Balkinization  

Wednesday, November 08, 2006

"Spoilers"

Sandy Levinson

I note that both Virginia and Montana have third party candidates for the Senate who have won enough votes possibly to affect the election. A so-called "Green" candidate in Virginia (not actually affiliated with the national Green Party) has won about 25,000 votes, more than the margin between Webb and Allen. There's no particularly way of knowing how her votes would break were she not in the race: she's a former Pentagon analyst who apparently ran on a platform of fiscal responsibility and a high-speed transportation network for northern Virginia. In Montana, the Libertarian candidate got 2.1% of the vote, far more than the margin between Treaster and Burns. I assume that most of those votes would have gone to Burns had it been a forced choice. And, of course, everyone remembers the fiasco of 2000. Might this not be a propitious moment for a bipartisan coalition to propose the Alternative Transferrable Vote as a way of 1) eliminating the role of "spoilers" and b) at the very same time, encouraging third party critics of the ossified two party system to run their races and make their pitches. The ATV allows voters to rank order their favorites. If their #1 choice comes in last (in a three-person race, for ease of analysis), then the #2 choice is counted. This results in a winner who has the most plausible claim to being the genuine choice of the majority, unlike the First Past the Post System, which guarantees the frustration of majority will with some regularity

The point is well illustrated in the Texas gubernatorial race, where the distribution of votes was

RICK PERRY (R) 39 percent

CHRIS BELL (D) 30 percent

CAROLE KEETON STRAYHORN (I) 18 percent

KINKY FRIEDMAN (I) 12 percent

JAMES WERNER (L) 1 percent

In a system where the loser drops out, one has to assume that almost all of Friedman's votes would have gone to Bell. Then, with Strayhorn now the lowest person, one might assume a relatively even split, since her insurgency candidacy (she had been elected to state-wide office as a Republican) was built on opposition to Perry. This could well have given Bell a majority. In a more complicated voting system, where each voter rank orders and one assigns initial votes on the basis, say, of 5 votes for #1 preference, 4 for #2 and so on through the five candidates on the ballot, Straynorn could easily have won, since most of the Bell voters would probably have had her as their second preference. What's interesting is that Perry would quite likely have lost in any system other than First Past the Post. In any event, this is ample proof that the system of voting matters with regard to producing winners, who may or may not plausibly be said to represent majority sentiment. Is there any good reason to prefer a system that results in some predictable incidence of disdain for majority sentiment over one that guarantees a more majoritarian result? I can't think of one (save for arguments made by ethnic minorities that they can occasionally win elections if the "majority" coalition is split between two whites).

I presume that none of my "republican" critics, who believe that I am too pro-democratic in my view of the Constitution, can plausibly criticize this proposal as the end of the Republican Form of Government.

Comments:

It's the end of the Republican form of government!!!

Oh wait, kneejerk reaction...
 

Perceptive Brit comments:

>From Juan Cole's Informed Comment
David Wearing said...
Whilst this election really ought to be none of my business, its sad but true the US effectively dictates UK foreign policy. So this election, in which I didn't get a vote, has more effect on the behaviour of my government in the field of foreign affairs than the elections I do vote in (and when it comes to the US my country is more independent than many, even most).

Hopefully this then permits me to make a couple of observations.

What is a "popular revolution"? Its seems to me that its a fundamental change to the very system or structure of government, moving control decisively from a one set of interests to another, and driven by the popular will. The election result only clearly fits the last of these three categories (the second only superficially, the first not at all).

As Noam Chomsky says in his recent book "Failed States", polling data shows the Democrats and the Republicans to be a good distance to the right of the US public on major issues like the economy, foreign policy and the environment. As always , there's a serious discrepancy between what the public voted for in this election and what they will actually get. The very point of state-capitalist managed democracy is to prevent popular revolution, or even control, of the decision making process. Irrespective of what US voters want, the operative view remains that those “who own the country ought to govern it”, in the words of John Jay.

There's certainly a measure of visceral enjoyment to be had from watching the contemptible Republicans get what's coming to them. However, in the US as here in Britain, you have two wings of the Business Party fighting for power, and this morning control over one part of US state-corporate-imperialist apparatus has simply passed to a less radical set of managers.

Of course small differences like this make for big outcomes in the real world when you're talking about an entity as powerful as the US. But amidst the euphoria its also important to recognise the limits of the difference that this will make. In respect of Iraq for example, we can hardly forget that hundreds of thousands of people, around half infants under the age of 5, were effectively murdered by our governments during the last Democratic presidency through sanctions. Will the welfare of Iraqis now become a driving concern in US Congressional politics?

Dems quibble about tactics. Their's is not a moral objection to the trashing of international law and institutions, the waging of a war of aggression, acts of straightforward terrorism like the assault on Falluja or a return to the age of colonialism (whatever the public makes of such things). They have never argued against the real moral issue: a “War on Terror” conceived not as the pursuit of Al Qaeda but as a generalised, aggressive military-based effort to impose Washington’s will on the world and make the unipolar moment permanent, irrespective of the (non-American) human costs. The Dems have simply argued that the policy isn’t being pursued successfully. Wehrmacht generals probably had stand-up blazing rows about the success of Operation Barbarossa, but none about whether German Imperialism was right or wrong. Ask yourself what would have happened if Iraq had been colonised successfully, with a client government, Vatican City–sized US Embassy, control over oil production, permanent military bases and the effective exclusion of Iraqis from the running of their country all secured. Would there have been a peep from the Democrats?

There was a change of management in the US Government after Tet, reflecting a common view that the US should get out of a losing situation. What followed was Vietnamization and carpet bombing (the latter helping to set the stage for the Khmer Rouge). In other words, the change was a tactical one in the interests of power, with the effects on the lesser races as irrelevant as ever.

Ultimately, whatever the US public may have hoped for when they voted, the political effect of this election in terms of the make up of your government constitutes a tactical revolution, little more. Its not insignificant, in fact its important in many of the ways that Juan describes, but its limits are considerable and should not be overlooked.
 

Professor,

Why does ATV seem to me like a violation of "one person, one vote"? I cannot currently explicate it myself, but I'm betting you've heard similar concerns better expressed and also have ready answers for same.
 

Sandy,

There are various mathematical ways to pick a single winner out of a field of more than two candidates, none of which can guarantee that all fairness criteria are met. Let us use your method.

We'll look at the following election:

1st place - Joe Schmoe

- 145 1st place votes
- 5 2nd place votes

2nd place - Rin Din

- 145 1st place votes
- 5 2nd place votes

3rd place - Nanc Fanc

- 10 1st place votes
- 290 2nd place votes

Followers of the two major candidate refused to vote for the other major candidate. Instead, they all voted for the third party candidate as a quasi-protest vote. Round one, no majority winner. Round two, Nanc Fanc, the third party candidate wins!

Obviously, I cooked the books, but that's the point. I don't think you really intended to elect Ralph Nader in 2000. It's difficult to say, however, that the election wasn't fair. You set it up. Nanc won it fair and square.

Generally, when there are more than two candidates, it is impossible to satisfy all election fairness criteria. However, there are various methods that have been developed mathematically: Borda Count, Plurality with Elimination, Pairwise Comparision. Again, none guarantees fairness. Fairness is mathematically an impossibility where there are more than two candidates.

Also, you'd have to do the analysis at the level of the candidate. That would mean for Virginia senate, you'd have to await all precincts before starting your analysis. It would have to be done at the state level. At the national level, right now, the method would be impracticable, since there is no national voting agency that aggregates all votes, prior to doing analysis on them.

I really enjoy your blog, one of my daily essentials. Finally I have something to post.
 

I should have assigned the first place candidate 146, and the second place candidate 144.
 

Actually, I now realize that my method is different from yours. The method you are suggesting is an actual mathematical method, the Borda count method. Same caveats apply, however.
 

Just think of how this would have worked in 1860. Abraham Lincoln got only 39.8% of the popular vote. Assuming that the various versions of Democrats had combined, we could have avoided the entire Civil War. And of course, we would still have all those slaves.
 

I think the reality of this voting method is that it is too complex in several ways.

First, voters (all voters) have to understand a new method: rank your choices.

Second, vote tabulation now is extremely important at every level. All votes must be counted before any decision can be made. Candidates will challenge the results of even minor candidates.

The problem is that this method is very logical, there are great reasons for voting this way, maybe similar to the way the Pope is selected. But the difference is that you have multiple ballots and the ability to discuss and adjust.

Third, the technical details of the ballot are increased, since you can't just fill in a bubble. Instead of a hanging chad, we have "is that a one, two or three".

I like the idea from a theoretical viewpoint, but reality may be difficult to achieve, and would probably require electronic voting equipment.
 

I agree with HGilbert, Sandy, although in a present tense. Consider the experience in San Francisco where supervisorial elections have adopted a plan much like your rollover ATV; the unanticipated outcome was precampaign secret alliances and subsequent outright slate campaigning. City supervisor elections there are nonpartisan.

I much prefer the stress which the two principal party organizations must feel when facing upstart splinter parties: to bring them under the umbrella, by compromise, and by giving the microconstituencies a voice. It is a way for the two majority parties to mature with the times and stay in touch with the people who vote.
 

A better option IMHO is to allow the candidate rather than the voter to transfer his or her votes to another candidate. Under this system, the candidates would bargin with each other to try to come up with a winning plurality. For example, if the Libertarian candidate has 5% of the vote, enough to swing the election, he could give his votes to either the Republican or the Democrat, depending on which one was willing to give him the most policy concessions. Alternatively, of course, the Republican and Democrat could reach a deal and leave the Libertarian out. This system of allowing the candidates to negotiate would allow the voters to recieve the maximum benefit for their votes. Note that the system does not need to end in a majority; a plurality would win, after allowing a short period of time for the candidates to negotiate. In this way, deadlocks are prevented.

This also has the advantage that no changes would be needed in the balloting. Each voter would still vote for the one candidate that best represents the voter and his/her interests.

As an embellishment, the system could also be made opt-in or opt-out. In addition to voting for one candidate, the voter could also check a box to indicate whether that vote is transferable. Under this system, the candidate would not be able to transfer all of his or her votes, only the ones that the voters designated as transferable. In this way, voters concerned about their votes being given to an unacceptable candidate can opt to make their votes non-transferable.
 

P.S., let me add another comment.

Under our current system, voters essentially have to choose between one of two candidates: the Democratic nominee or the Republican nominee. The Democratic nominee is chosen only by Democrats, and is significantly left of center. The Republican nominee is chosen only by Republicans and is significantly right of center. The net result is that we end up electing candidates at the extremes instead of the center.

An advantage of an alternative voting method is that we could eliminate primaries and have all candidates run in one election. This would tend to result in centrist candidates being elected, rather than exteme candidates.
 

If Canada and Great Britain will switch to some form of proportional representation, the U.S. will follow. The U.S. pays more attention to those two nations, than any others. So, Brits, you know what to do! Tony Blair promised to consider this years ago...what's going on?
 

But a significant number of people did choose to vote for Kinky Friedman knowing perfectly well that it was impossible for him to win. I think voting in protest is a valid and sensible thing to do sometimes. I don't know a lot about current Texas politics, but it looks to me as if a lot of people who would never vote for Perry very definitely did not want to vote for Bell. In other words, can we really say that the Friedman voters wanted Bell to win? Anyway, how exactly do we know that Bell would have received half of the Strayhorn votes in a two-candidate election?

If voters were allowed to rank their preferences, it seems they would be less likely to cast unenthusiastic votes for candidates like Bell and Perry, and more likely to throw caution to the wind and put someone like Kinky Friedman at the top of their ballots.
 

"Is there any good reason to prefer a system that results in some predictable incidence of disdain for majority sentiment over one that guarantees a more majoritarian result?"

I believe that it is mathematically possible and not even terribly unlikely that a candidate with an absolute majority of first-place votes will not be elected under a Borda count.

The psychological effects and tactical-voting ramifications of this kind of voting do look interesting. I don't think, however, you can look at the Texas results and determine the likely outcome of a Borda vote. A vote for Bell in the actual election might not equate to a first-place vote for Bell in a Borda election.
 

Sandy Levinson asks if it
"Might... be a propitious moment for a bipartisan coalition to propose
[Instant Runoff Voting IRV] as a way of 1) eliminating the role of "spoilers" and
b) at the very same time, encouraging third party critics of the ossified two party
system to run their races and make their pitches."

Unfortunately:
1) IRV does NOT eliminate spoilers, see
http://rangevoting.org/IRVpartic.html

for a proof. Indeed IRV can cause an even-worse kind of spoiler effect that was not previously
possible - you by voting, can actually cause the election result to get worse than if you had
not voted at all.

2) IRV does NOT help third parties. Every IRV country in the world is 2-party dominated in
all IRV seats. In contrast, voting with a top2 second-round "runoff" leads
to MORE than 2 powerful parties, as is demonstrated by
France, Argentina, Brazil, Central African Republic,
Chile, Colombia, Congo, Cyprus(meaning the Greek part),
Finland, Iran, Liberia, Macedonia, Mali, Nicaragua, Niger,
Peru, Poland, Portugal, Romania, Senegal, Slovakia.

It is therefore EXTREMELY SILLY
for US third parties to support efforts to convert elections from 2-round top2
runoff to "instant" runoff.

Levinson also uses the TX 2006 governor election as an example.
For an exit poll of this election conducted with range voting, see
http://rangevoting.org/Beaumont.html.

To learn about a voting system called RANGE VOTING which genuinely DOES eliminate spoilers
and which genuinely does help third parties, check
The Center for Range Voting website.

-Warren D Smith.
 

robert link write, "Why does ATV seem to me like a violation of "one person, one vote"?"

It is not. It provides "one person, one vote." Under the current system, a vote for a third party is give a person no power to choose between the top two contenders. With IRV, every citizen has a chance to express which candidate they prefer. In any case, how could it violate one-person-one vote if everyone has the same right to rank their choices?

russael writes: "Followers of the two major candidate refused to vote for the other major candidate. Instead, they all voted for the third party candidate as a quasi-protest vote. . . . Obviously, I cooked the books, but that's the point."

What more really needs to be said? Why would every conservative "protest" by selecting Nader? I scoff. It seems I must inform you that there are other 3rd party candidates. As a person who is generally liberal, I would not select Pat Buchanan over a more moderate Republican.

The current 3d party/independent spoiler is a very real problem. Why invent a fantasy scenario to fault IRV?

Alan writes, "A better option IMHO is to allow the candidate rather than the voter to transfer his or her votes to another candidate."
I would hate to see the litigation when a "deal" is brokered just one minute late. This system might also give a relatively small party tremendous power. Or if there are enough parties, you are really creating a second-stage proportional representation congress: each person votes for a party, then a coalition forms a "government" to appoint a "prime minister." Why distance voters even further from their elected representatives?

Richard Winger writes, "If Canada and Great Britain will switch to some form of proportional representation, the U.S. will follow."
This is a joke, right?

Mr. Warren D Smith supports range voting. He criticizes IRV because in some limited circumstances, a voter might vote strategically to get their second best choice. However, he admits that range voting encourages strategic voting as well: from the FAQs:
"Doesn't my vote for my second-favorite B perhaps hurt my first-favorite A? Yes, if it comes down to an A versus B battle. In that case, you might end up wishing you'd given B zero! But, in every other circumstance, it helps you."
Oops! Most Americans would choose the more moderate Republican or Democrat over Nader or a Libertarian candidate, but the smart way to vote is to give the preferred choice, A, 99 and the second-favorite, B, 0. I can't believe someone would spend so much time defending a system he already knows doesn't work. And, uh, we're not bees.
 

I have to go to Minneapolis tomorrow morning at 6AM, so I don't have time to respond to all of these interesting and thoughtful comments. I do want to make one observation about the Australian Senate, though. My understanding is that each of the six states has six senators AND THAT THEY ARE ELECTED IN BY A SYSTEM OF PROPORTIONAL REPRESENTATION that means that there are third (and maybe even fourth) parties in some of the state delegations. This strikes me as an excellent notion, where the second house purposely opens itself up to a wider range of represenatives than is available in the first house. (Incidentally, this requires that all the senators be elected simultaneously, rather than the staggered term system that we have.)

In any event, I am delighted that this discussion is taking on a comparative dimension.
 

Russael's example isn't the way I understood this to work. I thought that the candidate with the fewest votes was eliminated, and then we looked at the ballots of everyone being counted as a vote for that candidate, and looked at who their next choice was. So the first step to determining a winner is to look at the 10 ballots for Nanc Fanc and assign their votes to their second choice. So we'd add say 6 to one and 4 to the other. Then the next step is to eliminate the new bottom candidate, and check the ballots of the people who voted for that one to reassign them to their next choice, if they are still in the race. That would continue until there was only one candidate left, the winner.

I also wish there was a way to vote AGAINST someone. In the gov race, I was considering voting for Grace Ross, thinking that maybe enough of us would think Oatrick was so safe that we'd all vote for Ross, but then I worried that maybe Healey would lose by a whisker to Patrick, so I'd regret not voting for her. But if could have just voted against Patrick, then either way, my vote would have counted. Not that I don't like Patrick, mind you, I just like underdogs.

