Article Preview
TopIntroduction: The Rational Voter Model
In any system where participants’ preferences are aggregated to produce a decision, such as the electoral systems of modern democracies, the degree of public participation is a crucial metric. Not only does it bring legitimacy to the system – thus making decisions more accepted by the population – it also protects minorities against some forms of abuse.
Despite the importance of public participation for the political health of the country, the direct impact on the voter is low, giving rise to the paradox of voter turnout. If the voter’s interest lies only in their vote’s impact, then large countries should have extremely high abstention, which isn’t the case, with huge variability between countries and elections (Jackman, 1987; Blais et al., 2014). The original rational voter model supposes that the choice of whether to vote or not depends only the cost of voting, the probability of affecting the outcome, and the impact on the voter of each outcome.
This paradox has been extensively studied in the past five decades (Riker and Ordeshook, 1968; Overbye, 1995; Kanazawa, 1998), and multiple factors have been shown to have a strong impact on public participation. Here are some of the best-established factors:
- •
Population size, as an increase in population comes with a decreased probability of casting the decisive vote for each voter, hence a lower turnout.
- •
Population stability (defined as the rate of demographic exchanges), as people vote more frequently when they feel they – and their neighbours – will stay in the constituency for a long time.
- •
Closeness of the election, as a perceived sure win for one party lowers the turnout.
- •
Whether the vote is mandatory or not, with a fine in case of abstention (although the fine is not enough to explain the magnitude of the effect, and social costs also factor in).
Many different models have been developed (Grofman, 1993; Bendor, Diermeier, and Ting, 2003) to explain the discrepancies between the original model and the observed variability of behaviour, by incorporating risk aversion, externalities such as social costs, or by proposing imperfect information (Geys, 2006; Levine, 2007).
Moreover, the voters are often accused – legitimately – of being ignorant of the main issues at stake (Congleton, 2001), which can be explained using a similar model, under the term of rational ignorance (Downs, 1957; Martinelli, 2006). Under this model, if the expected effect of each person’s vote on an issue stays constant, while the cost of acquiring expertise on an issue increases, the proportion of knowledgeable voters will go down. As ignorant voters behave in a less predictable (and hence apparently random) manner, this imposes an externality on the knowledgeable voters as it further diminishes the effect of their own vote (Großer & Seebauer, 2016), which in turn lowers the interest in getting more information. In such cases, mandatory voting can in fact make matters worse.
Combining the different effects mentioned, we realize that large-scale democratic procedures are replete with adverse incentives, and leaving voters in direct control of policy has been decried as a suicidal move for any state for more than two centuries – Sieyès (1789), in the morrow of the French Revolution, stated that the average citizen’s lack of education and time doesn’t prevent him1 from choosing a representative, but that he shouldn’t concern himself with matters of state as it would lead to chaos. That was already an echo of Montesquieu (1748), for whom choosing the best suited for government was a task fit for the general population, unlike lawmaking which should be left to those in the government.