Diversity, Equity and Inclusion in the UK

Section 1

Bottom line up front

Both propositions are partly true, and they do not conflict

The two propositions in the question are usually treated as alternatives. The evidence says they are both partly true, and they are not in conflict.

1. Measurable disadvantage is real and has not gone away. Field-experimental evidence — the strongest method available for isolating discrimination — still finds substantial hiring discrimination against Black British and Pakistani applicants, with no detectable decline since the 1960s. Disabled people face a 29.5 percentage-point employment gap. These are not artefacts of framing.
2. The popular framing of "white men at the top, everyone else below" is not what the pay data shows. White British is not the highest-earning ethnic group in the UK. Several minority groups out-earn white British employees on raw pay. The picture only shows minority disadvantage once you control for occupation, qualifications, geography and age — and even then it is specific groups, not all minorities.
3. Documented, adjudicated cases of unlawful discrimination against white men exist in UK public bodies. The RAF, Cheshire Police and Thames Valley Police have all been found to have crossed the line. These are not rumours; they are tribunal findings and an official MoD review. They are, however, a small number of cases relative to the workforce, and in each the law found against the employer — which is evidence the legal framework is working, not that it is broken.
4. The largest low-performing demographic group in English education is white working-class children, and the DEI apparatus has systematically under-served them. The government's own expert panel noted in 2024 that this is "rarely a priority in D&I interventions". This is the strongest version of the "overstepped" argument, and it is a class argument rather than a race one.
5. The evidence base for most DEI interventions is weak. This is not a contrarian claim — it is the finding of the UK government's own Inclusion at Work Panel, whose chair concluded that diversity training in particular has repeatedly failed at its stated objectives.

Section 2

The case that DEI remains necessary

Evidence of persistent, measurable disadvantage

2.1 Hiring discrimination is measurable and persistent

The Nuffield College/GEMM field experiment sent applications to nearly 3,200 real advertised jobs between 2016 and 2017, randomly varying only the ethnic background of otherwise identical fictitious candidates. Minority-background applicants had to send 60% more applications than white British applicants to get a positive response.¹

The accompanying meta-analysis of British field experiments from 1969 to 2017 found no significant decline in discrimination against Black British and Pakistani applicants over roughly half a century.¹²

This matters because it is the one method that removes the usual objection to pay-gap statistics — that gaps reflect choices, sectors and qualifications rather than treatment. Here, qualifications are held identical by design.

2.2 Adjusted pay gaps

The raw picture is genuinely mixed (see 3.1). Once ONS controls for occupation, qualification, geography, age and sex, the picture changes:

The reversal is the substantive finding. Asian and Black UK-born employees earn more in raw terms largely because they are more concentrated in London and in professional occupations, and are better qualified — 63.9% of Asian employees hold a degree against 40.9% of White employees; 47.3% of Black employees live in London against 10.4% of White employees. Strip those advantages out and a like-for-like penalty appears.³

2.3 Disability

This is the least contested part of the evidence and the largest gap by some distance.

• Disabled employment rate: 52.8%; non-disabled: 82.3%. Gap: 29.5 percentage points (ONS Labour Force Survey, October–December 2025). The gap has widened by 2.0 points since mid-2024, with 218,000 fewer disabled people in work year-on-year.⁴
• Disability pay gap: 12.7% (ONS, 2023). Widest for autistic employees (27.9%) and those with epilepsy (26.9%).⁵

2.4 Sex

• Gender pay gap, April 2025: 6.9% full-time, 12.8% all employees, −2.9% part-time (women earn more part-time). Median full-time hourly earnings excluding overtime: £20.27 for men, £18.87 for women.⁶
• The gap is now near-zero for those in their twenties (0.9% at 22–29) and widens with age to 12.5% at 50–59 — a pattern consistent with career interruption around childcare rather than unequal pay for the same work at a single point in time.⁶

The gender pay gap has fallen by more than a quarter over a decade. That is real progress, and it is worth noting that this is one of the few areas where a specific DEI intervention — mandatory reporting since 2017 — coincides with sustained improvement, though causation is not established.

