A Political Legitimacy, Public Reason and the Constitution

Political legitimacy distinguishes the sort of government to which a democracy such as Australia aspires from rule by force or acquiescence. The democratic state and its framework of laws and government are legitimate when they are appropriate exercises of citizens’ collective power.^{Footnote 9} (We use the notion of the state and its framework of laws in a capacious sense, as including the private law, and so the legitimacy with which we are concerned is dependant (in part) on the private law appropriately constraining and channelling the conduct of private actors).^{Footnote 10}

To help understand what makes an exercise of power appropriate, it is necessary to contrast legitimacy with justification.^{Footnote 11} Complying with a framework of laws and government may sometimes be a good or right thing, and hence justified, even if that framework is not legitimate. A well-known example would be compliance with the law that prohibits murder, which (roughly) tracks a moral duty not to commit unjustifiable homicide. Similarly, a regulatory regime may generate independent moral duties, regardless of its legitimacy, because of the contribution it makes to coordination solutions.^{Footnote 12} More generally, morality or prudence may demand some compliance with an illegitimate framework because total non-compliance would produce worse consequences (eg, disruption or disorder).

Just as importantly, there may be cases where a framework of laws and government is legitimate, although not fully justified. ‘No human institution can be [perfectly just]’,^{Footnote 13} and thus ‘any feasible political procedure may yield an unjust outcome’.^{Footnote 14} Hence, as Rawls puts it, we must sometimes ‘[accept] as legitimate (even when not just) a particular statute or a decision in a particular matter of policy’,^{Footnote 15} and must ‘recognize unjust laws as binding provided that they do not exceed certain limits of injustice.’^{Footnote 16}

Rawls’ account of how those limits are established, and the role of the constitution in that respect, is of particular interest. In A Theory of Justice, Rawls says that, ‘We must rely on the actual course of discussion at the legislative stage to select a policy within the allowed bounds.’^{Footnote 17} The ‘allowed bounds’, in turn, are settled by what he describes as ‘the framework of the public conception of justice’:

Political Liberalism elaborates upon the idea of ‘the framework of the public conception of justice’, using the term ‘public reason’, and henceforth we adopt this usage. Rawls describes public reason as ‘the reason of citizens as such’, that is in their capacity as citizens rather than as members of private associations and bearers of personal concerns and commitments. The ‘nature and content’ of public reason is ‘given by the ideals and principles expressed by society’s conception of political justice’.^{Footnote 19} Thus, legislation and governmental decision-making are legitimate even if (arguably) not just, binding even those who disagree, if the latter cannot convincingly establish their objection within public reason. This is how the ‘allowed bounds’, the limits of injustice that are consistent with a legitimate framework of laws and government, are established.

The inability to establish an objection within the scope of public reason does not mean that public reason entails the contentious law or decision. Determining whether or not some institutional arrangement or policy conduces to justice ‘frequently depends upon speculative political and economic doctrines and upon social theory generally’,^{Footnote 20} and these are matters of reasonable disagreement. A law or decision that depends for its justification upon such ‘speculation’ will therefore not be entailed by public reason. Consider, for instance, a compulsory education law: it may be unequivocally aimed at providing all children with opportunities, but in its details may presuppose how certain sorts of compulsion (e, zoning, curriculum, etc) affect the development of those opportunities. Such presuppositions would be examples of ‘speculative’ doctrine. Those who disagree about how children develop opportunities in the context of their schooling might therefore oppose the law on the grounds that it will not in fact produce a just distribution of opportunities. But as long as the speculative doctrine itself conforms to public reason then the law cannot be refuted within public reason (e, so-called ‘race science’ would be excluded on this measure because inconsistent with the basic principle of justice of Australian or any other society; but mainstream psychology and sociology would not be).

We can now explain the contribution that a democratic society’s constitution makes to legitimacy. Rawls describes the principal demand on constitution-makers as being ‘to choose … the constitution that satisfies the principles of justice and is best calculated to lead to just and effective legislation’ given the ‘general facts about their society’.^{Footnote 21} The constitution contributes to legitimacy in two ways. The first is by establishing the legislative process: the legislature enjoys its power to generate authoritative law, that may be legitimate even if not just, because of its special constitutional position, of having been adopted as the best calculated means of achieving just and effective legislation.^{Footnote 22} It is this, and not (for example) special insight or expertise, that distinguishes the legislature from any other group of citizens that meets to discuss and debate political matters.

The second is by contributing to the ‘allowed bounds’. As Rawls notes, legislation ‘must accord both with the constitution’ and with the principles of justice to which the constitution aims to give effect.^{Footnote 23} For the reasons considered above, there is no way of guaranteeing that legislation will be just. But a society’s constitution is part of that society’s expression of its conception of political justice, and thus contributes to public reason.^{Footnote 24} We call this constitutional public reason. Constitutional public reason may be expressed in various ways, such as specifying rights, setting out directive principles (discussed below), or conferring powers (as we argue is the case for the Australian Constitution). A law that can be refuted within constitutional public reason is not legitimate: it is not defensible as an expression of the public conception of justice, and it cannot claim to be supported by the constitution. Thus, the special role that the legislature enjoys by virtue of its constitutional position cannot be called upon in support of such a law. From the perspective of legitimacy, the status of such a law is no different from an expression of the legislators’ private whims or interests. Whether crossing the bounds established by constitutional public reason ought to be remediable by way of judicial review is a further question of constitutional design. We agree with Rawls that philosophy provides no general answer to this question.^{Footnote 25}

This account of legitimacy puts the process of political argument, and political reasoning, at its centre, but does not accept that political argument and political reasoning are unconstrained: they are constrained by the requirements imposed by public reason, including constitutional public reason. The account is quite different, for instance, from Raz’s account of practical authorities, which grounds such authority, especially in the political case, in the authority’s capacity to resolve coordination problems and prisoners’ dilemmas, and to identify and proclaim ‘dependent reasons’ (that is, reasons that are meant to reflect the balance of reasons that already apply to those subject to the authority) that will better enable its subjects to conform to the reasons that already apply to them.^{Footnote 26} Our account, which grounds authority on occupying a certain institutional role in the constitutional scheme while having the appropriate orientation to public reason, is a better fit with Australia’s constitutionally mandated scheme of electoral democracy than one which grounds authority on a distinctive capacity. This is for at least two reasons: it does not demand any particular wisdom or expertise of political decision-makers; and it makes the special role of the legislature a matter of principle, rather than of being simply a salient provider of coordination and like solutions.

