The ‘Great Decarceration’: Historical Trends and Future Possibilities

During the 19th Century, hundreds of thousands of people were caught up in what Foucault famously referred to as the &#8216;great confinement&#8217;, or &#8216;great incarceration&#8217;, spanning reformatories, prisons, asylums, and more. Levels of institutional incarceration increased dramatically across many parts of Europe and the wider world through the expansion of provision for those defined as socially marginal, deviant, or destitute. While this trend has been the focus of many historical studies, much less attention has been paid to the dynamics of &#8216;the great decarceration&#8217; that followed for much of the early&#8208; to mid&#8208;20th Century. This article opens with an overview of these early decarceration trends in the English adult and youth justice systems and suggests why these came to an end from the 1940s onwards. It then explores parallels with marked decarceration trends today, notably in youth justice, and suggests how these might be expedited, extended, and protected.


Trends in Adult Incarceration and Decarceration, 1850s to 1940s
The English and Welsh prison population grew rapidly from its modern formation in the 1850s, reaching an average daily peak of 25,074 in 1879 (see Figure 1). However, thereafter, the rate of growth slowed dramatically. With the exception of a brief further rise in the early 1900s, it followed a broad downward trend from the 1880s onwards which lasted for over six decades. At the start of the First World War, the average daily population was just 10,326 -the lowest it has been since the prison estate was formed in the 1850s -and by the end of the Second World War the prison population was still under 15,000.
How can we explain the significant decline in prisoner numbers between the 1870s and the 1940s? Was it the product of a slow but sure 'great decarceration'? If so, was this decarceration by design or by default? An overview of key trends suggests that this decarceral trend resulted from the combined impact of four disparate factors: the operationalising of positive penological approaches; the increasing use of community disposals; the decline in prosecutions for certain kinds of offences; and efforts to cut public spending on prisons.
Following the end of convict transportation to Australia in the 1850s, a new prison estate was created to deal with those who could no longer serve their penal servitude overseas. The local prison system continued to be filled with more minor offenders as it had been decades before the establishment of the new convict prison system. The new punitive prison regimes brought in from the mid-1860s were widely viewed as having failed just two decades later. From the 1880s, penal reform focused on a duel track approach -an increased level of surveillance and long sentences for habitual and serious offenders, and an increasing extensive and innovative system of divergence for offenders who were deemed deserving of, and able to benefit from, training and reform. A raft of semi-carceral institutions grew up to house the latter, including reformatories and industrial schools, asylums, rescue homes, inebriates institutions. This operationalising of a new positive penology which sought to recalibrate punishment according to the characteristics of the offender, as well as the seriousness of their offence, was embodied in the report of the Departmental Committee on Prisons (1895) (often referred to as the 'Gladstone Committee'). The report recast incarceration as a form of rehabilitative individualised training and, in so doing, facilitated a significant shift in the purpose and practice of imprisonment and in wider historical cultures of control even if the conditions of imprisonment changed little (Bailey 1997(Bailey , 2019Garland 2002;Johnston 2015). As Wiener (1990) argues: 'the punitive impulse had by no means evaporated, it was being ever more attenuated ' (p.379). This attenuation, arguably, contributed to marked decarceral trends in the early 20th Century.
