My Moldbug Debate

Mencius Moldbug and I did debate futarchy Saturday.  I’m not sure when the vid will be posted; here’s audio (start at 1:30).  Moldbug also has a follow-on 3700 word post; his main complaint is manipulator payola:

Suppose player P stands to make $X from decision D. In our chess example, he might have a side bet, paying $X, that White will open with P-KB4. Therefore, the question is: what will be his expected loss, $Y, from buying enough P-KB4 bets that White opens with P-KB4?  If X is greater than Y, manipulating the market (ie, moving it intentionally) is profitable. …

Professor Hanson … addresses this problem in three ways.  First, he constructs a model in which Y is infinite and X is finite. … Second, he does sociological experiments with undergraduates. He sets up these markets such that Y is greater than X. …  Third, he finds actual markets in which actual manipulations fail. …  None of this goes even a millimeter toward proving what needs to be proved – namely, that in all markets, Y is always greater than X. …

Honestly, I think the greatest difference between my perspective and Professor Hanson’s is just that I have much higher standards. His entire argument proceeds from the position that, since government today is so bad, anything that could be somewhat less bad is worth a look. … Don’t people buy decisions now? Well, gee, they sure do. So there you go.

For me, government safety is like airplane safety. Not only do I want a watertight proof that Y is greater than X, I want two or three parallel and independent proofs. At least one of them will probably turn out to be wrong. Professor Hanson is a professor, and thinks like a professor. I’m an engineer, and think like an engineer.

I am also a student of history. … the European writers of the Victorian era, whose aristocratic governments worked much better than ours, and were thus appalled by government failures which to us seem trivial and not worth mentioning.

In the debate, I suggested we start by trying fire-the-CEO markets, and only gradually rely more on them in CEO decisions as such markets collect good track records.  Moldbug seems to accept wide trading in ordinary stock markets because he doesn’t think any decisions depend on them, but strongly advises against allowing non-employees to trade in fire-the-CEO markets, due to manipulation concerns.  But even a track record showing that firms which followed market advice do better on average than firms that do not would not persuade him.

In fact, Moldbug the “engineer” says no data anyone could collect in the lab or in any organization smaller than a nation would be relevant, and even with nations he doubts we’d see hidden manipulation. Nor does any data collected in the last century test his belief that the best governments are single rulers running city-sized polities with iron fists and complete discretion.  It is not even clear what prior data makes his case – apparently it can’t be summarized in any concise form; you have to just read dozens of books and have a feel for it.

Not only does Moldbug know such iron fists would rule best, allow emigration, not cheat their investors, and never ever accept manipulator payola, he apparently knows this deductively, as a noble philosopher, not like data-addicted corrupt pansy social scientists.  And he has no interest in improvements in the status quo below his philosopher-deduced-best pinnacle.

What more can one say to such a person?

GD Star Rating
Tagged as: ,
Trackback URL:
  • Carl Shulman

    “Not only does Moldbug know such iron fists would rule best, allow emigration, not cheat their investors, and never ever accept manipulator payola, he apparently knows this deductively, as a noble philosopher, not like we data-addicted pansy social scientists.”
    Why did you debate him anyway, given the unsatisfying previous blog exchange?

    • Josh Hall, conference organizer, at first offered to pay travel expenses so I could speak on econ of nanotech and AI. Then he requested I also do this debate. The additional time required was small, and I do love to argue.

  • “his main complaint is manipulator payola”

    Because current governance institutions are so resistant to manipulation by special interests.

  • psychologist

    Can’t wait to watch the debate. I’m always amazed when my disparate interests come together.

  • He seems to have imbibed some of his preference for deduction over empiricism from Ludwig von Mises. However, Mises claimed to have built a logical deductive system on top of apodictically certain axioms known to be true a priori, most famously “man acts”. Mencius, however, never makes clear any of the axioms on which he bases his deductive field of politics/governance. Lately, he has shifted away from the rationalism of the Austrians to the anti-economics reactionary romanticists such as Carlyle. But while Carlyle deployed much mockery against economics and the use of statistics in governance, I don’t think he made the claims to deductive certainty Mencius does. On the other hand, Carlyle makes my eyes glaze over so I can’t be certain.

    Mencius has also admitted an exception to his rule of benevolent iron-fisted exploiters: King Leopold’s Congo Free State. He argues that was a rare deviation which improved when it became the Belgian Congo and that post-colonialism has an even worse track record. Suffice to say, that is not a deductively certain argument!

