All Over by Eight

University of Newcastle economist Steve Easton emails with a tidbit from election night.

I thought you might be interested in the data presented in the attached spreadsheet. This is the first time to my knowledge that there has been active betting in [Australian] election markets immediately preceding and during the vote count. Betfair had a market for the NSW State Election but trading was virtually non-existent.

The attached data were captured for the Betfair betting market at 15 minute intervals starting at 5.00 ESDT. In short at 5.00 pm the probability of a Labor win was 75%, at 6.00 pm 77% (after some discussion of exit polling results on the various evening news services), at 6.15 pm 82% (after, amongst other things, Stephen Smith had confidently predicted a 20-seat gain to the ALP on the ABC telecast), at 7.45 pm 91%, and pretty much all over at 8.00 pm at 95%.

So the betting market “called the result” by about 8.00 pm.

Labor Bid
Labor Lay
Coalition Bid
Coalition Lay
  Total Amount Bet
Amount Bet During last 15 minutes
5.00 pm 1.3 1.34 3.8 4.2 0.75 $3,634,027  
5.15 pm 1.32 1.36 3.5 4.2 0.73 $3,646,359 $12,332
5.30 pm 1.27 1.34 3.6 4.2 0.74 $3,652,628 $6,269
5.45 pm 1.19 1.24 3.8 5.7 0.76 $3,670,071 $17,443
6.00 pm 1.25 1.28 4.3 6 0.77 $3,678,409 $8,338
6.15 pm 1.17 1.2 5.4 6 0.82 $3,716,876 $38,467
6.30 pm 1.18 1.22 5.4 6 0.82 $3,735,903 $19,027
6.45 pm 1.18 1.21 5.4 6 0.82 $3,763,468 $27,565
7.00 pm 1.2 1.21 6 7.8 0.83 $3,802,359 $38,891
7.15 pm 1.14 1.19 6.2 7.8 0.84 $3,812,776 $10,417
7.30 pm 1.13 1.15 7 9.4 0.86 $3,843,488 $30,712
7.45 pm 1.03 1.04 10 34 0.91 $3,868,311 $24,823
8.00 pm 1.03 1.04 18   0.95 $3,871,488 $3,177
8.15 pm 1.03 1.04 17   0.94 $3,893,539 $22,051
8.30 pm 1.02 1.03 23   0.96 $3,898,537 $4,998
8.45 pm   1.01 55 500   $3,917,569 $19,032
9.00 pm 1.01 1.02 55 300   $3,925,196 $7,627
9.15 pm   1.02 60 300   $3,929,166 $3,970


This entry was posted in Australian Politics. Bookmark the permalink.

12 Responses to All Over by Eight

  1. al loomis says:

    i am reminded that the ‘parties’ contending for control of the byzantine empire at one point were centered around the chariot racing. horses and politics are a mix ozzies are right at home with, indeed, can hardly distinguish.

    then there’s the very similar attitude to war and rugby..

  2. Dale Bailey says:

    What about the betting in Bennelong? I only looked longitudinally at one agency over the campaign, but John Howard was always in front of Maxine. The gap narrowed towards election day, for certain. I know it’s not over yet, but the pre-election odds don’t seem to reflect the outcome (close). What is different about this betting to the national odds?
    PS. Would have been nice to have some money on Maxine at the early odds of $4.20!

  3. Steve Easton says:

    There was little trading for the Bennelong seat on the Betfair market after the polls opened. You could have got a trivial 1.03 on Labor for this seat last night but I think its all over now.
    Best wishes,

  4. EconoMan says:

    Dale, I got $200 on @ $4.75, about a week after she announced. Can’t see him coming back on those postals. Needs a swing to him from 57% to over 60% of postal votes.

    All in all, I had a pretty damn good night: On the betting, on the desired result, and out partying.

  5. John Quiggin says:

    This election has been pretty strong evidence for the semi-strong efficient markets hypothesis and exceptionally strong evidence against the strong-form hypothesis, as it applies to betting markets.

    It’s clear (to me at least) that the betting markets didn’t know any more than the average intelligent person, unbiased by wishful thinking, who followed the polls and the news. When the polls shifted a year ago, the markets, like most people, saw it as a flash in the pan. It took until the opening of the campaign for the markets to catch up with the poll.s

  6. Steve Easton says:

    I agree fully, with the proviso that of course proving the strong-form inefficiency is problematic.

  7. Philg says:

    $3 was available on Betfair for Maxine McKew to win Bennelong at 5.30 when the exit polling started coming out. I loaded up.

  8. Dale Bailey says:

    This was what I was getting at, PhilG – the betting market was not on top of this one! Why so different to the overall outcome, where the betting markets performed so well? I guess the point about the predicting powers of the betting markets, in answer to John Quiggin, is that it is not just one intelligent person’s unbiased assessment of the signs out there, but a measurable, dynamic, quantitative assessment of many individuals. That to my mind gives it an objectivity that an individual would struggle to have. I’ll apologise now and admit that I know nothing about ‘semi-strong efficient’ and ‘strong-form’ market hypotheses.

  9. christine says:

    But it was John Howard in Bennelong. It wasn’t just any old person in there, it was a political mastermind and sitting Prime Minister. How could he lose?? One of the TV pundits at some point said they’d have called Bennelong for the ALP happily if it were just the numbers, but just couldn’t do it given it was Howard in there. Sentiment still matters, I think.

    Do Bennelong postal votes typically favour Libs more than non-postal votes do?

  10. Pingback: Polls, pundits and punters at John Quiggin

  11. The Doctor says:

    With Bennelong in particular, there was a running thought over at Poll Bludger, that some Liberal supporters may have been gaming the odds – in order to support the idea that Howard would not lose his seat.

  12. Andrew Leigh says:

    “there was a running thought over at Poll Bludger, that some Liberal supporters may have been gaming the odds”

    I’ve asked a couple of bookies about this, and they didn’t see evidence of it. As a practical matter, it’s quite expensive to keep the odds out of whack in a fairly thick market (as Bennelong was).

Comments are closed.