conflict z occurred in year T but did not last the entire year and it is 1 if conflict z endured the entire year T. The conflict dummy can thus take any value between and including 0 and 1 depending on the start and end month of the conflict. Moreover, a conflict dummy is only relevant for firm i if it’s domestic country j engages in that specific military conflict z. No dummy variable is included that counts the total number of military conflicts for a respective country each year, because it will be linearly related to the individual conflict dummies and cause collinearity. Following the hypothesis, the coefficients of the conflict dummies are expected to be positive or, in other words, the abnormal returns of the arms companies are positively influenced by the periods of conflict.
As described above the regression is split into two parts. The first part uses firm and
country factors along with period and cross-section effects to calculate the abnormal returns. The second part uses these abnormal returns to test whether there is a significant relationship between the abnormal returns and the periods of conflict. Since the first part of the model already adjusts for cross-section and period effects, the second part of the model, equation (5), will not include any cross-sectional or time dummies, because those effects are already
filtered out by equation (4). The reason to split the regression into two parts is to avoid collinearity between the period effects adjustment and the conflict dummies. Hence, the conflict dummies are based on time as their values are based on the number of months per year a conflict z lasts for a given year T. Therefore the conflict dummy is likely to be linearly related to the time trend adjustment. A second, but more redundant reason, is that it provides a better overview of the regression progress.
The next section continues by discussing the relevant firm and country factors,
the included military conflicts and the preliminary results.
4 THE TROUBLE WITH WAR IS DATA
Emphasized by Smith et al. (1985) and Brauer (2007), data availability is a major concern about the arms industry due to data shortage, incompleteness, manipulation or corruptness.
10 The methodology described in the text above requires firm factors, country factors and
periods of conflict for Western companies and countries. Data for the firm factors are
obtained through Datastream and data on country factors are retrieved from the World Bank. Furthermore, the Stockholm International Peace Research Institute (SIPRI) provides an annual top 100 of the biggest arms producing companies in the world and produces Trend
Indicator Variables (TIVs6) for the arms import and export values of countries. Note that
SIPRI clearly instructs not to use TIVs directly with economic or financial data since it
represents nonfinancial data. However, many researchers go against this advice, because there is simply no other option (except for constructing the data yourself). In this paper the net export values are expressed as returns to investigate whether stock price movement is related to a trend in the arms import and export. Therefore the SIPRI data are not directly compared with economic or financial data.
Before continuing with the description of the firm factors, country factors, periods of
conflict and preliminary results, the definition of Western countries should be given since there is no universal definition. This definition is important for selecting the relevant arms companies as described in the following subsection. In this paper Western countries are all
countries from continental Europe7 plus the United States, Canada, Australia, New Zealand
and Russia.
4.1 Selecting the companies
Data on arms producing companies can be obtained from SIPRI, which provides a top 100 of companies with the largest arms sales around the world from 2002 to 2014. Note that this shortens the sample period from 2000 to 2015 to a period from 2002 to 2014. The list contains annual data on arms sales, total sales, employment and ranking, among others. The companies that are selected have to fulfil three conditions: (i) the company’s domestic country has to be Western, (ii) the arms sales, total sales and employment data have to be complete from 2002 to 2014 and (iii) daily stock price data has to be available from December 31, 2001 to December 31, 2014. Furthermore, subsidiaries are deleted from the sample if the company group that owns the subsidiary is already in the sample. After trimming the SIPRI list, 31
6 Trend Indicator Value (TIV) is nonfinancial data. SIPRI states that “TIV is based on the known unit production
costs of a core set of weapons and is intended to represent the transfer of military resources rather than the financial value of the transfer”. For more information see section two of the SIPRI arms transfer database methodology: https://www.sipri.org/databases/armstransfers/background
7 Excluded Trans-European countries are Armenia, Azerbaijan, Cyprus, Georgia, Kazakhstan and Turkey. These
11 companies satisfy all conditions from which eighteen are from the United States, six from the United kingdom, three from France, two from Germany, one from Italy and one from
Sweden. Appendix A provides an overview of the selected companies, their domestic country and the stock market they are traded on.
4.2 Firm factors
After selecting the relevant companies, firm-specific factors are required to isolate the
abnormal returns that may be related to the periods of conflict. According to Fama and French (1993) three main factors explain stock returns: (i) the market risk premium, (ii) the book-to-market ratio and (iii) size. Each firm factor is explained in the text below.
(i) Market risk premium. The market risk premium is the market return of the market
the stock is traded on minus the risk-free rate of the domestic country of the company. The companies are linked to stock exchange markets in their domestic country, because these markets also reflect the economic situation in the country. The three month Treasury bill rate is used for each country on an annual basis as a proxy for the risk-free rate. Market returns are constructed by converting annual index values to returns using return equation (2).
(ii) Book-to-market ratio. The book-to-market ratio is the book value of the firm divided by the market value of the firm. According to Fama and French (1992) this is a good factor to explain cross-sectional stock returns. Fama and French (1993) also show that book-to-market equity increases from strong negative values for low book-book-to-market quintiles to strong positive values for high book-to-market quintiles. Figure 1 shows that most of the arms companies have a low book-to-market value which is often below one. The included
companies are large in size and the book-to-market ratio functions as an indicator of risk. Following Fama and French (1993) the companies are expected to have a negative relation with the book-to-market ratio.
(iii) Size. Arms sales, total sales and employment are scaled to cope with the
difference in size between the companies. The scalars are total sales, employment and market equity. First of all, arms sales is scaled by total sales and represents the companies’ arms sales intensity ratio. Large companies may have larger arms sales in absolute numbers, but that does not mean that they fully rely on their arms sales. Secondly, arms sales and total sales are scaled by employment and thus represent the dollars of total (arms) sales per head of
12
Figure 1: book-to-market values
The figure shows the book-to-market value of the companies in individual graphs with the book-to-market ratio on the y-axis and years on the x-axis from 2001 to 2014. The companies are ranked on alphabetic order from left to right, starting in the upper left corner. For all included companies see Appendix A.
