Last Updated on September 23, 2026 by Karl Thompson
Three official inquiries have now found institutional racism in British policing: Macpherson (1999), the Lammy Review (2017) and the Casey Review (2023). That is the state’s own conclusion about itself, three times, across twenty-four years.
The evidence and the argument, in short:
- Black people are stopped and searched at 3.8 times the rate of White people — and searches of Black people find something no more often than searches of White people.
- Black people are 4% of the population and 12% of the prison population. Among prisoners under 18, they are 25%.
- The disparity is not uniform. Asian people are searched at 1.4 times the White rate, and custody rates barely differ by ethnicity at all.
- It has also fallen sharply — the stop and search gap was 9.7 times in 2018/19.
Points 3 and 4 are why this is a genuine debate rather than a settled question, and why an essay that only presents the case for institutional racism will not reach the top band.
This post covers policing and the criminal justice system within the crime and deviance module, AQA Paper 3 (7192/3).
What “institutional racism” means
The term has a precise definition in Britain, and students lose marks by using it loosely.
Macpherson (1999), reporting on the Metropolitan Police investigation into the murder of Stephen Lawrence, defined institutional racism as the collective failure of an organisation to provide an appropriate and professional service to people because of their colour, culture or ethnic origin.
Two features of that definition matter.
It is about outcomes, not intentions. An institution can be institutionally racist without any individual in it holding racist views. The failure is collective, and it can operate through procedures, attitudes and behaviour that nobody has consciously chosen.
It includes unwitting prejudice and thoughtlessness, as well as ignorance and stereotyping. “Nobody meant to” is not a defence against the charge — it is part of what the charge describes.
This is why the argument that most officers are not personally racist misses the point. That can be entirely true and the finding still stand.
The evidence: stop and search
Stop and search is where the disparity is sharpest, and it is a power exercised on an officer’s suspicion rather than on evidence.
Home Office, Police powers and procedures, year ending 31 March 2025 (published 6 November 2025).
| Ethnic group | Searches per 1,000 people, relative to White |
|---|---|
| Black | 3.8× |
| Mixed | 1.8× |
| Other | 1.5× |
| Asian | 1.4× |
There were 528,582 searches in the year — 65% below the 2009 peak of 1,519,561. Two-thirds ended in no further action, and 61% were for suspected drug possession.
Some detail that does the analytical work:
- Black males aged 18 to 34 were searched at 125 per 1,000 against 24 per 1,000 for White males of the same age.
- Disparity varies by reason: section 60 “no suspicion” searches 8.8×, offensive weapons 7.2×, drugs 4.6×, criminal damage only 1.3×.
- Ethnicity is not recorded for 18% of searches, and is missing more often for minority ethnic people. Using officer-observed ethnicity where self-defined is absent raises the Black figure to 4.9×.
The find rates
This is the single strongest piece of evidence available, and almost no candidate has it.
| Something found | Arrest followed | |
|---|---|---|
| Searches of Black people | 29.9% | 16.7% |
| Searches of White people | 28.4% | 14.5% |
Near-identical find rates alongside a near-fourfold difference in search rates. If the extra searches of Black people were being made on better grounds, they would turn something up more often. They do not.
Pair it with the section 60 point: the power requiring no reasonable suspicion produces the largest disparity (8.8×) and the lowest arrest rate (3.6%, against 15.0% for ordinary PACE searches). The least evidence-based power produces the most unequal outcome.
The evidence: the rest of the system
Ministry of Justice, Statistics on Ethnicity and the Criminal Justice System 2024 (published 27 November 2025).
| Stage | White | Black | Asian | Mixed |
|---|---|---|---|---|
| Population (2021 Census) | 82% | 4% | 9% | 3% |
| Stop and search (2024/25) | 70% | 13% | 11% | 4% |
| Arrests (2024/25) | 79% | 7% | 8% | 4% |
| Prosecutions, indictable (2024) | 83% | 7% | 6% | 3% |
| Prison population (30 June 2025) | 73% | 12% | 8% | 5% |
Read the Black column down. The over-representation is severe at the entrance to the system and at the exit, and much weaker in the middle.
Where the disparity is worst:
- Children. 59% of prisoners under 18 are from minority ethnic groups. Black prisoners are 25% of the under-18 prison population and 21% of 18–24 year olds, against 12% overall. By contrast 84% of prisoners aged 50+ are White.
- Remand. 23% of White defendants were remanded in custody at magistrates’ courts against 30% of Black defendants. Across all courts, Black defendants were 29% more likely to be remanded.
