11
changes will produce opposite reactions, and an overall conclusion becomes more problematic.
If conflict is defined by the sum of Ah and Am, then it could continue to increase even in the
deterrence phase.14 It is unclear, however, whether the sum is a more meaningful measure than
some strictly quasiconcave transform of the two activism levels. For instance, consider the case
in which the number of militants by one side is so large that the other side is practically devoid
of activists of its own. There may be lots of mobilized militants but little street fighting.
The ultimate expression of a quasiconcave transformation is, of course, the Leontief specifi-
cation: overall conflict is related to the minimum of the two activism levels. In this case, overall
conflict will be maximal precisely when the two groups are equally aggressive. More generally,
one would expect that conflict — appropriately defined — would be maximal at some interme-
diate group configuration, in which one side is distinctly dominant, but not overwhelmingly
so.
4.2. Between-Group Heterogeneity. Consider three possible types of parametric changes that
can modify the heterogeneity between the two groups while preserving within-group hetero-
geneity: in radicalism, population, and wealth. By and large, the results are not surprising.
Let us start with the case in which radicalism increases across the board for a group. From
the first order condition (3) it is immediate that each individual of each type will increase his
contribution to conflict. Since for every degree of activism by the opponent the radicalized
group will contribute more resources and hence more activists, it follows that activism by that
group will increase, both as an equilibrium response and in equilibrium itself.
The case of an increase in population is even simpler. Each individual optimal contribution
will remain the same. The increase in the number of contributors will permit this group to
mobilize a larger number of activists.
The one change that does merit additional discussion is the case of a uniform increase in
wealth for a group, and by extension, the case of changing inter-group inequality. To study
this, suppose that all wealths or earning capacities within a group are multiplied by the same
proportionality factor. There are two effects. First, religious intolerance or radicalism becomes
an easier game to play: the marginal cost of contributions to the cause goes down. Fixing for the
moment the degree of activism by the opposing group, this change will increase the demand for
activists within the group in question, for any given compensation rate. At the same time, there
is a fall in the number of activists available at the old compensation rates: if earning capacities
are higher, then so must be the monetary compensation for activism. Ideally, the wealthy in such
groups would love to have continuing access to a low-wealth source of labor. While this desire
may be more easily fulfilled when within-group inequality is altered (see below), a proportional
increase in group wealth does not permit this luxury. Hence, the supply curve of activists will
be shifted upwards.
It is immediate that the response of societal conflict to inter-group changes in wealth inequal-
ity (that are evenly divided within groups) may be complex. The new intersection of the two
curves can produce either a higher or a lower number of activists as an equilibrium best re-
sponse. In particular, it is quite possible that increased inter-group inequality can bring down
equilibrium conflict.
In order to examine this in more detail, observe that if all incomes of one group have been
multiplied by a factor of λ so will the aggregate cost curve. Hence, we have to check whether
the supply of funds will also result multiplied by the same factor. Let us go back to (3) where
14It can be checked that this assertion is always true when p(Ah, Am) takes the special form Ah/(Ah + Am).