Diversity, equity and inclusion at work

In short

Diversity is the mix of people in an organisation, equity is the design of systems so that the mix has fair access to pay, progression and opportunity, and inclusion is the day-to-day experience of being valued and able to speak up. The agenda is enforceable in India through the Constitution, the POSH Act 2013, the RPwD Act 2016, the Transgender Persons Act 2019 and the four Labour Codes in force from 21 November 2025, and it is disclosed publicly by listed companies through SEBI's BRSR. MBA students meet it in HRM, OB, Business Ethics/CSR and Strategic HRM papers, and in placement interviews at TCS, Infosys, Accenture, Deloitte, HUL, Tata Steel and Mahindra, all of which publish diversity targets. The examinable core is the conceptual distinction between equality and equity, the bias and structural causes of under-representation, the Indian statutory framework, and the metrics used to measure and report progress.

Explained

Concepts, dimensions and classifications

• Diversity - the presence of differences in a workforce along identity and experience dimensions. It is a *fact* about composition, measured as a ratio or an index.
• Equity - proportional fairness: removing structural disadvantage so that outcomes are not predicted by identity. Equality gives everyone the same input (one ramp-free entrance for all); equity gives each person what they need to reach the same starting line (a ramp, a translated test, a crèche).
• Inclusion - the degree to which people feel respected, heard and able to influence decisions. Diversity without inclusion produces hiring followed by attrition.
• Belonging - the fourth element in the DEIB formulation; inclusion is what the organisation does, belonging is what the individual feels.
• Accessibility - sometimes added as DEIA; built environment, digital systems and processes usable by persons with disabilities.
• Dimensions, standard classification:
  • Primary / surface-level (largely visible, mostly ascribed): sex and gender, age, race/ethnicity, disability, physical appearance.
  • Secondary / deep-level (acquired, less visible): education, religion, language, marital and parental status, caste and socio-economic background, region or domicile, sexual orientation and gender identity, work experience, functional background, personality and cognitive style, neurodiversity, veteran status.
• India-specific dimensions that global frameworks miss: caste and community, mother tongue, rural/urban and Tier-2/Tier-3 origin, religion, North-East and inter-state migration, first-generation graduates, and disability in a context of poor physical accessibility.
• Levels of analysis: individual (bias, identity), team (faultlines, subgroups), organisational (policy, climate), societal (labour-market supply, education pipeline).
• Representational vs substantive diversity: headcount at the gate vs voice in decisions; tokenism is the former without the latter.
• Faultline theory - a team splits along *correlated* attributes (all the young women in operations, all the senior men in sales); correlated differences are far more damaging than uncorrelated ones.

Why organisations pursue it, and the theory behind it

• Legal and compliance driver - statutory duties (POSH, RPwD, Transgender Persons Act), disclosure duties (BRSR), customer and investor contract conditions.
• Talent-supply driver - India's female labour force participation, though rising in recent PLFS rounds, remains far below male participation; an employer that cannot retain women effectively recruits from half the market.
• Market and client driver - IT and GCC clients in the US and EU impose supplier diversity and pay-reporting conditions; EU pay-transparency rules flow into Indian delivery centres through contracts.
• Performance driver and its honest limits - consultancy studies (McKinsey's *Diversity Matters* series, Credit Suisse/Catalyst work) report correlations between executive diversity and financial performance. These are correlational, sometimes fail replication, and reverse causation is plausible (profitable firms can afford DEI). State the mechanism, not the magic number.
• Categorisation-Elaboration Model (CEM, van Knippenberg) - the examinable mechanism: diversity helps only when it triggers *elaboration* of task-relevant information; it harms when it triggers *social categorisation* into in-group/out-group. Inclusion is what tips the balance.
• Social Identity Theory (Tajfel and Turner) and Similarity-Attraction (Byrne) - people favour those like themselves, which is why "culture fit" interviewing reproduces the existing mix.
• Critical Mass Theory (Kanter) - below roughly 15% a minority member is a token: visible, stereotyped and isolated. Behaviour changes around 30-35%.
• Resource-Based View link - an inclusive climate is a socially complex, hard-to-imitate capability; the diversity *number* is imitable, the climate is not.