0 .0 0 .4 0 .8 1 .2 1 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Airbus Group 0 .0 0 .2 0 .4 0 .6 0 .8 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
Babcock International Group
0 .0 0 .4 0 .8 1 .2 1 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 BAE Systems - 0 .1 0 .0 0 .1 0 .2 0 .3 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Boeing 0 .2 0 .4 0 .6 0 .8 1 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 CACI Internation al 0 .3 0 .4 0 .5 0 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Cobham 0 .2 0 .4 0 .6 0 .8 1 .0 1 .2 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
Computer Sciences Co rporatio n
0 .4 0 .5 0 .6 0 .7 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Cubic Corporation 0 .4 0 .6 0 .8 1 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
Dassault Aviation Groupe
0 .2 0 .3 0 .4 0 .5 0 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
General Dyn amics
0 .2 0 .3 0 .4 0 .5 0 .6 0 .7 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Gen er al Electr ic 0 .0 0 .4 0 .8 1 .2 1 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 GKN plc 0 .2 0 .4 0 .6 0 .8 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Harris 0 .2 0 0 .2 5 0 .3 0 0 .3 5 0 .4 0 0 .4 5 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
Honeywell Intern ational
0 .0 0 .2 0 .4 0 .6 0 .8 1 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
Jacobs Engineering Group
0 .4 0 .6 0 .8 1 .0 1 .2 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 L3 Communications 0 .5 1 .0 1 .5 2 .0 2 .5 3 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
Leon ard o Finmeccanica
0 .0 0 .1 0 .2 0 .3 0 .4 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Lockheed Martin 0 .2 0 .4 0 .6 0 .8 1 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Moog 0 .2 0 .4 0 .6 0 .8 1 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Northrop Grumman 0 .0 0 .2 0 .4 0 .6 0 .8 1 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Precision Castparts 0 .2 0 .4 0 .6 0 .8 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Raytheon 0 .4 0 .8 1 .2 1 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Rh ein metall 0 .0 8 0 .1 2 0 .1 6 0 .2 0 0 .2 4 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Rockwell Co llin s 0 1 2 3 4 5 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Rolls Royce 0 .4 0 .6 0 .8 1 .0 1 .2 1 .4 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Saab - 0 .2 0 .0 0 .2 0 .4 0 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Serco Group 0 .0 0 .2 0 .4 0 .6 0 .8 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Textron 0 .2 0 .4 0 .6 0 .8 1 .0 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 Thales 0 .0 0 .4 0 .8 1 .2 1 .6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4 ThyssenKrupp 0 .2 4 0 .2 8 0 .3 2 0 .3 6 2 0 0 2 2 0 0 4 2 0 0 6 2 0 0 8 2 0 1 0 2 0 1 2 2 0 1 4
Un ited Technologies Corporation
13 two ratios show the total (arms) sales per US dollar market equity while the latter shows the number of employees per dollar market equity.
The formula below summarizes the firm factors discussed in the text above:
FFi,T = {R𝑀,j,T – rF,j,T;MVBVi,T i,T; ASi,T TSi,T; TSi,T Ni,T ; ASi,T Ni,T ; ASi,T MEi,T; TSi,T MEi,T; Ni,T MEi,T} (6)
were RM,j,T is the market return of market M in country j for year T, rF,j,T is the risk-free rate of
country j in year T, BVi,T is book value, MVi,t is market value, ASi,T is arms sales , TSi,T is total
sales, Ni,T is employment and MEi,T is market equity for firm i in year T. Note that the ratios
are in chronological order with the text. The next subsection continues with the relevant country factors.
4.3 Country factors
The remaining companies are located in only six of all the Western countries: the United States, United Kingdom, France, Germany, Italy and Sweden. Four country-specific factors of these countries are added to the model to isolate the returns that are related to macroeconomic conditions. In order to deal with country size, the factors are calculated as growth rates in the same way as stock returns are calculated using equation (2).
(i) GDP growth. Higher GDP growth is an indicator of increased economic activity,
which often results in higher demand. But the countries endured large economic contraction during the financial crisis of 2008, which falls in the second half of the sample period. The crisis has forced governments to run austerity programmes that include cutting defence budgets resulting in lower demand for arms and a drop in share prices. However, the conflicts continued despite the financial crisis thereby creating a floor for the defence budgets to drop to. Therefore the influence of GDP growth is unsure, although it normally is expected to be positive.
14 wants to expand its arms supply it would presumably support its arms industry in the form of subsidies, increased military spending or other policies.
(iii) Net export growth. Net export is calculated as the difference between exports of
goods and services and imports of goods and services. Since the large arms companies are internationally represented they are likely to be related to the export value of their country. Higher exports could thus mean more international arms sales. However, because exports may be too general, net arms export is included as well.
(iv) Net arms export growth. Net arms export is calculated in the same way as net
export. Note that import and export values are SIPRI trend values and that the growth rate prevents them from being directly compared to financial values. An increase in net arms export could indicate that the companies profited from higher international sales that year. Therefore the coefficient for arms export is expected to be positive.
The formula below summarizes the relevant country factors:
CFj,T = {gY,j,T; gMEX,j,T; gNE,j,T; gNAE,j,T} (6)
where gY,j,T is output growth, gMEX,j,T is military expenditure growth, gNE,j,T is net export
growth and gNAE,j,T is net arms export growth for country j in year T. The next subsection
discusses the relevant military conflicts.
4.4 Periods of conflict
There is no clear list that provides an overview of all military conflicts during the 21st century.
Therefore dummy variables for military conflicts are constructed by using data of the
Department of Peace and Conflict Research at the Uppsala University8 and going through an
almost infinite number of newspaper articles or encyclopaedias. Among the trouble is that often countries contribute troops and military equipment to organisations such as the North Atlantic Treaty Organization (NATO), International Security Assistance Force (ISAF) or the United Nations (UN) and then operate under their flag instead of their own. This can lead to unclear and imprecise start and end dates of the military conflicts for the contributing
countries since a particular country may already have withdrawn its troops while the mission of the umbrella organisation is still active. In order to overcome this problem the country is
8 Uppsala University: Department of Peace and Conflict Research. Uppsala conflict data program. Retrieved
15 considered to participate as long as its umbrella organisation participates or as long as the conflict lasts if there is no exact start or end date for the conflict. As explained in the
methodology section the conflict dummy counts the number of months per year of a country’s engagement in a specific conflict.