  • Moldbug or Modlbug?


  • Listening to the debate, I think David Friedman wins.

    Robin, you have said that Intelligent Design (or the existence of God generally) is unlikely but is dismissed to an unwarranted extent by experts/academics. Would you say the same is the case for Mencius’ theory that monarchy/neocameralism is superior to democracy?

    • It is far from crazy to suppose autocracy makes for more peace and prosperity than democracy, and yes that view is unfairly dismissed in academia. But we can and should look at data to see. For example, I recall recent democracies seem to have less income variance.

      • josh

        Which autocracies and democracies would you suggest comparing? Are there any modern autocracies based on the old concept of sovereign property with similar populations to modern wealthy countries? Perhaps Lichtenstein or Monaco; but even these have adopted democratic features over the past century. Further, how do you control for the fact that these countries are very, very weird by definition in today’s world.

        We really may just have to read dozens of old books and get a “feel” for it. The “goodness” of a society at some level is going to be an aesthetic preference no matter how much you try to quantify it. That doesn’t mean data is useless, its just that it it requires so much interpretation that the right old books may provide as good or better understanding. In the end, we have to use our own judgment no matter what evidence we are working with.

      • @Josh — Both Hong Kong and Singapore operate under the principles you state; Hong Kong does not sell land, only lease it (on fair terms for up to fifty years) thus preserving the idea of sovereign property, Singapore is nominally ademocracy, but there is only a single party who’s rule everyone seems content with thus far. I consider it a prime example of the democratic equivalent of “market dominance is not monopoly”

      • Vladimir

        Robin Hanson:

        It is far from crazy to suppose autocracy makes for more peace and prosperity than democracy, and yes that view is unfairly dismissed in academia. But we can and should look at data to see. For example, I recall recent democracies seem to have less income variance.

        Do you really believe that it’s possible to control for all the confounding and lurking variables in a way that would make such a comparison meaningful at all? This even assuming that enough good data is available, which is clearly not the case. Even just defining “democracy” and “autocracy” in a meaningful enough way for such a comparison is a hopeless task.

        A study that lumps together, for example, present-day U.S., Japan, Bulgaria, Iceland, and Mexico into one class called “democracies” and then proceeds to draw general conclusions about the problems and benefits of democracy shouldn’t even be allowed to pass the laugh test.

  • botogol

    Robin, you will be interested in this – derivatives with political outcomes as derivatives, a growing market. This isn’t so far from your ideas of how to futarchy ball rolling.

    Derivatives of Politics

  • Aron

    I considered building a market/neural-like collaborative chess website once, but decided very few would actually enjoy such a thing. I wasn’t about to subsidize them for playing. Plus the feedback mechanism to train the weightings of the participants seems rather tricky. Rather than mechanize the process, it would probably work better to let people figure out who they want to play with, and how to negotiate the weightings of their suggested moves between them. After all, people can be pretty good at this already.

  • Robin, you mean you didn’t debate for our entertainment? 😉

    I can’t wait to see the debate, although from yours and Patri Friedman’s accounts, I fear Moldbug’s debating skills did not live up to my expectations.

    I’ll have to read more about fire-the-CEO decision markets; strikes me as an excellent way of making shareholder decisions, although simple matters such as “is this CEO doing a good job or not” are far easier to understand and predict the implications thereof than some of the issues you suggest tackling with futarchy.

    But yes, the lack of empirical evidence re: the success of autocracy is a damning indictment of Moldbug’s neocameralism. I personally find his deductive arguments compelling, even though the inductive argument relies on novel subjective judgements of historical cases.

  • Harold

    Regarding Steve vs. the chimp, wouldn’t Google throw its financial weight behind the chimp in the last millisecond before the market closed? Sorry if there is an answer to this anyone with even a modicum of knowledge of such markets would know.

  • Jess Riedel

    I’m confuse. Was the debate about futarchy (vs. a republic) or about the best form of government in general? If the former, why was anyone talking about autocracy?

    • The debate was just pro vs con futarchy. I think one should compare a proposal to the status quo, but he preferred to compare it to his favored option.

  • tom

    Is it true that college student and other volunteer studies are a significant part of the proof the prediction markets work?

    I’ve been thinking about those types of studies this week because of (1) the quote from Moldbug above, (2) the Washington Post article on compassion in the “Telescope Effect” post on Tuesday and (3) Steve Sailer’s speculation on his blog on Saturday that black college students gamed studies on the ‘stereotype effect’.

