![Sample our Politics & International Relations journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5947/TOC-Subject-Sample-2021-Politics-International-Relations.jpg"})
ABSTRACT
Despite a principled commitment to assist people in need equally, the allocation of humanitarian assistance across conflict and post-conflict states shows remarkable variation that is not easily explained by differences in the level of recipient-need. This paper attempts to explain these “forgotten conflicts“ by analyzing the determinants of humanitarian aid to civil war and post-civil war states. Using cross-national panel data on humanitarian aid provisions, I show that the most important determinants of international humanitarian assistance are not always demand-side factors measuring humanitarian need – as the principals of humanitarian action would dictate – but often strategic factors that reflect donors’ political interests in providing humanitarian assistance. Although humanitarian aid to ongoing civil wars appears to be substantially more humanitarian than strategic in its allocation, humanitarian aid provided to post-conflict states in the aftermath of civil war tends to go to conflicts where donors perceive important strategic and political interests. These results suggest that one important explanation for why some conflicts are essentially ignored or gradually neglected over time is that strategic interests of donors can dominate humanitarian concerns over time.
Acknowledgments
I thank the United Nations High Commissioner for Refugees (UNHCR), Ed Mansfield and the Christopher H. Browne Center for International Politics, the Center for International Security and Cooperation (CISAC) at Stanford University, the Centre on Conflict, Development & Peacebuilding at the Graduate Institute Geneva, and the University of California Institute for Global Conflict and Cooperation for providing research and financial assistance. I also thank Michael Barnett, Mark Buntaine, Tim Büthe, Andrea Everett, Erik Gartzke, Peter Gourevtich, Stephanie Hofmann, Oliver Jütersonke, Miles Kahler, Keith Krause, David Lake, Brad LeVeck, Solomon Major, Robbie Narang, Daniel Tirone, Sarah Stroup, two anonymous reviewers, and Michael Colaresi for their helpful comments and support.”
Supplemental Material
Supplemental data for this article can be accessed on the publisher’s website.
Notes
1 The terms “forgotten emergencies” and “forgotten conflicts” generally refer to conflict areas—typically civil wars—that the international community has essentially ignored or gradually neglected over time (Darcy and Hofmann Citation2003; Oxfam Citation2000; Smillie and Minear Citation2003, Citation2004).
2 For recent examples of the literature on foreign aid and conflict, see Berman, Shapiro, and Felter (Citation2011); Besley and Persson (Citation2011); Crost, Felter and Johnston (Citation2012); Dube and Naidu (Citation2010); Narang (Citation2014, Citation2015); Nielsen, Findley, Davis, Candland, and Nielson (Citation2012); Nunn and Qian (Citation2013).
3 A separate literature explores the politicization of development aid. Marquette (Citation2003, Citation2004) documents how the World Bank attempted to reshape corruption as a development issue and argues that the World Bank should remain apolitical. On the other hand, Leftwhich (1994, 1995) claims that development is fundamentally a political matter and should be treated as such. Carothers and De Gramont (2013) consider the trend of making aid conditional on democratization and human rights “a fundamental advance.”
4 See Vaux Citation2006:4.
5 A table listing the top 15 recipients of humanitarian aid from 1999–2008 can be found in the Appendix A1.
6 The absolute increase in humanitarian aid to crisis areas (rather than the relative amount) may be less interesting, as the amount reported through the OECD increased nearly 1,400% in real terms, from US$796 million in 1989 to well over US$11 billion in 2008. This increase was especially dramatic from 1999 to 2008, when global humanitarian aid amounted to roughly US$98 billion in constant 2008 prices.
7 As described in the following, data used to estimate total humanitarian disbursements are taken from the OECD DAC2a Official Development Assistance, which utilizes specific criteria to classify aid as “humanitarian.”
8 The OECD defines aid (official development assistance) as nonmilitary grants and net disbursements of concessional loans that have at least a 25% grant element.
9 In their analysis of aid allocation, Collier et al. find that aid is more effective in augmenting growth in postconflict situations but that the effect is nonlinear over time.
10 We chose to explain aid disbursements rather than commitments (Berthélemy Citation2006; Tarp, Bach, Hansen, and Baunsgaard Citation1999; White and McGillivray Citation1995). In the OECD data, commitments represent “a firm obligation … to provide specified assistance to a recipient … ” Disbursements represent the actual “release of funds to or the purchase of goods or services for a recipient; by extension, the amount thus spent.”
11 For the 24 members of the DAC during the study period, see www.oecd.org/dac/dacmembers.htm.
12 The data include outflows from World Bank, regional development banks, and several UN agencies, including United Nations High Commissioner for Refugees (UNHCR), The Joint United Nations Programme on HIV and AIDS (UNAID), United Nations Development Programme (UNDP), United Nations Children‘s Fund (UNICEF), World Food Program, and several others. The list also includes 50 of the largest NGOs. See http://www.oecd.org/dataoecd/36/16/31724727.pdf.
13 Objections may be raised to aggregating these components since bilateral aid can be allocated through governments more often than multilateral. However, previous work has shown this is not necessarily the case postconflict, where bilateral provisions often bypass the state (Seybolt Citation2009). Nevertheless, we ran the estimations on each component and found similar effects across type. This is because both components are highly correlated due to conditionality by donors demanding allocation to needy populations.
