If you wanted more proof of the Wisdom of Crowds, you got it last week when the Citizens’ Assembly of BC, a representative group of 153 citizens selected at random from the voter rolls to identify the best electoral system for that province’s voters, issued its report on Thursday. By an overwhelming 90% plurality, the Citizens Assembly selected a system called PR-STV, which stands for Proportional Representation through Single Transferable Vote. Their recommendation will be put to a vote in a referendum as part of next Spring’s provincial election, and the ruling government has agreed in advance that if it is approved, PR-STV will come into effect in the following election in 2009. The opposition NDP, which has been a long-time supporter of PR, is almost certain to agree to be bound by the results if they are elected in the Spring election as well. The only real question now is whether existing power groups will lobby hard enough to convince the people to defeat the resolution.

PR-STV is a little known voting system used in Ireland, Malta and parts of Australia that combines elements of two other voting systems: Proportional Representation with Closed Party List (PR-CPL), the system used in much of Europe in which citizens vote for a party, and the top names on each party’s slate are elected in proportion to the percentage of votes each party receives; and Instant Runoff (or Alternative) Voting, in which citizens rank the candidates for office in their constituency, the candidate with the lowest votes is dropped off and that candidate’s votes are transferred to their voters’ second choice, continuing until one candidate accumulates a clear majority of the votes.

PR-STV is an attempt to accomplish three things reasonably well; and while other systems accomplish each better, no other system even attempts to achieve them all:

  • Elected candidates are representative of and responsible to specific geographic constituencies
  • Parties receive seats in government proportional to their share of overall popular vote
  • Small, new parties have a reasonable chance of electing at least one candidate

This is a difficult balancing act. PR-STV requires larger constituencies than the first-past-the-post system we in North America are accustomed to. It isn’t fully proportional: A party with a plurality (more than any other party) but minority (<50%) of the popular vote can still end up with a small majority of seats, and new parties must have significant support (at least 20%) if they hope to get candidates elected. Still, PR-STV does get much closer to the ‘perfect compromise‘ of the three objectives above, than any other system. In a recent study for the Blair government in the UK, whose fragile majority would disappear under PR-STV, the commission clearly recognized the superiority of this system over any other (and in the UK, at various levels, there are many options to choose from). More importantly, the commission’s report shows that PR-STV works in practice, and does so with a ballot for voters that’s ‘as easy as 1-2-3’.

Here’s how it works:

  1. The country, state or province is broken up into logical urban and rural agglomerations, with homogeneous economic characteristics and conforming to existing municipal boundaries (no gerrymandering allowed). The population of each agglomeration is computed and each is assigned a number of seats proportional to its percentage of the total number of eligible voters, so that the total number of seats remains unchanged. Any agglomeration with fewer than three seats by this calculation is combined with the closest similar agglomeration, and any agglomeration with more than seven seats by this calculation is split into two or more agglomerations, consistent with existing municipal boundaries, until each agglomeration has between 3 and 7 seats, These agglomerations become the electoral constituencies for PR-STV.
  2. Each party can nominate candidates up to the number of seats for grabs in each constituency. The ballot looks something like the picture above: candidates of each party in a column, with both the candidates of each party and the parties printed in random order from ballot to ballot (easy for printers to do in this day and age). The instructions on the PR-STV ballot are shown on the example above.
  3. Voters enter a 1 in the circle for their top choice, 2 for their second choice etc. They can rank all candidates in order, or rank fewer, or even just one, if they wish.
  4. Here’s where it gets complicated, and fair. The quota for election is the number of ballots cast, divided by (the number of seats to be filled +1). For example, if 180,000 ballots are cast in the constituency, and 5 candidates are to be elected from the 15 on the ballot above, the quota is 180,000/6 = 30,000 votes. Any candidate who achieves the quota in first-choice (‘1’) votes is declared elected.
  5. Next, the second choice votes of those candidates who exceeded the quota are tabulated, and reallocated to the remaining candidates in proportion to the number of votes in excess of the quota. So if a candidate gets 45,000 ‘1’ votes, each of that candidate’s second choices receive 1/3 of a vote for each ‘2’ vote they received, and the total of these 15,000 transferred votes are subtracted from that candidate, reducing his total to exactly the quota.
  6. Once all votes in excess of the quota have been transferred in this manner, if there still aren’t enough candidates who have reached the quota (five in our example), then one by one, the candidates with the lowest accumulated votes are dropped off, and their second (or third, fourth etc. if their second choice has already been elected or eliminated) choices are reallocated to the remaining candidates.

