Workers , Warri ors and Cr i mi nal s: So c ial Conflic t in
Gen er a l E q uilib r iu m
Ernesto Dal Bó
U.C. Berkeley
Pedro Dal Bó
∗
Brown Un iversity
September 22, 2004
Abstract
We analyze how economy-wide forces (i.e.shocks to te rms of trade, technology and
endowments) affect the intensity of social conflict. We see conflict phenomena such as
crime and civil war as involving resource appropriation activities. We show that not
all shocks that could make society richer will reduce conflict. Positive shocks to labor
intensive industries will diminish social conflict, while positive shocks to capital inten-
sive industries will increase it. The key requirement is that appropriation activities be
more labor intensive than the economy. Our model can explain the positive association
between crime and inequality, and the curse of natural resources; it predicts that aid
in kind to war-ridden societies will have perverse effects, and offers guidance on how to
integrate international trade policy and peacekeeping efforts. Including appropriation
activities into a canonic general equilibrium model introduces a social constraint to
policy analysis. Thus, we can also account for populist policies, apparently inefficient
redistribution and “national development strategies”.
JEL Classification: D72, D74, D78, F13, H23, K42, O1.
Keywords: conflict, civil war, crime, social constraint, populism, trade policy, inef-
ficient redistribution.
∗
We thank Rafael Di Tella, Herschel Grossman, Juan Carlos Hallak, Ben Hermalin, Daniel Heymann,
Edward Miguel, and participants in various conferences and seminars.
1
1 Introduction
One crucial aspect of social life is the conflict over the distribution of resou rces. The divi-
sion of wealth amon g individuals is not solely determined by a price system operating on
the basis of well defined and perfectly enforced propert y rights. In reality, expropriatory
efforts pla y an important role, and tak e various forms. Sometimes an agen t will engage in
criminal activities—either acting alone or as part of a group—with the undisguised intention
to expropriate goods from others. Some other times individuals may join a guerrilla that is
motivated b y ideology, but that in fact is (at least partly) fueled b y the group’s ability to
appropriate resources.
Economics has historically analyzed phenomena lik e crime, on the one hand, and revolts
and political instability, on the other, along separate lines.
1
These phenom ena, how ever,
recognize a common root: they are symptoms of social conflict in the sense that they express
individual and group pressures for the appropriatio n of resources. Th is paper offers a simple
general equilibrium framework to study ho w economy-wide forces affect the exten t of social
conflict seen as an appro priation -based phenom enon .
2
Both microfo u nded theoretical models and empirical evidence suggest that, all else equal,
a lower opportunit y cost in terms of w a ges in the labor mark et should increase the chance
that an individ ual engages in activities suc h as rebellion or crime.
3
This knowledge alone,
ho wever, is not sufficient to predict how aggregate economic shocks or policy int erven tions
will affect the extent of social conflict. One ke y reason is that usually all else is not equal:
real life shoc ks affecting the opportunit y costs of conflict also tend to affect the returns to
conflict and vicev ersa. Our model provides an integr ated view of ho w the costs and benefits
to conflict activities mov e in response to shoc ks (or policies). Thus, we can mak e predictions
on how suc h shocks will affect phenomena lik e crime, and accoun t for stylized facts of civil
wars.
1
Classic references in the literature on crime are Beck er (1968) and Ehrlich (1973). Various references to
work on revolts are given below.
2
Whether the fight for resources takes the form of atomistic criminals or large, politically organized
factions will depend on many factors shaping what we could call the industrial organization of social conflict.
In this paper we choose to abstract from all issues regarding the number and size of competing groups, as well
as from strategic interactions, to focus on what we deem to be more basic aspects, such as the determinants
of the relevant costs and benefits to the expropriatory efforts.
3
There is evidence that a higher income per capita is associated with a lower likelihood of civil confl ict
(Collier and Hoeffler, 1998, Elbadawi and Sambanis, 2002, Fearon and Laitin, 2003; Miguel, Satyanath and
Sergenti, 2004), and MacCulloch (2001) finds that higher household income diminishes the propensity to
express support for a revolt. Empirical studies on the relationship between wages and crime suggest that
higher wages deter participation in criminal activities (see Grogger, 1998 and Gould, Weinberg and Mustard,
2002).
2
In particular, our model shows that wealth-increasing shoc ks ma y both increase and
decrease conflict. W hat really matters in our model is the relative factor in ten sity of the
sectors in the economy that are initially affected. Our general equilibrium model of a small
open econom y comprises t wo productive sectors and a third sector w e call “appropria tion.”
This last sector expropriates a fraction of what is produced in the two productive industries.
Assum ing that appropriation is labor intensive relativ e to the whole econom y, we sho w that
an exogenous increase in the price of the capital in tensiv e good will cause the appropriation
sector to expand. (Conv ersely, a decrease in that price or an increase in the price of the labor
intensiv e good will lead to a smaller approp riat ion sector.) The reason is that an increase in
the price of the capital intensive good will expand the capital intensive industry, and contract
the labor in ten sive one. This makes labor relatively less scarce, resulting in a lo wer cost of
the appro priation activity relative to the amount of appropriable resources. Tech nolog ical
progress, whic h ma kes society ric h er, has similar effects: neu tral tech nical progress in the
labor in ten sive sector will decrease conflict but prog ress in the capital in te nsive sector will
increase it. The effects of c ha nges in endowments can also be analyzed: an increase in
the cap ital end owment, for instance, will increase conflict. We also show that the social
backlash of appropriation activities can be so strong that shock s that w ould make a conflict-
free economy ric h er will leave every body wo rse off.
The model is usefu l to explain empirical patterns in crime and civil wars. First, it can
account for the positive association between crim e and inequality.
4
Capital and labor in our
model can be interp reted to represent differently skilled labor forces. An y favorab le shock to
the skilled labor in t ensive sector will increase the wage gap across skill levels and will increase
conflict whenever appropriation is relativ ely inten sive in unskilled labor. Also, in our model
these mo vements are compatible with an increase in total production. Thus, our model can
explain why reforms that increase income per capita ma y be associated with increasing crime
and inequality. An example is that of Argen tina, where sw eeping market-friendly reforms
where introduced starting in 1991.
5
The model can also accou nt for st ylized facts of civil w ars that appear paradoxical. On the
one hand, there is the abo vementioned evidence that a higher income per capita is associated
with a lo wer likelihood of conflict. On the other hand, ho wever, not all wea lth-increasin g
forces appear to diminish conflict: Collier and Hoeffler (1998) report that a larger availabilit y
4
Fajnzylber, Lederman and Loayza (2002) document this connection across countries. There is also
evidence that income inequality increases individual propensity to express support for a revolution (see
MacCulloch, 2001).
5
The following years saw important increases in income per capita, inequality and crime in the country.
GDP per capita increased by 40% between 1991 and 1998. The Gini coefficient wen t from 44.7 to 49.5
between 1992 and 1998, and crimes against property increased by roughly 71% between 1991 and 1998.
3
of natural resources can increase the risk and duration of civil w ar. Our model can explain
these patterns. It predicts that a shock such as a drought—a negative shoc k to agricultural
productivity and income—will fuel conflict in countries where agriculture uses relatively little
capital, as in Subsa haran Africa (see Miguel et al, 2004). But our model also predicts that
tec h nical progress or new reserve disco veries in oil or diam on ds will increase conflict when
these extractive industries are relativ ely capital intensive, regardless of the fact that these
shocks should also increase income.
Over the last few decades, the develop m ent of Po litical Economics has allow ed us to
incorporate political constra ints into theory and policy analysis. In a similar spirit, our
framework brings a social constraint to bear on the analysis of economic policy. We find
that taxing capital and subsidizing productive labor can make both w orkers and capitalists
better off: although capitalists lose when pay in g taxes they may gain mo re from less in ten se
expropriation. When administering such tax-subsidy schemes is difficult (weak states may
ha v e trouble identifying who is really doing productive w ork), the promotion of unproduc-
tiv e public employment can be P areto-im p r oving. Th u s, some forms of populism, includin g
redistribution through w asteful ch an nels, can be accounted for as rational responses to envi-
ronments with social conflict.
