See the blog of the Evening Standard’s Paul Waugh.

Does the focus on marginals miss the nature of electoral reform?

There has been much discussion of the Lord Ashcroft commissioned poll, undertaken in the marginals by Populus, which seeks to examine what would have happened if 125 key marginals had used Alternative Vote (AV) instead of FPTP (First Past The Post) to elect their MP. The article on ConservativeHome can be found here, while analysis by Anthony Wells can be found here.

Headline results suggest that the Tories would have done better at winning Labour held seats, and the Lib Dems would have done better at holding against the Tory onslaught. Overall, the tenor of the original ConHome article by Lord Ashcroft was that Lib Dems should be careful of what they wish for - that electoral reform might favour the Conservatives over the Lib Dems in what we call key constituencies.

I have a number of quibbles with this. Firstly, as I’ve always said, I am suspicious of “polls of marginals” in general. Polling seats that happen (for manifold reasons) to be highly competitive between first and second place (be that 49-48-3 or 25-26-15-15-15-5, with a variety of different parties competing) seem not to be a natural class that you can poll separately. Scotland is, because it is the only area where the SNP run (so you could do a Scottish poll). A poll of urban centres similarly I could understand. Polling seats that happen to be ‘marginal’ in any given election seems to be to be like polling seats that happen to begin with the letter ‘C’.

Beyond that, I am not convinced that marginals are the right sub-set of seats to poll to show the differences between FPTP and AV.

Marginals, as we understand them, are seats where (under FPTP) first and second place are no more than a few thousand votes apart. You can win a marginal on less than 33% (Hampstead & Kilburn), or on closer to 45% (Tooting) - the difference means nothing in FPTP (they are both ‘marginals’) but could be important in AV (because of how the remaining vote is distributed).

I commented on ConHome and Iain Dale’s articles that in the 2010 General Election, looking only at the 632 seats in Great Britain (ie excluding Northern Ireland) only 150 or so were even vaguely competitive, and arguably fewer (closer to the 125 polled by Populus) were ‘marginal’. By contrast, only 212 seats were won with over 50% vote (75 Lab, 125 Tory, 12 LD) and only 84 of these were won with over 55% of the vote. There might be 450 safe seats at the moment, but that number will drop significantly if AV comes in.

Marginals are a key subset of the seats that are ‘competitive’ - under AV, a massive number of seats will become competitive for the first time. Partly for reasons of mathematics (enforced tactical voting if you like), but also due to the changes in the way seats are seen by electors: I think even the 212 seats which were won in 2010 with over 50% of the vote will come into play, because the reason of large shares of the votes is indicative of voters thinking there is no point voting against a ’safe’ incumbent. In short, I think the very idea of there being a handful of ‘key marginals’ will simply cease to be as relevant under AV.

This is not an argument in favour of electoral reform - I don’t know what I think about AV, and as much as making more seats competitive makes it more interesting for me as a pundit, I doubt the parties are well-financed enough or organised-enough to manage such distributed campaigns, meaning marginals might continue to matter in effect if not in theory for a few elections to come. But generally, I think that if we are to understand the effect that AV might have, polling 125 seats that happened to be marginal under FPTP is the wrong way to go about it.

Beyond the polling, there is also a question to ask about politics - if AV is to be doomed at the planned referendum, then breaking the already quivering enthusiasm of Liberal Democrat voters will be key. Purists for PR are already having to make difficult decision - by emphasising that AV might actually work for the Tories more than for the Liberal Democrats, I cannot help but wonder if this polling is more useful as a statistical cat being put amongst the yellow pigeons, than it is an examination of how AV might really affect the Tories.

Morus

But are the brothers really neck and neck?

There’ve been sharpish moves to EdM overnight on Betfair following a projected result published on the Left Foot Forward, the blog edited by Will Straw.

On Monday EdM was trading in the 3.75 - 3.8 range. This morning at 0430 it had tightened to 3.25 - 3.3. At the same time the DM price has eased.

The article sets out in detail how it has worked out its figures in the hope, no doubt, of provoking a debate that could produce other insights.

To my mind the weakest part of the model is its reliance on a self-selecting LabourList “poll” for the members section.

There is also the question of the extent to which union votes will be influenced by the endorsements of their executives - a big factor in the 2007 deputy race. Here we will will surely see a higher turnout which might mean that the unions’ recommendations matter less.

This is going to be down to the lower preferences and it’s the order in which supporters of Balls/Burnham/Knight rate Ed and Dave that will decide this election.

Today I’m having a smallish operation but should be fit again by the end of the afternoon.

By then I hope we’ll have the detailed data from the Michael Ashcroft funded Populus poll on the impact that AV might have on an election outcome. All we’ve got so far is the data that he’s selected and there are a number of unanswered questions. For instance it’s not clear whether the samples are past vote weighted or not. As it stands the numbers are not good for the Lib Dems.

All private polling is made public for a reason and we’ve seen time and time again that it’s wise to wait for the full data.

Mike Smithson