UK Election Results

With the results declared in 649 constituencies, here's a summary of the UK election results. First, let's look at the likely number of seats for each party (the remaining one will probably go Tory in a delayed election), with their likely final GB vote share in parentheses:

CON - 307 (37%)
LAB - 258 (29.7%)
LD - 57 (23.5%)
Other - 28 (9.8%)

Of the four polling averages that I calculated just before Election Day, the most restrictive one, counting only last-minute polls by established firms, performed best. Compared to that average, the Tories got +1.2%, Labour got +1.7%, while the Lib Dems got -3.5%. That 5-point Lib Dem loss relative to the other parties explains their woefully disappointing result - they actually lost 5 seats relative to 2005! This election can also be considered a failure for pollsters, especially Angus Reid, which completely screwed up their entry into the UK election polling scene by lowballing Labour's share at 23-24% throughout the campaign.

As it turns out, the uniform national swing (UNS) model would have indeed overstated Labour's seat count and understated the Tory one, though not quite by as much as I thought, and certainly be less than than FiveThirtyEight thought. However, because the Lib Dem vote share barely changed in the last election, UNS, like just about any sensible model, would have pinned down their seat number quite well.

The outcome of this election is right in the middle of the chaos range, where neither Labour+Lib Dem (315) nor Conservatives+Unionists (317) have a majority out of the 645 non-Sinn Fein seats. If the Tories don't get Lib Dem support, they will need some combination of SNP (6), Plaid Cymru (3), SDLP (3) and Green (1) backing. This will be very interesting, as all four of these parties are left-leaning. The excitement isn't over yet!

Of the 14 projections I noted, 10 accurately predicted a lack of working majority for any party, with the Tories as largest party. LSE was inaccurate in having Labour as the largest party, while The Times, Angus Reid and BC Iconoclast were wrong in projecting a Conservative working majority (with unionist support in the case of the Times).

Here are the 3 closest projections for each of the major parties:

CON: 1 (tie). YouGov and UK Polling Report (305); 3. Canadian Election Watch (310)
LAB: 1. Election Prediction Project (245); 2 (tie). YouGov and Electoral Calculus (235)
LD: 1. YouGov (80); 2. LSE (81); 3. The Times (82).

Adding up the absolute deviations for the 3 big parties gives the following ranking of projections:

1. YouGov - 48
2. Electoral Calculus - 62
3. ComRes - 63
3. UK Polling Report - 63
5. Riding By Riding - 67
6. Election Prediction Project - 68
7. Trendlines - 74
8. Canadian Election Watch - 84
9. The Times - 85
10. FiveThirtyEight - 105
11. LSE - 108
12. TNS - 126
13. Angus Reid - 161.5
14. BC Iconoclast - 212

Essentially, due to the Lib Dems' unexpected weakness and Labour's unexpected strength, projections relying more on UNS did better than those taking into account geometric vote transfers. With a middle-of-the-pack ranking, I can't say that I'm very happy, but at least I wasn't laughably bad...

The silver lining - and I think a rather important one - is that the most important number in these projections is the Tory seat count, since that dictates how easy/hard of a time David Cameron will have in forming a government. On that account, I ranked 3rd out of 14, just 1 off YouGov and UK Polling Report!

UK Election Tomorrow

Ten polling firms have published polls following the third and final debate last week. Here are the averages of the most recent poll by each firm, and changes from last week:

CON - 34.8% (+1%)
LD - 28.2% (-1%)
LAB - 26.7% (+0.7%)

Three of these firms are newcomers to UK political polling. Excluding them gives:

CON - 35.6% (+1.4%)
LAB - 27.9% (+1.1%)
LD - 27.4% (-1.1%)

If we instead exclude the three firms that did not conduct a poll in the past 3 days (so their most recent poll was either last weekend or conducted over the entire week), we get:

CON - 35.9% (+1.9%)
LAB - 27.4% (+1.1%)
LD - 27.3% (-1.3%)

Only including the six "old" firms that conducted a poll in the past few days gives:

CON - 35.8% (+1.7%)
LAB - 28% (+1.2%)
LD - 27% (-1.3%)

So, what's clear is that the Lib Dems have lost momentum in the past week, while the Tories and Labour have gained some. Also, the Tories are obviously in front, while Labour and the Lib Dems are very close.

