According to the CEVIPOF Political Trust Barometer of January 2026, the French express an adherence to democratic principles coexisting with a very negative evaluation of how institutions function. This gap between adherence and evaluation of functioning is at its highest level since 2009.
The Essentials
- Strong attachment to democratic principles coexists with a very negative evaluation of their functioning (CEVIPOF, January 2026).
- A majority of those surveyed wish citizens to be more involved in major political and budgetary decisions; a large proportion supports a citizen-initiated referendum.
- Comparison with Switzerland suggests that the absence of structured deliberative instruments can direct mistrust toward protest voting rather than toward institutional reform.
- The window for constitutional reform exists: a majority supports concrete changes, but the risk of slipping toward plebiscite or consultative paralysis remains real.
A Desire for Democracy, Not a Rejection
Let us begin with what the figures do not say. They do not say that the French have become hostile to the democratic system. Attachment to democratic principles has resisted for years the crises, mobilizations, and electoral scores of anti-system formations. Adhesion to the principle is robust.
What the figures say, by contrast, is of remarkable precision. A majority judges that democracy does not function well or does not function at all. A large portion asks that citizens be more involved in political decisions. 79% of respondents support more frequent recourse to referendums to decide on laws or important public policies. These measures point in the same direction: a demand for access, not a desire for rupture.
This disagreement between adherence to principles and evaluation of functioning signals an architectural inadequacy, not a democratic collapse. Citizens wish to participate more in major decisions, and current instruments offer them few opportunities to do so.
The malaise is real; its diagnosis matters.
The Swiss Lesson of 1848
Switzerland offers an instructive contrast, not as a model to transplant, but as proof of concept. Since 1848, 1874, and 1891, Switzerland has progressively instituted the mandatory federal referendum, the optional referendum, and the popular initiative, creating institutional channels for political expression. In Switzerland, 100,000 signatures allow for the launching of a federal popular initiative concerning a constitutional amendment, while an optional referendum against a federal law requires 50,000 signatures within 100 days. Political frustration does not accumulate due to a lack of channels: it flows through a regulated procedure.
The result is not idyllic. Swiss popular initiatives have sometimes resulted in decisions that economic or legal experts advised against. The country voted against membership in the EEA in 1992, limited immigration in 2014, and banned minarets in 2009. Direct Swiss democracy produces contestable choices. But it also produces a sense of belonging to the decision-making process that Swiss polls regularly measure as significantly higher than that of its neighbors.
Swiss mistrust exists; it takes a different form.
In France, the institutional situation differs. Article 11 of the Constitution allows the President to organize a referendum on the proposal of the Government or both assemblies, and also provides for parliamentary initiative supported by one-tenth of the electorate. The shared initiative referendum, introduced in 2008, requires 4.87 million signatures and the agreement of 185 parliamentarians: no procedure has ever been completed since its creation. Citizen conventions, experimented with since 2019, produce recommendations that the executive can choose to ignore. The French demand for participation runs up against an architecture that limits the channels for its direct expression.
Protest Voting as a Thermometer
When the demand for participation has limited institutional channels, citizens may seek alternative political bearers. Protest voting can function as a political signal: it is not stabilized by symbolic concessions or discourse about listening to citizens. Instruments modifying the relationship between citizens and decision-makers could absorb some of this pressure. In the absence of credible instruments, mistrust can find electoral expression in subsequent elections.
This blockage has a measurable consequence at the ballot box. Several political formations, presenting different orientations, have capitalized on demands for institutional change. Researchers who monitor protest voting in France note associations between feelings of exclusion from the decision-making process and protest voting.
Several major popular mobilizations, from the Yellow Vests to contestation of pension reform, have produced results perceived as unsatisfactory by their participants. This sequence coincided with variations in institutional confidence measured by polls.
Several democratic contexts show that when channels of participation remain limited, political formations presenting promises of systemic change can progress electorally. These phenomena are not specifically French.
Three Instruments That Work Elsewhere
A significant portion of French people favorable to a citizen-initiated referendum designates a specific tool, experimented with in several democracies with analyzable results.
The European Citizens’ Initiative, launched in 2012, allows one million citizens from seven member states to request that the European Commission legislate. It has experienced partial successes, notably on drinking water protection in 2020, and real limitations: the Commission is not obliged to follow through. But it has proven that the collection of signatures on a large scale can structure a political debate and force an institutional response, even if imperfect.
