Part 1 is here.
6. Random sampling will occasionally produce unrepresentative samples.
Significant deviation of a sample from the population sampled is in fact very rare. For example, in a population evenly split between men and women, the chance of having fewer than 40 women in a sample of one hundred people is less than 2%. The chance of having fewer than 30 woman is less than 2 in 10,000. And the chance of having 20 women or fewer is less than one in a billion. The current U.S. Senate (a body of 100 people) has 20 women. It is the highest number of women senators in U.S. history.
7. Since there are many population characteristics, the sample would be unrepresentative according to some of those.
Again, because the chance of significant deviation is so small, even if many characteristics are considered the chance that any of them would show significant deviation is small. For example, over one million characteristics would have to be considered before it would become likely that a group which is a minority of a third according any of those characteristics gains a majority in a sample of 200.
8. The lucky few who are selected will often serve personal or narrow interests rather than those of the people.
For policy to be approved by the allotted body, it would have to win a majority. Interests that are personal to one or to a few delegates would not be able to meet this criterion. By the time a proposal wins a majority is has to serve so many personal or narrow interests that it becomes representative.
There are special situations where the allotted body, as a body, has a narrow interest, such as when considering the compensation for the members. For those special situations some special arrangement needs to be set up. For example, such issues may be handled by a separate body or reviewed by an independent jury.
9. Many people would not want to put their lives on hold for a few years while they serve their allotted terms. If only some of the those sampled actually serve the sample is no longer statistically representative.
This is a question that should be settled by experience. It would certainly be a problem for sortition-based government if allotted seats are not generally accepted. The logistics of service should be set so as to be as inclusive as possible. In particular, the compensation for service should be generous enough so that for the large majority of people service would likely be financially rewarding. In addition, an attempt should made to accommodate specific needs that stem from the circumstances of the allotted.
It may very well be that, given the opportunity and generous compensation for their time and effort, few would give up the chance to devote a few years of their lives to changing their society for the better (as they see it), and at the same time to win the appreciation and respect of those who know them (and possibly of many other people as well).
10. Even among those who accept the nominations many would not put much effort into studying the issues facing the nation. They will not be able to come up with informed and considered decisions.
This is a version of objection #2. First, the assumption that the average person will not put much effort into their public service is not substantiated by the facts. Normal people who find themselves in a position to improve their community often go to great lengths to do so. But, again, even if we accept the negative assumption and take it for granted that the allotted delegates will not put much effort into their public service then it is still clear that their special position will allow them to reach much better informed and considered decisions than voters do.
Adam Cronkright
Ali
Ahmed R. Teleb
avanderven
avergne
Alex Kovner
azakaras
Brett Hennig
campbellwallace
Simon Threlkeld
Nick Coccoma
codyhipskind
Common Lot Sortitionist
Conall Boyle
constitutionalism
davidschecter
graytmatter
gtridimas
H.J. Hofkirchner
John Burnheim
jgarry20151
johnedwardtaylor
johnhuang865
Jonathan Crock
keithsutherland
Jacob Richter
lukewhito
SimonPek
Nathan El
Nicholas Gruen
Owen Shaffer
Patrick Chalmers
Paul Gölz
paulnollen
peterstone
Pierre-Etienne Vandamme
pronfeld
Emil
rcaze
richwbrown
Roger Knights
roslynfuller
Pierre-Etienne Vandamme
shaharavin
tbouricius
Ted Wachtel
Joshua Laferriere
theodorestathis
Wayne Liebman
Yoram Gat
Oliver Milne
Equality by lot 
>”For policy to be approved by the allotted body, it would have to win a majority. Interests that are personal to one or to a few delegates would not be able to meet this criterion.”
If policy trading is permitted then the nonrepresentative interests that every member holds will contaminate the overall results. It might only take a few votes to change the results of a vote on a major bill. It’s not that hard to buy a few votes.
Now, I have no problem with policy trading as a general rule. The problem is that when a trade involves nonrepresentative interests the overall results become nonrepresentative. Let’s say we have a general purpose national level allotted assembly. On the national level it’s members should, at least in voting on single issues, represent the people well. At the local level, not so much. There are around 30,000 towns in the US. I suspect most members will have some local issue they would like to see addressed. The pursuit of local issues (just like the pursuit of petty personal issues) by some members is very much nonrepresentative and the people harmed at the local level by the pursuit of a local issue are not likely to be represented in the sampling at all.
