LSE - Small Logo
LSE - Small Logo


January 6th, 2016

Compulsory voting results in more evenly distributed political knowledge

1 comment

Estimated reading time: 5 minutes


January 6th, 2016

Compulsory voting results in more evenly distributed political knowledge

1 comment

Estimated reading time: 5 minutes

Jill SheppardElectoral turnout in UK elections has been disappointing over recent cycles, despite a small uptick at the past two elections. One proposed solution is compulsory voting, though this measure is controversial in countries where voluntary voting is practiced. Jill Sheppard argues that countries which use compulsion to ensure turnout not only have higher turnout, but also more politically informed citizens – though the results of her research were not wholly straightforward in their implications.

Given interminably low rates of voter turnout across most Western democracies, looking to compulsory voting as a panacea for democracy’s ills seems sensible. Comparatively low – or declining – voter turnout is viewed generally as a symptom of civic disengagement from politics. Compulsory voting can also mitigate inequality in participation and representation. Citizens with the most resources and influence are typically the most likely to vote, and by voting for candidates and parties who reflect their interests, their participation can perpetuate systemic social biases.

Arend Lijphart has previously called on more countries to introduce compulsory voting. Currently, only 22 countries have some form of compulsory voting laws, and voting requirements, sanctions, and enforcement vary widely. Australia is arguably the archetypal case of compulsory voting. Voting has been mandatory since 1924, abstention is punishable (and usually enforced) with a fine of $20AUD in the first instance, and support for the law is almost ubiquitous. Australia enjoys consistently high levels of voter turnout. One consequence is the degree of inequality between voters and abstainers is lower than in similar countries, something Lisa Hill describes as the ‘great levelling’ effect of compulsion.

Compulsory voting also has consequences beyond the ballot box. As Malcolm Mackerras and Ian McAllister have observed, Australia has one of the highest rates of political party identification:

“Compulsory voting ensures that voters cast a ballot and the act of voting means that they are forced to think, however superficially, about the major parties”.

The assumptions made about first order (i.e. voter turnout) and second order (e.g. party identification) effects of compulsion are well founded, but require scrutiny. While the first order effects are well established (see Lijphart’s speech to the 1996 American Political Science Association annual meeting for a comprehensive overview), the second order effects are less certain.

My recent research seeks to address this gap. Starting with whether compulsory voting laws can reach beyond the ballot box to compel political engagement, I operationalise ‘engagement’ as political knowledge. Specifically, political knowledge refers to is measured by how many correct answers a survey respondent gives to three factual knowledge questions asked in the Comparative Study of Electoral Systems (CSES) project. Data come from 133 election studies, held in 47 countries between 1996 and 2013. The questions asked are specific to each country, but CSES oversight ensures comparable degrees of difficulty.

Compulsory voting laws in the CSES are categorised by in-country researchers. Three levels of compulsion are accounted for: compulsory voting with strong sanctions and enforcement; compulsory voting with moderate sanctions and enforcement; and compulsory voting with no sanctions or enforcement. A fourth category, voluntary systems, captures the remaining countries in the CSES dataset.

Table 1: Countries with compulsory voting within the CSES Modules 1-4 dataset.

Screen Shot 2015-10-06 at 16.49.51

Looking at the The CSES data reveals few differences in the mean numbers of correct answers, incorrect answers and ‘don’t know’ responses between compulsory and voluntary systems. If anything, weakly enforced systems appear to, prima facie, foster greater political knowledge than more strongly enforced compulsory systems. However, a cursory glance of the countries with ‘weakly enforced’ systems suggests that the two countries within this country – Greece and Mexico – likely have extenuating characteristics that can explaining this unexpected finding.

Fig 1: Mean answers by degree of compulsory voting enforcement (including voluntary systems).

Screen Shot 2015-10-06 at 16.51.36Note: Bars represent standard deviations.

Looking more closely at the distribution of political knowledge, Poisson regression analyses shows the role of compulsion on knowledge, holding other individual, district and national level factors constant.

Splitting the CSES sample by strength of compulsion does not reveal any obvious effects of compulsion, although small clues do emerge. The effects of party identification on knowledge are weakest in strongly enforced systems. Likewise, the effects of education are weakest in strongly enforced systems, although they are nearly as weak in voluntary systems.

Table 2: Poisson regression analyses.

Screen Shot 2015-10-06 at 16.56.49

Note: Cells represent regression coefficients with standard errors in parentheses.

To test whether the apparent differences in coefficients between voluntary and compulsory systems are significant, a Poisson regression model with both random and fixed effects sheds clearer light on how compulsory voting affects political knowledge. First, it shows strongly enforced compulsion – as in Australia – reduces the direct effects of education on political knowledge (when modelled as an interaction of education and compulsion level). Moderate enforcement reduces the direct effect of education on knowledge, but in almost imperceptibly small way. Weak enforcement appears to increase the effects of education on knowledge, even after controlling for the types of regime (with PolityIV scores) in those countries.

The results in Table 2 existing evidence that men tend to possess more factual political knowledge, using standard measures, than women. Looking at the interacted effects of gender and compulsion reveals the direct effects of gender are mitigated by compulsory voting across all three degrees of enforcement. In other words, the gender gap in political knowledge is smaller in countries with compulsory voting, even net of other individual and system-level factors.

Finally, previous research shows political knowledge, like other forms of knowledge, accrues with age. Here, the interaction of compulsory voting and age has no discernable effect, even with the large number of respondents. If anything, it has an almost invisibly small positive effect on political knowledge. This may reflect the accrual of knowledge over the lifecycle, as citizens in compulsory systems will have voted on average more times than those in voluntary systems and we expect participation in elections has some informative effects.

In conclusion, compulsory voting – particularly where it is enforced with sanctions or fines – does appear to result in more evenly distributed political knowledge than voluntary systems. But the effects are nuanced, and can be difficult to identify. If legislators want to increase knowledge, as a measure of broader political engagement, there are substantially less blunt instruments at their disposal. However, they would enjoy the other substantial benefits of compulsion.


Note: This article was originally posted on Democratic Audit and gives the views of the author only.

About the Author

Jill SheppardJill Sheppard – Australian National University
Dr Jill Sheppard is a Research Officer at the Australian Centre for Applied Social Research Methods (AusCen) at the Australian National University.



Featured image credit: Ian Britton / CC BY-NC

Print Friendly, PDF & Email

About the author


Posted In: Featured | Party politics and elections


Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported
This work by British Politics and Policy at LSE is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported.