STV when talking about voting systems stands for Single Transferable Vote.

Is it even possible to fairly elect a president via STV?

Representatives, yes, but a president?

Here is what I think would happen in a country with United States and a president where voting is via STV:

First off, there would be elections in every city of a county where people above a certain age rank the candidates when they vote.

For counting the votes there would be a threshold of 33% and votes would be counted the way it is in STV until there are 3 city electors left.

The process is repeated with the city electors when it comes to county elections resulting in 3 county electors.

Same with the county electors in the state elections and there being 3 state electors.

Then the process is similar with the states but every state votes for 2 candidates so the threshold is 50%.

Then the state electors' votes are counted and the one with the majority of votes becomes president and the one with the second largest percentage becomes vice president.

enter image description here

But would this be fair to the citizens(having a majority of the citizens prefer the president elected over the other presidential candidates and same for the vice president)? And if not, what is the best way to fairly elect a president and vice president because I have watched CGP Grey's videos on voting systems and I realized that STV has the most advantages without abolishing ranges.

  • 3
    STV with only one winner is just IRV, which is not a good system. politics.stackexchange.com/a/14850/10373
    – endolith
    Jan 26, 2018 at 5:14
  • I would have to ask... what would this accomplish? It sounds like a way to implement 'lesser of two evils'. A better solution would be not to promote seriously flawed candidates.
    – tj1000
    Jan 26, 2018 at 14:57
  • But first past the post aka FPTP, already implements the lesser of 2 evils. But that is not its only flaw. Vulnerability to gerrymandering, the spoiler effect, non-proportionality, minority rule etc. are all flaws of FPTP. IRV doesn't have the spoiler effect(in other words with a third party, the votes would go to the party they most agree with because votes are ranked unlike how with FPTP a third party guarantees a win for the party they disagree with). And by having the votes ranked on ballots, there would be way more people agreeing with the presidential candidate under STV than under FPTP.
    – Caters
    Jan 27, 2018 at 1:28
  • 1
    @Caters IRV does have the spoiler effect, it just mitigates it for fringe candidates who didn't have a chance of winning anyway, but if there are 3 or more competitive candidates, it still results in spoilers and chooses unrepresentative winners. (Also proportionality and gerrymandering have nothing to do with the president. You can't elect 2/3 of a president.)
    – endolith
    Jan 29, 2018 at 15:27

1 Answer 1



Single Transferable Vote (STV) is a multiple-winner system. It works in multi-member legislative races because more than one is elected at a time. A version is actually used in Australia now. It is fundamentally incompatible with a race with a single winner, like the United States presidential race.

What you actually seem to be recommending is that the electors be selected by STV. But electors are pretty irrelevant. In most states, they could be replaced by a ballot. Further, in a race with more than two candidates, your electors would not necessarily be able to reach a decision. It's possible that no candidate would reach a majority. One of the advantages of STV is how it splits the vote among more parties, but that would be problematic with electors who are supposed to produce a majority result.

The big change here would not be the STV, but rewarding electors within a state proportionally. States with competitive elections would have little influence on the election process. The electors would cancel out.

You've also added more indirection to the system. Now it's not the voters voting for candidates and being grouped by the statewide winner, but voters voting for municipal electors who vote for county electors who vote for state electors who vote for candidates.

The single race variant of the standard STV is called Instant Runoff Voting (IRV) or the Alternative Vote (AV). However, there are a number of similar methods preferred by many, e.g. the Condorcet-compliant methods, where a person who wins all the pairwise matchups is guaranteed to win. IRV does not guarantee that if someone is preferred to everyone else that that person wins.

The added complexity of your proposal doesn't seem to add anything to what IRV or a Condorcet-compliant method could do.

Vice President Loser

The US tried having the loser become the Vice President. It didn't really work. The problem is that you have these two people who disagree with each other. A lot. Then you make it so that they are expected to work together.

Consider Donald Trump as VP to Hillary Clinton. So we would lose any benefit from his executive experience. We would have to put up with his political inexperience, which would make it hard for him to be effective in the Senate. And of course, he would still tweet inanely. Only now he wouldn't face any real political cost to doing so. After all, he's only the VP. Worse, they would have to keep him informed in case of tragedy.

Also consider what this does to the chances of assassination. Assassinating Hillary Clinton with Tim Kaine as VP or assassinating Trump with Mike Pence as VP means that a partisan gets little or nothing. But if an assassination switched parties, then it would make assassination more valuable. It was bad enough in 1865 when Andrew Johnson replaced Abraham Lincoln. Think about what happens if every assassination were like that. President Nixon in 1963. President Dewey in 1945. President Cox in 1923. President Bryan in 1901. And that's just based on the assassinations that occurred under the existing rules.


If I were designing a voting system to replace the current US presidential election, I would have a range-ranked ballot in a non-partisan primary. I would then score the ballots using Range Voting, Ranked Pairs, and IRV. If a candidate won all three or two of the three, that candidate would make it to the second round. If no more than two candidates won at least one, then both those candidates would go to the second round. If three candidates each won one, then we'd go to more complex tiebreakers.

The second round (or general election) would thus have only two candidates. One or the other would have to be the majority winner.

Example range-ranked ballot, using 2016 candidates:

Score Candidate
100 Bernie Sanders
95 Jill Stein
70 Hillary Clinton
50 Donald Trump
45 John Kasich
40 Gary Johnson
10 Marco Rubio
1 Ted Cruz

For IRV and Ranked Pairs, this would be scored only ordinally. So the difference between Bernie Sanders and Jill Stein would be the same as the difference between Gary Johnson and Marco Rubio. Both would be one place ahead.

It would be easy to set the tiebreakers such that the Condorcet winner is guaranteed to be in the final two. Because the three methods have different strengths and weaknesses, this gives us the best of all worlds. We are guaranteed to have two good choices.

Comptroller General

One way to include those who do not get their vote for president would be to allow them to vote for Comptroller General of the United States. So if Hillary Clinton does not win, then that voter would get a vote for comptroller general. Meanwhile people who voted for Donald Trump would not get their vote counted for comptroller general.

The comptroller general would then have full access to the financial, legal, and classified data needed to audit the nation's budget, justice, and intelligence systems.

This would work more like STV, where votes that aren't initially satisfied roll over to other candidates. However, the comptroller general would not be the equivalent of president. More of a consolation prize.

  • Are you aware of Score-runoff voting? It's pretty similar to your proposal of range + top-2 runoff.
    – endolith
    Jan 29, 2018 at 15:46

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .