Votes for minor parties hurts Labor but Palmer overstates his influence

Almost a quarter (24.7%) of all formal votes cast in the House of Representatives at Saturday’s election were for a minor party or an independent. This is a new record, exceeding the 23.5% who ignored the major parties in 2016. The story about why Labor lost this election can be found in the composition of this one-quarter of the electorate, which is less Labor-friendly than the minor party voter base in 2016.

Neither major party’s primary vote changed by much. The Coalition’s primary vote has dropped by 0.6%, while Labor’s has dropped by 0.9%. Yet Labor’s two-party-preferred vote has dropped on average by about 1% across the seats where a swing has been recorded. This swing can be explained by which voters are voting for minor parties.

Clive Palmer’s United Australia party is currently polling 3.4% of the national primary vote in the lower house. This party did not contest the 2016 election, so all of these voters have swung from somewhere else.

Related: Senate results: Hanson-Young returns, but Hinch, Anning and Burston are gone

One Nation is currently sitting on 3% of the vote, a swing of 1.7% compared with 2016. One Nation only ran 15 candidates in 2016; this time it ran 59. The party actually suffered negative swings in nine of the 15 seats where it ran in 2016, so its “swing” is only explained by an ability to nominate more candidates.

There is no doubt that voters for these two parties are more favourable to the Liberal National party than was the case for minor party voters in 2016. But is there reason to think that these parties influenced voters who may have otherwise preferenced Labor to instead favour the Coalition?

I think there is some evidence that this may be the case for One Nation voters, although it may also be the case that Labor’s losses were simply more concentrated in areas with more One Nation voters.

I’m analysing the 135 seats where we have a two-party-preferred count between Labor and the Coalition, to allow for a clean comparison. The average swing against the ALP was about 2% in seats where One Nation ran, rather than a swing of just 0.1% in seats where they did not run. But these seats were already more conservative before this election, so I don’t think you can explain the difference by the presence of One Nation.

In the 57 seats in our sample where One Nation did run, there appears to be a negative correlation between the Labor two-party-preferred swing and the One Nation vote: the higher the One Nation vote, the bigger the swing against Labor.

We can go further and do the same analysis at the booth level, rather than the seat level, and again you can see a trend of a higher One Nation vote translating into a bigger swing against Labor.

We will eventually get detailed data which shows how many One Nation voters preferenced Labor or the Coalition, which will allow us to separate out their preferences from those of Greens voters and other minor party voters. But for now we have this information.

It does suggest that One Nation voters may have been influenced to preference the Coalition (in particular the LNP in Queensland), but there is an alternative explanation. It’s possible that Labor’s difficulties were simply concentrated in the kinds of seats where One Nation is more popular, such as in north and central Queensland. Some of the voters in this area may have switched their preference from Labor to LNP regardless of the presence of a One Nation candidate, but gave One Nation their primary vote where that was possible.

So what about the vote for Palmer’s United Australia party, who had agreed on a preference deal with the Coalition?

I can’t find much evidence of a similar trend for United Australia voters. The equivalent chart shows no particular trend:

There is no doubting that One Nation and United Australia voters lean towards the right and contributed to a swing against Labor in certain parts of the country. It’s not clear how much the parties influenced those voters’ preferences, or if it just reflected their pre-existing political inclinations.