Coalition fails to get post-budget boost predicted by commentariat

After the release of the Federal budget on Tuesday night, much of the political commentariat thought that the budget would be popular, and predicted a lift for the Coalition in the post-budget polls. Graham Richardson in The Australian said the government would "no doubt get a sugar hit from the budget".

All the regular post-budget polls are instead at least 53-47 to Labor, with little change apparent from the pre-budget situation. In Newspoll Labor gained a point, while in Ipsos the Coalition gained two points, leading to different commentary from Fairfax, which sponsors Ipsos, than The Australian, which sponsors Newspoll.

The last Ipsos was 55-45 to Labor in late March; this seemed an outlier at the time. The last Newspoll was 52-48 to Labor three weeks ago, and was probably influenced by the announcements on the citizenship test and 457 visas.

Here is the post-budget poll table. Two separate ReachTEL polls were conducted on 11 May, one for Sky News and one for Channel 7. They are the first public ReachTEL Federal polls since before the 2016 election. Only half of the Essential sample is post-budget, though this week's additional questions are based on the post-budget sample.

The Sky News ReachTEL was reported as 53-47 to Labor, and the Channel 7 ReachTEL as 54-46. However, both these results were based on respondent allocated preferences. To match polls that only give the previous election preferences, I am using Kevin Bonham's calculated two party vote from the decimal primaries of both ReachTELs. Since the rise of One Nation, ReachTEL's state polls have leaned to the Coalition, and this lean appears to be happening federally.

While individual budget measures, such as the bank levy and additional Medicare levy, are popular, the budget as a whole gets only a middling rating on a range of measures. Commentary suggesting that the overall budget would be very popular has been shown to be wrong.

While the budget allocated much spending to health and education, voters trust Labor more on these issues. A government that has tried to cut spending for three years, but suddenly has a poll-driven about-face strains credibility. Labor's fairness criticisms of the termination of the 2% deficit levy for high-income earners, and the now $65 billion for company tax cuts, are likely to be accepted by a large portion of the population.

Kevin Bonham's poll aggregate is at 52.7% two party preferred to Labor, a gain for Labor of 0.2 points since last fortnight.

Perceptions of this budget

After each budget, Newspoll asks three questions: whether the budget was good or bad for the economy, good or bad for the voter personally, and whether the opposition would have delivered a better budget.

45% thought they would be worse off and 19% better off, for a net of -26. 36% thought the economy would be better with this budget, and 27% worse, for a net of +9. Compared with previous budgets, neither of these scores are very bad nor very good.

Coalition governments do better than Labor ones on whether the opposition would have delivered a better budget. In this Newspoll, by a 47-33 margin, voters thought Labor would not have delivered a better budget. This 14-point margin is about the same as the last two budgets, but better for Labor than any budget in the Howard era, except the 2007 13-point margin, which came shortly before Rudd ousted Howard at the November 2007 election.

In other Newspoll questions, 45% said they would be prepared to see a reduction in taxpayer funded entitlements to pay down debt, while 41% thought otherwise. By 39-36, voters thought this budget was fairer than others under this government. As one of those budgets was the widely hated 2014 budget, this is not saying much. By 71-19, voters thought the banks would not be justified in passing on costs from the bank levy.

In Ipsos, by 45-44 voters approved of the budget, and by 42-39 they thought it was fair; these measure are much better for the government than following the 2014 budget. 50% thought they would be worse off with the budget, while 20% expected to benefit. By 58-37, voters supported increasing national debt to build infrastructure.

The Sky News ReachTEL found that 52% thought their family would be worse off with this budget, with just 11% for better off. 36% thought the government had done a good or very good job explaining its budget, 37% an average job and 27% poor or very poor. 34% of non-home owners thought the budget made it harder to buy a home, 13% easier, and the rest said there was no change.

The Channel 7 ReachTEL found that the budget was rated average by 38%, poor or very poor by 33% and good or very good by 29%.

In Essential, voters approved of the budget by 41-33, though 29% said it made them less confident in the government's handling of the economy, with 27% for more confident. On both questions, the strongest disagreement with the budget came from Other voters, not Labor and Greens voters.

