Oz Politics Blog

While the betting market stills sees the possibility of a 27 October 2007 election, the last Saturday of November or the first Saturday of December are firming as the favourites.

I remain convinced that the 24 November to 8 December window is the most likely timing.

A 27 October 2007 election would need to be called by next Monday. I don’t think that will happen.

The betting market has recovered a little for the Coalition since Sunday. Since then, the average probability of a Coalition win from the five bookmakers I am tracking moved from 26.7 per cent to 33.0 per cent. What is harder to explain is why. Was it Tuesday’s Newspoll? Was it yesterday’s tax gaffe? Or is there another factor at play?

The other graphs are here. You may need to hit the refresh or reload button on your browser to see the latest graphs.

If you don’t want to risk any of your well earned, you can place a tip on the 2007 election here

Update from my inbox:

Hi Bryan,

It was almost certainly the Newspoll. The movements (largely? totally?) happened before the tax gaffe. Not the most rational responses, but anyway…

EconoMan

The latest Newspoll is out. Today’s poll has been seen by a number of commentators as a reprieve for the Coalition, and the previous Newspoll has been cast as aberrant.

I am not convinced.

In respect of being aberrant, it was not just Newspoll that headed south for the Coalition. All of the major polls did. The weekly Morgan dropped 5.5 points on 2 September to 40 per cent, but only rose one point on 9 September 20 41 per cent. The fortnightly Newspoll dropped four points to 41 per cent on 2 September. The monthly ACNielsen dropped two points on 8 September to 43 per cent. I accept that all of these movements could have been random noise — let’s be honest, there was not a clear explanatory factor for the observed poll movement in early September. Nonetheless, that all three pollsters went south should have been worrying for the Coalition.

Is the latest result a reprieve? Well it is much better news for the Coalition than 41 per cent. But I were them, I would be worried about the 45 per cent ceiling that has been in play since May. Should 55-45 be repeated at the ballot box it would be a huge Labor win.

If you compare the 2007 trajectories with 2004 and 2001, Labor is riding significantly higher in the saddle. Given that we are 90 days out from an election at most, the Coalition would want to be hitting 47 and 48 as often as it falls below 45. Indeed, as we are approaching the short game of electoral golf, the odd 50 would not go astray at the moment.

The usual opinion poll graphs are here. You may need to hit the refresh or reload button on your browser to see the latest graphs.

PS: Antony Green’s Federal election web site should be up today on Friday.

The poor polls for the Coalition are showing up in the betting market. Last Sunday, the average from the five bookmakers I am tracking had the Coalition on a probability of 34.0 per cent of winning the next election. Today, that probability is down to 26.7 per cent.

The other graphs are here. You may need to hit the refresh or reload button on your browser to see the latest graphs.

If you don’t want to risk any of your well earned, you can place a tip on the 2007 election here

I have been using a weighted monthly aggregation for some time now to peer behind the random lumpiness in week to week polling.

The regular Reuters poll trend suggested another methodology to get beyond the random lumpiness. In a manner similar to the Reuters poll trend, every time a poll is published (typically weekly) I take the simple average of the last poll from each pollster (Galaxy, Morgan, Newspoll and ACNielsen). I do not weight the average for differences in sample size. To that rolling average I then apply a 13 term Henderson moving average to smooth out the lumps and bumps. I chose a 13 period average because the results from the 5 period moving average were pretty lumpy.

The next set of graphs depict the underlying trends, as revealed by the 13 term Henderson moving average.

These charts tell pretty much the same story as the monthly aggregation. Rudd’s election as Leader reinvigorated Labor in the polls. Nonetheless, between March and July this year, the Coalition was beginning to claw back its position. In the last month, that clawback has reversed. The graphs also point to a decline in the Green and minor party vote since Rudd’s elevation; however, that decline has now plateaued.