Seat-by-seat wagering currently favours the Coalition in 11 Labor-held seats
If Labor and the Coalition really are at 50-50 nationally then Labor ought to be on track for about 75 seats or so, right? Indeed, if Labor is going to pull this election out of the fire, then what seats will it take from the Coalition? Where does Labor improve on 2010?
The usual answer is "Queensland marginals", maybe New South Wales too. But there isn't a lot of public polling behind this point of view. Nor do the betting markets see Labor picking up enough seats to win. In fact, seat-by-seat, the betting market reckons the probability that Labor wins is just seven in 10,000. Moreover, the betting currently favours the Coalition in 11 Labor-held seats.
For the last couple of election cycles, betting markets have added an interesting dimension to Australian elections. In addition to the "headline" national-level markets, the major online bookmakers also offer seat-by-seat wagering.
We don't have a lot of polling in the public domain from marginal seats at this stage, although we can expect more once the formal election campaign is underway. In the absence of a steady stream of polling information, might the betting markets convey some information about what is going on in marginal seats?
In the graph below I show the relationship between the probability of the ALP winning a given seat implied by the betting market prices (vertical axis) and the ALP's 2010 two-party preferred result (or "notional" result, taking into account electoral redistributions in Victoria and South Australia since the 2010 election). The implied probability of an ALP win is computed by averaging over two online bookmakers: Centrebet and Sportsbet.
For each bookmaker and for each seat, I convert the prices into implied probabilities (factoring out the bookies' profit margins, a simple computation), then average between the two bookies.
The inverted, "S-shaped" curve in the graph summarises the relationship between the two quantities. Clearly, bookies (and maybe punters too) are not fools, with the prices on offer tracking the 2010 (or notional) results reasonably closely.
Maroon (yes, really!) coloured points indicate the 30 Queensland seats. These Queensland data points tend to sit above the S-shaped curve, meaning that the betting markets are pricing Labor's chances as slightly better there than in other states. In this way the betting markets are merely echoing the conventional wisdom, that whatever the national swing is, it will be a more ALP-friendly swing in Rudd's home state.
I've labelled some specific seats. Brisbane is the most marginal Coalition seat in Queensland (closest to the vertical line at 50% ALP 2PP), and would fall to Labor with a 1.13% 2PP swing. It is close to a 50-50 proposition at the betting markets (50.3% at Centrebet, 47.8% at Sportsbet), so well short of being considered a Labor pickup. Knocking off Teresa Gambaro — the LNP incumbent in Brisbane — will be no small chore. Other Coalition marginals in Queensland (Forde, Longman, Herbert, Dawson and Flynn) are priced at much more competitive odds than similar seats elsewhere in the country, with the implied probabilities of Labor wins ranging from 38% to 43%; seats on similar margins elsewhere are priced at around a 20% chance of a Labor win.
The story in Lilley is especially compelling, the kind of seat that will fall to the Coalition when a big swing is on. Wayne Swan lost the seat in 1996, but won it back in 1998. Swan was in grave danger of losing Lilley before Rudd's return. It is one of the more delicious morsels of this election that Rudd's return substantially boosts the likelihood that Swan stays in national politics.
But the bigger story here is that the betting markets see barely any Labor pickups at the moment: maybe Brisbane, and with slightly more confidence, Labor is tipped to take Melbourne back from the Greens (eg, Labor's currently at $1.50 to the Greens $2.45 at Centrebet).
Indeed, in the bottom right quadrant of the graph we find 11 seats that are currently held by Labor, but are tipped to fall to the Coalition! Three are in Victoria, two in Tasmania and six in NSW, although the betting markets have slowly become a little more optimistic about Labor's marginal seat showing in NSW in recent weeks.
Notice too that the inverted-S shape curve in the graph intersects the vertical 50% TPP line at around the 35% probability level. That is, a (hypothetical) seat that split 50-50 last election is thought to have just a 35% chance of being won by Labor in 2013. This closely echoes the position of the national-level, "winning party" betting markets, which prices Labor as about a 30% chance of winning government.
Averaging over the two bookmakers in this analysis, Labor is ahead in just 62 seats, well short of the 76 seats needed for a majority in the House of Representatives. We can also induce a probability distribution over Labor's seat count in the next parliament, given the probabilities implied by the betting markets. Some 61 or 62 seats for Labor are the single most likely scenarios, given the current state of the betting markets, with only a seven in 10,000 chance that Labor wins 76 or more seats.
The results of this analysis may reflect nothing other than the seat-by-seat betting markets simply flying below the radar for now, and hence somewhat uninformative as to the actual election outcomes. There is no way that Labor could be polling as well as it is nationally but wind up with only 62 seats. Labor will do better than that, if its TPP vote winds up in the 49-50% range.
Nonetheless, it is clear that people who are putting their money at stake on the election aren't convinced. And fair enough. There is a lot we don't know. If Labor is going to pull this election out of the fire, then which Coalition seats is it likely to pick up? Marginal seat polling out of Queensland and NSW would be pretty valuable right about now. Rest assured the major parties have plenty of it, and what it is saying will be critical in the decision about the election timing.