I expect the Tories winning by about 8 points, while Labour and the Lib Dems should be roughly tied for second. I expect the final seat count to be roughly:

CON - 310 (including Speaker and the delayed election in Thirsk and Malton)
LAB - 214
LD - 94
Other - 32

Around the web, here are some projections, including both the Speaker and T&M in the Conservative column (Con-Lab-LD-Oth; midpoint if only range is specified):

UK Websites
YouGov: 305-235-80-30
ComRes: 299-233-87-31
LSE: 251-286-81-32
The Times (based on Ladbrokes): 322-213-82-33
Electoral Calculus: 297-235-86-32
UK Polling Report: 305-225-85

International Websites
FiveThirtyEight: 312-204-103-31
Election Prediction Project: 284-245-90-31 (updated)

Canadian Websites
Riding By Riding: 300-230-90-31 (sums to 651)
TrendLines: 299-227-93-31
BC Iconoclast: 337-151-132-30 (updated)

Other Websites (via UK Polling Report)
TNS: 292-204-114
Angus Reid: 330-175-112.5

So the ranges are:
CON - 251 (LSE) to 337 (BC Iconoclast)
LAB - 151 (BC Iconoclast) to 286 (LSE)
LD - 80 (YouGov) to 132 (BC Iconoclast)

Excluding LSE and BC Iconoclast:
CON - 279+ (Election Prediction) to 330 (Angus Reid)
LAB - 175 (Angus Reid) to 242+ (Election Prediction)
LD - 80 (YouGov) to 114 (TNS)

Let's see how these 14 projections, some based on formal models and others (like mine) based on hunches, perform!

UK Election: 5 Days before Polls Close

Four polling firms have already completed surveys after Thursday's election debate in the UK. Here are the averages, with the change from the average of these same 4 firms' last results before the debate in parentheses:

CON - 35.5% (+1.75%)
LD - 28% (-0.5%)
LAB - 27% (-0.25%)

Note that, based on their last polls before the last debate, these firms had Labour about 1 point higher on average than all 11 firms (that I know of) combined.

According to the BBC's Election Seat Calculator, which assumes a uniform national swing, we would get CON 286, LAB 251, LD 84. Adjusting for problems with a uniform swing in an election with large changes, I expect something like CON 310, LAB 205, LD 105 if current numbers hold.

In the above scenario, the 30 seats belonging to small parties become very important, since the Tories might be able to govern with the support of small parties rather than the Liberal Democrats. A majority is nominally 326 seats this time, but there are complications:

- In the UK, not only the Speaker, but also three Deputy Speakers do not vote; in the previous Parliament, the Speaker and two Deputies were Conservative (the other Deputy was from Labour).

- Sinn Fein does not take its seats; in the 2005 election, they had won 5.

Thus, if the Northern Irish electoral results are replicated, there would be 641 voting members in the House, meaning that 321 votes would effectively form a majority. If the Tories still provide the Speaker and two Deputies, they would need 324 seats (or 323 if you count the Speaker separately) for a voting majority. If they gain the support of the unionist Northern Irish parties, which are right-wing, and if these parties win 10 seats like they did last time, then the Conservatives would need 314 seats (313 not counting the Speaker). All other small parties lean left, and are thus less natural partners for the Conservatives. Finally, the Tories may decide to provide fewer than 3 of the 4 non-voting members of the House, though they will almost certainly still provide the same Speaker, and probably at least one Deputy Speaker.

So this is how the situation for the Tories looks to me, counting the Speaker (and assuming Northern Ireland results remain unchanged from 2005):
324+: majority
322-323: majority through giving up on one or both Deputy Speaker positions
314-321: majority through alliance with Unionist parties
312-313: majority through both of the above
More than 298 or so: need support of at least some left-leaning small parties to govern without Lib Dem or Labour support
Less than 297 or so: require support of Lib Dems or Labour

Currently, the Tories may well be smack in the middle of that 298-323 ambiguous area, although my guess is that they will look for Lib Dem support before trying to woo over multiple small left-leaning parties. The last 4 days of campaigning will be crucial!

One Week from UK Election

The third and last debate in the UK election campaign has now taken place. In the week before this debate, 11 firms have polled UK electors, with the following results (numbers from 9 firms with polls wholly after the 2nd debate in parentheses; the other 2 firms had a small fraction of their polling before the 2nd debate):

CON - 34.0% (34.2%)
LD - 28.6% (28.3%)
LAB - 26.4% (26.4%)

Comparing these results to the average of the corresponding polls from a week earlier gives:

CON: +1.9% (+2.2%)
LD: -1.8% (-2.2%)
LAB: -0.3% (unchanged)

So basically the Lib Dems lost 2 points to the Tories in the past week, while Labour kept its ground. It will be interesting to see whether Bigotgate and the third debate generate some movement. An instant poll suggest that the 3rd debate is likely to have little effect on voting intentions.

A uniform swing with the current numbers would give something like CON 270, LAB 255, LD 95. As I've cautioned before, this likely overestimates Labour's count and underestimates the other two parties'. FiveThirtyEight takes into account and projects something around CON 300, LAB 200, LD 120, which looks reasonable.

UK Election After Second Debate

First, here's the average of the last poll before the second debate from 10 polling firms. Of these 10 firms, 7 have already conducted a poll after the second debate; the pre-debate average of these 7 firms is in parentheses.

