Abstract

In civil wars across the world, certain Islamist groups have competed exceptionally well against their rivals. The conventional wisdom points to either religion or ethnic politics to explain Islamist success. These ideological and identity-based explanations, however, tend to overlook the powerful economic influence that the local business class has over civil war outcomes. Civil war can be modeled as a market for security, wherein protection must be purchased from multiple substate rackets. Using this market model, a close investigation of the Somali case reveals why and under what conditions the interests of the profit-driven business class align with those of ideologically motivated Islamist groups. Security costs are of critical importance to businesses in a civil war, and Islamists are uniquely competitive in lowering these costs. The business-Islamist alliance is therefore driven by rational, economic considerations, which can contribute to the rise of Islamist power.

Introduction

Many intractable civil wars take place in countries with large Muslim populations.1 In these protracted conflicts, Islamists are often just one of many actors fighting in a complex landscape of ethnic, tribal, and political violence. Yet, certain Islamist groups compete exceptionally well in these conflicts. Why do Islamists sometimes gain power out of civil war stalemates? Although much of the existing research points to either ethnic or religious motivations, I argue that there are also hard economic reasons behind the rise of Islamist power.2 In this article, I offer a micro-political economy model of Islamist success in civil war that highlights the role of an important, but often-overlooked, class: the local business community.

The convergence of business and Islamist interests is relevant across a wide range of cases of contemporary civil war. In present-day Syria and Iraq, jihadi groups have built strong ties to smuggling and criminal networks to fund their war efforts. In Tunisia and Algeria, al-Qaida in the Islamic Maghreb has capitalized on illicit tobacco and narcotics trafficking across North Africa. In Pakistan and Afghanistan, smugglers in the border region have played a key role in financing Islamist power and were particularly involved in the early Taliban movement in 1994. In the 1979 Iranian Revolution, the alliance between the bazaar and the clergy proved to be a driving factor behind the rise of the Islamist regime. Yet despite evidence from both contemporary and historical cases, the relationship between business networks and Islamist groups remains undertheorized in the civil war literature.

What causes profit-driven business elites to support ideologically motivated Islamists? To begin to answer this question, I model civil war as a market for security, in which businesses purchase security from substate protection rackets. Within this security market, I posit that Islamists are uniquely competitive. Specifically, because they are able to work across ethnic and tribal divisions, Islamists can offer lower prices for providing security than rival groups that rely on a narrower base. These lower prices give Islamists an advantage in courting business support and monopolizing the domestic security market. To investigate this hypothesis, I use original field research to test the effects of security prices on political preferences in one of the most enduring civil wars in the world: Somalia.

Since the collapse of its government in 1991, Somalia has suffered more than two decades of civil war along clan lines.3 For fifteen years, these fierce tribal divisions were the main fault line of conflict.4 In 2006, however, a new Islamist movement emerged in the capital city of Mogadishu and quickly ousted the heavily entrenched clan warlords.5 Within the short span of six months, these newcomers accomplished what more than a dozen internationally sponsored peace processes could not: they centralized political control over the majority of the Somali countryside. What caused this change? In this article, I follow the money to demonstrate how the rational interests of the Somali business elite contributed to this rise of Islamist power. First, I present the results of an original survey of the business community that tested the relative weights of clan, Islam, and security prices on the preferences for four different systems of governance: clan-based, warlord-dominated, Islamist, and no government at all.6 Second, I provide data from extensive field interviews with Somali Islamists, clan warlords, civil-society actors, and business elites on the rise of Islamist power in Somalia. Both the quantitative and qualitative evidence indicates that security prices played a key role in the rise of Islamist power and that clan and Islamic identity may be overstated.

Although Somalia is an extreme case of enduring civil war, it offers a particularly good lens for investigating this hypothesis. First, given the intensity and duration of its clan-based social fragmentation, Somalia should have been an especially hard case for Islamist success.7 Second, the primacy of identity politics makes Somalia an unlikely case for a rational, economic explanation.8 Third, the case provides many years of unexplored data on different periods of civil war governance, both secular and Islamist.9 Despite its challenging field environment, Somalia offers a fruitful opportunity to conduct detailed within-case comparative analysis.10

The article has two key limitations in scope. First, I am specifically concerned with Islamist groups in civil war–affected states with large Muslim populations, not with political Islam in general. Second, I do not claim to offer a universally generalizable theory of political decisionmaking in civil wars. Whereas this article focuses on the interests of one elite class, more research is needed on other influential actors in civil war, such as traditional leaders and civil society organizations. What this article does provide, however, is insight into an important and understudied phenomenon: the relationship between business elites and Islamists in civil war.

For analytical purposes, I define an Islamist group narrowly, as a substate faction that utilizes Islamic ideas, identity, symbols, and rhetoric in its framing, and that espouses political order based on Islamic laws and institutions.11 I use the term “warlord” to denote a leader of a substate armed group that has control over a territory and constituency within a state.12 Although some warlords do engage in business activities, their primary function is to hold political and military power. In contrast, although some businesses hire private security to guard their consignments, I identify “businesspeople,” including those engaged in criminal activities, as actors primarily concerned with economic, rather than political or military, gain.13

The remainder of the article proceeds in five sections. First, I build a micro-political economy model of civil war competition and explicate the logic and scope of my theory of Islamist success. Second, I present a contextual narrative and the competing explanations for the rise of Islamist power in Somalia. Third, I discuss the field research and methodology. Fourth, I examine my theory against the competing explanations, presenting the quantitative and qualitative evidence. Fifth, I conclude with a discussion of implications of the theory for future research on other cases of Islamist power in civil wars.

A Security Market Model

Why do Islamists sometimes rise to power in civil wars? To address this puzzle, I develop an ideal-type micro-political economic theory. I begin by modeling the security architecture of a civil war as a competitive security market, involving two or more substate armed groups. According to Charles Tilly, the state is most basically a monopolistic protection racket, extracting taxes from its citizens in exchange for providing security.14 Hence, when a state descends into war between substate groups, no party has a clear monopoly on force, and security must be purchased from multiple and competing warlord protection rackets.15 Each of these rackets not only protects but also threatens other rackets and their own clients. This violence drives up costs for each racket, which are then passed on to the consumer as higher security prices. When there are multiple security providers engaged in active combat, protection is expensive and in demand.16

For businesses operating under these conditions, these security prices are factored into all transactions involving the transport of goods and services across privately held turfs. In addition to paying for trucks, drivers, and fuel, businesspeople must calculate the total price for security demanded from each strongman in each protection racket that they cross through, as well as the total price of militia checkpoints along major trading routes. These are the hard and measurable costs of moving from point A to point B in a civil war. In the absence of a functioning government, these compound security costs are constitutive of the business community's overall tax burden. These costs are then added to the price of goods, dramatically affecting supply and demand in target markets. Accordingly, if protection and checkpoint fees are too high, then businesses lose their profit margin, the trade engine breaks down, and the warlord will have undercut his own financial base.17 The warlord's optimal rate of extortion is therefore high enough that he can remain in power, but low enough to give businesses an incentive to compete in local and regional markets.18

These extortion rates, however, are also affected by local identities, which constrain the warlord's ability to be competitive. In identity-based conflicts, warlord protection rackets are often limited to selling security to their own ethnic and tribal constituencies to protect against rival groups. When a warlord primarily relies on a narrower pool of prospective buyers, he must charge a higher rate for providing security. The business elite's ability to switch support to a less expensive racket is also restricted by ethnic or tribal identities.19 Accordingly, when ethnicity or clan determines political affiliation, businesspeople are forced to accept suboptimal and coercive arrangements, in which warlords extort them with greater impunity. Businesses seeking expansion into new markets must find local partners on each warlord's turf that they cross through, or risk being brutally extorted or denied access. Not only does the business community pay higher security prices under these conditions, but these deep social divisions also increase transaction costs, lower trust, and block access to new markets.20

Even though civil wars create new opportunities for smuggling and other criminal activities, this social fragmentation can become very costly. Purchasing protection from multiple warlord protection rackets is an expensive risk-management strategy. Businesses can also hire their own security guards, but the price of salaries, weapons, and ammunition needed to build and maintain an adequately sized private army can greatly exceed the rates charged by warlords for protection.21 Security in a civil war always comes at a price. Under multiple and competing ethnic or tribal factions, these security prices are driven up.

When the business community is held hostage by expensive ethnic or tribal politics, Islamic movements can provide a convenient exit option.22 Islam offers a number of practical solutions to this costly social fragmentation: it has a common identity to nurture “feelings of loyalty and identity towards a large reference group that exceeds ethnic boundaries”;23 it provides reputational benefits and social capital to overcome a low-trust business environment;24 and it provides a ready-made religious legal system, Sharia, that can be used as an informal mechanism to govern and regulate social and economic life.25 Although seemingly strange bedfellows, the secular business community and political Islamists can make a highly effective marriage of convenience in the absence of a functioning government.

To hedge against ethnic and tribal protection rackets, the business community can simultaneously invest in Islamist security providers. Because Islamists are able to sell security across ethnic and tribal divisions, however, they can offer more competitive prices than groups that primarily sell protection to a narrower pool of prospective buyers. By charging lower margins across the board, Islamists have the ability to capture a larger share of the security market than ethnic or tribal groups. As their market share increases, they can lower their rates further to price out their rivals. Therefore, not only do Islamist groups give businesses the option to leave coercive relationships with ethnic or tribal protection rackets, but they also offer more competitive prices to do so. In the short term, Islamists offer businesses better value for their money, thereby attracting more and more buyers to their protection racket. In this way, Islamists have the ability to do what ethnic or tribal protection rackets cannot: monopolize the security market and subsequently reconstruct state power out of an enduring civil war.26

Turning to the Somali case, I present an in-depth empirical investigation of this ideal-type theory. In the following section, I provide a background narrative of the evolution of Somalia's civil war economy to contextualize the rise of Islamist power. I then test the effect of clan, Islam, and security prices on variation in political preferences, using both quantitative data and qualitative evidence. The empirical observations reveal how the economic interests of the business elite played a critical role in the rise of Islamist power in Somalia. These observations have implications for other cases across the Muslim world.

Islam and Business in Civil War Somalia

In 1991 Somalia's government dissolved, and a brutal civil war broke out along clan lines.27 As warring clan factions systematically looted government buildings and pillaged humanitarian aid convoys, the conflict gave birth to a dynamic and resilient war economy.28 A new business elite formed and quickly evolved into the wealthiest smuggling network in the Horn of Africa.29 This new business class became the chief conduit for all foreign trade, aid, and remittances, making it the backbone and lifeblood of the war economy. As the countryside broke up into an array of clan-based fiefdoms, warlords began extracting rents from businesses. While some warlords also received foreign support, virtually all criminal activity in the war economy flowed through the business elite. Protection rents from the local business class quickly became a key source of revenue for all armed groups.

To keep goods moving under these conditions, businesses cut deals with warlords. Business owners negotiated or were coerced into making multiple protection payments across different clan territories, which over time produced an intricate web of reciprocal agreements. While these deals enriched and entrenched the power of clan-based warlords, many businesspeople operated on a thin and unpredictable profit margin, making them sensitive to even marginal increases in security prices.

Clan politics defined the civil war in Somalia for more than a decade. Despite its near-universal Sunni Muslim population, pervasive tribalism and xenophobia actually made the country a relatively inhospitable terrain for Islamist activity.30 In the absence of a functioning state, however, religious institutions did offer practical social services; Islamic charities provided public goods and services such as hospitals and schools. Islam also provided an informal source of legal order in the form of traditional, clan-based Islamic courts. These courts used Sharia, as well as traditional law, to address everything from criminal charges to divorce proceedings, providing a valuable source of governance for local communities.31

The business community had particular need for the Islamic courts. As early as 1994, businesspeople from Mogadishu's dominant Hawiye clan began to invest in these courts, as a practical means of reducing transaction costs and resolving disputes with associates. Although each court exercised authority only over its own sub-clan constituency, these informal legal institutions provided businesses with a mechanism to mitigate uncertainty and facilitate contracts. Business owners continued to pay for security from warlord protection rackets, but increasingly relied on these clan-based courts to provide the rule of law. Buoyed by investment from the business community, the courts were able to create a small degree of order in the otherwise lawless capital.

Over time, the Islamic courts also became a channel for the business community to express its dissatisfaction with warlord rule. Ken Menkhaus characterizes the business community's 1999 investment in clan-based courts as a “coup” against Somali warlords: “Frustrated with having to pay tribute to militias that provided no security in return (and that were usually the source of insecurity and banditry), leading businesspeople in Mogadishu refused to pay taxes to the warlords associated with their clans. Instead, they bought out the militiamen from beneath the warlords and assigned the gunmen to the command of local sharia courts.”32

Although the Islamic courts of the late 1990s were still too divided to defeat the warlords, this was the business community's first real effort to exercise its political muscle.33 Its next serious attempt was through Somalia's official peace process. In 2000 the business community sponsored efforts to create the Transitional National Government (TNG), a clan-based power-sharing regime that became the first internationally recognized government since 1991.34 The prospect of a new Somali government increased consumer confidence and attracted investment and development aid.35 The TNG was supported by a range of traditional elders, Islamic organizations, and elites from the business class, but it excluded Somalia's strongmen. In response, the warlords rejected the TNG, ignored its leaders, and continued to exert rule over their fiefdoms. As the TNG failed, businesspeople found themselves supporting a powerless paper government, while continuing to pay protection fees to warlords.

The TNG's inability to govern forced a new peace process. In 2004 a new Transitional Federal Government (TFG) was formed in Kenya, which gave parliamentary positions to nearly every clan-based strongman with spoiler capability, resulting in a bloated parliament held in exile in Nairobi. Many warlords took up residence in Kenyan hotels to hustle donor funds and secure official titles in the TFG, leaving behind deputy commanders to run their fiefdoms in Somalia.

As these political processes floundered, the Mogadishu business community further subcontracted practical aspects of providing security to the Islamic courts. While continuing to pay fees to warlord protection rackets, these businesspeople also funded the courts to hire young boys with religious backgrounds to police their neighborhoods for freelance criminals and checkpoints. Preoccupied with their power struggles over high office, the warlords allowed the courts to arrest unaffiliated bandits and gangs, so long as they posed no challenge to their official clan militias.

Because these Islamic courts were divided along clan lines, however, their policing activity had limited influence, creating only small pockets of order inside powerful warlord fiefdoms. As a result, freelance criminals quickly learned to attack business shipments in adjacent communities and then run back to their own neighborhoods to claim clan immunity. To resolve these vexing jurisdictional issues, from 2004 to 2006 the head of a Hawiye-Abgal court, Sheikh Hassan Sheikh Sharif, strategized a merger of Mogadishu's twelve sub-clan courts.36 Through this collaborative effort, the courts were gradually consolidated into one organizational body, the Islamic Courts Union (ICU).

As a unified political body, the ICU pooled resources across clan divisions in a surprising and powerful way. Most notably, the merger joined the courts’ neighborhood militias into one force, transforming the ICU into a serious military contender. In 2006 the young, religiously motivated militias from each clan's court were brought under the multi-clan command structure of a militant wing associated with the ICU, named al-Shabaab.37

By January 2006, the united power of the Islamic courts had become a serious challenge to the divided TFG warlords. Following a series of clashes, by June 2006 the ICU had routed all of Mogadishu's warlords from their longstanding fiefdoms. Within six months, it had captured 90 percent of the Somali countryside and Sheikh Sharif became the de facto head of state. For the first time in fifteen years of uninterrupted civil war stalemate, the ICU and al-Shabaab had monopolized political and military control across clan and sub-clan divisions. The puzzle remains: What caused this dramatic rise of Islamist power?

Given the enduring legacy of identity politics in Somalia, most scholarship understandably points to the role of clan identity,38 or Islamic ideology,39 to explain the rise of the ICU. The first set of explanations correctly points out that eleven out of twelve of the Islamic courts in the original movement were from the Hawiye clan, and that the majority of the ICU leadership drew from the powerful Hawiye-Habr Gidir-Ayr sub-clan, whereas members of the Hawiye-Abgal and other clans were underrepresented.40 Given this composition, some analysts treat the ICU phenomenon as epiphenomenal of clan politics. This explanation, however, overlooks the role of key Abgal and non-Hawiye communities, without which the collective action would not have been possible. How tribalism affects preferences requires a more systematic investigation.

A second explanation naturally focuses on the role of religion in charting the rise of the ICU.41 This approach examines how religious ideas, including extremist elements, helped to mobilize the support for the Islamists. Somalia has had a very long history with Islam, however, so it is not easy to test variation of this variable. Despite centuries of religious homogeneity in Somalia, Islamic groups had little influence over the civil war for fifteen years. Why then was the ICU able to succeed?

I argue that while clan politics and religious ideology are important secondary considerations, the rise of Islamist power in Somalia was primarily driven by the rational, economic interests of the clandestine business class.42 By 2006, 70 percent to 75 percent of the Mogadishu business community had voluntarily contributed resources to the Islamic courts, in support of the unification effort.43 They also systematically withheld support for clan-based warlords. As the business community withdrew its funds from clan protection rackets, the warlords were unable to pay their foot soldiers, and the ICU absorbed their defecting militias into the swelling ranks of al-Shabaab. The logic behind this collective action was simple: the ICU lowered the price of security across clan lines. In the following section, I test the relative weights of clan, Islam, and security prices on preference for the Islamic courts, using both quantitative and qualitative data from Somalia.

Field Research and Methodology

This research presents a within-case comparison of competing systems of order in Somalia, using a combined method of statistical analysis and field interviews to uncover causal processes. First, I present a new survey of members of the Mogadishu business community that explores the effects of security costs, Islamic trust, and clan trust on political preferences. Virtually all nationwide trading firms, telecommunications companies, money transfer agencies, and other large industries have corporate offices in Mogadishu, so this sample includes businesses that operate across the country. This survey is a first effort toward exploring how the relationship between identities and interests affects the political preferences of this powerful elite class.44

Second, I present results from extensive interviews conducted from 2007 to 2013 in Somalia, Kenya, and the United Arab Emirates, with business elites, leaders of the ICU movement, clan warlords in the TFG, and civil society actors. I conducted more than forty in-depth interviews with the most powerful business owners in Somalia, from across clan and sub-clan divisions, on their interests, identities, political leanings, and business experiences. This qualitative research is the product of years spent in the field with these subjects. All interview respondents were offered anonymity, and survey answers were recorded on sheets that contained no information that could be used to personally identify the respondent. Considerable effort was made to assure that both the interviews and survey were conducted with balanced clan representation.

