Two office-seeking parties compete on an ideological line. Build up the Downsian model from its assumptions, explore who wins, and understand why both parties converge to the median.
Why assumptions matter. The Downsian model's prediction — that both parties converge to the median voter — follows logically from a set of specific assumptions. Understanding which assumptions are doing the work tells us when we should expect the prediction to hold, and when it might break down. Click each assumption to explore it.
The seven assumptions of the Downsian model
1
Two-party system
›
Two and only two parties (or candidates) compete for office. This is key: with more than two parties, the logic of convergence breaks down — a third party can always outflank a centrist, stealing votes from the extreme.
2
One-dimensional ideological space
›
All political conflict can be placed on a single left–right spectrum. In reality, policy is multi-dimensional (economics, culture, foreign policy), which undermines the unique median.
3
Parties are fully flexible
›
Each party can credibly commit to any position on the spectrum. In practice, parties have histories, reputations, and activists who constrain where they can plausibly move — limiting real-world convergence.
4
Parties are office-seekers, not policy-seekers
›
Their sole goal is to win. Policy positions are instruments — means to an electoral end. If parties actually care about policy (as most do), they may refuse to move to the median even if it would win votes.
5
Voters cast sincere proximity votes
›
Each voter supports the candidate whose position is closest to their own ideal point. Voters who are equidistant abstain. This rules out strategic voting, party loyalty, or valence considerations.
6
Perfect information
›
Parties know the full distribution of voter preferences; voters know exactly where each party stands. In practice, voters have limited political information, and parties know only noisy signals about preferences.
7
Voter preferences are fixed
›
No matter what parties say or do during the campaign, voter ideal points do not shift. If campaigns genuinely persuade — shaping what voters want — then parties may have incentives to move away from the median.
The Downsian prediction. Under these seven assumptions, both office-seeking parties will converge to the position of the median voter. Neither party has any incentive to deviate: any move away from the median immediately hands the pivotal voter — and with her, the majority — to the opponent. The median is the unique Nash equilibrium of the game.
Downs's insight. Before the Downsian model, political scientists explained parties’ behaviour in terms of their ideologies, histories, and social bases. Anthony Downs asked: what if we treated parties like rational, self-interested firms competing for votes? His answer: if those seven assumptions hold, both parties will converge to the median voter — regardless of history or ideology. The prediction is derived, not assumed.
Drag to compete. 101 voters are evenly distributed from -50 (far left) to 50 (far right). The median voter sits at 0. Drag Party A and Party B to any position — the model instantly calculates votes and tells you who wins.
The ideological battleground — drag the parties
The dividing line (midpoint) is shown as a dashed line. Voters left of it go to A; right go to B. A voter exactly at the midpoint abstains.
—
Votes — Party A
mid = —
Midpoint
—
Votes — Party B
—
Move the parties to see the explanation.
The key rule. Party A wins whenever its distance from 0 (the median) is strictly less than Party B's. In other words, A wins if |a| < |b|. The median voter is always pivotal: whoever captures her, wins.
Best-responding to each other. Start the parties anywhere. Watch what happens when each party rationally responds to the other's position — moving just enough toward the median to win. Press Next Step to advance.
Step log
Nash Equilibrium reached. Both parties are at position 0 — the median voter's ideal point. Neither can gain by moving: any deviation immediately hands the median voter (and the majority) to the opponent. This is the unique Nash equilibrium of the Downsian game. This is the Median Voter Theorem.