    I have doubts about how much study participants accept the ground rules of a study. For example, the Post article discusses how volunteers responded to questions about helping poor Africans and concludes: “Even when people are absolutely certain their money will not be wasted and will be used to save a child’s life, fewer people are willing to write the check than to leap into the pond [to save someone].” I really don’t believe that’s what is happening. Even though they were assured that they would be helping by sending money, I believe that participants discounted the value of sending a check because they don’t believe it is effective. Saving a drowning person right in front of you may be risky for you, but if it works, the result is clearly one life saved. Sending $1000 to Haiti may not really help anyone change their lives. They are totally different things. But the author accepts the results at face value. And Prof. Hanson accepted them too in his post, for purposes of thinking about near and far modes.

    The Sailer post (, raises a different problem with studies. Participants may have tried to help the researchers get the result they wanted.

    I think participants in lots of studies reject the stated premises and have the meaning of their responses misinterpreted as a result. I’m sure that Prof. Hanson tries to work around that. But isn’t that a huge problem?

    • Our experiment subjects are usually college students, but they have no idea what thesis we are testing and they are paid cash, more or less depending on their choices. So I believe they just try to make as much cash as possible.

      • tom

        I can see that volunteer studies may help confirm generally that manipulators may not poison a small market. But part of my comfort is that I think the burden of proof is pretty low to justify tests of things like CEO departure markets.

        Haven’t you thought that there are huge limits on the value of volunteer studies here when the incentives are necessarily so small (I’m guessing nobody ever got rich winning in your tests, or even paid his rent for a month) and so different in character from those of participants in a market with meaningful amounts of money?

        If people can make a lot of money in your studies, please sign me up. I may even see if Moldbug would pay me to act a certain way.

  • Lord

    Don’t bomb me?

  • Mencius Moldbug’s position seems crazy – he is arguing for the power of individual vs collective intelligence. IMO, the correct critique of Futarchy is that is a poor quality form of collective intelligence where the participants all fight with each other.

    Collective intelligence *usually* works best if the agents mostly cooperate. That is mostly how the brain works, for example. It is *not* a bunch of self-interested neurons fighting with each other.

    • ad

      Are you arguing against futarchy or democracy? Politicians in a democracy certainly fight against each other.

      On the face of it, traders in a futarchy merely bet against each other.

      • Futarchy. By “fighting” I just mean “competing” here. I made no mention of democracy.

    • nazgulnarsil

      I think various democracies with their checks and balances vs joint-stock corporations with a far less complicated decision hierarchy is a better comparison than to the brain.

    • mjgeddes

      Betting markets of any kind can easily be beaten by manipulation, simply by applying collusion. Take poker as an example. It’s only a fair game if each agent bets individually, but when two or more agents coordindate their betting they can easily rip off the other players. Agents simply have to coordinate to over-power any market and rip everyone off with ridicious else.

      Many on transhumanist lists seem to be so utterly blinded by Libertarian dogma that they have lost touch with reality. For example I see AI deisgners proposing ludicrious theories of mind in which the mind is treated as ‘a mini market’ for instance.of trading sub-agents. I tell you all again, Bayes is a bad basis for a rational foundation and AI designers which try to use prediction market-like systems (e.g CEV) are headed a tremenous beating.

      The reasons have been touched in Robin’s earlier thread. Not every infomation processing agent should be given equal attention so there needs to be a way of deciding ‘which sub-agents should be heard’. This deeper system is a system of coordination of agent behavior. As I mentioned, manipulating the attention given to each sub-agent totally warps the probabilistic reasoning of the system because of bias (missing info, selection effects etc).

      The deeper system of agent coordination is analogical inference, the real basis for mind.

    • Aron

      “poor quality form of collective intelligence”


      Are two people going to make a better decision if they write bets down on paper and reconcile the results with a market mechanism, or if they discuss the naturally fuzzy problem as people usually do? How about with 10 people, 100 people? Does any system enable 1000 people to make better decisions than 100 (once the participants are all moderately selected experts)?

      As the group size increases people stop talking face to face and fall into hierarchies, mechanisms for booting out uninformative people kick in, and various conscious and subsconcious mechanisms of how to participate in such an arrangement appear. People become constrained in what they can say, who they can say it to, etc. It’s no longer artificially simple, sadly. I suppose that’s the downside.

