Contemporary Conflict Analysis in Perspective 15latest article on this issue. 'Greed and grievance in civil war' contains an
interesting development pointing to the incorporation of both 'greed' and
'grievance' in a combined model." The authors move away from both con-
cepts of 'greed' and 'grievance' introducing the less controversial notions of
preferences and constraint^.^^ The concept of preferences and constraints
drastically changes the nature of the 'greed versus grievance' debate.s3
Nevertheless, the authors maintain their focus on 'the economic' rationale of
civil war implied in expected-utility: "both a greed theory and a universal
grievance theory predict that the risk of rebellion is increasing in the oppor-
tunities for rebel finance." While constraints-based theory is referred to in the
shorthand of 'greed', the authors recognise that, in contrast to preference-
based theories;"they do not necessarily literally imply that the motivation
for rebellion is exclusively, or even primarily, financial".@
For our purposes, a number of important conclusions stem from this. A
focus on the constraints that rebel organisations face to mount a credible and
effective rebellion is critical for the understanding of the onset and mainte-
nance of rebellion. By looking at the effects that changes in the levels of dif-
ferent types of constraints have on the probability of war occurrence, a clear-
er understanding of conflict life cycles may be achieved. The fmt constraint
considered refers to the size of rebel organisations in that "only large rebel
organisations generate casualties on the scale which defines civil war".ss For
this, rebel groups must mobilise people, as well as secure a large number of
weapons. They must raise finance either through extortion, donations from
diasporas or support from hostile governments. Following Collier's earlier
work, rebel groups' extortion is said to happen primarily through the plun-
dering of primary commodity exports, for the same reasons defined above.s6
Furthermore, rebel organisations face coordination costs quite different from
those that governments face.87
Turning to 'grievance rebellion', the authors focus now on preferences to
inquire the extent to which "the initiation of rebellion [is] determined by dif-
ferences in objective grievances". Here the authors look at inter-group hatred,
political exclusion and vengeance. The authors did not find that inter-group
hatred is greater in fractured societies than in homogenous ones. As before,
the crucial variable is polarisati~n.~~ Regarding political exclusion, the
authors used Ted Gurr's Polity Ill data set and concluded that "there is a veiy
large difference in the extent of democracy between conflict societies and
peaceful societies: on average, conflict episodes are preceded by a democra-
cy score less than half that which precedes peace episodes".s9 They also
investigate on the role of ethnic dominance (when one ethnic group consti-
tutes a majority, but not an overwhelming majority), where they found
unsignificant results. Marginalisation of the poor, which may be inferred by
a high degree of economic inequality, also showed no significant results in
that a survey of 15 violent civil conflicts concludes that "wars today are rarely