- Waiting times. Black defendants waited a median 24.9 weeks for Crown Court trial against 17.0 weeks for White defendants.
- Drug offences. Other ethnic groups 89%, Asian 58%, Black 44% and Mixed 26% more likely to receive custody than White offenders. The only offence category where the MoJ classes the disparity as both statistically and practically significant.
- In prison. 45% of Black male prisoners reported victimisation by staff against 30% of White prisoners (HMIP survey, 2024/25).
- The workforce. 91.5% of police officers were White in the year ending March 2025.
How it happens: Bowling and Phillips
Bowling and Phillips (2007) provide the mechanism, which is what an essay needs beyond the figures.
Their argument turns on discretion and low visibility. Police work, especially at street level, involves constant individual judgements about who looks suspicious — made quickly, alone, with no supervisor present and usually no record of the encounters that lead nowhere. Where discretion is wide and oversight is thin, stereotypes fill the gap.
They connect this to canteen culture: the informal occupational culture of policing, which they describe as suspicious, macho, socially conservative and racially prejudiced. Officers are recruited into it rather than arriving with it, which is precisely what makes the racism institutional rather than individual.
Their conclusion on stop and search is that the disproportion is the product of routine, low-visibility decision-making rather than deliberate targeting — and that this makes it harder to reform, not easier.
This is the paragraph that turns a list of statistics into sociology. Use it.
The three reports
Macpherson (1999)
The inquiry into the investigation of Stephen Lawrence’s murder. It found the Metropolitan Police institutionally racist, gave the definition above, and found that ethnic disparities in stop and search demonstrated racist stereotyping. It made 70 recommendations and reshaped British policing policy for a generation.
The Lammy Review (2017)
An independent review by David Lammy MP into outcomes for Black, Asian and minority ethnic people across the whole criminal justice system. 35 recommendations. It estimated the cost of over-representation at at least £309 million a year.
Its central mechanism is a trust deficit. BAME defendants plead not guilty more often because they do not expect fair treatment and prefer their chances with a jury to a magistrate — and a guilty plea attracts a discount of up to a third, so the lower plea rate feeds directly into longer sentences. In 2024, the Crown Court guilty plea rate was 63% for White defendants and 54% for Black defendants.
It also found the BAME share of young offenders in custody rose from 25% to 41% between 2006 and 2016, while the total number of young offenders fell to record lows — a disparity worsening while the overall problem shrank.
Lammy recommended the Relative Rate Index the MoJ now uses, and the principle “explain or reform”: where a disparity cannot be explained, it must be reformed.
The Casey Review (2023)
Baroness Casey’s review of the culture and standards of the Metropolitan Police, commissioned after the murder of Sarah Everard by a serving officer.
It found the Met institutionally racist, misogynistic and homophobic, and failing women and children. On policing specifically, it found the force could not explain clearly enough why it used stop and search on the scale that it did, and had not examined the effect of humiliating and traumatic searches on trust among young Black Londoners.
Commissioner Sir Mark Rowley accepted the substance of the findings but rejected the word “institutional”. That refusal is itself useful material — an organisation declining the description while accepting the facts is close to what Macpherson meant by collective failure.
The Met published a Stop and Search Charter in February 2025 in response.
The arguments against
A strong answer takes these seriously rather than listing them to dismiss them.
The wrong denominator. Waddington et al. (2004) argue that comparing searches to the resident population is a methodological error. What matters is the available population — who is actually on the street, in that area, at that time. Policing is concentrated in inner-city areas with younger, more ethnically diverse street populations. On the available-population measure, much of the disproportion disappears in some places.
The selective racism problem. Asian people are searched at 1.4 times the White rate against 3.8 for Black people. A racist police force ought to over-police both. This is left realism’s sharpest point and it has not gone away.
The middle of the system. The custody rate is 33% for White defendants and 35% for Black defendants. MoJ regression, controlling for offence, age, sex, plea and court, finds no statistically significant association between ethnicity and sentence length for indictable offences overall. A uniformly racist system should show its hand more consistently.
The MoJ’s own caveat. The department states explicitly that group-level differences cannot be read as evidence of bias, because offence mix, geography, age and prior record are not controlled in the headline figures. Using an official source’s own warning against its figures is a strong AO3 move.
It is getting better. The stop and search disparity fell from 9.7× in 2018/19 to 3.8× today. Whatever produced the 2018/19 figure has substantially changed.
