When the referendum really comes, the sovereign Parliament must go.  But whether for good or for evil, the referendum, in principle at least, seems to be coming.

– C. H. McIlwain, The High Court of Parliament and Its Supremacy (1910), xv.

[N]ow we are witnessing something that would have seemed almost impossible a few years ago, a serious discussion taking place in the United Kingdom about the possibility, and the desirability, of the introduction of a federal, or ‘quasi-federal’ system there.

–  M. J. C. Vile, ‘Federal theory and the “New Federalism”’ (1977) 12 Politics 1.

Recent discussion about the possibility of the United Kingdom adopting a written constitution and forming itself into a federal state is not as new as it may seem.  Certainly, the discussion has raised more questions than it has answered.  For present purposes, we may start with the proposition that a federal state exists where there is (1) a binding constitution which (2) provides for representation of the peoples of the regions and localities of the federation within a federal parliament, (3) distributes power among central and regional governments, and (4) cannot itself be altered unilaterally by either the federal or regional parliaments.  As such, there is still a long distance between the present arrangement in the U.K. and a fully-orbed federal state.  Even if the current debate about the grounds, scope and effect of a Scottish independence referendum are resolved, and if popular and political will was to consolidate in favour of some kind of federal state (‘devo-max’ to the nth degree), many more questions about how to create a federal constitution would remain, centred on the four characteristics of a federal state noted above.

The first of these large questions concerns how a binding and legally entrenched constitution could come into being in the U.K., noting that U.K. law still treats the Parliament as the highest authority in the land and the institution that has the most plausible capacity to initiate, if not consummate, a constitutional change of such magnitude.  A second big question concerns precisely how representation of the constituent regions and localities would be instituted, noting among other things the problem of bifurcating the present Parliament at Westminster into two institutions, a federal legislature for the United Kingdom and a regional legislature for England, alongside the existing legislatures of Scotland, Wales and Northern Ireland.  The third question concerns the breadth of power that would to be devolved to the regional legislatures and whether it is feasible for the current asymmetry to be replaced by a more symmetrical devolution of powers.  And, fourthly, there is the very important question about what procedures would be laid down for the amendment of the constitution in the future, a question that cannot be separated from the first one, about how the constitution is to be established as legally binding upon the legislatures.

Just as the constitutional experience of many former British colonies has proven instructive for the analysis of the powers of the Parliament following the European Communities Act, the Human Rights Act and the recent European Union Act, so the experience of former colonies that have formed themselves into federations suggests several different ways in which a federal constitution for the U.K. might be designed and instituted.  In this short comment, I want to suggest that there are at least three basic ways of proceeding, exemplified in the diverse means by which federal constitutions were established in the United States, Canada and Australia respectively.  As it turns out, there are already many interesting parallels between the experiences of these three countries and current developments in the U.K., especially the potentially very significant referendum on Scottish independence presently being proposed (see Grégoire Webber’s recent contribution on the Canadian parallels), not to mention the similarly contentious referendum requirement laid down recently in the European Union Act.

Revolution was the path that the United States took, but in that country there continues a very important but still unresolved debate over whether, when the revolutionary claim to autochthony was effectively made, independence was secured severally by the individual American States or jointly by the States acting together as the Second Continental Congress in 1776 (cf the views of Jack Rakove, Akhil Amar and Henry Monaghan, among many others, on this point).  While many may say that the controversy has been made irrelevant by the outcome of the Civil War, the issue goes very significantly to the nature, design and interpretation of the U.S. Constitution (as US Term Limits v Thornton demonstrated).  The view that I think is best supported by the evidence is that assertions of independence and of constitutive authority were exercised both severally and jointly, but in a way that meant that no State would be bound by either the Articles of Confederation (ratified 1777-81) or the U.S. Constitution (ratified 1787-9) unless it individually ratified the proposed arrangement.  This was clearly the case for the Articles of Confederation, but it was also the case for the Constitution, even though the provision in the Constitution for ratification by only nine States meant repudiating the requirement of unanimity for the amendment of the Articles (see Articles of Confederation, Art. XIII; U.S. Constitution, Art. VII).  This fact of separate ratification by each State was indeed emphasised by James Madison in Federalist No. 39 when he said that the ‘assent and ratification’ of the Constitution, although in the name of ‘the people of America’, was given by the people ‘not as individuals composing one entire nation, but as composing the distinct and independent States’.  Even Chief Justice John Marshall, who is famous for asserting that the American Constitution ‘derives its whole authority’ from ‘the people’, admitted that the people when ratifying the Constitution had ‘assembled in their several States’, noting that ‘[n]o political dreamer was ever wild enough to think of breaking down the lines which separate the States, and of compounding the American people into one common mass.’

Now, it is indeed quite unforeseeable at the present time that a federal constitution for Britain will emerge on the basis of a series of claims to revolutionary autochthony by the constituent people, or peoples, of the United Kingdom.  But the prospect of a Scottish referendum on independence as the basis upon which a new devo-max settlement might be negotiated, suggests that an analogy to the United States might not be altogether out of place.  And here, the making of the U.S. Constitution remains potentially relevant in at least one important respect, for it illustrates how the design of a federal constitution is related to the authority upon which it is conceived to be based.  As Madison pointed out, although the U.S. Constitution was thoroughly ‘federal’ in its foundation, the representative institutions, distribution of powers, direct effect of federal law, and means of amendment of the Constitution displayed both ‘federal’ and ‘national’ features.  To take the most obvious example, the U.S. Senate was chosen by the legislatures of the States on the basis of equality among the States (it is now directly elected by the voters in each State, but still on the basis of State equality), the House of Representatives was, and still is, elected by voters in a manner that is essentially proportional to each State’s population, and the President was and is elected through an electoral college which allocates to each State a number of votes corresponding to its total representation in both houses of Congress.  Similarly, the U.S. Constitution can only be amended through the consent of special majorities of the State legislatures or in conventions held in each State.  In both ‘representation’ and ‘amendment’, the federal principle is expressed in the special role and status of the States, while the national principle is expressed in the movement from unanimity among the States towards majority rule at a state and national level. Moreover, throughout, democracy is conceived essentially as representative democracy, even at the supreme constitutive moments of ratifying the Constitution and making formal amendments to it.

This much is fairly rudimentary, but it is the logic of the constitutional design that is important, for the prime questions to be addressed in constructing a federal system concern the many different ways in which Madison’s ‘federal’ and ‘national’ principles can be combined.  Other federal countries offer models of different combinations of these principles, but a tendency to move from unanimity among the constituent states to majority rule, and from control by the state governments towards popular involvement of some kind, is consistent across all federal systems that come into being on the basis of a negotiated agreement among several constituent states.  Some integrative systems go further in these directions than others, and all of them express the principles in specifically different ways, but the underlying principles are the same.

This brings us to the Canadian and Australian examples.  These two federal systems did not come into being through revolutionary assertions of autochthony.  Lawmakers in both instances were careful to ensure complete legal continuity with the then accepted authority of the British Parliament to legislate for the colonies.  But the Canadian and Australian ways of coming together and constructing a federation were significantly different.  The Canadian federation was designed in a manner that was consciously intended to avoid the supposed tendencies of the American system to disunity and dissolution, expressed most tragically in the Civil War.  Rather than begin with putatively sovereign states bargaining on the basis of a fundamental constitutive equality, the Canadian system was understood to rest, ultimately, on the authority of a sovereign Imperial Parliament which would through the British North America Act 1867 (BNAA) unite Ontario, Quebec and the Maritime provinces into a suitable form of union, modelled on the British system of parliamentary responsible government.  Accordingly, while political representatives of the Canadian colonies did participate in conferences in 1864 and 1866 at which the terms and structure of a proposed union were agreed in the form of a series of published resolutions, the colonies did not participate as equals (the Maritime provinces were treated, constitutionally, as a unit), and they did not presume to dictate to the Parliament the exact language of the statute under which they would be united.  Unlike the Americans, the Canadians thus wished to create a relatively unified federation, under which the legislative powers of the general government (the Dominion of Canada) would be plenary and the powers of the Provinces would be limited to certain specified topics – a significant departure from the American model, where the original and plenary powers of the constituent States were the very presupposition of the federal system and the powers of the United States Congress were therefore limited and specified.  The Canadian Provinces were thus conceived to be creatures of the BNAA (indeed, most of them still don’t have ‘constitutions’ of their own).  The provincial governments were presided over by Lieutenant Governors and ‘represented’ by Senators appointed by a Governor-General advised by the government of the Dominion of Canada as a whole.  Indeed, the very nomenclature was significant: Provinces, not States; Lieutenant Governors, not State Governors, and so on.  Moreover, the constitutive dependence of Canada on the Imperial Parliament was preserved in the fact that no local power of constitutional amendment was included in the BNAA.  The logic of Parliamentary sovereignty thus shaped the Canadian constitution of 1867 through and through.  In its ‘foundation’, the system was highly unitary, with the exception that the Provinces did negotiate the general nature of the system that would be adopted (but not as equals), and these unitary and unequal foundations shaped the fundamentals of the BNAA in terms of its distribution of powers, representative institutions, and lack of an amending provision.

