Aiding refugees, aiding peace?

• DOI: http://doi.org/10.1111/jpet.12461
• Author: M. Christian Lehmann
• Published date: 01 September 2020
• Date: 01 September 2020

J Public Econ Theory. 2020;22:1687–1704. wileyonlinelibrary.com/journal/jpet © 2020 Wiley Periodicals LLC

|

1687

Received: 6 August 2019

|

Accepted: 12 June 2020

DOI: 10.1111/jpet.12461

ORIGINAL ARTICLE

Aiding refugees, aiding peace?

M. Christian Lehmann

Department of Economics, University of Brasilia, Campus Universitario Darcy Ribeiro, Brasilia, DF, Brazil

Correspondence

M. Christian Lehmann, Department of

Economics, University of Brasilia, Campus

Universitario Darcy Ribeiro, Brasilia, DF

70910‐900, Brazil.

Email: clehmann@unb.br

Abstract

I add two novel features to the two‐player contest

model, the workhorse model of civil war: civilians

can flee (i.e., become refugees) and refugees receive

aid. I find that aid to refugees can promote peace or

fuel conflict, depending on the context.

1|INTRODUCTION

One strand of literature contends that foreign aid fuels civil war by increasing the pool of

resources that can be captured by armed groups, fostering a war economy, and increasing the

prize of capturing the state (henceforth prize theory).

1

Another strand argues that foreign aid

can mitigate conflict by strengthening a government's military capacity to fight insurgents

(military capacity theory).

2

Furthermore, aid may win “hearts and minds”of civilians, leading

them to cooperate with the government, which can reduce conflict by increasing the effec-

tiveness of counterinsurgency (hearts‐and‐minds theory).

3

The latter may backfire, however,

because it entices insurgents to sabotage aid delivery to avoid that civilian opinion sways against

them (sabotage theory).

4

I study an additional (so far neglected) mechanism: aid's effect on emigration. I add two

novel features to the two‐player contest model, the workhorse model of civil war

5

: civilians can

flee (i.e., become refugees) and refugees receive aid. The basic setup of the model consists of two

players contesting a territory. The prize of a military victory is “rents”—pecuniary (e.g., taxes,

1

For example, see Besley and Persson (2011), Nunn and Qian (2014), and Bluhm, Gassebner, Langlotz, and

Schaudt (2016).

2

For example, see Collier and Hoeffler (2002), De Ree and Nillesen (2009), Nielsen, Findley, Davis, Candland, and

Nielson (2011), and Ahmed and Werker (2015).

3

For example, see Berman, Shapiro, and Felter (2011) and Beath, Christia, and Enikolopov (2016).

4

See Crost, Felter, and Johnston (2014).

5

See, for example, Haavelmo (1954), Hirshleifer (1988), Hirshleifer (1989), Garfinkel (1990), Grossman (1991),

Skaperdas (1992), Hirshleifer (1995), Fearon (2008), and Besley and Persson (2011).

looting, etc.) and/or nonpecuniary (e.g., imposing a certain religious doctrine)—that the contest

winner can extract from civilians.

My analysis highlights that aid to refugees can decrease the prize (hence incentives to fight),

for three reasons: first, it entices fleeing, thus reducing the number of rent payers; second, it

forces players to reduce rent extraction to make fleeing less attractive; finally, it increases the

wage that players must pay their fighters (for rational fighters desert and flee if their expected

utility as refugees are higher). Yet, in some contexts, aid to refugees can also increase the prize

(hence incentives to fight), for example, when it entices civilians to flee and abandon their

assets (e.g., land).

Conversely, aid to civilians in the disputed territory can increase the prize (hence incentives

to fight) by increasing civilians' opportunity cost of fleeing, thus increasing the number of rent

payers.

6

For prize theories, my findings imply that aid to the disputed territory can fuel conflict, not

only because aid is a prize worth fighting over (the main argument of prize theories), but also

because it entices civilians to stay (my argument) hence increasing the number of rent payers. It

may also imply that the hearts‐and‐minds mechanism could be less effective in contexts where

aid entices civilians to stay.

7

Finally, for sabotage theories, my results imply that sabotage is less

likely to occur in contexts where successful aid delivery entices civilians to stay.

2|MINIMAL MODEL

The starting point of my analysis is the two‐player contest model, the workhorse of the

formal conflict literature.

8

Two competing players

∈

i

1,

2

, for example, the government and

an insurgent, represented by their respective leader(s), fight to gain control over some

territory with either pecuniary (taxes, looting, etc.) or nonpecuniary objectives (e.g., im-

posing a certain religious doctrine), that is, to extract rents of some sort from civilians. The

probability of military victory is increasing in the size of a player's army and decreasing in

the size of the other player's army. A player maximizes expected profit (expected rents

minus fighter wages) by choosing the optimal size of its army. My main departure from the

conventional contest model is to relax the assumption that civilians cannot avoid taxation

and that the tax rate is exogenous. In my model, a player must carefully choose the tax rate

because civilians flee if it is too high.

6

In practice, the lion's share of aid goes to conflict countries. For example, Nunn and Qian (2014) compile panel data

containing annual aid receipt and conflict data for 125 countries from 1971 to 2006. Using their data, I estimate that a

country with an ongoing civil war receives, on average, 140% more US food aid and 171% more aid from other donor

countries than a nonconflict country, even after controlling for recipient country GDP per capita, population, and

region. On the other hand, refugees receive comparatively little aid, which starts with the reluctance of many countries

to admit refugees (Lehmann, 2019), followed by opposition to giving them pecuniary forms of aid (e.g., Dahlberg,

Edmark, & Lundqvist, 2012; Dustmann, Vasiljeva, & 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).

7

On the one hand, aid increases civilians' cooperation with the government's counterinsurgency efforts, yet by enticing

civilians to stay it also increases the number of rent payers hence the insurgent's prize. This may explain why existing

empirical evidence for the hearts‐and‐minds theory is based mainly on aid in the form of infrastructure (e.g., roads),

which is arguably a less compelling reason to stay than, for example, humanitarian aid.

8

See, for example, Haavelmo (1954), Hirshleifer (1988), Hirshleifer (1989), Garfinkel (1990), Grossman (1991), Skaperdas

(1992), Hirshleifer (1995), Fearon (2008), and Besley and Persson (2011).

1688

|

LEHMANN