Abstract
International law provides that work by sentenced prisoners ought to prepare them for life after release. Yet reliable data on the nature, extent and impacts of prison work is lacking in much of the world, and what there is suggests that much prison work is far from voluntary, is unpaid or poorly paid, is unskilled or low-skilled, and does little to improve employment prospects on release. This article compares how prison work is defined and governed in law and shaped by policy, and how it is organised in practice. It describes the operationalisation of key legal provisions relating to prison work in three countries, and compares the purposes for prison work defined in law, the protections available to working prisoners, and the incentives or penalties used to secure participation. It also summarises the limited data available showing how many prisoners perform which kinds of work in each country. Finally, it argues that efforts to bring prison work provision into closer compliance with international standards should be sensitive to local variation, and suggests clarifying questions which could be used in such a reform process.
Keywords
Introduction
‘Work’ in prison is contradictory. Penal policy presents it as a means for ‘reform’ or ‘rehabilitation’, or for safer, better-functioning prisons; 1 by contrast, journalists and activists often represent it as the exploitation of prisoners for others’ benefit. 2 In both international and national law, it is separated from work in the wider economy, and ordinary labour rights do not apply.
Although work features in imprisonment everywhere, the forms of work prisons require prisoners to engage in vary substantially. So do the outcomes prisons use work to produce, and the incentives and penalties they use to secure compliance. Against this local variability, the applicable normative/human rights frameworks can seem vague in their wording, and difficult to operationalise. This raises a question: how can the standards they create best be interpreted and applied at the national level, since identical reforms in different penal contexts could generate unintended and undesirable outcomes?
In this article, we illustrate this argument comparatively, noting similarities and differences in the legal frameworks and practical realities of prison work in three countries covered by an ongoing research project. The article has four sections. First, we briefly describe the ongoing research on which we base our claims. Second, we describe key international legal provisions on prison work, and summarise evidence on its effects in practice; we suggest particularly that work capable of producing meaningful rehabilitative impacts represents only a tiny proportion of what is actually available to prisoners. Subsequent sections build from this. Third, we compare the three countries covered by the project—Brazil, the UK and the USA—comparing whether and how the law defines purposes for prison work, what protections and working conditions apply, what incentives or penalties are used to secure participation, and what is known about how many prisoners perform which kinds of work in each country. Finally, we draw together evidence from the preceding sections in a discursive conclusion, and suggest that efforts to improve the provision of work opportunities in prisons should proceed with care, paying careful attention to national circumstances and to specified questions.
The project
This article draws on research findings from an international comparative project, ‘Unlocking Potential: towards effective, sustainable, and ethical provision of work opportunities for prisoners and prison leavers’. The project examines prison work in law and practice in the UK, Brazil and the USA. As the latter are federal countries with considerable local variation, our research focuses on three contrasting states in each, along with the nations of the UK. 3 The project investigates how norms, values, and policies surrounding prison work are understood by practitioners, policymakers, prison monitors, partner organisations that provide work in prisons, and prisoners themselves; and to identify examples of good practice and effective collaboration.
In the project and this article, we use the following definitions:
The project thus far has consisted of a literature review and research on international and national law and policy on prison work. We have also conducted over 50 semi-structured expert interviews, with individuals who have knowledge of prison work in the UK, the USA and Brazil. 4 They include prison staff, government officials, policymakers, provider organisations, academic researchers, former prisoners, and prison standards monitors. These interviews have refined our emerging understanding of the topic. Where they are referred to, we have indicated this using a footnote attributing claims to ‘interviews with expert informants’. We also aim, via forthcoming fieldwork, to produce case studies of prison work as implemented in different prison facilities in each country.
Prison work in the law and in criminological research
Legal separation
Generally, work in prison is legally differentiated from work in the wider economy. This is largely a result of reforms which curtailed the exploitation of unfree labour, by public authorities and private capital. Before these reforms, exploitation took varied forms: both the British and Portuguese empires used forced convict labour in their colonies, 5 while in the US, prison labour was integral to capitalist industrialisation in Northern states, and segregationist oppression in Southern states. 6
Dating from the early-to-mid 20th century—the heyday of the ‘rehabilitative ideal’ and of welfarist policy orientations in punishment 7 —the International Labour Organisation's (ILO's) Forced Labour Conventions sought to eliminate more egregious forms of prison labour exploitation, while retaining the basic norm that states could direct prisoners to work, and penalise refusals. 8 Countries which ratify the ILO conventions may compel prisoners to work, provided the work is ‘directed and controlled by a public authority’ and prisoners are not ‘hired to or placed at the disposal of’ non-state entities. 9 The effect is that if prisoners work for any entity, besides the state, their labour can only be used if they have volunteered themselves for it, and if the terms and conditions (including pay) ‘approximate’ those available to free workers. 10 ILO rules are thus predicated on a sharp distinction between labour done for public authorities—which can be compulsory and is implicitly part of the punishment—and work done for non-public entities—which is voluntary and closer to ‘ordinary’ work. It is not clear that this sharp distinction still holds, given the interpenetration of the private and public sectors in the years since the ILO conventions were concluded. 11 This blurring underpins some of the arguments which follow, since it suggests that the risk of exploitation (which applicable normative standards sought to eliminate) has evolved over time.
Prison-specific human rights standards add extra layers to the ILO standards, further blurring the sharp distinction described above. 12 They require work not to add to the suffering of imprisonment, and set the overarching objective that it should, whenever possible, prepare prisoners for life after release. They state that prisoners should be able to choose work from what is available, and, where possible, gain vocational skills usable after release. They permit profit-making prison industries, but require prisoners’ interests not to be subordinated to the profit motive. They require conformity between the working hours and health and safety protections for working prisoners and those for free workers. Finally, they stipulate that prison work must be ‘equitably remunerated’, 13 albeit that prisons can make specified deductions from prisoners’ wages for some specified reasons. These standards, by contrast to the ILO ones, contain the expectation that prison work generally should be more voluntary, more ‘normalised’, and more reintegrative in its aims, whether it is directed by the state or another party. At the same time, they also preserve the norm that it can be compelled.
The international standards create ambiguity. Although work in prison is legally separate from work in the wider economy, may be coerced, and is exempt from key workplace protections, prison-specific rights standards emphasise voluntarism, and embrace aims such as resocialisation or preparation for release. They implicitly challenge the separation, by requiring prisons both to lower, and to preserve, the boundary between ‘prison work’ and ‘ordinary work’. This begs the question, which we address in Section 3 below: how do they balance these objectives in practice?
First, however, we briefly summarise research evidence on the effects of work in prison.
The effects of work in prison
Policymakers commonly claim that prisoners who work in prison will be less likely to reoffend after release. However, empirical evidence on this claim is inconclusive. It is hampered by several factors: by the fact that most studies were conducted in the USA (limiting the generalisability of the findings), 14 by methodological risks, and by the fact that prison work takes many forms, not all of them equally skilled or rehabilitative. Allowing for these limitations, some studies using quasi-experimental or meta-analytical designs to evaluate specific prison employment and employability programmes have found modest improvements in reoffending rates among participants who are followed up after release. 15 On the other hand, similar studies with similar designs are also sometimes inconclusive, 16 and a few even indicate increased risks of reoffending. 17
Regardless of these differences, there remain good reasons to promote ex-prisoners’ reintegration with employment markets after release, for reasons which are independent of recidivism. Although the direction of causation is not clear, there is still a clear relationship between employment and desistance from crime. Perhaps most importantly, though, positive outcomes are not always sustained over long periods: 18 the transition from prison to post-release employment is challenging. Ex-prisoners were often deeply excluded from labour markets before imprisonment, through poverty, educational shortfalls, and other factors, 19 and even if they gain qualifications and skills in prison, many face continued exclusion post-release. 20 Job quality also influences long-term outcomes: people who have participated in work programmes in prison are more likely to seek work after release, 21 but whether they sustain work is less certain. Low-paid, precarious, or unpredictable jobs are more associated with rearrest and return to prison, and may even produce worse outcomes than unemployment. 22 This is not simply a question of job permanence, nor of pay—even sporadic low-waged employment can protect against recidivism—but establishing any form of work record amidst destitution after release can be extremely difficult. 23 It also appears that ex-prisoners are pushed into low-quality work, both by financial insecurity and parole conditions which mandate employment. 24
As a result, the effects of prison work are nuanced. Prison authorities are not accountable for post-release outcomes and have little influence over post-release conditions, and in any case are hard-pushed to respond to individual need within prisons. Much is left to prisoners’ attitudes and motivations towards the work on offer. Ethnographic research has noted that prisoners will often seek out better-paid work in prison because it alleviates the immediate difficulties of prison life, but not necessarily because they seek, or expect to derive, longer-term benefits. 25 Some ethnographers have even concluded that unskilled but better-paid prison labour oncommercial contracts communicates to prisoners the norm that exploitation secures profit, thus reinforcing antisocial attitudes. 26 Generally, too, the low pay on offer can lead prisoners to conclude that even work which aims to offer skills and employability is simply punitive. 27
The variability evident in the literature raises two questions which guide the following sections. First, what does prison work actually look like, in the three countries we examine? What kinds of work are done, by how many people, and who directs it? Second, what incentives, rewards, and sanctions are used to secure compliance, and what does this tell us about the degree of ‘voluntariness’ or compulsion involved?
Prison work in the United Kingdom, the United States and Brazil
This section describes the legal basis, policy focus, and practical application of prison work in the three countries we examine in the project.
