Electoral Systems and Opinion Representation

^†Previous versions of this paper were presented at the Annual Meeting of the Canadian Political Science Association, Victoria, British Columbia, 2012, the Annual Meeting of the Southwest Political Science Association, San Antonio, 2014, the Annual Meeting of the Southern Political Science Association, New Orleans, 2015, and at Arizona State University, Columbia University, the University of Gothenburg, Lund University, and the University of Mannheim.

Abstract

Electoral systems are critical to the political representation of public opinion. Most existing work has examined the differences between proportional and majoritarian systems. The debate has centered on whether proportional representation produces greater ideological congruence between the positions of the government and the public. The focus has been on the immediate aftermath of elections, however—the existing literature on electoral systems has largely ignored the fact that representation occurs (or not) in the years between elections as well. Here, we argue for a shift in focus to the impact of electoral systems on representation throughout the electoral cycle, that is, during the tenure of governments. This shift, capturing the period during governments actually govern, leads to quite different expectations.

ACKNOWLEDGEMENTS

We are grateful to Clare Devereux for research assistance, and to Hanna Back, Lenka Bustikova, Ernesto Calvo, Peter Esaiasson, Mikael Gilljam, Staffan Kumlin, Jeff Lax, Paul Lewis, Staffan Lindberg, Johannes Lindvall, Mark Ramirez, Justin Phillips, Sarah Shair-Rosenfield, and Melissa Schwartzberg for helpful comments.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

FUNDING

The project was funded in part by the Fonds de recherche sur la société et la culture (FQRSC).

Notes

^† Previous versions of this paper were presented at the Annual Meeting of the Canadian Political Science Association, Victoria, British Columbia, 2012, the Annual Meeting of the Southwest Political Science Association, San Antonio, 2014, the Annual Meeting of the Southern Political Science Association, New Orleans, 2015, and at Arizona State University, Columbia University, the University of Gothenburg, Lund University, and the University of Mannheim.

1 This is just one of several types of representation, of course. Pitkin (1967: Chs 3–5) discusses four, the others being “formalistic”, “symbolic” and “descriptive”; Mansbridge (2003) offers another useful typology.

2 We are purposefully vague in our characterisation of levels of policy—we use “more policy” and “less policy”, but we intend for these to capture various ways in which we might think about more or less spending, or regulation, or commitments of other resources.

3 Note that parties in our (hypothetical) systems do not converge on the median voter, but rather hold distinct and in some cases far-right or far-left positions vis-à-vis voters. This follows convention, for example, Powell, and we adopt it partly for expository purposes—the dynamic we wish to illustrate is clearer when the parties are more easily distinguished—but also because we regard the positioning as realistic. We explicitly address party polarisation and its consequences in some detail below.

4 The positions of parties and governments (and public preferences) are purely hypothetical, and intended only to reflect the expectations of majoritarian and proportional systems in the existing literature on indirect representation. In the case of governments' policy positions, we determine the proportional-system placements by extrapolating from party placements alongside the distribution of public preferences, assuming that citizens will vote for the party closest to their position (and thus that parties closer to the thicker part of the distribution will receive greater vote shares). Of course, elections are not decided solely by policy preferences; for expository purposes, however, we assume that they are.

5 They do, however, show that proportional systems produce more representative legislatures.

6 This view of representation is closer to that in work focused on “dynamic” representation. See, for example, Erikson et al. (2002); Wlezien (1996, 2004); Soroka and Wlezien (2010).

7 Note that some of the same logic that applies to multiparty coalitions also applies to single parties—each party is itself a coalition, after all—if to a lesser degree. It follows that some parties will exhibit more “friction” than others; and some of these dynamics will operate behind the scenes, that is, in party caucus meetings, both in proportional and majoritarian systems.

8 For our illustration, the effects of party transitions on policy change are expected to be greatest early on and diminish over time, as policy approaches the party's ideal point—this is consistent with empirical research (see, e.g., Alt [1985]; redacted).

9 Though note that there is another possibility, namely, that majoritarian systems produce an incentive for governments not just to move in parallel with preferences, but to move towards public preferences over the electoral cycle—so that majoritarian governments are closer to preferences at the end of the electoral cycle than they are at the beginning of it.

10 In this instance, preferences would be stable much as is assumed in the standard, static model described above.

11 Data are drawn directly from Table 1 in Dalton (2008); a thorough account of the polarisation measure is provided there. We categorise the following countries as majoritarian (and all others as proportional): Australia, Canada, France, the UK and USA. For the 1996–2001 period, the difference in means is not significant different (p = .28). For the 2001–2006 period, the difference in means is statistically significant at (p = .09). Note that the N equals 28 for each period. Pooling data from the two-time periods produces a majoritarian-system mean of 2.35 and a proportional-system mean of 3.27, where the difference is significant (p = .06).

12 This does not mean that factors other than electoral systems do not matter for polarisation, and these can override the centrism of majoritarian systems. Consider the USA over the past 30 years.

13 Recall from our earlier discussion that this assumes that preferences trend up or down. If they do not trend, then we would expect little difference in the average congruence across systems but would expect a difference in the variation, which should be larger in proportional systems.

14 The difference between France and Germany on the one hand and the USA and UK on the other is based on our own interpretation of Eichenberg and Stoll's findings: we simply produce standardised beta coefficients from the raw coefficients for preferences (in Table 2 of their paper), and the standard deviations in both spending and preferences (from the raw data provided by the authors).

15 Though note that they do not actually assess responsiveness to public preferences for policy, and focus instead on public mentions of the “most important problem” facing the country.

Additional information

Notes on contributors

Christopher Wlezien

Christopher Wlezien is Hogg Professor of Government and Faculty Affiliate of the Policy Agendas Project at the University of Texas at Austin. He is coauthor of Degrees of Democracy and The Timeline of Presidential Elections, and coeditor of Who Gets Represented? and The Future of Election Studies. He was founding coeditor of the Journal of Elections, Public Opinion and Parties currently is associate editor of Public Opinion Quarterly, Research and Politics, and Parliamentary Affairs.

Stuart N. Soroka

Stuart N. Soroka is the Michael W. Traugott Collegiate Professor of Communication Studies and Political Science, and Faculty Associate in the Center for Political Studies at the Institute for Social Research, University of Michigan. His research focuses on the relationships between public policy, public opinion, and mass media. He is coauthor of Degrees of Democracy, and author of Negativity in Democratic Politics.