A Doctrinal Approach to Unconstitutional Constitutional Amendments: Judicial Review of Constitutional Amendments in Sweden

Abstract

The delegation theory of unconstitutional constitutional amendments, a limited theory of unamendability – Applying a doctrinal approach to justify unconstitutional constitutional amendments – Introducing a typology of unconstitutionality – The practical deficiencies and internal inconsistencies of the delegation theory – The limited role of the doctrine of unconstitutional constitutional amendments in the Swedish context – The redundancy of constituent power as a legal concept

Prologue

On 4 May 1912 the Speaker of the second chamber of the Swedish Riksdag (Parliament) refused to submit a proposal, put forth by a member of parliament, for consideration by a committee. ¹ According to the Speaker the proposal, to abolish the Swedish monarchy and to introduce a republican form of government, violated the 1809 Instrument of Government since it was irreconcilable with the basic objective principles of the constitution. In contemporary terms, the Speaker held that the proposal amounted to an unconstitutional constitutional amendment.

However, the second chamber was not convinced by the Speaker’s reasoning and used its power to refer the issue to the parliament’s Committee on the Constitution, for re-examination.

A narrow majority of a deeply divided Committee on the Constitution held that the rules on constitutional amendment in the 1809 Instrument of Government (§ 81), entailed that all constitutional provisions could be amended or even repealed. There were no exceptions for particularly important parts of the constitution. Hence, the proposal should have been submitted to a committee for consideration, in accordance with normal procedures.

The Committee’s minority opinions reveal that there were conflicting views on the matter. According to one member the proposal amounted to a revolutionary act against the Swedish form of government. Another member stressed that the provisions on constitutional amendment were intended to allow for improvements to the constitution, not its destruction.

The above related account seems to be the only instance when the possibility of unconstitutional constitutional amendments has been considered in Swedish jurisprudence or legal scholarship. ²

Introduction

We are experiencing what is often referred to as a global crisis for constitutional democracies. ³ This crisis is, among other things, characterised by populist movements abusing constitutional amendment provisions, using amendment procedures to weaken the constitutional protection for democratic and rule of law principles. ⁴

As a response to such democratic backsliding international scholars and national apex courts have engaged increasingly with the possibilities, and normative implications, of exercising judicial review of constitutional amendments; the doctrine of unconstitutional constitutional amendments. ⁵ The doctrine of unconstitutional constitutional amendments is, at least in one influential version (the delegation theory), premised on the distinction between constituent and constituted powers. That is, the power to amend the constitution is a limited power entrusted to constitutional organs, whereas the power to enact a new constitution resides with the people. ⁶ Under the delegation theory, this distinction necessarily entails that fundamental changes to the existing constitution – establishing a ‘new constitution’ – cannot be achieved through the regular amendment procedure but requires a ‘constituent process’. ⁷

In the first part of this article, the delegation theory of the doctrine of unconstitutional constitutional amendments is critically examined. It is argued that the delegation theory is best understood as a limited theory of unamendability, rather than a theory of unconstitutionality of constitutional amendments. By outlining a typology of unconstitutionality, it is shown that in most cases the unconstitutionality of an amendment can be established by relying on traditional doctrinal approaches. It is argued that, in these cases, the delegation theory’s crucial distinction between constituent and constituted powers is redundant.

It is further argued that even in those few cases where the doctrinal approaches are left wanting, the delegation theory only supports unamendability as a conceptual possibility, rather than a necessity. ⁸ It is also argued that other conceptual justifications for implicit unamendability are preferable to the delegation theory, and that the delegation theory can be subject to the critique of being internally inconsistent.

The second part of the article provides a Swedish perspective on the doctrine of unconstitutional constitutional amendments. The Swedish constitutional system and, in particular, the Swedish regulation and practice regarding constitutional amendments are introduced. Thereafter, the insights gained in the first part are applied to the Swedish constitutional setting. Whilst the possibilities of unconstitutional constitutional amendments seem never before to have been contemplated in Swedish legal scholarship, this article provides a first account of how a doctrine of unconstitutional constitutional amendments can be applied in Swedish constitutional law.

Finally, some concluding remarks are provided, drawing on the findings of the previous parts of the article.

The doctrine of unconstitutional constitutional amendments

The delegation theory: a brief recap

The power and legitimacy of courts to exercise judicial review of statutes enacted by the legislature is a never-ending topic of debate in legal scholarship. ⁹ However, in most jurisdictions it is acknowledged that courts have such power; the debate rather tends to focus on how that power is wielded. When it comes to the question of judicial review of constitutional amendments, the approach seems less uniform across various legal systems. ¹⁰ This is perhaps not very surprising seeing as the possibility for courts to review constitutional amendments, arguably, poses specific ‘counter-majoritarian’ challenges. ¹¹ Any doctrine of unconstitutional constitutional amendments is thought to respond to the question of whether a formal amendment to a written constitutional text can be unconstitutional. ¹² It can be added that if such unconstitutionality is to have any practical legal effect, the constitutionality of the amendment ought to be subject to review by a court. ¹³ It is common to make at least two distinctions with respect to different forms of unconstitutionality. ¹⁴ A first distinction is usually made between procedural and substantive unconstitutionality, where the procedural dimension relates to how the constitution may be amended and the substantive relates to what may be amended. The second common distinction is that between explicit and implicit forms of unconstitutionality. A typical example of the former is constitutional unamendability clauses, whereas the latter often refer to the idea of unwritten principles that are basic to the constitutional order. This classification yields four forms of unconstitutionality: explicit procedural; explicit substantive; implicit procedural; and implicit substantive.

It should be noted that the question of whether an amendment is unconstitutional is connected, but not equivalent, to the question of whether a specific constitutional provision or principle is unamendable. An amendment may, for instance, be deemed unconstitutional where the procedure used to amend the constitution does not comply with requirements found in the constitution. This aspect of unconstitutionality does not necessarily raise any questions relating to the unamendability of the constitution. Unconstitutionality is therefore a broader concept than unamendability, which can be viewed as one of several constitutional features that may impact the determination of whether a specific amendment is unconstitutional.

In recent years the doctrine of unconstitutional constitutional amendments has received ample interest from scholars of, in particular, comparative constitutional law. ¹⁵ One fundamental issue, particularly in relation to implicit substantive unconstitutional constitutional amendments, is how to account for the fact that an amendment that has been enacted in accordance with the rules prescribed in the constitution can be characterised as unconstitutional. ¹⁶ For this to be possible, the power to amend a constitution must be limited by factors that are not apparent from a textual reading of the constitution. ¹⁷ This is where arguments relating to popular sovereignty, and the concepts of constituent and constituted power, have entered the legal equation. ¹⁸

One influential theory of unconstitutional constitutional amendments, that has been put forth by Yaniv Roznai, is the delegation theory. ¹⁹ In brief, the theory makes a crucial distinction between the (primary) constituent power of the people to enact a new constitution and the thus constituted power of state organs to enact ordinary legislation etc. The power to amend the constitution is described as a secondary constituent power. In essence, the idea is that the power to amend the constitution is a delegated power, which is therefore limited. The power to amend can never entail the authority to change the essence, or the core, of the constitution, since this is a power that resides solely with the primary constituent power of the people.

In this context, it may be noted that the notion that constitutional amendments can be unconstitutional is not new. ²⁰ Nor are the concepts of primary and secondary constituent power novelties. ²¹ Furthermore, these concepts have been used by courts in various jurisdictions, when grappling with the question of invalidating constitutional amendments. ²² However, the delegation theory can be viewed as a novel and ambitious attempt to construct a coherent theoretical framework, building on the jurisprudence of courts across the globe.

