Against ranked-ballot systems: But why?

No, I am not against them, but Joseph Heath, writing at In Due Course, is a skeptic.

Heath’s main concern with “instant runoff” (IRV), also know as the alternative vote (AV), is that it does not guarantee a Condorcet winner.* This is true, and well known. It is one of several methods that will guard against a Condorcet loser, however. Of course, if you want to guarantee a Condorcet winner when there are three or more candidates, you might still use ranked-choice ballots (but a different counting rule), but that’s not what motivates me to write this response.

The motivation is the following passage, which comes after a stylized illustration based on the recent three-way race for Toronto mayor, after which Heath suggests that under his hypothetical distribution of preferences among the voters, Ford supporters would have been better off strategically giving first preferences to a different candidate. He then says:

So it is absolutely and categorically false to say that IRV eliminates the incentive for strategic voting. All it does is invert it. (This is something that everyone should know from history as well – in 2002 too many French voters failed to vote strategically in the Presidential election — which uses a run-off system — leading the Socialist Party candidate Lionel Jospin to be eliminated in the first round, forcing them all to vote for Jacques Chirac to keep the far right out of power.)

That’s some sleight-of-terminology there. Yes, France elects its president with “a runoff system”, but not with an instant runoff. It is a top-two runoff on a later date, and indeed, coordination failure on the left in 2002 prevented Jospin from beating LePen for the second slot. However, can anyone seriously doubt that with IRV/AV, Jospin would have been one of the last two standing when things reached the final count? I don’t think so.

I do not know who the Condorcet winner was in France in 2002. It might have been Jospin, and had approximately the same distribution of first preferences been rendered under an IRV system, then Jospin would have won–if he rather than Chirac was the Condorcet choice.

I agree with what I take to be Heath’s broader point that proponents of IRV oversell it at times. However, if one prefers a system that enhances the electability of a Condorcet winner, it is a pretty good choice. If one wants to guarantee a Condorcet winner, well, that is a different conversation, and one Heath does not enter in to. Most of the rules that would do so have their own pathologies. For single-winner contests, I’m still sticking with AV/IRV as being a good enough solution.

It is also possible that Heath would be happier with the Coomb’s Rule variant of AV.


* A Condorcet winner is a candidate who would beat each of the others in a pairwise competition.

20 thoughts on “Against ranked-ballot systems: But why?

    • It’s an AV fiesta! I am a little saddened, though, by the lack of replies to my own post…

      MSS, I think you might want to rename this post “Against AV: But why”, since Heath’s objection seems to be against AV specifically rather than ranked systems in general. In fact, as you point out, by his reasoning AV is certainly an improvement over plurality or two-round, which are the systems used for almost all single-winner contests in practice.

  1. I think a preferential ballot is one of the best ballots out there. Consumers do this every day when they go shopping, for example! Extremely interesting if a mixed system would use a preferential ballot for the district tier and a preferential ballot for the at-large tier, although there are many that think STV would be the most preferred choice. Think of it, a preferential ballot and no threshold (the voters would decide to make or break a party)

  2. 1. I’m missing something here. Is not Condorcet a “ranked ballot” system too? If Mr Heath is one of those Condorcetistas who prefer Approval because it will supposedly elect the Condorcet beats-all winner because people will vote sincerely (by ticking all, and only, their… uh.. pick a number?… highest-preferred candidates), then I fear he’s stuck on a desert island assuming a can-opener.

    2. Debates with zealous Condorcetistas tend to fall into the following, err, cycle…

    “AV is bad because it can elect White and defeat Green even though a majority of voters prefer Green over White, 52% – 48%!”

    “Very well. Let’s use Condorcet instead, then, and elect the candidate who defeats all others pairwise… but wait. Uh-oh, looks like a cycle exists. Green beats White pairwise, 52-48, and White beats Brown, 54-46… but Brown beats Green, 55-45. No candidate beats all others pairwise. What do we do now? Leave the Mayoralty vacant for four years?”

    “No, no, no need to panic. Just find the candidate who would need the fewest extra first-preference votes to beat everyone else one-on-one. In this case, White is the only candidate whose pairwise support never drops below 48% against another candidate. Therefore, we’ll elect White by default.”

    “Wait… so we end up electing White, even though a majority of voters prefer Green over White, 52 – 48%?”

    Relative too AV, Condorcet seems to involve (a) a lot of extra complication (intellectual and practical), and (b) the risk of electing a candidate with very low first-preference support – yet it still doesn’t rule out the same defect that its advocates think is fatal to AV.

    Refined versions of Condorcet include modifications like the Smith filter (eliminate every candidate who’s beaten pairwise by all other candidates, then any candidates who could only beat already-eliminated candidates, so that only the cycled candidates remain, before deciding which has the highest benchmark of minimum pairwise support) that make it less like White will be elected even though a majority of voters prefer Green over White pairwise.