Probably the most well-known and largest conflicts of the 21st century9 in which these
Western countries are involved are the wars in Afghanistan and Iraq. But many more, smaller and probably less notorious conflicts have occurred. The campaign against Somalian pirates is an international intervention that involves the navies of many Western countries. Furthermore, the military intervention in the Libyan civil war, meant to control the Libyan airspace and force a naval blockade, also involved military actions of multiple Western countries. The aftermath of the Libyan civil war led to violence in Mali and as a consequence its government requested Western military assistance. More recently, a coalition against the Islamic State and the Levant (ISIL) was formed, which rapidly intensified due to a number of terrorist attacks.
The selected military conflicts are well-known conflicts and are selected on the
involvement of at least one of the six countries in the sample. Furthermore, the conflicts show variability in duration and international involvement. All the considered conflicts, their start and end date and the role of France, Germany, Italy, Sweden, the United Kingdom and the United States are being discussed in chronical order in the text below.
(i) War in Afghanistan (2001-2014)10. In October 2001 the United States, supported by
the United Kingdom, started the War in Afghanistan as a reaction to the terrorist attacks on September 11, 2001. The aim of the war was to reduce the influence of the Taliban and to eliminate Al-Qaida. In August 2003 NATO members pledged their alliance to the United States under the ISAF flag. Although there have been military withdrawals of troops and equipment during the later years of the war, NATO formally ended their combat operations in
Afghanistan at the end of December 201411. France, Germany, Sweden and Italy are therefore
assumed to have kept their forces in Afghanistan from August 2003 until December 2014. The United Kingdom handed over their last military bases in October 2014, thereby ending
their military operations in Afghanistan12. The United States announced that they would leave
a small number of troops in the country to the end of 2016, although its combat operations
9 Note that the 21st century refers to the sample period from 2002 to 2014.
10 Council on Foreign Relations, 2016. U.S. war in Afghanistan. Retrieved from:
http://www.cfr.org/afghanistan/us-war-afghanistan/p20018
11 Rasmussen, S. E., December 2014. NATO ends combats operations in Afghanistan. Retrieved from:
https://www.theguardian.com/world/2014/dec/28/nato-ends-afghanistan-combat-operations-after-13-years
12 Shirbon, E., March 2014. Led by the Queen, Britain commemorates end of Afghan war. Retrieved from:
16
ended at the end of 201413.
(ii) Iraq War (2003-2011)14. In March 2003 the United States, supported by allies
from the United States-led coalition such as the United Kingdom, invaded Iraq in an attempt to overthrow the government of Saddam Hussein – who was later captured in December 2003. Although the government was quickly dismantled, the violence that followed in Iraq made the war last to the end of 2011. The international disagreement about the war caused France,
Germany and Sweden to oppose it15, while Italy assisted the coalition from June 2003 until
November 200616,17. In May 2009 the United Kingdom officially ended its combat operations
in Iraq18. The United States’ troops withdrawal was completed by December 2011.
(iii) Campaign against Somalian pirates (2008 - present). In June 2008, the Somalian
Transitional Federal Government requested the UN Security Council assistance in combating
piracy off the coast of Somalia19. Since then several counter-piracy campaigns have been
conducted such as Operation Ocean Shield20, Combined Task Force 15121 and Operation
Atlanta22. The campaigns include ships, aircrafts and other military equipment from the above
mentioned countries and many others. Although it is reported that piracy has declined steadily in the region and Operation Ocean Shield has just been terminated, anti-piracy campaigns remain active. Therefore the countries are considered to contribute to anti-piracy programs from June 2008 to the end of the sample.
(iv) Libyan civil war (2011). Although Germany opposed the Libyan intervention, a
coalition consisting of the United States, the United Kingdom, France, Italy, later followed by Sweden and others, intervened in the Libyan civil war by enforcing a no-fly zone, naval
13 CBS News, December 2014. U.S. formally ends the war in Afghanistan. Retrieved from:
http://www.cbsnews.com/news/america-formally-ends-the-war-in-afghanistan/
14 The Editors of Encyclopædia Britannica, 2011. Iraq war. Retrieved from:
https://www.britannica.com/event/Iraq-War
15 BBC News, March 2003. France and allies rally against war. Retrieved from:
http://news.bbc.co.uk/2/hi/middle_east/2821145.stm
16 BBC News, November 2003. Italy’s Iraq deployment. Retrieved from:
http://news.bbc.co.uk/2/hi/europe/3263505.stm
17 Global Security, 2016. Iraq coalition troops. Retrieved from:
http://www.globalsecurity.org/military/ops/iraq_orbat_coalition.htm
18 Hopkins, N., May 2009. UK's eight-year military presence in Iraq to end on Sunday. Retrieved from:
https://www.theguardian.com/world/2011/may/18/british-militarys-8-years-in-iraq-ends
19 UN Security Council, June 2008. Security council condemns acts of piracy, armed robbery off Somalia’s
coast, authorizes for six months ‘all necessary means’ to repress such acts. Retrieved from: http://www.un.org/press/en/2008/sc9344.doc.htm
20 North Atlantic Treaty Organization, December 2016. Counter-piracy operations. Retrieved from:
http://www.nato.int/cps/en/natohq/topics_48815.htm
21 America’s Navy, August 2009. New counter-piracy tack force established. Retrieved from:
http://www.navy.mil/submit/display.asp?story_id=41687
17
blockade and a NATO-controlled arms embargo23,24 starting March 2011. As of October 2011
this NATO mission has been officially terminated25.