Surveys in Somalia are, for obvious reasons, extremely difficult to perform.45 This study took advantage of a unique opportunity presented by a mass exodus of the Somali business community to Kenya from 2007 to 2009. The survey data were collected in 2009 from 1,003 members of the Mogadishu business community who were temporarily lodged in the Eastleigh neighborhood of Nairobi. All survey respondents were Somalis who had active business interests in Mogadishu under the reign of both the TFG warlords and the ICU. I developed the method of sampling and specific language of the survey questions through a month-long pilot phase, which involved meetings with business owners, Somali academics, policy analysts, and peacebuilding practitioners. An extended discussion of the pilot study, sampling method, variables, survey design, and ethics, as well as an alternative estimation, and the questionnaire text are presented in the online appendix.

Both the quantitative and qualitative research aimed at evaluating the relative effects of security costs, clan identity, and Islamic identity on the preferences of the Somali business elite. The dependent variable in the quantitative estimation is “political preference.” The survey asked business owners about their retrospective economic evaluations of the following four options: (1) the 2000–04 Transitional National Government; (2) the 2004–06 Transitional Federal Government; (3) the 2006 Islamic Courts Union; and (4) preference for no government at all. The question was crafted during the pilot study as the most ethical and safe proxy for capturing political preferences. The TNG was used as the residual category in the analysis.

The two primary independent variables in the estimation are “warlord security costs” and “Islamist security costs.” The survey asked respondents two questions: first, to estimate their total security costs under the reign of warlords from 2004 to 2006; and second, to estimate their costs under the Islamic courts in 2006. Both questions were measured on the same ascending five-point ordinal scale, ranging from none to very high.

Three additional independent variables were also included in the estimation to test the role of identity on political preference: “clan trust,” “Islamic trust,” and “support for Islamic law.” These variables are based on respondents’ reactions to different statements about clan and Islam in business. Businesspeople were asked to respond on a five-point ordinal scale ranging from strongly disagree to strongly agree; the responses were then recoded into binary variables for the analysis.

The analysis uses multinomial logit analysis in STATA. Logit, rather than probit, was used so that the analysis could produce clear and interpretable coefficients, specifically risk ratios. The independence of irrelevant alternatives problem was ruled out using the Hausman test, thereby making the utility of mlogit equal to that of mprobit.

The qualitative research involved scores of in-depth interviews in Mogadishu, Hargeisa, Nairobi, and Dubai with the most powerful businesspeople in Somalia, conducted during multiple field trips from 2007 to 2013.46 The later interviews included a review of the preliminary results of the survey. I spoke to leading Mogadishu businesspersons from those sub-clans that are commonly perceived as being pro-ICU, such as the Habr Gidir-Ayr, as well as to those from Hawiye sub-clans that have a reputation for having been less supportive of the ICU, such as the Abgal.47 I also conducted interviews with a number of powerful non-Hawiye CEOs and shareholders from large, nationwide corporations. Interviews with business elites were semi-structured and asked about clan networks in business, Islamic identity in business, the cost of security under different warlord protection rackets, the relationship between security costs and the price of goods in different in-land markets, the accessibility of in-land markets, and business investment in the Islamic courts. These data were cross-checked against each interviewee's sub-clan identity to determine how clan networks may have affected responses.

Interviews with ICU members, clan warlords, and members of civil society organizations were open-ended and aimed at building a narrative around the rise of the ICU; the respondents were, however, also prompted to comment on the roles of clan, Islam, and business in the Islamic courts movement. Many of these interviews were iterated over the course of several years.

Results from Quantitative and Qualitative Field Study

Both the survey and interview data reveal that security costs had a significant impact on support for the ICU. The quantitative results do not provide a clear explanation of the role of Islam in business, but the interview data indicate that the business community adopted Islamic identity for instrumental purposes. Clan identity failed to clearly explain preferences, and both the interview and survey data show evidence of a waning centrality of clan politics in Somali business.

SURVEY RESULTS

The estimation results indicate that the price of security was significant in driving the Somali business community's support for the Islamists over the warlords, and that both clan and Islamic identity may be overstated. Table 1 indicates the frequency distributions for questions on clan trust and Islamic trust. These results show that clan identity prompts a diverse range of responses, with a slightly negative attitude toward clan trust, whereas respondents were far more favorable to Islamic trust. In fact, all survey questions on Islamic trust demonstrated this bias. This pattern reflects the shift away from clan identity and toward a wider Islamic identity within the Somali business community.

Table 1.

Frequency Distributions for Identity and Trust Variables (in percentages)

Survey QuestionIt is more costly and difficult to conduct business with members of other clans than it is to work with my own clan.A person within my clan is more likely to respect my property rights than a person from another clan.A good Muslim is more likely to respect my property rights than a person who doesn't follow their religion.
Strongly Agree 18.44 16.65 80.66 
Agree 19.84 23.33 11.67 
Neutral 7.38 7.98 3.59 
Disagree 28.51 30.91 2.89 
Strongly Disagree 25.82 21.14 1.20 
Survey QuestionIt is more costly and difficult to conduct business with members of other clans than it is to work with my own clan.A person within my clan is more likely to respect my property rights than a person from another clan.A good Muslim is more likely to respect my property rights than a person who doesn't follow their religion.
Strongly Agree 18.44 16.65 80.66 
Agree 19.84 23.33 11.67 
Neutral 7.38 7.98 3.59 
Disagree 28.51 30.91 2.89 
Strongly Disagree 25.82 21.14 1.20 

The frequency distribution of responses for political preference are as follows: 54 percent of all survey participants felt they did best under the TNG; 9.5 percent felt they fared better under the TFG; 20 percent preferred the ICU; and 16.5 percent felt they did best with no government at all. Table 2 shows the results of the multinomial logit model, with risk ratios and the standard error in parentheses, using the TNG as the residual category. Negative relationships are indicated by values lower than 1. The regression results consistently show strong statistical significance for the security costs variables in determining the political preferences of members of the business community.

Table 2.

Political Preference: Multinomial Logit Analysis

Dependent Variable
Risk Ratio Multinomial Logit Coefficient (standard error)
Independent VariablesTransitional Federal GovernmentIslamic Courts UnionPrefer No Government
general social indicators    
Male 1.060 0.899 1.311 
 0.059(0.258) −0.106(0.194) 0.270(0.210) 
Size of Business 1.181 1.370** 0.773* 
 0.166(0.167) 0.315(0.131) −0.257(0.135) 
High School 0.410*** 0.600* 0.475*** 
 −0.889(0.332) −0.509(0.308) −0.743(0.269) 
University 0.799 2.006* 0.677 
 −0.224(0.537) 0.696(0.405) −0.388(0.457) 
Religious Education 0.574* 1.099 0.548* 
 −0.554(0.312) 0.095(0.280) −0.601(0.252) 
identity and ideology    
Clan Trust 0.679 0.808 0.673* 
 −0.386(0.261) −0.212(0.210) −0.395(0.206) 
Islam Trust 0.840 0.509*** 1.571 
 −0.174(0.326) −0.674(0.222) 0.451(0.305) 
Like Islamic Law 0.640* 0.846 0.676** 
 −0.445(0.241) −0.167(0.191) −0.391(0.193) 
taxes    
ICU Security Costs 1.243* 0.493**** 1.716**** 
 0.218(0.122) −0.706(0.109) 0.540(0.111) 
Warlord Security Costs 0.570**** 1.326*** 1.039 
 −0.561(0.116) 0.282(0.101) 0.039(0.106) 
Constant 0.410(0.757) 0.346(0.628) −2.582(0.774)**** 
Dependent Variable
Risk Ratio Multinomial Logit Coefficient (standard error)
Independent VariablesTransitional Federal GovernmentIslamic Courts UnionPrefer No Government
general social indicators    
Male 1.060 0.899 1.311 
 0.059(0.258) −0.106(0.194) 0.270(0.210) 
Size of Business 1.181 1.370** 0.773* 
 0.166(0.167) 0.315(0.131) −0.257(0.135) 
High School 0.410*** 0.600* 0.475*** 
 −0.889(0.332) −0.509(0.308) −0.743(0.269) 
University 0.799 2.006* 0.677 
 −0.224(0.537) 0.696(0.405) −0.388(0.457) 
Religious Education 0.574* 1.099 0.548* 
 −0.554(0.312) 0.095(0.280) −0.601(0.252) 
identity and ideology    
Clan Trust 0.679 0.808 0.673* 
 −0.386(0.261) −0.212(0.210) −0.395(0.206) 
Islam Trust 0.840 0.509*** 1.571 
 −0.174(0.326) −0.674(0.222) 0.451(0.305) 
Like Islamic Law 0.640* 0.846 0.676** 
 −0.445(0.241) −0.167(0.191) −0.391(0.193) 
taxes    
ICU Security Costs 1.243* 0.493**** 1.716**** 
 0.218(0.122) −0.706(0.109) 0.540(0.111) 
Warlord Security Costs 0.570**** 1.326*** 1.039 
 −0.561(0.116) 0.282(0.101) 0.039(0.106) 
Constant 0.410(0.757) 0.346(0.628) −2.582(0.774)**** 
*

Cells contain risk ratios with standard errors indicated in parentheses. Number of observations = 1,003. p < 0.10. Transitional National Government used as residual category.

