14 March 2011

The 5,000 Labour Voters Who Secured This Tory - Lib Dem Coalition

When the results came in after last year's general election, one thing became clear: the only stable coalition on the cards would be between the Conservative and Liberal Democrat parties. Given the Tories and Labour are too belligerent to work with each other, the one remaining alternative would be between Labour and the Lib Dems. But the election result didn't allow for this: Labour seats + Lib Dem seats only comes to 315, but 322 MPs are needed for a working majority (accounting for Sinn Fein MPs not taking their seats).

With 7 more MPs, the Lib Dems would have been able to form a coalition with Labour. So if 7 of the Con/LD marginals were won by the Lib Dem instead of the Conservative, a Labour - Lib Dem coalition would have been a real option.

So in these 7 seats that where the Lib Dems came closest to beating the Conservative, Labour voters who didn't vote tactically for the Lib Dems actually voted to ensure that Labour weren't in power. The totals shown are the number of Labour to Lib Dem switches needed to defeat the Tory:

1. Camborne and Redruth: 67
2. Oxford West and Abingdon: 177
3. Truro and Falmouth: 436
4. Newton Abbot: 524
5. Harrogate and Knaresborough: 1040
6. Watford: 1426
7. Montgomeryshire: 1185

That's a total of 4,852 Labour voters who ensured Labour were out of power and helped the Tories in.

If you think this is stupid, I agree. Under AV, Labour voters would have been able to vote Labour as their 1st choice and Lib Dem as their second, making sure their vote didn't do the exact opposite of what they intended. This is yet another reason to ditch our broken political system and vote Yes in the AV referendum on May 5th.


mark.e.scott said...

Nice one. Of course with AV, LibDems would potentially be perpetual kingmakers.

Duncan Stott said...

Not true, Mark.

There's data for how voters would have used their 2nd preferences going back all the way to 1983. AV would have delivered majority government in 1983, 1987, 1992, 1997, 2001 and 2005.

Only in 2010 would AV have resulted in a hung parliament, the same as when a hung parliament was delivered under the old system.

I have grave concerns about the Lib Dems' ability to get over the 50% mark given the unpopularity that the coalition is giving us. I think AV could well end up hurting us rather than helping.

But no-one knows for sure. Anyway, we should look at the systems and decide what is fairer for voters, not for the parties.

Anonymous said...

Using data from elections under FPTP and extrapolating hypothetical AV results from them is at best naive, at worst duplicitous.

Neal said...

The only study I've seen that modelled AV first choice votes as well as supposed 2nd prefs was the 2010 one. And the parliament was MORE hung than it is now.

Do you think it would have been right is Gorden Brown who (by any reasonable measure,lost the 2010 election) was propped up by the Lib Dems who gained only a couple of percent of the popular vote that they had before? The overwhelming swing was from Labour to Tory, not Labour to Lib Dem!

Lee Griffin said...

At best it's accurate, at worst it's inaccurate. Hyperbole isn't really necessary, anonymous.

ciphergoth said...

Isn't this rather incomplete without the seven closest Tory/Labour marginals?

Lee Griffin said...

It's not incomplete, it's just one view. You could, of course, look at the Lab/Tory marginals as well, or instead.

Fact remains that based on how everyone voted, it can be said 5000 labour supporters stopped their own chances of being in government.

Duncan Stott said...

ciphergoth, in Labour/Tory marginals, a Labour voter increases the chance of Labour getting elected and Labour getting power by voting Labour.

Lib Dem voters are the tactical voters in these seats. They could vote tactically for Labour if they wanted Labour in power, or Conservative if they wanted Conservatives in power.

Opinion polls gave a mixed picture of which way the Lib Dem vote would lean. Very generally, the sentiment for Lib Dem voters was that Labour were preferred over Conservatives, but Cameron was preferred over Brown.

So it is far less clear cut how a Lib Dem would vote tactically.