The delegation theory seems to have descriptive, conceptual and normative elements. The overarching aim is apparently to construct a coherent theoretical framework that is globally applicable. ²³ In Stone’s words, the theory ‘seeks to advance a general justification for the idea of unamendability founded in the idea of constituent power’ ²⁴ – something which Roznai himself claims to do by describing and explaining the practice of limits on constitutional amendment powers and by evaluating how they work against their own internal logic. ²⁵ Furthermore, he states that the theory ‘carries with it normative implications as to how one should conceive constitution-amending powers and the practice of judicial review of constitutional amendments’. ²⁶

The delegation theory, and in particular its reliance on the concept of constituent power, has been subject to critique on all three accounts – i.e. its descriptive, conceptual and normative aims.

As has been argued by Verdugo, the idea that the people as a constituent power are in any meaningful sense of the word the source of the constitution of any given state lacks empirical support. ²⁷ To be fair, it may be that Roznai’s descriptive account is not primarily related to the concept of constituent power in itself, but rather as to how limits to the constitutional amendment powers are construed in various jurisdictions. Even so, it can be argued that in the actual practice of legislators and courts around the world, the delegation theory’s justification for recognising limits on the power to amend the constitution finds only limited – albeit growing – support. ²⁸

Despite these descriptive shortcomings, there is force in the normative claim that constitutional legitimacy presupposes consent of the governed. Under Roznai’s account this translates to the idea that constitutions can be fundamentally changed only by invoking the constituent power of the people, using procedures that resemble the original constituent power of the people. ²⁹ In response to this normative element of the delegation theory, Verdugo has argued that the legitimacy of a constitution – which may or may not be relevant to its legal validity – is probably better assessed from other perspectives than the procedures used to amend it, and how these procedures conform to the concept of constituent power. ³⁰

Furthermore, the delegation theory of unconstitutional constitutional amendments can be viewed as conceptual, in that the proposed theory follows inductively from the conceptualisation of constituent (and constituted) power. That is, unamendability is presented as a conceptual necessity, which follows from the theoretical construct of (primary) constituent power as the inherently unlimited power of the people to enact a constitution, and the limited (secondary) constituent power of constitutional organs to amend the constitution. This idea of unamendability as a ‘conceptual necessity’ has been criticised by Stone, who contends that the delegation theory merely supports unamendability as a ‘conceptual possibility’. ³¹ As Stone notes, ‘constitutional theorists often make claims that are purportedly universal, but which turn out to embed assumptions that are particular to one (or a set of) jurisdictions’. ³²

A theory of unconstitutional constitutional amendments or of unamendability?

The essence of a doctrine of unconstitutional constitutional amendments is, as noted above, to respond to the question of whether a formal amendment to a written constitutional text can be unconstitutional. ³³ In principle, then, any theory of unconstitutional constitutional amendments ought to address questions of formal and substantive unconstitutionality, as well as explicit and implicit unconstitutionality.

However, it seems clear that the main focus of Roznai’s argument is to theoretically justify implicit (substantive) unamendability as a conceptual necessity, which follows from his explication of the notions of primary and secondary constituent power. ³⁴ To focus on the question of implicit unamendability is rational since, as many have remarked, it is the most controversial aspect of unconstitutional constitutional amendments, ³⁵ especially when used by courts exercising powers of judicial review (or preview). ³⁶

In my view, however, the emphasis on implicit unamendability tends to underestimate the relevance of other varieties of unconstitutionality, and the extent to which those forms of unconstitutionality can be doctrinally justified.

In the following sections I will argue that, from a doctrinal perspective, a vast majority of the instances of unconstitutionality that may arise can be subject to judicial review by relying on the familiar legal concepts of lex superior and the principle of legality. That is, without the need to embrace the delegation theory of unconstitutional constitutional amendments. Arguably, the practical usefulness of the delegation theory may in fact be limited to a rather small number of jurisdictions and instances. It should be noted that the argument advanced below is contextual, in that it presupposes the existence of a written constitution, and only applies to formal amendments to the constitutional text.

Legality and superiority

The very concept of constitutionalism implies that the legal system consists of a hierarchy of norms where the constitution is located at the top of this order, and that ordinary laws are valid on account of the fact that they have been enacted in accordance with the procedures laid down in the constitution. ³⁷

One way to articulate this feature of constitutionalism, is by reference to the principle of legality. There is no straightforward definition of the principle of legality (or legality principle) as a concept that would find acceptance across jurisdictions. ³⁸ What I mean when referring to the principle here, is the demand that a norm be authorised by a higher norm of the legal system, and that this demand binds also the legislature. Hence, the legislature is bound by the constitutional rules regarding norm-creation. ³⁹ Regardless of what this principle is called, it is a basic feature of any constitutional democracy and is closely related to the near universally reflected principle of the supremacy of the constitution. ⁴⁰

Another way of articulating the supremacy of the constitution is through maxims relating to conflicting norms. That is, the written constitution creates a hierarchy of law, and just as ordinary law prevails over decrees from the executive, a constitutional rule prevails over ordinary law. This is commonly referred to as the postulate lex superior (derogat legi inferiori). ⁴¹

The principle of legality would seem to be a logic corollary of the lex superior principle in any legal system with a hierarchical structure – which, of course, applies to all legal systems with a written constitution. ⁴²

The principle of legality: procedure, competence and substance

As defined in previous section, the principle of legality would seem to support judicial review of constitutional amendments that have not been enacted in accordance with the procedure set out in the constitution. A simple case of such formal, or procedural, unconstitutionality would be if the Swedish Riksdag amended one of the Swedish constitutional laws by a single decision, without adhering to the requirement that amendments must be adopted by two identical decisions separated by a general election to the Riksdag. ⁴³ In situations like these, the exercise of judicial review to enforce unconstitutionality can probably be accepted in many jurisdictions, even where the constitution does not expressly provide for judicial review of constitutional amendments. It can be added that the Venice Commission has voiced its strong support ‘for all systems that allow for effective and democratic supervision of the way in which the constitutional amendment procedures have been respected and followed’. ⁴⁴ Not all instances of formal unconstitutionality are, however, as straightforward as the example just mentioned.

Many constitutions provide for a number of different procedures to amend the constitution, with escalating levels of rigidity. Within this framework, it is quite common to place the power to amend the constitution in the hands of different organs, depending on the type of amendment being contemplated. ⁴⁵ For instance, under the Bulgarian constitution amendments are normally adopted by the parliament. However, amendments concerning issues of particular importance, or the adoption of a new constitution, can only be enacted by a grand national assembly, which is constituted for this specific purpose. ⁴⁶ In Bulgaria, then, the competence to amend the constitution is divided between the parliament and the grand national assembly.

