The Rationality of Serb Leaders in the Bosnian War

Abstract

This paper asks whether Bosnian Serb leaders’ choice to carry out a secession war in 1992–1995 was rational from the point of view of their stated goal of ethnic cleansing. We construct two indexes, one of ethnic purity and another of ethnic Serb concentration, and apply them to a counterfactual estimate of the outcome of ‘peaceful’ ethnic cleansing – what could have been achieved by population exchange based on pre-war territorial Serb power without war – in comparison to the actual outcome of the war. We find that the gross benefits of the chosen strategy of secession and war far exceed anything that could be achieved by the peaceful alternative. A conjectural assessment of perceived costs suggests that also net benefits were maximized by the war strategy. The implication for international deterrence policy is that credible judicial prosecution and punishment is the best way to alter the prospective perpetrators’ calculus.

Keywords:

• Bosnian war
• Ethnic cleansing
• Rational choice
• Cost-benefit calculus
• International deterrence

Acknowledgements

An earlier draft of this paper was presented at the Second Workshop on the Political Economy of Conflict, Catholic University of the Sacred Heart, Milan, 22 October 2013, whose participants provided interesting discussion. I am indebted to Stefano Costalli for making Tabeau et al.’s (2003) report available to me, and to a referee of this Journal for helpful comments.

Notes

¹ See Kaufman (2001, 2006) for a discussion and critique of the ‘rational choice’ theories of ethnic war and genocide. The case studies he addresses, however, include the dissolution of Yugoslavia and the war in Croatia, but not the Bosnian war.

² This paper is a contribution to a larger effort (Ferrero 2013a, 2013b) to place ethnic cleansing in a continuum of actions that a ruling power can take to remove an unwanted group from a society, and which include assimilation, deportation, exile and extermination. In that approach, the rulers are said to be rational in the sense that they will choose the combination of means that can achieve their goal of group removal at the minimum cost to themselves – hence the rationality of means rather than of ends.

³ This information, as well as the data in Table II, can be found in a Wikipedia (2014) entry (‘Demographic History of Bosnia and Herzegovina’), which however does not provide any reference or source; this 1996 census is also mentioned in several other Wikipedia entries on related subjects. I tried to find the original document underlying the entry by writing to the UNHCR archives, but my request was turned down on the grounds that their external access policy denies anyone access to any internal document that is less than 20 years old (documentation of this correspondence is available upon request).

⁴ Tabeau et al. (2003, 7, 12) explain that among the population eligible to vote in 1997 by the 1991 census (i.e. those born before 1980), 75% did register; of these, about 80% were ‘linked’ (i.e. their individual records in the voters register were matched with corresponding records in the 1991 census) and included in the data-set, while the remainder could not be linked mainly because of spelling mistakes in their names and had to be excluded. The researchers actually merged two voters registers, one from 1997 and another from 1998, to include a small number of new registrations that occurred in 1998; so the data-set essentially describes the situation in 1997 – hence the label used in the text.

⁵ In 1991, Bosnia and Herzegovina consisted of 109 municipalities. To make administrative divisions consistent with the post-war, inter-entity boundary line, the Dayton Agreement of 1995 split several pre-war municipalities into two parts, with one part belonging to the RS and one to the Federation; as a consequence, the post-Dayton municipalities numbered 149 as of 1997.

⁶ Population figures by municipality are from Zavod za Statistiku Bosne i Hercegovine (1991). The list of municipalities with per cent ethnic compositions from the 1991 census can also be found in Tabeau et al. (2003, 214–215, Table IV(B4)).

⁷ Excluding the Brčko district, for which we were not able to sort out who the returnees were.