UK Elections for Investors

I've been meaning to write something on the election for a while, so thought this poll (http://www.timesonline.co.uk/tol/news/politics/article7054655.ece) was as good an excuse to do it. Populus/The Times have polled 100 marginal constituencies (seats 51-151) putting both Labour and the Conservatives on 38% of the vote, representing a 6.7% swing to the Tories. The headline focuses on the Tories & Labour being "neck and neck in the marginals" (news of an inevitable victory doesn't sell papers!). However, this actually is pretty much where the Tories need to be in order to get an overall Majority. There are a lot of misconceptions about the national polls and the seats, majorities and hung parliaments that they supposedly translate into, and as someone who has been sad enough to watch the 1979 and 1992 election coverage start to finish (you can see it here if you're interested http://www.youtube.com/user/ajs41) I think it's worth highlighting these.

In particular, the Uniform National Swing (UNS) is the most frequently quoted approach in the Media and accepted by non-professionals. This basically does what it says on the tin: a swing in the survey group is extrapolated uniformly on a nationwide basis for the three main parties, and the rest (SNP etc.) are held constant. The simplicity of this approach is obviously quite attractive, but historically, it has not been a particularly good predictor of the eventual result - most notably in 1979 and 1992. This is a result of several phenomena which have been observed, particularly in relatively close elections. For example, the poster-child of the 1992 election was the "shy Tory" effect, where many voters are embarrassed to say that they will vote Tory as it is distinctly "un-cool" - I note Lily Allen dedicating her "F*** you" song to David Cameron last week.

The second phenomena is the discrepancy between marginal constituencies and core constituencies. As a result of a large amount of funding (usually from the Tories), there is a very large focus upon campaigning in marginal constituencies which tends to increase the swing within them. Indeed, the above Populus/Times poll suggests that this swing is to the tune of about 1.5-2% higher than recent nationwide surveys would suggest. As a result of desperation to "keep the other lot out", the core vote in the incumbent's constituencies also tends to increase, which serves to lower the national average but to no "useful" avail in terms of extra Parliamentary seats.

The third phenomena is the much-hyped "tactical voting" effect, which is harder to pin down. A study of the 1997 election (Impact of Tactical Voting in Recent British Elections. J.E. Curtis et al) suggests that 2% of the Liberal Democrat vote went to Labour and 1% went to the Conservatives - i.e. a 2-1 proportion. It is worth noting that the Liberal Democrats are polling 2-3% lower where they were in 2005 - this effect could be something of a wildcard in this election. Indeed, SNP/PC may play an important part in this. For a much more detailed and extensive discussion of these factors, I highly recommend reading UK Polling Report (http://ukpollingreport.co.uk/blog/) and Political Betting (http://www2.politicalbetting.com/) and in particular its articles on UNS (http://politicalbetting.com/index.php/archives/2010/03/07/andy-cooke-on-the-uns-part-2/ and http://politicalbetting.com/index.php/archives/2010/03/03/andy-cooke-on-the-uns-part-1/) and their seat model (http://politicalbetting.com/index.php/archives/2010/02/25/launching-the-andy-cooke-seats-calculator-final-version/)which takes these effects into account.

But back to the Populus/Times poll.

As a result of boundary changes, notionally, the Conservatives (based upon the 2005 election) will have 214 seats (vs. the 198 they won in 2005). That means that in order to secure an overall majority in the Commons they need to win an additional 112 seats (see http://ukpollingreport.co.uk/guide/conservative-target-seats). The 6.7% swing in this poll suggests that they will get 97 Labour seats, but that does not take into account the Liberal Democrat/SNP etc marginals in which the Tories can also expect to pick up some seats. Unfortunately, there has not been a poll of CON/LDEM constituencies since last summer (it showed a 5.5% swing to the Tories, which would indicate a Tory gain of 21 seats which would provide a clear majority), so until there is a new poll it is difficult to say just how many of these the Tories will pick up. However, even assuming that the Tories do not pick up any of the smaller-party seats (though they do look likely to take Perth & North Perthshire and Angus), in order to take the 15 LDEM seats required to reach the magic number of 326, they would need a swing of 4.25% from the Liberals. Including Perth & North Perthshire and Angus which look like a sure-win for the Tories, the swing required falls to just 3.5% which is not too different from the Liberal Democrat's national performance fall from 2005.

The point I'm trying to make is that the Tories do not actually need to perform that much better than they are currently in order to be sure of a working majority, and in fact, under the current poll results probably already notionally have one. This is in stark contrast to the market's pricing of a hung parliament in GBP and in Gilts.

Of course, all of the above misses probably the "key" issue: the election is as much a choice between Brown and Cameron as individuals as was the 2008 US Election between Obama and McCain, and the 1997 election one between Blair and Major. Faced with five more years of Brown - experience aside - when it comes to that choice at the ballot box, I think the Tories are on for a landslide. There is clear evidence that the Labour party is swinging to the Left, as evidenced by the rise and rise of Ed Balls and the placing of union backers in Labour safe seats (http://www.timesonline.co.uk/tol/news/politics/article7054578.ece). One of the most surprising outcomes of the financial crisis is just how little the political consensus shifted to the Left, globally (the US Healthcare issue is something completely different), and shows that the socialism vs. free market capitalism war was settled a long time ago. Labour looks like it's heading for the political nether-land.

UPDATE: The details of the Populus/Times poll are out (http://populuslimited.com/uploads/download_pdf-070310-The-Times-The-Times-Marginal-Seats-Poll---March-2010.pdf) and page 7 asks the question of LibDem, UKIP etc voters in constituencies where they have no chance of winning and it is clearly a two-horse race between Labour and the Conservatives whether they would vote for Labour or the Conservatives.

13% said they would vote Tory (unchanged from 2005), while 11% said they would vote Labour (down from 12% in 2005). So tactical voting looks as though it will help the Tories more than it will Labour.