A Federalism, Laboratory Federalism and Their Advantages

Federalism, and divesting legislative power to states or lower-level governments, is seen as having a number of theoretical benefits as a model of governance. It can facilitate policies that are tailored to the needs and community expectations of particular states.^{Footnote 3} Individuals have a louder ‘voice’ when engaging with state governments, who are therefore likely to be more responsive to constituents’ concerns. Further, individuals can choose to ‘exit’ particular jurisdictions if policies are not to their liking.^{Footnote 4} In a federal model, states can adopt different responses to the tensions between autonomy and security, liberty and privacy, dignity and efficiency; this potentially facilitates innovation and competition between jurisdictions,^{Footnote 5} as states compete ‘to attract desirable citizens and businesses’.^{Footnote 6} States may also be able to act more agilely to emerging social concerns, meaning policy change can be more immediate and responsive.^{Footnote 7} Finally, divesting power to the states can minimise the risk of tyranny^{Footnote 8} at the federal level. These represent substantial regulatory advantages over those offered by a unitary system of government.

While these benefits are associated with federalism — and not laboratory federalism specifically — they are likely enhanced and supported by jurisdictional experimentation. In the United States, Justice Brandeis’s long-quoted (perhaps optimistic) view of federalism is where ‘a single courageous State may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country’.^{Footnote 9} This serves as the basis for ideas of laboratory federalism; that is, where states can experiment and test novel policy solutions, learn from the innovations of other states, and where the risks of (failed) experimentation are contained within one particular jurisdiction. This model engenders an ‘image of policy percolating up and through the federal system’,^{Footnote 10} eventually reaching the federal level once it is tried and tested at the state level.

Experimentation is arguably critical for realising the posited benefits of federalism as a model of governance. It can ensure diverse policies across states, adapted to the needs of particular jurisdictions. It also ensures ‘exit’ offers residents meaningful choices between jurisdictions. Laboratory federalism may also be particularly useful in addressing ‘wicked’ problems. ‘Wicked’ problems emerge in policy areas that are highly resistant to resolution and have many interdependencies, where the problem is complex, unstable and difficult to define, and where there is no clear solution, thereby requiring a value-based management strategy,^{Footnote 11} ‘trade-offs between conflicting goals’,^{Footnote 12} and difficult normative decisions to be made.^{Footnote 13} Where there is no ‘right’ answer, devolving decisions to the local level can minimise and contain the damage caused by failed experimentation,^{Footnote 14} and can respond to differing values and priorities in each jurisdiction.^{Footnote 15} In contemporary society, many of the problems faced by government are ‘wicked’ and resistant to resolution; thus, a federal structure — and experimentation — is even more pertinent for addressing modern governmental and policy challenges, including those posed by growing inequality.

B Risks of Laboratory Federalism

These potential benefits of federalism, and laboratory federalism in particular, are balanced by the risks or potential limits of experimentation. It is questionable whether states will actually innovate; laboratory federalism may not occur to a desirable or optimal degree in practice. The risk of ‘exit’ may actually stymie social policy reform: states may be unwilling to introduce progressive reforms that risk putting business off-side, in case they choose to leave the jurisdiction.^{Footnote 16} States may lack the capacity or inclination to innovate: in the context of environmental policy, for example, Engel argues that states are better seen as ‘scale innovators’ than ‘policy innovators’; most ideas originate at federal or international level, so rather than creating new ideas, states just adapt and scale initiatives to their own circumstances.^{Footnote 17} Further, it may not be in states’ interests to innovate: if the costs and risks of an experiment are internalised, and it is easy to imitate successful experiments, then it is in states’ interests to let other states experiment, rather than doing so themselves.^{Footnote 18} Oates describes this as ‘a standard sort of incentive for free-riding’.^{Footnote 19}

This may just mean we need to adjust our expectations of what states as ‘laboratories’ means in practice. As Rom argues, the metaphor of states as laboratories should be taken ‘seriously but not literally’.^{Footnote 20} States mostly do not ‘experiment’: instead, they ‘innovate’, that is, adopt policies or programmes that are new to the states themselves, not necessarily new in a global sense.^{Footnote 21} Policy innovations are not developed or designed as laboratory experiments, seeking to find ‘good policy’: they are designed to ‘serve the interests of the politicians who design them’.^{Footnote 22} Policies are driven more by ideology, tradition and passion than by theory and evidence.^{Footnote 23} Further, innovations can be assessed for both their policy effectiveness (do they address the problem?) and political effectiveness (is it politically attractive enough to be implemented and sustained?).^{Footnote 24} Thus, assessing policy innovations against non-political criteria applied to actual laboratories is inappropriate.

Even if states are willing to experiment, and laboratory federalism occurs, it is unclear how successful those experiments will be. As Gelbach notes, ‘laboratories are only as good as the experiments their operators conduct’^{Footnote 25} (if the operators conduct any experiments at all), and those experiments then need to be evaluated in a meaningful way.^{Footnote 26} Successful policy experimentation and evaluation can consume significant resources (both fiscal and human): expecting states to bear this burden may be unrealistic, particularly where they are responsible for a wide range of policy areas, with limited resources and ways of raising governmental revenue.^{Footnote 27}

Conversely, too much innovation can jeopardise the ability to evaluate or study particular reforms, to see what works. As Gelbach concludes in the context of United States welfare reform,

Experimenting in multiple areas simultaneously can make programme evaluation difficult or impossible. Instead, effective evaluation requires a long-term, strategic and measured approach to policy reform, which may be uncommon (or unrealistic) in practice.

Further, if states are diverse in terms of their constituents’ needs and expectations, as laboratory federalism suggests, then they are also likely to have diverse institutions, histories and political groups. This may make it difficult or inappropriate to transplant innovations from one state to another,^{Footnote 29} or impossible to isolate the factors that contributed to policy success or failure.^{Footnote 30} Thus, serious questions must be asked about the transferability of policy lessons from one jurisdiction to another.

Experimentation also imposes additional administrative costs on businesses and those who operate across state borders, as they are forced to navigate and comply with different rules in each different jurisdiction.^{Footnote 31} This is particularly a challenge in small countries like Australia, where the population in each state is generally small, making economies of scale limited or non-existent.

Finally, and more fundamentally, there may be areas in which state experimentation is inappropriate. As Conkle writes, ‘these potential benefits of federalism [do not] outweigh the need for national protection and enforcement of fundamental constitutional rights’.^{Footnote 32} Even those who support experimentation recognise that it has normative limits: Azmy, for example, draws on the example of predatory lending reform in the United States at the state level to illustrate the practical and theoretical benefits of state experimentation.^{Footnote 33} However, Azmy distinguishes this economic and social policy context from that where ‘there [is] a need for the imposition of national moral norms [as] against racial discrimination’.^{Footnote 34} In that normative context, there is value in uniformity, including as a means of creating political bonds across the country.^{Footnote 35}

Thus, there is a live question as to when experimentation is (in)appropriate, and whether there are particular fields or policy areas (such as addressing inequality and discrimination) where uniformity is preferable and should be pursued. Indeed, in some fields, experimentation might actually lead to a ‘race to the bottom’.^{Footnote 36} If states are competing for businesses and investment capital, there is a temptation to become a low-regulation, ‘business friendly’ jurisdiction, secured via minimal protection for individual rights. Thus, competition can pursue both a race to the top; and to the bottom.^{Footnote 37} The question becomes, then, whether there are some fields in which minimum standards are required, and where experimentation is likely to undermine individual rights and liberties.

C The Practical Operation of Laboratory Federalism: Existing Comparative Studies

Building on these theoretical concerns, we can also question whether laboratory federalism and experimentation work in practice: are states effective ‘laboratories’?^{Footnote 38} Federalism may mean that states (‘the vanguard’) experiment with, test and evaluate particular policy ends, which can then be assessed and followed, avoided or improved upon, by other states (‘the laggard’),^{Footnote 39} and eventually adopted at a national level. However, as Althouse argues, it can be difficult to determine which states are in the ‘vanguard’ and which form the ‘laggards’: ‘The entire vanguard/laggard distinction creates a somewhat false, or at least sometimes false, picture of progress, depicting one right answer waiting for us in the future with the only task being to discover it and put the answer into action’.^{Footnote 40} ‘Laggards’ could instead be depicted as traditionalists, having an important role to play in federalism,^{Footnote 41} and, consistent with the notion of ‘wicked problems’, there may be no ‘right’ answer to complex policy dilemmas. Thus, laboratory federalism may present an overly simplistic picture of policy change and progress.

At a practical level, as Fenna points out, the potential for effective laboratory federalism is dependent on a number of pragmatic considerations, such as:

1. 1. The number of jurisdictions;

2. 2. The range of policy responsibilities exercised by the states;

3. 3. The resources at the states’ disposal;

4. 4. The level of difference and similarity between the states, such as to ensure policy divergence, but within the ‘spectrum of relevance’.^{Footnote 42}

To this, we might add:

1. 5. The political will and inclination of states to innovate;

2. 6. The degree to which states are able to implement and evaluate policy reforms (including through the provision of appropriate human and financial resources); and

3. 7. The presence of a long-term legislative and reform agenda, to ensure policies can be implemented and reviewed across an appropriate timeframe.

As Fenna notes, these characteristics are not well established in the Australian federal system, with a small number of jurisdictions, legislative powers of the states which have been ‘whittled away’, and limited state revenue.^{Footnote 43} Thus, Fenna concludes that ‘the potential for laboratory federalism to flourish in Australia is limited, and robust examples of laboratory federalism at work are disappointingly hard to find’.^{Footnote 44}

That said, even in a jurisdiction like Australia, there is still some evidence of policy experimentation and learning between states. For example, in his study of state betting laws, using a historical case study approach, Painter traces a shift across the Australian states from policy uniformity, to diversity, and back to uniformity.^{Footnote 45} Painter attributes this shift to political factors, shifting public opinions, and the changing position of key stakeholders. However, there was also evidence of mutual learning between and across jurisdictions, as evidenced by the series of inquiries and reviews conducted in each state.^{Footnote 46}

