Political Turbulence in the Middle East and Oil Company Returns

Political turbulence in the Middle East and oil company returns

Jee Young Bae[1], EunnyeongHeo[2]

Abstract: Due to the importance of Middle East oil, the political turmoil in the Middle East is the major disruption factor that influence on crude oil market. This study examines the effects of Middle East conflicts on the returns of national oil companies(NOCs) and international oil companies(IOCs). We apply the market model of event study method to assess the impact of such conflicts on the abnormal returns of 33 oil companies. We categorize 20 conflicts into three types—international, interstate, and intrastate—for the period of 1990–2013. Our results suggest that these conflicts have more positive effects on NOCs than on IOCs. The effects on IOCs and NOCs statistically differ by the conflict types. While international and intrastate conflicts have positive and long-lasting effects on NOCs, these affects are not significant for IOCs. Interstate conflicts present higher cumulative abnormal return after the event for NOCs than IOCs. Furthermore, we analyze the impact of wars in the Middle East on the stock returns of international oil companies in the countries engaged in the conflicts. We examine the stock returns of the top twenty oil companies in six countries during four wars from 1990 to 2011. By applying the market model of the event study methodology, we find that the abnormal returns of the companies in countries that did participate in the wars were higher than those of the companies in countries that did not participate. Overall, our research investigates the effect of major geopolitical events on oil company returns, such as the wars in the Middle East, from event study perspective. Empirical results show the magnitude and the direction of such effects varies according to the types of conflicts and companies. Also, these findings suggest that political factors, such as military engagement, can have a statistically significant impact on oil company returns.

Keyword: crude oil market, international oil company, Middle East conflicts, national oil company

(Part of this manuscript is a condensed form of the studypublishedin the journal of 'Energy Sources, Part B: Economics, Planning, and Policy'(2017), Vol.12, Issue 3).

1. Introduction

The outbreak of conflicts in the Middle East is a major uncertainty factor that influences the energy market. Geopolitical issues—such as Middle East conflicts—constitute a key factor that affects the supply and demand of oil through the rise and fall of oil prices, with speculative trading.Many studies considerthe effects of political conflicts in the Middle East on oil price behavior (Bae et al, 2017; Darbouche and Fattouh, 2011; Das et al., 1990; Fan and Xu, 2011; Rigobon and Sack, 2005; Schneider and Troeger, 2006;Toft, 2011; Zhang et al., 2009). A majority of empirical studies that focus on Middle Eastern conflicts apply the event study method to investigate the existence and magnitude of abnormal returns, including Guidolin and La Ferrara (2010), Fernandez (2008), Caplan (2002), Kollias et al. (2013), and Abadie and Gardeazabal (2003). For example, Guidolin and La Ferrara (2010) examines the impact of 101 conflicts between 1974 and 2004 on oil prices and the stock market and find that outbreaks of conflicts in the Middle East and Asia had a stronger impact on oil prices and the stock market. Caplan (2002) reveals that domestic wars produced shocks that had clear and significantly negative effects on real output, while foreign wars had a positive impact on actual production.

Middle East conflicts are expected to have a direct impact on oil companies, as they are key players in the oil market. However, most of the existing research onsuch conflicts examines how their occurrence influences the stock prices of individual companies, the extent of market integration, or economic indicators such as exchange ratesand the price of raw materials (Abadie and Gardeazabal, 2003;Amihud and Wohl, 2004;Guidolin and La Ferrara, 2007;Rigobon and Sack, 2005; Schneider and Troeger, 2006). Little research has been undertaken as to whether and how oil companies are affected, according to conflict type.Therefore, this study aims to analyze the effects of Middle East conflict outbreaks on oil companies’ returns.

In this study, these companies are classified as national oil companies (NOCs) or international oil companies (IOCs). The impacts of Middle East conflicts on the returns of NOCs and on those of IOCs are expected to differ for the following reasons. First, both types may respond differently to each conflict, as such conflicts are often caused by political disputes. Although NOCs and IOCs may belong to the same industry and the same market, the latter prioritize relationships with investors, commercial competition, crisis management, and profitability. On the other hand, NOCs take a greater interest in regional politics, bureaucratic operations, and access to capital, markets, and technology. For example, according to Hartley and Medlock (2008), given that NOCs are heavily influenced by politics rather than the interests of shareholders, IOCs act differently from NOCs in their pursuit of profits. Second, in circumstances where NOCs have a greater monopoly in a country’s oil sector, they could have the flexibility to promote projects in accordance with the country’s economic and political situation, beyond market mechanisms. Hence, if political conflicts occur in the Middle East, unlike with IOCs, NOCs can be expected to have “buffer policy” measures by which to respond to emergencies, and NOCs outside the Middle East could realize more returns. Furthermore,few studies have investigated the link between political positioning, such as “participation in wars,” and oil company returns.In this study, we investigate howMiddle East conflicts with foreign military intervention have influencedthe returns of IOCs and whether the returns of U.S. oil companieshave statisticallydiffered from those of non-U.S. IOCs.

2. Data and methods

While considering that conflict-research institutions generally categorize regional conflicts as extra-state, interstate, intrastate (internal), or internationalized–intrastate, we divided conflicts into three different types, on the basis of state—namely, internationalized (A), interstate (B), intrastate (C)[3]. First, a Type A conflict is an international-level conflict that involves a third country other than the parties in the dispute. Examples include the Gulf War in 1991, the Afghanistan War in 2001, and the Iraq War in 2003. A Type B conflict involves a conflict between countries in the Middle East region. Examples include the Iraq–Kuwait War in 1990, the Armenia–Azerbaijan War in 1991, and the Israel–Lebanon War in 2006. A Type C conflict is a civil war that is fought between forces belonging to the same state and which gives rise to more than 1,000 deaths.[4] Civil wars occur within the borders of a country and a major conflict of this type can have a destructive effect on that country’s economy and business environment. Representative examples of Middle Eastern civil wars include those in Iraq, Yemen, Syria, and Georgia. In categorizing the 20 selected events of Middle East conflicts from 1990 to 2013, one finds that 5 Type A, 4 Type B, and 11 Type C conflicts have occurred (Table 1).

Table 1. Middle East conflicts and number of oil companies in each event by type

Type / Date / Name / States involved / NOCs / IOCs
A / January 16, 1991 / First Iraq War / Iraq, Kuwait, US, 28 countries / 3 / 12
December 17, 1998 / The U.S./U.K. airstrike / Iraq / 7 / 17
October 7, 2001 / Afghanistan War / Afghanistan, NATO / 9 / 18
March 20, 2003 / Second Iraq War / US, Iraq / 10 / 18
March 20, 2011 / Libyan War / Libya, NATO / 12 / 20
A total / 41 / 85
B / August 2, 1990 / Gulf War / Iraq, Kuwait / 3 / 12
December 29, 1991 / Nagorno–Karabakh War / Armenia, Azerbaijan / 3 / 12
July 13, 2006 / Second Lebanon War / Israel, Lebanon / 10 / 18
August 8, 2008 / South Ossetia War / Georgia, Russia / 11 / 18
B total / 27 / 60
C / September 11, 1991 / Georgian Civil War / Georgia / 3 / 12
December 21, 1991 / Georgian Civil War / Georgia / 3 / 12
February 21, 1994 / Yemen Civil War / Yemen / 4 / 13
March 20, 1995 / Iraqi–Kurdish Civil War / Iraq / 5 / 13
May 18, 1998 / Georgian Civil War / Georgia / 7 / 17
April 13, 2004 / Turkey’s Kurdish conflict / Turkey / 10 / 18
June 15, 2004 / Yemen Civil War / Yemen / 10 / 18
September 2, 2007 / Lebanon conflict / Lebanon / 11 / 18
January 27, 2011 / Yemen Civil War / Yemen / 13 / 20
March 15, 2011 / Syrian Civil War / Syria / 13 / 20
July 3, 2013 / Egyptian conflict / Egypt / 13 / 20
C total / 92 / 181

Source: Data drawn from various sources.

Table 2. Regional distribution of the listed companies among the Top 50 oil companies

Middle East / Europe / America / Asia / Africa / Total
NOCs / 2
(11) a / 5
(6) / 1
(4) / 5
(7) / -
(1) / 13
(29)
IOCs / -
(-) / 10
(10) / 10
(11) / -
(-) / -
(-) / 20
(21)
Total / 2
(11) / 15
(16) / 11
(15) / 5
(7) / -
(1) / 33
(50)

Source: Energy Intelligence (2013).

a Numbers in parentheses indicate the number of companies per region that are among the top 50 companies worldwide. “NOCs” and “IOCs” refer to national and international oil companies, respectively.

The oil company stock and market index data used in this study were obtained using Datastream, provided by Thomason Reuters.We analyzed 33 listed oil production companies from among the top 50 oil production companies worldwide, as per “The Energy Intelligence Top 100: Global NOC & IOC Rankings” (Energy Intelligence, 2013)[5](Table 2). In this study, we include cross-border and intrastate conflicts that exceed a certain size and international conflicts but exclude terrorist attacks The number of companies considered in the analysis is shown in Table 2. For the additional analysis of the effects of Middle East wars on IOCs, the analyzed companies are the top 20 IOCs among the 50oil companies (Table 3).

