Previous Next

Double Disillusion

10

Changing Leaders, ‘Mediscare’ and Business as Usual: Electoral Behaviour

Clive Bean

In some respects, the 2016 Australian federal election had much in common with its recent predecessors. A leadership coup saw the sitting prime minister, Tony Abbott, ousted from office, by Malcolm Turnbull, months prior to the election. A bitter campaign between the major political parties, and with prominent participation by minor parties and Independents, culminated in a knife-edge result before the government was finally returned with the barest of majorities in the House of Representatives and without coming close to a majority in the Senate.

For much of the campaign, the Liberal–National Party Coalition government appeared to have an edge over the Labor Party opposition, but this advantage dissolved nearer to the election, as the campaign intensified and Labor adopted methods such as the so-called ‘Mediscare’ tactic. On the eve of the election, voters were targeted through their mobile telephones with text messages warning of the demise of the government health scheme, Medicare, under a re-elected coalition government, that appeared to come from Medicare itself (see e.g. Gillespie 2016). The government had made itself vulnerable to this claim by announcing during the campaign that it was contemplating outsourcing aspects of the administration of Medicare, and in so doing opened the door to claims by the Labor Party that the government would privatise the popular scheme. Labor introduced Medicare, a revamp of the original Medibank, in 1984 (Boxall and Gillespie 2013) and has always promoted itself as the champion of the scheme. Although the government argued very strongly that it had no plans to privatise Medicare, the damage done by loosely worded policy plans could not be reversed sufficiently to neutralise the opportunity that had been presented to the opposition.

This chapter explores sample survey data from the 2016 Australian Election Study (AES) to assess the impact of voter perceptions of the Medicare factor, leadership change and other issues pertinent to the 2016 federal election. The 2016 AES was conducted by Ian McAllister, Clive Bean, Juliet Pietsch, Rachel Gibson and Toni Makkai (McAllister et al. 2016; see also AES n.d.), using a grant from the Australian Research Council. The AES is a national post-election survey of voting and electoral behaviour, using a self-completion questionnaire that, in 2016, could be filled out either in hard copy and mailed back, or online. The sample comprised a combination of a systematic random sample of enrolled voters throughout Australia, drawn by the Australian Electoral Commission, and a sample drawn from the Geo-Coded National Address File. Non-respondents were sent several follow-up mailings and the final sample size was 2,818, representing a response rate of 23 per cent. The data were weighted to reflect population parameters for gender, age, State and vote, giving a final weighted sample size for electoral analysis of 2,711.

The voter and the campaign

Evidence across time has shown that many electoral trends are cyclical rather than secular, with indicators rising and falling according to how closely fought the particular election in question may be (see e.g. McAllister 2011). In the recent past, for example, the 2007 federal election, when the Labor Party under Kevin Rudd ousted the long-time Coalition government led by John Howard, was an election that captured the attention of voters in the way those either side of it did not (Bean and McAllister 2009, 2012). Yet, there has also been a sense that since that watershed, voter attitudes towards politics in Australia have been displaying greater levels of disaffection and disconnection than ever before (Cameron and McAllister 2016; McAllister 2011).

Table 10.1 displays an array of AES data focusing on voter orientations towards the election campaign in 2016. To provide appropriate context, it also shows the equivalent figures dating back five federal elections, to 2004 (for details of these earlier surveys, see Cameron and McAllister 2016). The table shows in the first line that 30 per cent of election study respondents reported taking a good deal of interest in the 2016 election campaign, which is well down on the level at the aforementioned 2007 election, but only slightly down on the 2010 and 2013 elections, and indeed the same as that recorded in 2004. However, fewer people cared about which party won the election in 2016 than in any of the previous four elections and the numbers paying attention to the election campaign through the traditional media were as low, or lower. By contrast, attention to politics through the new media continues to rise so that in 2016, for the first time in Australia, slightly more people indicated that they followed the election campaign through the internet than through radio. It is quite conceivable that, by the time of the next federal election, the internet may have become more heavily used for these purposes than print newspapers and second only to television. Table 10.1 also shows that viewing of the major debate between the party leaders was well down in 2016 and, interestingly, barely a fifth of the sample thought that the prime minster, Turnbull, performed better in the debate than his Labor Party rival, Bill Shorten.

Table 10.1. Voter engagement with the election campaign, 2004–16 (percentages)

2004

2007

2010

2013

2017

Took ‘a good deal’ of interest in the election campaign overall

30

40

34

33

30

Cared ‘a good deal’ which party won

72

76

68

68

65

Paid ‘a good deal’ or ‘some’ attention to the campaign

     –    in newspapers

57

61

62

50

50

     –    on television

69

77

77

70

68

     –    on radio

44

50

48

45

45

     –    on the internet

-

16

29

34

46

Watched the televised leaders’ debate

35

46

47

32

21

Thought prime minister performed better in the debate

25

13

37

30

21

Source. Constructed by author from data in AES 2004 (n=1,769), 2007 (n=1,873), 2010 (n=2,061), 2013 (n=3,955) and 2016 (n=2,711).

Further measures of voter orientations towards the election are contained in Table 10.2. The 50 per cent of 2016 AES respondents saying that it was only from around the start of the election campaign, or later, that they definitely decided how they would vote is easily the highest in more than a decade. However, it is a figure that has previously been equalled, or even surpassed, for example in the late 1990s (Bean and McAllister 2000). Similarly, over a third of the electorate said that they seriously thought of giving their first preference vote to another party and only 40 per cent said that before this election they had always voted for the same party. These latter two figures represent a new high and low for these respective measures in Australia since relevant data first started to be collected.

Table 10.2. Volatility and partisanship, 2004–16 (percentages)

2004

2007

2010

2013

2016

Decided definitely how to vote during campaign period

39

29

47

41

50

Seriously thought of giving first preference to another party in the House of Representatives during election campaign

25

23

29

30

34

Always voted for same party

50

45

52

46

40

Identifier with one of the major parties

77

77

78

73

67

Not a party identifier

16

16

14

17

20

Very strong party identifier

21

25

19

21

21

Source. Constructed by author from data in AES 2004 (n=1,769), 2007 (n=1,873), 2010 (n=2,061), 2013 (n=3,955) and 2016 (n=2,711).

Probably the most interesting data in Table 10.2, however, are to be found in the last three lines in the table, relating to party identification. The notion that most members of the electorate align themselves on an ongoing basis with one or other of the major political parties has long been a cornerstone of Australian politics (Aitkin 1982; McAllister 2011), even more so than for other comparable democracies. While patterns of decline in partisanship have been clear elsewhere (Dalton 2008; Dalton and Wattenberg 2000; Webb, Farrell and Holliday 2003), the expectation of similar trends in Australia has received little support from the evidence. Although assessments that such a decline had started or was imminent began to appear many years ago (Chaples 1993; Charnock 1997; Smith 1998), it is only in very recent times that sound supporting evidence has emerged. In 2013, the proportion of electors identifying with one of the major parties dipped to below three quarters of the electorate; in 2016, it had fallen another 6 percentage points to two thirds of the voting public. Based on the AES, major party identification has thus declined by more than 10 per cent over the last two Australian federal elections. At the same time, the number of citizens who claim to have no political party identification has now risen to a fifth of the electorate. Although the proportion of strong identifiers among those who still align with a party remains steady at 21 per cent, the complexion of the Australian electorate is now significantly different from the days when party identification stood at around 90 per cent (Aitkin 1982).

Social background and the vote

Now we turn to examine the impact of sociodemographic factors on the vote at the 2016 federal election. Social structure as an electoral influence has been less prominent in recent decades, although sometimes variable (McAllister 2011). The first variable in Table 10.3, gender, is a case in point. Over time, the so-called gender gap had virtually disappeared, but in 2010, when Australia’s first female prime minister, Julia Gillard, contested the election, it reappeared, albeit with its direction reversed (Bean and McAllister 2012). In 2013, the gender gap closed up again (Bean and McAllister 2015), but in 2016 we see a pattern similar to that of 2010 in which women are more likely to support Labor than men (by 7 per cent) and also more likely to vote for the Greens than men (by 4 per cent). One result of the gender pattern in 2016 is that women were equally likely to support the Coalition or Labor, whereas men were much more likely to support Labor than the Coalition.

Table 10.3. Vote by sociodemographic indicators in 2016 (percentages)

Lib.–Nat.