2.5 Representation at the top

• FTSE 100: 14 ethnic minority CEOs in 2025 — the highest ever recorded, but none of them women. Ethnic minorities hold 20% of FTSE 100 board seats but only 11% of UK senior management. Black representation declined slightly at both levels in 2025.⁷
• Nine FTSE 100 CEOs were women in 2025 (21 across the FTSE 350).⁸

Whatever the merits of the "white men are penalised" argument at the level of individual selection decisions, it is not visible in who actually holds the senior positions.

Section 3

The case that DEI has overstepped

Evidence of misapplication, weak efficacy and blind spots

3.1 White British is not the top-earning group

This is the single most under-reported statistic in the debate. Using the 18-category ethnicity breakdown for England and Wales (2022):

• White Irish employees had the highest median gross hourly pay at £20.20, against £14.42 for White British — a pay gap of −40.1% in favour of White Irish. For men specifically the gap was −50.7%.³
• Chinese and Indian employees also out-earned White British employees.³
• The lowest median earnings of any group were Mixed White and Black Caribbean (£11.75, an 18.5% gap), followed by Bangladeshi and Pakistani employees.³

The five-category "White" grouping used in most reporting conceals this by folding White Irish, White Other and Gypsy/Traveller groups in with White British. ONS says so explicitly. Any DEI framework built on a simple white/non-white binary is therefore working with a category that its own statistical authority describes as inadequate for measuring the thing it claims to measure.

3.2 White working-class educational outcomes

The Independent Inquiry into White Working Class Educational Outcomes, chaired by Sir Hamid Patel and Baroness Estelle Morris and published June 2026, is the most extensive research on this in recent years. Its findings:⁹

• 36% of white British pupils on free school meals achieved grade 4+ in English and Maths GCSE in 2025, against 72% of non-FSM pupils.
• 48% of white working-class children reach a good level of development by age five, against 75% of white British middle-class children.
• White working-class pupils miss 13% of school sessions against 7% for all pupils; they are twice as likely to be persistently absent and two-and-a-half times as likely to be severely absent.
• 52% of white working-class pupils thought they were likely to go to university, against 82% of peers.
• The Inquiry describes this group as "the largest low-performing demographic in the school system" and found "discomfort about how to engage with the issue", no shared definition, and a lack of robust data.

Comparing like with like on free school meals (Centre for Social Justice analysis of DfE data, 2026): 36% of white British FSM boys reached the expected GCSE standard in maths and English, against 39% of Black Caribbean FSM boys, 58% of Black African, 68% of Bangladeshi and 82% of Chinese FSM boys.¹⁰

Higher education is starker still. Just 14% of white British boys eligible for free school meals had started higher education by age 19 in 2023/24. More than half of all Black and Asian FSM-eligible pupils had done so, against 19% of White FSM pupils overall.¹¹

House of Commons Library analysis is unambiguous: "White males eligible for free school meals are less likely to go to higher education than any other group when analysed by gender, free school meal eligibility and broad ethnic group." White males not eligible for free meals are also below average.¹²

The sex gap is separate and also real: for every 100 young women entering higher education aged 19 or under in England, roughly 79 young men do; the shortfall was 35,475 in the 2024 cycle, up from 29,285 in 2015.¹³

3.3 Unlawful positive discrimination: the adjudicated cases

UK law permits positive action and prohibits positive discrimination. Under the Equality Act 2010:

• s.13 — treating someone less favourably because of a protected characteristic is direct discrimination and cannot be justified (except for age).
• s.158 — an employer may take proportionate action to address disadvantage or under-representation (outreach, targeted training, mentoring).
• s.159 — a protected characteristic may be taken into account in recruitment or promotion only as a tie-break between candidates who are "as qualified as" each other, only where under-representation or disadvantage is reasonably believed to exist, and only as a proportionate means. Quotas, reserved posts and preferring a less-qualified candidate remain unlawful.¹⁴¹⁵

Three cases where employers crossed that line:

