Wales, the Referendum and the Multi-Level State

Presiding officer, ladies and gentlemen, colleague and friends, I’m an academic whose main work over the last decade has been on devolution. And I’m someone who personally has been an avid supporter of the devolution of political power in the UK and elsewhere. Given that dual background it doesn’t really get much better than being asked to give a lecture such as this a few days after Welsh voters resoundingly endorsed a much fuller form of devolution for Wales.

The referendum was definitive in its result, but also, I think, in another sense: it marks a point from which we might rightly begin to talk of a distinctive Welsh political system, a system of authoritative decision-making for and by Welsh citizens with wide scope and genuine autonomy. I will come back to the idea of a Welsh political system in a little while.

My aim this evening is not to dissect the referendum result, but to put it in context and to try and establish the trajectory on which the referendum outcome places Wales. To do that I’m going to talk rather more to begin with about other places, and about how we – both academics and those involved in operating political systems – too often put places like Wales in the wrong analytical context and for that reason fail to understand fully the trajectory on which places like Wales find themselves.

Let me start with two propositions. The first is that statewide politics – that is here: UK-level politics centred on Westminster – is much more important than devolved politics. And second, that doing politics at the statewide scale is a qualitatively better form of politics than doing politics at a devolved scale. Statewide politics is both more important than, and superior to, devolved politics.

Now I don’t subscribe to either of these propositions, but I would say that the main, the dominant ways in which we – academics and practitioners – understand multi-level politics in states like the UK reinforce the assumption that devolved politics is less important than, and inferior to, statewide politics.

Before turning to talk about Wales, I want to spend a few minutes talking about how and why such assumptions take root, and some of the pitfalls we need to avoid in countering them.

How important is statewide politics?

So is statewide politics important? That’s a silly question, because of course it is. We need only to look at the lead roles taken by central governments in addressing the consequences of the financial crisis, or in grappling with the current dynamics of change in the Arab world, to see ample confirmation of how important the politics of central government is, both within and beyond the state. But that importance can also skew the way we view politics. It is easy to conclude that because statewide politics is clearly so important, devolved politics must by definition be unimportant.

There is much evidence, of course, to the contrary. The central state now does much less than it used to as European institutions, the private and voluntary sectors, and devolved governments have taken over many of the functions it used to perform. A recent, comprehensive survey of ‘the growth of regional authority’, covering 42 OECD states over periods extending back to 1950 showed that regional authority – that it, the number and powers of regional-scale political institutions – has grown steadily since 1970. Of those 42 states, 29 have become more regionalised since then, and only two (marginally) less regionalised.

Over a similar period and in an interconnected process we have seen a growth in the numbers, and the electoral strengthening, of political parties which compete in only one part of the state. There are now over 90 non-statewide parties in western Europe, and around 30 of them are significant players in regional party systems. And because parties which have ambitions limited to regions within states are in many places competing effectively to run the growing number of more powerful regional-level institutions, the statewide parties they compete with are forced to adopt distinctive, differentiated regional positions. As a result of these diverse patterns of regional party competition, all those more powerful regional institutions increasingly do different things with their powers, they produce different policy choices, they deliver outcomes for citizens different to those in the region next door.

All this is clear enough, seen from the vantage point of Cardiff Bay since devolution. It is clear enough in Spain, Belgium, Italy, Canada, the US, Germany – at least to observers like me. But let me give an example of how other observers come to a different understanding, from the place I have been studying now for over 20 years: Gerrmany. Here is a recent summary depiction of the German federal system by one of Germany’s most outstanding postwar political scientists, Fritz Scharpf. Scharpf wrote in 2008 that:

The post-war German polity is a federal state with a unitary political culture … there are no politically salient territorial cleavages … and no popular demands for  regional autonomy. Mass communication is dominated by nationwide media; political issues are defined and debated nationally; and public attention is focused on national parties even where they compete for office in the Länder [that is, the German regions]. By contrast, the political salience of policy-making at the regional level is quite low, and the 16 Land elections have the character of ‘second-order national elections’ as parties tend to fight over national policy choices and about the performance of the national government.”

