This situation reminds me of the Ultumatum game.
Say there are three candidates: Alice, Bob and Chandrakant. Everybody knows Chandrakant doesn't have a chance. Between Alice and Bob, you prefer Alice, but would like her to support the ban on transgenic oysters. So, along with a small vocal minority, you threatedn to vote for Chandrakant if she doesn't support the ban.
This plays out like this:
Round 1, Alice decides whether or not to add the ban on transgenic oysters to her platform.
Round 2, you decide whether to vote for her or Chandrakant.
Alice's decision in round 1 depends on how seriously she takes your threat to vote for Chandrakant in round 2. If you're the kind of person who is known for "not wanting to throw his vote away", then you'll end up still voting for her and eating transgenic oysters. If you (and people like you) are known for being stubborn as a mule, she may change her platform.
(note that none of this requires that you actually prefer Chandrakant to Alice - voting for a third-party guy can also be strategic voting!)
Wouldn't a better threat be to switch to Bob anyway? (in which case Alice effectively loses 2 votes instead of one)
In many cases, I would prefer an outcome where the wrong lizard wins, but my preferred 3p candidate gets a significant share of the vote. In this situation, I would vote 3p even if my vote mattered, and especially if I thought (by reflexive theory) that many other people were similar enough for me to also do so.
Assumption: Vote, with one winner.
-In one most extreme case, you write-in the option you most wish (even if, as far as you know, no one else has ever heard of it).
-In the other most extreme case, you vote where you most expect to create a swing.
-In the iterated case, your vote is also a signalling mechanism which influences the next vote. Does that change things?
-Is the answer necessarily one of the two extremes? Does it make sense to have a "middle ground" option?
-I think the extreme case "Vote where you most expect to create a swing in your favor" strategy is at least reasonable if not ideal. Can anyone give a reason why the other extreme, "write in your favorite option even if no one has heard of it", is not completely impractical?
My own current thought is that you should use the extremist "wherever the swing is" vote pattern in the non-iterated case/ For the iterated case, a "middle ground" option for signalling purposes, since your vote is unlikely to change the outcome anyway in most cases but guaranteed to generate signal.
TL:DR: In the iterated case, the ideal voting strategy is situation-dependent. When one clear winner exists, voting carries more importance as a signal than it does as an actual vote...so if you support a strong 3rd party, you should signal support for them. You probably shouldn't bother voting for your favorite-but-completely-obscure 23rd party unless you think your vote has a chance of pulling them out of obscurity. However, if you estimate the vote is clearly very close to swing, then you should switch strategies and sacrifice the chance to signal in exchange for the chance to swing the vote.
You might want a brief paragraph to explain to non-Americans (there are a lot of those here) the context. In Europe, for example, two-party systems are rare.
you see that your choice on voting is probably correlated with that of many other people.
I've heard this asserted on a couple of occasions (most memorably in Three Worlds Collide) and have a question about it. The case as I understand it is that my choices are correlated with others who have similar minds, so I should choose as if choosing for everyone. I don't really object to that, but it seems to me that the set of "similar minds" is quite small. Is my appropriate reference class the set of people who notice and act upon this theoretical correlation? Isn't the number of people who do so miniscule, to the point that even if the theory is correct, it still won't have a noticeable impact?
The number of people who have heard of UDT or Hoftstadter's super-rationality is miniscule, but informal versions, like "What if everyone failed to vote?" are common. Whether they actually drive action is a difficult question.
Many of the effects here depend on the voting system being used. Ka-Ping Yee has a great simulation of voting systems here.
Another consideration is the effects of your decision criteria on the lesser evil itself. All else being equal, and assuming your politics aren't so unbelievably unimaginative that you see yourself somewhere between the two mainstream alternatives, you should prefer the lesser evil to be more beholden to its base. The logic of this should be most evident in parliamentary systems, where third party voters can explicitly coordinate and sometimes back and sometimes withdraw support from their nearest mainstream parties, depending on policy concessions.
assuming your politics aren't so unbelievably unimaginative that you see yourself somewhere between the two mainstream alternatives, you should prefer the lesser evil to be more beholden to its base
The assumption you appear to be making here is that you're either between the two mainstream parties (and "unbelievably unimaginative") or further out than one of them along the axis running between them, somewhere near to where "their base" lives (and presumably not so unimaginative).