• Furlong v Chief Constable of Cheshire Police (2019) — a white, heterosexual male applicant was discriminated against by a misapplication of the positive action provisions. Cheshire had treated all candidates passing an interview as "equally qualified" so as to apply a tie-break at scale; the tribunal rejected this.¹⁶
• RAF (2023) — an MoD review found that pressure to hit recruitment targets for women and ethnic minorities had led to unlawful positive discrimination against white men.¹⁷
• Turner-Robson, Bishop and Horton v Chief Constable of Thames Valley Police (2024) — three white British officers with 19 to 26 years' service were blocked from applying for a detective inspector post after a superintendent, instructed to "make it happen", moved a minority-ethnic sergeant into the role with no advertisement and no competitive process. The sergeant had not yet been promoted to inspector. The tribunal found direct race discrimination under s.39(2), not protected by s.158 or s.159, and expressed "astonishment" that some staff had received no equality training in 21 years. A subsequent independent review found the episode had produced "a divided workforce".¹⁸¹⁹²⁰

3.4 The government's own verdict on DEI practice

The Inclusion at Work Panel, commissioned by the Minister for Women and Equalities and reporting in March 2024, is the most authoritative UK assessment. Its conclusions were markedly sceptical:²¹

• Definitions of diversity, equity and inclusion are "contested and can even be — legitimately — mutually exclusive". Equality (impartial, forward-looking) and equity (corrective, backward-looking) are in tension.
• Judgements about relative group advantage are "complex and likely beyond the capability of a corporate HR team".
• On diversity training specifically, it cites a body of evidence finding an absence of durable improvement in culture, failure to increase collaboration, and failure to improve hiring, retention or promotion of diverse candidates. A Harvard study of 800 US firms over three decades found no increase in minority representation in management five years after training became mandatory; representation of Black women and Asian-Americans actually fell.
• Only 38% of UK employers collect any equal opportunities monitoring data; only 25% consult data before planning new diversity activity; one in four describe their approach as entirely or mostly reactive, "in response to societal events".
• One in four employers with a DEI strategy do not evaluate its effectiveness at all.
• Some employers, "in an attempt to go 'above and beyond' the law… were actually breaking the law".
• Participants reported that definitions of bullying, harassment and discrimination are "becoming weaponised" in grievances, and described a censorious environment in which candid discussion of DEI efficacy was discouraged.
• Critically: both public and private sector representatives noted that data showing white working-class populations were often more disadvantaged than ethnic minority groups is "rarely a priority in D&I interventions".

The Panel recommended against creating a new inclusion accreditation scheme and instead proposed an Education Endowment Foundation-style evidence toolkit rating interventions by cost, impact and evidence strength.

3.5 Cost

• An FOI-based estimate cited by the Panel puts public sector DEI employment at roughly 10,000 jobs costing £557 million a year. Provenance matters: this figure originates with Conservative Way Forward, a campaign group, and should be treated as an advocacy estimate rather than an official statistic.²¹
• An estimated £13 million a year is spent on DEI roles in the NHS (reported November 2023).²¹
• A Treasury review found approximately 400 full-time-equivalent EDI roles within the civil service specifically, prompting a decision to fold them into HR and require ministerial sign-off for external EDI spending.²²

The wide gap between the 400 FTE (civil service) and 10,000 (all public authorities) figures reflects different scopes, not a contradiction, but the larger figure is the less reliable one.

3.6 Belief discrimination

Two rulings have established that the DEI framework itself cannot be imposed as an orthodoxy:

• Forstater v CGD Europe (2021) — gender-critical belief is protected under the Equality Act; over £100,000 awarded.²¹
• In 2023, an employment judge ruled that not subscribing to critical race theory is a protected belief under the religion-or-belief provisions.²¹

Section 4

Where opinion actually sits

The public splits the same way the evidence does

Ipsos, April 2025 (n=1,200, GB adults 16–75)²³

• Support for specific measures: flexible working 71%, gender pay gap reporting 65%, inclusivity training 64%.
• On DEI in general: 36% say it has gone too far, 31% about right, 19% more needed.
• Men 43% against women 29% say too far; the view is more common among over-55s and white respondents.

YouGov, 2025²⁴

• 25% say efforts on jobs access for ethnic minorities have gone too far; 26% say they have not gone far enough; 35% about right. Among ethnic minority Britons, 50% say more is needed and 8% say too far.
• 51% say more needs to be done for disabled people; only 6% say too much.
• 69% think disabled people face significant jobs discrimination; 62% say the same of older people.

The pattern is consistent: people support specific, procedural measures and are sceptical of the branded abstraction. That is close to what the evidence supports.

Section 5

Where policy is heading

The UK is moving opposite to the United States

• 25 March 2026: the government published its consultation response confirming mandatory ethnicity and disability pay gap reporting for employers with 250 or more staff, with draft primary legislation annexed. It will be delivered via the Equality (Race and Disability) Bill, promised in the July 2024 King's Speech and still awaited. No implementation date has been set. The Bill may also extend equal pay claims to ethnicity and disability and create an Equal Pay Regulatory and Enforcement Unit.²⁵²⁶
• April 2027: under the Employment Rights Act 2025, employers with 250 or more staff must publish an equality action plan on the gender pay gap; the government is encouraging voluntary publication before then. Reporting for sex, race and disability is intended to be harmonised into a single plan.⁶
• Meanwhile, an Institute of Directors survey of 605 UK business leaders (March 2025) found 71% do not intend to change their approach to DEI and only 11% anticipate scaling back.²⁷

The practical implication for employers is the reverse of a rollback: more measurement obligations, not fewer.

Section 6

Assessment

Answering the question directly

Is DEI still needed?

If "DEI" means enforcement of anti-discrimination law and attention to whether selection processes are fair and transparent — yes, on the evidence. Identical CVs still produce a 60% callback penalty for some minority applicants, with no improvement in half a century. A 29.5-point disability employment gap is not a statistical artefact. These are not solved problems.

Has it overstepped?

In three specific senses, yes, and each is documented rather than asserted:

1. Legally, in identifiable cases where public bodies substituted representation targets for competitive process and were found to have directly discriminated against white men.
2. Methodologically, in that the dominant interventions — particularly diversity training — have a weak-to-negative evidence base by the government's own assessment, are rarely evaluated, and are frequently adopted reactively rather than diagnostically.
3. In coverage, in that a framework organised around race and sex has largely missed the group with the worst outcomes in the education system. White working-class boys are the clearest case of severe, measurable, sustained disadvantage in modern Britain, and the DEI apparatus has not been the mechanism that identified or addressed it.

Does it penalise white British males as a class?

On the population-level evidence, no. White British men are not a low-earning group, are heavily represented in senior positions, and White British is not even the top-earning ethnic category in Britain — meaning the group is neither uniquely advantaged nor uniquely penalised in the way either side of the argument tends to claim.

But white working-class males are a different proposition from white British males. On education, higher education entry and early-years development, they are the worst-served large group in the country, and the deficit is severe. Conflating the two categories is the central error in this debate — made by advocates who treat "white male" as a synonym for advantage, and by critics who generalise a genuine class problem into a claim about an entire ethnic and sex group.

Section 7

Contested points and limitations

In the interests of not overstating the case

• The Sewell report (Commission on Race and Ethnic Disparities, 2021), which concluded Britain is not institutionally racist and that geography, family, socio-economic background, culture and religion matter more than racism for life chances, was fully endorsed by the then government but strongly contested. The Runnymede Trust, the Centre on Dynamics of Ethnicity and academic critics rejected its analysis, and the UN Working Group of Experts on People of African Descent condemned it. It should not be cited as settled.²⁸²⁹
• Adjusted pay gaps are contested in principle. Controlling for occupation removes discrimination that operates through occupational access. Whether adjustment reveals or conceals discrimination depends on assumptions the statistics cannot resolve.
• ONS ethnicity pay data are from 2022 and carry explicit sample-size warnings. There is no more recent comparable release.
• The £557m public sector DEI cost figure is an advocacy estimate, not official statistics.
• The Inclusion at Work Panel was commissioned under the previous government and its chair had a known critical stance on DEI orthodoxy. Its evidence citations are sound and largely academic; its framing is not neutral.
• Employment tribunal first-instance decisions (Turner-Robson, Furlong) are not binding precedent, though they are correct applications of settled law.
• The white working-class Inquiry was commissioned by Star Academies and funded by a private trust, though its panel spanned union leaders, former education secretaries of both parties and DfE non-executives, and its data analysis was conducted by UCL.