Given both Labour and Conservatives are open to the idea of entering coalition with the Lib Dems, but not with each other. This is what creates the imperative for tactical voting by Labour voters but not for Lib Dem voters.

Duncan Stott said...


I didn't extrapolate any results. I just said that AV stops this stupid situation from occurring.

Roger Moore said...

Do you have data on how many lib dem voters it would have taken to switch to Labour in order to cause a similar situation taking place?

Duncan Stott said...

Roger, the LD to Lab switchers needed for Labour to win enough seats to form a workable Lab/LD coaliton are as follows:

1. Warwickshire North 55
2. Thurrock 93
3. Hendon 107
4. Cardiff North 195
5. Sherwood 215
6. Stockton South 333
7. Broxtowe 390

That comes to 1,388.

But please note my comment to ciphergoth above. LD voters have LDs in government (albeit not in the form some may wish), and generally speaking didn't want Brown as PM, so is not an equivalent situation.

However, the other case that could be made is that LD voters would want to create a situation where both a Con/LD and a Lab/LD coalition were workable, giving the Lib Dems a stronger hand in negotiations. Therefore LD voters could have tactically voted Labour to engineer these circumstances. I think there is a strong case for this, but is harder to judge, as it requires voters to know what the balance between Labour and Conservative MPs will be before any votes are counted.

Lewis Baston said...

Colour me unconvinced.

1. It would have taken more than 7 more seats to make Lab/LD a viable option. Given backbench tendency to rebel, the need to get John McDonnell and David Laws into the same lobby not just once but every time... a government would need more padding than a bare majority of 1 to see out a parliament. Probability of an early election would affect behaviour, particularly among LDs in seats threatened by Tories, and become self-fulfilling prophecy.

2. The possibility of a technical Lab/LD majority would probably have strengthened LD hand in coalition talks with Con a bit, but this might not result in changes that Labour voters would necessarily like more than the alternative, and this leverage would vanish as soon as the coalition agreement was signed.

3. There is still an element of choice here, and the choice of direction was made narrowly - and probably unknowingly - by LD members in 2007. And there's a choice of external support/ confidence and supply or full coalition.

4. AV will still be open to hypothetical shuffling of votes affecting the macro result. Three-way marginals will probably become more common under AV, and the outcome will be affected by which party is second and third. A few votes changed might result in a Lab/LD rather than Lab/Con final round. AV is less susceptible to tactical voting than FPTP but it is still a factor.

5. It's not those Labour voters' fault that Con/LD was the only majority coalition option - there seems to be some intimation of blame here which would be wrong. Ex ante, there would be no way of knowing how things would end up in the constituency and nationally. Given the outcome in seats and % vote was expected to be better for LDs than it turned out, it would have been rational on the basis of expectations for a Labour supporter in Falmouth & Camborne (say) to assume the Lib Dems would win locally, and vote Labour because it seemed possible that Labour would need every vote to avoid coming third in the popular vote. I know your point was about the inadequacy and perverse effects of FPTP, but it sounds a bit accusatory.

Hywel said...

A slight flaw though - Watford was Labour held until 2010.

Anonymous said...

The count system in AV does not work
eg. Ballots for 3 candidates

45ballots for a with second c
35ballots for b with second a
20ballots for c with second b

b wins but a has more first votes than b and more second votes than b .

Anonymous said...

AV does not identify and select a "majority" candidate.
Another set of results can be used to show a mid placed candidate winning the AV count when they have less first votes and less second votes than another mid placed candidate.

Anonymous said...

I put " " around the word majority because AV does not acheive identifying and selecting a majority candidate.

Duncan Stott said...

I'm not suggesting AV is perfect. The solution to this is something called Borda counting. Look it up if you're interested. It is a far more complicated procedure at the count than counting AV. But for the voter, the way of marking your ballot is the same.

But despite this imperfection, AV is better then FPTP. AV does find a majority, but may miss the biggest majority. But FPTP fails to find a majority 2 out of 3 times.