CON - 31.9% (31.9%)
LD - 30.5% (30.3%)
LAB - 26.4% (26.7%)

Here's the post-debate average, with change from the pre-debate 7-firm average in parentheses:

CON - 34.3% (+2.4)
LD - 29% (-1.3)
LAB - 26.6% (-0.1)

These post-debate numbers give, according to UK Polling Report (change with respect to pre-debate 7-firm average in parentheses):

CON - 270 (+31)
LAB - 254 (-14)
LD - 95 (-17)
Other - 13 (0)

The Tories have taken the lead, but do not have a majority. The usual caveat still applies: Labour is likely to get fewer seats than predicted by a uniform national swing, while the Lib Dems and the Tories are likely to get more. It looks very likely that if an election were held now, the Conservatives would be the largest party, followed by Labour and the Lib Dems.

Could a Labour-Lib Dem coalition arise? The above prediction gives them 349 seats, against 270 for the Tories and 31 for others (13 from GB + 18 from Northern Ireland). But in reality, Labour-Lib Dem should get fewer than 349 seats. If they fall below the majority threshold (theoretically 326, but in practice around 323 due to Sinn Fein not taking its seats and the Speaker not voting) without the Tories rising above it, Labour and Lib Dems would need the support of small regional parties. Alternatively, the Tories could try to woo the Lib Dems or maybe the small parties into government.

The UK could also get a minority government like in Canada. I doubt Labour would try, or be able, to stay in power without a formal coalition if they finish third in popular vote. However, as the largest party, the Tories could try to do that.

This election will be fascinating to follow!

UK Update

I have now seen polls from 9 firms since last week's UK debate. Averaging the most recent one from each gives (change from Monday's post in parentheses, +/- 0.1% due to rounding):

CON - 32.3% (+0.2%)
LD - 30.6% (+0.2%)
LAB - 26.6% (-0.3%)

3 of the 9 polls considered are the same as Monday, while 3 firms have released updated polls; of course, 3 new firms were added to the mix. The corresponding seat count according to UK Polling Report (which, again, is probably too optimistic for Labour and too pessimistic for the Lib Dems):

LAB - 259 (-6)
CON - 246 (+3)
LD - 114 (+3)
Other - 13 (0)

Let's see what movement the second debate will yield... (And you've got to give it to the British: after not having any televised leaders' debates until now, they're going with 3 in 15 days!)

UK Election

6 different firms have completed polls following last Thursday's historic debate. Doing a straight arithmetic average of the polls gives:

CON - 32.2%
LD - 31.2% 30.3%
LAB - 26.8%

However, according to UK Polling Report, assuming a uniform swing gives the following distribution of the 632 British seats (18 more are allocated to Northern Ireland - none of them are expected to go for any of the three major parties):

LAB - 261 265
CON - 242 243
LD - 116 111
Other - 13

(The BBC's calculator gives the same result for each party, +/- 3 seats.)

If the popular vote pattern stays stable (which is of course a huge 'if', given the tremendous shift due to the debate), these seat projections are likely to be too pessimistic for the Lib Dems. The reason is that when a small party's support increases, assuming that the increase is uniform disadvantages it in a first-past-the-post system. Consider the following stylized example:

Before the surge
Seat A: Lab 40, Con 30, LD 20
Seat B: Lab 40, Con 30, LD 20
Seat C: Lab 35, Con 35, LD 20
Seat D: Con 40, Lab 30, LD 20
Seat E: Con 45, LD 25, Lab 20
Total: Con 36%, Lab 33%, LD 21% (Con 2.5 seats, Lab 2.5 seats, LD 0 seats)

After an uniform surge (+10 LD, -5 Con, -5 Lab)
Seat A: Lab 35, LD 30, Con 25
Seat B: Lab 35, LD 30, Con 25
Seat C: Lab 30, Con 30, LD 30
Seat D: Con 35, LD 30, Lab 25
Seat E: Con 40, LD 35, Lab 15
Total: Con 31%, LD 31%, Lab 28% (Con 2.33 seats, Lab 2.33 seats, LD 0.33 seats)

A uniform rise in Lib Dem fortunes allows them to finish second in ridings A, B, D and E, which were Labour or Tory strongholds. This of course doesn't give the Lib Dems any seats, so all they can hope for is that the 3-way race for seat C goes their way.

However, as we saw with the Conservatives in Québec during the 2006 general election, rising fortunes may well be distributed unevenly. Suppose the Lib Dems' support doesn't change in seats B and D, but their surge doubles in seats A and E. Then we get:

After a skewed surge
Seat A: LD 40, Lab 30, Con 20
Seat B: Lab 40, Con 30, LD 20
Seat C: Lab 30, Con 30, LD 30
Seat D: Con 40, Lab 30, LD 20
Seat E: LD 45, Con 35, Lab 10
Total: Con 31%, LD 31%, Lab 28% (LD 2.33 seats, Con 1.33 seats, Lab 1.33 seats)

In reality, the Lib Dem's increase in popularity is unlikely to be nearly as uneven as in the above example: if their popular vote stays where it is, they will likely end up with the most 2nd place finishes (and the fewest wins). However, it also probably wouldn't be as bad as the standard uniform swing model predicts.

With 2 debates left, this will be an interesting election to follow!