Alternative Vote
Apr. 19th, 2011 11:43 amOne thing I've found interesting is that a lot of people are talking about the Alternative Vote option in the forthcoming referendum as being an obvious choice due to being fairer. This is strange, because in itself it's no more fair on a constituency by constituency level than the existing 'most votes wins' system. In fact, no simple voting system can fulfill all the fairness criteria that exist, and this is mathematically proven, so it's always worth trying to understand the ways in which any new system can be unfair in ways that the previous system was not.
The UK is quite interesting in having 3 major parties, which is handy as it's with a 3 party system that you can most easily expose the flaws with a voting system. A common complaint with the current system is as follows: imagine 100 voters vote as follows: Conservative 35, Labour 33, Liberal 32. In the current situation, the Conservative candidate wins. Yet many would say that since the Labour and Liberal candidates have more similar policies (debatable right now, but that has historically often been the perception), the choice that would satisfy most people - up to 65 of them, maybe more if some Conservative voters are close to undecided - would be for either the Labour or Liberal candidate to be elected as a compromise choice. This is why Alternative Vote allows you to specify a second preference, and a third, etc., as an attempt to capture this sort of compromise. It also helps in that you can vote for a smaller party which is very unlikely to get elected, while still having some influence over the final result through your subsequent preferences. At the moment people are dissuaded from voting for the smaller party because they are unlikely to be elected and that vote is then 'lost' when it could have been spent more wisely in deciding a close contest between your 2nd choice and your 3rd or 4th choice (for example).
However. Imagine the voting went like this: Conservative 49, Labour 26, Liberal 25. Under the current system, the Conservatives win. Now take the Alternative Vote system, and the Liberal candidate is eliminated and anybody who voted for him or her gets their 2nd choice vote counted instead. If we continue to assume that Labour and Liberals both prefer each other to Conservatives, and that therefore the new result is Conservative 49, Labour 51, under AV the Labour candidate wins. Seems reasonable. Except what has happened here is that the Liberal voters have found that their 2nd preference votes are worth just as much as a Labour or Conservative 1st preference. That in itself seems unfair because the other voters haven't had their 2nd choice counted. Labour won't mind, as they've come off best. But what if you actually counted the Conservative 2nd choices, which have been ignored - they might be mostly Liberal! This is what the 'No to AV' campaigners mean when they say that some people's votes would count twice - the Liberal voters, perhaps knowing they were in 3rd place to begin with, got a 'free' vote. So, having established that a second choice can count for just as much as a first choice, how about we count everybody's first and second choices together? In this situation, now with 200 votes, you might get a score more like Conservative 49, Labour 51, Liberal 100. So the Liberals win.
One set of voting results, yielding 3 possible outcomes, all appearing fair in some sense and unfair in another.
Interestingly, the way "tactical voting" would operate under the Alternative Vote system is a bit different to now. Instead of deliberately picking your 2nd choice party if you think your 1st choice has no chance, people are best served by ordering their preferences from least likely to win to most likely - after all, if you back a winner from the start, your 2nd preference (and 3rd, and so on) are wasted, but if you back the first loser, your 2nd choice counts, then potentially your 3rd, all the way up. This means we're likely to see a lot more votes cast for fringe and extreme parties - what used to be just a protest vote will get augmented by the people who currently vote for the mainstream parties to avoid wasting their vote. This is good news for those who feel that the main 3 parties are not radical enough, but I think people might be dismayed when they see the strength of opinion on the far edges of political spectrum.
The UK is quite interesting in having 3 major parties, which is handy as it's with a 3 party system that you can most easily expose the flaws with a voting system. A common complaint with the current system is as follows: imagine 100 voters vote as follows: Conservative 35, Labour 33, Liberal 32. In the current situation, the Conservative candidate wins. Yet many would say that since the Labour and Liberal candidates have more similar policies (debatable right now, but that has historically often been the perception), the choice that would satisfy most people - up to 65 of them, maybe more if some Conservative voters are close to undecided - would be for either the Labour or Liberal candidate to be elected as a compromise choice. This is why Alternative Vote allows you to specify a second preference, and a third, etc., as an attempt to capture this sort of compromise. It also helps in that you can vote for a smaller party which is very unlikely to get elected, while still having some influence over the final result through your subsequent preferences. At the moment people are dissuaded from voting for the smaller party because they are unlikely to be elected and that vote is then 'lost' when it could have been spent more wisely in deciding a close contest between your 2nd choice and your 3rd or 4th choice (for example).
However. Imagine the voting went like this: Conservative 49, Labour 26, Liberal 25. Under the current system, the Conservatives win. Now take the Alternative Vote system, and the Liberal candidate is eliminated and anybody who voted for him or her gets their 2nd choice vote counted instead. If we continue to assume that Labour and Liberals both prefer each other to Conservatives, and that therefore the new result is Conservative 49, Labour 51, under AV the Labour candidate wins. Seems reasonable. Except what has happened here is that the Liberal voters have found that their 2nd preference votes are worth just as much as a Labour or Conservative 1st preference. That in itself seems unfair because the other voters haven't had their 2nd choice counted. Labour won't mind, as they've come off best. But what if you actually counted the Conservative 2nd choices, which have been ignored - they might be mostly Liberal! This is what the 'No to AV' campaigners mean when they say that some people's votes would count twice - the Liberal voters, perhaps knowing they were in 3rd place to begin with, got a 'free' vote. So, having established that a second choice can count for just as much as a first choice, how about we count everybody's first and second choices together? In this situation, now with 200 votes, you might get a score more like Conservative 49, Labour 51, Liberal 100. So the Liberals win.
One set of voting results, yielding 3 possible outcomes, all appearing fair in some sense and unfair in another.
Interestingly, the way "tactical voting" would operate under the Alternative Vote system is a bit different to now. Instead of deliberately picking your 2nd choice party if you think your 1st choice has no chance, people are best served by ordering their preferences from least likely to win to most likely - after all, if you back a winner from the start, your 2nd preference (and 3rd, and so on) are wasted, but if you back the first loser, your 2nd choice counts, then potentially your 3rd, all the way up. This means we're likely to see a lot more votes cast for fringe and extreme parties - what used to be just a protest vote will get augmented by the people who currently vote for the mainstream parties to avoid wasting their vote. This is good news for those who feel that the main 3 parties are not radical enough, but I think people might be dismayed when they see the strength of opinion on the far edges of political spectrum.