[EM] PR favoring racial minorities

[Raph Frank]

On 8/25/08, Kristofer Munsterhjelm <km-elmet at broadpark.no> wrote:
>  There's a problem with this way of thinking, as can be made general to
> explicit voting schemes (such as ones based directly on opinion axes), and
> that is that it's impossible to ensure perfect representation on all the
> axes, so one will have to make a tradeoff. For instance, if the system
> requires n% to be of group X, and all of group X are extremely radical
> (support a marginal left-wing party whose general support is an order of
> magnitude less than n), then one has to make a tradeoff between the n% and
> political composition. (In my opinion, that tradeoff should favor what the
> voters want.)

What you could do is divide the population up using various axes.
They would have to be objective to a certain degree.  For example,
gender, age bands, average taxable income and perhaps ethnic group.

Each voter would then be assigned to the appropriate group and elect
someone from the group.  The ages/income thresholds could be adjusted
so that each group is roughly equal in size.

Each group would then elect 1 representative via some kind of single
winner method.  Since the group would be reasonably homogeneous, this
may not be as big an issue as for a more heterogeneous constituency.
Also, gerrymandering would be relatively irrelevant.

It would be pretty complex as ach polling station would be collecting
votes for each race.  Also, if the voting register is public, it would
in effect give your income band for each voter.

There would thus be a representative for e.g. "Men aged 30-35 earning
between EUR 50-75k a year".  It would create more explicit dynamics in
the legislature on tax policy and how progressive it should be.