No to No to AV

So, Britain has a referendum on 5 May 2011 on whether or not to replace the first-past-the-post voting system for Westminster MPs with an alternative vote much like Australia’s (we call it “preferential voting”, although it’s actually one specific form of it). As an Australian political scientist living in Britain, I’ve been preparing myself for the inevitable disappointment when the forces of misinformation and apathy carry the day and the proposal fails, even though first-past-the-post is so obviously unfair at deciding a winner when more than two candidates are competing for a single seat, and even though its distorting effects on parliamentary representation are also obvious.

But even though I’ve tried to ignore the No to AV campaign, they insist on pushing their pamphlets through my door, and I insist on reading them, and I have to say something, even if it’s just “ARRRRGGGGGGHHH!”

Forget that they’re telling all of Britain that Australians want to change to first-past-the-post on the basis of a single problematic poll. Forget that they keep pushing racing analogies as if being able to run the fastest is what elections are supposed to measure. Forget the implication that Britons who support minor parties should accept effectively throwing away their votes as the price of that support.

No, the most irritating No to AV claim is that “the Alternative Vote ... gives some people more votes than others”, and that only first-past-the-post gives “one person, one vote”. The pamphlet I received went so far as to suggest that votes for “supporters of extreme parties” would be counted under AV more often than others, “again and again and again”. This only makes sense if counting means something other than everyone understands it to mean, as is easily shown with an analogy more meaningful than any horse-race:

Buy a bag of jellybeans, tip them out on the table, and sort them into piles by colour. Find the pile that has more than half the jellybeans, if there is one. How many jellybeans did you have to count to be able to tell? The answer, clearly, is all of them: the jellybeans in each pile, and the total number overall.

Now take the smallest pile and divide its jellybeans among all the others. Now look again for the pile that has more than half of them. How many did you have to count to be able to tell this time? Answer: all of them. You may have used a quick calculation as a short-cut, but you were still counting all of them. Not just the smallest pile.

Now imagine each jellybean represents a single vote which, if a distribution of preferences is required, is counted more than once. Counting second or third preferences doesn’t mean that some votes “count” more than others. The No to AV campaign is confusing different meanings of the word “count”, presumably in the hope of confusing everyone. AV is one person, one vote, all counted an equal number of times in any one electorate.

Which isn’t to say that first-past-the-post is never fair; in an election for a single seat with only two candidates, it’s as fair a system as any. But if we’re considering extreme-case scenarios (because how many seats in the last UK general election had only two candidates?), why not consider the opposite extreme?

Imagine an electorate with 100,003 voters and not 2 or 3 or 4 but 1000 candidates. Imagine an election under first-past-the-post where 998 candidates get 100 votes each, one candidate gets 101, and one gets 102. The candidate with 102 votes would win.

Now imagine that the election is run under AV, that these are all the first preferences, and that the candidate with 101 first preferences gets all of the others’ second preferences.

Nobody would win on the first count. The 998 candidates’ votes would be redistributed, and all the votes would be recounted. The candidate with 102 first preferences would lose to the candidate with 99,901 first and second preferences.

Which outcome would you prefer? Which would better reflect the will of the people? Which is based on “one person, one vote”?

The answer to the last question is obviously “both of them”—nobody has voted twice either way—which shows what a red herring it is. The answers to the others should suggest which system to support on 5 May. If it helps, think of the candidate with 102 first preferences (or first-past-the-post crosses) as “Nick”—Griffin or Clegg, whichever you least prefer.

27 April 2011 · Politics

Another thing that bugs me about the No to AV campaign is the assertion that FPTP prevents coalitions and all the horse-trading that implies, while completely ignoring that all a system of strong major parties does is move the horse-trading behind closed party doors. Actually, “horse-trading” is a needlessly negative term itself; the better term is “politicking”.

Even in a one-party state, plenty of politicking goes on: it’s just that you have to join the party to take part in it, which is why so many people in such states do. (Which is not to denigrate the principled few who reject the whole system and criticise from the sidelines, but neither should we forget those who try to change their systems from within: those are the very kinds of people who brought down the Iron Curtain.)

Even in a strong two-party system, parliament largely becomes an irrelevance, as all the policy decisions are made behind closed doors and presented to parliament as a fait accompli. If you want to influence those decisions directly, you have to make your way behind those doors.

A system of multiple parties where none commands an outright majority, with all the compromises and coalitions that result, offers more of us a meaningful say in government policy without having to give over half our evenings to local branch meetings or our entire lives to politics.

Not that AV will by itself produce such a system. In Australia, minor parties mainly have influence because of proportional representation in the Senate, not because of AV in the House of Representatives.

Added by Rory on 4 May 2011.