(v) Operation Serval (2012 - 2014). In December 2012 the Malian interim government
made an official request for French military assistance after Islamic groups had conquered the
major cities in Mali26. The operation was mainly a French mission to fight Islamic militants in
the north of Mali by using air force, the navy and the French army that provided military
assistance from January 11, 201327 to July 15, 2014. Germany28, Sweden29, the United
Kingdom30 and the United States31 assisted the French troops in the transport of French
military troops and vehicles, providing logistical and medical services and air refuelling. The United Kingdom provided also an air surveillance plane to assist French Forces and the United States provided flight support in some operations. Only Italy did not join the operation in the former French colony. Operation Serval was later replaced by Operation Barkhane.
(vi) Coalition against ISIL (2014 - present). With the rise of the Islamic State and the
Levant in Iraq and Syria and after multiple terrorist attacks on Western grounds, Western countries decided to form a coalition against ISIL. Most countries engaged in airstrikes or sent
military advisors, trainers or weapons to troops in Iraq or Syria32. As of August 2014 the
United States started with airstrikes and was quickly followed by the United Kingdom and
France in September33,34. Italy provides military trainers and weaponry since September 2014,
23 The Guardian Datablog, 2016. NATO operations in Libya: data journalism breaks down which country does
what. Retrieved from: https://www.theguardian.com/news/datablog/2011/may/22/nato-libya-data-journalism-operations-country
24 Traynor, I., Watt, N., March 2011. Libya no-fly zone leadership squabbles continue within NATO. Retrieved
from: https://www.theguardian.com/world/2011/mar/23/libya-no-fly-zone-leadership-squabbles
25 Charbonneau, L., October 2011. U.N. ends mandate for NATO operations in Libya. Retrieved from:
http://www.reuters.com/article/us-libya-un-idUSTRE79P6EC20111027
26 United Nations Security Council, December 2012. Security Council authorizes deployment of African-led
international support mission in Mali for initial year=-long period. Retrieved from: http://www.un.org/press/en/2012/sc10870.doc.htm
27 BBC, January 2013. France confirms Mali military intervention. Retrieved from:
http://www.bbc.com/news/world-africa-20991719
28 Stern, J., February 2016. German parliament backs new military intervention in Mali. Retrieved from:
https://www.wsws.org/en/articles/2016/02/01/mali-f01.html
29United Nations New York, February 2013. Swedish support for Mali mission. Retrieved from:
http://www.swedenabroad.com/en-GB/Embassies/UN-New-York/Current-affairs/News/Swedish-support-to-Mali-sys/
30 BBC, January 2013. Mali crisis: Situation ‘serious concern’ for UK. Retrieved from:
http://www.bbc.com/news/uk-21018382
31 Kurata, P., January 2013. U.S. supports French and ECOWAS intervention in Mali. Retrieved from:
https://geneva.usmission.gov/2013/01/18/u-s-supports-french-and-ecowas-intervention-in-mali/
32 Fantz, A., Pearson, M., February 2015. Who’s doing what in the coalition battle against ISIS. Retrieved from:
http://edition.cnn.com/2014/10/06/world/meast/isis-coalition-nations/
33 BBC, September 2014. RAF jets sent on Iraqi combat mission against IS. Retrieved from:
http://www.bbc.com/news/uk-29393379
34 BBC, September 2014. Islamic state: France ready to launch Iraq airs strikes. Retrieved from:
18
but refuses to engage in airstrikes at the time35. The same holds for Germany which provides
military aid since December 2014, but does not engage in airstrikes36. Sweden is left out of
this coalition since it only provides humanitarian aid. Unfortunately, this war still continues.
The next subsection describes the preliminary results by analysing the size of the arms
export and imports and providing an overview of the largest arms sellers.
4.5 Preliminary results
The use of returns and growth rates not only adjusts for size, but it also produces stationary variables in most cases. An Augmented Dicky-Fuller (ADF) test confirms that all firm factors are stationary. Furthermore, it concludes that all country-specific growth rates are stationary, except for military expenditure growth. However, the sample covers only a small period which includes the financial crisis of 2008 in the second half. Therefore the choice is made not to adjust military expenditure growth for its nonstationarity.
35 Drennan, J., November 2014. Who has contributed what in the coalition against the Islamic State? Retrieved
from: http://foreignpolicy.com/2014/11/12/who-has-contributed-what-in-the-coalition-against-the-islamic-state/
36 Connolly, K., December 2014. Germany joins anti-ISIS military campaign. Retrieved from:
https://www.theguardian.com/world/2015/dec/04/germany-joins-anti-isis-military-campaign
Figure 2a: GDP growth
The figure shows GDP growth in percentages on the y-axis and the years on x-y-axis for France, Germany, Italy, Sweden, the United kingdom and the United States from 2001 to 2014.
Figure 2b: Military expenditure growth
The figure shows military expenditure growth in percentages on the y-axis and years on the x-axis for France, Germany, Italy, Sweden, the United kingdom and the United States for 2001 to 2014.
-6 -4 -2 0 2 4 6 8 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 France Germany Italy Sweden
United Kingdom United States
G ro w th r a te ( % )
GROSS DOMESTIC PRODUCT GROWTH
Year -15 -10 -5 0 5 10 15 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 France Germany Italy Sweden
United Kingdom United States
G ro w th r a te ( % ) Year
19
Figure 2a and 2b show the impact of the financial crisis. The graph of GDP growth
shows an extreme harmonious drop for all countries during the second half of 2007 from which it starts to recover in 2010. In 2014 most counties have approximately the same GDP growth rates as before the drop in 2008. The graph of military expenditure growth shows that most countries increased their military expenditure independent of the crisis except for
Sweden. Overall the graph shows a slightly decreasing pattern in military expenditure growth. In 2014 only France and Sweden have a positive military expenditure growth, while the other countries seem to have reduced their military budgets.
4.5A The arms trade between countries
The United States, Russia, France, Germany and the United Kingdom lead the arms market. (see e.g. Brauer, 2007; Garcia-Alonso and Levine, 2007). The sample includes four of these countries, namely the United States, the United Kingdom, France and Germany. Table 1 shows the top five largest arms suppliers and recipients based on SIPRI TIVs. The largest suppliers are all Western countries with Russia and the United States combined owning more than 50 percent of the market. Italy and Sweden are the only countries which are not ranked among the top five suppliers. The largest exporters supply about three-fourth of the market. Furthermore, the table also shows that, when calculating the combined value of the exports of the six sample countries as a percentage of the world’s total, 50 percent of the arms markets is in the hands of these countries.
