The real reason Futarchists are doomed

post by lc · 2022-04-01T18:37:20.387Z · LW · GW · 2 comments

There are a lot of little ways political leadership is organized differently throughout the world. Looking just at outcomes, states themselves vary widely on a bunch of axes - you have command economies, like North Korea, and market economies, like Singapore and Switzerland. You have nation states with high state capacity, like North Korea, and nation states with low state capacity, like Somalia. And you have states that don't seem to have high per worker productivity, like North Korea...

But if you really look at how political leadership is chosen, there are only really two ways people govern. On the one hand you have representative republics, where citizens meet at a regular cycle to discuss and select rules by some kind of majority vote. On the other, you have autocracy, where a select group of individuals rule indefinitely and with relatively little checks on power, fighting (maybe bloodlessly) amongst themselves when any given ruler dies for the privilege. The quality of life of their citizens vary wildly and lots of the countries of the world must cope with very different cultural, economic, and geopolitical circumstances. Despite this, the means by which they do so tend to be mostly the same. 

This is curious. There's not a whole lot else that you can say about every country in the world. The fact that things seem to resist being organized any other way should give you pause. 

Despite this, rationalists, at least in my experience, tend to be big fans of alternative forms of policymaking. I don't find this strange. It seems like a natural side effect having a lot of nonstandard policy opinions and nerdily detailed models of economics and public choice that one might think that we need a better system of picking laws, perhaps even a radically different system. After all, even in Norway and Singapore, people seem to manage to commit gross inefficiencies all of the time. Surely there's some better, more ambitious political decision making process that sacrifices normal democracy in favor of jury voting, or prediction markets, or some other mechanism that is less biased towards the whims of the masses who have no particular personal incentive to get things right at the ballot box.

I like many of these proposals, but I am not so optimistic about their implementation. I'll clarify actually - I suppose no one is optimistic, but I have specific reasons to believe they would be very difficult to manage. 

To the daydreamer it may seem like constitutions are just arbitrary rules that civilized nations should be able to follow. Surely, if the founding fathers wrote something reasonable, like selecting members of a jury to vote for president, then it would have happened. If only Robin Hanson had been present at my nation's constitutional convention! However this mindset ignores (at least) one critical defining feature of a country's political organization, one America's founding fathers probably understood even if it wasn't elaborated upon in the transcripts or the document:  

All governments must put anyone capable of overthrowing the current regime in charge. Any government which wants to keep someone else in charge must give them the tools to organize a frightening counter-coup, and/or remove that capability from everyone else.. 

It's a simple rule - and easy to justify. If someone or something - whether it be the military establishment, a political party, a charismatic politician, or a religious leader, can organize a coup to overthrow the system you've designed, then your form of government is essentially reliant on that parties' goodwill. And there is only so much goodwill a nation can chance! If you want a country to last a few hundred years, it's going to have to reckon with someone or something that sooner or later does not hit the cooperate button. In this way the rule is a simple rephrasing of a Nash equilibrium as it applies to states - governments are the way they are because no single group has the power or reason to move them anywhere else. We can immediately see how the autocrat(s) sustain the equilibrium: they kill or oppress anybody else that could rise against them. They take control of the political and economic resources someone could use to mount a campaign. 

Yet even though this principle seems intuitive, what's not obvious to many people is how it applies to liberal democracies. Surely in democracies the point is that no one is able to organize a coup. You might have even thought that it's strange that democracies tend to be so stable. For all their citizens' worrying, the United States has lasted hundreds of years even while using the exact same founding document since 1786. They have managed to sustain strong civil liberties that should, at least under an autocracy like the Soviet Union, be destabilizing and open it up to agitators who want to seize things. Why hasn't our military seized power? There's a vague sense that it wouldn't be very successful, but surely the nation with the most powerful military in the world is somehow tempting fate.

I think the problem here can be resolved by using a different turn of phrase. Americans actually have coups all the time. The point of democracies is that coups are built into the system. They are writ large into the constitution as an indirect form of governance by the actual rulers, who organize their puppets like Joe Biden to be placed in power every few years. Who are these rulers?

