Randomized Governance

What if instead of electing our representatives in government, we simply chose them at random?
A new Rasmussen poll asked 1,000 likely voters exactly this question. Turns out, 43% thought that a random choice of people from the phonebook would do a better job than the current legislators, a plurality. Of course, these people were themselves chosen randomly from a phonebook, so I’m not sure they are entirely unbiased. 🙂
But why stop at the legislators? Why not just write random legislation using context-free grammars? We already have software that can automatically write scientific papers, so it doesn’t seem like a stretch. I guess that a lot of this random legislation would be better than SOPA.

19 Replies to “Randomized Governance”

  1. If legislators will vote according to the will of the majority (or else not be elected), wouldn’t random legislators make the same votes in expectation? Both would be equivalent to simply polling the general population on every issue, i.e., direct democracy. Random legislators would be more efficient, however, than polling on every vote.
    Your post suggests, however, that you don’t like the idea of rule by randomly selected people, which implies you don’t really like democracy. Perhaps you prefer “rule by the best” (a.k.a aristocracy)?

    1. This is a good idea, but I think you need to symmetrize a little further. Once I’m randomly chosen to be a Senator, I can pass the “Give Aram Harrow a billion dollars” bill. Somehow you want people voting as though they were behind a Rawlsian veil of ignorance.
      On the other hand, maybe it would suffice for each term of office to be a year, with no bills going into effect until the following citizen-parliament had a chance to repeal them. Of course, one could pass the “give all members of the 151th and 152nd senates a billion dollars” bill, so even this approach is tricky.

    2. Yes, the expected representation is indeed fair in this system, which is a nice feature. I definitely like that aspect of it. It would certainly boost representation of poor people in the US, for example.
      On the other hand, one must concede that there are times when the majority are simply wrong, perhaps due to the fact that some issue requires specific technical understanding, or maybe due to long-standing prejudice, e.g. civil rights. I think that civil rights is an important case to keep in mind. There is a quote about how democracy is a pack of wolves and a sheep voting on what’s for dinner. We would need to be vigilant on that front. Another problem I can think of is that the timescales involved before the sampling was nearly uniform might be longer than the lifetime of the constituents on some issues.
      One should be able to do much better than randomized representation in practice. Representative democracy is intended to do exactly that, right? If we subdivide the population and have them select their best and brightest to represent them, it seems like it should improve over the randomized case. Of course, there are all kinds of loopholes here: how you do the subdivisions (susceptible to gerrymandering), how you select the “best and brightest” (politicians are self-selecting on traits that are not necessarily aligned with representing the best interests of their constituency, to say the least).

      1. “one must concede that there are times when the majority are simply wrong”
        There are also times when politicians are wrong. To prevent the worse abuses, we rely on a legal framework. The same applies with a demarchy.
        This is how we use a jury in criminal cases. And yes, black people end up in a court of law all the time.
        Are you saying that the politicians are somehow more enlightened? Well. Consider: People who have been selected randomly to represent the interests of all tend to have less of an entitlement feeling. Yes, they will be prejudiced, but so can be politicians. The difference is that the politician *thinks* he is smart and entitled to lead. So he will impose his vision on others.
        Think about it. You have been selected, along with say, 12 others, to lead the USA, as co-presidents. Everyone is watching you. Everyone knows you got there through pure luck, there is no illusion. What are you going to do?
        Are you going to go all crazy? Probably not. You will probably (hopefully) take your job very seriously. Would a council of 12 people randomly selected take worse decisions than Obama? Hard to tell, but it seems likely that these randomly selected people are less likely to be susceptible to lobbying from powerful groups.
        Maybe, just maybe, if you have been randomly selected, and you know you’ll have your name in the history books, you’ll try to take the very best decisions you can without feeling too smug about yourself.
        And think that whatever you do, it is only temporary. It is not like you and your friends are in power for long. So if you have new ideas that you want to push through on long run, you have to convince a lot of reasonable people.
        It seems to me that the only major risk is that people will *not* agree and they will just stall. So demarchy might lead to a government that ponders every decision for a long time.
        This would probably please libertarians.

