Does Foreign Development Aid Trigger Ethnic War in Developing States?

Abstract

Can foreign aid trigger ethnic war? The quantitative conflict literature has produced mixed findings on the effect of foreign aid on civil war in developing states. One reason for the mixed results is that a subset of civil wars, ethnic wars, are more likely than other kinds of civil wars to be triggered by foreign aid. This is because large amounts of foreign aid can cause the state to become a prize worth fighting over, mobilizing ethnic identity and group-related rebellion. This article investigates this question by testing the separate impacts of total, bilateral, and multilateral aid given by state and nonstate actors on the onset of ethnic war, using a cross-national time-series dataset of 147 countries from 1961 to 2008. The findings show a very strong association of foreign aid with ethnic war, whether measured as total aid, bilateral aid, or multilateral aid.

Keywords

conflict resolution international relations minority issues origins/forms of conflict political science

Many developing states receive foreign development aid from donor states, organizations, or agents with the objective of improving their social and economic problems. Foreign aid is generally given from bilateral and multilateral sources and the conditions of aid differ according to the arrangements between donors and aid recipient states. Does foreign development aid help developing countries or lead to instability and war? A fair number of quantitative studies have looked at the impact of foreign aid on civil war in developing states, using different indicators of these variables. The findings in this literature are mixed and no study has tested the effect of different types of aid on ethnic wars quantitatively, specifically identifying whether a conflict is, at its core, an ethnically motivated war.

The objective of this article is to investigate the effects of foreign development aid from bilateral and multilateral sources on the onset of ethnic wars in aid recipient countries. Ethnic identity-based low-scale wars are common in developing countries. Foreign aid often makes up a significant portion of a developing state’s revenue and the allocation of foreign aid can influence the distribution of rents among different groups, increasing competition and conflict among them (Dimico, 2013; Esman, 1997; Ruggeri & Schudel, 2010).

While the impact of foreign aid on civil conflict in states has been extensively explored by many scholars, the majority of these studies are qualitative and many focus on humanitarian aid. Those that have analyzed the relationship between civil war and foreign development aid quantitatively have concentrated usually on regional level investigations, particularly in sub-Saharan Africa and parts of Asia (De Ree & Nillesen, 2009; Strange et al., 2017; Wood & Molfino, 2016; Wood & Sullivan, 2015). The few studies that have examined the impact of foreign aid on the onset of civil war quantitatively have attained mixed results (see, e.g., Collier & Hoeffler, 2002; Gutting & Steinwand, 2017; Nielsen et al., 2011; Savun & Tirone, 2011, 2012). There are no large-N quantitative studies that have tested the effects of total, bilateral, and multilateral foreign aid (excluding military aid) on the onset of ethnic war.

This study aims to improve our overall understanding of how the external economic force of development aid may impact ethnic war conditions in developing states. Relying on the Ethnic Power Relations (EPR) dataset, ethnic war is defined as low-scale and ethnically motivated war with an annual battlefield-connected death threshold of 25 people. External aid variables are separated as total, bilateral, and multilateral, to assess whether aid given from different sources has unique effects on the onset of ethnic war.

Total aid is aggregate level aid given by bilateral, multilateral, and non-Organization for Economic Cooperation and Development (OECD) states to developing states. Bilateral aid refers to state-to-state aid, usually flowing from a developed donor state to a developing state. Since France, Germany, and the United States appear as the largest donor states in the years examined here, this study expects that their aid can influence the rent-seeking politics among ethnic groups in developing states. Hence, I opted to examine the individual effects of each of these states’ bilateral aid to their recipient states. Multilateral aid is the aid given by international organizations and nonstate actors such as the World Bank and the United Nations. The analyses of the separate effects of total, bilateral, and multilateral aid can provide insight into whether aid from single donor states or by nonstate organizations have different impacts on ethnic war in developing states. The data for the aid variables are taken from the OECD’s official development aid sources. These variables are tested, using a cross-national time-series dataset of 147 countries from 1961 to 2008. The data and measures for all these variables are explained further in the Analysis section.

This study is organized as follows. First, the opposing arguments in the literature on foreign aid and internal war are reviewed. Subsequently, the main argument and expectations of this study on the relationship between foreign aid and the onset of ethnic war is explained with a brief discussion of Iraq and Sri Lanka cases. Following that, the research design including the measures and data on the main variables is described, and the regression analyses are presented. Then, the implications of this study for further research are discussed in conclusion.

Foreign Aid and Conflict

Much scholarly attention has been given to the effect of foreign aid on civil conflict with scholars reaching ambiguous and, at times, competing findings. Within the literature, scholars have advanced conflicting theoretical explanations which can largely be divided into two groups. The first group of scholars suggest that foreign aid can mitigate the likelihood of civil war as aid can strengthen a state’s government, institutions, and economy. The second group assert that aid can have a corrupting effect on the state, increasing inequality and grievances, and even making the control of the state more valuable and thus more attractive to challengers. In short, the competing explanations largely claim that foreign aid either insulates a state from conflict or exacerbates underlying issues within a state making conflict more likely. The literature with these competing views is summarized below.

The Pacifying Effect of Aid on Internal War

Studies envisioning foreign aid as a force reducing the risk of conflict in recipient states draw attention to the indirect benefits of aid on conflict by improving the economy. Collier and Hoeffler (2002) argue that foreign aid can be beneficial in stabilizing countries by reducing the recipient state’s dependence on primary commodity exports and improving economic conditions. Other scholars agree that foreign aid can indirectly reduce conflict by spurring economic growth. The creation of employment opportunities and more even income distribution would give less incentive for individuals to join rebel groups and participate in civil conflict (Burnside & Dollar, 2000; Collier & Hoeffler, 2002; Dimico, 2013; Hansen & Tarp, 2001).