The United Kingdom
Legal basis
Prison regimes and progression
The UK 28 stands out from comparable European countries, both in its high imprisonment rates and its use of long, indeterminate, and extended sentences. 29 Although it is committed to international standards which emphasise preparation for release as the normative purpose of the sentence, 30 the objectives of sentencing law are somewhat contradictory, 31 and ‘preparation for release’ is typically approached as a question for the final stages of a sentence. There is evidence that people serving long sentences deliberately focus their attention on goals within the institution, and consciously avoid looking to the post-release future. 32
Key legal provisions
Having ratified both ILO Forced Labour Conventions, 33 the UK can compel sentenced prisoners to work, provided the work is ‘supervised and controlled [by] a public authority’ and prisoners are not ‘hired to or placed at the disposal of private individuals, companies, or associations’. 34 Domestically, prison law derives from the respective Prison Acts for each jurisdiction, 35 and from secondary legislation (particularly the Prison Rules) passed using powers they create. The Rules say relatively little specifically about the purposes of prison work. Those for England and Wales and Northern Ireland require only that it be ‘useful’; 36 in Scotland, work (along with other ‘purposeful activity’) should aim to improve the prisoner's ‘morale, attitude, and self-respect’ and ‘prospects for […] successful resettlement’ after release. 37
Pay, sanctions, and incentives
Only the Scottish Rules entitle prisoners to payment for work, 38 but in practice, prison work is paid throughout the UK. In England and Wales, the minimum rate set by policy is £4 per week, or £2.50 for prisoners willing to work but for whom prisons cannot find work. 39 In Scotland, the minimum is £5 per week according to a 2012 policy; 40 the Northern Ireland Prison Service does not publish rates. Prison work is therefore paid well below the national minimum wage of £11.44 per hour for workers aged 21 and over, 41 with one exception: work done by prisoners on day release in England and Wales. In this case, employers pay at least the minimum wage to the Prison Service; the law provides that prison governors may deduct the net (i.e., post-tax) wage by up to 40%. Though the amount and purpose of these deductions are constrained by statute, they are discretionary and are not always made. 42
Wages are the only direct incentive for working: there is no earned remission of the sentence. Wage rates are responsive not only to merit, skill, or aptitude, but also compliance: behavioural incentives schemes include opportunities to earn higher pay, and the best-paid work is available preferentially to prisoners with higher earned privilege classifications. 43 Participation is therefore secured as much via incentives as via sanctions because work is the only way to pay for basic goods or phone credit without financial support from outside. In law, however, work remains compulsory: 44 prisoners can be penalised for not working, unless they are held pre-trial, past the statutory retirement age, or are medically certified as unfit. 45
Working conditions
Courts have repeatedly upheld the principle that prisoners are not employees and hence are not protected by employment law. 46 The Prison Rules and associated policies nevertheless mimic some basic protections for working conditions, for example by limiting working hours. Most other regulations relate to occupational health and safety, in which the Prison Service follows a policy of observing all relevant legal provisions. 47 However, this is a matter of policy: prisoners, who are not employees, do not have rights as such under the Health and Safety at Work Act 1974. This means that litigation for injuries at work can only be pursued on the basis that the prison owes prisoners a general duty of care, and must demonstrate the prison's negligence. 48
In some prisons, prison labour is used on contracts from public, private, and voluntary sector clients. Wages are supplemented by the customer, usually exceeding the ordinary prison wage but still falling far below the national minimum. The ILO has long criticised the UK government in relation to prison labour in contracts workshops, and all prison work in privately-operated prisons. 49 Its position is that if this work is done compulsorily and on terms and conditions not ‘approximating’ those available to free workers, 50 then it is illegal forced prison labour; the UK government demurs, arguing that this interpretation constrains innovation and prevents training opportunities being offered to prisoners via prison workshops, 51 and that contracted-out prisons should not be treated as private contractors for the purposes of ILO convention law.
In sum, UK law on prison work is highly pragmatic, declining to set clear aims and leaving much to policy and local discretion. The result is a complex patchwork of provision, varied in relation to prison function and to local management priorities.
Prison work in practice
Official data publications on prison work are limited, and do not facilitate a clear overview. Published figures describing prison work in Northern Ireland and Scotland are limited to incomplete snapshots via freedom of information (FOI) requests.
Data published for England and Wales is incomplete, discontinuous, and describes only prison labour (as we define it in section 1 above). The publication of data ceased in 2020 and has not resumed. 52 From a starting point of 10,500 in 2014/15 (12.2% of the England and Wales prison population in that year), the number grew to 12,500 in 2019/20 (15.5%). The average working week, calculated over this entire period, was 27 hours. Sales figures show that, between 2018 and 2021, over three quarters of contracts by value covered ‘general assembly and packing’, ‘recycling’, or ‘laundry’ work. 53 Contracts work was unevenly distributed, concentrating mostly in men's prisons holding populations close to release, and being largely absent from women's and remand prisons. Contracts work varies in terms of the skill and hours involved, but a substantial part—particularly the 46% of contracts by value which are marked as ‘general assembly and packing’ 54 —is unskilled and would probably (if paid at minimum wage) be uneconomical. 55 Generally, prison labour is paid better than other prison work, though governors sometimes steeply incentivise necessary but unpopular work performed on a small scale (such as biohazard cleaning). The highest wage described by our informants for England and Wales was £35 per week, for piecework in a contracts workshop.
England and Wales also publishes data describing the work done by prisoners on day release. An average of 305 performed this form of work in 2011/12—around 0.4% of the then prison population. This proportion remained roughly consistent until around 2018/19, before growing since then, reaching 1,088 or 1.2% of the prison population in 2022/23. Work on day release receives at least the national minimum wage, which is paid to the prison and then subject to deductions as described in section 3.1.1 above.