Citizen conventions selected by lottery offer a second model. France has experimented with this through the Citizen Convention for Climate from 2019-2020, then with the Convention on End-of-Life. The assessment is mixed. The Citizen Convention for Climate submitted 149 proposals; ministerial follow-up does not allow one to assert how many were taken up in full. Post-convention mistrust is real.
But the tool itself has demonstrated an ability to produce complex deliberations on technical subjects, with a popular legitimacy that parliamentary commissions struggle to achieve. The question of follow-up on recommendations remains central to the perception of legitimacy.
Ireland offers the third example, perhaps the most structuring. The country used citizen assemblies selected by lottery to prepare two major constitutional referendums: on marriage equality in 2015 and on abortion in 2018. In both cases, the assembly deliberated for several months, then a recommendation was submitted to Parliament, which convened the referendum. Both reforms were approved by clear majorities in referendums, without it being possible to establish that they were impossible without citizen assemblies. The architecture allowed deliberation to be transformed into decision.
The Drift Toward Plebiscite
The distinction between deliberation and validation is also a temporal distinction. A process that engages citizens upstream of the decision, before the position of the executive is settled, retains an authentic deliberative dimension. A process triggered to confirm an already-fixed orientation shifts toward legitimation. Explicit procedural rules defined in advance can help structure these processes in a way that preserves their deliberative character.
The risk exists. Naming instruments that work is not the same as ignoring those that drift. The recent history of European referendums offers several illustrations.
The referendum on the European constitutional treaty of 2005 in France and the Netherlands produced a negative vote. After this result, the Treaty of Lisbon was ratified by parliamentary means in these two countries. The sequence marked citizen perception in both countries. A plebiscite emerges when the referendum is used as validation of a decision already made, or as a tool for personal legitimation of the leader organizing it.
The difference between a structured citizen-initiated referendum and a Gaullist or Bonapartist plebiscite lies in a few precise architectural variables: who poses the question, on what perimeter, with what prior deliberation timeframe, and according to what rule of majority and minimum participation.
These variables are not details. They determine whether the process produces a legitimate collective decision or a disguised electoral victory.
Consultative paralysis is the other pitfall. A democracy that consults without ever deciding loses its credibility faster than a democracy that decides without consulting. The Citizen Convention for Climate produced 149 proposals, which the government had announced it would take up 146 of. Subsequent evaluations of follow-up diverged depending on the methods. Promising participation and failing to accept its consequences is politically more costly than honesty about the limits of the process.
The 2027-2030 Window and Its Bifurcations
France is entering a political sequence in which the constitutional question could become pressing. The next presidential election, scheduled for 2027, announces itself with recurring demands for citizen participation and institutional change. Several political formations, of different sensibilities, have inscribed institutional reform in their programs.
Two trajectories are taking shape, without anything today allowing one to prefer one with certainty.
The first is that of incremental reform. A government elected in 2027 could introduce a citizen-initiated referendum mechanism at a reasonable threshold, accompanied by a mandatory prior deliberation procedure and a perimeter excluding fundamental rights. This path assumes a sufficiently broad political agreement to revise the Constitution. In France, any revision text must be voted in the same terms by both assemblies; a government project can then be submitted to referendum or Congress by a three-fifths majority, while a parliamentary proposal must be approved by referendum. The difficulty is real.
Several constitutional revisions have been adopted since 2007, notably in February and July 2008 as well as in March 2024, but agreement between the two assemblies remains a major constraint.
The second trajectory is that of crystallization of mistrust. If the 2027-2030 sequence does not produce institutional change, the gap between adherence to principles and evaluation of functioning could persist. Demands for institutional reform would remain unsatisfied by existing channels.
What would allow us to distinguish between the two trajectories in the coming months: the capacity of governing parties to formulate a concrete institutional proposal before the presidential deadline, rather than after. Each time institutional reform has been put on the agenda in a situation of political weakness, in France as elsewhere in contexts of fragmented representative legitimacy, it has failed or been diverted. When put on the agenda in a situation of clarity and strength, it can succeed.
Before 2027, the signal to follow is less the level of mistrust itself than the political system’s ability to treat it as a useful diagnosis. A large portion of those surveyed judges that democracy does not function well or does not function at all and expresses a demand for institutional access and participation.
Sources
- CEVIPOF, Political Trust Barometer, January-February 2026 wave, democratie.direct