In her scholarship, Appleby has considered the extent to which leaving law and order within the Australian states’ legislative autonomy has allowed for local diversity and experimentation.^{Footnote 47} Appleby argues that the definitional difficulties that plague the Kable principle — which sets out the requirements of institutional integrity for state courts exercising federal judicial powers, as part of an ‘integrated court system’ — have unnecessarily hampered experimentation in the area of law and justice, and led to states ‘replicat[ing] known-to-be-valid constitutional provisions, at the price of developing new, locally tailored, regimes’.^{Footnote 48} Thus, requiring states to comply with federal (judicial) norms has limited experimentation in practice.^{Footnote 49}

Further, Appleby argues that Australian states adopt law and order measures due to political pressure, often before a policy has even commenced in another state, let alone been subject to evaluation and assessment.^{Footnote 50} Appleby therefore concludes that:

Thus, rhetoric and political pressure may be more influential than policy evaluation and mutual learning between states.

Even in a larger jurisdiction like the United States, scholars are increasingly doubting the practical effectiveness of laboratory or experimentalism federalism for advancing particular policy ends. Cunningham-Parmeter argues that Justice Brandeis (quoted in Part 2(a) above) poses two necessary conditions of effective experimentation: internalisation and replication.^{Footnote 52} First, internalisation requires that states internalise the costs of their policies, so that other states can effectively analyse and assess their success.^{Footnote 53} Second, replication requires that states take different approaches to addressing similar problems.^{Footnote 54} For Cunningham-Parmeter, immigration law in the United States does not satisfy either of these conditions, meaning it is an example of failed experimentation. In this context, states can only experiment within a narrow field of federally defined norms, limiting the level of diversity and experimentation that is possible.^{Footnote 55} Further, states do not internalise the costs of their policies, tending to shift the social, economic and human costs of immigration to other states.^{Footnote 56} Thus, Cunningham-Parmeter concludes that ‘not all subjects are ripe for local experimentation and not all tests produce valid results.’ Despite the appealing image of states as laboratories, today’s immigration experiments will not advance the nation’s ongoing search for sounder immigration policies’.^{Footnote 57}

Concerns about the effectiveness of laboratory federalism have also been raised in the context of United States health reform. In their study, Sparer and Brown conclude that states are generally poor policy laboratories: ‘although state and federal officials look curiously at approaches developed elsewhere, and occasionally replicate and emulate, such diffusion is rare’.^{Footnote 58} Thus, there is limited mutual learning and evaluation between states, undermining the scope for laboratory federalism to advance policy reform and development in practice.

Similarly, drawing on a case study of the legal recognition of same-sex relationships, Knauer notes that federalism can offer states freedom to experiment with innovative solutions to social issues.^{Footnote 59} However, she argues that laboratory or experimental federalism is inappropriate for securing minority rights: federalism is not a political philosophy, and experimentation can lead to both progressive and conservative outcomes.^{Footnote 60} Federalism is not an appropriate mechanism for securing broad-based rights, as rights are not portable from state to state, and rights can be overturned.^{Footnote 61} A lack of uniformity across states has resulted in same-sex relationships having ‘confusing and conflicting status’^{Footnote 62} which ‘weighs heavily on same-sex couples’.^{Footnote 63} As Knauer concludes, ‘same-sex couples who reside in a jurisdiction with some degree of relationship recognition must travel [between states] at their own risk’.^{Footnote 64} Knauer’s study, then, offers a practical example of the inappropriateness of experimentation in particular policy areas.

Overall, these studies illustrate the practical limits of laboratory federalism. In a wide range of policy areas, across both Australia and the United States, laboratory federalism is not occurring to a desirable or optimal degree. Democratic experimentalism might offer some suggestions as to why this is the case.

A The Federal Legal Framework

Age discrimination in employment is regulated in Australia by ‘a patchwork’^{Footnote 88} of age discrimination legislation at federal,^{Footnote 89} state and territory level,^{Footnote 90} and as both an equality issue generally, and in relation to employment particularly.^{Footnote 91} This complex legislative context has resulted in a proliferation of bodies that might deal with discrimination issues, including federal equality bodies such as the Australian Human Rights Commission (‘AHRC’), state equality bodies such as the Victorian Equal Opportunity and Human Rights Commission (‘VEOHRC’), federal employment bodies such as the Fair Work Commission (‘FWC’),^{Footnote 92} state administrative tribunals such as the Victorian Civil and Administrative Tribunal (‘VCAT’) and the federal Fair Work Ombudsman (‘FWO’).^{Footnote 93} Discrimination law in Australia is a ‘crowded space’^{Footnote 94} with ‘a lot of players in the field’;^{Footnote 95} it is ‘just very complicated’.^{Footnote 96}