Labor

Greens

Other

(N)

Gender

Male

47

31

8

14

(1,187)

Female

38

38

12

12

(1,223)

Age Group

Under 25

33

37

22

8

(235)

25–44

36

35

15

14

(794)

45–64

41

38

7

15

(799)

65 and over

57

28

3

12

(556)

Birthplace

Australia

41

34

11

14

(1,890)

British Isles

42

33

10

14

(175)

Europe

54

34

6

6

(86)

Asia

41

46

6

7

(139)

Other

50

33

14

3

(111)

Region

Rural

45

31

7

17

(633)

Urban

41

36

12

12

(1,755)

Religious denomination

Catholic

45

39

5

11

(557)

Anglican

49

31

6

14

(437)

Uniting

54

30

3

13

(229)

Other

42

31

6

21

(347)

No religion

33

36

20

11

(835)

Church attendance

At least once a month

45

29

4

22

(378)

At least once a year

46

36

9

10

(409)

Never

38

37

12

13

(1,146)

Source. Constructed by author from data in AES 2016 (n=2,711).

The age variable reaffirms the Coalition’s electoral advantage among the most senior members of the electorate. Some 57 per cent of those aged 65 and over voted for the Coalition parties, compared to only 33 per cent of those aged under 25. The Labor Party, on the other hand, could garner barely more than a quarter of the votes of the over 65s. But rather than being the beneficiaries of the Coalition’s weakness among the young, Labor saw many of those votes go to the Greens, who secured over 20 per cent support from the under 25s. The young, however, are somewhat disinclined to support the ‘Other’ minor parties. In Tables 10.3 and 10.4, the age group data also show that the level of support for the Liberal–National parties from the over 65s is greater than for any other sociodemographic group.

The immigrant vote has and continues to gain a good deal of attention in research that focuses on Australian politics (e.g. Jiang 2016; see also Jupp and Pietsch, Chapter 29, this volume). Patterns of party support by birthplace have changed over the decades since the main wave of immigrants to Australia in the middle of the twentieth century (McAllister 2011). Of particular interest in the twenty-first century are the partisan leanings of the increasing number of voters from Asia within the Australian electorate. The third section of Table 10.3 shows that in 2016 voters born in the British Isles largely mirrored the distribution of party support displayed by the Australian-born, while those from continental Europe disproportionately supported the Liberal–National parties. The Asian-born, by contrast, were more inclined to vote Labor, while both European and Asian voters tended to steer away from the minor parties. Rural voters favoured the Liberal–National parties over Labor in 2016, as they typically do, although the differences are relatively mild. The pattern of differences among the minor parties observed in 2013 (Bean and McAllister 2015) persisted in 2016, with rural voters less likely to vote for the Greens and more likely to vote for ‘Other’ minor parties and candidates.

It can no longer be said that the Labor Party is the bastion of Catholic support since, for the third election in a row, more Catholics voted Liberal–National than Labor. It remains the case, nonetheless, that Labor receives more support from Catholics than from any other denominational grouping, while the Coalition is most favoured by Protestants. For the Greens, its ‘religious’ champions are voters disavowing any religious adherence, while it is those who belong to ‘other’, nonmainstream, religious groups who most favour the Other minor parties. Church attenders support the Coalition and disproportionate numbers of frequent attenders also vote for Other minor parties, whereas the vote for Labor and especially the Greens tends to sag amongst such voters. Those who never attend places of religious observance are more evenly distributed between the major parties.

Turning now to Table 10.4, we see that the most noteworthy effect for education lies in the strong support for the Greens conferred by those with a university degree compared to those who do not have this level of education. This, of course, is one of the enduring features of the Greens’ support profile (Bean and McAllister 2002, 2009, 2012, 2015). Occupation is no longer the cornerstone of party support in Australia it once was. In 2016, the index of class voting (calculated as the Labor vote among manual voters minus the Labor vote among nonmanual voters) comes in at 8 per cent. While by no means high, this figure is not as low as at some previous elections in the last quarter of a century (McAllister 2011) and is slightly greater than the 6 per cent registered at the last election (Bean and McAllister 2015). So the traditional class/party divide, which sees Labor favoured by blue-collar workers and the Coalition by white-collar workers, continues in Australia into the twenty-first century, although only in muted form compared to half a century ago.

Table 10.4. Vote by socioeconomic indicators in 2016 (percentages)

Lib.–Nat.

