[EM] New election system in Hungary

Jameson Quinn jameson.quinn at gmail.com
Wed Jul 5 10:46:36 PDT 2017

2017-07-05 2:47 GMT-04:00 Magosányi Árpád <m4gw4s at gmail.com>:

> 2017-07-04 19:58 GMT+02:00 Jameson Quinn <jameson.quinn at gmail.com>:
>> Before you jump straight into designing mechanisms, it's important to be
>> clear about what you're looking for: the values you want the method to
>> fulfill. On that matter, you've said the method should be:
>>    1. not too shockingly new
>>    2. "proportional representation...
>>    3. ...and no entry threshold"
>>    4. The winning strategy for candidates is collaboration
>>    5. The winning strategy for voters is honest voting
>>    6. In the long run there is no two-party system
>> In the end, you're definitely going to have to compromise to at least
>> some degree on points 1, 4, and 5.
> #1 is not a requirement in my work. There are two proposals: one is a
> basis for discussion for parties, which should be conservative, the second
> is an ideal system as the proposal of the movement, which can bring in
> anything new, but if possible should build on the previous one. I am
> talking about the later now.
>> I'd also like to know more about #3 and #6. I understand that recently
>> Hungary's effective number of parties has been just under 2, and that from
>> that perspective increasing the number sounds like a great idea. But in my
>> opinion, the ideal ENP is between 3 and 4. That gives enough room for new
>> parties to grow and for once-major ones to die out, but still gives
>> incentives for building coalition-based parties able to articulate the
>> interests of more than one group of society.
> #3 is about the need to give opportunity for new parties.

There are two different questions here, which I think you're collapsing
into one, because traditional methods usually give the same answer to both:

   - Does a small (new) party have an opportunity to win seats?
   - Can a small (new) party get its fair share of votes without hurting
   its interests? In other words, is voting for a small offshoot party, rather
   than the closest larger party, strategically neutral, without risk of
   vote-splitting/spoiled elections?

In my opinion, it's important that the answer to the second question(s) be
"no", but that does not extend to the first question.

How does this work in GOLD voting? Say that the major parties are Orange
and Purple, and that Blue is a smaller party similar to Purple. As long as
Blue candidates ensure that Purple comes before Orange in their predeclared
delegation order, voters can safely give delegated votes to Blue
candidates. If Blue support is concentrated on a few candidates, then those
candidates can win fair and square. But if Blue support is diffuse, then it
could happen that all the Blue candidates get eliminated up-front because
none are in the top two locally. Still, those Blue votes were not wasted;
they not only help Purple win more seats, but they even help ensure that
it's the bluest of those Purple candidates who get those seats.

In the long run, these dynamics would allow the Blue party to grow
organically, and even to eventually supplant the Purple party if they
convince the majority of Purple's voters that they are a better choice.

Contrast that with, say, MMP with a 5% threshold. Under that method, if
Blue is below the threshold, all the Blue votes are simply wasted. Purple
politicians will tell Blue voters not to throw away their votes on Blue,
and the organic growth of the Blue party will be stifled. Orange could even
disingenuously fund various different versions of Blue parties — Navy,
Azure, Cobalt, etc. — in order to deliberately hamper the Purple party.

> #6 is based on our experience with Duverger's law: our political system
> quickly became a two-party system, and even that collapsed due to the
> underlying positive feedback loop. We now have a monoparty authoritarian
> regime, shockingly similar to communist dictatorships, but the style of
> political communication is even more ill. That communication style fucks up
> our everyday life.

Yes, I definitely agree with you here: the only thing that's worse than a
2-party monopoly is the 1-party state that it can collapse into, and even
the most cursory understanding of Hungarian politics shows that you guys
are suffering from those problems about as badly as anyone. But the
question is: do you want to go as far as possible from a 1-party situation,
to an Israel-style situation where no one party has even 25%? Or would you
rather settle on a medium path, with 2 or 3 parties in each size category
"large" (25-40%), "small"(10-24%), and "tiny"(1-9%)? In my opinion, running
from one extreme to the other is not a good way to solve the problem.

> This is why we put emphasis on the game theory part: we need collaborative
> behaviour from politicians, honesty from voters, and a rich political
> palette.
> As the stakes are high, we want these properties to be mathematically
> proven. One way of proving it (and a hint in the search) is to trace it
> back to the method, which have these properties proven: Condorcet.

I definitely understand wanting ironclad proofs. And there is some degree
to which that's attainable. But you'll never get all the way. For instance,
Gibbard-Satterthwaite shows that no deterministic, non-dictatorial method
avoids strategic ("dishonest") incentives for voters entirely.

Generally, then, instead of showing that a given method "promotes honesty",
you try to show that it doesn't promote a particular, problematic, version
of dishonesty. The discussion above about GOLD, unlike MMP, not having a
certain problem with wasted votes and spoiled elections is one example of
that kind of argument. This can be based on proof, on rigorous statistical
analysis and monte carlo simulations, or just on a thorough strategic
analysis. I would even say that of those three, proofs tend to be the least
convincing for me, because in order for a proof to work you almost always
need to restrict the problem and/or make assumptions in ways that are

In particular: Condorcet methods have NOT been shown to universally promote
honesty from voters or collaboration from candidates. In terms of scenarios
like chicken dilemma, whether or not a method is Condorcet compliant
doesn't even tell you that much one way or another about its strategic
properties, and there are certainly Condorcet methods that are not as good
at promoting honesty as something like 3-2-1 (empirically) or SODA
(provably). But all of that is single-winner, and of limited applicability
to the PR cases you're interested in anyway.

