Before we proceed with squeezing some last juice out of the post-budget polling, note that there are two other fresh posts below this one: one a guess post from Adrian Beaumont on Canada and the United States, the other a summary of recent state polling from New South Wales, Queensland and Victoria.
Part of the latter is a new DemosAU poll from New South Wales encompassing federal as well as state voting intention. The federal voting intention numbers have the Coalition leading 51-49, compared with 51.4-48.6 to Labor at the 2022 election, from primary votes of Labor 30% (33.4% in 2022), Coalition 38% (36.5%), Greens 12% (10.0%) and One Nation 9% (4.9%). The poll also finds Anthony Albanese leading Peter Dutton 39-38 on preferred prime minister, and 31% holding that Australia is headed in the right direction compared with 52% for the contrary view. The poll was conducted last Thursday to Saturday from a sample of 1013.
The full results from the Essential Research poll included results on the budget, including a finding that less attention was paid to it than the last two, with 36% saying they paid a lot of attention (down four on last year and nine on the year before) and 15% that they had paid no attention (up four on last year and three on the year before). Forty-one per cent felt the budget would be good for the well off (down five on last year and unchanged on the year before) and 27% felt it would be good for those on lower incomes (down three last year and fourteen on an unusual result the year before). A question on the Trump administration’s tariffs finds 37% holding that Australia should look for new trade relationships, 29% that retaliatory tariffs should be imposed, and 35% that the priority should be to remain on good terms with the US and seek exemptions.
Now for a closer look at Newspoll’s budget responses, which maintain a consistent set of such questions that the pollster in its various incarnations has been posing after each budget since 1988, encompassing 39 budgets overall. The minus 10 rating for last week’s budget in terms of its impact on the economy was the fourth worst result yet recorded, surpassed only by three successive Hawke-Keating government budgets in 1991, 1992 and 1993 (the latter was in a league of its own at minus 42, the budget in question being remembered for its breach of Paul Keating’s “L-A-W” tax cuts promise). The minus 19 rating on impact on personal situation, by contrast, rates around the middle of the field.
The chart below records how each budget scored on the two measures. While respondents invariably score budgets more favourably on economic than personal impact (last year’s two-point differential was the closest any had yet come to breaking the mould), the trendline points to a tendency for budgets to be generally perceived either as good or bad, reflected in relatively strong or weak results for both measures. Last week’s budget is the one marked in red – its placement below the trendline indicates that, as mediocre budgets go, respondents felt this one relatively better for themselves than for the economy.
Lest anyone overestimate the electoral significance of this result, the best result of plus 48 was at the budget preceding the defeat of the Howard government in 2007. Perhaps more to the point, the minus 11 rating on the question of whether the Coalition would have done better is part of the course for a Labor budget (overall average minus 10.4%) – consistent with the fact that the Coalition generally does better on economic management polling, their budgets tend to do better on this question than Labor’s (average of minus 17.3%, the overall overage being minus 14.2%). Another reason to doubt the budget’s electoral impact is the one just noted by Essential Research – that voters were unusually disinterested on this occasion, which does not factor in to Newspoll’s calculations.
Finally, some more on YouGov’s MRP poll, which didn’t get the attention it warranted amid the Sunday polling avalanche. Even more so than the first wave in late January and early February, the second wave from March was distinctive in suggesting that Labor is actually holding up well in Victoria – so much so that the lineball Liberal-held seat of Deakin is rated “toss-up Labor”, while the two seats rated more likely than not to be Liberal gains in the first wave, Chisholm and Deakin and now rated “lean Labor”. Even with five seats in the state now back in the their column, New South Wales continues to be rated the most troublesome state for Labor, being home to all four of the seats likely to be lost.
Since the post-stratification approach leans heavily on demographic variables in estimating seat-level results, I have made an effort to identify the underlying changes in the survey that have yielded the movement from the first wave to the second, which ranges from five points towards Labor (in Bean and Canning) to three points towards the Liberals (in Watson). This has been done through a linear regression analysis that uses the change in Coalition two-party preferred from one wave to the next as the dependent variable, with predictor variables of state/territory, AEC seat classification (inner metropolitan, outer metropolitan, provincial or rural) and four census demographic variables yielded through trial and error.
The demographic variables found to be highly significant (as indicated by the three asterisks) are Index of Relative Socioeconomic Advantage and Disadvantage, which is an Australian Bureau of Statistics measure of general affluence that avoids the pitfalls of income-based measures; the percentage of the population aged 55 and over; the percentage of the 18-plus population with trade certificates; and the percentage who primarily speak a Chinese language at home. The negative coefficients indicate that electorates scoring high on these measures tended to move most strongly to Labor from the first wave to the second. Conversely, the state/territory and AEC region classification variables prove not to be too illuminating. Keep in mind that what’s being measured here is change between wave one and wave two, not swing since 2022 – a potential subject for a future post.