Introduction
When five food delivery riders were killed on Australian roads within roughly two months in late 2020, the deaths gave sudden and grim visibility to a question that had circulated in industrial relations scholarship for a decade: who is responsible for the people who work through digital platforms? The question is more than rhetorical. Under the Fair Work Act 2009 (Cth), almost every workplace protection, from the national minimum wage and the National Employment Standards to unfair dismissal remedies, attaches to the status of “employee”. Workers’ compensation, paid leave and superannuation follow largely the same gateway. Platform businesses such as Uber, DoorDash and Menulog built their Australian operations on the other side of that line, engaging workers as independent contractors and thereby externalising costs that employment law would otherwise impose.
This essay traces the trajectory of Australian law from the High Court’s contractual turn in 2022, through the legislative definition reset of 2024, to the Closing Loopholes regime for “employee-like workers”. It argues that the 2024 settlement, which pairs a substance-based test of employment with Fair Work Commission power to set minimum standards for platform work, is a defensible and distinctively Australian architecture because it refuses to treat flexibility and protection as mutually exclusive goods. The settlement is nonetheless incomplete: until safety and insurance parity is addressed, the riskiest work in the platform economy will remain the least protected.
The Employee/Contractor Boundary and the High Court’s Contractual Turn
Australian law has traditionally policed the boundary between employment and independent contracting through a multifactorial test. From Stevens v Brodribb Sawmilling Co Pty Ltd (1986) onward, courts weighed control, integration, provision of equipment, capacity to delegate and mode of remuneration, examining the totality of the relationship rather than the label the parties attached to it. In Hollis v Vabu Pty Ltd (2001), the High Court applied that approach to bicycle couriers and found them to be employees notwithstanding their designation as contractors, because in practice they were closely directed, wore the company’s livery and had no realistic capacity to build independent goodwill (Stewart et al. 2016). The method assumed that reality trumps form: a business could not simply draft its way out of employment obligations.
That assumption was substantially unsettled in 2022. In Construction, Forestry, Maritime, Mining and Energy Union v Personnel Contracting Pty Ltd [2022] HCA 1 and ZG Operations Australia Pty Ltd v Jamsek [2022] HCA 2, the High Court held that where the parties have comprehensively committed their relationship to a written contract, characterisation is determined by the rights and obligations established in that contract, and the manner in which the parties subsequently conduct themselves is generally irrelevant. The immediate outcomes were mixed: the labourer in Personnel Contracting was held to be an employee because the contract itself conferred a right of control, while the truck drivers in Jamsek, who had formed partnerships and purchased their vehicles decades earlier, were confirmed as contractors. The methodological shift, however, favoured whoever holds the pen. Platform terms of service are standard-form documents drafted unilaterally and accepted by tapping a screen, so contractual primacy effectively handed classification power to the party with every incentive to avoid employment (Forsyth 2024).
The consequences appeared almost immediately. At first instance, the Fair Work Commission had found that Diego Franco, a Deliveroo rider, was an employee, relying heavily on how the work was actually performed. On appeal in Deliveroo Australia Pty Ltd v Franco (2022), the Full Bench reversed, holding that Personnel Contracting confined its analysis to the written terms, which pointed to independent contracting, while candidly observing that the result sat uneasily with the practical reality of Franco’s work. The case crystallised the central objection to the contractual turn: employment status had become an artefact of drafting sophistication rather than a description of economic life.
The 2024 Definition Reset and Its Limits
The legislative response arrived in the Fair Work Legislation Amendment (Closing Loopholes No. 2) Act 2024 (Cth). From 26 August 2024, a new interpretive provision requires that whether a person is an employee for the purposes of the Fair Work Act be determined by ascertaining the “real substance, practical reality and true nature” of the relationship, having regard to the totality of the relationship, including how the contract is performed in practice. The provision legislatively restores the pre-2022 orthodoxy associated with Hollis v Vabu, while an opt-out mechanism allows contractors earning above a high-income threshold to elect out of the new test, preserving genuinely commercial arrangements.
The reset was necessary, but it would be a mistake to regard it as sufficient for the platform economy. Even under a practical-reality test, many platform workers may still be classified as contractors: they typically choose when to log on, may refuse individual tasks without formal sanction and often work across competing apps simultaneously, features that tribunals have consistently weighed against the existence of a continuing work-wages bargain. Stewart and Stanford (2017) anticipated precisely this difficulty when cataloguing the regulatory options for gig work, observing that clarifying the definition of employment addresses misclassification at the margins but leaves genuinely novel work arrangements outside the protective net. A boundary, however honestly drawn, still produces an outside. What Australian law required was a mechanism for extending protection to those who remain, on any test, beyond the boundary.
