“Do I want it if we’ll lose it?” Democratic preferences of national minorities in federal states
“Do I want it if we’ll lose it?” Democratic preferences of national minorities in federal states (DOI: https://doi.org/10.1017/S175577392510009X) was written by Christoph Niessen, Sean Mueller, and Min Reuchamps in 2025. It was published in European Political Science Review.
The authors explore democratic innovations. Chiefly, direct democracy (i.e., referenda) and deliberative democratic institutions (e.g., citizen panels).
More specifically, the authors explore these in the context of subnational identities. Some context:
Belgium has not held a referendum since the 1950 non-binding referendum on retaining the monarchy. (Note Dutch majority in favor, French minority against.)
- Citizen panels are very common in Belgian national politics.
Unionists reportedly fear demographics changes, in anticipation of referenda they cannot win.
Referenda are very common in Switzerland, but 1992 referendum on joining EEC was narrowly lost, reportedly on ethnic lines. (Note German majority against, French minority in favor.)
The theory is that subnational groups will have preferences for or against these innovations. Deliberative democracy relies on consensus, so both majorities and minorities should be in favor of them. Direct democracy however favors the majority, so minorities should disfavor them. Representative democracy is expected to hold constant.
The authors administer a survey in three countries with attention to a main linguistic and national cleavage.
- Belgium, and the French and Dutch identities
Canada, and the French and English identities
- Switzerland, and the French and German identities
- About 1500 responses from Belgium and Switzerland each; about 1200 responses from Canada. All split evenly between the linguistic communities.
- Three outcome metrics: 11-point scales measuring support for...
- decisions by elected politicians
- decisions by sortitioned citizen panels
- decisions by referendum
Two concepts of 'minority': the objective measure which is linguistic group, and the subjective metric which is a 4-point scale measuring agreement with a statement about group efficacy.
- To Francophones, the survey asks how influential French speakers are in national politics. Opotions range from strongly disagree to strongly agree.
- Substantial matching between objective and subjective measures, but also a non-negligible number of 'mismatched'.
- Roughly two-thirds of surveyed Francophones (i.e., the objective minority) perceive low efficacy.
- Similar but higher proportions of the objective majorities perceive high efficacy.
- The authors expect both objective and subjective minority identification are related to disfavoring direct democracy and favoring deliberative institutions.
- They make more nuanced predictions for the mismatched.
Models are fit with both subject and objective metrics, and their interaction.
They use controls based on the literature for individual motivators.
- Satisfaction with democracy causes a preference for democratic innovations. Favor is found among the...
- very satisfied
- very dissatisfied
- "objective 'losers' of representative institutions"
- Discussion of instrumental motivations (i.e., those who expect to win a referendum also favor the referendum), but ultimately not a subject of study here.
- Political ideology causes a preference for different democratic innovations.
- Left and radical right favor referenda.
- Left favor citizen assemblies.
- Survey asks for the party they generally vote for, or is closest. The authors then classify these parties by an 11-point left-right scale.
- Lower income causes a preference for democratic innovations.
Also demographic controls: political interest, external political efficacy, education, age, gender, and country.
Across the board, 'Don't know' responses are treated as refusals/missing data.
Every model is fit thrice, for each of the three dependent variables. The authors first regress the interaction of the objective and subjective minority identities on these dependent variables. They then include the individual controls. Lastly they include the interaction of country and objective minority status.
- To make the model coefficients more compatable, the first ("baseline") model excludes responses that are missing controls.
- Some models featured heteroskedasticity so robust errors are reported.
On account of the interactions, this model is very difficult to interpret. Barring replication, the plots in figure 2 are just about the only thing I can validate.
- Authors interpretations:
- High group efficacy causes support for representative democracy.
- Citizen panels are supported most by "members of the majority who also perceive their group as influential as well as by minority members who do not. The first have nothing to fear, the latter everything to gain from this particular democratic innovation. A similar interpretation can be made for referendums."
- In the full model, "both democratic innovations gather stronger support among minority members if they have a low perceived group efficacy."
- My interpretation:
- In the baseline model, there are clear results as they described.
- In the full model, 95% confidence intervals for the objective majority/minority overlap.