The crisis in prisons

The Sunak government implemented several emergency measures to ease the crisis. Some prisoners have been released up to 10 weeks early, court hearings delayed and police cells used as overflow capacity. But it is clear these are insufficient and that the next government – if the polls are correct, a Starmer-led Labour government – will need to act immediately to begin to fix the problem. This short paper looks at how the UK criminal justice system got into this situation, and how it can find a way out of it.

How did we get here?

The prison population of England and Wales has doubled over the last 30 years, despite crime rates falling substantially. It is currently just over 87,000, up 13% just in the last three years, ⁵⁹ Ministry of Justice, ‘Prison population figures’, ‘Population bulletin: monthly May 2024’.  and the Ministry of Justice projects this to hit 99,300 by the end of next year. ⁶⁰ Ministry of Justice, ‘Prison population projections: 2023 to 2028’  Even with new prisons being built, capacity is not expected to grow anything like that fast: only around 4,400 new spaces are planned – but this is against an estimated 12,000 more prisoners. ⁶¹ Ministry of Justice, ‘Written evidence submitted by the Ministry of Justice’, UK Parliament, October 2023, committees.parliament.uk/writtenevidence/126249/pdf/

Prison population is determined by two things: how many people are sent to prison, and how long they stay there. And importantly, there are three groups of people in prisons: ‘standard’ prisoners, who are serving custodial sentences; people on remand, waiting either to be tried or to be sentenced after conviction; and people who have been recalled to prison after release for breaching the terms of their parole. Recognising the trends in all three of these groups is central to understanding why we have reached this crisis point.

Most of the people in prison are serving custodial sentences, and the bulk of the long-term growth in population has come from this group. In the 2000s, this was largely due to more people going through the courts and being convicted. ⁶² Ministry of Justice, ‘Criminal justice statistics: July 2011 to June 2012’, ‘Sentencing tables: June 2012’.  From the 2010s, however, it has instead been driven by longer sentences. In 2023, the average custodial sentence given at the crown court, which deals with more serious offences, was more than 25% longer than in 2012. ⁶³ Ministry of Justice, ‘Criminal court statistics quarterly, October to December 2023’, ‘Criminal court statistics quarterly: October to December 2023 (tables)’.  Some of this is due to more convictions for violent offences, which tend to attract longer sentences. But the same crimes are now attracting longer sentences as well. Sentences for robbery, for example, were 13 months longer on average in 2023 than in 2012, an increase of 36%. ⁶⁴ Ministry of Justice, ‘Criminal justice statistics quarterly: December 2023’, ‘Outcomes by offence data tool: December 2023’.

These longer sentences kept the prison population largely flat during the 2010s, even as the number of people tried and convicted in the courts was falling.

In the last few years, however, it is the other two groups – those on remand and those recalled to prison – that have principally caused the population to rise rapidly. The number of people on remand has risen 84% since 2019 and now accounts for almost 20% of the total prison population. Of these, two thirds are yet to be convicted of a crime. ⁶⁵ Ministry of Justice, ‘Offender management statistics quarterly, October to December 2023’, ‘Prison population 2024’.

This change has been driven by the growing backlog in the criminal courts, particularly the crown court, meaning people are waiting much longer for hearings. The number of people released on parole and then recalled has also risen sharply – up 72% since 2019, now accounting for 14% of all prisoners. ⁶⁶ Ibid.  The main reasons for this are people spending longer on parole after release from prison (the product of longer overall sentences) and serious problems within the probation service.

Some of these changes would have been difficult to see coming. The crown court backlog was already growing in 2019, but the pandemic unquestionably made the situation much worse, and very quickly. ⁶⁷ Ministry of Justice, ‘Criminal court statistics quarterly, October to December 2023’, ‘Criminal court statistics quarterly: October to December 2023 (tables)’.  Recall rates were fairly steady between 2016 and 2018, before rising rapidly. ⁶⁸ Ministry of Justice and HM Prison and Probation Service, ‘Offender management statistics quarterly, October to December 2023’, ‘Prison recalls: October to December 2023’.  However, the core of the problem lies in the longer-term trend for lengthier sentences and these were foreseen, and indeed predicted for years in the government’s annual prison population projections.

Prisons in 2024 are in a state of emergency, but one that has been brewing for years. Successive justice secretaries and prime ministers have chosen not to address the drivers of the crisis; some have added to them. Since 2010, budgets have been cut and supply of places underdelivered, while sentences have continued to grow. And since at least 2020, the government’s own prison population projections have implied that prisons would suffer a severe capacity crisis without corrective. That did not happen. 

As a result of this failure to manage capacity, the government was forced to take emergency measures in November 2023. The actions taken – over the last decade and particularly the last few years and months – have been profoundly inadequate. The crisis now facing the criminal justice system is one driven by sustained mismanagement, rather than unforeseen circumstances. It is a political and policy failure.

