Duverger’s Law is the principle that countries using winner-take-all elections tend to end up with two dominant political parties. First proposed by French political scientist Maurice Duverger in his 1954 book Political Parties, it remains one of the most widely cited rules in political science. The core idea is straightforward: when only one candidate can win each race, voters and parties naturally consolidate around two competitive options.
The Law and the Hypothesis
Duverger actually made two related claims, and they’re often conflated. The “law” states that plurality voting, where the single candidate with the most votes wins, favors a two-party system. The “hypothesis” (a weaker claim) states that proportional representation and runoff systems favor multi-party systems. Political scientists treat the first as a reliable pattern with strong evidence behind it. The second is considered more of a tendency than a rule.
The distinction matters. In a proportional representation system, like those used in much of Europe, a party that wins 15% of the national vote gets roughly 15% of the seats. Small parties survive and even thrive. In a plurality system like the United States or the United Kingdom, winning 15% of the vote spread evenly across the country gets you zero seats. That structural pressure is what drives the two-party pattern.
Why It Works: The Mechanical and Psychological Effects
Duverger identified two forces that push plurality systems toward two parties. The first is mechanical: the voting system itself filters out smaller parties. If three candidates split the vote 40-35-25, the candidate with 40% wins the seat and the other two get nothing. Over repeated elections, parties that consistently finish third simply stop winning seats and fade from power.
The second force is psychological, and it’s the more interesting one. Voters don’t want to waste their votes. If your preferred candidate has little chance of winning, you face a dilemma: vote sincerely for the long shot, or vote strategically for your second choice to prevent your least-favorite candidate from winning. Most people choose the strategic option. Researchers have confirmed this in experimental elections, where subjects consistently shift their votes away from trailing candidates toward the leading two, especially when they have information like polls showing who’s ahead and who’s behind.
This “wasted vote” psychology creates a self-reinforcing cycle. As voters abandon third-party candidates, those candidates poll even worse, which causes more voters to abandon them, which makes them even less viable. Over time, the political landscape consolidates around two major parties that each try to build the broadest possible coalition.
Real-World Exceptions
If Duverger’s Law were absolute, countries like the United Kingdom, Canada, and India wouldn’t have significant third parties. But they do, and understanding why reveals the law’s limitations.
The key factor is geography. Duverger’s Law operates most reliably at the district level, not the national level. A party that has broad but thin support across an entire country will get crushed by plurality voting. But a party with concentrated regional support can win individual districts even if it’s irrelevant nationally. Canada’s New Democratic Party, for instance, won 17% of the national vote in the 2006 elections. That support was spread very unevenly: in one-third of districts the party got less than 11%, but in six districts it won outright majorities. Regional and ethnic parties exploit this loophole regularly.
Research on competitive districts shows that strategic voting does suppress minor party performance, by roughly 15% in the most competitive races compared to less competitive ones. But most of the variation in third-party success comes down to political geography rather than strategic voting. Where a minority group or regional identity is concentrated in a specific area, plurality voting can’t eliminate the parties that represent them.
How It Applies to the United States
The United States is often cited as the textbook example of Duverger’s Law in action. The country uses plurality voting for nearly all elections, and the same two parties have dominated since the 1860s. Third-party candidates occasionally surge in presidential races (Ross Perot in 1992, Ralph Nader in 2000) but never win, and their supporters almost always migrate back to one of the two major parties by the next election cycle.
The psychological effect is especially visible in American politics. Voters who prefer a third-party candidate are constantly told they’re “throwing away their vote,” and polling data reinforces the message. This isn’t just folk wisdom; it’s the exact mechanism Duverger described. The structural incentive to consolidate is so strong that American third parties tend to function more as pressure groups that push a major party’s platform than as serious contenders for office.
Why Political Scientists Still Debate It
Duverger’s Law is unusual in political science because it makes a bold, testable prediction, and the field is generally skeptical of “laws” in the way physics has them. The basic relationship between plurality voting and two-party dominance holds up well across countries and time periods, and experimental research has confirmed that the strategic voting mechanism works as predicted.
The debate centers on scope and precision. The law works cleanly at the district level but gets messier at the national level, where regional parties complicate the picture. It also doesn’t specify how long the consolidation takes, or predict which two parties will dominate. Some scholars argue it’s better understood as a strong tendency shaped by additional factors, like the type of social divisions in a country, whether ethnic, religious, or economic, and how geographically concentrated those divisions are. Cross-cutting social cleavages, where different types of diversity overlap in complex ways, produce different party dynamics than simple regional divides.
Still, as a core principle for understanding why some democracies have two major parties and others have five or six, Duverger’s Law holds up remarkably well nearly 70 years after it was proposed. If you want to predict roughly how many viable parties a country will have, the single most useful piece of information is what kind of electoral system it uses.