A second factor contributing to the fall in prison numbers was a growing acceptance of community disposals (Mair and Burke 2013;Vanstone 2017;Whitehead and Statham 2006). In 1876, the Church of England Temperance Society established the London Police Courts Mission which offered The Howard Journal Vol 59 No 3. September 2020ISSN 2059-1098  support for habitual drunkards in the capital's magistrates' courts. Three years later, the Summary Jurisdiction Act 1879 enabled the lower courts to suspend punishment by binding offenders over to come up for judgment as a later date (if they reoffended). The principle that people could be kept out of prison so long as they were properly supervised within the community was slowly becoming established. In 1886, the Probation of First Time Offenders Act allowed courts outside of London to appoint their own 'missionaries', establishing a national system formalised by the Probation of Offenders Act 1907 (Whitehead and Statham 2006). After a slow take-up by sentencers, probation began to be used extensively. For those convicted defendants who were punished with fines rather than probation, another legislative change helped to keep them out of prison; paying court fines in instalments was made possible through the Criminal Justice Administration Act 1914, meaning that thousands of people who could not immediately pay their fine were no longer sent to prison in default of payment. The Home Office statistician in charge of judicial statistics, W.J. Farrant, noted the 'unparalleled' decrease in custodial sentences in his Introduction to the 1919 Judicial Statistics (Farrant 1921). As a result of these combined reforms, custodial court disposals decreased dramatically from 81% of all sentences in 1856 to 45% by 1938. Financial penalties replaced custodial sentences for most summary offences over the same period, to the point where fines made up around 80% of all punishments imposed by the courts. A third factor contributing to this change was a broader shift in the prosecution process. As Figure 1 shows, the decline in prisoner numbers from the 1880s onwards cannot be explained in terms of a marked decline in recorded crime but it may very well, in part, be explained in terms of the fact that the courts increasingly opted to punish those found guilty of these crimes with non-custodial sentences. That said, there was a decline in certain kinds of prosecutions in the early 20th Century which also helped to keep prison numbers down. Tougher regulation of alcohol sales, especially during the First World War may have reduced the number of prosecutions for vagrancy and drunkenness; the increased visibility of uniformed police in public spaces may have deterred public order offending (Gatrell 1990); while the increased police control of the prosecution process itself may have resulted in reduced levels of prosecuted violence (Godfrey 2008). While noting the decline of prosecutions for assault and drunkenness, it is important to stress that prosecution rates for other volume crimes, such as theft, remained relatively stable from the 1880s to the 1940s (see Figure 1). This is important because it shows that early decarceral trends were not tied to an overall fall in prosecutions.
The final factor likely to have shaped decarceral trends was the perceived need to reduce public spending on prisons -a need that was articulated by government almost as soon as those prisons were established. From 1856, every penal institution had to report estate and staff costs to the Prison Commissioners (and ultimately, the Home Office), with cost per inmate worked out to the exact pound, shilling, and pence (see https: //www.researchcatalogue.esrc.ac.uk/grants/RES-062-23-3102/read (accessed 1 December 2019)). Following the nationalisation of the prison estate in 1877 (when the central government took over control of the prison system from local authorities) there was a constant drive to reduce costs, and to obtain better value for money. This was never more evident than during the fiscal crises of the 1920s following the First World War and the later Wall Street Crash. In the early 1920s, 18 prisons were closed (a significant proportion of the prison estate) and the dwindling number of prisoners crammed into the remaining overcrowded prisons. Cost saving was a clear driver here, although other factors remained in play. The closures occurred under a reforming Prison Commissioner, Alexander Paterson, well known for his broader efforts to further rethink and humanise prisons (Paterson 1951). This extension of the positive penological agenda was also embodied in other decarcerally-oriented critiques of prisons of the period, notably The English Prison System by Sir Evelyn Ruggles-Brise (1921), the chairman of the Prison Commission; English Prisons Under Local Government by Sidney and Beatrice Webb (1922), leading members of the Fabian Society; and The Report of the Prison System Enquiry Committee authored by Stephen Hobhouse and Archibald Fenner Brockway (1922) (also published as English Prisons Today). Together they advocated a reduction in penal discipline, and an expansion in what we might now call the needs-based assessment of individual prisoners, in order to encourage rehabilitation and reintegration. This thinking remained influential into the 1930s but was challenged by the upheaval of the Second World War and its aftermath.
The 60-year decline in prisoner numbers came to an end with the Second World War (see Figure 1). In 1945, the average daily prison population was 14,708 (ten per 10,000 population). Five years later, it had risen to 20,500 (13.4 per 10,000 population); and by 1960, it was over 27,000 (15 per 10,000 population). New prisons were built in the late 1960s, although overcrowding would still be a significant problem in the 1970s when the prison estate reached 39,028 in 1970 (18 per 10,000 population) and over 42,000 by the end of the decade (17 per 10,000 population). Prison populations continued to grow in the 1980s, as did criticism of a tired, costly, overworked, and overwhelmed prison estate (for an overview see Crewe and Liebling 2017;Jewkes, Crewe and Bennett 2016). For a short period, in the early 1990s, prison numbers decreased (for four consecutive years) before rising steadily after Michael Howard's 'Prison Works' address at the Conservative Party Conference in October 1993, and continuing to rise after the New Labour government introduced a new raft of legislation designed to incapacitate dangerous offenders for longer and longer periods of time.
In 2004, the Home Office under New Labour published Reducing Crime: Changing Lives, a strategy underpinning the 'transfer' of much of the probation service to the authority of the prison service (Home Office 2004). A new overarching body, the National Offender Management Service, was established to further reduce costs in the criminal justice system, by opening up a 'market' in end-to-end offender management where private companies and voluntary organisations could bid for contracts to reform prisoners. This intervention, rooted in 'new public management' practices (McLaughlin, Osborne and Ferlie 2005) was a very unambitious approach, always unlikely to reduce a prison population largely acknowledged to be unsustainable. While the prison population has fallen since 2014 (when it reached a historic peak of 85,626 average daily population, it would have to continue its current downward trend for nearly a century in order to reach the lowest peacetime number of prisoners (9,638 in 1919). Since 2015, the coalition and subsequent Conservative governments have done little to address the problem of an unmanageable adult prison population, and tried to reduce costs, not by reducing the number of prisoners, but the number of prison staff. Efforts to 'transform rehabilitation' have focused on the further privatisation of the probation service and the incentivisation of penal performance through payment-by-results and selective social finance experiments (Burke and Collett 2015;Burke, Collett and McNeill 2018;Deering and Feilzer 2015;Fox and Albertson 2011). Notably, the privatisation of the probation system was declared a failure by the Justice Select Committee in 2019 (see https://publications.parliament.uk/pa/cm201719/ cmselect/cmjust/482/482.pdf (accessed 1 December 2019)).
However, this is still an interesting moment in the long history of decarceral trends charted here. New efforts to acknowledge the harms of incarceration and to drive decarceration are in train, particularly in devolved administrations in Scotland and Wales, and particularly in the female estate (see, for example, the 2007 Corsten Report at: https://webarchive.nationalarchives.gov.uk/+/www.justice.gov.uk/publications/ docs/women-justice-system-exec-summary.pdf (accessed 1 December 2019)). Abolitionist movements, enlivened by a new generation of activists, such as The Empty Cages Collective, continue to push for the closure or radical transformation of prisons and other institutions on the grounds that they reproduce and exacerbate repressive social division (Hart and Ginneken 2017; http://www.prisonabolition.org (accessed 1 December 2019)). Justice reinvestment pilots have explored the cost savings offered by community sanctions (Brown et al. 2016;Fox, Albertson and Warburton 2011) and recent government proposals are intended to limit the use of short prison sentences (see https://reform.uk/events/major-speech-rt-hondavid-gauke-mp-lord-chancellor-and-secretary-state-justice (accessed 1 December 2019)). Overall, however, prospects for large-scale decarceration of the adult prison population are still not promising. The scaling down of the adult estate would involve the decarceration of thousands of long-term prisoners in addition to a raft of new legislation to reduce sentencing tariffs, and a substantial investment in probation and community justice services. By contrast, the juvenile secure estate, which has often followed a different historical path, offers the greater promise of more rapid meaningful reform.

Trends in Youth Incarceration and Decarceration, 1850s to 1945
The dual track for juvenile offenders was established with the opening of Parkhurst Prison on the Isle of Wight in 1838. The prison held male offenders aged under 16 years in order to educate them with a moral, religious, and industrial education. Once ready for emigration, they were shipped to Western Australia from 1842. By 1849, this British colony had accepted 234 Parkhurst boys aged between ten and 21 years. By the mid-1850s, the Parkhurst apprentices were no longer sent overseas, and the function of the prison was broadened to include all boys sentenced to terms exceeding one year's imprisonment. In 1863, Parkhurst became a prison for young men; one with a poor reputation from the start, renowned and criticised for its harsh disciplinary regime. For all these reasons, the institution's young inmate population was scaled down and increasingly 'diverted' to newly-created juvenile reform institutions (Carpenter 1851) (see Figure 2).
