Societes, Governance and Conflict

The effectiveness of monetary incentives to vote

4 min


Mariella Gonzales, Gianmarco León and Luis Martinez

Low voter turnout raises concerns about the representativeness of elected officials and the legitimacy of elections. This column examines the impact of monetary incentives to vote on electoral turnout and representation. Evidence from Peru, where voting has been compulsory since the 1930s, indicates that while reducing fines for abstention decreases turnout, it has a negligible impact on electoral outcomes in settings with few barriers to participation. This finding is relevant even for countries with voluntary voting, which could provide similar incentives to vote through tax rebates, discounts on government services, or direct transfers.

‘It would be transformative if everybody voted – that would counteract money more than anything.’ So said former US president Barack Obama in a 2015 speech. In democracies, voting is citizens’ primary tool to ensure that government policies represent their interests and to hold public officials accountable. Episodes like the introduction of electronic voting in Brazil or the Voting Rights Act in the United States show that higher electoral participation significantly affects representation and policy outcomes.

Yet voter turnout has been dwindling worldwide for the last 30 years. In many countries in the Americas, for example, almost half of the electorate failed to vote in recent presidential elections. What’s more, we have limited knowledge of whether there are large-scale public policies that would be effective in mobilizing voters.

First, some of the most robust predictors of voter turnout (such as weather) are not immediately amenable to public policy. Second, although an extensive body of research on voter mobilization initiatives (such as Get-Out-The-Vote) has provided substantial evidence, most of it corresponds to field experiments with limited geographical scope and short duration. Large-scale policies are affected by various issues that are hard to study through these experiments.

So how can we increase voter participation at scale? And will this affect representation?

How about a fine for not voting?

Many policy advocates have suggested introducing compulsory voting, abstention fines, or monetary incentives to vote. In a recent study, we examine voters’ responses to government-provided monetary incentives to vote in Peru, a middle-income country with more than 20 million voters.

Voting has been mandatory in Peru since 1933, with those who fail to vote facing restricted access to government services and having to pay a fine. Unsurprisingly, turnout has regularly exceeded 80%. Until 2006, the value of the fine was the same throughout the country and also relatively high (at around $90).

A reform that year classified districts into three categories (high, medium, and low fine) and differentially reduced the fine to $45, $23, and $11, respectively. The reform was a compromise between the desire to preserve the high levels of voter turnout attributed to compulsory voting, and concerns about a high and homogeneous fine disproportionately affecting the more impoverished population.

A comparison of districts with different fine levels before and after the reform reveals that a lower fine decreases voter turnout. Figure 1 plots average turnout in high and low fine districts (relative to the medium fine districts in the 2006 presidential run-off). In the elections after the reform (2011 and 2016), we estimate that a 10 Peruvian Sol fine reduction (around $3) leads to a 0.5 percentage point drop in turnout.


Policies ‘in the wild’ do not work the same way as tightly run experiments

Our estimate is substantially smaller than the one from a field experiment in the same setting that involved a door-to-door information campaign about the fine’s modified value. It is not uncommon for experimental interventions to experience such a ‘voltage drop’ when scaled up. Often, this is a result of changes in the quality or incentives faced by personnel or the perils of very large-scale program implementation.

We assume (and provide evidence) that informational frictions play a large role – that is, voters not knowing about the change in regulation. This is a pervasive problem affecting policies that change the rules governing the interaction of citizens with the state without communicating the changes adequately.

Monetary incentives are not crucial for compulsory voting to work

People aged 70 or older are exempt from compulsory voting in Peru. Figure 2 shows voter turnout by age group (relative to age 69) in 2016. While people in their sixties vote at a roughly constant rate, turnout dramatically declines for those aged 70 or older, though only in Peru and not in neighboring Chile (which does not have compulsory voting).

Our estimates imply that a full reduction of the fine would lead to a drop in turnout only 18% as large as the one caused by this age exemption. This suggests that non-monetary incentives (for example, the ‘expressive function of the law’) are the main drivers behind the effectiveness of compulsory voting.


When voting is easy, decreasing the value of the fine hardly affects electoral outcomes

We find that a smaller abstention fine decreases blank or invalid votes. For every 10 extra votes caused by a larger fine, we see 8.6 more votes that are blank or invalid. Our findings suggest that in a setting where citizens face few barriers to electoral participation (automatic registration, voting on Sundays, etc.), abstention is driven by unsatisfied, uninformed, or uninterested voters. Forcing them to vote hardly affects representation.

Nevertheless, there are many countries in which turnout remains worryingly low, despite the absence of significant barriers (for example, just 54% in Colombia’s 2018 presidential election). For these countries, our findings suggest that the introduction of a mandate to vote can be a powerful tool to promote electoral participation, even with very low fines that do not overburden those who do not comply.


Mariella Gonzales
PhD student, University of Chicago
Gianmarco León
Affiliated Professor, Barcelona GSE
Luis Martinez
Assistant Professor, University of Chicago