Length of compulsory education and voter turnout—evidence from a staged reform

Abstract

It is possible that human capital produces positive externalities to the society indirectly, through non-market channels such as health or crime. Another such channel could be the effect of education on the functioning of democratic decision-making. Measures of the functioning of democracy are bound to be controversial, but one such measure -voter turnout -reflects the engagement of people to democracy, and also receives a considerable amount of attention from social scientists as well as the media.A vast body of empirical research supports the conclusion that educated people have a higher tendency to vote in political elections. However, since the level of education is not assigned randomly, but is a conscious choice, the possibility remains that both education and political participation are determined by unobserved personal qualities or parental influences. So far few studies have been able to find and utilise an institutional change that would have produced experimental variation in the amount of schooling across individuals or groups of people and match it to data on voter turnout or other variables of civic engagement. The few recent studies that attempt this, end up with partly conflicting results.This study estimates the effect of education on voter turnout in the long run. It contributes to the empirical evidence based on institutional changes by using the timing of a Norwegian staged school reform as an instrumental variable for education. In contrast to previous studies, the Norwegian reform created relatively large individual level variation in the years of schooling at lower levels of attainment, as the minimum number of compulsory years was raised from seven to nine. The timing of the reform varied across Norway in a scattered fashion, and this variation appears to be quasirandom as it is difficult to find socio-economic correlates for it.To provide a robust view of the effects of the reform, the analysis is carried out at two levels.Firstly, by using survey data, the impact of the reform on voting in parliamentary elections is tested using individual level data two to four decades after the reform. To the extent that the survey responses can be trusted, this provides an ideal setting for the evaluation of the impact of education on voting. Since some assumptions regarding accuracy of the survey data cannot be fully tested, the analysis is also carried out at the municipality level, using data from up to two decades after the reform started. Municipality level turnout data is representative and accurately measured and also caters for potential behavioural externalities in voting behaviour.

Voter Turnout, Information Acquisition and Education: Evidence from 15 European Countries

“…In line with previous research, we observe that overall voting rates emerging from survey data -in our case the European Social Survey -are higher than voting rates registered in general elections (see for example, Milligan et al 2004;Siedler 2007;Pelkonen 2009). Measurement error could be problematic for estimates of the effect of education on voter turnout because the bias in voting reports could be higher among the better educated (Bernstein et al 2001).…”

“…Measurement error could be problematic for estimates of the effect of education on voter turnout because the bias in voting reports could be higher among the better educated (Bernstein et al 2001). Pelkonen (2009) provides compelling evidence on the link between education and voter turnout in Norway and finds that results are similar whether individual-level survey data are used or whether estimates are based on municipal-level aggregate data on actual voter turnout. Empirical evidence, however, suggests that such bias does not affect empirical estimates in meaningful ways.…”

“…Information on voter turnout has been widely employed in the literature to compare political participation across countries and examine its determinants (Dee 2004;Milligan et al 2004;Pelkonen 2009;Siedler 2007). Less common is the use of information on political knowledge and the quality of political involvement even though democratic societies need democratically active citizens as well as a politically informed electorate (Galston 2001).…”

(De)Centralization and voter turnout: theory and evidence from German municipalities

“…Given the generally accepted importance of income and education on voter turnout (Verba and Nie 1972;Pelkonen 2012), we include the share of educationally highly-qualified (defined as the share of population with university degree) and low-qualified inhabitants (defined as the share of population without vocational training and without secondary school education) as well as the share of long-and short-term unemployed people in our model. We also include dummies equal to 1 when the local election takes place alongside EU or statelevel elections to control for the fact that concurrent (higher-level) elections tend to boost voter turnout (Geys 2006b …”

Can Raising Instructional Time Crowd Out Student Pro-Social Behaviour? Evidence from Germany

“…We contribute to two strands of literature: first, we contribute to the economic literature on the external, non-monetary effect of education on civic engagement, which focuses on the effect of years of education on predominantly political interest, information, and participation (Dee, 2004;Dhillon and Peralta, 2002;Milligan et al, 2004;Pelkonen, 2012;Siedler, 2010), as well as reciprocity (Fehr and Gachter, 2000;Kosse et al, 2014). 5 Here, the study most closely related to ours is Gibson (2001): the author uses a sample of twins to hold unobservable family characteristics constant, showing that more years of education are associated with a lower probability of volunteering and supply of volunteer hours.…”