Following some small-scale divergence schemes in the early 19th Century (Rogers 2014), the government established two major juvenile institutions in the mid-century designed to both protect, and correct, vulnerable and offending children (see Cox and Shore 2002). Reformatory schools, created in 1854, were reserved for children up to the age of 16 years convicted of a serious offence. Until 1899, children sent to a reformatory were first required to spend a fortnight in an adult prison in order to experience a 'short, sharp shock' of sudden imprisonment and metaphorically, and probably literally, to 'hear the clang of the prison gate'. They were then subjected to years in a reformatory system which combined harsh discipline with skills-based training and moral reform. The number of schools in England and Wales grew from eleven certified reformatories in 1854, to 53 certified reformatory schools in 1871 (Reformatory Schools Return 1854-5; Inspector of Reformatory Schools of Great Britain (1872)). By then they had been joined by a large number of Industrial Schools.
Industrial Schools were established in 1857 to hold children aged six to 16 years thought to be in need of protection, and also under-14-year-olds sentenced by magistrates and considered likely to benefit from the skillsbased training provided. Their admissions net was further widened in 1866 with new criteria allowing children to be admitted if they were orphaned or destitute, or if their parents were themselves imprisoned or unable to control them. Child-protection measures introduced in the 1880s (the Industrial Schools Amendment Act 1880 and the Criminal Law Amendment Act 1885) meant that children living in households linked to the sex trade, notably those 'found in houses used for immoral purposes' could also be admitted to an industrial school (Davin 1996, pp.163-4). As noted in our recent study of the longer-term impacts of historical juvenile incarceration (Godfrey et al. 2017), industrial and reformatory schools thus constituted the key custodial components of the emergent youth justice system [and] incorporated, from the outset, what later generations of criminologists would refer to as a 'justice model' (the punishment of deeds) with a 'welfare model' (the meeting of needs). Whereas approximately a thousand young people, mainly boys, were sent to reformatories each year from the mid-19th Century, this was small beer compared with the tens of thousands of boys and girls sent to industrial schools throughout the country. Industrial schools received fewer than 5,000 children from the courts in 1870; but nearly 14,000 just ten years later, and almost 25,000 by 1890. This 'great incarceration' of children was thus driven by the expansion of Youth Custody Rates 1856-1927 (under-21, and under-16  the 'welfare model' dimension of the Victorian juvenile justice system but, significantly, would reach a peak by the turn of the 20th Century and decline thereafter. The treatment of older youths followed a different, but connected path, with a focus on diverting them out of adult prisons. The 1895 Gladstone Committee proposed the introduction of new institutions for 16-to 21year-olds that would be 'educational rather than punitive', with a focus on routine, discipline and the enforcement of a strict internal authority (Departmental Committee on Prisons 1895, pp.30-1; Fox 1952, pp.329-31). Sir Evelyn Ruggles-Brise established the first of these institutions at Borstal Prison near Rochester, Kent, in 1902. The Prevention of Crime Act 1908 extended the Borstal system on a national basis, separated youths and adults out in the courts, and championed a system which was designed to reform children, provide them with skills which they could take to the marketplace, and ensure that they saw the benefit of a strict regime. Magistrates could still send a young person to prison in exceptional circumstances although convicted older youths were increasingly sent to reformatories and borstals (Godfrey et al. 2017).