I think this assumption is very wrong, and I don't see any reason for accepting it. It's wrong for at least two reasons. (1) Politics is not one-dimensional. You might be fairly near the middle on that primary axis but far from the two mainstream options on some other axis. (2) You may be further out than "between the two mainstream alternatives" but the nearest applicable "base" may be further out and highly undesirable. (For instance, you might consider that the centre ground between Republicans and Democrats is too far left while thinking the Tea Party is too far right -- or, for that matter, too far out on some other axis.)
Politics is not one-dimensional. You might be fairly near the middle on that primary axis but far from the two mainstream options on some other axis.
Statistically speaking, politics is largely one-dimensional, and there is no plausible case to be made for anything beyond two dimensions. The single axis covers >80%, the two-axis graph covers 95%+ of voting behavior, and the rest is more likely to be overfitting than any actual additional axis.
Two is enough to invalidate the assumption I think Oligopsony made, but in any case what's true "statistically speaking" is relevant only to those who are statistically fairly typical. LW is full of people who are, in various ways, unusual, and there is some reason to think that quite a few LW participants have unconventional political opinions. (And Oligopsony was specifically addressing the less-conventional, I think; hence "assuming your politics aren't so unbelievably unimaginative ...".)
you should prefer the lesser evil to be more beholden to its base
How would you go about achieving this? The only interpretation that occurs to me is to minimise the number of votes for the less-dispreferred main party subject to the constraint that it wins, thereby making it maximally indebted to (which seems an unlikely way for politicians to think) and maximally (apparently) dependent upon its strongest supporters.
To provide a concrete example, this seems to suggest that a person who favours the Republicans over the Democrats and expects the Republicans to do well in the midterms should vote for a Libertarian, thereby making the Republicans more dependent on the Tea Party. This is counterintuitive, to say the least.
I disagree with the initial claim. While moving away from centre for an electoral term might lead to short-term gains (e.g. passing something that is mainly favoured by more extreme voters), it might also lead to short-term losses (by causing stalemate and gridlock). In the longer term, taking a wingward stance seems likely to polarise views of the party, strengthening support from diehards but weakening appeal to centrists.
To provide a concrete example, this seems to suggest that a person who favours the Republicans over the Democrats and expects the Republicans to do well in the midterms should vote for a Libertarian, thereby making the Republicans more dependent on the Tea Party. This is counterintuitive, to say the least.
Is it? Again, I haven't done the math, but look at the behavior of minor parties in parliamentary systems. They typically demand a price for their support. If the Republican will get your vote regardless why should they care about you?
I agree that voting for a third party which better represents your ideals can make the closer main party move in that direction. The problem is that this strategy makes the main party more dependent upon its other supporters, which can lead to identity politics and legislative gridlock. If there were no Libertarian party, for example, libertarian candidates would have stood as Republicans, thereby shifting internal debate towards libertarianism.
Another effect of voting for a third party is that it affects the electoral strategy of politically distant main parties. If a main party is beaten by a large enough margin it is likely to try to reinvent itself, or at least to replace key figures. If a large third party takes a share of the votes, especially of those disillusioned with main parties, it may have significant effects on long-term strategies.
The problem is that this strategy makes the main party more dependent upon its other supporters, which can lead to identity politics and legislative gridlock.
Legislative gridlock is better than bad laws being passed.
Really? By whose definition of "bad laws"? There are an awful lot of laws that I don't like (for exaple, ones mandating death for homosexual sex) but that doesn't mean I'd like to screw up the governance of an entire country by not allowing any bills whatsoever to pass until a reform bill passed. That's a pretty good way to get a civil war. Look, for example, at Thailand, which is close to separating into two states because the parties are so opposed. Add two years of legislative gridlock and they'd hate each other even more; I am reasonably confident that gridlock in Thailand would lead to mass civil unrest and a potential secession of the northeast, which might well be violent.
but that doesn't mean I'd like to screw up the governance of an entire country by not allowing any bills whatsoever to pass until a reform bill passed.
