Better Voting Methods Through Technology: The Refinement-Manageability Trade-Off in the Single Transferable Vote

Abstract

The Single Transferable Vote (STV) is an attractive way of achieving representation that is proportional in terms of whatever characteristics of candidates voters value. Increasingly sophisticated methods of implementing STV have been advanced to overcome identified limitations of earlier methods. But every refinement comes at a cost of increased difficulty of understanding the vote-counting algorithm and increased cost of undertaking the count. This paper uses votes from actual elections to provide evidence about the frequency with which the choice of a particular STV method affects the outcome, and about the type of difference that different methods make. The most sophisticated STV method is CPO-STV, the comparison of pairs of outcomes by STV. This method avoids sequential exclusions and therefore overcomes the limitations of previous methods, that a paucity of votes in the early stages of a count can lead to the exclusion of a candidate who is the consensus choice of voters whose preferred candidates will be excluded at later stages of the count.