Darlings and Orphans: Interactions across Donors in International Aid

• Date: 01 January 2019
• Author: Stephan Klasen,Ronald B. Davies
• Published date: 01 January 2019
• DOI: http://doi.org/10.1111/sjoe.12261

Scand. J. of Economics 121(1), 243–277, 2019

DOI: 10.1111/sjoe.12261

Darlings and Orphans: Interactions across

Donors in International Aid*

Ronald B. Davies

University College Dublin, Dublin 4, Ireland

ronbdavies@gmail.com

Stephan Klasen

University of G¨ottingen, DE-37073 G¨ottingen, Germany

sklasen@uni-goettingen.de

Abstract

We examine how the bilateral aid flows from an individual donor to a recipient depend on the

aid flows from all other bilateral and multilateral donors to that recipient. Thereby, we assess

to what extent issues including donor coordination, free-riding, selectivity, specialization, and

common donor interests drive bilateral aid allocations. We find that others’ bilateral flowslead to

a significant increase in aid flows from a particular donor, but primarily within a givenyear. The

effects are particularly pronounced for large donors and so-called “darling” recipients. Overall,

the results suggest that herding is a dominant feature of aid inter-relationships.

Keywords: Aid; aid darlings; aid orphans; donor coordination

JEL classification:F35; F42

I. Introduction

Foreign aid continues to be one of the most important – if not the most

important – forms of capital flows to a large number of poor countries,

particularly in Africa. Foreign aid is given by an increasing pool of donors,

often allowing countries to choose from (or to be forced to deal with)

hundreds of potential donors in order to fund their budgets, programs,

and projects (World Bank, 1998). The number of donors has dramatically

increased in past decades, with new entrants including regional development

banks, new bilateral donors, global funds, philanthropic foundations, and a

myriad of non-governmental organizations (NGOs; Dreher et al., 2013).

With this expansion, there is increasingly less of a “cartel of good

intentions” (Easterly, 2002) and, instead, a highly fragmented aid landscape,

*Wewould like to thank Paul Collier, Axel Dreher, Michael Grimm, Anke Hoeffler,Adrian Wood,

two anonymous referees, and participants at workshops in Oxford, Namur, Marburg, Dublin,

Berlin, and The Hague for helpful comments on earlier versions of this paper.

C

The editors of The Scandinavian Journal of Economics 2017.

244 Darlings and orphans: interactions across donors in international aid

which is increasingly difficult to negotiate for both donors as well as

recipients (Harford et al., 2004). Although it is recognized that this

fragmentation can come with significant consequences, because aid choices

are made in an inter-related but uncoordinated environment, little is known

about the extent of the inter-relation between donor giving. We shed light

on this issue by examining bilateral overseas development assistance (ODA)

payments by 21 Organisation for Economic Co-operation and Development

(OECD) donors to 51 recipients over the period 1988–2013. We find robust

evidence of a net positive interaction in which, if all other donors increase

aid to a recipient by 1 percent in a single year, a given donor increases

theirs by 0.458 percent, with this effect essentially confined to that same

year. In particular, this is driven by aid “darlings”, that is, recipients who

receive higher than anticipated donations. This suggests that there exists

information-driven herding behaviour in aid.

There are a number of ways in which uncoordinated ODA can have

negative effects. First, it can lead to situations where recipients will be

favored by most donors and become so-called aid darlings, while others are

largely deserted by the international community and become aid “orphans”

(Levine and Dollar, 2005; Rogerson and Steenson, 2009; Marysse et al.,

2007; Utz, 2010). The recent focus on selectivity to improve aid

effectiveness could increase this problem, as some findings suggest that for

aid to be effective it should be focused on those countries with particular

characteristics, such as the combination of high poverty and a “good” policy

environment, better institutions, or poor achievement levels for millennium

development goals (Burnside and Dollar, 2000; Collier and Dollar, 2002;

Wood, 2008).1In fact, donors are now being ranked for their policy

selectivity (e.g., Knack et al., 2011), thereby providing explicit incentives

to focus aid on darlings with good policies, and to neglect orphans. If

the quality or need of a recipient is uncertain, then governments can

base their expectations in part on the choices made by other governments,

leading to herding (i.e., one donor’s aid follows that of others because of

the presumed information their donations convey). This then adds to the

selectivity problem.

Second, uncoordinated aid can lead to substantial aid volatility for

recipient countries. This has been found to be growth-reducing in recipient

countries by several studies (e.g., Bulir and Hamman, 2008; Arellano et al.,

1Note that there is a large body of literature suggesting that the findings from Burnside and

Dollar (2000) are not robust, with severalof these papers suggesting waysin which more selective

aid could be more effective.For a discussion, see Easterly et al. (2004), Roodman (2007), Nowak-

Lehmann et al. (2012) and Doucouliagos (2018).

C

The editors of The Scandinavian Journal of Economics 2017.

R. B. Davies and S. Klasen 245

2009; Kathavate and Mallik, 2012).2In addition, Nielsen et al. (2011) find

evidence that negative aid shocks can even induce internal conflict.

Third, as described in detail by the World Bank (1998) among others,

each aid program carries transaction costs for both the donor and the

recipient, as they work to manage the project and fulfil its terms.

Knack and Rahman (2007) show that such fragmentation indeed lowers

bureaucratic quality in recipient countries and Kimura et al. (2012) find

that a proliferation of donors in a country leads to lower growth.3

Naturally, these problems could be reduced if the aid coming from

numerous donors was well coordinated, predictable, and used joint

approaches to reduce the management burden. Recognizing the potential

damage of fragmented and volatile aid flows, the 2005 Paris Declaration

on Aid Effectiveness and the 2008 Accra Plan of Action called on donors

to coordinate their aid better, to make aid flows more predictable, to pool

aid flows in country-led programs, and to use country systems for aid

management. To facilitate this process of alignment and harmonization,

donors were encouraged to specialize by concentrating their aid on fewer

countries, and fewer sectors within those countries, in line with their

comparative advantage. For example, among bilateral donors organized in

the OECD’s Development Assistance Committee (DAC), lead donors have

been appointed for particular sectors in countries that should coordinate

the aid, and it has been proposed that the number of actors in each sector

should be reduced (OECD, 2009).

Unfortunately, most observers find that these goals remain largely

elusive. Aldasoro et al. (2011) find that, at least up until 2006, there was

little progress on donor coordination and specialization (using measures

of proliferation of donors at the country level). Nunnenkamp et al. (2013)

find that also in later years, aid fragmentation and the overlap of donor

engagement at the country and sector levels remain high; country studies

support these results (e.g., Nunnenkamp et al., 2015, 2016). In addition,

Bulir and Hamann (2008) have found that aid became more volatile in

the late 1990s and early 2000s when the first initiatives to improve donor

coordination and to increase its predictability were already underway. Donor

rankings based on 2007 data also suggest that much work remains to be

done here (Knack et al., 2011). At the same time, Angeles et al. (2008) have

2See Chauvet and Guillamont (2009) for a different view, which suggests that some aid volatility

might actually be desirable and that volatile aid is particularly problematic if it is pro-cyclical

with respect to other capital flows. See also Hudson and Mosley (2008) for further analysis on

the volatility–growth nexus, wherethey argue that the effects depend on the type of volatility and

the type of aid.

3See also Dreher et al. (2012) on this issue with related findings.

C

The editors of The Scandinavian Journal of Economics 2017.