Strategic Voting

I was talking to someone (whose name shall remain anonymous) about the whole voting for third parties thing, and although my response is long and rambling (the usual for me) I had fun with it, so I thought I would post it here as well.

I sent the following at January 8, 2004 06:44:53 GMT

On Dec 18, 2003, at 20:57, XXXXXXXXXXX wrote:

In a message dated 12/18/2003 8:35:06 AM Eastern Standard Time, abulsme@abulsme.com writes:

I don’t mind voting for people with little chance of winning, because they still better represent my views than the “major” candidates.  Some people say that is “throwing away your vote”, but they are very wrong.  You do not get bonus points for picking the winner.  The only way to truely throw away your vote is to vote for someone who is not really the candidate you think is best.

I don’t agree. I think the most important thing is to get Bush out of office.

Hmmm… well there is a difference right there. I think the MOST important thing is to ensure the outcome best reflects the popular will, as modified by the basic principles of federalism and republicanism. (As opposed to direct democracy, which is pretty crazy.)

So in other words, even if I think another person is the best, if the plurality of people disagree with me, then I *should* be overruled, and the person I disagree with should win.

And I want the results to reflect the true opinions of the voters (again, modified by federalist and republican [small r, not capital] principles), not those opinions distorted by strategic voting.

But then again, I am someone who every four years gets a bit agitated and upset that the original idea of the electoral college has been horribly distorted for over 200 years now.

The way it is SUPPOSED to work is that each state appoints electors (not necessarily by popular vote) and then those electors meet in the electoral college in early January to pick a president. I can even buy the notion that the electors should be elected by the public. But it should be the way it was intended, where they are not committed to any specific candidate until after they get to the electoral college.

But it has been screwed all up, and people are fooled into believing they are actually directly voting for the president rather than slates of electors as they really are. And the electors can only be elected by slate, and not individually, and states are winner take all, and the lectors are pre-committed to a specific candidate… URGH!!! All messed up!!

But anyway… that’s really an argument of the 1810’s… long ago lost and in the past… so we have to work with the system we have now…

But anyway, lets talk a little more about strategic voting…

He is doing a horrible job and I disagree with him on almost everything.

I’m with you on that part. :-) (Of course, I felt the same way about Clinton before him as well.)

If say, 49% vote for Bush (very unlikely, probably more like 70% but this is an example), 45% vote for Dean, and 6% vote for Kucinich, then Bush remains in office. (I know the election is not based on percents but this is an example!) If the 6% who voted for Kucinich had voted for Dean, Bush would be replaced by Dean. But since they didn’t, Bush would have won. If (for example) you like Kucinich the best, Dean the second best, and Bush the worst, then by voting for Kucinich you are getting the person you don’t like at all. If you vote for Dean, you would get your second choice rather than your first. HA I WIN!

OK. Good arguments. And many many people agree. But lets examine them and look at them with a different twist. And we’ll look for a second as if the popular vote was actually relevant, and then look at how it is given the way presidential elections really work…

Instead of the 49, 45, 6 split you hypothesize, Imagine it was Smith 32%, Dean 34%, Bush 33% and 1% other people. (Where Smith is someone made up, but whose views are COMPLETELY different from both Bush and Dean.) In this case Dean SHOULD be the winner. Although he does not have a majority, more people like him than either Bush or Smith. But… if people vote strategically regularly, then the Smith supporters would start thinking they didn’t have a chance, and would not vote for Smith. Instead some of them might vote for Dean, and some for Bush. Which benefits more would be somewhat determined by which views were closer to Smiths, but not entirely. And some people would vote for Smith anyway. In the end, either Bush or Dean could win, despite the fact Dean SHOULD win, because more people agree with him. But what is for sure is that the end result of the election will show support for Smith MUCH less than it really is, which will in turn hurt the chances of Smith (or those with similar views) in the next election. The views represented by Smith will end up marginalized and ignored, even though he represents 32% of the population!!!!

Now, if we had a true popular vote preferential voting system, where each person not only voted for their top choice, but ranked all their choices, with automated runoffs reassigning votes to 2nd and 3rd choices as the lower ranking candidates dropped off, then that would be a very good way of making your first choice, but still having your second choice be relevant. Unfortunately, we don’t have that sort of system.

In real life, what happens is that because EVERYBODY is thinking “I *must* pick one of the top two, or I am throwing away my vote” the whole system gets distorted and polarized, so that everything ends up clustering around more extreme views on various sides of issues, with less room left for middle ground and alternative views, because they are never given an opportunity to grow as minority views and eventually rise to that 2nd or 1st place status. Which is an aweful self-inflicted corruption of the system. If *everybody* actually voted for the people they thought were best, rather than who they thought could win, then the political debate would end up being much more vibrant and dynamic, which in the long term would lead to results which better represent the true opinions of the electorate.

Thus why I think strategic voting (voting for someone other than your true preference in hopes that the short term end result of that particular election will be better) is evil, and it is better to vote for who you think is best, plain and simple.

OK. I’ve probably lost you by now, but if not, lets switch to the other side of the argument and talk pure strategic voting… lets say I buy the idea and think strategic voting is good, and now want to vote in the way that will be most likely to result in Bush losing.

OK, first thing is, my strategy will be different depending on where I live. Different states have different dynamics, and different weights in the electoral process.

We’ll do PA and DC as examples.

First lets talk general election.

