Minding the gap? Blind spots in the ILO's and the EU's perspective on anti-forced labour policy

Abstract

This article critically examines the EU's recent proposal to ban products made with forced labour from its market, which adopts the ILO's definition of ‘forced labour’ as outlined in the Forced Labour Convention of 1930 (No. 29). The authors argue that the EU's endorsement of the ILO's approach is problematic due to two flawed assumptions: (1) the definition of ‘forced labour’ is universally accepted across the EU, and (2) it is well-suited to combat forced labour in contemporary supply chains. Through an analysis of ongoing debates between the ILO and its Member States, the article demonstrates a lack of consensus regarding the interpretation of the Convention, particularly in the context of hybrid public-private prison labour arrangements. Furthermore, introducing a new Hybrid Multi-Dimensional (HMD) model for analysing contemporary prison labour practices, the article reveals blind spots in the ILO's approach that may inadvertently allow the incorporation of prison labour into supply chains, contrary to the EU's objectives. The article argues that the EU's unequivocal endorsement of the ILO's definition disregards these fundamental issues and may hinder the effective implementation of its proposed ban. The authors suggest that the HMD model offers a more comprehensive framework for analysing the complex realities of modern prison labour and could provide a roadmap for resolving the ILO-States debate. The article concludes that the EU should reconsider its wholesale adoption of the ILO's approach in light of the HMD model's insights in order to fulfil the objectives of its proposed forced labour product ban.

1. Introduction: the EU's anti-forced labour proposal and the ILO

The fight against forced labour seems to be expanding today in all directions. Increasingly, it branches out, beyond labour law, criminal, and international human rights law, to trade and customs law. ¹ Likewise, it extends beyond the state, domestically and internationally, as non-governmental household companies adopt policies altogether excluding the sale of local and imported goods whose production incorporated some form of forced labour, ² in a similar vein to the comparable prohibition imposed by leading industrial countries, such as the US and Canada, on the importation of products created ‘wholly or in part’ by forced labour. ³ Internationally, significant bilateral and multilateral trade agreements growingly feature anti-forced labour provisions. ⁴

And now the EU is about to throw its considerable support behind the fight, as it is in the process of adopting a proposal to ban products manufactured with forced labour from the Union market (hereafter: the EU Proposal, or the Proposal). ⁵ The Proposal's ambit is undoubtedly extensive. It is projected to encompass all types of domestic or imported products incorporating – wholly or in part and at any stage of the production – forced labour. Correlatively, it arms Member States with the power to detain, seize, or order the withdrawal of a product manufactured, either wholly or in part, with forced labour. ⁶

Two aspects of the Proposal merit special attention. First, it explicitly adopts the definition given in the International Labour Organisation Forced Labour Convention 1930 (No. 29) (hereafter: the ILO Convention, or the Convention) to ‘forced labour.’ ⁷ Second, the Proposal specifically addresses supply chains, stating that any instance of forced labour within supply chains is detrimental to fair competition in both domestic and international trade, and it does so for a good reason: three-quarters of the forced labour instances are attributed to private entities, many of which operate in export-related sectors involved in supply chains. ⁸ Given the scale of the issue it seeks to address – with an estimated 28 million forced workers around the world – and the limited success in hitherto tackling forced labour throughout corporate supply chains, the Proposal seems to be a much-needed move in the right direction.

However, the Proposal has already drawn criticism, characterising it, for example, as another EU protectionist, rather than a rights-oriented, measure. ⁹ Its reliance on custom regulation with a view to upholding desirable labour standards outside of the EU has also been challenged as ineffective. ¹⁰ No academic attention, however, has yet been given to the Proposal's two basic, fundamental flaws that might render it not only ineffective but inoperative, whatever its motivation may be. Both emanate from the Proposal's gateway: its (actually, the ILO's) operative definition of ‘forced labour,’ as applied to convict labour – a pervasive yet understudied category of labour. ¹¹

Simply put, the Proposal seems to proceed on the following two implicit assumptions regarding forced labour: the ILO Convention's definition of forced labour is (1) universally accepted across the EU, ¹² and (2) well-suited to the current fight against forced labour, particularly in supply chain settings. Yet, this article will show that both assumptions cannot stand, even though ‘forced labour’ as defined in the Convention has universally become a benchmark, reverberating even into domestic legislation in the US. ¹³ The article will show, first, that the definition of ‘forced labour’ in the Convention is all but consensual; rather, several states – leading EU Member States, such as Germany and Austria, among them – have openly criticised and even rejected the ILO's approach thereto. ¹⁴ Second, the ILO's approach to forced labour might exempt some forms of supply chain production from the protection provided by the Convention, thereby undermining the EU's professed commitment to ‘levelling the playing field’ production across supply chains.