This is a classic depiction of German federalism. Scholars like Scharpf have always stressed the unitarist aspects of German federalism: the ways in which central and regional governments are required by the constitution to work together in making and implementing laws, the way tax bases are shared and equalised so that each regional government is in a position to pursue a constitutional goal of ‘uniform living conditions’ for all citizens, the limited scope the regional governments have for autonomous policy-making. The understanding of German federalism here is one of a single, integrated, statewide system that more or less needs to move in lockstep to get anything done. It is actually an entirely defensible understanding of German federalism as seen from the heart of statewide government in Berlin.

But it is deeply implausible when seen from any of the regional capitals, say Munich in Bavaria. Scharpf tells us there are ‘no politically salient territorial cleavages’, yet Germany has the most successful non-statewide party anywhere, ever in the form of Bavaria’s Christian Social Union, and another very successful one in what used to be the east German Party of Democratic Socialism, and is now part of Germany’s Left Party. Scharpf says there are ‘no popular demands for regional autonomy’. But a German survey that I along with Richard Wyn Jones and Dan Wincott of Cardiff University helped to design found in 2009 that around three-quarters of respondents in the German regions we sampled think that the regional government ‘should have most influence’ over politics in the region. ‘Political issues are defined and debated nationally’? Well, no: there is growing organisational and programmatic divergence of political parties across the regions, and there are strong territorial lobbies which, put crudely, polarise the interests of a rich south and a poor east.

Länder elections are second order, subordinate to the rhythms and issues of statewide politics? Well, no: if we bother to ask voters what they are thinking about in regional elections, we find that most regional elections in Germany respond to regional issues, with statewide issues secondary. And finally the ‘political salience of policy-making at the regional level is quite low’. Well, no, not for regional electorates. Policy in many fields is actually quite diverse and divergent from region to region; there is actually real scope for regional policy autonomy, and it therefore matters who wins regional elections because different parties do different things with the powers at their disposal.

So how can someone as brilliant as Scharpf get it so wrong? Well he hasn’t really. His Berlin perspective takes federalism as a single, statewide unit of analysis. The alternative perspective I have presented – the Bavarian perspective if you like – sees federalism as a system of 16 regions and therefore 16 units of analysis. Both perspectives are in a sense ‘right’; what they show is that there are different, co-existent modes of party competition and policy-making, some statewide, some region-specific, which vary by policy field and the territorial interest of the regions concerned.

The problem is that these two perspectives generally end up talking past each other. There is a nexus of academic analysis and policy practice around central government which reinforces the sense of importance attached to statewide politics. And there is a nexus of academic analysis and policy practice around regional politics which establishes how important regional politics also is. But the twain rarely meets – in Germany, in other places, including the UK – to  produce a systematic understanding of multi-level statehood, of how important political processes unfold simultaneously at different levels of the state.

So we end up with an ‘either-or’ understanding of multi-level politics. Either central government is really important, so we lose sight of what regional government does; or we focus on regional government, and lose sight of some of the statewide context and constraints in which regional government operates. What we need is a ‘both-and’ understanding of the multi-level state, recognising both what statewide and regional government do separately, and the connections and relationships which make them, to varying extents, interdependent. Sadly we get it all too rarely.

Is statewide politics ‘better’?

What about the second of my initial propositions, that statewide politics is somehow superior to regional or devolved politics? One reason that assumption can persist is precisely because statewide and regional politics often talk past each other, absorbed in their own senses of self-importance. But the value judgement about the superiority of statewide politics has a more deep-rooted source. It reflects a received wisdom about the form of statehood that evolved from the 17^th century and which became consolidated in the democratic welfare states of postwar western Europe. Democracy and welfare: we generally think of these as good things. In other work with Dan Wincott I have argued that the association of these ‘good things’ with that particular form of postwar statehood has often become a platform for arguments that public policy outcomes should be pretty much uniform for all citizens no matter where in the state they live; they should not vary from place to place.