Minimum Standards for Employee-Like Workers: A Third Pathway
That mechanism is the most innovative element of the Closing Loopholes reforms. The same 2024 Act inserted a new chapter into the Fair Work Act empowering the Fair Work Commission to regulate “employee-like workers” who perform work through digital labour platforms. A worker qualifies where at least two indicia are present: low bargaining power, remuneration at or below the level of employees performing comparable work, or a low degree of authority over the performance of the work. For such workers the Commission may issue binding minimum standards orders, and softer non-binding guidelines, dealing with matters including payment terms, deductions, record-keeping, insurance, consultation and representation. Two exclusions define the model’s philosophy: orders must not prescribe rostering arrangements or overtime rates, and they cannot convert contractors into employees. Protection is deliberately decoupled from reclassification.
The regime also confronts algorithmic discipline directly. Employee-like workers who have worked regularly through a platform for at least six months may challenge “unfair deactivation” before the Commission, which can order restoration of a worker’s account, with platforms expected to comply with a Digital Labour Platform Deactivation Code. A collective stream permits registered agreements between platforms and unions. The first applications for minimum standards orders, brought by the Transport Workers’ Union in road transport and on-demand delivery, were before the Commission within months of commencement (Forsyth 2024).
Three features make this architecture defensible. First, it targets the protection deficit directly instead of forcing workers into all-or-nothing status litigation whose outcome, as Franco demonstrated, may turn on drafting. Secondly, it is institutionally continuous with the Australian award tradition: standard-setting by an independent tribunal after public hearings is what the Commission and its predecessors have done since 1904, which lends the regime a legitimacy and administrative capability that a bespoke new agency would lack. Thirdly, the scale of the phenomenon justifies dedicated machinery. A national survey found that around 7 per cent of adults had undertaken digital platform work in the previous twelve months, with participation concentrated among younger workers and recent migrants (McDonald et al. 2019), while Australian Bureau of Statistics data show the independent contractor share of employment holding near 8 per cent across the past decade, suggesting that platforms are reorganising existing precarious work rather than creating a new entrepreneurial class (Australian Bureau of Statistics 2023). Notably, even the industry’s largest firm conceded ground: Uber and the Transport Workers’ Union jointly endorsed the creation of a federal standards-setting function for platform work in 2022, well before the legislation passed.
Safety and Insurance: The Unfinished Business
If minimum pay is the regime’s centrepiece, safety is its conscience. The Senate Select Committee on Job Security (2022) documented the structural transfer of risk in on-demand delivery: piece rates that reward speed, algorithmic metrics that penalise delay and a workforce riding through traffic in all weather. The 2020 deaths prompted a joint government taskforce in New South Wales and, subsequently, work health and safety regulations obliging food delivery platforms to provide training and high-visibility equipment to riders. Importantly, the model work health and safety laws already applied: their definition of “worker” extends to contractors, so platforms owe duties as persons conducting a business or undertaking regardless of employment status.
Compensation is a different matter. State and territory workers’ compensation schemes remain tied to employment, subject only to limited deeming schedules, so injured platform workers generally fall outside statutory no-fault cover and rely instead on group personal accident policies procured by the platforms themselves. The Actuaries Institute (2020) found such policies inferior on every relevant dimension: capped lump sums, short income-replacement periods, no structured rehabilitation and cover contingent on being logged into the app at the moment of injury. Families of riders killed at work received a fraction of what statutory death benefits would have provided (Senate Select Committee on Job Security 2022). The Victorian On-Demand Workforce Inquiry recommended aligning entitlements across work status for precisely this reason (James 2020).
Minimum standards orders can now require platforms to hold insurance, which is a genuine advance, but a tribunal-mandated private policy is a second-best instrument. Full parity requires the states to extend deemed-worker coverage to platform work, coordinated nationally through Safe Work Australia’s workers’ compensation policy functions. Until that occurs, the most dangerous work in the platform economy will continue to carry the weakest safety net, a distributional outcome that no defensible theory of labour regulation supports.
Flexibility Versus Protection: Objections and Rebuttal
The principal objection to this settlement is that regulation jeopardises the flexibility that attracts workers to platforms in the first place. In submissions to the Senate inquiry, platforms argued that most of their workforce performs short and variable weekly hours, values the ability to log on around study, caring responsibilities and other jobs, and would be harmed if minimum standards raised costs, dampened consumer demand and rationed available work (Senate Select Committee on Job Security 2022). Menulog’s attempt to trial direct employment in Sydney, wound back after the company struggled to fit on-demand scheduling within existing award structures, is cited as proof that employment obligations and platform logistics are incompatible, while Deliveroo’s withdrawal from Australia in 2022 is invoked as a warning that the market cannot absorb additional burdens. A second, more scholarly objection warns against intermediate categories as such: De Stefano (2016) argues that third statuses can become escape hatches, inviting firms to shift workers who would otherwise be employees into a cheaper tier, an outcome debated in the United Kingdom’s “limb (b) worker” experience even after Uber BV v Aslam (2021).
