Police in England and Wales conducted 586,591 stop and searches in the year ending 31 March 2026. The Home Office’s annual statistics, published on 1 October, put this 10.9%, or 57,494, above the 529,097 searches recorded in the year ending March 2025. The total combines searches under several statutory powers. Searches under section 1 of the Police and Criminal Evidence Act and associated legislation reached 580,068, also up 10.9% in a year. Yet the latest total remains well below the peak of 1,519,561 searches under all legislation in the year ending March 2009: the 2025 figure was 61.4% lower than that peak. The latest increase therefore describes a one-year change, not the whole longer-run pattern in police use of this power.

A rise in the number of searches is distinct from the narrower measure of what happened afterwards. Arrests following searches under all legislation increased by 9,202 to 88,944, and 15.2% of searches resulted in an arrest. The proportion was 15.1% the previous year and remains below the 17.1% peak recorded in the years ending March 2017 and March 2018. The arrest rate for section 1 PACE searches was 15.3%. Not every search is expected to result in arrest, and this ratio alone cannot establish the overall effectiveness of police powers. The outcome table records 64.3% of searches as requiring “no further action”. But since 2021, voluntary attendance, verbal warnings, seizure of property and other actions have been recorded in new categories; previously they were included under no further action. If those new categories were grouped as before, the 2026 share would be 71.7%.

The annual arrest total is different from arrests that followed searches: it covers arrests for notifiable offences, the categories counted in the official collection. Because Derbyshire could not provide data, the comparable total across 42 forces was 739,703, up 2.1% from 724,275 a year earlier. Violence against the person offences accounted for 45.7% of arrests. Arrests of people aged 21 and over increased 3%; those of children aged 10 to 17 fell 5.2%, while arrests of people aged 18 to 20 were nearly unchanged. The count is not a unique-person total for the year: it records arrest occasions under the collection rules. When several linked offences are involved in one occasion, generally the most serious offence is selected as the primary record. The Home Office stresses that arrest totals reflect police activity and priorities and do not indicate the overall level of crime committed.

Detentions under section 136 of the Mental Health Act, which applies in public places, totalled 32,465 in the year ending March 2026. Excluding Cumbria Police, which could not submit data for this year, the count was 2.7% higher than a year earlier. The power allows police to take a person from a public place to a place of safety when the officer considers that the person appears to have a mental disorder and needs immediate care or control; this is not the same as a criminal charge. Where the destination was known, 52.1% were taken to an accident and emergency department, 45.3% to a health-based place of safety and 2.5% to a police station. Section 135 concerns different circumstances, including entry to private premises with a warrant, and the quality guide rates its data as low quality and “Official Statistics in development”. That status should not be extended to the section 136 total.

The Home Office describes stop-and-search statistics as Accredited Official Statistics following regulatory review; arrests and section 136 detention statistics also carry accredited status. New or partial measures do not all share that status: data on force used, clothing removal, body-worn video and some linked outcomes are published as “Official Statistics in development” and require caution because of incomplete returns. Ethnicity information is also limited: self-defined ethnicity was unknown in 18% of search records in the year to March 2026. For some analysis, the Home Office combines self-defined ethnicity with officer-observed ethnicity, and readers need to keep that combined measure distinct from self-definition alone. The data come from force administrative records. Paper records, omissions and delays in uploading incidents can lead to undercounting in some forces. These annual figures are important official measures of recorded police activity, but they cannot by themselves establish every aspect of police practice or the level of crime.