Performance Tracker 2025: Criminal justice system methodology

General

Public services spending (Figures CJS 2.1, 3.1, 3.3 and 4.1)

To estimate the real cost of public spending, we deflate government spending figures using the GDP deflators published by the Office for Budget Responsibility in the Economic and Fiscal Outlook from March 2025. ¹ Office for Budget Responsibility, Economic and fiscal outlook – March 2024, 2024, https://obr.uk/efo/economic-and-fiscal-outlook-march-2024/  To better reflect the underlying inflation conditions present in 2020/21, we estimate our own figures by generating a mid-point that averages across values from 2019/20 and 2021/22. We deflate spending figures in our financial analysis across the entire report to 2025/26 prices.

In cases where we calculate real-terms changes in figures that relate to individuals – such as wages or the adult social care means test – we use the consumer price index rather than the GDP deflator. The CPI that we use also comes from the OBR’s Economic and Fiscal Outlook from March 2025.

Police

Police spending projection (Figure CJS 2.1)

We assume that police spending rises in line with projected police spending power in the 2025 Spending Review. We then deflate the figures to 2025/26 prices in line with the process set out above (see Public services spending, above).

Charges/summonses and positive outcomes recorded by police forces (Figure 2.18 and 2.19)

Greater Manchester Police only supplied data for Q1 in 2019/20 and Devon and Cornwall Police only supplied data for Q1 and Q2 in 2022/23. To represent the national trend as accurately as possible, we have imputed missing figures for these quarters. For Devon & Cornwall, we used the average of the same quarter from the previous and subsequent year. For Greater Manchester, for each positive outcome type (charge, diversionary activity etc.) we calculated the proportion of the national total that Greater Manchester Police is typically responsible for. We based this on figures from 2016/17 to 2020/21, excluding 2019/20 when data is missing. We took this approach because of substantial variation in the number of charges recorded by Greater Manchester Police over the period, driven partly by an increased emphasis on charging over other outcomes in recent years.

Data for this chart is drawn from two different sources. Values for 2003/04 to 2013/14 are headline national figures from Home Office, ‘Crime outcomes in England and Wales’ (Table B2), 2022/23. Values from 2014/15 to 2024/25 are from Home Office, ‘Police recorded crime and outcomes open data tables’, 2018/19–2024/25. Totals may vary slightly across different data sources.

Criminal courts

HMCTS spending projection (Figure 3.1)

We assume that HMCTS spending rises in line with the overall Ministry of Justice RDEL budget in the 2025 Spending Review from 2024/25 to 2025/26. For subsequent years, we assume that the additional £450 million in courts funding announced at the
Spending Review is 90% RDEL and 10% CDEL, and that this is increased in three equal stages in 2026/27, 2027/28 and 2028/29. We then deflate the figures to 2025/26 prices in line with the process set out above (see Public services spending, above).

CPS spending projection (Figure 3.3)

We assume that CPS spending rises in line with the Law Officers’ Department RDEL budget in the 2025 Spending Review. We then deflate the figures to 2025/26 prices in line with the process set out above (see Public services spending, above).

Complexity-adjusted cases received and closed by the crown court (Figures CJS 3.16, 3.17, 3.19, 3.20)

We use the same method to adjust complexity for both receipts (cases coming into the court) and disposals (cases closed). This involved three stages:

•    Classifying cases into sub-types
•    Generating a complexity weighting for each sub-type
•    Weighting cases by complexity, anchored to the real number of cases in 2018

Case classification

We classified cases into sub-types in line with HMCTS classifications, according to both their case type and offence group, as listed below:

Case types:

•    Indictable-only for-trial cases
•    Triable-either-way for-trial cases
•    Committals for sentence
•    Appeals
•    Unknown

Offence groups:

•    Violence against the person
•    Sexual offences
•    Robbery
•    Theft offences
•    Criminal damage and arson
•    Drug offences
•    Possession of weapons
•    Public order offences
•    Miscellaneous crimes against society
•    Fraud offences
•    Summary non-motoring
•    Summary motoring
•    Not known
•    Other

For 2016–24, data is available for all case types broken down by all offence groups, e.g. a triable-either-way for-trial case for a drug offence, a committal for sentence for a theft offence etc. For 2010–15, committals for sentence and appeals are not broken down by offence groups.

There are an extremely small number of receipts of ‘unknown’ case type (25 in 2024 and 0 in most years) or of ‘other’ offence group (under 100 in 2024). Summary for-trial case numbers are also small but are present in the data, despite these not typically being eligible for a crown court trial (c.300 receipts in 2024). For-trial cases when the offence group is recorded as ‘not known’ are now rare (c.450 receipts in 2024) but were previously much more common (c.2,800 in 2016). ² Institute for Government analysis of Ministry of Justice, ‘Criminal court statistics quarterly: April to June 2025’ (Crown court receipts, disposals and open cases tool), 2025.

Generating a complexity weighting

We created a complexity weighting for each case sub-type based on the average number of hearing hours they take to be closed. We did this by calculating a weighted average hearing time for each sub-type across 2016–19 and 2023–24. ‘This includes all cases closed, whether they end up in a trial, guilty plea or the case being dropped. We then divided that average by the weighted average for all cases to produce the complexity weighting.

