[EM] What's wrong with the party list system?

[James Gilmour]

Kathy Dopp  > Sent: Tuesday, July 05, 2011 2:30 AM
> 
> > On Mon, Jul 4, 2011 at 7:19 PM, James Gilmour 
> >
> > Kathy, your comments illustrate the fundamental problems with all 
> > party list voting systems: 1. you must have registered political 
> > parties;
> 
> As someone else noted in this thread already, registered 
> political parties are unnecessary to use the party list 
> system.  Candidates can simply put together their own lists.

But someone has to control, or take responsibility for, each list, even if it is only to submit it to the Returning Officer on
nomination day so that an agreed list will be printed on the ballot paper.  I am aware that in some jurisdictions otherwise
"independent" candidates can form "groups" for this purpose, but these groups are registered for the purpose of the election.


> > 2. each party must produce a list of candidates ordered in some way;
> 
> Each *list* is normally ordered - Yes.  But the list method 
> does not have to be done that way if it is an open list 
> system where voters can vote for candidates, and thus voters 
> determine the list order.  

In open-list systems the names of the candidates are printed on the ballot paper, under the respective party headings.  So the names
must be ordered in some way, even if it is alphabetical or random.  So the voters do NOT determine the order of the candidates in
each party's list.  The voters may vote in ways that determine which of the listed candidates is elected to one of that party's
seats, and determine the order in which the candidates in a list are elected, but that is all post-election.  The voters in the
public election do not in any way determine the order of the candidates in the parties' lists as those names appear on the printed
ballot papers.


> Most voters would disagree with 
> you and think it is a benefit to have the political party or 
> leading candidate put together the list order for them so as 
> to save the voters the time and effort it would take to 
> research all the candidates.

It is clearly "most voters" in some countries (because those voters appear happy with their present closed-lists), but others would
disagree and prefer to have some or a lot of choice in determining which of their favoured party's candidates should actually fill
the seats allocated to that party.  The voters could have a great deal of effect without having to research every candidate or
indicate a ranking for every candidate on their favoured party's list.


>  However, a less popular system, 
> would simply require voters to pick a candidate from the 
> list. 

I don't know what you mean by "less popular", but this (pick one candidate) is in fact a common version of open-list party-list.


> I suppose it's possible, as some have also commented 
> here, to allow voters to rank order a list, but that would be 
> administratively burdensome and probably not practical for 
> large national elections, as has been mentioned.

The practicality depends on the size of the electoral districts, and on how the candidates are presented.  In some countries there
is one national list for each party; in others, the votes are totally nationally, but the parties' list are presented on a regional
basis.


> > 3. voters are restricted (to a greater or lesser
> > degree) in how they can respond to the choices of representative 
> > offered to them.
> 
> Relative to some electoral methods that are less desirable in 
> other ways, perhaps.  However, the list system has many 
> benefits those other systems don't have, which is why it is 
> so popular in many countries - for nationwide legislative 
> bodies where other systems may not be practical or desirable.

Yes, even closed-list party-list delivers party PR in a way that some other systems do not, notably plurality in single-member
districts (UK, USA and Canada).  If there were no alternative, that would be an advance.  But we already know how to do better than
that.


> > All of these impose unnecessary limitations on
> > the PR of the voters that could be obtained by a less constrained 
> > voting system.
> 
> You might want to read up on the many studies of voting 
> behaviour - say American Voter Revisited or Controversies in 
> Voting Behavior. Most voters do not want to have to 
> investigate and individually rank hundreds of candidates, so 
> an open party list system where they are familiar with the 
> top ranked candidates on each list and have the chance to 
> vote for someone they prefer most to move them up the list.

I am certainly not recommending any voting system that would require voters to investigate and individually rank hundreds of
candidates.  That is both undesirable and unnecessary.  The number of candidates presented to voters in any one electoral district
is a function of electoral district size.

If the voter chooses a list headed by a familiar candidate and then has "the chance to vote for someone they prefer most", it MAY
move that candidate up the list, but very large numbers of such votes have no effect at all.  And such a system does not deliver
proportionality of representation WITHIN the party.


> When the number of candidates gets too big, the systems you 
> are proposing, are untenable.  However, in smaller local 
> election contests, voters may know more about more of the 
> candidates. However, that is also contrary to research that 
> shows that most voters pay more attention to national, then 
> downstream contests.

As stated above, I am not suggesting the use of any voting system that results in such large numbers of candidates that the systems
would be untenable.


> Political parties can be a good thing because people normally 
> take shortcuts when deciding whom to vote for - by selecting 
> the political party that agrees with their own ideology. 
> Especially under the party list system, parties can be 
> beneficial because smaller political parties that challenge 
> the major parties have a far greater chance of winning seats 
> in a large legislature, so that voters can vote for minor 
> parties without wasting their votes or causing a spoiler 
> effect, unlike with IRV/STV methods.

I am not at all opposed to political parties  -  if they did not already exist, we should have to create them.  They are an
essential part of the political system.  However, I am opposed to the parties having too much power in the system in comparison to
the voters.

You should not assume that votes for minor parties will not be wasted.  In large legislatures where the votes are totally
nationally, there is usually an artificially set threshold designed specifically to keep the smaller parties from winning any seats.
This has been done because of the very adverse political effects of having large numbers of small and very small parties represented
in the assembly.

As to spoiler effects, as has already been pointed out by others, the effect of STV-PR (multi-winner elections) is quite different
from IRV (single-winner elections).


> > I would also say that these restrictions are
> > undesirable, but that view reflects my political culture.  I do, 
> > however, recognise that these restrictions are accepted by many in 
> > continental Europe who happily use party-list PR voting systems 
> > without any clamour for change.
> 
> So you think the open list method is worse than the 
> nonproportional US system then of two-major political parties 
> and single member districts or entire states?

Straw man!!   There is absolute nothing in what I have written that could possibly lead to that conclusion.  Of course, in terms of
party proportionality, an open-list party-list system would be better than the present plurality in SMDs used in the USA and Canada
and to elect the UK House of Commons.  But it is not simply a choice between these two systems  -  there are alternatives that
deliver more than either.


> I definitely believe strongly that the IRV/STV methods are 
> far far worse than the current US plurality system.

IRV is not relevant in any discussion about multi-winner elections for which a party-list voting system might be used.

If you think STV-PR is worse than the current US plurality system, I can only say that you have not looked properly at STV-PR.


> The party list system does not have any downsides like 
> nonmonotonicity, and fundamental unfairness of the counting 
> method (as well as not solving any of the problems of 
> plurality) like STV and IRV have.

Monotonicity is desirable but is irrelevant in public elections.  We cannot have monotonicity without sacrificing compliance with
some other criterion (which many of us consider to be more important).

It is irrelevant to mention IRV in any discussion about party-list voting systems.  There is no "fundamental unfairness" in the
counting method used in STV-PR.  The system delivers proportional representation of the voters.  If the voters vote strictly on
party lines, then STV-PR will deliver party PR that matches any party-list system (given the same district magnitude - the key
determinant of proportionality).


> > Your comments also confuse what are essentially private 
> matters with 
> > public matters.  The candidates who can stand in the name of a 
> > registered political party must be decided by that party.  Some 
> > parties may decide that by centralised control; other may do it by 
> > very democratic (PR) elections ("primaries") of all party members.  
> > All parties are coalitions, some broad, some narrow.  It is in a 
> > party's interest to ensure that its list of candidates will appeal to 
> > the widest range of its potential supporters among the electorate.  
> > Thus all significant factions within a party are likely to be 
> > represented on its list.  If some faction within a party finds it 
> > candidates consistently excluded, that faction will almost certainly 
> > go off and form a new party.
> 
> Yes. And that new minor party has a good chance of obtaining 
> seats once its strength reaches the Droop quota - unlike with 
> IRV/STV methods where people quickly learn that the spoiler 
> effect means they still need to rank one of the two major 
> parties 1st and their 1st place candidate hurts the chances 
> of their 2nd place candidate to win because not all voters 
> 2nd and later choices are treated equally and a voters' 2nd 
> choice is eliminated often prior to the 2nd choice, ...

You really must stop trying to muddy the waters by repeatedly referring to IRV in a discussion devoted exclusively to multi-winner
elections  -  which is where party-list voting systems are used.

The comments about "STV" do not reflect how STV-PR really works in practice.


> > If some faction within a
> > party finds its candidates on the list, but always at the bottom (and 
> > so with little chance of election), that faction may well split off 
> > and form a separate party, when its candidates will automatically be 
> > at the top of its list.  That does happen, especially with closed-list 
> > party-list systems.  It is open for any group that can meet the 
> > requirements to be a registered political party to present a list.  In 
> > some jurisdictions, that can include individuals standing as 
> > "independent candidate".  But these are all "private" matters 
> > (within-party), determined by the respective parties before  the public 
> > election.
> 
> I agree that open list systems are better than closed, of 
> course since voters can change the list order.
> 
> >
> > At the public election a voter can choose one party from among the 
> > various parties, and in open-list systems make one choice (or a 
> > restricted choice) from among the candidates of that one party. The 
> > counting rules provide good proportionality among the parties (subject 
> > to various arbitrary thresholds).  But with the commonly used 
> > open-list systems, the counting rules do not provide PR within the 
> > parties.  Significant groups of voters who support a particular party 
> > can be seriously under-represented in terms of the within-party 
> > balance, either through piling up massive votes for some particularly 
> > popular candidates or through spreading their votes across too many 
> > candidates.  To overcome this defect, the votes must be transferable 
> > in some way.  And to ensure PR of the voters, those transfers must be 
> > determined by the voters, not by some party-list rule in the 
> > legislation.
> 
> The excess votes are transferrable in many countries in that 
> any unfilled seats are often filled by coalitions of minor 
> parties or minor and major parties who get together and 
> decide on the last remaining winners - again giving smaller 
> parties more power than under the current prevalent US system.

Maybe.  But I was referring to the lack transferability of the votes WITHIN one party.  That is the cause of the lack of
proportionality WITHIN each party.


> The party list system would be great for filling one of two 
> legislative bodies in each state. One body could be filled by 
> local single member districts and the other body via an open 
> party list system - making representation at the US state 
> level proportional and giving immediate greater 
> representation to the many minor political party interests.

If you have ANY legislative body that is supposed to be representative of those it is elected to serve, that body should be elected
by a voting system that will deliver a properly representative result.  No system based only on single-member districts can do that.


> > What you then end up with is a series of STV-PR elections  within each 
> > party list (or with something comparable for those who don't like 
> > STV).  The most complex open-list party-list systems go some way 
> > towards this.  But I have to say again, if you are going to  go to all 
> > that bother, why not  go the whole way and fully open up the voters' 
> > choice by removing all the restrictions of 'voting for a party' and of 
> > 'voting within one party list'?
> 
> Huh!?  That would be terrible if it were true IMO - 
> eviscerating election transparency, verifiability, and fairness.

Without some mechanism to transfer votes WITHIN each party's list you cannot obtain proportionality WITHIN each party.  However you
do that, it will be complex compared to marking only one candidate in the open list and so NOT obtaining that desired within-party
proportionality.  All I was saying was, that if you are prepared to entertain that degree of complexity to solve the problem
separately within each party, it would seem a lot simpler just to open the whole system up, allow the voters greater freedom of
choice, and apply the proportionality calculations across all parties and all candidates.

James Gilmour