Saturday, 1 November 2014

The Conservative battleground in November 2014

Yesterday I looked at the constituency betting markets as they currently stand from a Labour perspective.  Today I'm looking at the Conservatives.  So here's the table of all the seats where the Conservatives' chances are ranked by the bookies between 1/16 and 16/1, ranked in ascending order of implied probability:

As I explained yesterday, such a table assumes that in aggregate the constituency odds are our best estimate of what's going on, while noting that there may be individual anomalies (otherwise known as betting opportunities).  The constituency odds factor those matters in to the best judgement of bookies and punters. By arranging constituencies by order of odds rather than majority, we can see how many seats gamblers expect the parties to take - or what the odds are that each party will take a given number of seats.

The Conservatives are currently expected to lose seats if the constituency markets are to be believed.  The Conservatives are priced at 5/6 in Warwick & Leamington.  5/6 is the price that bookies often use for a 50/50 chance, allowing them to build in a margin for profit.  If the Conservatives take the seats up to and including Warwick & Leamington and no more, they will suffer a net loss of 45 seats, leaving them with 262 seats (I'm assuming for now that Rochester & Strood will be worse than a 5/6 shot for the Conservatives at the next election, which seems a safe enough assumption today).  With Labour's 5/6 shot giving them 307, the two main parties are currently assumed by the markets to have an almost exact reversal in seat counts.

How do the constituency prices compare with the overall majority and most seats markets?  If the Conservatives held all their 2010 seats and took every other seat that is shorter priced, they would have 343 seats (I shall assume for present purposes that Rochester & Strood will be priced at 3/1 from a Conservative perspective - it's likely to be shorter than Clacton but not by much).  To get an overall majority, the 326th seat (as ranked by odds) would be Waveney at 10/3.  That compares with 5/1 that you can get with Betfair that the Conservatives will get an overall majority and 4/1 that is available with Ladbrokes and Coral, among others.  When you also factor in the danger that you might select a constituency that bucked the trend, all the value if you think that the Conservatives might win an overall majority is simply to take that overall majority bet rather than go into the constituency markets.

How about the most seats market?  A lot depends how many seats you think would be needed for that tally.  If the Conservatives took Carshalton & Wallington and all the seats with shorter prices than that seat, they would have 300 seats.  Carshalton & Wallington is a 7/4 shot.  Everything up to and including Nuneaton would give them 290, and that's a 13/8 shot.  City of Chester is the 280th seat, and that's a 5/4 shot.  Meanwhile, you can get 6/5 with Paddy Power among others on the Conservatives having most seats and 5/4 on Betfair.

If Labour and the Conservatives perform as the markets currently expect, Nuneaton is the hinge seat, being the 290th in order of implied probability for both of them.  That 13/8 looks good as compared with 5/4.  If you think that 290 is indeed going to be the critical level, you probably could not do better than choosing Nuneaton or Bury North as your bet, since both are about as typical Labour/Conservative marginals as we get in these increasingly fragmented days.

But let's think a bit more about the relationship between the general markets and the constituency markets.  We saw yesterday that for Labour we would do better backing them at the overall majority market or the most seats market than by backing them in constituencies.  We see today that we would do better backing the Conservatives at the overall majority market rather than by backing them in constituencies.  Why is this happening on both sides of the fence?

The answer, I think, is because the contingencies between constituencies are thought to be becoming less correlated.  We don't actually know whether that is true.  It seems to me possible that both Labour and Conservative overall majority bets may offer better value than is presently appreciated.  Of course, value bets are often losing bets, and as a longterm backer of No Overall Majority, I'm not going to change my tune now when both major parties look in pretty ropey shape with six months to go before a general election.  But the point is worth noting.

There is a further point specific to the Conservatives.  We have seen that the constituency markets offer better value than the most seats market and worse value than the overall majority market.  This means that we should be very wary indeed of backing the Conservatives in any seat priced between the 5/4 of the Betfair most seats market and the 5/1 of the Betfair overall majority market, particularly at the longer end of that spectrum, since such seats seem much more likely to offer bad value, given the simpler bet that one can make on an overall majority.

I last looked at the Conservative position in June:

The 290th seat then was priced at 6/4, so slightly shorter.  The 326th seat (still Waveney then) was priced at 3/1 - again, slightly shorter.  So there has been a slight lengthening of Conservative odds, but nothing stark.

Are the constituencies becoming less correlated, and if so, why?  I think that they are, despite what I said earlier.  In 2010, the Conservatives had two sets of battles: with Labour and with the Lib Dems.  The dynamics of each were different.  In 2015, the Conservatives will add a third front with UKIP.  Meanwhile, in Scotland, the SNP are currently posing far more of a threat to Labour than they did in 2010, giving Labour a third front also.  These two new third fronts are only loosely related to the older fronts, and the table above caters much less well for those loose correlations than the differential correlations that previously operated.

I looked yesterday at Labour's exposed Caledonian flank and dealt with it effectively by taking the Scottish seats out of the equation and making a more general assumption about them.  That type of approach will not work with the Conservatives and UKIP.  First, UKIP pose a threat to the Conservatives across a lot more seats, albeit at a lower level and in all probability taking fewer seats.  Secondly, the impact of UKIP on Conservative chances is more indirect in most seats.  Waveney is a good example - the Conservatives have drifted slightly in Waveney because, with UKIP strengthening, what was already a difficult seat for the Conservatives has become in the punters' eyes still harder.  Labour have not seen their price shorten - the opposite, in fact.  But the arrival of UKIP has meant that both Labour and the Conservatives have seen their prices lengthen.

Anyway, do I have any other constituency tips?  Well, as it happens, I do.  To get an overall majority, Labour would need to take seats like Vale of Glamorgan, Rossendale & Darwen and Crewe & Nantwich.  Labour are priced at 5/2 to get an overall majority.  If, like me, you believe that to be rather too short, you can lay it by proxy by backing the Conservatives to take such straight Labour/Conservative marginal seats at prices like 8/15.  I like backing 2/5 propositions at 8/15 and I like backing 1/3 or greater propositions at 8/15 even better.

It is possible that Labour could take these seats and not get an overall majority (one set of circumstances might be that Labour lose a slew of seats to the SNP but perform well south of the border).  My take is that if Labour are performing that badly in Scotland, the chances of them doing well enough to get what would otherwise be an overall majority winning performance in England and Wales in standard Labour/Conservative marginals are low enough to make these bets still stand as decent proxies for that bet.  Make your own mind up, of course.

I will deal with the Conservatives' prospects against the Lib Dems tomorrow, and the impact of UKIP on Monday.

No comments: