Inside England and Wales's prisons crisis: What might be driving differences in performance?

All other things being equal, violence and protests are higher in overcrowded prisons 

A higher proportion of prisoners in crowded accommodation is associated with higher rates of violence and protesting behaviour. For assaults on staff, a 1 percentage point increase in the proportion of prisoners in crowded accommodation implies around 1.3 additional assaults on staff per 1,000 prisoners. The median prison has 11% of prisoners in crowded accommodation. Moving from this point to the 75th percentile, with 34% in crowded accommodation, implies an additional 32 staff assaults per 1,000 prisoners – an increase of more than a third. Incidents at height show a very similar pattern, with the same increase in crowding rates implying almost a 50% increase in the number of incidents at height per 1,000 prisoners (92 vs 62). 

While we cannot identify a causal relationship between crowding and violence or protesting behaviour based on this analysis, these are fairly strong relationships. 

For prisoner-on-prisoner assaults, the effect is less clear. They may also be slightly more common in more crowded conditions, but the effect is smaller – at less than one additional assault per 1,000 prisoners for every 1 percentage point increase in crowding, a finding that is not statistically significant meaning there may be no effect. It is possible that prisoner-on-prisoner violence is less likely to be reported or observed by staff in a very overcrowded prison, whereas staff assaults are more likely to be reliably recorded.

Higher participation in purposeful activity is strongly associated with lower violence and protest 

Purposeful activity has an even stronger relationship with violence rates, with prisoner-on-prisoner and staff assaults both lower when more prisoners are engaged in it. Each percentage point more prisoners (10 per 1,000) engaged in purposeful activity implies around four fewer prisoner-on-prisoner assaults and 2.5 fewer staff assaults per 1,000 prisoners. A prison with 80% of prisoners engaged in purposeful activity (the rate at the 75th percentile) will have an estimated 26% fewer prisoner-on-prisoner assaults and 17% fewer staff assaults than the median prison (where 67% of prisoners are engaged in purposeful activity). Incidents at height may also be less common where more prisoners are engaged in purposeful activity, though this effect is not statistically significant.

This finding is particularly interesting because purposeful activity involves prisoners spending time out of their cells mixing with other prisoners, which on the face of it might provide more opportunities for violence. The large drop in violence in 2020/21, for example, is typically attributed to prisoners spending almost all their time in their cells as almost all purposeful activity and time out of cells was suspended due to pandemic-era social distancing regimes. Indeed, some prison officers and leaders have been reluctant to return to a pre-pandemic prison regime for fear violence levels would increase. In 2020 the Prison Officers’ Association called for the extreme Covid restrictions on time out of cells to be made permanent in order to keep prisons “well ordered, less violent and more stable”. ²⁷ Fairhurst M, ‘National Chair: Returning to chaos is not an option’, undated, retrieved 23 December 2024, www.poauk.org.uk/news-events/news-room/posts/2020/august/national-chair-returning-to-chaos-is-not-an-option

Our findings suggest that time out of cells does not increase violence and may even have the opposite effect. This aligns with other evidence showing that purposeful activity can improve behaviour in prisons. ²⁸ McGuire J, Understanding Prison Violence: A Rapid Evidence Assessment, HMPPS, 2018, https://assets.publishing.service.gov.uk/media/5b8e434540f0b67d970b92d5/understanding-prison-violence.pdf  There are various reasons to think that providing more opportunities for purposeful activity could reduce violence. Prisoners are often keen to work to be able to buy goods from the canteen and make phone calls, so lack of access to earning opportunities may increase frustration. ²⁹ Institute for Government interviews, 2024.  Boredom and irritation if prisoners are spending long stretches of time in their cells with little activity may also increase the risk of violence. This may also increase the use of illicit drugs (which in turn can drive violence). ³⁰ Austin A, Favril L, Craft S, Thliveri P and Freeman TP, ‘Factors associated with drug use in prison: A systematic review of quantitative and qualitative evidence’, International Journal of Drug Policy, vol. 122, 2023, https://doi.org/10.1016/j.drugpo.2023.104248 ³¹ Taylor C, ‘Drugs and disorder: worrying times for prisons’, HM Inspectorate of Prisons, 25 March 2024, retrieved 7 February 2025, https://hmiprisons.justiceinspectorates.gov.uk/news/chief-inspectors-blog-drugs-and-disorder-worrying-times-for-prisons/

