Real Reform Begins In The Ballot Box
I remember about a month ago there was a lot of bruhaha about Pennsylvanian Republicans trying to change their contribution to the Electoral College by divvying up the votes based on Congressional districts, like Maine and Nebraska. Naturally, a lot of people got upset with that, with some (like Doug Mataconis and George Will) saying we should keep the Electoral College just like it is now, and many others saying that we should instead move to a National Popular Vote system. Now with Occupy Wall Street taking over our media senses, some of that talk has been pushed aside, with people instead focusing on Wall Street rather than Pennsylvania Avenue.
I would like to go back to the Avenue, however, for multiple reasons. First off, I actually think that a lack of serious political reforms is the reason for much of the discontent we’re seeing in Zucotti Park. Second, we have Congressional deadlock, as always—but in recent years, the vitriol and polarization we have seen has increased dramatically. Third, even with the 2010 GOP landslide in the House, we still have a very high incumbent reelection rate—although it was lower in previous elections, it still stood at 87%. Fourth, we have not seen any new ideas with regards to the major issues of the day: our debt crisis, our flagging economy, our eroding civil liberties, or our overburdening government.
Clearly, the emphasis is on the egg and not the noggin in the egg nog, here.
Very little can be done to change or institute major reform, even though we need it, badly. Part of that is by design. The Founders wanted a system where it would be difficult to radically change it, in order to preserve the liberty they had fought so hard for. In the modern era, that backfired. Instead of preserving liberty, the system is preserving the corrupt bog from which liberty is being drowned in.
There are numerous ideas floated in order to fix Washington. Term limits are a very popular one, and I agree with them. Then there are myriad forms of “campaign finance reform,” which I doubt will actually fix anything. Balanced budget amendments are another. So are independent redistricting commissions (which, I think, are a fantastic idea.) Unfortunately (or fortunately, in some cases), none of these will ever pass due to a much more fundamental problem being ignored, a part of the system that is at the very crux of our democratic system, even more basic than the three branches of government and separation of powers.
It’s what voting theorists call the “first-past-the-post system.”
Currently, when we go to a voting booth and are presented with a multitude of options (if that does happen, thank your lucky stars), we are allowed to choose one of the candidates to vote for. I always like to take the 2000 presidential election as an example, since its the most clear version (especially when simplified a bit.) So, you have Bush, Gore, and Nader. Now, the most sensible thing to do would be to vote Nader, so you could annoy just about everyone, but nobody truly votes sensibly anymore. So, if you’re fairly conservative, you vote for Bush, and if you’re fairly liberal, you vote for Gore—because he’s the only one with the chance to beat Bush. This system leaves Nader and other third party candidates out in the cold, a symptom known as the “spoiler effect.” In fact, the 2000 election is the quintessential example of it, as Nader “spoiled” the election for Gore by diverting just enough votes to throw the election into question. Nevermind that the only truly spoiled vote is the vote for someone other than your first choice, it’s still the leading and most popular argument against third party candidates and the main reason why they can’t get any traction.
But what about approval voting?
You have the same situation: Bush, Gore, and Nader. But this time, if you’re fairly liberal, you can vote for Gore and Nader; you can vote your conscience (if voting for the guy who killed muscle cars is your sense of “conscience”) while at the same time engaging in tactical voting. Neither candidate is truly hurt, and you win. On the flip side, in 2008, you might have written in Ron Paul and voted for him…while simultaneously voting for Bob Barr on the Libertarian Party ticket or McCain (or both!)
The thing is, first past the post system worked in a very different time. It functioned fine in the early days of the republic, when the electorate was very narrow and essentially restricted to wealthy, male, landowners. Even if Candidate A beat Candidate B in a tight race (that old “51% lording it over the 49%” saw), the interests of the two groups were still relatively close, at least as compared to, say, the general population, who couldn’t vote.
But that was 1791. Now it’s 2011. Our electorate has ballooned from a couple of million old white guys to well over 200 million whites, blacks, Hispanics, Asians, women, men, old, young, some who own property, others who are penniless, some with twelve degrees and an alphabet of postnominal letters, and some who didn’t even graduate high school. Opinions are not similar at all; they are wild and varied, with no hard and fast lines delinating one school of thought from another. Sure, you have liberals on one hand and conservatives on the other, but then there’s a whole murky purple haze in between where they start to blur together, sometimes leading to socialism, other times leading to libertarianism. And that’s before we get to the infinite variations in each camp! (Or the fact that the political “spectrum” is not based on two points, either.)
Trying to shoehorn this eclectic electorate into two ill-fitting boxes is producing unbelievable amounts of distress and discontent. Part of this is helped along by the media, who try to latch on to the “frontrunner” immediately, then ignore everyone else and push that frontrunner to the top because, well, the voters don’t know anyone else. Part of the Occupy Wall Street crowd are there because they thought they were getting a left-wing president who jived with their viewpoints…but really, thanks to this system, what they got was a slick salesman who was a mishmash of a bunch of ideas that didn’t work well together and certainly didn’t benefit anyone on the ground. I guarantee, no matter who wins the GOP primary and (with all likelihood) the general election in 2012, there will be similar discontent and unhappiness on the Right, because they will not receive a conservative president.
By switching from this antiquated, obsolete, and dsyfunctional electoral system, wholly in hock to Duverger’s Law, and moving towards this modern, responsive, and functional one, we can reap several benefits:
- “Consensus” Candidates: If group A strongly back Candidate A, and group B strongly back Candidate B, but large parts of both groups also like Candidate C, C will most likely get elected. This will go a long way towards tempering the heated rhetoric and hyperpolarization we’ve seen in the past couple of decades.
- No Spoiler Effect Means More Independents & Third Party Candidates: Many people would not run on a third party or independent ticket for any office (except maybe at the municipal level) because they fear the “spoiler effect,” which is that they will not get any votes since voters don’t want the other guy to win. Not so with Approval; there will be a much more diverse base of candidates, and more likely there will be someone people will like. The Center for Range Voting, which, admittedly, only prefers Approval Voting as a secondary measure, has this to say:
In any race with more than two candidates, plurality [first-past-the-post] may elect the candidate least acceptable to the majority of voters. This frequently happens in a three-way contest, when the majority splits its votes between two centrist candidates, allowing a more extreme candidate to win. Plurality also forces minor-party candidates into the role of spoilers, as we saw in USA 2000, which can be decisive in a close contest between the major-party candidates. That in turn causes many voters to refuse to vote for their true favorite unless it coincides with one of the two candidates that seem most likely to win – out of a (justified) fear of creating a “spoiler” or “wasted vote” scenario. And that in turn causes third parties to die out over time, causing 2-party domination, which severely reduces voter choice and thus makes democracy work badly.
With approval voting, spoilers do not happen; approving your true favorite is never strategically unwise; approving candidates unlikely to win is no longer “throwing away your vote.”
- Cleaner Campaigns: One of the worst aspects of modern political campaigns is the mudslingling and negative ads. Nobody likes it, but everyone does it. In a zero-sum game under first-past-the-post, you have to get the electorate to hate your opponent so they will vote for you. Under Approval, however, there is less of an incentive. Sure, they might vote for your opponent…but at the same time, they can vote for you too. There’s less of a need to go negative, and more of an incentive to stay classy and on the level—lest voters think you’re just a nattering nabob of negativity and disapprove of you.
- Better Candidates: With cleaner campaigns come better candidates. How many would really want to put their hat in the ring for public office, knowing that their every word will be scrutinized, their past will be dug up and aired out for potential scandals, that they will not only be misconstrued but actually demonized in front of millions of people. Some people are masochists who live for that stuff, but I would suspect that the majority of Americans really aren’t in for that sort of thing. They’d rather upload an offensive video to Youtube while using a pseudonym like “Dumbguy999”. With Approval, more will be tempted to run for office, knowing that their time and resources won’t be spent just making sure the public knows he or she isn’t a bat demon from Mars. Who knows: if we implemented Approval years ago, maybe Steve Jobs would have been President of the United States rather than President and CEO of Apple, Inc. (Whether or not that would have been a good thing, I leave up to you, dear reader.)
- More Libertarianism: Okay, this one is something of a joke, though not entirely. It’s essentially Point #1, but applied. Think about this: since the financial crisis and the non-cessation of the conflicts in Afghanistan and Iraq (plus the expansions in Libya and now Uganda of all places) there has been a not-inconsiderable shift in public opinion towards libertarianism. The problem is that neither party supports this view: they may be fiscally conservative, but also socially conservative, or socially liberal, but also fiscally liberal. What voters would need is some candidate that combines the best of both worlds. Under Approval, those who generally see themselves as more left-wing but supportive of libertarians would vote for both the Democrat and, mayhaps, the Libertarian Party candidate, while the same would happen for those who see themselves as generally right-wing but also supportive of libertarians. Now, true, the majority of people saying they are “libertarian” under these guidelines probably still want their government “services,” but not to pay for them, but with Approval, that might actually register with real libertarian politicians.
