Asymmetric ratification standards and popular perceptions of legitimacy

ABSTRACT

The ratification of EU agreements is characterized by the application of different democratic procedures across member states. Building on the demoi-cratic theory of legitimate global governance, I argue that citizens benchmark their national procedure against highly visible direct democratic ratification votes held in other member states. If citizens experience unequal influence on EU decision-making, the perceived legitimacy of the EU regime erodes. I test this argument with a research design that combines a population-based survey experiment and a quasi-experiment. First, a survey experiment in Germany reveals that information about asymmetric ratification standards decrease fairness perceptions and satisfaction with EU democracy. Second, a natural experiment around the 2005 French vote on the EU constitutional treaty shows that the referendum decreased satisfaction with EU democracy in states with pending and indirect ratification. These findings suggest that asymmetric access of citizens to EU decision-making can decrease popular support for EU governance.

KEYWORDS:

• Direct democracy
• European Union
• experiment
• legitimacy
• ratification
• referendum

Acknowledgements

For helpful comments on previous version of this article, I am indebted to Frank Schimmelfennig, Stefanie Walter, Fabio Wasserfallen, and all my colleagues at the European Politics Research Group at ETH Zurich. I am also grateful for the excellent comments by the two anonymous reviewers. Of course, all remaining errors are mine.

Disclosure statement

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

Notes

1 see Donovan (2019) for a recent overview.

2 Besides Germany, only the Belgian population is not able to vote directly. All other member states have some rules to enable direct democratic votes on EU agreements, see https://epthinktank.eu/2016/11/30/ratification-of-international-agreements-by-eu-member-states/referendum/ [accessed 05 August 2019].

3 See Online Appendix A for a complete presentation of the treatment and control text.

4 The ordered probit models in the Online Appendix C report a similarly small effect with an odds ratio of 0.78.

5 Alternatively, one could also argue that negative regime evaluations might emerge in states that already ratified the treaty, as the French rejection invalidates their approval. This expectation gives more weight to the negative outcome than the process. I cannot clearly disentangle the outcome form the process in Study II. However, I estimate the effect of the French vote across different ratification stages below. In contrast to this alternative argument, I do not find negative effects on EU regime support in countries with completed ratification (e.g., Spain).

6 Around the days of the French referendum, the Netherlands can be seen as a part of the treatment. After the French rejection, people across Europe learned that the French rejected the constitution and that the Dutch are about to vote on it (and probably reject it) as well. This is also because media reports frequently mentioned the Dutch referendum in connection to the French rejection. I therefore assume that the upcoming Dutch referendum is part of the benchmark.

7 A -/+1 day window also avoids the additional assumptions imposed by a window selection procedure with respect to the selection of covariates and p-value thresholds.

Additional information

Notes on contributors

Dominik Schraff

Dominik Schraff is senior researcher at the Center for Comparative and International Studies at ETH Zurich, CH.