Monday, November 4, 2024

Update: opinion polls and betting market

Before I begin, let me tell you a sad story. I went overseas for an extended period, and while I was away, the code I use to daily scrape the SportsBet website stopped working. So, I have a roughly six-week gap in the betting market data. In this time, the odds of a Labor victory have lengthened (with Labor still the favourite to form government after the next election). The betting market has the outcome of the next election pretty close to even money (reflecting the polls below). What has changed is the odds for any other result (which I ignore in the following charts). Those odds have shorted from \$101 to \$81. 



Turning to the opinion polls, we now have an extended period of some four or so months where the projected two-party preferred outcome is very close to 50-50 - the proverbial toss-up. 



At this point, I should note the way in which I model house effects has changed. For those pollsters that have changed their polling methodology (Essential and Resolve), I exclude the polls conducted under the earlier methodology from the requirement that all house effects sum to zero across the pollsters. I also exclude from the sum-to-zero constraint those pollsters that have fewer than five polls published over this period. For the record, the pollsters included in the sum to zero constraint above were: 'Newspoll', 'Resolve Strategic 2', 'Freshwater Strategy', 'Essential 2', 'Roy Morgan', 'YouGov', and 'RedBridge Group'. The estimated house effects for each pollster are as follows (note: these two charts are different ways of reporting exactly the same data).



Turning to the primary votes, some trends are clear. Labor's primary vote is down on the previous election and the Coalition's primary vote is up. The vote share for independents and other parties is down on the previous election. The vote share for the Greens is above where it was at the previous election.  I suspect this reflects some movement from the "Teals" back to the Coalition, and some movement of votes from Labor to the Greens. On these numbers, you would expect whichever major party forms government, they will rely on the support of others for the passage of legislation (yes, the polls suggest a hung parliament).







And for completeness, the attitudinal polling follows.




As usual, if you want to see how the charts were made: click here.

The data for these charts was sourced from Wikipedia.

Monday, August 26, 2024

Federal polls - quick update

The NT election result (convincing Coalition win) has not changed the betting odds at SportsBet for the next Federal election.






Monday, August 19, 2024

August polling update

The betting market remains largely unchanged this year. Labor remains the favourite to form government after the next election. Based on the latest polling, most pundits believe that will be a minority government, with the support of the cross-bench independents and/or the Greens. 



Labor is ahead in the Bayesian aggregation of the two-party preferred polling. It is just behind in the localised regression model. The localised regression can be overly influenced by data points at the end of the series. 






On the primary votes, the Coalition has improved over recent months. Its vote share is above that obtained at the previous election. Labor's vote share is slowly declining and is below that of the previous election.










The Greens and other parties primary vote has declined a little over recent months. The Greens are above their primary vote share at the previous election. Other are below their primary vote share at the previous election. 









The attitudinal polling follows. While Albanese remains the preferred Prime Minister, Dutton has better net favourability.





Tuesday, May 21, 2024

May polling update

Let's start with the localised regression charts. In aggregate, the two-party preferred (2pp) polls are relatively unchanged over the past month. Labor is ahead of the Coalition. Nonetheless, Labor's primary vote is at its lowest point since the previous election, Together, these polling observations suggests the most likely election outcome, if one was held now, would be a Labor minority government.





Turning to the Bayesian polling aggregations, we see a similar picture (albeit a touch more favourable to Labor). The Bayesian analysis considers the tendency for each polling firm's methodology to favour one party or another. The model assumes that collectively the polls are unbiased, even if an individual polling firm may have an unintended bias. 

Please note, I am treating the Resolve Strategic polls from this year as a different series from its earlier polls. I do this when a polling firm indicates that they have changed their methodology, or it looks like a firm may have changed its methodology. In this case, it is because it looks like Resolve Strategic may have changed. 

Also note, the charts marked with GRW are based on a Gaussian Random Walk, where the population voting intention on one day is assumed to be much like the voting intention on the previous day. The charts marked with GP are based on a Gaussian Process, where polls taken closer together are assumed to be more correlated than polls taken further apart in time. The GP approach to poll aggregation is (for me at least) a little experimental. It has a tendency to revert to a base state when polling frequency is low, and at each end of the series. I have set the base state to the average of the last 10 or so polls, but this does affect the early part of the series where polling was less frequent. 







The GRW primary vote charts follow. 













Finally, a quick update from Sportsbet, which I seek to monitor on a daily basis. Little has changed in the past month in respect of how the betting market sees the outcome of the next election. (Note: the specific question is which party will provide the Prime Minister following the next election.)



Please note: the polling data for this analysis comes from Wikipedia. The software that generates these charts can be found on my GitHub repo.