Professor Alex Sutherland, Co-Director of the Oxford Martin Programme on Decarceration and Professor in Practice of Criminology and Public Policy at the University of Oxford, explores the options for making early release and alternatives to prison safer.

When Brahim Kaddour-Cherif was mistakenly freed from custody, the case drew national attention. He later assaulted two police officers before being detained again. His case was not an isolated mistake. In the year to March 2025, 262 prisoners were released in error, more than double the number recorded the previous year.

It is against this backdrop of rising errors and public concern that major criminal justice reforms are now taking effect. Forced to act by prison overcrowding and a court backlog of nearly 400,000 cases (80k in Crown Court; 300,000 in Magistrates’ Court) the government introduced a raft of changes almost simultaneously. Included in these measures they have effectively scraped short-term prison sentences under 12 months, moved the early prison release threshold from 40% to 30% of time served,  and increased the use of alternatives to prison such as suspended sentence orders (SSOs), electronic monitoring and community supervision (see here and here). In progressing this criminal justice reform programme, the government is resting a good deal of public safety on the re-nationalised probation service and local policing.

There can never be ‘zero risk’ when it comes to releasing people from prison or placing them on community supervision, but all of this raises questions about how to enact these changes while balancing public safety. Risk assessments, pre-release planning and “through-the-gate” support are designed to identify and manage these risks before someone returns to the community.

However, when prisons are overcrowded, staff stretched, resources tight and the legislative and operational context changing rapidly, release planning and support may not happen as intended, despite the efforts of prison officers and staff. For example, to be released early prisoners need to complete groups or programmes in prison (e.g. education, training, rehabilitation). But overcrowding, a lack of prison staff in some prisons, waiting lists for programmes, and prisoners serving shorter sentences means that many may not be able to access services (see here). What happens next isn’t clear, but it’s likely that some will end up not being released early, and some may be released without having attended any in-prison programmes. Even those who do attend programmes may not benefit, depending on the quality of the programme itself and how well it has been delivered.

Releasing prisoners without services and support is almost guaranteed to increase the risk that many of them return to prison for breaching their community sentencing conditions or for further crimes, particularly when they face barriers to gaining even the most basic employment. But we also know that it is possible to support behaviour change, reduce crime and improve lives with rehabilitative intervention over and above monitoring and supervision. Knowing these changes are coming, questions about ‘how?’ must be approached using evidence from multiple perspectives and considers the wider social and societal consequences. There is also a need to view policy changes – and engage in public debate - from an ethical perspective, assessing the economic, human and societal costs of imprisonment and its alternatives. Failing to acknowledge trade-offs and the limits of public acceptability, and not being honest with the public about the trade-offs required, plays into the hands of simplistic story-telling about ‘how to solve’ crime and punishment.

It is not the first time that these issues have come to the fore in criminal justice. Prison capacity has been a regular feature of what was a highly politicised topic for the last 30 years and more. We are seeing more discussion about criminal justice and crime – including mis- and disinformation about ‘the crime problem’ in UK cities when some crimes, such as homicide, are at historic lows. That said, problems of release and supervision post-release are also not new – but the forthcoming changes represent a genuine opportunity to assess the effectiveness (or otherwise) of many of these changes and for research to inform both what needs to happen and how it might be done better in the future. In five years’ time we could be in a fundamentally different place in terms of knowledge about policy effectiveness and our ability to quickly and reliably test ideas out, but only if we push for it now.

This article was republished. The original version can be found here.

This opinion piece reflects the views of the author, and does not necessarily reflect the position of the Oxford Martin School or the University of Oxford. Any errors or omissions are those of the author.