Why is Turnout Higher in Some Countries than in Others?
1. Measuring Turnout
There are two methods in the literature for measuring turnout rates: either the population of voting age or the number of people on the lists of electors is used as the denominator.
Neither of these two indicators is perfect. The first is based on official national population censuses. It is intended to be more inclusive because it includes people who, for one reason or another, are not registered on the lists of electors. It is often used in countries where there is significant under-registration on lists of electors, such as in the United States. However, it can be too inclusive. The usual procedure is to estimate the population of voting age, although this estimate includes people who do not have the right to vote, particularly because they are not citizens.2 While some countries grant citizenship fairly easily to new arrivals, others are much more reluctant, and their total populations therefore include large numbers of non-citizens who are not entitled to vote. In addition, population censuses are conducted at different intervals in different countries (every five years in Canada, between seven and nine years in France) and also at times that do not necessarily coincide with election years. This makes adjustments necessary, which, especially for recent elections, can prove laborious and unsystematic.
The official turnout rate based on a comparison between the number of voters and the number of registered electors is also open to criticism. Its reliability depends on the quality of the methods for drawing up the lists of electors and the honesty with which these methods are followed, two factors that can vary enormously from one country to the next. There are no guarantees that all the people who are entitled to vote are on the lists of electors or that people who are not entitled to vote, or no longer entitled to vote, are not listed. Despite its acknowledged imperfections, though, this turnout rate is the one most used in official documents.
Both measurement methods have their biases. Calculations based on the population of voting age may underestimate voter turnout, because the denominator is artificially swollen by people who are not entitled to vote. On the other hand, calculations based on the number of registered electors may overestimate turnout (if the lists, and therefore the denominator, do not include all the people entitled to vote), or may underestimate turnout (if the lists are artificially swollen by duplicate or fictitious registrations). For this reason, both measurements were used in our analysis. Appendix C shows voter turnout calculated in both ways for every election included in this study. It should be noted that information on voter turnout using the population of voting age is more often unavailable, especially for the most recent elections.
2 A recent study (McDonald and Popkin, 2001) shows that calculating the turnout based on the population of voting age leads to a significant underestimation of the turnout in the United States. The same study indicates that there has been no decline in voter turnout in the United States when consideration is given to the growing number of Americans who are not entitled to vote because they are not citizens or are in prison.