**

Cells contain risk ratios with standard errors indicated in parentheses. Number of observations = 1,003. p < 0.05. Transitional National Government used as residual category.

***

Cells contain risk ratios with standard errors indicated in parentheses. Number of observations = 1,003. p < 0.01. Transitional National Government used as residual category.

****

Cells contain risk ratios with standard errors indicated in parentheses. Number of observations = 1,003. p < 0.001. Transitional National Government used as residual category.

Re-collection of high warlord security costs was negatively correlated with preference for the TFG and positively correlated with preference for the ICU. Other things equal, a one-level increase on the scale of warlord security costs made an individual 43 percent less likely to prefer the TFG and 33 percent more likely to favor the ICU. The security costs variables also show particular robustness on preference for and against the Islamic courts. A one-level increase in the level of re-collected ICU security costs on the quintile range made one 51 percent less likely to prefer the ICU and 24 percent more likely to support the TFG.

Despite the fact that the overwhelming majority of businesspeople stated that general insecurity was the greatest challenge facing their business,48 and that they would be willing to pay higher taxes in order to have a central government,49 those who felt they were paying higher rates under the ICU were considerably more inclined to prefer political anarchy. Most significantly, a one-level increase in ICU costs made the odds 72 percent higher that a businessperson would prefer having no government at all. This indicates that higher costs of warlord protection rackets resulted in increased support for the Islamic courts, whereas high ICU taxes led businesspeople to feel that they were better off without any government.

In contrast, and in a challenge to ideological interpretations of the rise of political Islam, there is much weaker evidence for the role of ideas and beliefs in explaining business support for the ICU. Surprisingly, neither “Islamic trust” nor “support for Islamic law” produced expected results with respect to preference for the ICU.50 In fact, those who felt most strongly that Islamic identity is a good measure of trustworthiness were also less favorable toward the ICU. Surprisingly, businesspeople who felt most strongly about Islamic trust were actually 49 percent less likely to prefer the ICU. Furthermore, there was also no statistically significant correlation between support for Islamic law and preference for the ICU. Respondents who strongly favored Islamic law as a way to regulate transactions among businesses were 36 percent less likely to support the TFG and 32 percent less likely to prefer no government at all. Strong feelings about Islamic law, however, did not correspond with increased support for the ICU.

Moreover, Islamic education had no obvious impact on support for the ICU. Religious education did make one 43 percent less likely to support the TFG and 45 percent less likely to prefer no government at all, compared to the TNG. There is no indication from the survey results, however, that religious training or strong religious sentiment explains why members of the business community might prefer the ICU. On the contrary, the data show that respondents who had a university-level education were twice as likely to feel positively about the ICU. This suggests that a university education, rather than religious education, is correlated with preference for the Islamic movement.51

Although the Islamic identity variables failed to produce a clear causal explanation for political preference, clan identity offered even fewer insights. The clan trust variable had no effect on preference for either the TFG or the ICU, and clan favoritism was negatively correlated with a preference for political anarchy. In fact, only the residual category, the TNG, attracted the support of those who felt positively about clan as a measure of social trust. Those who felt that clan identity was a good measure of trust were 48 percent more likely to prefer the TNG to the ICU, when “prefer no government” is used as the residual category. The TNG formally divided power on the basis of clan identity, thus drawing support from those who felt most strongly about clan.

Size of business also proved important in understanding preferences; moving one level up on the tertile range for size of business made one 37 percent more likely to prefer the ICU and 23 percent less likely to want no government at all. Interviews with large business owners involved in the ICU movement revealed that the wealthiest businesspeople not only experienced the highest rates of extortion under warlord protection rackets, but they also had the most resources to try to change their situation. This finding suggests that future research needs to explore the possibility of a curvilinear relationship between income and political activism in civil war competition. Moreover, despite much scholarly research on the effect of Islamic regimes on women's rights, the results showed no statistical correlation between gender and political preference. Somali businesswomen responded similarly to their male counterparts, both for and against the ICU.

The most significant estimation results are the strong correlation between security costs and political preference. Although the identity variables demonstrate interesting patterns, they do not produce a clear explanation of preference for Islamists. Rather, lower security prices best explained preference for the ICU. Equally important are the results for those individuals who preferred no government at all. These results highlight the powerful role of tax aversion among these elites, and suggest that as the cost of going it alone becomes less than the cost of paying taxes, the business community has an incentive to act as a spoiler. This finding suggests that policy interventions may be able to control spoiler activity by reducing formal and informal rates of taxation on this powerful class.

INTERVIEW DATA

The interview data provide surprising insights into the roles of clan and Islam in business life over the last decade. Since the mid-1990s, the Somali business community reported deliberately moving away from a clan-based business model, in an effort to expand operations into new markets. Some older businesspeople from the Darod-Marehan clan, who were powerful in the 1980s but lost out when the state fell, expressed bitterness toward the new Hawiye business community.52 All of the largest and more relevant Somali industries, however, including billion-dollar telecommunications and money transfer conglomerates, reported having multi-clan shareholder structures and a pan-clan approach to business development.53 Two businesswomen who own a prominent qaat-trading company sharply stated, “Clan has no place in business. We are all from different clans and we can work together for our interests.”54

In contrast, business owners reported actively using Islamic identity to facilitate trust across clan divisions. The chief legal counsel of the Islamic Courts Union elaborated, “With no government, Islamic adherence was a way to get the trust of the people. If you take the title of sheikh and grow a beard, then you will be more trustworthy than a person smoking a cigarette, with no beard.”55 The dean of a leading Mogadishu business school provided this insight: “Since there is no regulatory or law enforcement agency and everything is based on trust, this is where Islamists come in. Most companies are led by people who look like sheikhs.”56 In fact, one of the biggest firms in the country reportedly tests its prospective employees on their level of Islamic knowledge, as a risk-mitigation strategy. Because the biggest businesses in Somalia have nationwide offices, these elites have learned to utilize Islam to build trust across clan divisions.

Interviews with business elites also provided critical insights into security costs, markets, and politics in the civil war. One of the biggest traders in Mogadishu explained, “As the number of factions increased [in the mid-1990s], the number of checkpoints increased. … This was like taxation. Everywhere you paid more at each checkpoint. For example, at one checkpoint you used to pay 1,000 [Somali] shillings and two years later you would have to pay 100,000 shillings. This was because of both inflation and greediness.”57 By 2000, rates of extortion under warlord protection rackets were increasingly perceived as unfair by the business elite. One executive explained, “The warlords were demanding much from the business community. They sent freelancers to scare the business community into paying up. The warlords just demanded money and threatened violence. A businessperson would have to calculate what you have, what you would lose, and then pay up.”58

Businessmen and women across clan divisions complained that the warlords’ rates were too high and that their militia's checkpoints were frequent and unpredictable. The head of the Suuq Ba'ad market in Mogadishu elaborated on how the number of checkpoints directly affected the prices for goods: “During [the reign of the] warlords, there was an increase in the number of checkpoints, so we added this to the price of the commodity before sale. The difference between a sack of sugar in Mogadishu and Baidoa was approximately $2 more, including the cost of the car rental, fuel, and checkpoints.”59 Smaller businesspeople were extorted even more heavily. “The price of sugar per sack went up $4–$5 because of checkpoints on the Afgooye corridor,” explained a midsized trader. “I practiced. I took [sugar] from Mogadishu to Baidoa. The lowest price difference was $3.50. Lower than this was impossible due to insecurity.”60 Another frustrated businessman stated, “Under the warlords, the price of goods was high and the demand was low. It took two days to get the goods from the port to the market.”61 While driving to Afgooye, a prominent civil-society actor recalled: “This trip should only take us fifteen minutes, but it used to take five or six hours because of the number of checkpoints.”62

The head of Coca-Cola in Mogadishu simply gave up on in-land markets because the additional cost of checkpoints collapsed demand for his luxury good. “In the time of warlords, I couldn't sell Coke outside of Mogadishu,” he said. “My average production was 500 cases/day and 15,000 cases/month. But my actual production capacity was 285,000 produced per month, but I couldn't sell this because of the insecurity on the roads.”63 The chief executive of the El Ma'an Port Authority, one of the most powerful corporations in the country and a key financier of the ICU, complained that hiring private security was not a sufficient solution: “Just to provide security we needed 1,000 soldiers and 100 technicals. Eighty percent of our revenues were spent on security. We never saved anything from El Ma'an.”64

By 2004 the business elite had begun to organize a collective effort to support the Islamic courts to reduce these costs. Large and medium-sized businesses confirmed in interviews that the rates they were paying under the TFG were a crucial factor in their decision to support the ICU project.65 The leaders of Somali business associations played a key role in coordinating this effort across clan divisions.66 One prominent Mogadishu trader, who was a leader in financing the courts throughout the 2000s, described how the business community systematically worked together during this critical period.