The top five largest importers are all located in Asia. India and China are the largest
recipients, receiving approximately 20 percent of the total arms trades. The Western countries are relatively small arms importers while they are large exporters. Moreover, the sample countries together are responsible for less arms imports than China or India alone. Also, the import market seems to be less concentrated since the largest recipients are responsible for less than one-third of the total value of the arms deals.
It is not a wild guess to conclude that the United States leads the global arms market.
20 respectively. Apparently the arms business is a profitable international industry.
Table 3a and 3b show the main export and import partners of France, Germany, Italy,
Sweden, the United Kingdom and the United States. It is striking that these countries have many more export partners than import partners. Moreover, most import partners lie on the Western hemisphere while most export partners are located in Asia, Africa or South America such as the United Arab Emirates, Pakistan, South Korea, South Africa or Peru. Not
surprisingly, the United States has the most export partners followed by France and Italy. Furthermore, the United States has the most import partners as well, leaving the second position for the United Kingdom and the third position for France. The five largest import partners represent most of the imports to the sample countries. For Italy, the United Kingdom and Germany this import value is over 90 percent. However, the five largest export partners represent no more than approximately 50 percent or less of the export value in most cases. Only the largest export partners of the United Kingdom and France receive more than 50 percent of the total arms export. Note that exports are obviously more diverged due to the extreme difference between total export and total import partners.
Table 1: Largest suppliers and recipients
The table presents the largest suppliers and recipients based on the total exported or imported value from 2002 to 2014. The value in brackets represents the percentage of total world arms exports or imports. ‘Non-top five’ reveals the position of France, Germany, Italy, Sweden, the United Kingdom or the United States if they are not ranked among the largest top five and ‘Percentage sample countries’ shows the sum of the values of France, Germany, Italy, Sweden, the United Kingdom and the United States as a percentage of world total. Values are presented in billions of US dollars ($) based on SIPRI TIVs. UAE is short for the United Arab Emirates.
Suppliers (export) Recipients (import)
Rank Country Value Rank Country Value
1 United States 97.599 (30.5%) 1 India 35.767 (11.2%)
2 Russia 81.357 (25.4%) 2 China 26.541 (8.3%)
3 Germany 24.027 (7.5%) 3 UAE 14.407 (4.5%)
4 France 22.014 (6.9%) 4 South Korea 13.792 (4.3%)
5 United Kingdom 14.491 (4.5%) 5 Saudi Arabia 11.489 (3.6%)
Non-top five suppliers Non-top five recipients
7 Italy 7.775 (2.4%) 8 United States 10.204 (3.2%)
12 Sweden 5.671 (1.8%) 19 United Kingdom 5.852 (1.8%)
26 Italy 3.566 (1.1%)
36 Germany 2.456 (0.8%)
53 Sweden 1.068 (0.3%)
61 France 0.769 (0.2%)
World total arms export 319.930 World total arms import 319.930
Percentage top five of world total 74.9% Percentage top five of world total 31.9%
21
Table 2a: arms exports of the six countries from 2002 to 2014
The countries are ranked from highest total arms export to lowest total arms export. Values are based on SIPRI TIVs and presented in billions of US dollars ($).
Exports 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Total United States 4.958 5.616 6.790 6.827 7.481 7.800 6.799 6.806 8.098 9.104 9.163 7.687 10.470 97.599 Germany 0.902 1.660 1.121 2.071 2.696 3.244 2.378 2.539 2.745 1.349 0.816 0.722 1.785 24.028 France 1.474 1.441 2.324 1.842 1.702 2.402 1.991 1.918 0.898 1.752 1.025 1.511 1.734 22.014 United Kingdom 1.090 0.744 1.206 1.060 0.987 0.974 0.967 1.050 1.151 1.040 0.934 1.645 1.644 14.492 Italy 0.468 0.355 0.251 0.825 0.514 0.688 0.391 0.493 0.516 0.918 0.746 0.867 0.743 7.775 Sweden 0.165 0.518 0.297 0.534 0.390 0.336 0.454 0.411 0.658 0.699 0.478 0.390 0.342 5.672 Average 1.510 1.722 1.998 2.193 2.295 2.574 2.163 2.203 2.344 2.477 2.194 2.137 2.786 28.597
Table 2b: arms imports of the six countries from 2002 to 2014
The countries are ranked from highest total arms import to lowest total arms import. Values are based on SIPRI TIVs and presented in billions of US dollars ($).
Imports 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Total United States 0.499 0.592 0.560 0.520 0.641 0.819 0.951 0.968 1.111 0.995 1.180 0.802 0.566 10.204 United Kingdom 0.719 0.761 0.212 0.027 0.308 0.764 0.508 0.383 0.511 0.368 0.586 0.492 0.214 5.853 Italy 0.243 0.559 0.446 0.162 0.433 0.525 0.220 0.109 0.115 0.298 0.219 0.091 0.145 3.565 Germany 0.071 0.064 0.239 0.204 0.416 0.082 0.292 0.339 0.282 0.084 0.157 0.113 0.114 2.457 Sweden 0.075 0.064 0.047 0.078 0.122 0.041 0.044 0.054 0.051 0.191 0.206 0.052 0.043 1.068 France 0.043 0.057 0.093 0.002 0.067 0.074 0.005 0.076 0.103 0.033 0.091 0.107 0.017 0.768 Average 0.275 0.350 0.266 0.166 0.331 0.384 0.337 0.322 0.362 0.328 0.407 0.276 0.183 3.986
Table 2c: net arms exports of the six countries from 2002 to 2014
The countries are ranked from highest total net arms export to lowest total net arms export. Values are based on SIPRI TIVs and presented in billions of US dollars ($).
