Australian political opinion: from the 2019 election to COVID-19

Sarah Cameron and Ian McAllister*

Australian political opinion: from the 2019 election to COVID-19 (PDF 751KB)

In this paper we will cover 3 main topics. First, we examine the results of the 2019 federal election and discuss the 2 major issues that determined the outcome of that election, namely, policies on taxation and factors associated with leadership. Second, we provide an overview of long term trends in electoral behaviour, and how these trends may affect elections in the future. Finally, we look at the political implications of the COVID-19 pandemic.

Before we turn to these topics, some background about the Australian Election Study (AES) survey itself. We have been conducting the AES since 1987, completing 12 surveys after each federal election. We typically ask about 250 questions of each survey respondent; about 90 questions are ones that we ask consistently from election to election. We have therefore accumulated a huge amount of information about why people voted in each election, what they thought was important, and much else besides. This unrivalled database allows us to trace long term trends in electoral behaviour.¹

The survey we conducted in 2019 went into the field immediately after the election in May and it was in the field until September.² We surveyed just over 2,000 respondents nationally, with a response rate of 42%. More information and interactive charts are available at www.australianelectionstudy.org.

The 2019 election

The economy and taxation

Since 1996 the AES has consistently asked voters what they considered to be the most important factor that shaped their vote. From Figure 1 it is apparent that consistently across all that time period, policies are what determined a person's vote. There is a spike in 1998; this was the election in which the goods and services tax that the Liberal Party proposed was the major election issue. The importance of policies then declined somewhat in 2001, but it has been consistently increasing ever since. In 2019 a total of 66% of the AES respondents said that policies were the major factor which affected how they voted; this is similar to the proportion in 1998.³ Aside from policies, about one in 5 people say that it is the political parties that determine how they vote. These considerations relate to whether the political parties are seen as divided or united, as well as to leadership and party loyalties. At the other end of the scale less than one in 10 mention the party leaders or the local party candidates as determining how they vote.

Figure 1: Considerations in the voting decision

The AES also consistently asks people what they considered to be the most important election issues. Figure 2 shows the proportion of respondents who chose one of 10 issues as being their first most important. We find that the major issues that determine how people vote tend to be economic management, health, and education. The 2019 election was very similar to previous elections in that economic management and health were the 2 top issues. Although there were slight differences, the results in Figure 2 show that education declined to the last half of those issues. Similarly, 11% of people mentioned the environment as their most important issue, and another 10% mentioned global warming. If these 2 issues are added together, it means that the environment, broadly defined, was one of the top 3 election issues.

Figure 2: Most important election issues

In terms of how the party voters viewed these issues, Figure 3 shows that 3 in 4 Coalition voters thought that economic issues were the most important in the election. Labor voters, by contrast, were more diverse in the issues that they selected, covering health, the environment and the economy. Perhaps not surprisingly, two-thirds of Greens voters identified the environment as being the issue that they considered most important in the election.

Figure 3: Most important election issues by vote

We also ask voters which of the 2 major parties they prefer to manage these various issues (Figure 4). This shows a consistent pattern where the Coalition has an advantage on economic issues. For example, the Coalition has a substantial advantage over Labor as the preferred party on the economy and government debt, as well as on taxation. At the other end of the scale, Labor has a very strong advantage over the Coalition on issues such as health, education, the environment and global warming. On superannuation, the Coalition has an advantage over Labor of 14 percentage points. However, going back to 2016 the Coalition had put forward a series of policies to restrict superannuation and the amount that people could contribute. When we asked that same question in 2016 voters saw no difference between the 2 major parties. In 2019 the Coalition advantage opened out on the issue significantly.

Figure 4: Preferred party policies

The central economic issue in the election was taxation, and Labor proposed a series of significant policy changes. Labor argued that there should be major tax changes to dividends and tax imputation, as well as changes to the taxation of investment properties through capital gains tax and negative gearing. The AES asked the voters their views on these tax changes, and the respondents were very divided on them.⁸ Slightly more thought it was a good policy, slightly less thought it was a bad policy. The net effect, as Figure 5 shows, is that there was a major division between the 2 major parties on the issue of taxation. The Coalition had a significant advantage as the party most able to handle the issue. If we compare that to 2016 there was only a 2 percentage point gap between the 2 major parties; in 2019 that opened out significantly to a 13 percentage point difference. The other interesting part of Figure 5 is the proportion of voters who thought there was no major difference between the parties in 2019; at 14% this was the second lowest figure since 1990, the other one being the 1998 election, which was the election fought on the issue of the introduction of a goods and services tax.