There was a regular budget paid by the business community and the clans to [the courts]. At first, it was only to your own area, your own clan. Every clan took money to his own court only. Then upon centralization, taxes and fees were given at the district level. These were not taxes, but voluntary contributions. There was a committee of 65 elders who collected for them. In the beginning of 2004, [the business community] allocated $1,000 per month, per district. By the end of 2004 until mid-2005, we gave $3,000 per month, per district. From mid-2005 to 2006, we gave $10,000 per month, per district. In Yaqshiid District, near Suuq Ba'ad market, the amount of our contributions increased over time, as [the courts'] activities increased. In 2006, [the ICU] took the whole city, so there was no more need for voluntary contributions. We went from voluntary contributions to taxation, paid by all people for the security of their own houses.67

The former head of the Economics Department at Mogadishu University explained, “The reason that the business community supported the ICU was because of the illegal taxes.”68 Prominent businesspeople from every sub-clan faction in Mogadishu confirmed that cost of security under the courts was lower than it had been under clan warlords. The chairman of Suuq Ba'ad market succinctly stated, “For every $100 we were forced to pay to the warlords, we could give $35 to the Islamic Courts [to remove them].”69

After the Islamic Courts expelled the warlords in 2006, however, the business community reported a sharp decline in their power vis-à-vis the ICU. “The ICU came up with regulatory practices, such as stopping the qaat trade and cigarettes, that were not attractive at the time,” explained the owner of a multimillion-dollar cigarette business. Despite being a chief financier of the courts, he was unable to challenge these policies that were harming his company. “We could not stop supporting them because they were a power to be reckoned with,” he stated.70 Once the ICU had secured its total control, the movement had consolidated enough power and resources to be able to subjugate the rest of the periphery. The competitive security market had become a monopoly of force. The business community quickly found itself subjugated by the Islamist leviathan it had created.

The rise of Islamist power in 2006 had a transformative effect on the Somali civil war and a lasting impact on regional and international security. In January 2007, the United States backed an ill-devised Ethiopian invasion to overthrow the Islamic courts and reinstall the TFG warlords. The fragile new ICU government was crushed by the powerful Ethiopian offensive. The invasion, however, also galvanized al-Shabaab field commanders, who subsequently adopted a more extreme transnational agenda and welcomed an influx of foreign fighters into their ranks.71 When in 2009 ICU leader Sheikh Sharif finally negotiated a compromise with the TFG, al-Shabaab rejected his political leadership and labeled him a traitor and apostate. By 2010 al-Shabaab had established control over the majority of the countryside, generating tens of millions of dollars in taxation revenues from Kismayo port, as well as heavily extorting businesses along the Afgooye corridor.72 Two years later, the group declared its formal merger with al-Qaida, provoking another wave of fighting. After fifteen years of tribal conflict, Somalia had become home to one of the largest al-Qaida operations in the world. By overlooking the rational economic drivers of Somalia's Islamist uprising, the international intervention inadvertently helped to create a virulent ideological insurgency.73

To contain the crisis, the international community expanded African Union operations and financed the creation of the Federal Government of Somalia, which has regained control over a number of key cities, including the capital.74 With the influx of foreign support for the new Somali government, the Mogadishu business elite has enjoyed a new wave of economic investment. Yet, when asked whether or not they would be willing to pay for their government through taxes, members of Mogadishu's new Chamber of Commerce immediately replied in unison, “It depends on how much it will cost.”75 Somalia's most powerful and wealthy made it clear that if the taxation rate under the government proves too high, then support for the new government may decline.76 Whether the new Somali government succeeds in creating a level of order similar to that of the ICU will depend on its ability to generate support from its domestic business base.

Conclusion

The research presented in this article offers an initial examination of the effect of security costs on political preferences of an important and often overlooked class: the local business elite. A plethora of factors can affect individual political preferences, both for and against Islamists. This research suggests, however, that both ethnic and Islamic identity politics may actually be overstated in the existing literature, and that more work needs to focus on the pragmatic cost calculations of the business elite. By modeling civil war as a market for the provision of security, this article offers a practical, economic explanation for why Islamists may prove more competitive than groups defined by narrower identities. I argue that because certain Islamist groups have the ability to sell security to prospective buyers across ethnic or tribal divisions, they are able to charge the business elite lower rates than protection rackets that rely on a more limited ethnic or tribal base. By courting the valuable support of the business class, Islamists are thus able to monopolize the market for providing security. The Somalia case provides a useful lens to explore this ideal-type theory, and the results of the in-depth empirical study illustrate how the business elite's need for lower security costs gave the Islamic courts a competitive advantage in the civil war.

Dissecting the relationship between business networks and Islamist groups may also help to explain conflict dynamics in other cases. In North Africa, illicit business activity along the Tunisian-Algerian border is a chief source of funding for militant Islamist groups in that area. The International Crisis Group reports that “criminality and radical Islamism gradually are intermingling in the suburbs of major cities and in poor peripheral villages. Over time, the emergence of a so-called islamo-gangsterism could contribute to the rise of groups blending jihadism and organised crime within contraband networks operating at the borders—or, worse, to active cooperation between cartels and jihadis.”77 In fact, these networks are so intertwined that top al-Qaida in the Islamic Maghreb commander Mokhtar Belmokhtar was nicknamed “Mr. Marlboro” for his extensive involvement in North Africa's $1 billion cigarette-smuggling industry.

Competition for control of the business community has also been a source of conflict among rival Islamists. In Pakistan's restive tribal region, Tehrik-e-Taliban Pakistan (TTP) is the dominant Islamist faction in the tribal belt and operates as an umbrella organization of different loosely allied tribal militias.78 Some leading TTP commanders have profited heavily from their relationships with criminal and smuggling networks, whereas other tribal commanders within the organization have fought against criminal organizations. There is strong evidence that unequal access to extortion revenues from the local business class was the source of serious conflict within the TTP leadership, causing the group to split. The Islamist split in Pakistan has had an important effect on the war economy. Businesspeople in the border town of Dera Ismail Khan report that whereas they used to only have to pay off one group, since the Islamist in-fighting began, they are now being extorted by multiple Islamist protection rackets.79 Understanding which tribal factions within TTP benefit from criminal business and which tribes have been excluded from these economic opportunities is necessary to explain the fissures within Pakistan's Islamist insurgency, and its relationship to powerful criminal networks.

Criminal networks have also played a key role in the rise of Islamist power in Syria and Iraq. In both countries, Islamist insurgents have reportedly built strong ties to regional smuggling networks and are predominantly involved in looting; arms trafficking; and the black market sale of oil, humanitarian aid, and antiquities.80 Preliminary research in Syria has found that these criminal networks are related to Arab tribal structures in rebel-held areas.81 Iraq's Islamist militants have also allegedly grown rich through oil and antiquities smuggling. The relationship among local tribal organizations, transnational Islamism, and regional criminal networks in both Syria and Iraq, however, remains understudied.

Given the international community's security concerns surrounding these Islamist groups, further investigation into their relationships with the local business class is warranted. This research has shown that there are practical reasons why local business elites may favor Islamists over ethnic or tribal groups. This observation speaks to many cases of complex civil war in majority-Muslim states, especially where a dynamic war economy already exists. When violent ethnic or tribal divisions ratchet up security costs for local businesses, Islamists will be uniquely positioned to court the support of powerful economic elites. In fact, the less bound to ethnic or tribal politics an Islamist group is, the more likely it will be able to sell security to businesses across these local divisions.

The business community's alliance with Islamists in civil wars is driven by rational, economic considerations. These material interests are not, however, inherently in conflict with the non-Muslim world, nor are these businesspeople intrinsically favorable toward Islamist groups. Rather, the business class clearly responds to economic incentives. Taking seriously the cost calculations of the business elite, including those involved in smuggling activities, is therefore critical to developing effective policy in these types of conflicts.

Most important, this research suggests that both formal and informal taxation rates have a significant effect on the preferences of powerful elites. This finding has implications for the design and construction of postconflict governments in other parts of the world. For example, in Afghanistan the current national government is too large in size and scope to be effectively sustained through domestic taxation revenues anytime in the near future. If international support for the government dwindles, the local business class may be unwilling to invest in maintaining these institutions, especially if its members can buy protection for a lower rate elsewhere. Given the international community's limited resources, designing postconflict governments that can be sustained by a willing domestic tax base is essential for the long-term success of peacebuilding projects. If this market power is harnessed effectively, however, it may actually provide new opportunities for cooperation and peacebuilding.

Acknowledgments

The author is especially grateful to outstanding colleagues in Kenya and Somalia, including Jabril Abdulle, Sahal Abdulle, Dr. Ibrahim Farah, Dr. Yahya Amir Ibrahim, and Dr. Deqo Mohamed. The author also thanks Theodore McLauchlan, Stephen Saideman, Stuart Soroka, and the anonymous reviewers for comments on earlier drafts, as well seminar and workshop participants at the Belfer Center and McGill University who provided helpful feedback. The research for this study was carried out with funding from the Social Sciences and Humanities Research Council and the International Development Research Centre.

Notes

1. 