Unfortunately Condorcet is primarily a single seat method, and does not
> have a party list version. Though it does have a version for committees,
> which can help in the proof.
>> I'd urge you to take a look at GOLD voting
>> <http://wiki.electorama.com/wiki/Geographic_Open_List/Delegated_(GOLD)_voting>
>> (further discussion here
>> <http://wiki.electorama.com/wiki/Geographic_Open_List/Delegated_(GOLD)_voting>
>> and here
>> <https://medium.com/@jameson.quinn/worthwhile-nah-this-canadian-voting-initiative-is-golden-d93717a88221>
>> ).
> I don't think this is an ideal system for your use case; it was designed
>> as a proposal for replacing FPTP, primarily in places like Canada, US, and
>> UK. But some of the mechanisms it uses are actually quite powerful building
>> blocks for a PR method. In particular, it:
>>    - Uses an STV-like mechanism as an underlying process.
>>       - This is a familiar, well-understood mechanism, yet it's easy to
>>       add flexible vote transfer mechanics on top.
>>    - Offers delegation to a candidate's pre-declared list
>>       - this gives a great combination of simplicity and voting power.
>>    - Makes delegation optional; voters can use an open-list-like voting
>>    style if they prefer
>>       - This prevents political insiders from getting unearned
>>       horse-trading power to effectively set the party list order through
>>       back-room deals.
>>    - Eliminates candidates with inadequate direct support before they
>>    can receive transfers
>>       - This helps prevent tiny splinter parties from getting more than
>>       one seat, without wasting votes (or incentivizing favorite betrayal) for
>>       those inclined to support such splinter parties.
>>    - Simplifies ballots by explicitly listing only local candidates,
>>    leaving non-local candidates as a write-in option
>> By using these mechanisms, GOLD voting does quite well on points 2, 4, 5,
>> and 6 above; reasonably on point 1; and is arguably OK even on point 3.
> Any proof or at least reasoning for #4,#5 and #6 ?

#4: promotes party collaboration

This is largely because of the pre-elimination phase, which as discussed
above, disfavors smaller parties without a geographic nucleus, except in
cases of a high-profile candidate who can pull in many votes from outside
of their district. Since these parties will still get votes, their
predeclared preference order will have real clout. That means that
politicians from larger parties will have a strong incentive to listen to
the interests of smaller parties, especially if they are relatively
compatible ideologically to begin with.

#5: promotes honest voting

There are actually two ways in which votes can fall short of honest
expressiveness: strategy and laziness. GOLD's ballot simplicity helps
enormously with the laziness problem; by making it relatively easy to cast
an expressive ballot, it makes it more common.

As for strategy... well, like any voting method, GOLD is subject to some
strategy. In particular, as with many PR methods, free-riding can be an
issue. But free riding is an essentially self-limiting problem; it works
exactly insofar as you are in the minority who's doing it, but as soon as
everyone is trying to do it, it doesn't work at all. Furthermore, GOLD's
ballot format makes free riding difficult to pull off even in the best of
circumstances. So I don't see that as a major concern. And really, I think
that as minor as it is, other strategic issues are even smaller. If I were
voting in a GOLD election, I would vote honestly, even in the kind of
manichean good-versus-pure-evil situation that would make me inclined to be
strategic under many other voting methods (including Condorcet).

Again, I encourage you to think beyond just criteria and proofs. I've done
that stuff — my unpublished proof that SODA is monotonic is the longest
voting criterion passage proof that I, as a serious student of the
literature, am aware of — but I think that when it comes to proportional
methods, pragmatics of the ballot format and comprehensibility of the
process can be more important than passing abstruse criteria.

>> I think that by recombining mechanisms like these, you'll be able to
>> build something that will have better appeal for the average voter. You
>> don't want the instructions on the ballot to end up more complicated than
>> the rules for Settlers of Catan (with the Cities and Knights expansion).
>> GOLD's instructions ("Choose one candidate or write one in, then if you
>> wish you may choose one of the two transfer methods") are about the limit
>> of complication you should be going for.
>> All in all, I'd be happy to hear more about your group and your plans for
>> making a difference in Hungary. I'm certainly rooting for you.
> The agenda is something like this:
> - parties give their inputs until aug 10.
> - agreement on a system until sep 20
> - the new system is legislated until oct 23
> If the ruling party does not meet with the last milestone (and we now they
> won't), we will force the change through nonviolent civil disobedience.
> There are a lot of activities around learning and teaching these methods,
> and building movements for this purpose. These are very interesting times
> in Hungary. And we need all help.
> I'm definitely rooting for you!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20170705/c5182bff/attachment-0001.htm>

More information about the Election-Methods mailing list