Strategic or tactical voting is voting in consideration of possible ballots cast by other voters in order to maximize one's satisfaction with the election's results. Gibbard's theorem shows that no voting system has a single "always-best" strategy, i.e. one that always maximizes a voter's satisfaction with the result, regardless of other voters' ballots. This implies all voting systems can sometimes encourage voters to strategize. However, weaker guarantees can be shown under stronger conditions. Examples include one-dimensional preferences (where the median rule is strategyproof).
With large electoral districts, party list methods tend to be difficult to manipulate in the absence of an electoral threshold. However, biased apportionment methods can create opportunities for strategic voting, as can small electoral districts (e.g. those used most often with the single transferable vote). Proportional representation systems with small districts often involve large-scale vote management operations, which are common in countries using STV-PR such as Ireland.
Common types of strategic voting
Some types of strategic voting described in the literature are:
(sometimes called favorite betrayal)
A voter ranks or rates a candidate they dislike – a lesser evil – higher in the hope of getting them elected, thus preventing an even worse candidate from being elected. This type of strategic voting includes the first-past-the-post strategy of voting for the lesser of two evils among the top two frontrunners, as well as the more general favorite betrayal strategy. A sufficiently strong compromising incentive, possibly strengthened by a strategic exit incentive, can lead to or reinforce two-party rule as voters will abandon minor-party candidates to support stronger major-party candidates. This effect is called Duverger's law. Most affects: instant-runoff (RCV), two-round system, and (especially strong for) plurality. Also affected: Borda, Score, approval voting. Immune: Coombs' method, antiplurality, rated voting rules (e.g. score voting)
A voter ranks or rates a candidate lower in the hope of defeating them. For example, a voter may insincerely rank or rate a perceived strong candidate last in order to help their preferred candidate win. Game theory arguments suggest that if burial incentive is sufficiently severe, a method may induce a race to the bottom. Each voter follows a strategy that, in isolation, helps their candidate win, but the result is that a candidate that nobody likes gets elected instead. Most affects: Borda, antiplurality. Also affected: Most Condorcet methods, with the exception of IRV-Condorcet hybrids. Immune: Instant-runoff voting and plurality voting.
(sometimes called pushover or pied-piper. Includes raiding.) A voter gives a high rank to a weak (i.e. pushover or pied-piper) candidate, but not with the intent of getting them elected. Instead, the voter intends for the weak candidate to eliminate a strong alternative, who would otherwise keep the voter's preferred candidate from winning. Party raiding is a well-known example of such a strategy.Negrin, Matt. "Could Democrats Decide the Winner in Michigan?" ABC News, 28 Feb. 2012. Retrieved 9 Dec. 2017. Most affects: Multi-round rules like instant-runoff, two-round, and primary elections. Immune: Plurality and all commonly-used rated voting systems, including score voting and approval voting. (sometimes called leveling. Includes bullet voting, exaggeration, and truncation.) Compression is a strategy where a voter refuses to disclose which of two candidates they honestly prefer (i.e. both candidates are given the same rating or ranking). Compression is unique in that it does not involve a rank reversal: in compression, a voter who prefers A to B doesn't claim to prefer B to A, only that they're indifferent between them. This is most common in rated voting rules, where the optimal strategy often involves exaggerating the differences between candidates, assigning high-quality candidates a perfect score and low-quality candidates a score of zero. Most affects: Rated voting. Immune: Random ballot and methods that require strict rankings. Plurality voting is immune to truncation.
(sometimes called quorum-busting or a walkout) Technically a form of strategic "non-voting", strategic abstention can be used in no-show paradoxes. It is an especially widespread kind of strategy in the case of quorums, where a group of voters can prevent a law from passing by ensuring the turnout is too low to meet a certain threshold. Most affects: Common in case of participation quorums, resulting in quorum-busting. Immune: Point-summing systems (i.e. score voting and positional voting) or Rawls's method and its converse.
Frequency and susceptibility
Strategic voting is highly dependent on the voting method being used. A strategic vote which improves a voter's satisfaction under one method could have no effect or be outright self-defeating under another method. Gibbard's theorem shows that no deterministic single-winner voting method can be completely immune to strategy, but makes no claims about the severity of strategy or how often strategy succeeds. Later results show that some methods are more manipulable than others.
Michel Balinski and Rida Laraki, the inventors of the majority judgment method, performed an initial investigation of this question using a set of Monte Carlo simulated elections based on the results from a poll of the 2007 French presidential election which they had carried out using rated ballots. Comparing range voting, Borda count, plurality voting, approval voting with two different absolute approval thresholds, Condorcet voting, and majority judgment, they found that range voting had the highest (worst) strategic vulnerability, while their own method majority judgment had the lowest (best). Further investigation would be needed to be sure that this result remained true with different sets of candidates.
