The elasticity of voter turnout: Investing 85 cents per voter to increase voter turnout by 4 percent☆
Introduction
The extent to which voters turn out is a recurring theme in the public debate before and after important elections and ballot decisions. For example, the recent presidential elections in the US and France or the close ballot result in the UK to leave the EU, known as Brexit, sparked lively discussions on voter turnout and how it relates to the legitimacy of an election or ballot result. Also the academic literature has focused on voter turnout and its implications for the workings of democracy. Famously, Lijphart (1997) argued that low voter turnout biases electoral influence in favor of the already better-off citizens.
In the last decades countless reforms to increase turnout, for example, by reducing the hurdles to participate in elections have been implemented in western democracies.1 A very prominent example is the Voting Rights Act of 1965 in the US, which removed hurdles that discriminated against minorities, even though, constitutionally, all citizens had the same fundamental political participation rights (e.g., Alt, 1994). Other examples are the introduction of postal voting, as for example in all elections and ballots in Switzerland or (partially) in the US, or (trial) projects to implement online voting systems, as for example in Estonia, the US, or Switzerland (e.g., Solop, 2001, Luechinger et al., 2007, Gronke et al., 2008, Alvarez et al., 2009, Funk, 2010, Gerber et al., 2013, Hodler et al., 2015, German and Serdült, 2017).
The differences in voting procedures across countries are vast. The US, for example, requires voters’ active registration before the actual act of voting. In Switzerland, in contrast, all Swiss citizens of the age of 18 and above automatically receive all the required materials to participate in elections and popular votes. Also the procedures on Election Day differ. While a growing part of the US population has access to postal voting in all-mail, absentee, and early voting procedures (with and without excuse), still a majority of voters in the US cast their vote on a specific day, at a specific location (Gronke et al., 2008, Giammo and Brox, 2010, Gerber et al., 2013). In contrast, all Swiss voters automatically have the option of postal voting. In this case, they fill out their voting materials – which they receive by postal mail at least 3 weeks prior to the ballot – and send it back to the election authority.
As there are vast differences in voting procedures across democracies, it comes as no surprise that there exists a vast academic literature exploring these differences. It analyzes how people vote and why they vote at all, given the well-known paradox of voting.2 According to the Downsian model of electoral participation voter turnout decreases with voting costs (Downs, 1957, Tullock, 1967, Riker and Ordeshook, 1968; Aldrich, 1993, Fedderson and Sandroni, 2006). Voting costs are affected by many factors: information costs, time costs, travel costs, inconveniences such as burdensome voter registration procedures, queuing on ballot day, short or inconvenient opening hours, finding polling stations, social pressure, or the weather, etc. (E.g., Niemi, 1976, Fedderson and Pesendorfer, 1999, Highton, 1997, Highton, 2004; Knack, 1995, Blais, 2000, Haspel and Gibbs Knotts, 2005; Gomez et al., 2007, Gerber et al., 2008, Hansford and Gomez, 2010, Spencer and Markovits, 2010; Brady and McNulty, 2011, Fraga, 2011, Meier et al., 2016; Potrafke and Roesel, 2016). A rich literature searches for the institutional drivers of voter turnout, but there is still much debate on which factors have the most consistent impact and are the most important (e.g., Besley and Case, 2003, Blais, 2006, Geys, 2006, Hill, 2006, Smets and van Ham, 2013, Cancela and Geys, 2016). Often, however, studies have to rely on cross-country variation and the exploited institutional differences are not very sharp,3 or changes in voting cost are not precisely specified.
Our aim is to contribute to this literature and evaluate the impact of a very precise reduction of voting costs on voter turnout, while all else remains constant. We answer two simple yet important questions: Does a small change in the costs of voting affect voter turnout and, if yes, how does it affect party support in popular ballots?
While evidence on the influence of differences in voting procedures – for example, the introduction of postal voting – is not new, we assess the impact of the introduction of prepaid postage, which reduced voting costs by CHF 0.85 today (about USD 0.85), or the price of a stamp. Clearly, this change of voting costs is very precise from the perspective of public authorities. From the perspective of voters, however, prepaid postage might as well have other advantages affecting voting costs for individual citizens. For example, prepaid postage might also reduce transaction costs, and turnout is possible if stamps are not readily available at home and purchasing stamps at a postal office seems too burdensome. Hence, the costs of the policy are very well specified, while the actual cost reductions of voting might be higher from the perspective of actual voters.