Burnside and Dollar (2000) and Collier and Dollar (2002) find that foreign aid has a positive effect on economic growth, but this effect is confined to a set of states with good institutions and policies, which can mitigate incentives to engage in civil war. They argue that foreign aid can be beneficial in many ways if it is directed toward strengthening institutions. Even if the aid revenues are used for consumption and not for improving governance, they can still enable governments to allocate more resources to security, which reduces the risk of attack by rebels (see also Ezcurra & Manotas, 2013; Gurr, 2000; Gurr & Moore, 1997).

Savun and Tirone (2011) also suggest that aid can help states avoid civil conflict under certain conditions. Specifically, their study of development aid eligible countries from 1990 to 2003 shows that democratizing states are quite vulnerable to civil war, but when they receive high levels of democracy aid, they are less likely to experience civil war onset than those which lack such support. Nielson et al. (2011) specifically investigate how aid shocks can affect the likelihood of violent armed conflict in recipient states between 1981 and 2005, utilizing AidData, a research lab developed at William and Mary University in Virginia, which reports development aid commitments to recipient states (Tierney et al., 2011). With these data, Nielson et al. (2011) show that negative aid shocks (sudden drops) significantly increase the likelihood of civil war onset while positive aid shocks (sudden increases) have an inconclusive effect on the probability of war onset. Examining foreign aid recipient states for the period 1990–2004 from OECD sources, which is a more commonly used aid data set since it reports actual disbursements rather than commitments, Savun and Tirone (2012) revealed similar results to Nielson et al.’s (2011), finding that aid was likely to decrease the probability of the onset of large-scale civil war (1,000 battle death threshold).

Gutting and Steinward (2017) offer the insight that donor fragmentation can help overcome the severe consequences of negative aid shocks. If a state receives a substantial proportion of aid from only one donor (total concentration), this reliance makes them susceptible to the volatility that may result from a negative aid shock (i.e. Nielsen et al., 2011). However, states with more donors can, in effect, insulate themselves from the effects of negative aid shocks and the subsequent risk of civil conflict. In theory, more donors moderates the consequences of negative aid shocks, and Gutting and Steinward (2017) find support for their argument as recipient states with more donors are less likely to see negative aid shocks that may cripple them to the point of conflict.

How Aid May Increase the Likelihood of Internal War

The second group of scholars in the literature on foreign aid and conflict claim that aid may not be directed at policies for economic growth. Instead, money spent on building up institutions and improving policies can also fail in achieving desired outcomes. For example, focusing exclusively on sub-Saharan Africa, De Ree and Nillesen (2009) found that foreign aid did not impact the probability of conflict initiation within this region. Thus, foreign aid does not appear to insulate states from conflict. However, they show that increasing levels of foreign aid can assist in decreasing the duration of an ongoing civil war.

Some scholars contend that foreign aid can increase the likelihood of civil war because rebel groups and government forces struggling over power are willing to exploit development assistance for personal gain. According to these scholars, both governments and rebels are in a constant “game” over resources with each aiming to maximize the outcome in their favor (Arcand & Chauvet, 2001; Azam, 1995). Grossman (1992) and Gershenson and Grossman (2000), for example, offered that aid can lead to an increased probability of political instability as control over the government becomes a valuable payoff for opportunistic rebellions. In essence, foreign aid can result in rent-seeking behavior as each actor attempts to maximize their expected payoffs from engaging in conflict. In addition, Addison et al. (2001) argue that development assistance that fails to reduce poverty can actually aggravate and even intensify inequalities within society, thus increasing the likelihood of conflict and even inhibiting the opportunity for conflict resolution in ongoing wars. The greater the potential of territory or power to be won, the greater the incentive to engage in conflict.

Knack (2001) argued that higher aid levels actually served to undermine the quality of governance in recipient states. His cross-national investigation for the period 1982–1995 revealed that higher levels of development aid as a percentage of gross national product and total development aid as a percentage of government expenditures both decreased the quality of governance as captured by the International Country Risk Guide—an 18-point scale considering corruption, bureaucratic quality, and rule of law. Zürcher (2017), in his review of qualitative studies on various forms of aid and civil war, noted that there is limited evidence to suggest that aid can result in a reduction in violence, and aid in conflict zones only had a violence reducing effect in areas that had already been stabilized. Furthermore, among different types of aid, humanitarian aid never had a “violence-dampening effect.” On the contrary, violence actually increased when humanitarian aid was disbursed (Zürcher, 2017, p. 511).

It is difficult to exclude members of local militia groups from being direct recipients. Qualitative studies widely showed that humanitarian aid is likely to generate violence and promote civil conflict in developing countries (Anderson, 1999; De Waal, 1997; Narang, 2015; Nunn & Qian, 2014; Polman, 2010; Wood & Molfino, 2016; Wood & Sullivan, 2015). For example, Nunn and Qian (2014) found that U.S. food aid increases both the occurrence and duration of civil conflicts, mainly in states with a recent history of civil war. They highlight the ease with which armed factions and opposition groups acquire humanitarian aid that is transported through weakly controlled territories. They also found that even if the aid reached its intended recipients, it can still be confiscated by armed groups with little effort. Examining twenty sub-Saharan African countries after the Cold War era, Wood and Molfino (2016) demonstrated that areas where aid accumulated are likely to see increased frequency of violent engagements between governments and rebel forces, while nonhumanitarian aid showed no significant effect on the frequency of battles.

Several scholars in this literature drew attention to the fungibility of foreign aid and how the fungible development aid can also affect internal conflict. When aid is fungible, it can be redirected by recipients so as to be used for multiple policies such as to fund development projects in a different sector, to aid military expenditures, or may be hoarded by corrupt officials (Arcand & Chauvet, 2001; Kolstad, 2005; Morrissey, 2006). Burnside and Dollar (2000) found that bilateral aid was in particular fungible and exhibited a positive and significant impact on government spending across all sectors, not just where the donor intended for the aid to be spent.