Labor

Greens

Other

(N)

Education

No postschool qualifications

43

38

8

11

(577)

Nondegree qualifications

43

33

7

16

(910)

University degree

40

34

15

11

(892)

Occupation

Manual

37

40

8

15

(681)

Nonmanual

44

32

12

12

(1,454)

Employment

Self-employed

56

22

6

16

(331)

Government employee

38

39

11

11

(519)

Trade union membership

Union member

23

54

11

12

(420)

Not a union member

46

31

10

13

(1,948)

Source. Constructed by author from data in AES 2016 (n=2,711).

Clearer patterns of political division are evident by employment status: the self-employed were much more likely to give their vote to the Liberal–Nationals; Labor, at least in relative terms, fared well among public sector employees. Trade unionists are traditional supporters of the Australian Labor Party and the 2016 data reinforce this alignment. After an apparent dip in union support in 2013 (Bean and McAllister 2015), Labor again received more than half the votes of union members this time while the Coalition gained less than a quarter.

Perceptions of parties and leaders

Closer to the forefront of voters’ electoral decisions are their views about the political parties and their leaders. Given some of the evidence we have already seen concerning voter detachment from the political parties, it is not surprising that the data in Table 10.5 paint a fairly negative picture of the parties and leaders generally. On a scale where 10 represents an extreme like of a party or leader, zero represents an extreme dislike and 5 is a neutral or middle score, none of the parties or leaders managed to register a score on the positive side of the mid-point, a similar scenario to 2013 (Bean and McAllister 2015). The Labor Party had a slightly better mean rating than the Liberal Party (4.9 versus 4.8). There is a different story for the leaders, however. Prime Minister Turnbull scored slightly better than his party and substantially better than the leader of the opposition, Shorten, who, on a score of 4.2, was well behind his party. While clearly more popular than his rival, however, even Turnbull had a mean rating of less than 5 (4.9). At most previous elections since relevant data began to be collected, one or other or both of the major party leaders has recorded a score on the positive side of the ledger, in other words greater than 5 (McAllister 2011), but this has changed in the last three elections. Returning to the data for 2016 in Table 10.5, the National Party, at 4.4, was a little more popular than its leader, Barnaby Joyce (4.1), while Richard Di Natale, the Greens leader, rated slightly above his party: Di Natale scored just above 4 while the Greens Party scored just below 4.

Table 10.5. Party and leader evaluations in 2016 (means on 0–10 scale)

Party

Mean

Std Dev.

Leader

Mean

Std Dev.

Liberal

4.8

3.1

Malcolm Turnbull

4.9

2.8

Labor

4.9

2.9

Bill Shorten

4.2

2.8

National

4.4

2.7

Barnaby Joyce

4.1

2.7

Greens

3.9

3.0

Richard Di Natale

4.1

2.4

Tony Abbott

3.6

3.1

Source. Constructed by author from data in AES 2016 (n=2,711).

Probably the most interesting figure in Table 10.5 is the score for former prime minister Tony Abbott, which at 3.6 is low by any standards (and, importantly, much lower than that for Turnbull) and provides some vindication of the decision of the Liberal Party parliamentary party to replace him in September 2015. Interestingly, a question in the AES about whether voters approved or disapproved of the move to replace Abbott revealed that a narrow majority disapproved (51 per cent, to 49 approved). This finding stands in contrast to when the Labor Party replaced Rudd with Gillard before the 2010 election, after which some three quarters of the AES sample disapproved (Bean and McAllister 2012).

Table 10.6 looks in more detail at the images of the two main leadership rivals, Turnbull and Shorten. When the AES respondents were asked how well a set of nine attributes described each leader, Turnbull outscored Shorten on most, as the sitting prime minister tends to. Thus, 84 per cent said that the term ‘intelligent’ described Turnbull extremely or quite well, 78 per cent rated him as knowledgeable, 67 per cent as competent and 64 per cent as sensible. These attributes were also perceived to be among Shorten’s best, although his scores were substantially lower on each. In fact, the only characteristic on which Shorten did better than Turnbull was on the quality of compassion, for which 55 per cent said that described Shorten extremely or quite well, compared to only 44 per cent for Turnbull. Of more concern would be that neither leader was rated strongly for being trustworthy, inspiring or honest. Shorten’s scores on each of these, however, were clearly lower than Turnbull’s and, in particular, not even a quarter of the sample saw Shorten as inspiring. These assessments of the major party leaders’ personal qualities serve to reinforce the summary ratings in Table 10.5, showing Turnbull to be more well regarded by the electorate—or, more accurately, less poorly regarded—than Shorten.

Table 10.6. Perceived leadership attributes of Malcolm Turnbull and Bill Shorten in 2016 (percentage saying attribute describes leader extremely well or quite well)

Quality

Malcolm Turnbull

Bill Shorten

Intelligent

84

70

Compassionate

44

55

Competent

67

54

Sensible

64

56

Provides strong leadership

52

48

Honest

45

37

Knowledgeable

78

64

Inspiring

38

24

Trustworthy

39

33

Source. Constructed by author from data in AES 2016 (n=2,711).

Issue agendas

While the issue agendas of the political parties tend to dominate the election campaign, it is increasingly becoming apparent that voters have their own issue agendas (Bean and McAllister 2015; Goot and Watson 2007; McAllister, Bean and Pietsch 2012). For example, voters repeatedly emphasise their concerns about policies relating to health, education, taxation and the economy in general, irrespective of the extent to which the major parties focus their campaigns on these issues. Interestingly, the first two of these, health and education, consistently tend to work in favour of the Labor Party, while taxation and the economy generally favour the Coalition.

The data in Table 10.7 suggest that not too much was different in 2016. On the surface, taxation appeared to be the exception, for it did not feature strongly as a concern for voters, nor did it generate a clear distinction between the major parties. But health, education and management of the economy all featured strongly in importance and in differentiating Labor from the Coalition. As it so often is, health and Medicare were the number one concern for voters, with 70 per cent of the AES sample describing them as extremely important when they were deciding how to vote. Education came next, with 60 per cent saying it was extremely important, and management of the economy was close behind in third place, with 57 per cent calling it extremely important. None of the other issues in the list presented to the respondents had half the sample saying it was extremely important, with the next in line being the related issues of immigration and refugees and asylum seekers, on 46 and 45 per cent respectively. The environment, government debt, superannuation, and only then taxation, came next, with global warming coming last on the list of concerns to voters, attracting only 34 per cent of respondents saying it was extremely important.

When we look at which issues mattered more or less to which party supporters, we see that health and education were of considerably greater concern to Labor voters than to Liberal–National voters, while management of the economy was of much greater concern to the latter than the former. Other issues that divided the major party voters were government debt, which concerned more than twice as many Coalition voters as Labor, the environment, which concerned more than twice as many Labor voters as Coalition, and global warming, which fewer than a fifth of Liberal–National voters regarded as extremely important compared to over two fifths of Labor voters. Nearly three quarters of Greens supporters, expressed concern about global warming, and even more did about the environment, naturally enough. Greens voters were also concerned about health, education and asylum seekers, while being relatively unconcerned about the economy, government debt, superannuation and in particular taxation. Supporters of ‘Other’ minor parties largely mirrored the overall sample in their concerns, with the biggest discrepancies being the greater level of concern they expressed on immigration and refugees and asylum seekers.

Table 10.7. Importance rating of election issues (percentage describing issue as extremely important) and party differential (percentage saying Liberal–National closer on issue minus percentage saying Labor closer) in 2016

Importance

Issue

All voters

Lib.–Nat. voters

Labor voters

Green voters

Other voters

Party differential

Taxation

37

41

39

16

38

+2

Immigration

46

50

43

40

52

+12

Education

60

50

73

65

61

–18

The environment

42

26

54

77

42

–14

Government debt

38

58

25

19

37

+28

Health and Medicare

70

60

84

68

73

–18

Refugees and asylum seekers

45

42

45

62

50

+15

Global warming

34

18

43

74

35

–12

Superannuation

38

42

38

18

38

+1

Management of the economy

57

76

48

20

57

+27

Source. Constructed by author from data in AES 2016 (n=2,711).

Immediately following the question on the importance of issues, the AES respondents were asked which of the major parties’ policies came closest to their own views on each of the issues. The far right column of Table 10.7 shows the calculation of the difference in the percentages opting for each party. As can be seen, the largest differentials favoured the Coalition, with 28 per cent more voters favouring it than Labor on government debt and 27 per cent more favouring it on management of the economy. The government parties also had an advantage on refugees and asylum seekers and immigration. The Labor Party was clearly favoured on health and Medicare, education, the environment and global warming.

Having a larger percentage of the electorate favouring one party over the other on a particular issue does not, however, necessarily translate into that issue influencing the overall vote in that party’s favour. The issue of government debt, for example, is a case in point. While the weight of voter opinion clearly favoured the Coalition on the issue, the level of concern expressed by voters was quite low. It is thus quite possible that this issue did not significantly affect the way people voted, meaning that the apparent advantage to the Coalition would come to nothing. In order to test the effect of this and other issues on voting decisions, we turn to multivariate analysis in the final stage of the investigation in the next section.

Which factors mattered most?

Multivariate analysis allows us to see which of the factors examined earlier in the chapter significantly influenced the vote in the 2016 Australian federal election, taking account of all the other potential influences. Particulars of the method employed are given in the Appendix, but in essence the multivariate procedure reveals which of the sociodemographic factors, party identification, leadership ratings and issue variables assessed above had a statistically significant influence on voting behaviour. Table 10.8 displays the results, with only those variables registering statistically significant effects at the .01 level being shown.

Table 10.8. Factors influencing the vote in 2016 (multiple regression)

Unstandardised regression coefficient

Standardised regression coefficient

Age

.001

.05

Party identification

.57

.53

Malcolm Turnbull

.15

.09

Bill Shorten

.11

.07

Taxation

.13

.10

Education

.09

.07

Health and Medicare

.08

.07

R-squared

.69

Note. Entries in the table are statistically significant at p < .01 or better. Further methodological details can be found in the Appendix.

Source. Constructed by author from data in AES 2016 (n=2,711).

Of the 29 variables included in the multivariate equation, only seven proved to have a significant impact according to the criteria used. With all the other variables controlled, age was the only sociodemographic factor to record a significant effect. As the data in Table 10.3 suggested, older voters are more likely to give their electoral preference to the Liberal–National parties than to Labor. Turning to the next variable in the table, we see that even amid the signs of declining partisan affiliation with the major Australian political parties that was revealed earlier in the chapter, party identification nonetheless continues to dominate the electoral landscape in Australia. The unstandardised regression coefficient for party identification in Table 10.8 shows that, net of all other factors, Liberal–National identifiers were 57 per cent more likely to vote for the Coalition than Labor identifiers. This is by far the biggest influence in the model.

As evidence from previous elections would predict, the two major party leaders both recorded significant effects on the vote in 2016. Research has shown that the leaders of the two main party groupings repeatedly feature as influences on Australian electoral behaviour (Senior and van Onselen 2008), albeit with the size of their effects varying depending on the individual personalities involved. In 2016, the leader effects were reduced compared to recent elections (Bean and McAllister 2012, 2015), but were still far from trivial. For Turnbull, those who most liked him were 15 per cent more likely to vote for the Coalition than those who most disliked him, while for Shorten, voters who regarded him most positively were 11 per cent more likely to vote Labor than those who regarded him most negatively.

Three issues recorded statistically significant effects—namely taxation, education and health and Medicare. While the latter two will come as no surprise, the appearance of taxation is more intriguing. The bivariate evidence in Table 10.7 suggested that voters were not terribly concerned about taxation in this election. On the basis of the multivariate evidence, however, it would appear that for those who did regard it as an important issue it really did matter, to the extent that those who said taxation was extremely important and that the Liberal–National Coalition’s policies were closer than Labor’s policies to their own views were 13 per cent more likely to vote for the Coalition. While, of the four consistent policy influences listed earlier, economic management did not record a significant effect on this occasion, once again, taxation, health and education did. In this instance, the effect for taxation is presumably attributable to the ongoing debate about options for tax reform being considered by the government during the latter part of the electoral cycle (see e.g. Barnes 2015).

To work out how much these issues, and the party leaders, really mattered for the outcome of the election, we need to examine the impact of each one on the balance of party support. While none of these effects is very large, we saw in Table 10.7, for example, that the balance of opinion towards health and education favoured Labor in both instances and this could make a difference to the election outcome at the margin. If the effect of each variable on individual voters (the regression coefficient) is combined with the net balance of voter opinion (derived from the mean of the variable), the overall effect of each variable on the balance of the two-party vote can be estimated. Details of how these calculations are made are contained in the Appendix.

As we saw in Table 10.5, neither major party leader was popular with the electorate, but Shorten was markedly less so than Turnbull. The relevant calculations show that Turnbull was a small drag on the Coalition vote, to the extent of around 0.15 per cent, while Shorten dragged the Labor Party vote down by nearly 0.9 per cent. The net effect for the leaders was thus about three quarters of 1 per cent in favour of the Coalition. Perhaps more intriguingly, it is possible to make a hypothetical assessment of the impact of replacing Abbott with Turnbull. Assuming the same level of effect on individual voting decisions if Abbott had remained, and taking account of Abbott’s much lower rating in Table 10.5 (3.6 compared to Turnbull’s 4.9), the benefit to the Coalition vote of replacing Abbott with Turnbull was just under 2 per cent. Given the closeness of the final result (a two-party preferred margin of only 0.72 per cent), one clear implication of this calculation is that if Abbott had remained prime minister, the Coalition would have lost the election, all other things being equal.

Turning to the policy issues, taxation narrowly favoured the Coalition while education and health both favoured Labor to a considerable degree. Because the balance of opinion on taxation was not divided strongly, this issue resulted in only a small benefit to the Coalition, of around 0.25 per cent. By contrast, education and health each swung the electoral balance towards Labor by around 0.8 per cent. When these are added together, even with the counteracting effect of the taxation issue, the net electoral benefit from campaign issues in the 2016 Australian federal election amounted to over 1.3 per cent of the vote in Labor’s favour. Considering how close the election was in the end, the benefit from these issues could be argued to be what allowed Labor to push the Coalition to the brink of defeat.

Conclusion

The evidence that the issue of health and Medicare played a significant role in bringing Labor so close to taking office away from the Coalition brings us back to the introductory discussion about the so-called ‘Mediscare’ campaign tactic employed by Labor. Unfortunately, however, the results of this analysis do not allow us to say whether it was the ‘Mediscare’ strategy that caused the health issue to have the electoral impact it did. Previous research showing that the issue of health has had similar effects in other recent Australian elections (Bean and McAllister 2009, 2015) means that it is entirely possible that the health effect in 2016 had more to do with the issue agendas of the electorate than the campaign tactics of one of the major political parties. Whatever the basis, the issue of health mattered in the 2016 election, as did the change of leadership in the Liberal Party. But there was also a strong sense of ‘business as usual’, with an intensely fought contest leading to a very narrow victory, as has become a feature of recent Australian federal elections.

Appendix

The equation in Table 10.8 was estimated by ordinary least squares regression analysis with pairwise deletion of missing data. The dependent variable, first preference vote for the House of Representatives in the 2016 federal election, was scored 0 for Labor, 0.5 for minor parties and 1 for Liberal–National. Party identification was likewise scored 0 for Labor, 0.5 for minor parties or no party identification and 1 for Liberal–National. Apart from age, scored in years, all other independent variables were either 0–1 dummy variables or scaled to run from a low score of 0 to a high score of 1.

The issue variables were computed by combining the importance ratings with the party closer to the respondent, so that at one end those who rated the issue as extremely important and felt closer to the Labor Party on the issue were scored 0 and at the other end of the scale those who rated the issue as extremely important and felt closer to the Coalition on the issue were scored 1.

For both the issue and leader variables, the calculations showing the impact of these factors on the party balance were made by multiplying the unstandardised regression coefficient from the first column in Table 10.8 by the amount by which the mean of the variable deviated from the neutral point of 0.5. For taxation, the deviation was +0.02, for education it was –0.09 and for health and Medicare it was –0.10. In the case of the leadership variables, the deviation for Turnbull was –0.01, for Shorten it was –0.08 and for Abbott it was –0.14.

References

Aitkin, Don. 1982. Stability and Change in Australian Politics. 2nd edition. Canberra: Australian National University Press.

Australian Election Study (AES). n.d. Available at: www.australianelectionstudy.org/voter_studies.html

Barnes, Terry. 2015. ‘Tax reform: Stop talking and just get on with it’. The Drum, ABC, 4 November. Available at: www.abc.net.au/news/2015-11-04/barnes-tax-reform-stop-talking-and-just-get-on-with-it/6910724

Bean, Clive and Ian McAllister. 2000. ‘Voting behaviour’. In Marian Simms and John Warhurst (eds), Howard’s Agenda: The 1998 Australian Election. St Lucia: University of Queensland Press, pp. 174–92.