Indeed, pleas for regional distinctiveness are by contrast all too easily seen as unwelcome throwbacks, the death throes of ‘primordial local culture’ (Flora, Kuhnle and Urwin 1999: 283), ‘irritating anachronisms’ (Rokkan and Urwin 1982: 1). Take Bavaria as an example. It is hard to imagine a place more fully and successfully integrated into the highest tech end of the global economy, yet the imagery attached to Bavaria in Germany is often that of back-country rednecks swilling beer from unfeasibly large mugs. It has to be said that Bavarian politicians at times foster such perceptions … but then that’s all part of an outstandingly successful form of political mobilisation which trades on tradition and mobilises it for the contemporary and hard-nosed purpose of pursuing distinctive regional interests.

The tendency to lionise statewide politics and to disparage regional political mobilisation has been especially strong in our understanding of the welfare state. There, it is bound up with the assumption that the statewide scale is the one at which citizens ought to express solidarity with one another. As a result, when regional political institutions pursue policy goals different to those that operate at a statewide scale, these are often portrayed as damaging or regressive, undermining social solidarity, rather than as alternative and legitimate outcomes of democratic processes at the regional scale.

For example, it has been argued in the US, Canada and elsewhere that regional differentiation of policy goals runs the risk of a ‘race to the bottom’ as regions compete on (lowering) welfare standards to attract inward investment (e.g. Peterson 1995; Harrison 2006). The consequence – ‘obviously’ to quote a German colleague, Dietmar Braun (2000: 15, my emphasis) – is that ‘social policy and health policy are areas which require a high degree of standardisation and harmonisation’ at statewide scales. The Canadians Keith Banting and Stan Corbett (2002: 19) push the point further by tugging on heartstrings in the rhetorical question: ‘Does a sick baby in one region have access to the same level of care on similar terms and conditions as a sick baby in another region of the same country?’ We are left in no doubt as to what our answer should be: all citizens in the same state – especially sick babies –should have uniform services provided by the state.

But there are alternative answers to questions like that of Banting and Corbett. A region might conceivably decide to do more for sick babies, or older people, or students than its neighbour, or than statewide government might provide for, on the basis of a political programme legitimized through regional-scale electoral competition. A regional government might believe that by withdrawing from statewide measures and doing things itself, it can better serve the interests of its citizens. Indeed, there is evidence that regional governments often seek to enhance welfare standards rather than be drawn into a damaging, downward spiral of inter-regional competition (e.g. Keating 2009: 506-10). There is also evidence that regional-scale innovations in enhancing welfare standards lead to emulation elsewhere within the state (Schmuecker and Adams 2005: 48-9), echoing ideas that the US states act as ‘laboratories of democracy’ and open up the possibility of an inter-regional ‘race to the top’ rather than inevitably one ‘to the bottom’.

Why, in other words, should we imagine that the postwar state is necessarily a superior ‘framework for public action’ (Keating 2009: 504) than other scales for public action? Clearly it is not (Some might say, in parenthesis, that the current UK central government provides ample evidence on this point). But that does not and should not mean a presumption that regional-scale solutions are necessarily better. It means simply that some regions seek to gain, and have the resources and confidence to use, regional-level powers within the state to make different political choices. Whether these are better or worse choices than those made by central governments is not a property of the level of government at which decisions are made. Whether they are better or worse choices is a matter for public debate within the region and ultimately something for regional electorates to decide.

The UK as a Multi-Level State

I hope by now to have dismantled the propositions that statewide politics are inevitably more important and better than devolved politics. If we want to understand the contemporary multi-level state we need to recognise the significance of what both statewide and devolved governments do, and to avoid value judgements that one scale for collective public action is inherently better than another.

