[EM] Sainte-Lague vs d'Hondt for party list PR

[Juho Laatu]

On 12.7.2012, at 7.58, Michael Ossipoff wrote:

> On Tue, Jul 10, 2012 at 3:14 AM, Juho Laatu <juho4880 at yahoo.co.uk> wrote:
> On 2.7.2012, at 13.58, Raph Frank wrote:
> 
> > For example, 26 parties at 1.5% and one party at 61% for a 49 seat parliament would split the seats, 20 for the large party and 29 for split between the micro parties.  The micro parties get 59% of the seats for 39% of the vote.
> 
> I only now checked the numbers in this Webster / Sainte-Laguë example. It seems that even with first divisor 1 the largest party would get 23 seats. The example is still valid, just the numbers seem to be inaccurate.
> 
> Divisor methods are based on a fixed seat allocation order (based on the number of votes of each party) that does not depend on the number of seats. This means that if there are N small parties of equal size and one large party, there will be some fluctuation in the results when the number of seats grows. There is some number of seats (S1) that will all go to the large party, and next N seats will all go to the small parties (assuming no other ties than those between the equal size small parties). Size S1 thus favours the large party, and size S1+N favours the small parties. This is what I called fluctuation above.
> 
> In the given example S1 = 20 (all seats to the large party) and S1+N = 20+26 = 46 (one seat to every small party, still 20 to the large one). In the S1 seats case the large party gets 100% of the seats with 61% of the votes
>  
>  
> [endquote]
>  
> Then the small parties surely wouldn't do a splitting strategy such as was described.

Yes. Small parties should strategize only if their size is large enough to get two seats by splitting and only one without a split. In most cases that is too risky and not recommended. (And SL can also defend agaist such strategies by using a modified first divisor.)

Btw, this strategy is much less risky when we talk about splitting districts, since the population of districts is very stable when compared to possible risky changes in party support.

>  
> But what if it isn't the result of splitting strategy? What if those are the natural spontaneous parties and their vote totals? Does the result look wrong? It isn't wrong. SL, by doing what it does, is minimizing the deviation of each party's s/q from the ideal equal s/q value.
>  
> It might _look_ wrong at first glance, because it "violates quote", but it isn't wrong,in terms of fair s/q.

Yes, some people might look at the s/q values, but I'd expect someone to notice also the unfair use of the quotas / votes.

>  
>  
> . In the S1+N seats case the large party gets 43.48% of the seats with 61% of the votes. Or in other words, all 20 seats with only 12.2 quotas (7.8 extra seats), or only 20 seats with 28.06 quotas (8.06 seats too little).
>  
> [endquote]
>  
> That's ok if the parties are genuine, natural and not the result of splitting strategy. As I described above. 

I'm afraid some people might get upset if they think they were entitled to 8 seats more but will get none, and will lose a very clear (28 seats vs. 18 seats) majority. There are thus many approaches to measuring the fairness of the results, and the quota based approach may be a very natural one to check first.

Juho


> 
> In real life party sizes usually vary more, and as a result the fluctuation is not as radical as in this kind of extreme examples. Modified first divisors can be used to eliminate strategic splitiing of parties.
>  
> [endquote]
>  
> Yes. (I call them "denominators), to distinguish them from "divisors" as the word is used in divisor-method terminology).
>  
>  
> Use of the second divisor for strategic purposes is more difficult than using the first divisor, so there may be no need to modify the second divisor
>  
>  
> Yes, and splitting involving the 2nd denominator could only multiply s/q by 4/3, instead of doubling it.
>  
> Mike Ossipoff
>  
>  
> ,
> 
> 
> 
> 
> ----
> Election-Methods mailing list - see http://electorama.com/em for list info
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120712/4563ffc9/attachment-0004.htm>