In late November I had a look at how the constituency markets were assessing the parties' chances in aggregate - not just by looking at how the favourites stacked up in aggregate but also at how many seats in aggregate were in practice in play:
I thought it was about time, on the eve of the campaign, to look at this again. Here are the aggregates of the best prices favourites on the seat markets:
Conservatives 275
Labour 278
Lib Dems 31
SNP 40
UKIP 3
Plaid Cymru 3
Greens 1
Respect 1
DUP 9
SDLP 3
Sinn Fein 5
Lady Sylvia Hermon 1
We now have Northern Irish seats, so we have a full house. I have treated the Speaker's seat as being in the Conservative column, though different "most seats" markets treat his seat differently, so check this.
Since the end of November, the SNP have risen dramatically in the best-guess seat tally. Correspondingly, Labour have dropped substantially. Interestingly, the polls haven't actually changed much since December. Sentiment has changed. It seems to be a dawning realisation that the Scottish polls reflect an underlying reality.
Otherwise, the aggregate of the favourites has stayed very stable. That is also interesting, because the constituency markets assume a level of recovery by the Conservatives and the Lib Dems which doesn't so far seem to be happening.
As you can see, both main parties are thought to be quite a long way from an overall majority.
Just as importantly, Labour are thought to be slightly ahead of the Conservatives in the battle for most seats. Even if you take the view, as I do, that the seat markets still underestimate the SNP's chances in Scotland, the two main parties are neck and neck. You'd never guess that from the "Most Seats" market, where Labour can be backed on Betfair as of this morning at 2/1. This price is crazy and backing Labour heavily on the Most Seats market should be an essential part of any portfolio of political bets.
Just as importantly, Labour are thought to be slightly ahead of the Conservatives in the battle for most seats. Even if you take the view, as I do, that the seat markets still underestimate the SNP's chances in Scotland, the two main parties are neck and neck. You'd never guess that from the "Most Seats" market, where Labour can be backed on Betfair as of this morning at 2/1. This price is crazy and backing Labour heavily on the Most Seats market should be an essential part of any portfolio of political bets.
How stable is that projected outcome?
As before, I've looked at how strongly fancied the favourite is. And as before, I've worked on the basis that if there is a second favourite at or below 2/1, this is a true marginal. I've presented the results in a tabular format, with the parties' range shown below:
And again we see a substantial change from late November in both the Labour and the SNP columns. The Conservative band has edged up slightly, but the Labour band has shifted down roughly 20 seats. Meanwhile, the SNP band has moved forward by roughly 15 seats.
The consequence is that even at the outer edges of this measure of the parties' safe seats and marginals, neither party will get to an overall majority. Neither party looks very likely to get to 300 seats - everything within this bound of probability would have to come right for the Conservatives for that to happen for them and most things within this bound of probability would have to come right for Labour for that to happen for them.
For those that are interested, here are the Northern Irish seats presented on this basis:
These will not affect the outcome of the main battle, of course.
Taking a broader view of what makes a marginal
Let's push the boundaries further.
For those that are interested, here are the Northern Irish seats presented on this basis:
|
|
DUP
|
SDLP
|
Sinn Fein
|
UUP
|
Independent
|
|
DUP
|
8
|
1
|
0
|
1
|
0
|
|
SDLP
|
1
|
2
|
0
|
0
|
0
|
|
Sinn Fein
|
0
|
0
|
4
|
1
|
0
|
|
Independent
|
0
|
0
|
0
|
0
|
1
|
|
UUP
|
1
|
0
|
1
|
0
|
0
|
|
Total
|
8-10
|
2-3
|
4-5
|
0-2
|
1
|
These will not affect the outcome of the main battle, of course.
Taking a broader view of what makes a marginal
Let's push the boundaries further.
As previously, I've also had a look at the position if we use the test of marginality as a seat having a second favourite that is priced at 5/1 or less. This table is complicated by the presence of 13 three way marginals (hence the extra row).
(The independent is Claire Wright in East Devon, who is a localist environmental type, who has apparently been quite fancied at the bookies.)
The ranges of outcomes are now much wider than when I used a 2/1 cut-off for assessing marginals, but these wider ranges barely differ from late November. The betting public aren't that sure of the SNP's chances.
This stability of range is itself noteworthy. These markets aren't shifting very much. I haven't changed my views on the question of how correlated the results in different constituencies will be. In my opinion, the answer is different for different parties. The SNP will see very correlated results, since the election in Scotland is either going to be all about their turf (as seems likely right now) or Labour will have reframed it to some extent on UK-wide terms. At the other extreme, Labour and the Conservatives are fighting multiple different battles and so the correlation will be much less. To get a fuller version of my views on this subject, I went into this in much more detail last time, so have a look at the link at the top of the page if you want.
Again,
for those that are interested, here are the Northern Irish seats presented on
this basis:
https://docs.google.com/document/d/1EVhvEvf3wZDQmJbF55wNavDpAHUMRn0n6Jf5GGyPSwA/edit?usp=sharingIt is a more stable position than the rest of the UK, even at this level of marginality.
If the seat markets are to be believed, there is only a slim chance of either main party achieving an overall majority. Even on the most generous interpretation of what makes a marginal, each main party is going to have to get close to taking everything thought to be in play to get there.
Incidentally, it would be nice to be able to produce a neat graph which shows the exact implied probabilities of a hung Parliament, but the degree to which the results in individual constituencies will be correlated is uncertain, so it will be a matter of guesswork rather than science.
Are the seat markets to be believed? Good question. Two considerations push me in opposite directions.
On the one hand, the two main parties have been regularly polling in the low 30s in the national polls. On those national vote shares the winners in many seats will do so on relatively low local vote shares, meaning that the winners may be considerably less predictable at an individual level than the constituency markets suggest. (The corollary to that is that some of the odds against shots may well offer good value.) Set against that, the unpredictability is quite likely to cancel out to a considerable extent at an aggregate level.
So on balance I probably do believe the overall story of the seat markets. That leads me to three conclusions:
1) Don't bet against a hung Parliament unless you get long odds.
2) Labour most seats is a terrific value bet at 2/1.
3) Since the Conservatives look unlikely to get above 300 seats (I'd guess at least 3/1), the next Government is very likely to be a Labour-led government. Lay David Cameron as Prime Minister after the next election at anything close to, or even not very close to, current prices. He's odds on at present to keep his job. This is a crazy price. Enjoy.
Incidentally, it would be nice to be able to produce a neat graph which shows the exact implied probabilities of a hung Parliament, but the degree to which the results in individual constituencies will be correlated is uncertain, so it will be a matter of guesswork rather than science.
Are the seat markets to be believed? Good question. Two considerations push me in opposite directions.
On the one hand, the two main parties have been regularly polling in the low 30s in the national polls. On those national vote shares the winners in many seats will do so on relatively low local vote shares, meaning that the winners may be considerably less predictable at an individual level than the constituency markets suggest. (The corollary to that is that some of the odds against shots may well offer good value.) Set against that, the unpredictability is quite likely to cancel out to a considerable extent at an aggregate level.
So on balance I probably do believe the overall story of the seat markets. That leads me to three conclusions:
1) Don't bet against a hung Parliament unless you get long odds.
2) Labour most seats is a terrific value bet at 2/1.
3) Since the Conservatives look unlikely to get above 300 seats (I'd guess at least 3/1), the next Government is very likely to be a Labour-led government. Lay David Cameron as Prime Minister after the next election at anything close to, or even not very close to, current prices. He's odds on at present to keep his job. This is a crazy price. Enjoy.
No comments:
Post a Comment