The EU's unequivocal and unreserved ‘cut-and-paste’ endorsement of the ILO's definition of forced labour therefore disregards a profound definitional disagreement, whose exposure requires a deep understanding of the Convention and its treatment by the ILO, especially in the prison labour context. This article, dedicated to a critical engagement with the Convention's ‘forced labour’ definition as it relates to prison labour, will explain the terms of the disagreement, put it in context, and provide indeed a deeper understanding of the difficulties embedded in the ILO's understanding of ‘forced labour.’ While the debate in the EU has touched upon recurrent discussions surrounding the Convention, for example, its underlying rationales (protectionism/labour rights), or whom it is meant to primarily protect (local producers/forced labourers), ¹⁵ it has thus far overlooked the foundational issues identified in the article. Still, our analysis will be relevant not only to the EU, for – as we will argue – the ILO's current approach advances a somewhat incomplete, and even anachronistic view that disregards contemporary practices of inmates’ production, especially in the context of supply chains, in the EU and beyond.

How will the article debunk the two assumptions referenced above, and what conclusions will it draw? It will handle the issue of disagreement on the proper interpretation of the Convention by detailing a line of controversies on the proper interpretation of ‘forced labour,’ especially in the context of hybrid prison labour, where both public and private entities are involved in various ways in the production or service provision. Hybrid prison labour, in particular, divided the ILO and several of the most prominent Convention's signatory States - again, several key EU Member States included. The argument to be followed from this finding is that in light of the global spread of such hybrid schemes and the challenges in their regulation per the Convention, the ILO-States controversies might be rehashed within the EU as it moves to implement the Proposal (should it be adopted), thus bringing home to the EU the overlooked fact that the Proposal is based on unsettled foundations - so much so, that it may argued that it is not clear what the Proposal actually holds, especially in the context of hybrid prison labour. Alarming as it may sound, this argument stands, largely notwithstanding the substance of the ILO-States controversies. It is valid regardless of the question of whether the ILO is in the right in the controversies. As a matter of simple legal draftsmanship, deep disagreement regarding the main operative term of a norm does not bode well for its future.

But is the ILO approach to hybrid prison labour defendable on its merits? Assessing the ILO's treatment of evolving configurations of private contractors’ involvement in prison labour – brought into sharp relief by supply chain production schemes – will require a bolder methodological approach. We will turn to groundbreaking publicness theories to scrutinise the ILO's normative perspective on prison labour. Publicness theories call into question the conventional binary between public and private organisations, positioning them on a spectrum stretching between these two poles. It argues (paraphrasing Tolstoy) that while all organisations are alike since all organisations are hybrid in nature, each organisation is hybrid in its own way, namely, different organisations’ constitutions and operations result in different calibrations of their public/private dimensions. Drawing on Publicness and comparable non-binary conceptualisations of prison labour modi operandi, the article will outline a comprehensive framework for the empirical and analytical study of the dynamic, growingly hybrid realities of prison labour, our Hybrid Multi-Dimensional (HMD) model.

As we shall demonstrate, HMD unmasks important blind spots in the ILO's outlook on private involvement in prison labour, notably, the involvement of prison labour in supply chains, thus offering eye-opening insights into the EU Proposal. In general, HMD may meaningfully advance the extant discussion surrounding prison labour – in the ILO and elsewhere – due to its realistic outlook, which allows for a fuller engagement with the contemporary kaleidoscope of prison labour. HMD may also advance the ongoing ILO-State polemic by pointing to the harbingers of critical areas of accord between the ILO and Member States, concerning the methodology of the inquiry into the nature of the activity concerned – its (spatial, organisational, and economic) details and arrangements. Further, HMD may also take the debate forward by laying out a self-conscious, full methodological model on which the divided parties, and especially the ILO, can draw in their ongoing search for the appropriate framework for the regulation of prison labour.

The article will follow this line of argument. The first section will briefly explore the Convention's definition of ‘forced labour’ and present the debate surrounding its treatment by the ILO. The second section will introduce Publicness theories and our HMD model and re-evaluate the ILO's handling of the private-public binary. Thus, the first section will bring to the surface an acute definitional disagreement, thereby undermining the EU's first assumption. The next will reveal, based on HMD, that the ILO – soon to be the EU – approach to ‘forced labour’ does not protect prisoners who are compelled to work from involvement in the supply chain business, contrary to the EU's intentions – thus undercutting what is presumably the EU's second assumption. The second section will therefore show that, although prisoners and prison labour are not mentioned in the Proposal, they may be adversely affected by it – in a particularly vulnerable setting. The final section will conclude.