Table 3. List of Middle Eastwars and their participants (states and their IOCs)

Middle East
wars[6] / Category / States / Listed IOCs in these states (among top fiftyglobal NOCs and IOCs)
Gulf war
in 1991 / Participated / US UK Netherlands Italy Spain Australia Canada / Exxon Mobil, BP, Shell, Chevron,
Eni, Conoco Phillips, Repsol, BG, Apache,
Devon Energy, Anadarko, BHP Billiton, Chesapeake, CNR
Did not participate / France Russia / Total, LukoilSurgutneftegas, TNK-BP, Novatek
Afghanistan war in 2001 / Participated / US UK Netherlands Italy Spain Australia Canada / Exxon Mobil, BP, Shell, Chevron,
Eni, Conoco Phillips,
Repsol, BG, Apache,
Devon Energy, Anadarko,
BHP Billiton, Chesapeake,
CNR
Did not
participate / France Russia / Total, LukoilSurgutneftegas, TNK-BP, Novatek
Iraq war
in 2003 / Participated / US UK Netherlands Italy Spain Australia / Exxon Mobil, BP, Shell, Chevron,
Eni, Conoco Phillips, Repsol, BG, Apache,
Devon Energy, Anadarko, BHP Billiton, Chesapeake
Did not participate / France Russia / Total, LukoilSurgutneftegas, TNK-BP, Novatek
Libyan war
in 2011 / Participated / US UK Netherlands Italy Spain Australia Canada / Exxon Mobil, BP, Shell, Chevron,
Eni, Conoco Phillips, Repsol, BG, Apache,
Devon Energy, Anadarko, BHP Billiton, Chesapeake, CNR
Did not participate / France Russia / Total, LukoilSurgutneftegas, TNK-BP, Novatek

The use of the event study method became common following the work of Fama et al. (1969). In this study, the market model of event study method is used to assess normal returns when events have notoccurred as well as when abnormal return(AR)s arise following an event occurrence. In order to obtain a normal rate of return (i.e., as if the event had not occurred), we choose a 100-day estimation period from 120 to 21 days before the event.We assessthe ARby specifying a 40-day event window from 20 days before to 20 days after the event(MacKinlay, 1997). For the analysis on the returns of IOCs by Middle East wars, we assess the abnormal return by specifying a 10-day event window from 5 days before to 5 days after the event and choose a 20-day estimation period from -25 to -5 days before the event. In addition to this, we separate event windows into several phases fora statistical test of cumulative average ARs (CARs), for example, into periods of 5, 10, 15, and 20 days after the event.

The AR is estimated as follows by using the difference in the ex post return (), which includes the impact of the event and the normal return ().

is the actual returns of firm at day , and is the normal return during the event window. In this study, we used the market model to calculate the normal rate of return. is the stock return of the market at day .

,

where is the stock price of a firm at and is the market index to which each firm belongs. is an error term.

,

where is the average AR and n is the number of firms.In addition,

where CAR is the cumulative average AR, is the starting date of the event window, andis the ending date of the event window. The null hypothesis to be tested is , and it indicates thatthe cumulative average AR of the event window is zero. The t-statistics for assessing statistical significance are calculated as follows.

3. Results and discussion

First, we analyzed CARs to understand the impact that conflicts of TypesA, B, and C have on the returns of IOCs and NOCs, by applying the event study method. Figure1 is a CAR graph of IOCs, which representsthe 20days before and after the occurrence of the three types of Middle East conflicts. The most characteristic is Type Bconflict where the effect on stock prices changed from negative to positive. On the other hand, forType A conflicts, the CAR of an IOC is a value close to zero. By contrast, for Type C conflictsin the Middle East, prior to the event, the CAR tends to be close to zero and after the event, negative ARs tend to occur rather. Figure2 shows NOCs’ CARs in the 20 days before and after the three conflict types. The NOCs showed the highest CARs in the event of Type A conflicts. This result implies thatType A conflicts in the Middle East do not have a direct negative impact on the share prices of NOCs. Overall, since NOCs can control their own oil reserves and production in the event of an oil supply crisis, they seem to optimize their profit by adjusting supply and hence generatepositive CARs compared to IOCs.

Figure 1. IOCs’ CARs, by types of conflict

Figure 2. NOCs’ CARs, by types of conflict

The IOCs' CARs in countries participating in the war showed an upward trend even before the outbreak of war, and positive returns occurred after the outbreak of war (Figure 3). In addition, abnormal returns continued to rise five days after the war. By contrast, IOCs in countries that did not participate in the wars had some negative returns before the outbreak of the war. The abnormal returns of the non-participants also showed an upward trend, but the returns remained be negative even after the war. Nearly four days after the conflicts, the returns approached zero. In addition, we classified IOCs in countries participating in the wars into two groups, U.S. companies and non-U.S. companies, and explored the difference in the magnitude and the direction of abnormal returns (Figure 4). As a result of the graphical analysis, positive CARs with an upward trend were generated for the two groups of IOCs, but the U.S. companies experienced higher CARs than the non-U.S. group.

Figure 3. CARs of IOCs in the states participating/not-participating in Middle East wars

Figure 4. CARs of U.S. IOCs and non-U.S. IOCs

After this graphical inspection of CAR paths, we test the significance of the CARs. Table 4 presents the test results of the null hypothesis, 5, 10, 15 and 20 days after the event announcement. These results presentthe t-statistics and indicate whether the aforementioned CAR tendencies were statistically significant. Significant values for CARs would indicate events of that time had a positive or negative impact on the stock prices. We can thus verify whether the aggregate conflict impact is reflected in the stock index and whether it had a significant impact on the post-event return.Table 5 shows the statistical significance of the CARs of IOCs from five days before to five days after the conflicts. As a result, the average CARs of the war-participating countries are positive, and the result of the t-test is statistically significant at the 1% level. On the other hand, the IOCs in countries that did not participate in the wars had negative CARs during the same period, and this finding is also statistically significant at 1% level.

Table 4. Cumulative abnormal returns of IOC and NOCsand t-test results

IOCs / NOCs
Type / Date a / Mean / t-statistic / p-value / Mean / t-statistic / p-value
A / 5 / -0.008 / -7.670 / 0.001 / 0.004 / 4.516 / 0.006
10 / -0.005 / -4.162 / 0.002 / 0.009 / 3.721 / 0.004
15 / -0.006 / -6.170 / 0.000 / 0.010 / 5.550 / 0.001
20 / -0.006 / -5.587 / 0.000 / 0.010 / 6.660 / 0.001
B / 5 / 0.005 / 2.324 / 0.068 / 0.014 / 4.933 / 0.004
10 / 0.006 / 4.084 / 0.002 / 0.015 / 9.471 / 0.001
15 / 0.010 / 5.449 / 0.000 / 0.017 / 10.879 / 0.001
20 / 0.007 / 3.562 / 0.002 / 0.017 / 13.372 / 0.001
C / 5 / 0.002 / 6.657 / 0.001 / 0.010 / 8.572 / 0.001
10 / 0.001 / 2.436 / 0.035 / 0.013 / 10.083 / 0.001
15 / 0.000 / -0.365 / 0.721 / 0.014 / 12.936 / 0.001
20 / -0.001 / -0.898 / 0.380 / 0.015 / 15.258 / 0.001

aDate indicate number of days following the event.

Table 5. Cumulative abnormal returns of IOCs and t-test results

IOCs in countries that participated in wars / IOCs in countries that did not participate in wars
Event
period / Mean / t-statistics / P-value
Pr(T>t) / Mean / t-statistics / P-value
Pr(T<t)
-5
to
+5 days / 0.020 / 5.728 / 0.001
*** / -0.008 / -4.877 / 0.001
***
-5 to 0
days / 0.012 / 3.602 / 0.015
** / -0.011 / -4.947 / 0.004
***
0 to 5 days / 0.009 / 3.073 / 0.027
** / 0.010 / 5.197 / 0.004
***

4. Conclusion

The analytical results of this study can be summarized as follows. Depending on the type of Middle East conflicts, the returns of IOCs differed from that of NOCs. The outbreak of internationalized conflictsled to negative CARs for IOCs, but these were not substantial. Meanwhile, internationalized conflictsgenerated significantly positive CARs among NOCs.At the interstate conflict level, NOCs and IOCs commonly experienced an upward trend of CARs.However, fluctuations in IOCs’ CARs were large, while those of NOCs tended to move in a positive direction in a relatively stable manner. We also found that the returns of IOCs affiliated with countries that participated in the wars were higher and showed a constant upward trend even before the wars. In addition, considering that the U.S. participated in all four Middle East wars and had a leading role, we classified IOCs into U.S. IOCs and non-U.S. IOCs and found positive and significantly higher abnormal returns for U.S. IOCs.

This study holds several implications for policy makers and oil companies. Its findings would enable a deeper understanding of investors’ perceptions about the emergency response capability of oil companies to Middle East conflicts. Oil companies need to build crisis-response strategies through examining market reactions to conflict outbreaks. It is also essential that one considers differences between IOCs and NOCs in analyzing the reactions of stock market participants to conflicts and understands that political decision to intervene in a war is tightly correlated with the stock returns of a country’s oil companies.