If you could replace the US "first past the post" voting system with a better one, which voting methodology would you choose, and why?

google "A Government ...

New Zealand has had MMP since 1996 and is currently undertaking a review: http://www.mmpreview.org.nz/ Some of the issues vexing voters there are the threshold (in NZ, a party is proportionally represented if it wins one electorate seat - decided by a traditional First Past the Post vote - or gains 5% or more of the total, nationwide, vote). Whether it should be raised, lowered or even abolished is one matter under review.   Seemingly of less importance to voters but an issue I see as a key flaw of NZ's system is the ranking of party lists. Parties are not required by law to follow any democratic process in ranking their lists; that in itself is a flaw, but worse still is that there's nothing to stop a party telling the public that their list repesents the votes of hundreds of party members when in fact it is drawn up in secret on the basis of the personal likes, dislikes, and even romantic attachments of a small cabal associated with the leader.   For a small party with all, or a majority, of List Members this means MPs sit in the Parliament having been "voted in" by just two or three people. Some argue that this is acceptable because the Party Vote is just that - a vote for a party, not specific people. My response to that is to question why, when the process could be democratic, it isn't - there seems no valid excuse.   In NZ at least, political parties don't exist in any recognised form. They are not associations in the way a club must be, nor are they businesses, trusts or anything else. They might establish those bodies for a specific purpose - to raise funds or to hold property, for instance - but the organisation which selects candidates and runs them for election and sits in the Parliament is seen, in law, as just a list of people who've opted to be called party members.   That has the handy result - for the parties, though not the democratic process - of meaning they cannot be called to account in the courts. So, like David Hood, I'd like to see some serious modernisation of the political system before worrying about who was elected. And that would include making political parties accountable through the law to members and others, and setting minimum levels of democratic behaviour (via electoral law) on any body which wished to be known as a political party.   When MMP was introduced in NZ it was in a run-off against Supplementary Member and STV. At the time I was on radio and advocated strongly for STV as the best of those choices at converting the wishes of the greatest number of people into a representative Parliament. Having studied it further I, like Jameson Quinn above, favour STV with multi-member districts.   However I'm drawn to the idea of an acquaintance, Bruce Simpson, called Recoverable Proxy: http://www.aardvark.co.nz/rproxy.shtml While it has elements of recall, it works differently, as Bruce explains: an MP representing an electorate of 35,000 voters may find that 20,000 of those voters have chosen to recover their proxy, leaving that MP with just 15,000 proxy-votes able to be cast. The 20,000 voters who recovered their proxies will be able to cast their own votes to be counted along-side the remaining proxies exercised by the MP. With the technology available now, the concept of direct democracy is theoretically viable. We could, if we wished, exercise our votes issue-by-issue, individually, in a vast "online Parliament". Most of us, however, don't want to; we're happy with representative democracy. Recoverable proxy offers the opportunity for those who want to exercise their right to vote directly to do so; for others to choose a representatives once per cycle and then play no further active role, as they do now; and for "rogue" MPs to have their influence reduced or eliminated without necessitating the by-election that occurs after a successful recall.   I awed by the knowledge of various systems displayed by other commenters, and await your reactions to Bruce's idea with interest.

For single-winner elections, I would implement Score Voting (aka Range Voting), or even its simplified 0-1 scale variant known as Approval Voting. These systems are both simpler for voters and election administrators than essentially any ranked method. They also achieve better average voter satisfaction than most other systems, according to an objective economic measure known as Bayesian Regret (see http://ScoreVoting.net/BayRegsFig.html). Condorcet methods (e.g. Schulze) have competitive Bayesian Regret performance, but are much more complicated. We use a comparatively simple ranked system here in San Francisco, known as Instant Runoff Voting (aka Alternative Vote, which is the single-winner form of Single Transferable Vote) — and virtually no one understands how it works. Not even smart software engineers I've worked with and polled. You can see lots of evidence about this here: http://www.electology.org/approval-score-sf#TOC-Intuitive-understanding The Berlin branch of the German Pirate Party uses the Schulze form of Condorcet, and there have been complaints about its complexity. There was this resolution, which nearly passed, to replace it with Approval Voting: http://lqpp.de/be/initiative/show/1377.html A Pirate Party member summarized it to me as follows: that thing says, they do not want schulze any more, because they dont want to explain it to new people over and over again. most members of the pirate party berlin (as far as they take part in LQFB) would like to use approval voting in elections within the party. So we certainly don't expect that there would be good transparency if Condorcet or other complex ranked systems were adopted for general government elections. Whereas Score Voting is immediately intuitively obvious to most people. Here's a small Score Voting exit poll I conducted in 2006 in the fairly blue collar town of Beaumont, Texas. It seems pretty clear that people understood it. http://ScoreVoting.net/Beaumont.html For multi-winner elections, I would implement a system of Proportional Representation. My preferred systems would be largely based on this analysis, by Warren Smith: http://ScoreVoting.net/WarrenSmithPages/homepage/multisurv.pdf From best to less good: Asset Voting - http://ScoreVoting.net/Asset.html Proportional Score Voting - Candidates are rated and we elect the n candidates with the greatest sum of scores where we may only count one score per ballot (the highest score for any of those n candidates). Single Transferable Vote - Needs no introduction. Riddled with problems, but still pretty good at picking proportional multi-winner outcomes. Reweighted Range Voting - http://ScoreVoting.net/RRV.html Clay Shentrup San Francisco, CA The Center for Election Science

This is easy... For Congress: MMPR - I moved from the US to South Africa 15 years ago and am really impressed with how the MMPR system works here.  People complain about the government here too, but it's not nearly as dismal as the discussion around Congress in the US.  MMPR does a brilliant job of incorporating the broad spectrum of society in Parliament while representing the "median" citizen.  I often wish we had the same system in the US and think things would really improve if they did. For President:  http://en.wikipedia.org/wiki/Schulze_method - Since you can only have one winner, this provides an elegant method of selecting the least objectionable option.  It's a bit math-y, so it may be harder for people to get their heads around, but if you look at the results and method in detail, it all makes sense.  Wikipedia uses this for internal decision making and it has worked out well for them according to my understanding. There are plenty of good Condorcet methods out there, but this is my favorite.

