16
an escalation of conflict. The empirical findings are far from clear. Our paper suggests that the
lack of clarity may arise from at least three different sources.
First, given any changes in the relevant parameters, whether or not conflict escalates on both
sides depends critically on whether the group experiencing the change is the less conflictual of
the two to start with. Therefore, we should not expect significant empirical results unless this
factor is explicitly controlled for. Indeed, we have attempted to argue that evidence of escalation
or deterrence can tell us something about the identity of the party that experiences the initial
change.
Second, in a multidimensional world we have to start by identifying the particular dimen-
sion that is key to each particular conflict. Specifically, even though wealth always plays an
important role in conflict, it does matter whether a specific conflict is fundamentally played out
along class or ethnic lines. In particular, while a reduction of within-group inequality in wealth
might increase group cohesion in class conflict, it might actually weaken aggressive impulses in
ethnic conflict. It follows that we cannot have the same kind of relation between inequality (or
income-based polarization) and conflict for all types of conflicts.
Finally, our model shows that even in the case of an ethnic conflict, wealth plays a dual role.
On the one hand, wealthier individuals may be predisposed to contribute more money to the
cause. However, wealthier individuals are harder to mobilize. The net balance between the
two opposing effects in the event of a an overall increase in wealth critically depends on the
distribution of wealth gains and not just the average gains.17
These last two points are related to a literature on modernization and conflict. There is evi-
dence — Melson and Wolpe (1970), Olzak and Nagel (1986) and Tellis, Szayna and Winnefeld
(1998) — that the process of modernization might ignite rather than moderate conflict. Our
model allows to examine the implications of a general increase in wealth. In particular, we ar-
gue that the distribution of the benefits of this process of modernization turns out to be critical to
the direction of the change in conflict.
The distributional issues have two facets: across groups and within groups. Wintrobe (1995),
Robinson (2001) and Stewart (2002) put the emphasis on across-group inequality as a major
explanatory factor for an escalation in ethnic conflict. Yet, as suggested by our analysis, a very
uneven distribution of the benefits across groups, but very egalitarian within the beneficiary
group, might have a limited effect on conflict. The additional resources might fail to mobilize a
larger number of activists because their individual opportunity cost will be higher too.
On the other hand, Horowitz (1997) and Bates (1999), argue that increased within-group in-
equality has to have a major role in igniting conflict. Our results are aligned with Horowitz and
Bates in that an increase in within group inequality will give more resources to the contributors
without increasing (and perhaps even decreasing) the cost of mobilizing people. It follows that
in these circumstances conflict will go up.
Finally, our model underlines a neglected factor: the correlation between religious passion
and wealth. Higher correlation precipitates more conflict. It seems interesting to investigate
whether the data confirm this prediction.
Our model can be improved in many directions. In the first place, we need an explanation for
the origin of the very alternatives that are put forward to the population. Why might an ethnic
goal come to the forefront, rather than the more “classical” class objective of seizing economic
power? Or is it economics all along, except that in some cases economic goals are better served
17
17In addition, it depends on the exact form of individual preferences, a point that we also note in the paper.