Scarcity and Surfeitdata, 1816-1992 (Computer file), Ann Arbor, Michigan, lnter-University
Consortium for Political and Social Research, 1994.
86 The dependence on extortion of primary commodity exports by rebel groups also
affects their size in the sense that because "primary commodities need defence of
a large physical space best suited to an army", "a viable extortion racket itself
needs considerable military power': However, the authors point out that "the
threshold of rebel force required for survival is increasing in the government's mil-
itary expenditure. As a result, the viability of rebellion need not be continuously
increasing in the endowment of primary commodity exports. Beyond some point,
the increment in potential rebel revenue may be more than offset by the increased
rebel expenditure needed to survive against augmented government forces': Ibid,
p 4. The descriptive statistical investigation that preceded regression analysis con-
firmed previous results on the role of primary commodity exports in the occurrence
of civil war as well as regarding the structure of income. Applying regression meth-
ods to these variables in order to test their validity as regards a 'greed-model' they
found a similar non-linear result as to the effect of primary commodity exports.
87 As Collier and Hoefller posit, "the government army has two advantages over a
rebellion. It can spend many years building a sense of unity, whereas if a rebel
force fails to achieve unity quickly it will presumably perish. Additionally, the
government can use the powerful rhetoric of nationalism: with this imagined
identity already occupied, a rebellion cannot afford diversity". Collier & Hoeffler,
op cit, p 7.
88 In line with J M.Esteban and D Ray, On the measurement of polarisation,
Econometrica, vol 62, no 4, pp 819-851 and M Reynal-Querol, Religious conflict
and growth: Theory and evidence, London School of Economics and Political
Science, mimeo, 2000. A non-monotonic result was found leading to the conclu-
sion that "highly fractionalised societies are no more prone to war than highly
homogeneous ones" while polarised societies have around a 50% higher proba-
bility of civil war than either homogeneous or highly fractionalised societies.
P Collier & A Hoeffler, On economic causes of civil war, pp 7-8.
89 Ibid, p 9.
90 Based on M B Anderson, Do no harm, Lynne Reinner Publishers, 1999. Collier &
Hoeffler. Greed and grievance in civil war,op cit, p 10.
91 Ibid.
92 Ibid, p 15.
93 Le Billon, op cit, p 28.
94 Ibid, p 30.
95 For an in-depth discussion of the development of the level-of-analysis problem in
international relations refer to inter alia, B Buzan, The level of analysis problem
in international relations reconsidered, International relations theory today, K
Booth & S Smith (eds), Polity Press, Cambridge, 1995. Also J D Singer, The level
of analysis problem in international relations, The lnternational system.
Theoretical essays. K Knorr & S Verba (eds), Princeton University Press, 1961.
Reprinted by Greewood Press Publishers, 1982; W B Moul, The level of analysis
problem revisited, Canadian Journal of Political Science, vol 61, no 1, 1973, pp
494-513. For an application of this framework in the discussion of the leading