Voting matters — To advance the understanding of preferential voting systems

Issue 26, January 2009

(Complete issue in PDF, 420KB.)

Editorial

There are three items in this issue:

• A Buhagiar and J Lauri: STV in Malta: A crisis?
  Anton Buhagiar and Josef Lauri consider a critical problem in the use of STV in Malta. Although STV is well established in Malta, the two-party system has the effect of expecting a higher degree of proportionality than STV can actually deliver. The authors propose a very interesting solution to this problem.
• Brian Wichmann: Review — Mathematics and Democracy.
  Brian Wichmann provides a review of the book Mathematics and Democracy by Steven J Brams. Although the book has almost nothing useful to say about STV, it does contain some interesting analysis of preferential voting and also techniques for resolving conflicts.
• R. Sewell, D. MacKay and I. McLean: Probabilistic electoral methods, representative probability, and maximum entropy
  Roger Sewell, David MacKay and Iain McLean have produced a paper which introduces an aspect of voting not previously covered in Voting matters. This is the use of a non-deterministic algorithm using random selection. This paper is substantially longer than our usual, but it covers more — the relationship with Arrow's Theorem and also the mathematical background (as an appendix).

  The tutorial presented in the paper may well encourage people to consider the concepts in the paper seriously, in spite of the obvious problems in its use in real elections.

The actual application of preferential voting is quite widespread. To analyse the practical implications of specific counting algorithms needs appropriate voting data. The best data are from those elections for which the complete profile of the votes are available. Over the years, I have collected such data. To encourage its use, I have now placed the data at http://stv.sourceforge.net/ for free download.

Another aspect of the openness of a preferential voting process is the algorithm used. Due to the difficulty in specifying algorithms precisely, it is surely best to provide the algorithm as open-source software. The Australian Capital Territory passes both tests — the voting data is available and the counting program is open source. How can we encourage others to follow?

Voting matters — editorial understudy wanted

Dr Brian Wichmann is intending to step down shortly as editor of Voting matters. The McDougall trustees are grateful to Brian for his voluntary efforts in ensuring that it has been published one or, usually, twice every year over the past decade or so.

The trustees are seeking a replacement, and have welcomed Brian's suggestion to look for someone to "understudy" him as editor with the expectation that they would assume the editorship of Voting matters in due course.

That person should preferably be someone with relevant academic interests and/or experience, who would be able to make contact in person with the Trust at least annually. Good computer skills would be an advantage.

An interest in developing the scope of Voting matters would also be welcome; one suggestion is that it should seek to cover the preferential voting opportunities that exist in some list forms of proportional representation.

For expressions of interest, please contact: Paul Wilder, Voting matters, McDougall Trust, 6 Chancel Street, London, SE1 OUX or email admin@mcdougall.org.uk by 28 February 2009.

Papers with citations

1. Anton Buhagiar and Josef Lauri: STV in Malta: A crisis? (p1-12, PDF 324KB)
2. Brian Wichmann: Review — Mathematics and Democracy. (p13-15, PDF 80KB)
3. Roger Sewell, David MacKay and Iain McLean: Probabilistic electoral methods, representative probability, and maximum entropy. (p16-38, PDF 568KB)