That data is published only on prison labour (as we define it in section 1) is puzzling, since most prison work involves ‘domestic’ or ‘prison housework’ tasks relating to prison functioning (such as cleaning and catering). A government strategy in 2018 estimated that over 13,000 prisoners performed such tasks in 2016, against the 11,000 active in prison industries at the same time. 56 The government told Parliament in 2023 that it did not hold data centrally to compile equivalent figures to these, nor to describe the wages paid for different kinds of prison work. 57 Wages appear to vary widely, according to local discretion: 58 a 2018 Freedom of Information request showed (for example) that wing cleaners in eight prisons earned wages ranging between £5.40 and £13.50 per week, while kitchen workers received between £7 and £21.84 per week. 59 Recent policies have sought to make ‘prison housework’ improve prisoners’ longer-term employability by offering accredited qualifications for prisoners performing tasks of this kind, for example in prison kitchens. 60 However, these initiatives are delivered by external providers using external funding, 61 and not always as core provision via prison budgets. Recent policy has also focused prominently on increasing the number of ex-prisoners in work six weeks after release. This has led to investment in employment hubs in resettlement prisons, 62 and funding to foster links between them and local employers. 63
Scotland and Northern Ireland have followed similar paths, and there are prison workshops doing contracts work in both countries. 64 Some prisoners in Scotland and Northern Ireland do leave prison to work on day release, but they are not paid the minimum wage (with deductions) because relevant legislation has not been implemented.
Overall, the UK picture is patchy, appears locally variable, lacks detail, and recent that prison overcrowding risks compromising earlier policy progress. 65 Though figures are incomplete and difficult to obtain, it appears that substantial numbers of UK prisoners—perhaps as many as half—do not work. Wages are well below those in the outside world, but appear to incentivise participation effectively, so that penalties are not generally used. 66 Prison labour is typically paid more, and concentrates in lower-security ‘resettlement prisons’. Elsewhere, prison work is used to support institutional functioning and prison order, with ‘preparation for release’ left until late in the sentence.
The United States
Legal basis
Prison regimes and progression
The US is known for ‘mass’ imprisonment, but state prison populations have been declining since around 2010, while remaining high by global standards. They have also followed a process of ‘bifurcation’, 67 whereby places in state prisons are increasingly reserved for people serving long (often indeterminate) sentences for violent offences. 68 Non-violent offences are handled increasingly at the local level, sometimes via non-carceral alternatives. 69 In many states, the state prison population (which our work focuses on) increasingly comprises people in the first group. 70 This is important for prison work, because it means prisoners are often a) incarcerated for extremely long, and often indefinite, periods; b) released conditionally if at all, and often when elderly; and c) under parole conditions which may threaten recall to prison for those who do not find or sustain work. This holds ex-prisoners near the bottom of the employment market. 71
Key legal provisions
The US has ratified only one of the ILO's Forced Labour Conventions. 72 Thus it has committed to abolish forced or compulsory labour only if it is imposed as punishment in five specified circumstances, 73 not in any circumstances. US constitutional and domestic law carries more weight in this regard. The Thirteenth Amendment to the US Constitution abolished slavery and involuntary servitude, except as punishment for a crime. Because of the exception, prison administrators—barring in states which have abolished it in their state constitutions—can implement compulsory prison work at their discretion. Work is therefore mandatory in most American prisons, and the failure to work if directed to do so is a disciplinary offence which can result in sanctions, including the loss of property or earned privileges, confinement to cells, and the loss of contact with family. 74
Federal law regulating the purposes and aims of prison work is limited, so any guidance derives from the purposes of punishment itself. The Eighth Amendment to the US Constitution requires that punishment not be ‘cruel and unusual’, but this does not entail clear positive aims. Sentencing law is markedly retributive, and likewise does not specify aims for imprisonment or for prison work. In Arizona, the law aims only to allocate punishment proportionately to the seriousness of the offence; 75 California and Texas both list multiple aims but do not indicate how work in prison is to contribute. 76 In any case, 90% of sentences in the US are set not by courts, but via plea bargains; 77 rehabilitative aims (including for work) are left to correctional authorities to work out. State law tends to present work as an end, not a means to an end. 78
Although federal employment law does not specifically exempt prison work from its protections, federal courts have consistently declined to extend employment protections to imprisoned workers. 79 These protections include the right to a minimum wage and a 40-hour working week, 80 the right to associate and form trade unions, 81 the right to decline dangerous work, 82 the right to protection against occupational hazards, 83 and (in some parts of the US) to protection against discrimination in the allocation of work. 84
Forty-nine states operate state-owned prison industries corporations, which manufacture and supply goods only for consumption by public sector customers. 85 It is also common, in many states, for prisoners to be assigned work on public works projects outside prison perimeters, including firefighting, disaster response, groundskeeping, and janitorial work for state and municipal agencies. Private companies use prison labour on a small scale, and federal law has restricted the sale of prison-made goods across state borders since the 1930s. Reforms initiated in 1979 created a certification scheme whereby companies can use prison labour if they pay the relevant minimum wage, but this scheme operates on a tiny scale. 86 Since federal restrictions apply only to inter-state trade, some states, including Arizona, have opted instead to hire prisoner labour out to companies headquartered within the state. 87 The hiring of prison labour by state and municipal authorities is largely unrestricted, and is common in many states. 88
Pay, sanctions, and incentives
Prison work in seven states, all in the South, is almost or entirely unpaid. 89 In Texas, for instance, the only incentive for working comes via credits: reductions in sentence length which accrue for each day of work, at different rates depending on the prisoner's classification by ‘conduct, obedience and industry’. 90
In states where prison work is paid, prison wages fall well below prevailing minimum wages. Structured incentive schemes also apply. Arizona, for instance, classifies jobs as ‘unskilled’, ‘semi-skilled’ and ‘skilled’, determining wages according to a matrix which cross-references the job with the prisoner's classification in an incentive scheme. 91 Similar arrangements are specified by state law across the US; credits are typically offered at a rate proportional to each day of work completed, but accrue differentially according to incentive schemes. The full range of incentives is not available to all: for example, the 32% of sentenced prisoners in California who are serving life sentences cannot earn remission credits, 92 and exclusions based on sentence or offence type are common.
Meanwhile, the best-rewarded and most skilled work is restricted to select groups, with only some restrictions relating to aptitude or qualifications (for example, via literacy requirements). For example, Californian prisoners serving life-without-parole (LWOP) sentences and those more than five years away from an earliest release date are ineligible for state prison industries jobs. 93 Similarly, in Arizona, foreign nationals and people convicted of some violent or sexual offences are ineligible for better-rewarded prison work, 94 including work done outside prisons.
Working conditions
Federal and state law do not consistently apply the international human rights standards requiring working hours and health and safety protections not to be worse than those in the wider economy, 95 and in practice, protections are few. Accidents and injuries are unevenly reported; research by the American Civil Liberties Union (ACLU) documents examples of serious injuries sustained by working prisoners, based on media reports and litigation. 96 Given the many barriers to litigation, cases which reach the courts will represent the tip of the iceberg, but reliable overview figures are lacking. 97 Some states offer more protections against injury than others: California is among the states which allow prisoners to claim from workplace injury compensation schemes, whereas Texas explicitly excludes prisoners from all protections under its Labor Code. 98
Summing up, the law in US states generally requires prisoners to work in whatever programmes, roles, and conditions they are assigned to, unless prison authorities determine that they are unfit or unsuitable; 99 those who decline can face formal sanctions. Some states (including Colorado, Nebraska and Utah) do not require compulsory work, having abolished the slavery exception in their state constitutions, but they are the exception.
Prison work in practice
A 2022 estimate by the ACLU, based on a nationally representative prisoner survey, put the proportion of prisoners who performed some kind of work in prison in 2016 at approximately 65%. At 2021 imprisonment rates, this would mean around 791,500 people worked in custody. 100 Of these, around 80% performed tasks supporting the functioning of prisons—for example, janitorial work, cleaning, catering, prison maintenance, or otherwise helping prisons to function. The ACLU estimated that a further 8% performed public works assignments outside prison walls, for example, groundskeeping or janitorial work for states and municipalities. Some 6.5% worked in state-owned prison industries, around 2.0% in work-release programmes mandated by the sentence, 2.2% in agriculture, and just 0.6% in private sector prison industries. 101
The ACLU estimated that prison wages are lowest for work supporting prison functioning: a low average of $0.13 and a high average of $0.52 per hour across states. Wages in prison industries assignments were generally higher, averaging between $0.30 and $1.30 per hour. 102 Such wages, then, are far below the applicable federal or local minimum wages. 103 Most states also deduct them, for a range of reasons: fines, court-ordered restitution payments to victims, taxes, family support, and payments towards the cost of imprisonment. Typically, these deductions are graduated according to the original wage, with deductions at higher rates from higher wages. In most states, prison wages do not count towards unemployment or social security insurance schemes. 104
A recent analysis of data from earlier waves of the survey used by the ACLU, supplemented by interviews with former prisoners and prison administrators, adds important longitudinal detail to this picture, showing how work in state prisons has changed since the 1970s. 105 Most importantly, it suggests that, on average, prisoners now work less than ever before, with a greater proportion doing low-paid, low-status, and often short-hours ‘prison housework’. 106 Simultaneously, prison work has become more unequal: the best-rewarded and most skilled/demanding assignments now pay more, relatively, than in the past. The incentive structures described above offer an explanation: since more favoured assignments are scarce, and allocated based on compliance and satisfactory evaluations, then the work helps incentivise prison order. A steep incentive structure also offers a basis on which to ration more resource-intensive programmes, by allocating them preferentially based on compliance to those seen as most likely to benefit. 107
Two points should be noted in summary. First, prison work functions to subsidise state institutions: for example, where prisoners perform maintenance tasks and reduce operating costs, or produce goods and services (via prison industries) for public sector consumption, or allow states and municipalities to use cheap or free labour on public works projects. Second, and simultaneously, much of it is inefficient, economically unproductive, and results in underemployment, in ways that would not be possible if prisoners were paid a market wage. Put simply, US prison work offers instances whereby prisoners’ low-waged labour is used to generate private profit. But if anything, the more representative case involves states mobilising prison work to lower the costs of imprisonment. Strong positive aims which would emphasise benefits to prisoners (e.g., that work must pay a minimum wage, or prepare them for release) are lacking.