We analyze the impact of prepaid postage on voter turnout in postal voting in Switzerland. In the Canton of Bern, some municipalities distribute prepaid return envelopes for voters who use the option of postal voting, while some municipalities do not prepay the postage. We use this simple and low-cost intervention to analyze the impact of prepaid postage on voter turnout in nationwide ballots. Using a difference-in-differences approach, we find that the introduction of prepaid postage increases turnout by about 1.8 percentage points. Our data also permit us to estimate the impact of (continuous) changes in postage costs – due to increases in stamp prices, the introduction of prepaid postage, and inflation – on voter turnout. Estimating these effects provides a notion of the cost elasticity of voting. We find that a 1 Cent (CHF) increase in postage costs reduces voter turnout by 0.022–0.031 percentage points. Furthermore, we study the effect of the increase in voter turnout due to prepaid postage on the relative party support in popular ballots, something we refer to as “voter-party alignment”. We compare the municipal voting results in nation-wide ballots with the officially announced voting recommendation of the major parties. Based on these observations we construct a measure of voter-party alignment for each of the five major parties. We find that an increase in turnout negatively affects the alignment of voters with leftist party positions.
Closely related to our research question, Luechinger et al. (2007) find that the introduction of postal voting (per se) in Switzerland increased turnout by about 4.1 percentage points on average. Hodler, Luechinger and Stutzer (2015) document a turnout increase of about 5 percentage points and an altered composition of the active voting population. Postal voting reduced the average years of education as well as the average knowledge of the ballot propositions in the voting population. Bechtel and Schmid (2016) also find an increase in turnout of about 5 percentage points and differential effects on specific groups of voters depending on income, education and genuine interest in politics. Funk (2010) finds that the increase in turnout due to postal voting was modest in the aggregate and that especially small and close-knit communities even saw a negative effect on turnout. She attributes these results to a reduced incentive to vote, as the social control at the polling station disappeared with the introduction of postal voting.4
Section snippets
The setup
Today, all Swiss citizens have the option of postal voting. They receive the voting materials with a return envelope. However, the cantons have adopted different rules with respect to whether or not they prepay the postage. In 2003, only the Canton of Geneva provided stamped return envelopes, while the Cantons of Thurgovia and Grisons required municipalities to take over the cost of postage (Federal Chancellery, 2003, Federal Chancellery, 2010). The Canton of Bern introduced postal voting on
The data
Because there is no official information on the municipal practices with respect to whether or not postage is prepaid, we collect the information directly from the 325 municipalities for the period 1989–2014.5
Empirical strategy
First, we are interested in identifying the average causal effect of paying postage and introducing Easyvote on voter turnout. We estimate a two-way fixed effects model, which is the panel data application of a difference-in-differences model. We estimate variants of the following basic model:where y is voter turnout in a municipality i on ballot day t. Prepaid postage and Easyvote are dummy variables indicating that postage is prepaid or
The effect of changes in nominal and real postage costs on voter turnout
So far we have focused on the extensive margin of whether or not a municipality introduced prepaid postage. To strengthen the link between voting cost and voter turnout, it is useful to also study changes at the intensive margin. In a refinement we take advantage of the fact that the costs of postage changed over time in nominal as well as real terms and that these changes differ for municipalities with and without prepaid postage. This allows us to estimate the price elasticity of voter
Voter turnout and voter-party alignment
In our last step we want to study whether or not political parties are systematically affected by increases in voter turnout due to the introduction of prepaid postage. In other words, whether or not some segments of the voting population are relatively more responsive to the introduction of prepaid postage.
Conclusions
We estimate the elasticity of voter turnout and show that a reduction of voting costs of as little as 85 cents per voter can have a substantial influence on voter turnout. We find that the introduction of prepaid postage in postal voting increases voter turnout by about 4 percent. A second intervention is the introduction of Easyvote, which sends out specific and easy-to-understand voting materials to young voters. We do not find a significant impact of this initiative.
Obviously, the
References (58)
- et al.
Explaining voter turnout: a meta-analysis of national and subnational elections
Elect. Stud.
(2016) Electoral consequences of declining participation: A natural experiment in Austria
Elect. Stud.
(2014)Explaining voter turnout: a review of aggregate-level research
Elect. Stud.
(2006)- et al.
Compulsory voting, turnout, and government spending: evidence from Austria
J. Public Econ.
(2017) Random group effects and the precision of regressions estimates
J. Econ.
(1986)- et al.
The embarrassment of riches? A meta-analysis of individual-level research on voter turnout
Elect. Stud.
(2013) - et al.
Only conservaties are voting in the rain: evidence from German local and state elections
Elect. Stud.
(2016) Rational Choice and turnout
Am. J. Political Sci.
(1993)When is it rational to vote?
The impact of the voting rights Act on black and white voter registration in the South
Internet voting in comparative perspective: the case of Estonia
PS Political Sci. Polit.
Mostly Harmless Econometrics. An Empiricist's Companion
Swissvotes, die Datenbank der eidgenössischen Volksabstimmungen
Does compulsory voting increase support for leftist policy?
Am. J. Political Sci.
Direct democracy, postal voting, and the composition of turnout
How much should we trust differences-in-differences estimates?
Q. J. Econ.
Political institutions and policy choices: evidence from the United States
J. Econ. Literature
To Vote or Not to Vote
What affects voter turnout?
Annu. Rev. Political Sci.
Turning out to vote: The costs of finding and getting to the polling place
Am. Polit. Sci. Rev.
Direct and Indirect Effects Based on Difference-in-differences with an Application to Political Preferences Following the Vietnam Draft Lottery
An Economic Theory of Democracy
Rational Choice theory and the paradox of not voting
J. Econ. Perspect.
Abstention in elections with asymmetric information and diverse preferences
Am. Political Sci. Rev.
A theory of participation in elections
Am. Econ. Rev.
Kantonale Normen zu den Portokosten brieflicher Stimmabgabe
Kantonale Normen zu den Portokosten brieflicher Stimmabgabe
Voting costs and voter turnout in competitive elections
Q. J. Political Sci.
Turnout matters: Evidence from compulsory voting in Australia
Q. J. Polit. Sci.
Cited by (20)
The heterogeneous effects of internet voting
2023, European Journal of Political EconomyExtending term limits, constitutional referendums and elections in francophone Africa
2023, Social Sciences and Humanities OpenPrepaid postage using pre-stamped envelopes to affect turnout costs
2021, Electoral StudiesCitation Excerpt :These replication results show an effect size varying between 1.1 and 1.3 percentage points of higher voter turnout. Like Schelker and Schneiter (2017), we find a positive and significant effect of the treatment, however, this effect is somewhat smaller than the original estimate (which varied between 1.8 and 2.1 percentage points). This difference is due to the larger sample that we examine by including municipalities in four more cantons.
Compulsory voting and political participation: Empirical evidence from Austria
2020, Regional Science and Urban EconomicsCitation Excerpt :An important issue is whether electoral institutions influence voter turnout. Research has shown that extended opening hours at polling stations and prepaid postage serve to increase voter turnout (Garmann, 2017; Schelker and Schneiter, 2017; Potrafke and Roesel, 2020). An alternative policy measure to increase voter turnout is compulsory voting.
Could rainfall have swung the result of the Brexit referendum?
2018, Political GeographyCitation Excerpt :Therefore, the question of whether postal voting could be suppressing the effect of rainfall requires further investigation, as postal votes are not affected by polling day weather. The question is of wider significance to scholars of voting patterns, since postal voting has been shown to eliminate some of the costs associated with voting in person (Karp & Banducci, 2000; Schelker & Schneiter, 2017; Wass, Mattila, Rapeli, & Söderlund, 2017). The findings we present below contribute tentatively to this body of evidence.
Electoral reforms and the representativeness of turnout
2021, Political Science Research and Methods
- ☆
We would like to thank Michaela Slotwinski, Jo Thori Lind, and Francesco Zucchini for helpful comments and discussions. The paper has greatly benefited from comments made by the editors and referees as well as by conference participants of the Annual Meeting of the European Public Choice Society in Budapest, the Beyond Basic Questions Workshop in Gargnano, and the Annual Conference of the Swiss Society of Economics and Statistics in Lausanne.