Saturday, 9 July 2011

Probably Perfect...

Earlier this year the electorate of the UK proved themselves to be ~70% idiots by rejecting the opportunity to make a small step to reforming the nation's archaic voting system via the adoption of AV. The debate was palpably poor on both sides with many misleading claims and some outrageous lies, the ultimately successful "NO" campaign being particularly scurrilous with their literature.

Anyway, as result of this debacle, I resolved to come up with a voting system which addressed all the criticisms of both sides. The result of this consideration is the voting system I've designated "Pickering's Probably Perfect Vote" system - though in the interests of brevity I'm happy to drop the "probably" and shorten it to PPV.

The intrinsic benefits of this system are:
  • Gives proper PR for parliamentary representation
  • Totally eliminates safe seats - guaranteed!
  • Eliminates tactical voting- guaranteed!
  • One person, one vote
  • Voters just put a simple X (none of this tricky counting to 3)
  • Compatible with independent candidates (no party lists)
  • Cheap to implement (cheaper than FPTP!)
  • Combats voter apathy
  • Ensures all votes count equally
  • Defeats gerrymandering
So How Does It Work?
Under PPV an election is carried out just like under the current FPTP voting system. Each voter simply puts an X next to the one candidate they'd most like to win. The "magic" happens once the ballots have been cast: the ballot papers are all placed into one gigantic hat, shuffled well, and then a single paper drawn. The named candidate selected by that paper is the winner. Job done.

This technique works because the probability for drawing one particular candidate's name from the hat is of course proportional to the number of votes cast for them. As a candidate, getting most people to vote for you increases your chances of being selected, but unlike FPTP that doesn't guarantee you'll win. This may at first seem unfair, but in fact it's the crucial factor for achieving the balanced parliamentary representation that's so conspicuously deficient in the UK's current voting system.

While it's possible that a candidate could get selected with a relatively small vote over a more popular one, it's improbable that that will happen. And when the random element is averaged out by effectively repeated sampling over many constituencies, a remarkable picture emerges: from a political party point-of-view, the overall number of MPs selected by PPV matches amazingly well with the proportion of votes cast for that party. Computer simulations show the degree of agreement to be generally within 2% - a much better match than with FPTP.

So by randomly sampling the constituency votes, rather than by systematically counting them, the overall result is actually a whole lot fairer. This may at first seem counter-intuitive but is actually quite logical. The point is that it's clear that the principle of representing the choices of a particular constituency and the principle of representing the overall choices of the electorate are at odds with each other. Under FPTP the former is given complete precedence and the latter given no consideration whatsoever. It only appears that a reasonable result is achieved because of the natural variations due to cultural differences and clustering of political viewpoints to different constituencies. To appreciate this, imagine that PartyA had 34% of the vote, whilst PartyB and PartyC had 33% each. If this balance was distributed evenly across the whole country, then PartyA would actually have a 100% majority - a total travesty of fairness! Under PPV the parliamentary representations would be an almost perfect reflection of the voters's wishes.

So in fact the only reason FPTP works is by clustering the electorate into regions of social and political identity, but in doing so the system effectively disenfranchises anyone with minority views and at the same time creates the "safe seats" that have been blamed for so many MPs's indiscretions. Under PPV, every vote is equal and carries the same chance as the others of selecting the winning candidate.

And of course the geographical dependence of FPTP makes it open to manipulation by adjustment of constituency boundaries to favour a certain outcome - the process known as gerrymandering. Whilst that could still be attempted under PPV the act of increasing a party's selection in one constituency would similarly reduce their chances of selection in the neighbouring region thereby foiling any overall advantage.

The Big Bonus
One huge advantage of PPV over FPTP is that without the burden of all that laborious counting it will be much cheaper. While this shouldn't really be the prime concern in matters of democracy, it was an issue that was given much weight by the AV "NO" campaigners, so they all should really be delighted by the savings that PPV has to offer. However, it seems fair that this efficiency should be shared back with the electorate, so it's therefore proposed that as well as the draw for the election result, each legitimate voter is also allowed to enter a ballot (with their contact or bank details) for a second draw, the winner of which receives a prize of £5k. This addresses concerns of voter apathy with a carrot rather than a stick: a carrot bestowed by the adoption of a voting system which is finally truly fair for all...