Explaining why Shorten did not mention punitive measures against the unemployed in his budget reply speech, a crushing 76-14 supported payment reductions for jobseekers who fail to attend appointments, and 69-22 supported a drug trial for jobseekers. The second airport in Sydney was supported by 54-18.

By 51-27, voters agreed with the statement that the budget was more about improving the government's popularity than the economy. 56% thought higher income earners should bear a greater share of the cost of funding the National Disability Insurance Scheme, while 27% thought applying the Medicare levy for all taxpayers is the right approach. Scott Morrison was favoured over Chris Bowen as preferred Treasurer by 26-22 with 52% undecided.

There was strong support for the bank levy (68-21 in Newspoll, 62-16 in the Sky News ReachTEL, 60-18 in the Channel 7 ReachTEL, 68-29 in Ipsos and 66-19 in Essential). The additional Medicare levy was also well supported (54-36 in Newspoll, 48-34 in the Sky News ReachTEL, 51-28 in the Channel 7 ReachTEL and 49-39 in Essential).

Primary votes, leaders' ratings and other polling

Primary votes in Newspoll were 36% Coalition (steady), 36% Labor (up 1), 10% Greens (up 1) and 9% One Nation (down 1). 33% (up 1) were satisfied with Turnbull's performance and 53% (down 4) were dissatisfied, for a net rating of -20, up five points. Shorten's net rating was -22, down two points.

In Ipsos, primary votes were 37% Coalition (up 4), 35% Labor (up 1) and 13% Greens (downs 3 from an unrealistic 16%). 45% approved of Turnbull's performance (up 5) and 44% disapproved (down 4), for a net rating of +1, up nine points. Shorten's net approval increased a sizable 13 points to -5. Turnbull's ratings in Ipsos have been much better than in other polls. Ipsos skews to the Greens, but less this time than in their first two polls of the new parliamentary term.

The Sky News ReachTEL had primary votes of 37.8% Coalition, 34.2% Labor, 10.3% Greens and 10.2% One Nation. In the Channel 7 ReachTEL, assuming the 9.2% undecided are excluded, primary votes are 37.1% Coalition, 35.0% Labor and 10.8% for both the Greens and One Nation.

Primary votes in Essential were unchanged on last week at 38% Labor, 37% Coalition, 10% Greens, 6% One Nation and 3% Nick Xenophon Team.

In the Channel 7 ReachTEL, both leaders' ratings tanked from the final survey prior to the 2016 election. Turnbull's (total good) minus (total poor) score fell 18 points to -24, his record lowest, just ahead of Tony Abbott's ratings before Abbott was replaced. Shorten's rating was down 17 points to -21, his lowest since March 2016.

38% preferred Turnbull as Coalition leader, followed by 29% for Julie Bishop, 17% for Abbott, 11% for Peter Dutton and 6% for Scott Morrison. Among Coalition voters, it was 61% Turnbull, 18% Bishop and 14% Abbott.

For preferred Labor leader, Tanya Plibersek had 31% with Shorten and Anthony Albanese tied on 26%. Labor voters had Shorten leading with 40%, Plibersek on 33% and Albanese on 20%. Plibersek was strongly favoured by the Greens, with 51% support from them.

Turnbull led Shorten as better PM by 47-35 in Ipsos and 44-31 in Newspoll, but only 52-48 in the Channel 7 ReachTEL. ReachTEL uses a forced choice question, and this method usually benefits opposition leaders.

ReachTEL's respondent allocation problem

As noted at the beginning of this article, ReachTEL's respondent allocated preferences are over a point more favourable to Labor than using the previous election method. It appears that some of this difference is explained by ReachTEL asking National voters which of Labor or Liberal they prefer.

This is a mistake, as in most cases the Nationals are not opposed by a Liberal, and so their preferences are not distributed. In the few cases where National votes were distributed, 22% leaked to Labor at the 2016 election. Applying this rate to the 3.5% National vote in the Sky News ReachTEL would mean that Coalition leakage would increase Labor's two party vote by 0.8 points; the actual Coalition leakage is worth only about 0.1 points to Labor.

Ipsos also asked for respondent allocated preferences, and had Labor ahead by 53-47 on this measure, the same as when using the previous election method.

Adrian Beaumont, Honorary Associate, School of Mathematics and Statistics, University of Melbourne

This article is republished from The Conversation under a Creative Commons license. Read the original article.