Why should the other members care? Agreeing to address the local issue (or the petty, self-serving, arguably corrupt issue) does them no personal harm. In fact, it brings them closer to achieving their own goals if the local issue is addressed as part of a policy trade deal that includes the things they want. However, the people harmed in the trade have no one to argue on their behalf. They have no representation in government. There’s no clear dividing line between representative interests and nonrepresentative interests and deals will no doubt include both. Ideally they should only be made between interests large enough to hold reliable representation in the sampling.
There’s nothing sinister about addressing local concerns. In fact, I would argue that is is very much necessary. Any national level policy is going to have local consequences. But since the representation of such interests in the sampling is purely random, we have a problem. I’d be tempted to suggest an aggregative advocacy as a natural solution.
LikeLike
> Why should the other members care? Agreeing to address the local issue (or the petty, self-serving, arguably corrupt issue) does them no personal harm. In fact, it brings them closer to achieving their own goals if the local issue is addressed as part of a policy trade deal that includes the things they want.
It may very well be easier to promote the general interest rather than somehow generate policy that promotes the interests of hundreds of localities at the expense of all others. Besides, blatant exploitation of power for narrow interests may result in a personal price to pay in future litigation.
That said, I agree that if the prevailing attitude in the population is a no-holds-barred readiness to exploit any opportunity for personal or narrow gain, then sortition-based government – and indeed any government – cannot be realistically expected to generate good policy. (I referred to this situation as a state of low background representativity.)
The indication is, however, that despite the hostile environment in many modern societies, background representativity is rather high in many of those societies.
LikeLike
Yoram,
>”It may very well be easier to promote the general interest rather than somehow generate policy that promotes the interests of hundreds of localities at the expense of all others.”
I’m not necessarily sure we can separate the general interest from the interests of various subdivisions of the population (all the way down to the individual level) in aggregate. On this note, I’m a big believer in policy trading. I suspect that policy trading can get us closer to an ideal set of policies than taking every issue one at a time. People will give up the things that don’t matter much to get what they really want. Both numerical strength and strength of preferences come into play to find a natural equilibrium. But there’s no reason to believe that the overall results will be favorable in aggregate if the trades involve both representative and nonrepresentative (random) interests. There’s nothing corrupt or shameful about seeking to address serious local concerns. But local concerns are very much nonrepresentative in a national level body of several hundred allotted members.
Here’s a simple example. My sister works for a small social work firm in Missouri. I know they are wildly underfunded for the work they are expected to do. If I were a legislator I could push for a funding increase for social services in her area. This would surely benefit her. But would it be an example of corruption? I know social services are underfunded in her town. Probably most places too, but I don’t know for sure. I don’t think most people in this situation would feel they were acting in a corrupt fashion. Other towns would lose out because they didn’t have a similar insider to push for their interests. If the members act rationally, and pursue their interests, we can expect these random, nonrepresentative interests to be pursued by the individual members in much the same way fully representative interests are pursued.
LikeLike
Naomi,
> I suspect that policy trading can get us closer to an ideal set of policies than taking every issue one at a time. People will give up the things that don’t matter much to get what they really want.
I agree.
> If I were a legislator I could push for a funding increase for social services in her area.
It would be much easier for you to get your fellow legislators to increase the funding in a specific area by promoting this as a general issue – evaluating and addressing the funding needs of such activity country-wide. This is how the narrow interest become representative.
> I don’t think most people in this situation would feel they were acting in a corrupt fashion.
I agree. And this is another reason why promoting this issue on a principled basis would have much higher chance of being effective.
LikeLike
Naomi and Yoram,
I am not convinced that policy trading (vote swapping) works like an idealized market leaving everybody better off (getting what they want most). Instead we get a ratcheting up of deficit spending and relatively inefficient and unproductive pork. Of course, the evidence for this only exists in ELECTED legislatures (and the evidence is massive), so I can’t assert with much certainty that the same dynamic would apply in an allotted legislature. But the logic is the same… and I think the best chance of getting people to genuinely seek the common good is to isolate issues into stand-alone decisions where the way to satisfy a minority interest is through an appeal to fairness and empathy. I have seen that work in real life community organizations and seen horrible vote swaps for unrelated issues when I was in the legislature.
LikeLike
Terry,
I believe the reason vote swapping causes problems in conventional legislatures is due to the agency loss inherent in the electoral accountability mechanism. If the will of a representative were well and truly indistinguishable from the will of her representatives, would it not be the case that she would act to promote their preferences and interests through vote swapping? Agency loss allows elected members to pursue nonrepresentative interests (such as their own reelection or their pet issues) and I don’t think there’s any reason to believe that deals meant to further nonrepresentative interests will generally be favorable in aggregate.
LikeLike
Naomi,
It works like this… We are in a legislature (with no concern for re-election or the like) People in your town genuinely want a new road, (though the traffic level doesn’t really warrant it). People in my town likewise want a new top of the line civic center. People in another town want to allow strip-mining for more employment in town. Looking at each issue individually no rational legislative process would adopt these three items, but through the magic of vote swapping all three “local interests” will pass.
LikeLike
>”but through the magic of vote swapping all three “local interests” will pass.”
… iff the members of the body value their local projects over the additional cost of all the local projects in aggregate as paid for through additional taxes/deficit spending. Conventional pork barrel spending reflects that the first concern of an elected representative is reelection. Naturally they take advantage of the limitations of the electoral accountability mechanism to promote themselves and load their portfolios with items that sound good in campaign ads. If an allotted sample values a collection of unrelated local projects (that would never stand on their own) over the taxes needed to pay for them after a few days of well-moderated debate/discussion… so be it. In any case, the need to pass through the sample limits trading to interests large enough to be reliably represented in a sample of several hundred. So, national-level interests only. There’s room and incentive for big-picture compromise and coordination across the board. A narrow majority is not guaranteed to always get their way in every single minute detail of every single policy as would be the case in a single issue system.
LikeLike
Naomi,
You do raise an interesting issue, of how to assure that a majority that becomes aware of itself, doesn’t ride rough shod over a permanent minority. Now, in a non-electoral system without partisan identification of what majority might exist, there will be many different majorities on different issues (nearly all members would be part of SOME natural majority on some issue.) But your scheme requires all of society to pay a heavy toll in order to allow a minority to defeat the majority on some issues (through vote swapping). The quandary is what is the fairest result if a majority slightly prefers A over B, but a minority desperately and passionately prefers B over A? For deliberative democrats this is where empathetic discussion comes into play, which would be lost in Keith’s “presentation and vote” scheme (simply because the presenters would be expected to make the strongest case possible for their side). In that system, vote swapping among multiple issues may be the only way for a passionate minority to “convince” the body to let them have one. But remember those vote swaps are allowing several different minorities to cooperate to defeat the actual majority on each issue…the deal is not with the majority (who have no need to make any deals), but among minorities (who need a deal to prevail). So a multi-issue legislature can end up with the natural majority losing on every issue. Not very democratic, nor epistemically good.
LikeLike
Terry,
From a democracy standpoint, what good is majority approval of the individual details if the details in aggregate can’t pass the same test?
We can’t segregate issues into neat policy domains. In the end there are almost always tradeoffs. Why should we believe the overall results will be considered acceptable in the eyes of the average person if those tradeoffs cannot be made consciously? It’s very likely any mechanism that would allow for coordination across policy domains would also be used to facilitate nonmajoritarian compromise.
Where’s the “heavy toll”? Any advocate could attempt to undermine a deal by bringing forward a proposal to strip out components individually. A deal can only stand if the majority of the sampling consistently resist attempts break it up. If there is reason to believe a component of a package deal is epistemically poor, then the case should be made in debate before the sampling. It’s not, in principle, any different from a single issue proposal which may just as easily include foolish components. Just because a package deal addresses more than one issue does not exempt it from due scrutiny. Not every advocate will be happy with it.
LikeLike
I agree with Naomi. I would also point out that there is no way to formally define what’s a “single issue” and what’s “vote swapping on multiple issues”.
LikeLike
[…] Part 1 Part 2 […]
LikeLike
[…] 1 Part 2 Part […]
LikeLike
[…] years ago I wrote a set of posts refuting several standard arguments against sortition (1, 2, 3, […]
LikeLike