Brazil
Legal basis
Prison regimes and progression
Brazil's Penal Code stipulates that prison regimes be classified by security level and structured to serve the overarching aim of resocialisation. 108 Prisoners can progress through three separate regimes—closed, semi-open and open. Closed regime prisoners are held in maximum or medium security facilities. Semi-open regime prisoners are required to work or attend supervised training or education during the day, either on- or off-premises, but return to the prison at night. The open regime requires prisoners to attend courses, work, or engage in other authorised activities unsupervised, and to stay in hostels (casas do albergado) outside working hours.
Initial regime placement is prescribed by the sentencing court and depends partly on sentence length. Sentences of more than eight years must begin in the closed regime. Those of between four and eight years generally begin in the semi-open regime; and those of less than four years, in the open regime. Progression is decided by a penal execution judge, 109 based on criteria that vary according to the offence, previous convictions, and conduct in custody.
Undermining the effectiveness of these measures is Brazil's relentless prison population growth, leading to overcrowding and a chronic shortage of prison places in all three regimes. The open regime envisioned by federal law is now largely defunct: 110 there are very few hostels, and most prisoners are released instead into ‘residential custody’—a system of conditional freedom and (often) electronic monitoring, accounting for around 20% of the sentenced population. 111 Even progression from closed to semi-open regimes is sometimes substituted by release to ‘residential custody’, owing to the lack of space in semi-open facilities.
Key legal provisions
Brazil has ratified the ILO's Forced Labour Convention and Abolition of Forced Labour Convention. The federal constitution of 1988 outlawed forced labour as a punishment.
The aims of imprisonment in law are various, and somewhat contradictory. The Penal Code requires sentences to be proportionate to the seriousness of the offence, and to meet the aims of deterrence and retribution. But the federal Supreme Court has held that in addition to these statutory objectives of punishment, ‘its resocializing character cannot be forgotten, and the State should therefore be concerned with rehabilitation of the convict’. 112
The Penal Code structures regimes and progression, but most law on prison work derives from the federal Penal Execution Law (‘LEP’), 113 which requires every sentenced prisoner to work to the extent of their capacity and aptitude, and states that ‘work must have an educational and productive purpose’. Allocation of work, and its remuneration, are listed as specific rights of all prisoners alongside essentials such as food and communication with the outside world. Beyond these broad criteria, there is no statutory definition of prison work.
Pay, sanctions, and incentives
Working prisoners must be paid no less than three-quarters of the national statutory minimum wage, currently 1,212 reais per month. 114 Deductions can be made for reparation or restitution for victims; family support; and to reimburse state expenses in maintaining the prisoner. Prisoners may access what is left to meet personal expenses in prison, with any remainder transferred to them on release.
A further entitlement—and, given the uneven implementation of the minimum prison wage, perhaps the major incentive for working—is sentence remission. Prisoners in closed or semi-open regimes are entitled to remission of one day of their sentence for every three days worked. 115 This operates as an incentive for overcrowded prison systems to provide work to as many prisoners as possible. Refusal to work can result in sanctions, including being held in isolation or denied progression to a less secure regime. In practice, however, the scarcity of work and competition for assignments means this rarely happens.
Working conditions
The LEP excludes working prisoners from national employment legislation, but prison working conditions must be safe and hygienic. Prisoners are required to work for between six and eight hours per day, excluding Sundays and public holidays; but those engaged in maintaining prison establishments may have to work different hours.
Prisoners who work in the closed regime must either do so inside the establishment, or on public works projects outside it (e.g., building roads or public hospitals, in which case the work can be managed by private entities on state contracts).
Federal, state and municipal governments may contract with the private sector to establish workshops, including as necessary to maintain the prison itself. The provision of labour to any private entity requires the prisoner's express consent.
As with the regime progression laws, prison population growth has rendered many of the LEP's aims and protections inoperative. When it came into force, the total prison population stood at around 88,000. Today, it is around 840,000. 116 Investment and staff recruitment have not kept pace with numbers. As a result, containing unrest, violence and criminality, rather than providing work, is the first priority for many prisons—especially larger, higher-security facilities.
Prison work in practice
Numbers working
Increasing the number of prisoners working and receiving training has been a major federal policy focus since at least 2009, with the Começar de Novo (New Start) programme, which also aimed to incentivise private sector involvement. In 2012, the first significant federal programme, PROCAP, 117 funded state prison departments to establish prison production facilities in sectors including civil construction, food processing, technology, agriculture, and recycling. Further initiatives have followed: for example, the 2018 national prison work plan, which encourages states to give subsidies or tax incentives to private companies which hire prisoners or prison-leavers. 118 In 2021 the federal government committed to increasing the number of prisoners engaged in work by 185% by 2030. 119
All prisons are expected to record data on the numbers of prisoners working and on their remuneration; figures are aggregated by a federal body, SENAPPEN, which publishes them biannually. 120 We understand that inconsistencies may exist around whether prison maintenance jobs are counted as work—and that some states may deploy remission-for-work as a tool to control prison overcrowding. 121
In June 2023, the number of prisoners in physical custody reported as working was 154,531—around 24%. 122 Of these, around 80% worked inside prisons and 20% outside. SENAPPEN's data show considerable variation by state. Maranhão currently has top ranking with 64% of prisoners working, compared to 34% in Santa Catarina and under 20% in São Paulo. Population composition is a key factor: fewer prisoners work where (as in São Paulo) larger proportions are held pre-trial. Also significant is the type of establishment: higher rates of engagement are seen in ‘colônias agrícolas’ and ‘colônias industriais’ (agricultural or industrial colonies), facilities designed around production which are often semi-open, and do not exist in every state.
The types of work offered, and whether partnerships are formed mainly with public or private sector organisations, also varies regionally (as does pay: see section on remuneration, below). Unlike São Paulo and Santa Catarina in the more prosperous south, Maranhão is less industrialised, with very few companies using prison labour. Most prison work there is for the public sector: for example, producing school uniforms and desks for the state education department or hexagonal blocks to pave urban pedestrian areas. 123 Santa Catarina, by contrast, is highly industrialised, and private companies engage substantial numbers of imprisoned workers. This affords them political clout: when the state legislated in 2021 to require its commercial partners to pay energy and water costs at prison production sites, many abruptly withdrew from these partnerships in response, causing the state to rescind the law. 124
Remuneration
Official data reveal that many working prisoners receive either no pay, or less than the statutory minimum mandated by the LEP. Over 67,000 prisoners (or 43%) received only sentence remission for working in June 2023. Some 26,300 were paid an amount below the legal minimum, while just over 34,000 earned above this level. 125 Again, state-level variation is significant: in Maranhão 74% of working prisoners received no remuneration, only remission; in Santa Catarina and São Paulo the proportion was 10% and 12%, respectively.
Discussion
Comparison of the different countries described in this article reveals some striking contradictions and similarities. We now move to discussing these, in relation to three themes.
Purposes of prison work
Our first observation relates to the stated purposes of prison work in law. Brazil, alone among the countries we examine, specifically states a rehabilitative purpose for prison work. ‘Educative and productive purposes’ are prioritised and promoted ‘as a condition of social dignity’. 126 Also uniquely, national law includes the right to allocation of work among prisoners’ basic rights, and entitles working prisoners to receive a proportion of the national minimum wage, though this entitlement is not fully realised. UK law requires only that prison work be ‘useful’, and of a kind directed by the government; US state and federal law generally does not direct clear aims for prison work, leaving this instead to correctional agencies. US constitutional law offers normative cover for work imposed as punishment; and both the UK and the US organise some forms of work to prepare prisoners for release, tending to do so for pragmatic reasons and to offer these opportunities only to some prisoners.
Operational issues and incentives
A second comparative observation relates to operational and policy priorities, which powerfully shape the provision of work opportunities regardless of formal legal aims. In Brazil, chronic prison overcrowding has been a long-standing challenge, and managing prison space among the most pressing concerns for penal practitioners. In overcrowded, inactive closed-regime prisons, internal prison governance has been ceded to prison gangs or ‘factions’. In some states, such conditions have produced riots and massacres, for example in Maranhão in 2014. Even if wages are not paid (though prisoners are entitled to them), the entitlement to remission is both an effective incentive for prisoners, and an effective means by which to ease overcrowding by accelerating progression through the system. The development of a large-scale public sector-led public works programme via prison workshops can best be understood in this context.
Some US states remain committed to large-scale prison labour schemes organised ‘in-house’. Texas, for instance, operates large-scale agribusinesses and state prison industries which produce food, clothes, and other goods for consumption within prisons and the wider public sector. 127 These occupy prisoners and defray costs, but some lose money and have been criticised for their expense compared to procurement via the open market, 128 and would certainly cease to be viable if prisoners working in were paid. Revenue-boosting, meanwhile, was a major factor driving Arizona's reorientation of its prison industries programmes towards hiring prisoners’ labour to external parties, rather than operating production efforts ‘in-house’. 129
Other US states—and the UK—operate a more ‘mixed economy’, in which engaging prisoners in activity is a relatively high priority, but ensuring that this activity is economically productive work is less so. Low-risk prisoners are permitted to leave prison daily to work and to earn a ‘normal’ wage with deductions, and private sector contracts are taken on to raise revenue where this is practical and investment can be made, but many prisoners appear not to work, and inactivity, not exploitation, is the norm. It is telling that (for example, in both California and England and Wales) policymakers vaunt the vocational training and qualifications available to a select few prisoners, but publish less data on the large and indistinct mass of inactivity and ‘prison housework’ which appears to be the reality for rather more.
What is notable throughout is an inherent tension: between prioritising activities which occupy prisoners in prison, and those which prepare them for work after release; or put differently, a tension between in-prison and post-prison outcomes.
The role of outside organisations and markets
This leads to a third observation: much depends on local political economy, and particularly the role of market logics in prison work provision. In the US, where statistics permit the longest view on the issue, work opportunities as currently offered discipline prisoners into conformity with an unequal and highly exclusionary employment market: the behaviours incentivised relate to competitiveness, compliance with expectations and rewards for taking whatever employment is on offer, 130 and these market logics pervade both work provision within the public sector and work provision involving partnerships with outside organisations. Ethnographic research offers glimpses of the subcultural hierarchies which prisoners construct under these conditions, showing that they are based on the virtues of working hard within such a system. 131
The shortage of available data means we cannot trace the issue over the same periods of time in the other countries we examine. But while the current US position is not inconsistent with parts of the UK and Brazil, it differs from others. In England and Wales and (to a lesser extent) Scotland, prison labour partnerships with outside organisations are mobilised to steepen the incentives—training, qualifications or pay—which can be offered to those who have ‘earned’ these opportunities through compliance; but these partnerships are offered mainly in lower-security facilities and are patchily available by region. Similarly, private sector partnerships are evident in the highly industrialised, market-oriented Brazilian state of Santa Catarina. But they are largely absent in Maranhão, where the public sector is more economically important, and where large-scale and ostensibly more inclusionary state-building public works projects are prioritised. Thus, political economy powerfully shapes the kinds of work opportunity available.
Conclusions
History suggests that it is difficult to organise economically productive work in prisons while also sustaining parallel commitments to labour that is not forced, and meaningfully prepares prisoners for release. 132 There are inherent security and administrative challenges, and anything more than low-skilled work presupposes the political will to invest in addressing prisoners’ typically low educational attainment, and histories of social exclusion. Capital and human investment are prerequisites, and so are sources of revenue. Their scarcity, in a context of high imprisonment rates and often overcrowded, dilapidated prisons, means that preparation for release is most often a privilege, ‘earned’ by prisoners held in less restrictive forms of custody, where the aforementioned administrative and security challenges are less pronounced.
The involvement of external providers is a necessary but insufficient condition for meaningful preparation for release to be offered to more prisoners. However, their involvement is patchy, and many barriers exist. Serendipitous wider market conditions can provide incentives: in the UK, labour shortages relating to Brexit and the pandemic, and in Brazil, labour shortages in highly industrialised regions, have each produced this effect. But where these opportunities are absent or fleeting, incentives such as low wages or tax breaks (as in Brazil's policy framework) may be necessary to secure outside involvement, at least by the private sector.
For most prisoners in all three countries, wages for prison work are far below the minimum wage, regardless of who they work for. This appears exploitative, most strikingly where (as in some US states, or some parts of Brazil) no wage is paid. Sentence remission is often framed—including by human rights courts 133 —as an acceptable alternative incentive, and may be seen as such by prisoners. But it is not available to all, and in the UK is not available at all.
ILO protections against exploitation are unsatisfactory, 134 because their protections against exploitation are premised on a false binary between state and private use; the ILO's requirement that wages for private prison labour must ‘approximate’ a market wage is openly ignored by some states. International prison standards, which stipulate that work should prepare prisoners for release, require individualised provision and are difficult to operationalise given the scale at which prison systems currently operate. They set norms which are also hard to realise for those whose release is far distant. All of this is consistent with the picture described above, in which only a select few prisoners benefit from more ‘normalised’ or ‘market-like’ work opportunities, and then often not through the payment of a ‘real’ minimum wage.
Reform via major legislative change—for instance, a blanket restoration of the right to a minimum wage for all prison work—is not a straightforward remedy. Given the current scale of prison systems, if all prison labour were paid at the prevailing minimum wage, the resulting monetary cost would likely force prisons to reduce the numbers of prisoners in work, by finding ways to increase productivity among those who remained. This would harm prisoners’ short-term interests (and their longer-term ones), and would likely further increase the allocation of the ‘best’ prison jobs to those already most qualified to do them. Similarly, if all prisoners working in ‘prison housework’ jobs had to receive ‘fair pay’, even if below the country's minimum wage, prisons at their current scale would not operate without billions in state support.
Our analysis suggests that several questions on appropriate legal reforms must be tackled, with attention paid to custodial realities, as well as to the nationally and locally specific forms taken by exploitation through low pay. A principled debate—based on an understanding of the ‘adjacent possible’ 135 —should focus on several questions. How can current levels of involvement by external parties in providing work opportunities for prisoners be maintained or expanded? Connectedly, how can these opportunities be better tailored to meet the needs of more prisoner groups—not only those close to release, but also those held in longer-term custody? A third key question, given that human rights standards on imprisonment require ‘equitable remuneration’ without defining this, is how work by prisoners should be rewarded. Is sentence remission alone, and not also a wage, an acceptable reward for work? Should prisoners serving longer sentences be offered remission in lieu of wages, and higher quality or more skills-focused jobs, given that these are currently in short supply and usually offered to those nearer to release? Finally, given that prisoners face fewer living expenses than free workers, what link, if any, ought to exist between prisoners’ wages and rates of pay for similar work outside prison; and what (if anything) can justifiably be deducted?
Footnotes
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Porticus Foundation [grant number GR-075674].