Net exports 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Total
22
Table 3a: main export partners
The table shows the main exporting partners of France, Germany, Italy, the United Kingdom and the United States based on the total exported value from 2002 to 2014. ‘Total top 5’ represents the total exported value by the five largest exporters, ‘Percentage of export total’ represents the export value by the five largest exporters relative to the total exported value from 2002 to 2014 and ‘Total trading partners’ shows the total number of trading partners of the respective country. Export values are represented between brackets. Values are based on SIPRI TIVs and presented in billions of US dollars ($).UAE is the United Arab Emirates.
Rank France Germany Italy Sweden United Kingdom United States
1 UAE (4.320) Greece (2.498) Peru (.706) South Africa (0.684) Saudi Arabia (3.202) South Korea (10.333) 2 China (2.621) Turkey (2.068) Pakistan (0.454) Pakistan (0.412) United States (2.412) UAE (7.773) 3 Singapore (1.798) South Africa (1.768) Turkey (0.419) Thailand (0.405) India (1.439) Australia (5.793) 4 Saudi Arabia (1.684) South Korea (1.697) India (0.406) Czech Republic (0.387) Chile (0.638) Israel (5.517) 5 Morocco (1.476) Spain (1.437) UAE (0.384) Finland (0.381) Canada (0.617) Japan (5.270) Total top 5 11.899 9.486 2.369 2.269 8.308 34.686
Percentage of export total 54.05 39.48 30.47 40.00 57.33 35.54
Total trading partners 84 75 76 38 54 107
Table 3b: main import partners
The table shows the main importing partners of France, Germany, Italy, the United Kingdom and the United States based on the total imported value from 2002 to 2014. ‘Total top 5’ represents the total imported value by the five largest importers, ‘Percentage of import total’ represents the imported value of the five largest importers relative to the total imported value from 2002 to 2014 and ‘Total trading partners’ shows the total number of trading partners of the respective country. Import values are represented between brackets. Values are based on SIPRI TIVs and presented in billions of US dollars ($).
Rank France Germany Italy Sweden United Kingdom United States
1 United States (0.306) United States (1.243) United States (1.757) United States (0.581) United States (3.568) United Kingdom (2.412) 2 Austria (0.108) Netherlands (0.615) Germany (1.014) Germany (0.144) Germany (0.649) Canada (1.780) 3 Spain (0.107) France (0.177) United Kingdom (0.445) Canada (0.080) Netherlands (0.514) Germany (1.194) 4 Italy (0.079) Switzerland (0.110) Israel (0.134) Finland (0.076) France (0.360) Switzerland (1.142) 5 Netherlands (0.043) Sweden (0.095) France (0.110) France (0.056) Sweden (0.260) Norway (0.685) Total top 5 0.643 2.240 3.460 0.937 5.351 7.213
Percentage of import total 83.72 91.17 97.05 87.73 91.42 70.69
23 4.5B The biggest arms sellers
Table 4a and 4b provide an overview of the top arms sellers in absolute and relative numbers, respectively. Table 4a clarifies that only six of the 31 companies control the top five of highest arms sales from 2002 to 2014. Boeing is on top from 2002 to 2007 and passes the stick to Lockheed Martin, who takes the lead from 2009 onwards. The United States is the domestic country to both companies. Lockheed Martin had the highest total arms sales in the whole period from 2002 to 2014 and is followed by Boeing, BAE Systems, Northrop Grumman and Raytheon. Those five players made arms deals for a total of 1647.20 billion US dollars which represents more than 52.77 percent of the total arms sales by all companies during 2002 to 2014. It is striking that the total arms sales are much higher than total arms exports or imports of the respective country for 2002 to 2014. For example, Lockheed Martin produced arms sales for a total of 397.9 billion US dollars which is four times the total arms export value of the United States. And that is just one company! Most arms are presumably sold to domestic players such as the government. Most arms deals were made in 2009 followed by 2010, 2011 and 2008. It seems that arms deals have not suffered much from the financial crisis of 2008.
Table 4b shows that being among the largest arms sellers does not necessarily mean the
company also relies the most on arms sales since Boeing is not in the top five at all and
Lockheed Martin is only represented in 2003 and 2014. However, BAE Systems and Raytheon are among the largest arms sellers and the companies with the highest arms sales intensity ratios. Taken over the full period, British BAE Systems - which takes the lead from 2006 onwards - has the highest arms sales intensity ratio followed by CACI International, Raytheon, L-3
Communications and Saab. On average 85.52 percent of their total sales consisted of arms deals from 2002 to 2014 while the average over all 31 companies is 46.09 percent. The arms sales intensity ratios are the highest in the period from 2009 to 2011 and that is almost the same period during which most arms deals were made according to table 4a.
Figure 3a and 3b show the individual graphs of the arms sales and total sales of the arms
24
Table 4a: ranking arms sales
25
Table 4b: ranking the arms sales intensity ratio
The table provides an overview of the top five companies with the highest arms sales intensity ratio (i.e. arms sales per US dollar sales) for each year from 2002 to 2014. ‘Total’ at the last column represents the arms sales intensity ratio over the full period based on the total arms sales versus total sales during the full period, ‘Average’ means the average arms sales intensity ratio of the top five companies that respective year and ‘Average year’ represents the average arms sales intensity ratio of all 31 companies that respective year. The arms sales intensity ratios are represented in brackets. Values are presented in percentages. The company codes can be looked up in Appendix A. In this table: BAE Systems (BAE), CACI International (CAC), General Dynamics (GDY), L-3 Communications (L3C), Lockheed Martin (LMT), Northrop Grumman (NGR), Raytheon (RTN) and Saab (SAB).
26
Figure 3a: individual companies’ arms sales
The figure shows the arms sales of the companies in individual graphs with sales in billions of US dollars on the y-axis and years on the x-axis from 2002 to 2014. The companies are ranked on alphabetic order from left to right, starting in the upper left corner. For all included companies see Appendix A.