Now it is of course very true that constitutional politics in Canada has seen the country shift very dramatically in the direction of much greater autonomy for the Provinces.  This due to several factors: most notably, the pressures of linguistic and cultural diversity expressed in Francophone Quebec and calls for secession; the addition of several new Provinces to the federation by way of carefully negotiated agreements between the parties; and the unintended consequence that specifying the legislative powers of the provinces in the BNAA provided the Privy Council and Supreme Court with a textual ground upon which to limit expansionist interpretations of federal power.  But in the 1890s, when the Australian colonies were contemplating federation, the Canadian model appeared much too Imperial and centralist for politicians and a voting public that had become quite accustomed to exercising substantial powers of local self-government and constitutional self-determination (cf Colonial Laws Validity Act 1865).  The Australians wanted to follow the American model, and they did everything they could to reproduce an American-style federation subject only to the dictates of a continuing (but oftentimes grudging) willingness to acknowledge the authority of the British Parliament to legislate for Australia.  And it was in this respect that the referendum proved very significant indeed.

Following the American example, Australian politicians generally refused and resisted British encouragements towards federation until they themselves, as elected representatives of the colonies, thought that it was expedient and right to do so.  Accordingly, federation did not proceed in Australia until the Premiers of each colony supported it.  With this support, secured at a conference held in 1890, Enabling Acts were passed in each of the colonial Parliaments which set up a U.S.-style federal convention at which a draft Constitution Bill was to be debated, drafted and submitted to each of the colonial legislatures for their approval.  Such a convention, at which each colonial Parliament was necessarily equally represented, was duly held in 1891.  And the draft bill that emerged was inspired deeply by the American example.  The existence, powers and mutual independence of the constituent colony-states was taken as a presupposition of the whole system rather than as a product of it, and it was thought quite improper to make any provision at all for the governing institutions of the States within the federal constitution.  It was enough that the State constitutions should ‘continue’ as they had, subject only to the conferral of certain limited powers on the federal institutions of government.  Thus, the limited and specific distribution of legislative powers to the federal Parliament presupposed the original and plenary legislative powers of the colonial Parliaments, and these same Parliaments were also equally represented in the federal Senate.  Moreover, federal executive authority, although formally vested in the Crown, was to be exercised by a Governor-General acting on the advice of a Prime Minister and Cabinet responsible to a Parliament in which the Senate had equal power with the House of Representatives except in relation to financial bills; and even here the power of the Senate to refuse to pass supply was conceded, making the government potentially responsible to both houses (as famously occurred in 1975, leading to the controversial dismissal of the Whitlam government by Governor-General Kerr).  And, finally, again influenced by the American example, but also following the particular federating logic of the Australian system, provision was made for the amendment of the constitution by specially elected conventions held in each constituent state.

As it turned out, the Constitution Bill of 1891 did not secure the support of the colonial governments of the day, and federation languished for another four years, until a second convention was proposed in 1895.  This convention was duly held in 1897-8 and a second Constitution Bill was drafted, approved and finally enacted into law by the British Parliament in 1900.  What distinguished this constitution from the earlier draft was a slightly different federating logic.  This time it was thought important for the federal convention itself to be directly elected by the voters in each colony, and for the draft Constitution Bill prepared by the convention to be submitted to the colonial legislatures for their comments, subsequently revised at a second sitting of the convention, next submitted to the voters in referendums held in each colony and, only to the extent thus approved, finally sent to Westminster for enactment into law.  The principles embodied in this constitutive process dictated in the minds of the Constitution’s drafters that while the principles of unanimity and equality among the colony-states must be preserved, the principle of direct, popular, constituent authority should also be expressed throughout the system, particularly in the direct election of the Senate by the voters in each State, and in the provision for amendment of the Constitution by a referendum at which a majority of the voters in a majority of States would be needed, in addition to the support of a majority of voters in the nation as a whole.

