Changing the American election system

[Squall83]

Hello everyone,

first there were SOPA and PIPA, then came CISPA, then ACTA and then NDAA. No matter how many people are protesting, the government just keeps on trying to put through some law that disrespects individual rights. It looks like the people can only defend (but not attack) against a lobby-controlled government and I believe that it doesn't really matter whether the democrats or the republics are in charge.

In my opinion, the core problem is one thing: The winner-takes-all-principle that happens during elections.
You see, here in Germany we have a 5% hurdle, i.e. any party that gets at least 5% of all votes gets to be in the parliament. That's why we now have a pirate party – the first party that actually knows about internet and how to go about this topic properly – which also thoroughly shakes up the other politicians, and I hope they'll be more successful than our Green (environmental) party 20 years ago.

But if you have a winner-takes-all-principle, new parties have no chance to enter the government unless they get more votes than any other party in a state. That's a goal which neither our Green party nor the pirates have reached at any point in time for the parliament, and they're my 2 favorites.

Therefore I propose that the US protest against this winner-takes-it-all-principle and then starts to create new parties that haven't been completely corrupted, instead of just fighting law after law.
In my opinion, every party that gets enough votes in one state to get 1 seat should be allowed to keep that seat.

What do you think about this? Is there already a movement like this? I read that there are only 2 states that don't do winner-takes-all and imho that's not enough.

Btw: We also have local governments for each federal state (e.g. Bavaria) as well, all of which have their elections at different points in time, so a new party doesn't have to wait 4 years to be seen somewhere in a government. Therefore our pirate party – which didn't make it into the Germany-wide parliament in 2009 – can be seen in a few local parliaments already, being consistently at 8-9%. So we can react quicker if a party doesn't do what we want it to do. Do you have some means of interacting with the government in between a 4-year-period as well? I heard Obama had a minority after 2 years, but I didn't really read up on what actually happened and I don't know what to Google to find out.

Squall83

[Dr. Diaphanous]

In the UK there was a referendum last year to change from first-past-the-post to alternative vote (instant runoff voting). That's probably not as extreme as what you're proposing, but might go some way to give hope to smaller parties and prevent tactical voting.

It got defeated 70-30.

[Squall83]

Yes, that change is less radical, that's true.

In Germany, everyone has 2 votes. One vote goes to a person that is a candidate for an electoral district and the 2nd one goes to a party. The 2nd vote determines the minimum amount of seats a party gets. Those seats are filled beginning with persons who got the most 1st votes or - if none of those remain - members from a list the party made. If there are more 1st-vote-winners than seats, the party gets extra seats.

The thing is, in more than 99% of all elections, 1st-vote-winners come from one of two parties (CDU and SPD), which are similar to rebublics and democrats. The smaller parties only get 2nd votes. That's why I don't like election systems where you elect persons only, because votes for other people go to waste. If our voter's 2nd votes go to e.g. 32% CDU, 28% SPD, 10% Green, 9% Pirates, 7% FDP, 6% Lefties and 8% other then that's how the seats are distributed (except for the 5% hurdle and possible extra seats from the 1st-vote-winners). So almost no votes go to waste.

Well, even though we have that system we still have a lot of corruption, but recently a lot more demonstrations happened all around the world, petitions were signed and people are starting to actually care for politics and take responsibility for their land's fate and I believe that lands with a system that doesn't let votes go to waste has better chances of freeing themselves than systems where persons win an election and all votes against that person go to waste.

[Ormurinn]

In the UK there was a referendum last year to change from first-past-the-post to alternative vote (instant runoff voting). That's probably not as extreme as what you're proposing, but might go some way to give hope to smaller parties and prevent tactical voting.

It got defeated 70-30.

Though that has more to do with the massive barrage of tory-funded propaganda and lies that flooded almost every media channel, than genuine hostility to the idea.

[jianmin]

After reading up on Germany's election system, I'm somewhat confused on how yours differs from ours in the US except that you vote for a regional representative *and* a party.

How do you reconcile the difference between the regional vote and the vote for the party? What if nobody from party X wins in local districts, but on the national level 15% voted for party X?

Do you force some districts to be represented by someone who didn't get the majority of the votes in that district? Or do you have a separate group of slots in your legislative chamber that are selected by the party vote? (Or am I completely off?)

[Diadem]

After reading up on Germany's election system, I'm somewhat confused on how yours differs from ours in the US except that you vote for a regional representative *and* a party.

How do you reconcile the difference between the regional vote and the vote for the party? What if nobody from party X wins in local districts, but on the national level 15% voted for party X?

Do you force some districts to be represented by someone who didn't get the majority of the votes in that district? Or do you have a separate group of slots in your legislative chamber that are selected by the party vote? (Or am I completely off?)

Not an expert on German election law, but if I'm not mistaken:
Half of the seats (299 out of 598 in total) in the Bundestag (The German parliament, which has only one house) are elected via district elections. One member from each district, whoever gets the most votes. The other half are elected from party lists. You vote for a district to determine who represents your district, and then for a party. The party-seats are filled in such a way to make the total Bundestag representative to how people voted for parties. So if party A got 30 districts and 20% of the popular vote, then they'd get 20% * 598 = 120 seats in total, but they already had 30 seats from districts, so they get 90 of the remaining 299 seats

If a party has more seats from districts than it would deserve based on popular vote, then no districts are taken from it, but rather extra seats are added to the Bundestag to make the distribution proportional. Wikipedia says the current Bundestag has 622 seats, so they seem to have added 24 seats in this way.

I don't know what happens if one party got 100% of districts and another party got 100% of votes. Presumably they'd add an infinite number of seats to the Bundestag

[Dr. Diaphanous]

I think CGP Grey on YouTube has some pretty good videos explaining voting systems.

Problems with first past the post
Alternative vote
Mixed-member proportional representation (As in Germany)
The US electoral college

[Squall83]

After reading up on Germany's election system, I'm somewhat confused on how yours differs from ours in the US except that you vote for a regional representative *and* a party.

How do you reconcile the difference between the regional vote and the vote for the party? What if nobody from party X wins in local districts, but on the national level 15% voted for party X?

Do you force some districts to be represented by someone who didn't get the majority of the votes in that district? Or do you have a separate group of slots in your legislative chamber that are selected by the party vote? (Or am I completely off?)

Not an expert on German election law, but if I'm not mistaken:
Half of the seats (299 out of 598 in total) in the Bundestag (The German parliament, which has only one house) are elected via district elections. One member from each district, whoever gets the most votes. The other half are elected from party lists. You vote for a district to determine who represents your district, and then for a party. The party-seats are filled in such a way to make the total Bundestag representative to how people voted for parties. So if party A got 30 districts and 20% of the popular vote, then they'd get 20% * 598 = 120 seats in total, but they already had 30 seats from districts, so they get 90 of the remaining 299 seats

If a party has more seats from districts than it would deserve based on popular vote, then no districts are taken from it, but rather extra seats are added to the Bundestag to make the distribution proportional. Wikipedia says the current Bundestag has 622 seats, so they seem to have added 24 seats in this way.

I don't know what happens if one party got 100% of districts and another party got 100% of votes. Presumably they'd add an infinite number of seats to the Bundestag

That's correct.

If we didn't have the vote that goes to the party we'd most likely have only 2 parties in the Bundestag, like America, because everyone who votes on a loser essentially wastes his vote and so it's risky to vote on anyone but the 2 biggest parties.

Concerning the infinite seats: As you said: 299 seats go to district elections, so that'd be 299 extra seats.

[Derek]

I despise party list systems. It makes the candidates primarily responsible to the party to retain their position on the party list, and not to the electorate. I would much rather have the US two party system with direct elections of all seats than a proportional system with no direct elections.

There was a pretty long thread a few weeks ago about better voting systems for direct elections than FPTP.

[Dr. Diaphanous]

There was a pretty long thread a few weeks ago about better voting systems for direct elections than FPTP.

This?

[Silknor]

first there were SOPA and PIPA, then came CISPA, then ACTA and then NDAA. No matter how many people are protesting, the government just keeps on trying to put through some law that disrespects individual rights.

Of which I believe just one (NDAA) has passed.

I read that there are only 2 states that don't do winner-takes-all and imho that's not enough.

You're thinking of the electoral college (which weighs and aggregates the state-level vote in determining who wins the presidential election). Maine and Nebraska give one electoral vote to the winner of each congressional district, and two to the state-wide winner. Every Congressional election, as far as I am aware, is filled by the plurality winner of a single-member district (there are different primary systems though, for example California is switching to a "jungle primary" where the top-two, regardless of party, go to a heads up runoff, unless someone gets a majority in the first round).

Do you have some means of interacting with the government in between a 4-year-period as well?

The entire House of Representatives is elected every 2 years, as is a rotating 1/3rd of the Senate.

2 broader points though:
1. Are you so sure that these policies are in fact unpopular? There's a big difference between unpopular somewhere like reddit and actually unpopular with likely voters. I don't have any evidence one way or another on this question, I looked for NDAA polling and found just one poll, but it was online and more than 1/3rd had no opinion, so I'm not inclined to take it seriously. It did have NDAA at 23% support and 38% against, though. Link.

2. Assume you're right that the Democrats and Republicans are controlled by corporate lobbyists and are hostile to individual rights (a claim which I do not believe to be accurate, though I might be over reading into what you said in your first post). Why would replacing 10-15% of those Members* with, say, Pirate party, members make a difference?

*Granting here that you'd expand the number of Senate states per state along with this, because two-member proportional districts most likely wouldn't elect any third parties.

[WilCSUN]

Speaking as a citizen of the United States, I think that our Electoral system is far from perfect.

It's true that first-past the post is is extremely hostile to third parties, and Duvenger's law posits that it is the natural course that all FPTP system lead to a two-party system.
We have some significant problems when it comes to gerrymandering, which we usually try to solve with bi-partisan committees, but as CPG Grey so elegantly points out, leads to uncompetitive elections, which has been probably the biggest criticism of American democracy that I've heard.

http://en.wikipedia.org/wiki/Gerrymandering#United_States



It's also becoming clearer every election cycle that the Electoral College doesn't really fulfill any purpose other than making election day easier to follow on the major news networks and mucking up the results every once and a while.... sigh... I wonder what President Gore would have been like...

The reason I don't think it will change though is that,
1. Democrats and Republicans have a monopoly (oligopoly?) on political power in the United States, and from their point of view, the Electoral system is working just fine.

2. This would be kind of a radical re-haul of Article 1 and Article 2 of the Constitution, and in American politics, everybody gets really touchy when you bring up any alterations to the Constitution (although we have altered it before with the 17th Ammendment which stipulates the direct election of Senators).

3. For the most part, our elections carry on just fine. We have the occasional hiccup like in 2000 (My fellow Democrats are still kind of getting over that one) 1888, and and 1876, but usually, whoever wins in the Electoral College wins the popular vote, and something like a 5% error over how ever so many decades just doesn't get people motivated enough to change it.

That's just my two cents though.

[aldonius]

2. Assume you're right that the Democrats and Republicans are controlled by corporate lobbyists and are hostile to individual rights (a claim which I do not believe to be accurate, though I might be over reading into what you said in your first post). Why would replacing 10-15% of those Members* with, say, Pirate party, members make a difference?

*Granting here that you'd expand the number of Senate states per state along with this, because two-member proportional districts most likely wouldn't elect any third parties.

My take (I know, opinion with no backup, sorry SB) is that it would narrow each party's Overton window a bit. There are two distinctive types of pandering by the politicians that I see - in Australia, it tends to be more to the apathetic outer suburbanites, and from what I see of the USA, it's more to the party base.

Since the party base tends to be further from political 'centre' than the average person who simply voted for that party on the day (and even further from the average overall), we see the problems Romney (and even more so, Huntsman) have faced in the Republican primaries. The response to those is to veer right-wing-wards to get the nomination and then back toward centre for the general election. You yourself noted this (on the liberal side):

There's a big difference between unpopular somewhere like reddit and actually unpopular with likely voters.

In the general case we have a candidate/party representing an arbitrary range of ideology. It follows that there is an optimal magnitude of said range. I contend that in the specific case of the US, the candidates/parties try to represent overly wide ranges (base + edge groups vs mainstream) in the leadup to election, but due to pragmatism/party discipline/lobbying/sheer bloodymindedness they ultimately only represent a small range of positions when in Congress. The edge groups can't effectively split off because of a) spoiler effect (arguably Ralph Nader),or (to a lesser extent and not FPTP specific) b) lose attention. So they clump awkwardly with a major party (e.g. the New Party, the Ronpaul).

Note also that in the case of closely balanced major parties, the minor parties/independents can stir things up in relatively constructive ways, because now they have to talk policy in addition to politics.
I'm particularly fond of independents because they can simply be for their constituency and not be beholden to the party machine.

TL;DR: With first-past-the-post, what should be smaller edge groups a) act as spoilers or (to a lesser extent and not FPTP specific) b) are ignored or c) are either clumped somewhat awkwardly with a main party.
So those edge groups make the primary process interesting, but they distort the public discourse around the mainstream of the group. Rather, they need to be represented separately, *in Congress*.

[Squall83]

@Silknor:

Why would replacing 10-15% of those Members* with, say, Pirate party, members make a difference?

Our pirate party made a difference even when they had 1-2%, because they were telling the politicians what they were trying to do to the internet.
Also through petitions people are making a difference with 0%. Therefore I believe it will make a difference indeed.

2. Assume you're right that the Democrats and Republicans are controlled by corporate lobbyists and are hostile to individual rights (a claim which I do not believe to be accurate, though I might be over reading into what you said in your first post).

Yes, I believe that in a party system with only 2 parties can be controlled way more easily than a system which has 5 parties, especially when there's a pirate party that wants to do everything through internal democratic votes.

Oh and I forgot to mention: The 1st time Bush was elected, actually less than 50% of all Americans wanted him. He just focused on the big states. In theory, you can win an election by getting 25,1% of all votes if these votes let you steal another 25%.

@WilCSUN:

1. Democrats and Republicans have a monopoly (oligopoly?) on political power in the United States, and from their point of view, the Electoral system is working just fine.

Yes, the politicians probably don't want to change that system easily. You just have to demonstrate long enough. If you feel that a system has to change, just spread that opinion and if the majority agrees something will happen, unless you're in China where everyone just gets shot for having an opinion.

[Qaanol]

This topic is quite similar to the thread An election system. That thread includes a substantial discussion of single-winner systems. I am nearly done with my election simulator program, and it would be a shock if the results were anything other than what we all concluded in that thread.

Namely, for single-winner elections, the Range Voting and Majority Choice Approval systems are the best, with Approval Voting very close behind. For practical purposes, especially in light of how people actually vote, Approval Voting makes the most sense right now. It is by far the simplest alternative system, and it uses the exact same ballots we already have. (Approval is actually the simplest form of both Range and MCA, and to get any theoretical benefit beyond that, more complicated ballots would be needed.)

The way Approval Voting works is, you vote for the candidates you approve, and against the rest. Whoever gets the most “yes” votes wins. You can think of this as treating each candidate as a separate question, “Do you approve of this person for the job?” That means the candidate with the highest approval rating wins. Furthermore, everyone gets the same number of votes: one per candidate, either for or against.

Approval Voting solves the problem of the vote-splitting “spoiler effect”, because there is never a reason to vote against the candidates you like. Unlike many voting systems, with Approval Voting the best “strategy” is to vote sincerely. Another effect is that politicians have to appeal to a wide base in order to get elected.

For the United States, the Approval Voting method is good for Democrats, because it prevents a repeat of 2000 when Nader was a spoiler. It is good for Republicans, because it ensures they will never suffer from that same type of vote-splitting at the hands of a Libertarian for example. And it is good for third parties and independents, because it lets people vote for them without fear of those votes being “wasted”.

Thus, Approval Voting is beneficial to all political organizations. It is also good for the country as a whole, because it makes election results reflect the true will of the people. All votes count equally, so it is fair to everyone. And the switch to Approval Voting is extremely simple, just a matter of making the instructions say, “Vote for the candidates you approve.”

Any sort of proportional representation system would require a constitutional amendment in this country, because the constitution defines how senate and congressional seats are assigned per state. However, the constitution also says each state can decide the manner in which it holds elections. That means, any given state legislature could pass a law to use Approval Voting in that state, including for Senators and Representatives. And since each state legislature decides how its presidential Electors are chosen, states can even use Approval Voting for President.

I don’t know if proportional representation can do better than Approval Voting at electing the best will-of-the-people candidates (minimizing Bayesian regret), because I have not studied proportional multi-winner systems. But even if one of them could do better than Approval Voting, the US would need a constitutional amendment to use it. Approval Voting is a vast improvement over our current election system, extremely simple to switch to, and just as easy to use as our current system.

[Silknor]

Approval Voting solves the problem of the vote-splitting “spoiler effect”, because there is never a reason to vote against the candidates you like. Unlike many voting systems, with Approval Voting the best “strategy” is to vote sincerely.

Solving the spoiler effect with regards to minor parties is not the same as creating a system where the best strategy is to vote sincerely. Take as an example a liberal state, where 20% of voters are Republicans. The liberal side is split between Democrats and Greens, who we'll say a quite similar. It should be obvious that for liberal not entirely indifferent between the two, that they should approve only their first preference, even though they "approve" of the other.

