3 ms·

Approval voting: The political reform engineers – and voters – love

dane-pgp 5y ago
Approval voting should be supported wherever it is an option to replace FPTP, but I'm always worried that if a voting reform is adopted, and an election produces a strange result, some people will blame the voting system and want to return to FPTP (poisoning the well for future reforms). The way I can imagine an approval voting election going badly is a situation where a major candidate - let's call him John Barron - is so disapproved of by supporters of the opposing party, that those supporters feel compelled to approve a "lesser of two evils" member of Barron's party, who ends up winning with an apparent landslide of votes. In such a situation, if voters had been able to express the relative strength or conditionality of their support for this "less of two evils" candidate, the election winner might have been a better compromise candidate, or, indeed, the opponents of Barron could have had their first choice elected if they had stayed ideologically pure and not tried to hedge their bets. This may seem like an edge case, and maybe in the long run the political culture of a country would change and new parties would emerge which would stop this result from happening, but I still feel that there is an inherent problem in asking voters to say which candidates they "approve" of, without really specifying what that means (or what the tactical implications are). It effectively requires ranking all the candidates in order, and choosing a threshold above which all candidates are "approved", and below which none are, but how that threshold is set is left as an exercise to the voter - an exercise which in practice has a right and wrong answer, which can only be determined after the result has been announced.
- jfengel 5y ago
  I don't think it would take a case as exotic as that to put voters off voting reform. The odds are that any election is still going to result in one of the two main parties winning. There may be occasional exceptions, just as there occasional upsets in FPTP, but for the most part there isn't any single Third Party that has the quiet approval of a huge range of voters. Rather, there will be many other parties, and the voters' second choices are likely to be either one main party, or blank. Even if they check "everybody but the main party candidates", it's unlikely to produce a non-main-party winner. I have a feeling that after voting reform, the strongest advocates will take note that nothing actually changed. They might declare fraud, or demand more reforms, but I don't see them accepting "Well, turns out that the people who compromise hardest assemble the biggest tent" as the answer. Maybe I'm wrong, and I don't think it would do much harm to try. Sure, some voters will be confused, but they're already confused so that's just a drop in the bucket. But I don't believe it will make anybody happy, either.
  - dane-pgp 5y ago
    I admit that most modest (and thus politically achievable) voting reforms would be unlikely to result in immediate wins for small parties, but that sort of stability is probably a desirable property of any electoral system. The main benefit in the short term would likely be that large parties (which are themselves coalitions of various factions, with internal dynamics that play out behind the scenes and usually without popular support) would split into smaller, more representative units. For example, if the Democrats split into a Hillary party and a Bernie party, the Bernie party might never have one of their members become the Speaker of the House, and two parties would formally or informally act as a coalition, introducing and passing pretty much the same bills as the (combined) Democrats do today, at least initially. Rather than advocates complaining that "nothing actually changed", though, I think they'd notice that the supporters of these new separate parties were reinvigorated, enjoying the ability to have a party that truly speaks for them and doesn't have to compromise at the stage of choosing its leader and identity. These new parties would allow the political discourse to change, and shift the Overton window, while also allowing parties to appeal to non-voters who had previously been faced with a choice of two evils, or uncompetitive districts.
dane-pgp 5y ago
> Unlike its competitor, ranked-choice voting, approval voting doesn’t require new machines, ballots or hardware. ... It’s a sensible, intuitive and tested approach that every voting machine in America could execute tomorrow. I understand the desire to be pragmatic and accept the reality of voting machines as they exist in America today, but I also think that any proposal which entrenches a dependency on voting machines is potentially a dead end for democracy. Even if you could guarantee that all the software and hardware used in voting machines was open for security researchers to audit, and there were no bugs or vulnerabilities, you would still be introducing the attack surface of people being able to lie and claim there are vulnerabilities and that the election had been stolen. Once we accept the premise that votes should be cast without using a machine (at least for 99% of voters who are physically capable) and counted by hand, consideration needs to be given to what that counting process looks like. Certainly approval voting is vastly preferable to ranked systems, since approval ballots can be split between different counting stations whose sub-totals can be easily summed, but the actual management of those physical ballots can be tricky, since counters either need to keep track with a tally system or group ballots into one of 2^N different piles (given N different candidates). An alternative system, which has all the benefits of approval voting (same ballot design, solves the spoiler problem, summable sub-totals), is Asset Voting. Basically it works the same as FPTP, except that after the initial count, candidates are eliminated (from least to most votes) and they can choose to redistribute their votes to another remaining candidate. In terms of counting, this can be done by hand using only N piles for N candidates, the way that countries have done it for hundreds of years and many still do to this day.