14 This approach is standard in dealing with high monetary values, including in the literature on campaign finance (see Erikson and Palfrey Citation2000), to provide fairly symmetric errors. We confirmed this in our data by plotting the residuals and the bivariate relationship between ln(aid+k) and the independent variables.
15 The population of civil wars is from Cunningham (Citation2006), which is based on the Armed Conflict Dataset.
16 The universe of peace spells is from Fortna (Citation2008), which is adapted from (Doyle and Sambanis Citation2006).
17 We thank an anonymous reviewer for noting that some component of aid provisions may simply reflect states’ status as developing countries. However, in addressing the puzzle of “forgotten conflicts,” we are limited in our ability to expand the study by data availability, as the most important indicators of need in conflict settings have not been collected for the larger universe of developing countries—largely because they do not apply to such countries. For example, the number of conflict-related deaths and refugees/IDPs would not be possible to collect for LDCs like Tuvalu or Madagascar. Thus, while a broader analysis would be interesting, it would obscure the most important conflict-related elements of the puzzle addressed here.
18 With respect to bilateral aid from major donors like the United States, Kevlihan et al. (Citation2014) cite statements in the 2002 and 2010 US National Security Strategy that aid will be “based on needs alone.” Similarly, the decision-making process around multilateral aid is guided by UNGA resolution 46/182, stating that assistance be provided based on need and in accordance with humanitarian principles. With respect to private aid, Büthe and Major (Citation2012) note that ICRC statutes call for giving to recipients “solely guided by their needs,” while Save the Children, World Vision, CARE, and other NGOs also emphasize a focus on needs.
19 The CAP provides a framework for humanitarian organizations—including UN offices and specialized agencies (Food and Agriculture Organization of the United Nations [FAO], Office for the Coordination of Humanitarian Affairs [OCHA], Office of the United Nations High Commissioner for Human Rights [OHCHR], United Nations Development Programme [UNDP], United Nations Children’s Fund [UNICEF], United Nations High Commissioner for Refugees [UNHCR], World Food Program [WFP], World Health Organization [WHO]), international organizations (IOs) and NGOs—to prepare the appeals and monitors the receipt and use of contributions.
20 Because variables are on different scales, we first transformed the regression inputs by subtracting the mean and dividing by two standard deviations. This places scalar variables on the same scale as each other and approximately the same scale as the binary variables (Gelman Citation2008).
21 The BIC supports this conclusion more strongly: The constrained humanitarian model is actually preferred to the full model, and the full model over the strategic model after penalizing complexity.
22 Barnett (Citation2011) makes this point about the increasing politicization. In charting the history of humanitarian action, Barnett distinguishes a second age of “neohumanitarianism” from WWII through the end of the Cold War from a final age of “liberal humanitarianism.” A defining feature of the latter was that “humanitarian organizations that once sought to use states for humanitarian action now found themselves being explicitly used by states as a tool for their political and strategic objectives” (p. 32).
23 Chow test results: GDP per capita, F(2, 887)=9.57, p<=.0001; infant mortality rate, F(2, 887)=18.93, p<.0001; UN Voting Affinity with P-5, F(2, 887)=1.60, p=.2016.
24 In the full population and Cold War sample, only Infant Mortality gains significance while P-5 Contiguity loses significance. In the post-Cold War era, the once-significant coefficient on Former Colony of P-5 becomes insignificant. This suggests even less support for the claim that strategic interests of donors substantially dictated the allocation of humanitarian aid to ongoing civil wars. Controlling for time in analyzing bilateral versus multilateral humanitarian aid further supports the inference that bilateral aid is more strategic than multilateral provisions, as Oil Exporter and UN Voting Affinity become significant for bilateral giving while the level of Democracy becomes insignificant for multilateral giving.
25 Chow test results: GDP per capita, F(2, 1833)=20.06, p<.0001; infant mortality rate, F(2, 1833)=7.38, p=.0006; refugee-IDP, F(2, 1833)=56.54, p<.0001.
26 Chow test results: P-5 colony, F(2, 1833)=17.59, p<.0001; P-5 voting affinity, F(2, 1833)=11.47, p<.0001; oil, F(2, 1833)=17.59, p<.0001, prior 5-year democracy average, F(2, 1833)=7.8, p<.0004).
27 In comparing the Cold War to Post Cold War period, only Oil Exporter and P5-Contiguity change to become significant during the Cold War, adding further strength to the inference made that the Cold War was a period of more strategic humanitarian provision to postconflict states. In comparing bilateral to multilateral humanitarian aid, only the number of conflict-related deaths changes to become insignificant for the bilateral aid, adding strength to the inference that bilateral aid appears to be more strategic.
28 The effect of Infant Mortality and refugees-IDPs diminishes over time, while the effects of Oil and P-5 Contiguity increase. This pattern of humanitarian factors waning and strategic factors increasing is more pronounced in the Cold War and less in the post-Cold War period. Two humanitarian factors become less significant over time for bilateral disbursements (GDP per capita and conflict related deaths), while one becomes more significant for multilateral, suggesting the former is less influenced by humanitarian factors.
