Diverse neighbors and post-conflict recovery at the village level: Evidence from Iraq after ISIL
| Author | Lloyd Lyall |
| DOI | http://doi.org/10.1177/00223433211039115 |
| Published date | 01 July 2022 |
| Date | 01 July 2022 |
| Subject Matter | Regular Articles |
Diverse neighbors and post-conflict recovery
at the village level: Evidence from Iraq after ISIL
Lloyd Lyall
Stanford University & Harvard Law School
Abstract
Why do some towns recover faster than others after intrastate conflict? Many important decisions about post-conflict
recovery are made at the substate level, but little empirical work has investigated what causes differences in recovery
outcomes within a country. This article suggests that proximity to ethno-religiously diverse neighbors slows a town’s
post-conflict recovery. A town has ‘diverse neighbors’ if towns with different plurality ethno-religious groups are
nearby. This hypothesis is tested by exploring variation in recovery speed among Iraqi towns after the 2014–17
Islamic State insurgency (ISIL). The article constructs 81-month panels of economic activity for 379 Iraqi settle-
ments occupied by ISIL by using satellite-observed nighttime light emissions as a proxy for economic activity. The
panels reveal large variation in post-conflict recovery among towns during the first year of peace. Village-level survey
data are then used to construct a measure of neighbor diversity, which is combined with lighting-based recovery
scores in spatial autoregression. The results show that greater neighbor diversity is robustly associated with slower
settlement recovery. The neighbor diversity penalty cannot be fully explained by cleavages between groups ‘on
opposite sides’ of the conflict; proximity to out-group neighbors appears to slow recovery even between wartime
allies. Several explanations are considered, and this article suggests that the types of post-liberation controllers that
arise in diverse areas – which tend to be substate militias rather than the government – may be one important
mechanism.
Keywords
ethnic diversity, ethno-religious, Iraq, ISIL, luminosity, post-conflict recovery
Introduction
Why do some towns recover faster than others after
intrastate conflict? Most of the literature focuses on
cross-country comparisons, but many important post-
conflict reconstruction decisions are made at the substate
level: rebuilders must decide how to allocate funds
among different conflict-stricken settlements, which
strategies to use, and how interventions should vary
among villages with different characteristics and
experiences.
Using satellite-observed nighttime lighting as a proxy
for economic activity, this article examines village-level
variation in post-conflict recovery speed among Iraqi
settlements following the 2014–18 Islamic State in Iraq
and the Levant (ISIL) insurgency. It first identifies large
variation in recovery among Iraqi settlements during the
first year of peace. It then seeks to explain this variation.
The bulk of this article’s attention concentrates on test-
ing the core hypothesis that having ethno-religiously
diverse neighbors slows a town’s recovery.
The article’s first and second sections organize existing
literature on post-conflict recovery, motivate the neigh-
bor diversity hypothesis, and introduce the case. The
third and fourth sections explain how the core depen-
dent variable (recovery) and independent variable (diver-
sity of neighbors) are measured and present descriptive
statistics. The fifth section explains and implements a
causal test. The remaining sections interpret the results
and discuss causal mechanisms.
Corresponding author:
lloydl@alumni.stanford.edu
Journal of Peace Research
2022, Vol. 59(4) 543–561
ªThe Author(s) 2021
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/00223433211039115
journals.sagepub.com/home/jpr
Post-conflict recovery: Phoenixes
and conflict traps
This article focuses on post-conflict economic recovery. At
the state level, it typically follows one of two paths: some
countries rise phoenix-style from the ashes of war with
dramatic post-conflict growth; others remain trapped in
poverty and resurgent violence for years after conflict ends.
Solow’s (1956) model of economic growth was
among the first to provide insights into why some states
recover rapidly after conflict. At the end of war, the
marginal return to capital is enormous. The first few
investments in postwar capital reconstruction therefore
result in tremendous economic growth, returning coun-
tries to their pre-conflict GDP path. Organski & Kugler
(1977), Koubi (2005), and Chen, Loayza & Reynal-
Querol (2008) have all observed state-level empirical
support for Solow’s theorized recovery after conflict,
with Organski & Kugler remarking that war-torn coun-
tries rise ‘like a phoenix from the ashes’ with dramatic
post-conflict growth. Empirical support for phoenix
recovery has also been observed at the substate level.
