Receiving Foreign Aid Can Reduce Support for Incumbent Presidents

• DOI: 10.1177/1065912918798489
• Published date: 01 September 2019
• Author: Ryan C. Briggs
• Date: 01 September 2019
• Subject Matter: Articles

798489PRQXXX10.1177/1065912918798489Political Research QuarterlyBriggs
research-article2018
Article
Political Research Quarterly
2019, Vol. 72(3) 610 –622
Receiving Foreign Aid Can Reduce
© 2018 University of Utah
Article reuse guidelines:
sagepub.com/journals-permissions
Support for Incumbent Presidents
https://doi.org/10.1177/1065912918798489
DOI: 10.1177/1065912918798489
journals.sagepub.com/home/prq
Ryan C. Briggs1
Abstract
Foreign aid is thought to be useful, and therefore desirable, to recipient governments because it allows them to
increase their support through the provision of goods or services. However, the effect of the provision of aid on vote
choice has rarely been directly tested. I examine the effect of receiving foreign aid on incumbent electoral support
in three African countries using a spatial difference in differences design. Surprisingly, receiving aid lowers support
for incumbent presidents. I test two mechanisms that could produce this result and find support for a mechanism
where aid reduces one’s opinion of the incumbent, perhaps because aid does not live up to expectations. I discuss the
implications of this result for research on the political effects of foreign aid and electoral accountability.
Keywords
foreign aid, electoral politics, distributive politics, retrospective voting
Foreign aid is thought to be useful, and therefore desir-
and Fiala (forthcoming) finds that participating in an anti-
able, to recipient governments because it allows them to
poverty program increases opposition support in Uganda
increase their support through the provision of goods or
and De Kadt and Lieberman (forthcoming) find that
services. This claim assumes that voters will respond to
improvements in service provision lead to declines in
the provision of aid-funded goods or services by increas-
incumbent support in southern Africa. Thus, while the
ing their level of support for the government. This
present paper’s results are surprising, they contribute to a
assumption is reasonable and common and has rarely
growing number of surprising findings. When read
been tested. I test this assumption by examining how vot-
together, this collection of work suggests that, at least in
ers respond to the initiation of an aid project near their
the context of low-income countries, the causal effect of
location using geotagged aid data and multiple geotagged
economic improvements on political support may be
survey waves from Nigeria, Senegal, and Uganda. I use a
more variable than is typically assumed.
spatial difference in differences strategy, as in Knutsen
et al. (2017). Causal identification in this strategy rests on
Literature
the assumption that people who live near aid projects that
were started just before the survey should be very similar
Foreign aid is a transfer of resources to recipient govern-
to people that live near aid projects that will be started
ments or organizations in recipient countries. Recipient
just after the survey. Contrary to expectations, I find that
governments want aid because the additional resources
receiving aid lowers the probability of someone express-
that aid provides allow them to increase their support. In
ing a desire to vote for the incumbent by between 8 to 10
some regimes, this occurs when aid is appropriated by
percentage points.
elites. For example, in Malawi, in the 1990s the country
I propose three mechanisms that might cause aid to
received aid to build schools, but procurement fraud by
reduce incumbent support and am able to test two of
associates of members of the ruling party meant that few
them. I find little support for an explanation where aid
schools were built (Briggs 2015). Instead of funding pub-
empowers citizens and in doing so boosts opposition sup-
lic goods like schools, this aid was converted into private
port. I find stronger support for a mechanism where aid
fails to meet voter expectations and so leads to lower trust
1University of Guelph, ON, Canada
in the president and ruling party. Although I find that aid
decreases trust, I find no evidence that people who receive
Corresponding Author:
aid are exposed to more corruption.
Ryan C. Briggs, Guelph Institute of Development Studies and
Department of Political Science, University of Guelph, 046 MacKinnon
Although this paper’s core finding is unexpected, it
Building, 50 Stone Rd. E, Guelph, ON, Canada N1G 2W1.
has some precedent. For example, Blattman, Emeriau,
Email: rbriggs@uoguelph.ca