      2. Civil rights is indeed a good example. If the majority doesn’t like some minority group, then the majority is free to screw them over. Some minorities are now protected by law, but certainly not all (e.g. gays).
        Another one that comes to mind is from the first democracy in classical Greece: ostracism! Every year, they would vote on whether to ostracize someone. If yes, they would all vote for a person and someone (the plurality?) would be ostracized. He would be thrown out of the city and presumably all of his property taken away.
        And as you said, the majority can simply be wrong. If you want to see some examples, read The Myth of the Rational Voter. His thesis is that the reason democracies have bad policies is that the majority of voters want bad policies. For example, the majority is anti-immigration and pro-tariffs despite unanimous disagreement from economists.
        As to trying to select the best and brightest, I’ve never seen any evidence that the majority will accept any policy that seems intuitively wrong to them, regardless of how many experts say that is the right policy. See the examples above.
        I think this is just how democracy works, whether direct or representative. It’s a system that enacts the will of the majority. It ensures that the majority is not screwed over, but it allows the majority to screw over whomever they want. It improves upon aristocracy and any other political system, though, in that guarantees the fewest number of people are screwed over. But it certainly doesn’t protect everyone.
        If you want to see an example of how a political system can allow almost everyone to be screwed over, look at North Korea.

        1. Greek ostracism was not random though. It was more of a referendum, something which is common to this day. Of course, it would be kind of odd today to hold a referendum on the banishment of an individual for 10 years, but it appears that the president of the USA has now the power to order a death sentence regarding a citizen without any court proceedings. So ostracism was hardly very harsh all things considered.
          And yes, the majority tends to be anti-immigration is western democracies… but so are the governments and the politicians! The “wall” between Mexico and the USA wasn’t built by a crowd, but by the order of politicians. The TSA is not searching people because the average people asked for it: it came through political decisions.
          It really comes down to whether a ruling class is necessary. Though we have abolished monarchies (mostly), we are still under the rule of a class of politicians. There is a lot of rationalization for why this is necessary…
          Some time ago, I wrote a blog post on this topic:
          Demarchy and probabilistic algorithms

      3. I’m not sure civil rights is a good example. In the US, the problem is that the majority of the nation was in favor of civil rights, but the majority of people in the South were not. So this is perhaps an issue of federalism, but not of democracy.

        1. The majority’s opinion changed with time. I’m sure there was a large period of time between emancipation and the passage of the voting rights act where the majority did not approve of voting rights for african-americans. (Lincoln himself made some questionable statements about this issue.)
          In any case, the fact that voting rights did not pass even when it had support from the majority suggests that our democracy is not even as good as direct democracy, at least in this respect.

          1. The majority also approved of slavery at one point. But the question is whether elected representatives are generally more enlightened than public opinion at the time.

  2. @Daniel, I definitely *don’t* think politicians are “more enlightened”. However, I do believe that some people are more capable than others. Generally speaking, I think it is better to have those people in positions of responsibility and leadership.
    The basic reason that I think representative democracy should perform better than true democracy is that it is easier to recognize when someone makes good decisions rather than to make the good decisions in the first place. (Classic P-vs.-NP-type situation!) Having people elect their representatives ought to allow the cream to rise to the top.
    Let me again reiterate the caveat that I mentioned above: politicians are self-selected for reasons that are *other* than the best interest of their constituencies. Am I right that you are asserting that in a demarchy the citizen-politicians would be less susceptible to serving their own interests? I’m skeptical… see Aram’s comment, as well as kczat’s comments about immigration and tariffs.
    What are the best modern examples of demarchy? I feel like I need a lot more empirical input to decide how well a demarchy would work in practice.

    1. “However, I do believe that some people are more capable than others.”
      Let us be precise. Given a task, some people are more capable than others at it. If you are going to pass a bill which aims to regulate technology, I don’t think the members of congress are more able to assess the bill than I am. If you want to help struggling entrepreneurs, maybe a struggling entrepreneur will have better insights than a billionaire who made his way in congress.
      This being said, I am not advocating that anyone be elected with equal probability. You know how the slashdot moderation system works? It is basically a demarchy, but not everyone is selected with the same probability. I don’t know the exact algorithm they use, but you can think that people who have been long time members and who posted comments that have risen to the top are more likely to be selected as moderators.
      I imagine a system where you earn karma for doing your civil duty… and then you have *very small* probability of being selected for office. And when you are, it is only for a short time, with limited powers. Like jury duty.
      “Am I right that you are asserting that in a demarchy the citizen-politicians would be less susceptible to serving their own interests? ”
      Yes. I am. The reason is that they are selected for civil duty. Meanwhile, a politician “earned” his position.
      There is also a matter of moral stature. A politician can always say “I have a mandate to do this, I was elected”. Someone who was randomly selected cannot say the same.
      “What are the best modern examples of demarchy? I feel like I need a lot more empirical input to decide how well a demarchy would work in practice”
      The thing is, we don’t know… but before Wikipedia really took off, many learned people predicted its immediate demise.

      1. +1 to this portion of the thread which says much of what I’d say on this subject (both Steve and Daniel).
        It’s interesting to ponder exactly how to create systems of governance in which we eliminate much of the current crazy incentives from our current systems. Chance certainly adds something to the conversation, as do systems that reward past behavior. But it still feels to me like there are things missing from this mix. As Daniel says in another portion of the comments, democracy feels like a good local optimum, but its certainly never been clear to me that it’s the global optimum, especially as the world changes (I’ve previously argued, for example, that the scale invariance of our governance is broken due to the large growth of our population with a similar scaling in the properly optimized governance hierarchy.)

  3. And one last comment… Despite any flaws we may find in either representative democracy or demarchy, I think Churchill’s quote is appropriate: “It has been said that democracy is the worst form of government except all the others that have been tried.” When I look at e.g. North Korea, as kczat suggests, I definitely count myself lucky.

    1. “It has been said that democracy is the worst form of government except all the others that have been tried.”
      This is the local extrema fallacy.
      Go back a to 1500. People could say “It has been said that monarchy is the worst form of government except all the others that have been tried.”
      Go back to -1500. People could say “It has been said that tyranny is the worst form of government except all the others that have been tried.”
      And so on.
      Basically, we have found something, and by tweaking a it a bit, we can’t improve it much. And all other solutions tried are worse. So we conclude that we are done and that we have the end of history.
      But in 20 years, some country could adopt a political model vastly different from what we know and they could end up living in what is, compared to us, an utopia.
      Of course, may be this is not true, and we have, indeed, reached the end of history and representative democracy is as good as it gets. We do, indeed, need a ruling class made of politicians. We have no choice.
      But, still, maybe it is not true. Maybe we could do ten or 100 times better without a ruling class… the same way wikipedia is vastly better than any encyclopedia that came before it.
      Go back to 1990 and tell people about wikipedia, they will tell you that you are mad. No such thing is possible. Yet it came to be.

  4. @Daniel, It seems we agree that some filtering will improve the results. I think your idea of “slashdot”-like (or mathoverflow-like) reputation points to lend greater weight for selection is a cool idea.
    If there are no nations which are demarchies, then do you know of professional societies or something similar where your ideas have been implemented (even approximately)? That would be the next natural place to look. A similar situation has occurred with various alternative voting systems. They have been implemented in some professional societies and this can offer an empirical window into how they perform in practical situations. Perhaps the same is true for some variant of demarchy.

    1. I’m unaware of the use of demarchy in a professional society. Well. The Amish use it with great results: http://vimeo.com/25005520 Does that count? Mind you… the idea of allowing women to vote was similarly challenged: nobody does it so it can’t be done. Someone, somewhere, must start.
      I can see a reason why it might not be interesting with a professional society. In most cases, there is no ruling class to begin with. So your really want to replace professional politicians. Of course, can you guess *who* is precisely never going to support demarchy? Politicians! So the reason we rarely see might have little to do with effectiveness and a lot to do with the difficulty to make it happen. You really need a major constitutional overhaul with politicians working against politicians for it to happen. Do you see either the republican or democratic parties in the USA proposing this? They would be effectively proposing their own retirement.
      Of course, how did democracy come about? At least in recent history, we had crazy Americans, savages really, who decided to do away with monarchy… then the French followed suit with a bloody revolution while the Brits slowly migrated to a similar system… it really was not easy, so I don’t suppose the path to demarchy can be easy.
      BTW this is not “my idea”, just an idea I’m trying to popularize. Google “sortition” and “demarchy”. There are entire organizations dedicated to this idea. Books have been written (e.g., by Brian Martin). It is typically dismissed as ridiculous.

  5. Just about the only thing I ever (almost) agreed with that William F. Buckley said was that he’d rather be governed by the first 500 names in the NY phone book than the faculty of Yale (or was it Harvard?). Pretty funny line. 😉

Leave a Reply

Your email address will not be published. Required fields are marked *