Election Maths: Proportional Representation and Iterative Proportional Fitting
Friedrich Pukelsheim (Augsburg University)
Friday 29th June, 2012 15:00-16:00 Maths 203
Many electoral systems aim at what may be called double proportionality, to represent political parties in proportion to vote counts as well as to represent geographical districts in proportion to census figures. We show that the Iterative Proportional Fitting (IPF) procedure may be adapted to achieve both goals simultaneously. Since parliamentary seats are inherently discrete rather than infinitely divisible, discretized versions of the IPF procedure are needed, such as the Alternating Scaling (AS) algorithm or the Tie-and-Transfer (TT) algorithm. Double proportionality had its world premier during the Zurich City Parliament election in 2006, has been adopted by other Swiss Cantons to date, and may prove attractive for European Parliament elections in the future.