Though I have spoken about other places, I think you will have recognised that much of what I have said applies just as well to the UK, and to Wales. There is ample evidence that UK-level and devolved politics in the UK talk past each other, with the former locked inside a Westminster bubble in which devolved politics are ultra-peripheral (how much, for example, did the Westminster bubble actually notice there was a really quite important referendum here last week? Barely at all as far as I could see).

There is ample evidence too that a Westminster perspective holds in great suspicion the idea that devolved governments should be able to do things differently than prevailing wisdom at Westminster. I have particularly enjoyed all the spleen-venting in the Daily Telegraph about the very idea that Welsh (and Scottish) students will receive public subsidy for their tuition that is unavailable to English students. It does seem strange, though, the Telegraph has pursued Leighton Andrews for committing to maintain public subsidy for the Welsh rather than Vince Cable and David Willetts for abolishing it for the English.

But there are also bubbles in Holyrood and Cardiff Bay (and all the more so but for different reasons in Stormont) which are absorbed in devolved politics and often none too interested in or engaged with statewide politics. I’m always struck by how very insular Scottish politics is. Quite remarkably there was no coverage in the two main Scottish newspapers of the result to the Welsh referendum last week, none at all. And while at a more general level we’ve been having for the last four years a really quite vibrant constitutional debate, in both its unionist and nationalist variants that debate is extremely insular. The SNP has by definition a frame of reference limited to Scotland and focused on loosening Scotland’s relationship with the rest of the UK.

And though unionist perspectives typically have a UK dimension, they focus narrowly on the scope of Scotland’s autonomy and its bilateral relationship with UK-level institutions. They generally do not consider the other component parts of the UK, Scotland’s relation to them, the implications of ideas on constitutional change in Scotland for the rest of the UK, or the implications of what others in the UK – notably in England – might think of calls for an even stronger Scottish Parliament.

That insularity is structural. It is a feature of the peculiar form of multi-level state the UK is. The expansion from its English core was never a systematic one, never assimilationist as for example in France. It was in many senses quite relaxed about accommodating Welsh, Scottish and Irish distinctiveness, always asymmetrical and bilateral in institutionalising that distinctiveness, never terribly bothered about coordinating across the UK’s component parts.

Prior to devolution the government of Wales, Scotland and Northern Ireland was integrated to the extent that the Secretaries of State and the distinctive policies for which they were responsible fell under the collective cabinet responsibility of the UK government. Devolution essentially transferred those policy responsibilities to the devolved administrations. What it didn’t do was establish routinised mechanisms of coordination across the post-devolution governments of the UK that might replicate those that had existed within UK cabinet government.

So the relations between the UK’s post-devolution governments are – comparatively speaking – extremely ad hoc, informal, weakly institutionalised. They do not provide a framework – as they do in most other devolved and federal states – for central and devolved governments to define shared, statewide purposes and to take coordinated action to pursue those purposes. Central and devolved governments work in highly compartmentalised ways. When they come together it is much more to deal reactively with the points of interface and friction that arise between the different compartments of government in the UK than to work proactively to secure shared interests.

Indeed, those interests are inexorably diverging, both as between UK and devolved governments, and among the devolved governments. The dynamics of party competition in each political arena – UK, Wales, Scotland, Northern Ireland – are different. The electoral systems are each different, translating differences in party competition into different government formations. And the wider policy communities – the organisation and priorities of interest groups, of civil society, of the media – are each different, compartmentalising policy debate and fostering policy divergence [more so than anywhere else, I should stress, in the case of England. It is there we have seen the most radical policy changes since 1999; the devolved administrations have been in many respects conservative administrations, rejecting changes introduced by the UK government for England].