The incidence of religious civil wars in majority-Muslim countries is striking. For further explanation of religious civil wars, see Monica Duffy Toft, “Getting Religion? The Puzzling Case of Islam and Civil War,” International Security, Vol. 31, No. 4 (April 2007), pp. 97–131.

2. 

On clan politics in Somalia, see Ken Menkhaus, Somalia: State Collapse and the Threat of Terrorism, Adelphi Paper No. 364 (Oxford: Oxford University Press for the International Institute for Strategic Studies, 2004); and Ken Menkhaus, “There and Back Again in Somalia” (Washington, D.C.: Middle East Research and Information Project, February 11, 2007), http://www.merip.org/mero/mero021107. On ethnic and tribal politics in Afghanistan and Pakistan, see Rasul Bakhsh Rais, Recovering the Frontier State: War, Ethnicity, and the State in Afghanistan (Karachi: Oxford University Press, 2008); and Ahmed Rashid, Descent into Chaos: The World's Most Unstable Region and the Threat to Global Security (London: Pan, 2008). On Islamist ideology in Somalia, see Andre Le Sage, “Prospects for Al Itihad and Islamist Radicalism in Somalia,” Review of African Political Economy, September 2001, pp. 472–477; and Robert I. Rotberg, ed., Battling Terrorism in the Horn of Africa (Washington, D.C.: Brookings Institution Press, 2005). On the ideology of al-Qaida and affiliate groups, see Assaf Moghadam, “Motives for Martyrdom: Al-Qaida, Salafi Jihad, and the Spread of Suicide Attacks,” International Security, Vol. 33, No. 3 (Winter 2008/09), pp. 46–78.

3. 

For further reading on the Somali civil war, see Terrence Lyons and Ahmed Ismail Samatar, Somalia: State Collapse, Multilateral Intervention, and Strategies for Political Reconstruction (Washington, D.C.: Brookings Institution Press, 1995); Stig Jarle Hansen, Al Shabaab in Somalia: The History and Ideology of a Militant Islamic Group, 2005–2012 (Oxford: Oxford University Press, 2013); and Benjamin Powell, Ryan Ford, and Alex Nowrasteh, “Somalia after State Collapse: Chaos or Improvement?” Journal of Economic Behaviour & Organization, Vol. 67, Nos. 3–4 (September 2008), pp. 657–670.

4. 

On the relationship between tribalism and Islamist groups in Somalia, see Roland Marchal, “Islamic Political Dynamics after the Somali Civil War: Before and After September 11,” in Alex de Waal, ed., Islamism and Its Enemies in the Horn of Africa (Indianapolis: Hurst, 2004), pp. 114–145; Roland Marchal, “Warlordism and Terrorism: How to Obscure an Already Confusing Crisis? The Case of Somalia,” International Affairs, Vol. 83, No. 6 (November 2007), pp. 1091–1106; and Ken Menkhaus, “The Crisis in Somalia: Tragedy in Five Acts,” African Affairs, July 2007, pp. 357–390.

5. 

Further discussion of these events can be found in Hansen, Al Shabaab in Somalia; Harry Verhoeven, “The Self-Fulfilling Prophecy of Failed States: Somalia, State Collapse, and the Global War on Terror,” Journal of Eastern African Studies, Vol. 3, No. 3 (October 2009), pp. 405–425; and Christian Webersik, “Mogadishu: An Economy without a State,” Third World Quarterly, Vol. 27, No. 8 (Winter 2006), pp. 1463–1480.

6. 

The appendix and dataset are available at doi:10.7910/DVN/28760.

7. 

Clan divisions have been the defining feature of Somali politics for centuries. Although Somalia's population is almost entirely Muslim, clan, rather than Islam, has been the dominant form of social organization throughout its history. For fifteen years, the civil war was fought along clan lines. Islamic groups had little to no influence, and the population was so xenophobic that there were almost no foreign fighters in the theater before 2006. As a result, the majority of existing research on Somalia gives primacy to clan politics. The best works on the history of Somali clan politics are Lee V. Cassanelli, The Shaping of Somali Society: Reconstructing the History of Pastoral People, 1600–1900 (Philadelphia: University of Pennsylvania Press, 1982); David D. Laitin and Said S. Samatar, Somalia: Nation in Search of a State (Boulder, Colo.: Westview, 1987); Abdi Ismail Samatar, The State and Rural Transformation in Northern Somalia, 1884–1986 (Madison: University of Wisconsin Press, 1989); and Ioan Lewis Understanding Somalia and Somaliland (New York: Columbia University Press, 2008).

8. 

The study of rare events, such as Islamist revolutions in civil wars, presents challenges for case selection. See Barbara Geddes, “How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics,” Political Analysis, Vol. 2, No. 1 (1990), pp. 131–150. Given the dearth of rigorous field research, however, in-depth investigation of Islamist insurgencies is needed for the generation of hypotheses. Moreover, selecting “hard cases” for analysis presents the opportunity to test other seemingly obvious explanations against evidence that should readily confirm them. For a discussion on the use of case studies for hypothesis generation, see Andrew Bennett and Colin Elman, “Case Study Methods in the International Relations Subfield,” Comparative Political Studies, Vol. 40, No. 2 (February 2007), pp. 170–195; John Gerring, “Is There a (Viable) Crucial-Case Method?” Comparative Political Studies, Vol. 40, No. 3 (March 2007), pp. 231–253; and Jack S. Levy, “Qualitative Methods in International Relations,” in Michael Brecher and Frank P. Harvey, eds., Millennial Reflections on International Studies (Ann Arbor: University of Michigan Press, 2002), pp. 432–454.

9. 

Since 1991 a handful of dedicated scholars have produced a small but rigorously researched body of literature on Somalia. See Matt Bryden and Jeremy Brickhill, “Disarming Somalia: Lessons in Stabilization from a Collapsed State,” Conflict, Security, and Development, Vol. 10, No. 2 (May 2010), pp. 239–262; Ken Menkhaus, “Governance without Government in Somalia: Spoilers, State Building, and the Politics of Coping, International Security, Vol. 31, No. 3 (Winter 2007), pp. 74–106; Menkhaus, “The Crisis in Somalia”; Webersik, “Mogadishu”; Marchal, “Islamic Political Dynamics after the Somali Civil War”; and Hansen, Al Shabaab in Somalia.

10. 

Within-case comparison is useful here because clan, religion, geography, and other variables can be held constant. Longitudinal study allowed for a more careful examination of the small number of variables that may have affected outcomes. For a discussion of within-case analysis, see Alexander L. George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences (Cambridge, Mass.: MIT Press, 2004).

11. 

Given the wide range of definitions in the existing literature, the term “political Islam” can include everything from nonviolent student groups to terrorist networks. For a discussion of the diverse types of Islamic movements, see Quintan Wiktorowicz, Islamic Activism: A Social Movement Theory Approach (Bloomington: Indiana University Press, 2004). For an analysis of whether Islamist terrorism differs from other forms of terrorism, see Martha Crenshaw, “The Debate over ‘New’ vs. ‘Old’ Terrorism,” in Ibrahim A. Karawan, Wayne McCormack, and Stephen E. Reynolds, eds., Values and Violence: Intangible Aspects of Terrorism (Dordrecht, Netherlands: Springer, 2008), pp. 117–136.

12. 

The term “warlord” is the subject of much controversy. I use it here without any pejorative connotations. Roland Marchal argues that the term warlord is not only poorly conceived, but is also overused in the Somali context. See Marchal, “Warlords and Terrorism.” Kimberly Marten defines the term as “individuals who control small pieces of territory using a combination of force and patronage.” See Marten, Warlords: Strong Arm Brokers in Weak States (Ithaca, N.Y.: Cornell University Press, 2012), p. 3. I discuss the meaning of the Somali term for warlords used in the field study in the appendix. For discussion of warlord politics in Africa, see also William Reno, Warlord Politics and African States (Boulder, Colo.: Lynne Rienner, 1998).

13. 

Andre Le Sage observes that the distinction between a businessperson and a warlord is not always perfectly neat. See Le Sage, “Somalia: Sovereign Disguise for a Mogadishu Mafia,” Review of African Political Economy, Vol. 29, No. 1 (March 2002), pp. 132–138.

14. 

Charles Tilly, “War Making and State Making as Organized Crime,” in Peter B. Evans, Dietrich Rueschemeyer, and Theda Skocpol, eds., Bringing the State Back In (Cambridge: Cambridge University Press, 1985), pp. 169–191.

15. 

Seminal works on civil war economies include Peter Andreas, Blue Helmets and Black Markets: The Business of Survival in the Siege of Sarajevo (Ithaca, N.Y.: Cornell University Press, 2011); Paul Collier and Anke Hoeffler, “Greed and Grievance in Civil War,” Oxford Economic Papers, Vol. 56, No. 4 (October 2004), pp. 563–595; and Paul Collier and Nicholas Sambanis, eds., Understanding Civil War, Vol. 1: Africa: Evidence and Analysis (Washington, D.C.: World Bank, 2005).

16. 

Mancur Olson describes these types of actors as “stationary bandits.” See Olson, Power and Prosperity: Outgrowing Communist and Capitalist Dictatorships (New York: Basic Books, 2000). For a discussion of how economic shocks created warlord politics in Africa, see Reno, Warlord Politics and African States.

17. 

If the rate of extortion in stateless societies is too high (such that it incites rebellion) or too low (causing warlords to engage in predatory behavior), then political order falls apart. For a game-theoretic model of the protection-extortion relationship between warlords and citizens, see Robert Bates, Avner Greif, and Smita Singh, “Organizing Violence,” Journal of Conflict Resolution, Vol. 46, No. 5 (October 2002), pp. 599–628. Bates, Greif, and Singh outline the consequences of extortion rates that are too high or too low on p. 615.

18. 

As Robert H. Bates explains, political order occurs at an equilibrium point at which neither party has a strong enough incentive to defect from the political status quo. See Bates, When Things Fell Apart: State Failure in Late-Century Africa (Cambridge: Cambridge University Press, 2008), p. 17.

19. 

For discussion of how identity politics constrain side-switching in civil war, see Theodore McLauchlin, “Loyalty Strategies and Military Defection in Rebellion,” Comparative Politics, Vol. 42, No. 3 (April 2010), pp. 333–350. For further discussion of complex ethnic civil wars, see Halvard Buhaug and Kristian Skrede Gleditsch, “Contagion or Confusion? Why Conflicts Cluster in Space,” International Studies Quarterly, Vol. 52, No. 2 (June 2008), pp. 215–233; Donald L. Horowitz, Ethnic Groups in Conflict (Berkeley: University of California Press, 1985); and David A. Lake and Donald S. Rothchild, eds., The International Spread of Ethnic Conflict: Fear, Diffusion, and Escalation (Princeton, N.J.: Princeton University Press, 1998).

20. 

Ronald Rogowski, “Causes and Varieties of Nationalism: A Rationalist Account,” in Edward A. Tiryakian and Rogowski, eds., New Nationalisms of the Developed West: Toward Explanation (Boston: Allen and Unwin, 1985), pp. 87–108; Douglass C. North, Institutions, Institutional Change, and Economic Performance (New York: Cambridge University Press, 1990); Francis Fukuyama, Trust: The Social Virtues and the Creation of Prosperity (New York: Free Press, 1995); Deborah Brautigam, “Substituting for the State: Institutions and Industrial Development in Eastern Nigeria,” World Development, Vol. 25, No. 7 (July 1997), pp. 1063–1080; and Jean-Phillipe Platteau, “Institutional Obstacles to African Economic Development: State, Ethnicity, and Custom,” Journal of Economic Behaviour and Organization, Vol. 71, No. 3 (September 2009), pp. 669–689.

21. 

Hiring private security has a number of financial advantages to consider. For example, a businessperson could both use a road containing a large number of checkpoints and own some of the checkpoints on that road, turning insecurity into another source of profit by taxing other users of the road. Businesspeople can also choose to hire their own militias to protect their businesses, in order to pay less at checkpoints than their competitors.

22. 

In a competitive security market, the business class does retain a degree of voice and exit power in these coercive relationships; in a civil war, however, both voice and exit options are always risky. On exit power, see Albert O. Hirschman, Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States (Cambridge, Mass.: Harvard University Press, 1970). On the role of identity in Islamic movements, see Kathleen Collins, “Ideas, Networks, and Islamic Movements: Evidence from Central Asia and the Caucuses,” World Politics, Vol. 60, No. 1 (October 2007), pp. 64–96; and Wiktorowicz, Islamic Activism.

23. 

Platteau, “Institutional Obstacles to African Economic Development,” p. 39. For a discussion of how Pakistan has attempted to use Islamic identity to allay ethnic conflict, see Stephen Philip Cohen, The Idea of Pakistan (Washington, D.C.: Brookings Institution Press, 2004). See also Seyyed Vali Reza Nasr, Islamic Leviathan: Islam and the Making of State Power (New York: Oxford University Press, 2001).

24. 

Howard J. Rotblat, “Social Organization and Development in an Iranian Provincial Bazaar,” Economic Development and Cultural Change, Vol. 23, No. 2 (January 1975), pp. 292–305.

25. 

The utility of the business-Islamist alliance is well established in the literature on Iran. Although there are significant differences between the Iranian Revolution and the emergence of political Islamism in collapsed states, the Iran case provides useful insights on the role of economic factors in Islamic movements. See Ervand Abrahamian, “Structural Causes of the Iranian Revolution,” MERIP Reports, No. 87 (May 1980), pp. 21–26; Said Amir Arjomand, “Iran's Islamic Revolution in Comparative Perspective,” World Politics, Vol. 38, No. 3 (April 1986), pp. 383–414; and Mehdi Mozaffari, “Why the Bazaar Rebels,” Journal of Peace Research, Vol. 28, No. 4 (November 1991), pp. 377–391.

26. 

The utilization of religious credentials by elites is a strategy for outbidding rival groups and generating a broader base of support. For a thorough discussion of religious outbidding, see Toft, “Getting Religion?”

27. 

A number of critical factors precipitated the collapse of the state. For a discussion of Cold War interference in Somalia, see Peter Schwab, “Cold War on the Horn of Africa,” African Affairs, January 1978, pp. 6–20. For a discussion of economic mismanagement during the Siad Barre regime, see Jamil A. Mubarak, “The ‘Hidden Hand’ behind the Resilience of the Stateless Economy of Somalia,” World Development, Vol. 25, No. 12 (December 1997), pp. 2027–2041.

28. 

In response, the United Nations (UN) Security Council in 1992 sent to Somalia its first ever peace enforcement operation, with a broad and ambitious mandate. The UN mission quickly became embroiled in clan politics, however. For further analysis of the UN mission, see Lyons and Samatar, Somalia; and John L. Hirsch and Robert B. Oakley, Somalia and Operation Restore Hope: Reflections on Peacekeeping and Peacemaking (Washington, D.C.: U.S. Institute of Peace Press, 1995).

29. 

Although businesses must pay fees to warlords who own airstrips, ports, and roads, state failure provides savvy businesspeople with a valuable comparative advantage in tax-, duty-, and regulation-free trade. Individuals who can navigate the high-risk, high-reward game of conducting business in the context of political anarchy can do extremely well under conditions of state failure. The trade, transit, and smuggling of goods into neighboring states can turn a healthy profit. See Peter D. Little, Somalia: Economy without State (Bloomington: Indiana University Press, 2003); Webersik, “Mogadishu”; and Stig Jarle Hansen, “Civil War Economies, the Hunt for Profit, and the Incentives for Peace,” AE Working Paper No. 1 (Bath, U.K.: Department of Economics and International Development, University of Bath, 2007); Aisha Ahmad, “Agenda for Peace or Budget for War: Evaluating the Economic Impact of International Intervention in Somalia,” International Journal, Vol. 67, No. 2 (Spring 2012), pp. 313–331; Lyons and Samatar, Somalia; Hansen, Al Shabaab in Somalia; and Powell, Ford, and Nowrasteh, “Somalia after State Collapse.”

30. 

Even Somalia's most fervent and radical Islamist organization of the 1990s, al-Ittihad al-Islamiyya, eventually failed on account of debilitating tribal politics. Indeed, clan militias were so xenophobic that few foreign fighters were able to safely operate in the countryside.

31. 

As Islam spread through the Horn of Africa, Islamic law melded with existing cultural institutions, such as the Somali customary legal code known as Xeer. See Michael van Notten, The Law of the Somalis: A Stable Foundation for Economic Development in the Horn of Africa, ed. Spencer Heath McCallum (Lawrenceville, N.J.: Red Sea, 2005). For a survey of Somali clans, religion, and culture, see Lewis, Understanding Somalia and Somaliland.

32. 

Menkhaus, “Governance without Government in Somalia,” p. 88.

33. 

Islamic courts have existed in Somalia for hundreds of years, serving as a religious institution that provides social and legal services to the community. After the fall of the Somali state in 1991, these courts took on a larger social role to fill the political void; however, earlier attempts by these courts to politically organize were repressed by clan-based warlords. For more on the historical evolution of Islamic courts in Somalia, see Marchal, “Islamic Political Dynamics after the Somali Civil War.” See also Andre Le Sage, “Stateless Justice in Somalia: Formal and Informal Rule of Law Initiatives” (Geneva: Centre for Humanitarian Dialogue, July 2005); and Ken Menkhaus, “Political Islam in Somalia,” Middle East Policy, Vol. 9, No. 1 (March 2002), pp. 109–123.

34. 

Marchal, “Islamic Political Dynamics after the Somali Civil War.”

35. 

The TNG was also formed during a time of global economic prosperity, before the crackdown on Somali remittances after the September 11, 2001, terrorist attacks.

36. 

Sheikh Sharif was the head of the Sii-Sii court in Mogadishu's Yaaqshiid District in the north of the city.

37. 

Al-Shabaab translates to “the youth” or “the boys.”

38. 

Menkhaus, “There and Back Again in Somalia.”

39. 

Cedric Barnes and Harun Hassan, “The Rise and Fall of Mogadishu's Islamic Courts,” Journal of Eastern African Studies, Vol. 1, No. 2 (July 2007), pp. 151–160; Saul Shay, Somalia between Jihad and Restoration (New Brunswick, N.J.: Transaction, 2008); and Jon Abbink, “The Islamic Courts Union: The Ebb and Flow of a Somali Islamist Movement,” in Stephen Ellis and Ineke van Kessel, Movers and Shakers: Social Movements in Africa (Leiden, Netherlands: Brill, 2009), pp. 87–113.

