Disproportionate Democracy

The results of today's election in the UK may look a little bit funny and it calls to mind a myth I was taught in college – the myth is that multi-party parliamentary democracies better represent the will of the people than the American two-party system.  I think this is a myth and the results from the UK bust that myth. I can't be 100% sure because I'm writing this post a day in advance of actual election results. But if those results look like the polls, there will be a hung Parliament, and two parties will need to form a coalition government. Indeed, projects overall looks like the Liberal Democrats got really screwed -- their popular polling will have been far in excess of the number of seats in Parliament which they actually won. Fivethirtyeight.com is running this projection as of the time I write:
 

Party	Popular Vote	Seats in Parliament	Disparity
Tory	34.2%	308	+13.18%
Labour	26.3%	198	+4.16%
Liberal Democrats	27.5%	113	-10.12%
All Others	12.0%	31	-7.23%

This may look a little, well, undemocratic in the sense that the results of representation in Parliament are significantly disproportional to the way the British electorate is predicted to have behaved. If the predictions are pretty much accurate, the Liberal Democrats will have outpolled the Labour Party but have significantly fewer members in Parliament than the just-tossed-out-of-power Labourites. And I suppose that in the pure sense of the word, this is indeed counter-democratic. If the Tories get 34.2% of the vote, they should get 34.2% of the seats in Parliament, right? If there were strict proportional representation, today's results (as projected yesterday by fivethirtyeight.com) would look like this:
 

Party	Popular Vote	Seats in Parliament	Disparity
Tory	34.2%	222	-0.05%
Labour	26.3%	171	+0.01%
Liberal Democrats	27.5%	179	+0.04%
All Others	12.0%	78	even

Now, the disparity between the popular vote and the representation in Parliament here is well within limits that any reasonable political analyst would find tolerable. But such a system would leave a badly-fragmented Parliament, with any two of the three major parties able to form a government if they could agree on how to do so – the Conservatives, despite winning a plurality, could find themselves "on the outs" as Labour and the LD's jam together a 25-seat majority. That, too, would be an undemocratic result, and is proof enough that strictly proportional democracy does not guarantee a democratic result.
 
How does this happen? It's a combination of the relative popularity of the minor parties, the regional and single-issue parties and the Greens, lumped together in the "All Others" category. It's also the "first past the post" rule, which really means that the plurality in a given district is going to be the one seated in Parliament. If the UK had, effectively, a two-party system with the Tories and the Labourites as the only realistic choices, it would look a lot like the USA. (More about that below.) But they don't, so in each district the minor parties drain votes from the major parties and diminish the margin of victory for the eventual winner, potentially even changing the identity of that winner. 

For U.S. Readers unfamiliar with the British political system, a few quick concepts to keep in mind.  First, the UK is divided into geographical districts, just like the various states and Congressional districts of the U.S.  The number of districts varies from election to election; this year, there are 650 districts so it takes 326 seats to have a majority of Parliament and form a government. Individual candidates from the various parties stand for election in the various districts.  It is not required, but helpful, that the candidate be a resident of the district.  Suffrage is universal, with a few exceptions -- most interestingly, members of the House of Lords are not allowed to vote, and while there is no law disenfranchising members of the royal family from voting, they traditionally do not vote.
 
My expectation is that the 538.com projections are going to be pretty good; they line up roughly with what the BBC is projecting too: by the time this article posts to the blog, the Tories will have won a plurality but not a majority of seats in Parliament. Thus, the Queen will invite the leader of the Tories, David Cameron, to form a government of some kind and serve as Prime Minister for the next term of Parliament.  So what I'm pretty sure is going to be standing out today is that the Liberal Democrats polled so well but got so few seats. But in fact, this is nothing new for the UK. In 2005, the British election results were as follows:
 

Party	Popular Vote	Seats in Parliament	Disparity
Labour	35.3%	308	+19.81%
Tory	32.3%	198	-1.65%
Liberal Democrats	22.1%	113	-12.50%
All Others	10.3%	31	-5.66%

Note that in 2005, there were 646 seats in Parliament, not 650 as in today's election. Although Labour won that election instead of the Conservatives, the same thing that was projected for today's election happened in 2005 – thanks to the rule that the individual candidate who gets the plurality of votes within the geographic district is the one who goes on to sit in Parliament, a very modest advantage in polling produces a significant positive disparity in legislative representation.  And it was at the expense of the LD's, who did pretty well at the polls but did not get anything like the reward this would seem to indicate in seats in the Commons.
 
So, is the UK an undemocratic place despite being the "Mother of all Democracies"? (Greek and Icelandic Readers may now justifiably bristle at this phrase.) Let's compare the UK with the last round of elections in the nearest analogue to the House of Commons that we can find in the United States – the House of Representatives, using figures from the 2008 Congressional elections:
 

Party	Popular Vote	Seats in House	Disparity
Democrats	53.06%	257	+6.02%
Republicans	42.44%	178	-1.52%
All Others	6.00%	0	-6.00%

The Republicans should have got proportionally more seats than they did, but not that much and they certainly were better treated by their system than the Liberal Democrats were by the UK's. It's the minor parties that get trod upon by the majority party in the U.S.'s two-party system. I would expect to see similar kinds of disparities in historical Congressional elections, looking back to the days when Republicans held the majority in the House – whichever party gets the most votes is going to do so at the expense of representation of parties getting the fewest. But the point here is that the U.S., with a two-party system, still suffers from disproportionate representation. It gets closer, though, to the popular preference than the UK system.
 
That suggests to me that the insertion of a third party skews things in favor of the plurality party and warps the translation from voting to representation.  To confirm that, let's take a look at a third parliamentary democracy, one in a multi-party system with geographic districts rather than a national proportional representation system, and one in which there are a lot of parties in play. 

France's National Assembly held elections in 2007, with 577 seats up for grabs between two major parties and a scad of minor parties. In France, a candidate must get an absolute majority to win a seat in the National Assembly, and if no candidate gets a majority in a particular district, there is a runoff election. In all but one case in 2007, the runoff election was a two-way contest between the top two candidates from the first round of elections. The results, on the same scale of comparison used for the UK and the USA, look like this:
 

Party	Popular Vote	Seats in National Assembly	Disparity
Union for Popular Movement (Sarkozy)	39.54%	313	+14.71%
Socialists (Royal)	24.73%	186	+7.51%
Democratic Movement (aligned with Sarkozy)	7.61%	3	-7.09%
Communist Party (aligned with Royal)	4.29%	15	-1.69%
New Centre (aligned with Sarkozy)	2.37%	22	+1.44%
Greens	3.25%	4	-2.56%
All others, both left and right	18.21%	34	-12.32%

So the French system is at least as skewed as the UK's. The proliferation of parties significantly distorts popular support for the various parties' platforms as compared to support for those platforms in the legislature.  

The US system seems to produce results closer to the preferences of the people, perhaps in part because the U.S. effectively narrows the choices down to two parties. Now, this isn't an in-depth comparison of a lot of systems. But it appears that the more parties are in play in a system whereby representatives are elected from specific geographic areas, the more the plurality party benefits in the form of disproportionately high representation in the legislature. The U.S., with its two-party system, winds up coming the closest to actual proportional representation of the three countries I looked at here.


Sigh.  Another myth from college busted. And too bad for Nick Clegg and his Liberal Democrats (barring an upset that I can't predict as I'm writing a day in advance of the actual results). Perhaps the LD's will be invited to form a government with the Tories, although it looks to me like the Tories will wind up being pretty close to a majority and therefore able to maybe throw a few bones to the regional or minor parties and put together a majority that way, which looks like the easiest thing for David Cameron, the likely new Prime Minister, to do.  After all, the "going it alone" strategy seems to have worked, albeit inelegantly, for Stephen Harper in Canada.