Now, there are a few niggling questions. The first is, “But what about ‘one person one vote?’” A very good question, and the answer is: same as it ever was. Under our current system, a “vote” is construed as your entire ballot. Under Approval, you will still have one vote—for each candidate. If you don’t mark a candidate for approval, you don’t get to “save” that vote and use it on another; instead, you’ve effectively voted yourdisapproval on that candidate. All the ballots are weighed the same, nobody receives more votes than another. The Center for Election Science (well, that’s what they call themselves) have a better explanation on the subject than me.
Second, you may have heard of “Instant Runoff Voting,” or IRV. You might be asking: “Why don’t we use that?” Simple: it sucks. No, wait, let me rephrase: it is needlessly complicated, inherently flawed, and extremely susceptible to tactical voting shenanigans. It led to ridiculous absurdities in the 2009 mayoral election in Burlington, Vermont, and the official results of other elections are impossible to decipher for the ordinary voter (thus leading to increased suspicions of “cheating” or election fraud, which is never a good thing.) Approval voting, meanwhile, is practically impervious to tactical voting, and is extremely simple. In fact, the machines and calculations don’t have to be changed at all.
When you look at all this evidence, it’s clear our own voting system has failed us. It can’t cope with all the myriad variations out there, leads to increased antagonism and polarization, and rewards the same tired old views without allowing any new blood in. The real question we must face is: “So how do we get that?” Unfortunately, right now, I do not know. We are in such a hole, with our finances, with our economy, with our national security and foreign policy, it seems difficult to imagine how we will be able to make this a reality.
But maybe that’s the key, right there. Things seem to be breaking down all around us. The old paradigm is being washed away. We saw big flips in 2006 and 2010, as our electorate grows increasingly dissatisfied with the choices presented. Jay Cost wrote back in 2009 that we may be entering another “Jacksonian moment” (bolded emphasis mine.)
We might be on the verge of another Jacksonian moment: a time when the people awake from their slumber, angrily exercise their sovereign authority, and mercilessly fire the leaders who have for too long catered to the elites rather than average people. The first time this happened was in 1828 - when the people rallied to the cause of Old Hickory to avenge the “Corrupt Bargain” of four years prior. It’s happened several times throughout the centuries. Most relevant to today, it happened time and again in the 1880s and 1890s, as the people hired then fired one Republican and Democratic majority after another in search of leaders who could attend to the people’s interests instead of the special interests. That age saw the birth of the Populist Party. It was a time when so many felt so disgruntled by the political process that young William Jennings Bryan - just thirty-six years old and with only two terms in the House - came within a hundred thousand votes of the presidency.
I wonder if we’ve returned to that kind of dynamic. In true Jacksonian fashion, the country fired the Republicans in 2006 and 2008 because they bungled the war in Iraq and allowed the economy to sink into recession. They might soon have another Jacksonian moment, and fire these equally useless Democrats for hampering the recovery, exploding the deficit, and playing politics with health care.
Similar sentiments are expressed in Nick Gillespie and Matt Welch’s new book, Declaration of Independents, where they argue the duopoly of the two parties is not stable, and will eventually give way to something new (and, they additionally argue, libertarian.) I think this is pretty much on the money, mainly because we don’t have any money to pay for either Republicans’ defense programs or Democrats’ social welfare projects, and so these things will go by the wayside one way or another. Maybe in this new environment, we’ll see the plurality vote system go as well, replaced by Approval Voting and cleaner, calmer, easier elections.
May we live in exciting times.
United Liberty








Great post overall. You are right that it is deeply important to focus on voting reform, because it would break the current deadlock and enable other necessary change.
I would suggest that even more than the single-winner systems you discuss, proportional representation is the true holy grail of voting reform. Without proportional representation, minor parties will still only win scattered local victories, and not show their true, fair strength of support. I understand that you may be making a tactical choice to tackle single-winner reform first; still, you should at least mention the deeper strategic goal of proportionality.
You do make one misstatement which I have to call you on. I strongly support approval voting, but it is not true that it is “practically impervious to tactical voting”. As James Green-Armytage showed at http://www.econ.ucsb.edu/~armytage/svn2010.pdf , approval voting is actually more subject to tactical voting than the average voting system. The point I think you were trying to make is that tactical voting under approval is both honest, in that you don’t have to reverse any true preferences, and socially beneficial, in that it tends to elect a candidate who makes the average voter happier. Tactical voting might be rarer under IRV; but that matters little if IRVs results are worse with or without tactical voting.
I concur completely with proportional representation. In fact, I think that if the House was 50% PR and 50% geographic (I can’t see the House going completely over to PR, there still need to be SOME districts to directly represent people) we would see a far more responsive Congress that would actually reflect the populace. Changing to a PR system in the House, however, is far more difficult and far away than approval voting, which is why I focused on the latter, but you’re right: it deserves its own blog post.