Here’s an example of how the ballot counting under PR-STV would work, courtesy of the UK Electoral Reform Society (another organization that speaks highly of PR-STV). If you want to play around with it, there’s a computer program that registers all your test votes and does all the calculations for you here. I’ve tried it, and it works, as long as you have a reasonalbly large number of voters in your test. And it’s really hard to explot to a party’s advantage: If a major party publishes lists telling party members what order to list its 3-7 candidates in, that can backfire and put the lower-ranking candidates on their list out of contention before the second and third-place votes get counted.

What are the drawbacks of this system? Other than the less-than-perfect way it addresses each of the three objectives bulleted above, these are the criticisms and challenges:

  • Elected officials need to be more focused on the needs of their constituency; a major party nomination isn’t enough to ensure a win. As far as I’m concerned, if they have to neglect their constituency to focus on ‘larger’ issues, they don’t deserve to be elected.
  • Under PR-STV, majority governments are rare, so coalitions are needed to pass legislation. Experience indicates, however, that minority governments are more responsive and responsible to voters, less likely to take extreme positions, and just as able to act decisively when an emergency arises.
  • Politicians hate it, and politicians with safe seats under the current system will probably do everything they can to sabotage its introduction. They’d actually have to work to be re-elected.
  • There are concerns that the ballot is too complicated for voters to understand. I don’t get this concern — if a voter can’t write 1, 2, 3… in order of preference then they probably aren’t capable of voting. I have greater concerns that the tabulation algorithm, or any machine-readable ballot, could be tampered with: It would be essential to use paper ballots and take the time to count them manually. If it takes an extra day, so be it — better than taking an extra month or two to deal with disputes over irregularities. I’m sure the areas that have been using PR-STV for years can give us some pointers. Canada uses manually-counted paper ballots, and election results here come out faster than they do in the US.
  • One concern I have is that wingnuts with a hard core of 15-20% support in their constituency could end up being elected under PR-STV, and even holding the balance of power in a coalition, but I guess that’s the cost of true democracy; I’m sure some people would have the same concerns about my friends in the Green Party.

If this summer’s Canadian election had been conducted using PR-STV, the Liberal minority would have been even smaller than it was, but the separatist Bloc QuÈbecois would have won many fewer seats, and the Liberals and NDP, along with a handful of Green Party electees, would be able to govern as a coalition without the need for support from either the anti-federalist Conservatives or Bloc QuÈbecois. The opposition parties all favour some form of electoral reform for Canada, and the NDP has made a referendum on reform a condition of its continued support for the Liberals. The BC report is likely to give this national initiative tremendous momentum, and make PR-STV the preferred alternative for consideration for the country as a whole.

If the 2000 US election had been conducted using PR-STV, Gore would be president (there’s no room for an anachronistic electoral college under PR-STV), since he would have received the lion’s share of second-place votes when Nader was eliminated. Democrats would probably also have won more seats than Republicans in Congress (which under PR-STV would be combined into a single ‘House’), though it’s possible that some third parties and independents would have held the balance of power, reining in the House from some of its more flagrantly partisan and pork-barrelling actions. What’s more important, a lot of the old hacks would have been turfed out in favour of more moderate, competent candidates from their own party, bringing new ideas, youth and collaboration to the House.

The study for the Blair government looked at the entire history of various PR voting around the world. Here’s what they found:

  • 94% of PR-STV voters voted for more than one candidate; 8% ranked all the candidates
  • 81% ‘split’ their vote, voting for candidates of more than one party; 49% voted for candidates of three or more parties
  • PR-STV did foment some intra-party feuds, with older, established candidates objecting to having to run against younger candidates with new ideas from their own party (heh, heh, poor guys)
  • Contrary to expectations, alas, PR-STV had no significant impact on voter turnout

I know it’s hard for Americans to think about electoral reform with the critical election pending next week. But I predict that, whoever is ultimately declared the winner in that election, controversy, litigation and acrimony will ensue for months, even years. The sad truth is that the US electoral system is simply incapable of accurately reflecting the will of the electorate, either in the vote for president or the votes for Congress. If Americans finally wake up after a second consecutive electoral debacle and vow “never again”, modern electoral systems like PR-STV that are immune to most of the abuses that have destroyed the credibility of the American system, might merit a closer look.

This entry was posted in How the World Really Works. Bookmark the permalink.


  1. Dermot Casey says:

    Davewe don’t use closed party lists in Ireland as part of PR-STV. And PR as we fondly refer to it does mean that people get elected for ‘minority issues’. But that’s one of the wonderful parts of democracy. Twice it was put to the people in a referendum to change the voting system and twice it was rejected. If you’ve any questions drop me an emailDermot

  2. Derek says:

    Hey, how can I get elected as a wing-nut? I’m a big hardware freek.Seriously, this sounds like a pretty good system. Would it work at the county level? or would the state level be the smallest level it’d be appropriate for?

  3. Rajiv says:

    Proportional representation, and the lack of it was the sole reason for the partition of India into India and Pakistan — and resulted in a host of problems — including the rise of terrorism on that region of the worldIndian Congress leaders were too enamoured of the British system to make the necessary adjustments. Even now, an Indian system based on proportional system would work much better

  4. Dave Pollard says:

    Dermot: Thanks. We in the Green Party are committed to PR, and the leaders seem poised to embrace PR-STV as the preferred solution for Canada. We’ll need lots of help explaining it to voters.Derek: It’s been used at all levels where there are at least 3 people to be elected. In situations with only 1 person to be elected (like the vote for the US President), it can still be used, but turns out to be identical to Instant Runoff / Alternative voting, i.e. the bottom candidate keeps dropping off and votes redistributed to second choices until one candidate accumulates >50%. It still works better than the current system, but not as well as PR-STV.Rajiv: Useful history lesson, and links — thanks.

  5. What? No punch cards or paperless electronic voting machines? Why can’t we enjoy the fun of being leaderless for up to a month after the election just like those guys to the south.In all seriousness though, I think it is time that we look at reforming how we elect our leaders and hopefully force them to be more accountable to the people (and not their lobbyists and political donators).

  6. Chui says:

    I’m not sure if you can characterise the coalition government in Israel as not been extreme. A minority government can be extremely susceptible to high pressure groups. In Australia, Senator Harradine, an independent for years held the deciding vote. Often he’d only vote for a bill if some other conservative issue was included in the bill, this can be about sex-education, discrimination of gays, etc.The current system, bad as it is, allows the government to be swept out of power without any backroom manouevering to try to stay in power.

  7. Life Tenant says:

    Dave, this is a great post, and it addresses some of my skepticism about the ‘wisdom of crowds’ thesis you’ve been advancing recently. It seems to me that crowds have as much madness as wisdom. Traders on the U.S. stock market weren’t ‘wise’ when they drove up the market in September 1929, and they were no wiser when they drove it down in October. The trick is to create a system or a procedure that maximizes the amount of wisdom and minimizes the chances of madness. It’s a hard trick, witness the constant evolution of securities regulation, and the continued inflation, and bursting, of bubbles. The voting system you describe shows great promise.

  8. Daniel Grice says:

    The system is looking real strong. The concern over backroom politics isn’t that strong, as voters determine who their representatives will be, and can reject MLAs who don’t represent their

  9. David Barry says:

    I live in London, in England, but grew up in both parts of Ireland. In Northern Ireland all elections are held by stv, including elections to the European Parliament. Only elections to the Westminster Parliament in London are held by the first past the post system familiar to North Americans. In the Republic of Ireland all elections are held by STV. Not only does the system work but I find voting in London elections very frustrating as I am only allowed a single choice; I cannot rank the candidates, which once you are used to it, seems the obvious way to vote as it reflects the reality of most people’s opinions.It is however so different from the closed list system used here for electing Euro MPs that calling them both PR just confuses. I prefer First past the Post to closed list, and STV as my first preference (You see ranking is natural!)I would have thought that it would be a really good way of electing the House of Representatives in the USA but I dont see how you would get the reform through against vested interests. You see politicians hate it. It makes them directly accountable.

Comments are closed.