How about if we had CHANGABLE votes? We could vote on-line, staring say November 1st. Totals would be posted, and we'd see that Patrick had a huge lead. And he'd reveal some character flaw, and we could switch or vote to Ross. More people would see that Ross was truly gaining, and they'd switch to Ross too. This would go on for a week or two, and we could see what the effect of our vote is, realtime, and have time to think about changing it. Or, most likely, we could vote once and forget about it. Then at the end of two weeks, voting would end, and we'd instantly see who we decided on. Technically possible?
 

Let me respond to some previous comments.

Sandy Levinson:
"I do want to make one observation about the Australian Senate, though. My understanding is that each of the six states has six senators AND THAT THEY ARE ELECTED IN BY A SYSTEM OF PROPORTIONAL REPRESENTATION that means that there are third (and maybe even fourth) parties in some of the state delegations."

This is all true. However the Australian HOUSE is elected via IRV. IRV is not proportional representation, it is single-winner. Australian
third parties want to get rid of IRV and switch to PR:
http;//rangevoting.org/AusIRV.html.

Russael: gives an election example in which he says Nanc Fanc wins. However, according to the rules of IRV, Nanc Fanc actually is eliminated in round 1.
According to the rules of Borda voting, Nanc Fanc
does win. Borda voting is however an extremely bad voting system, see
http://rangevoting.org/rangeVborda.html.
Levinson's initial post was about Borda but
elsewhere he was advocating IRV. (I actually
originally was responding to a post attributed to Levinson at this blog, but which was posted on a different blog.)

C&d writes:
Oops! Most Americans would choose the more moderate Republican or Democrat over Nader or a Libertarian candidate, but the smart way to vote is to give the preferred choice, A, 99 and the second-favorite, B, 0. I can't believe someone would spend so much time defending a system he already knows doesn't work. And, uh, we're not bees.

To respond, I nowhere insisted that c&d vote in
the way he considers to be non-smart, with range voting. By all means, be smart. In that case,
c&d would vote as follows with range voting (I assume):
major party canddt A he least-hates: 99
major party canddt B he most-hates: 0
minor party candidate C he likes more than A: 99
minor party canndt D he likes more than B but less than A: 52?
minor party candidate E he likes less than B: 0.

Now, if every voter acted this way and most felt similarly to my critic
[except for switching roles on A vs B], then C would be elected.
This would be good if indeed most voters
liked C the most.

However, suppose we instead were using
Borda voting. Then my critic's
smart voters would vote
A=max, B=min (or the reverse) and C=middle
(forced by the rules of Borda in a 3-canddt election)
and C could never win. Not so good.
Under the same voter behavior, C also could nver win using IRV. Also not so good.
 

sorry, I mistyped a URL last post:
http://rangevoting.org/AusIRV.html
 

Eerst Europa Doelstellingen: De Ci2i Verzekering (Ci2i) zal het nummer een gebrandmerkte pan Europese commoditized online verzekeringsmakelaar door 2010 zijn.
 

free credit report" REL="nofollow">free credit report credit repair" REL="nofollow">credit repair credit reporting" REL="nofollow">credit reporting credit rating" REL="nofollow">credit rating credit check" REL="nofollow">credit check credit reports" REL="nofollow">credit reports credit history" REL="nofollow">credit history credit score" REL="nofollow">credit score credit report" REL="nofollow">credit report online credit history" REL="nofollow">online credit history web credit report" REL="nofollow">web credit report internet credit reports" REL="nofollow">internet credit reports internet credit report" REL="nofollow">internet credit report free credit reporting" REL="nofollow">free credit reporting instant credit report online" REL="nofollow">instant credit report online credit history online" REL="nofollow">credit history online
 

What is a popular revolution? Its seems to me that a fundamental change in the system or the structure of government to move decisively control a set of one interest to another, and led by the popular will. The outcome of the election clears the last of these three categories (only the second surface, the first not at all).WOW GoldWOW Items GoldBuy WOW ItemsCheap WOW ItemsTera Goldbuy tera goldTera GoldBuy WOW Gold
 

A simple Internet search shows a number of specialized online resources that sell Cheapest WOW GoldWorld of Warcraft. As a dedicated gamer, you know that the Buy World Of Warcraft Gold credit cards are used to buy stocks in order to improve your gambling skills.
 

seorang istri terkadang ingin lebih merasakan saat berhubungan dengan pasangannya lebih memuaskan maka dari itu kami memperkenalkan produk obat kuat tahan lama yang khusus untuk pria yang tidak ada gangguan ereksi.

 

greetings all.
I read this article very interesting, so I really want melajutkan read everything on this blog.
greetings blogging.
alat bantu sex alat sex obat kuat
 

Post a Comment

Home