Distortions in Aid Allocation of United Nations Flash Appeals: Evidence from the 2015 Nepal Earthquake

Originally published
View original



On April 25, 2015 an earthquake of magnitude 7.8 struck central Nepal.1 With more than 100 subsequent aftershocks, of which the largest reached an intensity of 7.3 on May 12, it was one of the most destructive natural disasters in the history of Nepal. The disaster triggered sub‐ stantial international attention. It is estimated to have killed 8,800 people and affected more than 5.6 million (Guha‐Sapir et al., 2016a), which constituted almost a fifth of Nepal’s inhab‐ itants.

Four days after the earthquake, the United Nations Office for the Coordination of Human‐ itarian Affairs (UNOCHA), in collaboration with the Office of the Humanitarian Coordinator (OHC), the Nepalese government, and humanitarian partners, issued a strategic response plan for Nepal, a so‐called flash appeal.2 The 2015 UN Nepal Earthquake Flash Appeal iden‐ tified 184 projects and requested US$ 422 million in order to provide life‐saving assistance and protection for the Nepalese people in the five months after the earthquake (AidData, 2016a).3 Although 77 donor organizations responded to the Flash Appeal, a third of the re‐ quested amount remains unmet as of May 2018 (UNOCHA, 2018).4 However, the spatial het‐ erogeneity is large. This holds with respect to the number of aid projects, the proposed finan‐ cial amount, and the degree to which proposed projects obtained funding commitments by international donors. This raises the question: what explains the selection of project locations and the provision of the requested funds?

This paper is the first to investigate the allocation of proposed and funded humanitarian aid projects in the framework of a United Nations flash appeal. Flash appeals have become important components of humanitarian relief in emergency situations, as demonstrated by the 76 flash appeals the UN has issued over the 2005‐2016 period. In sum, the international donor community has spent US$190 billion over this period alone to satisfy humanitarian needs and stimulate economic reconstruction. As can be seen from the list of the 20 largest events in Table 1, the 2015 Nepal earthquake triggered the seventh largest flash appeal in terms of its financial size. Common hope and expectation is that disaster affectedness and the specific vulnerabilities of municipalities are the criteria for choosing the location of aid pro‐ jects. A meaningful aid allocation is particularly salient since 9 of 10 flash appeals are under‐ funded (UNOCHA, 2018).

Studying the allocation of emergency aid is important as these flows are intended to im‐ prove the humanitarian situation of the population living in disaster‐affected areas. Beyond the mere humanitarian aspect, empirical findings suggest that post‐disaster aid can boost economic growth (Bjørnskov, 2013), speed up the recovery process of microenterprises (Mel et al., 2012), and play a role in reducing the likelihood of escalating government repression in democracies (Wood and Wright, 2016). Even critics of ‘general’ development aid support the continued provision of emergency relief following devastating disasters (Moyo, 2009). Since humanitarian aid flows are surprisingly small in comparison to the damages caused (Becerra et al., 2014, 2015), a need‐oriented aid allocation is particularly salient.

By analyzing the humanitarian response triggered by the 2015 Nepal earthquake, this study addresses two (sequentially) related aspects of flash appeals. We study the municipali‐ ty characteristics that influence the number of emergency aid projects and financial amount committed to a particular municipality. First, we analyze what determines the choice of pro‐ ject‐locations in the design stage of the 2015 UN Nepal Earthquake Flash Appeal across mu‐ nicipalities. Second, we investigate which proposed project locations obtain funding from international donors.

