Facilitating the Electorate: A Multilevel Analysis of Election Timing, Registration Procedures, and Turnout

Abstract

This study suggests that certain election administration arrangements can have a modest but significant impact on the probability of voting. Using data from Module 3 of the Comparative Study of Electoral Systems, supplemented by contextual data from the ACE Electoral Knowledge Network, I explore the effects of election timing and voter registration procedures on turnout in 30 democracies. I show that when voter registration is in the hands of an independent electoral management body, there is a greater likelihood of an individual voting. Contrary to conventional wisdom, election timing has little impact on the probability of voting. Weekend voting, polling hours, and the timing of the election during the year have no significant impact, even among young people, supposedly the beneficiaries of such arrangements. The results show the potential for certain election arrangements to stimulate turnout. They also reveal that facilitating voters by lessening voting costs is not a cure to the problem of growing abstention. These findings have implications for our understanding of voter turnout and the means by which declining electoral engagement is addressed.

Keywords:

• turnout
• weekend voting
• voter registration
• polling hours

Acknowledgement

The author is grateful for the advice and suggestions of Andre Blais, Magda Gircucanu, Peter Loewen, Theresa Reidy, and Fiona Buckley and to two anonymous referees. Any errors are the responsibility of the author.

Notes

1. There have been two occasions when Ireland has voted on a weekend: the first was the second referendum on the Nice Treaty in October 2002, and the second was the Children's Referendum in November 2012. On both occasions, the referendums took place on a Saturday. Previous to this, the Tipperary South by-election in 2001 was the first time an election was held on weekend in Ireland, also taking place on a Saturday (Lyons & Sinnott, 2003).

2. In the 19 general elections that have been held between 1948 and 2011, voting has taken place on a various weekdays. Wednesday has been the most popular choice, but Friday's have become the norm recently.

3. Young people are classified as citizens under the age of 30.

4. Even if an individual is working the maximum week permitted under EU law, which is 48 hours per week, the maximum their working day could be, assuming a 5-day working week would be 9 hours 35 minutes.

5. In a sample of 32 countries included in the Comparative Study of Electoral Systems data set, the range of polling hours in non-compulsory countries was between 8 and 15 hours.

6. All permanent residents of Germany are required to register their place of residence with local government authorities.

7. Compilation of the electoral register is the responsibility of the 31 local authorities. Each local authority compiles a draft electoral register each year (published in November) having conducted house-to-house inquiries, where registration can be completed by anyone in the household on behalf of all residents. The draft register is then made publicly available for inspection in libraries, post offices, Garda stations, and local authority offices. After a period of adjudication on claims of errors, the final register is published the following February.

8. For access to the data and more information on the CSES data set, see http://www.cses.org.

9. In circumstances were multiple election data were available for one country, the survey which contained the greatest number of observations was chosen.

10. Countries were judged to have democratic, free, and fair elections if they scored 2.5 or above on the Freedom House Political and Civil Rights index in the year the CSES survey was carried out. The Freedom House scores were obtained from the macro component of the CSES data set.

11. The Netherlands (1967), Italy (1993), and most recently Chile (2012) have all abandoned compulsory voting over the past 50 years.

12. Accordingly, Australia, Brazil, Peru, and Uruguay are excluded from the study.

13. Latvia and Taiwan are omitted from the analysis because there was no data available for party identification or campaign involvement respectively, deemed crucial covariates to include in models exploring the likelihood of voting (Blais & Achen, 2010). Switzerland is also excluded given the idiosyncratic nature of electoral participation and electoral administration in the country (Franklin, 2004).

For the analysis on polling hours, poll timing varies across provinces/states in the Canada and the USA. Thus, the values for these countries represent the mean number of polling hours.

For the analysis on voter registration responsibility, data were unavailable for Croatia and Israel. The USA are also excluded from this analysis as registration responsibilities vary by state.

The full list of countries included in the analysis are: Austria, Canada, Chile, Croatia, Czech Republic, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Israel, Japan, Latvia, Mexico, Netherlands, New Zealand, Norway, Philippines, Poland, Portugal, Romania, Slovakia, Slovenia, South Africa, South Korea, Spain, Sweden, and the USA.

14. For access to the data and more information on the ACE data repository, see http://aceproject.org/.

15. The Czech Republic votes across two days: Friday and Saturday. For the purposes of this analysis, it is classified as a ‘weekend voting’ state.

16. I also explored the possibility the effect of polling hours might not be the number of hours the polls open but that the polls remain open later in the evening. Reconstituting the models though indicated, there was no greater likelihood of individuals voting if polls were kept open beyond 7 pm local time.