The introduction of reformatories, industrial schools, and borstals was, therefore, both part of a net-widening process that brought children, young people and young adults into the 'care' of the State who would otherwise not have been subject to any form of State action, but also prevented them from experiencing the full disciplinary rigour of adult imprisonment. The establishment of this extensive 'semi-carceral' juvenile system was very much aligned with the rise of the positive penological practice outlined above. However, net-widening had its limits. Reform and industrial schools saw a collapse in admissions, and mass closures, in the early 1900s (as Figure 3 shows). Figure 3 clearly illustrates the scale of juvenile decarceration in this period. The numbers of children and young people sent to reformatories and industrial schools had risen rapidly from the 1860s onwards, and, unlike the adult prison population which peaked in 1879, continued to rise until the early 1900s. Very much like the adult prison population, however, numbers then fell sharply after that point in a downward trend that also lasted until the Second World War. Nearly 100 industrial schools closed in the 1910s and 1920s. How might we explain this? The establishment of better welfare provision, higher real wages, smaller family sizes, State education, and the gradual replacement of the poor law with new forms of social protections and income maintenance for the low paid and unemployed, meant that fewer children were growing up in extreme poverty by the 1910s. This, arguably, did more than anything else to reduce the demand for industrial school places. Reformatories experienced a similar, although less steep, decline (Cox 2003, pp.80-2, 11-12;Radzinowicz and Hood 1990, p.182). The vast network of wider statutory and charitable child-saving institutions was also scaled back in the same period, although children's homes run by both sectors remained a part of the welfare landscape (Hendrick 2003). As with the adult decarceral trends, the rise of community disposals played a significant part in keeping young people out of Committals to Reformatories, Industrial Schools andBorstals 1893-1927 (Source: figures taken from Judicial Statistics, 1856-1927.) custody, for example juvenile probation, recognisance and fines expanded in the 1910s and 1920s. Birching (or corporal punishment) remained an option for boys but was less routinely used and was banned in 1948. The combined effect of these factors produced a dramatic change in the youth justice landscape, and their later amalgamation with reformatory schools under the Children Act 1933 within new streamlined 'approved schools'. During the 1930s, when the population of England and Wales reached over 40 million, only a few hundred children were being sent to some form of juvenile justice detention each year. This would all change, however, after the Second World War.

Trends in Youth Incarceration and Decarceration, 1945 to the Present
Wartime conditions were conducive to a growth in youth crime for a number of reasons, including reduced surveillance and social control over children and increased opportunities for looting, scavenging, mischief, and criminal damage (Adey, Godfrey and Cox 2016;Emsley 2011;Smithies 1982). After the war ended, young men in their twenties seemed to dominate prosecutions for robbery and violent crime (Hale 1999;Meier 2011). Concern about a perceived rising tide of youth and young adult offending prompted a 'return' to a more robust approach by the authorities (Lee 1998;Page 1950). The urge to reform delinquent youth (through reformatories, youth clubs, special schools, and so on) was not displaced, but it now ran alongside more punitive impulses. The incoherence and inconsistencies of these twin 'welfare' and 'justice' approaches can be seen in legislative innovations such as the passing of both the Children Act 1948 which empowered local authorities to take children into care, and the Criminal Justice Act passed in the same year which set up attendance centres and detention centres (Worrall and Hoy 2005). The new institutions were supposed to still 'advise, assist and befriend' the young person, but also to instil a strict regime of military-style drill and hard labour (Muncie 1984). The militarisation of many aspects of civilian life in a period of total war carried over, not surprisingly, into the post-war juvenile justice system. Youth custody numbers were fairly stable through the 1960s but rose again in the 1970s with the media stoking concern about deviant youth cultures (Cohen 1972) and a racialised 'mugging crisis' (Hall et al. 1978). In ways that echoed the mid-Victorian 'garrotting' crisis (Godfrey 2018;Sindall 1990), and the late Victorian 'hooligan' crisis (Pearson 1983), a small number of incidents were amplified by the press to create a disproportionate moral panic around 'disorderly' and 'dangerous' young men. The 'mugging' crisis was framed in a way that won public and political support for a more punitive approach (Hall et al. 1978, p.14;Williams and Godfrey 2018). As can be seen in Figure 4, following the initial spate of 'muggings' in 1972, the number of detention centre orders rose significantly, especially for boys and young men from black and minority ethnic backgrounds who were, and still are, heavily over-represented among the young incarcerated population (Cunneen, Goldson and Russell 2018;Lammy 2017). Yet, in the 1980s, the youth justice system continued to try to balance the punishing of  (Goldson 2015;Lotti 2016). The Criminal Justice Act 1982 abolished borstals and merged youth imprisonment and borstals into youth custody centres (YCCs) for those aged under 21 years, restricting the use of custody as a last resort, but reaffirming that YCCs were meant to provide a 'short, sharp shock' for their inmates. Eight years later, YCCs and detention centres were merged together to form young offender institutions (YOIs). Restrictions were placed on magistrates' use of youth custody, as they were when Article 37 of the UN Convention on the Rights of the Child 1989 stipulated that custody should be limited to the shortest possible time. Imprisonment was removed as an option for those aged under 15 years, and curfew orders were introduced for those aged over 16 years, but, in 1994, the Criminal Justice and Public Order Act once again allowed the imprisonment of children aged between twelve and 14 years (if they were persistent offenders). So, although new institutions for young offenders introduced in the post-war period were not meant to be as punitive as previous historical regimes, imprisonment for children was still very much an option for sentencers (see Coleman and Warren-Adamson 1992;Muncie 1990Muncie , 2009). In the 1990s, it was an option that was increasingly used (see Figure 5) (Cullen and Minchin 1999). Indeed, the rate of growth was such that it led to predictions in 2000 by the Ministry of Justice, that by 2007, the youth custody population would be nearly 13,000. In fact, it turned out to be a fraction of that number and, instead, we have seen a sharp decline in juvenile incarceration.