In DC, the place is so solidly Democratic it is a gimme. There is absoluelty no chance of a Republican getting DCs electoral votes, and there are not that many anyway. In DC in the General election, if you think a 3rd candidate is better than either Bush or the Democrat, but you hate Bush… you can safely vote for the 3rd person you like. Your view will be represented in the popular vote, the Democrat will still win DC and get the electoral votes, and you will not have effected the balance of Bush vs Democrat in the slightest. There are many states that are so solidly one way or the other that this is the case.

Where I am in PA however, it is a whole different story. PA, like a number of other states, are “swing states”. They are closely divided, and easily swing one way or the other in elections depending on the situation of the moment. In this case, if one buys the whole strategic voting thing, then the argument you gave holds sound, and someone wanting to make sure to get rid of Bush, should vote for the Democrat, even if they really would prefer a 3rd candidate. Because, as you said, their choice has no chance of winning, but if they give their vote for their 2nd choice (the Democrat) then that MIGHT make the difference in making sure Bush loses.

OK, now lets talk Primaries.

Again, it matters very much which state you are in.

Now a few things matter here. First of all is when your state has its primary. To put it bluntly, states that have primaries early matter, states that have them late are completely irrelevant. The only exception is in cases where the primary race is very close… it would be great if that happens again, but it has been a long time since it has.

For instance, DC’s primary is on Jan 13th… but no delegates will be awarded there to let Iowa and New Hampshire go first, so the actual delegates will be decided in the caucus on February 10th… by that time Iowa, New Hampshire, Delaware, Missouri, South Carolina, Arizona, New Mexico, North Dakota, Oklahoma, Michigan, Washington (state) and Maine will already have voted (in that order). By that time there is a good chance someone will be so far ahead that everybody is sure they are going to win… but there is a chance that some interesting things will have happened and things are still close. By contrast, Pennsylvania will not vote until April 27th. Chances are very strong that by the set of Primaries of March 2nd the nominee will be decided already. It *might* not be the case… if so, it will be a very interesting year, and that would be cool. But chances are, by the time Pennsylvania gets to vote, it will pretty much be irrelevant. It might not be mathematically impossible for other people to win at that point, but anybody who was not the frontrunner at that point will probably have “suspended” their campaign to wait for the convention, or dropped out entirely. They may still be on the ballot, but they will not campaign, and will not get results worth mentioning. So if I am in Pennsylvania, my opinion on who should be the nominee matters a LOT less than someone who lives in the states that vote early.

This is another case where the system is somewhat corrupted… what is really happening is that people are electing delegates to the party conventions, and once those delegates are at the convention THEN they should decide who the candidate of the party is… really delegates should be chosen uncommitted to any candidate… in fact there should NOT BE ANY candidates until the convention… then those people chosen to be delegates should decide on someone… but again, now all delegates are pre-committed for the first ballot, and only become free agents if a second ballot is needed, which has not happened for decades… Plus… they really should hold all the primaries nationwide on the same day, so people in all states have some say in the matter, rather than just the early states…

In any case, the pattern for at least the last couple of decades is that anybody who doesn’t win in the first handful of primaries drops out and then whoever is in the lead gets “momentum” because everybody wants to vote for a winner (rather than who they think is best) and everything is decided well before the last states have their primaries.

This is again a place where people actually voting for who they agree with most, through the whole primary process, would lead to better results in the end (and much more interesting conventions!) but, that’s not what tends to happen.

But to be strategic? Well, if the goal is to beat Bush, then one should not even consider how well or not one agrees with the candidates, but ONLY how well they would do against Bush. Who would have the best chance of beating him? All the polls seem to have Dean at way behind Bush if the two are placed head to head. Are any of the others better? I’m not sure they are. Maybe Clark. My personal opinion is on a head to head, he probably has the best chance. I’d have to look at the most recent polls showing all the head to head match ups right before my states primary though, and just pick the one who was doing the best vs Bush. If the goal is truely JUST to beat Bush, then if I agree with the person or not is pretty much irrelevant. Just pick the one that has the best shot against Bush.

Again though, only in primaries before the result is already decided. If by the time my state comes around, the nominee is already a done deal, then I should vote for the candidate I agree with most, because they won’t win anyway, but their delegates may end up having some influence in the convention on marginal issues, or if the candidate who is ahead dies before the convention or some such, in which case they may get to vote on something where the result is not pre-determined, and I want people as delegates who best represent my own views. (Again, a situation where you wish you knew the ACTUAL delegates you were voting for, rather than just who they were committed to… but usually you don’t.)

In any case…. to sum up…

#1) I think strategic voting is something to be avoided in all cases, because while it may give results which are preferred in the short term, in the long term, anything other than people voting for the people they actually think are best (as opposed to people they think can win) results in massive distortion of the process, which is bad for all of us.

#2) If you DO go for strategic voting, it is a very complicated matter since we do NOT have a system based on popular vote in any way shape or form. (And that is for the most part a good thing… there are good reasons for things being structured the way they are.) In the general election, which state you are in and the dynamics of that PARTICULAR STATE have to be looked at both in terms of likely results in the state, and how much weight the state has in the electoral college. The country as a whole is irrelevant. It is only state by state that should be looked at. In the primaries, the dynamics of the individual states are very important, but even more important is where your state is in the primary schedule. What the best strategy for voting is will depend on if you are in an early state or in a late state and what the status of the delegate count is by the time they get to you and how much of a difference your state could make.

Anyway, it is all good fun to look at!

Even better is looking at alternative constitutional structures and examining how they affect both strategic voting strategies, and the expected results of how governments look and how they will (or will not) reflect the views of the electorate, or different parts of the electorate.

Fun stuff!!!