Hood (1992). The classic sentencing study found Black defendants had a five percentage point greater probability of custody after controlling for legally relevant factors. Note this carefully: five percentage points, not “five times”. The “5×” version circulates widely in revision material and is wrong.
Evaluation
| ✓ The police are institutionally racist | ✗ Against that view |
|---|---|
| Three official inquiries have found it: Macpherson (1999), Lammy (2017), Casey (2023) | Waddington et al. (2004): the resident population is the wrong denominator for stop and search |
| Equal find rates alongside a 3.8× search rate — the extra searches detect no extra crime | The disparity is not uniform: Asian 1.4× against Black 3.8× |
| Section 60, the power needing no suspicion, has the highest disparity (8.8×) and lowest arrest rate (3.6%) | Custody rates barely differ (33% White, 35% Black) and regression finds no ethnicity effect on sentence length |
| Bowling and Phillips (2007) identify the mechanism: wide discretion, low visibility, canteen culture | The MoJ states its own figures cannot demonstrate bias |
| Black children are 25% of the under-18 prison population against 4% of the population | The gap has narrowed sharply — 9.7× to 3.8× since 2018/19 |
| Self-report studies find near-identical Black and White offending (Graham and Bowling, 1995) | Left realists argue part of the difference is real offending caused by deprivation |
| 91.5% of officers are White; 45% of Black male prisoners report victimisation by staff | Ethnicity is unknown for 18% of search records and a third of court records |
Has anything actually changed?
The most interesting question, and the one that separates a good answer from a very good one.
Yes, measurably. The search disparity has more than halved since 2018/19. Total searches are 65% below their 2009 peak. Section 60 use is far below its mid-2000s level.
But the deep pattern holds. Black children are still a quarter of the under-18 prison population. Remand disparity has not moved. Drug-offence sentencing disparity is large and significant across every minority group.
And the category may have relocated. Criminologists have argued that the “gang” now does the work that “mugging” did in the 1970s — a label applied overwhelmingly to young Black men, linked only weakly to serious violence, and used to justify database entries and police attention. Gilroy’s framework predicts exactly this: the category outlives the policy.
That is a strong closing line for an essay. Reform changed the numbers; it is less clear that it changed the thing producing them.
Exam application
A 10-mark question on criticisms of the police-racism explanation is common. The two best-evidenced are the selective racism point (Asian 1.4× against Black 3.8×) and the available-population argument (Waddington et al.). Both have current data behind them.
In a 30-mark essay, the structure that works is stage by stage: policing disparity large, court disparity small, prison disparity large again and worst for children. That shape is the argument — it tells you discrimination is concentrated where discretion is widest.
Date everything. “Home Office, year ending March 2025” earns more than “recent statistics show”. Examiners notice.
For a worked answer see this 30-mark essay on ethnic differences and the criminal justice system.
Crime and Deviance Revision Bundle
If you like this sort of thing, you might like my Crime and Deviance Revision Bundle, which contains 32 pages of revision notes, seven colour mind maps and 12 exam practice questions with exemplar answers, written for the AQA specification.
Better value is the A-Level Sociology Revision Mega Bundle.
Signposting and Related Posts
This post has been written for students studying crime and deviance as part of the AQA A-Level Sociology course, Paper 3 (7192/3).
For the figures in full, see official statistics on ethnicity and crime.
For the theory that the statistics are constructed rather than discovered, see Paul Gilroy and Policing the Crisis.
For the argument that some of the difference is real, see left realist explanations and cultural factors.
Institutional racism also appears in the education module — see are schools institutionally racist?
Please click here to return to the homepage – ReviseSociology.com
Sources
Home Office (2025) Police powers and procedures: stop and search, arrests and mental health detentions, England and Wales, year ending 31 March 2025.
Ministry of Justice (2025) Statistics on Ethnicity and the Criminal Justice System 2024.
Macpherson, W. (1999) The Stephen Lawrence Inquiry.
Lammy, D. (2017) The Lammy Review.
Casey, L. (2023) An Independent Review into the Standards of Behaviour and Internal Culture of the Metropolitan Police Service.
Bowling, B. and Phillips, C. (2007) ‘Disproportionate and Discriminatory: Reviewing the Evidence on Police Stop and Search’, Modern Law Review, 70(6).
Hood, R. (1992) Race and Sentencing. Oxford: Clarendon Press.
Waddington, P.A.J., Stenson, K. and Don, D. (2004) ‘In Proportion: Race, and Police Stop and Search’, British Journal of Criminology, 44(6).