Minneapolis is trying out ranked voting in their upcoming local elections.
Here is how it works, we’ll use the mayoral race, with 11 candidates as an example:
Each voter gets to vote for a first, second and third preference.
The first choice votes are tallied.
If one candidate gets more then 50% of the votes, then a winner is declared.
If no one has 50%+, then the candidate with the lowest number of votes is dropped out, and the ballots with that candidate as first are recounted using the second choice.
Still now 50%? Drop another candidate, and retally.
Repeat until someone has 50%+
On the plus side, this should make people more willing to vote for non-major party candidates. No more “I’d be wasting my vote” arguments. Also, this should remove the chances of a minority vote win, if you look at all choices as votes.
On the down side, you could see it as some people getting up to 3 chances to vote. Well, I voted for candidate A, but they got very few votes, so my second choice was candidate B, but they didn’t do much better, so my vote ended up counting for candidate C.
So is this a good thing or a bad thing? Will we see third party candidates running as a “vote for me first, and you still can vote for the other guy”. Does this violate one person, one vote?
I don’t think people will really have second choices - it’s tantamount to not believing in your first choice - but I suppose it depends on the breadth of the political spectrum being encompassed in total.
On the first point - This works for me if the voter may limit their vote to only one or two candidates. And that the voter may cast a vote for the same candidate in all 3 positions.
On the second point - By your own description, all (rather than some) people would get up to 3 chances to vote. Because of this it is applied fairly to all voters.
I think it’s better. There’s a theorem out there someplace that explains why there can be no completely fair or thorough voting system, or something like that (no doubt it relies on a reasonable explanation of itself, in stark contrast to what I just wrote).
Imagine the choices for what you want to drink with your pizza: there are 8 different kinds of root beer, and milk. 5/6 of the people would prefer root beer with the pizza, but the vote gets split over all 8 kinds, which nobody can really tell apart. A sixth of the voters prefer milk, which wins, so that by far the least satisfactory of 9 options is the one chosen.
Ranking systems pretty much always fix that problem.
Here’s the wikipediasite about it. It works viably in a number of countries, most notably Ireland, so the Irish dopers might have some more to tell you about it.
It obviously does not violate ‘one person, one vote’, all it does is make the one vote that the voter gets contain more information about what they want. I’m not sure by which mechanism this would lead to a 3rd party per se, but on the level of individual candidates, this system tends to gravitate towards the more moderate candidates that are more broadly acceptable (in terms of game theory:the condorcet winner). Of course, the electorate might be such that a more moderate candidate is not conceivable (say one third says a>b>c, one third says b>c>a and one third says c>a>b). However, examples are conceivable where this system would lead to the more obvious choice.
You may be referring to Arrow’s impossibility theorem, which holds that no decision mechanism that meets a number of ‘democratic’ criteria (such as that there’s no dictator, or that people can hold every conceivable ranking of preferences (Universal Domain)), that can guarantee that individually rational voters will attain a collectively rational outcome. In the example I mentioned above:
1 2 3
A B C
B C A
C A B
you will get cycling: first a majority will vote for A over B, then another majority will vote for B over C, and than a majority will vote for C over A.
Pretty much, sure, but it’s impossible to guarantee (hence, the *impossibility *theorem).
It’s nothing like ‘not believing in your first choice’ - except maybe in a very fundamentalist universe*. No matter how bad you want to have milk with your lunch, you can still also say that you would prefer coffee over tea if there was no milk, without that detracting in the least from your preference for milk.
I’m thinking of situations in which people say that if you are not completely indifferent to any alternative to the ideal, you are not a true believer. Like saying ‘I find Stalinism is preferable to Nazism’ takes away from your genuine preference for democracy.
I can’t imagine there ever being much agonizing over order. People always know who they want most, if they know anything at all about their politicians. You know your guy, and you know the guy you prefer who actually stands a chance of winning. It’s not that hard - particularly if you’re not required to fill in all three slots. (Filling in one name in more than one slot is equivalent to leaving a slot empty, of course.)
For exactly this reason, I don’t know that I like the look of approval voting. If you’re going to vote for more than one guy, you know which one you’d rather have. The rest of the ones you check are just backups. Why not encode and use that information?
