uncrediblehallq:

nostalgebraist:

OK, I am curious:

I think I have at least some followers who like Robin Hanson.  Does anyone reading this think futarchy is a good idea?

To be explicit, the big problem I see with it is that once prediction markets get plugged into actual government behavior, the betters gain a new incentive in addition to “be correct so I can make money” – namely, “cause the government behavior that I believe will benefit me.”  So if someone thinks a policy would be bad for the nation but good for them, they may bet on it being good for the nation.  After it’s implemented and turns out to be bad for the nation (according to the metric being used), they’ll lose money, but they’ll have a new policy working in their favor, which may outweigh the cost.

There obviously should be a limit on the amount of money any individual entity can invest in the market, to prevent rich people and organizations from directly causing policy by investing in the way described above.  Even then, we would have to carefully design the rules to close loopholes (such as an organization creating large numbers of spin-off organizations, or a large company mandating that all its employees invest in a particular way).  And if any loophole were left open, some entity would exploit it, after which it could gain enough control that it would be difficult to close the loophole afterwards.  Admittedly, the existing political system has these problems too, but at least we have the appeal of “principled” candidates as a possible counterweight.

Even if big money doesn’t gain control of things, individual investors would still have an incentive to “vote” for policies that benefit them, even if they’d lose money by doing so.  Depending on the size of this effect compared to the size of the incentive to make money in the market, futarchy could end up being a lot like direct democracy with votes counted in proportion to individual wealth – something I don’t imagine anyone wants.  This could be made less severe by making the cap on individual investments very low, but that would correspondingly make the incentive to make money weaker, and in the low-cap limit we just get unweighted direct democracy.

(It should be noted that even if people are investing in the general interest rather than in self-interest, they are still the same general public who are currently able to vote, more or less.  Hanson seems to assume that if people consistently lost money they would leave the market, but it’s not at all clear to me that this is true – one can always keep “voting for” the same “bad” policies, arguing that they only failed due to other confounding factors rather than intrinsic badness, which is similar to what people already do in politics today.  Note also that the average person doesn’t necessarily invest in a way that makes them money, even if their only incentive is to make money.  A lot of people end up buying high and selling low.)

So for futarchy to work as intended, we need (1) very good rules that foresee all possible big money loopholes, and (2) individual incentives that are dominated by making money in the market rather than “spending money to vote,” and (3) investors who use strategies that actually make money, or who at least leave the market if they don’t do so.  (1) seems impossible, but admittedly is also a problem for our current system.  (2) is an empirical matter and seems impossible to know until we implement futarchy, in which case it might be too late to go back.  And (3) is very unlikely.  (I would imagine that people currently interested in prediction markets know an unusually large amount about finance, while futarchy would produce wide public interest in these markets.)

---

This all seems so obvious that I am sure Hanson has written about it, but I hate his writing style and so I haven’t looked into his writing on the subject.  Does he have solutions to these problems?  Do you?

Did you mean to have the second to last paragraph be “or”s instead of “and”s?

I suspect (1) is likely to be counterproductive, but some combination of (2) and (3) might obtain. How likely is it that they would obtain? Well, we can reason by analogy with the stock market… which there’s no consensus about.

I think how you feel about futurarchy depends on how “efficient” (in the technical sense of reflecting available information accurately) you think financial markets are. Scott Sumner is a hard-core efficient marketeer, and has loudly promoted what he calls NGDP futures targeting, which is basically futurarchy applied to monetary policy. I’m less of an efficient marketeer than Sumner, but looking at recent Federal Reserve policy, lean towards thinking NGDP futures targeting might be an improvement.

They were supposed to be “and”s – I was saying that if any of 1-3 fails, we will have one of the problems I described earlier in the post.

Efficiency of financial markets is complicated (I read a book called The Myth of the Rational Market once – which was less reductive than the title makes it sound – and remember almost nothing from it except “wow, this is complicated”).  But I don’t really see how we can extrapolate from the efficiency (or lack thereof) of currently existing markets to the efficiency of futarchy markets.

The possible incentive to “vote by taking a loss” isn’t nearly as big in the stock market as it would be in a futarchy market.  Some people do, say, invest in specifically in companies they like (”socially responsible investing”).  But this doesn’t generate nearly the level of general interest that government policy does.  And no one feels like they have a civic duty to invest in the stock market, while people might stay in futarchy markets despite losses because they continue to believe their favored policies are good and feel a duty to keep supporting them.  (If the real-world impact of their failure is unlikely to change their minds, will an additional monetary incentive do so?)

Likewise, this general interest and sense of duty would bring many people into the futarchy market who have no interest in investing otherwise, and so the proportion of investors who “know what they’re doing” would be lower than in the stock market, even if we ignore the considerations of the previous paragraph and assume everyone just wants to make money.

Mostly, my worry is that a market hooked up to political outcomes, of the sort many people care a lot about, would be much less efficient than currently existing markets.  And I don’t see how to test this, except by actually implementing futarchy and seeing what happens.  (Might be possible to do lab experiments where the market is tied to something like fairness of distribution among the participants at the end?)

(via uncrediblehallq)