Binding or advisory?
When a citizens’ assembly finishes, what happens to its recommendations? In almost every case, they’re advisory: the assembly proposes, and elected institutions decide whether to listen. Hélène Landemore argues this is a deeper problem than it looks.
The trouble with advisory-only
Section titled “The trouble with advisory-only”A purely advisory assembly is structurally weak. It’s toothless by design, easy to ignore, and — worse — prone to participation washing: the appearance of empowering people without giving them any real control. It lets a government claim it “consulted the citizens” while keeping every actual decision. The more assemblies stay advisory, Landemore argues, the more they’re prone to capture and to creating an illusion of voice.
It can be binding — it already has been
Section titled “It can be binding — it already has been”Binding citizen power isn’t hypothetical:
- The Michigan Independent Citizens Redistricting Commission — randomly-selected citizens who drew their state’s binding electoral maps, a power voters took from the legislature by constitutional amendment.
- The French climate convention functioned as de-facto citizen legislators: President Macron outsourced legislative initiative to it and promised to pass its proposals. The constitution didn’t formally recognise the move — and, tellingly, the assembly itself declined the referendum that could have made its work binding, mistrusting a public that polling later showed largely agreed with it.
The frontier
Section titled “The frontier”Going from advisory to binding means institutionalising assemblies with real power — which immediately runs into resistance from elected officials (who question their legitimacy) and into constitutional barriers. It’s a hard step. But it’s the only way out of participation washing, and the work of legitimising and embedding these bodies is where much of the field’s energy is now going.
Sources
Section titled “Sources”- Hélène Landemore — DemocracyNext (2026): youtube.com/watch?v=sgFUtZCgAqI.