2. Mapping the disputed terrain: the debate surrounding the ILO's ‘State exemption’ approach

Hybrid configurations of prison labour, where both private contractors and public authorities play a role in inmates’ labour schemes, are ubiquitous nowadays. Yet, their regulation is far from settled. For some years now, the appropriate framework for regulating hybrid prison labour has divided the ILO and several of its leading Member States, particularly, those considered industrialised – EU members numbered among them. In previous scholarly work, we have elucidated the understudied and acrimonious debate, ¹⁶ whereby States characterised the ILO's approach to prison labour as outdated, pointing out that the Convention was drafted almost a century ago. In the following lines we demonstrate that, as anticipated in our prior scholarly work, the debate is still very alive, kicking, and branching out, drawing in new players and novel arrangements.

Let us briefly summarise the debate, starting with the essentials of the ILO's approach. The pertinent provisions of the Convention regarding what constitutes ‘forced labour’ in carceral settings are anchored in Articles 1 and 2 of the Convention. Article 1, paragraph 1 declares that all members ‘undertake to suppress the use of forced or compulsory labour in all its forms.’ Article 2 further clarifies that ‘forced or compulsory labour’ refers to ‘all service which is exacted from any person under the menace of any penalty and for which the said person has not offered himself voluntarily.’ But the Convention quickly moves to exempt several labour activities from that definition, thus excluding them from its purview.

‘[A]ny work of service from any person as a consequence of a conviction in a court of law’ ¹⁷ is thus set aside by the Convention, an exemption commonly interpreted as enabling prisoners’ participation in public endeavours, such as constructing and maintaining essential infrastructure. ¹⁸ However, as provided by the Convention, the exemption is subject to two specific conditions: (1) the excluded work must be conducted under the supervision and control of a public authority, and (2) it must not not involve hiring out or placing prisoners at the disposal of private individuals, companies, or associations. ¹⁹ Taken together, the Convention's provisions dealing with prison labour seek to protect inmates working for private entities while reserving a (public) sphere of ‘State Exemption’ in the employment of prisoners. ²⁰

Analytically, the ongoing ILO-States debate is twofold: comprising a definitional (essentially, empirical) stage and an outcome (normative) stage, both of which are intricately shaped by a sharp dichotomy between the public and private spheres. In the initial stage, the ILO determines whether a specific arrangement or organisation of prisoners’ work belongs in the State exemption category, giving States’ public agencies free rein in regulating work exacted from incarcerated individuals. Subsequently, in the second stage, the consequences of falling outside the purview of the State exemption are fleshed out. The ILO outlines demanding requirements for work practices that do not fall within this exemption due to their private elements, insisting that all such work arrangements remain within the Convention's purview and therefore should be conducted on a fully consensual, voluntary basis. ²¹ Moreover, the ILO stipulates that the provision of ‘market conditions’ should serve as an indication of voluntariness, effectively rendering their provision a requirement that must be met lest the employment at issue be considered ‘forced labour.’ ²² Now, ‘market conditions’ do not come in cheap. They revolve mainly around labourers’ compensation and other ‘cash standards.’ Herein, we can understand why the State exemption debate between States and the ILO has become so intense.

Thus, it is in the first stage that States typically portray various formations of prison labour where private elements figure as public (e.g., on the argument that public bodies oversee the operation), and therefore argue that they should be (leniently) regulated as such. To States’ frustration, the ILO typically rejects the classification offered by the States, moving to demand that inmates should be afforded (greater) protections suitable to employment by private contractors. The two parties appear to talk past each other as the classificatory dispute rages on. In these circumstances, again, it is truly surprising that the ILO's approach would be endorsed by the EU et al.

Numerous States have rejected the ILO's approach, asserting that it is rigid, antiquated, and incongruous with the current trend of expanding private sector involvement in prison labour. ²³ They posit that the ILO's public-private distinction is unrealistic, given that the two domains have become increasingly intertwined in their use of prison labour. ²⁴ Consequently, the ILO has been blamed for hindering private entities’ beneficial and productive involvement in prison labour. As suggested above, however, Member States are not alone in taking a critical stance towards the ILO's approach. Employers’ organisations, too, had contended that the Convention failed to address the privatisation of prisons, as ‘there had been hardly any private prisons in existence when the Convention had been adopted.’ ²⁵ Challenging the importance given to voluntariness and market conditions as an indicator thereof, ²⁶ in the general discussions of the Conference Committee of 1998 employers’ organisations further argued that, at the time of the Convention's drafting, ‘neither the consent of the prisoner nor (almost) normal conditions of work were indispensable prerequisites for allowing the performance of labour by convicted persons for private enterprises.’ ²⁷ Therefore, in concert with Member States, employers’ organisations have urged the ILO to revise its approach, advocating for greater flexibility when addressing forced labour by private entities and States.