It is not surprising given its compartmentalised structure, and the different political dynamics at play in each governmental arena in the UK, that the UK’s governments each operate in rather self-referenced ways. A Fritz Scharpf would not find much evidence of an integrated, statewide system of government in the UK. The constitutional and political logic of the UK is rather a centrifugal one, a disintegrative one. I suspect the longer term logic of constitutional development for the UK is a confederal one, with quite disconnected and highly autonomous governments acting for each part of the UK, with the UK government acting both as England’s autonomous government and the contracting partner for the Welsh, Scottish and Northern Irish governments in delivering some set of UK-wide services.

Let me say a little about that centrifugal logic in the case of Scotland, before turning, to conclude, to Wales after the referendum. I mentioned a moment ago the insularity of both the nationalist and unionist cases in the Scottish constitutional reform debate. I now want to turn to the commonality of those cases. In a sense unionists in Scotland are now markedly nationalist. That is, while endorsing continued union, they see a growing rationale for a more self-contained Scottish political system more fully demarcated from the UK-level political system centred on Westminster. They conceive of politics in Scotland increasingly in distinctive national terms.

To an extent this nationalist turn has been disguised by the way in which the Calman Commission on Scottish Devolution and the Scotland Bill it informed were the products of cooperation both across the party boundaries of Labour, the Conservatives and the Liberal Democrats, but also the jurisdictional boundaries of Holyrood and Westminster. There was perhaps more consensus across parties at each level than there was within parties across levels. Key voices in all three parties in Scotland were in favour of significantly strengthened devolution. Powerful voices among the Conservatives and Labour at Westminster were sceptical about the need for further devolution. Only the Liberal Democrats had similar pro-devolution instincts at both levels.

What has emerged in the Scotland Bill is therefore a relatively low cross-border common denominator which understates the strength of opinion among the unionist parties in Scotland about further-reaching devolution. There are certainly advocates in Scotland of more ambitious ideas on legislative powers than those in the Bill – for example concerning the devolution of aspects of social security, so as to connect with current powers in health, education, housing and social services in a joined-up anti-poverty strategy.

And where the Scotland Bill did introduce a radical departure – in its recommendations on the financing of Scottish devolution – it was again probably behind the curve of debate in the unionist parties in Scotland, which favours more fiscal autonomy than the Bill in the end recommended, either as a way of increasing the accountability of the Parliament’s spending decisions, or as a lever for achieving economic or environmental goals [indeed, the report of the unionist-dominated Scottish Parliament Committee on the Scotland Bill, published last week, pushes for additional fiscal powers to those in the Bill both immediately and in the future].

Together with the openness to more legislative powers this commitment to even greater fiscal autonomy suggests a unionism focused in Scotland on establishing an increasingly distinctive, national political system for Scotland within the UK union.

How far from the SNP this is, is a moot point; over the last few years the SNP has firmed up a position which might be called a unionist nationalism. For some time now the SNP’s mantra has been ‘independence of course, but if not that, then maximum devolution within the UK’. ‘Devolution-max’ is now part of the political lexicon of SNP nationalism, with rather more thought given to this in Scottish Government interventions in the constitutional debate over the last few years than to independence.

But even on full independence the SNP has given increasing attention to a ‘British dimension’: a union of crowns; retaining the pound sterling (and with it acceptance of UK monetary policy), at least until a referendum on the introduction of the Euro; cooperation with the UK on defence, including retention of bases in Scotland; shared services with the UK, ranging from vehicle licensing to diplomatic representation abroad; and a number of institutions of ‘partnership between the Scottish Government and the UK Government’ (Scottish Government 2009, p. 112).

All this suggests a nationalism in Scotland focused on establishing as distinctive a national political system for Scotland as possible, but in all variants – devolution-max, or what some have called ‘independence-lite’ – within some kind of continuing union with the rest of the UK.

Where exactly the point of demarcation lies between the SNP’s nationalism-within-union and the new nationalism of the unionist parties in Scotland is unclear. The traditional dichotomy of ‘union’ and ‘independence’ appears less and less useful; there is a substantial area of overlap, largely unacknowledged because of partisan ritual. But look behind the ritual and you will see a large middle ground.

I see this middle ground as a shared subscription to a distinctive national political system for Scotland. I use the terminology of ‘political system’ pointedly. The Scottish Political System was the title of the book published by James Kellas in 1973 and the term he used over three decades ago to categorise the government of pre-devolution Scotland. Others felt then that this was too bold a claim for a territory integrated, inter alia, through a UK government department and a UK-level electoral process into the wider UK political system (Midwinter, Keating and Mitchell 1991).

It hardly seems too bold a claim now. The Scottish Parliament is a powerful institutional locus for a distinctive variant of UK-level party competition. The Scottish specifics of party competition and government formation, and the features of a largely distinctive policy community in Scotland have opened up growing distinctiveness of policy substance and style. The absence of a more systematic approach to coordination between Scottish and UK jurisdictions underlines the scope for continued and wider distinctiveness. Finally, overwhelming public support for devolution – and indeed for further-reaching devolution – suggests this distinctive political system has popular legitimacy.

The referendum and the Welsh political system

Can we speak of Wales in the same way? A political ‘system’ refers to structures and processes of collective decision-making for groups of people within particular territorial boundaries. To use the term ‘political system’ to describe the structure and process of devolved politics is not standard practice. The term is one that has attached predominantly to statewide politics, though has been used to capture the systematic, or ‘patterned actions relating to political decisions’ (Almond 1956: 393) in international relations and in EU politics. I see no reason why it should not be attached to devolved politics.

While contained within a UK political system, and both subject to and part of the collective decision-making of that system, Scotland clearly has a distinctive political system in which substantial collective decisions are made by and for Scots. But to claim that Wales too might have a distinctive ‘political system’ is a far bolder claim to make than it is for Scotland. Wales is frequently measured against Scotland, and generally comes up short. It was joined with England earlier than Scotland, and on more encompassing terms. The Anglo-Welsh union did not recognise and entrench institutions of law, religion, education and administration which reproduced national distinctiveness within the union, as did the Anglo-Scottish union. Wales was assimilated to the English pattern of law, religion, education and administration. Welsh national identity nonetheless persisted, but in ways – through language and religion – which drew boundaries between, and divided people in Wales. Scotland’s institutional focal points for national identity were in principle open to all Scots.

Wales did develop what by the 1990s was an extensive portfolio of administrative devolution, but much later than Scotland, and always within the confines of a legal jurisdiction of ‘England and Wales’. There was no body of Welsh law, as there was Scots law in Scotland, but rather the same body of law as in England, though with a varying, sometimes wide, often limited scope for Welsh deviation. Welsh nationalism was never as powerful a political movement as in Scotland, and was long torn by the dilemma of whether to pursue a nationalism for Welsh-speakers, or for all the people in Wales. Scottish nationalism evolved more or less from the outset as a civic movement open to all people in Scotland, and had fewer obstacles to growth as a popular movement. The form of devolution adopted in Wales after 1997 was a lesser form, limited to ‘secondary’ legislative powers rather than the gleaming, bespoke primary legislation available to the Scots. And so on.

It is easy to depict Wales, against this background, as stuck in an interminable game of catch-up with Scotland. It is also lazy to do so. The trajectory of Welsh politics since devolution is not really one of emulation. As Alan Trench (2009: ???) has put it: ‘A more nuanced explanation than “Wales follows where Scotland leads” is that Scotland opens doors to what is possible within the United Kingdom, and that Wales is then able to consider how to use the opportunities that arise as a result.’ How it has done so is home grown rather than Scotland-led.

It is striking how John Osmond of the Institute of Welsh Affairs described the first year of the National Assembly as a ‘constitutional convention by other means’, ‘a time when a relatively powerless institution began the process of feeling its way, testing its powers, and defining the issues which were to be the most important for the future’ (Osmond 2000: 76). Precisely because Wales lacked the strength and stability of national institutional tradition that Scotland does, there was a more open canvas than in Scotland. Over the last 12 years that canvas has been – to use Rick Rawlings (2003) term – ‘delineated’.

Delineating Wales, Rawlings monumental study from 2003 of the early years of Welsh devolution, is by some way the most meticulous dissection of post-devolution developments in any part of the UK. What he set out was a process of ‘constitution-building by stealth’ (Rawlings 2003: 2), the ways in which Welsh political actors began to make sense of the curious institutional structures established by devolution, to craft them to define and pursue collective purposes in Wales, and to look to reform them where they were dysfunctional – and they clearly were dysfunctional in many ways.

Last week’s referendum was a landmark event in this constitution-building process. The two-thirds vote in favour signed off the critiques made in the internal review of procedure in the first term of the Assembly and Ivor Richard’s seminal report in 2004 as spot on. That two-thirds vote dismissed as inadequate and unfit for purpose the complex interdependencies with Westminster that bedevilled the initial variant of devolution and its successor form in the 2006 Government of Wales Act. The dismissal of that complex interdependence with Westminster in favour of autonomous law-making in 20 fields of public policy signals a robust commitment among the Welsh to seek home-grown solutions to the policy challenges that face Wales. It signals a belief that Westminster is not an appropriate forum for dealing with those challenges.

All in all the referendum result was a remarkable statement of confidence in consolidating a Welsh political system. But it was also a statement that is double-edged. If you remove Westminster from those 20 fields of public policy in Wales, then Wales becomes even more marginal to the Westminster bubble. Politics in Wales becomes even more compartmentalised from politics in Westminster. I see the Presiding Officer’s calls for the abolition of the position of Secretary of State for Wales as entirely right and logical in this context; after the referendum, and without the role in linking Welsh legislative process in Cardiff and Westminster, it is a non-job, just as the Secretary of State for Scotland has for years been a non-job.

The corollary of all this – the other edge of a double-edged sword if you like – is that Wales has to become more self-reliant. If Wales is marginal in Westminster, then Westminster will from time to time do things – like the near wholesale withdrawal of state funding for tuition in English universities – that have tremendous knock-on effects in Wales, without feeling a need to consult anyone in Wales about the policy and its effects. Get used to it.

Likewise, if Wales has its own political system producing home-grown policy solutions, it would not be wise to expect much recognition in Westminster of arguments about ‘fair funding’, however well-grounded they might appear. Indeed, we know from survey work on English political attitudes that the English think by a big majority that the Scots should themselves levy the taxes to finance the policies of the Scottish Parliament, and that the Scots currently get more than their fair share of UK public funding. So do many MPs, across all the major British parties. Expect the same kind of attitude for Wales. Expect more movement in the coming years on the Holtham Commission’s recommendations on fiscal autonomy than those on ‘fair funding’. Expect to hear refrains to the effect of ‘Well done on the referendum, but you’re on your own now, hope it goes well’. That, I think, is the political logic of the UK’s multi-level state.

I think, though, to conclude, this is a logic that can and should be embraced. It places an enormous challenge on the National Assembly, not least in making sure that logic is understood across the regions, by the local authorities, by the main interest groups, by the electorate. Self-reliance means making tough decisions, and those who lose out will complain. But the centrifugal and compartmentalised logic of the UK’s multi-level state also presents a powerful opportunity. And it has to do with my opening propositions about how important different scales of politics are, and how they should be judged. It is the opportunity to mark out what is important to Wales, it is the opportunity to generate Welsh answers to Welsh questions, it is the opportunity to continue Rick Rawlings’ process of ‘delineating Wales’, it is the opportunity for the people of Wales and their representatives to define just what kind of political system Wales should have, what values it should embody, what outcomes it should pursue. That is a rare opportunity; I look forward to seeing Wales and its National Assembly take it.