Application to the Faculty Research Committee for a Faculty Research Leave

 

 

 

Title:                     

The Turnout Puzzle

 

 

 

Submitted by :     

Mark N. Franklin

Reitemeyer Professor and Chair

Department of Political Science

 

 

Abstract:

 

Why do people vote in democratic elections? Why is turnout lower in some countries than in others? Why does turnout appear to be declining in many countries around the world? These questions provide a set of interlocking puzzles that I have been addressing in published and unpublished work for several years. I am requesting a Research Leave that would extend a scheduled sabbatical so as to enable me to spend a full year completing this research and drawing it together into a book-length manuscript.

 

BACKGROUND

Until my research findings started to appear in print, the conventional wisdom among American political scientists was that people vote because they are well-educated, of high social status, well-established in their communities, and perhaps also motivated by a strong campaign (Rosenstone and Hansen, 1993; Verba, Schlozman and Brady, 1995). This conventional wisdom, however, comes primarily from the American experience. Its general applicability is put into question by the fact that America itself contains one of wealthiest and most highly-educated populations on earth, yet turnout in American elections is generally less than half what is found in other countries. By contrast, the world record for high turnout is held by little Malta – a poor country which did not even have a university until the 1990s. My own work addresses the question of why some countries have higher turnout than others and uses the findings to illuminate the problem of low voter turnout, attempting in this way to transcend cultural straightjackets that can limit our understanding of such problems. Only thus can we hope to address the problem with solutions that work. This is underlined by recent studies of the new U.S. “motor voter” law which show a very small effect on turnout, suggesting that the reform was perhaps misconceived.

 

In my published work I have resurrected a theory proposed by the great Swedish social scientist, Herbert Tingsten (1937), to the effect that people vote because they are motivated to do so. This theory had come into disrepute in the United States because it does not accord with contemporary ideas of rational behavior. If people only voted because of measurable benefits to them (it has been several times argued) then no-one would vote. This is because the probability that any one person would affect the outcome in a national election is miniscule, yet the effort required of that person is not. Hence, in terms of pure self-interest, voting is often considered irrational. Indeed, turnout has been referred to as “the paradox that ate rational choice theory” (Groffman, 1993). My own attempt to break through this supposed paradox involves the proposition that people vote not because of measurable benefits to them but out of solidarity with other people who have the same political concerns as themselves. In other words, voting is a social act performed in the hope of reciprocation by like-minded individuals. This theory overcomes the paradox of voting but brings with it the need to catalog the different ways in which benefits of voting might vary between countries and over time. It also requires that someone demonstrate empirically how these variations trump the differences in social conditions that had previously been thought to be fundamental.

 

I have been working to fulfil these requirements through a number of interlocking research projects conducted over the past twelve years. Scientific advances (both small and large) generally follow from the availability of new data sources or new measurement techniques. Previous turnout researchers have focussed almost exclusively on explaining cross-sectional turnout differences – differences between one individual and another or between one country and another at one point in time.  New data sources have enabled me to focus on the question why turnout varies with the passage of time.

 

My main data source has been the studies of European voters conducted after elections to the European Parliament held in 1989, 1994 and 1999 (the last one funded in part by a grant from Trinity College). Although these are not 'real' elections (in that no government is formed on the basis of the parliamentary majorities established in these elections) this has the advantage of removing much of the 'noise' that makes it hard to study real elections – noise (such as might be generated by an untimely scandal or an unflattering photograph) that otherwise would have to be measured and accounted for. So these venues are paradoxically well-suited to studying the mainsprings of voting behavior. Over three successive elections, involving first 12 and more recently 15 different countries, I and my collaborators have been able to catalog the effects of differences in electoral arrangements and political conditions on turnout and other electoral phenomena (van der Eijk, Franklin et al., 1996; van der Brug, van der Eijk and Franklin, 2000; Franklin, 1996a; 1996b; 2001). These findings have implications beyond the European arena which can be tested using a second dataset: one relating

 to the outcomes of democratic national elections held in the 22 countries that have held such elections continuously (or virtually continuously) since World War II. The United States is, of course, included in this second dataset.

 

These two massive data collection enterprises have reached a point where the data are sufficient to address the questions that arise from my attempted solution to the turnout puzzle. The paper that I presented to the American Political Science Association annual meeting in Washington DC at the beginning of September (Franklin, Lyons and Marsh, 2000) constitutes the last big piece of the puzzle. A number of small pieces still need work, but the general picture is now in place and the time has come to bring it to fruition as a single extended argument supported by the necessary empirical findings. It is for this purpose that I am requesting a research leave to extend my sabbatical.

 

RESEARCH OBJECTIVES

I shall use my time for four things:

1) to refine the rational choice basis of my social solidarity theory of election turnout (see above);

2) to extend the dataset of election outcomes since World War II by adding a number of additional countries (especially in Latin America) that have held elections through most of this period;

3) to replicate and refine the findings from the two datasets (one for the two European elections and the other for national elections in 22 countries) with my new datasets for three European elections and national elections in 30 (or so) countries so as to arrive at a parsimonious but well-specified model of electoral turnout; and

4) to present my theory and findings in a book-length work that will definitively establish the reasons for turnout variations in the United States and elsewhere, together with the reasons why some countries see lower turnout at national elections than other countries.

 

RESEARCH PROGRAM

The chapter outline for this book is supplied in Appendix 3. Essentially the whole book is an expansion of my chapter "Electoral Participation" in Leduc, Niemi and Norris (1996) which is now in the process of redrafting for a second edition of that book. The original chapter is to be reprinted in Niemi and Weisberg's Controversies in Voting Behavior (4th Edition, 2001), the standard work for graduate students of political science which summarizes the 'state of play' (by reprinting cutting edge articles) in the areas of major concern to voting specialists. My book will have a number of secondary co-authors whom I wish to recognize on the title page for their contributions to this work. However, the actual writing of the book is entirely my own responsibility. Insofar as I reproduce text from previous articles, this text will have been written by me.

 

The most important point that needs fleshing out is the way in which differences in electoral arrangements give rise to motivational differences. To take a trivial example, compulsory voting generally motivates turnout by threatening non-voters with a fine. The effects of postal voting, Sunday voting, electoral responsiveness (see below), and the proportionality of the electoral system need to be linked to the motivations that cause turnout to increase when voting is held on a Sunday, when postal ballots are more easily obtained, when legislative majorities are more easily overturned, and when the electoral system causes outcomes in terms of seats in the legislature to more nearly reflect the proportion of votes cast.

 

The second most important point that needs fleshing out is the way in which differences in electoral arrangements trump individual differences in determining the level of turnout at an election. Intuitively it is easy to understand that if midterm congressional elections in this country see lower turnout than presidential elections, this is not because the adult population has different characteristics in these elections. The population of the U.S. does not change its composition every two years! But this insight needs to be developed into a proper argument with implications for other countries that can be tested.

 

My previous work will also be extended in several minor ways. One is by increasing the number of countries included in the country-level dataset from the present 22 to something more like 30 to give better global coverage, even though this would mean including some countries that have only held elections intermittently since 1945. Another is by refining my measures of electoral responsiveness. For example, one of these measures now counts the number of times party control of the legislature changes. A more appropriate measure might be the average length of time between changes of party control. A third is by 'tweaking' the method used for obtaining predicted levels of turnout in each country from the historical record of turnout variations in other countries. There are some unusual interaction terms in my prediction equation, and I have not settled on the best way to handle these. I will be consulting with statisticians and other methodologists in an effort to improve my estimates.

 

My preliminary findings suggest that there has been no general decline in turnout unexplained by changes in electoral arrangements over the past fifty years. The observed decline appears to be largely due to abandonment of compulsory voting and other procedural innovations with predictable consequences for turnout in certain countries. In the United States, recent reductions in turnout seem to be largely explicable on the basis of increases in the size of that part of the voting-age population not entitled to vote (mainly felons and immigrants). Evidently, conclusions like these will be controversial; so they need to be carefully documented before they are published.

 

In summary, the year will be spent in extending my research (by collecting new data and by performing new computer analyses of the extended datasets), in revising existing papers for inclusion in the volume, and in writing a small number of new chapters (see Appendix 3). I intend to apply for a one year Research Expense Grant to provide me with funds to pay for data collection.

 

BIBLIOGRAPHY

van der Brug, Wouter, Cees van der Eijk and Mark Franklin (2000) "The Economy and the Vote: Electoral Responses to Economic Conditions in 15 Countries." Paper presented at the American Political Science Association annual meeting.

van der Eijk, Cees, Mark Franklin et al. (1996) Choosing Europe? The European Electorate and National Politics in the Face of Union. Ann Arbor: Michigan University Press.

Franklin, Mark (1996a) "Electoral Participation" in Laurence Leduc, Niemi and Norris (1996), 214-233.

Franklin, Mark (1996b) "European Elections and the European Voter" in Jeremy Richardson (ed.) European Union: Power and Policy-Making. London: Longman, 188-199.

Franklin, Mark (2001) "Electoral Participation" in Niemi and Weisberg (2001)  (in press).

Franklin, Mark, Patrick Lyons and Michael Marsh (2000) "The Tally of Turnout." Paper presented at the American Political Science Association annual meeting.

Groffman,Bernard (1993) "Is Turnout the Paradox that Ate Rational Choice Theory?" in Bernard Grofman, ed., Information, Participation and Choice: An Economic Theory of Democracy in Perspective. Ann Arbor MI: University of Michigan Press.

Leduc, Laurence, Richard Niemi and Pippa Norris, eds. (1996) Elections and Voting in Global Perspective. Thousand Oaks CA: Sage.

Niemi, Richard and Herbert Weisberg, eds. (2001) Controversies in Voting Behavior (4th Edition). Washington D.C.: CQ Press (in press).

Rosenstone, Steven and Mark Hansen (1993) Mobilization, Participation, and Democracy in America. New York: Macmillan.

Tingsten, Herbert (1937) Political Behavior: Studies in Election Statistics. Totowa, N. J.: Bedminster Press, 1963 (first published in 1937 as volume 7 of  the Stockholm Economic Studies).

Verba, Sydney, Kay Schlozman and Henry Brady (1995) Voice and Equality: Civic Voluntarism in American Politics. Cambridge MA: Harvard University Press.