Where are we now?

The Prison Governors’ Association has said prisons are days away from becoming full. Without action taken by the government to free up capacity, the implications would be stark. The courts would have nowhere to place serious offenders who present a real risk of the public. The police’s ability to make arrests would be constrained. This would call into question one of the most basic roles of government: keeping the public safe.

The situation calls for further emergency measures. These will be unavoidably short term: longer-term, systemic reform – like revising sentencing policy, expanding prison capacity or offering more options for community rather than custodial sentences, among others – will all take months or years to enact and take effect. The next government is likely to have weeks or even days. 

Emergency measures instituted so far have included releasing prisoners up to 10 weeks early, using police cells as overflow capacity and delaying court hearings. This last initiative, launched in May under the title Operation Early Dawn, has probably been most effective, but carries significant downsides. People who would have been remanded to custody, for example because they are at high risk of committing further offences or interfering with witnesses, have instead been released on police bail after being charged. Longer waits for trials and sentencing leave victims, defendants and their families in limbo, and increase the number of victims dropping out of trials.

The current system of early release is haphazard, with confusion reported among inmates and staff on when prisoners are due to be released. ⁷² Institute for Government interviews, 2024.  Many prisoners are released without homes to go to, putting additional strain on the probations service's already limited resettlement support. Little data is available, but there are reports that many of those released early have been recalled almost immediately. ⁷³ Hymas C, ‘Up to half of prisoners released early recalled within days at some jails’, The Telegraph, 30 May 2024, retrieved 01 July 2024, www.telegraph.co.uk/news/2024/05/30/up-to-half-of-prisoners-released-early-recalled/#:~:text=Up%20to%20half%20of%20prisoners%20released%20early%20hav…

Confusion around release dates has also increased the pressure on probation, which is struggling to supervise new parolees effectively. Prisoners are identified for early release by the central prison and probation service according to specific criteria, and the risk assessment processes in place are opaque. While prison governors can raise concerns if they think specific individuals pose a risk, they do not have a veto over who gets released. ⁷⁴ Institute for Government interview, 2024.

How can the next government begin to solve these problems?

The next government will need to act, and fast, but it has limited options available to it. Before beginning to design its plan to fix prisons, it will first need to take into account several important constraints.

What is standing in the next government’s way?

Prison space

There are simply not enough physical cells to support further growth in the population. Building new spaces would take too long, and other options – such as using army barracks or barges for additional accommodation – would not offer enough space, even if they could be made secure and meet human rights requirements. There is also a risk of suddenly losing a few hundred spaces, either due to a riot or an unexpected safety issue, as happened in March when more than 180 cells were closed at HMP Dartmoor due to the discovery of radon gas. ⁷⁹ Telegraph reporters, ‘Radioactive gas detected at HMP Dartmoor forces removal of prisoners’, The Telegraph, 10 March 2024, retrieved 1 July 2024, www.telegraph.co.uk/news/2024/03/10/radioactive-gas-dartmoor-prison-prisoners-removed/

Staffing and problems with ‘crowding’

Accommodating more inmates in existing prisons (‘crowding’) is, in theory, an option but would require more staff to ensure this is done safely. Recruiting and retaining prison officers is already challenging, while crowding itself presents practical, and legal, issues too. Increased crowding and fewer staff per prisoner makes violence more common: violence levels have already been rising, including serious assaults, and the risk of riots and unrest will be elevated over the hotter summer months. Around 10% of prisoners are already in accommodation not rated ‘decent’ by the prison service, ⁸⁰ Ministry of Justice, ‘Prison population figures: 2024’.  and Prison Governors’ Association have threatened judicial review at the prospect of further crowding. ⁸¹ Syal, R, ‘Prisons in England and Wales will be at ‘breaking point’ in July, governors told’, The Guardian, 20 June 2024, retrieved 1 July 2024, www.theguardian.com/society/article/2024/jun/20/prisons-in-england-and-wales-will-be-at-breaking-point-in-july-governors-told

Probation capacity

Probation is dangerously overstretched and already rationing supervision. Most prison leavers are now only supervised for the first two thirds of their licence period, the part of their sentence they spend in the community. ⁸² Ministry of Justice, Written Ministerial Statement, ‘Update on foreign national offenders, prisons and probation’, 11 March 2024. questions-statements.parliament.uk/written-statements/detail/2024-03-11/hcws332  Only the most serious offenders will be supervised for their whole sentence. Probation capacity to support offenders and manage risk is likely to be a major limiting factor if moving offenders from prisons to community supervision.