In the event that such procedures are transgressed, it is possible to frame this as a case of formal unconstitutionality; the amendment has not been enacted in accordance with the procedure provided for in the constitution. However, it is also possible to frame the issue as a case of substantive unconstitutionality. Turning again to the example of Bulgaria, it can be argued that if the parliament enacts an amendment within a subject matter that falls under the competence of the grand national assembly, it is the substance of the amendment which is at the heart of the matter. It is only possible to ascertain whether the procedural requirements have been met, if one first examines the substance of the amendment, so the argument goes. In other words, the procedural unconstitutionality is dependent on how the subject matter of the amendment is classified, which is a substantive review. ⁴⁷

One actual example of this bifurcation between form and substance can be found in the (in)famous headscarf decision – and related jurisprudence – from the Turkish constitutional court. ⁴⁸ In the 2008 judgment the Turkish constitutional court struck down a set of amendments designed to open the door for removing the ban on wearing headscarves in universities. According to the court the amendments violated the principle of secularity, which is an unamendable principle under the Turkish constitution (Articles 2 and 4). This was accomplished despite the fact that the constitution only authorises the constitutional court to exercise a formal review of constitutional amendments (Article 148). The court used the notion of competence in order to justify its intervention. ⁴⁹ In essence, it argued that the competence to amend the constitution did not entail the competence to alter the unamendable parts of the constitution, which according to the court the proposed amendments did. According to the constitutional court, this approach was consistent with the constitution’s limitation of its powers to exercise strictly formal review of amendments, since formal review includes reviewing the condition of whether the amendment has been put forth as a ‘valid proposal’. ⁵⁰ The judgment of the constitutional court has been criticised for giving the notion of ‘form’ an inadequate meaning, conflating it with substantive review. ⁵¹

However, as argued above, the difficulty of separating form and substance, in cases such as the Turkish one, seems inherent in all cases where the alleged unconstitutionality is related to violations of unamendability clauses or to the existence of several, mutually exclusive, procedures for amending the constitution – such as the example of Bulgaria.

It is suggested here that the issue, as was implied by the Turkish constitutional court, can be addressed by labelling these instances as questions of competence. ⁵² In legal theory a distinction is sometimes made between regulative and constitutive norms, where the former refers to rules that prescribe certain types of behaviour and the latter includes power-conferring rules (or norms of competence). ⁵³ It seems accurate to describe the demand of the principle of legality, that all norms should be authorised by a higher norm, as closely corresponding to the concept of norms of competence. With regard to the Bulgarian example, it could thus be reformulated in the following way: the Bulgarian constitution confers competence to the parliament to enact ‘ordinary’ amendments, and to the grand national assembly to enact amendments that have been deemed to be of special importance. Should the parliament enact an amendment that is reserved for the grand national assembly, it seems succinct to label this as a breach of competence. Rather than trying to classify these instances as either formal or substantive forms of unconstitutionality, an exercise that seems destined to fail, the label competence-based unconstitutionality is proposed here.

Lex superior: a normative hierarchy within the constitution

It could be argued that the lex superior maxim has little to offer in a discussion regarding unconstitutional constitutional amendments. If two norms on the same hierarchical level are in conflict, one will in principle have to solve this conflict using other critera, such as lex posterior (derogat legi priori) or lex specialis (derogat legi generali). ⁵⁴ However, quite a few constitutions are structured in ways that make it possible to speak of a normative hierarchy within the constitution.

Indeed, Roznai himself argues that ‘unamendable provisions create a normative hierarchy between constitutional norms’, ⁵⁵ the effect being that an unamendable provision will prevail over a future constitutional amendment. Of course, under the delegation theory this follows from the fact that the unamendable provision is established by the primary constituent power, whereas the amendment is established by the secondary constituent power. This division is, however, in Roznai’s own account, governed by the principle of lex superior. ⁵⁶

It can be noted that the Venice Commission has argued along similar lines. Although the Commission, in principle, seems to be of the opinion that substantive review of constitutional amendments is a ‘problematic instrument’, ⁵⁷ it has in recent opinions held that unamendable provisions indicate that there is an internal hierarchy of constitutional provisions and that such clauses helps to strengthen the justification of exercising substantive review of constitutional amendments. ⁵⁸

Arguably, there is no reason to limit this argument to unamendable provisions. It could also be applied to constitutions that have several procedures to amend the constitution, with escalating levels of rigidity. For example, if a constitutional norm (A) can be amended by the parliament with a two-thirds majority and another constitutional norm (B) in addition requires confirmation in a referendum, it can be argued that this indicates an internal hierarchy between these norms where B will prevail over A in the event of a conflict. ⁵⁹

Returning to the example of Bulgaria, the distinction between constitutional provisions that can be amended by the parliament and provisions that must be submitted to the grand national assembly would seem to indicate an internal hierarchy between these constitutional norms. ⁶⁰

Furthermore, some constitutions make a distinction between procedures for amending the constitution and the enactment of a new constitution. Another variant on this theme is the distinction between amendments and total revisions of the constitution. Where such distinctions exist, they arguably invite the judiciary to make – very difficult – determinations of whether a specific constitutional change should be classified as an amendment or if it is in fact so fundamental that it ought to be categorised as total revision or a new constitution.

With reference to the preceding section, it should be emphasised that the just-related examples could be framed either as cases of competence-based unconstitutionality, or as cases of substantive unconstitutionality.

In fact, it could be argued that all of the instances described above are best understood as competence-based forms of unconstitutionality. Possibly, the lex superior principle should be reserved for instances where a lower ranking norm and a higher ranking norm have both been established in accordance with the competence criteria laid down in the constitution, but have conflicting substance. For instance, it is often the case that the legislature and the executive have overlapping or shared competencies to enact legislation and decrees within a certain area. Should a statute and a decree have conflicting contents under such circumstances, it would seem apt to use the lex superior principle in order to give precedence to the statute. If the lex superior principle is given this more limited scope, it would only in exceptional circumstances be applicable when reviewing constitutional amendments.

A typology of unconstitutional constitutional amendments

In the previous sections it was argued that the principle of legality and the principle of lex superior can be utilised to justify the unconstitutionality of constitutional amendments. It was also shown that in many cases it is untenable to make the common distinction between formal and substantive unconstitutionality.

Only where a competent organ, using the correct amendment procedure, fails to comply with the stipulated procedural requirements, is it accurate to use the term formal unconstitutionality – for instance, where an amendment has been adopted by a simple majority, but the constitution prescribes a qualified majority.

In almost all other cases the formal-substantive dichotomy seems untenable – for instance, where more than one amendment procedure exists, and the wrong procedure has been used to adopt an amendment. In order to determine whether the procedure used to enact an amendment was the correct one, it is inevitable that this evaluation will have to take into account the substance of the amendment. That is, the correct form is determined by the amendment’s substance. The same logic applies to amendments that are in contravention of a rule or principle that is protected by an unamendability clause. In such instances the lack of competence to infringe on the unamendable part of the constitution can only be ascertained by reference to the substance of the amendment in question. As demonstrated in the previous section, these instances are more accurately described as competence-based unconstitutionality.

It is suggested here, therefore, that instead of categorising certain amendments as unconstitutional along the formal-substantive dichotomy, it might be more elucidating to formulate a typology of unconstitutional constitutional amendments. The underlying rationale for this approach is to ground potential justifications for unconstitutionality in the specific constitutional text that is being amended. As has been noted by Barak, ‘a natural standard for examining the constitutionality of a constitutional amendment is to examine the requirements in the constitution’. ⁶¹ In other words, the extent to which an amendment can be unconstitutional is, first and foremost, dependent on constitutional design.

A somewhat similar approach has been outlined by Pfersmann, who makes a distinction between what he labels monomorphic and polymorphic conceptions of formal constitutional law. ⁶² Constitutional monomorphism refers to the situation that there is only one superlegislative (or constitutional) form, whereas polymorphism means that the constitution opens alternative ways of producing constitutional amendments with equivalent status (equivalent polymorphism) or that the constitution provides for the production of amendments that are hierarchically differentiated (non-equivalent polymorphism). ⁶³

According to Pfersmann, constitutional monomorphism and equivalent polymorphism exclude ‘any hypothesis of “unconstitutional constitutional amendments”’. ⁶⁴ In my view, Pfersmann seems to be neglecting the possibility, mentioned above, that amendments may be unconstitutional on strictly procedural grounds. This version of unconstitutional constitutional amendment is not dependent on the constitution having an internal hierarchical structure. The more interesting instances of unconstitutionality are, perhaps, those that Pfersmann labels cases of non-equivalent polymorphism, which would in essence seem to correspond to what I have referred to as competence-based forms of unconstitutional constitutional amendments.

It is suggested that a typology of doctrinal variants of unconstitutional constitutional amendments may be organised into four main categories: procedural unconstitutionality; unconstitutional constitutional replacements; tiered unconstitutionality; and unamendability.

The first category, procedural unconstitutionality, refers to instances where a competent organ has employed the correct amendment procedure, but has failed to comply with the stipulated procedural requirements. This type of unconstitutional constitutional amendment can appear in any constitutional jurisdiction where the legislature, in accordance with the principle of legality, is bound by the constitutional rules regarding norm-creation.

The second category, unconstitutional constitutional replacements, refers to instances where the constitution makes a distinction between constitutional amendments and the enactment of a new constitution, or where it distinguishes between constitutional amendments and a total revision of the constitution (a de facto new constitution). This type of unconstitutionality is conceivable both where the constitution establishes a specific procedure for the enactment of a new constitution, or where the constitution does not make explicit how a new constitution is to be enacted. For instance, Article 44(3) of the Austrian constitution provides that ‘any total revision of the Federal Constitution shall … be submitted to a referendum by the Federal people’. This provision certainly implies that any amendment that could be interpreted as a ‘total revision’ may be deemed unconstitutional, unless it is has been approved in a referendum. Consider, on the other hand, Article 146 of the German Basic Law, which provides that it ‘shall cease to apply on the day on which a constitution freely adopted by the German people takes effect’. Even if the Basic Law does not establish the procedure for the enactment of a new constitution, it would seem reasonable to argue that what is envisaged is something other than the procedure for amending the Basic Law, as set out in Article 79(2).

The third category, tiered unconstitutionality, refers to instances where the constitution provides for two or more, mutually exclusive, amendment procedures, where each can be used only in relation to specific parts of the constitution. As noted above, such escalating procedures of constitutional entrenchment are often used to signal that certain parts of the constitution are of a more fundamental character than other parts. As such, these distinctions may also serve as justifications for declaring an amendment unconstitutional, where it has been enacted using a less rigid amendment procedure than provided for in the constitution.

Of course, the fourth category, unamendability, refers to such cases where the constitution provides that certain principles or rules may not be subject to amendment at all. As noted above, there is strong support for the notion that unamendability can serve to justify judicial review of constitutional amendments, as to their compatibility with an unamendability clause.

It follows, then, that where one or more of the following traits is found in a constitution it is possible to doctrinally justify a doctrine of unconstitutional constitutional amendments:

• a specific entrenched procedure for the production of constitutional norms (procedural unconstitutionality);

• distinctive procedures for amendments and the enactment of a new constitution, or a total revision of the constitution (unconstitutional constitutional replacements);

• two or more, mutually exclusive, amendment procedures with escalating levels of rigidity (tiered unconstitutionality);

• unamendable rules or principles (unamendability).

This indicates that a limited doctrine of procedural unconstitutional constitutional amendments could be justified in any constitutional state that has a distinctive procedure for the enactment of constitutional norms – which presumably applies to all jurisdictions with a written constitution. Furthermore, it would appear that most jurisdictions could find justifications for further variants of unconstitutional constitutional amendments, based on their respective constitutional design.

As an example, it can be noted that of the 27 states that are members of the EU, eight states have constitutional unamendability clauses, and eight states have constitutional escalators. ⁶⁵ Furthermore, four states have constitutions that explicitly recognise and establish distinctive procedures for the enactment of a new constitution, or a total revision of the constitution. ⁶⁶ The constitutions of Sweden, Finland and Slovakia also explicitly envisage that the existing constitution can be replaced by a new constitution, but that this can be accomplished using the regular amendment procedure. ⁶⁷ However, most of the constitutions of the EU member states do not explicitly address their own replacement. ⁶⁸ Thus, they only have provisions regarding constitutional amendments, not constitutional replacements. Notably, only ten states would seem to lack all of the traits identified above, that could be used to justify competence-based unconstitutionality; Belgium, Croatia, Denmark, Finland, Hungary, Ireland, the Netherlands, Slovakia, Slovenia and Sweden. ⁶⁹

For these ten states, and others with similar constitutional structures, it may be argued that the delegation theory could fill an important gap. But before addressing that issue, I will high-light some peculiar effects that the application of the delegation theory may have in those states where the doctrinal approach can adequately justify competence-based unconstitutionality.

Practical defects of the delegation theory?

It can be noted that the delegation theory does not seem to support the notion that provisions of the original constitutional text can be unconstitutional – provided that the constitution is a product of the primary constituent power. ⁷⁰ In the same vein, the delegation theory’s justification for holding amendments that alter unamendable provisions unconstitutional, is the fact that the unamendable provision derives from the primary constituent power. According to the internal logic of the delegation theory, this seems to imply that unamendable provisions cannot be enacted by way of constitutional amendment, which derives from the secondary constituent power. ⁷¹

In light of this, the question arises how one should view such amendments as the one to the German Basic Law of 1968, where ‘the right to resist any person seeking to abolish this constitutional order’ was inserted to Article 20 of the Basic Law. Under Article 79(3) an amendment to the Basic Law affecting the basic principles laid down in Article 20 is inadmissible. The amendment to Article 20 would, thus, seem to have added an unamendable principle to the German Basic Law.

However, under the delegation theory it is difficult to view this amendment as being unamendable, in the same sense as the original parts of Article 20. ⁷² Indeed, it has been argued that the amendment to Article 20 of the German Basic Law, having been established by the secondary constituent power, cannot bind that same power. In consequence, it has been argued that the German Constitutional Court would have no basis to review an amendment which alters or annuls the right of resistance in Article 20. ⁷³

It could, of course, be argued that the inclusion of a new unamendable provision in the German Basic Law is merely an example of the secondary constituent power tying its own hands. However, as Roznai argues, under the delegation theory it should not be possible for the secondary constituent power to bind its successors, and to hinder democratically legitimate reform. That power should rest solely with the primary constituent power of the people. ⁷⁴ However theoretically sound this reasoning may be, according to the internal logic of the delegation theory, I find it unconvincing for at least two reasons.

First, such a conclusion is difficult to reconcile with a textual reading of the Basic Law. It would appear that the Basic Law does not rule out the addition of unamendable principles by way of amendment, provided that these amendments do not affect the already unamendable elements of the Basic Law in the sense provided for in Article 79(3). Second, I believe this example shows that the delegation theory’s strict dichotomy between original constitutional provisions and amendments rests on an unrealistic view of how constitutions are in fact created and transformed. ⁷⁵

A further example, where the application of the delegation theory would seem to lead to curious results, relates to the existence and application of constitutional provisions that allow for the enactment of a new constitution or the total revision of the existing constitution. Roznai does recognise that such legally regulated processes for constitutional replacement may carry important benefits, in that they may enhance legal certainty, continuity and legitimacy. However, he also states that this method of regulation ‘seems bizarre’ since the primary constituent power is not bound by any constitutional rules. ⁷⁶ In this vein, Roznai asserts that where a constitution stipulates the means by which it can be replaced, the primary constituent power ‘does not have to abide by it, although it can act accordingly if it so wishes’. ⁷⁷ Whilst this statement makes sense according to the internal logic of the delegation theory, it begs the question of how a court should reason when faced with the question of whether a constitutional replacement enacted in accordance with the relevant constitutional rules is unconstitutional or not.

Roznai’s answer to this question seems to hinge on the extent to which the procedure used to replace the existing constitution is able to ‘imitate … constitutional moments in which the primary constituent power is incarnated’. ⁷⁸ One essential aspect is, thus, whether the power to replace the constitution rests with government organs, or whether the procedure caters for popular involvement and other deliberative processes. ⁷⁹ Relatedly, Roznai outlines a standard of review for courts, which in part is to be determined on the basis of the amendment procedure. He suggests a disproportionate violation standard of review when assessing governmental amendment powers, and a fundamental abandonment standard of review when assessing popular amendment powers. ⁸⁰

Without discussing the merits of these proposed standards of review, it can be noted that they seem to be designed for cases where the question is whether an amendment is in violation of an (explicit or implicit) unamendable provision or principle. It is questionable if they give any guidance when the explicit purpose is to enact a new constitution. Surely, one legitimate reason for enacting a new constitution is a perceived need to abandon some of the principles that have previously been seen as fundamental to the constitutional order. For instance, when the Swedish 1809 Instrument of Government was replaced in 1974, the political powers of the monarch were reduced to a bare minimum.

Furthermore, it arguably makes little sense to rely on the level of popular involvement in the amendment procedure in order to evaluate whether there is cause for judicial review of constitutional amendments. Roznai’s argument for doing so appears to be twofold. First, a high level of popular involvement more closely resembles the exercise of constituent power, as compared to amendment procedures that are more similar to ordinary legislative procedures. Second, popular involvement – and other inclusive and deliberative processes – are, according to Roznai, more likely to have stronger human rights protection. ⁸¹ The first argument would seem to be somewhat circular. That the amendment process resembles an exercise of original constituent power (a concept which is not all that clear), ⁸² is really only legally relevant if the delegation theory’s reliance on the distinction between constituent and constituted powers is accepted. The second argument would seem to indicate that the aim of securing human rights protection should be assessed indirectly, by looking at the amendment procedure’s level of popular involvement. However, if the aim of judicially reviewing constitutional amendments is to promote democracy, rule of law and human rights, it might be more fruitful to ground such an enterprise in other theoretical concepts than the ones proposed by Roznai. As argued by Verdugo, it is possible to justify the unconstitutional constitutional amendment doctrine on other grounds – such as using the core elements of a democratic regime or supranational instruments – both making the concept of constituent power unnecessary and giving judges more guidance in their review of amendments. ⁸³

Regardless, it would seem that the question of whether an amendment should rightfully be regarded as falling within a specific constitution’s definition of being a total revision or a new constitution, can be assessed without resorting to theoretical concepts such as the delegation theory. In essence, such a determination is a question of positive law, which can be disentangled using traditional sources of law – that is, by applying a doctrinal approach to the question of unconstitutionality.

What remains for the delegation theory?

So far it has been argued that most forms of unconstitutional constitutional amendments can be justified by relying on a doctrinal approach to unconstitutionality. The proposed typology of unconstitutionality shows how the concept of unconstitutional constitutional amendments is intimately related to the design of specific constitutions. With this in mind, is there any room left for the delegation theory, or does the doctrinal approach make it redundant?

This question can be addressed by recognising that the delegation theory of unconstitutional constitutional amendments is best understood as a limited theory of unamendability, which has in principle been devised in order to account for instances of implicit substantive unconstitutionality. As has been shown above, the delegation theory does not add much of value to the legal determination of whether an amendment is unconstitutional, provided that the constitutional text can support a doctrine of unconstitutional constitutional amendments. Of course, the textual support can be more or less open to interpretation. It is, for instance, an open question whether an amendment can be characterised as a total revision of the constitution, or whether an amendment contradicts open-ended unamendability clauses protecting principles such as human dignity. Regardless, these are in principle doctrinal questions of law, which can be adequately addressed without relying on the distinction between constituent and constituted powers.

There are, however, instances where the delegation theory would seem to provide answers that cannot be fully addressed by relying on the doctrinal approach. For example, many constitutions contain unamendability clauses whose existence can be explained by historical factors, but that do not accurately reflect the most fundamental or core values of the contemporary constitutional order. It may seem arbitrary (or even absurd) that the constitution would permit amendments that abolish human rights guarantees, whilst it would be unconstitutional to change the form of government from a monarchy to a republic. ⁸⁴

It can of course be argued that if it is deemed that new principles ought to be afforded unamendable status, there is nothing to stop the addition of new unamendable provisions to the constitution. As mentioned above, this was the case when the German Basic Law was amended in 1968. In some jurisdictions, where the constitution is very difficult to amend, such a course of action is less realistic. In these jurisdictions it is, on the other hand, perhaps equally unrealistic to envisage constitutional amendments being enacted that would impair core constitutional values. A pragmatic answer may therefore be that in constitutional settings where it is feasible to introduce new unamendable clauses, it may indeed be seen as a conscious choice not to do so, and that this is a constitutional path that ought to be respected by courts exercising judicial review of constitutional amendments.

A more principled answer would be, as noted above, that is quite possible – and probably even preferable – to justify implicit unamendability on other grounds than constituent power theories. ⁸⁵ In the words of Stone, the delegation theory is merely a conceptual possibility, rather than a necessity. ⁸⁶ In order to reach the conclusion that the delegation theory is a conceptual necessity, it would have to be the case that the validity of a constitution is dependent on it being a product of an exercise of the constituent power of the people. Even if legal validity is a contested concept, ⁸⁷ there is no standard conception of legal validity that is easily reconciled with such a requirement. ⁸⁸ Indeed, there are convincing arguments to the effect that the use of constituent power, in this ‘strong version’, is redundant. ⁸⁹

Internal inconsistencies of the delegation theory

Another example where the delegation theory would seem to be able to fill a void, is where the constitutional text provides that the constitution can be amended or replaced by a new constitution, using the same procedure – as is the case in Sweden. ⁹⁰ It may seem unreasonable that a small technical amendment to the constitution should be evaluated by the same standards as the adoption of a new constitution. Indeed, it seems to make sense that a process which results in a new constitution should be subject to more demanding deliberative and legitimising processes than are minor amendments. It is not self-evident, however, how this translates in relation to the delegation theory’s premise that fundamental constitutional change should be reserved for an exercise of the primary constituent power.

In part, this is due to the confusion as to what constitutes an exercise of constituent power. Duke has argued that the widely supported idea of popular sovereignty as an accepted source of constitutional legitimation can refer to the fact that the people are the bearer of constituent power, that the people exercise constituent power, or both. He maintains that identifying the people as the bearer of constituent power says very little about any required level of popular participation by citizens when constitutions are enacted or amended. ⁹¹

Duke has further argued that the people’s exercise of constituent power can be read either as a claim about their capacity to effect constitutional change, or as a claim that elected representatives should always engage in processes of constitutional change on behalf of the people. ⁹² His conclusion, which I tend to agree with, is that the first proposition does not withstand close scrutiny, which means that constituent power is exercised by representatives on behalf of the people. ⁹³

If the proposition is accepted, that constituent power is exercised by representatives, it seems clear that this process can have many shapes and forms. One way of recognising this aspect of the exercise of constituent power is to acknowledge, as Roznai does, that ‘all constitutions can be considered as imposed to some extent’. ⁹⁴ However, by stating the matter in those terms, the presupposition would seem to be that the exercise of constituent power is within the capacity of the people, rather than an exercise of representatives on behalf of the people. To my mind, Duke is more convincing in arguing that the notion of constituent power – if it is to be utilised – should be grounded in the normative requirement that representatives exercise such power on behalf of all citizens. ⁹⁵ How this requirement ought to be understood is, of course, open to debate. One interesting proposition has been put forth by Beckman, who argues that popular sovereignty is a commitment to legitimate exercise of power, which is met only if power is exercised in ways that all citizens can reasonably accept. ⁹⁶ Beckman’s understanding of how constituent power is legitimately exercised can of course have repercussions for how constitutional change ought to be engineered. However, it is normatively grounded in rational principles, rather than reliant on the idea of the people as an extra-legal power.

Regardless, the link between procedure and the legitimate exercise of constituent power is not straightforward. As Duke points out, there is a spectrum of constitutional models that has varying ways of constructing how representatives can actively exercise constituent power. As an illustration, he refers to the examples of the British model – where there is no active exercise of constituent power beyond parliamentary representation – and the ratification of the US constitution, where constituent power was exercised by a constituent convention. ⁹⁷

It would seem that the delegation theory is open to the criticisms of conflating legitimate exercise of constituent power with the people as the bearers of constituent power, of subscribing to the fallacy that the people – in any meaningful sense – have the capacity to effect constitutional change, or both. One ambiguity of the delegation theory, in this regard, is that it fails to demonstrate how the people can exercise their primary constituent power, yet it maintains that the more similar the characteristics of the secondary constituent power are to those of a primary constituent power, the less it should be bound by limitations and judicial scrutiny, and vice versa. ⁹⁸ On the one hand it is emphasised that ‘process matters’, ⁹⁹ but on the other hand no ‘original process’ for the exercise of primary constituent power is offered. To my mind, this implies that the underlying, but unarticulated, reason for advocating that fundamental constitutional change should follow genuinely deliberative processes – rather than ordinary law-making procedures – is in fact related normative principles of rational decision-making, and not the people conceiving ‘themselves as a single sovereign in order to attribute the Constitution to a single act of will’. ¹⁰⁰

In summary, even if the delegation theory’s premise is accepted, that there is a relevant distinction between the primary constituent power to institute fundamental constitutional change (residing with the people), and the secondary constituent power to amend the constitution, this distinction is not very useful when it comes to assessing how the primary constituent power can be legitimately exercised within a specific constitutional setting. Indeed, the lack of clarity with regard to how the primary constituent power is ‘typically’ exercised, seems inconsistent with the very firm ideas regarding how it cannot be exercised. If the delegation theory’s answer to these questions is that the primary constituent power is an extra-legal power that cannot be constitutionally regulated, ¹⁰¹ then, it is submitted, this should also preclude it as a relevant ground for courts to invoke when exercising judicial review – which must, if the phrase is to retain any meaning, relate to the law.

Summarising the doctrinal approach

In this first part of the article it has been argued that doctrinal aspects of the doctrine of unconstitutional constitutional amendments have been somewhat neglected in scholarly debates, and that a dogmatic approach to the question of unconstitutionality can both provide new insights and reveal shortcomings of the more theoretical approaches that have dominated the field.

One such insight is that the common distinction between formal and substantive unconstitutionality is not sustainable. As an alternative, it has been proposed that a typology of unconstitutionality, based on features of constitutional design, may be used to elucidate how a doctrine of unconstitutional constitutional amendments can be justified in specific constitutional settings. The approach has also demonstrated that an application of the delegation theory to concrete cases of constitutional change exposes practical shortcomings of the theory.

It has, furthermore, been argued that even in instances of implicit substantive unconstitutionality – which is the core application of the delegation theory – the delegation theory can be subjected to the critique of being a mere conceptual possibility, and of facing internal inconsistencies.

The second part of this article will explore the feasibility of developing a doctrine of unconstitutional constitutional amendments within the Swedish constitutional setting.

A Swedish perspective on the doctrine of unconstitutional constitutional amendments

Brief introduction to the Swedish constitution

Sweden is a liberal constitutional democracy. One distinguishing feature of the Swedish constitution is the existence of not one but four constitutional (or fundamental) laws: the Instrument of Government [Regeringsformen], the Freedom of the Press Act [Tryckfrihetsförordningen], the Freedom of Expression Act [Yttrandefrihetsgrundlagen] and the Order of Succession [Successionsordningen].

The oldest of the constitutional laws is the Order of Succession from 1810. In the Order of Succesion the line of succession of the monarch is set out, and it is also provided, for example, that the monarch must be of the Lutheran faith.

The current Freedom of the Press Act was enacted in 1949, but the original act dates back to 1766. The Freedom of the Press Act contains detailed provisions regarding freedom of expression in printed media, mainly books, newspapers and journals. In 1991 the Freedom of the Press Act was supplemented by its sibling the Freedom of Expression Act, which mirrors the content of older act but is applicable to the use of freedom of expression via tv, radio and (to a limited extent) the internet. The Freedom of the Press Act also provides for a general right to access public documents.

The Instrument of Government, from 1974, contains most provisions that one would expect to find in a constitutional document. The 15 chapters of the Instrument of Government set out the basic rules governing parliament and government, constitutional rights, law-making, budgetary and treaty-making powers, the functions and competence of the courts and the administrative agencies, constitutional control mechanisms and emergency powers.

In Sweden, formal constitutional change has not come about as a result of revolutions or the collapse of constitutional continuity. Instead, constitutional change has, in principle, followed the established procedures for constitutional amendments laid down in the Instrument of Government and its predecessors. The Swedish experience can, therefore, be described as evolutionary rather than revolutionary. ¹⁰²

Formal amendment rules

Rules regarding the enactment of fundamental law (i.e. constitutional laws) are found in Chapter 8 Articles 14-16 of the Instrument of Government. Under these provisions a fundamental law is enacted by the Riksdag by two decisions of identical wording. The two decisions must be separated by a general election to the Riksdag, and at least nine months shall elapse between the first submission of the proposal to the Riksdag and the general election. ¹⁰³ There are no requirements pertaining to qualified majority or quorum rules. Hence, fundamental laws can be enacted by a simple majority vote. ¹⁰⁴

However, a referendum shall be held regarding a proposal to enact fundamental law if at least one-third of the Members of Parliament vote in favour of such a motion. A motion for a referendum must be submitted within 15 days of the initial decision of the Riksdag.

The referendum shall be held simultaneously with the general election separating the two Riksdag decisions. The proposal is rejected if a majority of those taking part in the referendum vote against it, and if the number of those voting against exceeds half the number of those who registered a valid vote in the election. Otherwise, the proposal goes forward to the Riksdag for final consideration.

Under Chapter 8 Article 18 of the Instrument of Government the same procedure applies to proposals concerning amendments to, or abrogation of, a fundamental law.

