UK General Election Odds: Electoral geography makes a hung parliament a superb value bet

General Election Betting / Paul Krishnamurty / 12 December 2008 / Leave a Comment

Election betting expert Paul Krishnamurty tells us why betting on no party having an overall majority at the next election is the best political punt he's seen for years...

Having looked almost over as a betting contest just a few months ago, the next election is beginning to hot up. The Conservatives remain [1.42] favourites, but their lead in the polls has shrunk considerably. The most recent Comres/Independent poll had David Cameron's party just one point ahead, and the four point lead projected by Populus this week was their worst showing with that pollster for several months.

Everything points to yet more volatility in a range of political markets directly affected by each swing in public opinion. Besides the main battle to win the most seats, the 'Election Date' market can be expected to ebb and flow on the basis of rumours until Brown names the date. And we now also have the excellent 'Date and Most Seats' market, for which the Conservatives in 2010 or later is favourite at [2.2].

However, while these markets offer endless trading opportunities, success will rely on correctly predicting trends in UK public opinion over the next 18 months. Clearly the politics of recession are beginning to make an impact, and few could confidently predict exactly how each party's fortunes will fare as economic conditions change. Supporters of each of the three main parties could easily construct an argument to say their charges will benefit most from the recession, but there is no definitive answer.

Alternatively there is one market where the outcome remains comparatively unaffected by short-term issues, and which has a very strong chance of producing the same result no matter which party wins the most seats, so long as the result is not much more one-sided than current trends suggest. Ever since the 2005 General Election, I've been telling anyone who'll listen that the most likely result next time around is 'No Overall Majority', or in other words the first hung parliament since the 1970s.

Before getting to the analysis, a quick explanation and rundown of the electoral maths. To gain an overall majority, each party needs to win a minimum of 326 seats. Note this is different to winning the 'Most Seats', which could plausibly be achieved with 40 fewer. Due to boundary changes favouring the Conservatives, we're starting from a slightly different position to the 2005 result. Academic analysis estimates the Conservatives already have a 'notional' seat-total of 211, compared to Labour's 345, and the Lib Dems 63. (These include by-election results in this Parliament).

The problem, or opportunity, for traders in the 'Overall Majority' market is that its driven by fundamentally flawed criteria; media analysis of national opinion polls. Such polls can tell us a great deal about the big picture, but they are often misused to make seat-total projections that assume these swings in opinion are evenly distributed across regions. When we hear for instance, that the Tories have a 10 percent lead, pundits reach for statistical models that tell them how many seats that will translate into, assuming a uniform national swing of 10%.

However, trends over the last few decades show that UK politics is no longer just a two-party affair, and that there are significant political differences across regions. So when Labour are unpopular in Scotland, the likeliest beneficiaries are the SNP rather than the Tories. Similarly, the Tories can draw little recent encouragement from results in Wales, or the North of England, where they must compete for the anti-Labour vote with Plaid Cymru and the Lib Dems. So a perfectly plausible scenario might involve the Tories enjoying a national swing of 10%, consisting of 20%+ swings across the Southern marginals but little improvement in the North, Scotland or Wales.

And make no mistake, in order to reach their target of 115 extra seats, they will need to win beyond the Southern and Midlands 'hyper-marginals'. Comprehensively winning all 60 or 70 of these would be enough to become the largest party, but in order to win a majority, Cameron will need to break through in parts of the UK which the Conservative Party hasn't reached in generations. Even in 2005 when advancing their nationwide share of the vote and seat total, Tory support in the North of England actually declined.

I've studied each of the 200-odd seats that will decide the next election, and drawn the following conclusions. Firstly, bounce or no bounce, Labour's majority will disappear. Whatever reservations the public may have about David Cameron, he has made the Tories more electable than any of his three predecessors as opposition leader. They will, without doubt, improve significantly on their 2005 performance. Labour cannot afford to lose more than 20 seats, yet must defend 40 with majorities of less than 2,000.

There are several other seats where I expect the Tories to overhaul vulnerable Labour majorities. If there was an election tomorrow, assuming for example a 7 point Tory lead, I'm confident they would win 60 of the 99 constituencies where they trail by less than 5,000 according to the notional figures, and that are in potentially fertile areas of the country. This should be enough to make them the largest party.

This still leaves them with 55 short of a majority. Its also safe to assume that Cameron will enjoy a net gain over the Lib Dems, though again the effect will be limited. Taking an extremely one-sided assessment of marginals concerning these two, it is possible to see as much as a 25-seat swing to the Tories.

However, I expect 15 is more realistic, and that could still under-estimate the Lib Dems. Historically, their vote share tends to go up during election campaigns, as these are the only times they are given a comparable amount of media coverage. Furthermore, they are notorious as a local campaigning force when defending areas, relying on a strong activist base.

These localised factors usually fly under the radar of headline predictions; on the 2005 election night for instance, the BBC initially predicted them to total less than 50 seats, yet they actually got 62. In the current Baxter electoral calculus (which actually predicts a hung parliament), the Lib Dems are predicted to get 23 seats. In my view, you can double that, and it would still represent a poor result.

Anyway, assuming they do get 15 off the Lib Dems, Cameron still needs to snatch 40 Labour targets and its hard to find any that look easy pickings. Consider the following seats, for example, all of which are well within the top-100 Tory targets from Labour; Halifax, Dewsbury, Tynemouth, Bradford West, Cleethorpes, Keighley, Poplar and Limehouse, Stirling, Morecambe and Lunesdale. If I were to price these individual seats up right now, I wouldn't make the Tories shorter than [1.8] in any of them. The same applies to the next 50 odd seats that could compensate failures in these areas.

So in short, backing a Tory majority at [1.85] equates to backing them to win 40+ seats where they have roughly an even chance at best. Backing a Labour majority involves an unpopular government, in the middle of a recession, losing less than 20 seats to a rejuvenated opposition. Both are possible, but both are very tough asks. Whereas 'No Overall Majority' requires things to stay very broadly where they are right now, and allows for either party to take the lead as long as they don't completely run away with it. At over [3.0], I doubt we'll see a better political bet in the next couple of years. In fact, I'd make it odds-on.