The Political Economy of Political Survival: External Debt Accumulation and Default during Civil Conflict

Abstract: 

Building on prior national case studies of the economic consequences of civil conflict, this paper articulates and analyzes a theory of external debt accumulation by states during and immediately after civil conflict. The demands of military and public goods provision on state leaders are a matter of political survival during civil conflict, which drives an economically unwise but politically indispensable increase in the external debt burden. The composition of this debt burden is affected by the presence and intensity of conflict, which then affects bilateral lending, multilateral finance, and aid in distinct ways. Default is also considered as a means of conflict financing by states, given the importance of conflict financing. The findings broadly support the importance of disaggregating the dynamics of debt and default during civil conflict.

Table of Contents: 

    Introduction

    What is the impact of civil conflict on external debt accumulation by sovereign states? Case studies on this question exist, but as of now, no cross-national examinations of this question have been undertaken. Developing country debt is an economic issue of great salience for the long and short-term economic prospects of states with the greatest risk of civil conflict. The existence of a class of highly indebted poor countries, many of which have experienced civil conflict, is undisputed, but the dynamics of debt accumulation and the role of external debt in the “conflict trap” are unclear.

    From both a theoretical and practical perspective, understanding the full extent of civil conflict’s costs is essential to explaining how conflicts are begun, sustained, and resolved. While most states in the midst of civil conflict are unable to access bond markets, their external debt stocks often increase. This irregularity may be explained by the presence of other sources of financing such bilateral or multilateral concessional lending and foreign bank debt, but the analyses to date are unclear on this question. Such alternatives, along with the economic costs of civil war (Collier, 1999; Fitzgerald, 2001b; Murdoch & Sandler, 2002), mean that states have the motivation and means to take on a greater debt burden. Indeed, debt accumulation may play a key role in the survival of state leaders during civil conflict.

    This paper will lay out an analysis of debt accumulation by states in the midst of civil war from 1970 to the 2012, beginning with a review of the literature on sovereign debt as well as the economic caU.S.es and consequences of civil conflict. After that, a theory of debt accumulation and the prospects for default will be articulated and subjected to a quantitative analysis.

    Sovereign Debt and the Economic CaU.S.es and Consequences of Civil Conflict

    The Theoretical Underpinnings of Sovereign Debt

    The central puzzle in the literature on sovereign debt is how states in an anarchic international system are able to borrow funds in the first place. Unlike civil bankruptcy, the creditors of a government have no formal legal avenue of recourse should the state decide not to honor its obligations or change the terms of the contract, and this would at first seem to be an insurmountable barrier to loaning funds to a state. Despite the decay from absolute immunity to partial immunity for states (Bulow & Rogoff, 1989b) under U.S. law in 1976, legal challenges are still unlikely to succeed (Panizza, Sturzenegger, & Zettelmeyer, 2009). States can now face lawsuits they do not consent to hear in U.S. courts. Even then, default on sovereign obligations is a relatively uncommon phenomenon with, by one measure, 248 external defaults of varying durations between 1820 and the present day. Defaults also tend to occur in waves as shifts in the credit market impact a large group of states simultaneoU.S.ly (Beers & Chambers, 2007: Reinhart & Rogoff, 2009).

    Could sanctions or the implicit threat thereof provide recourse for private and public creditors of states? Some scholars argue that coercion is the key force undergirding sovereign debt markets, and point to incidents such as the 1902 blockade and bombardment of Venezuela by British, French, and German vessels and repeated U.S. interventions in the Caribbean (Mitchener &Weidenmier, 2005). The most recent empirical evidence for this theory predates the First World War, and is subject to debate (Tomz, 2007). Another possibility is that the influence of private creditors on their legislatures leads to the implementation of trade sanctions (Bulow & Rogoff, 1989a). Such sanctions face a collective action problem (Wright, 2005) and the empirical evidence for trade sanctions is divided. See Tomz (2007) or Rose (2005). In place of external sanctions, some scholars explain the domestic economic ramifications of default as a deterrent for default (Panizza, Sturzenegger , & Zettelmeyer, 2009; Wright, 2011). Default or repayment, however, create distinct sets of winners and losers (Hartzell, Hoddie, & Bauer, 2010; Tomz, 2003) among actors with varying interests (Tomz, 2004), meaning the political impact of default may not be uniform.

    States have also defaulted under widely varying economic conditions (Tomz &Wright, 2013), which implies that economic concerns cannot fully explain why states repay their debts so regularly. With this empirical puzzle in mind, the most plausible explanation for debt repayment is advanced by Tomz, (2007), who points to the importance of states’ concerns for their reputation in capital markets as the key to explaining repayment. According to the reputational theory of debt repayment, states understand that default will damage their reputation in international markets, thus precluding access to capital or increasing the interest they must pay. Such financing is a positive good, inasmuch as it increases the capacity of the state to provide goods for the population and finance other initiatives. For example, differing levels of access to credit can help explain the relative power of states in the international system (Schultz & Weingast, 2003). As such, states have a strong incentive to maintain access to credit in one form or another, and can do so only by maintaining a good reputation for repayment in the credit markets.

    Despite the importance of capital market access, default can be sparked by political (Hatchondo & Martinez, 2010; Saiegh, 2009; Tomz, 2003; Van Rijckeghem &Weder, 2009) or economic (Reinhart & Rogoff, 2009) shifts. External debt accumulation can lead to economic issues such as a debt “overhang” which reduces growth and diverts fiscal resources from other initiatives (Eichengreen, 1991), which may help explain default. Underdevelopment and conflict can preclude access to the bond markets on the part of states. However, these same states can still accumulate debt from bilateral loans at concessional rates, international development assistance, and multilateral financing (Addison & Murshed, 2003). So while developing nations can take on external debt, the externalities are not exclusively positive. What, then, are the linkages between civil conflict and debt default? To answer this question, we must first understand the economic causes and ramifications of civil conflict.

    An Economic Logic of Civil War

    Economic considerations have a long history in the literature on the causes of civil war. According to one school of thought, if actors face sufficient deprivation relative to other groups in society or their perceptions of what is possible for them to acquire, violence is more likely (Gurr, 1970). Once such grievances become sufficiently acute, actors are more willing to engage in violent challenges to the state. Under this model, changes in the state or international system (Gurr, 1994) or a lack of coercive capacity (Fearon & Laitin, 2003) leave states vulnerable to challenges by aggrieved or disempowered actors. Irregular transitions, even in democratic systems, can also provide an opportunity for rebellion, regardless of the exact motives of the rebels (Gleditsch & Ruggeri, 2010).

    These grievance-centered arguments may lack some empirical support (Collier & Hoeffler, 2004), and are vulnerable to criticism based on the misrepresentation of grievances by actors, who may overplay the role of grievance in justifying their resort to violence (Hirschliefer, 2001). The theory of “relative deprivation” does, however, point to the economic incentives that can spark and perpetuate conflict, and this “greed” based theory of civil conflict has a great deal of explanatory power. Ethnic fractionalization, the availability of natural resources (Ross, 2004), and income per capita all affect the prospects for a group seeking to start a violent challenge to the state (Collier & Hoeffler, 1998). The interaction between these macro characteristics and the capacity of the state to maintain both a redistributive capacity and the rule of law (Fjelde & De Soysa 2009; Gurr, 1994) can explain much of the international variance in patterns of civil war. Growth shocks, coupled with weak institutions, could also play a role in precipitating civil conflict (Miguel, Satyanath, & Sergenti, 2004).

