Blueprint for a Constitutional Convention

In a new report published today, Alan Renwick and Robert Hazell examine options for the design of a constitutional convention in the UK. The report identifies and examines twelve key design features that need to be decided. These are summarised here.


Proposals for a UK constitutional convention are made by several parties in their 2017 election manifestos and have been prominent on the political agenda ever since the Scottish independence referendum in 2014. Such proposals are intended to address both widespread disillusionment with the state of democracy and deep constitutional challenges, such as those posed by Brexit and uncertainty over the future of the Union. But there has as yet been little detailed thinking about the form that a constitutional convention should take. In our new report, we seek to fill that gap. We examine the issues, explore the lessons to be learned from constitutional conventions elsewhere, and identify the pitfalls to be avoided.

Most supporters of a constitutional convention argue that it should not be a commission of the ‘great and the good’ and nor should it be composed solely of politicians. Such approaches may have been viable in the past, but expectations for democracy have moved on and more direct forms of citizen engagement are now widely advocated. Where fundamental questions about the country’s future form and direction are at stake, the voices of members of the public should be clearly heard. This attracts many to the citizens’ assembly model of a constitutional convention.

A citizens’ assembly is a body of citizens who are selected at random from the population at large. Stratification is used to ensure that, so far as possible, the assembly’s membership reflects the diversity of the population in terms of criteria such as gender, age, and place of residence.  The assembly meets over multiple weekends. First, the members learn about the options that are available and get the chance to quiz experts and discuss initial ideas among themselves. Then they hear from advocates of a wide variety of views – from politicians, campaigners, and members of the public who wish to be heard. Finally, they reflect on all they have heard, deliberate in depth among themselves, and agree conclusions. Those conclusions are written up in a report, which is submitted to government and parliament.

Citizens’ assemblies were first held around a dozen years ago in British Columbia, Ontario, and the Netherlands. The most recent official assembly of this kind is working at present in Ireland: it agreed proposals for the liberalisation of Ireland’s highly restrictive abortion rules in April and it will shortly move on to consider a number of other issues.

There is clear evidence that such assemblies work well: the quality of members’ engagement is very high and they can develop conclusions that are reasoned and coherent.  At least in Ireland, they have also done much to encourage wider public debate and shape decision-making.

Equally, such successes are achieved only through careful design: a sample that is broadly representative of the public is not easily secured; ensuring that the learning programme is balanced requires great care; genuine, well-grounded deliberation does not take place spontaneously. Our report identifies and examines twelve key design features that need to be decided:

1/ The purposes of the convention. A constitutional convention can be proposed for multiple reasons, but it will work best if its purpose is simply to develop well-grounded proposals. The success or otherwise of such a convention can be gauged in terms of four criteria: how far it represents the wider population; how far it deliberates effectively – in a manner that is reasoned and reflects the interests and values of all parts of society; how far it influences wider public discourse positively; and whether its proposals are implemented.

2/ Scope and terms of reference. While some activists would like to see an overarching constitutional review, there is good reason to think this would be too complex and controversial to yield useful results. Limiting the convention to one aspect of the constitution is likely to be better. The breadth of this aspect of the constitution should be determined in light of the time and resources that are available: the tighter the resources, the narrower should be the focus.

3/ Membership. A citizens’ constitutional convention should ideally consist of ordinary members of the public only, who should be chosen through stratified random sampling from the population as a whole. The only reason to include politicians or representatives of organised civil society as some have proposed would be to encourage them to take the convention process seriously – but there may be better ways of achieving this. A unitary convention would best have around 100–150 members; a federal body would, in total, need more.

4/ Selecting citizens. For an official convention, selection should take place from adult citizens on the electoral register and the selection process should include local meetings where potential members can learn about the convention’s design. For an unofficial convention, cheaper methods, including sampling from an existing online panel can be considered. Stratified random sampling is needed to provide broad representativeness of the wider population.

5/ Structure. A federal structure is complex, so should be avoided on topics that do not need it. But it is likely to be necessary if the agenda relates to devolution. If a federal structure is chosen, careful consideration should be given to its implications for the size of the convention, the character of its deliberations, and the timetabling of its work.

6/ Operating methods. The convention’s work should be divided into three phases: learning, consultation, and deliberation and decision. The learning phase should be supported by a learning programme that is carefully worked out to maximise accuracy, breadth, impartiality, clarity, and accessibility. Consultations should be as wide and open as possible. Deliberation should characterise all of the convention’s work and should be carefully structured and supported by trained facilitators.

7/ Duration and schedule. The convention should have a minimum of two weekends to consider each topic on its agenda. But we strongly recommend that more time than that is better – and essential for any convention with official status. Convention meetings should generally be spaced two to four weeks apart. They should run from Friday evening to Sunday lunchtime. The time should be carefully structured, though organisers should also be flexible to developments in the course of the convention’s proceedings.

8/ Support for members. Members should be treated well. They should be given good accommodation, food, and refreshments, their expenses should be paid fully and quickly, and they should receive a small honorarium if perhaps £150 per meeting day. Individual members’ special needs should be catered for so far as possible, including support for those with caring responsibilities. Steps should be taken to minimise any harassment of members via traditional or social media.

9/ External engagement. A citizens’ convention needs to reach out from the start to engage with politicians, the media, and the public, to explain its role and to interest people in its work and recommendations. This might include briefings with relevant select committees, interim reports, extensive media engagement, and the maintenance of a good website.

10/ Staffing. An official constitutional convention requires a full-time staff of around a dozen, plus key figures such as the Chair, Academic Director, and Chief Facilitator. An unofficial convention could get by with more limited support. Roles, relating, for example, to recruitment and the testing of learning materials in focus groups can be contracted out.

11/ Budget. A citizens constitutional convention would not be cheap. An official, UK-wide, unitary convention would likely cost somewhere in the region of £5 million. A federal body would be more expensive. An unofficial convention could be run on a tighter budget – but it would be unlikely to work well with a budget less than around £1 million. While these sums may seem high, they would represent money well spent if, as is likely, they helped to strengthen our constitutional and broader democratic system.

12/ Role in decision-making. A government setting up a constitutional convention needs to think ahead to what it will do when the convention reports. It may commit to giving a public response to each recommendation from the convention within a certain time. It may commit to holding a parliamentary debate, or referring certain recommendations to a parliamentary committee. If the proposals need to be put to referendum, the government needs to think ahead about how that will be done. The government could allow the convention to remain in being after it has reported so that its members can be advocates for its proposals.

You can read the full report, Blueprint for a UK Constitutional Convention, here.

About the authors

Dr Alan Renwick is the Deputy Director of the Constitution Unit.

Professor Robert Hazell is the founder and former Director of the Constitution Unit.

One thought on “Blueprint for a Constitutional Convention

  1. Pingback: What an English Parliament might look like – and the challenges of giving it proper consideration | The Constitution Unit Blog

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s