Since the Wales Act 2017, the Assembly has been given power to set its own name, size and electoral arrangements, as well as much more freedom to control its own procedures. In anticipation of exercising these powers, Elin Jones, the Llywydd of the Assembly (the word is linked to the word for steering a boat – so “Helmsperson” rather than “Speaker”), decided to commission an Expert Panel to look at the question of size, to the necessarily linked issue of electoral method, and to the cognate questions of diversity and the voting age. The Panel was to make recommendations that could be implemented for the next Assembly elections in 2021.

Interestingly, only two of the Panel were Welsh – the Chair, Laura McAllister of Cardiff University, and myself. Our other colleagues are experts who will be well-known to many members of the Hansard Society: Rob Clements, a former Council Member of the Society and Director of Research at the House of Commons, David Farrell of University College Dublin, Alan Renwick of the Constitution Unit and (as job-shares) Rosie Campbell and Sarah Childs of Birkbeck. The Panel met regularly during 2017, engaged with many other experts (including the Hansard Society), and worked with a political reference group containing nominees of the five political parties represented in the Assembly.

We concluded unequivocally that Assembly Members (AMs) are overstretched. There are simply too few of them to do the job they are supposed to do. No fewer than seven AMs sit on three principal Committees, for example – and (quite bizarrely, but inevitably) many Committee Chairs must sit on more than one Committee, with three of them sitting on two others. Concentrated and effective scrutiny cannot be delivered that way.

The workload was already too great in 2004. Since then, it has grown enormously, with full legislative powers in 2011, considerable taxation powers after the Wales Act 2014 and increased responsibilities (as well as a move from conferred to reserved powers) after the Wales Act 2017. And Brexit, whether a power-grab or a power-enhancer, will add its own complexity to the workload.

We concluded that the Assembly needs between 80 and 90 AMs, with the higher end of that range preferable. An increase of this size will give it the capacity it needs to do the job citizens want it to do. Comparative evidence is on our side: while there is no accepted formula for calculating the ideal size of a legislature, the Assembly is much smaller than its international peers (or its comparators in the UK).

We were given estimates of what this would cost as a straight-line projection, but we recognised that the acceptability of our recommendations would increase if the argument of affordability could be neutered. So we also set a tough challenge to the two bodies responsible for Assembly costs – the independent Remuneration Board who are responsible for AM salaries and allowances, and the Assembly Commission with responsibility for the running costs of the establishment, including its staff. We asked them to work imaginatively to ensure that costs are kept to an absolute minimum.

What about electoral method? We arrived at a set of principles (proportionality; government stability and accountability; Member accountability; equivalent status; diversity; voter choice; equal mandates; acceptable boundaries; simplicity; sustainability and adaptability). No system delivers all these perfectly, but three systems would, in our view, meet these tests.

Our least preferred system among these three was the current system – Mixed Member Proportional (MMP). We were not aware of any system where the number of list Members exceeds the number of constituency Members. That means, in effect, that an enlarged Assembly elected in 2021 by MMP would be limited to the 40 existing constituencies (there is no time for a boundary review) and so to 80 AMs in total – the very low end of what we thought acceptable. A two-stage process would be possible (80 elected by MMP in 2021, and perhaps 90 in 2026 following a boundary review), but we thought this would be confusing and undesirable.

Our preferred system was Single Transferable Vote (STV), but we strongly believed that gender equality – an issue on which the Assembly was something of a pioneer – should form part of the electoral system. One chapter of our Report makes key recommendations for promoting diversity of representation. These are so important that we decided that if they were not implemented – whether through lack of political consensus or the limits of the Assembly’s legislative competence – STV could not be guaranteed adequately to achieve the election of a diverse Assembly. In those circumstances, we believed that a Flexible List system, with its potential encouragement for parties to implement their own arrangements to prioritise candidates with different protected characteristics,would be more likely to deliver a more representative Assembly, and would therefore be preferable to STV.

Either STV or a Flexible List requires new constituencies. To do this by 2021, existing administrative units need to be used. We considered a number of possibilities, but eventually concluded that only two models were really possible. One was to double up the existing parliamentary constituencies to make 20 new constituencies. The other was to use local authority areas, combining some of them and dividing Cardiff to make17 constituencies. We wrestled with a number of issues here. There are merits in both approaches. But our modelling suggested that the 20-constituency model was preferable because it will enable the election of an Assembly towards the upper end of our recommended size bracket, thereby achieving meaningful and sustainable capacity gains, while also minimising the variation in district magnitudes between constituencies which is inherent in the local authority model.

Our final recommendation is for the minimum voting age for Assembly elections to be set at 16, and for resources to be given to citizenship education for young people as they achieve the franchise.

We are not so naïve as to believe that our report will be acclaimed by all politicians, let alone those who remain disdainful about the Assembly. Political parties will be tempted to look at our models and to see whether they work to their own advantage (they will find a lot to interest them in the modelling we publish). Some opponents of the Assembly will revive the hoary myths about Welsh legislative and administrative failure. We will be told that money should be spent on the NHS or schools or infrastructure rather than on democracy. But we have a legislature and we must empower it so that it can work for Wales as well as it ought. That is beyond narrow questions of party or personal advantage. And it is precisely to improve people’s lives that effective democratic institutions are needed.

To be truly effective a legislature needs to be representative of the community, to be fairly elected, and to have enough Members of high quality. If the Expert Panel’s recommendations are embraced by Wales’s politicians, then we will be on our way to an institution that can do the job that Welsh citizens want it to do. On this occasion, I hope that our current political leaders do not emulate Rhodri Morgan.