    For multi-choice referenda (with two or more competing, mutually-conflicting motions), it would be better first to eliminate any that are supported by fewer than 50% of the valid ballots (and some quorum, say 33%, of all persons eligible to vote) against the “NO CHANGE” option, and then use AV if two or more remain after this first cull. But I could live with Condorcet for this second stage – in referenda. However, I think that, if ever adopted for elections of living bodies, Condorcet would end up being repealed as discredited within a few years.

  3. I just realised AV might offer a better way of electing a vice president than having a running mate or electing the VP separately. Running mates are often chosen strategically to represent another branch of the party or coalition, or to make the ticket more diverse descriptively, rather than necessarily for being a good replacement (in terms of capability or ideology) in case the president dies or resigns. When the VP is elected separately, candidates have to choose one race or the other. Under AV, once a presidential candidate has more than the requisite 50% of the vote, the whole count can be restarted, while excluding the candidate just elected president. The resulting second candidate elected could in a sense be said to be the voters’ real 2nd choice (this is what the US constitution originally tried to do by making the second-place candidate VP).

    • JD, I believe that is the system used to elect the mayor and deputy mayor in at least one and possibly all three of NSW, Victoria and Tasmania, alongside a PR-STV elected council. However, if I recall NSW ballot format correctly, the candidates for the two mayoral positions can also have themselves grouped in ranked order on the ballot (like the quasi-party-list STV systems used for the mainland upper houses), so it is usually foreseeable that the higher candidate of the two will get mayor and the person in the lower slot will get Deputy Mayor. Nonetheless, the possibility always exists for voters to do a “Bush/ Bentsen” as was occasionally mooted in 1988.

      There’s a case for an option, along the lines suggested by either Cass Sunstein or Akhil R Amar in relation to the US Vice-Presidency, that voters get an option of voting for “mayor’s running mate” for Deputy Mayor, if it’s very important to you to not have a divided executive. (Perhaps less important in local government where parties tend to be looser, ad hoc groups – “Residents’ Rally,” “Progress for Boggabilla,” whatever – divided over non-ideological issues such as where to put the bridge).

      A simpler version would be that any valid ballot that showed a voter for Mayor/ Governor/ President, but that is blank (or exhausts) for Deputy Mayor/ Lieutenant Governor/ Vice-President, is deemed credited to the candidate nominated as running-mate by the Mayor/ Governor/ President. If there is none, that ballot stays exhausted. So voters could elect Bush I as president and Bentsen as Veep, but the latter would need 50.01% to beat Quayle.

    • I wonder how long it would have taken the party of Vice-President Bush to impeach President Clinton in 1993? Are we talking weeks or days?

      Or the party of Vice-President McCain to impeach President Obama in 2009?

      Or the party of Vice-President Temer to impeach President Roussef in 2015?

      Guaranteed cross-party vice-presidencies are a terrible temptation to the losing party.

      • I see Alan’s point – although success depends on having strong party discipline, not just reaching 50% + 1 but a two-thirds majority.
        That said, having a VP who is in effect hand-picked by the President means that if the President is removed for perceived serious corruption or wrongdoing, there will be a perception that the Veep is somehow implicated. The USA lucked out in 1974 because Agnew had already gone and Ford had been chosen precisely because he was viewed as a cleanskin. But imagine if he (rather than George McGovern) had still been Vice-President when Nixon resigned.
        One or another law prof (Ackerman, perhaps, or Balkin) suggested that for this very reason the Veep should not attend Cabinet meetings or take part in executive policy-making. However that may mean a very steep learning curve if the VP succeeds, not because the President has been impeached and deposed for high crimes, but because the President has died in office. (“Manhattan Project? What the hell you talking about?” – Truman, 1945)

      • A hands off vice-presidency works if and only if a president who is otherwise corrupt and criminal is suddenly moved by some shred of respect for the law to honour the Chinese wall separating them from the vice-president. The better solution is not to have a vice-president.

      • > “The better solution is not to have a Vice-President”

        Alan, that is heresy on this blog. I recollect our host and his co-author proposing in their 1992 book that Presidential democracies allow not just two names on the Presidential ticket but three, to make cross-party coalition-building easier. And designate them as “co-Presidents” rather than as Senior Vice-President, or Second Vice-President, etc.

        Mind you, I think they were also advocating “closed-list” Presidential tickets, not a separate vote for each slot, which also rules out my proposal.

    • There are feasible distributions of votes where the second scrutiny would produce an opposition vice-president and feasible distributions of votes where it would not. There would be a strong tendency to nominate vice-presidents only from the largest states to try and get them elected over the opposition presidential candidate.