The way it leads to third parties is by removing the penalty for third parties in our current system. Consider the “Nader effect”: In 2000, George W. Bush ®, Al Gore (D), and Ralph Nader (Green) ran for the Presidency. A few percent voted for Nader, who never really had much chance, and Gore ended up narrowly losing to Bush. But most of the people who voted Nader preferred Gore to Bush, and if they had all voted for Gore instead, he would have won easily. So because of the way they voted, the Nader voters ended up with what, to them, seems like the worst possible situation.
Now, everyone knows this, so most folks don’t cast their votes for third-party candidates (especially lately, having just had an object lesson), for fear of a situation like this. And of course, since most folks don’t vote for third parties, the third parties remain unelectable. But if we had instant-runoff voting, the Nader voters could have put Nader as their first choice, and Gore as their second. They’d still be putting their top vote on the person they think is best, but when he’s eliminated from the race, their votes still aren’t going to waste. So now there’s no penalty for voting for Nader like there was in 2000, and so all those people who really thought that Nader was better than Gore but voted out of expedience could instead also put their first choice on Nader.
Oh believe me there is agonizing! It comes into play when you’ve ranked the candidates from the 3-4 recognised political parties, and then have to assign an order to the 4-5 parties you’ve never heard/are jokes/one issue parties/etc.
Does this type of voting have any other advantage besides improving the prospects for third party candidates? I don’t see why that alone is an important enough goal to change the way we vote now.
actually I’m still not convinced that it will even in all circumstances do that - but you may want to read my posts about how this systems gravitates towards more moderate candidates.
Just in case anyone gets the idea that this is a Democratic ploy to gain control of the voting system, an instant-runoff vote in 1992 would have probably seen George H. W. Bush in the White House, not Clinton.
What happens if you rank the top four, and leave the rest blank?
From the voter’s POV, it offers a way out of the “wasted vote” problem. E.g., suppose in 2000 you wanted to vote for Pat Buchanan, but you knew that was just going to make Al Gore likelier to win. Frustrating, yes? Under IRV, you could make Buchanan your first choice and Bush your second. (Or substitute Ralph Nader if you’re left-wing.)
From the body politic’s POV, even if IRV does not change who actually wins, it makes the election results a more accurate barometer of public opinion; i.e., we would have a clearer picture (clearer than polls can give us) of who, when it comes right down to it, really prefers Buchanan, Nader, Perot, etc.
It gives the winner a more solid mandate. There would be no “minority presidents” (or whatever office) – under IRV, the winner would always have more than 50% endorsement, even if not first-choice endorsement.
I disagree. Voting systems for electing single candidates, such as the one for US presidents, but also simpler systems such as (in the US situation) for mayors or senators or what have you, or hypothetical systems such as the alternative vote being discussed here, will only ever affect (if anything*) which single candidate may be elected. On the level of single candidate elections, it is not common to speak of party-systems, because they can only ever speak of one party. Only for the aggregate level is it possible and, in fact, common, to speak of party systems. There is no way of telling how the election system that picks the individual candidates affects the composition of the aggregate level parliament (note that in the case of the US president, or the mayor of Minneapolis, there is no aggregate level, meaning that strictly speaking, there is no party system.
Of course, very often people do make generalizations, the most commonly known of which is Duverger’s law which states that in first past the post election systems, there will be a two-party system. Many of the FPTP systems bear this out, and there’s plenty of theoretical arguments to back it up (Anthony Downs’s ‘An Economic Theory of Democracy’ (1957), for instance), but it is not *necessarily * the case, as any Canadian doper will be able to tell you.
*if everybody wants A>B>C, you’d have to try work pretty hard to devise a system that does not pick A - in fact, Kenneth Arrow, whom I mentioned above, excludes any system that does not elect A in this case.
I don’t think that’s the case. We just had a mayoral election in Memphis. I favored one candidate (who won by a landslide, so the scenario from above would not have obtained), but I did have a second and third choice and could easily have chosen that way.
This voting method has been in use in Australia since the 1920s.
There have been plenty of examples of strong, local independent candidates getting elected in Australia as a result of preferential voting. I had one in my own electorate for a numbers of years. He never scored an absolute majority on first preferences, but he gained sufficent second preferences from the major party voters to win ultimately.
An example from the most recent NSW state election in March 2007 was in the inner city seat of Sydney, where the successful independent candidate won on the back of preferences from both left-wing and right-wing voters.
It comes into play when you’ve ranked the candidates from the 3-4 recognised political parties, and then have to assign an order to the 4-5 parties you’ve never heard/are jokes/one issue parties/etc.