Party-list proportional methods typically show less strategic voting, although the existence of electoral thresholds can lead voters to vote strategically to avoid wasted votes.
Mitigation Switching from a method that is highly manipulable to one that is more resistant would help discourage widespread strategic voting, all else equal.
From a voting method, it's possible to design another voting method that attempts to strategize on behalf of the voter. Such methods are called declared strategy voting methods, and seek to improve a method's strategy resistance by moving the strategy into the method itself. DSV methods have been proposed for plurality voting, approval voting, and score voting.
Coordination
The nature of strategic voting requires some level of coordination. Because it is a system of voting that assesses outcomes before casting one's vote, coordination towards the outcome is a prerequisite for strategic voting. There must be some kind of coordination device in order to vote strategically. Coordination devices, though not the only device, are typically opinion polls. For example, if there was two similarly appealing majority candidates and a minority candidate and voters were given the choice between the three, polls are imperative to coordinating votes for either majority candidate. Polls are then a tool used by voters to ensure that the minority candidate doesn't obtain office, which would be adverse to Duverger's law.
Coordination may also arise through social influence, affected from observing others' voting behavior. Computational models of strategic voting suggest that repeated interaction and imitation within social networking can cause particular voting strategies to spread among voters. The process may contribute either to greater electoral consensus or to persistent divisions between groups of voters.
Tactical voting may occur in isolation or as part of an organized campaign. In the former situation, electors make their own judgement as to the most effective way to (typically) prevent the election of a specific candidate or party. In the latter, one or more parties or groups encourage their supporters to vote tactically in an effort to influence the outcome.
The form that coordinated tactical voting takes depends largely on the electoral system of the polity. For example, vote thresholds in a proportional representation electoral system are found to prompt voters to coordinate their votes. For example, in the 2018 Swedish election, thresholds led voters who identified with a small party to consider the outcomes of their votes on the system of government as well as the other parties' outcomes. The voting history of the polity also influences coordination efforts: in proportional representation systems with little voting history, voters tend to vote more sincerely than engage in strategic voting due to the unknown nature of what party will be most viable.
Organized tactical voting in which a political party mounts a campaign calling on its supporters not to vote for their own favored candidates, but for those of a party which it perceives as more likely to defeat a common opponent, is less common. An example is the 1906 United Kingdom general election, where the Liberals (incidentally, the predecessors of the Liberal Democrats from the previous example) and the then-insurgent Labour Party (founded in 1900) agreed on the Gladstone-MacDonald pact, under which certain Liberal candidates would stand aside in favor of Labour ones, again to ensure that the Conservative candidates would not win on the basis of a split anti-Tory vote.
An intermediate case also exists, where a non-party campaign attempts to coordinate tactical voting, typically with the goal of defeating a certain party. Cases of this include the Canadian Anything But Conservative campaign, which opposed the Conservative Party of Canada in the 2008 and 2015 federal elections, or the Smart Voting campaign organized by Russia's Anti-Corruption Foundation with the goal of opposing and weakening the United Russia party in the 2021 Russian legislative election. This also occurred in the UK 2019 General Election, where voting groups such as Tactical Vote and Turf Out The Tories, promoted tactical voting to prevent a Conservative majority.
Examples in real elections Canada The observed effect of Duverger's law in Canada is weaker than in other countries.Duverger's Law of Plurality Voting The Logic of Party Competition in Canada, India, the United Kingdom and the United States, 2009, André Blais, Bernard Grofman, Shaun Bowler In the 1999 Ontario provincial election, strategic voting was encouraged by opponents of the Progressive Conservative government of Mike Harris. This failed to unseat Harris but succeeded in suppressing the Ontario New Democratic Party vote to a historic low.
In the 2004 federal election, and to a lesser extent in the 2006 election, strategic voting was a concern for the federal New Democratic Party (NDP). In the 2004 election, the governing Liberal Party was able to convince many New Democratic voters to vote Liberal to avoid a Conservative government. In the 2006 elections, the Liberal Party attempted the same strategy, with Prime Minister Paul Martin asking New Democrats and Greens to vote for the Liberal Party to prevent a Conservative win. The New Democratic Party leader Jack Layton responded by asking voters to "lend" their votes to his party, suggesting that the Liberal Party was bound to lose the election regardless of strategic voting. This failed to prevent the Conservatives from winning the election, although they did not win a majority of seats.