In essence then, the Swedish constitution can be amended, or completely revised, by two simple majority decisions by the Riksdag, separated by a general election. It should be noted that the possibility to subject a proposal for constitutional amendments to a referendum has never been activated since its inception in 1980. ¹⁰⁵

It is often remarked that the Swedish constitution is comparatively easy to amend. ¹⁰⁶ Whilst this is essentially correct, it should be noted that under a long-established praxis changes to the constitutional laws should have a broad political support in the Riksdag. Hence, even if the formal decision-making procedure only requires a simple majority, almost all changes to the Swedish constitutional laws, in fact, find support that would easily meet a requirement of two-thirds qualified majority. ¹⁰⁷ Regardless, it is a fact that the Swedish constitutional laws are amended with startling frequency. ¹⁰⁸

Notably, between the adoption of the 1974 Instrument of Government and 2003, almost half of the provisions were subject to amendments. The election in 2006 was the first which was not followed by amendments to the Instrument of Government – the reason for this was, perhaps, the upcoming thorough revision of the Instrument of Government that followed the 2010 elections. ¹⁰⁹ In essence, then, the Instrument of Government seems to be amended as often as the decision-making procedure allows for, i.e. following every general election.

To a certain extent it is fair to say that the high amendment rate is a product of the idea that the constitution should be up to date and accurately reflect what happens in practice. ¹¹⁰ Thus, formal amendment has become the preferred way of introducing even minor changes, that in other jurisdictions might have been brought about by way of interpretation.

Unconstitutionality: a doctrinal approach

As noted by Ojanen, the concept of constitutional unamendability appears almost weird in a Nordic constitutional context, and the issue has attracted only limited attention in Nordic constitutional scholarship. ¹¹¹ The same is, naturally, true for the idea of unconstitutional constitutional amendments.

Two potential reasons for this disinterest is the Swedish historic attachment to the concept of democracy, interpreted as parliamentary supremacy, and the restrained attitude towards judicial review. It should, however, be noted that both of these intertwined features of Swedish constitutional law are of declining importance – to a large extent due to the influence of European law. ¹¹²

But what scope does the Instrument of Government allow for courts to review constitutional amendments? The starting point for any such discussion must be the provision on judicial review in Chapter 11 Article 14 of the Instrument of Government, which reads: ¹¹³

If a court finds that a provision conflicts with a rule of fundamental law or other superior statute, the provision shall not be applied. The same applies if a stipulated procedure has been disregarded in any important respect when the provision was made.

In the case of review of an act of law under paragraph one, particular attention shall be paid to the fact that the Riksdag is the foremost representative of the people and that fundamental law takes precedence over other law. ¹¹⁴

The first section of the article provides that constitutional review shall be exercised in two different situations: (1) if there is a material, or substantive, conflict between a provision and a fundamental law; and (2) if a provision has been enacted in a formally, or procedurally, flawed manner. The second section, which highlights the importance the Riksdag, can primarily be seen as a reminder to the judiciary that constitutional review should be exercised with caution.

It is clear that the main gist of the article is to open up for courts to review whether a piece of ordinary legislation, or any other inferior legal norm, is in conflict with the substance of the constitution (e.g. introduction of the death penalty in violation of Chapter 2 Article 4 of the Instrument of Government), or suffers from procedural flaws (e.g. the requirement that legislative proposals must be remitted to relevant authorities and other stakeholders has been disregarded, in violation of Chapter 7 Article 2 of the Instrument of Government). But does Chapter 11 Article 14 permit courts to review constitutional amendments? There is no clear answer to this question in the preparatory works to the Instrument of Government and there are no court decisions dealing with the issue.

When it comes to procedural unconstitutionality, a plain reading of the article would seem to imply that review of constitutional amendments can be justified. Arguably, the question hinges on the meaning of the word ‘provision’ (in Swedish föreskrift) in Chapter 11 Article 14. It could be argued that the word refers to a provision of ordinary law (or another inferior norm) and not to a provision of constitutional law. This seems to be how the word provision is used in the first sentence of Chapter 11 Article 14 (‘a provision conflicts with a rule of fundamental law or other superior statute’). Were this to be the case, it would seem that formal review of constitutional amendments is ruled out.

However, the word provision ought to be interpreted in a structural manner, taking other parts of the Instrument of Government into account. In Chapter 8 Article 1 of the Instrument of Government it is provided that ‘provisions are adopted by the Riksdag by means of an act of law and by the Government by means of an ordinance’. It is further stated that the Riksdag and the Government may authorise state and local authorities ‘to adopt provisions’. In this article, it would seem that the word provision is used as a collective description for all legal norms within the Swedish internal legal system – including constitutional laws. Such an interpretation finds support in the preparatory works to Chapter 8 Article 1, where it is stated that in the Instrument of Government binding rules of law are referred to as provisions. ¹¹⁵ With this understanding of the word provision, it would seem that Chapter 11 Article 14 of the Instrument of Government does allow for judicial review of formal unconstitutionality.

Even if the issue has not been addressed in Swedish legal scholarship or in the jurisprudence of the Swedish Supreme Courts, it can, therefore, be argued that Swedish courts could refuse to apply a constitutional amendment on the basis that the amendment has not been enacted in accordance with the procedure set out in the Instrument of Government. It would indeed be unsatisfactory if the courts could not review the validity of a constitutional amendment which had been enacted in violation of the stipulated decision-making procedure, for instance if the Riksdag decided to amend the constitution by a single decision.

Notably, cases of competence-based formal unconstitutionality are not an issue when it comes to the review of constitutional amendments in Sweden. This is so, whilst the Instrument of Government allows for the enactment or amendment of constitutional laws under one unified procedure. Or, using the terminology of Albert, Sweden has a ‘comprehensive single track’ procedure for the enactment and amendment of constitutional laws. ¹¹⁶

Aside from the episode related in the prologue, the idea that a constitutional amendment could be conflict with the substance of the constitution seems never to have been contemplated in Swedish jurisprudence. Perhaps for good reason, seeing as there is scant support for reaching the conclusion that Swedish courts can set aside a constitutional amendment as being materially unconstitutional.

The wording of Chapter 11 Article 14 of the Instrument of Government makes explicit that substantive judicial review can only be exercised when an inferior legal provision is in conflict with a hierarchically superior legal provision. The article, in this respect, restates the essence of the lex superior principle. This entails that in order for substantive review of constitutional amendments to fall within the scope of Chapter 11 Article 14, one would have to argue that there exist an internal hierarchy within the Swedish constitutional provisions. Such an argument is, however, difficult to demonstrate.

First, the Swedish constitutional laws do not contain any provisions or principles that are explicitly unamendable. Second, there is only one procedure available for adopting constitutional amendments. Whereas several procedures to amend the constitution, with escalating levels of rigidity, can give credence to the argument that some constitutional provisions are hierarchically superior to others, no such distinctions exist in Swedish constitutional law. Third, it is explicitly stated in the Instrument of Government that the Riksdag can both amend existing constitutional laws and enact new constitutional laws using the same procedure. This implies that constitutional amendments have the same hierarchical position as the ‘original’ constitution.

If the words of the Swedish constitution are to be given serious consideration, this makes it rather difficult to argue that certain constitutional amendments ought to be out of bounds for the Riksdag.

Using the typology set out above it would seem that the Swedish constitution only recognises procedural unconstitutionality.

Unconstitutionality: the delegation theory

Turning to the delegation theory, it could be argued that it is still possible to find implicit limits regarding the kind of amendments that the Riksdag should be considered authorised to adopt – regardless of what follows from a textual reading of the Instrument of Government. The basis for the recognition of such implicit limits is, of course, the distinction between the primary constituent power of the people to enact a constitution, and the delegated secondary constituent power of the Riksdag to amend the constitution. In essence then, the Riksdag’s power to amend the constitution should, according to the delegation theory, be limited to such amendments that do not change the core, or the basic structure, of the constitution.

Any such approach would, as noted above, have to disregard the fact that the Instrument of Government makes no distinction between the primary constituent power to enact a constitution, and the secondary constituent power to amend the existing constitution. Rather, the sole power of constitution-making and amendment rests with the Riksdag.

Such an approach would also have to disregard the particular features of the Swedish evolutionary version of constitutionalism, which traditionally has been founded on ideas of a political (rather than legal) view of the constitution and a conception of democracy as parliamentary government. In the Swedish constitutional context, then, the idea of popular sovereignty – or constituent power – is intimately connected to the concept of representative government. ¹¹⁷ This, among other things, manifests itself in the fact that the Riksdag has the power both to enact a new constitution and to amend the constitution, using the same decision-making procedure.

Arguably, it makes little sense to speak of constituent power as a legal criterion for the validity of constitutional provisions within the Swedish context, at least in the meaning that is afforded the concept in the delegation theory. Rather, the Swedish view of the Riksdag as having unlimited powers to enact (and fundamentally change) the constitution seems to bear a resemblance to the doctrine of parliamentary sovereignty as found in the UK. ¹¹⁸ In a sense, then, the idea of a constituent power unbound by any prior constitutional rules is prevalent in Swedish constitutional law. But instead of vesting the people with this power, in accordance with the delegation theory, the Swedish view seems to be that constituent power lies in the hands of the Riksdag.

However, in my view, the notion that the Riksdag would enjoy unlimited power is open to similar conceptual critique that I have voiced against the delegation theory’s notion that the people have extra legal constitution-making powers. ¹¹⁹ The reliance on sovereignty and constituent power as legal concepts appear to be dead ends.

Summarising the Swedish perspective

One characteristic feature of Swedish constitutional law is its evolutionary development, from autocratic rule to a modern liberal constitutional democracy: a development that has predominantly taken place within the formal boundaries of the existing constitutional laws, that is, without any significant legal rupture. ¹²⁰

Another essential feature of the Swedish constitution is its focus on popular sovereignty in the form of a parliamentary form of government and – as an implication thereof – a restrained attitude of Swedish courts towards judicial review, which is fading in the face of the influence of European law.

In brief, the Swedish constitution would seem to grant the Riksdag unlimited power to amend the constitution, or even to enact a new constitution. Furthermore, the prescribed procedure for constitutional amendments, or enactment, is not very rigorous: two simple majority decisions by the Riksdag, separated by a general election (with the, never used, possibility of a referendum).

The Swedish constitution seems to envisage the possibility for courts to review procedurally unconstitutional constitutional amendments, but not to review other forms of unconstitutionality. The delegation theory of unconstitutional constitutional amendments seem foreign to the Swedish constitutional system, which may be connected to fact that the Swedish constitution makes no distinction between the primary constituent power to enact a constitution and the secondary constituent power to amend a constitution. Both of these powers can, in the Swedish context, be exercised on identical terms by the Riksdag.

Concluding remarks

Theoretical justifications for legal doctrines are essential. The delegation theory of unconstitutional constitutional amendments provides an impressive first comprehensive account of how to justify the position that procedurally immaculate constitutional amendments may still be viewed as unconstitutional. However, as has been demonstrated in this article, the delegation theory’s aspirations of providing a globally applicable and coherent theoretical framework for the doctrine of unconstitutional constitutional amendments can be questioned. ¹²¹

The main aim here has been to subject theory to constitutional law’s doctrinal realities. First and foremost, this means paying due regard to the constitutional text as found in specific constitutional settings. This approach has revealed that, to a substantial part, the doctrine of unconstitutional constitutional amendments can be justified without relying on the delegation theory’s distinction between constituent and constituted powers. The proposed typology of unconstitutionality highlights the importance of paying attention to constitutional design when applying the doctrine of unconstitutional constitutional amendments in specific jurisdictions.

Even so, it cannot be disregarded that the question of implicit substantive unconstitutionality will not find any conclusive answer by relying on constitutional exegesis. Indeed, the concept itself implies that answers will have to be found elsewhere. In this respect, some potential shortcomings of the delegation theory have been addressed here. It has been demonstrated that the delegation theory’s reliance on the concept of constituent power can be criticised as being both unnecessary and internally inconsistent: unnecessary, because it is quite possible – and probably even preferable – to justify implicit unamendability on other grounds than constituent power theories; internally inconsistent, because the delegation theory seems to conflate the people as being the bearers of constituent power with the question of how constituent power can legitimately be exercised.

Using the constitutional setting of Sweden as an example, it has further been suggested that the distinction between constituent and constituted powers may appear artificial in constitutional democracies whose historical experiences can be characterised as evolutionary, rather than revolutionary, and where the concept of democracy is closely related to a parliamentary form of government.

However, the Swedish view, that parliament has unlimited powers to enact (or fundamentally change) a constitution, seems theoretically as unconvincing as the delegation theory’s view that the primary constituent power of the people is an extra legal authority, unbound by any legal rules. Whereas the Swedish view seems to completely negate the existence of implicit unamendability, the delegation theory provides an account of implicit unamendability that is both doctrinally and theoretically unconvincing. This begs the question: what is the alternative?

Ordinarily, the doctrinal approach outlined in this article will suffice to disentangle questions of whether constitutional amendments are unconstitutional. That is, under ordinary circumstances there should be wide acceptance of the position that constitutional amendments are valid, provided that they have been enacted in accordance with the procedure set out in the constitution and do not contradict explicitly unamendable rules or principles.

The thorny question, of course, is what to make of amendments that can be said to violate core democratic or rule of law principles without transgressing any explicit limits provided for in the constitutional text. There is no doubt, in my mind, that there is in principle a strong case for accepting at least a limited doctrine of unconstitutional constitutional amendments under such circumstances. Whether, and how, that case can be translated into legal doctrine is bound to be subject to reasonable disagreement. In other words, it is unlikely that any one ‘ultimate justification’ for implicit unamendability will find acceptance across jurisdictions.

In response to this, it could be argued that the delegation theory is simply one justification, which may find support in some jurisdictions but not in others. To that I would respond that as a legal doctrine the delegation theory is both arbitrary and unnecessary – which I have tried to demonstrate in this article. Arguably, one major shortcoming of the delegation theory as a legal doctrine is that it claims to provide a ‘definitive’ answer to the question of what gives the constitution legal validity, by relying on a specific and highly contested theory of constituent power. Essentially, what the delegation theory proposes is a criterion for validity of the legal system. On such a profound legal philosophical issue, there is bound to be reasonable disagreement. It could also be argued that reliance on any such contested theory exposes legal doctrine to inevitable critique, since there is no way to determine whether any such theory is ‘correct’. ¹²²

Bearing this in mind, it is preferable to ground a legal doctrine of implicit unamendability on other factors. One such approach has been outlined by Dixon and Landau. ¹²³ They propose that a unconstitutional constitutional amendment doctrine should be limited to cases where an amendment threatens to erode democracy, and that this determination can be aided by engagement with transnational constitutional law – that is, taking on board institutional practices and jurisprudence in other democratic constitutional systems. On a similar note, it has been suggested that norms constituting ius cogens, or international or regional human rights law, could be used to justify implicit unamendability. ¹²⁴

These approaches are, much like the delegation theory, conceptual possibilities (not necessities), and need to be the subject of further elaboration within specific constitutional contexts. In my view, however, they share the advantage of having the potential to find support in existing legal doctrines in most democratic constitutional systems, in the sense that such systems have far-reaching constitutional commitments to democracy and fundamental rights – something which is arguably not the case when it comes to the concept of constituent power proposed by the delegation theory.