The global economic burden of violent conflict
| Date | 01 March 2022 |
| DOI | 10.1177/00223433211046823 |
| Author | Anousheh Alamir,Olaf J de Groot,Tilman Brück,Carlos Bozzoli |
| Published date | 01 March 2022 |
| Subject Matter | Regular Articles |
The global economic burden
of violent conflict
Olaf J de Groot
Office of the UN Resident Coordinator
for Jamaica, Bahamas, Bermuda, Turks
and Caicos & Cayman Islands
Carlos Bozzoli
Fundacio
´n Bunge y Born
Anousheh Alamir
European Center for Advanced Research in
Economics and Statistics, Universite
´Libre de
Bruxelles
Tilman Bru
¨ck
Natural Resources Institute, University of
Greenwich & ISDC - International Security and
Development Center & Leibniz Institute of
Vegetable and Ornamental Crops & IZA &
Households in Conflict Network
Abstract
Calculating the consequences of global public bads such as climate change or pandemics helps uncover the scale,
distribution and structure of their economic burdens. As violent conflict affects billions of people worldwide, whether
directly or indirectly, this article sets out to estimate its global macro-economic repercussions. Using a novel methodology
that accounts for multiple dimensions of war, the article finds that, in the absence of violent conflict since 1970, the level
of global GDP in 2014 would have been, on average, 12% higher. When disaggregating these results by conflict type, civil
conflicts are estimated to have been the costliest by far. Income growth is found to be altered up to four years following the
end of a conflict, although the direction of this relationship depends on the intensity and type of conflict. Countries also
suffer significantly from fighting in neighbouring countries, thereby showing the importance of mitigating spillovers
rapidly. The largest absolute losses associated with violence emanate from Asia, while many high-income economies are
found to benefit economically from participating in conflicts on foreign soil. This analysis thus shows that, despite some
evidence of a faster post-conflict growth and possible benefits for external participants, violent conflict leads to net global
losses that linger long after peace is achieved, reducing the peace dividend. The article concludes by discussing public
policy options to strengthen the benefits of peace as a global public good.
Keywords
conflict spillover, costs of conflict, growth, internal armed conflict, international war, peace dividend
Introduction
Public bads cost society: HIV/AIDS (Veenstra &
Whiteside, 2005), malaria (Sachs & Malaney, 2002), alco-
holism (Rehm et al., 2009) and diabetes (Yach, Stuckler &
Brownell, 2006) reduce economic growth, as do air pollu-
tion (Larsen, 2014), climate change (Stern, 2007; Hsiang
& Jina, 2014; Kalkuhl & Wenz, 2020), soil degradation
(Pimentel et al., 1995), financial crises (Cerra & Saxena,
2008; Inklaar, de Guevara & Maudos, 2012) and, recently,
COVID-19 (Fernandes, 2020; McKibbin & Fernando,
2020).
Likewise, violent conflict burdens society both eco-
nomically and socially (Collier, 1999; Verwimp, Justino
Corresponding author:
t.brueck@greenwich.ac.uk
Journal of Peace Research
2022, Vol. 59(2) 259–276
ªThe Author(s) 2022
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/00223433211046823
journals.sagepub.com/home/jpr
&Bru
¨ck, 2009, 2019; Mueller, 2013; Serneels &
Verpoorten, 2015; Bove, Elia & Smith, 2017; Costalli,
Moretti & Pischedda, 2017), though the aggregate, global
scale of this burden remains unclear (IEP, 2016; OCHA,
2017).
Past studies have shown that a country’s economic
growth can be influenced by conflicts occurring both on
its own territory (Polacheck & Sevastianova, 2012) and on
its neighbours’ (Murdoch & Sandler, 2002a,b, 2004; de
Groot, 2010; Dunne & Tian, 2014), and this influence
can be prolonged long after the conflict is ended (Organski
& Kugler, 1977; Collier, 1999; Kang & Meernik, 2005).
The direction, period and factors determining these rela-
tionships are, however, still debated. Furthermore, little is
known about the differential consequences for the various
types of conflict participants.
We thus set out to fill this knowledge gap. Using a
panel of 190 countries from 1970 to 2014, we estimate
the average consequence of various dimensions of con-
flicts on yearly GDP growth. First, we categorize all the
major conflicts
1
that took place during the observed
period into three types: civil (between a state and a
non-state actor), interstate (between two states) and
non-territorial (participationinacivilorinterstatecon-
flict taking place entirely on foreign territory). Second,
we analyse the number of years following the end of
each type of conflict, during which a significant rela-
tionship continues to be observed with growth. Third,
we introduce a spillover variable that measures the pos-
sible influence of each type of conflict taking place up
to 1,000 km away from a country’s border. Finally, as
we posit that the economic consequence of each of
these conflict dimensions depends on the intensity of
the episode, we use an index that quantifies conflict
intensity on a scale from 0 to 10, based on a variety
of factors.
Using a country-fixed effects estimation and cross-
checking the robustness of the results with other spec-
ifications, we find that an additional unit of civil or
internationalconflict intensity reduces the ‘host’ country’s
yearly GDP growth by an average of 0.9% during the
conflict years. On the other hand, we find a U-shaped
relationship between conflict intensity and GDP growth
for externalparticipants. Secondly,civil conflicts are found
to have a U-shaped influence on growth up to four years
following the end of a conflict, while international
conflicts seem to have a positive influence up to two years
later. This thus provides some evidence for Organski &
Kugler’s (1977) ‘phoenix’ effect, whereby post-conflict
nations grow faster than their peaceful counterparts.
Finally, spillovers from civil and international conflictsare
found to be strongly detrimentalto growth, while having a
neighbour participating externally in a conflictis found to
be economically beneficial.
As a second step, we provide a methodology to esti-
mate what the yearly growth rate would have been in the
absence of conflict. We then use this estimated rate to
provide a year-by-year counterfactual GDP level for each
country. We consequently find that, had there been no
conflict throughout the world since 1970, global GDP
would have been 12% larger in 2014.
Disaggregating our main results, we find that Asia has
suffered the most from conflict; while many developed
economies in North America, Europe and Oceania are
found to have benefited from their participation. A
potential explanation for this is that external state actors
intervene in armed conflicts primarily out of their own
economic and national security interests (Aydin, 2012;
Bove, Gleditsch & Sekeris, 2016; Bove, Deiana &
Nistico
`, 2018).
Our inquiry into the global economic burden of vio-
lent conflict is structured as follows. First, we illustrate
the conceptual framework used for the analysis. Next, we
identify the datasets and methodology used for our
empirical estimation. We then present our results, dis-
cuss their robustness and compare them with previous
findings before concluding.
Conceptual framework
Although there is a clear consensus that ‘hosting’ a conflict
is damaging to growth in the short term (see e.g. Bozzoli,
Bru
¨ck & Sottsas, 2010 for a comprehensive survey on the
topic), the growth reactions to being near a conflict, parti-
cipating externally and ending conflict remain unclear. We
thus analyse the global dynamics surrounding conflicts to
estimate their net global consequences over 45 years of
observation. While this is an empirical exercise, we first
consider what can be learned from the theoretical literature
to help construct our empirical model.
As the literature is mostly conclusive regarding con-
flicts occurring within a country’s own territory, we
focus on this case to graphically illustrate our theoretical
framework. We discuss details of the concepts intro-
duced here below and alternative scenarios in the follow-
ing subsections.
1
As per our data source (Marshall & Cole, 2014), we define conflict
as having at least 500 directly related fatalities throughout the
conflict, and 100 deaths per annum.
260 journal of PEACE RESEARCH 59(2)
To continue reading
Request your trial