[EM] Sociological issues of elections

[Kristofer Munsterhjelm]

On 09/11/2013 10:00 PM, Fred Gohlke wrote:
> You mentioned a town meeting arrangement, and I'd like examine that for
> background on the inclusiveness of a political system.  Although a town
> meeting structure is often thought of as the purest form of democracy,
> it has flaws.

You have a point there. Generally, different political systems have 
different dynamics: in a given system, some things are plainly 
impossible, other things degrade if you try them, and which things these 
are depends on the system in question.

If you try to force direct democracy on a very large scale, I suspect 
the information overload leads to either a breakdown of the system, into 
a coalescing into some kind of representative system, or - particularly 
if there are any influential demagogues around - the kind of mass 
dynamic that can lead to collective insanity (I've mentioned the Terror 
and different cult-of-personality dictatorships before).

So we shouldn't then be surprised that small scale direct democracy has 
its dynamics too. And I can definitely see situations where preexisting 
social oppression can spread into the social domain in a community where 
everybody knows everybody else.

As I also said in another post, we've had quite some history with that 
in the social realm in the small towns here (in Norway). It could get 
quite ugly: stand out from the crowd and you would find yourself shunned 
and nasty gossip would start to spread about you. Fortunately, this 
social conformity effect is no longer as strong, among other reasons 
since the people who live in such places can move more easily. (And as 
an aside, many do. Not just because of the oppressive small town effect, 
but because the jobs are to be found in larger cities -- though I 
imagine Norway is doing better than say, Sweden, in this respect: we 
have a deliberate element of decentralization in our policies.)

> Town meetings tend to favor assertive individuals, and the decisions
> made in the meetings tend to be made by cliques, usually under the
> guidance of the same assertive individuals.  Understanding how and why
> this happens is important if we are to construct an inclusive system.

Yes. Perhaps I was quick to say that town hall democracy was the 
standard by which representative democracy measured itself. But the 
argument that I mentioned has an intuitive appeal: "we'd all get 
together and discuss things if we could, but there are so many of us 
that we can't, so we have to elect people to do the job". And I think 
the argument would feel intuitive whether or not direct democracy was 
the ultimately best system. There's an intuition that direct democracy 
is the natural state, and that it is definitely better than 
representative democracy, in particular where the representative 
democracy does a bad job of actually representing the people.

Maybe it would be better to say that direct democracy, while not 
perfect, is a good thing to try to emulate. Then we can deal with the 
problems of direct democracy when we get there - to the extent those 
dynamics also show up in whatever we're using to emulate it.

As an aside, I once read about an economic system that proposed the 
creation of consumer and producer councils so as to balance power. Since 
the consumer councils would have known what the individual people who 
were part of it were buying, some critics brought up the potential 
pressure to conform -- they called it the "kinky underwear problem" (or 
"banned books problem"). So others are aware of this problem, though in 
differnt domains, as well.

> Jane Mansbridge examined such an arrangement in a town she gave the
> fictional name of Selby, Vermont, and described her findings in Beyond
> Adversary Democracy[1].  Her descriptions matched my personal experience
> in a village in a farming community in Western New York State in the
> 1930s and 1940s so accurately that I'll use excerpts from her book to
> describe the weaknesses in 'town meeting' democracy (showing the page
> references in brackets).
>
> Over the two hundred twenty six years of the United States' existence,
> our political system has gradually broken down.  We are now experiencing
> a political situation foretold by Jean Jacques Rousseau in a passage
> quoted by Mansfield:
>
>    "Finally, when the State, on the eve of ruin, maintains
>     only a vain, illusory, and formal existence, when in
>     every heart the social bond is broken, and the meanest
>     interest brazenly lays hold of the sacred name of 'public
>     good', the general will becomes mute, all men, guided by
>     secret motives, no more give their views as citizens
>     than if the State had never been, and iniquitous decrees
>     directed solely to private interest get passed under the
>     name of laws." [page 19]
>
> Rousseau's description accurately fits the present state of politics in
> the United States.  The need to envision an alternative to the present
> system is compelling, but, before we can do so, we must consider the
> human traits that deter us from achieving a truly democratic form of
> government, so we can devise a process that lessens their impact.

I've sometimes thought that the United States is suffering the political 
equivalent of the infrastructure problem where "older" countries have to 
maintain their comparatively inefficient infrastructure whereas "newer" 
ones can go right to what is known to be very efficient. That could 
explain quite a bit of the US becoming corrupt. It was one of the first 
lasting democracies, so it couldn't know what kind of attacks that would 
be attempted against that democracy, and so was much less prepared for 
them. If the founders had magical hindsight from the future, perhaps 
they would have used Approval, which (even though it definitely has its 
flaws) is pretty hard to beat in the "bang for the buck over Plurality" 
category.

> Here, we see that although the process is nominally open to everyone in
> the community, and thus is deemed democratic, the reality is that only
> about one quarter of the people actually participate in any part of the
> process.  To give us a little insight into why this is so, I will cite
> the comments of two citizens:
>
>     [Page 61] Elizabeth Hurley, who "has not attended a
>     town meeting in the last ten years", said:
>
>     "I don't care to -- well, to tell my part, you know,
>      right agin a whole mess of people ... I don't know,
>      I don't like to get up in town meeting and say, well,
>      this and that ... well, everybody's looking, or doing
>      something, and they'll say [whisper], "She's a fool!"
>      There's one man in particular [Bedell], that's up on
>      this road here, boy oh boy, he's into hot water all
>      the time.  [JM: He talks up in town meetings?]  Oh!
>      Gracious to Betsy, I guess he did.  [JM: Do people
>      pay attention to him?]  Hah, hah, no they don't, boy,
>      we just, ah ..."

Anonymity should be integrated into the system, but finding the exact 
point where the identity is unveiled would be difficult. On the one 
hand, individuals should be given the capacity to reveal and bring 
issues to attention that the community might not like to talk about; on 
the other, when politicians gain power, their apparatus, which makes 
decisions on the behalf of the people, should be as transparent as 
possible. Somewhere in between, the need for protection of unpleasant 
expression turns into the need for transparency - but there's no 
algorithm that can tell us where that crossing point should be.

Though perhaps one could say that advisory systems could be secret while 
the decision systems are public. But that doesn't really fit with the 
secret ballot: it is a "decision system" - where the people decide who 
gets into power - yet it is anonymous and secret.

> Another flaw in the town meeting concept is the tendency of the process
> to resolve public issues in private, outside the process.

Perhaps that could be seen as an ad-hoc manner of handling the anonymity 
problem. Ask people that you know (and trust) first, so that if they 
agree, you can all speak as a group.

The point about anxiety is useful to note as well: it may not be that 
the community is all that hostile, but some feel as if it would be. Then 
having some "protection" works to alleviate that anxiety, whether or not 
what one fears is real.

> This material tells us that, if we are to increase the inclusiveness of
> the political process, if we are to empower every member of the
> electorate by letting them participate in the practice of politics to
> the full extent of each individual's desire and ability, we must devise
> a process that does not expose the people to ridicule for expressing
> their views.
>
> One may argue that voting does not expose anyone to ridicule, but the
> efficacy of the vote is open to question.  The right to vote would be an
> essential element of democracy if the people were able to select the
> people and the issues for which they vote.  However, in the United
> States, the people's right to select the candidates and decide the
> issues on which the people vote has been usurped by the political
> parties.  They have arrogated to themselves the right to name the
> candidates for public office.

 From a control perspective, voting happens too infrequently. It would 
be like trying to keep a temperature by adjusting the power to the 
heater once every four (or two) years.

But the control perspective isn't complete. At least in multiparty 
systems that I am familiar with, there is a political debate between the 
actual elections. The parties try to position themselves - adjust what 
they're "selling" - even though the actual "purchase" only happens every 
n years.

