Revised Predictions for 2015 General Election

In May last year, I had a stab at predicting how many seats each party would win in the coming general election.

The unforeseen (by me) rise of the SNP in Scotland has completely thrown out my predictions.

I also accept that my broad brush approach to universal swings was too rough an instrument as perhaps universal swings are too generous to Labour. Lord Ashcroft's constituency polling gives some more pointers as to what is happening in the marginals.

I gave predictions for scenarios ranging from a 6 point Tory lead to a 6 point Labour lead. The current Labour 2 point lead suggested a 40 seat overall majority for them (LAB 345 CON 245). A tad optimistic for Labour perhaps.

I rounded each party to a nearest 5 seat total and worked on the assumption of a combined "others" of 60 seats (Lib Dems with 25 of those). With Labour and Tory combined splitting 590 seats between them.

Now we are closer to the election, I think I can narrow that down and also incorporate more data from spread betting predictions, which tend to overestimate the Tories and underestimate Labour. In 2010 betting predicted the Tories getting 13 more seats than they actually got and Labour 40 less.

Bearing all of this in mind, my predictions are now as follows (with net change from 2010 result);

Using the current polling of Labour with a 2 point lead over the Tories in the UK but 17 points behind the SNP in Scotland.

LAB 279 (+21)
CON 269 (-37)
SNP 47 (+41)
LDEM 25 (-32)
DUP 8 (-)
UKIP 6 (+6)
SF 5 (-)
PC 4 (+1)
SDLP 3 (-)
GRN 1 (-)
RES 1 (-)
AP 1 (-)
OTH 1 (-)

Behind these figures, I have predicted Labour losing 33 seats to the SNP. The Lib Dems losing 13 seats to the Tories, 10 to Labour, 8 to the SNP and 1 to Plaid Cymru. And the Tories losing 44 seats to Labour and 6 to UKIP.

Assuming Sinn Fein with their 5 seats continue to not take up their seats, then 321 is needed for a majority.

I predict a Labour minority government with the SNP, SDLP, Plaid Cymru and Green providing confidence and supply (a total of 334 seats, a majority of 26 excluding Sinn Fein). Obviously this would be a very precarious arrangement unlikely to last anywhere near 5 years.

My Predictions for the 2015 General Election

In the 2010 general election, a 7 point Tory lead in votes over Labour (36%-29%) led to a seat allocation of;

CON 306 (47% of seats)
LAB 258 (40%)
LD 57 (9%)
OTH 29 (4%)

For every 1 point swing to Lab from 2010, Lab gain around 10 seats from the Tories.

So a 2 point Tory lead is a 5 point swing to Lab.

The LDEM vote is likely to be halved from 2010. This will reduce LD seats by around half, possibly 30 seats, 25 going to CON, 5 to LAB.

A UKIP vote of below 15% doesn't affect this formula. UKIP are likely to win less than 3 seats, possibly none...

So my predictions for 2015 are as follows (rounded to nearest 5 seats). 323 seats are needed for a majority as Sinn Fein do not take up their 5 seats in Westminster.

Tory vote leads;
2pts= CON 285 LAB 305
4pts= CON 305 LAB 285
6pts= CON 325 LAB 265

TIE= CON 265 LAB 325

Lab vote leads;
2pts= CON 245 LAB 345
4pts= CON 225 LAB 365
6pts= CON 205 LAB 385

LDEMS 25, OTHERS 35 for all above predictions.

If you want to allow for the effects of Scottish Independence, just deduct 40 from the Lab total, 10 from LDems and 9 from Others. 294 would then be needed for a majority.

Without Scottish Independence a 2pt Labour lead will give them a comfortable 40 seat majority. Even with Scotland going independent, Labour would still have a 20 seat majority. Tories need a 6pt lead for a much smaller majority or 4pt lead if Scotland leaves.

I'm sure you must be thinking; why do Labour get more seats than the Tories with the same voteshare?

There are 2 real reasons for this and 1 fake. The only reason you are likely to hear in the media is the fake one.

The real reasons are called "geographical concentration" of more of their votes in marginal seats rather than super safe or 'no hope' seats and "differential turnout" - different turnout of voters between seats you win and seats your opponents win.

Any boundary based voting system will always have these 2 biases to varying degrees, no matter how carefully you draw boundaries....

Basically Labour votes are more concentrated in seats they need to win, whereas Tory votes are more in seats they either cannot win or concentrated in seats they already win easily anyway. This geographical bias generally hurts smaller parties more than the bigger parties, but it can also affect 1 of the big 2, as in this case.

The other reason is that turnout in poorer urban seats that Labour win is lower, so are won with lower vote totals than high turnout Tory seats.

The fake reason you will hear in the media is that seats are "not equal" in terms of numbers of electors.
While true that seats do vary in numbers of "registered" voters and Labour seats are fractionally smaller by this measure. This has only a negligible impact on the seat results. Also it has to be stated that by "voting age populations" i.e. including all those eligible to vote whether registered or not, Labour seats are actually bigger than Tory seats.

The drawing of boundaries in our system is extremely important. It can be the difference between a landslide or a defeat. Yes it can easily make that much difference! Both Thatcher and Blair landslides relied heavily on favourable boundaries. In the UK the drawing of boundaries is "notionally impartial" and the responsibility of national boundary commissions. However both that local process and government guidelines are heavily party politically influenced.

And I have to say, impartiality has gone out of the window in my opinion. With increasingly both main parties looking for technical changes that bring party political advantage both in the drawing of boundaries and the registration process. This is the most worrying aspect of our system. It makes a mockery of democracy. The USA tells us where this is heading and it is frightening. We need a system where our representatives in parliament actually reflect the votes cast for them. We need to move away from systems where the drawing of political boundaries has such a big impact on the result.