Approval voting should be supported wherever it is an option to replace FPTP, but I'm always worried that if a voting reform is adopted, and an election produces a strange result, some people will blame the voting system and want to return to FPTP (poisoning the well for future reforms). The way I can imagine an approval voting election going badly is a situation where a major candidate - let's call him John Barron - is so disapproved of by supporters of the opposing party, that those supporters feel compelled to approve a "lesser of two evils" member of Barron's party, who ends up winning with an apparent landslide of votes. In such a situation, if voters had been able to express the relative strength or conditionality of their support for this "less of two evils" candidate, the election winner might have been a better compromise candidate, or, indeed, the opponents of Barron could have had their first choice elected if they had stayed ideologically pure and not tried to hedge their bets. This may seem like an edge case, and maybe in the long run the political culture of a country would change and new parties would emerge which would stop this result from happening, but I still feel that there is an inherent problem in asking voters to say which candidates they "approve" of, without really specifying what that means (or what the tactical implications are). It effectively requires ranking all the candidates in order, and choosing a threshold above which all candidates are "approved", and below which none are, but how that threshold is set is left as an exercise to the voter - an exercise which in practice has a right and wrong answer, which can only be determined after the result has been announced.

jfengel 5y ago
I don't think it would take a case as exotic as that to put voters off voting reform. The odds are that any election is still going to result in one of the two main parties winning. There may be occasional exceptions, just as there occasional upsets in FPTP, but for the most part there isn't any single Third Party that has the quiet approval of a huge range of voters. Rather, there will be many other parties, and the voters' second choices are likely to be either one main party, or blank. Even if they check "everybody but the main party candidates", it's unlikely to produce a non-main-party winner. I have a feeling that after voting reform, the strongest advocates will take note that nothing actually changed. They might declare fraud, or demand more reforms, but I don't see them accepting "Well, turns out that the people who compromise hardest assemble the biggest tent" as the answer. Maybe I'm wrong, and I don't think it would do much harm to try. Sure, some voters will be confused, but they're already confused so that's just a drop in the bucket. But I don't believe it will make anybody happy, either.
- dane-pgp 5y ago
  I admit that most modest (and thus politically achievable) voting reforms would be unlikely to result in immediate wins for small parties, but that sort of stability is probably a desirable property of any electoral system. The main benefit in the short term would likely be that large parties (which are themselves coalitions of various factions, with internal dynamics that play out behind the scenes and usually without popular support) would split into smaller, more representative units. For example, if the Democrats split into a Hillary party and a Bernie party, the Bernie party might never have one of their members become the Speaker of the House, and two parties would formally or informally act as a coalition, introducing and passing pretty much the same bills as the (combined) Democrats do today, at least initially. Rather than advocates complaining that "nothing actually changed", though, I think they'd notice that the supporters of these new separate parties were reinvigorated, enjoying the ability to have a party that truly speaks for them and doesn't have to compromise at the stage of choosing its leader and identity. These new parties would allow the political discourse to change, and shift the Overton window, while also allowing parties to appeal to non-voters who had previously been faced with a choice of two evils, or uncompetitive districts.

dane-pgp 5y ago

> Unlike its competitor, ranked-choice voting, approval voting doesn’t require new machines, ballots or hardware. ... It’s a sensible, intuitive and tested approach that every voting machine in America could execute tomorrow. I understand the desire to be pragmatic and accept the reality of voting machines as they exist in America today, but I also think that any proposal which entrenches a dependency on voting machines is potentially a dead end for democracy. Even if you could guarantee that all the software and hardware used in voting machines was open for security researchers to audit, and there were no bugs or vulnerabilities, you would still be introducing the attack surface of people being able to lie and claim there are vulnerabilities and that the election had been stolen. Once we accept the premise that votes should be cast without using a machine (at least for 99% of voters who are physically capable) and counted by hand, consideration needs to be given to what that counting process looks like. Certainly approval voting is vastly preferable to ranked systems, since approval ballots can be split between different counting stations whose sub-totals can be easily summed, but the actual management of those physical ballots can be tricky, since counters either need to keep track with a tally system or group ballots into one of 2^N different piles (given N different candidates). An alternative system, which has all the benefits of approval voting (same ballot design, solves the spoiler problem, summable sub-totals), is Asset Voting. Basically it works the same as FPTP, except that after the initial count, candidates are eliminated (from least to most votes) and they can choose to redistribute their votes to another remaining candidate. In terms of counting, this can be done by hand using only N piles for N candidates, the way that countries have done it for hundreds of years and many still do to this day.