So yes, there's never a reason to vote against a candidate you "like" and the best strategy is always to vote "sincerely." But this is only true in the strictly rational sense! It's no longer true once you use those terms with their general as opposed to technical meanings. There is a major disconnect between the rhetoric as understood by the general meaning, and the empirically supportable claims that supporters make.

Also I have to disagree with the idea that "Approval Voting is beneficial to all political organizations. It is also good for the country as a whole." It's clearly not good for everyone. Lets grant that you're right that AV is "fair" to everyone. Regardless of if this is true, the clear (and correct) implication, is that the current system is unfair, ie. that some are hurt by it and some benefit from it compared to AV.

Why might say, the Democrats, want to have all of the Representatives from liberal districts be Democrats instead of mostly Democrats and a few Greens, with 0 possibility of a spoiler effect electing a Republican? I can think of a few, and I'm sure there's more:

1. It makes the representatives more dependent on the party leadership. In this case, you'd expect that to mean the Would-Be Greens (WBGs) to be more loyal, to vote closer to what the party leadership would like, and more willing to compromise with the rest of the party.

2. It makes the party leadership/apparatus more powerful. Given that we are a government of people, not angels, the appeal of this to those in party leadership should be self-evident.

3. It keeps the public discussion closer to the preferred issues of the party leadership/core party members (both public and representatives).

I'm not passing judgement on any of those, simply stating that a switch to Approval Voting, even ignoring the problems with the voting system itself, is not Pareto-optimal. Someone loses, and it matters who. Supporters of Approval Voting do themselves no favors when they pretend otherwise.

[JohnSmith1]

The 1st time Bush was elected, actually less than 50% of all Americans wanted him. He just focused on the big states.

16 US presidents won with less than 50% of the popular vote, including Lincoln and Kennedy, but please don't let the facts stand in the way of your Bush bashing.

In theory, you can win an election by getting 25,1% of all votes if these votes let you steal another 25%.

This is true for all representative democracies, not just the electoral college. If you want to keep democracy, the only other alternative is direct democracy, where everyone pretends to spend a few hours everyday reading bills.

[Qaanol]

Approval Voting solves the problem of the vote-splitting “spoiler effect”, because there is never a reason to vote against the candidates you like. Unlike many voting systems, with Approval Voting the best “strategy” is to vote sincerely.

Solving the spoiler effect with regards to minor parties is not the same as creating a system where the best strategy is to vote sincerely.

Correct, but Approval Voting does both.

Take as an example a liberal state, where 20% of voters are Republicans. The liberal side is split between Democrats and Greens, who we'll say a quite similar. It should be obvious that for liberal not entirely indifferent between the two, that they should approve only their first preference, even though they "approve" of the other.

Not only is that not “obvious”, it is flat-out wrong. Suppose the same candidate were running in FPTP: some voters who prefer greens would insincerely vote for the dems. Now switch back to Approval Voting, and consider those strategic voters who compromised to prevent a bad outcome: do you think they are going to suddenly switch to voting only for the greens, when they didn’t even do that under FPTP? Of course not! They are going to vote for both greens and dems.

So yes, there's never a reason to vote against a candidate you "like" and the best strategy is always to vote "sincerely." But this is only true in the strictly rational sense! It's no longer true once you use those terms with their general as opposed to technical meanings. There is a major disconnect between the rhetoric as understood by the general meaning, and the empirically supportable claims that supporters make.

FPTP, and every system where you have to rank candidates from best to worst, necessarily has situations where you as a voter can see that your best strategy is to betray one or more candidates you like a lot, in favor of some candidate(s) you think are worse. That is what “insincere” voting means.

The strictly rational strategy is to estimate how likely you think each candidate is to win, and how much you agree with each candidate. Multiply those together for each candidate, add them all up, and that gives yours expected-value outcome for the election. A perfectly rational actor will vote for all candidate he or she likes better than that expected outcome.

Also I have to disagree with the idea that "Approval Voting is beneficial to all political organizations. It is also good for the country as a whole." It's clearly not good for everyone. Lets grant that you're right that AV is "fair" to everyone. Regardless of if this is true, the clear (and correct) implication, is that the current system is unfair, ie. that some are hurt by it and some benefit from it compared to AV.

Major parties: Approval Voting makes sure they do not unfairly lose elections from the spoiler effect.
Minor parties: Approval Voting makes sure their supporters can vote for them without wasting their votes.
Quod erat demonstrandum.

Why might say, the Democrats, want to have all of the Representatives from liberal districts be Democrats instead of mostly Democrats and a few Greens, with 0 possibility of a spoiler effect electing a Republican? I can think of a few, and I'm sure there's more:

1. It makes the representatives more dependent on the party leadership. In this case, you'd expect that to mean the Would-Be Greens (WBGs) to be more loyal, to vote closer to what the party leadership would like, and more willing to compromise with the rest of the party.

2. It makes the party leadership/apparatus more powerful. Given that we are a government of people, not angels, the appeal of this to those in party leadership should be self-evident.

3. It keeps the public discussion closer to the preferred issues of the party leadership/core party members (both public and representatives).

The problem you are describing is Gerymandering, and it affects every single-winner election system. No matter what voting system you use, there is always some way of drawing the district boundaries that helps one party and hurts another. This has nothing to do with the specific voting system.

For any given set of electoral districts, however, major parties want to make sure they don’t lose to spoiler effects, and minor parties don’t want to be wasted votes.

I'm not passing judgement on any of those, simply stating that a switch to Approval Voting, even ignoring the problems with the voting system itself, is not Pareto-optimal. Someone loses, and it matters who. Supporters of Approval Voting do themselves no favors when they pretend otherwise.

Approval Voting is much more Pareto efficient than FPTP—indeed more so than any single-winner system except for Range and MCA, which are just more complicated versions of Approval Voting.

Specifically, anytime a candidate gets elected who is not the best representative of the people, there is deadweight loss. Real, actual, economic and social harm. The benefit to those who benefit, minus the harm to those who are harmed, is a smaller value than it could have been by electing the most-approved candidate.

That happens a lot in FPTP, and much less in Approval Voting. Every other single-winner system (besides Range and MCA) is less efficient than Approval Voting.

[Silknor]

Suppose the same candidate were running in FPTP: some voters who prefer greens would insincerely vote for the dems. Now switch back to Approval Voting, and consider those strategic voters who compromised to prevent a bad outcome: do you think they are going to suddenly switch to voting only for the greens, when they didn’t even do that under FPTP? Of course not! They are going to vote for both greens and dems.

I don't think you read the example. 20% of voters are Republicans in this district. By construction there is no spoiler effect under FPTP, and no reason for voters who slightly favor D over G or vice versa to approve both. In this situation it is rational to vote for one. This highlights the disconnect between the rhetoric "just vote for who you like!" and the reality. If it helps, imagine that the Republicans aren't there: you have a bunch of liberal voters, some of whom are indifferent between Dems and Greens, some who slightly prefer Dems, and some who slightly prefer Greens. It's clearly rational, as long as utility is a function of who is elected alone, for those with a slight preference to approve only one candidate, even if the rhetoric suggests they should approve both.

The strictly rational strategy is to estimate how likely you think each candidate is to win, and how much you agree with each candidate. Multiply those together for each candidate, add them all up, and that gives yours expected-value outcome for the election. A perfectly rational actor will vote for all candidate he or she likes better than that expected outcome.

Lets apply this to the above example, for Voter Alfred.
U(Green): 101
U(Dem): 100
U(Rep): 0
P(Green): .4
P(Dem): .6
P(Rep): 0

Expected utility=100.4

Alfred votes for Green only, as I said above. This can obviously be true even if P(Rep)>0, though it has to be low if U(Green) is not notably greater than U(Dem).

Major parties: Approval Voting makes sure they do not unfairly lose elections from the spoiler effect.
Minor parties: Approval Voting makes sure their supporters can vote for them without wasting their votes.
Quod erat demonstrandum.

I will treat this as a serious claim for the purpose of illustration, though I hope it wasn't. Approval voting is better for some precisely because FPTP benefits others. You can qualify that last part as unfairly benefiting others, and in many cases I would agree.

When a spoiler happens, it is silly to pretend that it is a pure loss for major parties. Rather it should be evident that one party gains substantially! Nader's spoiler in Florida hurt Gore/Dems, but it helped Bush/Reps. A simple model which ignores perceived legitimacy would suggest this spoiler effect is zero-sum: any losses by one major party are reaped by the other, I don't intend to defend that model, but do point out that from the perspective of major parties, it should be roughly zero sum if each side is equally likely to be spoiled. That is another unspoken assumption though, if one side is more likely to be spoiled, then it's clear that the other side on whole probably benefits from preservation of the spoiler effect through FPTP.

There are, as I've discussed elsewhere, other advantages to a 2-party system to those two parties as well.

The problem you are describing is Gerymandering, and it affects every single-winner election system. No matter what voting system you use, there is always some way of drawing the district boundaries that helps one party and hurts another. This has nothing to do with the specific voting system.

For any given set of electoral districts, however, major parties want to make sure they don’t lose to spoiler effects, and minor parties don’t want to be wasted votes.

What I described has nothing to do with gerrymandering, I'm uncertain how you drew that connection. The arguments above would hold just as well when considering a hypothetical Senate wherein each state has just one Senator.

Approval Voting is much more Pareto efficient than FPTP—indeed more so than any single-winner system except for Range and MCA, which are just more complicated versions of Approval Voting.

I'm not sure we're using Pareto in the same way. It has nothing to do with the overall gain or loss to society, rather it asks if everyone is better off.

Obviously there's lots of times when FPTP sucks and AV would be better. I am leaning towards thinking that, assuming a preclusion of any other voting system besides AV in single-member districts, we'd be better off with AV than FPTP.

But this does not mean that AV benefits everyone or that AV never produces situations where it's irrational to vote for people you'd be content with, something that certainly is relevant when considering if AV is superior to some non-FPTP system.

[TrlstanC]

This article makes an interesting case, based on two proposals, that a good way to make election results better represent the interests of the electorate isn't to increase the number of parties/options, but to decrease the number of people voting. If a small percentage of the electorate was picked at random to be a representative sample we would expect both turnout and the amount of time voters spend reviewing the issues to increase. It would also decrease the impact of ad spending and door-to-door campaigning, thus decreasing a campaigns dependance on fund raising.

It's certainly something I'd sign up for, assuming of course that there were protections included to keep wealthy/specialtry interests from un-duly influencing the voters (at least not anymore than they do now).

WIRED - Why Fewer Voters Can Mean Better Elections

[Griffin]

So wait, people other than me are voicing their support for a sort of electoral jury pool system?

Whoo boy, never thought I'd see the day!

[Dr. Diaphanous]

Sadly I can't see the jury pool being adopted in the near future, because
a) People fear the unknown as much as they complain about the current system.
b) People would feel they didn't get a say in the election if they were not picked (though they could still influence the result by spreading propaganda).
c) Many people would not want to spend a few days learning politics & economics (although that seems to be an issue of how they get reimbursed financially).
d) It is in the interest of those in power not to shake things up too much.

What are the actual reasons against a jury system (if any)? It seems like it would mean that those who vote really understand the issues, without taking power from any subset of the population.

[PeteP]

What are the actual reasons against a jury system (if any)? It seems like it would mean that those who vote really understand the issues, without taking power from any subset of the population.

Of the top of my head:
For one manipulation, it would probably be easier to hide a manipulated selection of voters.
Then is the question of anonymity, if it is truly anonymous who do you ensure hat the voters are actually better informed, though they might be motivated by the lower numbers.
If it isn't anonymous they will be directly catered to, and if you take the 100000 mentioned in the article, then bribery actually becomes a possibility. I found this figure „$310 million: Total spent by John Kerry in the 2004 election“ , you could give 10% of the voter 30k each or equivalent gifts. Naturally that would be illegal, but in comparison to a system with a few hundred million voters it is at least mathematically possible.
Oh and if you actually educate them about matters, their opinions won't be representative of the population — or rather only of what the population might think if they had been educated in a similar way. Which is pretty much the goal, but if this has an effect, then people will point out that the vote results differ significantly from opinion polls and question whether that is democratic.

Edit: But with a system like that more direct votes about some issues might be feasible.

[Dr. Diaphanous]

Oh and if you actually educate them about matters, their opinions won't be representative of the population — or rather only of what the population might think if they had been educated in a similar way. Which is pretty much the goal, but if this has an effect, then people will point out that the vote results differ significantly from opinion polls and question whether that is democratic.

Why is that a problem? The voters are drawn randomly so they are representative of the population, and educating them can only lead to better choices (for the country's interests, and the interests of the voter and those close to them). Only problem I can see is if the education is one-sided/brainwashing.

[Qaanol]

Approval voting is better for some precisely because FPTP benefits others. You can qualify that last part as unfairly benefiting others, and in many cases I would agree.

When a spoiler happens, it is silly to pretend that it is a pure loss for major parties. Rather it should be evident that one party gains substantially! Nader's spoiler in Florida hurt Gore/Dems, but it helped Bush/Reps. A simple model which ignores perceived legitimacy would suggest this spoiler effect is zero-sum: any losses by one major party are reaped by the other, I don't intend to defend that model, but do point out that from the perspective of major parties, it should be roughly zero sum if each side is equally likely to be spoiled. That is another unspoken assumption though, if one side is more likely to be spoiled, then it's clear that the other side on whole probably benefits from preservation of the spoiler effect through FPTP.

In hindsight, or from a viewpoint of counterfactual definiteness, yes, any given instance of a spoiler effect must benefit some party to the detriment of another. And collectively over many election cycles, the spoiler effect is likely not to be perfectly balanced among parties. But when looking forward to future elections, consistency and fairness are valuable in their own right.

There is benefit to everyone from eliminating the spoiler effect, because then each candidate knows that by running the best campaign and garnering the most popular support, they can win the election.

Imagine you are part of an NBA team, and the referees are notorious for making bad calls. On any given night, or over the course of a whole season, your team might actually benefit from those officiating errors. But looking ahead to next season, you would still prefer the referees become more accurate in their calls.

In order for some other team to actually want to retain bad referees, that team would have to acknowledge that it needs unfair help to win, and it would also need to be sure the refs are on its side more often than not. But if the refs are simply incompetent, there is no way to be certain who will benefit from their mistakes in the future.

[omgryebread]

Okay, problems with the jury system:

1) Wow, really, Wired? You didn't notice this? I know you're a tech magazine, but come on, this is a painfully obvious flaw.

Cryptographer and voting-technology expert David Chaum has developed his own elegant model—which he calls a “random-sample election”—that prioritizes the anonymity of the voter. A randomly selected voter receives a ballot in the mail and is directed to an election-commission website that features candidate debates and activist statements. This would allow voters time to deliberate in their home, without being identified publicly.

This restricts informed voting to people who have computers and internet connections. And no, libraries or other public access terminals are not a valid thing here, as that's inconvenient, public (anyone who sees you at the website could bug you, and a lot would), and out of the question for poor rural people (who are already the least likely to have decent internet.)

2) Random selection does not always produce accurate results. It's a hotly debated topic in polling, but everyone at least agrees that simple random selection can produce some very unrepresentative results. This is fine in polls, since no one serious uses one poll, and averaging many polls will get you a more accurate result, and since polls don't actually matter. Juries are also different. They are not random, since lawyers have significant influence on the juries. Juries also require unanimous decisions. This system lacks the safeguards of the legal system, and elections are hugely important. If polls get it wrong, oh well. If elections get it wrong, that's a really big deal.

3) Why is our current election system broken? Because people aren't voting? Compulsory voting, then. If people don't vote now, why would you assume they would want to put more time and effort into voting and would do it under the new system? Or are you concerned that people are voting ignorantly? Why would you assume they would be less ignorant if they were part of a smaller sample? Paying them wouldn't change much, unless you had some sort of test, but polling tests are extremely bad.


4)It allows for marginalization of blocs. If a senator casts a vote that say, overwhelmingly hurts blacks, right now he would have reason to fear reprisal, since that vote could energize black voters and lead to increased turnout from them. Same with GLBT voters, immigrant voters, even (ugh) evangelical voters. Under a jury system, the politician could expect consistent turnout from all those groups. (This is also true under compulsory voting, but I see that as less of an issue, since compulsory voting is forcing all blocs to utilize 100% of their political power, rather than artificially limiting the political power of all blocs, hopefully consistently).

Sadly I can't see the jury pool being adopted in the near future, because
a) People fear the unknown as much as they complain about the current system.
b) People would feel they didn't get a say in the election if they were not picked (though they could still influence the result by spreading propaganda).
c) Many people would not want to spend a few days learning politics & economics (although that seems to be an issue of how they get reimbursed financially).
d) It is in the interest of those in power not to shake things up too much.

What are the actual reasons against a jury system (if any)? It seems like it would mean that those who vote really understand the issues, without taking power from any subset of the population.

A is for a good reason. A few experts advocating something should be a far cry from sufficient evidence for changing an entire election system this drastically.

I can't give a personal example for C, since I do learn politics (and get graded on it), but I can give a somewhat similar situation. Schools tell you to do work, then reward you with a grade if you do it. My senior year of high school, I knew that I could get 100% in my super-easy AP English Class on a paper without reading the book. I had no desire to read the Scarlet Letter, so I didn't. The size of the reward (I suppose how much the grade mattered) didn't affect my decision making process, because I knew I would receive the reward. In other words, you have no way of ensuring that people actually learn politics and economics, and the financial reimbursement doesn't change that.