40. 

Menkhaus, “There and Back Again in Somalia”; and author interview with Matt Bryden, Nairobi, Kenya, October 2009.

41. 

See de Waal, Islamism and Its Enemies in the Horn of Africa; Barnes and Hassan, “The Rise and Fall of Mogadishu's Islamic Courts”; Shay, Somalia between Jihad and Restoration; and Abbink, “The Islamic Courts Union.” See also Verhoeven, “The Self-Fulfilling Prophecy of Failed States,” for a critique of the allegations that the ICU was a terrorist organization.

42. 

Author interviews with early members of the ICU movement and civil society actors suggest that the diaspora and other non-Somali Islamist actors were late players, supporting the Islamists after the latter had already purged the clan warlords. There continues to be a dearth of academic research on the impact of Somali diaspora support on the civil war, however. For an assessment of money transfer agencies and diaspora politics in Somalia, see Khalid Mustafa Medani, “Financing Terrorism or Survival? Informal Finance, State Collapse, and the US War on Terrorism,” Middle East Report, Summer 2002, pp. 2–9.

43. 

This estimate is drawn from author interviews with leaders of the Mogadishu business community conducted in October and November 2009. Stig Jarle Hansen's 2007 survey found that 69 percent of the business community admitted to voluntarily contributing to the Islamic courts. See Hansen, “Civil War Economies, the Hunt for Profit, and the Incentives for Peace.” See also Webersik, “Mogadishu.”

44. 

This survey is the first and largest of its kind. Two previous studies have presented simple statistical data based on smaller samples of the Mogadishu business community, but they provide descriptive statistics rather than regression analyses. See Hansen, “Civil War Economies, the Hunt for Profit, and the Incentives for Peace”; and Roland Marchal, A Survey of Mogadishu's Economy (Nairobi: European Commission, 2002).

45. 

A number of scholars have managed to conduct surveys in failed states. See, for example, Jason Lyall, Kosuke Imai, and Graeme Blair, “Explaining Support for Combatants during Wartime: A Survey Experiment in Afghanistan,” American Political Science Review, Vol. 107, No. 4 (December 2011), pp. 679–705; and Andrew Beathe et al., “Randomized Impact Evaluation of Phase II of Afghanistan's National Solidarity Programme (NSP): Estimates of Interim Program Impact from First Follow-Up Survey” (Washington, D.C.: World Bank, July 2010).

46. 

Preliminary interviews were conducted as early as 2004, with additional interviews in 2007 and 2009, and follow-up interviews in 2013.

47. 

Intra-Hawiye clan competition defines much of the political landscape in Mogadishu, and considerable effort was made to include respondents from a range of Hawiye sub- and sub-sub-clan groups. I also interviewed leading chief executive officers (CEOs) from non-Hawiye clans, including the Darod and Isaaq.

48. 

Nearly 79 percent (78.86 percent) of all respondents said that general insecurity was the biggest challenge for their business, followed by 11.47 percent who said that banning or regulation of industry was the worst for business, and 9.57 percent who said that high taxes were the worst.

49. 

Ninety-seven percent of all respondents agreed to some extent that they would be willing to pay higher taxes to remove checkpoints and have a peaceful central government

50. 

Several diagnostic tests were conducted to rule out problems of multicollinearity with all independent variables in the model.

51. 

It is possible that respondents who indicated they had a university-level education include those who pursued Islamic studies at the university level. Islamic universities across the Middle East are a popular destination for wealthier Somalis seeking advanced education. It is not possible to decipher this nuance from these survey results, however, and further investigation is needed to ascertain the relationship between education type and political preference.

52. 

Author interviews with Darod-Marehan business elites, October 2009. These business elites were pushed out of Mogadishu when the central government fell and Hawiye clans became more dominant in the region. Notably, Siad Barre was from the Darod clan, so older elite businesspeople who benefited from his clan privilege expressed resentment about their losses after the collapse of the government. This reaction, however, was not clearly based on clan. Some Hawiye respondents who lost a lot of their businesses after the Barre regime collapsed made similar statements about newcomers being criminals, which suggests a resentment between winners and losers in the new civil war economy.

53. 

Author interviews with CEOs of the largest Somali, businesses in Nairobi, Kenya, and Dubai, United Arab Emirates in 2009; in Hargeisa, Somalia, in 2011; and in Mogadishu, Somalia in 2013. Conducting an exhaustive survey of the clan compositions of the governing boards of major Somali corporations was impossible. I did, however, conduct interviews with owners and shareholders of these corporations from a wide range of clan backgrounds. These CEOs self-reported that multi-clan governance structures originated because, in order for a business to expand into the territory of another clan, it was necessary for the business owner to establish a partnership with a prominent businessperson from that clan.

54. 

Author interviews with prominent qaat traders, Nairobi, Kenya, November 2009. Qaat is a popular narcotic drug that is consumed widely in the Horn of Africa. In Somalia, the qaat industry is dominated by women.

55. 

Author interview with Avocat Zakaria, ICU legal counsel, Nairobi, Kenya, November 2009. Some clan warlords and their supporters were more assertive, calling the business community “Wahhabis” and “hypocrites.”

56. 

Author interview with a former dean of a Mogadishu-based business school, November 2009.

57. 

Author interview with a leading business trader, November 2009.

58. 

Author interview with the head of one of the largest multi-conglomerates in Mogadishu, November 2009.

59. 

Author interview with the chairman of Suuq Ba'ad market, November 2009.

60. 

Author interview with a midsized trader, November 2009.

61. 

Author interview with a businessman, October 2009.

62. 

Author interview with the chief executive of a humanitarian organization in Lafoole, Somalia, February 2013.

63. 

Author interview with the regional head of Coca-Cola in Mogadishu, 2009. He added, “With the checkpoints, we lost both time and money. Delays on the road also reduce profits because you make only one trip when you could have made two or three trips back and forth with the same truck.”

64. 

Author interview with the chief executive of El Ma'an Port Authority, November 2009.

65. 

Author interviews with multiple sources from various levels of business practice throughout Somalia and Kenya, 2009 and 2013.

66. 

Some business associations in the major markets of Mogadishu (such as Bakara, Hamar Weyne, and Suuq Ba'ad) are organizational bodies that bring together major players in these markets. These business community associations are chaired by the most prominent businesspeople in these markets, and these networks function as forums for sharing information that affects the business community as a whole. I held meetings with members and chairmen of these Mogadishu-based business associations, and discovered that these associations are critical for coordinating the efforts of the business community as a whole.

67. 

Author interview with a Mogadishu-based businessman working mostly in imports, November 2009.

68. 

Author interview with a professor of economics, Mogadishu University, November 2009.

69. 

Author interview with the chairman of Suuq Ba'ad market, November 2009.

70. 

Author interview with the owner of a large cigarette and trading business, May 2007.

71. 

For a discussion of foreign fighters, see Thomas Hegghammer, “The Rise of Muslim Foreign Fighters: Islam and the Globalization of Jihad,” International Security, Vol. 35, No. 3 (Winter 2010/11), pp. 53–94; and Hansen, Al Shabaab in Somalia.

72. 

Author interviews with the chief executive of a humanitarian organization in Lafoole, Somalia, January and February 2011; and author interview with Matt Bryden on taxation levels at Kismayo port, Toronto, Ontario, February 2014.

73. 

Verhoeven, “The Self-Fulfilling Prophecy of Failed States”

74. 

At the time of this writing, the majority of the countryside was divided among rival clan factions allied with or included in the new government and al-Shabaab. For further reading on the expelling of African Union forces, see Matt Bryden, “Somalia Redux? Assessing the New Somali Federal Government” (Washington, D.C.: Center for Strategic and International Studies, 2013).

75. 

Author interviews with members of Somali Chamber of Commerce, Mogadishu, Somalia, February 2013.

76. 

For a discussion of the economic advantages of statelessness in Somalia, see Peter Leeson, “Better Off Stateless: Somalia before and after Government Collapse,” Journal of Comparative Economics, Vol. 35, No. 4 (December 2007), pp. 689–710; Webersik, “Mogadishu”; and Little, Somalia.

77. 

International Crisis Group, “Tunisia's Borders: Jihadism and Contraband,” Middle East and North Africa Report No.148 (Brussels: International Crisis Group, November 2013), p. i.

78. 

TTP commanders both participate in criminal activities and act as a protection racket in extorting the business community. See Gretchen Peters, “Crime and Insurgency in the Tribal Areas of Afghanistan and Pakistan,” ed. Dan Rassler (West Point, N.Y.: Combating Terrorism Center at West Point, October 2010).

79. 

Katharine Houreld and Jibran Ahmad, “Crime Spree Helps Pakistani Taliban Squirrel Away Cash before Raids Begin,” Reuters, July 10, 2014.

80. 

The black market for antiquities is a multibillion-dollar global industry.

81. 

For further discussion, see Reinoud Leenders and Steven Heydemann, “Popular Mobilization in Syria: Opportunity and Threat, and the Social Networks of the Early Risers,” Mediterranean Politics, Vol. 17, No. 2 (Summer 2012), pp. 139–159.