[EM] Preferential Party-List Proportional Representation (PPLPR)

[Vidar Wahlberg]

On Mon, Nov 10, 2014 at 12:24:38AM +0200, Juho Laatu wrote:
> PPLPR / PPLSPR does not fall very well in this category since it is
> not really short of being proportional, but it adds a new feature to
> proportionality, namely the tendency to favour compromise parties. It
> is quite far from the winner takes it all / big parties take it all
> approach. For this reason I'd like to categorize it in some other
> group than those "semi-proportional" methods. It could be classified
> based on its tendency to elect compromise candidates from compromise
> parties, to make the results less extreme and more moderate, to seek
> balanced centrist allocation (instead of direct proportional
> allocation). But it's your method, so you pick the best way to
> describe it. Also "semi-proportional" is ok if it's ok to you.

Personally I'm more concerned about how well the method would perform
in a real election, than the words used to describe the system.
I've called it a PR system as it, as you say, starts off as a
proportional system while adding features (namely weighting in later
preferences).

There is however one thing I'd like to point out. One concern about the
method may be that it would coalesce the result into some centrist
party. Surveys giving some figures on voters' second preference may be
used to give some clues regarding this. Unfortunately these are not so
common, but I've found some figures for the nine largest parties in
Norway.
The question asked in the survey was "if you couldn't vote for your most
preferred party, who would you vote for?", and was asked shortly after
the municipality election in 2011 (parties ordered in typical left-right
axis):
Party | 2011 | 2nd pref | Short party description
------+------+----------+-------------------------------------
R     |  1.5 |      2.7 | Radical, marxistic socialism, left
SV    |  4.1 |     12.4 | Radical, socialism, left
AP    | 31.7 |     17.6 | Conservative, socialism, left
MDG   |  0.9 |      2.0 | Radical*, green, centrist
SP    |  6.7 |      6.6 | Conservative*, agrarian, centrist
--- Left block above, right block below ---
KrF   |  5.6 |      5.4 | Conservative, Christian, centrist
V     |  6.3 |     16.0 | Radical, social liberalism, centrist
H     | 28.0 |     22.9 | Conservative, liberalism, right
FrP   | 11.4 |     12.6 | Radical, liberalism, right

(Numbers won't add up as smaller parties are excluded)

* SP and MDG are slightly difficult to categorize as radical or
  conservative. MDG claims block independence, but are largely
  considered to be a left party.

There's not enough information from the survey to predict the exact
outcome with my suggested system, but we can use the numbers for some
educated guesses:
The dominating parties AP and H would lose some influence. SV and V
would gain influence. The other parties would see minor changes.
SV and V are indeed "compromise" parties in the sense that neither
specialize in a field, and therefore are likely to be ranked 2nd by many
voters. FrP is to a certain extent this as well, but the party is
causing some controversy and is often criticised by the other parties.

There are many variables and too little data to draw any conclusions,
but this example indicates that a compromising party isn't necessarily a
centrist party.


-- 
Regards,
Vidar Wahlberg