With elections coming up in November, I'm once again looking at our current voting system (Plurality, aka first-past-the-post), and wondering what it would take for States to switch to something better. With various third-party candidates declaring their candidacy, I worry whether they are going to steal votes. I wish I could vote for some of them, but they have no chance.

Here are some of the things I look at when considering a better system:

• Majority - If there exists a majority that ranks (or rates) a single candidate higher than all other candidates, does that candidate always win?
• Monotonicity (aka Sincere Voting) - Is it impossible to cause a winning candidate to lose by ranking him higher, or to cause a losing candidate to win by ranking him lower?
• Clone Independence - Is the outcome the same if candidates identical to existing candidates are added?
• Voter Comprehension - How easily would the typical voter understand how to vote? How easily would the typical voter understand how the results were tallied?
• Ease of Replacement - How easily and inexpensively could a State switch to the new system.

and others can be found here.

Questions for debate:

What are your reasons for selecting the poll choice you did? Does it make sense to change?

Which of the above considerations are most important? What others should be included?

What are greatest obstacles to switching?

(Edited to get poll to work)
(Edited again to add these links:)
Plurality
Instant Runoff
Approval
Condorcet

Actually, there are a number of variations in our systems in the US depending on what office we are discussing. Are you wanting to discuss only the election of the US President, or something else/more?

I was in fact referring to US President, yes...sorry for the confusion.

I picked "other".

I have to admit, in theory, I really like the original concept of presidential electors, elected locally and empowered to convene with no preconceived notions and pick a president. Many of our forefathers--James Madison comes to mind--believed popular democracy of one man, one vote on all matters equated to anarchy and they concocted several different ways to elect federal officials, with the most unusual being the electoral college. While it is all but dead today, except in the way if affects the math, I really liked the original concept.

I like the college because it attempts to eliminate long campaigns with nominees pandering to the electorate. I guess the slates of electors could pander, but presumably it would at least be limited to regional issues. It is certainly possible that presidential aspirants could contact electors and lobby, bribe, etc., but if the electors were sequestered like a jury, perhaps it could be avoided. I like the idea simply of picking people of good character and letting them pick a leader in a quiet, deliberative way.

In any event, I would not argue with a straight face that the electoral college as originally conceived would really work, as history proves me wrong. I just like it as a novel, if failed, experiment.

As to other options, I'll just have to say I am content with the current system. It at least provides some barrier to straight one man, one vote, marjority rule, with all the potential for abuse of minority viewpoints that entails, and if worse comes to worst, an elector could still switch his vote to prevent electing a card-carrying Communist, for example, who somehow snuck through the system.

Switching now would require a Const amendment, which would be difficult, and this process could open other Const'l issues to debate, which I would like to avoid.

I like IRV for primaries but don't think I like it for the general election. I don't know WHAT I like for the general election, but I know I don't like the current system because of the unfair advantage it gives rural/less populated areas. I understand the reasoning, but don't think any one person's vote should count more than another's and that's, in essence, how our current system works.

I've been a fan of IRV for a while now...I remember reading the website with the Muppet IRV election in 2000, maybe earlier. I used to think it would be great for a general election, but I have since changed my mind. Sure, Gore probably would have won instead of Bush in 2000 with IRV in place using very conservative estimates of the independent/third party voters, but I guess I see too many potential problems with it for the general election. I think it would be too easy to sabotage the other side with IRV in a general election.

I'm pretty sure we'd be looking at an Edwards/Obama run-off if IRV was in place for primaries. Edwards was certainly the 2nd choice guy for most Obama supporters and was probably the 2nd or 3rd choice for Hillary supporters. Hillary would likely be the LAST choice for a great many people.

A different methodology doesn't necessarily have to require dropping the electoral system. Each state could pick one of the above methods, or stick with Plurality, in conjunction with the electoral college. If that's the case, no amendment would be necessary.

I do agree, however, that the electoral system has its problems. It's been the rare election when Idaho's electoral votes have made any difference one way or the other. True, when combined with Wyoming, Utah, Montana, and other conservative mountain states, they add up to a reasonably significant chunk, but it's quite the disconnect from my vote to the Prez being sworn in.



IRV is certainly an improvement over Plurality in terms of fairness. I can think of two elections where it would have honored the popular vote. Bush, Sr. lost in 1996 to Clinton because Ross Perot took a lot of the conservative vote. Gore vs Bush (2000) is the other one, as you mentioned kmsouthern, in which Ralph Nader took enough votes from Gore.

The two problems I see with it are:
1) occasional need for tactical voting. Suppose you have three candidates, A, B, and C. If 58% prefer B to A, and 58% prefer B to C (assuming separate head-to-head matchups), B can still be the first one out in IRV. It seems wrong to me that the candidate preferred by a majority over any other candidate can still be the ultimate loser. If that seems like it's going to be the case, it would be in some voters best interest to move one of their two favorites to the bottom of the ballot, which defeats the purpose of IRV.
2) More complex to administer, and switch to. You've got to re-design the ballot. You've got to update all your voting machines. And you've got to have some way to tally votes manually, which is hard to do with a ranked ballot. And if that weren't enough, you've got to keep in mind all the people who don't read instructions, and rank their ballot 3-2-1, when what they really intended was 1-2-3.

I liked that muppet IRV election demo, though, that's one of the clearest explanations I've seen of the benefits of IRV.

