Statistics on Preference Flows between the Coalition and One Nation

This post is additional information on preference flows related to a post I published earlier today, “Liberal, National and One Nation – Friends, Enemies or Frenemies?”

With One Nation’s surge in support in recent polls, pollsters and election modelers have been looking around for information on preferences. Especially for examples of how Coalition preferences might flow when excluded in contests between Labor and One Nation.

I’ve managed to pull as much information as I can from recent Federal and Queensland elections. I hope you find it useful.

One Nation Preferences to the Coalition 1998-2025

One Nation’s first Federal election was in 1998, shortly after its success at the Queensland election. Nationally the party polled 8.4%, less than expected before the campaign. The late poll surge that occurred earlier at the state election was not repeated federally and One Nation lost a third of its vote in Queensland, down from 22.7% to a still significant 14.4%.

In other states in 1998, One Nation polled 9.0% in NSW, 3.7% in Victoria, 10.1% in SA and 9.3% in WA. The party polled 4.8% in inner-metropolitan seats, 6.9% in outer metropolitan, and 5.9% overall in metropolitan seats. This was less than half the 11.9% One Nation polled in provincial and rural seats. In Labor seats One Nation polled 6.1%, 8.4% in Liberal seats, and 14.9% in National seats. The pattern of One Nation support has been the same ever since, strongest in Queensland, WA and SA, stronger in non-metropolitan seats, stronger in National seats.

One Nation largely disappeared as an electoral force after the 2004 federal election before returning in 2016 with a limited House campaign in Queensland. At elections since the party contested 59 seats in 2019, 149 in 2022 and 147 in 2025.

The left-hand side of the chart below shows One Nation’s national vote in these two periods. The bars on the right show the national flow of One Nation preferences to the Coalition at the same elections.

The right-hand side of the chart shows the slow growth of One Nation preference flows to the Coalition over time. Having been denied preference recommendations by the Coalition in its early years, One Nation tended to recommend preferences against sitting member, producing relatively even splits of preferences. In 2019 One Nation normally put the Coalition ahead of Labor, though not necessarily at 2, and preference flowed around 65% to the Coalition. In 2025 One Nation put the Coalition as 2nd preferences in many more seats and preference flows increased to 74.5%.

The same dramatic change shows up when you compare the flow of One Nation preferences by division in 1998 and 2025.

Last week’s Newspoll had One Nation on 22%, the Coalition on 21%, Labor 32% and the Greens on 12%. Adding them together, One Nation and the Coalition is on 43%, Labor and the Greens 44%.

But you can’t add together the vote of competing parties. At elections this is done by the preferences voters mark on their ballot papers. Given Labor and the Greens achieve preference flows of around 85%, their support merges to around 42%. With only 75% flows between parties of the right, the merged right total is only around 38%.

Victory by some combination of the Liberal, National and One Nation parties requires higher preference flows between them than we have seen in the past. While we have data on how One Nation preferences flow, there are fewer examples where Liberal, National and indeed Labor preferences flow in seats where One Nation finishes in the final pairing of candidates.

The scatter chart below plots One Nation’s preference flow to the Coalition by electorate against One Nation’s primary vote in the same seats. The chart shows some variability in flow, but the trend line is virtually flat pointing to 75% being the default flow of One Nation preferences to the Coalition. There’s a few very low One Nation votes in ‘teal’ Independent seats with surprisingly high preference flows to the the Coalition. I suspect it was a very hard core One Nation support in these strongly small-l liberal seats.

The chart shows that One Nation flows to the Coalition are consistently higher in Liberal and National seats than they are in Labor seats. Which raises a conundrum for the Coalition that preference flows from One Nation have tended to be lower in the seats that the Coalition needs to gain from Labor, and higher in its own seats which it must defend against One Nation.

Coalition and Labor Preferences in One Nation Federal Contests

At the 2025 federal election there where only two seats where One Nation finished in the final candidate pairing after preferences.

In the NSW seat of Hunter, easily retained by Labor, One Nation reached second place and received 83% of National Party preferences. As I explain below, this figure is similar to flows seen in regional seats at the 2020 Queensland election, but higher than the flow seen elsewhere.

In David Littleproud’s Queensland seat of Maranoa, he won with more than 50% of the primary vote, but Labor slipped to third place on preferences. Labor had recommended preferences for Littleproud, but only 57.9% of Labor voters followed the recommendation. This low figure for Labor preferences to the Coalition has been common in the sort of rural seats where Labor can finish third.

Labor previously slipped to third in Maranoa at the 2019 Federal election when Labor preferences flowed 65% to the LNP.

The 2017 Queensland Election

After many years as a minor player, Pauline Hanson’s One Nation returned and polled strongly at the 2017 Queensland election. The 2017 election was also the first Queensland election in 28 years to use full rather than optional preferential voting.

One Nation polled 13.7% state-wide, 20.8% in the 61 of 93 seats contested. The party won only a single seat, the central Queensland seat of Mirani, but finished in the final pairing against Labor in 13 seats and against the LNP in eight. The party polled above 20% in 35 electorates and 40% in four.

We do not have actual preference flows for the 2017 election, but we can use final preference distribution data for electorates. There were 33 electorates where One Nation candidates were the last excluded, 12 where LNP candidates were last excluded, and eight with Labor candidates last excluded.

The final preference distributions for these pairings were –

• The 12 seats where One Nation candidates were the last excluded and the party recommended preferences to Labor, preference distributed 42.1% to Labor and 57.9% to LNP. The flow to the LNP ranged from 49.1% to 65.2%.
• The 21 seats where One Nation candidates were last excluded and the party recommended preferences to the LNP saw preference distributions of 31.4% to Labor, 68.6% to LNP. The range of flows to the LNP was 63.3% to 78.0%.
• There were 10 seats where the LNP candidate was last excluded and preferences were recommended to One Nation. The total preference distributions were Labor 29.2%, One Nation 70.8%. The range of flows to the LNP was from 63.1% to 79.2%.
• In two seats the LNP candidate was the last excluded and the party recommended preferences to Labor. Preferences split 50% to Labor 50% to One Nation.
• Labor recommended preference for the LNP over One Nation in all seats. In the eight where Labor’s candidate was the last excluded, the total preferences distributed were 65.0% to LNP and 35% to One Nation. All eight were LNP safe rural seats and the flows to the LNP ranged from 58.4% to 71.4%.
• There were 27 electorates where Greens candidate were the last excluded and the flows were 75.6% to Labor and 24.4% to the LNP. The flows were above 78% to Labor in all seats where the Greens polled above 20%.

In short the preference recommendations of the LNP and One Nation impacted the preference flows, but neither managed to deliver flows that matched the Greens’ preference flows to Labor.

The 2020 Queensland Election

One Nation contested 90 of Queensland’s 93 electorates in 2020. Despite the increase in candidates, the party’s vote halved to 7.1%. There were only two seats where the party finished in the final pairing. One Nation’s David Ashby noted the difficulty his party faced campaigning with the closure of local newspapers during the first year of Covid.

For the 2020 election we have reliable preference flow data collected on election night. In 70 Labor/LNP contests, Green preferences flowed 80.1% to Labor, 19.9% to the LNP. In 69 Labor/LNP contests, One Nation preferences flowed 33.8% to Labor and 66.2% to the LNP. The 2020 election is another example of how much stronger Green preference flows to Labor are compared to One Nation’s flows to the LNP.

With One Nation having polled so well in many seats in 2017, there were a number of seats where on election night the Electoral Commission tallied LNP and Labor preferences.

LNP preferences were counted out in 12 electorates, of which eight were in contests between Labor and One Nation. The data for all 12 is shown below, the first eight being the Labor-ONP contests. The flows were election night counts only, shown as a total in the “Sample Votes” column. The final primary count and percentage for LNP candidates is shown in the final two columns.

Only in the regional seats of Maryborough and Mirani did the flow of LNP preferences to One Nation exceed 80%. It was between 67.7% and 76.5% in five listed urban Labor seats, Bundamba to Logan. These seats surround the Ipswich district where One Nation had polled so strongly in 1988 and again in 2017. Across the eight seats where LNP preferences were counted out, the flows were in total 76.1% to One Nation.

Note that LNP preferences were stronger to Katter’s Australian Party in the two seats where counted out. Also the 2020 Queensland election was the first since the 2010 Federal election where one of the Coalition parties recommended preferences for the Greens over Labor, being largely responsible for defeating Jacki Trad in South Brisbane.

There were also six seats where Labor preferences were counted out in LNP/ONP contests.

Across the six electorates, only 60.2% of Labor voters followed the party’s recommendation of sending preferences to the LNP. Combined with the low rate for Maranoa in 2025, this suggests that in regional seats where the Nationals are challenged by One Nation, Labor preference flows will be weak and not be enough to lift trailing National candidates ahead of One Nation. Labor voters were much more likely to follow their party’s recommendations of preferences for Katter’s Australian Party ahead of the LNP.

2024 Queensland Election

One Nation contested all 93 electorates. Preference flow data is available for 87 Labor versus LNP contests, and in line with results at the 2025 Federal election, flows of One Nation preferences to the LNP increased, the levels 74.3% to the LNP and 25.7% to Labor.

The Greens contested all 93 electorates and polled 9.9% of the vote. There were 86 contests where Green preferences were distributed in Labor/LNP contests with preferences flowing 81.2% to Labor and 18.8% to LNP.

There were no seats where Labor or Liberal candidates had their preferences distributed in contests versus One Nation.

Conclusion

A number of pollsters have begun to calculate nation-wide after preference counts between Labor and One Nation in addition to the traditional two-party preferred estimate between Labor and the Coalition.

I’m not really sure how useful these competing numbers are.

In 2025, only 115 of the 150 electorates finished as two-party preferred contests between labor and the Coalition. There were 14 Coalition-Independent contests, 7 Labor-Independent races, nine Labor-Greens, one LNP-Green race, one LNP-ONP, one Labor-ONP, and one contest each involving Katter’s Australian Party and the Centre Alliance.

If One Nation draws level with the Coalition at the next election, the national result will be an even bigger mix and match of electoral contests. Which makes it difficult to rely on a nationwide count after preferences.

Initially one polling company used the 83% National to One Nation preference flow from Hunter to estimate a Labor-One Nation nationwide preference count. Others are now using Senate preference flows as a guide.

While it is clear that support for Labor and the Greens coalesces with an easily assumed preference flow above 85% between the parties, the evidence of the 2025 Federal and recent Queensland elections shows the similar figure between the Coalition and One Nation is around 75%.

It’s why in my previous post I pointed out that the chances of support for the Coalition and One Nation coalescing into an election winning bloc requires stronger preference flows than we have seen in the past.