Figure 5: Preferred party policy on taxation

Labor's economic policies in the election were predicated on the idea that asset ownership was concentrated among relatively small groups of voters who were affluent and able to pay the extra tax that was being proposed. What our research shows, is that this was not necessarily the case. While some voters are affluent and would not have been unduly affected by the tax changes, other asset owners were not particularly well-off. This latter group are using investment properties and share portfolios in order to provide an income in retirement or to build up superannuation.¹⁰

A major change that we have observed in the electorate over the last 20 years is a significant increase in asset ownership across the electorate. Figure 6 shows that two-thirds of people say that they are homeowners, evenly divided between those who own their properties outright and those who are paying them off. This represents a slight decline over the last 20 years or so, largely caused by younger people not being able to enter the housing market because of the cost of property.¹¹

Figure 6: Asset ownership across the electorate

The second bar in Figure 6 shows that around one in 3 directly own shares on the Australian Stock Exchange.¹³ That is also a decline, because it was around 20 percentage points higher in the early 2000s when there was the privatisation of Qantas, CSL (Commonwealth Serum Laboratories), Commonwealth Bank, and other government-owned entities. There was a significant decline in share ownership after the global financial crisis. But even with 34% of the AES respondents saying they directly own shares, this represents one of the largest proportions of direct share ownership in the world, more than the United States (US), the United Kingdom (UK) and Canada.

The final assets in Figure 6 are ownership of an investment property and a self-managed superannuation fund; in each case about one in 5 people said that they owned such an asset. This estimate is higher than the proportion reported by the Australian Bureau of Statistics (ABS).¹⁴ The reason for this discrepancy is that the ABS uses a strict legal definition of asset ownership, while the AES question identifies people within the household. For example, there could be a household of 4 or 5 people but only one or 2 people might technically own that asset. However, the taxation applied to the asset will affect the income of the total household. We believe that this was one of the key mistakes in Labor's policy towards superannuation and property investment. The proposed changes would have affected a larger proportion of people than Labor assumed. For example, the AES finds that

3 in 4 voters own at least one of these assets. When we compare voting between 2016 and 2019, we find that 19% more Coalition votes in 2019 came from people who owned

self-managed superannuation funds compared to the previous election. So, there was a significant movement in votes between those 2 elections, with a shift of about 8% towards the Coalition among investment property owners alone.¹⁵

Leadership

Apart from taxation and economic policy, the other factor that emerged in the election as important was leadership. Two aspects of leadership were important. One was the relative unpopularity of Bill Shorten as Labor leader, while the second was the fourth change of prime minister outside of an election in the space of 8 years, from Malcolm Turnbull to Scott Morrison. What we find from the data is that Bill Shorten's unpopularity was a factor that harmed Labor, while the change in prime minister was relatively unimportant.

To place leadership popularity in long term perspective, the AES consistently asks the respondents to rate the party leaders on a scale from zero to 10. Figure 7 shows the figures for all of the major party leaders from 1987 through to 2019. The figures at the end of the bars are the mean value of the score from zero to 10. At the top are the most popular leaders over the past 30 or so years; these include Kevin Rudd in 2007 and Bob Hawke in 1987. Bob Hawke was probably even more popular in 1983 but we lack earlier comparable data. At the bottom of Figure 7, Bill Shorten is the least popular leader over the period, with the exception of Andrew Peacock in 1990. Shorten was also relatively unpopular in 2016.

Figure 7: Leader popularity 1987-2019

Leadership is electorally significant because there is a large body of research which looks at how voters evaluate leaders based on their qualities.¹⁷ The AES consistently asks a question about how appropriately 9 particular qualities apply to the various leaders. Figure 8 shows that Scott Morrison led Bill Shorten on all of these 9 characteristics, with the exception of compassion, where they are rated equally. The research in Australia and internationally shows that the most important quality that voters look for is integrity; in practice voters are seeking leaders who are honest and trustworthy.¹⁸ Voters also want to see leaders who exhibit leadership, as reflected in strength and inspiration. From the characteristics in Figure 8 it is clear that Bill Shorten fell very far behind Scott Morrison on these qualities.

Figure 8: Leader characteristics

Leadership played out in the election in that only about 4% of Labor voters said they were motivated by leadership in the election (Figure 9). If we compare the trend back to 2007, around 20% of Labor voters said they were motivated by leadership, when the highly popular Kevin Rudd led Labor. Even in 2010 and 2013 the results show that leadership figured significantly for Labor voters. Liberal voters were more motivated by leadership in 2016 and 2019 than Labor voters, but again if we look back to 2001 and 2004, the heyday of the Howard Coalition government, it is clear that leadership was much more important at that time.

Figure 9: Voting based on the party leaders

The other aspect of leadership is the consistent changes in prime minister, 4 of them occurring outside of elections since 2010. Each of the AES surveys has asked the respondents what they thought of these leadership changes. It is clear from Figure 10 that for the most part voters were not impressed. They particularly disliked the change from Rudd to Gillard in 2010. They also disapproved of the change from Turnbull to Morrison in 2018. The remaining 2 changes are more ambiguous, but in each case there was a majority who disapproved of it.

Figure 10: Attitudes towards the leadership changes

The 2019 AES survey asked the respondents if they would have changed their vote if Malcolm Turnbull had remained Liberal leader. These results show that while there would have been an exodus of voters from the Liberal Party, it is almost exactly matched by voters who would have been attracted to the Liberals. In effect, the fourth change in prime minister, from Turnbull to Morrison, did not affect the outcome of the election.

A divided electorate?

The 2019 election can be situated in the context of long-term trends in Australian political behaviour and attitudes. Following the 2019 election result, which few expected, there was commentary about the existence of 'two Australias'—an increasingly divided electorate contributing to the unexpected outcome.²² The AES data allows us to unpack some of the longer-term divisions that have emerged in the electorate, how they fed into the election result, and what this means for the trajectory of democratic politics in Australia.

Gender

Starting with gender differences, in 2019 there was a substantial gender voting gap as shown in Figure 11. Ten per cent more men than women voted for the Coalition, and the Greens attracted a good deal more support among women compared to men. Placing the gender gap seen in 2019 in long-term perspective shows that this was not the usual state of affairs. Figure 12 shows that the gender gap in voting behaviour has actually reversed over time. In the 1990s women were slightly more likely to vote for the Liberal Party, and men were more likely to vote Labor. Over time this has gradually reversed so that women now prefer Labor and men prefer the Liberals. In 2019 there was the biggest gender gap in voting behaviour on record.

Figure 11: Gender and vote choice

There are a number of factors underpinning this transformation of gender and voting in Australia. This includes tremendous societal changes that have taken place over this time. An increasing proportion of women undertake higher education, which is associated with greater support for parties on the left. Moreover, women have greater representation in the labour force and as union members.²⁴

Figure 12: Gender differences in voting for the Liberal and Labor parties

Changes within Australia's major political parties have also contributed to this shift in gendered voting patterns. Back in the early 1990s women were similarly underrepresented in both the major parties. Just 13% of parliamentarians in 1990 were women.²⁶ Since then, Labor has dramatically increased their proportion of women in Parliament, to 47%, through introducing voluntary party quotas.²⁷ The Liberal Party on the other hand has made slower progress, with just 23% of Liberals in federal parliament being women.²⁸ There has also been a shift in parties' issue priorities over time. In particular, Labor has shifted from a focus on working class issues to incorporate a broader set of priorities including progressive social issues. So, there are factors contributing to the gender gap both within the electorate and within political parties. This reversal of the gender gap is not unique to Australia – it has been observed in other democracies including in Europe and North America.²⁹

Age and Generation

The AES data also shows generational differences in political attitudes and behaviour. Younger Australians are primarily voting for Labor and the Greens, while a majority of those over 65 are voting for the Coalition as shown in Figure 13. The Greens have a lot of support amongst younger Australians, while older Australians are primarily casting their ballots for the major parties.

Figure 13: Age and vote choice

We can look at the trends over time to see whether this age gap is unique to the 2019 election or a continuation of what has happened in the past. What the results in Figure 14 show is that younger Australians have always been further to the left of older Australians, although the division is growing over time. The current generation of young people are much less likely to vote for the Liberal Party compared to previous generations when they were young. The gap between the voting behaviour of younger and older Australians was greater in 2019 than at any other time on record. The AES surveys have typically shown that as voters get older they shift further to the right in their political preferences. As Millennials and Generation Z are further to the left to begin with, this has potential long-term ramifications for the preferences of Australian voters, as these generations get older.

