Disproportionality in UK electoral system

[armb]

A neat interactive demonstration from the BBC:
http://news.bbc.co.uk/1/hi/uk_politics/election_2010/8609989.stm

With the three main parties getting equal shares of the votes, Labour very nearly have a majority of the seats. (Though there are clearly some slightly strange assumptions at the limits, because "Other" still get 24 seats if you set their share to 0.0%, and Northern Ireland only explains 18 of them.)

(via http://www.boingboing.net/2010/04/19/lib-dems-soar-in-uk.html)

Comments

[bohemiancoast.livejournal.com]

It's a foolish way of looking at it because individual seats don't swing uniformly. As long as people have good information about what people *in their local area* are doing, so they can make informed voting judgments, all sorts of fantastic things can happen in the real world.

[keristor.livejournal.com]

As the site says, if you read it, "this is a crude model - in reality every seat is unique". Since there is nothing which can determine which way voters will go in any particular seat -- that's why we have elections, if we could predict it then we could just appoint the people mechanically -- it's no worse than many other models and it does show how drastically biased FPTP is.

I'm pretty certain what people in my area will be doing, and so my vote won't make any difference at all because I'm in a (Tory) 'safe' seat. It would need all of the Labour voters in the seat to swing to Lib Dem to do anything about it and that won't happen, anyone who votes Labour here is pretty much hard wedded to that party.

[keristor.livejournal.com]

The others are hard core Plaid Cymru and SNP. Assuming that they don't swing (i.e. you drop the 'other' votes to zero without giving any one of the 3 main parties a noticable boost) they are places which will stay PC/SNP whatever happens (on past performance). Similarly, however, increasing the 'other' vote does very little to the actual number of seats won by 'others, it simply dents the other parties from which they have swung.

[the-magician.livejournal.com]

how drastically biased FPTP is.

All systems are biased in some way, it's whether you agree (broadly) with the bias that's important.

If we keep local MPs, but use national voting to determine the makeup of Parliament, then 5% voting BNP will give them around 32 seats in the House of Commons. There's not a constituency in the UK where the BNP have a majority, so 32 constituencies would have a local MP that was BNP and not the local choice.

In the US the two houses are determined in different ways, House of Representatives is basically by population, Senate is two per state (regardless of size or population, Alaska is big and empty (under 700k people in 660,000sq.miles, Rhode Island is small and much fuller, 1M people in 1,000sq.miles ... so Alaska is basically one person per square mile, Rhode Island is 1,000 people per square mile! But each state gets two senators ... worse compare Wyoming (500k people) with California (37million) and so you see the ratio of people to senators varies by 70:1)

[armb.livejournal.com]

My objection was that logically if the model has some seats which (realistically) stay PC/SNP, then it shouldn't let you (unrealistically) drop "Other" to zero, and vice versa.
But that's a detail, and the whole thing is an approximation, obviously. (That doesn't make it foolish, it would just be foolish to read more into it than is there.)