Complexity weighting = [weighted average hearing hours for sub-type across 2016–19 and 2023–24]/[weighted average hearing hours for all cases across 2016–19 and 2023–24]

We chose these years as 2016 is the first year for which revised data is available and we wanted to draw on data across a range of years, while excluding years during the Covid-19 pandemic, which seriously affected the kinds of cases courts were dealing with.

Case type and offence group are not the only determinants of how long a case takes to deal with: a large number of factors like the number of defendants, volume of evidence, age and vulnerability of witnesses or defendants and complexity of legal issues at play may all have a big impact. But case type and offence group do have a big impact on hearing time, especially case type: for example, the average indictable-only for-trial case took 8.7 hours to deal with (across 2016–19 and 2023–24) while the average committal for sentence took just 46 minutes. ³ Institute for Government analysis of Ministry of Justice, ‘Criminal court statistics quarterly: April to June 2025’ (Waiting and hearing time tool), 2025.  This makes them reasonable proxies for overall complexity. Figures CJS 5.1 and 5.2 set out the complexity weights of cases by case type and of for-trial cases specifically by offence group. There is minimal variation across offence groups for committals for sentence or appeals.

Since data on committals for sentence and appeals is not broken down by offence groups for 2010–15, the average complexity weighting for the offence group is used instead. This has a minimal impact on overall estimates.

The complexity weighting is fixed and does not change across years. This means it does not capture changes in complexity within a given sub-type over time. For example, the average hearing time for a triable-either-way sexual offence for-trial case has gone from 7.8 hours in 2016 to 6.4 in 2024. For the same reason, it does not capture any changes in complexity that affect all cases equally. Instead, it reflects how the overall make-up of cases has changed. However, there is not much evidence that there have been significant changes to case complexity within case sub-types, with changes in average hearing hours over the period generally fairly small.

How a defendant pleads in a for-trial case has the biggest impact on how long it takes to deal with, but we did not adjust for changing plea rates for two reasons. Firstly, plea rates are only known once a case has been completed, so cannot be used to adjust receipts. Secondly, plea behaviour is also affected by how the court system is operating, and so should not be thought of as an indicator of the level of demand entering the system, which is what we were principally interested in.

Finally, it is important to note that the complexity weighting only accounts for hearing hours in court, and not how long it takes to review the evidence, build a case file or otherwise prepare for trial. While there is likely to be a strong relationship between time in court and preparation out of court, some cases are likely to involve disproportionately more time for lawyers outside the courtroom – such as domestic abuse or sexual offences that involve applications for special measures for victims.^* If these cases are increasingly common, that would increase the overall workload for criminal lawyers and may have knock-on implications for courts.

_{*    Special measures are additional support or protection for victims or witnesses testifying in court, such as recording a cross-examination in advance or giving evidence via video link. The CPS has to apply to the court for permission to use special measures, which are then decided on by the judge. See Victim and Witness Information, ‘Special measures’, UK Government, no date, retrieved 16 October 2025, https:// victimandwitnessinformation.org.uk/the-court-process/special-measures}

Weighting cases by complexity

We created a complexity-adjusted value for receipts and disposals for each sub-type by multiplying the actual number of receipts and disposals by the complexity weighting for that sub-type. We then summed these sub-types to create complexity-adjusted totals for each case type and an overall total for all cases. Finally, we adjusted the complexity- adjusted number of cases to equal the actual number of cases in 2018, to provide an anchor point.

We chose 2018 because it was a year in the middle of our data series unaffected by the pandemic, and in which the average case complexity weighting was already close to zero, so the required adjustment was fairly small. However, the specific index year selected has no impact on the trends in the complexity-adjusted figures.

Complexity-adjusted outstanding cases in the crown court (Figure CJS 3.28)

We used a different method to model the complexity of outstanding cases. This is because jury trials were disproportionately likely to be delayed during the pandemic, so outstanding cases (cases in the ‘backlog’) are more likely to require jury trials than the average case. Jury trials are more complex and require much more hearing time in courts, so adjusting for this allows us to better compare the current backlog with pre- pandemic levels.

The number of outstanding cases (the ‘backlog’) is the cumulative difference between cases entering the court system (receipts) and cases completed (‘disposals’) each quarter. We adjust for the greater complexity of the backlog in several stages:

•    We calculate a pre-pandemic baseline for the proportion of case receipts that result in a jury trial by dividing the number of trials by the total number of disposals. This is based on data from Q2 2019 to Q4 2019 and we include both effective trials (those that go ahead as planned and result in a verdict) and cracked trials (those that are cancelled on the day of the listing and not rescheduled, for example because of a late guilty plea or discontinued prosecution).

•    We assume that the proportion of cases coming into the crown court that end up as jury trials is the same as before the pandemic and use this to estimate the number of jury trial receipts. We then subtract the number of jury trial disposals to estimate the number of outstanding jury trials. We use the same process to calculate outstanding non-jury trial cases, by subtracting other disposals from estimated non-jury trial receipts.