Neither objection survives close scrutiny of the Australian design. The flexibility argument mistakes the content of the regime: minimum standards orders are expressly precluded from prescribing rostering and overtime, so no order can compel a shift, and the freedom to log on at will is untouched. Australia has, in any case, been running the relevant experiment for more than a century. Casual employment couples complete scheduling flexibility with an award wage floor and a 25 per cent loading, demonstrating that flexibility and minimum standards are administratively compatible (Stewart & Stanford 2017). The autonomy said to be at stake is also overstated. Labour process research on Australian food delivery shows that platforms exercise granular control through dispatch algorithms, acceptance-rate metrics and information asymmetries, which amounts to control without the correlative obligations of employment (Veen, Barratt & Goods 2020). Stanford (2017) situates the point historically: gig work is piecework and outwork digitised, and Australian law successfully regulated piece rates for over a century without abolishing task-based work. As for Deliveroo, its exit occurred in November 2022, before any of these reforms commenced and under some of the lightest regulatory settings in the developed world; competitive consolidation, not regulation, explains the departure.
The levelling-down objection is more serious, but it is answered by the settlement’s dual structure. Because the employee-like regime operates alongside the restored substance-based definition of employment and the Act’s sham contracting prohibitions, a platform cannot lawfully park de facto employees in the cheaper tier: the boundary is policed even as the floor beneath contractors rises. The Australian third pathway adds rights to workers who would otherwise have none and subtracts nothing from employees. The relevant comparison is therefore not employee-like status versus employment, but employee-like status versus the unregulated void that preceded it.
Conclusion
The trajectory from Personnel Contracting to the Closing Loopholes reforms records a legal system correcting itself. The contractual turn demonstrated that a judicially drawn boundary, once anchored to the written instrument, could be captured by the stronger drafting party; the 2024 definition reset restored substance; and the employee-like regime did something more important than either, accepting that some work will lawfully sit outside employment and that basic protections should reach it anyway. This essay has argued that the resulting architecture is defensible precisely because it dissolves the supposed choice between flexibility and protection: the Fair Work Commission may set floors, but it may not touch the scheduling freedom that gives platform work its value to workers. The unfinished business is compensation parity for injured platform workers, which lies substantially with the states, together with adequate enforcement resourcing and honest empirical evaluation of the first minimum standards orders. Australia has, characteristically, chosen institutional adaptation over either deregulation or prohibition. If the early orders hold pay floors in place without measurable loss of flexibility, the model will repay close attention well beyond Australian shores.
References
Actuaries Institute 2020, The Rise of the Gig Economy and its Impact on the Australian Workforce, Green Paper, Actuaries Institute, Sydney.
Australian Bureau of Statistics 2023, Characteristics of Employment, Australia, August 2023, ABS, Canberra.
De Stefano, V 2016, ‘The rise of the “just-in-time workforce”: on-demand work, crowdwork and labor protection in the “gig-economy”‘, Comparative Labor Law & Policy Journal, vol. 37, no. 3, pp. 471-504.
Fair Work Act 2009 (Cth).
Fair Work Legislation Amendment (Closing Loopholes No. 2) Act 2024 (Cth).
Forsyth, A 2024, ‘Closing the loopholes? Regulating employee-like work and the digital platform economy in Australia’, Australian Journal of Labour Law, vol. 37, no. 1, pp. 1-28.
James, N 2020, Report of the Inquiry into the Victorian On-Demand Workforce, Industrial Relations Victoria, Melbourne.
McDonald, P, Williams, P, Stewart, A, Mayes, R & Oliver, D 2019, Digital Platform Work in Australia: Prevalence, Nature and Impact, Queensland University of Technology, Brisbane.
Senate Select Committee on Job Security 2022, First Interim Report: On-Demand Platform Work in Australia, Commonwealth of Australia, Canberra.
Stanford, J 2017, ‘The resurgence of gig work: historical and theoretical perspectives’, The Economic and Labour Relations Review, vol. 28, no. 3, pp. 382-401.
Stewart, A, Forsyth, A, Irving, M, Johnstone, R & McCrystal, S 2016, Creighton and Stewart’s Labour Law, 6th edn, Federation Press, Sydney.
Stewart, A & Stanford, J 2017, ‘Regulating work in the gig economy: what are the options?’, The Economic and Labour Relations Review, vol. 28, no. 3, pp. 420-437.
Veen, A, Barratt, T & Goods, C 2020, ‘Platform-capital’s “app-etite” for control: a labour process analysis of food-delivery work in Australia’, Work, Employment and Society, vol. 34, no. 3, pp. 388-406.