Winners and Losers under Proportional Representation

In elections under proportional representation (PR), proportionality between votes andseats is an ideal never achieved in reality. Discrepancies between votes and seats are mostprevalent under the most widely used PR seat allocation formula, D’Hondt/Jefferson. Thisarticle provides the first comprehensive model to explain the discrepancies between thenational seat allocation and the vote distribution under the D’Hondt rule. The model identifies six systematic factors that drive these deviations, ranging from the bias in the seatallocation rule to differences in the electoral rules (e.g. malapportionment) and party competition across districts.Thus, this model goes beyond the well-known rule that large partiesare overrepresented under D’Hondt by providing a quantitative point estimate of this deviation and of between-district inequalities. The model identifies the (expected) beneficiariesand losers of the seat allocation and predicts the magnitude of these penalties and benefits.At the level of political parties, the model predicts the gains and losses in terms of seats,relative to the parties’ proportional share of seats. Empirical tests on 156 elections from 23democracies, covering 3,653 political parties and 2,139 constituencies, are used to assess itsvalidity and to calculate the resulting substantial effects.