•    We calculate a complexity value for jury trial cases by dividing [trial cases share of total hearing time] by [trial cases share of total cases]. We multiply the number of outstanding cases by this complexity value. We repeat this process for non- jury trial cases, calculating a separate complexity value by dividing [non-jury trial cases share of total hearing time] by [non-jury trial cases share of total cases] and multiplying it by the number of outstanding non-jury trial cases. We then add them together for a total complexity-adjusted increase in the backlog.

•    Even before the pandemic, outstanding cases were more complex than the average of all cases processed. We adjust for this to make our backlog consistent with the pre-Covid backlog. We calculate a complexity value for pre-pandemic backlog cases by dividing [average hearing time of pre-pandemic backlog case mix] by [average hearing time of all pre-pandemic cases]. We multiply this by the complexity- adjusted increase in the backlog for a comparable complexity-adjusted backlog figure.

Prisons

Spending on prisons projection (Figure CJS 4.1)

We have assumed that spending on prisons changes in line with the overall Ministry of Justice budget, minus spending specifically committed to courts and probation. This is calculated slightly differently for different years. In 2024/25, we use the HMCTS outturn figures and assume probation spending increases in line with the overall MoJ budget, which gives an estimated nominal year-on-year increase of 8.9% for prisons. For 2025/26, we assume both HMCTS and probation spending increase in line with the overall budget.

For subsequent years, we include specific spending commitments on courts and probation announced in the 2025 Spending Review. We assume that the £450 million additional courts spending and £700 m additional probation spending by 2028/29 are 90% RDEL and 10% CDEL, and that the RDEL increases are in three equal stages in 2026/27, 2027/28 and 2028/29. This causes small decreases in projected nominal prisons spending in 2027/28 and 2028/29.

We then deflate the figures to 2025/26 prices in line with the process set out above (see Public services spending, above).

Prison capacity and population (Figure CJS 4.4)

We use month-end figures for both actual and projected population and useable capacity. Total spaces is the operational capacity, the total number of prisoners that an establishment can hold taking into account control, security and the proper operation of the planned regime. Useable capacity is the operational capacity minus an ‘operating margin’ that reflects the constraints imposed by the need to appropriately accommodate different classes of prisoner (for example, by sex, age, conviction status, single cell risk assessment, geographic distribution and security category). Decent accommodation capacity is in-use Certified Normal Accommodation, set by HMPPS, and is the number of prisoners that can be held in non-crowded conditions.

In any given monthly data release, either the total operational capacity or useable capacity is given, along with the operating margin in the notes. In cases where the total operational capacity is published, the useable figure is calculated by subtracting the operating margin from the total, and vice versa.
 

Prison capacity and population projections (Figure CJS 4.6)

The 2023 and 2024 ‘SDS40’ population projections are the Ministry of Justice’s (MoJ) central estimate for prisoner numbers over the five years, starting in 2023 and 2024 respectively. These projections incorporate the recruitment of the additional 20,000 police officers and the estimated impact of other policies, including new offences and changes in agreed sentencing, such as the Police, Crime, Sentencing and Courts Act and the Release of Prisoners Order 2020. They do not account for emergency early release measures such as the End of Custody Supervised Licence scheme set up under the Sunak government, and only the change to the point of automatic release (SDS40) is factored in to the 2024 projection; it does not include further measures around home detention announced in December 2024.

The IfG population projection after new reforms accounts for changes to home detention and recall that came into effect in June 2025, and for the proposed reforms in the government’s Sentencing Bill (2025). The June 2025 changes are estimated to reduce demand on prisons by 1,400 places by early 2028, compared to the status quo when they were announced (which was roughly the 2024 SDS40 projection). The immediate drop in the prison population in June 2025 was around 800, but this was reversed over the next two months before a drop back to June levels in September. This makes it extremely hard to identify the impact of this measure; we therefore simply assume that 500 of the 1,400 spaces saved by 2028 will have been achieved by December 2025, but that otherwise the population increases in line with the 2024 SDS40 projection.

We estimate that reductions from changes in the Sentencing Bill and other proposals following the Independent Sentencing Review (such as expanding intensive supervision courts) begin to be felt from January 2026. This is an optimistic timeframe, given the current progress of the Sentencing Bill, but some measures that do not require primary legislation may begin to have an impact more quickly. The government’s impact assessment for the Sentencing Bill estimates it will save 7,500 spaces, relative to the status quo in September 2025, by 2028. The total reduction in expected demand is therefore 8,900 spaces, which we assume to have been realised in full by April 2028. We assume that this will be evenly distributed between January 2026 and April 2028, resulting in a total estimated prison population of 89,600 by April 2028 (compared to 87,336 in September 2025).

The capacity forecast is based on the MoJ projection from December 2024. This is based on plans to expand the estate, including applying a “reasonable contingency to deliver timelines”, and also includes cells expected to be unavailable for maintenance work or lost permanently to dilapidations, and an operating margin to run prisons safely. See Ministry of Justice, Annual statement on prison capacity, 2024 for more detail.