    One key consensus in the literature is that, as levels of political and economic development increase, the risk of civil war decreases (Elbadawi & Sambanis, 2002; Collier et al., 2003; Stewart & Fitzgerald, 2001). Economic development is especially important as it raises the opportunity cost of engaging in rebellion for individuals (Collier & Hoeffler, 1998). In states with a large population, especially ones with a high number of unemployed young men, the risk of civil war is increased as potential rebels have a greater pool of recruits to draw from and form the nucleus of an insurgency (Fearon & Laitin, 2003). This lowers the costs of war, which is the essential consideration of potential rebels considering a violent challenge to the state (Keen 2001a).

    Conflicts over the redistributive role of the state can play a role in sparking civil conflict (Azam, 2001) and revenue sharing has been a key component of several successful peace negotiations (Aning & Atuobi, 2011). Such issues often turn on ethnic identity and fractionalization, which is another key risk factor in explaining the incidence of civil conflict. Ethnic cleavages can be utilized by political entrepreneurs to form a winning coalition, but when this division disempowers a significant ethnic minority, the risk of civil conflict increases significantly (Fearon, 2006; Azam, 2001; Bridgman, 2007). Furthermore, vertical income inequality is not associated with incidents of civil conflict, but horizontal or inter-group inequality is a much better predictor of the odds of civil conflict (Øsby 2008). Policies such as structural adjustment, which invariably create groups of “winners” and “losers” within a society, can further exacerbate these divisions (Hartzell & Hoddie, 2010).

    Ethnic divisions also exacerbate the risk of civil conflict by retarding development. Competitions over power and rent-seeking behavior tied to ethnic identity have helped stunt economic progress in Africa, or at least have exacerbated existing economic challenges such as a lack of infrastructure development (Easterly &Levine, 1997; Bridgman, 2007). Ethnic divisions cannot completely explain the lack of economic development and poor policies that have left many states at risk of civil conflict (Stewart & Fitzgerald, 2001; Sarkees, Wayman, & Singer, 2003), but they do help clarify why a large number of developing states have a high risk of civil war (Collier et al., 2003). Civil conflict itself can also help explain these economic difficulties and the risk of renewed conflict, as conflict is a negative economic externality both sub-nationally and regionally.

    Economic Consequences of Civil War

    The economic ramifications of civil conflict are myriad, but can be broadly divided into the economic distortions and destruction wrought by conflict. The destruction of infrastructure and capital is both tangible and pernicious (Collier, 1999; Kang & Meernik, 2005). Social and human capital also suffers from the increased mortality and forced migration caused by conflict (Collier, 1999). Overall, short term GDP growth and the post-conflict GDP growth paths also suffer, leading to a decline in output (Murdoch & Sandler 2002). Conflict also changes the composition of a state’s economy by instigating sectoral shifts in a manner that is particularly prejudiced against capital-intensive industries (Collier 1999). Even then, agricultural production invariably suffers (Collier, et al., 2003; Sen, 1992). The prevalent economic uncertainty and insecurity created by conflict also favors easily transportable tradable goods at the expense of capital goods even when a civil conflict has ended (Collier & Gunning 1995).

    States in the midst of civil conflict face a choice between the production of economic goods or conflict goods (Hirshleifer, 2001). Individual groups within society must also choose an optimal amount of conflict goods to acquire in order to defend their interests and deter predatory behavior (Anderton, 2000), and this represents a further diversion of resources from otherwise productive economic activities. The prosecution of conflict by the state also creates fiscal imbalances as leaders deal with contractions in output and revenue (Fitzgerald, 2001a), lost income from the export sector (Anderton & Carter, 2000), and increased military spending (Stewart, Huang, & Wang 2001). All of these changes, in conjunction with the costs of post-conflict recovery and reintegration (Ghobarah, Huth, & Russett 2003; Kang & Meernik, 2005; Wennmann, 2011), mean that civil wars, especially those in Sub-Saharan Africa, represent an enormous impediment to sustained development (Sen, 1994).

    Conflict is also an important force in shaping the dynamics of particular regions, as there are negative externalities for neighboring states as well (Gleditsc, 2007), particularly in the form of lost international trade and investment. Neighbors of countries in the midst of civil conflict do face a modest but significant decline in output (Murdoch & Sandler, 2002), and often shift resources to strengthen their military capacities due to a reasonable fear of contagion (Sen, 1994). Direct or indirect intervention can also change the regional dynamic as other states strive to profit from or contain a civil conflict in their region (Rosenau, 1964). Conflict is thus a key determinant of the dynamics in any region, and also an outgrowth of those same dynamics (Gleditsch, 2002) as it creates positive or negative cycles.

    How do these economic consequences of civil conflict impact external debt accumulation by the state? Several case studies are present in the literature, and the general conclusion is that foreign debt burdens increase (Stewart, Huang, & Wang, 2001), albeit at a slightly slower rate, relative to other states without civil conflict (Fitzgerald 2001b; Alvarez-Plata & Brück 2008). Much of this debt is from development agencies or bilateral concessional lending (loans from one state to another at a lower interest rate), which has led some scholars to argue for debt forgiveness (Addison & Murshed, 2003; Alvarez-Plata & Brück, 2008), but the effectiveness of that policy in reducing the debt burdens of these highly indebted poor countries (HIPCS) is disputed (Easterly, 2002). The arguments in favor of debt forgiveness emphasize that debt forgiveness will increase the ability of the state to provide goods to groups within the society. While conflict, therefore, likely precludes access to bond markets, there is an overall dynamic of external debt accumulation. Conflict creates an additional, potentially existential, imperative for states to seek financing, to which we will now turn.

    Theory: The Need for Finance during Civil Conflict

    Civil conflict is a potentially existential threat to the leaders of a state. Insurgent victories are rare, but this outcome always results in the transformation of the regime or government. Incumbent leaders, should they desire to retain their position and possibly avoid losing their head, must provide public and private goods to their supporters to ensure that they are not deposed (Bueno de Mesquita et al., 2003). This provision of goods is made much more difficult by the risks civil conflict present for lenders of any type, leading to a supply and demand imbalance wherein sovereign demand for credit is higher than the supply. Furthermore, should conflict intensity increase sufficiently, it is almost certain that states would lose access to most external financing, owing to additional risk.

    According to the selectorate theory of Bueno de Mesquita et al. (2003), leaders form a winning coalition of supporters from a selectorate of actors within the polity. Selectorate members are enfranchised electorally or otherwise in determining the selection and survival of a leader in office. Democracies have larger selectorates and winning coalitions, necessitating the provision of public goods to supporters, while autocracies have a small selectorate and winning coalition which instead favors the provision of private goods to members of the winning coalition. Rebels can come either from the selectorate or the disenfranchised sector of society, but they are rarely part of the winning coalition. This exclusion comprises the core of the existential threat to the leadership of the state. With a rebel triumph, the incumbent loses access to both the private goods associated with control of the state and the ability to provide public or private goods to supporters. During separatist rebellions, the incumbent might lose access to revenues or resources from the disputed territory, thus weakening their position relative to future challengers.

    How would we expect the leaders to behave in the face of civil conflict? A common first response to a real or perceived armed threat is an increase in military spending (Stewart, Huang, & Wang 2001). This increase has two functions: ensure the support of the armed forces and increase the odds of securing a military victory. However, military spending is only one part of the state’s obligations during civil conflict, as private and public goods must be provided to ensure the support of other actors, be it the populace at large or key political figures (Bueno de Mesquita, et al. 2003; Hirshleifer, 2000). Otherwise, the promised public or private goods provision by either the rebels or an internal challenger might prove sufficient to induce defections from the winning coalition, thus threatening the leader’s survival.