4 8 12 16 20 2002 2004 2006 2008 2010 2012 2014 Airbus Group 0 1 2 3 4 2002 2004 2006 2008 2010 2012 2014 B abcock Interna tiona l Group
10 15 20 25 30 35 2002 2004 2006 2008 2010 2012 2014 B AE Sys te ms 22 24 26 28 30 32 2002 2004 2006 2008 2010 2012 2014 B oe ing 0 1 2 3 4 2002 2004 2006 2008 2010 2012 2014 C AC I Inte rna tional
0.5 1.0 1.5 2.0 2.5 2002 2004 2006 2008 2010 2012 2014 C obha m 1 2 3 4 5 2002 2004 2006 2008 2010 2012 2014 C omputer Scienc es C orporation
0.2 0.4 0.6 0.8 1.0 2002 2004 2006 2008 2010 2012 2014 C ubic C orporation 0.8 1.2 1.6 2.0 2.4 2002 2004 2006 2008 2010 2012 2014 Da s s ault Avia tion Groupe
5 10 15 20 25 2002 2004 2006 2008 2010 2012 2014 Ge ne ra l Dyna mics 2 3 4 5 2002 2004 2006 2008 2010 2012 2014 Ge ne ra l Elec tric 0.5 1.0 1.5 2.0 2.5 2002 2004 2006 2008 2010 2012 2014 GKN plc 0 1 2 3 4 2002 2004 2006 2008 2010 2012 2014 Ha rris 3.5 4.0 4.5 5.0 5.5 2002 2004 2006 2008 2010 2012 2014 Hone ywell Interna tional
0.4 0.6 0.8 1.0 1.2 2002 2004 2006 2008 2010 2012 2014 J a cobs Engine ering Group
0 4 8 12 16 2002 2004 2006 2008 2010 2012 2014 L3 C ommunic a tions 0 4 8 12 16 2002 2004 2006 2008 2010 2012 2014 Le onardo Finmec ca nica
15 20 25 30 35 40 2002 2004 2006 2008 2010 2012 2014 Loc khe ed M a rtin
0.2 0.4 0.6 0.8 1.0 1.2 2002 2004 2006 2008 2010 2012 2014 M oog 16 18 20 22 24 26 28 2002 2004 2006 2008 2010 2012 2014 Northrop Grumman 0.4 0.6 0.8 1.0 1.2 2002 2004 2006 2008 2010 2012 2014 Prec is ion C a s tpa rts
8 12 16 20 24 2002 2004 2006 2008 2010 2012 2014 R a ythe on 1.5 2.0 2.5 3.0 3.5 2002 2004 2006 2008 2010 2012 2014 R he inme ta ll 1.0 1.5 2.0 2.5 3.0 2002 2004 2006 2008 2010 2012 2014 R oc kwe ll C ollins 2 3 4 5 6 2002 2004 2006 2008 2010 2012 2014 R olls R oyc e 1.2 1.6 2.0 2.4 2.8 3.2 2002 2004 2006 2008 2010 2012 2014 Sa a b 0.5 1.0 1.5 2.0 2.5 3.0 2002 2004 2006 2008 2010 2012 2014 Se rc o Group 1 2 3 4 5 2002 2004 2006 2008 2010 2012 2014 Te xtron 6 7 8 9 10 11 2002 2004 2006 2008 2010 2012 2014 Tha le s 0.8 1.2 1.6 2.0 2.4 2002 2004 2006 2008 2010 2012 2014 Thys s e nKrupp 4 6 8 10 12 14 2002 2004 2006 2008 2010 2012 2014 United Technologies C orpora tion
27
Figure 3b: individual companies’ total sales
The figure shows the total sales of the companies in individual graphs with sales in billions of US dollars on the y-axis and years on the x-axis from 2002 to 2014. The companies are ranked on alphabetic order from left to right, starting in the upper left corner. For all included companies see Appendix A.
20 40 60 80 100 2002 2004 2006 2008 2010 2012 2014 A irbus Group 0 2 4 6 8 2002 2004 2006 2008 2010 2012 2014 B abcock Interna tiona l G roup
15 20 25 30 35 2002 2004 2006 2008 2010 2012 2014 B AE Sys te ms 50 60 70 80 90 100 2002 2004 2006 2008 2010 2012 2014 B oe ing 0 1 2 3 4 2002 2004 2006 2008 2010 2012 2014 C AC I Inte rna tional
1.0 1.5 2.0 2.5 3.0 3.5 2002 2004 2006 2008 2010 2012 2014 C obha m 10 12 14 16 18 2002 2004 2006 2008 2010 2012 2014 C ompute r Sciences C orporation
0.4 0.6 0.8 1.0 1.2 1.4 1.6 2002 2004 2006 2008 2010 2012 2014 C ubic C orporation 3 4 5 6 7 2002 2004 2006 2008 2010 2012 2014 D as s a ult A viation Groupe
10 15 20 25 30 35 2002 2004 2006 2008 2010 2012 2014 Ge ne ra l Dyna mics 120 140 160 180 200 2002 2004 2006 2008 2010 2012 2014 G e ne ra l Elec tric 6 8 10 12 14 2002 2004 2006 2008 2010 2012 2014 G KN plc 1 2 3 4 5 6 2002 2004 2006 2008 2010 2012 2014 Ha rris 20 25 30 35 40 45 2002 2004 2006 2008 2010 2012 2014 H one yw ell Interna tional
4 6 8 10 12 14 2002 2004 2006 2008 2010 2012 2014 J ac obs Enginee ring G roup
0 4 8 12 16 2002 2004 2006 2008 2010 2012 2014 L3 C ommunic a tions 5 10 15 20 25 30 2002 2004 2006 2008 2010 2012 2014 Le onardo Finmec ca nica
25 30 35 40 45 50 2002 2004 2006 2008 2010 2012 2014 Loc khe e d M artin
0.5 1.0 1.5 2.0 2.5 3.0 2002 2004 2006 2008 2010 2012 2014 M oog 20 24 28 32 36 2002 2004 2006 2008 2010 2012 2014 N orthrop Grumman 0 2 4 6 8 10 12 2002 2004 2006 2008 2010 2012 2014 Prec is ion C a s tpa rts
16 18 20 22 24 26 2002 2004 2006 2008 2010 2012 2014 R a ythe on 4.0 4.5 5.0 5.5 6.0 6.5 2002 2004 2006 2008 2010 2012 2014 R he inme ta ll 2 3 4 5 2002 2004 2006 2008 2010 2012 2014 R oc kwe ll C ollins 5 10 15 20 25 2002 2004 2006 2008 2010 2012 2014 R olls R oyc e 1.5 2.0 2.5 3.0 3.5 4.0 2002 2004 2006 2008 2010 2012 2014 Sa a b 0 2 4 6 8 10 2002 2004 2006 2008 2010 2012 2014 Se rc o G roup 8 10 12 14 16 2002 2004 2006 2008 2010 2012 2014 Te xtron 10 12 14 16 18 20 2002 2004 2006 2008 2010 2012 2014 Tha le s 30 40 50 60 70 80 2002 2004 2006 2008 2010 2012 2014 Thys s e nK rupp 20 30 40 50 60 70 2002 2004 2006 2008 2010 2012 2014 United Technologies C orpora tion
28 largest drops are Computer Science Corporation, General Electric and Thales and GKN plc suffered an extreme drop in arms sales around 2004.