The criteria that are important to me are: Reduce conflict between strategy and sincerity.  If voting in the way that is most in your interest feels dishonest, that is a bad thing.  Although all voting systems can be "gamed" to some extent, some of them make it extremely hard to get any significant advantage by doing so. Reduce strategic value of parties.  Our present system forces people into parties, otherwise similar candidates would split the vote and cause a huge disadvantage.  This results in polarized government, and a polarized constituency. Increase tendency to elect centrist candidates. Elected officials that are centrist...that is, those at the center of the ideological bell curve...have the best chance of serving the will of the people, and of cooperating well with other elected officials to actually get things done. Reduce surprises on election day.  Surprises are fun with sports, but elections are too important for this.  We don't want a government that jerks us back and forth between right and left, we want one that drives right down the middle of the road with only the tiniest and smoothest of corrections. That said, I prefer ranked choice systems, with ones that meet the Condorcet criterion being the best.  The Schulze method is possibly the best of these:  http://en.wikipedia.org/wiki/Schulze_method  As notes in the comments, Majority Judgement and SODA would work quite well too.

I presume I was invited to answer this question because I've built a series of ranked choice balloting sites (http://AmericanQuorum.com, for example) that support visualizations of Borda, IRV, and Condorcet results. Those sites didn't generate much interest, but I continue to believe that the call for a better system than "first past the post" is bound to be taken seriously some day... hopefully sooner rather than later. Jameson Quinn has already provided a lot of information about many of the alternative voting systems that might be considered. If the goal is to build better tools for better democracies, any kind of decision making process refinement that helps a community engage in efficient problem solving and broader consensus building is of course a Good Thing.  But sustaining and fortifying the legitimacy of the process, whatever it may be, should also be a priority. The bigger problem in the US right now is how often the margin of error in the ballot counting of an election exceeds the margin of victory. I'd be reluctant to mandate a transition to some sort of ranked-choice, multi-option ballot unless much more is done to raise confidence in the accuracy, speed, and auditabllity of existing registration, voting, and counting processes. That said, I believe it's important to keep experimenting with online RCV for the sake of improving UX, and also in order to build up a knowledge base of how to explain these more advanced systems to the voting public. Condorcet is very impressive, but I'm a Ph.D. and it took me a while to get it. Folks will rightly refuse to adopt a new system if they think it's too tricky. That's a key reason why I think development of visualization tools is so important.

I would reform our system with proportional representation.  I don't think there is any system that can both not be gamed and that gives a voice to minorities as well as majorities. Promoting diversity in representation is the path I would find most favorable, and least disruptive to the existing system. Watching the Pirate Party obtain some access to government is intriguing, simply because similar efforts here generally fail.   Our current system can be better leveraged through movement politics, either vote one party, no vote allowing others who do to make the call, or vote the other party, right? There is also taking over a party, which does require a sustained and fairly resonant civic effort.  Timelines on this are long however, and that undesirable element is why I would favor proportional representation. As far as basic mechanics, I do not favor electronic recording of the vote, preferring the chain of trust between voter and vote cast be intact.  That's only possible with physical media and a human readable vote record.  The population can always count it's votes too, but using machines for that is ok, given we have solid audits, and a chain of evidence process that insures the record of the vote cast is accessible in it's entirety for said purposes. Public financed elections on fixed dollars, and the removal of the provision for unlimited special interest group spending would be welcome additions to a proportional representation system.

I preface this answer with the disclaimer that I think some serious modernization needs to happen to the U.S. government. Before I changed the actual voting system I would push for certain reforms to open the electorate up to be more answerable to voters and modernize certain offices: Overturn the Electoral College all together in favor of direct popular vote for federal elections with a required threshold of 50% to win I would then lift the 435 member cap for the House and base each seat on a firm 200,000 person districts (that means even one person over that number would trigger a new seat) which would force the government to address actually representing local/regional concerns; which is what the body is suppose to do anyway. For me with a country the size of the U.S. having a "peoples house" small enough to fit in one room is a crime. Require that all redistricting be done via an independent commission similar to Iowa's I would get rid of the Vice President in favor of Primer Minister elected by the joint House/Senate from one of its members to serve the traditional role of the Vice President. This would further force coalition building, and provide a stronger check against the President. I would make the Parliamentarian a permanent role in Congress and to be a person nominated by the MP, and confirmed by majority and minority leaders. I would limit all members of congress to two terms in each chamber. Then I would put in Alternative Vote Plus It is important to put a system in that is familiar so maintaining single-member constituencies is important while at the same time introducing new ideas into the voters brains like coalitions. Alt+ also has the potential of killing the notion of a "spoiler"  which is a huge barrier to having more proportional voting in the U.S. Alt+ also does a good job of keeping out fringe parties (like say a KKK Party) while at the same time also makes it hard to have safe seats; its a good balance. Couple that with a much larger lower House that would likely have to set up regional chambers it would be a huge breath of fresh air into a very stale system. The wiki article is helpful if you want more info: http://en.wikipedia.org/wiki/Alternative_vote_Plus

Preferential voting is a good way to get the most preferred candidate in an election. If the US had used a preferential voting model in past elections, people could have voted for Ralph Nader and Ross Perot and then chosen which of the other candidates received their second preference, which means their vote wasn't wasted.