      Parties would face exactly the same problem as under the original Jefferson/Burr rules where they would need to try and assign votes between their 2 candidates for maximum effect. AV, which claims to end strategic voting would now have stategic voting on steroids.

  4. I’ve been crafting a proposal that would allow for voters to give a first preference to a candidate and also give them all the runoffs possible. This could help meet the Condorcet winner criteria but will voters be delighted if they have to see who’s better between a Democrat and a Republican in a runoff or between a Libertarian and a Green?

  5. It is true that STV-AV can REWARD tactical voting – as can every other voting system ever devised.

    But it’s highly debatable to claim that STV-AV can ENCOURAGE our OFFER AN INCENTIVE for tactical voting.

    An incentive or encouragement is a function not just of the existence, and attractiveness, of a reward, but also of the difficulty of achievability of attaining it.

    Robbing Fort Knox of its gold would be a lucrative reward if you could carry it if successfully – without being stopped, shot, or arrested. But no one would say that any incentive or encouragement exists to try robbing Fort Knox. On the contrary, there are quite substantial deterrents to making any such attempt.

    Likewise with tactical voting under STV-AV. In certain very limited circumstances, it’s possible to identify situations where you might have gotten a better result from your point of view – ie, elected more of your N highest preferences, or obtained a result higher in lexical order (ie, earlier in alphabetical order if you assign your first-choice candidate the letter A, you second-preference candidate the letter B, your third-choice candidate the letter C, and so forth), had you voted other than sincerely.

    But this is a very long way from saying that this is a rational strategy to pursue (and to exhort your voters to pursue), for two reasons:

    1. There’s only a very narrow band where that will help.

    2. Correlated to 1, it becomes self-correcting. Yes, in hindsight, directing between 50 and 65 of your supporters to switch their first preferences to another candidate, XYZ, might have helped you, because you would fare better facing XYZ in the final tour decisif than facing ABC, who would otherwise have run second. BUT with a secret ballot and mass universal suffrage it is very difficult to ensure that exactly the desired number switch their votes as desired. Too few, and it doesn’t make a difference. Too many, and you might lose the final runoff (your own supporters’ votes helping to elect XYZ over you), or even drop to third place and miss the runoff.

    By contrast, under almost non-transferable systems (other, perhaps, than List-PR with a district magnitude in three figures, Hare or Sainte-Laguë allocation and no supra-quota threshold, or any two of these three), “your first-choice party or candidate is polling too low to be viable. Switch your support to your least-hated choice among the larger, viable parties/ candidates” is a winning strategy that keeps on winning. It’s not a loop, it’s a centrifuge. It perpetuates the exclusion of smaller parties by making the job of aggregating their support a coordination problem for millions of isolated individuals, voting separately by secret ballot.

    • Tom

      Your analysis is true for electing legislators with the same responsibilities. That is not the same as electing a chief executive and a non-executive with the duty of regularly inquiring after the chief executive’s health.

      A two-Round Alternative Burr vice-presidency would be vulnerable not only to tactical voting by the parties, but to cross-party unannounced tactical voting of the kind that almost elected Aaron Burr. The truly ghastly thing about AB voting is that it not only allows side-deals, it encourages them. I am uncomfortable with electoral proposals that reward and encourage treachery against the electorate as well as the party.

      • Just for greater clarity, if Tom and I ran for election as MPs for the two-seat district of Upper Middle Bogan it would make very little difference which of us become the first and which the second Member for Upper Middle Bogan. If we were running in an executive election we could well care slightly more which of us became President and which Vice-President of the Republic of Boganland.

      • Alan, apologies if my comment [@13/01/2016 at 3:26 PM] looked as if it was aimed at your statement; it was actually aimed at Mr Heath’s original article (I began writing it some days ago and saved it as a draft), but it ended up positioned directly under your comment…

        I think your argument in favour of a single closed, ranked ticket to elect a President/Vice-president or Governor/Lieutenant Governor is strong and probably persuasive. I’d prefer a parliamentary executive for every sovereign state (national or subnational) so the details of electing the top two (or, for devout Shugartis, three) is a second-order issue. But for local and regional governments within a state, I’d advocate direct election of the mayoralty – partly (1) because parties are too weak (or non-existent) in many small cities and shires to sustain a responsible Westminster-style parliamentary executive, also (2) because local councils are usually unicameral so an elected mayoralty (with one or two slots) gives voters a second “estate” to vote for.

        (And yes it is possible for a unicameral, largely non-partisan body to be too disciplined to sustain an executive yet disciplined enough to have legislative spasms that need restraining. Supporting an executive requires positive agreement on budgets and Cabinet portfolios. Passing ill-advised legislation merely requires you to persuade the extreme left and the extreme right that the centrists are conspiring to do them over…)

Leave a Reply

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

You are commenting using your 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