Ruggeri and Schudel (2010) also argue that bilateral aid is most likely to be fungible, reflecting the interests of the donor state; therefore, it is likely to afford recipients more discretion when compared to multilateral aid. Multilateral aid limits the ability of a government to spend at their own discretion and rather comes with specific provisions and stipulations. Thus, multilateral aid is more likely to be appropriated for its specific intent and at times can even bypass the central government (Ruggeri & Schudel, 2010, p. 11). Their findings show that bilateral aid has no meaningful effect on conflict duration, and multilateral aid appears to be the only form of assistance that can reduce the duration of civil war in the world’s poorest countries. However, their study does not consider how aid might impact the initiation of civil conflict.

Foreign Aid and Ethnic Conflict

While it is clear that many scholars have quantitatively examined the impact of developmental aid on civil war, less attention has been given to examining how foreign aid might impact a specific form of intrastate conflict—ethnic war. Some studies have investigated the effect of foreign aid on economic development in ethnically divided states, finding that ethnic divisions matter. For example, Easterly’s (2001, 2003) work on varying effects of foreign aid in ethnically diverse societies suggests that more foreign aid can increase corruption, whereas more aid in ethnically homogenous societies increases social capital, resulting in faster growth and development. As mentioned, the objective of this study is to investigate whether foreign aid increases the probability of the onset of ethnic war. It is common that certain ethnic groups gain advantages over others by means of accessing states’ resources. Accordingly, larger amounts of foreign aid (especially as a percentage of gross domestic product [GDP]) may trigger battles among ethnic groups, considering that rent-seeking along ethnic lines is likely in ethnically divided states.

Esman and Herring (2001) noted that while developmental aid projects can be a positive occurrence for developing states, they can also usher in significant distributive consequences that can be exacerbated when ethnic dimensions enter the picture. In particular, they state that even if aid is administered with the best intentions, it still is given in a political context, which can spell significant problems in societies divided by ethnic power relations, territoriality, and patronage systems. In these environments, it is common to see inequity and discrimination in how aid is managed and disbursed (Esman & Herring, 2001, p. 3). While recipient governments can become “gatekeepers” of aid regardless of the ethnic dimensions within a state, the ramifications can be more severe when ethnic tensions within a state are already volatile.

Hence, development aid in ethnically divided societies may be disproportionately distributed, at times being used to maintain patronage networks instead of being used for its intended purposes. This often means that the politically dominant ethnic group reaps the benefits of developmental aid while those on the outskirts of this community are left empty-handed. In these cases, where the disbursements of foreign aid benefit only those in power or are systematically skewed in a way that ostracizes ethnic communities, it is likely that resentment and grievances will be intensified (Esman & Herring, 2001, p. 7).

Svensson (2000) contends that increases in government revenue through foreign aid is likely to be associated with increased rent-seeking in ethnically diverse societies. In these societies, foreign aid can both increase corruption at the government level and the grievances of social groups at the microlevel. Svensson further demonstrates that foreign aid is associated with higher levels of corruption, particularly in states with competing social groups. Covering 66 countries from 1980 to 1994, his study presented findings that aid can also increase the degree of instability in ethnically fractionalized states.

In summary, it appears that the motivations behind ethnic armed conflict may not be the same as those for all armed conflicts, and the impact of foreign aid in states ripe with ethnic tensions might trigger more volatility, increasing the risk of violence. As suggested here, foreign aid is more likely to affect group incentives to rebel, and this is especially likely for ethnically identified groups who are excluded from the benefits of foreign aid. Under these conditions, the prize worth fighting over—the state—is far more valuable. The argument of this study on the role of foreign aid in stimulating ethnic violence and war can be illustrated with cases such as Iraq and Sri Lanka. Below, I briefly discuss these cases and then propose the hypotheses to be tested. Following that, the research design and regression analyses are presented.

The Cases of Iraq and Sri Lanka

Iraq and Sri Lanka have been among the states that have received large amounts of foreign aid, and the latter is often linked to the onset of ethnically motivated wars in these countries. Since the U.S.-led war in 2003, Iraq has been one of the largest recipients of U.S. foreign aid, providing Iraq a wide variety of economic and development aid. For example, Iraq received around $29 billion from the United States in the post-2003 reconstruction years, and in 2010, still remained among the top aid recipients of U.S. aid, receiving more than $1 billion (Tarnoff & Lawson, 2011). As the reconstruction took place, these large amounts of foreign aid given to the Iraqi state could have been perceived by ethnically divided political leaders as a substantial source of income, intensifying their competition over state rents.

As is known, the 2003 war caused a shift in power in Iraq from the Sunni minority to the Shia majority, and an outbreak of sectarian tensions between them (another major ethnic group, the Kurds, gained autonomy). This tension escalated to an ethnic civil war, which was subdued only after the United States arranged for payments to Sunni insurgents and their leaders between 2007 and 2011, a program known as the “Sunni Awakening.” Indicative of the potential role of foreign aid in civil conflict, after the United States withdrew in 2011, the Shia-led government under Prime Minister Maliki chose not to follow promises to employ and pay the minority Sunnis. Tensions resumed, and in early 2013, tens of thousands of Sunnis participated in anti-government protests in Ramadi, Fallujah, Samarra, Mosul, and Kirkuk, accusing Maliki of exclusionary sectarian policies. The civil war resumed, with many Sunnis joining the extremist terrorist organization, the Islamic State in Iraq and Syria (Wilbanks & Karsh, 2010).

Sri Lanka is another case in point. There are two main ethnic groups: the Sinhalese majority and the Tamil minority. Upon independence from the British colonial rule in the 1940s, the Sinhalese majority retained control of the parliament, and politics became polarized based on ethnic identity. Sri Lanka (formerly Ceylon) had been receiving aid from the UN and donor states such as the United States in order to implement development projects under a multipurpose national program, called the Mahaweli Development Program. Several scholars argue that foreign aid given to Sri Lanka for their development purposes reinforced patronage linkages and corruption along ethnic lines and increased rent-seeking among ethnic politicians (Easterly, 2001; Mason, 2003).

In the 1970s, tensions among groups increased when the majority Sinhalese government aggressively sought development assistance for the Accelerated Mahaweli Project, designed to bring irrigation and power to dry parts of the country. However, the Sinhalese government planned and channeled aid into areas that put the Tamil populated lands at a disadvantage, increasing discontentment along ethnic cleavages. In the early stages of the project, one of the planned canals to bring water to Tamil areas was canceled, and between 1977 and 1982, some Tamil populated areas were outright excluded from foreign aid support. The minority group, the Tamils, believed that the government aimed at discriminating the Tamil population from the funded projects by excluding the dry areas where they were a majority. War broke out in 1983, and many scholars link it with political competition over foreign aid (Easterly, 2001; Mason, 2003; Nielsen, et al., 2011).

As these cases and the literature above suggest, this study expects a linkage of foreign aid with ethnic war, and quantitatively tests the following hypotheses. In order to distinguish the diverse effects of bilateral and multilateral aid, they will be tested separately in addition to the yearly total aid amounts given to recipient states.

Hypothesis 1: Higher levels of total foreign development aid increase the likelihood of ethnic war onset.

Hypothesis 2: Higher levels of bilateral aid from the three largest donors per capita (France, Germany, the United States) increase the likelihood of ethnic war onset.

Hypothesis 3: Multilateral aid increases the likelihood of ethnic war onset.

Research Design

To examine how the mentioned foreign aid variables affect the onset of ethnic war in states, I constructed cross-national time-series dataset with the country year as the unit of analysis. To gauge our dependent variable, I draw on the EPR dataset, which disaggregates the Uppsala/PRIO Armed Conflicts Dataset (ACD; N. P. Gleditsch et al., 2002). ACD defines armed conflict as any armed and organized confrontation between government troops and rebel organizations, or between army factions, that reaches an annual battlefield-connected death threshold of 25 people. Massacres, genocides, and communal riots are not included. To identify Ethnic War Onsets, the EPR dataset considers the aims of the armed organizations as well as their recruitment patterns and alliance structures. Wars are defined as ethnic wars when armed organizations recruit fighters predominantly from their own ethnic group or forge alliances on the basis of ethnic affiliation. Further, armed groups must explicitly be pursuing ethno-nationalist aims, motivations, and/or interests in order for a conflict to be categorized as an ethnic war.¹ The lower death threshold of 25 deaths is arguably more informative than the lead alternative of 1,000 deaths, as the latter threshold omits many lower intensity ethnic conflicts (Nielsen et al., 2011). Of the 94 armed conflict onset years in the ACD data in the sample period observed here, the EPR data set identifies 38 as Ethnic War Onset years. I omit ongoing (continuation) years with an ethnic war after the year that ethnic war started from the sample to distinguish the onset of the ethnic war from its duration, since the factors that affect the onset of ethnic wars may not be the same as those that affect its duration.

As mentioned, previous studies analyzed both aggregate and disaggregate measures of development aid, a majority of them relying on aggregate forms of aid (Dimico, 2013; Ruggeri & Schudel, 2010; Savun & Tirone, 2012; Subhayu et al., 2014). I analyze aggregate (total) aid as well, but to reflect the fungible aspect of the aid and to see separate effects of aid by multiple sources (i.e., nongovernmental organizations [NGOs]) and by one donor, I also disaggregate it into bilateral and multilateral aid.

I obtained data on total aid, bilateral aid, and multilateral aid from the OECD, which defines development aid as “government aid designed to promote the economic development and welfare of developing countries” (OECD, 2018). Development aid may be “provided bilaterally, from donor to recipient, or channeled through a multilateral development agency such as the UN or the World Bank. Aid includes grants, ‘soft’ loans (where the grant element is at least 25% of the total) and the provision of technical assistance” (OECD, 2018).

The OECD data have several advantages over alternatives. First, it is commonly been employed in quantitative large-N studies (see, e.g., Gutting & Steinwand, 2017; Savun & Tirone, 2011, 2012). Second, loans and credits for military purposes are excluded from the OECD data. My argument is that aid affects group incentives to rebel. While military aid can make a state a greater prize worth fighting over, it can also reduce the incentive to rebel since it can reduce the odds of rebel success; it also signals that a foreign patron supports the regime and thus may cut aid in the event of regime change. Third, the OECD data tabulate actual disbursements of aid only, excluding commitments of aid. It is disbursements that can affect the incentive to rebel, not commitments, since commitments can change with rebel success in capturing the state or otherwise altering the political landscape.

Aggregate Total Aid includes aid received from bilateral, multilateral, and non-OECD states. It includes bilateral aid from 23 donor states, as well as combined multilateral aid from IGOs and NGOs, and assistance from non-Development Assistance Committee (DAC; donors other than the members of the DAC of the OECD) states. Bilateral Aid is state-to-state aid. I examined the three largest sources of bilateral aid: France, Germany, and the United States. Over the time span of this study, these states contributed more than all others when gauged as per capita for recipient states, which is the figure that matters most given our expectation that higher amounts of such aid are more likely to impact a society compared with lower amounts of such aid. Measured per capita per recipients, the next highest donors, Japan and the United Kingdom, had far lower amounts of development aid and thus are less likely to have affected the risk of ethnic war.² Multilateral Aid is from 35 international organizations, including the World Bank, United Nations Children’s Fund, World Health Organization, and regional development banks.³ While the measure for bilateral aid provides insight on state-to-state assistance, multilateral aid provides the opportunity to explore the effects of development and welfare assistance from nonstate organizations. As far as I know, multilateral aid has not been tested in any of the mentioned studies.