“Prisoners who don’t receive money from outside can only get it by working. If they’re not able to work [because there are no jobs available], that means they can’t pay for phone calls to family or buy things to make the prison environment bearable.” – senior prison officer 

Alternatively, it may be that high levels of violence make it challenging for staff to deliver purposeful activity – a causal relationship running in the other direction – or that both lower violence and higher levels of purposeful activity are signs of a prison that is performing well in other ways, such as strong relationships between staff and prisoners or effective leadership and culture.

Purposeful activity is not associated with operational staffing levels or cost per prisoner 

Perhaps surprisingly, we do not find a significant relationship between the proportion of prisoners in purposeful activity and either the number of operational staff per 1,000 prisoners or average cost per prisoner. This deserves attention as staff shortages are frequently cited as a reason for a lack of purposeful activity, including by the prisons inspectorate in its thematic review on time out of cell. ³² HM Chief Inspector of Prisons, Purposeful prisons: time out of cell, 2024, https://cloud-platform-e218f50a4812967ba1215eaecede923f. s3.amazonaws.com/uploads/sites/19/2024/09/Purposeful-prisons-time-out-of-cell-web-2024.pdf  However, we find no relationship between total operational staff and rates of purposeful activity. While teachers and trainers are not included in the staffing figures we look at here, prison officers and operational support staff are responsible for running the prison regime, unlocking prisoners and escorting them around the prison as necessary, so this is still a surprising finding. We also looked separately at the effect of more band 3 and 4 prison officers, and did not find a statistically significant effect on purposeful activity participation. 

We did find a significant relationship between purposeful activity and non-operational staff. Non-operational roles vary widely, ranging from administrative support to psychologists. They include probation officers working in prisons and other offender management roles, which involve developing sentence plans, advising and supporting prisoners and developing effective relationships with them to support rehabilitation. Having more of these roles per prisoner may encourage more prisoners to participate in purposeful activity by providing additional motivation and accountability, or may simply be associated with more availability of activities.

The same is true on cost per prisoner: the purposeful activity participation rate is associated with a slightly higher cost per prisoner, but again this is not statistically significant. In fact, controlling for staffing levels, higher rates of purposeful activity are significantly associated with lower costs per prisoner, suggesting that savings associated with purposeful activity participation (such as lower violence levels) outweigh the costs. In both cases, it may be that more staff do support a more comprehensive regime with more prisoners participating in purposeful activity, but this is outweighed by the stronger relationship between low purposeful activity participation and high violence, which also goes along with high staffing levels and high costs (discussed below).

These findings require an important caveat: we only have data on the proportion of prisoners in at least some purposeful activity, not on the amount of purposeful activity delivered. Two prisons, each with 100 prisoners, could each be delivering 200 hours of purposeful activity, one giving 100 prisoners two hours each and the second giving 20 prisoners 10 hours each. The first would have a purposeful activity rate of 100% and the second just 20%, even though they are delivering the same amount of activity. These findings do not therefore imply anything about the relationship between the amount of purposeful activity delivered and staffing or costs.

