such information only emerges after several election rounds and the development of an effective press. (Cox 1997 69-92; Dawisha and Deets 2006; Duch 2002; Filippov, Ordeshook, and Shvetsova 1999, 8-10; Kostadinova 2002; Reich 2004; Tavits and Annus 2006; Turner 1993) A third group of scholars also found transitional party systems to be more fragmented than established ones but is skeptical that political learning will uniformly and swiftly give rise to strategic voting. They pointed out that party system disequilibria can in certain cases be a permanent feature and requires consideration of non-institutional factors. (Birnir 2004; Cox 1997, 225-65; Mainwaring 1999; Moser 2001; Stoner-Weiss 2001)
Historical Induced Equilibria: A number of historically sensitive scholars point out that voters and politicians already can hold important mutual expectations of each other prior to the first election; they show how long-term historical legacies (i.e. historical cleavages, prior regime type) and short-term starting conditions (i.e. type of democratic transition, timing and sequencing of founding elections, prominence of external actors) shaped such ex ante expectations. (Benoit 2004, 384; Cox 1997, 203-37; Kitschelt et al. 1999, 19-42; Mainwaring 1999, 63-88; Ordeshook and Shvetsova 1994; Reich 2001; Stepan and Linz 1992) Legacies and starting conditions most directly influence the (re-) formation of parties and thus generate important ex ante information about their electoral viability prior to the first national election. Such information determines the effective number of electoral contestants (ENEC). ENEC differ from ENEP or ENPP in that their electoral viability is shaped by a wider array of factors than just prior election outcomes.
Historical factors have received less attention than institutions in part because they pose methodological problems. In varying a great deal, they confront methodologist with the many variables, small N problem. This makes them unsuitable for theoretical parsimony and large N, variable-centered research designs. Instead of raising questions about the strength of their co-variation, historical factors raise issues about which ones actually matter, specifying how they are transmitted across time and determining how long their effect lasts. (Abbott 1988, 177; Grzymala-Busse 2002, 20-23; Kitschelt et al. 1999, 12-13) Paul Pierson, for example, points out that not all initial conditions are as likely to shape subsequent developments; conditions where large set-up costs, learning