PV: An electoral wildcard

I must confess to being a little obsessed with electoral systems and the degree to which they are proportional. This comes in part from my experience of having been largely disenfranchised by the system presently used for UK general elections and local elections in England: First Past the Post (FPTP). I have never voted for a successful candidate in a parliamentary election, and that’s not because I’ve always voted for ‘fringe’ parties or for none, which I have on occasions but not every time. For most of my adult life, I’ve lived in Tory ‘safe seats’: unwinnable under FPTP even by Labour (for which I’ve never voted) or the Lib Dems (OK, I admit it).

So I’d like to see the introduction of a genuinely proportional voting system such as STV or a Party List system; although this question should not be dealt with in isolation from consideration of the English Question, as I pointed out in a comment to an article elsewhere. Realistically, it may be the case that, for elections to an English parliament or to the House of Commons under a reformed system, the present single-member constituency system will be retained, at least initially, although possibly in combination with a party-list element; whereas the proportional systems I favour require multi-member constituencies or county / regional / national lists.

If single-member constituencies are retained, it’s important to ensure that the voting system used is as fair and proportional as possible. I recently advocated what I now know to be called the Approval System, which involves the ability to vote for as many candidates / parties as you like without giving them a ranking. I think this has some advantages over the Alternative Vote (AV), which is the system Gordon Brown wants to hold a referendum about if Labour win the forthcoming election, in that it turns every expression of preference for a candidate into a genuine endorsement of them as MP if they are elected; whereas, under AV, most candidates are elected on the basis of a sizeable number of lesser-preference votes (people’s second, third or even lower choices).

I would argue that the result that the various single-member systems produce are broadly as follows:

  • FPTP: the winner is the first-choice candidate of most, but not necessarily the majority of, voters
  • AV: the winner is the candidate obtaining a majority of relatively high-preference votes if no candidate obtains a majority of first-choice votes
  • Supplementary Vote (SV – another single-member alternative to FPTP): of the two candidates gaining the largest number of first-choice votes, the winner is the one obtaining a majority of votes once second-choice votes for either candidate are added to their total
  • Approval Voting: the winner is the candidate enjoying the broadest base of support, but not necessarily universally strong support nor that of the majority.

In summary, AV and SV are attempts to overcome one of the main deficiencies of FPTP – the fact that the winner is frequently not supported by a majority of voters – by determining the winner on the basis of a majority of high-preference votes: guaranteed under AV and relatively likely under SV. By contrast, the Approval Vote aims to determine merely the candidate who most people are willing to give some level of support to, but without any attempt to ensure that this is a majority – although it often will be.

I would like to propose a new system, which I’m calling the Popular Vote (PV), which guarantees that any candidate who is the first choice of a majority of voters is automatically elected, while the winner of any poll that doesn’t produce an overall majority for any candidate is, quite simply, the most popular candidate among all voters. My system is a type of Borda Count, but with a revision that overcomes the main problems with that particular method.

It would work as follows:

  • PV is a form of preferential voting and, like a Borda Count, it assigns a number of points to voters’ preferences. The maximum number of points that can be obtained by any candidate is determined by the overall number of candidates. For example, if there are five candidates, the highest number of points you can give to any candidate is five (i.e. they would be your first choice).
  • However, unlike a classic Borda Count, voters would not be obliged to rank all of the candidates (e.g. from five to one); nor would they be obliged to award their top candidate five points. They could, for instance, decide to award their favourite candidate any number of points from one to five: a voter’s first-choice candidate would simply be the one to whom they gave the most points. In most cases, voters would give their preferred candidate five points, but it would be entirely up to them how many points they decided each candidate merited, the only restriction being that no more than one candidate could be given the same score; i.e. you couldn’t award the same number of points to more than one candidate, meaning that you’re obliged to demonstrate your preferences in the points you award to each candidate.
  • When the vote is counted, note is taken of the first preference of each voter, and if a majority of voters selects one of the candidates as their first choice, that candidate is elected. However, if no such majority is attained, the result is then determined by the number of points each candidate has obtained – the winner being the candidate with the most points.

This system provides a fail-safe way of ensuring that the will of the majority prevails, but that if there is no majority, the most popular candidate wins. The fact that preferences are assigned a given number of points guarantees that the candidate with most points is the most preferred candidate. And the fact that, unlike a classic Borda Count, voters are not obliged to rank all of the candidates prevents the winner from being the consensus candidate (e.g. someone who averages around three points out of five for most voters, beating candidates enjoying a high degree of preference from a smaller share of voters) because the number of points voters assign to candidates is a much more accurate reflection of the degree of support they really give to each of them.

This also overcomes one of the main objections to the Borda Count: that it can be rigged by factions or parties fielding more than one candidate to reduce the points value that their supporters will give to opposition candidates. My system also overcomes the vulnerability of the Borda Count to voting tactically by, for instance, giving a higher points score to weaker candidates in order to defeat the stronger candidate you don’t want to win. Both of these problems only exist if voters are obliged to give every candidate a ranking / points score. But if you can simply just leave the box next to the names of candidates you dislike blank, then the number of points you decide to give to each candidate you like is a true reflection of the degree of support you give them.

Let’s take my constituency, where there are five candidates at the forthcoming election: the three main parties plus UKIP and the Greens. Under a Borda Count, I would be obliged to rank all of the candidates from five (my favourite candidate) to one (least favoured). Under PV, I can give any number of points, or none, to any candidate; but once I’ve used any number from one to five for any candidate, I can’t give the same number of points to any other candidate.

The way I’d be likely to vote under PV would be: Lib Dem five points, UKIP four and Greens two – but no points for either Labour or the Conservatives. By contrast, under FPTP, I’m probably going to end up voting UKIP. The reason for giving five points to the Lib Dems under PV would be because the Lib Dem candidate would have a chance of winning under that system, given that the incumbent Tory MP polled only 47% of the vote last time round (not enough to win automatically under PV) while the Lib Dem candidate polled around 32%.

I would want the Lib Dem candidate to win only to help bring about a hung parliament, which creates circumstances in which constitutional and political reform could be favoured. A hung parliament also brings the West Lothian Question into play more than ever, as whichever party was in government would need the support of Scottish and Welsh MPs (and even parties) to pass English laws, which could exacerbate English people’s grievances about the WLQ and might focus the minds of the other parties to do something about it.

Under FPTP, by contrast, the Lib Dem candidate stands virtually no chance, unless the Lib Dems start to push the barrier of 30% support across England. In this context, it costs nothing to use my vote to register a protest against the Lib Dems’ venal abandonment of their previous support for a referendum about the UK’s membership of the EU: hence UKIP – and hence why I would still give UKIP four points under my PV system.

Under a Borda Count, my voting choice would be far more tactical: probably, Lib Dem five (for the same reason as under PV), Labour four, UKIP three, Greens two and Conservative one. The reason for giving four points to Labour would be purely tactical: to harm the Tory and enhance the chances of the Lib Dem candidate without any risk of Labour winning, because they came a poor third last time. But this is a travesty of my real feelings and opinions (loathing Labour, and detesting the Tories’ similar self-serving complacency about the political system and refusal to acknowledge the English Question in any form). There would still be an element of tactical voting on my part under PV, but the system itself still enables my real preferences to be expressed quite accurately: giving a greater priority to getting a hung parliament than registering support for UKIP’s position on the EU and, to a lesser degree, immigration; but nonetheless still showing strong support for UKIP by awarding them four points.

Under AV, by contrast, I’d probably vote UKIP first choice and Lib Dem second, on the basis that if the Tory didn’t win an outright majority, then my second preference for the Lib Dem would get a chance of being registered; but that there’s no point listing any further preferences, because the Lib Dem is going to finish second (or, very outside chance, first) anyway.

I think there can be no question that PV, as I’m proposing it, would produce much more representative results at a constituency level than either FPTP or AV, because if you don’t get an outright majority of first-choice votes for any candidate, the system ensures that the most popular candidate overall wins – and that that candidate is genuinely popular, as the number of points obtained is directly related to the real strength of support for each candidate.

PV wouldn’t necessarily produce a proportional election result at a national level, however, as it shares the deficiencies of any single member-constituency system in this regard. However, logically, it would produce a more proportional and representative result nationwide than either FPTP or AV in that it genuinely determines the most popular candidate in each seat. Therefore, by extension, the aggregate, national result would also be a better reflection of the degree of popularity enjoyed by each party across England and Britain as a whole.

And PV would definitely alter the result in each constituency, compared with FPTP, to a much greater degree than would AV. For example, in my constituency, it would be conceivable that the Lib Dem could win under PV, whereas, under either FPTP or AV, this is virtually unimaginable. In order to win, under PV, voters giving candidates other than the Lib Dem or the Tory their highest score would have to also give a large number of points to the Lib Dem, and a low score or no points at all to the Tory. You could say that this artificially engineers victories for candidates with a relatively low level of first-preference support. But I would argue that PV simply determines the candidate people most want to win, as opposed to the candidate most people want to win, which is what FPTP does. And remember, if a majority of people select any candidate as their first choice, the result under PV and FPTP would be the same.

Up to my readers to decide which system they think is best, and to measure their choice against how the different systems would affect their voting decisions at the upcoming election. I know which system I’d vote for: a system that empowers my vote rather than returning the same party with no real alternative time after time, even without an overall majority.


2 Responses

  1. […] my own invention that I previously termed ‘PV’ (or the ‘Popular Vote’). I previously described how this works as […]

  2. Hello. Your “popular vote” is very similar to “range voting” also called “score voting.” See
    http://rangevoting.org = http://scorevoting.net
    (which contains over 1000 subpages & a search engine…)

    The differences between your “popular” system & range voting are
    * Your system more complicated because you make the #possible scores be the #possible candidates rather than a fixed range such as “integers 0 to 99” or “reals 0 to 1”.
    I think this decision by you can cause certain paradoxes which plain range voting does not suffer from, such as “deleting a losing candidate can reverse his two rivals’ finish order.”

    * Your system also is more complicated because
    it has a special “majority-top winner” clause.

    Plain range voting is capable of refusing to elect a candidate, if one exists, whom a majority of voters rate unique-top.
    Your system (thanks to this clause) will never do that.

    Is this a “defect” in range voting that needs to be overcome via such a special rule, or is it a “feature”?

    Interesting question. You can work thru some example elections you design to make range voting look good, or make it look bad, vs the modification.
    I did this and it seems to me I can accomplish both design goals, but overall it looks like when plain range voting fails to elect a majority-top winner, it usually does so in a way which does not cause a lot of harm and which can avoid a lot of harm.

    E.g. if 75% of the voters want to elect Hitler who’ll kill the other 25% and use their money to reduce taxes, that’s a majority-top winner, but if the 75% think Schmoe is almost as good as Hitler, and the 25% of course say Schmoe is fine but Hitler is zero… then plain range voting elects Schmoe, avoiding a LOT of harm caused by electing the majority-top winner Hitler.
    Note in this example all votes were honest self-interested votes. If enough voters strategically exaggerated, plain range might still elect Hitler, in which case it’d agree with the majority-top rule, so no difference there.

    You can try to construct some example that goes the other way to try to make the majority-top winner look better than the plain range winner… but it seems hard to make it look a LOT better for society.

    But anyhow, this question (of whether to add a “majority-top special clause”) – is worth more study than I have so far given it.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: