Five lessons from Iceland’s failed crowdsourced constitution experiment.

Iceland Tried to Crowdsource a New Constitution. It Didn’t Work.

What's to come?
July 31 2014 8:22 AM

We, All of the People

Five lessons from Iceland’s failed experiment in creating a crowdsourced constitution.

Reykjavik from Hallgrímskirkj church tower, April 2013.
Iceland’s experiment in redrafting its constitution has challenged the assumption that a constitutional process is always exclusive and opaque. Above, Reykjavik in April 2013.

Courtesy of Tanya Hart/Flickr

Who should write the constitution of a democratic country and, indeed, any country? The answer seems obvious: its people. Yet the constitutions of existing states, including democratic ones, have usually been written by small, rather unrepresentative subsets of individuals. Solon is supposed to have single-handedly laid out the foundations of democratic Athens. The U.S. Constitution was penned by a few dozen white men. More recent examples of constitutional processes involve the usual elites: professional politicians and state bureaucrats. But even elected or otherwise democratically authorized constitutional drafters are at best metaphorically, “We, the People.”

Not only are typical constitutional processes rather exclusionary and elitist, but they also tend to be characterized by an utter lack of transparency. The American Founding Fathers purposefully kept their deliberations hidden from the public in an attempt to protect themselves from popular passions. Even contemporary political theorists such as Jon Elster insist that the ideal constitution process is hourglass-shaped, with widely open consultative moments upstream and downstream of the drafting, but a tiny waist, corresponding to the exclusive and closed moment of actual writing by a select few.

Iceland’s recent experiment in redrafting its constitution has challenged the assumptions that a constitutional process needs to be exclusive and opaque. In 2013 the country came close to passing into law the world’s most inclusively and transparently written constitutional text. This experiment—sometimes dubbed the “crowdsourced constitution”—should prove inspirational for people around the globe intent on writing, or re-writing, their own social contract.


The Icelandic constitutional process included three original features. The first one was a so-called National Forum—an upstream consultation of a demographically representative minipublic of 950 quasi-randomly sampled citizens. These citizens were gathered in a one-day meeting and asked to list the principles and values they would like to see embedded in the Icelandic constitution. They listed, among others, human rights, democracy, transparency, equal access to health care and education, a more strongly regulated financial sector, and public ownership of Icelandic natural resources.

The second unusual feature was an assembly of constitution drafters selected from a pool of 522 citizens that purposely excluded professional politicians (the latter having been discredited in the eyes of the public during the 2008 financial crisis). The resulting council was characterized by relative gender balance—including 10 women and 15 men—and diverse professions beyond the usual doctors and lawyers, including a farmer, a pastor, an art museum director, a radio presenter, a trade union chairman, a consumer spokesperson, a student, and a filmmaker. The presence of Freyja Haraldsdóttir, a human rights activist affected by glass-bone disease, strikingly illustrated that popular sovereignty need not be represented only by able-bodied, middle-aged men in suits and ties.

The third unusual feature was the decision by these 25 constitutional drafters to use social media to open up the process to the rest of the citizenry and gather feedback on 12 successive drafts. Anyone interested in the process was able to comment on the text using social media like Facebook and Twitter, or using regular email and mail. In total, the crowdsourcing moment generated about 3,600 comments for a total of 360 suggestions. While the crowd did not ultimately “write” the constitution, it contributed valuable input. Among them was the Facebook proposal to entrench a constitutional right to the Internet, which resulted in Article 14 of the final proposal.

Finally, the inclusiveness of the process was enhanced by a limited but still unprecedented level of transparency. For example, the National Forum was streamed online for all to see. Similarly, while the work of the Constitutional Council members was mostly done in closed sessions, it included open meetings that were filmed, recorded, and disseminated as PDF files on the council’s online platform.

The resulting constitutional proposal was approved as the basis of a constitution by two-thirds of the voters in an October 2012 referendum, but the bill based on it ultimately stalled in Parliament the following spring. This outcome, though disappointing, far from proves that democratic constitutional process designs are bound to fail. As hopefully more of them will be tried in the future, the question becomes: What, if anything, can be learned from the Icelandic experience? Five lessons seem particularly prominent for any country intent on trying this at home.

1. Plan carefully.
This does not go without saying, as some aspects of the Icelandic experiment seemed a little too improvised. For example, when the elections to the constitutional assembly were held, it was unclear what would happen once its work was done. Would the Parliament make further revisions to the text? Would there be a binding referendum on it? It is harder for the population to understand and take the process seriously if it seems poorly planned. Similarly, when settling on a design choice like crowdsourcing, resources must be made available accordingly. Though the crowdsourcing moment could have led to a virtuous deliberative feedback loop between the crowd and the Constitutional Council, the latter did not seem to have the time, tools, or training necessary to process carefully the crowd’s input, explain its use of it, let alone return consistent feedback on it to the public.

2. Publicly justify (and if possible debate) procedural design choices.
To the extent that a constitutional process aims to be inclusive and transparent, so must the reasoning behind key design choices. In the Icelandic case, various choices seemed a little arbitrary. It is, for example, not clear why the National Forum favored aggregative practices of opinion elicitation as opposed to deliberative ones, or why the constitutional assembly included 25 elected members as opposed to, say, a randomly selected 100. More efforts at justifying design choices ahead of time would have likely increased the quality as well as the legitimacy, legibility, and credibility of the process. In that respect, it might actually have been a good idea to make a discussion of the process part of the National Forum or use crowdsourcing at that early stage too.

3. Be aware of the conditions for change.
In the Icelandic experiment, the popular referendum was just one of the three hurdles the Icelandic proposal had to overcome to be passed into law, and its positive result was not even binding. The other two hurdles—approval by both then current and post-election Parliaments—were probably one too many. Given that these conditions for constitutional change are themselves generally not up for debate (being dictated by the existing constitution), it might have been wiser to go for the less ambitious goal of revising a few key articles than rewriting the whole document. The situation would be different for a country writing a constitution for the first time.

4. Do not try to bypass entirely other existing representative institutions.
The members of the Constitutional Council saw themselves as outsiders to the system fighting entrenched elites. That stance may have played out well just after the crisis, but on the long run it harmed their legitimacy and jeopardized their work. The Icelandic process thus suffered from the antagonism of several key players: The Supreme Court, which struck down the first elected constitutional assembly in January 2011 for what seemed like minor procedural irregularities in the elections; the liberal party, which decried the whole process as illegitimate from the beginning; the academic community, which kept voicing skepticism about the process as “too complicated”; and the media, who ignored it for the most part. The success of any constitutional process is largely dependent on rallying major political actors and maintaining legitimacy throughout.

5. Use experts wisely.
The role of expert consultants in the process should be clarified and a sensible division of labor worked out with, in particular, the members of the constitutional assembly. Involving lawyers in the wording of the Icelandic constitutional proposal was arguably a good idea (to ensure vocabulary accuracy and compatibility with international treaties), but on several occasions it ended up distorting the drafters’ intent. Not only were such expert interventions a violation of popular sovereignty, but they objectively worsened the quality of the proposal.

Although it didn’t result in any actual constitutional change, the Icelandic experiment has definitely challenged the view that a constitutional process must be exclusionary and secretive, creating a precedent for a more democratic design. Let us hope it will inspire more experiments of the kind in the near future.

This post originally appeared on Democratic Audit. For a longer discussion of this topic, see the author’s recent article in the Journal of Political Philosophy.

Hélène Landemore is an assistant professor of political science at Yale. Follow her on Twitter.

  Slate Plus
May 22 2015 11:14 AM What Happened at Slate This Week? Assistant interactives editor Andrew Kahn on the best of Slate, from the history of slavery to Mad Men.