——. 2002. ‘From impossibility to certainty: Explaining the Coalition’s victory in 2001’. In John Warhurst and Marian Simms (eds), 2001: The Centenary Election. St Lucia: University of Queensland Press, pp. 271–86.

——. 2009. ‘The Australian election survey: The tale of the rabbit-less hat. Voting behaviour in 2007’. Australian Cultural History 27: 205–18. doi.org/10.1080/07288430903165360

——. 2012. ‘Electoral behaviour in the 2010 Australian federal election’. In Marian Simms and John Wanna (eds), Julia 2010: The Caretaker Election. Canberra: ANU E Press, pp. 341–55. Available at: press-files.anu.edu.au/downloads/press/p169031/pdf/ch27.pdf

——. 2015. ‘Documenting the inevitable: Voting behaviour at the 2013 Australian election’. In Carol Johnson and John Wanna with Hsu-Ann Lee (eds), Abbott’s Gambit: The 2013 Australian Federal Election. Canberra: ANU Press, pp. 411–24. doi.org/10.22459/AG.01.2015.24

Boxall, Anne-Marie and James A. Gillespie. 2013. Making Medicare: The Politics of Universal Health Care in Australia. Sydney: UNSW Press.

Cameron, Sarah M. and Ian McAllister. 2016. ‘Trends in Australian Political Opinion: Results from the Australian Election Study, 1987–2016’. Canberra: School of Politics and International Relations, The Australian National University.

Chaples, Ernie. 1993. ‘The Australian voters’. In Rodney Smith (ed), Politics in Australia. 2nd edition. Sydney: Allen & Unwin, pp. 319–32.

Charnock, David. 1997. ‘How secure is the Coalition’s majority? Partisan alignments at the 1996 federal election’. In George Crowder et al. (ed.), Australasian Political Studies 1997: Proceedings of the 1997 APSA Conference, volume 1. Adelaide: Department of Politics, Flinders University of South Australia.

Dalton, Russell J. 2008. Citizen politics: Public opinion and political parties in advanced industrial democracies. 5th edition. Washington, DC: CQ Press.

Dalton, Russell J. and Martin P. Wattenberg (eds). 2000. Parties without Partisans: Political Change in Advanced Industrial Democracies. Oxford: Oxford University Press.

Gillespie, Jim. 2016. ‘Labor’s “Mediscare” campaign capitalised on Coalition history of hostility towards Medicare’. The Conversation, 5 July. Available at: www.theconversation.com/labors-mediscare-campaign-capitalised-on-coalition-history-of-hostility-towards-medicare-61976

Goot, Murray and Ian Watson. 2007. ‘Explaining Howard’s success: Social structure, issue agendas and party support, 1993–2004’. Australian Journal of Political Science 42: 253–76. doi.org/10.1080/10361140701320018

Jiang, Liang. 2016. ‘Valence politics and immigrant voting in the 2013 Australian election’. Australian Journal of Political Science 51: 418–35. doi.org/10.1080/10361146.2016.1202890

McAllister, Ian. 2011. The Australian Voter: 50 Years of Change. Sydney: UNSW Press.

McAllister, Ian, Clive Bean and Juliet Pietsch. 2012. ‘Leadership change, policy issues and voter defection in the 2010 Australian election’. Australian Journal of Political Science 47: 189–209. doi.org/10.1080/10361146.2012.677002

McAllister, Ian, Toni Makkai, Clive Bean, Juliet Pietsch and Rachel Gibson. 2016. Australian Election Study, 2016: Study Documentation. Canberra: Australian Data Archive, The Australian National University.

Senior, Philip and Peter van Onselen. 2008. ‘Re-examining leader effects: Have leader effects grown in Australian federal elections 1990–2004?’ Australian Journal of Political Science 43: 225–42. doi.org/10.1080/10361140802035754

Smith, Rodney. 1998. ‘Australia: An old order manages change’. In John Kenneth White and Philip John Davies (eds), Political Parties and the Collapse of the Old Orders. Albany: State University of New York Press.

Webb, Paul, David Farrell and Ian Holliday (eds). 2003. Political Parties at the Millennium: Adaptation and Decline in Democratic Societies. Oxford: Oxford University Press.


Previous Next