U.S. Refugee Aid and Civil Conflict

ABSTRACT

I present evidence that U.S. aid for refugees mitigates civil conflict in their origin country. My main result is that a 10 percent increase in U.S. humanitarian aid for refugees reduces conflict deaths in their origin country by 1.5 percent. Presumably, aid for refugees entices civilians to flee from the location of conflict, thus depriving armed groups of resources and targets.

KEYWORDS:

• Civil conflict
• refugees
• foreign aid

Acknowledgments

I thank Daniel Masterson for collaboration at an earlier stage of the project. I am also grateful to Nathan Nunn, Jesse Shapiro and several anonymous referees for thoughtful comments and suggestions.

Disclosure Statement

No potential conflict of interest was reported by the author.

Notes

1. Nunn and Qian (2014), Besley and Persson (2011), Crost, Felter, and Johnston (2014) and Weintraub (2016) suggest that aid can exacerbate conflict. Bluhm et al. (2016) find that receiving bilateral aid raises the chances of escalating from small conflict to armed conflict (but no evidence that aid ignites conflict in truly peaceful countries). On the other hand, Berman, Shapiro, and Felter (2011) finds that U.S. military spending on small- scale infrastructure projects in Iraq reduced insurgent violence. Collier and Hoeffler (2002), De Ree and Nillesen (2009), Nielsen et al. (2011), Berman et al. (2013), Crost, Felter, and Johnston (2016), and Beath, Christia, and Enikolopov (2016) also find conflict-reducing effects of aid. Most recently, Gehring, Kaplan, and Wong (2019) find that aid by both the World Bank and China does not increase outright conflict nor any type of citizen protest (Chinese aid, however, is associated with more government repression).

2. Nunn and Qian (2014) highlight the ease with which armed factions and opposition groups appropriate humanitarian aid, which is often physically transported over long distances through territories only weakly controlled by the recipient government. Reports indicate that ‘up to 80 percent of aid can be stolen en route’ (1631). Even if aid reaches its intended recipients, the authors argue, it can still be confiscated by armed groups (against whom the recipients are typically powerless) and that it is difficult to exclude members of local militia groups from being direct recipients if they are also malnourished and qualify to receive aid. ‘In all these cases, aid ultimately perpetuates conflict’ (1631).

3. A prominent historical example where aid arguably played a role in enticing emigration and reducing conflict deaths is the Mariel boatlift (e.g. Card 1990). The U.S. government responded to Cuban relaxation of restrictions on emigration by allowing Cuban-Americans to send up to 2,000 USD to an emigrating relative, and declared that it would grant refugee status to all Cubans fleeing to the U.S. This U.S. aid to Cuban refugees arguably greatly contributed to the ensuing mass emigration from Cuba, including numerous opponents of Fidel Castro’s regime. It seems plausible that this exodus of opponents reduced conflict deaths in Cuba, for émigrés could not be persecuted anymore by the Cuban government and its militias, and it arguably reduced the risk of an organized rebellion (e.g. a guerrilla war).

4. See also Azam and Berlinschi (2008), Bermeo and Leblang (2015), Lanati and Thiele (2017), Camarena (2017), and Dreher, Fuchs, and Langlotz (2019).

5. In 2014 the U.S. made contributions worth more than 2.2 billion USD to WFP and more than 1.28 billion USD to UNHCR, constituting more than 50% of 5.9 billion USD in U.S. humanitarian assistance that year. (Sources: WFP, https://www.wfp.org/funding/year/2014; UNHCR, http://www.unhcr.org/536c960a9.pdf; http://www.globalhumanitarianassistance.org/countryprofile/united-states).

6. Source: https://www.prio.org/Data/Armed-Conflict/Battle-Deaths/.

7. The measure is smaller than the total number of people killed by conflict, for it does not count the ‘indirect’ victims of conflict, i.e. people killed by deterioration of the economy, infrastructure of health and human services, and public safety systems (Lacina and Gleditsch 2005).

8. A meta-analysis of 60 primary studies show that deviations from normal precipitation and mild temperatures systematically increase the risk of conflict: In the post-1950 era, a one standard deviation change in climate toward warmer temperatures or more extreme rainfall increases the frequency of interpersonal violence by 4% and intergroup conflict by 14% (Hsiang, Blurke, and Muguel et al. 2013).