Section 8

References

Sources in order of first citation

1. 1Di Stasio, V. and Heath, A. (2019), Are employers in Britain discriminating against ethnic minorities?, Centre for Social Investigation, Nuffield College, Oxford. nuffield.ox.ac.uk
2. 2Heath, A. and Di Stasio, V. (2019), 'Racial discrimination in Britain, 1969–2017: A meta-analysis of field experiments on racial discrimination in the British labour market', British Journal of Sociology 70(5): 1774–1798.
3. 3Office for National Statistics (29 November 2023), Ethnicity pay gaps, UK: 2012 to 2022. ons.gov.uk
4. 4Work Foundation, Lancaster University, analysis of ONS Labour Force Survey A08 (October–December 2025). lancaster.ac.uk
5. 5Office for National Statistics (17 October 2024), Disability pay gaps in the UK: 2014 to 2023. ons.gov.uk
6. 6Office for National Statistics, Gender pay gap in the UK: 2025 (ASHE, April 2025 provisional); House of Commons Library briefing SN07068. commonslibrary.parliament.uk
7. 7Parker Review Committee, annual report, March 2026 (data as at 31 December 2025).
8. 8Statista, Number of female CEO positions in FTSE companies, UK, 2025.
9. 9Independent Inquiry into White Working Class Educational Outcomes, final report, June 2026. educationaloutcomes.org.uk — summary at publicfirst.co.uk
10. 10Centre for Social Justice, analysis of DfE attainment data, 2026. centreforsocialjustice.org.uk
11. 11House of Commons Library, Higher education student numbers (CBP-7857), citing DfE Widening participation in higher education 2023/24. commonslibrary.parliament.uk
12. 12House of Commons Library, Equality of access and outcomes in higher education in England (CBP-9195). commonslibrary.parliament.uk
13. 13UCAS end-of-cycle 2024 data, England, 19 and under.
14. 14Equality Act 2010, ss.13, 158, 159. legislation.gov.uk
15. 15House of Commons Library, The Equality Act 2010 and positive action (SN06093). commonslibrary.parliament.uk
16. 16Mr M Furlong v The Chief Constable of Cheshire Police, ET 2405577/2018. gov.uk
17. 17BBC News (2023), 'RAF diversity targets discriminated against white men'. bbc.co.uk
18. 18Turner-Robson, Bishop and Horton v Chief Constable of Thames Valley Police (2024), Employment Tribunal.
19. 19Field Seymour Parkes, 'Positive action or positive discrimination?' fsp-law.com
20. 20Personnel Today (2025), 'Thames Valley Police "positive discrimination" led to "divided workforce"'. personneltoday.com
21. 21Inclusion at Work Panel (20 March 2024), Report on recommendations for improving diversity and inclusion (D&I) practice in the workplace, Equality Hub / Cabinet Office. gov.uk
22. 22HM Treasury (2023), 'End to Civil Service expansion and review of equality and diversity spending announced in productivity drive'.
23. 23Ipsos UK (April 2025), polling of 1,200 GB adults aged 16–75. ipsos.com
24. 24YouGov (2025), 'Do Britons think DEI initiatives have gone too far in the UK?' yougov.com
25. 25Cabinet Office (March 2026), Equality (Race and Disability) Bill: mandatory ethnicity and disability pay gap reporting — government response, CP 1542. gov.uk
26. 26Littler (April 2026), 'UK Government Confirms Commitment to Introduce Mandatory Ethnicity and Disability Pay Gap Reporting for Large Employers'.
27. 27Institute of Directors survey of 605 UK business leaders, March 2025.
28. 28Commission on Race and Ethnic Disparities (31 March 2021), The Report.
29. 29UN Working Group of Experts on People of African Descent (19 April 2021), statement on the CRED report. ohchr.org

Changes

1. 1.0September 2026First published.

How to cite

Allen, P. (2026). Diversity, Equity and Inclusion in the UK (version 1.0). Occam’s Shaving Kit. https://occamsshavingkit.uk/papers/dei-in-the-uk/