The judiciary

Not everything is within the control of the Ministry of Justice or the prison service. Judicial independence means any changes affecting court decisions is dependent on support from judges, including who gets remanded, what sentences are given and the cases listed in the courts. The judiciary is likely to be responsive in the context of an emergency, but it cannot be taken for granted.

Most effective options

Given the scale of the problem, the new government will need to free up thousands of prisons places. Unfortunately, there are few options that will have sufficient impact, can be implemented quickly and pose low risk in terms of both public protection and miscarriages of justice. But there are some that would make a meaningful impact on prisoner numbers, albeit with various practical and political challenges. At least one of these will likely be required immediately. 

Lower the point of automatic release for most offenders from 50% of their sentence to 40-45%

This is the lever that delivers the greatest capacity the quickest, as it can be done in weeks. While it would likely require new legislation, it would be fairly straightforward to draft and deliver. Lowering automatic release below 50% would however have a symbolic impact, as it would mean criminals spend more time on licence in the community than in prison, which would likely damage public confidence in the system. But it could be moderated – for instance, it need not apply to the most serious violent and sexual offenders, who currently serve two thirds of their time in prison and would continue to do so. This would limit the public safety risk and free up capacity for serious offenders in prison. 

Those released would be supervised by probation, and if they committed further offences or breached conditions, would be recalled to prison. This unavoidably places some additional strain on an already creaking probation system and high recall rates, as have been seen in the current early release scheme, could see a chunk of any capacity freed up being used for recalling the same prisoners. This option is likely to be necessary to deliver sufficient space in the available time – even if other options are also used.

Introduce a ‘queuing system’ for immediate custodial sentences, with low-risk prisoners starting their sentence on house arrest

Some prisoners already spend the last 12 weeks of their sentence in home detention, with a curfew enforced by an electronic tag. The government could choose to take the same approach to lower-risk offenders at the start of a prison sentence, until prison space becomes available. The time spent in home detention would count towards a prisoner’s sentence, to avoid a growing backlog. If they breached conditions, they could be sent to prison immediately. If they do not, and a space does not become available for the duration of their sentence, they could have served their sentence without entering prison.

This would prioritise prison places for the highest risk offenders, while taking a more lenient approach to those who are lower risk (a third of prisoners in March 2024 were serving time for a non-violent offence, for example). ⁹⁰ Ministry of Justice and HM Prison and Probation Service, ‘Offender management statistics quarterly, October to December 2023’, ‘Prison population 2024’.  The challenge in delivering this is time: it would require legislation to be drafted and passed, new probation capacity and electronic tags, and finally detailed design on who would be eligible and how breaches would be managed. 

This option would be unlikely to work either in isolation or if the pressure is so acute that there is no time to fully design and deliver it. There are different ways to approach the scheme to be less complex to design and pass through parliament, but these might limit the impact on the prison population.

Allow sentences of up to three years to be suspended

Currently sentences of up to two years can be suspended – that is, the offender does not have to go to prison as long as they do not commit any further offences and abide by any conditions attached. Raising this to three years could deliver a significant and sustained reduction in the prison population, while also improving outcomes. Sentences served in the community typically have lower reoffending rates, ⁹¹ Mutebi N and Brown R, The use of short prison sentences in England and Wales (UK Parliament POSTbrief 52), UK Parliament, July 2023.  and relatively few suspended sentences result in the offender breaching their conditions and entering custody. ⁹² Ministry of Justice and HM Prison and Probation Service, ‘Offender management statistics quarterly, October to December 2023’, ‘Prison recalls: October to December 2023’.  Most people receiving a sentence of 2-3 years have committed a non-violent offence (62%) ⁹³ Ministry of Justice, ‘Criminal justice statistics quarterly: December 2023’, ‘Outcomes by offence data tool: December 2023’.  so the public safety risks are relatively small, and the impact on public confidence should also be limited compared to other options. This option would also require legislation, but the practical implementation would not be complex.

However, any impact is highly dependent on how judges respond when handing out sentences. To have a rapid impact, this measure would likely need to be combined with a presumption to suspend sentences for certain types of offences, such as drug or theft offences, or sentences up to at least two years. This would be substantially more complicated to introduce, and would still be less certain than the other two options. If judges respond by increasing the length of sentences to avoid suspending them, it could even make the situation worse.

Supplementary options

There are a number of other promising options that will not have sufficient impact on their own, but may help buy more time for some longer-term measures to have an impact.

