National picture

The Relative Rate Index compares the rate at which an ethnic group experiences a decision with the rate for White children. A value of one is parity; above one, the group is more likely to experience the decision. The cascade follows children through four decision points, from the first police contact to the court.

Black children are about 6% of the 10 to 17 population of England and Wales, and about 26% of children remanded to youth detention accommodation in the year ending March 2025. The bars below set each ethnic group's share of the child population against its share of custodial remand.

Not every child subject to a remand decision is held. Most are bailed; the rest are placed on community remand, remanded to local authority accommodation, or remanded to custody. The bars below show how remand decisions divided for each ethnic group in the year ending March 2025. The custodial remand share, the darkest segment, is larger for Black children and children of Mixed Heritage than for White children.

The cascade's first stage varies sharply by place. Across the 40 police force areas shown, the stop-and-search Relative Rate Index for Black children runs from 0.98 to 6.68, with a median of 2.38. The highest values are in forces with small Black child populations, where a modest number of searches produces a high rate against a small denominator; those values are real but less stable than the large-force figures. In 2 forces the Black search count falls below the disclosure threshold, so no value is shown.

The remand population is the product of how many children enter remand and how long each remand lasts. A 25% reduction in the stock can therefore be achieved by remanding fewer children, by shortening remands, or both; and a reduction achieved by shortening is assessed for whether the shortening is equal across ethnic groups. The published episode data shows it currently is not. In every year since 2019, remand episodes for children from ethnic minority groups have lasted substantially longer at the median than those for White children: 74 nights against 40.5 in the year ending March 2025, and 70 against 48 in the provisional 2025-26 year.

The road ends in custody, and the only published account of what custody is like for remanded children as a group is HM Inspectorate of Prisons' annual survey. In 2025-26, 315 children in the four young offender institutions and Oakhill secure training centre completed its questionnaire, 86% of those offered one, at five inspections between April 2025 and March 2026. Of the 312 who answered the question, 128, or 41%, were on remand or waiting to be sentenced. The table sets that group against the sentenced children on the survey's own questions. The figures are HMIP's, quoted as published; they are not PRISM-R analysis.

Three differences are statistically significant on HMIP's test, and all three are about what is planned for a child whose stay may end at the next hearing: remanded children were less likely to say somebody was helping them prepare to leave, less likely to find a nurse easy to see, and less likely to be able to speak to a chaplain in private. That is the same finding the inspectorate's short-term remand thematic reached from casework, arrived at here from the children's own answers.

The survey also records the custody end of the disproportionality this page traces. Children from an ethnic group other than White were 63% of respondents in 2025-26, against 33% when HMIP first asked in 2001, and Muslim children 38% against 21% in 2019-20. The inspectorate's own reading is that diversion from custody has worked better for some groups of children than for others. On the survey's binary ethnicity split, children from other ethnic groups were more likely to report having been kept locked up as a punishment, 61% against 47%, and threats or intimidation from staff, 31% against 18%, and less likely to think the rewards system fair, 30% against 44%; HMIP's test does not flag those three, so PRISM-R reports them as differences in the survey, not as findings.

PRISM-R quotes these figures and does not ingest them. The survey is self-reported, and it describes respondents at five survey dates rather than the custody population; "on remand or waiting to be sentenced" includes convicted children awaiting sentence and 18-year-olds, so its 41% is not the remand share in the Youth Custody Service data the target tracker follows, and the two are not reconciled. It carries no home geography and no community denominator, so nothing in it can enter the Relative Rate Index or the geographic explorer.