In our previous work, we have pointed out how central the private-public distinction has been to the ILO's policies, and anticipated that this debate would persist unless the ILO changed course. ²⁸ Unsurprisingly, as already noted, reviewing recent ILO supervisory system documents confirms our predictions. For example, the recent ILO-Australia dispute demonstrates Australia's unwavering belief that the benefits of prisoners working for private entities, even absent the safeguards anticipated by the ILO's position and when its standards of voluntariness are not met, outweigh their drawbacks. ²⁹ As it considers legitimate and even welcomes settings where carceral institutions involve private entities, finding little reason to raise the bar in respect of voluntariness on that account, Australia gave little indication that it would reform its law to conform with the Convention. ³⁰ As before, the ILO remains unconvinced and expresses its ‘firm hope’ that Australia will take appropriate measures to ensure conformity with the Convention. ³¹

Turning to EU States, Germany, too, openly defies the ILO's position on the compulsory work of prisoners in privately operated workshops, situated in many German state prisons. The latter arrangement has sparked an ongoing discussion, whose resolution is not in sight, as the very recent, 2024 ILO Report makes clear. In the 2024 Report, the ILO posits that Germany's failure to legally require that all work performed for private undertakings is based on the prisoners’ freely given, formal and informed consent, and done under conditions ‘approximating a free labour relationship’, constitutes a violation of the Convention. ³² To the ILO, Germany is in a large-scale violation of this standard, and the ILO provides a list of the large number of prisoners working for private enterprises, but not in compliance with the Convention's terms, throughout Germany. ³³ Germany's reply follows a now-familiar argument, pointing to the limited legal authority granted to private entities to instruct prisoners, as well as to contractual agreements specifying that prison authorities are obligated to prevent prisoners from complete integration into the operations of the private enterprise. Predictably, the German Government's position is supported by the Confederation of German Employers’ Associations, arguing that the compulsory labour of prisoners serves as a method of resocialisation, and that prison authorities maintain responsibility for prisoners and cannot delegate their custody to private entities. ³⁴ The Confederation further suggests that, as a practical matter, the State is unable to employ inmates, making their engagement in the private sector a necessity, provided that their work conditions are not exploitative.

If history serves, the debate will rage on and on, unresolved. ³⁵ The parties seem entrenched, and their differences unbridgeable. Now, since the EU Proposal does seem to presume consensus on the Convention's forced labour definition or assumes compliance among EU members, it should take these recent examples as a cautionary tale. Obviously, EU States’ confrontation with the ILO over the very definition of ‘forced labour’ – that is, their lack of consensus on the matter – is liable to encumber, if not entirely frustrate, any EU attempt to translate the Proposal, as it relates to prison labour, into a reality.

The following section will suggest a potential way to move forward productively. First, we introduce the HMD model and elaborate on its usefulness in analysing practices of prison labour. As we hope to show, given the centrality of the private-public divide in the ILO debate and generally, in the global discourse surrounding prison labour, the promise HMD holds is considerable. Second, we will point to implicit use of HMD-bound consideration by both the ILO and States, thus exposing what could become a common ground for the parties to fruitfully explore with the help of HMD. Third, we will touch upon the normative level of the discussion, ³⁶ as we briefly outline what normative challenges and opportunities may emerge following HMD.

3. From binary to multi-dimensional perspective of prison labour

A. Introducing HMD

What is HMD about, then? Drawing on the Dimensional Publicness school, we have previously constructed a comprehensive framework with which to roundly explore extant diverse prison-labour practices and organisations and their challenges. ³⁷ A traditional doctrinal legal framework, which satisfies itself with a singular, dichotomous ‘core’ public-private criterion, such as work provider ownership or formal legal status, leaves out pivotal sides of prison labour practices, certainly in their contemporary hybrid forms. Following Dimensional Publicness, our approach, too, replaces the public-private binary conceptualisation of social activities with an ‘empirical publicness grid.’ The grid is charted along the two axes of Economic and Public Authority, reflecting the simple fact that organisations operate, albeit to varying degrees, within both the private and public spheres. ³⁸ Furthermore, our research indicates that Dimensional Publicness has to be supplemented with another layer of considerations to more fully capture the peculiarities of modern prison labour schemes. Ultimately, we have introduced a unique framework for studying prison labour, consisting of dimensions identified in Publicness theories alongside additional dimensions specifically applied to the business of prison labour. The resultant model, we have argued, may greatly advance legal conceptualisations of contemporary forms of prison labour organisations. ³⁹ In what follows, we present the HMD model, which is an upgraded version of our previous somewhat embryonic framework, revised to reflect insights from our recent research on the growingly hybrid nature of prison labour systems. Hopefully, this section will illustrate the intricacy and profundity of the HMD model and its advantages as an empirical analysis framework.

The HMD model focuses on six dimensions (or indicators) of the publicness of prison labour practices, each dimension representing a continuum. We will now present the six dimensions, and explicate their relevance for the inquiry into the nature of prison labour schemes. As we shall see, viewed through these dimensions, the publicness of pertinent organisations and the operation of prison labour worldwide could be insightfully reassessed.

We begin with the dimension of (1) ownership, which asks: who owns the organisation? The carceral facility? The state? Private individual, political communities, etc.? Answering this question is paramount in countries implementing private-ownership models for their correctional institutions. In accordance with ILO standards, prison labour conducted in these institutions would breach Convention No. 29, unless it meets ILO's exacting criteria for voluntary work. ⁴⁰

Private prisons, however, are merely one setting of private prison labour. Public ownership of prisons by no means ensures the publicness of prison work, especially considering the myriad manners in which private contractors are today weaved into schemes of prison labour. ⁴¹ Further, our research reveals that once incorporated into the operation, private pressures are often brought to bear on public prison (labour) agencies, namely, those administrative agencies responsible for the operation of prison labour, which, as state agencies, are presumed to be directed by public considerations. And this is so even when responsibility for prisoners’ care, custody, and discipline is reserved for public prisons – and herein enters the second dimension, (2) the locus of responsibility for prisoners’ custody.

While surely noteworthy, a divergence between these first two dimensions is not rare – for although ownership typically leads to responsibility for the prisoners’ care, custody, and discipline, this is not always true – nor is it a new phenomenon. In fact, it is quite prevalent in the history of incarceration in the West. For instance, in the past, publicly owned prisons in the United States would lease prisoners to commercial companies for extended periods, making the latter businesses responsible for providing basic needs such as food, care and disciplinary supervision. ⁴² Though these practices have been mostly abandoned, at least in industrialised countries, it is important to recognise that responsibility for inmates’ custody and discipline remains a crucial issue even today.

After all, experience shows that even public prisons, whose daily operation is reserved to public agencies, are usually not private-free zones. For example, as is frequently the case, public prisons may contract for private services, giving rise to ambiguities regarding the publicness of these facilities. ⁴³ This is even more so the case if private contractors are featured in the prison's employment scheme, as often, for prisoners, the rule of ‘what happens at work stays at work’ simply does not apply. For instance, in countries with ‘work-release’ programmes, inmates often reside in separate ‘no drugs’/rehabilitative facilities, which may offer prisoners a more supportive environment, where they are encouraged to stay in contact with family members and have regular, direct contact with people outside the prison. Additionally, prisoners residing in such facilities are also provided with better living conditions, and offered assistance with drug abuse or mental health issues. ⁴⁴ Crucially important for our purposes, the input of a private employer in charge of such a rehabilitation-oriented programme may significantly influence inmates’ admission to, and stay over time in, such privileged facilities. ⁴⁵ Thus, the private employer may wield considerable influence over a great many aspects of the prisoner-worker's life. In light of the ubiquity of such arrangements, a multifaceted examination, as advocated here, is indispensable for comprehending the interdependencies between various aspects of the prisoners’ work and overall living conditions.

But the argument could be further pushed. In fact, there are good reasons for arguing that the institution of the modern prison is a hybrid private-public institution by its very nature. After all, prisons are designated as prisoners’ ‘homes’ for the duration of their imprisonment, while remaining public institutions. ⁴⁶ Within classical liberalism, the public-private distinction reverberated beyond the divide between the private and public spheres (the market and the family versus the state), well into the private sphere. The latter move grants the family greater private credentials than the market so that the family is considered private vis-à-vis the market (while the two - the market and the family - are considered private vis-à-vis the state). ⁴⁷ Thus viewed, the inmates reside in a public institution, which resembles the quintessential private institutions of one's family and home. ⁴⁸

Moving on, irrespective of the type of ownership, another critical consideration to be taken into account is (3) the funding sources that support the prison system and its labour practices. Specifically, it is crucial to ascertain the relative proportion of funding from public sources, such as taxes or incentives given to prisoners’ employers in the form of tax exemptions, and private sources, such as payments made by businesses that purchase prisoner-manufactured goods. ⁴⁹ Varied as funding sources may be, they all relate to the allocation of costs and benefits within the prisoners-state-private contractor triangle. This is illustrated by the tax exemptions example: when available, they amount to a transfer of costs incurred by the employment of prisoners from prisoners-workers, whose reduced compensation may effectively subsidise their employment, to the state. ⁵⁰

Funding is therefore a decisive component in the business of prison labour. It directly impacts prisoners-workers and may reveal how (and in what ways) the state/private sector is implicated in the operation and economic/social rationales underlying it (e.g., is revenue-generating prioritised over inmates’ compensation?). Finally, it should be borne in mind that, as with other dimensions, the funding dimension is also clearly not binary, for state subsidies and private funds may come in different forms.

The (4) market dimension scrutinises the presence of prisoner-produced goods or services in the market. States are divided on this. Some perceive it as problematic on the argument that prison production enjoys competitive edge, since it is often allowed to disregard market conditions and minimum working conditions customary in the market. This concern is widespread and, as already noted, has led some key global players to embrace policies excluding from their markets products and services originating in forced labour. ⁵¹ Yet, in some jurisdictions, such products flow, freely or under some restrictions, to the market. ⁵² Naturally, with globalisation, the question of whether prison products and services may be available in international markets is also relevant: it may set production volumes and necessitates compliance with a complex net of national and international regulations. Generally, the lower the barrier set for prison products and services to flow into the market, and the larger the volume of production dedicated to the market, the more robust the operation's exposure to the pressures exerted by the private market – and thus the stronger the pull of the private side of the continuum.

We should note already here that it follows that, according to this dimension, the effective and widespread implementation of the EU Proposal certainly pushes prison labour schemes toward the public side of the Publicness continuum for it aims, as noted, to bar forced labour products from the market. Indeed, it does not mean that excluded schemes automatically become ‘public’ (and thus, paradoxically, able to enter the market as non-forced labour). The market is only one among six dimensions. Excluded products and services may still be engulfed in private dimensions, as we further illustrate below. But the point remains: viewed from the prism of the HMD model, the EU Proposal directs prison labour toward public production.

Next is (5) the ‘control’ dimension, holistically looking into public-political and private-market pressures exerted on the operation and moulding it. It proceeds on the understanding that public correctional institutions are restricted not only by political demands and public regulations, but also by competitive constraints. It endeavours to differentiate between the types of pressures and norms that actually steer prison labour schemes. ⁵³ We should not look far for the fingerprints of (private) competitive pressures in prison labour, as they usually impact remuneration policies. True, in this context, too, de facto control may be mixed. The Israeli Prison Service (IPS) provides a telling example: it has recently contended in court, in an effort to justify its undergraded wage policies, that they should be judged in comparison to the wages offered by its business competitors in the Occupied Territories and China. ⁵⁴ Consider further another IPS episode illustrating the influence private entities may exert on public-run operations: in 2018 the IPS considered wage increases for incarcerated workers, and sought input from private entities employing prisoners within prison facilities regarding the magnitude of the proposed increase. Based on these consultations, the IPS determined the specific remuneration hike (7%), as proposed by the private entities. The rationale provided was that higher rates would purportedly hinder the sustainability of these operations within the prison facilities. ⁵⁵

Likewise, while the norms governing the setup and operation of prison labour seem to belong squarely in the public side of the ledger, the reality on the ground in this context, too, may defy easy classification. Take, for example, the standards of accountability and transparency. Normally, they [the standards] push in opposite directions when applied to public as opposed to private bodies, ⁵⁶ which raises a host of questions in our context: how should they apply to hybrid prison labour? Should private contractors be exempted from divulging information regarding their joint ventures with prison agencies? Could the latter agencies rely on their private counterpart's interests to deny information they would otherwise be obligated to disclose? ⁵⁷ Essentially: should such hybrid employment arrangements be considered public and fall within the private end of the continuum?

(6) Control over the production process, our last dimension, requires in-depth information on the nuts and bolts of the production process. While the previous dimension of control relates to the external, general legal and economic framework governing the operation, this dimension takes an internal perspective, examining the realities on the ground, on the assembly line. The importance of the latter dimension lies, inter alia, in the fact that even in publicly owned factories, such control may be shared with private entities. This is the case, for example, when prisoners work for privately-owned factories in public prisons, regardless of the supervision that prison guards may provide. ⁵⁸ Control over the production process may also come into question whenever publicly owned factories are sub-contractors of private entities. In these situations, the relevant private entity, even when physically removed, may substantially impact the production process. ⁵⁹ This dimension underscores, for example, the significance of examining whether prisoners receive compensation based on an hourly wage or work under a piecework system often favoured by private entities. ⁶⁰ In that and related scenarios, private organs dictate work structures, product design and production methods, as well as working environment and conditions. Indeed, it is crucial to recognise that even public prison factories and operations often retain only limited control over the production process under quite standard arrangements of cooperation with private contractors.

Taken together, the HMD framework strongly suggests that, globally, in the contemporary business of prison labour, a clear-cut division between the public and private based solely on ownership dimension, while conceptually feasible, is no more than a phantom. Any legal standpoint neglecting these perspectives embraced by international and regional institutions, such as the ILO and the EU, might be regarded as phantom pains.

Next, we will illustrate what insights our HMD model provides into contemporary practices of prison labour. Following the EU's concern around abuses of supply chains, we will illustrate, first, how the ILO's approach – soon to be the EU's – does allow for the incorporation of prison labour in supply chains. Second, as promised, we will briefly point to a surprising area of agreement that the bitter disagreement has remarkably produced – it revolves around the methodology of the inquiry, rather than its substantive, normative conclusion. As such, it may offer a possible way out of the deadlock in which the States and the ILO are stuck, especially if they turn to the assistance of HMD.

B. Where does HMD take us?

Back to supply chains

The ILO advances a somewhat misguided, and even anachronistic, view that disregards contemporary practices of inmates’ production, especially in the context of supply chains. In order to realise this, consider, for example, a state-run prison production facility contracted to produce an item to be integrated into a private company's global supply chain. ⁶¹ As we have seen, the ILO satisfies itself with making sure that the work is conducted under the supervision and control of a public authority, and that it does not involve hiring out or placing prisoners at the disposal of private companies. Both conditions are met in our example, and according to the ILO's approach the operation falls on the public side of the divide and is worthy of State exemption. Hence, it is not ‘forced labour’ at all, regardless of the labour conditions and even if – as is often the case – labouring prisoners are paid well below a reasonable standard. Moreover, the item referenced in our hypothetical example, produced in a prison under sub-market conditions, can circulate freely in global markets without restrictions.

Therefore, the current ILO approach allows for prisoners to be incorporated into a supply chain (again, say, when they work in a public prison, which has taken it upon itself to be a subcontractor of a supply chain), a conclusion that undermines the Convention's very raison d’être. ⁶² The Convention's regulation of prison labour, particularly the emphasis on limiting its private use, stemmed from a dual rationale. Historically, a primary motivation for the Convention was to address concerns about the unfair competition that could arise between forced prisoner labor and free-market workers. ⁶³ Complementing this unfair competition rationale, another crucial impetus behind the Convention was the acknowledged imperative to protect labouring prisoners from exploitation and guarantee their well-being. The current ILO approach fails to fully consider both of these rationales in the context of supply chains. Equally significantly, it contradicts a primary objective of the EU Proposal itself, which is to address the eradication of forced labour within supply chains. ⁶⁴ Whether the EU's motives lean towards protectionism or a focus on worker welfare, ⁶⁵ it is now clear that endorsing the existing ILO-based definition of forced labour fails to fully realise both objectives.

Encouragingly, HMD may readily expose this blind spot in the ILO's approach. How does HMD's multidimensional approach bring to the surface the fact that the Convention legitimates prison labour integrated into a private supply chain? Mainly, in this context, via the market dimension, deliberated upon and contested in ILO proceedings, ⁶⁶ a testament to the pertinence of this dimension to the discussion. Omitting any reference to the market dimension legitimates hybrid situations where prisoners toil under state supervision and even within state-owned prisons but produce products that become part of a (private) market supply chain. Yet, the ILO has explicitly repudiated the market dimension by allowing public authorities to compel prisoners to work in circumstances where the products ‘may be sold to the market by the public authority.’ ⁶⁷ But the rationale of the market dimension must be fully engaged to be taken seriously. While it may be true that the mere act of a public entity selling products to the market does not necessarily erode its publicness, experience shows that in the context of prison labour, practices such as this can create a growing dependence on market conditions, particularly if correctional entities manage their budgets as closed units. Such reliance could potentially impact the conditions of labour for the working prisoners in a significant manner. ⁶⁸

Seeds of agreement

Where do we find seeds of agreement among the ILO-State infighting? In a word, in their methodology of inquiry, which contains the seeds of a HMD-bound, realistic examination of prison labour schemes.

We should first note that the very language of the Convention appears to endorse several dimensions of the HMD model and be greatly attentive to the complex realities of prison labour. For example, the provision that is critical of ‘placing [prisoners] at the disposal of’ non-public entities clearly relates to legal custody, HMD dimension 2. Its severance from other inmate work dimensions suggests Convention No. 29's framework goes beyond the formal question of ownership of the prison facilities, thus transcending a simplistic public-private binary.

Similarly, per the Convention's provisions, the exemption from forced labour is contingent upon the work being performed under the ‘supervision and control’ of a public authority, regardless of the formalistic categorisation of the institution providing the work or where its formal ownership lies. It can thus be interpreted as acknowledging the importance of control over the production process (HDM dimension 6) in judging the credentials of a given prison labour scheme, irrespective of ownership or any comparable formal criteria. ⁶⁹

In addition, in assessing recent ILO-State exchanges, it should be noted that Australia, Germany and Austria also urge the ILO to look beyond the formal prism of ownership of the operation. What they argue for is the separation of the dimension of control and supervision over work processes from the dimension of formal classification of the employing entity as either public or private. They, too, insist on transcending formalistic ownership in the regulation and evaluation of prison labour; they, too, gravitate toward, if not directly invoke, HMD dimensions.

Other controversies provide a similar impression, including, for example, the debate over the meaning of the term ‘hired to [private individuals, companies or associations]’: ⁷⁰ does it equally denounce situations where the private entity pays the state for the prisoners’ services as well as situations where the state pays the private entity for operating a private prison? ⁷¹ The latter situation, several States argue, should be considered less detrimental to the well-being of prisoners than the former. ⁷² Again, States seek to highlight considerations beyond the formal classification of the job provider, and their argument surely addresses the funding side of the operation (HMD dimension 3).

In sum, both States and the ILO seem to appreciate the importance of scrutinising the subtleties of prison labour practices, such as the actual supervision and control of the detainees, alongside the formal proprietorship of the prison or the business that employs them. In other words, neither the ILO nor the States insist on a narrow and rigid interpretation of the Convention, as both seem to agree that formalistic doctrines of ownership cannot capture and be the decisive guide to regulating modern prison labour. It follows that they both, in fact, endorse a realist methodology in evaluating schemes of prison labour. ⁷³ Lastly, and more specifically, they both appreciate considerations other than ownership in their evaluations of prison labour operations, considerations which correspond nicely to the HMD framework.

The framework of the HDM can offer structure and give coherence to this ongoing and contentious debate. It reveals that, at the end of the day, the crux of the debate may not revolve around formal definitions or a definitive understanding of public and private realms. HMD therefore has the potential to provide a much-needed roadmap for the complex terrain and challenges involved, especially when the parameters of State exemption are hotly debated.

4. Conclusion

Today's growing currency of the ILO forced labour approach presents us with an auspicious opportunity to revisit its handling of the Convention's definition of ‘forced labour’ and the disagreement it stirred in the context of prison labour. ⁷⁴ It should certainly serve as an invitation to reengage with some of its sticky challenges. First among them is the division the ILO draws between private work providers and public occupation of inmates: while the former are generally frawned upon and associated with arduous conditions, the latter are regarded salutary by definition and exempted from any comparable conditions. In a previous study, we focused on the public side of the divide, making the argument that the ILO's dichotomous view disregards the dangers posed to labouring prisoners by the state, especially now that the state and the private sectors are barely distinguishable from one another in doing business. ⁷⁵ Therefore, we have argued, the state may infringe upon the rights of prisoners just like – or even more than – a private contractor. Having introduced HMD, we turned, in this article, to the other side, and focused on the ILO's understanding of the private sector's potential involvement in the business of prison labour. We argued that, despite the ILO's explicit aim to shield prisoner workers from the private sector, its approach to the Convention falls short of achieving this objective, particularly in the context of supply chains. This is a major blind spot in its approach, given the ubiquity of supply chains in the current, global economy and the concerns they raise, for example, for the EU. Yet, the EU is about to sign up to that same ILO approach, if it adopts the EU Proposal discussed here. At a minimum, then, the EU should reconsider its position on the definitional, methodological and substantive controversies highlighted in this article.