6
Similarly, policy in terven tio ns that shield the economy from
certain shoc ks to term s of trade and technology may become desirable. Suc h interven tions
can include (i) trade protection of labor intensive industries and (ii) sub sidizin g techni-
cal progress in labor intensive industries while discouraging the adoption of foreign-based
innovations that migh t be biased to wards the capital intensiv e sector. Th ese results ma y
help rationalize the political support often enjoyed b y policies that protect labor intensive
industries. They also resonate with proposals b y dev elopment theorists aligned with the
Latin America n structuralism (see for example Prebisch, 1959, on how national dev elop m e nt
strategies could use selectiv e interven tions).
The policy applications of the model reach the int ernatio nal efforts to stop violen t con-
flic t in the third world. One imp lication of our model is that aid in kind may have perverse
effects, because it does not affect the returns to labor, but it increases the mass of lootable
resources. Also, peace effo rts may need to be coordinated with international trade agree-
ments. Otherw ise, the trade policies ch osen by industrial powers may undermine their peace
efforts abroad. To the best of our kno wledge, our model pro vides the first framew ork with
which to analyze the integration of trade and peace policies.
Our approach abstracts from sev eral factors that ma y affect conflict. Exam ples are income
effects, un em ployment, and additional connections bet ween natural resources and conflict.
7
6
See Coate and Morris (1995) and Acemoglu and Robinson (2001a) for different explanations of inefficient
redistribution mechanisms.
7
Extractive activities, for example, tend to be associated with monopolic rents. These in turn are as-
4
Previous papers on conflict(seeforinstanceGarfinkel, 1990; Grossman, 1991; Skaperdas,
1992; Hirsh leifer, 1995; Grossman and Kim , 1995; Acemoglu and Robinson, 2001b; Powell,
2004) ha ve studied a variet y of problems w e do not deal with. Examples are the determinants
of militar y expenditures in repeated in tera ction s among coun tr ie s, the failure of cooperation
in set ups where coercion is possible, and the determination of in vestment in production
versus offensive and defensive capabilities.
Our model can be though t to hav e implications for ren t-seeking. In particular, our
appropriation sector could be though t to capture ren t-seeking efforts under exogenous tariffs
(on this topic, see for instance Krueger, 1974; and Bhagwati and Srinivasan, 1980). Besides
differences in focus and modeling choices, our theory differs from Kruger’s and Bhagw ati and
Srinivasan’s in that our theory reserv es a role for the relativ e use of factors across industries.
The plan for the paper is as follows. Section 2 presents the model. Section 3 char acter izes
the equilibrium after proving its existence. Then a compa rison is made with the equilib riu m
in a conflict-free society. Section 4 studies ho w economic shoc ks affect the extent of social
conflict, thus establishing our centra l results. Section 5 establishes further results and dis-
cusses policy implication s. Section 6 extends the basic model to the case of industry-specific
factors. Section 7 concludes.
2 The model
Consider an economy comp rising two productiv e sectors along the lines of the canonical 2x2
international economics model.
8
The productive sectors involve many firms which maximize
profits and use tech nologies cha racterized b y constan t returns to scale. In a competitiv e
equilib r iu m profits are driven to zero. The t wo productiv e sectors or industries are labeled 1
and 2, and they use two inputs, capital and labor, respectively labeled K and L.Allfirm s in
each industry share the same production function with the property that industry 1 is more
capital in ten sive than industry 2. Wedenotewithr and w the respective rental prices of
capital and labor. The given primitives of the model are: the factor endowments, a vailable in
fixed amounts
K and L; the tech no logies; and the prices of output, which are in ternation ally
determ ined and are labeled p
1
for industry 1, and p
2
=1for industry 2. (Good 2 is the
numeraire.)
In add ition to the productiv e sectors, there exists an app rop riation sector. This sector
sociated with higher corruption, and the latter can be expected to damage the quality of institutions and
economic performance. The end result may well be higher conflict. For the connection between rents and
corruption see Ades and Di Tella (1999), and for the role of institutions see Acemoglu, Johnson and Robinson
(2001). Ross (2003) lists various connections between natural resources and conflict.
8
See Stolper and Samuelson (1941) and Jones (1965).
5