People in power love shaking things up for good reason like more power, or more of the reason they got power. I'd say that people who are in power because they like power are the minority. Most politicians want to actually do something they believe in, whether that's provide universal health care, or oppress women and minorities.

[Griffin]

Some counterpoint, hopefully decent:

Okay, problems with the jury system:

1) I agree, here. I think anonymous voters is incredibly stupid for a host of reasons. We don't have anonymous juries, after all. I think we should enact safeguards so we don't know what positions they were called /for/, mind you.

2) Of course, perfectly random isn't the desired outcome - a somewhat representative but ultimately higher quality group of voters is. If we go by registered voters, we're already doing significantly less-than-random, after all. And at least some people would need to opt out. Considering we already disallow vast swathes of the population from voting at all, and have since the country was founded, I'm not sure if "perfectly random" or representative has ever been the desired goal.

3)

Because people aren't voting? Compulsory voting, then. If people don't vote now, why would you assume they would want to put more time and effort into voting and would do it under the new system?

Because they get paid, because of social expectations, and because of legal requirements to actually be present - like a real jury. Sure, they can still slack off and dial it in - but this makes it significantly less likely.

Or are you concerned that people are voting ignorantly? Why would you assume they would be less ignorant if they were part of a smaller sample?

Because they are given time, resources, and potentially even access a normal person wouldn't have. It would actually even be possible to speak to the candidates directly under various systems. In addition, they would see their own votes as more important than usual (because they are), likely increasing their internal evaluation as to the merits of further research on the topic compared to the effort involved.

Paying them wouldn't change much, unless you had some sort of test, but polling tests are extremely bad.

The purpose isn't to make them less stupid, it's to make them less ignorant. Education, the whole purpose of the exercise, would go a lot further than polling tests.

4) You point out that this "problem" is just as much a problem under compulsory voting (though I don't understand exactly why it gets more of a pass from you there). However, countries with compulsory voting don't seem to be much more worse off on this front. Why don't you think alienating a section of the populace and having their votes go to an opponent isn't enough, by itself, to sway opinion? Why do you think small but vocal groups should have so much more power and influence than the size of their group would indicate? Do you think it's truly important to achieving the best government possible?

[Dr. Diaphanous]

I mostly agree with Griffin, but why don't we want the selection to be perfectly random? Wouldn't that invalidate the election in terms of statistics?

[Griffin]

Our current elections certainly aren't even remotely random. Do you think that's a major issue? I'm not saying it's a bad goal, just that perfect statistical representation has never been our goal in the past, and I don't know why we would inherently seek it after this change.

[Dr. Diaphanous]

With a jury system you'd have to purposely go out of your way to make it not random.

[Griffin]

Simply letting people opt out (or only pulling from people who volunteer) would make it no longer perfectly random. That's not particularly out of one's way.

[omgryebread]

I mostly agree with Griffin, but why don't we want the selection to be perfectly random? Wouldn't that invalidate the election in terms of statistics?

Yes and no. Yes because random samples produce accurate results on average. No because "on average" isn't good enough for elections.

Random sampling is variable. It produces a statistical representation by averaging large numbers. An outlier one way has an equal chance of occurring as an outlier the other way. The problem in polling (where this is discussed) is that polls are not meant to be right "on average." A poll is a snapshot of the public mood right now. Truly random polls have a pretty good chance of producing some wonky statistics. Therefore, most pollsters go with some sort of quota system (geographic is popular. Random dial people a certain number of people from each state.)

This is especially important in elections. Sure, if 2012's jury happens to have an unusual amount of Communists, it will be balanced out in future elections. In the meantime though, we have 4 years of the guy those people helped elect.

4) You point out that this "problem" is just as much a problem under compulsory voting (though I don't understand exactly why it gets more of a pass from you there). However, countries with compulsory voting don't seem to be much more worse off on this front. Why don't you think alienating a section of the populace and having their votes go to an opponent isn't enough, by itself, to sway opinion? Why do you think small but vocal groups should have so much more power and influence than the size of their group would indicate? Do you think it's truly important to achieving the best government possible?

They should have voting power equal to their portion of the voting-eligible population. The cool thing about our election system is that every bloc already has that. Guaranteed, not subject to flukes of random selection. Higher turnout from a group is them using their voting power, which is perfectly healthy.


Basically, yeah, right now the voting people are not representative of the US as a whole. But the population eligible to vote is representative of the population eligible to vote (they could vote for the Tautology Party.) Their decision not to vote is just that, their decision. In a way, it's a vote. By systematically disallowing people from voting at all (even at random) you necessarily make the population eligible to vote non-representative.

[Griffin]

Right. And I don't get why that matters.

But I think we may be operating from fundamentally different opinions about what representative democracy is, what its for, why it matters, and what's important about it. I'm more of a traditionalist on those fronts.

And most people probably agree more with you than me nowadays, which is why I'm pretty sure this will never happen. :/

Ah well.

[Dr. Diaphanous]

I mostly agree with Griffin, but why don't we want the selection to be perfectly random? Wouldn't that invalidate the election in terms of statistics?

Yes and no. Yes because random samples produce accurate results on average. No because "on average" isn't good enough for elections.

Random sampling is variable. It produces a statistical representation by averaging large numbers. An outlier one way has an equal chance of occurring as an outlier the other way. The problem in polling (where this is discussed) is that polls are not meant to be right "on average." A poll is a snapshot of the public mood right now. Truly random polls have a pretty good chance of producing some wonky statistics. Therefore, most pollsters go with some sort of quota system (geographic is popular. Random dial people a certain number of people from each state.)

This is especially important in elections. Sure, if 2012's jury happens to have an unusual amount of Communists, it will be balanced out in future elections. In the meantime though, we have 4 years of the guy those people helped elect.

That's just basic statistics: those problems fade away as you increase the sample size. If you're taking a sample size of a few thousand to chose between 6 candidates/parties (for example) I think it would be very unlikely to get a freak event (e.g. communist party winning 50% of the vote when only 10% of the population would vote for them).

[Griffin]

And if you're doing a group per election, rather than one group to handle ALL elections, the impact is lessened even further. (And I think it should be a group per election)

[Squall83]

1. Are you so sure that these policies are in fact unpopular? There's a big difference between unpopular somewhere like reddit and actually unpopular with likely voters. I don't have any evidence one way or another on this question, I looked for NDAA polling and found just one poll, but it was online and more than 1/3rd had no opinion, so I'm not inclined to take it seriously. It did have NDAA at 23% support and 38% against, though.

I'm not sure about NDAA, but for the others it seemed to me as if the people generally didn't like to be leashed by them. I mean, even xkcd blacked out because of them.

@JohnSmith1

The 1st time Bush was elected, actually less than 50% of all Americans wanted him. He just focused on the big states.

16 US presidents won with less than 50% of the popular vote, including Lincoln and Kennedy, but please don't let the facts stand in the way of your Bush bashing.

I am not bashing Bush, I'm bashing the winner-takes-all system. Bush is the only president of whom I knew that he won like this (because I did no research about American history but I followed that election when it happened) and that's why I'm naming him as an example. If 16 presidents won like this, it just shows the problem of that system: President candidates consciously abuse the weakness of winner-takes-all. Yes, I don't like Bush, but if he had had a majority I would've accepted that most Americans had wanted him.

The 1st time Bush was elected, actually less than

In theory, you can win an election by getting 25,1% of all votes if these votes let you steal another 25%.

This is true for all representative democracies, not just the electoral college. If you want to keep democracy, the only other alternative is direct democracy, where everyone pretends to spend a few hours everyday reading bills.

And? Does that mean you don't want a change because all representative democracies are like that, too? There is always the voting system I introduced in my opening post or - as Qaanol pointed out - the Approval Voting system.

About the jury pool election system:
I can understand how it makes a difference if the voters are selected randomly, if
1. the people would actually vote
2. if it weren't for the Gaussian distribution, which simply states that there is a chance of getting a very unrepresentative result.
3. if it can be guaranteed that the voters are selected randomly
4. if the voters are not bribed

Without these problems it probably would improve the German system as well, because without it we have the problem: "the fewer people vote, the higher are the odds of bad (e.g. racist) parties to make it past the 5% hurdle"

But it also kinda feels bad to strip more than 90% of all people of their vote. If less than 50% of our people vote (and in some parts this is the case) then it's a problem we must solve socially and not mathematically, by exploring why people don't vote and by solving these problems.

[JohnSmith1]

The 1st time Bush was elected, actually less than 50% of all Americans wanted him. He just focused on the big states.

16 US presidents won with less than 50% of the popular vote, including Lincoln and Kennedy, but please don't let the facts stand in the way of your Bush bashing.

I am not bashing Bush, I'm bashing the winner-takes-all system. Bush is the only president of whom I knew that he won like this (because I did no research about American history but I followed that election when it happened) and that's why I'm naming him as an example. If 16 presidents won like this, it just shows the problem of that system: President candidates consciously abuse the weakness of winner-takes-all. Yes, I don't like Bush, but if he had had a majority I would've accepted that most Americans had wanted him.

I assumed you were familiar with the 16 number, so it follows you must have cherry picked 1 particular president out of the 16. That assumption turned out to be wrong, but it was a reasonable assumption to make of someone who started a topic on problems of the winner-takes-all system in American politics.

The 1st time Bush was elected, actually less than

In theory, you can win an election by getting 25,1% of all votes if these votes let you steal another 25%.

This is true for all representative democracies, not just the electoral college. If you want to keep democracy, the only other alternative is direct democracy, where everyone pretends to spend a few hours everyday reading bills.

And? Does that mean you don't want a change because all representative democracies are like that, too? There is always the voting system I introduced in my opening post or - as Qaanol pointed out - the Approval Voting system.

When did I say I'm against change? I'm all for removing winner-takes-all, but I also realize any other voting system (except direct democracy) also share the 25.1% problem. Specifically the voting system you introduced in your opening post and the Approval Voting system.

I was just pointing out the logic flaw in your argument:
A has the 25.1% problem, so let's go with B and C instead.
Even though B and C also has the 25.1% problem.

[Squall83]

It's good to hear that you're against winner-takes-all, too. The misunderstanding is, that I didn't propose an entirely different system. The German system was an example. I'm quoting my first post now:

In my opinion, every party that gets enough votes in one state to get 1 seat should be allowed to keep that seat.

That's all I want to see changed. No more vote-stealing via winner-takes-all. Everything else can remain the same for now.

[Kain]

Yay, now that I won't be necroing the thread, I can finally respond to JohnSmith1!

Re the 16 US presidents with less than 50% of the vote: I am not sure about the other 13 (I am assuming Bush was included in the 16 count), but both Lincoln and Kennedy had the plurality of the popular vote, unlike Bush in the 2000 election, where neither main candidate had the majority, but Gore had the plurality (again, not saying that this didn't happen before, just that your cherry picking of Lincoln and Kennedy was rather poorly considered). Honestly, I am was surprised to learn that there were so many elections with an actual majority win.

[JohnSmith1]

Yay, now that I won't be necroing the thread, I can finally respond to JohnSmith1!

Re the 16 US presidents with less than 50% of the vote: I am not sure about the other 13 (I am assuming Bush was included in the 16 count), but both Lincoln and Kennedy had the plurality of the popular vote, unlike Bush in the 2000 election, where neither main candidate had the majority, but Gore had the plurality (again, not saying that this didn't happen before, just that your cherry picking of Lincoln and Kennedy was rather poorly considered). Honestly, I am was surprised to learn that there were so many elections with an actual majority win.

I cherry picked Lincoln and Kennedy because they're the most famous of the 16 and nothing else. A stink was raised in this thread about not winning the popular majority, so in that sense all 16 were equally "guilty" of the same thing.