Between 2006 and 2016, the number of children and young people sentenced to custody fell by 74%. Rather than the positive penology, which had stimulated previous decarceral trends, the dramatic fall in juvenile incarceration shown in Figure 6 has, arguably, arisen from new public management strategies (Goldson 2015) and the need to reduce public spending, although the rise of restorative justice and referral orders as new community disposals has also clearly been significant.
In 1996, the Misspent Youth report published by the Audit Commission for Local Authorities and the National Health Service in England and Wales (1996) set the tone in criticising the youth justice system as too costly, too inefficient, and too ineffective, and two years later the Crime and Disorder Act 1998 placed greater emphasis on diversion from custody (Hindley, Lengua and White 2017, p.36) although it produced an initial net-widening effect, as Figure 5 indicates. Subsequent legislation increased post-release support to prevent reoffending (early release schemes, with curfews and electronic tagging replacing custody). However, the greatest change came in 1999 when the Youth Justice and Criminal Evidence Act created referral orders. First-time offenders, who pleaded guilty in court, received an initial judicial hearing, and were then sent to referral panels (made up of local YOT members, lay members, and sometimes the victim) within a restorative justice framework. In 2015, just over 12,000 referral orders were imposed, over 10,000 youth rehabilitation orders, and fewer than 3,000 detention and training orders (youth custody) (see the review of referral orders by HM Inspectorate of Probation (2016); see also Taylor (2016)). The 'decarceral' impact of referral orders and restorative justice initiatives has been rightly questioned (Crawford and Newburn 2003;Cunneen and Goldson 2015;Goldson 2000) but must be factored into this discussion of broad decarceral trends in youth justice. Today, the number of young people in custody is less than 1,000. The youth justice system therefore stands at a pivotal moment in its history, with modern youth incarceration rates back at the level they were in the late 1980s. This is arguably the moment to break with our long historical reliance on youth custody.

Re-imagining Juvenile Decarceration for Our Times
Current political economy models of punishment and justice are based on particular kinds of commodified contractual models that developed with the historical rise of capitalism yet persist today in late-capitalism. They imagine the serious young offender as a person owing a contractual 'debt' to society that must be 'repaid'. At present, incarceration remains the most powerfully symbolic means by which the youth courts can demonstrate that a serious young offender has been required to 'pay the price' for their crime. This is one of the factors impeding full youth decarceration today. However, this 'debt'-based view is at odds with an earlier 19th Century reformist view of the young offender as a potential worker who could become economically productive, both to reduce the future 'burden' they might place on the community and to improve their character. The creation of new industrial and reformatory schools in the 1850s was an attempt to recalibrate the political economy of punishment by removing children from symbolic incarceration within prison and replacing it with a new social investment in their rescue, reform, and training. In other words, children were placed in these institutions not only because they owed a debt to society but because society also owed a debt to them. It is ironic, if not surprising, that -over time -these institutions should have evolved into the symbolically punitive spaces of last resort that survive today. What tools are available to us to help us to reimagine the young offender as a subject of late-modern governance today? How can we combine a political economy of reintegrative punishment with a political economy of rights-based care? Current child rights frameworks demand that we offer children (including those who have been convicted of serious offences) the right to protection, provision, and participation (CYC-Online 2000). If we imagine young offenders as citizens with rights, we pay greater attention to their needs. Further, if we imagine them as relational citizens whose basic needs must be met by others (their families, communities, and wider society), we pay greater attention to the social networks around them and to the factors that inhibit and enhance these (Cox, P. 2018). The US 'smart decarceration' movement offers exciting possibilities here, with advocates calling for forms of decarceration that are 'effective, sustainable and socially just', and that 'catalyze a paradigm shift in justice, punishment, rehabilitation and safety' by 'chang [ing] the thinking that led to mass incarceration in the first place' (Epperson and Pettus-Davis 2017;Johnson, Hoelter and Miller 1981;Sherraden 2017, p.viii).
Recent work which has assessed longitudinal evidence of the impact of institutional care for vulnerable and criminalised young people concluded that reform and industrial schools offered certain protective effects and were likely to have contributed to the very low post-release reoffending rates of those children sent to them (Godfrey et al. 2017). Two particular features were prominent among these protective effects: first, all these institutions offered basic education and through-the-gate employment skills for all young people, which made them employable (and equipped with the skills demanded by local employers with whom they had already had placements). Second, many young inmates benefitted from the long-term caring relationships with staff which were continued past their release from the institution. However, these protective effects came at a substantial and questionable cost. They were rooted in Draconian practices of child removal and were enacted within institutions that, taken as a whole, had a poor or hidden record of child maltreatment. Recent studies of historical institutional abuse (Bingham et al. 2016) indicate that everyday emotional, physical, and sexual violence -between some staff and children, as well as between children -flourished in such settings, and still do (Smaal 2013). In other words, the protective effects offered by these institutions must be offset by their harmful effects.
Overall, then, the social conditions that sustained the 'success' of early youth justice institutions are not replicable today. Further, we should no longer need to look to youth justice institutions to provide the protective effects -caring relationships, education and employment skills -that are now widely acknowledged to prevent young people from (re)offending and to encourage desistance. Such protections are now recognised as rights owed to all children and are best realised within families, schools, and communities. Social policies fully informed by child rights operating within a more fully redistributive economy can yet play a key role in further reducing the volume of first-time entrants into the youth justice system. Enhanced safeguarding, diversion, restorative justice, and therapeutic interventions can yet further reduce the dwindling demand for youth custodial places. A 'social model' of youth offending is required here, drawing on a 'social model' of child protection (Featherstone et al. 2016;Goldson and Muncie 2015), that further roots the remedies of troubled youth within families, networks of care, and communities, rather than custody.
Some will argue that some minimal secure facilities must be retained for those young people who, despite early intervention, have a high risk of harming themselves or others. We would agree with this argument, but would also suggest that such facilities should be decoupled, as far as possible, from YOIs and secure training centres (STCs) because these institutions are failing to deliver such provision in a way that keeps vulnerable young people safe from harm. Recent evidence here is shocking and substantial (Goldson 2015).
A recent report by the Standing Committee for Youth Justice (2016), an organisation with member organisations drawn from across the sector, has described our remaining YOIs and STCs as 'violent, intimidating and … not conducive to helping troubled children to change', while noting that secure children's homes (SCH) offer a much 'higher level of care' to 'some of the most vulnerable children in the country' (p.3). Even then, a teenager living in a SCH or other kind of children's home is almost six times as likely to be criminalised as a looked-after child in another form of placement, and 20 times more likely to be criminalised than a non-looked-after child (Standing Committee for Youth Justice 2016). Further, recent inspections of Medway STC by the Care Quality Commission, HM Inspectorate of Prisons and Ofsted (2017) judged it to be 'inadequate' across several criteria: 'overall effectiveness', 'safety of young people', 'promoting positive behaviour', and 'effectiveness of leaders and managers'; and to 'require improvement' in several others: 'care', 'achievement', 'resettlement', and 'health' of young people. This is a damning indictment of the performance of a youth justice estate that is evidently unfit for purpose and has since prompted further calls from 36 leading child rights advocates for urgent action (Children and Young People Now 2019). The impact of that poor performance is felt more keenly by young black and minority ethnic inmates who make up a rising proportion of the remaining 1,000 young people in youth custody, now accounting for up to 40% of this group (Cunneen, Goldson and Russsell 2018;Lammy 2017, p.4).