There is surprisingly little relationship between post-prison outcomes and conditions inside – though open prisons again do much better on employment 

We find little relationship between short-term outcomes after release and conditions or activity in prison, although there is some other evidence that these do affect reoffending rates. ³⁷ van Ginneken EFJC and Palmen H, 2022, ‘Is There a Relationship Between Prison Conditions and Recidivism?’, Justice Quarterly, vol. 40, no. 1, pp. 106–128, www.tandfonline.com/doi/full/10.1080/07418825.2022.2040576 ³⁸ HM Inspectorate of Prisons, Purposeful prisons: time out of cell, September 2024, https://cloud-platform-e218f50a4812967ba1215eaecede923f.s3.amazonaws.com/uploads/sites/19/2024/09/Purposeful-prisons-time-out-of-cell-web-2024.pdf  Open prisons have much higher rates of employment six weeks after release than any other prison type. Within prison categories, however, there is no identifiable relationship between employment and levels of purposeful activity, violence rates or overcrowding. This suggests that, excluding open prisons, employment rates after release are largely driven by external factors, rather than conditions within prisons – at least in the short term and in current circumstances, where resettlement support tends to be limited everywhere. 

Similarly on housing, there is little relationship for individual prisons between conditions in prison and the proportion of prisoners housed on the first night after release. While higher performing prisons might be expected to have more time and attention to spend on resettlement work, this does not appear to be significant in terms of immediate accommodation outcomes, though this does not rule out a longer term impact. More crowded prisons and reception prisons both do worse on first night housing, but this is likely to reflect geographic and economic characteristics of the areas: more crowded prisons are likely to be in urban centres, particularly reception prisons, where there may also be more intense pressure on housing generally. This is also a very small effect, with only a few percentage points of variation in the proportion of prisoners housed for their first night after release. 

Controlling for other factors, more violent prisons have more staff 

Rates of prisoner-on-prisoner and staff assault are both associated with more operational staff per 1,000 prisoners. This includes all staff in operational roles, but the majority of these roles are band 2 operational support staff and band 3–4 prison officers. The median prison has 388 operational staff per 1,000 prisoners. A prison with 532 (the rate at the 75th percentile) will have an estimated 26% more prisoner-on-prisoner assaults and 44% more staff assaults than the median prison. The same pattern applies even more strongly for incidents at height, which are an estimated 64% higher in the prison with more staff. 

These effects are likely due to a combination of factors. Prisons that have previously had high levels of violence are likely to recruit more staff to manage that violence. ³⁹ Institute for Government interviews, 2024.  Some will also have physical infrastructure – such as limited central visibility of wings or large distances between blocks – that make it harder to maintain order and security, which may lead to both higher staff requirements and more violence. ⁴⁰ Institute for Government interviews, 2024.  Higher staffing levels may also increase the chance that violence is directly observed or reported by a prisoner, increasing recorded assaults even if the actual level of violence is not different. 

Resignation rates are higher in more violent prisons, particularly where assaults on staff are more frequent. This higher turnover rate potentially mean that staff are less experienced and less likely to have developed relationships with prisoners, which could exacerbate violence. 

“Lots of band 3 prison officers working the wings is a good thing – that’s the traditional view. [But] if they are very inexperienced, then they are not really working the wings, they’re sitting in an office terrified”– voluntary sector organisation

We also looked at officer seniority. A 1 percentage point shift in the proportion of operational staff that are senior officers (band 4 supervising officers and band 5 custodial managers) rather than junior prison officers is associated with lower levels of violence, particularly against staff. Staff assaults are around 19% lower in a prison where 13% of operational staff are senior officers, compared to the median prison, in which 11% are senior officers. Prisoner-on-prisoner assaults also appear to be lower; this finding was not significant based on 2023/24 data but was when we looked across multiple years.* We found similar results when looking at the number of senior officers per 1,000 prisoners, rather than the proportion.** This may reflect either benefits of having more managers and supervising officers, or it may be a proxy for officer experience: these officers are typically more experienced than band 3–4 officers. As such, they may be better able to de-escalate situations that might otherwise deteriorate into violence and may have better relationships with or more respect from prisoners. We also find a small but significant relationship between more non-operational staff per 1,000 prisoners, and lower rates of assaults on staff. Again, this may be associated with stronger relationships between prisoners and non-operational staff such as offender managers.

_{* We are unable to control for all variables across multiple years because of data gaps. See Methodology for details.} 

_{** We ran a number of different regressions to understand the relationship of staffing levels to violence and purposeful activity. See Methodology for details.}