I notice that Labor has firmed quite a bit at Centrebet, coming in from 3.30 to 2.55 in the past week or so. This is good news for Labor supporters, but not, I think, such good news for the view that betting markets outpredict polls and pundits. Throughout the election, the polls and pundits, while leaning to a Liberal victory have generally presented the view that this is going to be a close race. Until now, the betting markets have treated a Liberal win as a foregone conclusion.
I’d argue that the most efficient test of claims about the informational efficiency of betting markets for political events would be to test the prediction that movements in polls should follow movements in betting odds (since we know that election-day polls give reasonably good predictions of actual outcomes). For this election at least, the opposite seems to be true – the betting markets have moved into line with the polls.
John Quiggin 
POLLS, PUNDITS AND PUNTERS
The polls have mostly showed the ALP to be infront over the past year. Centrebet correctly predicted the campaign season swing to the LN/P.
I think that the $3.30 figure might have represented a little post-Jakarta bubble for the LN/P. Centrebet has shown the LN/P as moderate favourites for most of the year. In the CoW Trifecta I reported (Sept 6 2004) I reported Centrebet showing
I will stick by my prediction, which dates back to mid-2003, that the CoW parties will all win their respective elections due to, low-interest rate financed, house-price/consumption booms.
This will be in spite, not because, of the relatively poor results in Iraq. I do not think Iraq will be a big factor in the election since voters now perceive all ME nations as basket cases.
The AUS voters will (correctly) not punish Howard on Iraq as they will instinctively know what the pundits have missed: the ADF in Iraq was payback for the USMC in Timor. The AUS-US alliance has been strenghthened by the Iraq venture. So far, jihadists have not attacked AUS on account of Iraq.
If Kerry were a more effective candidate Bush might have been more vulnerable on IRaq. But I dont think US voters want to be reminded that they were all conned into this embarassing venture.
The suggestion that betting is a good guide is probably based on US elections.The US would have much larger betting pools than we have here with less chance of distortion from big bets.On Tuesday of this week it was reported that a bet of $200,000 was placed on the coalition to win at Centrebet.Labor’s price went from $2.70 to $3.30 in one go because the total pool was only around 600,000 at the time.If that bet had not been made the market would look a whole lot different now.In my opinion that bet was placed to keep the Coalition price up where they think it needs to be.Of course there has been no shortage of pundits pointing to the betting market as a guide which I think has been one of the big cop outs of the whole campaign.Who ever are favourites on Friday night will most probably win the election but I believe the betting markets will look nothing like they do now.
Arrgh! The problem with the betting market as a reliable predictor is that bookies have a vested interest in manipulating the odds to ensure they stay in business. Bookies love favourites and odds on bets. This bias will prejudice any informational efficiency. I’m biased of course, I’ve never won a bet on the Melbourne Cup yet!
The prices for Howard and Latham victories are about the same as the prices for Bush/Kerry victories ($1-45, $1-65) respectively. But it seems to me that the Australian polls are closer. Labor looks like the value bet to me given that some recent polling gives it to Labor (e.g. SMH yesterday on marginal seats).
I think a lot of Liberal supporters are angry with Howard for not just cutting taxes further rather than providing handouts. Some of these people might vote Labor if they feel ambivalent on other issues (e.g. Iraq) or just want a change to a younger leader.
I will vote Liberal this week but have $100 on Labor to win. I think its a reasonable hedge.
Sorry $2-65 on Kerry not $1-65.
Another point to remember about these elections is that electoral boundaries favour the Right wing parties. The ALP won the 1998, and the GOP won the 2000, elections on the popular vote. So the Left parties need to do better than 51-49 distribution of votes in order to win.
Given the invariablly tightly marketed struggle for median voters, pork-barelling advantages of incumbency and absence of recession one would have to favour the Right wing parties to win on the day.
Maybe the betting market simply reflects the fact that gamblers are more likely to be greedy, self centered and affluent, with more money than brains – exactly the type that would prefer Howard’s style of politics.
“Maybe the betting market simply reflects the fact that gamblers are more likely to be greedy, self centered and affluent, with more money than brains – exactly the type that would prefer Howard’s style of politics”.
John S. What a noble way of describing half the Australian population. And wrong too — gamblers according to Productivity Commission and other reports are overwhelming blue collar and therefore more likely to vote Labor. That the workers love to gamble is the reason economists see reliance on gambling taxes as a very regressive form of taxation.
Don’t put me in the Austrian camp (please), but do we know enough of what’s happening in the marginals to question the betting? After all, a couple of big bets from party pollsters may be skewing the market towards the (gulp) truth …
Well the marginal betting has been going towards Labor for a week.So on Tuesday Labor’s price tightened up in the marginals at Centrebet but blew out on the main result betting.
Hurrah,jack stroppy knows something!
However,despite claims that the bookies generally get it right,they got the AFL final wrong.
Please remember the betting market in the 1999 victorian state election. LNP $1.02 ALP $10 on the friday before the election. The bookies have been known to take a bath before.
Looking at Centrebet just now, I notice that the price on Labor has gone out again to $3. A reaction to the latest AC Nielsen poll?
dear harry, do the “Productivity Commission & other reports” also show that blue collar workers are NOT “greedy, self centered and affluent, with more money than brains”?
marklatham at October 4, 2004 10:15 AM likes to blow rasberries from the safety of a nom de cyber.
If he wants to be a real smartie he might lay a bet on the ALP to win. I will give him Centrebet odds as we comment. I am sure that Pr Q wouldnt mind acting as the warrant holder for a little psephological flutter. Come on Markie, put some money where your big fat mouth is.
But I am guessing Markie is all hat and no cattle.
Very good, I’m impressed with all this, will go online and read in detail sometime.
and harry… where’s the evidence that blue collar workers vote labour – don’t they make up the bulk of aspirationals in the mortgage belt?
Hey,jack stroppy scores again!
Since last friday I have considered placing a bet on labor,however one thing constrains me….
I have never placed a bet in my life!!!!
My son,of thai heritage,unfortunately makes up for me.
He bets on AFL football at the TAB compulsively.
Luckily he works part time at the casino and can not gamble there anymore,you may have surmised by now that gambling is not my favourite pastime.
My ex wife,a thai,contributed copious amounts of money whilst I was engaged in building the hotel and other areas at the perth casino.
I just find it interesting that the bookies can get it wrong,which elections have they got wrong-kennet,goss?
Harry I think you’ll find that while the PC profile of gamblers you cite is an accurate picture of those who gamble on lotto, pokies etc, in my experience it is most certainly not true of internet gamblers.
Furthermore, a small number of gamblers who more closely fit John Scrivener’s description tend to skew the odds by placing large bets ($200000 at $1.30 is not a shrewd bet in my view).
I have been watching these Centrebet prices with interest since John pointed them out to us. Wednesday at 6-30pm Labour was paying $4-00 for a $1 bet and the Coalition $1-20. These are the longest odds for Labor since the campaign started.
This was after Howard’s speech on logging in Tasmania and after Latham’s press luncheon. I stated a few days ago that I thought Labor at
$2-65 seemed a good value bet but have steadily watched the odds on Labor winning lengthen. If they were a bargain at $2-65 they should be a real bargain at $4-00 but, perhaps judging quality by price, I am now getting cold feet.
Still the election seems closer to me than these prices suggest.
Fear, loathing #19
OK, we’ve now turned. We’re in the home straight. The ALP’s odds are shortening, but no one reallly has a clue who’s going to win. One way I can take a gauge is by comparing my ’04 writing with ’01,…