What Citizens’ Assemblies Can Teach Legislatures
Ahead of the Athens Democracy Forum, Marjan H. Ehsassi reflects on the promise and limitations of citizens’ assemblies, and on how they might strengthen representative institutions
BY BEATRIZ REY
This conversation is part of our series ahead of the Athens Democracy Forum, where questions of democracy’s future take center stage. Citizens’ assemblies have emerged over the past two decades as one of the most visible experiments in democratic innovation, praised for creating space for genuine deliberation but also criticized for their uneven impact on policy.
Few people are better placed to speak to both their strengths and their limits than Marjan H. Ehsassi, Executive Director of the Federation for Innovation in Democracy (FIDE) North America, and author of Activated Citizenship: The Transformative Power of Citizens’ Assemblies. With over 20 years of experience in governance and democratic reform, Ehsassi has served as guarantor, adviser, and evaluator of assemblies in France, Belgium, Canada, and the United States.
Marjan H. Ehsassi is the Executive Director of the Federation for Innovation in Democracy (FIDE) North America, and author of Activated Citizenship: The Transformative Power of Citizens’ Assemblies.
In our conversation, Ehsassi underscores that assemblies cannot be judged solely on whether their recommendations become law. Their value must also be measured at the individual, social, and institutional levels: giving participants voice, building social cohesion, and even reshaping how elected officials view their citizens. At the same time, she cautions against “participatory whitewashing” and stresses that assemblies must be designed with integrity – through sortition, sufficient time, and meaningful deliberation – if they are to deliver meaningful outcomes.
Perhaps most intriguing for readers of this series is her call to build stronger bridges between assemblies and legislatures. Assemblies, she argues, can give legislators the political cover to act on difficult issues, and legislators in turn can learn from the collaborative, problem-solving ethos that assemblies model. Ehsassi is currently working to bring governors and state legislators in the United States into direct contact with citizens’ assemblies, an experiment that could offer important lessons on how to strengthen representative institutions themselves.
Beatriz: You’ve worked across both international and domestic contexts. How do you see the relationship between democratic innovation efforts in the US and those happening abroad?
Marjan: Yeah, it’s an interesting question. In fact, there has been something relevant to this on my mind this past week that I will share with you. I was on a call earlier this week with a national organization that works with local elected representatives. They told me they are setting up an international team to foster learnings from abroad. This is quite fascinating. Three or four years ago, when I shifted from international work to more domestic work, I would talk about my experiences in France or Brussels and what I had learned there. In fact, I intentionally launched FIDE, a European-based organization in the US because it was really important to me to send the message that we have much to learn from other parts of the world. At the time, though, I received a lot of resistance such as:
“Why FIDE?”
“It’s a difficult name! How do even you pronounce it?”
“It’s not very American.”
“I’m tired of hearing about France’s example or Ireland’s example.”
But even in the past few years, there’s been a shift. It’s still subtle, but there’s now this underlying recognition that maybe there are things we can learn in the democratic innovation space from abroad — this development has been interesting and is one we welcome.
Beatriz: That’s very interesting. I have a question about your work with FIDE. You have comparative experience in this area, which I find really valuable. From all this work, what key lessons have you learned that could inform democratic innovation? For instance, if someone wanted to start a citizens’ assembly, what would you tell them? There’s often debate about how effective they can be, and I get asked about that a lot.
Marjan: This is a question that we could talk about for two hours. We’re in the process of putting together a theory of change for assemblies. We’ve developed it collectively, through workshops with both practitioners and donors.
What has surprised me, in the six years since I started learning about assemblies and working in this space, is how little consensus there is around defining success – what makes an assembly effective. I think that’s part of the problem. Often, when people talk about whether an assembly was successful, the focus is on whether the policy recommendations led to actual policy change. But there are multiple ways to measure the success of an assembly.
For FIDE – North America, the impact of an assembly can be assessed on three fronts:
the individual level,
the societal level, and
the institutional level.
When you hear about assemblies “failing,” it’s because they haven’t been successful on one of these dimensions – most often because the recommendations didn’t become law. But there are many other outcomes we should consider.
I often say that we need to be clear about the problem we’re trying to solve. To give you some background: I didn’t come into this work through a purely policy lens. I came to CAs after 16 years of working internationally on governance and rule of law, and realizing that my adopted country and where I was raising two kids was not practicing democracy the way I thought it should. For me, at its core, the problem is voice and the strained relationship between citizens and electeds. This is what I’ll be speaking about in Athens, what I wrote about in my book, and what we’re also highlighting in a report we’ll release before Athens on voice insecurity.
I believe a central ailment of our democracies is the sense that people – even those who are engaged – don’t feel their voice matters, that policies don’t reflect their preferences. I frame this as three deficits:
Minimalist democracy. In the US especially, people equate their voice with their vote. Democracy becomes: you vote every few years, then you go home. There are few ways to hold leaders accountable between elections. Scholars like David Van Reybrouck have done remarkable work urging us to think beyond elections and electoral cycles.
Policy preference misalignment. Data shows growing gaps between people’s preferences and the policies being enacted. Citizens don’t feel that policies reflect their views, and that gap is widening.
Hearing from the usual suspects. This is a point that resonates most with elected officials in the US. They comment on the fact that they hear from the same groups again and again, not from a truly representative cross-section of their constituents.
When we talk about success or failure of assemblies, it really depends on which of these problems you’re trying to solve and how the process is designed to then respond to these challenges.
I’m often called a purist – I believe citizens’ assemblies have key components that must be respected: sortition and representation, sufficient time, meaningful learning, real deliberation. You can’t just do an “assembly” and skip these steps. We have to remember that the practice of assemblies and their potential are two very different things – thoughtful design and iterative implementation are crucial.
Right now, in the US, I see assemblies being designed and implemented in ways that don’t set them up for success. That worries me, because if we want this field to grow, we really need solid proofs of concept. For instance, we should not run an assembly with a small sample of participants or a mandate that does not represent a recognized public problem. Or, if participants deliberate and draft proposals in a single day, those proposals won’t be actionable or well-justified. And then, when you present them to a mayor or the city council, the final report won’t come across as a worthwhile investment.
I’d say this: yes, people look at assemblies and say they’re not very successful. But we need to ask: what are you measuring against? On which dimensions? And how can we correct the design and implementation? We’re still experimenting.
Beatriz: But how do you make these institutions work in a way that is meaningful for polities with representative democracies?
Marjan: You give them real power. There are many things you have to do. One of the most important is what I call “workshopping” a citizens’ assembly before it even starts. Too often, assemblies are launched around a topic that isn’t truly a public issue but rather something a politician wants to put forward. That becomes participatory whitewashing, and we see this.
Even in the French Citizens’ Convention on the End of Life, where I spent four and a half months as one of four guarantors, this was a concern. In many ways, I believe that process was a gold standard for what assemblies can achieve. However, while we were deliberating euthanasia and end-of-life issues, massive pension reform protests were happening outside the building. So why wasn’t pension reform selected as a potential topic?
I argue that the selection of the topic should come from the people. What I recommend to mayors and other elected is to work with policy planners when identifying possible topics. Too often, community engagement staff are tasked with design and implementation. This is important but policy staff also need to be at the table. They are the ones who can say, “Yes, this topic has real potential for policy change.” The worst outcome is to hold an assembly and have nothing change, because that can backfire with participants who feel that you have wasted their time.
Good design and implementation process is complicated. This is why in our theory of change includes a full section on risks and mitigation measures, identifying the various risks and suggesting how to address them to enhance impact. One practical approach is for the government to put forward three topics it is genuinely willing and able to move the needle on, and then let citizens vote on which one resonates most. That way, people have agency in the selection process while ensuring the topic has real potential for impact.
Beatriz: I'm very curious about why you said that the French experience was the golden model. Could you expand a little bit on that?
Marjan: Sure. I think the governance structure was complex but generally strong. What made that process successful were several things. First, all of us wanted it to succeed and there was a lot of flexibility. Whenever problems arose, we tried our best to work collaboratively and correct course. There was a shared commitment to give power to the 184 participants. That wasn’t the case from the beginning, but we moved in that direction gradually.
For example, the guarantors weren’t involved in the initial design. At the very first session, participants asked for co-responsibility. We agreed in principle that they should be, so we introduced a sortition-based model that allowed members of the participant body to join us every weekend as part of the governance committee.
Another example was the press conferences after each session. After three or four sessions, the president of the governance committee from the CESE was the one leading exchanges with the press. The participants wrote a letter, signed by many of them, essentially questioning the legitimacy of others to speak on their behalf. We responded immediately by providing media training and having participants share that role and be present at each press conference.
Step by step, they had more co-responsibility and power. For me, this matters because at the core of our democratic deficits is the fragile relationship between citizens and government. That needs to be addressed by giving people real power and voice. I believe that when assemblies are designed to shift the locus of power to participants, through co-responsibility and co-governance, they can be truly successful. The French convention was not perfect but provides a strong example of this.
Importantly, the minority voice also felt empowered. At the final session, one of the respected participants from the minority position thanked the majority, around 70 percent, for giving the 26 percent equal speaking time. That was a powerful moment and showed how inclusive the process had become.
Beatriz: From your description, it sounds like citizens’ assemblies create real space for debate and participation, which legislatures often seem to lack. In many countries, parliaments are constrained by strict rules, and now social media makes it even harder, since representatives are focused on producing content to attract attention and votes. How do you see the flexibility of assemblies fitting with the more rigid structures of representative institutions? After all, whatever assemblies decide still has to be funneled back through those institutions.
Marjan: I’m not sure I fully understand your question. What an assembly can do for legislators is give them the buy-in to tackle issues that are otherwise extremely problematic for them. Often, we say: don’t choose the low-hanging fruit. Choose topics that are contentious and provide space for real deliberation around tension and disagreement.
But there are really difficult, stubborn problems. Take the Irish models: what those assemblies did was bring together a representative sample of people to tell their representatives, essentially: we have weighed the pros and cons, considered the tradeoffs, and this is where and how we want policy to change to better represent our communities. No expert group can do that. I’m not opposed to having expert consultations on the side, but there is no other process in our democracies — or in any kind of citizen engagement — that allows you to bring together a representative sample of people in person to learn together and provide considered judgement.
I am not a real fan of online deliberation, though it can be a useful add-on for the learning phase of an assembly. This connects to what you were saying about social media being problematic. Many times, I’ve seen participants in citizens’ assemblies say, “I’ve learned that my sources of information were false, that they were feeding me invalid information.” One of the attractions of the assembly model for me is precisely this learning phase.
Of course, curating evidence is very difficult. Do we always get it right? No. It is complicated. But when you do it well, when you have 46 people or 184 people together learning before they start disagreeing with each other, that is incredibly powerful. The set of proposals that emerge and are then submitted to representatives are incredibly rich. Are they completely unique in the sense that no one else could have thought of them? No. But they are grounded in what the community truly wants and believes best represents their collective interest.
Beatriz: Why don't you believe in online deliberation? Or why do you think it's weaker?
Marjan: Because I haven’t seen anything that works really well. Again, the key question is: what problem are we trying to solve? If the goal is simply to tabulate people’s preferences, maybe online platforms offer solutions. The problem with our democracy is not a lack of spaces to provide input. Rather, it is a lack of connection between input and government response. I’m not sure that I have seen large scale online deliberative platforms deliver on government response and policy impact.
In addition, one of the big problems in the United States, and also in other parts of the world, is this pervasive epidemic of loneliness and isolation. There is nothing like bringing people together in person.
People often say cost is the issue, that assemblies are too expensive. Yet we spend millions. In fact, in the US, we spend billions on federal elections. So, in my view, this concern is not one that is grounded in trust. We are a resource-rich country but we prioritize away from citizens’ voices and interests. For those of us who have participated in assemblies repeatedly, there is nothing more powerful than seeing people from different neighborhoods—who would otherwise never sit at the same table—come together and form lasting relationships.
We would like to set up an alum network for assembly participants. Between Canada and the US, just in the past couple of years, we already have about 400 people who have gone through the process. It is still a small number, but these are 400 people who want to stay engaged with their democracy and each other. It would be great to create a speakers’ bureau for them so that they – not us – are coming to Athens to talk about assemblies and their experience.
These are some of the reasons I believe in-person convenings are far more powerful. I have seen it firsthand. For example, I did a case study in Brussels and another in Ottawa. In Brussels, participants didn’t stay in the same hotel, so after deliberations they went home, and the sense of community was not as strong. In Ottawa, even though they came from different parts of Canada, they stayed at one hotel for five days, walked to the convening each morning together, and the friendships that formed were incredibly strong.
So yes, I recognize that online participation improves access for some people and can address certain barriers. I am completely open to including online processes as part of deliberative forums. But in my view, in-person convenings create something very different and much more powerful.
Beatriz: I wanted to ask about scale. I know you have experiences in various scales, but I’m thinking about countries like my own, Brazil. Would you envision something like that working at the national level?
Marjan: Having been involved in national assemblies, I see both advantages and drawbacks. In my view, the main drawback is that at the national level it can be hard to achieve policy impact, unless the mandate is very clear and specific. For example, in France the assembly on euthanasia and assisted suicide has taken a while but, because of its specific mandate, it is on track for legal reform. Without that kind of precise mandate, policy impact is more complicated. Climate change, for instance, would be difficult to tackle through a single national assembly. That said, for instance, a national assembly on housing affordability in the United States could influence national legislation. Still, the policy impact piece is complicated.
Where I think national assemblies are especially powerful is their symbolic impact. Bringing together 150 to 200 people from across an entire country, over multiple sessions, can be incredibly meaningful. It has the potential to change the narrative around democracy, the social contract and reshape the expectations that ordinary citizens have of their national government.
Of course, media has an important role here too. Journalists are not always eager to cover these processes, their promise, and their potential. But when they do, it can amplify the impact and help shift the way we think and talk about democracy.
Beatriz: I often think we are a bit unfair toward representative democracy in terms of what we expect it to do. It doesn’t give people a direct voice in the way that folklore theory suggests. Instead, it gives us a voice to the extent that we elect someone and send that person to represent us. But as you’ve been pointing out, there are several problems with that relationship.
Marjan: Yes. In a country like the US, that relationship is very weak and the power of interest groups gets in the way. If I want to meet my representative, honestly, my best chance is at a fundraiser. I rarely see elected officials walking around my neighborhood. By contrast, my brother is a member of parliament in Canada, and he walks the streets of his riding in North Toronto most weekends.
In Washington, DC, I can attend a town hall, but again, they arrive, deliver remarks, and leave. I don’t feel like my presence or my perspective matters. It is easy to understand why Americans are so turned off by politics. Why should it matter? Electeds don’t seek real input and they don’t respond in a meaningful way. There is no real channel for it.
One thing I have been reflecting on a lot is how much we talk about trust. So often, we measure the health of democracy through the lens of trust, and there is constant talk about trust levels being too low. This seems quite paternalistic. As a citizen, the government’s job is to deliver, regardless of whether I trust it or not. We should focus on people’s trust in their own ability to effect change and solve public problems as members of society.
I have seen something interesting in my years working with citizens’ assemblies. If they are done well, you can see trust increase on five different levels:
First, people can gain more trust in government. Second, government begins to trust citizens and their abilities. That is something we rarely talk about, but it is crucial. Governments often do not trust citizens, which is why they prefer to consult experts and interest groups rather than average people. Yet I have seen again and again that when elected officials come to observe a citizens’ assembly, they are surprised. They say, “the participants are taking this seriously and are committed.” Or, “participants have good ideas.”
Third, as an average citizen, you begin to trust your fellow community members more. You get to know people from different backgrounds and who live in different neighborhoods, you see that you are trying to solve problems together, and this builds social cohesion.
Fourth, and for me most importantly, you begin to trust yourself more. You develop confidence in your own capacity to be an active member of democracy.
Fifth, and this happens sometimes but not always, elected officials can also begin to trust each other more. One of my case studies was in Brussels, where deliberative committees of the French Parliament of Brussels included members of parliament. Typically there were about 40 participants plus 16 MPs. They learned together and deliberated together. Several MPs shared, and I include this in my book, that they had not worked so collaboratively across political parties in a long time. In parliament, they were often disagreeing, but in the assembly they were problem solving. Even in that small space, MPs started to trust one another.
Beatriz: I think that was my point when you were describing the functioning of a citizens’ assembly. It sounded like the way an ideal legislature would work, focused on problem solving. And I think that is something legislatures often lose, for many reasons that vary across countries. But maybe there is institutional learning possible, both from legislatures to citizens’ assemblies and from assemblies back to legislatures.
Marjan: I think that’s a great idea. I would love to have a program or fellowship for members of the legislature to observe and learn about assemblies.
Beatriz: That would be great.
Marjan: We are bringing together a bipartisan coalition of governors, six to eight in total. We have started building our Deliberative State Governance advisory council. It has been very challenging, as you know the US is complicated — mapping red states, blue states, and purple states, and trying to balance the composition of the coalition but initial responses have been positive. Initiatives like Governor Cox and Governor Moore’s “Disagree Better” are the types of programs we need to expand.
I have also been doing presentations for other state bodies like state legislators who have full-time jobs doing other things. They are not well-resourced. Introducing them to assemblies will be important and, for every state that we invite join our Deliberative State Governance learning journey about assemblies, we will also invite one Democratic and one Republican state legislator from that state to participate.
Beatriz: That is a great idea. What is the role of AI for citizens’ assemblies? Do you see potential there, whether positive or negative?
Marjan: That is a good question, but I am not sure I am the best person to answer as I am quite uneducated when it comes to AI. So far, I have not seen AI being used really effectively at the intersection of technology and democracy.
I can imagine AI being helpful in certain areas, such as supporting the learning phase of a CA. Not for deliberation itself, but perhaps for compiling proposals so that the integrity of participants’ ideas is preserved while the proposals are made more actionable and well justified. AI could also play a role in connecting assemblies to the broader public. There may be opportunities there.
My fear, though, is that we are focusing on the wrong things. Scale and cost come up in every conversation about citizens’ assemblies – especially in the US, but as one of my colleagues eloquently puts it, their smallness and slowness is precisely what makes them unique. That is why I am cautious about scaling through technology. I think we still have work to do in perfecting the core formula for a successful assembly. Until we get those components right, I do not see a reason to introduce too many additional elements.
Of course, if someone can show me that AI can help us improve the core practice, I am open to it. I have participated in several workshops on the promises and perils of AI. I have not come away with a sense of real value yet. I have not seen a program or tool that can genuinely support deliberation or the kind of critical thinking needed to process input effectively without taking away the important role of participants. But I would welcome the day when such a model exists.
Beatriz: In my experience focusing on legislative institutions, I find that the most useful tools are those that support staffers and members in their work without making them dependent on the tool. They remain independent actors making decisions. For instance, if I want a summary of a meeting, I can upload the audio and get a concise recap, but the ideas for writing remain my own. I do not want to delegate that part to AI.
This is what works best for legislatures. Brazil’s Senate, for example, has developed an internal ChatGPT for procedural rules. It is very useful because going through rulebooks and finding precedents is time consuming. The tool helps, but the staffers still decide what to do with the information. I share your concern, though. I worry that we are moving too quickly into adopting the technology without having figured out the political side of it.
Marjan: I agree with you completely. As you were speaking, one thought occurred to me. There might be a role for AI in the early stages of designing an assembly. For example, in Boulder, we are supporting a CA on 15-minute neighborhoods. AI could help generate an initial list of experts to invite — those who support the idea and those who oppose it to provide testimony. That preliminary list could then be scrutinized by participants and curators. Using AI as a starting point, while ensuring human judgment and oversight, could be one way to integrate it constructively.
Beatriz: At least that’s where I am with AI. So let me move to my final question, which is about the Athens Forum. It is going to take place in Athens, and you are a specialist in citizens’ assemblies. I find that very beautiful in symbolic terms for democracy. What message are you planning to bring to the forum about the role of citizens and these institutions in the future of democracy?
Marjan: That is a really good question. I haven’t fully decided what I am going to say. A few things come to mind. I am on a panel on populism, and I think that sometimes, especially given the way populist movements have developed, the term itself frightens people.
To me, what we are often missing is a real understanding of people’s discontent. Many people simply want to feel that they have a voice, that their concerns matter, and that their leaders are listening to them. When that does not happen, there is space for mal-intentioned actors and strongmen to step in and co-opt that energy. We have seen how disruptive this can be in mobilizing people.
What I would like to add at the ADF is that we bear some responsibility for not finding ways to channel this energy in a more constructive way. We may not always agree with ideas or policies that emerge, but we need to recognize and value the power of validation that people get when they feel heard and included.
As someone who is a transplant in the United States, I feel that, for decades, citizens of this country have been given two binary choices: defend a democracy that does not work well for them, or destroy democracy altogether. This is also true in other parts of the world. What is missing, and this is a failure of imagination on all our parts, is a third option: to reform democracy in real and meaningful ways, to introduce institutional changes that are inclusive and respond to people’s concerns.
Voice insecurity is real. We could tap into it to strengthen democracy, but too often we fail to do so.
Modern Parliament (“ModParl”) is a newsletter from POPVOX Foundation that provides insights into the evolution of legislative institutions worldwide. Learn more at modparl.substack.com.