Legitimacy, and the lack of it, may be matters of degree, in two senses. First, it may be that most but not all of what the state does and demands — its laws, the decisions made and actions taken by its government, etc — may be legitimate. So illegitimacy may be local rather than widespread or total. Second, failures of legitimacy may be more or less egregious.^{Footnote 27} Constitutional public reason may play a role in making these assessments.

B Political and Legal Constitutionalism, and Constitutional Public Reason

The relationship between political constitutionalism, legal constitutionalism, and constitutional public reason, is a complicated matter. Using Rawls as our foil, we identify a number of ways — several more than Rawls envisages — in which they may interrelate.

Rawls identifies one possible set of such relationships when he describes the Supreme Court as an exemplar of public reason.^{Footnote 28} He says that:

This argument rests on four premises. One of them — that a constitution contributes to public reason — we agree with, for the reasons we have already given. But the other three premises may be rejected.

The first of these is that political actors do not have to justify their actions by reference to an understanding, or at least a systematic understanding, of the constitution. This rests on Rawls’ assumption that a constitution will not have content extending beyond protection of civil and political rights, and the basic machinery of democratic government (ie elections, office-holders, etc).^{Footnote 30} Rawls assumes that although principles of economic justice are part of public reason,^{Footnote 31} they will not be part of the constitution,^{Footnote 32} and hence will not inform constitutional public reason (he notes the 16th Amendment to the Constitution of the United States, permitting a national income tax, but describes it as a dealing with a ‘basic [matter] of policy’^{Footnote 33} without considering its possible implications for the role of public reason in non-judicial political discourse). This is not because securing social and economic justice is less important than the matters dealt with by the constitution, and nor is it for anti-perfectionist reasons:^{Footnote 34} Rawlsian political liberalism is fully committed to ensuring an adequate distribution of all-purpose means, of just the sort that many (not all) constitutional directive principles are concerned with.^{Footnote 35} It is because securing social and economic justice requires knowledge of particular and concrete social and historical circumstances, which (according to Rawls) makes its attainment harder to ascertain without controversy;^{Footnote 36} in other words, it is said to be a consequence of the fact, discussed above, that determining whether some institutional arrangement or policy conduces to social and economic justice ‘frequently depends upon speculative … doctrines’.^{Footnote 37}

Rawls asserts that the content of rights is not contentious in the same fashion as ‘speculative’ doctrine, but this is an empirical claim for which no evidence is provided beyond the dubious claim that the list of protected rights is much the same in all democratic constitutions.^{Footnote 38} Furthermore, Rawls includes within the scope of constitutional rights not just formally equal political liberties, but political liberties of equal worth,^{Footnote 39} and these depend (in part) upon the distribution of wealth,^{Footnote 40} and a ‘social minimum’.^{Footnote 41} It is quite unclear how a constitution can secure the equal worth of political liberties, and a social minimum, if it leaves questions of social and economic justice to be resolved via an unguided process of legislation. We therefore reject Rawls’ arguments in favour of such thin constitutional content.

A further reason why Rawls assumes that politicians do not have to justify their actions by reference to a systematic understanding of the constitution is that he envisages the constitution being protected by judicial review.^{Footnote 42} This relieves political actors from the need to systematise their own understanding of constitutional requirements. However, this assumption simply begs the question in favour of legal rather than political constitutionalism. We therefore treat it as an open question whether political actors are obliged to have regard to constitutional public reason. This will depend on the details of a society’s constitution and political culture. We take this up below, first in relation to constitutional directive principles, and then in relation to Australia.

The second premise that may be rejected is that constitutional public reason plays the role that it does (at least in part) in virtue of its legal meaning. The third is that legal meaning, which the judiciary is obliged to draw upon, develop and apply, is itself informed by elements of public reason that lie outside constitutional public reason (that is, the broader public conception of justice).

Accepting these three premises yields a picture of constitutionalism very similar to Ackerman’s:^{Footnote 43} the people adopt a constitution, a ‘higher law’ that expresses and is informed by their public political conception of justice, which then governs ‘ordinary’ politics by setting up (i) the democratic institutional machinery, and (ii) agreed boundaries to protect democratic citizenship in the form of fundamental rights, which are enforced by the courts. The courts, in playing this role, develop the legal meaning of the constitution by reference to the public conception of justice. On this picture, the Supreme Court is thus

Thus, says Rawls,

On this understanding, as the content of public reason changes, so the legal meaning of the constitution and the content of constitutional public reason changes (or should change) with it, without necessarily requiring formal amendment.^{Footnote 46}

Rawls identifies two other possibilities: a system of parliamentary supremacy with no bill of rights, in which public reason has no constitutional expression at all; and a German-like system, in which certain rights are entrenched and unamendable, and hence beyond the control of the people.^{Footnote 47} But these are not exhaustive. A constitution may contain content that goes beyond basic democratic machinery and protection of civil and political rights. It may also contain content that is addressed not to the courts, to be given legal meaning and applied via a process of legal reasoning, but rather to politicians, to help establish the ‘allowed bounds’ of legislation and other political decision-making. Constitutional directive principles, discussed in the next sub-section, exemplify both these possibilities. By making economic justice a subject of constitutional public reason, such arrangements may do better at helping ensure a just constitutional order than Rawls’ own picture.