      • mjgeddes

        The irony is that ‘Less Wrong’ (with many enthusiastic Libertarians) deploys a democratic voting system to ‘direct focus of attention’ to sub-agents supporting LW dogma, which actually runs contrary to the ‘free market of ideas’. This voting system awards points (Karma) to order to perform categorization of sub-agents. Undoubtably an AI would have to deploy something similar, and it would no doubt take precedence over Bayesian inference.


      • Peter Twieg

        So does this argument that we should replace Wall Street with two people trying to work out what asset prices should be?

        Which two people?

  • Felix

    There is an easy and “canonical” way to solve the decision market manipulation problem. Suppose we want to decide between a number of options. Then create a contract for each option that obliges to pay all harms and entitles to receive all benefits that come from taking that option. Then make an auction where the winning bidder can decide which of these contracts he wants to get. Then decide in favor of the option that the highest bidder wanted. This way, it makes sure that when a bidder wants to “manipulate” in favor of an inferior option, he has to take all the losses that come from this.

  • Can there be a historical analysis of the approximate effects of the “oust the CEO” market and of applications to this to government ?

    1. A majority of Shareholders can enact changes to the board of directors and thus indirectly achieve control of who is CEO and whether to fire (it is difficult but possible)

    2. There is the California voter recall system to recall a Governor.

    Successful recalls
    1916 recall of J. W. Robinson, mayor of Boise, Idaho
    1921 North Dakota recall of Governor Lynn Frazier
    1983 recall of Michigan state senators Phil Mastin and David Serotkin due to their support for a state income tax hike. Loss of these two Democratic lawmakers, along with two special elections won by Republicans, flipped the state senate to GOP control, where it has remained ever since (as of January 2010.)
    1987 recall of Mike Boyle, mayor of Omaha, Nebraska.
    1994 recall of officials in River Vale, New Jersey: Mayor Walter Jones, Councilwoman Patricia Geier, and Councilman Bernard Salmon
    1995 recall of California State Assemblyman Paul Horcher
    1995 recall of California State Assembly Speaker Doris Allen
    1996 recall of Wisconsin State Senator George Petak
    2002 recall of Woodrow Stanley, mayor of Flint, Michigan.
    2002 recall of multiple Milwaukee County, Wisconsin, elected county officials including Executive F. Thomas Ament (resigned before election); Board Chair Karen Ordinans; and Board Supervisors Penny Podell, LeAnn Launstein, David Jasenski, Kathy Arciszewski, James McGuigan, and Linda Ryan. All were recalled due to a retirement pension controversy.
    2003 recall of Gray Davis, governor of California
    2003 recall of Wisconsin State Senator Gary George
    2005 recall of James E. West, mayor of Spokane, Washington.
    2006 recall of Neil Marko, mayor of Roosevelt, New Jersey.

    Unsuccessful recalls
    1978 Cleveland Recall Election of Mayor Dennis Kucinich
    2008 recall of California State Senator Jeff Denham and Paula Flint

    Unsuccessful attempts to qualify recall elections
    United States Senator Frank Church of Idaho was the subject of an unsuccessful recall effort in 1967. Courts ruled that a federal official is not subject to state recall laws. See also the similar unsuccessful effort in 2009 to recall Anh “Joseph” Cao, U.S. representative for Louisiana’s 2nd congressional district.
    Evan Mecham, Governor of Arizona, was scheduled for a recall election on May 17, 1988 after a successful petition drive (301,000 signatures). However, the Supreme Court of Arizona canceled the election, since Mecham had already been impeached and removed from office by the Senate on April 4
    In 2009 a petition failed to garner sufficient signatures to oblige an election for recall of Eddie Price III, mayor of Mandeville, Louisiana.
    In 2009 a petition for recall of Stacy Head, New Orleans city councilwoman, likewise failed to gain the requisite number of signatures.

    3. I will repeat my claim that there are ballot propositions (made at the debate -the point was dismissed) which mimick Futarchy. Financial issues like bond measures. In California, bond measures can have thousands of dollars per year impact on each voter. This is far more than I have ever gambled in Vegas. I have not directly invested in the stock market for decades (buying individual stocks) but use mutual funds. I have far more skin in the game of California money related propositions. Also, there were several measures that directly relate to a developer getting or not getting permissions to do something that would be worth a lot of money to them. There is a bias to vote down financial measures that spend money.

    I put forward that analysis of relevant measures would be far more indicative than the CEO firing test.

    the interesting analysis would be trying to project whether California ended up better off with Arnold or would have done better with Grey Davis.

  • Hugh

    “For me, government safety is like airplane safety. Not only do I want a watertight proof that Y is greater than X, I want two or three parallel and independent proofs.”