Similar to previous studies (see, e.g., Nielsen et al., 2011), I examine the ratio of total official development aid to GDP as a measure for each of the three forms of aid. However, aid-to-GDP ratio captures the importance of aid to a recipient state and is not necessarily a reflection of how aid impacts the people in a state (Knack, 2001). My argument is not that aid affects state behavior but rather that it affects the incentives groups have in a country to control the state: The more the aid affects individual incomes, the more the control of a state receiving the aid becomes a prize worth fighting for. Aid-recipient states of similar incomes can vary in population size, so at identical aid-to-GDP ratios, larger populated states should be less affected by foreign aid than less-populated ones. I therefore, and in similar fashion to earlier scholars (i.e., Collier & Dollar, 2002; Gutting & Steinwand, 2017), examine the risk of armed ethnic conflict with consideration of foreign aid as a portion of income per person in an economy.

In addition, my argument implies a nonlinear relationship of aid with ethnic conflict: that very high levels of aid-to-GDP ratios per capita are far more likely to provoke group conflict than medium and lower levels of such aid. This was confirmed in preliminary tests (unreported), with the squared functional form more robust than the linear one. I thus report all analyses of all types of aid with the squared functional form.

Control Variables

Foreign aid is not the only factor affecting ethnic war in states, and there is risk of specification bias if we do not consider all known extraneous factors that could be associated with foreign aid. In the literature, these are democracy, income, oil exports, mountainous terrain, presence of another civil conflict, population size, and the extent to which a country is ethnically fragmented.

Income/Economic Development

Consistent with the literature, I also include a control variable for level of economic development (income). I use GDP per capita drawn from the World Bank. GDP per capita has been used to proxy for a diverse set of factors including level of development, weak states, and wealth distribution. Previous findings largely suggest that higher levels of GDP can reduce civil conflict as the quality of life in these countries improves with the level of development. Consequently, more development deters people from joining rebellions as the incentives to do so are minimized (i.e., Barbieri & Reuveny, 2005; De Ree & Nillesen, 2009; K. S. Gleditsch, 2007). To address issues of skewness, I take the natural logarithm of this variable and lag it for 1 year.

Democracy

Compatible with the studies in conflict literature, the variable of democracy is also controlled in this study. Levels of democracy is measured using the Polity IV dataset from the Center for Systematic Peace, this indicator ranges from −10 (the lowest level of democracy) to +10 (the highest level of democracy; Marshall & Jaggers, 2003). As demonstrated in previous literature, it is generally expected that increases in levels of democracy will reduce the grievances that make civil war onset more likely to occur. Many scholars agree that democratic states are believed to be more responsive to the people’s demands and less likely to experience civil war (Hegre et al., 2001). On the other hand, some authors found that freedom creates opportunities of association and movement that would make rebellion easier (Rajan & Subramanian, 2008).

Oil Production

As the conflict literature suggests, I control for oil production (lagged by 1 year) to measure the impact of oil production on ethnic war onset since these resources create competition and rent-seeking among groups, increasing the likelihood of conflict. The oil production per capita variable hails from the aforementioned (Collier & Hoeffler, 2004; Ross, 2004) dataset (Wimmer et al., 2009). Previous literature has found that the availability of natural resources affects civil conflict (Morelli & Rohner, 2015; Ross, 2004). It is expected that when rebels can obstruct the extraction of natural resources, as with oil, the likelihood of secessionist movements increases (Collier & Hoeffler, 2004; Ross, 2004). Oil reserves may create grievances over the distribution of revenue and become a target for rebels to gain power or use as funds to finance war, all of which should increase the likelihood of civil war (Barbieri & Reuveny, 2005). Buhaug (2006), on the other hand, argues that oil only matters in conflicts over an existing state because oil resources are usually controlled by the central government; this increases the incentives to capture a state.

Mountainous Terrain

This study also controls for the “mountainous terrain” variable, for which the data are taken from the EPR dataset as well. Mountainous terrain indicates the share of national areas that are mountainous, the assumption being that difficult terrains provide an advantage to rebels over government forces, raising the likelihood of civil war onset (De Ree & Nillesen, 2009; Fearon & Laitin, 2003; Hammond, 2017; Wimmer et al., 2009). In addition, similar to the studies in this literature, population size is introduced as a control variable.

Ongoing War

This variable controls for a situation whether a state is active in any kind of an ongoing civil war, including an internationalized war that started in the previous year to separate its effect on ethnic war from the other independent variables. First, scholars have previously suggested that states active in civil war vary in their likelihood of being aid recipients and vary in the types of aid they are likely to receive (Alesina & Weder, 2002; Blattman & Miguel, 2010; Savun & Tirone, 2012). Secondly, the existence of any war other than ethnic may conflate the relationship of our main independent variables with the ethnic war onset, since existence of another war may also lead to the initiation of an ethnic war. Thus, I include a lagged dichotomous measure as to whether or not a state is active in civil war to avoid conflating the main dependent variable (ethnic war onset). The data for this variable are taken from the armed conflict dataset provided by the PRIO.

Ethnic Fractionalization

The variable of ethnic fractionalization is also included to measure the level of ethnic heterogeneity in a country. Fractionalization index provided by the EPR dataset gives the likelihood of two randomly drawn people in a country belonging to different ethnolinguistic groups. This variable is supposed to capture the social and cultural cohesion of a country by expressing which state has linguistically assimilated their population over the course of time. Studies expected that the less cultural cohesion a country has, the more likely that country will experience civil war onset (Wimmer et al., 2009).