    These additional pressures to provide goods might be easily realized in peacetime, but civil conflict places tight economic constraints on most states. As we have seen, economic growth suffers during civil conflict as capital and infrastructure are destroyed, exports suffer, and foreign investment retreats. This implies that borrowing is necessary for the state to both prosecute the war and maintain consumption, as a fiscal deficit would be all but inevitable under these conditions. Even in the event of a government victory, the costs of reconstruction (Ghobarah, Huth, & Russett, 2003) and pernicious economic legacies of conflict (Collier et al., 2003) will likely drive an increase in the debt burden. As such, the threat of deposition is sufficient to justify decisions that are unwise from an economic perspective, such as taking on greater debt while the export sector declines, or expropriating industry and other domestic assets to finance conflict.

    The experience of Mozambique prior to the implementation of a peace accord with the Renamo in 1993 provides anecdotal evidence for this theory (Brück, 2001) as the state’s debt burden increased during conflict. Nicaragua’s decade-long civil war is also a useful case study, as the conflict between the Sandinista government and Contras led to a fiscal deficit that peaked at 20 percent of GDP and an output contraction. This fiscal deterioration can be attributed to macroeconomic conditions, increased defense expenditures, and the implementation of rural development programs (Spalding, 1987). The development projects represent an important example of public goods provision by a state during civil conflict. U.S. support for the Contras restricted access to aid and other financing from Western states leaving Nicaragua facing an acute shortage of external financing (Fitzgerald & Grigsby, 2001). It was only with the end of the conflict that Nicaragua regained access to large-scale financing.

    H1a: As civil conflict duration increases, external debt stocks will increase.

    H1b. As conflict intensity increases, external debt stocks will increase.

    H1c: During the first 5 years after conflict termination, external debt stocks will increase.

    Beyond this broad portrait of external debt accumulation by states, it can be hypothesized that conflict impacts debt in three distinct ways. First, the presence of civil conflict is enough to create potential fears for the political survival of leaders, as described above. Second, the intensity of conflict impacts these calculations of survival and the risk perceived by potential creditors. Last, the economic legacy of conflict and return of peace place new demands on states (Ghobarah, Huth, & Russettt 2003; Kang & Meernik, 2005) to rebuild their economy and address the deterioration of human wellbeing that conflict leaves as a legacy. These three aspects of financing, namely the presence and duration of conflict, the intensity of conflict, and the post-conflict legacy, impact different types of financing  in distinct manners.

    As we have seen, conflict financing cannot be predicated on the economic situation of the debtor state alone, as external capital flows would be greatly diminished. As such, the focus should shift to financing that are less tied to economic performance. Bilateral loans are one such alternative. Bilateral lending is particularly attractive for states, as such financing often has geopolitical connotations, and states can appeal for loans from friendly governments. The implication of this political rationale is that bilateral funding is least likely to be impacted by increased conflict intensity, and that such loans will increase in the aftermath of conflict.

    H2a: As conflict duration increases, bilateral lending will increase.

    H2b: As conflict intensity increases, bilateral lending will increase.

    H2c: During the first 5 years after conflict termination, bilateral lending will increase.

    Multilateral funding is less likely to be continued during conflict (Stewart, Huang and Wang 2001), especially during the most intense conflicts, which represent a greater political risk. Instead, multilateral funding should increase in the aftermath of conflict as states seek to finance rebuilding efforts. The political costs of conditionality agreements imply that multilateral financing is an unlikely source for states in the midst of conflict, but is more likely in the aftermath as states seek funds for rebuilding.

    H3a: As conflict duration increases, multilateral funding will decrease.

    H3b. As conflict intensity increases, multilateral funding will decrease.

    H3c. During the first 5 years after conflict termination, multilateral funding will increase.

    Development or humanitarian aid represents another source of financing that is not tied to robust macroeconomic conditions. Aid is also fungible and can be used to support increased military expenditures (Collier & Hoeffler, 2007), making it an attractive source of financing for states. Recipients of aid also have another means of providing public goods such as health services and thus reducing their immediate fiscal obligations.  Indeed, aid played a key role in the conflict in Sierra Leone, as it allowed the government to prosecute the conflict while aid agencies provided public goods such as health services (Keen, 2001b). While aid is not traditionally thought of as a source of financing, it remains an important aspect of debt accumulation in the HIPCs (Beers & Chambers, 2007) and thus should be considered. Furthermore, one could expect that aid might actually increase to a certain extent as conflict intensity increases, and will certainly increase in the period immediately following the cessation of hostilities (Collier et al., 2003).

    H4a. As conflict duration increases, net official development assistance received will increase.

    H4b. As conflict intensity increases, net official development assistance received will increase.

    H4c. In the first five years after conflict termination, net official development aid received will increase.

    Under this theory of civil conflict and external debt accumulation, what is the relationship between civil conflict and the possibility of default by a state? While a state at peace might default under a variety of economic (Reinhart & Rogoff, 2009; Eichengreen, 1991) and/or political conditions (Hatchondo & Martinez, 2010; Van Rijckeghem & Weder, 2009), the pressures of political survival during conflict mean that access to financing is a possibly existential concern. A state might temporarily free up some resources via default, but the long term costs in additional financial instability (Reinhart and Rogoff, 2009) and a destroyed reputation for repayment (Tomz, 2007) would likely outweigh the immediate benefits. Default would almost certainly preclude further access to the bond markets for a state, and this would further limit the ability of the state to prosecute the conflict and maintain consumption.

    Default on “bank debt” (loans directly from private banks to governments), however, has been more common since 1960 (Beers & Chambers, 2007), and was widespread during the 1980s debt crisis. Governments have taken to treating bond debt as senior to bank debt, which helps explain this irregularity, but empirical research on this question remains limited. Regardless, default on bank debt would free up resources and help maintain a favorable balance of payments. Inflation and a weakening domestic currency also increase the burden of debts in foreign currency, increasing the likelihood of default. Hence, during civil conflict, states would be less likely to default on bond debt, but bank debt denominated in foreign currency, would be more vulnerable to default. Local currency default, given the control of state over the banking system, would also be vulnerable to default.

    H5a: Civil conflict presence decreases the likelihood of sovereign default on bonds.

    H5b: Civil conflict presence increases the likelihood of default on bank debt in foreign currency.

    H5c: Civil conflict presence increases the likelihood of default on debt in local currency.

    We can anticipate that civil conflict will lead to an increase in a state’s debt holdings via international aid and bilateral lending, while multilateral funding simultaneously decreases. Under the logic of political survival during civil conflict, one would expect these patterns to deepen as conflict intensity rises, but only to a point. The most intense civil conflicts, such as Afghanistan’s civil war, can destroy the basic functions of the state and the formal financial sector (Marsden & Samman, 2001). Such conflicts also augment the risks to creditors, and are likely to further restrict the supply of finance, leaving only states with foreign sponsorship with the ability to access additional financing. Intense, destructive civil conflicts will also be more likely to induce default as leaders prioritize the short-term exigencies of survival over long-term market access and stability and are less able to roll-over (i.e., use new loans to repay previous debt) existing debt.

    H6a. As civil conflict intensity increases, debt service on external obligations will decrease.

    H6b. As civil conflict intensity increases, the likelihood of bond default will increase

    H6c. As civil conflict intensity increases, the likelihood of bank default will increase.

    H6d. As civil conflict intensity increases, the likelihood of local currency default will increase.

    It would be possible to argue the converse of this theory, namely that civil conflict is caused by debt accumulation and its effects on economic growth. Such an argument neglects the relative economic costs of debt accumulation and civil conflict and the motivations by which civil wars begin. A debt overhang becomes most problematic in the period immediately before a possible default as growth contracts, but there is no empirical evidence linking conflict onset with growth contractions induced by “debt intolerance.” Debt tolerance is the concept that individual economies can handle varying levels of debt before growth affects the risk. Furthermore, no threshold of debt relative to GDP has been conclusively found to cause the growth shocks many scholars believe to impact the odds of civil conflict occurrence. Debt is often just one component of factors such as poor policies, ethnic divisions, and competitions over power that leave states economically underdeveloped and vulnerable to civil war (Easterly & Levine, 1997; Sen, 1994). So far, scholarly work on growth contractions and civil conflict onset has focused on the role of instrumental variables such as rainfall in Sub-Saharan Africa to proxy for declines in income (Miguel, Satyanath & Sergenti, 2004). This is unsurprising as developing states with a low-per-capita income are unlikely to have robust and diffuse financial systems, thus limiting the direct impact of any crisis sparked either by default or debt accumulation. Indeed, debt may lack enough political salience in developing nations to provide any grievance, or opportunity for greed (Saiegh, 2005).

    In summary, we can expect that conflict creates serious difficulties for servicing both new and old debt, by damaging a state’s export capacity and economic growth. Regardless of these difficulties, the logic of political survival dictates that states seek financing to prosecute conflict and provide public or private goods to supporters. This financing can take multiple forms, including aid and bilateral lending.  Regardless, the disequilibrium between supply and demand outlined in this theory points to the importance of finance in sustaining states in the midst of civil conflict.

    How then to analyze these linkages between debt and civil conflict? The following section details the methods by which this question will be addressed.

    Data and Methodology

    This analysis draws on a universe of cases comprised of 110 developing nations from Latin America, Africa, the former Yugoslavia, the Middle East and Central Asia, as well as South and East Asia. The sample of developing nations was drawn from the World Bank’s classification of low and lower middle income countries in the above regions, in order to keep geographic and economic consistency among the states to be analyzed. There are several notable exceptions to this pattern, especially the exclusion of states with distinctive economic and demographic characteristics that influence their capacity for debt accumulation even during civil conflict, such as India, China, and the Koreas. Qatar and the United Arab Emirates are excluded for similar reasons. Chile’s recent ascension to middle income status means that it is included for historic purposes, and also to ensure comprehensive coverage of Latin America and Central America. The above regions, particularly Africa, have also seen the highest incidence of civil conflict, but still contain a large portion of states that have not experienced civil conflict. Temporally, the analysis extends from 1970 to 2013, owing to the lack of reliable and consistent cross-national data on debt service and other macroeconomic variables from earlier time periods. In summary, the unit of analysis for this study is the country-year, and the data in question is pooled panel data.

    Debt Service and the Dependent Variables

    The complexities of debt accumulation during civil conflict dictate that several dependent variables be used during the course of this analysis. These variables include change in external debt stocks, bilateral lending, multilateral debt, official development assistance, debt service and the likelihood of default. All of these economic variables, excluding default, are drawn from the World Bank’s World Development Indicators Databank (2014).

    First, the primary variable of interest, change in the external debt stock, is operationalized as the change in external debt stocks in current U.S. dollars from year to year. This captures the accumulation of debt regardless of its maturity. A common measure such as the debt to GDP ratio is more comprehensive, but many states have not kept or reported accurate statistics on their debt/GDP ratio. Other measures, including debt service as a ratio of exports or GNI, are vulnerable to distortions caused by the income and export declines during civil conflict. Hence, while this operationalization is imperfect, it is the best means of capturing the overall change in the debt burden for states in the midst of civil conflict.

    The other dependent variables are more straightforward. Bilateral lending is captured by the World Bank data on Public and Publicly Guaranteed (PPG) loans, which includes both bilateral loans to the state, and bilateral loans the state has chosen to guarantee. This captures the full inflow of credit and extent of state obligations, as the state can facilitate credit inflows by guaranteeing loans. Multilateral debt is operationalized as the percentage of total external debt. Another hypothesized means of financing for states in conflict is official development assistance, which is operationalized as a percentage of GNI, and captures the importance of development assistance to the macroeconomy. It is also the most complete of the applicable World Bank measures of ODA.

    To examine hypothesis 6a, debt service is operationalized as total debt service on external obligations in current U.S. dollars. The data on default is based on a variety of sources, but is primarily drawn from Standard and Poor’s data (Beers & Chambers, 2007; Chambers, 2012) with some supporting data from Tomz and Wright (2010).The data is then disaggregated into foreign currency bond defaults, foreign currency bank debt defaults, and local currency defaults. In all three cases, default occurs either when an interest or principal payment is missed, or a rescheduling reduces the value owed (Reinhart & Rogoff, 2009). As noted in the theory section, default on debt denominated in bonds is the least frequent of these. Default on obligations to banks is more common among the two forms of foreign currency default. In this particular temporal and spatial domain, instances of bond and local currency default represent less than 4 percent of the sample. Lastly, state default on obligations in local currency is less common than bank default, which may reflect both a lack of data and other irregularities in reporting (Reinhart & Rogoff, 2008). The local default data covers from 1975 to the present, which will shift the temporal domain of that particular analysis. All three types of default are coded as binary variables, 1 for a year of default, which can be contiguous, and 0 for years in which a particular type of default has not taken place.

    Civil Conflict

    The independent variables of interest for this study are the presence and intensity of civil conflict. These variables are operationalized using the UCDP-PRIO Armed Conflict Dataset (Themnér & Wallensteen, 2014). The authors of this data define intra-state conflict as the use of armed force between two parties, one of which is the state, the other of which is a non-state actor, which is a sufficiently robust definition to prevent the inclusion of less violent types of conflict. Furthermore, the UCDP-PRIO data disaggregates a range of conflict intensities, instead of using a single threshold, allowing for the empirical examination of the hypotheses described in the theory section.

    Civil conflict presence is operationalized via the use of conflict years from 1970 to the present. For years in which a civil conflict is active at any level greater than 25 battle-related deaths per year, the variable is coded as a one, but if no such conflict is present, a zero is entered, creating a binary variable. This is applied based on the presence of war types 3 and 4 from the same data set, which are internal conflict and internationalized internal conflict, respectively. This variable allows for the examination of hypotheses measuring civil war presence and duration, although it does not capture the nuances of civil conflict, particularly the type of conflict, be it separatist or otherwise.

    The UCDP-PRIO data measures two levels of conflict intensity, namely low intensity conflicts (those ranging from 25 to 1000 battle-related deaths in a given year) and high intensity conflicts (those with more than 1000 battle-related deaths in a given year), which allows us to test hypotheses 6a-d. This variable is coded with a 1 for low intensity conflicts and a 2 for high intensity conflicts. The UCDP-PRIO data also includes a variable for the cumulative intensity of a conflict, which is a binary variable coded 1 if the aggregate battle-related deaths from a conflict surpass 1,000. The low threshold of the cumulative intensity variable and the lack of differentiation beyond the 1,000 death mark necessitates the use of the annual intensity variable to analyze Hypotheses 6a-d.

    Post Conflict Peace

    The 5 year period of post-conflict peace is operationalized using the episode end variable from the UCDP-Prio Armed Conflict Dataset (Themner & Wallensteen, 2014) and the war presence variable. The end of episode variable deals with the resolution of a dyad, and thus may code for the end of an episode between one rebel dyad and the state, but not the end of civil conflict as a whole. Hence, the variable for post-conflict peace is only coded for the five years after the end of an episode if civil conflict is also absent. The five year window for this variable was selected in order to omit cases wherein conflict recurrence occurred relatively quickly, while also giving a temporal horizon that is sufficiently long for the dynamics of debt accumulation to shift, but short enough to differentiate the immediate effects of conflict from the long-term economic conditions within each state.

    Control Variables

    In order to distinguish the causal effect and direction of the relationship between civil conflict and debt default, it is necessary to control for a several variables. The effect of regime type on debt accumulation (Saiegh, 2009; Schultz & Weingast, 2003; Van Rijckeghem & Weder 2009), wherein democracies have greater access to the bond markets than autocratic regimes, necessitates the inclusion of the Polity IV variable, which classifies regimes on a scale from -10 (full autocracy) to 10 (full democracy). As natural resource production affects both the odds of conflict occurrence and the fiscal prospects of the state, oil rents as a percentage of GDP is included as a control. Both Nigeria and Venezuela have used oil warrants as a debt instrument in international markets (Beers & Chambers 2007), which has allowed both states an additional degree of access to financing, albeit not without complications (Chung & Mahtani 2007). In Angola, the MPLA was able to attract lucrative foreign direct investment in the oil sector, which helped finance the war effort once support from the U.S.S.R. waned (Frynas & Wood, 2001).

    GDP per capita impacts both conflict risk (Collier & Hoeffler, 2004) and debt accumulation (Panizza, Sturzenegger, & Zettelmeyer, 2009; Reinhardt Rogoff, 2009). Furthermore, while it has been hypothesized that growth slows during conflict, there is variance across cases (Stewart & Fitzgerald, 2001), and GDP growth also plays a key role in supporting the fiscal base of the state (Reinhart & Rogoff 2009; Tomz & Wright, 2013). The ramifications of these two variables make GDP per capita and GDP growth essential controls. Inflation is another key control that is relatively straightforward to measure, and it is operationalized as the annual percentage change in consumer prices. Lastly, the effect of debt forgiveness may have a role in at least temporarily reducing sovereign debt (Addison & Murshed, 2003; Alvarez-Plata & Brück, 2008; Easterly, 2002). Debt forgiveness is operationalized as the total debt forgiveness grants for a year in current U.S. dollars. All the economic variables are based on the World Bank Development Indicators and are denominated in current U.S. dollars (World Bank 2014).

    To help differentiate the risk of civil conflict from debt accumulation, population is used as an instrumental variable in the relevant regression analyses. While population has been shown to have a direct relationship with the risk of civil conflict (Fearon & Laitin, 2003), no such empirical relationship has been established between population and debt accumulation. Instead, scholars of debt and default have emphasized a confluence of economic (Tomz & Wright, 2013) and political factors (Hatchondo & Martinez, 2010) as explanatory variables.

    Data on ethnic wars from the Ethnic Power Relations (EPR) (Wimmer, Cederman, & Min 2009) dataset is also used as a robustness check, given the economic impacts of ethnic divisions (Bridgman, 2007; Easterly & Levine, 1997) and the conflict risk associated with ethnic polarization (Gurr, 1994; Azam, 2001). In this case, the EPR binary variables for the presence of ethnic civil war and high intensity war will be analyzed. Ethnic conflict should not have a markedly different effect from other forms of civil conflict on external debt accumulation by states.

    Analytical Techniques

    In order to examine the first four hypotheses, a linear regression model is most appropriate, owing to the hypothesized relationships and continuous nature of the dependent variables. The use of population as an instrumental variable means that Two-Stage Least Squares instrumental variable regression is the specific statistical tool used. Each model includes the control variables described above. For hypotheses 5a-c and 6b-d, a logistic regression model is used, as the variable is coded as a binary. However, both bond and local currency defaults, measured by year, comprised less than 3 percent of the sample, necessitating the use of a rare events logistic regression model to ensure the validity of the results for hypotheses H5a, H5c, H6b and H6d.

    Results

    Presence of Civil Conflict

    The following results are organized by the independent variables of war presence, conflict intensity, and the post-conflict period. The logistic regression data on the impact of war presence and intensity on the odds of default is combined into one section. For the first three regression models, population is used as an instrumental variable to differentiate between the risk of civil conflict and debt accumulation. First, the effects of war presence on the change in external debt stocks (H1a), bilateral lending (H2a), multilateral financing (H3a), and development assistance (H4a) are seen below. It was hypothesized that civil conflict presence would increase external debt stocks, bilateral lending, and development assistance, while decreasing multilateral funding. The results of that analysis, an instrumental variable two stage least squares regression, can be found below in Table 1.

    The presence of civil conflict has a statistically significant effect on all of the above aspects of external debt. Conflict presence causes a significant increase in total debt stocks and bilateral financing, while also reducing the share of multilateral financing. However, contrary to Hypothesis 4a, net official development assistance as a percent of GNI declines during conflict.

    In this first series of models examining the impact of the presence of civil conflict on debt accumulation, several control variables fail to achieve significance, including GDP growth and inflation. Otherwise, the controls for institutions and debt forgiveness are marginally significant, and the sign for the institutional coefficient is negative. Otherwise, greater GDP per capita is a significant predictor of increased debt stocks and bilateral funding as well as decreased multilateral funding and development assistance, while oil revenues decrease the external debt burden and reliance on bilateral and multilateral funding.

    Conflict Intensity

    The next series of hypotheses to be examined dealt with the impact of conflict intensity measured in battle deaths on the four types of debt accumulation; overall debt stocks, bilateral debt, multilateral debt as a percentage of total debt, and net official development assistance as a percentage of GNI. Hypotheses 1b, 2b, and 4b posited that greater conflict intensity would increase external debt stocks, bilateral debt, and official development assistance, respectively. Hypotheses 3b and 6a anticipated an inverse relationship between conflict intensity, multilateral financing and debt service. The results of the examination of these hypotheses are summarized below in Table 2.

    Civil conflict intensity has a statistically significant impact on all of the forms of debt accumulation and debt service. In the case of changes in external debt and bilateral financing, the signs of the coefficients are positive, as expected. Multilateral financing as a percentage of total debt also decreases significantly. However, official development assistance has an inverse relationship with conflict intensity, which was unexpected, but is again understandable. The direct relationship between conflict intensity and total debt service was much more surprising. Despite greater conflict intensity, which presumably indicates a greater threat to the survival of state leaders, developing states actually increase their debt service, which means that resources that could finance military expenditures or consumption are instead used to pay debt.

    Otherwise, the control variables behave in a similar manner to the war presence models. Debt forgiveness only serves to reduce multilateral financing. Inflation and GDP growth do not have a significant effect on debt accumulation or service. The effects of GDP per capita and oil rents are similar to the war presence model in their respective direct and indirect relationships with the four types of debt accumulation. This contrasting effect of these two forces on debt service is notable, as GDP per capita increases debt service, but oil rents decrease it. Under this model, institutions lack any statistically significant explanatory power.

    Post-Conflict Peace

    Changes in external debt accumulation during the five years after the end of conflict were also analyzed using a two stage least squares instrumental variable regression. Hypotheses 1c, 2c, 3c and 4c posit that external debt stocks, bilateral financing, multilateral financing, and official development assistance will all increase during this time period, as peace returns  and the risks of conflict subside. The results of this analysis are presented below in Table 3.

    The pattern of external debt accumulation from conflict, as noted above, runs contrary to the expectations of the theory. While the post conflict variable has a significant effect on all four types of financing, the coefficients for both multilateral financing are negative, indicating that even in the 5 years after the end of conflict, these two types of financing retain a secondary importance. However, debt forgiveness has a statistically significant role in reducing the total and bilateral debt burden of a state in the post conflict environment. GDP per capita and oil rents remain significant, and the substantive role of oil rents in reducing the total external debt of a state increases.

    Odds of Default

    A logistic regression analysis was used to analyze Hypotheses 5a-c and 6b-d, which cover the odds of bond, bank, and local currency default under the presence of civil conflict and greater conflict intensity. Given the distribution of bond defaults and local currency defaults in the sample, a rare events logistic regression model was used as a robustness check, which did not result in any substantial changes in the results presented above.  It was posited that the presence of civil conflict decreases the likelihood of bond default, wherein reputational concerns are paramount for states, while the odds of bank and local currency would increase. Greater conflict intensity in addition to the presence of conflict is expected to increase the odds of all forms of default as leaders seek whatever funding is immediately available to ensure their survival. The results of these analyses are presented below in Table 4.

    The above models indicate that neither civil conflict presence nor intensity have a significant effect on the odds of bond or local currency default. These same forces, however, significantly increase the odds of default on debt to banks that is denominated in foreign currency. This split may be caused by the low incidence of bond and local currency default among the 100 countries in this sample. Regardless, the bank default finding is intriguing, if inconclusive. The lack of differentiated default data for many developing nations, and the likely underreporting of local currency defaults mean that this finding is preliminary, and requires more comprehensive examination.

    Robustness

    These above results for the impact of civil conflict presence and intensity on debt accumulation were also subjected to a robustness check using the Ethnic Power Relations Data Set (Wimmer, Cederman, & Min, 2009). When a similar instrumental variables regression model was run, replacing the presence of civil conflict with the presence of ethnic conflict, the results retained the same substance and statistical significance. The results also held for high intensity ethnic conflicts, which helps substantiate the above findings. Tables for the results of this robustness check can be found in Table 5 and Table 6.

    Discussion

    The above findings provide important, but by no means conclusive, support for the case studies that have detailed the economic consequences of civil conflict (Stewart, Huang, & Wang 2001; Keen, 2001b; Frynas & Wood, 2001; Brück, 2001; Marsden & Samman, 2001; Spalding, 1987). Unsurprisingly, some of the pertinent details of these individual cases are lost to the aggregation of civil conflicts across 40 years and several regions. Regardless, those aggregate findings have several interesting implications that broadly support the theory of external debt accumulation as a key component of political survival during civil conflict while also reinforcing the difficulty states face in accessing such funding.

    We have seen that conflict drives an increase in external debt stocks, but the sources of this financing are of particular interest. The most important source of financing under the above models is bilateral lending, which implies that the geopolitical importance of the conflict may help a state secure financing. While regional and global powers can sponsor rebel groups, support for friendly states is also common, and was more so during the Cold War, which was characterized by proxy wars. Fears of regional contagion, either economic or political, may also leave other states predisposed to provide financing, even when the presence and intensity of a civil conflict might otherwise discourage lending.

    This is particularly important given the negative impact of conflict presence and intensity on multilateral financing and official development assistance. While it was hypothesized that the exigencies of conflict would lead to increased development assistance, as states aggressively sought such aid to offset the costs of war, and that donors and lenders would respond in kind, the risks of conflict may well be too great to justify development assistance projects. An alternative measure of aid, however, might show different results, as some development assistance has a longer temporal horizon than meeting the immediate needs of a state in the midst of civil conflict.

    The implications of the drop off in multilateral financing, even in the post-conflict environment, indicate a probable lack of both supply and demand for such lending. Conflict represents a serious economic and political hurdle to multilateral financial institutions, which, in the event of a rebel victory, might see their loans repudiated as “odious debt” or face repudiation by leaders with a greater interest in survival than repayment. The obligations of the previous regime are considered oppressive by the victors, and are thus not paid. The Bolshevik Revolution provides an early and well-known case of this. For their part, leaders may be unwilling to risk backlash from implementing a conditionality agreement to secure multilateral financing, which invariably favors certain segments of society (Hartzell, Hoddie, & Bauer 2010). Rebuilding from conflict is enormously costly, but the political costs of multilateral funding may preclude its use in the post-conflict environment.

    This analysis also represents a preliminary step in understanding what relationship, if any, exists between civil conflict and default. While there is a clear relationship between default on bank debt and the presence and intensity of conflict, the other cases are much more muddled. The insignificance of bond default could be attributable to a lack of sufficient bond market access on the part of developing nations to accumulate a sufficient debt burden for default to take place. Or it may be that reputational concerns are sufficient to all but extirpate bond defaults on what debt is held by the countries in this sample. Local currency default is more complicated, as while the relationship is statistically insignificant, the reason for this is difficult to discern. Underreporting of domestic debt and default is one possible cause that is unlikely to be remedied anytime in the immediate future. However, the lack of local currency defaults in the sample may reflect that other means of default or “sovereign theft” (Tomz and Wright 2010) are used by states in the midst of conflict. The resolution of these two unsettled questions represents a promising avenue for future research.

    This project represents a first step in the cross-national understanding of the political economy of political survival during civil conflict, and, as such, it is not a comprehensive analysis of external debt accumulation during civil conflict. Private lending and bond debt were not part of the analysis undertaken here, which instead focused on the sources of financing emphasized by several case studies and considerations of which types of financing were most likely to be present during conflict. More comprehensive analyses might include bonds and private lending, for example.

    Conclusion

    The above findings, tentative though they are, accentuate several important matters for academics and policymakers to consider. The first is how states meet the obligations created by the increase in external debt stocks during civil conflict. We have seen here that conflict presence and intensity lead to an increase in external debt stocks, and I argue that this can be attributed to leaders seeking financing in order to perpetuate their tenure in office. However, external debt is certainly just one facet of the fiscal picture for states during conflict. The impact of inflationary pressures, whether caused by state policy or conflict itself, remains somewhat unclear. The role of expropriations and local currency debt is also an open question.

    In summary, while the literature currently understands reasonably well how rebel groups finance their operations and perpetuate their activities, the state’s economic options are more complicated and less well understood. Hence, as scholars and policymakers seek to understand how civil conflict is perpetuated, the ability of the state to sustain conflict must be more fully understood. Otherwise, policies such as economic sanctions may fail to bring the state to the bargaining table, should that be the aim. See Rogers (1996) for one example of the literature on sanctions and internal conflict management. The importance of bilateral lending to this model may imply that geopolitical concerns provide a rationale for the financing to sustain and aid states facing civil conflict, but such conclusions are preliminary. The possibility of continued financing from foreign sources may even keep the state from entering into negotiations, or encourage bellicose behavior towards marginalized internal groups. This draws further into focus the importance and motivations of the suppliers of credit to states in conflict.

    The pursuit of political survival by state leaders during conflict almost certainly leads to the implementation of economic policies that are unsustainable in the long term. As noted earlier, conflict creates acute difficulties for external debt service, but the evidence of this analysis suggests leaders are heedless of such complications in their pursuit of financing. It may be that states at peace are able to take on greater external obligations than those that are embroiled in civil conflict, but peace ensures a greater capacity to service such debt. By understanding better the means of financing and fiscal policies of states in the midst of civil conflict, we can move beyond simple correlations between conflict and debt accumulation and clarify the role of external debt in the perpetuation of civil conflict.

    References

    • Addison, T. & Murshed, S.M.  (2003). Debt relief and civil war. Journal of Peace Research, 40(2), 159-176. Retrieved from http://www.jstor.org/stable/3648409
    • Alvarez-Plata, P. & Brück. T. (2008). External debt in post-conflict countries. World Development, 36(3), 485-504. doi: 10.1016/j.worlddev.2007.03.002
    • Anderton, C.H. (2000). An insecure economy under ratio and logistic conflict technologies. Journal of Conflict Resolution, 44(6), 823-838. Retrieved from www.jstor.org/stable/174592
    • Anderton, C.H., & Carter, J.R.  (2000). On disruption of trade by war: A reply to Barieri and Levy. Journal of Peace Research, 38(5), 625-628.
    • Aning, K. & Atuobi, S. (2011). The neglected economic dimensions of ECOWAS’s negotiated peace accords in West Africa. Africa Spectrum, 46(3), 27-44.
    • Azam, J.P. (2001). The redistributive state and conflicts in Africa. Journal of Peace Research, 38(4), 429-444. Retrieved from http://www.jstor.org/stable/424895
    • Beers, D.T. & Chambers, J. (2007). Default study: Sovereign defaults at 26-year low, to show little change in 2007. S & P Credit Research. September 18, 2006.
    • Bridgman, B. (2007). Why are ethnically divided countries poor? Journal of Macroeconomics, 30, 1-18. doi: 10.1016/j.jmacro.2006.08.007
    • Brück, T. (2001). Mozambique: The economic effects of the war. In Valpy Fitzgerald and Frances Stewart (eds.), War and Underdevelopment Vol 2. Oxford: Oxford University Press.
    • Bueno de Mesquita, B., Smith, A., Siverson, R.M., & Morrow, J.D. (2003). The Logic of Political Survival. Cambridge: The MIT Press.
    • Bulow, J. & Rogoff, K. (1989a). A constant recontracting model of sovereign debt. Journal of Political Economy, 97(1), 155-178. Retrieved from http://www.jstor.org/stable/1831058
    • Bulow, J., & Rogoff, K. (1989b). Sovereign debt: Is to forgive to forget? American Economic Review, 79(1), 43-50. Retrieved from http://www.jstor.org/stable/1804772
    • Chambers, J.B. (2012). Default study: Sovereign defaults and rating transition data, 2012 update. S&P Credit Research. Retrieved from http://www.standardandpoors.com/ratings/articles/en/U.S./?articleType=HTML&assetID=1245350156739#ContactInfo
    • Chung, J. & Mahtani. D. (2007, February 2). Nigeria’s warrants problem. Financial Times. Retrieved from http://www.ft.com/intl/cms/s/0/479a97bc-b262-11db-a79f-0000779e2340.html#axzz35r7KGWZE
    • Collier, P. (1999). On the economic consequences of civil war. Oxford Economic Papers, 51(1), 168-183. Retrieved from http://www.jstor.org/stable/3488597
    • Collier, P. & Hoeffler, A. (1998). On economic causes of civil war. Oxford Economic Papers, 50(4), 563-573. Retrieved from http://www.jstor.org/stable/3488674
    • Collier, P. & Hoeffler, A. (2004). Greed and grievance in civil war. Oxford Economic Papers, 56(4), 563-595.
    • Collier, P. & Hoeffler, A. (2007). Unintended consequences: Does aid promote arms races? Oxford Bulletin of Economics and Statistics, 69(1), 1-27. doi: 10.1111/j.1468-0084.2006.00439.x
    • Collier, P. & Gunning, J.W.  (1995). War, peace and private portfolios. World Development, 23(2), 233-241. doi: 10.1016/0305-750X(94)00126-J
    • Collier, P., Elliot, V.L., Hegre, H., Hoeffler, A., Reynal-Querol, M.  & Sambanis, N. (2003). Breaking the Conflict Trap: Civil War & Development Policy. Washington, D.C.: The IBRD and World Bank.
    • Elbadawi, I. & Sambanis, N. (2002). How much war will we see? Explaining the prevalence of civil war. The Journal of Conflict Resolution, 46(3), 307-334. Retrieved from http://www.jstor.org/stable/3176229
    • Easterly, W. (2002). How did heavily indebted poor countries become heavily indebted? Reviewing two decades of debt relief. World Development, 30(10), 1677-1696. doi: 10.1016/S0305-750X(02)00073-6
    • Easterly, W. & Levine, R. (1997). Africa’s growth tragedy: Policies and ethnic divisions. The Quarterly Journal of Economics, 112(4), 1203-1250. Retrieved from http://www.jstor.org/stable/2951270
    • Eichengreen, B. (1991). Historical Research on International Lending and Debt. Journal of Economic Perspectives, 5(2), 149-169. Retrieved from http://search.proquest.com/docview/208983025?accountid=7113
    • Fearon, J.D. (2006). Ethnic mobilization and ethnic violence. In Barry R. Weingast and Donald A. Wittman (eds.), The Oxford Handbook of Political Economy, pp. 852-868. Oxford: Oxford University Press.
    • Fearon, J.D. & Laitin, D. (2003). Ethnicity, insurgency, and civil war. American Political Science Review, 97(1), 75-90. Retrieved from http://search.proquest.com/docview/60646450?accountid=7113
    • Fitzgerald, V. (2001a). Paying for the war: Economic policy in poor countries under conflict conditions.  In Frances Stewart and Valpy Fitzgerald (eds.), War and Underdevelopment, pp. 21-28. Oxford: Oxford University Press.
    • Fitzgerald, V. (2001b). The international political economy of conflict in poor countries. In Frances Stewart and Valpy Fitzgerald (eds.), War and Underdevelopment, pp 204-224. Oxford: Oxford University Press.
    • Fitzgerald, V. and A. Grigsby. (2001). Nicaragua: The political economy of social reform and armed conflict. In Frances Stewart and Valpy Fitzgerald (eds.), War and Underdevelopment, pp. 204-224. Oxford: Oxford University Press.
    • Fjelde, H. & De Soysa, I. (2009). Coercion, co-optation, or cooperation? State capacity and the risk of civil war, 1961-2004. Conflict Management and Peace Science, 26(1), 5-25. doi: 10.1177/0738894208097664
    • Frynas, J.D. & Wood, G. (2001). Oil and war in Angola. Review of African Political Economy, 28(90), 587-606. Retrieved from http://www.jstor.org/stable/4006839
    • Gleditsch, K.S. (2002). All International Politics is Local. Ann Arbor: University of Michigan Press.
    • Gleditsch, K.S. (2007). Transnational dimensions of civil war. Journal of Peace Research, 44(3), 293-309. Retrieved from http://www.jstor.org/stable/27640512
    • Gleditsch, K.S. & Ruggeri, A. (2010). Political opportunity structures, democracy, and civil war. Journal of Peace Research, 46(3), 299-310.
    • Ghobarah, H.A., Huth, P., & Russett, B. (2003). Civil wars kill and maim people – long after the shooting stops. American Political Science Review, 97(2), 189-202. doi: http://dx.doi.org/10.1017.S0003055403000613
    • Gurr, T.R. (1970). Why Men Rebel. Princeton: Princeton University Press.
    • Gurr, T.R. (1994). People against states: Ethnopolitical conflict and the changing world system. International Studies Quarterly, 38(3), 347-377. Retrieved from http://www.jstor.org/stable/2600737
    • Hartzell, C., Hoddie, M., & Bauer, M. (2010). Economic liberalization via IMF structural adjustment: Sowing the seeds of civil conflict? International Organization, 64(2), 339-356. doi: http://dx.doi.org/10.1017/S0020818310000068
    • Hatchondo, J.C., & Martinez, L. (2010). The politics of sovereign default. Economics Quarterly, 96(3).
    • Hirshleifer, J. (2000). The macrotechnology of conflict. Journal of Conflict Resolution, 44(6), 773-792. Retrieved from http://search.proquest.com/docview/224563802?accountid=7113
    • Hirshleifer, J. (2001). The Dark Side of the Force: Economic Foundations of Conflict Theory. Cambridge: Cambridge University Press.
    • Johnson, T.H., Slater, R.O., & McGowan, P. (1984). Explaining African coups d’etat,1960 1982. American Political Science Review, 78(3), 622-640.
    • Kang, S. & J. Meernik. (2005). Civil war destruction and the prospects for economic growth. Journal of Politics, 67(1), 88-109. Retrieved from http://www.jstor.org/stable/3449706
    • Keen, D. (2001a). The political economy of war. In Frances Stewart and Valpy Fitzgerald (eds.), War and Underdevelopment, pp. 39-66. Oxford: Oxford University Press.
    • Keen, D. (2001b). Sierra Leone: War and its functions. In Frances Stewart and Valpy Fitzgerald (eds.), War and Underdevelopment, pp. 155-175. Oxford: Oxford University Press.
    • Marsden, P. & Samman, E. (2001). Afghanistan: The economic and social impact of conflict. In Valpy Fitzgerald and Frances Stewart (eds.), War and Underdevelopment Vol 2. Oxford: Oxford University Press.
    • Miguel, E., Satyanath, S. & Sergenti, E. (2004). “Economic Shocks and Civil Conflict: An Instrumental Variables Approach.” Journal of Political Economy 112(4): 725-753.
    • Mitchener, K.J. & Weidenmier, M. (2005). Empire, public goods, and the Roosevelt Corollary. The Journal of Economic History, 65(3), 658-692. doi: http://dx.doi.org/10.1017/S0022050705000240
    • Murdoch, J.C., & Sandler, T. (2002). Economic growth, civil wars, and spatial spillovers. The Journal of Conflict Resolution, 46(1), 91-110. Retrieved from http://search.proquest.com/docview/224562200?accountid=7113
    • Øsby, G. (2008). Polarization, horizontal inequalities, and violent civil conflict. Journal of Peace Research, 45(2), 143-162. Retrieved from http://www.jstor.org/stable/27640647
    • Panizza, U., Sturzenegger, F., & Zettelmeyer, J. (2009). The economics and law of sovereign debt and default. Journal of Economic Literature, 47(3), 651-698.
    • Reinhart, C.M., & Rogoff, K.S.. (2008). The forgotten history of domestic debt. National Bureau of Economic Research. doi: 10.3386/w13946
    • Reinhart, C.M. & Rogoff, K.S..  (2009). This Time Is Different: Eight Centuries of Financial Folly. Princeton: Princeton University Press.
    • Rogers, E.S. (1996). Economic sanctions and internal conflict. In Michael E. Brown (ed.), The International Dimensions of Internal Conflict pp. 411-434. Cambridge: MIT University Press.
    • Rosenau, J.N. (1964). International Aspects of Civil Strife. Princeton: Princeton University Press.
    • Ross, M.L. (2004). How do natural resources influence civil war? Evidence from thirteen cases. International Organization, 58, 35-67. Retrieved from http://www.jstor.org/stable/3877888
    • Saiegh, S. (2005). Do countries have a “democratic advantage?” Political institutions, multilateral agencies, and sovereign borrowing. Comparative Political Studies, 38, 366-387. doi: 10.1177/0010414004273204
    • Saiegh, S. (2009). Coalition governments and sovereign debt crises. Economics and Politics, 21(2), 232-254. doi: 10.1111/j.1468-0343.2009.00344.x
    • Sarkees, M.R., Wayman, F.Wm  Singer, & J.D. (2003). Inter-state, intra-state, and extra-state wars: A comprehensive look at their distribution over time, 1816-1997. International Studies Quarterly, 47(1), 49-70. Retrieved from http://www.jstor.org/stable/3096076
    • Schultz, K.A. & Weingast, B.R.. (2003). The democratic advantage: Institutional foundations of financial power in international competition. International Organization, 57(1), 3-42. doi: http://dx.doi.org/10.1017/S0020818303571065
    • Sen, A. (1994). Economic regress: Concepts and features. Washington, D.C.: The World Bank.
    • Sen, A. (1992). Wars and famines. In W. Isard and C.H. Anderton (eds.), Economics of Arms Reduction and the Peace Process. Amsterdam: North-Holland.
    • Spalding, R.J. (1987). The Political Economy of Revolutionary Nicaragua. Boston: Allen and Ulwin.
    • Stewart, F. & V. Fitzgerald. (2001). War and Underdevelopment. Oxford: Oxford University Press.
    • Stewart, F., C. Huang, & M. Wang. (2001). Internal wars in developing countries: An empirical overview of the economic and social consequences. In Frances Stewart and Valpy Fitzgerald (eds.), War and Underdevelopment, pp. 67-103. Oxford: Oxford University Press.
    • Themner, Lotta &Wallensteen, Peter. (2014). Armed conflicts, 1946–2013. 2014. Journal of Peace Research 51(4): 541-554. doi: 10.1177/0022343314542076
    • Tomz, M. (2003). Domestic attitudes toward debt repayment: Public opinion and economic sophistication in Argentina. (Unpublished). Retrieved from http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CB8QFjAA&url=http%3A%2F%2Fweb.stanford.edu%2F~tomz%2Fworking%2Fargsvy-old.pdf&ei=gEnhU9mtKcKe8gGT8oCIDg&U.S.g=AFQjCNGyIMkAN7mwgw6UX1gKduGQngoxFw&sig2=GKFCzZBlM0OcoDmyzF3hLA&bvm=bv.72197243,d.b2U
    • Tomz, M. (2004). Interests, information, and the domestic politics of international agreements. (Unpublished). Retrieved from http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CB8QFjAA&url=http%3A%2F%2Fweb.stanford.edu%2F~tomz%2Fworking%2FTomz2004a.pdf&ei=2EjhU9eTMOKj8gGm9IDwAw&U.S.g=AFQjCNF2q5b4eaX1d2KW3e1GQ9aVXqspiA&sig2=FTJMI1Z2w2RJLWB9ZBLo5g&bvm=bv.72197243,d.b2U&cad=rja
    • Tomz, M. (2007). Reputation and International Cooperation: Sovereign Debt Across Three Centuries. Princeton: Princeton University Press.
    • Tomz, M. & Wright, M.J.. (2013). Empirical research on sovereign debt and default. Annual Review of Economics, 3, 1-36. Retrieved from http://www.chicagofed.org/webpages/publications/working_papers/2012/wp_06.cfm
    • Tomz, M. & Wright, M. (2010). Sovereign theft: theory and evidence about sovereign default and expropriation. In William Hogan and Federico Sturzenegger (eds.), The Natural Resources Trap: Private Investment without Public Commitment, pp. 69–110. Cambridge: MIT Press.
    • Van Rijckeghem, C. & Weder, B.. (2009). Political institutions and debt crises. Public Choice, 138 (March), 387–408.
    • Wennmann, A. (2011). Economic issues in peace negotiations. In The Political Economy of Peacemaking.  Routledge: London, pp. 34-51.
    • Wimmer, A., Cederman, L.  & Min, B.  (2009). Ethnic politics and armed conflict. A configurational analysis of a new global dataset. American Sociological Review, 74(2), 316-337.
    • World Bank. (2014, June 20). World development indicators. World Bank DataBank. Retrieved from http://databank.worldbank.org
    • Wright, M.L.J. (2005). Coordinating creditors. American Economic Review, 95(2), 388-392. Retrieved from http://www.jstor.org/stable/4132853
    • Wright, M.L.J. (2011). The theory of sovereign debt and default. Annual Review of Economics, 3. doi: 10.1146/annurev‐economics‐061109‐080443.

    Table 1: Effects of Civil Conflict on External Debt Stocks, Bilateral Lending, Development Assistance, and Multilateral Funding

    Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

    Table 2: Effect of Conflict Intensity on External Debt Stocks, Bilateral Debt, and Official Development Assistance; and Effect of Conflict Intensity on Multilateral Financing and Debt Service

    Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

    Table 3: Post Conflict Effects on  External Debt Stocks, Bilateral Financing, Multilateral Financing, and Official Development Assistance

    Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

    Table 4: Effects of Civil Conflict and Conflict Intensity on Odds of Bond, Bank, and Local Currency Default under the Presence of Civil Conflict and Greater Conflict Intensity

    Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

    Table 5: Robustness Check: Effects of Civil Conflict Presence on Debt Using the Ethnic Power Relations Data Set

    Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

    Table 6: Robustness Check: Effects of Civil Conflict Intensity on Debt Using the Ethnic Power Relations Data Set

    Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1