Davis & Weinstein (2002), Brakman, Garretsen &
Schramm (2004), and Miguel & Roland (2011) all pro-
vide empirical evidence of cities or regions that were
more extensively destroyed during conflict growing faster
after conflict than their less-destroyed neighbors.
However, not all cases recover as ph oenixes. Some
countries, cities, or regions experience long-lasting neg-
ative consequences from war (Kang & Meernik, 2005).
Collier (2003) argues that the qualities that make states
vulnerable to new domestic conflicts – such as low eco-
nomic growth, poverty, and the availability of weapons –
are also qualities that result from previous conflicts. Post-
war states can become mired in a ‘conflict trap’ and fail to
realize phoenix-style postwar growth.
The phoenix effect and the conflict trap describe well
the two major categories of post-conflict recovery. Flores
& Nooruddin (2009) show that the category a state will
fall into is often cemented by the end of the first post-
conflict year: a post-conflict country’s chance of relap-
sing to conflict jumps from 27% to 50% if it does not
fully recover in the first year of peace. However, a polit-
ically interesting question remains: why do some states
recover quickly while others do not? Some argue the
answers are economic: it may depend on the relative
levels of capital and labor destruction (Barro & Sala-I-
Martin, 2004), the per-capita income at the start of
peace (Collier, Hoeffler & So
¨derbom, 2008), or the vol-
ume of capital that fled the country during conflict and
how quickly it returns (Collier, 1999). Others advocate a
political story. Autocracies may be better rebuilders than
democracies (Collier, Hoeffler & So
¨derbom, 2008),
negotiated settlements may lead to slower growth than
outright military victories (Flores & Nooruddin, 2009;
Atlas & Licklider, 1999), post-conflict democratic tran-
sitions may slow growth (Flores & Nooruddin, 2009),
and post-conflict elections may slow growth and increase
the risk of resurgent violence (Flores & Nooruddin,
2012; Collier, Hoeffler & So
¨derbom, 2008).
These state-level explanations do not easily scale to
the settlement level: why might outcomes be different
within a country? This article posits that social influences
are key to understanding substate variation in recovery.
A robust development literature suggests that ethno-
religious diversity slows economic growth in peacetime
(see e.g. Easterly & Levine, 1997; Alesina & La Ferrara,
2005), and ethno-religious identities can become espe-
cially salient after conflict (see e.g. Simonsen, 2005).
However, ethno-religious diversity has received little
attention in post-conflict empirical literature. The lim-
ited existing discussion is confined to cross-country com-
parisons and divided: Kang & Meernik (2005) find that
ethnic fractionalization has a significant negative impact
on post-conflict growth rates, but Collier, Hoeffler &
So
¨derbom (2008) find that ethnic diversity is insignif-
icant with respect to the duration of post-conflict peace.
This article suggests that having ethno-religiously
diverse neighbors slows a town’s post-conflict recovery.
A town has ‘diverse neighbors’ if towns with different
plurality ethno-religious groups are nearby. The effect is
strong and unambiguous when the groups involved were
‘on opposing sides’ of the conflict, but this effect may
extend – albeit perhaps less powerfully – to groups that
were wartime allies. The discussion section accesses sev-
eral explanations for this result and suggests that the
types of post-liberation controllers that arise in diverse
areas – which tend to be substate militias rather than the
government – is one important mechanism.
Two caveats are in order. First, this article does not
directly examine the effect of within-town diversity. Its
claims are limited to the effect of towns being in diverse
areas. Second, there are many reasons some post-conflict
towns rebuild quicker than others. Proximity to diverse
neighbors is just one piece of the puzzle. This article’s
primary contribution is empirical; however, the final
section discusses mechanisms.
Case selection: Post-ISIL Iraq
To test its hypothesis, this article considers Iraq’s recov-
ery from the Islamic State insurgency. This section
544 journal of PEACE RESEARCH 59(4)
To continue reading
Request your trial