Hacker News new | comments | show | ask | jobs | submit login
Politics is bad because we use an 18th century voting system (80000hours.org)
357 points by robertwiblin 5 months ago | hide | past | web | favorite | 349 comments

A quick scan of the transcript didn't let me find what makes his plan "viable". The system is so bad that almost any plan is a massive improvement, but such efforts often end in squabbling over which system is the best.

Here in Canada one of Trudeau's platforms was to get rid of first-past-the-post. But he nixed it when he found out the committee was going to recommend mixed-member-proportional (which would probably hurt his parties chances) rather than ranked ballots (which would definitely help his party).

I think the only way to avoid this brand of self-interest is to, prior to any discussion, agree that the changes will only come into effect 20 years into the future. It's hard to predict who it'll benefit that far into the future. 16 or 12 years might be doable. 8 is definitely too close. 4 is a train wreck, everyone is on campaign already.

The downside of doing so is that it still gives plenty of time to cancel the changes at year 12 or 16, once it becomes apparent who will stand to benefit/lose from the change. This type of hacking future benefits is evidence in many of the tax overhauls, which use predicted future savings to offset short-term costs. And often those future savings are legislated away before they come to fruition.

(Of course, sometimes those future savings wouldn't have happened anyways, due to unreasonably optimistic assumptions about the performance of the economy.)

Cancelling such a scheduled change is harder though than simply blocking a change in the first place.

Exactly. Sunset provisions have large effects even though in principle they can always be removed if the majority in the future wanted to.

I encourage you to look into the history of copyright extension laws.

It's not that simple. First, I'm not sure that anyone knows who would benefit 8 years out. There will be guesses, sure, but nobody's really going to know how it plays out until after the first election under the new system. Even that won't really tell us much, because everyone will adjust their tactics after the first experience.

Second, even if everybody knew who was going to be the loser from the changes 8 years in advance, that's not enough. That party would have to be strong enough under the current system to undo the changes. In practice, that would require controlling the House, the Senate, and the Presidency... unless both parties decide that they would lose power to third parties, and strike a deal.

I was thinking they need to start recording and widely publicizing exit polls using the approval system so that people can get a sense of what the alternative 'feels' like. The big problem is that it'd take at least ~three election cycles for most people to appreciate the difference it'd make.

E.g., in the first cycle, extreme candidate A wins the actual election while the moderate, B, is shown to be widely tolerated under the approval system. Nearly half the voters--those who voted for the opposite extreme, C, end up loathing the results of the actual election and wishing in retrospect that the approval system had been in place. Perhaps in the second election cycle, candidate A wins re-election because they're a known-quantity/incumbent.

Finally, in the third cycle, history starts repeating itself and people get tired of A's party. E.g., the Overton window slides back towards C's perspectives leading to C being seated. This time the people who still prefer A are disgusted with the outcome and prefer the outcome of the approval system.

At this point, half the voters have long argued that the approval system makes more sense and the other half has also come around to thinking maybe it could bolster their position as well. It'd only remain for less partisan activists to e.g., get a petition going towards a referendum.

> I think the only way to avoid this brand of self-interest is to, prior to any discussion, agree that the changes will only come into effect 20 years into the future. It's hard to predict who it'll benefit that far into the future.

It's hard to predict the short-term beneficiary parties when it goes into effect, sure, but its quite easy to predict the long term ideological benefit of increases in proportionality on modern democracies.

Ironically, that may make near immediate implementation more viable, if their is a divergence between the expected short-term and long-term impacts, and those expected to benefit believe they can use that to preempt the naturally expected long term benefit but those who would expect that long term benefit don't believe their opponents will succeed at that.

Or to (infuriatingly for the general public) simultaneously pass a package that ensures present-day personal benefits to all current incumbents, removing all risk from their calculations of personal benefit and thus inducing them to do what is simultaneously best for them and for the future polity.

That can result in last minute scuttling too.

While politicians are notoriously bad at planning 20 years ahead, I wouldn't put it past them to try, especially when that timeframe makes it easier to project the impact of demographic trends rather than merely political ones.

"...there's actually a really good track record in the US for passing ballot initiatives on single winner voting methods, so we expect the likelihood of winning some to be pretty high.

The way that we look at it is instant runoff voting has been passed as a ballot initiative in a number of cities, but we see approval voting as producing better outcomes, and having better political dynamics compared to instant runoff voting. Approval voting is also so much easier, and it avoids a lot of the problems.

If instant runoff voting can win, then surely, a simpler voting method that produces good outcomes and has good dynamics should also be able to do it."

Approval voting is worse than even IRV because of inconsistent meaning of ballot markings, an effect which the naive mathematical analyses which support the claims of it's superiority ignore.

This problem is negated when the is a consistent meaning to approval or disapproval markings, which can happen with non-secret ballots tied to concrete commitments. Approval is, for that reason, an excellent voting method to decide group activities in a social group, where an approve vote is a binding opt-in to the activity if it is chosen (or if a disapprove vote is a binding opt-out.)

For the same reason (the lack of a concrete definition of what “approve” or “disapprove” means), approval is not really simpler than IRV (or other ranked ballots) methods, even if the space of possible ballot markings is narrower.

The problem is voting-theoretical good outcomes has no relation with "good outcomes" that lets voting methods to win. cf. Canada case in GP.

I think that political calculations regarding proportional representation were a factor in the Liberal decision to scrap reform but the stated reason for not wanting a MMP system is a pretty good one. In Europe MMP has supported a lot of fringe parties that have used the funding that comes with having full-time representatives on government payrolls to build strength to the detriment of the political environment.

Folks cite the fact that fringe parties only receive 1%ish of the vote in Canada as evidence that this is not a valid concern but it is easy to see how voter behaviour would change as campaign mechanics changed due to a PR system. (These arguments rather conveniently ignore the number of Green Party votes that Canadians in most ridings cast knowing they are throwing their vote away when arriving at the 1% number.)

"has supported a lot of fringe parties that have used the funding that comes with having full-time representatives on government payrolls to build strength to the detriment of the political environment."

-- Like the "pirate party", which is really doing a good job on pushback against all kinds of corporate enrochement against copyright laws. (like the stupid Spain link tax).

Sometimes these 'fringe' parties bring ideas and raise concerns on issues that the major parties are not considering due to their self interests.

Sometimes they are down-righty looney, but I think the positives on having small parties represented outweighs the negatives.

Scotland, whose Parliament uses proportional representation via AMS/Additional Member System (one of the main drawbacks of which is that it takes an age to explain...), has had representatives from socialist/green parties win seats they wouldn't have had a chance in hell winning under FPTP ("First Past the Post"), which I think many people feel has lead to some pretty positive policies being debated and implemented which otherwise might not have happened.

AMS/PR has had the effect of forcing parties to work together much more in Scotland than has ever happened in the FPTP UK Parliament. I would dearly love for the UK Parliament to learn some positive lessons from the Scottish experience, but I doubt I'll see it in my lifetime.

I really like AMS - it retains the direct constituency link of a member for a region elected via FPTP, with the fairness of additional regional members selected proportionally resulting in a pretty proportionate outcome. The downside is that it creates two classes of member - those elected via FPTP, those elected via what AMS calls the 'list', and of course the complexity of giving people two votes instead of one that behave in different ways. The first vote is a classic FPTP, the second vote applies to the list. Your party's performance in the FPTP vote affects the "weight" of any list votes you receive, which is how the system remains proportional. As I said, it's a pain to explain, but one of the fairest systems I've experienced.

Like "Golden Dawn" & Jobbik - are we going to pretend that the good that the "Pirate Party" is doing is outweighed by literal fascists with seats in government? Sometimes these fringe parties represent things best not funded by parliamentary salaries.

That's why I think Holland has quite a progressive voting system with such a small theshold and 17 parties ending up in the Parliament.

Many countries (including Germany, where MMP was first used) have a minimum bar of votes for a party to make it to parliament, e.g. 5% in Germany. Where exactly you put the cutoff is a bit arbitrary, but this solves the problem efficiently. There are parties in parliament in Germany that I'd rather not have there, but none is fringe.

But "fringe" parties are also the parties to change the status quo, and it's not always (or often) a bad thing.

Plus, you don't already see the Republican party in the US as the type of extremist party you worry a "fringe party" could become?

My opinion regarding the Republican party is consistent with the above statement and gerrymandering was a key part of building a party system that incentivized the same sort of bad behaviour that an MPP system incentivizes just within the Republican Party instead of as a fringe party acting autonomously.

Canada has mostly solved the gerrymandering problem and where geography makes solving it impossible the strong party system makes MPs that want to act irresponsibly accountable to a party that has battles to win in other places that make their positions untenable.

I am open to some examples of fringe parties that have made significant positive contributions to the status quo, I would like to see some examples that are of more significance than the negative effects pseudo-fascist parties rising Europe have had.

Yeah, the stated reason is pretty much BS.

There's nothing stopping the government from saying "A party must have 5% of the vote to be awarded any PR/list seats" in order to keep the fringe parties from grabbing individual seats here and there. BC just released their suggestions for an October referendum on reforming provincial elections, and that was part of all three suggested systems.

I don't think it would be much of a challenge to get a party committed to explicit racism to 8 or 10 percent in Canada with a charismatic leader and decent organizing.

If lightning strikes once and you get an MP into parliament after that as a private member that MP is going to get a lot of media attention, it will immediately change politics because with large parties member discipline is important whereas in an MMP system a party that has no hope of ever winning it all has the reverse incentive - they are 100% committed to whipping up their "base" for their pet issue.

For reference, that's basically the case in Germany, introduced after remembering horrible bikeshedding in the 1920s.

Do the voters of the fringe parties think that it is to the detriment of the political environment? It sounds almost like saying those groups don't deserve representation because of a ideological difference, which is not a justification to deny a certain voting system.

I am perfectly comfortable saying that there are groups that don't deserve a specific type of representation in parliament because of ideological differences between them and the mainstream of the Canadian public.

They are welcome to participate in democracy and do the hard work to get their ideas into the mainstream of Canadian public opinion but they are not "entitled" to a system that gives them representation in a specific way. One should be honest with oneself about how representative democracy works and not pretend that the purpose is to have every hair-brained idea at the table.

Democracy doesn't start and stop with elections.

I don't think the stated reason of "voters want something, but I don't think they should have any representation" is a good reason.

In other words, the best system of government is a sham democracy where the public are mollified by frequent elections, but prevented from changing the status quo.

Here is what I think of representative democracy:


But yes, at the very least, use the Approval vote. It has way better statistical properties than Ranked Voting.

Maine switched.

The primary parties will never approve of IRV nor proportional representation because it hurts their chances of winning. The partisan divide is a good motivator for forcing people into one camp or the other.

It was not just some cynical reason the effort failed. MMP is a party list system. Canadians don’t want it. Over and over again Canadians vote down any change that doesn’t make representatives accountable to voters. Our government system is called Responsible Government for a reason.

Also Canadians do not like minority governments. Proportional representation is not that popular on Canada.

Also we like simple systems with simple voting strategies. MMP is very complicated. Lots of weird tactics like voting for a small party but a candidate from a bigger party.

> Also Canadians do not like minority governments. Proportional representation is not that popular on Canada.

Proportional representation doesn't need to result in minority governments. Proportional voting is very common in Europe but minority governments aren't.

Just means parties need to find common themes on which they agree, and compromise on some their more controversial goals. Finland has had up to six parties in a single government to make sure tough decisions can be passed.

Coalitions are minority governments. They are super common in Europe.

Only when the coalition doesn't have a majority of the parliament seats. I don't think minority governments are common at all in Europe, as the whole point of a coalition is usually to get majority.

Personally, I prefer the minority governments. They have to play nice with the other kids, at the risk of being chucked out of power. As a minority, they have to keep at least one other party reasonably content. Harper's minority years were more reasonable than once he got a majority.

The NDP/Green coalition in BC is new, but seems OK so far, and that same minority aspect probably tempers the behaviour of both parties.

I do need to do some homework before the referendum though.

Disclosure: I'm a Liberal Party member.

To me the answer to Trudeau's concerns was so obvious; just do mixed-member-proportional, but limit the number of top-up seats to 2x or 3x the number of riding based seats won by the party.

That way we don't have to put up with the lunatic fringe, but we still fix the warping effects of FPTP.

Yes, but as others have mentioned, Trudeau's stated concerns were total B.S.

The "fringe" party he's concerned about is the NDP.

Meanwhile Italy is still struggling to form a government. I prefer Canada's system to Italy's

Successive one-party governments, where the Liberals or the Conservatives have absolute power, from now, till the end of time, despite neither party having ~35% of the vote?

> which system is the best

Can't this be reasoned scientifically?

Yes, and Arrow's Theorem states that there is no "best" - you basically get the choose-two-of-three: https://plus.maths.org/content/which-voting-system-best

Arrow's Theorem only considers single-winner systems. But most politics is done by the Parliament, not the President.

There is no need to use small, single-winner voting districts to choose the members of the Parliament. Make voting districts 5-10 times larger and use proportional voting. Arrow's Theorem is not valid for that kind of systems.

Arrow's Theorem only applies to ordinal voting systems.

There are other voting systems that get around this. Kenneth Arrow himself was into non-ordinal voting systems.

Approval voting wasn't considered in that article.

"there is no perfect system" != "no system is any better or worse than any other system"

http://ncase.me/ballot/ and featured here: https://news.ycombinator.com/item?id=13161396

This interactive graph predicting outcomes for different voting methods convinced me that this is much harder to do than it would seem. Very much worth checking out the link. I think maybe I found it on HN.

From blog:

The further to the right a voting method is, the more it "maximizes happiness" for the voters

Why should group satisfaction with the winner be the important metric? Surely good governance should be the most important metric. If the quality of government is too hard to measure then length of time before overthrow or revolution should be a good proxy.

Good governance is a loaded term. Good for who? Different stakeholders will frequently have opposing goals, desires, and needs. Cutting foodstamps program helps some by lowering taxes, but hurts others who need help. Then the argument moves into secondary effects, cutting foodstamps will make poor people better off because the economy will grow, or cutting foodstamps will hurt taxpayers because poor nutrition and financial instability fewer people will escape poverty. The problem becomes complex enough that no one really knows the answer and so even the smartest often just end up restating their preheld belief.

Markets are supposed to be efficient they match demand with production efficiently, or some such. Political demand needs to be matched with equal efficiency. Imposing good governance into a voting system will very likely create distortions.

Good for the country, obviously. If you can't define what good governance is, how can you possible come up with a way to maximize it?

The purpose of elections is not to be happy with the outcome of the election, it's to be happy with the policy that results from the election.

You seem to ignore my point about multiple stakeholders. I believe that the happiness measure is not a measure of outcomes but happiness with getting a representative closer to his/her choice or preference...meaning one that may best represent their views, desires, and needs in government. Tha seems like good governance.

You seem to ignore my point about the collective good. By definition measuring happiness is measuring feelings. Rule by emotion. When has that ever worked out well?

If closer to personal preference is better, shouldn't direct democracy - absolute majority rule - be the ideal? But it's clearly not the case, at least not in any sizeable country.

what does "good for the country" mean?

If welfare helps people in the short term, but reinforces cycles of poverty is that good for the country or not?

If nationalized medicine helps people in the short term, but eliminates the incentive for medical innovation is that good or bad?

> Good for the country, obviously. If you can't define what good governance is, how can you possible come up with a way to maximize it?

It's almost like this is a really complicated problem that the best and brightest have been thinking about for centuries.

It's almost like "happiness" with who won one election is a gross oversimplification and fairly meaningless metric. Sounds like we're in agreement.

People disagree on what the purpose of government even is, so... no.

And this is probably the origin of most political disagreements.

Only if you can define your criteria for "best".

Approval voting is biased toward "moderate" candidates. This has all the problems of first-past-the-post in terms of electing representatives who do not allow for a wide range of political opinion. It empowers status quo politicians beholden to special interests.

Is it really? Or is that your take? You could just as easily say that it's biased toward candidates who work effectively with a broader range of people.

We have observed reality to draw conclusions from. Regardless of political affiliation, would anyone really argue that special interests haven't taken over the political process in the US, and made significant inroads in most if not all of the industrialized democracies throughout the world?

Sure, I’ll argue that. “Special interests” is just code for “people whose opinions I don’t like.” But at the end of the day, our government very closely reflects what you’d expect. If you account for the fact that likely voters skew richer, older, and more conservative than the population as a whole, what exactly do you think would be different if “special interests” were not in charge? Would we spend less on defense? 1 in 3 Americans say we don’t spend enough. Would we spend less on social security, more, etc?

For the quote of mine being referenced, I'm content with the dictionary.com definition of "special interest group":


a body of persons, corporation, or industry that seeks or receives benefits or privileged treatment, especially through legislation.

Removing wealthy 'special interests would probably have some interesting effects. Off the top of my head, I expect we'd have better regulations on banks and the like. And perhaps Mickey Mouse (and everything else more than 75 years old) would be public domain.

> “Special interests” is just code for “people whose opinions I don’t like.”

It's code for above-the-median-wealthy, enough to influence opinion outside of the populist opinion. Why you choose to go partisan over a systemic issue, is beyond me.

"Special interests" includes many powerful groups that aren't especially wealthy:

- religious groups wanting faith-based educational programs

- farmers wanting crop subsidies (hugely powerful in the US)

- internet users wanting net neutrality

- conservation groups wanting to preserve parks or endangered species

- unions wanting occupational safety regulations

- retirees wanting increased social security

You might even agree with some of these.

The common theme here is "concentrated benefits and diffuse costs".

The guilty flee where none pursueth. You'll note my original post was carefully non-partisan. Your rush to bring conservative talking points into the discussion suggests some kind of bias, or perhaps a guilty conscience.

I think if you had a national legislature with "moderates" (as defined by California and Massachusetts) and "moderates" (as defined by Utah and Mississippi), you'd have a fairly wide range of political opinion represented. You wouldn't have the extremes of California represented, and you wouldn't have the extremes of Mississippi represented, but that may not be that much of a loss...

> Approval voting is biased toward "moderate" candidates.

Wouldn't that only be true if preferences are roughly normally distributed (or some other fat-middle distribution)?

And if that's the case, shouldn't the outcomes of the voting system tend towards a moderate middle?

A wide range of political opinion is sometimes useful (e.g. in one of the houses of a legislative body). Moderate candidates are also useful. Rather than choose one voting system to use for everything, it might be appropriate to use a couple voting systems that have different strengths and weaknesses. So you might imagine one house of a legislature elected with approval voting, and a different house elected with a proportional method.

If a moderate candidate is the one who is acceptable to the largest proportion of the voting public, why shouldn't they win?

Approval voting allows voters to vote for 3rd party candidates without "throwing their vote away". That would be a huge improvement over first-past-the-post, as 3rd party candidates would have a decent chance of winning if they can make a compelling case to voters.

Bias toward moderation is a trade-off. Attempting to minimize it seems naive.

At the extreme end, a legislature consisting of communists, anarchists, facists, theocrats, and hippies wouldn't accomplish anything but igniting civil war.

Or perhaps it would keep the really extreme and unworkable BS out of the law and government policy. Such a diverse body of legislators is going to have a much smaller set of policies they agree is a good idea, and that would result in a smaller body of law. I see a smaller, simpler body of law that more people can think and agree with as a Good Thing.

> communists, anarchists, facists, theocrats, and hippies

I would think all those groups together would only make what, max 20% of the US electorate? In a proportional system, they would still be a minority of the Parliament. And the internal disagreements within that minority of the representatives would prohibit them from joining forces and using even their 20% of the power. The rest 80% could mostly just ignore those fringe groups.

Every time I discuss voting with Americans, or when I read articles like this, they only focus on how to make a better system for single-winner elections. I think that is missing the main point.

I think the main problem is the single-winner election system itself, and even the best voting mathematics can do only very little to help that. They all will still lead to a two-party system.

Only the presidential elections need to be single-winner. But for all other political bodies, a proportional system where at least 5 to 10 representatives are chosen from each voting district, would be better. Choosing e.g. 10 winners from a single voting district with a proportional voting system would set the election threshold to 10%, so any party with at least 10% support would get at least one representative. This is how almost every European country runs their elections. (Only UK and France still use single-winner systems.)

This is the only way to bring diversity and options to the political landscape. And by having more than 2 viable parties, you would have more diverse political discussions, too.

And even for the single-winner presidential elections, more than 2 parties would have the existing organization and funding structures to plausible run campaigns and candidates, so as a byproduct you would get diversity and options for the presidential elections, too.

I do recognize that it is problematic to organize proportional voting for (a) the Senate and (b) for those states that have less than 5 House representatives. For example, you'd need to pool Idaho, Montana, Wyoming, N Dakota and S Dakota into one voting district. But even if you went with proportional voting only for the House, you'd still have more than 2 viable parties in each state, who each could make serious attempts to run for the 2 senators. So you would still get more than 2 parties represented in the Senate, too.

One downside of proportional representation is that it gives a lot more power to the parties to choose the winners. I'm not sure if voters would be comfortable with that. In the present U.S. system, a person can run as a Republican or a Democrat without the blessing or support of their party and win. I think that's a good feature to keep the party turning into some sort of "this guy and all his friends" club.

Granted, if there are more parties to choose from there's more opportunity for voters to reject dysfunctional parties, but it's not inconceivable that you might end up with, say, five dysfunctional parties instead of two.

I suppose a sensible hybrid system might be to have voters vote for their party in the regular election, and then for each party with one or more seats, that party runs a primary (via a multi-winner variant of approval voting or STV or something) to elect the people who actually serve in those seats. Does anyone already use a system like that?

> I think the main problem is the single-winner election system itself, and even the best voting mathematics can do only very little to help that. They all will still lead to a two-party system.

Two-round elections in which a candidate that doesn't get 50%+1 in the first round must run against the runner-up (and no one else) in the second round is a significant improvement over the single-round First Past The Post structure in the U.S.

> Two-round elections in which a candidate that doesn't get 50%+1 in the first round must run against the runner-up (and no one else) in the second round is a significant improvement over the single-round First Past The Post structure in the U.S.

No, it's not. It's one of the two systems in the US that are both referred to a FPTP (the two are more specifically known as plurality and majority/runoff, and both forms are widely used in the US.)

You'd still need to combine that with getting rid of FPTP, you can look at California this election cycle for an example. The multiple Democratic candidates running against 2 Republican candidates might lead to 2 Republicans in the runoff even if way more Democrats turn out to vote. For the Democrats running dropping out is a prisoner's dilemma situation. It's giving people a bad opinion of that system.

The California “jungle primary”effectively creates a new system—top-two/runoff—distinct from either form of FPTP, especially as usually practiced in the US where those follow a partisan primary.

I know I've seen some California elections have a 51% primary. I believe the Los Angeles Mayoral race was one. I wasn't sure if that applied to these jungle primaries or not since it's very doubtful anyone will get 51%.

Either jungle or runoff system, with FPTP in place, seems to leave a bad taste in people's mouth when they see the top two in the same party. I think it's getting a lot more bad attention because the likely minority party has a good chance of getting the top two spots.

> Only the presidential elections need to be single-winner.

An interesting thought experiment is having 3 equally powered heads of state (or indeed any small odd numbered group).

Most people's instant reaction is that it wouldn't work because we don't have it now.

After some reflection most people think it wouldn't work for other reasons that they can't adequately explain.

I think if you're going to contemplate nation-state wide administration reform it's worthwhile contemplating all kinds of changes.

(I'm in AU, where we have some semblance of 2pp, though in our most recent federal election some 25% of people did not vote for one of the two major parties/coalitions. Our system is clearly broken, but not as badly as the USA's. Small comfort.)

Interesting thought experiment, but how can one argue the point when there is a "people who disagree are ignorant/stupid" premise?

> Most people's instant reaction is that it wouldn't work because we don't have it now.

Straw man

> After some reflection most people think it wouldn't work for other reasons that they can't adequately explain.

Condescending generalization

Did you note the context of my message?

Parent had said that only the presidential elections need to be single-winner.

In Australia the head of government is the Prime Minister - curiously the specifics of that role are absent from our constitution, as it was evidently assumed to be a given. Head of State remains the reigning monarch of the Commonwealth, but that's a slight aside.

Prime Minister is not elected by the gen pop, despite many citizens of AU believing it to be the case. The 150 elected representatives in the parliament actually choose amongst them who the Prime Minister shall be -- a potentially very civilised approach to leadership elections.

Anyway, I'm not sure that I can defend myself against your claims about my personal observations.

Perhaps you could give it a go, and see if you get the same trends of responses that I described.

> Only the presidential elections need to be single-winner.

It sounds to me like you're thinking of the US as more monolithic than it is. If we only had the federal government, sure, only the president would really need to be single-winner. But pooling congressional seats logistically cannot happen, because congress is not parliament. The entire point is that senators represent their states, and House members represent their districts. Pooling them would be like pooling the EU Council and having all Europeans vote together on national ministers.

Remember that states are semi-autonomous, with their own governments following their own laws according to their own constitutions. Obviously our federal government is more powerful and rather more cohesive than the EU, but only up to a point.

(Note: I'm not saying this is a good or sensible system. In many ways, I would love for our resources and our voices to be pooled. The autonomy of individual states is a huge part of why the right wing is disproportionately represented at the federal level. And the fact that only states—not territories—get representation is a travesty.)

You're absolutely right, though, that the single-winner aspect is why we have essentially always had a two-party system (though _which_ two parties has changed several times over the years).

The way I once heard someone put it is that in parliamentary democracy, people vote for representation and their MPs form coalitions, while in the American system, the people themselves have to be the ones to form coalitions, in the form of the two dominant parties.

Pooling increases district representation, not decrease it.

The US Congress is far too small for the size of the population it represents, so many constituencies are ignored. Having larger voting districts takes power away from gerrymanders and gives it to the constituents. The population would create their own districts via voting blocs, instead of being trapped in gerrymandered districts.

I think uncapping the size of the house of Reps and going back to original apportionment would effectively do a similar thing.

Not necessarily. If one assumes that political affiliation is more or less uniformly distributed within a district—say, 60% party A, 30% party B, and 10% party C—dividing each district into 10 uniform subdistricts just results in 10 winners from party A and no representation at all for the remaining 40% of the population. This is an issue with winner-take-all districts, regardless of their size. A proportional system would instead select 6 representatives from party A, 3 from party B, and 1 from party C.

If you listen to the podcast, I think—like nearly every other comment here—that you'll find this is addressed.

One part of the 18th century that we left behind that maybe we shouldn't have was the level of representation in the house. Back then, each US rep had about 10,000 constituents, but today each has close to 1,000,000.

Going back to the original proportions, and having ~35,000 representatives could solve a lot of the money in politics issues.

On the other hand, a 35,000-member House would undoubtedly be too large for more than a small handful of important members to actually debate anything. Think of the PRC National People's Congress (setting aside that it would probably be powerless at any size..!), but 10 times as large: all the work in the 3,000-strong NPC is done by a standing committee of only 150.


At that scale, they would have to abandon traditional oral parliamentary procedures and switch to a text-based threaded comment medium. Reddit and HN show that it's at least possible to have hundreds and occasionally thousands of people "talking at once" without collapsing into chaos.

By comparison, I think most of us have had the experience of being in a conversation with a small group of people (or even one other person) and wanting to say something but not being given an opportunity to do so without seeming rude.

This could have other advantages as well. No reason why any one bill should block any other bill; anything with N/2+1 upvotes is passed by the House. It would also be possible for representatives to remain in their home districts among their constituents.

You think a bunch of lawyers who got into their positions because they speak well will start writing in comment threads?

Oh no... we'd have to more closely approximate a direct democracy in the part of the federal government that is supposed to be the most answerable to the people...

But really, just because the extremely autocratic CCP decides to concentrate power, doesn't mean that that's your only option.

It might be something structural that is general to all such situations; maybe when there is a very large number of 'ordinary legislators' there is a high probability of them becoming, over time, de facto powerless compared to party leadership.

The US already had de facto individually-powerless legislators -- The Republicans especially vote as a bloc whichever way party leadership declares.

Good. The Federal Congress has too much power anyway. Power should devolve to the several States.

Interesting to think about it from this perspective. Just off the top, I wonder if we could have a second level. Then each "Advocate" (just picking a word) that represents 10K people would then be the constituency for the representative. This is a model like the Rep is the CEO and the Advocates are the C suite. The Advocates never go to DC or the state capital, then only remain local. Then to make things simpler and cleaner, outlaw lobbying at the Rep level and make the Rep elected by most votes among the Advocates. Now to get a quorum on the Rep or an issue, lobbyists have to try to convince a majority of the Advocates. Hopefully this increases the cost and decreases the payoff. It could also be that the role of the representative is to just help with the negotiations and wrangling for the bills but doesn't actually get to vote. The act as a Super-Advocate based on the position of the majority of their Advocates and in the end, the must cast their vote on a bill based on a sub-vote of their advocates.

Don't take the above too seriously, I'm sure there are holes. Just a thought experiment in my brain right now.

The problem is that such a system is actually more easy to game and have a larger proportion of voters not support the overall policy.

If 51% of districts are 51% yellow while 49% are 100% pink, yellow wins.

With a second level, you can run that game twice, and 25% are enough to win.

That’s roughly how the Cuban system works although I think representatives are chosen by groups much smaller thank 10k and there are more of them.

This is a great start, but there's room for improvement. What if we could scale it up to roughly one representative per 10K constituents AND have each constituent represented by someone whose views closely resemble their own?

This is possible if you eliminate congressional districts altogether and have people "subscribe" to a representative. It removes the problem of a single congressman "trying" to represent constituents with widely disparate views equally, and abolishes gerrymandering in the process.

Pardon my GenZ analogy:

Like YouTube, you would end up with superstar representatives that would capture an even larger share than the most powerful do now. And the fact that they wouldn't even have a home district would mean that they would be accountable to absolutely no-one besides those who helped them achieve their superstar status. So yes, you would have "your rep" but he would be like a small you-tuber compared to PewDiePie—4k subs vs. 100,000,000.

That's a feature, not a bug. The system grants more weight to the voices of those who are thoughtful and contemplative in their selection than it does to lemmings. PewDiePie may have eleventy billion constituents, but he still only gets one vote.

It definitely would be different because the focus would shift from the popular candidates who are clearly going to get in to the ones who are right at the cut off. Depending upon the spread, if a lot of candidates are at the 'barely enough' line, it would mean major swings in policy driven by a small population shift, as long as the general population doesn't shift as well.

I'm guessing the numbers won't be known until after the voting is done, meaning a lot of effort will be spent finding out where the line is at.

I kinda want to simulate it for a few elections to see what happens, but there isn't any sufficient simulation.

No, it's a bug. I liked your original idea, let's have that one plus a 10k constituents cap per representative as suggested in the other thread.

I like this idea. Do you know of the potential downsides of it?

That's straightforward in the case of massive defederalisation.

50 states. 300 representatives in each state's assembly, almost all power devolved to those states. 15000 representatives. OK, it's not 35,000, but I doubt that makes much difference.

That would result in Wyoming congressmen having less than 2,000 constituents each, while Californian congressmen would have over 127,000 each. Wyoming voters would have about 65x as much power as California voters in the house.

In their house, yes.

> Going back to the original proportions, and having ~35,000 representatives could solve a lot of the money in politics issues.

As much as I like the idea of proportional representation, I don't think getting 35,000 people to agree on anything is a viable solution.

Perhaps a better approach is one that decentralizes power as much as possible, giving more voice to local constituencies versus the Federal government -- and similarly, giving power to continuously smaller communities such as county vs. state, city vs. county and so on and so forth.

An improvement on top of that is having the superior legislature write framework laws that more local communities can elect to follow, should they prefer that over drafting their own legislation, for whatever reason. While this isn't applicable to every legislative issue, the theoretical argument for this approach is quite compelling¹, at least according to one UVA Law professor²


1. http://www.law.virginia.edu/pdf/faculty/hein/kitch/70va_l_re...

2. https://content.law.virginia.edu/faculty/profile/ewk/1180712


Even if we upped it to 5,000-6,000 reps, it would be a huge improvement.

> Going back to the original proportions, and having ~35,000 representatives could solve a lot of the money in politics issues.

How? When people complain about money in politics issues these days, they're usually complaining about 3rd party money. I don't see how having 100x more representatives would reduce 3rd party money. It's not like it's hard to target 3rd party ads to small geographic areas.

I'm talking more about the borederline quid pro quo "I donated $10,000 to your campaign so here's my lobbyists, and more importantly if you don't vote the way they want, that $10,000 is going to your most difficult opponent next round". At the same funding levels, that $10,000 is now $100.

The per-person limit is already only $2,700, and only people can give money (not companies), so direct contributions are not particularly compelling as a path for corruption. Plus, quid pro quo is already illegal.

Candidates are far more concerned about 3rd party money than direct donations these days, because 3rd party spending is unlimited. Quid pro quo is not necessary because it's all done in public. You don't need a secret meeting with a lobbyist to know if the NRA likes your position on guns. Just watch the ads.

Of course not every House member has to worry about this; in fact most don't. Most of the big spending on House races is concentrated in the few districts that everyone expects to shift the balance. This would still be true in a House of 35,000. While most races would probably get less expensive, there would still be a small set of "tipping point" races that would attract most of the effort and money.

Campaign donations are only one form of bribery. There are in-kind gifts, and cash gifts to family and friends, and "consulting contracts" given after the politician resigns.

But it would take a lot more races to tip the House.

Not necessarily; it only takes a one-vote margin to pass legislation in the House.

Quite a lot of House races are on "auto pilot", politically--that is, there's not much suspense about which party's candidate is likely to win. This is in part because of gerrymandering that produces districts favorable to one party vs another.

Tiny districts would be like gerrymandering on steroids. It wouldn't take much fancy geometry to produce a map in which the vast majority of House races remain easily predictable. And it is the marginal (unpredictable) races that atract the bulk of the big spending.

But the cost of campaigns downs basically linearly with number of voters too. (TV ads are proportional to eyeballs.) There are some non-linear effects here, but I'm not sure how they operate.

I believe their point is that you'd have to buy 350 reps to get the same 1% coverage you can get today with 4.

You could probably get the same result by blocking funding external to the district and killing citizens united. And/or shorten campaigns to 60 days so they just don't cost as much. And/or, provide public funding and election spending caps.

I don't think that trying to scale out of it would be that effective but I see what they are getting at. My guess, is that would raise the inequality of the playing field again as the richest would be able to spin up an organization to manage 10k reps, decide whose relevant who to fund, and so on, people who didn't have lobbyists before won't be able to afford that even if the per rep cost gets cheaper, the total influence cost will rise and price additional people out of the market.

> But the cost of campaigns downs basically linearly with number of voters too.

What makes you say that? Hell at 10,000 constituents you borderline don't even need to run ads, just go door to door, personally.

Exactly. Seems to me it would be more plausible for a candidate without deep pockets backing them to be able to run a competitive campaign on issues the district cares about. Simply because they could just go around knocking on doors. A candidate and nine helpers each talking to 100 people a week for ten weeks could theoretically cover the entire district. Not to mention being able reach a good chunk of them by just hanging out at a local grocery store or visiting local bars and other hotspots. Throwing money at a hotspot race would have rapidly diminishing returns compared to a race with several hundred thousand people who can only be reached in numbers by broadcast ads or a truly massive campaign force.

EDIT: Although I suppose the tactic could at that point be turned to buying voters more directly. Spend a ton of money on something that benefits everyone in the district and make it clear that it'll keep coming only if they vote for your candidate, and you might have the same effect.

It wouldn't be the House of Representatives anymore, it would be the Colosseum of Representatives or Sports Arena of Representatives.

Like in Star Wars. I think it can work.

Even if you go down one representative for ten citizens, chances are that four out of ten won't be represented at all, given enough polarization. Regional representatives, on any level of granularity, solve underrepresented regions, but not underrepresented opinions.

And as an added benefit, in Presidential elections, low populations states wield less power than they do now. With the current number of representatives, votes in CA and TX count less than those in SD and MN.

I find that to be an added benefit.

That's what the Senate is for.

No, the house is supposed to represent the population and the senate is supposed to represent the state. Each member of the house should represent the same number of people, the senate should not. That's not the case though. CA/TX/WA/NY reps represent more people than less populous states.

The last census included non-citizen residents of those states, so this may be less true than you'd assume just based on census data. A sizable portion of the population in CA and TX are not eligible voters but they are represented in the electoral college anyways.

California has a total population (census 2010) of about 37,000,000 [1]. And about 87% of those are citizens [2]. CA has 55 electors, so roughly 585,000 citizens per elector.

Wyoming has a total population (census 2010) of about 563,000. And about 98% of those are citizen. WY has 3 electors, so roughly 187,000 citizens per elector.

Citizen in WY have roughly 3x the voting power as citizen in CA, due in large part to a system designed to appease slave-holding states 200 years ago.

1 - https://en.wikipedia.org/wiki/List_of_U.S._states_and_territ... 2 - https://www.kff.org/other/state-indicator/distribution-by-ci...

>Citizen[s] in WY have roughly 3x the voting power as citizen in CA

Which is the point, so as to prevent high population urban centers from dictating everything to the rest of American that doesn't share their values.

>due in large part to a system designed to appease slave-holding states 200 years ago.

I assume you're referring to the 3/5 Compromise here. While true that this was a consideration, the underlying reasoning is still sound. It makes no more sense for San Francisco to rule over Sac City Iowa than Sac City Iowa to rule over San Francisco. That's the beauty of the system the Founders designed. Otherwise you just drift further and further toward Direct Democracy, which is a euphemism for "chaos".

Sure. But, by capping the total number of representatives, the large, urban states are slowly losing power in relation to the small, rural states (at least when it comes to presidential elections).

> A sizable portion of the population in CA and TX are not eligible voters but they are represented in the electoral college anyways.

This is intended behavior. IIRC it was to address the question of whether slaves counted as members of the population for the purposes of determining number of congressmen.

money? the real issue is party control of the system. money is the only chance anyone has at breaking through the control each party exerts over elections.

if we want to take money out we need to remove the ability of the political parties from spending money on who they chose and require them to fund a set number of candidates equally. Not all candidates as its obvious you can get some kooks, but as it stands now elections are more decided by the party than the voters. the parties decide who gets to run for the seat and it becomes big news when a maverick usually with their own funds upsets this quagmire.

so please by all means try to minimize the money influence elections but understand that most people's views are being heavily manipulated by operatives of both parties to make sure that the only money in politics is money they have full control over.

Interesting that ~30,000 was also the number of Athenian citizens. The US framers aimed to avoid some of the problems of their system, but this is also evidence that that number is not unworkable.

You would need a lot of money to pay these reps though :k

You can greatly reduce the money spent on support staff to compensate for the additional reps.

Paying representatives is totally worth it. Are you saying that we shouldn't pay even the current number of representatives, or benefit of representatives sharply decline as their numbers increase?

I am saying currently we have 600 reps making 150000 or something like that per year. Having 35000 at the same salary would just cost a little more. In addition 35000 is like a decent town. TO get anything done probably would need some elections to form a council that makes decisions.

The representatives will elect representatives. Problem solved!

While I think the parent was referring to how much a company would need to spend to buy off enough reps for their legislation, it is a valid point that increasing the current $100M or so we currently spend on these guys by 2 orders of magnitude would eat a lot of the national budget. Then again, it would also lower unemployment.

Personally, I think we should go back to first principles to understand the meaning of a representative democracy. The aim is to get a representation of the population. Currently, we get the representation of the people who vote - but this is not the objective. The aim should be to get a representation regardless of vote.

There are different ways to get representation. One way is to force everyone to vote like Brazil or Australia but this might not be practical in a country like India where elections go on for 1.5 months. But, there are other ways - where you draw a representation of the entire population and make them vote aka Sortition [0]. We already use this concept in juries and other aspects of life. It would be worthwhile to give this a serious try for alternative voting systems.

We could perform alternate experiment systems in parallel and then compare their outcomes to current systems - we don't need to adopt them. But, this has to start somewhere. We could have a debate on "HOW" to draw this representative sample but we should acknowledge that the current system rarely achieves the objective of representation.

[0] https://en.wikipedia.org/wiki/Sortition

We could fix some of the barriers like automatically registering people at age 18 or making voting day a national holiday so people aren’t stuck at their jobs unable to vote.

That alone would help.

But every change produces winners and losers, and the losers really don’t want that change to happen.

Automatic voter registration and making voting day a national holiday is such a no-brainer improvement that it's a pity USA can't get it done. As a South Korean voter, this has been standard for decades.

Our system is the way it is for whatever haphazard reasons got us here. My impression is there wasn’t every any real planning outside of Jim Crow laws to keep minorities out of the electorate.

But we can’t automatically register people because young people disproportionately support one of the two major parties, and that would be ‘unfair’.

And we can’t make Election Day a holiday because the people who work so much going voting is a hardship disproportionately support one of the two parties. And it would ‘hurt business’.

On the other end we can’t require IDs to vote people that would disproportionately hur the other party. And we can’t spend the resources to fix that because it would be ‘too expensive’.

Americans are so concerned about how other people vote but why not just vote for who you want if you want to? If you really really want some party to win, then take the day off work somehow, however difficult. If you can't be bothered with that level of effort, then you care about your day at work more than the election, so don't vote and be happy with whoever wins. In my country, for me, voting was complicated so I didn't. Now the winner is doing things I don't like so I've decided to vote next election. I'm not going to go waving banners and insulting people who want to vote for someone else. I'll just privately find a party I like and go vote for them. I don't yet know which party that is! Americans should try that too - research what they actually want and who's likely to do that, instead of voting with their tribe. Once you become locked to a tribe, your vote loses its power because your tribe's party can do anything it wants without losing your vote. If you want election reform, then you'd better not have voted for either other two main parties because it's in their interests not to allow others in.

>making voting day a national holiday so people aren’t stuck at their jobs unable to vote.

early voting? postal ballots?

I'd actually go in the opposite direction, and making voting more difficult.

I'd much prefer a system where you have to "earn" your say in government. This would, I think, show a deeper commitment to the system.

Edit: if you’re downvoting me, irony aside, please detail why making voters more committed to the government they’re informing is a bad idea. Seems fairly reasonable to me.

Yes! We use juries for criminal trials not because they're especially good fact-finders but because they're very resistant to regulatory capture. Similarly, Sortition is the real answer to money in politics problems, because there's no such thing as running for re-election to raise money for. You get nice side effects in representation too - 50% of your legislators will be women, for example, but more significantly 50% of them will come from below-average incomes.

In the US at least, we already have a representative body laying around with super-majoritarian rules and long, overlapping terms where citizen-legislators could build institutional knowledge. Conveniently it has both the most anti-democratic setup right now and one that's already been subject to revision via constitutional amendment. Selecting Senators at random is the simplest fix with the biggest potential impact on our democracy.

And that's just first-order effects, not counting the potential civil society benefits of encouraging every citizen to think of themselves as a future Senator, or having that structure mirrored in micro in smaller jurisdictions.

And 50% of them will have below average intelligence and 50% will be the laziest people in the country. 50% will be the most hateful and 50% will be the most racist.

We should pay people $500 for voting (in the bi-annual national elections). Turnout would be >99% (?). ID and fraud problems would have to be handled correctly, and the people would equate voter fraud with stealing (and thus care about it.) National ID participation would voluntarily increase. This would help with other issues like homelessness, illegal immigration, AWOL soldiers, dead-beat parents, etc... The redistribution of wealth as a direct payment could also serve as the foundation for UBI administration.

Finally, it seems uniquely American to address this problem with money.

I have a hunch you'd get better policy results if people were paid to consciously abstain from voting – so that the least-informed, now making their decisions based on last-minute TV ads or mass-mailers with crude messages, voluntarily opt-out – delegating the choice to others.

I too miss the days when only landowners could vote.

Sounds fantastic. But I can already see a problem of poor people being disproportionately incentivized not to vote. Maybe the payment would be a tax reduction instead that affects everyone proportionally to their income.

Let's temporarily assume for the sake of argument that 'poor people' could wind up significantly better off under the decisions of a system where low-information, low-motivation voters are discouraged – but not disenfranchised. There'd be fewer (Hugo) Chavez- or Trump- type winners.

If under such a system, the 'voluntary, compensated abstentions' are disproportionately 'poor people', what principle is being violated? Is the option-to-vote or the actuality-of-voting the more important value?

Is maximal voting participation a religious goal, to be pursued even if it delivers poor results, such poor results sometimes including the collapse of democratic-processes entirely? Or is voting just one part of a system to be evaluated based on how well it delivers welfare and protects rights?

I think it's arrogant to assume that you know better than (other) voters. If you believe a demagogue is bad, what happens if he turns out to be just what was needed (Churchill perhaps?). Some people say Trump is good for poor people by wanting to reduce welfare which would pressure more people off their ass and into work. Welfare does have that incentive problem so there is a tradeoff and it's not obvious how much welfare is the optimum amount. If you believe democracy itself is too important to leave up to the will of the people, what happens if that turns out to be wrong? China is doing fine without it, while South Africa is going down the gurgler with it and Egypt fell on its face attempting it.

I would lean a little bit in the direction of actually voting being more important than being able to vote. Anything less than 100% turnout is bound to have some bias in how the non-voters are selected, even if they're self selected.

I'm not sure about solving the other issues you mentioned, but I've often thought that a ~$100+ payment for voting in a national election would fix the voter turnout problem.

Interestingly, that solution would motivate the rich less than it would motivate the poor.

Even more American? make it a tax deduction.

People who don't vote generally don't have strong opinions, or don't care at all, or haven't really thought about the issues. Why do we think representing the "meh, whatever" component of the population will help?

Stronger: If you don't care enough to learn about what the issues are and where people stand on them, I'm not sure that I want you to vote. In fact, I think I want you not to. You're going to vote either for a label, or for whoever had the most ads. Neither is a good basis for choosing who to vote for.

Those who do not vote generally skew towards certain groups and classes. These groups are not specifically any less knowledgable than the groups that do vote. This adds a skew to the results in favour of the groups that vote.

Systems like this are described and evaluated in Fishkin, James (2011). When the People Speak. Oxford University Press. ISBN 978-0-19-960443-2.

Why should the government care about people who don't care enough to vote?

They should because it benefits themselves and the country.

In the long run, the rulers of a country don't condescend to pay attention to voters because it's the right thing to do - rather because it prevents chaos, revolutions and guillotines. Voting is fundamentally an accounting mechanism that is justified by its provision of information about what is best for the country.

It's not so rational for people to participate in voting, because (virtually) no election is ever decided by a single vote. Thus, the electorate, as long as voting is voluntary, is going to inherently be biased towards irrational people motivated by emotion. That's why compulsory voting is a good idea. Voting isn't a privilege, it's a form of feedback that is vital to maintaining a healthy state, and having it be voluntary makes no more sense than voluntary taxes.

Sounds like an interesting idea. In the end the US needs a system where if a party gets a certain percentage of votes it gets some level of representation in Congress. For example in the 90s Ross Perot got 18% of votes but these voters got 0 percent representation anywhere. This is just extremely unhealthy. The extreme partisanship would get reduced quickly if there were some people in Congress who would vote with one of the big parties one and then with the other another time.

We have a federal, not national, government. National percentages of votes earned are meaningless.

I'm not saying that's good or bad, but that's what it is.

Agreed. But I hope my point makes sense. If 18 percent of people vote for someone and get no representation out of it they quickly will get disillusioned about voting and the country's politics will not be representative of the people.

Is it worse that Clinton got 48 percent of the popular vote or McCain got 46 percent and lost - that is a much bigger disenfranchised proportion.

Yes. Because by definition in a presidential election there can be only one winner. In the House or Senate there is no reason to not have proportional element.

Presidential elections can't be proportional, true, but they could be made less polarizing. The goal should be consensus, the election of a candidate acceptable to the vast majority of the population, not someone loved by 51% but loathed by the other 49%.

In the good old days the winner became president and the loser vice president. Maybe we should back to that. I would love a Trump/Clinton presidency.

Those Clinton voters still have their representatives in the House (and probably in Senate, too), so they are represented in the political system, just not in the White House.

The argument might better applied to state-level elections for legislative bodies. They are usually FPTP by district, so in a 51/49 split district, a significant portion of the electorate is left out in the cold.

We could adopt some flavor of split system... FPTP for district reps, with some number of positions set aside state-wide for proportional representation.

I believe Germany does something like this, though I'll have to dig up the details.

The german Bundestag (federal level) election rules can be roughly summarized like this:

Each voter gets two votes, one for a candidate and a second for a party list.

Half of the members of the Bundestag are elected by direct vote. The candidate in the election district that gets the most votes is elected.

The other half are distributed so that the total distribution matches the distribution for the party list. A party that fails to accumulate at least 5% doesn't get any seats from the second vote, but gets to keep any direct seats they gained, unless they win a minimum number of direct candidates.

If a party wins more direct candidates in a state than their share of votes per party list would allow for, the Bundestag grows and more seats are distributed to balance the proportion.

The system is designed to allow for direct candidates, but still keeps smaller parties relevant since any party that gets beyond 5% or manages to win direct candidates is part of the Bundestag.

(grossly simplified, but probably good enough for here)

Thanks for the summary!

Yes. That's what Germany does. If your party gets 5 percent you are in parliament even without winning a single district.

The lack of proportional representation is a feature of single winner districts, as opposed to multi-winner districts, not national vs federal.

I am not even saying we need full proportional representation but just some element so a party that gets voted for by a certain percentage of people can get some foothold. Right now you can't maintain a third party because the threshold to have an impact is way too high.

The Scandinavian countries have something like this, called leveling seats:


The Scandinavian countries have voting systems that are approximately proportional in the core. Then they use those leveling seats to fix the sub-5% differences in the vote proportions and the seats. But that's just a small correction if we are comparing voting systems that are not proportional in their design at all.



I disagree. I think it is a feature that Congress does not reflect electorates. Even under first-past-the-post Congress is much saner than electorates.

For years Unification Church tried to get representation in South Korean congress. They failed, but votes show that they would have succeeded under proportional representation. Proportional representation enables extremist parties.

If an extreme party gets maybe 5 percent they deserve representation if you like them or not.

I disagree. The idea of approval voting is that if majority is against for you to be a representative, you shouldn't be a representative. This is "the will of the people", per representative basis.

You know the term "tyranny of the majority"?


For some definition of "work". It's really just a different set of tradeoffs.

Continental Europe (the UK uses FPTP) is characterised by countries that have governments that are frequently deadlocked or missing in action, because proportional representation results in lots of parties getting into power with none having a majority. Therefore they spend lots of time building coalitions that can agree on very little except preserving the status quo.

These coalitions often result in absurd political configurations, like alliances between parties that are supposedly of the left and right, resulting in an inability to create policy. You also see bizarre compromises that can cause extreme or weird policies to appear that hardly anyone would have voted for, due to the need to bring on board smaller parties to make the coalitions work.

A few examples: in the Netherlands, the ruling coalition has a working majority of 1. It contains many tiny parties and took months to create, because they were determined to ensure that the politician who runs the second largest party in Parliament (Geert Wilders, who is anti-EU) would be entirely frozen out of power.

Germany is ruled by a coalition of two parties that both lost votes in the last election, and whose leaders hate each other. They both believe repeating the same coalition again will severely damage both parties, as it's a coalition of both the left and right meaning ... again ... essentially the coalition has no coherent policies that anyone campaigned on. However they did it anyway to freeze out the second largest party, AfD, again because AfD is anti-EU and that is seen as so bad almost anything is worth it to avoid having to compromise with them. Before the current coalition formed, Merkel tried to create the "Jamaica coalition" ... a truly laughable attempt to break the basic laws of politics by uniting the far left Green party with a libertarian party. Not surprisingly the quasi-communists and Ayn Randians couldn't agree on anything and that attempt failed, but it goes to show how strange proportional representation can get.

Italy has just voted in a coalition of two parties, again, a left and right wing party who have, as a result, economically incoherent policies. The "right wing" party wants devolution of power within Italy and more regional self rule, the 5-Star movement wants the opposite, with more subsidies for southern Italy.

Basically PR gives you messy governments that frequently collapse or deadlock. This may or may not be a better arrangement than FPTP which tends to yield strong governments with clear policies, which frequently see-saw between polar opposites as left and right parties take turns in power.

"frequently deadlocked or missing in action,"

That's exactly what I see in the US right now. The country has no ability to address issues anymore, be it health care, immigration, infrastructure, housing or education. there is a lot of hot air and screaming and yelling but they can't find compromises and act on them anymore. See Obamacare. Instead of working together and accommodate each others wishes first one party went into full strike and now it has been effectively abandoned at the detriment of most (non-rich) citizens.

I think the problem there is that they have to form a coalition. (I think this is because the coalition elects the leader of government, but I'm not certain of that.)

But what if the parties didn't have to form a coalition? What if they, as individual legislators, just voted on bills? Then you'd have to write a bill that could peel away some votes from some members of some parties other than your own. For a bill to pass, it would have to be a good enough bill that you could sell it to people outside your party. That's not such a bad thing...

Unless all the parties maintain strict party discipline. I think that increasing party discipline is one of the problems that is making Congress so dysfunctional. But hopefully, if you had more than two parties in Congress, not all of them would have strict party discipline.

They don't strictly have to form a coalition. The leader of government is elected with a normal vote, so a party could propose somebody as chancellor and if members of other parties vote for them too they get the job. But for now, no party has believed that this would work better than forming an (uneasy) coalition, especially since the parties sadly are fairly strict about voting en bloc. It'd be an interesting experiment though.

Lack of party discipline just makes it hard to know what you're voting for. Voting would become about perceived personalities of the candidates rather than party manifestos. That's a problem because for all their faults, the amount of thought that goes into a party's political positions is usually higher than that which an individual puts into theirs. Parties are also a way to combine communication strength.

Considering how diverse and complex the portfolio a politician has to decide on you can't really upfront know what you are voting for. I am fine if my rep deviates from party line if there is some thought behind it. Actually i prefer it that way,

> uniting the far left Green party with a libertarian party. Not surprisingly the quasi-communists and Ayn Randians couldn't agree on anything and that attempt failed, but it goes to show how strange proportional representation can get

Those characterizations have very little to do with the actual positions of those parties, so it's no wonder it seems comical to you.

>Not surprisingly the quasi-communists and Ayn Randians couldn't agree on anything and that attempt failed, but it goes to show how strange proportional representation can get.

That's a bit of a misinterpretation. One of the leading figures in the FDP party broke up the jamaica band when everyone was about to get together and form the coalition. (Naturally his party was pissed about this but because he won an election for them they just rolled over and went with that).

While the german government can sort of deadlock it's not a situation like in the US where everything shuts down. The government continues to function, although with minimized power, until the new parliament is formed. They can't put out any major new laws but they're not completely dead-in-the-water.

>Germany is ruled by a coalition of two parties that both lost votes in the last election, and whose leaders hate each other.

It's more complicated than that. The SPD-CDU/CSU coalition (which is three party not two depending on how you look at it) is getting a reputation with the voters, under this coalition Merkel has been chancellor for a long time now. They two sides (other than CDU/SPD resenting the CSU for sitting in the backseat of federal politics and screaming incoherent right-wing garbage (I'm bavarian, I know what I'm talking about here)) don't hate eachother but to my knowledge it's a mutual understanding that they cannot continue this coalition any further.

>because proportional representation results in lots of parties getting into power with none having a majority.

Note that Germany has a 5% clause for this; you need 5% of the votes to get into parliament, otherwise you have no say.

This is mainly a result of the previous government turning into the third reich after several parties deadlocked the parliament for real.

> Second largest party, AfD

Citation needed

Whoops, sorry, it's the third largest but the formal opposition party. If you consider the coalition to be "the ruling party" then AfD would be the second largest, but that would be an abuse of terminology.

It'd have been better to say the second largest power bloc.

In France we have two rounds. Each voter chooses one candidate during the first round and then only two candidates are selected for the second one.

Having to choose amongst one candidate only is terribly bad because some candidates may have lot of similar ideas. However they may diverge on some points, such as they decide to make two parties. The votes get then divided between these two candidates, and they may not reach the second round, even though their idea may be more popular...

The winning parties has never wanted to change the system since that makes them win I guess.

Any other system would be a better idea than that one.

Yeah the french system is very strange. On the one hand, it forces a second round of "pragmatic voting", which so far has kept out the various extremists. On the other hand, you get a feeling that people only vote for these extremists as a kind of protest, because they know there's a second round. And one day maybe we'll end up with a second round with two extremists (imagine LePen vs Melenchon...).

I think what also makes things strange in France is the rule that all parties must be given equal airtime, which leads to a few bizarre weeks every election cycle where complete loons are on TV and Radio 24/7

Are the two candidates of the first round chosen by who gets the most votes? If so it has the exact same problem, as the votes from the 3rd+ most popular candidates will get syphoned off to the two predicted winners.

Yes, see [1] for details. The "syphoning" exists, but it's not as bad as in a one round system where each vote on a small party is in practice lost. With two rounds, we usually have two "big" parties, flanked by two "small" parties at the extreme. People can push to one side by voting on the extreme side to extreme dissatisfaction with their closer mainline party, and up to a point it won't change the result of the second round with the two big parties. That's typically what happens, and it adds some variety with 4 large enough parties (and more very small ones of course) in the media / national discussions.

Then you have exceptional circumstances with many disgruntled people pushing for the extreme. Even then, two rounds can help. A one round system makes it very hard for the fringe to take control, but once this happen they're one of two. With a two round system only one side can go radical, with a likely result of the other side mainstream party winning.

Both sides extremes getting stronger is possible, and it's really what happened during the last presidential elections. But we got lucky and one "center" candidate went through (by a very small margin), and as expected won the second round. But as in other places we had about half people disgruntled, it's just that it was evenly split between left and right extreme parties, and only one went to second round (but the other was pretty close).

[1] https://en.wikipedia.org/wiki/Two-round_system

Isn't this the "Jungle Primary" process that's described in the article?

The difference is that jungle primaries can have multiple candidates from the same party. In the French system, the individual parties have primaries and send their winners on to the first round of voting.

Approval voting seems like a great way to make sure we only ever elect bland, inoffensive candidates with carefully focus-group calibrated opinions and policies, whom once in office, should they care for reelection, will steer clear of touching any even mildly controversial subjects.

As some commenters point out, it's not clear whether that's a good or bad thing. But approval voting would broaden out the conversation by allowing more heterodox parties to attract a significant fraction of the vote, reflecting their actual support, rather than just an artificially low 0-3%.

But that's exactly what's missing from all this discussion of voting systems, if the US and UK are your examples of "failed" democracies these experts obviously haven't considered the full range of how voting works. Living in a city that has instant run off voting it is clear that first past the post has some important advantages that voting system gurus seem to discount--namely it forces pre-negotiation and coalitions before questions are put to the voters. It is impossible for a voting system to inquire what the will of the body politic is on all questions as all times like it is an oracle--but if you construct a system that poses choose A or B question and there's education (campaigning) around that choice, the electorate can make a reasonably informed decision about whether A or B is better. Viewing voters as a static oracle to be consulted misses the whole point which is to actually get the electorate to make a decision since all things are not possible and government inevitably picks policies that favor some interests over others. In fact, suppression of heterodox policy party ideas in policy making (while protecting their expression through a strong 1st amendment) seems like one of the greatest features (not a bug) of US democracy.

I think that the citizens have kind of lost the whole idea of "representative".

Voters now want to send people to Congress who will carry out a specific agenda. But the idea of a representative is that you were sending someone with (more or less) your general views, and they would use their own judgment.

This whole "pre-negotiation and coalitions before questions are put to the voters" might be more the problem than the solution.

Your second and third paragraphs seem to be at odds with each other? Negotiating (whether pre or post) and forming coalitions is exactly what you'd expect proper representatives to do.

Representatives, yes. Political parties before elections, not so much (which is the point of the bit I was quoting in the third paragraph). Especially not so much if people expect the representatives to be bound by those negotiations for the duration of their terms, rather than actually using judgment as their term unfolds.

> if the US and UK are your examples of "failed" democracies these experts obviously haven't considered the full range of how voting works.

"Obviously?" Perhaps they have considered it properly and have reasonably concluded both systems are utterly failed in terms of representing the views of the electorate. As an outsider looking in US politics appears equally broken, just sometimes in different ways.

So now let me stay on ground I am more sure of, the UK. It's one of the least representative "democracies" around.

For most of the electorate it simply doesn't matter who they vote for, it's pointless. Most seats are "safe". Where I currently live, if you have anything but Brexit-Tory views don't bother as the seat hasn't changed hands in years. So consider yourself entirely disenfranchised. We get especially poor candidates as a result - from both parties. A cardboard cutout would probably get elected if wearing a blue rosette.

If you have Tory views, and live in one of those nice £1m+ warehouse flat conversions you are probably now in a permanently safe Labour inner-city seat, and are equally disenfranchised. cardboard equally electable, just give it a red rosette.

The percentage in a given constituency who get the government they voted for is often quite shocking. CGP Grey did a video on this if I recall correctly.

What does that do to politics? Well, it's total war, total victory and total annihilation only. Parties only treat the electorate as even vaguely interesting in the 10% of seats that might change hands.

No one negotiates. Beforehand or otherwise, unless every other avenue to grab power has failed. System working as intended then as it tends to reinforce the two party status quo.

"Woo, we got elected, just." Now it's simple. Every policy of the previous government was crap. Even the really good ones that were proven to be working. No matter, we didn't think them up, so get rid of the policy with extreme prejudice. We'll rename the same policy when writing our next manifesto so that we can imply we did invent it next time.

> namely it forces pre-negotiation and coalitions before questions are put to the voters.

When did the US last have a coalition, or even pre-negotiation?

In the UK we had one, successfully, during WW2. Since then there has been one, but a selection of minority governments. The recent coalition is widely considered to have been a failure. Essentially we got Tory govt, with a few rough edges knocked off. Tories got a convenient kicking partner.

> the electorate can make a reasonably informed decision about whether A or B is better.

Gosh that's a positive view of how it works. In our media-first, FPTP, system it is much more about the media friendly face, nice smile, soundbites and dog whistles. Interviews become an excruciating exercise of answering the question you'd like to have been asked like it's a Monty Python sketch.