---

Briggs
611
goods (cash) held by members of the ruling clique and
expected to cause an increase in support for the incum-
their close associates. Because aid can be stolen or repur-
bent government.
posed through fungibility, it could increase the durability
Although the claim that aid will increase incumbent
of autocratic regimes in ways that are similar to an unen-
support is commonly assumed, its universality has
cumbered resource like oil (Morrison 2007, 2011).
recently been challenged. De Kadt and Lieberman (forth-
Although there is at best mixed empirical support for the
coming) find that increases in service provision lead to
equivalence of aid and oil (Bermeo 2016), there is good
decreases in incumbent political support in South Africa,
evidence that aid increased the durability of autocratic
Botswana, Lesotho, and Namibia. Blattman, Emeriau,
regimes during at least the Cold War (Bermeo 2016;
and Fiala (forthcoming) examined the electoral effects of
Dunning 2004).
an aid-supported antipoverty program in Uganda and
Much aid, however, is not captured by elites and
found that recipients of the program where economically
instead funds the creation of additional goods and ser-
better off but were also more likely to campaign for and
vices that are enjoyed by the citizens of recipient coun-
vote for the opposition party.
tries.1 Consider again Malawi, where health aid has
The present article examines the effect of aid on
reduced malaria prevalence and improved self-reported
incumbent support using individual-level survey data
health care quality (Marty et al. 2017). In this case, aid
from multiple time periods and multiple countries. In line
should help democratic leaders remain in power because
with the results of De Kadt and Lieberman (forthcoming)
it funds improvements in people’s lives and so should
and Blattman, Emeriau, and Fiala (forthcoming), I find
increase incumbent support through a retrospective vot-
that receiving aid lowers the likelihood that one expresses
ing mechanism.2 Donors often have an interest in ensur-
a desire to vote for the incumbent president. In the fol-
ing that their aid effectively helps people in recipient
lowing section, I propose a number of mechanisms that
countries (Bermeo 2017). To this end, they sometimes
may explain this result.
channel aid around recipient governments with weak
institutions (Dietrich 2013). Citizens in at least one recip-
Aid and Incumbent Support
ient country seem to be aware of donor efforts, and prefer
development projects funded by donors to those funded
The main contribution of this paper is to present an unex-
by (and possibly captured by) their own government
pected result. In proposing possible explanations for this
(Findley et al. 2017).
result, it is worth considering both those that are unique
There is some evidence that aid helps leaders remain
to foreign aid and those that are due to a surprising rela-
in power, possibly by increasing citizen support for
tionship between improvements in welfare and voting
incumbents. Briggs (2015) finds that increases in aid
behavior more generally. It should be emphasized that
before elections help incumbents win reelection. this section is speculative, and was written in response to
Jablonski (2014) finds that receiving aid increases votes
the puzzling results below. Thus, the tests of these mecha-
for the incumbent in Kenya. Licht (2010) finds that dem-
nisms should be understood as exploratory rather than
ocrats benefit from aid, with the effect being most pro-
confirmatory.
nounced when democrats first take office. Yuichi Kono
I first focus on a mechanism that is unique to foreign
and Montinola (2009) find that aid improves leader sur-
aid and then explain two mechanisms that may apply to
vival for both dictators and democrats, though the mecha-
other forms of service delivery as well as aid. The aid-
nisms are somewhat different and current aid helps
specific mechanism is that of foreign dependence. Aid is
democrats most. Although not examining the effect of aid
unlike other service improvements because it is foreign-
directly, Harding (2015) shows that road improvements
funded. It thus seems possible that receiving aid could
increase incumbent support in Ghana.
delegitimize incumbent leaders or parties if it leads people
Finally, leaders of both donor and recipient countries
to view their government as incapable of providing for
target aid as if it is a politically useful resource. Recipient
their citizens without outside support. If receiving foreign
governments take care to target aid to places where they
aid causes people to view their government or leaders as
think it is most politically valuable to them (Briggs 2014;
less legitimate or more dependent, then it may reduce an
Jablonski 2014; Masaki 2018). Donor countries give
incumbent’s electoral support. Despite the coherence of
more aid to recipient...