It may also be possible to have a regime in which legal meaning, even of the constitution, is insulated from public reason. This would be a system, at least in aspiration, of ‘strict legalism’. Rawls, in a relatively early article, sets out a general framework for understanding why this might be desirable.^{Footnote 48} Rawls points out that what it is that justifies an institution or practice (for instance, that it tends to secure worthwhile ends) differs from what it is that justifies an action within an institution or practice (for instance, that it conforms to the rules that govern the role that a particular actor occupies within that institution or practice). Sometimes, the best way to ensure that an institution or practice serves the purpose for which it has been instituted is to govern those within it according to rules that do not have express regard to that purpose. A simple example is the institution of promising: it is at least arguable that this institution serves a general social purpose, of facilitating trust and hence cooperation among people; but what helps promising achieve this purpose is precisely that promisors are precluded from having regard to broader social considerations in deciding whether and how to keep their promises. Rather, the only relevant considerations (except in the direst circumstances) are those which are personal to the giver of the promise and the person to whom it was given.^{Footnote 49}

In a legalist system, therefore, the constitution might conceivably play two distinct functions. It might act as a source of constitutional public reason for political decision-makers, contributing to the establishment of the ‘allowed bounds’. At the same time, it might be interpreted and applied by courts, in accordance with its legal meaning, which is determined by the application of technical legal methodologies which do not treat public reason as weighing on the scales. This, we assert, is the basic character of the Australian Constitution. We make the argument for this claim in Sections III and IV.

C The Example of Directive Principles

Constitutional directive principles are best known perhaps from the Irish and Indian constitutions but are also found in many others. They require the political mechanisms of government (particularly the legislature) to pursue certain goals,^{Footnote 50} and ‘are not designed to be given effect by direct judicial enforcement’.^{Footnote 51} Thus, s 37 of the Constitution of India states that ‘provisions contained in this Part’ — entitled ‘Directive Principles of State Policy’ — ‘shall not be enforceable by any court, but the principles therein laid down are nevertheless fundamental in the governance of the country and it shall be the duty of the State to apply these principles in making laws’. Section 38 states the following principle: ‘[t]he State shall strive to promote the welfare of the people by securing and protecting as effectively as it may a social order in which justice, social, economic and political, shall inform all the institutions of the national life’.

Khaitan and Weis both identify directive principles as primarily a political constitutional device.^{Footnote 52} As Weis observes, constitutional directive principles:^{Footnote 53}

Khaitan says more about what this complication consists in. Constitutional directives demonstrate that political constitutionalism, as an institutional mechanism for the pursuit of constitutional government, need not entail that the constitution does nothing but establish institutional machinery:^{Footnote 55}

Furthermore, Khaitan recognises that constitutional directives directly inform constitutional legitimacy. Being posited (either expressly or by implication) by the constitution,^{Footnote 56} they ‘set the agenda for the political discourse, and are used by political actors to legitimise or criticise certain acts and agendas.’^{Footnote 57}

The framework we have set out above straightforwardly explains these features of directive principles. As noted above, they are an example of constitutional content that goes beyond basic democratic machinery and rights protection, but that is addressed primarily to politicians. They are a part of public reason, and moreover of constitutional public reason, in those societies in whose constitutions they are found. But their primary purpose is not to contribute to the legal meaning of the constitution as applied and interpreted by the courts.^{Footnote 58} Rather, they contribute to the ‘allowed bounds’ of political, including legislative, decision-making, by creating obligations, ^{Footnote 59} thereby setting boundaries which cannot be crossed if the state and its framework of laws and government are to be legitimate.

The Australian Constitution contains statements of obligation addressed to the legislature. There are provisions mandating certain institutions (eg, the Inter-State Commission)^{Footnote 60} that cannot come into being except by legislation,^{Footnote 61} and also express imperatives to legislate, most prominently s 88 which states that ‘[u]niform duties of customs shall be imposed within 2 years after the establishment of the Commonwealth’. There is no adjudicative mechanism for enforcing these provisions: it is a basic principle of Australian constitutional law that Parliament is immune from (purported) orders mandating that it enact a certain law.^{Footnote 62} During the Convention that finalised the Australian Constitution, George Reid (then Premier of New South Wales, later the fourth Prime Minister of Australia) nevertheless described s 88 as a ‘legal compulsion’ to establish a uniform tariff.^{Footnote 63} On the other hand, Quick and Garran describe it differently:^{Footnote 64}

Our argument does not require determining whether or not these provisions impose legal obligations on the Parliament. What is relevant for our purposes is that they do not contribute any substantive content to constitutional public reason. For similar reasons, Khaitan would not characterise s 88 of the Australian Constitution as a directive: while it requires Parliament to enact legislation, it does not specify any substantive content.^{Footnote 65} As is well-known, s 88 represented a decision to postpone the making of any substantive decision about the content of the tariff law, and to allow it to be determined by the political process that the Australian Constitution was to establish.^{Footnote 66}

There are features of the Australian Constitution, however, that are not neutral as to the course of action to be pursued. Thus, while Khaitan identifies the Australian Constitution as the ‘best approximat[ion]’ to ‘the purest form of proceduralist political constitutionalism’,^{Footnote 67} we disagree.

D Conferrals of Power and Constitutional Public Reason

Goldsworthy and Burton Crawford state that political constitutionalism in the context of the Australian Constitution ‘uses political mechanisms to control government’ and observe that:^{Footnote 68}

However, the absence of legal constraints, including constraints amenable to judicial review, does not mean that the constitution imposes no requirements on politicians. Constitutional directive principles are a counter-example to such an assumption. So too, we argue, are the powers in the Australian Constitution with which we are concerned.

It may seem puzzling to suggest that a conferral of power can impose requirements and generate constraints. Khaitan, for instance, illustrates his claim that ‘[d]irectives are … distinct … from constitutional provisions that are non-obligatory’^{Footnote 69} by reference to the conferral of power on Canadian provinces to ‘make Laws in relation to Education’.^{Footnote 70} However, while the textual difference between a conferral of power and an imposition of a duty is apparent, this does not preclude a conferral of power from generating a constitutional requirement. A conferral of power need not be a bare power and may bring with it an implication as to whether and how it is to be used.

A relatively pedestrian example is s 64 of the Australian Constitution. This confers a power on the Governor-General ‘to appoint officers to administer … departments of State’ who must be drawn from the Parliament and who ‘shall be members of the Federal Executive Council’ that is to advise the Governor-General.^{Footnote 71} Although in strictly textual terms a power, s 64 establishes a constitutional requirement that the Governor-General appoint a government that enjoys the confidence of the House of Representatives (the core of responsible government).^{Footnote 72} It is debatable whether this is a justiciable, or even legal, requirement;^{Footnote 73} but it is clearly a constitutional one, and departing from it can — and in 1975 did — provoke a significant crisis of legitimacy.

Khaitan recognises that constitutional conventions may establish non-justiciable and non-legal constitutional obligations, but takes them to differ from directive principles because they are concerned with procedural questions (eg, about assenting to legislation and appointing governments) rather than substantive goals of state action.^{Footnote 74} Expressed in our terms, procedural conventions do not contribute substantive content to constitutional public reason. But these sorts of conventions are not the only requirements that can be generated by the constitutional conferral of powers: a constitution may confer powers in circumstances where there is a clear political understanding, associated with the introduction of the power into the constitution in which it is found, that such powers are to be exercised, or that the purposes for which the powers were conferred are otherwise to be pursued. This will give rise to a corresponding requirement within constitutional public reason.

We are not suggesting that such a requirement implies any obligation to exercise the power as a matter of legal meaning. But as we explained above, constitutional public reason need not be a source of legal meaning in order to contribute to the ‘allowed bounds’. The United States Constitution provides two candidate illustrations of such powers: the 14th and 15th Amendments provide for civil rights and voting rights respectively, and each concludes with a statement that ‘[t]he Congress shall have the power to enforce [it] by appropriate legislation’. These seem to be not merely conferrals of power upon Congress. They establish an expectation that Congress shall use its legislative powers, including but not limited to those conferred by the Amendments,^{Footnote 75} to ensure that civil and voting rights are secured. A parallel Australian example would be the 1967 referendum, which established the welfare of, and perhaps justice for, Aboriginal and Torres Strait Islander peoples, as a requirement within constitutional public reason.^{Footnote 76}

In a case of this sort, failure by the legislature to exercise the powers conferred upon it will generate two bases for criticism. First, laws that do not conform to the expectation generated within constitutional public reason by the conferral of the relevant legislative powers will thus be liable to refutation within public reason, falling outside the ‘allowed bounds’, and therefore lacking legitimacy. Second, and more profoundly, such failure may cast doubt on the legitimacy of the legislature itself, either because it has forfeited its claim to be legislating within the ‘allowed bounds’ in general, or because its failure to pursue justice as is constitutionally required raises doubts about the legitimacy of the political process that the constitution provides for, severing the connection between the constitution and justice. Keeping the example of the 1967 referendum in mind, there is an argument that one or both sorts of failure have occurred in Australia, given that Australia’s framework of law and government does not secure either welfare or justice for Aboriginal and Torres Strait Islander peoples. This article does not further explore the implications of the resulting lack of legitimacy for the Australian constitutional order,^{Footnote 77} but we do note that our framework clearly enables it to be identified.

A Income Justice

As the introduction explained, our analysis of the income justice powers takes place entirely within the domain of political constitutionalism. The income justice powers conferred by the Australian Constitution, we assert, establish expectations that the Parliament will use its legislative powers to secure income justice, and hence a requirement in constitutional public reason that it do so, although that requirement does not inform the legal meaning of the Australian Constitution.

In his closing address to the drafting Convention, Edmund Barton spoke in support of the draft constitution. He adduced ‘evidence of the want of foundation of accusations against this Bill on account of its alleged illiberal character’ and this included ‘some very important … powers’ such as ‘the power to legislate with reference to invalid and old-age pensions’ and ‘the right to legislate for the appointment of courts of conciliation and arbitration in industrial disputes which extend beyond the limits of one state’.^{Footnote 78} Barton thus identifies these powers as part of what makes the Australian Constitution a ‘liberal’ one, by which he clearly intends that it aims at justice; and part of what ensures that a framework of laws and government established under the Australian Constitution will fall within the ‘allowed bounds’ is this capacity of the Parliament to legislate in pursuit of income justice. This understanding is reinforced by what seems to be these provisions’ obvious logic in the constitutional scheme: s 92 provides for a national free trade area; s 51(xxxi) protects property rights, reinforcing the character of the society as a liberal market one; and the national legislature is conferred with the power to ensure that income justice within that national market economy can be secured.

Furthermore, Barton adduces these powers as fundamental features of the Australian Constitution, despite having voted against their inclusion,^{Footnote 79} and he went on to emphasise both in his policy as Australia’s first Prime Minister.^{Footnote 80} This demonstrates an acceptance of their place in constitutional public reason, as part of the shared conception of justice to which the Australian Constitution contributes.

The place of these powers in the public conception of justice is something that has endured since Barton’s time. This article does not canvass in detail the Australian political history that establishes this point, but points to two key pieces of evidence. First, wage-fixing by way of industrial arbitration was central to Australian economic policy from the celebrated Harvester judgment until what Kelly describes as the end of the ‘Australian Settlement’ in the 1980s,^{Footnote 81} which saw the emergence of new legal mechanisms for establishing minimum wages. Second is the 2007 election, in which the central issue was the industrial relations regime known as ‘WorkChoices’ and that regime’s effect on those new mechanisms. Within the legal constitutionalist domain, the bulk of WorkChoices comfortably survived a constitutional challenge that turned on technical legal questions as to the legal meaning of key constitutional provisions (in particular s 51(xx) (the corporations power), including whether its legal meaning is affected by that of s 51(xxxv)).^{Footnote 82} But in the domain of political constitutionalism, the debates around WorkChoices showed the enduring importance, within public reason, of expectations about the Parliament’s role in securing income justice, an expectation which has its foundation in the contribution of the income justice powers to constitutional public reason.

This also suggests a contrast with the political dynamics of constitutional directive principles. Weis and Khaitan associate directive principles with ‘transformative’ constitutionalism’.^{Footnote 83} Khaitan, in particular, contends that: ‘a directive will also lack normative force in politics if its goal has largely been achieved … [D]irectives are politically useful only when they add extra weight to a political agenda by borrowing some of the constitution’s legitimacy’.^{Footnote 84}

Whether or not Khaitan’s statement is true of directives, it does not seem to be true of those elements of constitutional public reason that follow from conferrals of power such as the income justice powers. As Rawls says, it is the role of fundamental institutions to ‘secure’ and to ‘maintain’ justice.^{Footnote 85} In Australia, this is done primarily by legislating. Once legislation is enacted, if its repeal could be shown to be contrary to constitutional public reason, then that repeal would not fall within the allowed bounds, and the extant legislation would enjoy the support of public reason even were its function one of maintaining rather than transforming. This is one way in which legislation may take on constitutional significance: it can give effect to a constitutional requirement grounded in a conferral of power, such that its repeal (in the absence of a suitable replacement) would be illegitimate.^{Footnote 86}

Our final point in relation to the income justice powers concerns both the methodology and the conclusion of our argument. As the introduction stated, this article does not defend a particular conception of economic justice in general, nor income justice. We therefore do not attempt to specify where the ‘allowed bounds’ might lie in relation to income justice in Australia beyond what was said in the introduction: legislation will fall outside those bounds if it allows differences of income that undermine political equality or fair equality of opportunity. Furthermore, recall our introductory remarks about the importance of social processes to the realisation of constitutional requirements: when constitutional public reason operates in the domain of political constitutionalism, it is a task for actors in the political realm to articulate and deploy the content of constitutional public reason in defending and criticising legislation. What we have shown is that the Australian Constitution contains provisions that make a distinctive contribution to this, by keeping the demands of income justice at the forefront of debates in Australia about whether political decisions fall within the ‘allowed bounds’.

B Taxation in the Domain of Political Constitutionalism

The importance of taxation to securing economic justice has long been recognised. Tax provides the primary means of funding to secure income justice in the context of a liberal market economy and is an important method of breaking up concentrated holdings that undermine democracy.^{Footnote 87} However, the failure to translate this recognition into requirements (including constitutional requirements) for the design of the tax system is widely acknowledged.^{Footnote 88} Sugin notes that ‘Rawls wrote a great deal about economic justice generally, but very little about taxation in particular, and what he did say is puzzling’.^{Footnote 89} Moreover, much of the ‘speculative’ doctrine in taxation derives from economics which, over the course of the past half-century, has been dominated by a mode of analysis that favours tax policies that are as neutral as possible to market-based actions and outcomes, on the presumption that markets are the best way to maximise individual and therefore collective welfare. Therefore, tax policy has moved away from a broad and progressive income tax, arguing that it is too distortive, and towards taxing consumption, such as through a value-added tax, on the assumption that this interferes less with market outcomes.^{Footnote 90} This has had profound distributional implications — it is no coincidence that wealth and income inequality has increased in Australia over the near three-decade-long period of economic growth up to 2018.^{Footnote 91}

As the introduction said, we consider the taxation power from the perspective of both political and legal constitutionalism. This illustrates how the Australian Constitution involves an interplay between political constitutionalism, in which constitutional public reason operates, and legal constitutionalism, which is a domain of technical legal reasoning largely insulated from public reason. From the political point of view, taxation serves three main functions: raising revenue; breaking up holdings of wealth; and influencing behaviour. The last of these functions may support or undermine economic justice, depending on the behaviour in question. The first two are essential in securing economic justice: revenue provides the means to fund social services spending, including direct income support, public health, and education. Breaking up concentrations of wealth reduces the distorting effect these have on political and civic equality. Our previous work has argued that, under contemporary conditions, the income tax is the most effective way to realise these two functions because, properly designed and implemented, it is the only mass revenue-raising tax (in contrast to a wealth tax) that is progressive in its impact before considering public spending (in contrast to the value-added tax).^{Footnote 92}

In Australia, taxation is a concurrent power of the Commonwealth and the States, but the Commonwealth dominates the raising of revenue. This largely results from a series of decisions about the legal meaning of the Australian Constitution that have given the Commonwealth exclusive de facto power to levy income taxes^{Footnote 93} and expansively interpreted the Commonwealth’s exclusive power to impose customs and excise duties so as to effectively exclude the States’ levying a general sales tax.^{Footnote 94} This centralisation of revenue-raising power has led to the well-documented problems of vertical fiscal imbalance whereby, the Commonwealth dominates the raising of revenues while the states are responsible for much spending.^{Footnote 95} For our purposes it is sufficient to note that this state of affairs makes proper use of the Commonwealth’s legislative power to impose taxation, and particularly income taxation, central to economic justice in contemporary Australia.

This does not, however, mean that the taxation power contributes to the content of constitutional public reason in the same way as the income justice powers do. As Section II explained, the content of public reason is a matter of political history and culture. As the previous sub-section explained, the significance of the income justice powers in this respect is established by the Convention debates and the prominence given to these powers by Barton, which has endured since then. We do not know of any evidence that the tax power has played, or does play, the same role. It is true that in both 1993 and 1998, the central issue in the national election was a proposal to introduce a broad-based consumption tax.^{Footnote 96} However, we think that this illustrates a different, though still significant, relationship of the tax power to constitutional public reason within the domain of political constitutionalism: a regime of tax legislation that is not adequate to ensuring legitimacy in respect of income justice, and perhaps economic justice more broadly, will itself be able to be refuted within public reason, precisely because of that inadequacy. Thus, the successful 1998 campaign for the GST was supported by other policy changes such as increases in certain welfare payments (which entailed public debate over the adequacy (or otherwise) of the increases in these payments, that is, whether they fell within the ‘allowed bounds’).^{Footnote 97} (The same point would apply to a fully confiscatory tax system, which would be refuted in constitutional public reason by reference to those features of the Australian Constitution, noted above, which clearly mandate a liberal market economy.)

This point about the instrumental significance of taxation within constitutional public reason extends beyond who pays what rates. Consider, for instance, legislation granting tax concessions. While some tax concessions might foster economic justice,^{Footnote 98} many do not. Estimates suggest that the top 10 tax concessions alone amount to roughly one-third of federal government revenue, with the gains accruing mostly to high income/wealth taxpayers.^{Footnote 99} Granting such concessions may thus undermine the capacity of taxation to perform its economic justice functions. That does not mean that such laws are constitutionally invalid. Legislation enacted by the Australian Parliament can be struck down, upon judicial review, if it does not fall under a head of legislative power,^{Footnote 100} or if it contravenes an express or implied, legal, constitutional limitation or prohibition.^{Footnote 101} As a general rule, however, judicial review does not consider whether a piece of legislation conduces to constitutional ends or imperatives, where those are given primarily in the domain of political constitutionalism.^{Footnote 102} Thus legislation may fall under a head of power and hence be valid in the domain of legal constitutionalism, as determined by the application of legal meanings in the context of adjudication, even if it contradicts requirements of constitutional public reason that are established in the domain of political constitutionalism. (Any illustration of this point is apt to be controversial, but we nominate the interpretation and application of the (post-1967) race power in Kartinyeri v Commonwealth.)^{Footnote 103}

Monitoring tax expenditures is, therefore, not an aspect of judicial review. It is primarily the Executive’s responsibility to identify and measure these concessions (for example as part of the budget process),^{Footnote 99} and to appropriately administer them; and Parliament is obliged to ensure that such concessions are defensible within public reason, including constitutional public reason as it applies within the domain of political constitutionalism, and to repeal them if not.

C Taxation in the Domain of Legal Constitutionalism

In the domain of legal constitutionalism, the validity of a tax regime is hostage to two exercises of adjudication: the interpretation of the taxation power itself; and the interpretation and application of the statutes enacted pursuant to it. In this sub-section we show how the High Court interprets the taxation power by way of a legalistic approach, insulated from public reason, of the sort that we outlined at the end of sub-section II B. We further show that the resulting broad interpretation of the power contributes to the legitimacy of the Australian framework of laws and government, because it enables Australian tax laws to satisfy the demands of constitutional public reason within the domain of political constitutionalism. In Section IV we argue that attempting to break down this divide between Australia’s legal and political constitutionalism would risk undermining that legitimacy. The interpretation of taxation statutes, and how this relates to public reason and potentially affects legitimacy, we leave as a matter for future research.^{Footnote 104}

Over the course of more than a century of jurisprudence, the High Court has adopted a relatively broad understanding of what constitutes a tax, for the purposes of s 51(ii) and related constitutional provisions (including ss 53, 55, 81, 88, 92 and 99). The limits these establish on Commonwealth taxing power, including (i) the primacy of the House of Representatives over the Senate in the enactment of tax laws and restrictions on the content of such laws,^{Footnote 105} and (ii) requirements of federal uniformity in relation to taxation,^{Footnote 106} do not bear upon our contention that the Constitution creates a framework for economic justice.

The Court has not adopted a universal or exhaustive definition of taxation; its conception has evolved over time and in response to changing contexts.^{Footnote 107} The starting point is generally Latham CJ’s statement in Matthews v Chicory Marketing Board (Vic), that a tax is ‘a compulsory exaction of money by a public authority for public purposes, enforceable by law, and … not a payment for services rendered’.^{Footnote 108} The High Court has used this statement to develop a number of positive and negative elements to guide the approach to identifying a tax.^{Footnote 109} (Consistent with the Communist Party Case, the legislative description of an impost as a tax is not determinative.^{Footnote 110})

Positive elements include ‘a compulsory extraction of money’ (but ‘there is no reason in principle why a tax should not take a form other than the exaction of money’^{Footnote 111} and so might include something measurable in money’s worth), ‘by a public authority for public purposes’ (although neither is an essential requirement),^{Footnote 112} ‘enforceable by law’ (requiring that the impost not be arbitrary^{Footnote 113} and be open to judicial challenge).^{Footnote 114} Public purpose has been interpreted broadly,^{Footnote 115} and is more readily satisfied if the impost, firstly, raises revenue and, secondly, if that revenue is directed into consolidated revenue. As Windeyer J observed in Fairfax, ‘[t]axes are ordinarily levied to replenish the Treasury, that is to provide the Crown with revenue to meet the expenses of government. That is the primary purpose of the income tax.’^{Footnote 116} However, the mere fact that an impost is paid into consolidated revenue is not determinative of its status as a tax.^{Footnote 117} The High Court has additionally recognised that an impost can be a tax even if it lacks the intention to raise revenue or raises negligible revenue.^{Footnote 118} The Court has also recognised the redistributive possibility of taxation by acknowledging that ‘[i]mposing a financial burden on one group in society for the benefit of another group in society will often constitute a tax’.^{Footnote 119}

Negative elements which might determine that a compulsory extraction of money is not a tax include being ‘fees for services’ or ‘payment for services rendered’, quid pro quo arrangements for which the recipient, or some other identifiable person receives some benefit in proportion to the fee.^{Footnote 120} Further excluded by virtue of section 53 of the Australian Constitution are ‘the imposition of fines or other pecuniary penalties’ (which generally entail a liability that flows from a ‘failure to discharge antecedent obligations’),^{Footnote 121} as well as laws ‘for the demand or payment…of fees for licences’.^{Footnote 122} The list of exceptions is open and further includes royalties, fees for a privilege, and fees for the acquisition or use of property.^{Footnote 123} We consider a further possible exclusion below.

Consistent with its generally permissive approach to characterisation,^{Footnote 124} the High Court has adopted a broad approach to determining whether a law can be described as one ‘with respect to’ taxation.^{Footnote 125} Provided that the subject matter of an Act is properly described as taxation, it is not fatal to the validity of the Act if, in substance, the Act is designed to achieve another (economic or social) end other than the raising of revenue for public purposes.^{Footnote 126} In Fairfax, Kitto J stated that it is ‘beyond serious question that a tax does not cease to be valid merely because it regulates, discourages, or even definitely deters the activities taxed.’^{Footnote 127} This reinforces the conclusion that an impost can be a tax even if raising revenue is a secondary consideration, or the revenue raised is negligible.^{Footnote 128}

The combination of a broad interpretation of ‘taxation’ and a permissive approach to characterisation permits Commonwealth taxes to perform all three functions identified above, not merely the raising of revenue. The function of influencing behaviour is enabled by the recognition of the validity of tax expenditures (tax concessions or exemptions designed for purposes other than revenue-raising),^{Footnote 129} including: the use of tax concessions and exemptions to incentivise investment in Commonwealth government securities;^{Footnote 130} to encourage conformity with industrial relations obligations including those supporting income justice (through incentivising employers to meet minimum training requirements,^{Footnote 131} employee pension/superannuation requirements,^{Footnote 132} or (possibly) minimum wage requirements).^{Footnote 133} The Court has also enabled the function of breaking up holdings, by confirming the validity of additional or progressive tax burdens to achieve an end beyond revenue-raising. In Osborne v Commonwealth (‘Osborne’) the Court upheld the validity of a progressive land tax that included higher rates on absentee landlords, imposed to break up land holdings and force non-residents to sell.^{Footnote 134} We can imagine a tax similar to that in Osborne where all three functions are complementary: a steeply progressive land tax on unused land above a certain value to encourage owners to sell, thus engaging the third function while supporting the second if the targeted landholders do in fact sell, and fostering the first (of raising revenue, assuming broad compliance) in the event they do not.

This review of the tax power within the domain of legal constitutionalism has shown that the High Court adjudicates the power without any express regard to the requirements of economic justice that arise in constitutional public reason. The Court applies technical legal methods, and to the extent that these methods cloak more substantive considerations, those pertain primarily to the proper scope and concern of judicial review, rather than the elements of public reason that are the focus of this article.^{Footnote 135} This insulation of the legal constitutional domain from considerations of public reason has not prevented that domain serving the ends of constitutional public reason, however; it has in fact facilitated it. In Section IV we argue against suggested changes to the Court’s interpretive approach, which would reduce that insulation, because they risk undermining the contribution made to legitimacy by the operation of the tax power within the domain of legal constitutionalism.

A Taxation and s 51(xxxi)

In political theoretical discussions about taxation, and also in political debates about it, the relationship between taxation and private property is a recurrent theme.^{Footnote 136} This might seem to be a difficult problem in Australian constitutional law, because s 51(xxxi) of the Constitution establishes a general requirement that any acquisition of property by way of Commonwealth statute must be on just terms. This has been described as a ‘guarantee’ that ‘is there to protect private property’.^{Footnote 137} However, the High Court’s interpretation of s 51(xxxi) — which occurs within the domain of legal constitutionalism — recognises certain exceptions, of which taxation is one. The High Court has held that ‘what is validly within s 51(ii) is outside s 51(xxxi), for the two powers are mutually exclusive’.^{Footnote 138} As Mason CJ explained:

This sort of approach to the relation between the taxation power and the protection of private property by s 51 (xxxi) has been criticised, because it ‘proceeds by examining whether a law has the formal characteristics that qualify for exemption, rather than the reasons for exemption’, thereby giving rise to a risk that constitutional protection will be ‘categorically denied in cases that, at the very least, appear to require further analysis to ensure that the confiscation of property is not arbitrary’.^{Footnote 140} However, we do not accept this criticism. The High Court’s technical, legalistic approach to the interaction of s 51(xxxi) with the tax power supports the use of the latter power for the three functions identified above.

B Retrospectivity and s 51(xxxi)

A further feature of the Commonwealth taxation power, as interpreted by the High Court, is that it includes power to enact tax laws with retrospective effect, provided these are not arbitrary or capricious.^{Footnote 141} In this sub-section, we defend this feature of the Court’s jurisprudence against suggestions that some retrospective taxation may amount to an unjust acquisition of private property that raises s 51 (xxxi) concerns.

Retrospectivity can be important for achieving both the first and second purposes of taxation — raising revenue and securing justice in holdings — often by addressing tax evasion or aggressive tax minimisation. (If retrospective legislation deters future avoidance or minimisation, it may also serve the third function of influencing behaviour.) Because of the importance of the first two functions to economic justice, it follows that retrospective tax legislation can be a necessary component of a constitutionally legitimate framework of laws and government.

There are systemic reasons why taxation may need to operate retrospectively to achieve outcomes within the ‘allowed bounds’. As we have seen, in constitutional adjudication the High Court uses formal legal reasoning. The same is true of private law adjudication, which develops the judge-made private law. And in Australia’s liberal market economy, itself a constitutional imperative, the holdings and transactions that are the subject matter upon which taxation operates are defined, at least in part, by this private law. Thus, there is always the prospect of ingenious use of private law concepts and devices to circumvent extant taxation laws, no matter how well drafted these might be. (As noted above, this article sets aside the question of the degree to which the High Court’s approach to the interpretation and application of tax laws tends to facilitate such ingenuity — something others have argued forcefully elsewhere).^{Footnote 142} This (relative) autonomy of legal reasoning, in a legalistic constitutional order, from the demands of public reason that operate in the domain of political constitutionalism, means that there is always the prospect of a need for retrospective taxation legislation.

Justice Gordon, however, writing extra-crucially, puts forward a case where a law that relates to the imposition of taxation may also enliven s 51(xxxi), and hence be invalid for violation of the constitutional guarantee, arguing that ‘the retrospective alteration of taxation liabilities may (I do not say must) reach a point where some question of the acquisition of property arises..^{Footnote 143} She has in mind ‘laws that, on enactment, apply retrospectively and, on one view, apply differently to different taxpayers’ which might therefore give rise to ‘[a]dditional questions’ such as those ‘that may arise from what can be seen to be the potential destruction, or degradation of longstanding rights’.^{Footnote 144} Her focus is on a retrospective law that:^{Footnote 145}

She suggests, without definitively concluding, that such a law may be an acquisition by the Commonwealth of property in the form of established rights.^{Footnote 146}

Her Honour similarly argues that such retrospective tax laws might fail for arbitrariness, being ‘laws that treat particular taxpayers in a different way depending on the way in which they legitimately ordered their affairs in the past’ that is, ‘solely because of the change in the applicable criteria as a result of the retrospective legislation’.^{Footnote 147} However, this seems not to add to the argument with respect to s 51(xxxi): if a law failed to be with respect to taxation because arbitrary in character, then presumably it would cease to fall outside of the s 51(xxxi) guarantee, which has the prevention of arbitrariness as a key function.^{Footnote 148} Finally, Her Honour suggests that such laws might fail for want of due process (an idea developed by reference to US Fifth Amendment jurisprudence) but it is not clear how this would operate in Australia independently of the ‘just terms’ requirement of s 51(xxxi).^{Footnote 149})

In Chevron Australia Holdings Pty Ltd v FCT (‘Chevron’), the full Federal Court considered Justice Gordon’s argument in the context of legislation retrospectively validating the extant use by the Commission of extrinsic materials (the relevant OECD model convention) in applying the legislative framework so as to value intra-company transactions in the course of determining the tax liability of the corporate group.^{Footnote 150} The taxpayer ‘accepted that the Commonwealth parliament has power to enact tax legislation with some retrospective effect’,^{Footnote 151} but contended that the retrospective law in question ‘imposed an arbitrary and incontestable tax’ because it was not ‘ascertainable and sufficiently general in nature’ and because a taxpayer could not ‘[know] the criteria for liability in the tax year in question’.^{Footnote 152} Pagone J (with whom Allsop CJ and Perram J agreed) rejected this argument, stating ‘that a person may not be aware at an earlier point in time what the (sic) criteria of liability may subsequently be made to apply to that earlier point in time does not make the legislation arbitrary or incontestable’.^{Footnote 153}

We agree with this conclusion. But Chevron does not address the s 51(xxxi) argument. This argument pits two constitutional requirements against one another: the demands of economic justice (which arise in political constitutionalism) and the constitutional guarantee of private property (which operates in both political and legal constitutionalism in the distinct ways described earlier). We now explain why Chevron was correct not to countenance this argument.

We accept the possibility that a retrospective tax liability that is so targeted in its operation that, in effect, it reallocates particular privately-owned assets may generate a legal question. Perhaps the strongest example would be a near confiscatory land tax levied at a particular address or on a particular individual. However, if a retrospective tax is expressed as a general norm then its validity should be affirmed in the orthodox manner (ie that there is no right or entitlement to be shielded from retrospective taxation, and hence no s51(xxxi) question), even if it alters the tax liability and thus unsettles expectations based on a previous legal state of affairs, of only a small, known group of taxpayers. (This is consistent with the Rawlsian method of this article: Rawlsian justice is not concerned with concrete allocations to known individuals but with the ongoing operation of an institutional scheme.^{Footnote 154})

The political need for retrospectivity may often be generated by the actions of a few who have the resources and inclination to engage in complex arbitrage involving tens or hundreds of millions of dollars (as Chevron demonstrates). In such circumstances, the general norm (eg, the levying of a broad-based and progressive income tax), consistently with its intended operation, fosters economic justice via the first and second functions of taxation. There is thus no warrant, from the perspective of political constitutionalism responsive to the demands of constitutional public reason, to expand the operation of s 51(xxxi) so as to strike down such a law. And ex hypothesi there is no such warrant within the technical reasoning of the domain of legal constitutionalism. Expanding the operation of s 51(xxxi) to strike down such a law would represent a departure from the legalistic approach the High Court has adopted to the relationship between the tax power and s 51(xxxi). Considerations of constitutional public reason, arising from the presence of s 51(xxxi) in the Constitution, would be allowed to affect legal meaning, resulting in a purposive or functional approach to the application of the guarantee.^{Footnote 155} This would have a distorting effect upon Australian constitutional law: private property interests would gain the additional protection of substantive, public reason-based consideration in adjudication; but there would be no corresponding change to the legal meaning of the powers that permit economic justice to be pursued, given that their existing legal meaning already permits them to operate in an expansive fashion.^{Footnote 156} This change, therefore, would simply undermine the support that the existing legalistic approach to legal meaning provides to the demands that constitutional public reason imposes within the domain of political constitutionalism. This is why we support the Court’s current approach to the relationship between the taxation power, s 51 (xxxi), and the retrospective imposition of tax liabilities.