The Australian anti-discrimination regime has been developed in the context of a federal division of powers between state, territory and Commonwealth legislatures.^{Footnote 97} The Commonwealth Parliament has no express power to legislate in relation to discrimination matters. Instead, the Age Discrimination Act 2004 (Cth) (‘ADA’) (and other federal discrimination legislation) has been created on the basis of a number of heads of law-making power, including the external affairs power (to implement international treaty obligations);^{Footnote 98} and powers to make laws for the territories,^{Footnote 99} Commonwealth employees,^{Footnote 100} foreign, trading and financial corporations,^{Footnote 101} banking,^{Footnote 102} insurance,^{Footnote 103} trade and commerce,^{Footnote 104} and incidental matters.^{Footnote 105} This hodgepodge of powers is reflected in s 9 of the ADA, which specifies its geographical application, and s 10 of the ADA, which links the application of the Act to the relevant constitutional powers.^{Footnote 106}

Given this complicated division of powers, it is difficult to determine the outer reaches of the federal discrimination Acts, and there has been limited case law on this issue.^{Footnote 107} In many instances, state and territory discrimination statutes will operate as a parallel scheme to the federal law, with similar (though not identical) wording to the federal legislation. This will likely minimise any difficulties with determining which legislation applies in any given case.^{Footnote 108}

Where the state and federal schemes are inconsistent, Commonwealth legislation will generally take precedence under s 109 of the Australian Constitution. However, this is not always clear cut. For example, in Jemena Asset Management Pty Ltd v Coinvest Ltd,^{Footnote 109} which related to long service leave provision, it was held that a Victorian state scheme that provided for pension portability in the construction industry was not inconsistent with federal awards and agreements; rather, the two schemes were complementary.^{Footnote 110} The state scheme did not alter, impair or detract from the federal provisions; further, there was no evidence that the federal scheme intended to cover the field of pension portability.

In the context of age discrimination law, while compulsory retirement is prohibited by ADA s 18(2), some state and territory legislation still contains mandatory retirement as an exception to the principle of age equality.^{Footnote 111} This is likely a direct inconsistency; the passage of the ADA has therefore likely rendered those state and territory exceptions invalid (and almost certainly in the case of the Northern Territory legislation),^{Footnote 112} and employers rarely rely on the exception in practice.^{Footnote 113}

Beyond these examples of direct inconsistency, and following Jemena, there is perhaps an argument that state equality laws do not detract from federal laws; rather, the two schemes can be seen as complementary and mutually supportive, with potentially different scopes and tests. This should minimise any scope for inconsistency. There is also an argument (again, as in Jemena) that ‘it is possible to infer from the beneficial nature of the federal [statute] that the Commonwealth legislature did not intend to exclude a compatible State law’.^{Footnote 114} That said, there is far more substantive overlap between state and federal equality laws than was the case in Jemena, raising a real risk of a substantive conflict between the regimes.^{Footnote 115}

Further constitutional complexity in this space arises from the use of state administrative tribunals to hear age discrimination complaints in the states and territories.^{Footnote 116} Under Chapter III of the Constitution, only ‘courts’ may be vested with the capacity to exercise judicial power with respect to federal matters.^{Footnote 117} Ananian-Welsh concludes that, while the Queensland Civil and Administrative Tribunal (‘QCAT’) is a ‘court’, the New South Wales Civil and Administrative Tribunal (‘NCAT’), VCAT and South Australian Civil and Administrative Tribunal (‘SACAT’) are not, leaving their constitutional character ‘open’.^{Footnote 118} This means, then, that ‘disputes that raise constitutional questions, involve parties from different states, or are otherwise listed in ss 75 or 76 of the Constitution, may be determined by the courts of the states but not by state administrative tribunals’.^{Footnote 119}

Given the potential overlap between state and federal discrimination law, it is critical that tribunals are capable of resolving questions relating to the interplay between state and federal statutes; and between discrimination law and constitutional law.^{Footnote 120} In the disability discrimination case of Citta Hobart Pty Ltd v Cawthorn,^{Footnote 121} though, the High Court held that the Tasmanian Anti-Discrimination Tribunal (‘ADT’) could not consider any potential inconsistency between state and federal discrimination law under s 109. The majority concluded that:

The claim — in its entirety — therefore had ‘the characteristics of a matter’ described in s 76 of the Constitution, and had to be referred to the Supreme Court for determination, even if the constitutional defence under s 109 was ultimately dismissed.^{Footnote 123} The strengths and merits (or otherwise) of the defence were largely irrelevant: ‘it is enough that the claim or defence be genuinely raised and not incapable on its face of legal argument’.^{Footnote 124} This has the potential to severely constrain the ability of state tribunals to hear and determine matters relating to discrimination law.

Further, state tribunals are likely unable to hear cases where the federal government is a party,^{Footnote 125} or those that cross state borders. Thus, in Commonwealth v Anti-Discrimination Tribunal (Tasmania),^{Footnote 126} Kenny J opined that the Tasmanian ADT lacked jurisdiction to resolve a discrimination complaint against the federal department Centrelink;^{Footnote 127} regardless, the Commonwealth was not bound by the Anti-Discrimination Act 1998 (Tas).^{Footnote 128} Instead, then, these matters involving federal jurisdiction will typically need to be transferred to a properly constituted ‘court’, occasioning additional expense and delay, and likely inhibiting access to justice.^{Footnote 129}

To address these complex issues of jurisdiction, Ananian-Welsh recommends that state tribunals be reconstituted as Chapter III ‘courts’, by designating these bodies as courts of record.^{Footnote 130} This would not necessarily mean reducing the flexibility and informality of tribunal proceedings (which is a major advantage of the current state tribunal system): as Ananian-Welsh concludes, ‘The powers, practices and procedures of a tribunal will play a secondary and distinctly non-determinative role in ultimate determination of constitutional character’.^{Footnote 131}

Further complexity arises from the constitutional differences between the states and territories, particularly as they affect the status of territory tribunals. In the territories, the field is also shaped by s 122 of the Constitution. Ananian-Welsh confines her discussion to state tribunals and does not consider in detail how territory tribunals might differ;^{Footnote 132} it is possible that the same Chapter III rules apply to territory tribunals,^{Footnote 133} or that s 122 of the Constitution brings territory tribunals beyond the reach of Chapter III.^{Footnote 134} In Bottrill v Sunol ^{Footnote 135} (decided prior to the High Court’s decision in Burns), the ACT Civil and Administrative Tribunal (‘ACAT’) rejected the argument that s 75(iv) of the Constitution prevented ACAT from determining a matter between a resident of the ACT and a resident of NSW.^{Footnote 136} Section 75(iv) refers to residents of ‘different States’; and the ACT is not a state.^{Footnote 137} Since Burns, then, ACAT has adjudicated vilification disputes between parties from the ACT and NSW^{Footnote 138} and Queensland and the ACT.^{Footnote 139} This arguably does not resolve the question of whether ACAT is capable of exercising judicial power in federal matters when, for example, adjudicating discrimination claims that raise constitutional issues.^{Footnote 140} The character and powers of ACAT and the Northern Territory Civil and Administrative Tribunal (‘NTCAT’) are therefore still matters of significant contention.

The experience in the ACT and NT is likely also shaped by the risk that the Commonwealth Parliament might disallow or overrule territory legislation.^{Footnote 141} In the NT Anti-Discrimination Amendment Act 2022 (NT), for example, the new vilification provisions^{Footnote 142} are drafted to mimic s 18C of the Racial Discrimination Act 1975 (Cth), despite most states adopting a different approach,^{Footnote 143} and a Victorian inquiry putting forward models to overcome perceived shortcomings of s 18C.^{Footnote 144} It may be, then, that legislative reform in the NT is being constrained by federal legislation or the fear of federal intervention; or, perhaps, that there is limited knowledge of developments in other jurisdictions (see below).

The federal system also poses a risk of ‘double-dipping’, where individuals who experience discrimination might make claims under both state/territory and federal discrimination legislation, or under both employment and discrimination legislation. This is expressly prohibited by s 12 of the ADA, which disentitles individuals from making a claim under the ADA if a claim has already been made under state or territory legislation. The FWA also prohibits ‘double-dipping’ in ss 724 and 734.

B Individual and Organisational Impacts of the Federal Structure

Before considering the extent to which there is mutual learning and experimentation in this field, it is important to recognise the organisational and individual risks arising within this complex federal structure. These are not problems of laboratory federalism as such; but they are important counterpoints when considering the practical impact of jurisdictional diversity in a federal country.

At an individual level, for those who experience discrimination, the most noticeable practical impact of the federal structure relates to the need to choose where proceedings or a complaint should be brought.^{Footnote 145} This may have particular significance in practice, particularly in relation to relevant time limits, costs that might be incurred and any limit on damages.^{Footnote 146} Once a claim is filed, it can be difficult (or impossible) to shift between jurisdictions.^{Footnote 147} Claimants who have quality legal representation can use this ‘choice’ strategically to maximise chances of success, minimise exposure to costs orders and maximise the efficiency of the process.^{Footnote 148} This duplication and variety of avenues for making a claim may therefore be a healthy sign of democracy,^{Footnote 149} enhancing access to justice.

For claimants who are unrepresented, though, or represented by practitioners unfamiliar with equality law, it is difficult to make a meaningful ‘choice’ of jurisdiction.^{Footnote 150} State and federal overlap is ‘fairly confusing for a lot of people’.^{Footnote 151} Even with skilled representation, it can take hours to decide which route to pursue,^{Footnote 152} and the decision is ‘not easy’,^{Footnote 153} often requiring a full assessment of a claim’s merits^{Footnote 154} that is highly context-specific.^{Footnote 155} Vulnerable individuals may be more inclined to use ‘local’ state or territory processes,^{Footnote 156} especially in small jurisdictions where there are personal connections between equality agencies and referring NGOs,^{Footnote 157} or where there is scepticism of the federal government.^{Footnote 158} Where claimants are more resilient (or legally represented) they might pursue a complaint to the more ‘scary’ federal system, where the threat of proceeding to a federal court can ‘keep an employer on track’.^{Footnote 159} Overall, though, confusion regarding where claims can and should be made might lead to claims not being brought at all.^{Footnote 160}

