Government performance and dissatisfaction with democracy in Australia

ABSTRACT

Satisfaction with democracy has declined rapidly in Australia, reaching the lowest level recorded since the 1970s in 2019. Whereas Australian citizens used to be among the most satisfied democrats in the world, there is now evidence of widespread dissatisfaction. What explains this rapid decline in political support? Comparative studies emphasise the role of government performance, both political and economic, in shaping citizens’ attitudes towards the political system. This paper examines the role of government performance in shaping satisfaction with democracy in Australia, using Australian Election Study data from 1996 to 2019. The results demonstrate that frequent changes of prime minister, which a majority of voters disapproved of, and rising economic pessimism contributed to the decline of democratic satisfaction in Australia.

在澳大利亚，对民主的满意度正快速下滑，2019年达到1970年代以来最低水平。澳大利亚公民一向是世界上最心满意足的民主人士，而如今的证据却表明民主正大失人心。如何解释政治支持的快速式微呢？比较研究强调政府的政治、经济表现左右了公民对政治制度的态度。本文采用1996年至2019年澳大利亚选举研究的数据，考察了政府的表现如何造成人民对民主制度的失望。研究结果显示，为多数选民所不满的总理频繁更换，以及经济上日益悲观的情绪造成了澳大利亚民主满意度的下降。

KEYWORDS:

• Australian politics
• satisfaction with democracy
• political trust
• leadership
• political parties
• economic performance

Acknowledgements

Many thanks to the Australian Election Study teams 1987–2019 for collecting the data used in this article. Full details are available at www.australianelectionstudy.org. Thanks also to the University of Sydney’s School of Social and Political Sciences for supporting this research, in particular Professor Lisa Adkins. I am grateful to two anonymous reviewers from this journal for their constructive feedback.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes on contributor

Dr Sarah Cameron is a Research Fellow in the School of Social and Political Sciences at the University of Sydney and a contributor to the Australian Election Study. Her research focusses on elections, comparative political behaviour and Australian politics.

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 The research to date has found that retrospective congruence matters more than prospective congruence (Ferland 2017); and egocentric evaluations of congruence – the connection between an individual citizen’s ideology and that of their representative(s) – matters more than sociotropic congruence.

2 For an overview see Lewis-Beck and Stegmaier (2007).

3 Comparative data on satisfaction with democracy is sourced from the Comparative Study of Electoral Systems Module 4 (2018), fielded over 2011 to 2016. The average level of satisfaction with democracy in the 26 OECD countries incorporated is 57%, ranging from 13% (Greece) to 94% (Norway). See Cameron and McAllister (2019a, 15).

4 Bivariate correlations between relative leader popularity (whether the respondent liked the new leader better than the old leader) and approval of the leadership change are positive and reasonably strong: 2010: r = 0.39, n = 2165, p < 0.001; 2016: r = 0.52, n = 2698, p < 0.001; 2019: r = 0.54, n = 2131, p < 0.001. Leader popularity alone does not account for the differences as disapproval of the 2018 change was high (74%) even though Scott Morrison was evaluated more favourably than Malcolm Turnbull in the 2019 AES.

5 This is the figure based on votes not seats (Döring and Manow 2019).

6 In 2016 and 2019 the survey has also used a randomly selected sample from the Geo-Coded National Address File (G-NAF) (Cameron and McAllister 2019b, 150). Further details on methodology can be found at www.australianelectionstudy.org.

7 See Cameron and McAllister 2019b, 150.

8 Weights are used in 1993 and 2010 onwards. This corrects for subpopulations that were oversampled in particular surveys, and the weights also ensure representativeness in surveys where response rates have been lower (Cameron and McAllister 2019b, 150; Kohut et al. 2012). The AES technical reports show that even before weighting survey respondents are broadly representative on many characteristics including vote, gender, and state (Lethborg 2017, 29).

9 The minimum number of group level units suitable for multilevel modelling is a topic of debate (see Bryan and Jenkins 2015; Robson and Pevalin 2015, 26–27). The number of elections included here is towards the lower end, although Robson and Pevalin (2015) argue that even with a smaller number of groups it is preferable to use a multilevel model than not, as not to ignore group-level variance.

10 Data is not available for the 2013 leadership change (when Kevin Rudd replaced Julia Gillard) in the AES, so it cannot be incorporated in regression models. The question was asked in the ANUPoll (McAllister 2018), providing the descriptive results as shown in Figure 2.

11 In years that the Liberal-National Coalition won the election, those who voted for either the Liberal or National party are considered winners.

12 The small uptick in economic evaluations in 2019 was counterbalanced by higher levels of disapproval of the latest leadership change in 2018.