Today–one day before UK election day–the UK Polling Report average of polls shows the Tories on 35%, Labour on 28%, and Liberal Democrats on 27%. Needless to say, this would not be an outcome that matches Duverger’s Law, were it to materialize in the actual voting. But just how uncommon would such a result be?
In the database I assembled for my chapter in the edited volume by Andre Blais, To Keep or Change First Past the Post, I have 210 elections under FPTP. These include three or four decades worth of elections from the UK, Canada, New Zealand (before it changed from FPTP), India, Botswana, all but the smallest Caribbean island states, Belize, and all Canadian provinces.
In these 210 elections, how many saw the largest party have less than 36% and the third party have more than 25%?
One. Nova Scotia, 1998. (In the comments, Joffré reminds me that there has been one more since the period for which I collected data: Quebec in 2007. And it’s the best example of all, with the three parties ranging from 33.1% to 28.3% of the vote!)
If we expand the band a bit further, to something that might be more reflective of the final result (just a hunch, not a prediction), to less than 38% for the first party and still over 25% for the third, we pick up just two more: Nova Scotia 2003* and Ontario 1975.
In all of these three, the top two parties were within two percentage points of one another, whereas in the UK election it is likely that the second and third parties will be closer than the top two. Each of these three produced a no-majority situation (a hung or balanced parliament): The top two parties tied in seats in Nova Scotia 1998, while in 2003 the leading party (Conservative) won 48% of the seats on 36% of votes (whereas the Liberals managed 23% on 31%). In Ontario 1975 the Conservatives and Liberals were 41%-29% on seats despite being 36%-34% on votes.
In none of these did the third largest party by votes gain more than 30.4%, or less than 26.9% of seats. This election almost surely will see the third party do considerably better or worse than that. If the third party in votes is Labour–still a distinct possibility–it will almost certainly obtain over a third of the seats.** If it is the LibDems, the third party is likely to have not much more than 15% of seats. (Earlier projections that had Labour third in votes but first in seats no longer look likely, but such a result certainly remains imaginable.)
Looks like a lot of us over here will be staying up rather late on Thursday.
* There was an intervening election in 1999 that does not make the cut.
** In addition to UK Polling Report, see Politics Home. On the other hand, FiveThirtyEight is projecting Labour on 198 seats, which would match the 30.4% won by the third-place NDP in Ontario in 1975. (The FiveThirtyEight projections for the LibDems are on the high side–currently 113.)
Travel note: This was composed and uploaded somewhere south of Preston, on a train between Glasgow and London. Nice countryside (despite that junkyard we just passed), and farther north we saw lots of itty bitty lambs.