9. Bazzi and Blattman (2014) examine price shocks from 65 commodities in all developing countries from 1957 to 2007, and find that positive commodity price shocks lead to shorter, less deadly wars. Dube and Vargas (2013) find that, in Colombia, violent deaths increase when the coffee price falls but decrease when the oil price falls. Nunn and Qian (2014) argue that U.S. production is unlikely to affect world wheat prices since the United States does not dominate global wheat production and because U.S. price stabilization policies are effective in eliminating any correlation between U.S. production and wheat prices. Accordingly, they find no correlation between total production and average wheat prices between 1971 and 2006.

10. Like NQ, I assume that a country experiences conflict if there are at least 25 conflict-related deaths during the calendar year according to the UCDP/PRIO data set. The countries for which the UCDP/PRIO data set records civil conflicts between 2000 and 2006 are Afghanistan, Algeria, Angola, Armenia, Azerbaijan, Bangladesh, Burundi, Central African Republic, Chad, Colombia, Congo (DRC), Congo (Rep.), Cote d’Ivoire, Eritrea, Ethiopia, Georgia, Guinea, Guinea-Bissau, Haiti, India, Indonesia, Iran, Iraq, Kyrgyz Republic, Lebanon, Liberia, Mali, Namibia, Nepal, Niger, Nigeria, Pakistan, Philippines, Russia, Rwanda, Saudi Arabia, Senegal, Sierra Leone, Somalia, Sri Lanka, Sudan, Tajikistan, Thailand, Turkey, Uganda, and Uzbekistan.

11. I obtain similar results with logged conflict deaths (see appendix).

12. The quotes are from https://www.unhcr.org/5a0c00537.pdf (accessed on 21 March 2020).

13. I adopt the region definition of the World Bank: East Asia and Pacific; Europe and Central Asia; Latin America and Caribbean; Middle East and North Africa; South Asia; Sub-Saharan Africa; and other.

14. Recent estimates of the Syrian conflict’s death toll are nearing 500,000 individuals. The pre- conflict population of Syria is 22,850,000. The death risk for the average citizen, therefore, is 500,000/22,850,000 = 2.2%.

15. The dependent variable is average conflict deaths across countries hosting refugees from origin country o in year t.

16. WFP operations in a given country include food aid to residents affected by drought or other natural disasters, to orphans, to civilians infected by HIV/AIDS or other diseases, in addition to food aid to refugees.

17. Salehyan and Gleditsch (2006), for example, argue that refugee camps ’often provide sanctuary to rebel organizations, a base of operations, and fertile recruitment grounds’ (p. 324). Zolberg, Suhrke, and Aguayo (1992) describe refugee camps as potential military bases for ‘refugee warriors’ to continue opposition activities. Recent work, however, finds no evidence that aid turns refugees into warriors (Masterson and Lehmann 2020).

18. Dahlberg, Edmark, and Lundqvist (2012) and Dustmann, Vasiljeva, and Piil Damm (2018) exploit exogenous variation in the allocation of refugees to Swedish and Danish municipalities, and both find negative effects on residents’ support for redistribution. Vasilakis (2018) reports similar results for Greece. Eger (2009), Jofre-Monseny, Sorribas-Navarro, and Vázquez-Grenno (2016), Runst (2018), Harmon (2018), Edo et al. (2019), Senik, Stichnoth, and Van der Straeten (2009), Alesina, Miano, and Stantcheva (2018), Alesina, Murard, and Rapoport (2019) suggest that immigration reduces electoral support of parties favoring redistribution in Sweden, Spain, Germany, Denmark, France, and other European countries. The U.S. also admits only a relatively small number of refugees (Lehmann 2019), which perhaps shows residents’ lack of support for redistribution. Sometimes opposition to refugees even takes the form of violence (Falk, Kuhn, and Zweimueller 2011; Onoma 2013; Benvcek and Strasheim 2016; Jaeckle and Koenig 2017; Lehmann and Masterson 2020).

19. A literature in political science, for example, finds that voter attitudes can be changed through canvassing (e.g. Gerber and Green 2000). If canvassing can make voters more willing to aid refugees, remains unstudied to my best knowledge.