Figure 14: Age differences in voting behaviour

The generational divide is also evident in where Australians' place themselves ideologically from left to right. Figure 15 shows the average placement of voters on a left-right scale from zero to 10. This shows that the electorate as a whole has been gradually shifting to the left over time. Young people have consistently been further to the left of the electorate as a whole, although there are further indications of generational change in ideology, with the current generation of younger Australians moving a lot further to the political left.

Figure 15: Left-right ideology

In the lead up to the 2019 election there was discussion about it being a 'climate change election'.³³ Examining attitudes on climate change shows further evidence of generational change on this issue. Although Labor, the preferred party on environmental issues, lost the election – there is some support for the idea that 2019 was a climate election. Figure 16 shows that one in 5 voters identified the environment or climate change as their top issue priority in the election. This is a greater proportion than at any other time on record. Young people in particular saw environmental issues as important – around half identified the environment or global warming as the most important election issue. This heightened concern about climate change took place in the context of a wave of global climate change protests in 2019, including in Australia, led by young people.

Figure 16: The environment and global warming as most important election issues

Social class

Scott Morrison declared the 2019 election a victory for the so-called 'quiet Australians' a somewhat ill-defined group that has been compared to John Howard's battlers.³⁵ A question frequently raised in these discussions, is whether the working class, traditionally considered Labor voters, are shifting their votes to the Coalition. The AES data provides a number of ways to look at the voting behaviour of different social classes, although understanding social class is not so straightforward. Previous understandings of class based on occupation no longer reflect the complexities of social class in modern Australia, with the increasing importance of asset ownership.³⁶

One way of examining social class is how people identify themselves. The AES asks respondents to identify themselves as either working class, middle class, or upper class. Very few Australians see themselves as a member of the upper classes, and about half and half say they are working or middle class, respectively. Figure 17 shows how the self-identified working class have voted over time. This shows that the working class are still more likely to vote Labor than Liberal, although the trends over time show a gradual erosion of Labor's working-class base. In the late 1980s, 60% of the working class voted Labor, by 2019 this had dropped to 41%. The drop in support for Labor was in favour of minor parties more so than the Liberal Party. Similarly, the data shows that asset owners are much more likely to vote for the Liberal Party.³⁷ Class therefore remains an important influence on voter behaviour. Although some of the traditional patterns are eroding, while new class divisions, in particular based on assets, are of greater importance.³⁸

Figure 17: Working class vote choice

Voting volatility

Another major change that has emerged in the Australian electorate over time is increasing voter volatility. Back in the 1960s around 70% of voters would always vote for the same party, by 2019 less than 40% always voted the same way (Figure 18). There are a number of other indicators that point in a similar direction. More voters are making up their mind about how they are going to vote during the election campaign, rather than far in advance.⁴⁰ An ever-growing proportion of voters do not align with any of Australia's political parties. Partisanship has reached record lows – one in 5 voters have no partisan alignment.⁴¹ Combined, these factors are resulting in more unpredictable elections. This presents both opportunities and challenges for political parties, who can no longer rely on particular groups of voters for support, increasing the importance of the election campaign and leadership to shift votes.

Figure 18: Voter volatility

Citizen disaffection

A final factor where we are seeing major shifts in the electorate is in citizens' attitudes towards democracy. Various indicators show record levels of citizen disaffection with democratic politics in Australia. Satisfaction with democracy in 2019 reached its lowest level since the 1970s Whitlam dismissal; fewer than 60% of Australians were satisfied with the performance of democracy (Figure 19). Trust in government reached its lowest level on record, with just one in 4 voters believing people in government could be trusted.⁴³

Figure 19: Satisfaction with democracy

Placing Australia's level of democratic satisfaction in international comparison in Figure 20 shows that in 2007, when Labor won the election led by Kevin Rudd, Australians would have been among the world's most satisfied democrats alongside Norway and Switzerland. Since then, Australia has dropped down to the middle of the pack among democracies in the Organisation for Economic Co-operation and Development (OECD). Comparing the downward trend in satisfaction with democracy in Australia to other democracies around the world shows that this downward trend is not universal (Figure 21).

Figure 20: Satisfaction with democracy in OECD countries

There has been a decline in satisfaction with democracy in the UK, particularly following the Brexit referendum, and in the US, following Trump's 2016 election win. Canada and New Zealand on the other hand, have not experienced this decline in democratic satisfaction. Some have argued that political disaffection in Australia is simply a reflection of global trends, such as the rise of social media or increasingly disaffected young people. Cross-national trends suggest these factors do not explain the decline in democratic satisfaction, as Canada and New Zealand also have social media and young people however have not experienced a decline. Rather, the evidence suggests that the steep decline in satisfaction with democracy in Australia has been driven primarily by government performance.⁴⁶ The merry-go-round of Australian prime ministers during the 2010s undermined democratic satisfaction with democracy, alongside other dimensions of government performance.

Figure 21: Satisfaction with democracy in Anglo-American democracies

The COVID-19 pandemic and political opinion

So far, we have covered the 2019 election, and the long-term trends up to 2019. Of course, a few things have changed since 2019 with the onset of a major global crisis from the COVID-19 pandemic. The pandemic has significant implications for democratic politics around the world including in Australia. The COVID-19 pandemic has brought about a multifaceted crisis, combining a major public health crisis with an economic crisis. Research on previous crises gives us some indications as to how crises can shape democratic politics. Previous studies have found that in times of certain crises—particularly military threats—there is a tendency for voters to rally around the flag, where people give greater support to incumbent leaders and governments at a time of crisis.⁴⁸ An example of this effect could be seen in 2001 following 9/11, when George W. Bush's approval ratings skyrocketed after the terrorist attack.⁴⁹ On the other hand, there is a lot of research on how economic conditions influence voter behaviour and attitudes – with the central idea being that people punish governments for poor economic performance. Although in the context of a global crisis, benchmarking can become important – how well the crisis is being handled in comparison to other countries.⁵⁰ Given the unique nature and unprecedented scale of the COVID-19 crisis, this raises questions about its effects on citizens' attitudes.

To investigate the rally 'round the flag' effect we can examine approval data for leaders—Scott Morrison in Australia as well as Boris Johnson as a point of international comparison. Polling data shows that there was a 'rally round the flag' effect in both Australia and the UK at the beginning of the pandemic, in March to April 2020 when many countries around the world first went into lockdown. Both Scott Morrison and Boris Johnson received a boost in support of around 20 percentage points (Figure 22). In Morrison's case, before the pandemic his approval ratings were exceptionally low as a result of the bushfire crisis which he was perceived as handling poorly.⁵¹ After the initial jump in support at the beginning of the pandemic, Morrison's support remained exceptionally high throughout the pandemic, whereas for Boris Johnson, the 'rally round the flag' effect was relatively short-lived. We could expect this divergence stems from the 2 countries experiences in handling the pandemic, with Australia doing exceptionally well, whereas the UK was one of the most affected countries in terms of COVID-19 cases and deaths.

Figure 22: Approval ratings of Scott Morrison and Boris Johnson

ANUPoll data on confidence in state and federal governments in Figure 23 shows further evidence of the 'rally round the flag' effect in Australia. Before the pandemic in 2019 political trust had reached record lows. The closely related measure of confidence in government shows that there has been a huge improvement in citizen attitudes towards government during the pandemic. Differentiating the data by state shows that Victorians lost some confidence in the state but not the federal government, at the time of the second wave of COVID-19 in Victoria. This is consistent with the state government's responsibility for hotel quarantine, which was the source of the outbreak in July 2020. In Queensland, the data shows that confidence in Annastacia Palaszczuk's government increased when she closed the border to New South Wales (NSW) and Victoria in August 2020.

Figure 23: COVID-19 and confidence in government in Australia

While 2020 started with record low levels of trust, the COVID-19 pandemic increased confidence in government and support for incumbents – as a result of voters rallying round the flag at a time of crisis. This increased support for incumbents provides an electoral advantage. Speculation about the possibility of an early election in Australia stems from the high levels of support the government enjoys in early 2021 and the potential to capitalise on that.

Conclusion

To sum up, the AES long-term trends show an increasingly volatile electorate. Old divisions such as class and gender are changing over time. The gap between younger and older voters has never been greater, and today's generation of young people are much further to the left than their predecessors. Partisanship has reached record lows. All this contributes to greater unpredictability for electoral politics in Australia.

Following record levels of disaffection in 2019, the COVID-19 crisis, and Australia's relative success in handling the crisis as of February 2021, has ushered in a tremendous boost in support for incumbent governments. The crisis has become the salient issue at the expense of other priority areas. We could expect that continuing support for the current government is conditional on its handling of health and economic dimensions of the COVID-19 crisis.