I disagree with the premise that countries need a constant supply of new laws to function.
That's a pretty good way to get a civil war. Look, for example, at Thailand, which is close to separating into two states because the parties are so opposed. Add two years of legislative gridlock and they'd hate each other even more; I am reasonably confident that gridlock in Thailand would lead to mass civil unrest and a potential secession of the northeast, which might well be violent.
The problem in Thailand is not gridlock itself, it's that different factions have very different ideas about what the laws should be.
I disagree with the premise that countries need a constant supply of new laws to function.
In many countries, the budget or appropriations bills must be passed as a new law every year for things to keep moving. In the US, not having appropriations for even a month was considered unpleasant, and it was getting increasingly so as the effects percolated through the various layers of buffering between the US Treasury and actual paychecks.
Traditionally, in a parliamentary system, if the budget fails, that's a vote of no confidence and requires new elections -- things can't just keep going.
Abstracting, it seems like constitutions are deliberately designed so that the government can't go on autopilot -- if it were possible for the country to run without new legislation for years, that would weaken the legislature compared to the executive, and it's widely believed that unchecked executive power is dangerous to liberty.
Con 1) Voting for a third party is throwing your vote away.
We have all heard this argument before, and it is true. [...] If you vote for a third party, the same candidate would be elected as if you did not vote at all.
Note that this argument also goes through for voting for a second party. Or voting for any party that doesn't tie for first place or win by exactly one vote. So this is less an argument against voting for third parties specifically and more of an argument against voting at all.
Con 2) Actually if you use a reflexive decision theory, you are much more likely to change the election, so you should vote like it matters.
Looking at the problem like a timeless decision agent, you see that your choice on voting is probably correlated with that of many other people. You voting for a primary party is logically linked with other people voting for a primary party, and those people whose votes are logically linked with yours are more likely to agree with you politically.
That looks like magical thinking. In practice, if I silently decide to vote for party Y instead of party X, that has no effect on anyone else's vote (at least at that election).
Does this mean it's a clean sweep for the four Pros mentioned in the post? I wouldn't say so, because I disagree with at least two of them. Still, I agree with Eliezer on the basic answer to your original question: if you think a third party is the most meritorious party, vote for it.
If you always vote for one of the two most popular parties instead, you're at risk of experiencing a ratchet effect. Suppose you prefer the unpopular party X to the popular party Y, and the popular party Y to the popular party Z. For fear of handing an advantage to Z, you vote for Y instead of X. Now suppose Y & Z both shift a little away from where you stand politically. You end up voting for Y again because it's still a bit closer to you than Z. Now suppose Y & Z both shift a little further away again. You still vote for Y because it's still a little closer than Z. This can repeat indefinitely, and if so you wind up voting for Y even as it becomes arbitrarily awful to you.
I think the probability of changing the election is generally underestimated.
The probability distribution of votes approximately follows a bell curve. If you're ignorant of the election, it peaks at 50%. If there's a uniform distribution, the expected value of your vote is also the average effect on one citizen if you do make a difference, multiplied by a constant to take into account that not everyone votes. If it peaks at 50%, then that's a higher probability density than uniform distribution, so you'll make a bigger difference.
If you're less lazy, and you look into the expected results, it will get more refined. It might still peak at 50%, but with a much narrower peak, in which case your vote matters orders of magnitude more than before. The much more likely result is that it will peak somewhere else, and have a sharp decline. In this case, your vote matters orders of magnitude less.
As a result, it looks like it's not worth voting in any given election, but unless you've actually done the research, it's worth voting.