Results

To strengthen causal inferences, I lagged all independent variables that vary across time by 1 year. The result yields a cross-sectional time-series sample that includes 147 official development aid recipient states from 1961 to 2008. Since the dependent variable is binary, I estimate the risk of having an Ethnic War Onset in a given year with a maximum likelihood logistic function for rare event data, with time dependence controlled using Beck et al.’s (1998) cubic spline variables (appropriate for analyses time-series cross-sectional data with a binary dependent variable). All standard errors are clustered by country. Summary statistics can be seen in Appendix, and all data, and the Stata do-file constructing them, are available at https://sciences.ucf.edu/politics/person/demet-mousseau/.

Model 1 in Table 1 tests the base model of control variables only, without consideration of foreign aid. As can be seen, Income (−0.48) is negative and significant. This is consistent with most studies, which report increasing levels of income as among the most robust factors reducing the risk of civil conflict in states (e.g., Barbieri & Reuveny, 2005; Collier & Hoeffler, 2002, 2004; cf., Nielsen et al., 2011). It appears that Income (GDP per capita) also has a pacifying effect on the ethnic war onset. Also significant in expected directions are the coefficients for Ethnic Fractionalization (1.58) and Population (0.19), suggesting that countries with higher levels of ethnic divisions and populations are at higher risks for ethnic wars.

Table 1.

Economic Aid and the Onset of Ethnic War, 1961–2008.

Ethnic War Onset	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Total Aid _{t − 1 × 100}	—	0.57***	—	—	—	—
Total Aid _{t − 1 × 100}		3.60
French Aid _{t − 1 × 100}	—	—	3.55***	—	—	—
French Aid _{t − 1 × 100}			15.79
German Aid _{t − 1 × 100}	—	—	—	35.46***	—	—
German Aid _{t − 1 × 100}				4.81
USA Aid _{t − 1 × 100}	—	—	—	—	6.12*	—
USA Aid _{t − 1 × 100}					1.88
Multilateral Aid _{t − 1 × 100}	—	—	—	—	—	4.03***
Multilateral Aid _{t − 1 × 100}						2.9
Income_{GDP/person ln t − 1}	−0.48***	−0.48***	−0.48***	−0.48***	−0.48***	−0.48***
Income_{GDP/person ln t − 1}	−2.65	−2.61	−2.63	−2.64	−2.63	−2.61
Democracy _t _{− 1}	0.01	0.01	0.01	0.01	0.01	0.01
Democracy _t _{− 1}	0.57	0.53	0.59	0.53	0.54	0.54
Oil production_{/person ln t − 1}	0.42	0.41	0.37	0.42	0.42	0.43
Oil production_{/person ln t − 1}	1.51	1.47	1.34	1.53	1.52	1.55
Mountainous terrain	0.15	0.16	0.16	0.16	0.15	0.16
Mountainous terrain	1.43	1.47	1.50	1.46	1.42	1.48
Ethnic fractionalization	1.58***	1.59***	1.56**	1.57***	1.59***	1.58***
Ethnic fractionalization	2.62	2.63	2.58	2.59	2.62	2.62
Population_ln	0.19**	0.21**	0.20**	0.20**	0.20**	0.21**
Population_ln	2.15	2.27	2.22	2.23	2.19	2.31
Ongoing war	0.71*	0.71*	0.72*	0.71*	0.71	0.72*
Ongoing war	1.65	1.66	1.67	1.67	1.65	1.67
Constant	−3.21*	−3.41*	−3.29*	−3.31*	−3.30*	−3.47*
Constant	−1.82	−1.91	−1.87	−1.88	−1.84	−1.94
Pseudo R²	.10	.10	.10	.10	.10	.10

Note. N = 5,043. Rare event logistic coefficients with t values in second rows calculated with robust standard errors clustered by country. Peace years with three cubic splines included but not shown (Beck et al., 1998). GDP = gross domestic product; # events = 68; # clusters = 147.

***p < .01. **p < .05. *p < .10 (two-tailed tests).

All the remaining control variables are not significant predictors of ethnic war onset: democracy (0.01), oil production (0.42), mountainous terrain (0.15), and ongoing war (0.71). The insignificance of democracy is not surprising, as most studies have yielded little support for it as a factor of civil conflict (see, e.g., Fearon & Laitin, 2003; Gutting & Steinwand, 2017; Hegre et al., 2001). The insignificance of the other factors may be more surprising in light of some prior studies, but the causal processes of ethnic war are not the same as for all civil wars. It could be that the war-inducing effect of oil exports as noted in prior studies (Morelli & Rohner, 2015; Ross, 2004) is moderated in the context of ethnic groups, since ethnicity can serve as a clear channel through which a state can distribute oil revenues in order to buy off ethnic group support. Similarly, mountainous terrain would seem relevant for ethnic war only where it correlates with ethnic dwelling. Regarding ongoing war, it could be that states facing rebellion may make extra efforts to accommodate ethnic groups, to avoid having to fight multiple wars at once.

Model 2 reports the results for total official development aid. The coefficient for total aid (0.57) is positive and highly significant. This corroborates Hypothesis 1: The greater the foreign aid as a portion of income per person in an economy, the greater the risk of ethnic war onset. This result is highly robust, yielding a t value greater than 3, even with consideration of all the control variables. This is an important finding, as it supports the expectation that higher levels of foreign aid can reinforce perhaps already existing rent-seeking structures and/or grievances along ethnic lines, as some groups gain more than others from the foreign development aid.

Models 3–6 disaggregate total aid into its bilateral and multilateral components. Models 3–5 examine the individual effects of bilateral aid from the three largest donors: France, Germany, and the United States. As can be seen, all three coefficients are positive and significant: French Aid (3.55), German Aid (35.46), and USA Aid (6.12), corroborating Hypothesis 2. It seems that with regard to the onset of ethnic war, the source of aid makes little difference. This again is consistent with theoretical expectations: The war-inducing effect of aid appears to be about conditions in recipient states, not about peculiarities of donor conditionalities and procedures. This is confirmed again with Model 6, where multilateral aid (4.03) too appears to have a war-inducing effect.

Overall, regardless of the specification of different types of foreign aid, it seems that developmental aid of any kind generally seems to increase the risk of ethnic war, whether it is gauged in the aggregate form or unpacked into its bilateral or multilateral components. This is consistent with the idea that foreign aid of any kind has the effect of making the state receiving the aid more of a prize worth fighting for, increasing the risk of identity-based armed conflict in a state.

Conclusion

This study investigated the impact of various forms of foreign aid on the onset of ethnic war. Prior scholarship has examined how various forms of foreign aid can affect civil war and its dynamics more generally, but scant attention has been given to the onset of ethnic war. Ethnic conflicts are more likely than others to be exacerbated over perceptions of inequitable treatment by the state, perceptions of unfairness, and access to state rents, so foreign aid may trigger unbalances among ethnic groups in a state, increasing the risk of ethnic war.

Examination of 147 countries between 1961 and 2008 yielded robust support for all expectations, as higher levels of total foreign development aid appear to increase the likelihood of ethnic war onset. Disaggregation of development aid into bilateral and multilateral components makes no difference in results, as all examined sources of aid appear to significantly impact the odds of ethnic war in a state. This lends some degree of support to arguments which suggest that foreign aid can exacerbate tensions in ethnically divided societies, leading those in marginalized populations to rebel. These findings are also coherent with how foreign aid plays a role in stimulating ethnic violence in Iraq and Sri Lanka, the cases briefly discussed here as examples. Further studies may pay attention to these and other cases for more detail analyses of this question.

While foreign aid is inherently designed to improve stability within a state, this study shows that its disbursement is not a guarantor of peace and rather may likely be a culprit of conflict. Thus, policy makers would do well to consider a state’s demographics, internal composition, and relations among ethnic groups before committing to deploying development aid. Further, when designing aid packages, donors should follow the advice of Esman and Herring (2001) and focus on the distributive inequalities that might undercut the intention of developmental aid. Consequently, this study’s findings on foreign aid offer a fairly pessimistic outlook for ethnic conflict onset, thus invites donors to pay attention to ethnic differences and their domestic politics while formulating aid packages to developing countries.

Footnotes

Appendix

Table A.

Summary Statistics.

Variable	Mean	SE	Min	Max
Total Aid _t _{− 1}	0.05	0.22	0.00	9.34
French Aid _t _{− 1}	0.01	0.12	0.00	7.68
German Aid _t _{− 1}	0.00	0.02	0.00	0.80
USA Aid _t _{− 1}	0.01	0.06	0.00	3.58
Multilateral Aid _t _{− 1}	0.01	0.03	0.00	1.09
Income_{GDP/person ln t − 1}	8.23	1.29	5.08	11.10
Democracy _t _{− 1}	0.84	7.55	−10.00	10.00
Oil production_{/person ln t − 1}	0.33	0.72	0.00	4.23
Mountainous terrain	2.09	1.43	0.00	4.42
Ongoing war	0.05	0.22	0.00	1.00
Population_ln	9.15	1.34	6.05	14.09
Ethnic fractionalization	0.39	0.29	0.00	0.93

Note. N = 5,043. GDP = gross domestic product.

Acknowledgment

The author is grateful for the thoughtful and constructive comments of the anonymous reviewers.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Demet Yalcin Mousseau

Notes

References

Addison

Le Billon

Murshed

(2001). Finance in conflict and reconstruction. Journal of International Development, 13, 951–964.

Alesina

Weder

(2002). Do corrupt governments receive less foreign aid? American Economic Review, 92, 1126–1137.

Anderson

(1999). Do no harm: How aid can support peace—or war. Lynne Rienner.

Arcand

Chauvet

(2001). Foreign aid, rent-seeking behavior, and civil war. CERDI-CNRS, Université d’Auvergne.

Azam

(1995). How to pay for the peace? A theoretical framework with references to African countries. Public Choice, 83, 173–184.

Barbieri

Reuveny

(2005). Economic globalization and civil war. Journal of Politics, 67, 1228–1248.

Beck

Katz

Tucker

(1998). Taking time seriously in binary time-series-cross-section analysis. American Journal of Political Science, 42, 1260–1288.

Blattman

Mguel

(2010). Civil war. Journal of Economic Literature, 48, 3–57.

Buhaug

(2006). Relative capability and rebel objective in civil war. Journal of Peace Research, 43, 691–708.

10.

Burnside

Dollar

(2000). Aid, policies and growth. American Economic Review, 90, 847–868.

11.

Collier

Dollar

(2002). Aid allocation and poverty reduction. European Economic Review, 46, 1475–1500.

12.

Collier

Hoeffler

(2002). Aid, policy, and peace: Reducing the risks of civil conflict. Defense and Peace Economics, 13, 435–450.

13.

Collier

Hoeffler

(2004). Greed and grievance in civil war. Oxford Economic Papers, 56, 563–595.

14.

De Ree

Nillesen

(2009). Aiding violence or peace? The impact of foreign aid on the risk of civil conflict in sub-Saharan Africa. Journal of Development Economics, 88, 301–313.

15.

De Waal

(1997). Famine crimes: Politics and the disaster relief industry in Africa. Indiana University Press.

16.

Dimico

(2013). The evolution of conflict and effectiveness of aid. Munich Personal RePEc Archive No. 47050. Munich, Germany.

17.

Easterly

(2001). Can institutions resolve ethnic conflict? Economic Development and Cultural Change, 49, 687–706.

18.

Easterly

(2003). Can foreign aid buy growth? Journal of Economic Perspectives, 17, 23–48.

19.

Esman

(1997). Can foreign aid moderate ethnic conflict? Peaceworks (Vol. 31), US Institute of Peace.

20.

Esman

Herring

(2001). Carrots, sticks, and ethnic conflict: Rethinking development assistance. University of Michigan Press.

21.