Youth decarceration today is contingent on the closure of our remaining youth incarceration facilities and the creation of alternatives to youth custody. This has now, arguably, become more of a mainstream view in the sector. The Standing Committee for Youth Justice (2016) calls for the development of intensive therapeutic models as alternatives to youth custody. They are supported in this by the Howard League who, in turn, back the conclusions of the UN Committee on the Rights of the Child on the pressing need for alternatives to child custody in the UK (see https://howardleague.org/news/ governmentfailingchildren/ (accessed 1 December 2019)). A 2013 consultation on the youth justice system completed for the Ministry of Justice, with input from the Magistrates' Association, invites us to imagine a 'radically different youth custody' where community-based educational alternatives would play a much greater part (Ministry of Justice 2013, p.5).
Those alternatives are already established and are ready to be scaled up. In many ways, they are redolent of the positive penology and therapeutic impulses that underpinned the 1895 Gladstone Committee and the first phase of de facto decarceration practice outlined earlier in this article. The Family Centered Treatment (FCT) model, for example, offers a 'home-based service for juvenile court-involved youth' that has been shown to be more effective than group care (GC) in reducing reoffending (Bright et al. 2017). New specialist community forensic child and adolescent mental health services (FCAMHS) exist in some parts of the UK, although 'there is still no secure mental health in-patient provision for young mentally disordered offenders (or those with mental health needs outside the youth justice system who present high risk of harm to others) in Scotland, Wales or Northern Ireland' (Hindley, Lengua and White 2017, p.36). There is potential for much 'greater co-ordination between different providers of services for high-risk young people, including multisystemic therapy (MST), multidimensional treatment foster care (MTFC) and mental health in/out-reach teams working in specialist residential settings and special educational settings' (p.36). Criminological critics of decarceration have tended to regard any form of residential restraint as ultimately incarceral and punitive (see Hudson 1984). We recognise that possibility -and also note important critiques of 'alternative provision' and 'secure colleges' (George 2018) -but also argue that 'net-widening' arguments like these need to be more nuanced: our duty of care to some young people will, in certain circumstances, require the use of rights-based, therapeuticallyoriented residential facilities.
Such facilities could be co-ordinated as part of a broader 'smart decarceration' programme of the kind being developed in the US (Epperson and Pettus-Davis 2017) that seek to effect 'criminal justice transformation' by reducing 'excessive, racially biased, ineffective and unaffordable' mass incarceration (Sherraden 2017, p.viii). In England and Wales, the FCT and FCAMHS projects have the potential to become part of a 'smart decarceration' initiative that could help to push the decarceration pendulum further by posing hard questions about the 'inevitability' of continuing policies of child incarceration and child removal.

Conclusion
The juvenile decarceral trends of the late 1980s and early 1900s and, again, since 2008, are particularly significant moments within the longer history outlined here, because they marked distinctive shifts in respective cultures of youth justice and a collective questioning of youth institutionalisation as a response to social exclusion of many different kinds.
In the early 20th Century, a policy drift towards de facto youth decarceration dissipated for a number of distinct reasons: because of the challenge of the Second World War and its aftermath; because it was not articulated as a strategic aim by any major players; because progressives believed in their ability to improve remaining institutions in line with positive penological ideals; because the practice of child removal and institutionalisation remained central to other areas of social policy, notably in child protection; because therapeutic alternatives and more psychologically-informed approaches to youth justice remained relatively marginal and/or too costly. Most importantly, it dissipated because of an underlying faith in the 'last resort' symbolic place of total institutions in criminal justice policy.
In the early 21st Century, the current move towards youth decarceration is much stronger because it is more orchestrated and more clearly articulated by the core agencies charged with the care and control of young people. It connects more readily with new public management discourses which prioritise value for money, performance measures, and actuariallyinformed interventions. Put simply, youth decarceration may happen because youth incarceration does not appear to offer a return on public investment.
Since the early 19th Century, subsequent generations of reformers have tried and failed to improve conventional juvenile justice institutions. We argue that it is time to stop believing that we can do this, and time to start thinking more creatively about real alternatives to youth incarceration.