Investigations of a Dog

Still very busy with the book, but if you’re registered to vote in the UK please consider voting Yes to the Alternative Vote system in the referendum on Thursday 5th May. (I know that lots of my readers have nothing to do with the UK. For you this post will be boring and irrelevant. The good news is that if we lose the referendum I’ll want to bury this post as quickly as possible by posting lots of other things. But if we win it’ll probably stay on top until October.)

Our current system for electing members of the House of Commons is known as “First Past the Post”, but it shouldn’t be. Some people have pointed out that there is no post. There are no fixed absolute criteria for winning a seat. The share of the vote needed to win is always relative to the share of the vote gained by the candidate who comes second. (Conservatives are usually against all forms of relativism, real or imagined, but the No campaign supports a relativist system!) First Past the Post should really be called First Past the Second. In practice it leads to some very strange results. I could tell some hypothetical stories about made-up parties with euphemistic names, but my inner empirical historian demands concrete examples of real things that really happened in reality. Below are some results from the 2010 general election (I’ve omitted lots of third placed and lower candidates who aren’t relevant to my argument). For the purposes of this discussion I’ve assumed that no-one voted tactically. This is a big, unfounded and probably wrong assumption, but the election results make it impossible to tell the actual preferences of the voters. What use is a voting system that doesn’t tell us what the voters want?

I’ve cherry picked some of the most extreme examples here, but if these results are even theoretically possible then the system is seriously broken. Gordon Brown and John Mann have overall majorities, but the other three winners don’t. Gordon Brown’s share of the vote is almost twice Glenda Jackson’s but they both win the same thing: a seat in the House of Commons. Antony Calvert has a bigger share of the vote than Glenda Jackson but came a close second to Ed Balls; Keith Girling has a bigger share of the vote than Glenda Jackson but came a more distant second to John Mann. Chris Philp and Edward Fordham, who didn’t win a seat, each have a bigger share of the vote that Simon Wright, who did win a seat. This is a bizarre effect of the First Past the Second system: losing candidates in some constituencies get a bigger share of the vote than winning candidates in other constituencies. In 2010 even someone who came third got a bigger share of the vote than someone who won a seat! It’s hard to find any reason why it’s fair, obvious or natural that Glenda Jackson or Simon Wright should win a seat with less than a third of the votes.

Horse racing is a terrible analogy for parliamentary elections but let’s try it anyway. If Hampstead and Kilburn was the Grand National, Glenda Jackson, Chris Philp and Edward Fordham all fell at the 20th fence but Glenda Jackson was declared the winner because she hit the ground a fraction of a second later. If Morley and Outwood was the 2,000 Guineas, Ed Balls was declared the winner because he was leading after six furlongs. How about football then? First Past the Second is like every match being sudden death right from the kick-off. Instead of everyone playing for 90 minutes, the first team to get one more goal than the other instantly wins. But what if the other team would have scored more goals before full-time? Doesn’t matter. The team that scores first should win. Obviously. (All this paragraph really shows is that sporting metaphors are potentially misleading and open to lots of different interpretations. They can’t be used to prove that either voting system is better than the other. Please don’t anyone quote these sporting examples out of context, because they aren’t actually good arguments.)

Glenda Jackson, Simon Wright and Ed Balls all won seats even though far more people in their constituencies voted against them than for them (in First Past the Second a vote for one candidate is always effectively a vote against all the others, even if it isn’t intended that way). Maybe all the people who voted for Chris Philp and Edward Fordham hated Glenda Jackson and wouldn’t want her as their MP under any circumstances; maybe Edward Fordham’s supporters still think that Glenda Jackson would be better than Chris Philp; maybe voters’ views are more diverse than that. We can’t possibly know unless we ask them. AV is a simple and efficient way to do just that. Voters can rank as many or as few candidates as they want in order of preference by writing numbers next to their names from 1 to whatever (but note that the referendum itself uses the traditional method of putting a cross next to one choice). If one candidate gets more than 50% of first preferences they automatically win. Gordon Brown and John Mann would win under both systems as long as all their votes were genuine first preferences. If the same votes were cast under AV there would be no possibility that Keith Girling or Douglas Chapman could have moved into first place. The winning post is at 50% and once a candidate has passed it the result cannot change any further. In the other three constituencies there is no clear winner in the first round. In this situation the least popular candidate is eliminated and their votes are redistributed to the voters’ next preferences. This continues until someone does have more than 50% or there are no more preferences left. Everyone gets their vote counted in every round unless they choose not to specify enough preferences. Supporters of small parties will be less likely to waste their votes but will have to make do with second, third or fourth preferences when supporters of big parties are still voting for their first choice. The winner will usually have to gain at least grudging acceptance from a majority of the voters. This will be impossible for the BNP, who under the current system have at least some hope of winning with 29% like Simon Wright.

As long as we have single member constituencies, AV is the best way to select the candidate who is most representative of the constituency as a whole. It would level the playing field by making the winning post the same in every constituency. Candidates could win with less than 50% of the vote if there are no more preferences to be redistributed, but even then the difference between the highest and lowest winning share of the vote should be much less extreme. AV would make it much less likely that a losing candidate in one constituency could get a bigger share of the vote than the winner in another constituency. First Past the Second can sometimes produce the same results as AV, but it doesn’t guarantee anything more than representation of the biggest minority. In Hampstead and Kilburn the biggest minority is only marginally bigger than the second and third biggest minorities but the winner takes all without even passing the post.

The fundamental problem of electoral reform is that there are really two problems: how to select candidates who are representative of their constituencies, and how to make the distribution of seats representative of each party’s share of the national vote. AV solves the local problem but not the national one. First Past the Second solves neither. The national problem could be solved by a second chamber elected by Proportional Representation. The coalition is committed to implementing this without a referendum and regardless of the result of the referendum on AV for the House of Commons. This will make a solution possible, although not inevitable. One house for local representation and one for national representation would be better than one elected house and one unelected house. The referendum doesn’t give us a choice of PR. We have to choose whether we are for or against electoral reform. You should only vote No if you are completely happy with the current system and its results.