Your point about the people being able to select the people shows this. 
If they can select the people, then there is an adjustment between the 
elections, where the people decide which candidates are to run - or 
which candidates have a chance of winning when they do run. A better 
election method makes this work better because it decreases the chances 
that artifacts in the system itself obscures the selection process. For 
instance, if everybody spent lots of time deciding which of many 
candidates would be good ones, and then the system just picks the 
candidate a dictator approves of, then the selection process is 
pointless. The more accurate the election method is, the better it would 
be at letting the accumulated opinions of the people decide which 
candidate is (or candidates are) to rule.

> You made the critical point that, with an electorate in the
> multi-millions of people,
>
>    "... we need somebody to act for us, but few enough that
>     that they can still govern by deliberation.  Hence,
>     representative, elected democracy."
>
> In the United States, the number of representatives in the two Houses of
> Congress seem (to me) reasonable.  That leaves the question of how the
> representatives are to be selected.  You mentioned that delegable proxy
> lets the deliberation happen "on the outside", presumably before the
> representative is chosen.  However, I'm not sure if that approach
> requires or allows political campaigning, which is a corrosive activity
> that demands huge expenditures to manipulate the voting public.

It might support campaigning in general, whether or not it's produced by 
a political party. But there is a balance here. Say famous person X says 
"delegate your vote to me". If you think X is particularly good at 
politics, you might do so. For that matter, plenty of people would give 
their vote to X just because X is famous - actors have become 
politicians after all.

But with delegable proxy in particular, the balance is that the relation 
is supposed to be two-way. If you delegate your vote to a celebrity, and 
there are thousands that also do so, and you say "hey, I'd rather prefer 
you to support an extension/limitation of copyright", then he doesn't 
have to listen to you. He has thousands of others. But if you give your 
vote to a proxy along that has a hundred others, or fewer, then he's 
going to take your suggestions much more seriously. In turn, if you're a 
proxy and you have a hundred voters subscribed to you, then you can 
influence greater proxies more than if you're just a single voter.

Thus there's a point where campaigning fails to do much good. The voters 
will say, emotionally, that they feel more familiar with smaller 
proxies, and instrumentally that they have more of a say when subscribed 
to a smaller proxy than to a large one. You could say that proxy-proxy 
interactions happen near their own scale: large proxies talk to large 
proxies, small proxies to small. There would be less of the kind of mass 
"one to many" manipulation that parties, particularly in two-party 
systems, do.

> re: "But on the other extreme, consider you have an emperor (or
>       an elected king).  If you suspect corruption from power is
>       a real problem, then you have to set up some form of
>       oversight.  The oversight doesn't automatically arise from
>       the system itself, but rather has to be separately
>       implemented."
>
> Personally, I don't consider either a viable option, but you make the
> excellent point that neither automatically accommodates oversight.  In a
> democracy, oversight should be exercised 'before' a candidate is elected
> to public office.  The electors need a way to assure themselves that the
> candidate's internal gyroscope is aligned with their objective(s)
> 'before' they make their choice.  Therefore, the electoral method must
> ensure that candidates are carefully examined by people with a direct
> interest in uncovering any aspect of a candidate's intentions that is at
> odds with the public interest.

What I meant was that systems that derive their authority from the 
people combine the governing and oversight aspects. The people 
themselves provide the oversight. If the system is inaccurate (has a bad 
election method, say, or gives too much power to a single individual for 
too long a time), then that will effect both the governing and the 
oversight aspects. In contrast, aristocratic systems need explicit 
systems to provide oversight because the two tasks aren't implicitly 
connected.

Still, you're right. You'd need explicit oversight in two respects even 
in a system based on authority from the people. First, between elections 
- hence the checks and balances of the US system. Second, when it's 
possible for the apparent "popular will" to be manipulated or to diverge 
from what the people really wanted (had they had enough time and clarity 
of mind), then the implicit feedback can be weakened.

> re: "(In a multiparty democracy, the different parties are
>       supposed to check one another, and in a coalition system,
>       the opposition is supposed to check the position, where
>       the respective coalitions may change from time to time,
>       again lessening the chance that corruption can set in.)
>
> Does it actually work this way in practice?  I have no experience with
> multiparty democracy, and cannot make an informed comment on the
> practice.  However, systems based on the organized pursuit of power seem
> (to me) susceptible to corruption.

Consider a comparison to the market. If you have only one business, 
that's a monopoly (one-party state). Monopolies regularly charge too 
much for low quality goods. If you have two, then you have a duopoly, 
which is better but still not ideal, and there's the definite risk of 
collusion. As you increase the number of businesses providing the same 
goods, competition pushes their prices down.

A multiparty system thus solves more of the corruption problems by 
providing more alternatives. Like a market with many companies, it's not 
perfect. You still need regulation (in the political case, good systems 
and rules); but in my opinion, multiparty democracy seems to be 
considerably better at avoiding corruption than two-party ones.

The culture of the society in question of course also matters. Some 
countries have multiparty systems but still have only two parties of 
importance (e.g. Mali), or even one (e.g. South Africa). And yet others 
are unfamiliar with the democratic process and have so many fragmented 
parties that the system doesn't really work.

In any case, the point of the market comparison is that every 
participant in the market (political system) wants to maximize profit 
(power). Yet by having enough of them, the idea goes, they will keep 
each other in check. It seems to work at least enough that I would say 
multiparty systems are better than two-party ones.

> re: "I'm imagining the election method for the hybrid to be
>       proportional, also, so that if 10 of the 500 think that
>       advocates for position X should be on the legislature,
>       then 3 (same proportion) will, assuming they vote according
>       to that opinion."
>
> I'm not clear on this point.  By proportional, do you mean the number of
> random choices will be proportional to party size?  That is, if party
> 'A' is 23% of the electorate, 23% of the total candidates selected will
> be chosen, at random, from party 'A'?  If so, may I suggest that
> non-partisans be treated in the same way?  If 57% of the electorate is
> non-partisan, 57% of the candidates are chosen, at random, from the
> non-partisans?  (I will avoid consideration of the treatment of
> advocates of a particular position until I have a clearer understanding
> of your idea.)

Not per party, but per opinion. Say that in the initial stage, 23% will 
rank people who express a liberal foreign political stance ahead of 
everybody else. Then 23% of the second council will consist of people 
who expressed a liberal foreign political stance.

(And such preferences might be nested. If 50% prefer X-ists, and half of 
these 50% prefer [X,Y]-ists, the other half [X,Z]-ists, in the sense 
that they'll rank these groups ahead of everybody else, then you would get

25% [X,Y]-ists
25% [X,Z]-ists
50% everybody else

in the legislature, subject to that it's actually possible to do so. For 
instance, the above would be impossible if there were only three members 
in the second stage.

> re: "A system can be pushed more towards "not alienating those
>       further away" by increasing the threshold for action (e.g.
>       supermajority rule), and that's what I noticed."
>
> At the risk of diverting attention from the critical issues we are
> discussing, I'd like to suggest that, instead of using a supermajority
> rule, we consider the sunsetting of legislation, that is, varying a
> law's life depending on the percentage of legislators supporting the
> law, with all laws subject to repeal at any time by a simple majority.
> Perhaps, something like:
>
>      Approval Rate      Term of Legislation
>      -------------      -------------------
>      Less than 52%      law expires in one year
>      52% to 60%         law expires in two years
>      60% to 75%         law expires in five years
>      75% to 90%         law expires in ten years
>      over 90%           no automatic expiration
>
> These terms are, of course, only for illustration.  The actual terms
> should be determined by study.  Given the harm done by bad legislation,
> this might be a topic worthy of thought and discussion.

That may also work. If you want less fine-tuning, you could use a 
variant of Simmon's consensus algorithm, something to the effect of: 
"everybody gives both a compromises vote and a favorite; if the 
compromise has sufficient support, it wins with a long term sunset, 
otherwise the favorite wins with a short term". But I expect that in a 
small council like a legislature, there would be negotiations ahead of 
time anyway, so the compromise would win an ordinary vote as well. 
Perhaps it would be useful for time-limited matters where the body can't 
afford to deliberate for months on end?

(Note that when I say a "compromise vote", the vote itself may suggest 
multiple candidates.)