District magnitude and representation of the majority’s preferences—a reply and new perspectives

Abstract

We reply to the comment of John Carey and Simon Hix on our original contribution entitled "District Magnitude and Representation of the Majority's Preferences: QuasiExperimental Evidence from Popular and Parliamentary Votes" in Public Choice 151: 585-610 (2012). District magnitude does not necessarily affect deviations between political representatives and their district voters in a strictly monotonic way but monotonicity is upheld for deviations between representatives and the national majority. We provide new perspectives and caution against evaluating electoral systems by focusing on individual politicians' behavior but neglecting aggregation effects. CH group together districts (Swiss cantons) with different magnitudes by means of separate dummies and regress one of our two original measures of divergence between representatives and voters on these dummies. The results show that individual politicians from small but multi-member districts represent the respective majority of their cantonal voters not only better than politicians from large multi-member districts (our original result) but also better than individual politicians from single-member districts (CH's refinement). Keywords

Voting against the separation of powers between legislature and administration

“…We pursue a straightforward empirical approach to disentangle the behavior of parliamentary representatives from the preference of the constituents regarding the separation of powers between the legislature and the administration. As argued already by Schneider et al (1981) and discussed more recently by Portmann et al (2012Portmann et al ( , 2013, among others, constituents in Switzerland reveal their preferences for policy proposals in referenda. The proposals are implemented immediately if the majority of voters approves them, i.e.…”

“…Only very few studies have performed a referendum-by-referendum comparison between voters and representatives in the past (Hersch andMcDougall 1988 or Garrett 1999). However, the match between referendum decisions and legislators' roll call votes has been revived by Stadelmann et al (2012 and Portmann et al (2012Portmann et al ( , 2013 who all rely on comparing the voting behavior of representatives to the Swiss parliament with the voting behavior of citizens in referenda on exactly the same issue. 5 Recently, Brunner et al (2013) employed a broad set of referenda in California to investigate effects of income on representation and Potrafke (2013) analyzes parliamentary votes and referenda on concert halls in Germany cities.…”

“…A referendum is mandatory for any amendment to the constitution and citizens may demand a popular referendum on parliamentary decisions when laws are concerned. Importantly for our case, citizens may also propose constitutional amendments by demanding an initiative (for details see Portmann et al 2013). Referendum decisions reflect revealed preferences for policies, as they permit constituents to rank referenda against the status quo (see e.g.…”

“…The results of Portmann et al (2012) regarding the influence of district magnitude on the quality of representation and the possibility of diverging from constituents' preferences were extended by Carey and Hix (2013) and reanalyzed by Portmann et al (2013). Hix (2011, 2013) suggest that the leeway of representatives does not increase linearly in district magnitude but that there is an electoral ''sweet spot'' in small but not single-member districts within proportional electoral systems where congruence between members of parliament with their constituents' may be expected to be highest.…”

More federal legislators lead to more resources for their constituencies: Evidence from exogenous differences in seat allocations

“…4 It is noteworthy that increasing the number of representatives from the same constituency makes it more difficult for each of them to be recognized for their personal commitment (e.g. Lancaster 1986;Carey and Shugart 1995;Lizzeri and Persico 2001;Portmann et al 2011;Carey and Hix 2011). Thus, the law of 1/n is more likely to hold when overall electoral district magnitude is small and individual responsibility of legislators more clearly defined (Milesi-Ferretti et al 2002;Edwards and Thames 2007;Primo and Snyder 2008;Carey and Hix 2013).…”