As for your citation, well, I’m not sure what to say. I guess maybe “tactical voting” is a misnomer; I was looking for voters being disingenuous and voting for people they don’t prefer. At least with approval, there’s no incentive to do so, unlike under plurality and other voting systems. That’s what I’m trying to eliminate or at least cut back on.
And let me welcome you to United Liberty. Thanks for commenting!
Getting Score Voting or Approval Voting is probably (almost certainly) a PREREQUISITE to getting proportional representation in the US, beyond a handful of cases in progressive areas.
http://asitoughttobe.wordpress.com/2010/07/18/score-voting/
Jameson,
I strongly disagree that Approval Voting is more subject to tactical voting. The metric that James Green-Armytage was using to assess this was extremely misleading, because for instance it said that IRV was not very susceptible to tactics. Yet Approval Voting performs as well with 100% tactical voters as IRV with 100% honest voters. You can see this fallacy depicted graphically here:
http://www.electology.org/tactical-voting
Warren Smith (Princeton math Ph.D. behind ScoreVoting.net) had a lengthy discussion with Green-Armytage here:
http://groups.yahoo.com/group/RangeVoting/message/10439
It appears that what Green-Armytage is talking about is the probability that a situation will arise in which some voters could have changed the result with strategy. That may be academically interesting, but it’s irrelevant to the issue of voting method performance. Performance is about frequency *times severity*, which is what Bayesian Regret gives you.
http://ScoreVoting.net/BayRegsFig.html
We are actually agreeing here. Approval voting may cause more “tactical voting” than IRV, but it never causes dishonest tactical voting, and, as you say, even the worst tactical result under Approval can be better than the best honest result under IRV.
It is not correct that it never causes dishonest tactical voting. There are cases when you prefer A>B>C>D and want to approve A and C only.
http://scorevoting.net/RVstrat2.html
Glad you’re thinking structurally, Jeremy. And glad in responding to Jameson, you recognize the great value of proportional voting/fair voting. That’s certainly what we think at fairvote.org
Clay, Jameson and I often spark on the relative merits of the less important conversation about different single winner methods. But here’s something to consider: majority systems based on the runoff logic (if no 50% majority, then go to a second round, generally between the top first-round finishers) are a tested way of allowing voters to have more than two choices. Most presidents are elected that way. It’s definitely better than plurality voting.
So to trash runoffs and instant runoff voting in favor of ideas with essentially no proven and review-able track record seems tactically counterproductive. For instant runoff voting, if you were a voter in in Ireland’s presidential election that’s taking place right now, you would be feeling very free to vote your conscience. I don’t see the value of criticizing that and all the other advances for IRV in the US, including for mayoral races this fall in San Francisco, Portland (ME), Telluride (CO) and more. I’d suggest not to let the perfect be the enemy of the good.
That’s especially true when you consider that approval voting is likely to get very messy, very fast. Jameson privately has agreed with me that in the much-discussed Burlinton mayoral race in 2009, approval voting likely would have elected the plurality leader, but Condorcet loser (among the top three candidates) and very likely finished with support of well under 50% of voters. For our skeptical take on approval voting, see www.approvalvoting.blogspot.com — take a particularly close look at how the similar “bucklin” system worked (or, rather didn’t work work). I’m hoping approval voting advocated spend a lot more time researching that history and reporting back to us on lessons learned.
But bottomline — think structural, as you are.
> Jameson privately has agreed with me that in the much-discussed Burlinton mayoral race in 2009, approval voting likely would have elected the plurality leader, but Condorcet loser (among the top three candidates)
By which you mean the Republican, Kurt Wright. You’ve been making this claim over and over again without any evidence. Here’s a nearly bullet-proof argument that actually Montroll (the Democrat, and Condorcet winner) would have won with Approval Voting.
http://scorevoting.net/Burlington.html#UngerApp
I have pointed this out to Rob Richie in something like 10 different comment areas like this one, but he has never addressed it, nor provided any evidence that Wright would have won.
> Clay, Jameson and I often spark on the relative merits of the less important conversation about different single winner methods.
There’s no evidence that it’s less important. Single-winner posts (e.g. mayor, governor, senator, president) have an enormous amount of influence in government. Bayesian Regret calculations show that Score Voting and Approval Voting would roughly DOUBLE the democratic-ness of those single-winner elections. There is no comparable way of measuring the effectiveness of multi-winner systems like STV. It could be that a Congress comprised of single-member districts could be just as representative as with a proportional system, if the winners were decided with GOOD single-winner systems like Score or Approval Voting. It’s highly speculative, but all we know is that you’re putting the cart before the horse. Right now we have single-member districts, so we must focus on getting the best system in place for those elections. As we weaken the two-party stranglehold on American politics, it will become easier to enact systems like PR.