The limited case law on age discrimination in Australia^{Footnote 161} shows that some claimants have (unsuccessfully) attempted to bring both human rights and FWA claims.^{Footnote 162} Others have sought to bring claims at both state and federal levels,^{Footnote 163} or to challenge federal legislation under state discrimination law.^{Footnote 164} The complexity of these jurisdictional decisions is reflected in the Tasmanian case of B v Naval Reserve Cadets.^{Footnote 165} In that case, part of the alleged conduct occurred on federal government property and part of it did not. Further, the Tribunal had to consider whether the Naval Reserve Cadets were an emanation of the Commonwealth or whether the Commonwealth might be covered by state discrimination legislation. Resolving these questions necessarily entailed further submissions and an additional hearing^{Footnote 166} (and, now, would likely be held to be beyond the ADT’s jurisdiction).

Overall, then, this raises the question of whether a ‘choice’ of jurisdiction is meaningful, if it can only be exercised effectively with skilled, detailed (and expensive) legal advice, accessible to only a small proportion of claimants.^{Footnote 167} Indeed, most claimants who file and pursue an age discrimination complaint with an equality agency in Australia are not legally represented.^{Footnote 168} There are serious questions, then, about whether the possibility of state-based experimentation, which might strengthen legislative protection in some jurisdictions, is sufficient to balance these risks of federalism for individuals. While well-resourced claimants may be able to optimise the benefits and opportunities of experimentation, the practical impacts of regulatory diversity for less fortunate claimants are far less positive.

Multiple systems and statutes across jurisdictions may also result in a duplication of enforcement,^{Footnote 169} fragmenting and undermining protection from discrimination: ‘you’ve got a relatively small population base but within that a ludicrous … amount of duplication’.^{Footnote 170} It might fragment the funding of statutory equality agencies, undermining the effectiveness of enforcement in small jurisdictions.^{Footnote 171} Having multiple state, territory and federal structures might also undermine the collection of data in relation to claims, and therefore the evidence base for effecting change in this area.^{Footnote 172} A lack of age discrimination complaints, for example, may fall into the ‘Bermuda triangle’ between state and federal agencies,^{Footnote 173} and not be noticed. These are issues of federalism, not laboratory federalism as such. However, they are potentially compounded by the differences in the legal frameworks between jurisdictions, which make understanding the law, and monitoring its implementation, more complex. For organisations and employers, experimentation creates a ‘minefield for organisations to navigate’, as different exceptions and exemptions apply at different levels.^{Footnote 174} This is a particular problem for companies operating across jurisdictions.^{Footnote 175} This legal complexity is illustrated by my previous study of exceptions to age discrimination law in Australia and the United Kingdom.^{Footnote 176} Most Australian jurisdictions adopt similar, albeit differently worded, exceptions to age discrimination law. However, the exceptions often vary in their breadth across jurisdictions, and additional exceptions exist in some states and territories and not others. This represents a complex legal field for both organisations and claimants to navigate.

That said, this diversity and multitude of players within a federal system can enhance protection from discrimination: state, territory and federal equality bodies address equality issues ‘from quite different angles’ and in different legislative frameworks.^{Footnote 177} Multiple bodies and agencies can therefore fulfil complementary roles; state and territory bodies are more connected with their communities and can therefore focus on targeted capacity building particular to that community.^{Footnote 178} Given the cultural, normative and structural differences between Australian jurisdictions, a ‘one size fits all’ approach to advancing equality is inappropriate.^{Footnote 179} Agencies with limited resources can also address some of these constraints by ‘building on the back of what other Commissions have done and recognising their work’,^{Footnote 180} though this essentially outsources costs of enforcement to other jurisdictions. Understandably, then, ‘the four smaller Commissions — WA, Tasmania, NT and SA — … are much more inclined to work collaboratively together. Because none of us have any resources’,^{Footnote 181} though both large and small equality agencies can learn from each other’s experiences and processes.^{Footnote 182} This flags, then, the potential for mutual learning and experimentation across and between the states, both in relation to the legislative framework and in their approach to enforcement.

C Experimentation and Policy Percolation?

To what extent, then, is there evidence of experimentation and mutual learning in this space? Can the different approaches and structures in each state and territory drive innovation and produce better solutions for addressing age discrimination?^{Footnote 183} The broad similarity between age discrimination statutes in some states and territories implies that jurisdictions are at least taking legislative ‘inspiration’ from other jurisdictions.^{Footnote 184} Inter-state comparison can play a key role in equality law reform processes: ‘you know the question that they’re going to ask in the Cabinet room or Parliamentary Council or something is, “What are the other states and territories doing?”’^{Footnote 185} This ‘copycat’^{Footnote 186} approach may support policy percolation and diffusion over time.