Ezcurra

Manotas

(2013, January). Globalization and civil war: An empirical analysis. (Working paper). Department of Economics, Universidad P´ublica de Navarra.

22.

Fearon

Laitin

(2003). Ethnicity, insurgency, and civil war. American Political Science Review, 97, 75–90.

23.

Gershenson

Gossman

(2000). Civil conflict: Ended or never ending? Journal of Conflict Resolution, 44, 808–822.

24.

Gleditsch

K. S.

(2007). Transitional dimensions of civil war. Journal of Peace Research, 44, 293–309.

25.

Gleditsch

N. P.

Wallensteen

Eriksson

Sollenberg

Strand

(2002). Armed conflict 1946–2001: A new dataset. Journal of Peace Research, 39, 615–637.

26.

Grossman

(1992). Foreign aid and insurrection. Defense and Peace Economics, 3, 275–288.

27.

Gurr

(2000). Peoples versus states. United States Institute of Peace Press.

28.

Gurr

Moore

(1997). Ethnopolitical rebellion: A cross-sectional analysis of the 1980s with risk assessment for the 1990s. American Journal of Political Science, 41, 1079–1103.

29.

Gutting

Steinwand

(2017). Donor fragmentation, aid shocks, and violent political conflict. Journal of Conflict Resolution, 61, 643–670.

30.

Hammond

(2017). Maps of mayhem: Strategic location and deadly violence in civil war. Journal of Peace Research. Advance online publication. https://doi.org/10.1177/0022343317702956

31.

Hansen

Tarp

(2001). Aid and growth regressions. Journal of Development Economics, 12, 375–398.

32.

Hegre

Ellingsen

Gleditsch

N. P.

Gates

(2001). Toward a democratic civil peace? Democracy, political change and civil war, 1816-1992. American Political Science Review, 95, 33–48.

33.

Knack

(2001). Aid dependence and the quality of governance: Cross-country empirical tests. Southern Economic Journal, 68, 310–329.

34.

Kolstad

(2005). Direct budget support and corruption (U4 Issue Paper No. 1). CMI.

35.

Marshall

Jaggers

(2003). Polity IV project. University of Maryland.

36.

Mason

T. D.

(2003). Structures of ethnic conflict: Revolution versus secession in Rwanda and Sri Lanka. Terrorism & Political Violence, 15, 83–113.

37.

Morelli

Rohner

(2015). Resource concentration and civil wars. Journal of Development Economics, 117, 32–47.

38.

Morrissey

(2006). Fungibility, prior actions, and eligibility for budget support. In Koeberle

Stavreski

Walliser

(Eds.), Budget support as more effective aid?: Recent experiences and emerging lessons, (pp. 333–343). World Bank.

39.

Narang

(2015). Assisting uncertainty: How humanitarian aid can inadvertently prolong civil war. International Studies Quarterly, 59, 184–195.

40.

Nielsen

Findley

Davis

Candland

Nielson

(2011). Foreign aid shocks as a cause of violent armed conflict. American Journal of Political Science, 55, 219–232.

41.

Nunn

Qian

(2014). US food aid and civil conflict. American Economic Review, 104, 1630–1666.

42.

Organization for Economic Cooperation and Development. (2018). Net ODA. OECD data. https://data.oecd.org/oda/net-oda.htm

43.

Polman

(2010). The crisis caravan: What’s wrong with humanitarian aid? Metropolitan Books.

44.

Rajan

Subramanian

(2008). Aid and growth: What does the cross-country evidence really show? Review of Economics and Statistics, 90, 643–665.

45.

Ross

(2004). What do we know about natural resources and civil war? Journal of Peace Research, 41, 337–356.

46.

Ruggeri

Schudel

C. W.

(2010, March 16). Pacifying aid: The effects of foreign aid on civil war duration [Paper presentation]. Annual International Studies Association, Montreal, Canada.

47.

Savun

Tirone

(2011). Foreign aid, democratization, and civil conflict: How does democracy aid affect civil conflict? American Journal of Political Science, 55, 233–246.

48.

Savun

Tirone

(2012). Exogenous shocks, foreign aid, and civil war. International Organization, 66, 363–393.

49.

Strange

Dreher

Fuchs

Parks

Tierney

(2017). Tracking underreported financial flows: China’s development finance and the aid–conflict nexus revisited. Journal of Conflict Resolution, 61, 935–963.

50.

Subhayu

Sandler

Younas

(2014). Foreign direct investment, aid, and terrorism. Oxford Economic Papers, 66, 25–50.

51.

Svensson

(2000). Foreign aid and rent-seeking. Journal of International Economics, 51, 437–461.

52.

Tarnoff

Lawson

M. L.

(2011). Foreign aid: An introduction to U.S. programs and policy. Congressional Research Service Report. https://www.everycrsreport.com/files/20110210_R40213_270577bfb8efb9044920eae5ec56d650fce6ca54.pdf

53.

Tierney

Nielson

Hawkins

Roberts

J. T.

Findley

Powers

Parks

Wilson

Hicks

(2011). More dollars than sense: Refining our knowledge of development finance using AidData. World Development, 39, 1891–1906.

54.

Wilbanks

Karsh

(2010). How the sons of Iraq stabilized Iraq. Middle East Quarterly, 17, 57–70.

55.

Wimmer

Cederman

Min

(2009). Ethnic politics and armed conflict. A configurational analysis of a new global dataset. American Sociological Review, 74, 316–337.

56.

Wood

Molfino

(2016). Aiding victims, abetting violence: The influence of humanitarian aid on violence patterns during civil conflict. Journal of Global Security Studies, 1, 186–203.

57.

Wood

Sullivan

(2015). Doing harm by doing good? The negative externalities of humanitarian aid provision during civil conflict. The Journal of Politics, 77, 736–748.

58.

Zürcher

(2017). What do we (not) know about development aid and violence? A systematic review. World Development, 98, 506–522.