Your argument about Bucklin voting is that a great number of people “bullet voted” with it, only for one candidate. This fails as a criticism of Approval Voting for several reasons. Most notable is that “bullet voting” has been, in practice, a worse problem for IRV than for Approval Voting.
http://www.electology.org/bullet-voting#TOC-Real-world-data
Moreover, FairVote’s own critique of Bucklin voting argues that on average 13% of voters opted to support a second candidate. That’s a significant number which counters your criticism that Score/Approval would just “degenerate into Plurality” in practice.
Lastly, to say that Score Voting and Approval Voting are untested is simply absurd. Many large organizations with more members than most US cities (e.g. the American Mathematical Society) have used Approval Voting for their internal elections for many years, and have found the system to work well. The German Pirate Party, which just won about 10% of the seats in the Berlin parliament, has used Approval Voting extensively, included for the election of its party leader, Sebastian Nerz. The ballot data is publicly available, and it shows that many people voted for more than one candidate. Here’s more on examples of Approval Voting use:
http://www.electology.org/approval-voting#TOC-Where-has-Approval-Voting-…
The bottom line is that we have a clear game theoretical understanding of rational Score/Approval Voting tactics. Under a realistic model, these systems perform far better than IRV with tactical voters. In fact, they perform as well or better with 100% strategic voters than IRV does with 100% sincere voters. So even if we make assumptions of voter behavior which are incredibly unrealistically biased in favor of IRV, IRV is still worse.
The only point of looking at real world data is to see whether there is in fact some kind of weird behavior going on that is intuitive, but not tactically wise, which is flawed in some way that it makes Approval Voting behave differently than we’d rationally expect. After looking at massive amounts of empirical data, we see no such pattern.
And in terms of simplicity for voters and election workers, Approval Voting is far and away superior to IRV. Plus it can be sub-totaled in precincts, unlike IRV, making it also more resistant to fraud than IRV.
Rob Richie, you and your FairVote associates are a desperate, dishonest, and reckless group when it comes to these complex issues of voting theory. You ignore empirical and theoretical analysis, and make numerous specious criticisms which don’t hold up to scrutiny. People intelligent enough to have already understood the merits of Approval Voting vs. alternatives like IRV will see through you misleading rhetoric.
Check the link, the approval voting sounds interesting, but a better system has actually been utilized in some areas. Even if you sense an approval of some candidates, you may approve of some more than others. For addressing ‘ranked’ approvals, the best solution is instant-runoff aka ‘proxy ballot’ voting. In this case you not only express those you approve of or disapprove of, but ‘rank’ the order of preference.
So, for example if you want to primarily vote for a third party candidate, but wish to express your approval for another third party candidate as a second alternative and approval for one of the two mainstream parties as a last resort, you could rank them 1, 2 and 3 respectively. Then when tabulating the votes, the counters first look at all the #1 choices from all the ballots. If the top vote-getter does not have a super majority (>50%) they take whoever’s name is on the ‘bottom’ of the list and look at those ballots to find all of their #2 choices (if any) and apply them to the totals. If the top vote-getter still doesn’t have a super-majority, they again remove the bottom candidate and look at the #2 (or #3 where appropriate) choices until a candidate has >50% of the vote. If in fact, there is not a super-majority candidate, it would require an additional election (or some other pre-determined alternative such as reducing the pool until there were two candidates remaining and awarding the office to the highest percentage)
Also on term limits, I have been hesitant to support term limits as technically it is a restriction on voter choice. But, at the same time, the founding fathers did not create the representative system to be a mainstay of career politicians. I thought on this and came up with a different alternative - no consecutive terms. Allow any candidate to be a career politician if they so desire, but do not allow ANY candidate to hold the same elected office for more than one sitting term. Senators would have to either go back to a real job or run for a congress seat, or a state office, etc. if they wanted to be full-time politicians.
This would address the ‘choice’ issue allowing the electorate to re-elect senators or congress people they liked, but would prevent the entrenched incumbency that has been so problematic. And best of all, it might get some of these fossilized has-beens out of politics for a while forcing them to get real jobs!
See the attached link for more details on these two ideas as well as a few other budget and representative salary ideas (especially in the comments section)
Actually, I already addressed the problems with IRV in the post above. It’s a nightmare to decipher and the results themselves come out of nowhere.
Post new comment