There is also growing evidence of experimentation in equality law among the Australian states and territories, and some evidence of mutual learning. For example, statutory equality duties have been adopted in equality statutes in Victoria and Tasmania, and equality duties under human rights statutes have been implemented in Queensland, Victoria, and the ACT (see further below).^{Footnote 187} The ACT has made statutory provision for dual or multiple discrimination,^{Footnote 188} and a shifting burden of proof has been adopted in the ACT and under the FWA.^{Footnote 189} Diversity and experimentation may therefore breed innovation and a race to the top.^{Footnote 190}

A process of inter-state learning and engagement was also reflected in the original emergence of age discrimination legislation in Australia. Age was first added to the grounds of discrimination in the Human Rights and Equal Opportunity Commission Act 1986 (Cth) from 1 January 1990;^{Footnote 191} and prohibited in South Australia from 1 June 1991. Compulsory retirement was progressively abolished in NSW from 1 January 1991, and other states soon followed.^{Footnote 192} What is notable about this history, however, is that it was the federal government that led developments in this area; states and territories were following the central government’s lead, the opposite to what is envisaged by laboratory federalism.

Indeed, in relation to discrimination law generally, one of the justifications for the 1984 consolidation of equality law in South Australia was that the Commonwealth had indicated that it would amend federal equality laws to not apply to South Australia ‘[i]f the Commonwealth is satisfied that the legislation of this State implements the objectives of their Acts’.^{Footnote 193} This makes explicit the influence the federal government has had over the development of state discrimination law, and the limited field within which states have been free to experiment. Similar concerns drove the passage of equality law in Western Australia:

Thus, fear and mistrust of Commonwealth interference — and a Commonwealth threat to legislate — drove equality legislation in Western Australia.

In a contemporary setting, discrimination law appears to be entering a new and renewed period of experimentation and cross-jurisdictional learning. Since 2020, substantial reviews and inquiries into discrimination law have been conducted in the NT,^{Footnote 195} ACT,^{Footnote 196} Queensland,^{Footnote 197} WA^{Footnote 198} and federally.^{Footnote 199} As Painter argues, these inquiries and reviews provide an important source of evidence regarding the degree to which ‘[i]mitation, emulation, or learning are important in shaping the evolution of policy in each jurisdiction’.^{Footnote 200} While the process of legislative change is ongoing, and is not limited to age discrimination law specifically, these reviews and their recommendations illustrate a strong degree of cross-jurisdictional influence.

A critical example of this is the percolation and spread across jurisdictions of the Victorian positive duty to eliminate discrimination.^{Footnote 201} In Victoria, this duty applies to all protected grounds, including age, and requires those who have a duty not to engage in discrimination, sexual harassment or victimisation to ‘[t]ake reasonable and proportionate measures to eliminate that discrimination, sexual harassment or victimisation as far as possible’.^{Footnote 202} Through this series of inquiries, reviews and reforms, a positive duty like that in Victoria has been proposed in the NT, ACT, WA, Queensland and federally; and, to date, adopted in the NT,^{Footnote 203} ACT^{Footnote 204} and federally in relation to sex discrimination and sexual harassment.^{Footnote 205} These provisions have largely replicated the framing and wording of the Victorian positive duty. Indeed, the AHRC’s Respect@Work report specifically cited a case study of the Victorian provisions in informing its recommendation for adopting a positive duty in the context of sexual harassment and sex discrimination.^{Footnote 206} In its Free and Equal position paper, the AHRC further recommended extending a federal positive duty, modelled on that in Victoria, to cover all protected grounds.^{Footnote 207} The WA Law Reform Commission also recommended the adoption of a positive duty, again modelled on and closely reflecting the Victorian provisions,^{Footnote 208} and referring to the Victorian provisions and the Respect@Work report as comparative examples.^{Footnote 209} The Respect@Work recommendations^{Footnote 210} and Victorian positive duty^{Footnote 211} also informed the Queensland review of discrimination law: the options canvassed in the Queensland Human Rights Commission’s discussion paper relating to the implementation of a positive duty closely reflect the legislative framework in Victoria.^{Footnote 212} The Queensland Human Rights Commission ultimately recommended a positive duty be introduced, replicating that in Victoria and as recommended by Respect@Work.^{Footnote 213} In the ACT, the discussion paper on equality law reform drew extensively on the Victorian experience in considering the adoption of a positive equality duty;^{Footnote 214} with a positive duty ultimately being adopted in the Discrimination Amendment Act 2023 (ACT).

This provides a clear and documented example of state-based experimentation percolating across the states and territories, and up to the federal level, as envisaged by laboratory federalism. State and federal inquiries and reviews have provided a forum, albeit an ad hoc one, for jurisdictions to engage in peer and policy review, analysis and sharing. Victoria’s decade-long ‘experiment’ with a positive duty has provided other jurisdictions with a degree of certainty that the risks of experimentation are minimal, allowing other states to simply imitate the successful Victorian experiment^{Footnote 215} and ‘allaying some of their worst fears’.^{Footnote 216}

That said, these developments have not come from nowhere. The AHRC has been recommending the adoption of positive equality duties in Australia since at least 2008, though it has also repeatedly noted that there is a need for further consideration of how such a duty might be drafted and implemented.^{Footnote 217} Victoria’s experimentation has likely helped to resolve some of the practical difficulties of adapting positive duties to the Australian context, offering a workable example of a positive duty that can be replicated in other jurisdictions.^{Footnote 218} There are also numerous examples of positive equality duties internationally, as in Great Britain,^{Footnote 219} Northern Ireland,^{Footnote 220} and Canada,^{Footnote 221} meaning the states — even Victoria — are better seen as ‘scale innovators’ than ‘policy innovators’,^{Footnote 222} adapting international ideas to their own circumstances, rather than creating new ideas as such.^{Footnote 223} This also reflects Rom’s argument that states are unlikely to ‘experiment’ with policies that are new in a global sense; their ‘innovation’ involves adopting policies or programmes that are new to the states themselves.^{Footnote 224}

Given this long history, the impetus for change in 2020 perhaps came from the telling findings of the AHRC’s Respect@Work report, which chronicled the perilous state of workplaces for women, and the chronic failings of discrimination law to achieve meaningful change. Successful experimentation, then, is likely not enough to prompt mutual learning; replication across jurisdictions likely requires a compelling catalyst for change. Further, learning is not so much about finding the ‘right’ solution, but about sorting convincing from unconvincing arguments.^{Footnote 225} Again, Respect@Work offered a critical positioning for these reviews and inquiries, by cataloguing a compelling argument that existing discrimination law — relying on individual complaints and individual enforcement — had failed to achieve change.^{Footnote 226} In this context, advocating against a positive duty became unconvincing.^{Footnote 227}

There is a risk, though, that these developments reflect the replication and duplication of interstate developments, without mutual learning or evaluation. For example, the enforcement mechanisms for the positive equality duty in Victoria were watered down in 2011,^{Footnote 228} removing the VEOHRC’s powers to ensure compliance via compliance notices, enforceable undertakings and public inquiries.^{Footnote 229} The VEOHRC can now only enforce the positive duty via persuasion; or, in exceptional circumstances, by conducting an investigation,^{Footnote 230} and thereafter taking any action it sees fit, including entering into an agreement with the organisation.^{Footnote 231} These lesser enforcement powers no longer reflect what was recommended in the Gardner review;^{Footnote 232} the VEOHRC has repeatedly called for its original, broader suite of enforcement powers to be reinstated.^{Footnote 233} Despite these concerns, these lesser enforcement mechanisms have been largely replicated in the NT, where the duty will be enforced by the Commission undertaking an investigation, and taking any subsequent action considered appropriate.^{Footnote 234} However, unlike in Victoria, in the NT the Commission can enter into an enforceable undertaking.^{Footnote 235} In the ACT, the positive duty is not to be enforced by the Commission; instead, compliance with the duty can be taken into account in conciliating and adjudicating discrimination complaints.^{Footnote 236} At the federal level, the AHRC will be empowered to inquire into compliance with the positive duty;^{Footnote 237} in cases of non-compliance, this can lead to a compliance notice,^{Footnote 238} or enforceable undertaking.^{Footnote 239} The WA Law Reform Commission goes further, recommending both escalating compliance powers for the WA Equal Opportunity Commission and individual enforcement mechanisms for a positive duty.^{Footnote 240} This may illustrate mutual learning and experimentation across and between jurisdictions, particularly in rectifying some of the enforcement gaps in Victoria. It remains to be seen whether the individualised, complaint-driven approach to the enforcement of positive duties in the ACT proves more effective than the investigative, Commission-driven approach in Victoria, the NT and federally. Thus, even in adopting successful experiments from other jurisdictions, there is still evidence of diversity, evaluation and innovation in this space.

This, then, flags the importance of enabling mutual learning and evaluation across and between jurisdictions, particularly in the absence of a structured process of peer review or a central unit that facilitates mutual learning, as envisaged by democratic experimentalism. In a contemporary setting, processes to facilitate information exchange and learning between the states have been established through the government agencies operating in each jurisdiction: the Australian Council of Human Rights Agencies (‘ACHRA’) provides a forum for engagement between human rights bodies at the federal, state and territory level, to consider issues of mutual concern and emerging developments.^{Footnote 241} ACHRA has become a forum for the sharing of resources and the development of joint submissions, creating broader synergies and efficiencies across jurisdictions.^{Footnote 242}

The Australian Discrimination Law Experts Group (‘ADLEG’), a group of academics with a shared interest and expertise in equality law, has also come to represent a critical enabler of mutual learning across jurisdictions. Through scholarly submissions to government inquiries, reviews and consultations, ADLEG has facilitated the transmission of ideas and ‘experiments’ across Australian jurisdictions, including positive equality duties. In the WA Law Reform Commission review of discrimination law, for example, ADLEG’s submissions were cited 42 times.^{Footnote 243} By working collectively, ADLEG has been able to pool scholarly knowledge across jurisdictions, identify and communicate critical experiments to policymakers, and encourage innovation in the reform process. That said, relying on the individual goodwill of time-poor academics to overcome capacity constraints of individual states and territories is potentially not sustainable in the longer-term, especially given the sheer number of reviews, consultations and inquiries being conducted in the field. A long-term, structured process to enable mutual learning is required.