As far as the other two:
Condorcet - this is by far my favorite, but if you think elections in Florida are complicated with Plurality, wait until you can't even summarize the results except in one big monster table. And you need to have special training to understand the formula for reaching the results. The reason I like it though, is that it is guaranteed to pick the one candidate who a majority prefers over every other candidate, regardless of similar candidates running (clones). The only exception is in the case of an exact tie, which will never happen in a general election.
Approval - This seems to me to be the ideal balance between fairness and feasibility (and the one I voted for). You use the same ballot, it's just multiple choice instead of single answer. In almost all cases, it will pick the Condorcet winner. And even though there is some degree of tactical voting ("should I approve of my #2 or not?"), it still allows you to order your candidates sincerely. To me that's the biggest hindrance to giving 3rd-party candidates a fighting chance. All the Perot supporters need not have feared costing Bush Sr. the election. (and the same for Nader/Gore supporters )

I like the Supplementary Vote sub-type of the Contingent Vote system, which itself is closely related to IRV.

The difference is that only the top two candidates on first preference votes ever get through to the final round (which eliminates the potential problem of the least objectionable candidate who is nobody's first preference getting through), and each voter only ranks a set number of candidates (not all of them). If the leader on first preference votes does not win by a predefined margin, then second preference votes of the losing candidates get counted up and added to the first preference votes until a winner emerges.

In this way, the opinions of people who didn't vote for either of the two leading candidates as their first preference are still counted. *

The London Mayoral elections have just been concluded under this system with two leading candidates (and a number of trailing others) more or less neck and neck all the way. The Conservative, Boris Johnson , got in on second preference votes.

*The London Assembly was elected at the same time, and this is done on a full proportional representation system. IRV systems only seem to be workable when electing single winners; over a whole assembly/council/parliament this type of voting still excludes the views of many people because the candidates they like best never get a sniff of power.

With the Electoral College system in the USA, I don't see how any such system could work. Perhaps States could use PR to elect the College, and the collated colleges nationally could then use IRV? In any case, the Founders appear to have deliberately made anything that allows public opinion to directly influence the selection of a national President quite difficult, which has it's own advantages.

However, with proportional voting systems for election to the Senate and House, I think that the tyranny of the majority they so feared (rightly, if the UK's Parliamentary sovereignty and FPTP constituency system is anything to go by) would be almost neutered.

So in summary, I think that the Presidential system does appear from the outside to need some kind of reform, but that it should not be considered in isolation - the Senate and House systems should be reviewed at the same time to ensure that the overall balance intended in the Constitution is maintained.

Sorry if I cause any headaches here

So here's a scenario that demonstrates the weaknesses in IRV, Supplementary (thanks for the info Julian), and especially Plurality. I'm basing it on four real candidates, but with numbers that, while made up, aren't far-fetched.

Disclaimer: the results of this scenario don't necessarily reflect my preferences.

Suppose in the presidential election, that there wasn't a limit of one candidate per party. Also suppose that Obama, Clinton, McCain, and Romney are all campaigning still. Further suppose that voter preferences, in some State, are thus:

• Liberals:
  • O-C-M-R: 22.5% (favor Obama)
  • C-O-M-R: 19.5% (favor Clinton)
• Moderates:
  • M-O-C-R: 9.0% (left leaning)
  • M-R-O-C: 10.0% (right leaning)
• Conservatives:
  • R-M-O-C: 19.5% (prefer Obama)
  • R-M-C-O: 19.5% (prefer Clinton)

For example, 22.5% of voters would prefer Obama, Clinton, McCain, and Romney, in that order. Assuming everyone voted sincerely, here would be the results under each system:

Condorcet:
Comparing each candidate against every other candidate,

1. McCain wins (58-42% vs Obama, 61-39% vs Romney, 58-42% vs Clinton). Note that these are not minor victories.
2. Obama is second (51-49% vs Romney, and 61-39% vs clinton). Narrow victory over Romney, and clobbered Clinton.
3. Clinton is third (51-49% vs Romney).
4. Romney loses (beats no one head-to-head)

Plurality:

1. Romney: 39.0% - What the? he wins, even though every single candidate would have beaten him one-on-one.
2. Obama: 22.5%
3. Clinton: 19.5%
4. McCain: 19.0% - Dang loser. You were supposed to win this thing.

IRV:
1st round (same results as Plurality):

1. Romney: 39.0%
2. Obama: 22.5%
3. Clinton: 19.5%
4. McCain: 19.0%

• no winner yet. McCain...eliminated

2nd round (McCain's votes re-distributed)

1. Romney: 49.0%
2. Obama: 31.5%
3. Clinton: 19.5%

• no winner yet. Clinton gone.

3rd round (Clinton's votes re-distributed)

• Obama beats Romney 51% - 49%

Supplementary:
1st round (same results as Plurality):

1. Romney: 39.0%
2. Obama: 22.5%
3. Clinton: 19.5%
4. McCain: 19.0%

• no winner yet. McCain and Clinton eliminated, Romney and Obama advance

2nd round (McCain's and Clinton's votes re-distributed)

• Obama beats Romney 51% - 49%

Approval:
This one is hard to set up a scenario for, but let's assume:

• Liberals approve Clinton and Obama. Half also approve McCain.
• Moderates approve McCain. 1/3 also approve Romney, another 1/3 approve Obama and Clinton (slightly more to Obama)
• Conservatives approve Romney. Half also approve McCain.

Approval Ratings:

• McCain: 59.5%
• Obama: 48.4%
• Clinton: 48.3%
• Romney: 45.3%

So while, IRV and Supplementary are light-years ahead of Plurality, neither one gives as good a results as Condorcet. Approval is the only one that does (in this situation, anyway, but I haven't found a reasonable one yet that doesn't), and it doesn't have the computational baggage that Condorcet does. I would, however, be interested in seeing a case that does vary from Condorcet.


this is probably true.