I'm pretty sure I don't live in a battleground state, which means that I already know I have an extra low probability of making a difference. It still matters a little, since the politicians know they have to cater to the voters to keep winning. In this case, I figure I might as well vote for a third party, since that will make my position more clear. I can only give one bit of feedback if I vote for a major candidate, but I can give two or three by looking at the third parties.
I think the probability of changing the election is generally underestimated.
But the analysis that follows seems to suggest a different conclusion: (1) If you are ignorant, then your subjective probability of changing the outcome is more than people often think, (2) If you aren't ignorant, then your subjective probability of changing the outcome is very small (but, I note, it is probably comparable to the probabilities estimated by Gelman in the study linked here), and (3) the high-probability case is a bit like that where someone goes into a casino thinking he has a winning system and (correctly, given that belief) estimates a substantial probability of coming out richer.
It seems to me that the "informed" probability is the relevant one here. I'm not sure how to tell whether this is generally overestimated or underestimated; among people who have followed Coscott's link to Gelman's paper, I'd guess it's estimated reasonably well unless Gelman's analysis is bad.
If you are uninformed, the informed probability is probably low, but it might be high. By conservation of expected evidence, the expected value is always equal to the uninformed probability.
Margins matter. Looking at the paper there, they used tons of margins data to do their calculations. They talked about how groups should direct their efforts in shifting thousands of votes, based on margins... and somehow missed that that meant that margins matter.
Now, that doesn't apply as directly to third party vs wrong lizard, it may seem. But just keeping the voting rolls up (especially but not just in your area and demographic) matters, as it impacts where attention is directed. You're saying, "I vote and neither of you is doing a good job and here's what you can do" and the bigger that number is the more of a prize they see. Every tenth of a percentage point makes you more and more attractive. Making the less terrible lizard better is a good thing.
So, there isn't just one threshold that matters a lot, but there are also many thresholds that matter less but still can count enough to add up.
Plus, there's normally more than one election per ballot. So that multiplies the impact of whatever decision you make. And your leverage is greatly increased in many of these elections. Last election I may not have been able to impact national matters, but two municipal elections came down to single votes. Either candidate in that case could have kicked themselves for not courting that one other voter.
Pro 5) The more people vote for a third party, the more likely it is that one of the main parties will adopt some of their policies in an attempt to get those votes.
I think that many people if given the choice between changing the next president between the two primary parties or being paid an amount of money equal to the product of the amount of gas they spent to drive to vote and 10^7 would take the money.
Just pointing out that this is a lot of money. Using the first value I could find for the mean road distance to a polling place and a 20mpg car, it comes to just under 1.5 million (1). I don't think that it's much of a conclusion to say "most people would prefer 1.5 million dollars to deciding whether Candidate 1 or Candidate 2 wins". I suspect many here could argue that they could do more with the money than their choice would in expectation (depending on the candidates).
(1) Mean road distance to assigned polling place in Atlanta, Georgia divided by 20 miles per gallon, plugged into WolframAlpha for price, * 10^7 =1.416 million US dollars
On the other hand, it's being weighed against a big effect. If you believe that your disfavoured candidate has an extra 10% chance of getting the US into a futile $10B war (note: this is at least two orders of magnitude less than Iraq, even if we consider nothing but financial costs) and consider money spent on war to be wasted (rather than e.g. as a form of redistribution, with impoverished defence contractors as beneficiaries) then preferring $1M for oneself at that cost amounts to a 1000:1 preference for self over others in the same country. Far from implausible, alas, but it's still quite extreme.
(Note 1: The above calculation is really simplistic. I know. I don't think a less simplistic version would give drastically different results. Note 2: I am not assuming that all wars are futile. Note 3: There are plenty of other ways in which one candidate could be much better than another; e.g., 10% chance of a difference of $10B in GDP growth would look rather like 10% chance of a futile $10B war; GDP growth under past US presidents has varied by much more than that.)
I think one of the key things to consider about voting, at least voting in the United States, is that geography matters. If you are in a swing district of a swing state, your potential rewards for voting are going to be substantially different than if you are in a gerrymandered district of a solidly one party state. The gerrymandering of your district might make it much less likely for your vote to influence the district elections, and your state being solidly one party might make it much more difficult for you to influence national presidential elections, since in most states the Presidential winner takes all of the electoral college votes (Although again, this isn't ALWAYS the case: For Nebraska, you can pick up some electoral college votes and not the whole state.)
Also, if you do attempt to follow state and local elections, it takes substantially more effort: People are going to pick up the highlights of presidential political runs and positions if they follow any national media coverage. By comparison, you may need to spend hours doing web searches to even be able to get a vague idea of which of the 100 or so people running for local offices are the 20 you care about. (And some local offices aren't even partisan in my states elections, so there isn't even a single party insignia you can choose, if you wanted to - whereas apparently some areas allow straight ticket selection, so you can just pick D or R and vote for the default candidates of that party across the whole ballot.)
Except, if you are going to be spending a few hours considering the procedure, that's also a few hours you could theoretically spend doing things other than only voting for a national Candidate. Nothing stops someone from deciding "I don't care about my local elections, I'm going to spend that research time calling people and telling them to vote for National Candidate X (or alternatively, I don't care about the national elections, I'm going to go volunteer for local Candidate Y.)"
There are probably other concerns that I haven't even mentioned. Voting and political offices can get much more complex when you start looking at the details.
The probability that your vote changes the election or future elections is negligible. The primary personal benefit for voting is the personal satisfaction of voting.
If this is true, I would advise not voting and completely ignoring politics, freeing up a lot of your time, removing a lot of stress and negative emotions (in my case), and allowing you to associate with people regardless of their political affiliation. Wouldn't you be happier if you literally didn't know anything about politics - what your president's name is, what their policies are?
Voting is not limited to voting for a President of the United States. Local officials (in the USA) of a third party are more common than high-level federal officials (in the USA) of a third party.
If you find yourself saying that a recommendation for political behavior doesn't have any political implications, it's likely that you miss the point.
Insightful political commentary should have political consequences.
In how many US presidential elections did the 3rd party voting affect the outcome? The '2000 US election where Nader played the spoiler is the most recent one, and it has certainly affected the world in profound ways (for example, the Iraq war likely would not have happened, 9/11 or no, the market regulation might have been tighter, etc.).
I would say that if you estimate the number of likely 3rd party voters in your state to be insufficient to change the election outcome in your state with, say, 95% probability, then by all means, vote a 3rd party, otherwise vote strategically.
Perot was a spoiler in 1992. Teddy Roosevelt was arguably a spoiler in 1912. Or (depending on how you look at it) Taft was a spoiler for Roosevelt.
It is easy to study the likelihood that you vote changes the election. One study finds that it roughly varies from 10^-7 to 10^-11 in America for presidential elections.
Given American presidential elections, the likelihood clearly varies by state. In particular, in non-battleground states it is basically zero.
If you follow the link where I say "one study" you can see a graph by state with values ranging like I said between 10^-7 and 10^-11
Ah, yes. Guilty of not following the link, I'll retract the comment.
The study is interesting. Gelman and Silver are highly respected people, but I'm not sure of real-life applicability of far-in-the-tails estimates. The basic issue is that a situation in which an event in the tails is viable is so different from "normality" that the assumptions of the initial analysis no longer apply.
For example, in the study there is a non-zero probability that California would be tied. The state of the world in which this happens doesn't resemble the state of the world in which the polls (on which the analysis is based) were conducted.
It is an unfortunate consequence of the plurality voting system. Plurality is horrible and there are all better alternatives, but it is what we are stuck with for now.
I think this sentence both adds nothing to your conclusion and is also be unnecessarily off-putting to people who do not oppose this system.
I was acting under the assumption that anyone who thought about it did not like plurality. If I am wrong, I am very curious what the reasons are. Is anyone willing to justify plurality to me?
I personally prefer a Condorcet method, but I also think STV is strictly better than plurality, and much more likely to ve accepted. I guess I do prefer plurality to Borda count and anti-plurality.
