Print Friendly and PDF
only search openDemocracy.net

How to design a constitutional convention for the UK

There are numerous options for how to hold a convention. Let's not rush in and instead make sure we get the very best option for lasting democratic change in Britain. Here are some things to consider.

Flickr/Kathleen Tyler Conklin. Some rights reserved.

Politicians from across the political spectrum have come out in the days since Scotland’s independence referendum in favour of a constitutional convention to determine the future shape of the UK. Ed Miliband, Nick Clegg, Nigel Farage, and Green Party leader Natalie Bennett have all said this is the best way to resolve the constitutional anomalies created by further devolution to Scotland and to seek a revival of our democracy. Some Conservatives apparently also share that view – though the party’s leadership stands against it.

There should be no doubt that a constitutional convention is what is now needed.  Even the most innocuous-sounding of the possible reforms that are being floated—English votes for English laws—is far from the simple adjustment that some of its advocates portray it as.  If, as is quite possible, the next election delivers a Conservative majority in England but a Labour majority in the UK, this arrangement would effectively require the establishment of separate UK and English governments – with enormous implications for the structure of government and the civil service. Even if this were successfully achieved, a federation in which one of the constituent units was so dominant—England holds 84 per cent of the UK’s population—would be so unbalanced that it could well fall apart after a few years. All of this needs to be thought through. And the other possible path forward, devolution within England, raise even more questions. What units should power be devolved to? What powers should be devolved?

So we need a process that will identify the various options, gather a wide range of views on them, subject them to expert scrutiny, and develop well reasoned proposals. In the old days, that might have been done by a royal commission. But such an exclusive approach is no longer tenable: the feeling that the voices of ordinary people are excluded from decision-making is part of the problem that our democracy now faces. Trying to solve it with a committee of the great and the good just won’t wash.

So an inclusive constitutional convention is required. But what form should it take?

The most basic question concerns the membership of a constitutional convention. As I explored in depth in a recent pamphlet, if we want a body that is inclusive and likely to foster public legitimacy, there are four options worth considering: a civil society convention (the model of the Scottish constitutional convention of 1989–95); a directly elected assembly (the Icelandic model); a citizens’ assembly (the Canadian model); and a mixed assembly of politicians and ordinary citizens (the Irish model).  The last of these, I will argue, is the best – though that does not mean we can simply import every aspect of the Irish experience into our own context.

Directly elected assembly – the Icelandic model

The easiest of these four to reject is the directly elected assembly. Only one such assembly has ever been established in a functioning democratic state: the Icelandic Constitutional Council of 2011. In many ways it worked well. Comprising only independents, it was free of entrenched party interests.  It reasoned effectively and produced a cogent draft constitution that was endorsed by 67 per cent of voters in a referendum in 2012.

But, for three reasons, it is not a serious option for the UK.  First, it just won’t happen: the House of Commons will not agree to establish a rival body with legitimacy equal to its own. The Icelandic case, remember, is unique, and occurred in the wake of a massive financial meltdown that caused a deep crisis of the Icelandic political system. Our democracy faces chronic inadequacy, not crisis.

Second, it would not work in the UK as it did in Iceland. If a directly elected assembly is not to be just another talking shop for politicians, its members need to be independents. That worked in Iceland, with fewer than a quarter of a million eligible voters. But the UK has an electorate over 45 million.  The only independents who could connect with voters on such a scale would be those who already have strong media profiles – people who could hardly claim to represent the population as a whole. 

Third, even in Iceland the Constitutional Council, for all its merits, has delivered no actual constitutional change. Though the draft constitution was passed at a referendum, governments have since failed to implement it. At least in part, that is because politicians felt excluded from the process and under no compulsion to carry it forward.

Civil society convention – the Scottish model

An option that is likely to attract more support in the UK is a civil society convention, comprising politicians and representatives of groups such as trade unions, business federations, and churches.  We have a precedent for this in the Scottish Constitutional Convention that operated between 1989 and 1995 and that paved the way for devolution to Scotland and Wales in 1999. This Convention was successful in finding common ground among initially disparate views. It was excessively dominated by the politicians, but a new convention could be balanced differently in order to avoid that.

Nevertheless, a civil society convention is also an unsatisfactory model. The reason is simply that it cannot adequately represent the population as a whole. Who would decide what groups should be included or excluded and on what basis? If the idea is that people should be represented as members of trade unions or churches or whatever other groups, why is that a good idea, when our whole democracy is organized around representation of individuals? If, by contrast, it is intended as a way of representing all individuals, then why choose such an indirect method? What of groups, such as the unemployed, that are not well organized – groups that are often the most alienated from politics as it functions today?

For all these reasons, a civil society convention is just not up to scratch. Indeed, those who originally designed the Scottish Constitutional Convention always saw this model as no more than second best. They advocated it only because their preferred option—a directly elected assembly—would be possible only with government support, which would not be forthcoming so long as the Conservatives remained in power.

Citizens’ assembly – the Canadian model

Fortunately, two further options have been invented since the Scottish Constitutional Convention was established. The first is a citizens’ assembly, as used in British Columbia, Ontario, and also the Netherlands between 2005 and 2007 to explore possible electoral reforms. Such assemblies are composed of ordinary citizens. Potential members are randomly selected from the electoral roll and invited to participate.  They are not required to participate and, indeed, most of those invited choose not to take part. Among those who are interested, a further random sampling takes place, which can be stratified to ensure proportionate representation of genders, age groups, and any other categories as may be deemed appropriate. The members then meet at weekends to  learn about options, canvas views, deliberate, and make proposals.