To conclude, the tables provide evidence for the importance of the arms industry for Western countries. Most Western countries are large net exporters and have many more arms export than import partners. Moreover, most of the arms import partners are Western
countries while the export partners are mainly located in Asia. Even more striking is the fact that total arms sales is much higher than total arms exports or arms imports taken over the full period. Clearly, the arms industry is both domestically and internationally profitable.
The next section continues with the regression analysis and discussing the results.
5 RESULTS
This section discusses the results of the panel regressions. The first subsection describes a pooled panel regression, shows the first results and carefully draws the first conclusions. The second subsection describes the cross-section and period fixed effect adjustment. The third subsection shows a robustness check. The final subsection discusses the implications of the results and explains a number of limitations.
5.1 A first glimpse: pooled panel regression
The dataset does not have any missing values for any of the variables over time which means that the panel data is balanced. The first step is to run a standard pooled balanced panel regression by regressing the stock price returns on the relevant firm and country factors as equation (3) suggests. The residuals of that regression (i.e. the abnormal returns) are then regressed on the conflict dummies. The model is estimated using the usual least squares method. The regression provides a first overview of the significance and the explanatory power of the independent variables.
Table 5 shows the first results of three different regressions. The left column shows
the coefficients, standard errors and p-values of the pooled panel regression and the bottom row presents the R-squared, adjusted R-squared and Watson statistic. The Durbin-Watson Statistic is close to 2 and a correlogram, including a standard twelve lags, shows low Q-statistics accompanied by high probabilities for the residuals indicating the absence of serial correlation. However, an R-squared of 0.3980 and an adjusted R-squared of 0.3794 suggest low explanatory power of the pooled panel regression.
It appears that most variables are significant at the one or five percent level. Scaling
29
Table 5: panel regression with firm and country factors
The table presents three balanced panel regressions. The left column presents a standard pooled panel regression, the middle column presents the same regression with cross-section and period fixed effects and the right column presents a robustness check of the panel regression with cross-section fixed effects, cross-section generalized least squares and White period robust coefficient covariance. The dependent variable, stock return (Rs), is regressed on the independent variables: the market risk premium (RM – rF), book-to-market ratio (BTM), arms sales intensity ratio (AS/TS), arms sales per employee (AS/N), arms sales per US dollar ($) market equity (AS/ME), total sales per employee (TS/N), total sales per US dollar ($) market equity (TS/ME), employment per US dollar ($) market equity (N/ME), gross domestic product growth (gY), military expenditure growth (gMEX), net export growth (gNE) and net arms export growth (gNAE). Standard errors are represented between brackets. Values are represented in percentages. The R-squared, adjusted R-squared, Durbin-Watson statistic and F-statistic are shown in the last row. The reported F-F-statistic is a Wald test of the hypothesis that all coefficients excluding the constant are equal to zero. Significance at the one, five and ten percent level is indicated by ***, ** and * respectively.
Pooled panel regression Fixed effect adjustment GLS and White period
Variable Coefficient P-value Coefficient P-value Coefficient P-value
30 since both arms sales and total sales per dollar market equity are insignificant. However, scaling by total sales and employment does seem to be appropriate since those ratios are all significant. The insignificance of military expenditure growth contradicts the statement of Smith et al. (1985) that government policy drives the arms market. Furthermore, net export growth is also insignificant, but net arms export growth is significant.
Table 6 provides an overview of the variables’ coefficients, standard errors and
p-values for the same regressions after eliminating the insignificant variables. In most cases the coefficients and standard errors change slightly, but that does not affect the significance level of the variables. Only the standard error of the employment per dollar market equity has almost halved. The market risk premium and book-to-market ratio are consistent with the expectations. Referring back to the discussion in the data section, Fama and French (1993) estimate the book-to-market value to be negative for low book-to-market quintiles, which seems applicable to the arms companies as they often have book-to-market ratios below one (see figure 1). The arms sales intensity ratio is negative, suggesting a 0.1 percent drop in stock returns if more arms sales are made per US dollar of total sales. Furthermore, both efficiency ratios for arms sales and total sales seem to have an almost negligible effect on the stock returns since their coefficients are very low. More surprising is the negative sign of total sales per employee, meaning that stock returns decline when sales per employee increases. Also net arms export growth has a very low coefficient, but its sign is positive as expected. Finally, the coefficient of GDP growth suggests a 0.06 percent decline in stock returns. The negative relation with stock returns can be a consequence of the financial crisis of 2008 as discussed in the data section.
The residuals of the regressions in table 6 are used to estimate equation (5) (i.e. regress
the residuals on the conflict dummies). Table 7 presents the coefficients, standard errors and the p-values of the six conflict dummies. All variables are significant at least at a ten percent significance level, except for the Libyan civil war. However, the signs of the conflict
dummies for the Iraq War, the campaign against Somalian pirates and the military coalition against ISIL are negative which contradict the hypothesis that the (abnormal) stock price returns are positively influenced by military conflicts. Only the conflict dummies for the Mali intervention and the Afghanistan War support the hypothesis. According to the results the Afghanistan War positively influenced the stock returns with approximately 0.02 percent and the military assistance in North Mali appears to have increased the stock returns with