In this, and in numerous other more specific ways, the Australian founders gave effect to a particular form of federating logic, similar to the American (and the Swiss) federations, and somewhat different from the Canadian.  In particular, through the referendum, the constitutional logic of the Australian federal system appealed to a kind of political sovereignty in the plurality of peoples of the constituent States as a means of asserting as much autochthony as was possible without altogether repudiating the authority of the British Parliament to legislate for Australia.  Indeed, one of the powers conferred upon the Australian Parliament, acting (significantly) with the consent of all of the State Parliaments concerned, was a ‘catch-all’ or ‘residuary’ capacity to exercise the legislative powers of the British Parliament with respect to Australia (see Australian Constitution, s.51 (xxxviii)).  As Andrew Inglis Clark, one of Australia’s leading constitutional lawyers, said at the time: ‘[the draftsmen] knew what they were doing.  …  They told the Convention what they were doing, and it agreed with them.  …  They did not hold anything back.  They faced the position that they were going in for absolute legislative independence for Australia as far as it could possibly exist consistent with the power of the Imperial Parliament to legislate for the whole Empire when it chose’.

In the 1980s, the constitutional ties between the British Parliament and Australia and Canada were decisively brought to an end.  But within Australia at least, opinions about precisely when constitutional independence effectively occurred, and what it has amounted to, have turned, in part, on views about the referendum – both as the means by which the federal Constitution was first approved by the voters, and as the only regular means by which it can formally and legitimately be amended in the future.  The statutory confirmation of Australia’s constitutional independence from the U.K. in 1986 was in fact secured in two separate Australia Acts, one enacted by the British Parliament following Australia’s request and consent pursuant to the Statute of Westminster, the other enacted by the Australian Parliament following the consent of the State legislatures pursuant to s. 51(xxxviii) of the Constitution.  Which of these statutes is the actually effective one, and by what authority the Australian Constitution is now binding, have been questions that have intrigued constitutional lawyers in Australia ever since, and in the ensuing discussions, the existence of the referendum has played a central conceptual and normative role.

It is exactly in this sense that the use of referendums in the United Kingdom has the potential to be of very great significance.  The referendum, even if only used as an ‘indicative’ device, has the capacity to be much more than a means by which the popular will is ascertained.  Whether it will in fact do so depends on numerous political and legal factors.  The referenda that have accompanied the current devolution arrangements have not led to such a conclusion, of course, but that is a function of the limited nature of devolution itself.  The proposition asserted by the Scottish Parliament that it has the power to define and hold a referendum of the Scottish people on the question of independence, if it is upheld, has the potential to be interpreted not simply as an appeal to public opinion, but as an appeal to an alternative basis of ‘sovereignty’, in much the same way that the referendum has functioned in the Australian debate.  For, as the Australian experience shows, even if independence (or devo-max, or a federal system) is formally established by an Act of the Parliament, the existence of a referendum initiated by local authority can be interpreted as the basis upon which the whole system rests, permanently limiting, or even displacing, the authority of the Parliament.  To be sure, such a fundamental realignment will only occur if it has fairly general support among the political and legal branches of government, but it can happen.  And the fact that the ultimate grounds of the Australian and Canadian federal systems are still debated shows that these things can take a long time to work themselves out.

In any case, many steps would need to be taken before anything approaching a federal state in the contemporary Australian or Canadian senses could be said to have developed in the U.K.  And yet, there are several respects in which the current scheme of devolution is not all that far removed from the Canadian system as it existed in the second half of the nineteenth century.  While the details of both systems were to an extent negotiated between the parties, each scheme rested (or still rests) on British parliamentary enactment.  The powers of the centre also appear in both instances to be original and plenary, while the powers of the regions are conferred from the ‘centre’, or from ‘above’.  And both systems have been asymmetrical in respect of the powers devolved and/or the degree of regional representation in the central legislature.  Even the development of the Sewel Convention and various constitutional concordats between Westminster and Holyrood recalls the way in which the fundamentals of the Canadian system have evolved through intergovernmental negotiations and constitutional agreements in a way and to extent that simply has not occurred in Australia.  Moreover, the legal capacity to make ‘constitutional’ changes to the scheme in both the U.K. and Canada rests, or once upon a time rested, with the British Parliament.

What nonetheless distinguishes the U.K. from mainstream federal systems is the extent of the powers devolved, the way in which those powers are conferred and the grounds upon which they are conferred.  This is because the logic of devolution is fundamentally different.  Devolution works as a grant from a superior legislature to formally subordinate ones.  The Scotland Act 1998 thus affirms the continuing legal authority of the Parliament at Westminster to legislate for Scotland generally and confers on the Scottish Parliament what are in principle subordinate and limited powers.  However, contrary to the scheme envisaged by the Scotland Act of 1978, the Act of 1998 confers general legislative power subject to a (long) list of specified reservations.  Such a scheme presents questions of interpretation that are intriguingly different from those presented by either the Canadian or Australian federal systems, because in Canada the powers of the Provinces are limited to specific topics in a manner similar to that envisaged for Scotland in 1978, whereas in Australia, like the United States, the powers of the States are treated as original and plenary and the powers of the federation specified and limited.  Interesting, the majority of the U.K. Supreme Court in Martin and Miller v Lord Advocate, following a line of Privy Council decisions arising out of Canada, Ireland, and India, seems to have adopted a ‘pith and substance’ theory of characterisation, an approach which tends to open up consideration of the scope of legislative power distributed to both levels of government, instead of focussing attention on whether, simply and literally, a particular enactment can be characterised as having a minimally sufficient connection with a conferred competence.  This marks the British and Canadian approaches off from those which have shaped Australian and American jurisprudence, where the courts have tended to interpret the specific and formally limited powers conferred upon the federation in the widest terms possible and have found sufficient connections to federal heads of power in a manner quite contrary to the framer’s intentions and expectations.  And yet, even these jurisdictions, there have been strong dissenting voices.  Similarly, the highly complex constitutional structure established by the devolution statutes is capable of radically different interpretations based, ultimately, on competing conceptions of the U.K., as the divergent judgments in Martin and Miller intriguingly demonstrate.  Indeed, whether a purposive, ‘pith and substance’ approach will continue to hold in relation to the Scottish Parliament’s asserted power to define and hold the referendum is an interesting question which approaches a kind of petitio principii.  This is because the requisite connection between a Scottish Act and a reserved matter depends on the purpose and practical effect of the law, and such purpose and effect, it seems from Martin and Miller, must in some sense be ‘legal’ and not merely ‘political’; but a Holyrood enactment which authorised a referendum on Scottish independence would only have the effect of legally authorising the holding of the referendum, and would have no legal effect on the existence or terms of the Union – that is, unless one accepts the theory that such a Holyrood-initiated referendum, if answered in the affirmative, would be the expression of the legally-effective sovereign will of the Scottish people to separate from the U.K.  Thus, absurdly, a Holyrood-initiated referendum is legal if it is of no legal effect, and illegal if it has a legal effect.  This is what happens when we play logical games with Austinian sovereignty!  The established federal systems have had to learn to be very careful with such volatile substances.  Most political scientists today say that sovereignty – of either the states or the union – is simply the wrong way to think about federalism.

This brings us, finally, to the identity of the U.K. Parliament as both ‘central’ legislature and as ‘sovereign’ legislature – a point of profound difference from the Canadian, let alone the Australian and American, systems.  This is not quite the same thing as the ‘West Lothian’ problem, although the fact that non-English representatives vote on matters concerning England within the Westminster Parliament – ie, the corollary of the fact that England has no legislature of its own – is one of the consequences of the dual nature of the Parliament and the asymmetrical structure of the U.K. system.  Interestingly, the two functions fused within the one institution (‘central’ legislature and ‘sovereign’ legislature) makes the U.K. look more like what the Germans are accustomed to calling a zweigliedrig or two-tier system, rather than a dreigliedrig or three-tier system, noting that the former conception entails a highly centralised federal system, in which the ‘states’ are subordinate to the ‘federal’ level of government, rather than both the ‘states’ and the ‘federation’ being equally subject to the order of the ‘federal state’ as a whole, as defined by the constitution.  It is not necessary to buy into the metaphysics of German state-theory to see the point.  For the U.K. to become more like a federal state in the dreigliedrig sense, a way to separate the ‘central’ legislature and ‘sovereignty’ functions would have to be found, and the formation of a written British Constitution, resting on the authority of the peoples of the U.K., is one way in which this might be achieved.

In drawing attention to all of this – about the four large questions that would have to be addressed if the United Kingdom were to become a federal state, with comments along the way about the nature of devolution and the question of the referendum – I am conscious that this is all a matter of very lively political debate, and it is not my intention to take sides (in this short piece, at least!).  But I suggest that the American, Canadian and Australian examples (and many other ‘federal’ models besides) can at least help us think through what U.K. devolution is, what it is not, and what it might become.  And because a similar analysis can be undertaken of the practice of many of the member states of the European Union to make ratification of European treaties subject to referendum approval (including the U.K., especially since the European Union Act), such comparisons may also shed light on what the troubled European Union is and what it also might yet become.

Nicholas Aroney is Professor of Constitutional Law, The University of Queensland and Visitor, Centre for European Legal Studies, Cambridge University.