PolSoc: Electoral Turnout

What are two important things to note about the turnout literature?

Out of 127 distinct variables found in Frank and Martinez’s (2021) literature review, only 44% appear more than once. Authors rarely include the same control variables in their models which possibly leads to underspecified models and spurious inferences (Smets and van Ham, 2013).

Many variables explaining individual-level turnout are interconnected, either because of path-dependency or because they measure closely related concepts, so findings for each independent variable are highly dependent on whether its covariates or more proximal causal factors are included in the model as controls or not.

What is the rational choice model of voting?

The rational-choice perspective of voter behaviour conceptualizes the decision to vote as the result of a personal cost-benefit calculation (Downs, 1957).

Riker and Ordeshook’s (1968) ‘Calculus of voting’ conceptualizes the decision of whether to vote or not as vote if BP – C + D > 0 where B is the difference in utility depending on the outcome, P is the probability that the vote will be pivotal, C is the cost of voting and D is the utility gained from voting that is unrelated to the outcome.

What does this model lead us to conclude?

Individual political participation does not (unless in the most extreme cases) have any effect. Gelman et al. (2012) calculate that the probability of a single vote being decisive in the US as, at most, 1 in 60 million. Because this probability is miniscule, the prediction from this model without D is zero turnout.

This suggests that psychological variables, which induce a sense of utility from voting unrelated to the outcomes, are the main factors affecting political participation.

What is there to say about civic duty?

Firstly, turnout is strongly correlated with civic duty to vote. People participate in politics simply because they believe in the normative value of voting as something that everyone in society has an obligation to do.

In 2015, for example, 84% of people in the UK who agreed that “It’s everyone’s duty to vote” turned up to the ballot box compared to only 54% of those who chose the option “People should only vote if they care who wins.” (Curtice, 2016). Civic duty consistently turns up in research as a driver of political participation.

What is one issue with this finding?

Estimating the true effect of civic duty is difficult as there is an inherent selection bias. Survey respondents who do vote are more likely to want it to seem like they are civically responsible, rather than self-serving, and those who abstain will not want to seem as they are ignoring a civic responsibility, so they will advocate against voting being a civic duty (Karp, 2008)

How does civic duty explain a decline in turnout?

Blais and Rubenson (2013) find that generational differences disappear once civic duty and external political efficacy are controlled for.

Studies find that, in contrast to older generations, new generations consider citizenship more as a right than as a duty and increasingly prefer direct forms of participation, such as protests or boycotts, to voting. Kostelka and Blais (2021) attribute this to the cultural changes which accompany economic development. One of these such changes is a decline in deference to authority and norms, including the norm of civic duty.

What do theories of mobilization say?

Theories of mobilization view voting essentially as a social behaviour guided by norms and sanctions, and argue that citizens go to the polls just because their family and peers do so (Smets and Van ham, 2013), or even simply because they are asked to vote by campaigners.

Social networks reduce the costs of political participation by providing information about parties, candidates and the electoral process. They also provide utility as voting brings positive reinforcement from friends, family, and co-workers.

What evidence supports social network theory?

There is statistically significant evidence that living with someone who votes increases turnout, especially for young people (Fieldhouse and Cutts, 2012). A first-time voter in an average turnout area who lives with another voter has an 83% chance of voting compared to only 27% for one who does not.

Does this relationship work the other way?

Furthermore, this relationship also works the other way. Dahlgaard (2018) found that parents were more likely to vote by 2.8 points if their child had become recently eligible to vote compared with if their child was only slightly too young to vote, and this finding only held if the child still lived with their parents, further supporting the in-house mobilization theory.

How do social norms change with age?

First-time voters spend their time within a peer group which consists almost entirely of other non-voters. However, as voters age, their peer cohort has increasingly higher participation rates and they eventually move into new environments, such as workplaces or community organizations, where the average levels of political knowledge and turnout are higher. Thus, citizens will be subject to more intense participation norms.

What is there to say about the decline in turnout due to parents?

If voting is a habit, which the literature indicates, then the initial engagement of first-time voters in the electoral process will resonate throughout those voters’ lives. (Fieldhouse, 2012). Plutzer’s (2002) study of voting-habit and socialization finds that the largest effect on initial turnout is parent’s reported turnout.

Thus, parents who do not vote leave their children substantially behind in the acquisition of voting-habit. As increasing numbers of parents fail to vote, this abstention gets transferred to many of their children as they have not been socialised into the habit.

What is there to say about the decline in voting due to other social networks?

Younger generations are less likely to have or experience vital cues for engagement from traditional sources such as work, civic groups, religion, or community. Young people take longer to join full-time employment, are less likely to join a political party or trade union, and increasingly lack cues from traditionally salient collective identities such as class, community and religion.

What is there to say about habit?

Past turnout is one of, if not the, strongest indicators of voting behaviour. Once individuals become politically and electorally socialized, their propensity to vote tends to remain relatively stable throughout the rest of their lifetime (Blais, 2021).

Some argue that this observance reflects stability in attitudes that drive the decision to vote, such as consistent political interest, however most works interpret this stability as a sign of habit. People who vote become accustomed to voting and perhaps even acquire a taste for it.

Evidence for habit:

Meredith (2009) shows that Californians who were just over 18 at the 2000 US presidential election were more likely to vote four years later than those who were just under 18 and therefore not eligible to vote in 2000.

Franklin (2011) finds that survey respondents in 27 European countries are significantly less likely to vote if their first election was a European Parliament election, which tends to draw low turnout.

What implications does habit have, and how may it explain a decline in turnout?

If voting is habitual then one uninspiring, uncontested election may disrupt voting habits and lower turnout in future election cycles. Turnout dropped to a historic low of 59% in the 2001 UK General Election as Blair was expected to retain power easily and the result was a foregone conclusion. Consistent with the habit hypothesis, turnout has still not recovered and no election in the 21^st century even reached 70%. Every election since 1922 managed to achieve such a feat.

How have the number of elective institutions increased?

In European Union member states, the number of elected institutions has increased by 34% on average from the 1960s to the 2010s. This is due to a combination of factors including decentralization, regional (European) integration, frequent use of referendums, introduction of directly elected presidencies, and shorter terms of office.

France case study:

In France, in the three years before the 1978 legislative election, citizens could participate in three elections on average. In the three years before the 2017 legislative election, they could participate in nine elections not to mention presidential primaries which were organised by several parties. Interestingly, the 2017 legislative election recorded the lowest turnout in France’s post-1945 history.

How do the number of elections link to habit and declining turnout?

Frequent voting opportunities leads to voter fatigue and, importantly, disrupts habitual voting patterns. The routine of regular voting creates a psychological commitment, and once this habit is disrupted, it becomes easier to miss subsequent elections, as the sense of obligation weakens and the barrier to non-participation becomes lower.

Analysis from Kostelka and Blais (2021) supports the hypothesis that the rise in the number of elective institutions has directly led to reduced turnout.

Why are the costs so high for first-time voters?

When young people confront their first election, all the costs of voting are magnified: they have never gone through the process of registration, may not know the location of their polling place, and may not have yet developed an understanding of party differences and key issues. (Plutzer, 2002)

What is the resource model?

The resource model suggests that political participation is an act driven by resources, particularly time, money and skills (Verba, 1972). Those with jobs, a high income, and a high socio-economic status are more likely to have the resources which enable them to participate and to overcome the costs of participation.

What are examples of these resources in action?

For example, having the leisure time to read the news, and the skills to follow media discourse, helps foster political engagement. You also need the time to research candidates, register to vote, and the means to physically travel to and from the polls. Higher levels of education enable citizens to process complex information, such as sorting through policy choices come election time (Dalton, 2022).

Class turnout disparity stat:

For instance, in the 2017 General Election, turnout was 69% for social grades AB and 53% for social grades DE.

Does this participation gap extend outside of elections?

Inequalities extend to other forms of political participation. More affluent and educated citizens are more likely to write to their representative, lobby, join parties and even more likely to protest and engage in other contentious forms of political action.

There is no limit on the number of letters a citizen can write to their MP or the number of meetings they can attend. Therefore, inequality in non-electoral activities shows a participation gap that is nearly three times larger than the (already widening) participation gap in elections (Dalton, 2022.

What change discredits the resource model?

In Britain, the working class have been consistently less likely to vote than the middle class. However, this gap has widened substantially since the 1990s. Evans and Tilley (2017) assert that this provides compelling evidence against the resource model of turnout because resources have not changed but non-voting rates between the classes have dramatically altered.  

What has happened in the UK recently? (particularly to Labour)

As levels of education and the numbers of people with middle class jobs have increased, the average voter’s views have become more economically right-wing and socially liberal.

This led to a convergence of policy stances between the Conservatives and New Labour in the 1990s, with both adopting more right-wing economic policy positions.

Labour began explicitly targeting the middle-class and ceased referring to the working class in speeches or policy documents, preferring to use more neutral terminology such as “hard-working families” (Evans and Tilley, 2017).

Furthermore, the number of working-class Labour candidates dropped substantially (Heath, 2007).

What did this lead to?

All these changes altered voter perceptions of Labour so that the working class now view both main parties as representing the middle class and offering similar policies.

Working class individuals now feel they lack a genuine political choice and have thus increasingly chosen not to vote. This has produced the kind of class non-voting that has long been prevalent in the US.

What further implications does this have?

The less educated, the poor, and members of the working class are turning away from elections and losing their voice. Sadly, this may further erode the participation of other lower-status individuals who conclude the political system is unconcerned with their needs (Solt, 2008).

What is there to say about education and voting?

A vast literature asserts that higher levels of education correlate to higher levels of turnout (e.g Smets and van Ham, 2013). This is regarded as almost a ubiquitous finding.

However, the existing empirical literature focused almost solely on the effect of education within liberal, western democracies.  

What throws a spanner in the works regarding education?

In Zimbabwe, an electoral authoritarian regime, Grossman et al. (2016) find strong and statistically significant evidence that better-educated citizens are less likely to vote, contact politicians, or attend community meetings. What could explain this result?

What could explain this result?

Well, education increases critical capacities and political awareness. Education is also associated with greater socioeconomic status and therefore better educated citizens may have the resources to access more critical foreign media. All in all, education increases the likelihood of recognising that participation in an authoritarian regime is futile and serves to legitimize autocrats. It allows you to recognise that the outcome of the election is already predetermined and, it seems, this may cause people to abstain.

Final point against resource model:

Western democracies continue to become wealthier and better educated over time, and yet turnout is dropping. This alone provides compelling evidence to discredit the resource model.

What is there to say about the weather?

Bad weather has a proven, negative effect on individual-level turnout as it increases the cost of voting and decreases the net utility of voting. Rainfall on election day has been found to reduce turnout by around 0.95 percentage points per centimetre.

Hansen (2023) further shows that turnout is also receptive to positive weather, with sunshine increasing turnout.

Young and inexperienced voters are deterred from voting by bad weather up to six times more than mature voters.

Institutional factors: what is there to say about the electoral system?

Much of the literature supports the view that proportional representation (PR) increases turnout. For example, Frank and Martinez (2021) find that PR election systems are associated with 5.2% higher turnout than states without this system.

Caveats to PR:

Whilst most of the literature supports the view that PR fosters turnout, there is no compelling explanation of how and why.

Research dealing with Latin America reports no association between PR and higher turnout (Fornos et al. 2004), and an analysis that incorporates both established and non-established democracies concludes that the electoral system has a weak effect (Blais 1998).

What is there to say about compulsory voting?

The fact that compulsory voting raises turnout is the literature’s most commonly accepted finding (Frank and Martinez, 2021). Jackman (1987) estimates that compulsory voting increases turnout by around 13% but Blais (2003) observes that compulsory voting needs to be accompanied by enforced sanctions to guarantee success.

Political factors: differences between parties. What is there to say about the UK?

Turnout levels in Britain have fallen dramatically over the last 50 years, from around 80% in the 1950s to around 60% in the 2000s, and one of the most prominent explanations for this decline emphasizes the ideological convergence of the Conservative and Labour party.

In the 1990s, new-Labour shifted their attention towards the middle-class and began to adopt more centre-right policies. The two main parties therefore now cover a far less extensive ideological range than was ever the case over most of the twentieth century (Evans and Tilley, 2017). When both parties share similar policies, voters may despair that outcomes will look very similar regardless of whom gets into power, and may therefore abstain.

What evidence supports this?

Such a theory is supported by the fact that the two general elections involving socialist Jeremy Corbyn (2017 and 2019) and therefore two radically different policy stances from the major parties, yielded the highest two turnouts of the century.

What stat supports the ideological distinctiveness hypothesis?

How much difference voters see between parties has been proven to be a powerful predictor of participation. 83% of those who thought there was a big difference between the parties voted in the 2017 British election, while only 59% who felt there were no major differences voted (Curtice and Simpson, 2018.)

How does competitiveness effect the election?

Blais (2006) observes that the closeness of an election has been found to increase turnout in 27 of the 32 studies that has tested the relationship, in many different settings and diverse methodologies. He described it as “the most firmly established result in the literature” and states “I cannot see how this finding could be wrong.”

Is this really true?

However, Smets and van Ham’s meta-analysis (2013) found closeness of the election to consistently have no effect on turnout.

Frank and Martinez (2021), in their extreme bounds analysis, which ran over 15 million regressions to determine which variables are robustly associated with voter turnout, found that higher levels of competitiveness actually decrease turnout.

Why is turnout declining in Western democracies? Which hypothesis is not true?

Political commentators often assume that the growing number of abstainers reflects discontent and dissatisfaction with politics. However, the political science literature does not support this view theoretically or empirically (Kostelka and Blais, 2021).

What is there to say about globalization?

Mahoney (2022) contends that the driving force behind declining turnout is globalization, particularly economic globalization.

Such a relationship is supported analytically (Frank and Martinez, 2021) although no compelling evidence has demonstrated causation or vindicated a particular explanatory mechanism.

Mahoney contends that economic globalization, especially capital mobility, limits national autonomy and government power in areas of economic policy. As a result, citizens increasingly view their vote as having less capacity to influence economic policy.

What is there to say about Partisan dealignment?

Previously, people would simply vote on the basis of their party-allegiance, giving very little weight to any other contextual consideration. However, stable party-links have been declining over time in advanced democracies. In the UK, for example, the proportion of the electorate identifying strongly with a party dropped from 45% in 1964 to just 10% in 2005 (Fieldhouse, et al., 2020).

What implication does partisan dealignment have on voting?

Therefore, Heath (2007)  argues, that the act of voting nowadays, more than ever, is a response to more narrow concerns related to short-term factors associated with the political context.

Reher (2014) similarly conceives of party identification as a stabilising force that makes turnout less dependent on voters’ and elites’ short-term policy agendas. Party-identifiers will turnout at high rates regardless of short-term context.

What is there to say about inequality?

Polacko (2021) finds that higher levels of income inequality significantly reduce turnout and widen the turnout gap between rich and poor. In the United States, a country with huge inequality, the difference in voting between the highest and lowest income quintile sits around 40 percentage points.

Furthermore, the income gap in turnout has increased most notably in countries that have experienced the largest rises in inequality, such as the United Kingdom (Birch et al. 2014).

Relative power theory suggests that increased wealth for high-income individuals translates into more political power which in turn causes low-income earners to disengage from the political process when their interests are not responded to.  

Do referendums increase turnout?

Referendums – in countries where used sparsely – often yield huge turnouts. Not a single general election in history received as high a turnout as the Scottish Independence referendum in 2015, and the turnout for the EU referendum was higher than any election this century.

Interestingly, the disparity in class turnout narrowed for the Brexit referendum (Heath, 2016.

However, it is an exaggeration to suggest that referenda always lead to higher levels of turnout. For example, the 1998 referendum on a London mayor and assembly had a turnout of 34%.

What is one final point on the ability to affect outcomes?

The highest levels of turnout are reported in small countries such as Malta. One interpretation is that voters are more likely to feel that their vote could be decisive in a small country (Blais, 2006).