28 January 2014

In the unlikely event of a draw

Mike Smithson’s PoliticalBetting.com styles itself as “Britain’s most-read political blog – and the best online resource for betting on politics” and “2004-2014: The view from OUTSIDE the Westminster bubble”. Smithson comes 33rd on PeerIndex’s list of the most influential people on Twitter aged over 50, and, if there were a list for the over 65s, he would no doubt be second to Hilary Clinton. No sour grapes here – I read his blog and follow him on Twitter, where I was alerted to this on 21 January:

and it’s been making me think about some of the issues it raises, in particular what happens if the two main UK parties were to get the same percentage share of the vote in the 2015 general election.

I looked at the outcomes of the 18 general elections between 1945 and 2010:

It can be seen that, firstly, in only 3 of the 18 were the percentages within 1% and none between 1 and 2%:

so it isn’t a common occurrence, and,secondly, the percentages of votes given to the two main parties have been reducing over the period. This comes out in the next chart which shows that since the 1950s the “Market share” of the two main parties has been declining and has usually been below 75% for the last 40 years, and that the difference between the two main parties has been over 5% for most of the last 30 years. 

The final greyed out columns are not for an election but the recent position as concluded by the School of Politics and International Relations at the University of Nottingham blog, Ballots and Bullets (whose work I have quoted here before). I’ve taken their recent chart and extrapolated it (linearly!) to the assumption made by Smithson in his tweet relating to 7 May 2015 of Conservative and Labour drawing at 35%:

I thought it might be useful to make use of Electoral Calculus to conduct some sensitivity testing of Smithson’s result. The next chart (a bit of an eye-glazer for those who aren’t chart-minded) explores what happens in terms of seats won if the Lib Dem share of the vote is fixed at 8%, and the two main parties tie at 33 to 38%. UKIP gets squeezed from an 18% share at the 33% each level to only 8% if they get 38%. (The percentage of the vote given to Greens, Nationalists etc is kept at 8% level Smithson used.)

For reference I’ve added the Smithson point (35%/35% for the main parties, 12% UKIP, 10% Lib Dem). This, and other sensitivity testing which I haven’t plotted here, reveals that for likely situations in which the two main parties tie, ie in the 35% region and without a total collapse in the Lib Dem vote or a major upsurge in UKIP’s, the outcomes lie inside the dotted ellipse.

The final chart puts these outcomes in the broader context of securing a majority in a 650 seat House of Commons. For interest, I’ve added the results achieved in the two previous elections where the main parties had very close shares of the vote, 1964 and 1974, (these shares and the Lib Dem percentage shown) and were close to the number of seats to form a majority (1974 below, 1964 just above). Further elections followed soon afterwards.

If there is any precedent for a draw in 2015, it seems to be 1974 but such close shares of the vote are uncommon.


  1. " ... and follow him on Twitter, where I was alerted to this on 21 February:" Here I assume the month to be January. Apologies for being picky - I wish I had the expertise to comment on these very interesting posts at the same high level achieved by their author. David Martin

  2. Once again, many thanks for your kind and helpful comments - correction made!