Our study makes use of a new, and so far unexploited, geo‐referenced aid dataset from AidData that contains information on proposed and ultimately funded aid projects that have been a part of the 2015 UN Nepal Earthquake Flash Appeal (AidData, 2016a). These data cover 156 out of 184 projects in more than 850 locations. We combine these aid data with data on nighttime light intensity and rainfall, survey data, and electoral statistics at the local level, to evaluate whether the allocation and subsequent financing of humanitarian aid projects in Nepalese municipalities are based on actual disaster impact and the population’s specific vulnerabilities or rather biased by particular interests. To assess the disaster impact, we use peak ground acceleration maps provided by USGS (2017b) combined with damage functions, which—as we argue below—provide a suitable indicator for potential destruction from earthquakes to short buildings up to seven stories (USGS, 2017c). To evaluate whether aid allocation decisions also reflect socio‐economic and physical vulnerabilities of the affected population, we add measures of municipalities’ level of development, their exposure to rain‐ fall, and their distance to the Nepalese capital Kathmandu, amongst others. Finally, we test for ethnic, religious, and political distortions in aid giving by analyzing the role of a munici‐ pality’s share of Hindus and privileged caste population, as well as the vote share of Nepal’s two dominant parties in the 2013 Constituent Assembly elections.

By studying the allocation of UN flash appeal aid, we contribute to the broader literature on the allocation of humanitarian aid. A first strand investigates aid allocation across coun‐ tries and emergencies, where emergency aid has been shown to increase with disaster severi‐ ty but is also driven by media coverage and is prone to political bias (Drury et al., 2005; Ei‐ sensee and Stromberg, 2007; Fink and Redaelli, 2011; Raschky and Schwindt, 2012; Fuchs and Klann, 2013; Bommer et al., 2018). A much smaller second strand analyzes the allocation of disaster relief within disaster‐affected areas. For example, Benini et al. (2009) and Wiesenfarth and Kneib (2010) study relief supply to earthquake‐affected communities in Pakistan after the 2005 earthquake and find that needs and logistical convenience of locations affect aid delivery. Francken et al. (2012) investigate politico‐economic factors underlying aid alloca‐ tion across communities in Madagascar in the aftermath of cyclone Gafilo in 2004. They un‐ cover that domestic aid is provided to regions where governments have stronger incentives to respond, specifically those with higher radio coverage and with stronger political support from the ruling administration. At the same time, their results suggest that foreign aid is dis‐ tributed to poorer areas and to those that are more easily accessible. Our study adds to this smaller and less developed second strand of the literature by examining the role of UN flash appeals. To the best of our knowledge, no existing study analyzes the geographic pattern of proposed and funded projects following UN flash appeals despite their importance in the immediate aftermath of a disaster.

Our empirical results show that aid allocation in the framework of the 2015 Nepal Earth‐ quake Flash Appeal lacks need orientation and shows ethnic and political biases. At the de‐ sign stage, the location choice is not guided by municipalities’ level of development and shows little regard for other socio‐economic and physical vulnerabilities—the exception be‐ ing that less urban municipalities receive more aid projects. Municipalities populated by up‐ per castes receive more projects and the strongholds of the two major Nepali parties benefit from larger aid amounts in the design stage. On the positive side, the initial appeal project proposals correlate positively with the extent of earthquake damages. Similarly striking, the funding decisions of the international donor community show little regard of socio‐economic and physical vulnerabilities.

We conclude that the geographic selection of aid projects is distorted at all levels of deci‐ sion‐making. Therefore, the need orientation of geographic project selection and funding should be strengthened. This would involve different actors: UNOCHA, the OHC, and the national government during the design of flash appeals, as well as donor countries, multilat‐ eral donors, and non‐state donors during the funding and coordination phase.

The remainder of the paper proceeds as follows. In Section 2, we outline the decision‐ making process that underlies the 2015 UN Nepal Earthquake Flash Appeal. We also discuss the factors that should (not) guide the selection and funding of project locations according to the flash appeal document and from a humanitarian perspective. Section 3 presents our re‐ search design and the data. We use various proxies for the needs and vulnerabilities of Nepalese municipalities to analyze whether these factors guide aid distribution. We then explore alternative allocation rules, which could explain the lack of need orientation. Specifi‐ cally, we test for ethnic, religious, and political distortions. In Section 4 we present and dis‐ cuss our empirical results. Finally, Section 5 summarizes our findings and outlines potential avenues for future research.