[EM] Largest-Remainder

[Juho Laatu]

On 8.6.2012, at 20.44, Michael Ossipoff wrote:

> As I was saying, I can understand the justification for using d'Hondt to
> allocate seats to parties. And I know why you don't use d'Hondt to allocate
> seats to districts: Proportionality is much more important when you're
> allocating seats to districts.

In Finland people don't claim that D'Hondt would be fair. All agree that it favours large parties (slightly in the allocation of the last seats). D'Hondt was included also in the recent reform proposal. The argument was that Finland might use D'Hondt again simply because there is a long historical tradition of unsing it. Many people know the algorithm, have used it themselves in various elections, and therefore are quite happy to keep it.

The bias of D'Hondt is not very big, so traditions might be a stong enough reson to continue with it. Actually the Finnish districting system causes larger bias in favour of large parties. And the reform proposal included also a threshold for getting the first seat (I opposed this since it makes the idea of favouring large parties an explicit target, not just an accidental result) that also is much more radical in favouring large parties than D'Hondt.

One "ideal" explanation behind D'Hondt is that it is fair in the sense that if parties could divide their votes evenly between the optimum number of their caniddates, then electing those candidates that have largest support would give the same result as D'Hondt. D'Hondt is thus one exact approach to measuring the strength of the parties. Maybe not the best one, but in some special elections this approach might be the ideal solution that we want.

> We used Largest Remainder too, and it was abandoned due
> to its paradoxes.

And I of course think that it was unnecessary to do so since also the paradoxical allocations can be seen to be correct and ideal results.

> Of course, given the other things that work against equal
> voting power per person here, it makes not the slightest difference which
> apportionment method we use. In fact, our only real problem is Plurality.

Yes, definitely. Differences between the most well known allocation formulas are minor when compared to many other possible factors that may make the result biased.

> So, seats should be allocated to districts by Sainte-Lague.

Well, I accept that too, but I'm fine with the natural Largest Remainder.

Juho