Augury 2022

No birds were disemboweled in the making of this forecastEven if I wanted to do actual bird-augury I would not do it in Australia.

A model of the upcoming 2022 Australian federal election using various inputs such as voting intention polls, leader approval polling, economic data and historical performance to model the election. We then run 100,000 simulations to forecast the election.

If you’re interested in how first-preference vote shares convert to a Coalition-vs-Labor 2-party-preferred, we have a calculator which lets you input primary vote and preference flows to derive a 2-party-preferred estimate >>

Now that you have that 2pp estimate, what would the electoral map look like if it was replicated across the country on election day? Use this tool to simulate your very own election given a particular 2pp >>

Has Australian polling gotten worse over time?

What would happen if the 2019 polling error repeats itself?

If you like models and analyses like these, please consider supporting us on Patreon or donating to us using the PayPal button on the sidebar.

Labor is somewhat favoured to form government

We currently estimate that the Coalition will win between 44% and 52% of the two-party-preferred (2pp) vote on election day, with a median 2pp prediction of 47.9%. This range will likely narrow if we get more polling closer to election day.

What would the forecast have said at past elections?

As part of model development, the model was backtested on past elections by hiding all data from before each election and getting the model to “predict” that election.

Here’s what the 2pp forecast from the backtests on the four most recent elections would have been at the same point before each election – actual result plotted as a circle:

Seat totals forecast

The forecasted distribution of seats for each party/grouping. Higher bars indicate outcomes which happened more frequently in our simulations. Dashed lines represent the number of seats needed for an outright majority.

We’d expect the total number of seats won by each party/grouping to end up between the 1st and 3rd quartile about half the time. The seat total for each party should only fall outside the Lower/Upper ends once every 64 elections; roughly equivalent to flipping five coins and getting five heads.

Probability of various outcomes

Randomly simulate an election:

(If you’re on a mobile device, I highly recommend turning your device landscape.)

Division-by-division probabilistic forecast

This table shows how often each party/grouping won each district in every 100 simulations.

Note: not a forecast of the vote each party is expected to win in each electorate.

(if you’re on a mobile device, scroll right for full data or turn your device landscape)