Populists or nativist authoritarians? A cross-national analysis of the radical right

ABSTRACT

Populism is widely considered to be one of the most significant political phenomena of the past decade. Yet for all the scholarly and media attention it receives, how important it is in driving support for populist radical right parties is debatable. Scholars have long theorised that nativism and authoritarianism are likely to be equally if not more important than populism in driving support for populist radical right parties, but the empirical evidence to support this argument has been limited. We conduct a cross-national analysis on a representative sample of voters from the United States, United Kingdom and Australia (n = 4650) to test this proposition. We demonstrate that rather than populism, it is primarily nativism driving support for populist radical right parties. Populism is, therefore, shown to be less important than often suggested.

人们普遍认为民粹主义是过去十年最重要的政治现象。尽管受到学界和传媒的关注，民粹主义对极端右翼民粹政党有多大的支持却众说不一。学者一直有种理论：本土主义和威权主义对极端右翼民粹政党的支持等于甚至大于民粹主义。这种观点缺少实证数据的印证。我们根据来自美国、英国和澳大利亚的有代表性的样本（数目为4650））做了一项跨国分析以检验这个观点。根据我们的分析结果，主要是本土主义而非民粹主义支持了极端右翼民粹政党。民粹主义的影响不像过去以为的那样大。

KEYWORDS:

• Populism
• radical right
• voter attitudes
• cross-national survey-data

Disclosure statement

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

Notes

1 Despite this, Axelrod (1967) is said to be the first to have suggested this as a potential approach.

2 One recent exception to this was Gest, Reny, and Mayer (2017) who examined the attitudes of White voters in the US and UK and argue that nostalgic deprivation was a driver of support for the populist radical right.

3 In the appendix, we outline the question order and direction of questions, so that other researchers can understand our approach in more detail. In building on the work of Akkerman, Mudde, and Zaslove (2014), and to ensure that the question items were tapping populism, as opposed to its opposites – pluralism and elitism – we included questions which tapped these attitudes. For more on this discussion about the opposites of populism see Akkerman, Mudde, and Zaslove (2014) and Hawkins and Kaltwasser (2018, 3–4).

4 One recent exception is Vasilopoulos and Lachat (2017) who looked at voting for the Front National in France. The recent study of Aguilar and Carlin (2018), which set out to explore the relationship between an authoritarian predisposition and populist voting provides an interesting contrast with our approach.

5 While we focus on the relationship between attitudes, institutions and voting, we acknowledge that there are other forces at work in these democracies and that includes the role of media in shaping agendas.

6 It is worth noting that some scholars disagree with classifying UKIP as a populist radical right party (Mudde 2016, 15), however, a great many others do see it as a populist radical right party (Goodwin and Dennison 2018).

7 Referred to as One Nation from this point.

8 The party was aided further by the fact the quotas for election to the Senate in 2016 were halved as a result of both chambers being dissolved due the government calling a ‘double dissolution’ election.

9 In the Australian analysis, it is important to note that the ‘Coalition’ is made up of the Liberal Party of Australia and the Nationals who are in a formal coalition at the federal level and in a number of states, as well as having merged into a single force in the state of Queensland, the Liberal National Party. Due to the complexity of this relationship, especially in disaggregating results, we decided to pool them together for analytical purposes.

10 In this model, all of the items in our dataset were coded in the same direction, with responses to each question that reflected a populist, authoritarian or nativist response coded with a higher score and the alternatives a lower score. For instance, ‘immigrants take jobs away from people born in <their country>’, strongly agree was given a value of ‘4’ and strongly disagree ‘1’. Moreover, responses to how important it was ‘To limit the power of the police to protect citizen’s freedoms’ were coded ‘5’ for not at all important and ‘1’ for extremely important. Shown in Table 1, we restrict how items load onto each dimension in these models, with constraints placed on each of these so that they load onto the latent set of attitudes to which the literature indicates they belong, and constrains of zero on the others. While we place significant restrictions on dimensional loading, we reduce constrains on other parts of our models. As nativism and authoritarianism, or populist and nativist attitudes, may be related, there are good reasons to believe that estimates derived from a model measuring latent dimensions from survey items intended to capture these preferences will be correlated. We relax restrictions on our model allowing for correlation to exist between these dimensions.

11 For this descriptive analysis the survey weights provided by YouGov were used to ensure samples were representative of the Australian, British and American populations. Data were weighted by age, gender and region for Australia; age, gender, social class, region, education, 2017 general election, the EU referendum and political interest for the UK; and for the American sample, 2016 Presidential vote choice, gender, age, race and education. For the regression models fit below these weighting variables were incorporated as controls.

12 Penalised logistic regression was used for the UKIP model, as the small sample number of UKIP voters in the sample (N = 36). For a discussion on the problem of fitting logistic regression to rare events, see King and Zeng (2001).

Additional information

Notes on contributors

Glenn Kefford

Glenn Kefford is a Lecturer in the School of Political Science and International Studies at the University of Queensland and an Australian Research Council Discovery Early Career Researcher Award Fellow (2019–2021).

Shaun Ratcliff

Shaun Ratcliff is a lecturer in Political Science, United States Studies Centre, University of Sydney.