^aCevipol, Université Libre de Bruxelles (ULB), Avenue FD Roosevelt, 39, B-1050 Brussels, Belgium. E-mail: jpilet@ulb.ac.be

Abstract

Most of the time, electoral reforms are seen as decisions taken by political players aiming at increasing their seat share.¹ According to Benoit's model, parties rank reform plans according to whether they maximize their share of seats (Benoit, 2004). In that respect, the decision of the Belgian government to change district boundaries for the 2003 federal elections is rather strange, as the reform adopted has no mechanical impact on the allocation of seats among parties. This does not, however, mean that this redistricting is one of the few reforms not driven by strategic considerations. Belgian parties are driven by other forms of strategy lying under the surface. As a matter of fact, four power-related elements have played a part in the decision to change district boundaries. Firstly, the change to province constituencies was made to reduce the uncertainty of parliamentary careers caused by the two-tier system in use up to 2003. Secondly, promoters of redistricting tried to introduce a new campaign structure, which they hoped would be more appropriate for their party. Thirdly, large parties required to introduce a 5% threshold jointly to the redistricting. The threshold was meant to block smaller parties and consequently to favour larger ones. And finally redistricting was also pushed by centralized parties expecting this reform to increase their control over decentralized party structures. All these strategies show that in the study of electoral reform, political science will have to go beyond merely considering its impact on each party's share of seats.