May gave us a comical example of doing soundbites and interviews wrong with the "strong and stable government" line that everyone, including the media, were sick of on the very first day of campaign. After a term or two "it's about time we gave the other guys a chance" and "I'm sick of these idiots" becomes a significant influence.

In short I am firmly in favour of electoral reform and adoption of PR (and not the second-class incarnation of it we were given a referendum about during the coalition). I'd quite like to figure out a way to weaken party politics too, but I digress. In my twenties I was quite in favour for FPTP as I felt it enabled more to be done. Nowadays I'd like a lot less done and a system that a) encourages consensus day to day and b) provides a parliament representative of the far wider range of views commonly held than two diametrically opposed parties. This became far, far too long so I'll spare you the deeper reasoning in favour of PR. :)

My own politics is a coalition - I like some parts of Tory, Labour, Green and LibDem. Even during my most focused support of one party I found some of their policies were utterly bonkers. I think that's the case with most aside from the tiny, tiny minority who actually join parties.

Consistency and planning seem like an improvement from the current system of reversing every policy that the last guy put in every 4-8 years.

Sounds like an improvement.

Of course it sounds like an improvement when those controversial issues have never affected you. Who cares about all the people they do affect though?

Voting methods should reflect the will of people. Voting method is not the tool to get your pet issues done.

That sounds amazing.

carefully focus-group calibrated opinions and policies, whom once in office, should they care for reelection, will steer clear of touching any even mildly controversial subjects

This happens nowadays anyway.

This already happens in decidedly Red/Blue states. They're full of incumbents that nobody really cares to keep in but, more importantly, nobody cares to vote out.

I am fine with bland candidates. Much better than extreme.

Candidates who don’t take strong stances or don’t take stances at all are unlikely to be acceptable to many voters. A consensus candidate of the kind approval voting would prefer is not the same as a candidate who fills the lowest common denominator.

If you want to fix politics in the US, bring back earmarks.

It sounds like I'm advocating corruption, and I kind of am. But at the same time, earmarks served the purpose of providing a good reason for parties to work together, across the aisle. If a little bit of corruption accomplishes that higher purpose, I'm glad to have it.

Earmarks are orthogonal to corruption. An earmark simply allocates money to a specific project; a member of Congress who steers money to their district is not "corrupt," he or she is just looking out for the citizens they represent--which BTW is their job.

An earmark is only corrupt if it is bought--that is, if a citizen or company bribes the member of Congress to steer an earmark to them.

When earmarks were killed, there was actually little evidence that was happening. Earmarks were killed as a talking point about reducing spending. Which is silly, because earmarks allocated funds, they didn't appropriate them. That is, they typically divvied up the pot, not increased its size.

It's sort of like the debt limit. Politicians lie to people about things work, in order to attract votes. I don't know if that's corruption so much as cynical opportunism.

> Earmarks are orthogonal to corruption.

Only in theory. In practice, they are 1:1 - You initiate pork A here, I help you add pork B there. You own pork B company, I own pork A. We both win money, forget the constituents.

That's fair enough. Maybe stupid is a better word than corrupt. The "bridge to nowhere", grants for absurd causes, that kind of thing. They might seem stupid, but if they would allow a fix to Obamacare to pass (say, smoothing out rough edges, rather than a repeal or nothing) then the bridge is just a monument to getting things done.

A slight counterpoint: earmarks are correlated to a time when there were less strict party lines, so you did have people in office who were more moderate politically.

How about less gerrymandering? It results in more extreme candidates on both sides (as you get much more "pure" blue/red districts and fewer purple).

I'd argume that's true, but that earmarks are part of the cause of the less strict party lines. "Sure, I might be moderate. But you can be sure that year in and year out, I can bring West Eastport the bacon."

why would bipartisanship be good?

Are you saying it's bad?

If someothing can't be seen as worthwhile by both sides, without payoffs, there's probably a strong argument for not doing it. Gridlock is generally a good thing at the Federal level.

What happens, then, when one "side" takes it as an article of faith that any plan proposed by the other side needs to be rejected out of hand? That's the circumstance we're in right now, and have been since 2008 at least.

To wit, Merrick Garland would like a word.

This is fine as a theoretical construct, but the reality is that a lot of the best features of American government were the result of quid pro quo deal-making. A huge example is the entire Bill of Rights, which was the price of creating a stronger central government. And the movie "Lincoln" illustrates (with only a little exaggeration) the deal-making that went into passing the 13th Amendment.

As an aside, I happen to think that there is an important difference between "bipartisan" and "deal-making," and the latter is really where it's at in governance. Earmarks are tools for deal-making.

Framing things as "both sides" inappropriately elevates the 2 major political parties as a central feature of our system of government. They're not; in fact the largest political affiliation in the U.S. today is "indepdendent"!

And on any serious legislation, there are typically more than two "sides" with an interest in a particular outcome. That's why making deals is so important.

Why do these discussions always assume rational or good faith actors, when that so obviously isn't the case?

No, just asking why it would be good. The claim was that it's so good that it's worth increased corruption.

First thing to know is that the system was _designed_ 200 years ago to have many of the "bugs" that we notice today. That is : this isn't so much a case of a system that was fine 200 years ago not suiting the modern world but rather a system designed deliberately to have deficiencies; for the benefit of a subset of people, from the beginning. Knowing this should make it more clear why the bugs haven't been fixed yet.

care to explain this a little further? Genuinely curious what these deficiencies are and who they benefit.

There have been many books and articles on this over the years. Re-searching just now I found this one:


e.g. "The Framers worried about demagogic excess and populist caprice, so they created buffers and gatekeepers between voters and the government. Only one chamber, the House of Representatives, would be directly elected. A radical who wanted to get into the Senate would need to get past the state legislature, which selected senators; a usurper who wanted to seize the presidency would need to get past the Electoral College, a convocation of elders who chose the president; and so on."

If I can track down any of the book titles I'll follow up.

I'd argue that many of these "features" really ended up being poorly conceived in the long run and caused all kinds of unforeseeable side effects. In particular they completely failed to anticipate the explosion in urban population which has caused the voting power of those in urban areas to be greatly diminished. As a result of the electoral college and winner-take-all systems the interests of urban voters are generally grossly underrepresented at the federal level with respect to population size.

The founding fathers were smart, but nobody in the 18th century was smart enough to anticipate the requirements of a modern electoral system. Their real mistake though was making the electoral system both incredibly complicated and very hard to change.

Well, at least 2 out of 3 of those components are failed designs. We may as well figure out a system that doesn't make dumb compromises instead.

Single member districts is another reason. Even with good voting system single member district is bad for politics.

Proportional representation should be the minimum requirement. Adding additional voting system goodness on top of it is great.

I've been thinking this for almost all my life: the voting systems used in almost every country are plain ridiculous. The only country I know that uses a system that sounds reasonable is Australia.

Yep, STV: Single Transferable Vote.


Closer to home, it's also used by the Apache Software Foundation for the annual election of the Board of Directors.

When the election is only for a single position, this is exactly equivalent to "instant runoff".

Yeah, that video explains our Senate voting process almost exactly - each state gets a number of Senators, and you vote preferentially either at the party level (about 15 - 30 choices and you can preference up to six), in which case the party decides which candidate gets the first votes, or you can preference every candidate in the whole state (can be over one hundred - the voting paper is comically large[1,2]!)

For the House of Representatives, each electorate is much smaller and parties only have one candidate each, so you only have five or six choices, and you preference all of them.

1. https://www.smh.com.au/content/dam/images/z/r/7/i/p/image.re... 2. http://farm8.static.flickr.com/7296/9569414534_b3ee06fd84.jp...

The German proportional system seems quite good.

Should be noted that it's not perfect. It encourages a rather slow pace in politics. There is also a lot more than simply a proportional system to it (for one there is a 5% minimum for getting into parliament), there are laws and rules around voting such that the government doesn't deadlock because of a single party or complicated political spats

It's okay IMO but could be improved somewhat.

As far as I'm aware Australia just uses the same voting system as the UK, Ireland and New Zealand–i.e. almost the entirety of the non-American Anglo-sphere. Is there something else unique about how Australia does it?

The Open List system used in most of Europe (with France, parts of Germany, and Spanish congress being notable exceptions) seems like a very good, successful one also.

In the Republic of Ireland we use single transferable vote with proportional representation (which I think is a really good voting system, although a bit complex to understand the operation of).

UK is still first past the post for MPs. Australia seems to be different again but not first past the post.

STV is complex when there are multiple winners. It still produces a broadly representative result, but it is hard to understand how the mechanism worked near the margins that determine which candidates prevail near the lower threshold.

When there are is only one winner and STV is exactly the same as instant runoff, it is not so difficult to understand.

> UK is still first past the post for MPs.

This surprised me but you're right. NI have(/had?) STV, but not Westminster. Very odd.

> Australia seems to be different again

This was my question really. Australia uses IRV, afaik, which is somewhat equivalent to STV except that it's slightly simpler, though I don't see any major disadvantage to the extra complexities in STV.

Broke: 18th century voting systems

Woke: 13th century voting systems http://www.hpl.hp.com/techreports/2007/HPL-2007-28R1.pdf


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact