[EM] Record activity on the EM list?

[James Gilmour]

> Juho Laatu  > Sent: Thursday, August 04, 2011 5:12 PM
> > On 4.8.2011, at 14.21, James Gilmour wrote:
> > There is only one real issue in elections: representation of the 
> > voters.
> > 
> > If in a single winner partisan election the voters vote 51% for A and 
> > 49% for B, we have a major problem in representation.
> 
> Ok, 49% of the voters without representation.

This throws the problem into its sharpest perspective.  There are related, difficult problems when there are three, four or more
candidates for the one seat.


> If one uses single-member districts to elect multiple 
> representatives, then this means also some randomness in the 
> results. This is not really a problem of single-winner 
> methods themselves but a problem in how they are used (as 
> multi-winner methods).

I agree.  It is fundamentally wrong to use any single-winner, single-member district voting system to elect the members of a
"representative assembly" (e.g. city council, state legislature).


> > But if the voters vote in the same way (51% to 49%) in a two-member 
> > election, any sensible voting system will give one seat to A and one 
> > seat to B.
> > 
> > Compared to that difference in providing "representation of the 
> > voters", all the other differences between single-winner and 
> > multi-winner elections are trivial.
> 
> From this point of view single-winner methods are more 
> "problematic" than multi-winner methods (at least when used 
> to elect multiple representatives from single-member 
> districts).

No  -  not just when (improperly) used to elect the members of a "representative assembly".  THE problem is inherent in the
single-winner election.   As you go on to say in your next comment.

> This problem of single-winner methods is quite 
> impossible to fix (most single-winner methods respect the 
> will of the majority).

The extreme problem (51% to 49%) is impossible to fix and so it is the greatest challenge in electoral science to obtain the "most
representative" outcome.  In the two-candidate election, the best we can do is to guarantee representation to the majority.

 
> The 51% vs. 49% problem is present also in accurately 
> proportional representative bodies since also those bodies 
> may make majority decisions. One way to alleviate this kind 
> of narrow majority related problems is to seek compromise 
> decisions.

I have to part company with you here.  It should NOT, in my view, be part of the function of the voting system to manipulate the
votes to obtain any outcome other than "representation of the voters".  It is not part of the function of a voting system to "seek
consensus".

If the voters want to vote for candidates who will seek consensus, that's fine  -  but that is very different for making "seek
consensus" an objective of the voting system.

The function of the voting system should simply be to return the "most representative" result in terms of representing the voters,
as expressed by the voters' responses to the candidates who have offered themselves for election.

"Seeking consensus" and "not seeking consensus" are aspects of how the elected members will behave within the elected assembly.  And
of course, the voters may rightly take such views into account in their assessments of the candidates before they cast their votes.
But that is just part of candidate appraisal.  Given a sensitive voting system, the outcome (seats won) will reflect the views of
the voters, which may include views on "seeking consensus".

James


> That is what in principle happens e.g. in 
> coalition governments. Coalition governments may represent 
> well over 50% of the voters. Let's assume that this is the 
> case. The program of the government may contain multiple 
> topics that would be 51% vs. 49% questions in the 
> representative body or among the voters, but probably all 
> coalition members will get more than they lose. Let's assume 
> that the coalition is heterogeneous so that it does not agree 
> on all the 51% vs. 49% decisions that is has to make. Maybe 
> there are two 51% vs. 49% topics that go the right way 
> against every one such topic that goes wrong. In that way we 
> don't have a narrow majority that always makes 51% decisions 
> but a supermajority that has considerably higher support behind
>   everything it does (although all parties of the coalition 
> do not like all the decisions).
> 
> In two-party systems the balance is based more on two 
> alternating policies. Often both parties have quite centrist 
> policies since both try to meet the needs of the median 
> voters. In some topics they may however have also clearly 
> opposite positions. I guess the overall policy and results of 
> two-party system governments are typically more 51% majority 
> driven than in multi-party governments. (Coalition 
> governments may however also have only narrow majority and 
> the coalitions may be quite fixed, e.g. left vs. right, and 
> as a result their decisions may follow the 51% majority style.)
> 
> My point is just that in addition to multi-winner methods and 
> proper PR one may need "the art of compromise decisions" to 
> get rid of the strongest 51% vs. 49% . This discussion went 
> already quite far from the technical properties of the 
> single-winner methods, but maybe this kind of compromise 
> making related problems can be considered to be one key 
> problem that the different methods and their use in societies 
> should try to address (if the case that one wants to replace 
> "the dictatorship of narrow majority" with "horse trading 
> deals of larger majorities").
> 
> Juho