Interrogating the popularity of direct democracy among European national publics

ABSTRACT

Research to date into the drivers of public support for participative democracy has offered limited insights into different national contexts. The present article addresses this gap in the literature. It empirically examines the sway of citizens’ appraisal of the state of democracy in the country, their political interest and personal socio-economic positions as well as the objective national socio-economic environment on public attitudes towards direct democracy. Using European Social Surveys (ESS) data, the results in this paper demonstrate that citizens’ political interest explains more about the public attitudes towards direct democracy than citizens’ assessment of the performance of the political regime, institutions and actors, and education. They also show that individuals in countries with integral proportional representation voting systems are more likely to support the use of referendums for key issues than those in countries without such systems. Education and income are found to have only a relatively modest impact on this support. The findings finally suggest that apart from a country’s level of press freedom, no other objective national socio-economic indicator tends to sway popular support for participative democracy. In the light of these findings, implications are drawn.

KEYWORDS:

• participative democracy
• public opinion
• political interest
• press freedom
• integral proportional representation voting systems
• education
• income
• public satisfaction with the political system

ACKNOWLEDGEMENTS

The author thanks the reviewers for making suggestions that improved the article.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author.

Notes

1 One of the key requests of this movement is the referendum on popular initiative.

2 Throughout, we use the terms ‘direct democracy’, ‘participative democracy’ and ‘referendums’ interchangeably. Although the first two terms can involve more than the mere use of referendums, the latter tend to be construed as an important element of both direct democracy and participative democracy. A significant part of the research literature on public attitudes towards referendums uses them likewise for this very reason. Direct democracy should be interpreted here as a system in which people decide on policy issues directly by means of referendums and initiatives (for a description of this system, see Matsusaka, 2005).

3 Radical parties such as France Unbowed/La France Insoumise (LFI) and National Rally/Rassemblement national (RN) (France), Alternative for Germany/Alternative für Deutschland (AfD) (Germany), 5 Star Movement/Movimento 5 Stelle (M5S) and the League/La Lega (Italy), VOX (Spain), and the Austrian Freedom Party/Freiheitliche Partei Österreichs (FPÖ) (Austria).

4 The latest research (Schuck & De Vreese, 2015; Werner et al., 2020; Rose & Wessels, 2021; Bowler & Donovan, 2019) has started to address this aspect.

5 A referendum (control) variable is used in the analysis. This enables us to see whether the findings hold controlling for the related variable and if the related control variable has its own unique effects on public approval of participative democracy. As shown in Table 3, the referendum variable is not statistically significant. This means that citizens in countries where referendums are frequently organized (in other words, Switzerland) are not more (or less) supportive of direct democracy than citizens in other countries.

6 This being said, being on the winning or losing side of a group vote or even the possibility of winning or losing the argument through the group decision-making (referendum for political/policy issues) could in itself sway individual support for group decisions or referendums (Werner, 2020).

7 A research limitation in these studies is that they do not integrate in the empirical analysis the potential effect of the objective socio-economic and political (electoral) context on public support for referendums. Bowler and Donovan (2019) incorporated a few socio-economic variables (namely unemployment and Gini index (inequality)) and only trust in politicians in their analysis. Our analysis here is more comprehensive.

8 Switzerland, Norway and Iceland are not members of the EU. However, many EU laws apply to them. Switzerland, Norway and Iceland are, for example, part of the single European market and signatories of the Schengen agreements.

9 Two studies (Rose & Wessels, 2021; Ferrin & Kriesi, 2016) have integrated a dummy (household) income variable into their analysis (and, moreover, have tested it with a binary dependent variable). Such variable operationalization (lowest or highest income quintile against the rest) represents a loss of valuable information for the analysis. This situation is remedied here, and two income-related variables are included in the analysis.

10 Radical political parties such as M5S, Brothers of Italy/Fratelli d’Italia (FdI) and La Lega (Italy), LFI and RN (France), FPÖ (Austria), VOX (Spain), the UK Independence Party (UKIP) (UK), Danish People's Party (Dansk Folkeparti, DF) (Denmark), AFD (Germany), Kukiz ’15/Kukiz ’15 (K ’15) (Poland), Jobbik Movement for a Better Hungary/Jobbik Magyarországért Mozgalom (Jobbik) (Hungary) and Sweden Democrats/Sverigedemokraterna (SD) (Sweden). Mainstream parties or factions within them such as The Republicans/Les Républicains (LR) (France), the Labour and Conservatives parties (UK), Go Italy/Forza Italia (FI) (Italy), ÖVP (Austria), the Law and Justice/Prawo i Sprawiedliwość (PiS) party (Poland), and Fidesz (Hungary).

11 The EU is considered as a key exponent of globalization and as such is seen as pursuing a free trade policy stance.

12 Citizens consider press freedom as an essential, defining aspect of democratic societies (Ferrin & Kriesi, 2016).

13 As one examines how the objective national socio-economic context shapes individual differences in attitudes toward direct democracy, a two-level hierarchical linear model is employed. This method is appropriate since the concern is with explaining variation at both the individual and national levels. A multilevel model enables us to explore causal heterogeneity and test the generalizability of findings across different national contexts (Steenbergen & Jones, 2002). This model also corrects for the dependence of observations within countries – intra-class correlations – and makes adjustments both within and between the parameter estimates for the clustered nature of the data (Snijders & Bosker, 1999).

14 Such as M5S and Lega (Italy) and LFI and RN (France). The referendum popularity among the masses exceeds the political weight of these political movements, but the latter have contributed to publicize it.

15 Similar results were obtained with the ‘satisfaction with the way democracy works (in the country)’ variable, 0.172** (0.013). The latter was used in earlier research.

16 Due to space limitations, the full results are not presented here for the models integrating the disproportionality index. These are, however, available from the author upon request.

17 This may also help us to clarify the divergent statistically insignificant and significant (income) results in earlier research (Rose & Wessels, 2021; Ferrin & Kriesi, 2016).

18 This is especially true about the EU markets’ opening to non-EU competition and the level of government spending.

19 In a large part of the extant literature on models of democratic, political decision-making, direct democracy is often opposed to representative democracy (Lijphart, 2012; Barber, 2004). In some quarters, the decision to organize a referendum on Brexit and abide by its results has thus even been portrayed as an infringement of the British constitution and its model of parliamentary democracy. However, the confrontation between direct democracy and representative democracy is not necessarily unavoidable (Rose, 2020). Current research (Font et al., 2015; Ganuza et al., 2017), moreover, shows that while citizens want to be directly involved in policymaking, they also tend to support at the same time representative democracy. Combining both direct democracy and representative democracy, though, requires a willingness from both politicians and the electorate to share political power on actual policymaking and a process model that can make this happen in a complex, EU multilevel system.