    If you’re on a modern airliner, moving to a aeroplane designed by a social scientist that is statistically pretty good at staying aloft is a bad idea. But if you’re currently on a plane whose main certification is that people have a “good feeling” about, changing might be a good idea.

  • rob

    Assume Y is less than X. Who is taking the other side of the X trade and what are their motivations? A side-market would also function as a market. Arbitrage between markets would make it such that it doesn’t matter if Y < X.

  • sconzey, in the linked Moldbug post Hanson is quoted as saying “We’re both entertainers”. So maybe it wasn’t JUST for our entertainment.

    Harold, I believe Hanson has said that the price has to stabilize around a certain range, so you can’t disrupt it at the last minute without anyone being able to correct it.

    tom, because lab experiments are limited Hanson also bases his case on field data regarding attempts to manipulate larger markets.

    Tim Tyler:
    what’s a cooperative alternative to futarchy?

    That is mostly how the brain works, for example.
    That reminds me of a bit in Bertrand de Jouvenel’s “On Power”. Herbert Spencer used biological/anatomical analogies to justify quasi-anarchist libertarianism. But of course our bodies operate more along the lines of fascist principles!

    individual vs collective intelligence
    Oddly enough, I think that widely-linked critique of the Phantom Menace helps make the point. When George Lucas had complete control, he did a much worse job than when he was hampered and had to compromise with others.

    Brian Wang, recalling governors is an interesting analogue to firing CEOs. However, David Friedman’s point about how bidding wars are different from decision markets remains.

  • From an engineering perspective, Moldbug is correct. If you have a logical problem in your software design – ie a problem you have diagnosed via deduction – you do not ship that code. Even if the code is working on your staging environment, you do not ship it. If you accidentally release the code, and it works, you still fix the logical problem. You have to assume that you are just getting lucky and that it will break at the worst possible time.

    So even if futuarchy worked on small scale systems that would not convince me that it was a good design. It would both have to work, and the actual working of it would have to reveal flaws in my logic and deduction.

    Of course, if you have and idea that works via logic, you must also test it in the real world, in a safe contained environment before releasing it at large scale.

    So neither deduction or experiment results are enough alone. For any well engineered system, you must have both.

    It is not even clear what prior data makes his case – apparently it can’t be summarized in any concise form; you have to just read dozens of books and have a feel for it.

    This is how actual management works in the real world, when extensive data is not available. And in the case of both futurachy and Moldbug’s formalism, good data is not available, so “feel for it” is all you’ve got right now. Also, books are a form of data, a much richer form of data than straight statistics where complicated problems have been reduced to a sterile number. First hand accounts are also far richer than lab experiments with college students.

  • Prakash

    To draw an analogy, I think mencius is arguing in favour of horse driven carriages because they have worked well in the past and the present era is the earliest days of automobile (algorithmic government) design. I’m pretty sure that the earliest cars were also pretty bad compared to good carriages.

    The horse is organic. The monarchy with the big man feels natural to most of humanity.

    The horse is tried and tested technology. The automobile is relatively new. Some have done it well, most are not doing it that well.

    There are limits to a horse’s power or even multiple horse’s power.
    Similarly, there are always limits to the sovereign’s thinking power. A well designed automobile will beat any number of multiple horses one day, but we don’t have such a design yet.

    I think Friendly AI might be able to create such a design.

  • Hanson, Moldbug, most of the commenters here – you are all would-be fascists, technocrats, nerds clutching for power. The idea of either of them or any of you having any power over anyone is truly repulsive.

    “Friendly AI might be able to create such a design” Typical.

  • Adam Wallace

    “Nor does any data collected in the last century test his belief that the best governments are single rulers running city-sized polities with iron fists and complete discretion. It is not even clear what prior data makes his case.”

    It’s nice to hear you actually go on the offense and criticize his favored system, however, your criticism isn’t necessarily relevant to his claims against Futarchy (which can stand on their own against your system, albeit critically and not constructively). You also haven’t proven why inductions are better that deductions, which, by nature, deal with rules of a generally applicable nature (if not, then the logic becomes unsound). During the debate you were too busy defending your own views to really investigate his. If you actually looked into how he defends his neocameralist formalism, he does cite plenty of historical examples of his system working out. It’s not hard to think of them. Off the top of my head, Monaco is a great one. Singapore is another.

  • Pingback: Alt-Right: USA’s nye radikale højre – Critique()