Remove supervision or abolish recalls post-release for offenders serving sentences of under 12 months

This would roughly halve the number of recalls. The impact on the prison population would likely be fairly small, as most of these recalls are only 14 days, but it would free up some probation capacity for more impactful measures, and greatly reduce the burden on prisons from entrants and releases. However, some of this prioritisation is already happening under ‘Probation Reset’, ⁹⁴ Ministry of Justice, Written Ministerial Statement, ‘Update on foreign national offenders, prisons and probation’, 11 March 2024, questions-statements.parliament.uk/written-statements/detail/2024-03-11/hcws332  so the impact is hard to assess and could be small. Reduced probation supervision could be replaced by commissioned support from the third or private sector to meet offenders’ housing, mental health and other needs in the community. This would help address reoffending and could be established more quickly than expanding probation capacity.

Reintroduce the requirement for magistrates’ courts to approve standard recalls

This should reduce the number of recalls without having to raise the thresholds for recall, simply by adding an extra administrative step. Fixed-term and emergency recalls would not be affected, preserving flexibility for emergency situations. However, the additional burden this would place on magistrates’ courts and probation practitioners could be significant, given both are struggling to cope with existing demand.

Prioritise defendants on remand for case hearings

Many crown court defendants spend at least six months on remand and almost 30% do not receive an immediate custodial sentence. ⁹⁵ Ministry of Justice, ‘Criminal justice statistics quarterly: December 2023’, ‘Remands data tool’.  Accelerating the passage of these cases through the courts would free up capacity and prevent long periods in custody for those who should not be there.

Introduce a presumption to suspend sentences for certain offences or up to a certain length

A presumption to suspend sentences of up to 12 months would likely only reduce the prison population by a few hundred, but the evidence suggests it would cut reoffending and improve other outcomes as well. ⁹⁶ Ministry of Justice, ‘Sentencing Bill Impact Assessment – Short Sentences’, November 2023, assets.publishing.service.gov.uk/media/65537ccdc684c4000db64d21/Sentencing_Bill_-_IA_-_Short_Sentences.pdf  Though again this would result in more pressure on probation.

These are not competing options, and the next government will need to consider a response that involves a combination of measures to free up capacity while delivering on other policy objectives. For example, lowering the point of automatic release only slightly, to 45%, and reducing post-release supervision for those on short sentences could create both time and capacity to implement a queuing system. Alternatively, a more radical change to automatic release would provide more breathing room to do an in-depth review of sentences. 

Real-time prison capacity data is not perfect, and nor are population projections. Given that, the new government will need to create a sufficient capacity buffer, whichever options are pursued, to avoid further crises and emergency measures being needed. The system will need a sustained period outside of crisis to recover and reset.

Options to be avoided

There are some measures that should be avoided as part of a short-term response, either because they would not have a large or fast enough impact or because they pose significant risks. The first group includes measures which rely on changing judges’ and magistrates’ behaviour, such as encouraging community sentences or reducing the use of remand. These are likely to be needed in the medium to long term, but are hard to do quickly and so risky to rely on for immediate crisis management. The second group includes expanding the current early release scheme or further delays to hearings. Both likely carry greater public protection risks than the more managed schemes suggested above, and delaying hearings in particular undermines the quality of justice and exacerbates existing court delays. Measures to encourage guilty pleas, such as offering unconditional release for people on remand in return for a guilty plea if they have served more than their likely sentence, might be effective, but would risk miscarriages of justice.

First steps for the next government

The next government will need to decide quickly what steps it will take to address the crisis in prisons. It will need model all available options to get an accurate understanding of the effect on the prison population, most fundamentally, but also on probation services, the wider criminal justice system and the public. 

The options likely to have the greatest impact, soonest, are also likely to substantially increase the probation caseload. This is essentially unavoidable, so consultation with the probation service and providers of other services such as drug and alcohol support should be prioritised to see what is possible. This may involve relaxing supervision for some offenders, providing support for low-level offenders via the third sector, or prioritising certain groups for electronic tags, for example. 

A combination of options is likely to be necessary. These should be evaluated and modelled as a package, as well as individually, to understand the full implications across the system – including on probation and the criminal courts as well as prisons.

The government will also need to keep a close eye on implementation – being aware that it will take some time to ensure effective implementation plans are in place, including the communications to the public and the criminal justice system. Rushing measures, or announcing them too hastily, only risks undermining them. This will do little to repair the low confidence the public already has in the system.

At the same time, the government should develop a long-term strategy to reduce pressure on prisons in a more systemic, considered way than in recent times. There are more choices available in the longer-term: most obviously, prison capacity can be increased by building more prisons and recruiting more staff, while sentencing can be reformed to reverse some of the increases of recent years and changes made to the use of remand and recall. The recent growth in prison and probation workforces is also likely to improve things over time – if staff are retained – by improving the quality of support and supervision and increasing access to education and other purposeful activity. 

It is vital that planning for this happens quickly. All of the more comprehensive and long-term solutions require time and sustained commitment from the government to work, and the emergency measures discussed here, while necessary, will only buy so much time.