Well, in theory, the people. More accurately, the majority coalition. The group that, in a democracy with strong civil liberties - perhaps even a law guaranteeing private ownership of firearms - by numbers should be the group most motivated and capable of overthrowing the government in the first place. The beauty of a democracy is that they don't have to. 

This is the fundamental reason why democracies tend to be so much more stable than autocracies, have more peaceful transitions of power, and have so many fewer civil wars. The democratic system automatically, as a mechanical mechanism, assesses and then gives power to the most powerful political party in the country. The tension that exists when the nash equilibria is out of whack simply isn't there; when it works well, any rivals that are actually capable of organizing a popular uprising to Joe Biden out of power know that the system will grant them that power within a few years. There's no reason for the majority coalition to organize a coup to express that power and take the risk when they can just sit back, relax, and watch that power be handed to them.

This is also the reason why the "more" democratic a nation gets the more it tends to support civil rights and civil liberties. The closer a nation gets to a true democracy, run indirectly by the majority coalition, the more that majority coalition will vote and organize for the tools and means to monitor (and potentially insurrect against) the rogue agents inside its government that want to take power from that majority coalition and give it to some other group. Civil liberties are not just some cultural artifact, present in some countries that "want to fight for them" and not in others; they're also the expression of the majority coalition's will to rule.

It's an imperfect abstraction, but it shows why jury voting (along with the more radical proposals, like statically skewed sortition and futarchy) will never be implemented. The reason reforms like ranked voting are politically plausible is because they can be sold to the majority coalition as a more accurate signal of their will. However, if jury voting for political positions is to be more effective than representative democracy, then ideally it's not necessarily representative of the preferences of the average voter. And if the jury, because of better insights or more time to think or anything else, doesn't vote like the majority coalition does, then that's a problem. The nash equilibrium is now unstable. There's no one, by default anyways, with the personal or psychological motivation inside the 'futarchic' system to oppose the majority coalition when it decides to act against the will of Metaculus.

I have heard of many proposed alternatives to the two systems, some of which I even find preferable, but I have yet to hear one that satisfies this criteria. I suspect it's probably because it's really hard to come up with one. Most people who grow up in liberal democracies in the west are very attached to the ideas of civil liberties, myself included. But civil liberties exist to protect the rights of the majority coalition. If you want your nation to be ruled by some other entity, then you have to do the hard work of figuring out how to ensure the survival of that entity's political power first. 

2 comments

Comments sorted by top scores.

comment by Jackson Wagner · 2022-04-07T07:25:23.026Z · LW(p) · GW(p)

I agree that one of democracy's core features is that it reduces conflict, and any future system of government should be designed with that in mind. Nevertheless, this post strikes me as foolish for several reasons:

  • Democracy rewards the majority of the people. But is the side with more raw numbers (even if just 51%) really the group best positioned to win a coup? I propose a new form of government, Money-Democracy, where you get votes equivalent to LN(your_net_worth). Now the rich and powerful are properly weighted as being more influential. Wouldn't this be a better way of always crowning the faction most likely to win a fight? Does this mean that Money-Democracy would be a more stable form of government, and we should expect it to sweep the world any day now?

  • Futarchy means "vote on values, bet on beliefs". Since you're still voting on values, couldn't you still do all your coalitional fights over who gets to determine the values that the prediction markets are optimizing for?

  • You say, "There's no one, by default anyways, with the personal or psychological motivation inside the 'futarchic' system to oppose the majority coalition when it decides to act against the will of Metaculus." But by that logic, nobody in the USA would oppose the majority coalition if it acted against the Supreme Court? Couldn't a prediction-market branch of government be respected by the various factions, in the same way that the Supreme Court or Congressional Budget Office are reasonably respected by both parties in the USA? Your theory of democracy seems to predict total majoritarian dictatorship, but that isn't what happens. A futarchy could similarly be supported by a system of checks and balances and societal norms.

  • Imagine that one country implements futarchy despite the fact that (I will admit) this is a difficult trick to pull off. But imagine that futarchy works great and that country soon develops unparalleled state capacity, economic growth, etc. Soon, other countries would be forced to follow suit -- they'd have to either adopt futarchy themselves or be consigned to irrelevance. In my view, this is basically the process by which democracy took over from other forms of government -- by growing faster and being more stable and providing citizens a better life. Citizens of nondemocratic nations see what they're missing out on, and they agitate for change.

  • Empirically, not all democracies continually ratchet up to higher levels of civil rights. Some countries go in a very majoritarian direction as an off-ramp from democracy, like Hungary, Turkey, or India. Even among fully democratic countries, civil rights are at a high level but I'm not sure they're all currently increasing. So I'm not sure what to make of that idea.

  • Imagine living in a country run by futarchy. The president (elected by the majority coalition) is supposed to handle foreign diplomacy, be the leader of the military, respond to crises, set the cultural tone for the nation, and help define the values function that the prediction markets implements. But one day the president tries to go against a futarchy ruling, saying "actually this law is null and void, the conditional prediction market that favored it is rigged, anyone can see that this would be a bad law!" Wouldn't people freak out and the stock market drop (because the stock market knows that ignoring the prediction markets and making authoritarian power grabs is bad for business)? Wouldn't people then get mad that the economy was doing poorly, and then vote out the president? Thus, "respect the prediction markets" would quickly become a well-reinforced norm in the society, since no president would want to tank the markets and get voted out. This is just the same as why Biden today doesn't try to ignore congress or the supreme court.

Replies from: lc
comment by lc · 2022-04-07T08:57:32.829Z · LW(p) · GW(p)

First: these are a lot of good points, some I did not think of as an intellectual test of my ideas beforehand. Any responses that I come up with now are a little suspect because I had not anticipated several of them. 

Democracy rewards the majority of the people. But is the side with more raw numbers (even if just 51%) really the group best positioned to win a coup? I propose a new form of government, Money-Democracy, where you get votes equivalent to LN(your_net_worth). Now the rich and powerful are properly weighted as being more influential. Wouldn't this be a better way of always crowning the faction most likely to win a fight? Does this mean that Money-Democracy would be a more stable form of government, and we should expect it to sweep the world any day now?

The majority coalition organizes conditions to defend itself from illegal coups, but that doesn't mean it's the most capable group to form one itself. The point is that by passing civil liberties the majority coalition ensures the system is stable. Numbers serves as a good enough proxy under a highly libertarian democracy for power that it tends to be able to reliably maintain order as the current incumbent, though not always!

I agree that it's an imperfect abstraction to say that the more your system puts powerful people in charge the more things are stable, because obviously, some more overt ways of putting the powerful person in charge galvanizes opposition by offending people's moral or ideological sympathies. However, it's IMO the most important deciding component, and the other components tend to be technically surmountable by the ruling faction, once power is achieved (via propaganda, lying, etc.). They simply don't tend to be as much of a binding constraint on behavior in practice.

Futarchy means "vote on values, bet on beliefs". Since you're still voting on values, couldn't you still do all your coalitional fights over who gets to determine the values that the prediction markets are optimizing for?

I have no response to this. It actually just occurred to me that Futarchy can be packaged as a more effective way to enforce majority coalition values, even if it doesn't immediately enforce their beliefs about the necessary actions. In that sense it, and jury voting, can be seen as a little analogous to ranked voting. It's not necessarily true that you have to give literal power to the majority coalition to placate it. This throws the entire posts' thesis into question.

You say, "There's no one, by default anyways, with the personal or psychological motivation inside the 'futarchic' system to oppose the majority coalition when it decides to act against the will of Metaculus." But by that logic, nobody in the USA would oppose the majority coalition if it acted against the Supreme Court? Couldn't a prediction-market branch of government be respected by the various factions, in the same way that the Supreme Court or Congressional Budget Office are reasonably respected by both parties in the USA? Your theory of democracy seems to predict total majoritarian dictatorship, but that isn't what happens. A futarchy could similarly be supported by a system of checks and balances and societal norms.

This ignores the implementation details of acting against the Supreme Court overtly, from the perspective of a coalition. You of course still need some institution to resolve legal disputes about the laws passed by congress. What will that institution be? A designated party man, like Joe Biden? Now you're giving him almost unlimited power. What if he turns against the majority coalition? 

Perhaps supreme court cases could be decided by majority vote, but you have to remember the point you brought, namely that the majority coalition also might genuinely believe that the current system advances their values better than a simpler system. Democrats and Republicans (rightly IMO) have decided that the best way to influence the supreme court is by putting party-hardliners on the bench when they can. The fact is that coalitions inside democracy do try to influence the composition and power of the supreme court through normal means of law, pretty transparently. It's just that they don't do that by attempting to overthrow the court by force, because that's risky and opens up a can of worms of failure scenarios.

Imagine that one country implements futarchy despite the fact that (I will admit) this is a difficult trick to pull off. But imagine that futarchy works great and that country soon develops unparalleled state capacity, economic growth, etc. Soon, other countries would be forced to follow suit -- they'd have to either adopt futarchy themselves or be consigned to irrelevance. In my view, this is basically the process by which democracy took over from other forms of government -- by growing faster and being more stable and providing citizens a better life. Citizens of nondemocratic nations see what they're missing out on, and they agitate for change.

I have a similar but not exactly similar view about democracy, and agree this is a plausible scenario. If the majority coalition decided its interests would be better served by Futarchy through the power of a salient example, then certainly this could happen.

Also important is that if the countries that founded Futarchy adopted a universalist ideology about their political system, not unlike some other countries we know (wink wink), then the wild success of their economy could enable them to project power and attempt to "support" other Futarchist revolutions around the globe. Of course, well-functioning domestic institutions, strong economic growth, etc., isn't 1:1 with international power, but the correlation is there.

Empirically, not all democracies continually ratchet up to higher levels of civil rights. Some countries go in a very majoritarian direction as an off-ramp from democracy, like Hungary, Turkey, or India. Even among fully democratic countries, civil rights are at a high level but I'm not sure they're all currently increasing. So I'm not sure what to make of that idea. 

I think you mean authoritarian instead of majoritarian. And here I mostly disagree. If you look at the history of a solidly democratic regime like America, it starts with a system of more wacky and less democratic control (political bosses of cities like New York, the electoral college, etc.) and then ends up gradually moving further and further towards majoritarian control. My perspective on flawed democracies is that they're either one, just pulled both ways; instances where the majority coalition in those countries is not quite powerful or inclined enough to oppose the power grabs by authoritarian cabals, but for circumstantial reasons the authoritarian cabals can't grab any more power than they currently have. Or, two, and more commonly IMO, the countries simply in an interim position and are actually unstable, in a Nash sense. Those specific examples you mentioned are nations that have been on a pretty clear trend-line toward authoritarianism, and I would place them in bucket #2. Nothing about this model suggests things can't change, especially as time moves forward and circumstances do; just that things try to move to an equilibrium. For full democracies that equilibrium is pretty solidly giving more power to the public, from the perspective of the public, to effectively advance their agenda.

Imagine living in a country run by futarchy. The president (elected by the majority coalition) is supposed to handle foreign diplomacy, be the leader of the military, respond to crises, set the cultural tone for the nation, and help define the values function that the prediction markets implements. But one day the president tries to go against a futarchy ruling, saying "actually this law is null and void, the conditional prediction market that favored it is rigged, anyone can see that this would be a bad law!" Wouldn't people freak out and the stock market drop (because the stock market knows that ignoring the prediction markets and making authoritarian power grabs is bad for business)? Wouldn't people then get mad that the economy was doing poorly, and then vote out the president? Thus, "respect the prediction markets" would quickly become a well-reinforced norm in the society, since no president would want to tank the markets and get voted out. This is just the same as why Biden today doesn't try to ignore congress or the supreme court.

Maybe?