Detailed research has shown that such assemblies—if given the necessary time and resources, about which I shall say more in a moment—work very well. Members show great capacity to understand the issues they face and develop reasoned proposals. They are free of the vested interests of established politicians. They tend to be trusted by other citizens.

One big problem, however, is that, as in the Icelandic case, none of the citizens’ assemblies yet established has led to actual institutional change. That is partly because all discussed electoral reforms in which there was little interest among either politicians or the wider public. But a second reason is that, again as in Iceland, politicians felt excluded from the process and free, therefore, to rubbish its conclusions. That is a possibility that needs to be avoided in the UK.

Mixed assembly – the Irish model

That leads us on, then, to the final model – the Irish model. Following the 2008 economic crash, widespread anger over the failure of the political system led to calls for political reform. A Constitutional Convention was established in late 2012 and continued its work until early 2014. It shared much with the Canadian and Dutch citizens’ assemblies: two thirds of its members were citizens chosen at random (but with the possibility of refusal) from the electoral role. The final third, however, comprised politicians nominated by the various political parties.

At first, it was widely feared that this mixed composition would not work: that the politicians would dominate the proceedings and prevent reasoned deliberation. But research shows this was not the case. In fact, all members felt able to contribute and the quality of deliberation was high.

And, crucially, what sets the Irish case apart from the Canadian and Dutch cases is that politicians are now taking the Convention’s proposals seriously. Three referendums have already been promised and more might follow. Clearly, it is too early to know the outcomes of those referendums, or how politicians will behave during the campaigns. What is striking, however, is that, as the same research paper as above reveals, the politicians who belonged to the Convention have often acted as ambassadors for the Convention’s recommendations during Dáil debates. They feel buy-in to the proposals and want to take them forward. This contrasts with the absence of any such pro-assembly lobby among politicians in Canada.

But don’t import the Irish model wholesale…

All the evidence we have therefore suggests that the Irish Constitutional Convention offers the best model for the composition of a convention in the UK. Inclusion of a majority of ordinary citizens will promote deliberation that is free from partisan vested interests. But incorporating a minority of politicians will make subsequent implementation of the convention’s proposals more likely.

That is not to say, however, that all aspects of the Irish Convention should be adopted for the UK. In Ireland, the Convention was given a random shopping list of eight topics to consider, chosen for no better reason than that they were the issues that the parties of the governing coalition could not agree on when they negotiated their common programme. And the Convention was given just one weekend to discuss each of these. Though these weekends were packed full of talks from experts and pressure groups and sessions allowing the Convention members to deliberate carefully, the time was simply not available to develop deep understanding.

When it comes to working patterns, the Canadian examples offer a much better way forward. Here, the work of the assemblies proceeded through three phases. In the ‘learning phase’, the members received basic training in the issues before them by a team of academic experts. This support was supplemented by sessions with specialists flown in from all over the world. In a ‘public hearings phase’, assembly members travelled around the province attending public meetings; citizens were also able to submit ideas online, in writing, and through other media. Finally, in the ‘deliberation phase’, the members reflected on all they had heard, considered the options, and reached decisions.  At all stages, while the assemblies met for some of the time in plenary, much of the hard work was done in small groups supported by trained facilitators who ensured that all voices could be heard.  This work was spread over ten to twelve plenary weekends. Such support is essential if a convention with a majority of ordinary citizens is to deliberate well and deliver reasoned conclusions.

Application to the UK

A constitutional convention for the UK will have a particularly complex set of tasks before it. It should not be asked to devise a full written constitution: that would open up too many controversies and would only raise the barriers to successful implementation of the key changes that are needed.  Rather, it should focus on the structure of the Union. Everyone agrees that there should be devolution within England, but in what form and to what geographical areas? Should there also be an English Parliament and, if so, should this operate within the UK Parliament under the ‘English votes for English laws’ scheme or as a separate institution, perhaps located outside London? How, if at all, should the units of this federation be represented in Westminster? These are all big questions, and they will be plenty to keep a convention very busy.

The learning phase should allow the convention to learn about the options and hear arguments from experts and pressure groups in relation to all these issues. Convention members should be free to raise related questions and seek further guidance on them. Given the need to consider the status of the nations and the nature of the units within England to which power will be devolved, more short-lived conventions in the different parts of the UK will also be needed to feed in to the process. They should take part in a public hearings phase that also seeks contributions from as many people and groups across the UK as possible. During the deliberation phase, the convention members should continue to receive support from experts and facilitators to help them reason towards final proposals.

All of this will take time to set up and then to operate. Labour’s proposal that a convention should start its work in autumn 2015—much as it has been maligned by those who want to make cheap political hay—in fact gets it about right. Thereafter, the convention will need a good year of meetings—dispersed over weekends—to do its work.

That may not satisfy some who want quick change before the next election. But what we are discussing here is the most radical reform of the structure of the Union since its creation in 1707. To push it through to fit someone’s short-term political timetable would be appalling. It is remarkable that it is Conservatives who are apparently willing to play so fast and loose with the fundamentals of our constitution.

If we get this right, we have some reason to hope that our democracy will be strengthened and the Union will be rendered fit for purpose for the decades to come. If we get it wrong, democratic disillusion will grow further and the Union will collapse within a decade amidst acrimony and a sense of betrayal.


This article is part of the Great Charter Debate series. If you want to support this project, you can donate to OurKingdom here. Thank you.

About the author

Alan Renwick is Associate Professor in Comparative Politics at the University of Reading.


We encourage anyone to comment, please consult the
oD commenting guidelines if you have any questions.