I have found this assumption to be correct. Also, especially when it comes to economists and other who specialize in voting, I have found that "plurality sucks" and variations thereof are almost universal in academia.
I suggest changing the wording to something like "Plurality voting has many problems and there are alternatives that avoid some of those problems, but plurality is what we have. The question I'm addressing here is to how to vote given that reality." which I think even someone who thinks plurality voting is best overall couldn't reasonably object to.
(For what it's worth, I agree that plurality voting is a terrible system and I would expect the great majority of people who think clearly and honestly about the matter to agree.)
Cross Posted on my blog, By Way of Contradiction.
Anti-Trigger Warning: There is not really any politics in this post. I doubt it will kill your mind.
If your favorite candidate in an election is a third party candidate, should you vote for him?
This question has confused me. I have changed my mind many times, and I have recently changed my mind again. I would like to talk about some of the arguments in both directions and explain the reason for my most recent change.
Con 1) Voting for a third party is throwing your vote away.
We have all heard this argument before, and it is true. It is an unfortunate consequence of the plurality voting system. Plurality is horrible and there are all better alternatives, but it is what we are stuck with for now. If you vote for a third party, the same candidate would be elected as if you did not vote at all.
Pro 1) The probability that you vote changes the election is negligible. All your vote does is add one to the number of people who voted for a given candidate. Your vote for the third party candidate therefore matters more because it is changing a small number by relatively more.
This argument is actually an empirical claim, and I am not sure how well it holds up. It is easy to study the likelihood that you vote changes the election. One study finds that it roughly varies from 10^-7 to 10^-11 in America for presidential elections. However, it is not clear to me just how much your vote affects the strategies of political candidates and voters in the future.
Pro 2) The probability that your vote changes the election or future elections is negligible. The primary personal benefit for voting is the personal satisfaction of voting. This personal satisfaction is maximized by voting for the candidate you agree with the most.
I think that many people if given the choice between changing the next president between the two primary parties or being paid an amount of money equal to the product of the amount of gas they spent to drive to vote and 10^7 would take the money. I am not one of them but any of those people must agree that voting is a bad investment if you do not consider the personal satisfaction. However, I think I might get more satisfaction out of doing my best to change the election, rather than placing a vote that does not matter.
Con 2) Actually if you use a reflexive decision theory, you are much more likely to change the election, so you should vote like it matters.
Looking at the problem like a timeless decision agent, you see that your choice on voting is probably correlated with that of many other people. You voting for a primary party is logically linked with other people voting for a primary party, and those people whose votes are logically linked with yours are more likely to agree with you politically. This could bring the chance of changing the election out of the negligible zone, where you should be deciding based on political consequences.
Pro 3) Your morality should encourage you to vote honestly.
It is not clear to me that I should view a vote for my favorite candidate as an honest vote. If we used the anti-plurality system where the person with the least votes wins, then a vote for my favorite candidate would clearly not be considered an honest one. The "honest" vote should be the vote that you think will maximize your preferences which might be a vote for a primary party.
Pro 4) Strategic voting is like defecting in the prisoner's dilemma. If we all cooperate and vote honestly, we will get the favorite candidate of the largest number of people. If not, then we could end up with someone much worse.
The problem with this is that if we all vote honestly, we get the plurality winner, and the plurality winner is probably not all that great a choice. The obvious voting strategy is not the only problem with plurality. Plurality also discourages compromise, and the results of plurality are changed drastically by honest vote splitting. The plurality candidate is not a good enough goal that I think we should all cooperate to achieve it.
I have decided that in the next election, I will vote for a primary party candidate. I changed my mind almost a year ago after reading Stop Voting for Nincompoops, but after recent further reflection, I have changed my mind back. I believe that Con 1 is valid, Con 2 and the other criticisms above adequately respond to Pro 1 and Pro 2, and I believe that Pro 3 and Pro 4 are invalid for the reasons described above. I would love to hear any opinions on any of these arguments, and would love even more to hear arguments I have not thought of yet.