Barriers: bias, structure and climate

• Cognitive biases in HR decisions: affinity/similarity bias, confirmation bias, halo and horn effects, anchoring (on current salary - which perpetuates a gap), recency in appraisal, contrast effect, stereotyping, and group-serving attribution bias / the ultimate attribution error (a woman's success called luck, a man's called ability).
• Second-generation bias - no hostile intent; disadvantage embedded in neutral-looking practice: referral hiring, late-evening meetings, travel-heavy "stretch" roles, relocation as a promotion condition, "potential" judged by self-promotion.
• Stereotype threat (Steele and Aronson) - awareness of a negative stereotype depresses performance in evaluative settings, such as aptitude tests.
• Structural metaphors: glass ceiling (invisible barrier below senior levels), sticky floor (stuck in entry grades), leaky pipeline (proportional losses at each stage), glass cliff (minorities promoted into failing units), broken rung (biggest drop at the first promotion to manager), maternity wall / motherhood penalty and the mirroring fatherhood premium.
• Microaggressions - small, repeated slights (mispronouncing names, "you speak English so well", commenting on a transgender colleague's appearance) that drive attrition without producing a single formal complaint.
• Covering (Yoshino) - downplaying a stigmatised identity to fit in: anglicising a name, hiding a disability, hiding a same-sex partner. High covering is a symptom of low inclusion.
• Supply-side constraints in India - safety and commuting, care work, hostel and sanitation facilities, engineering-intake gender mix in some branches, inaccessible campuses.
• Backlash and fatigue - perceived zero-sum competition, "reverse discrimination" claims, and resentment where selection standards are suspected to differ. Managed by transparency about process, and by fixing criteria rather than lowering them.

The Indian legal and regulatory framework

• Constitution - Articles 14 (equality before law), 15(1) (no discrimination on religion, race, caste, sex, place of birth) with 15(3) (special provision for women), 16 (equality of public employment, with 16(4)/16(4A) reservation for SC/ST/backward classes), 17 (abolition of untouchability), 23 (forced labour). Articles 15 and 16 bind the State; private employment is reached through statute and contract, not directly.
• Reservation is a public-sector and public-education instrument in India; there is no general statutory quota for private-sector employment, which is why private DEI uses *goals* and *slates*, not quotas. Some state laws on local-candidate quotas in private jobs have been challenged in court - treat their status as unsettled.
• Sexual Harassment of Women at Workplace (Prevention, Prohibition and Redressal) Act, 2013 (POSH) - codifies the Supreme Court's *Vishaka* guidelines (1997). Internal Committee mandatory where 10 or more workers are employed, presided over by a senior woman, with at least one external member from an NGO or with legal knowledge; Local Committee at district level for smaller units and the informal sector; inquiry normally within 90 days; written policy, display, annual training and an annual report to the district officer. The Act protects women (any workplace, any employment status, including interns and visitors). Men and transgender persons are protected through gender-neutral internal policy, not this Act.
• Rights of Persons with Disabilities Act, 2016 - 21 specified disabilities; prohibits discrimination in employment; requires reasonable accommodation; 4% reservation in government establishments across four categories (and 5% in higher-education institutions); every establishment, including private, must publish an equal-opportunity policy and maintain records of employees with disabilities; accessibility obligations; Chief Commissioner and State Commissioners for Persons with Disabilities as grievance authorities.
• Transgender Persons (Protection of Rights) Act, 2019 and Rules, 2020 - prohibits discrimination in recruitment, promotion and conditions of service; every establishment must designate a complaint officer and adopt an equal-opportunity policy and infrastructure such as gender-neutral facilities. Follows *NALSA v. Union of India* (2014), which recognised gender identity; *Navtej Singh Johar* (2018) read down s.377 IPC. In *Supriyo* (2023) the Supreme Court declined to recognise same-sex marriage, so partner benefits for LGBTQIA+ employees remain an employer policy choice.
• The four Labour Codes, in force from 21 November 2025:
  • Code on Wages, 2019 - subsumes the former Equal Remuneration Act 1976; prohibits discrimination on ground of sex in wages and in recruitment for the same or similar work, and bars reduction of anyone's wages to comply.
  • Code on Social Security, 2020 - subsumes the former Maternity Benefit Act; 26 weeks paid maternity benefit (12 weeks for the third child onward and for commissioning/adopting mothers as specified), nursing breaks, and a crèche where the establishment has 50 or more employees. Gig and platform workers are brought within social-security schemes.
  • Occupational Safety, Health and Working Conditions Code, 2020 - women may be employed in all establishments and in all types of work, including before 6 a.m. and after 7 p.m., subject to consent and prescribed safety, transport and facility conditions; also repealed the former Mines Act and Dock Workers Act.
  • Industrial Relations Code, 2020 - Grievance Redressal Committee in establishments with 20 or more workers, with adequate representation of women.
• HIV and AIDS (Prevention and Control) Act, 2017 - bars discrimination in employment, bars mandatory HIV testing as a precondition of employment, and requires larger establishments to have a complaints mechanism and designated officer.
• Mental Healthcare Act, 2017 - non-discrimination on the ground of mental illness; relevant to fitness criteria and medical-insurance design.
• Company and securities law - Companies Act, 2013 s.149(1) requires at least one woman director for prescribed classes of companies; SEBI LODR Reg. 17(1)(a) requires at least one independent woman director for the top 1,000 listed entities by market capitalisation. SEBI's BRSR (Principle 5, plus Section A) requires disclosure of workforce and board composition by gender and disability, wages by gender, POSH complaints, accessibility and human-rights due diligence; it applies to the top 1,000 listed entities and assurance of BRSR Core is being phased in by market-cap rank - check the current SEBI circular, as the phasing has been revised.
• Data protection caveat - diversity data is personal data and often sensitive (health, caste, sexual orientation). Self-identification must be voluntary, purpose-limited and reported in aggregate. Under the DPDP Rules, 2025 the Data Fiduciary duties (notice, consent, breach intimation, erasure) take effect only from 13 May 2027; until then the IT Act s.43A and the SPDI Rules, 2011 govern sensitive personal data.
• Voluntary global frameworks: ISO 30415:2021 (*Human resource management - Diversity and inclusion*, guidance, not certifiable as a management-system standard), GRI 405 (Diversity and Equal Opportunity) and GRI 406 (Non-discrimination), UN Women's Empowerment Principles, UN Guiding Principles on Business and Human Rights, ILO Conventions 100 (equal remuneration) and 111 (discrimination) - India has ratified both, SDGs 5, 8 and 10.

Designing and running a DEI programme

1. Diagnose - representation by level, function and location; flow rates (apply → shortlist → offer → join → promote → exit); pay-equity regression; inclusion survey by segment; accessibility and policy audit; POSH and grievance data.
2. Set goals - time-bound *aspirational representation goals* benchmarked to available supply, not quotas; separate hiring-mix goals from stock goals, because stock moves slowly.
3. Fix the hiring process - structured interviews with a scored rubric, job-relevant criteria, work-sample tests, diverse interview panels and diverse slates (the "two in the pool" effect), anonymised CV screening at shortlist, de-biased job adverts, no pay-history anchoring, campus mix beyond the usual institutes, returnship programmes for career-break candidates, PwD and transgender hiring partnerships (for example, Lemon Tree Hotels' large share of employees with disabilities, Tata Steel's transgender and women-in-mining recruitment).
4. Fix progression - calibration of ratings with bias checks, promotion-parity monitoring, sponsorship (an advocate who spends political capital on you) alongside mentoring (advice), transparent job posting, succession slates with diversity checks, succession-blocker reviews.
5. Fix the experience - inclusive leadership behaviours and manager training, psychological safety, flexible and hybrid work, parental and adoption leave, crèche and transport, accessible premises and assistive technology, gender-neutral washrooms, insurance covering partners and gender-affirmation where offered, zero-tolerance enforcement.
6. Employee Resource Groups / affinity networks - voluntary, employee-led, executive-sponsored groups (women's network, PwD network, PRIDE network, regional networks) with a charter, budget and consultation role; risk is becoming a social club or an unpaid second job.
7. Govern and report - DEI council chaired by the CEO, metrics in the balanced scorecard and in leaders' KPIs, annual public reporting through BRSR and the sustainability report.
8. Maturity ladder: compliance → awareness/training → representation targets → systemic equity (process redesign) → inclusive culture embedded in business decisions.
9. What the evidence says works: structured selection, transparency and accountability, sponsorship, targeted recruitment. What works poorly alone: one-off unconscious-bias training (effects decay in weeks and can trigger backlash), mandatory grievance-only approaches, and diversity statements without process change.

Measuring it: metrics and their traps

• Stock metrics - headcount share by segment, overall and *by level* (the level cut is what reveals a glass ceiling); board and senior-management share; share of employees with disabilities.
• Flow metrics - diversity of applicants, shortlist, offers and joiners; offer-to-join (declination) rate; promotion rate; voluntary attrition; internal mobility - all split by segment.
• Equity metrics - unadjusted (raw) and adjusted (like-for-like) pay gap, mean and median; equal-pay-for-equal-work gap within grade and role; access to training hours and to high-visibility projects.
• Inclusion metrics - inclusion index from survey items, favourability and the gap between segments, eNPS by segment, covering rate, POSH/grievance volumes (a *rise* can mean greater trust, not more harassment).
• Statistical tools - Blau and Shannon indices for mix, four-fifths (adverse impact) rule for selection fairness, chi-square for independence of selection and group, Oaxaca-Blinder decomposition or multiple regression for pay gaps, cohort/funnel survival analysis.
• Traps: aggregating across levels hides structure (a Simpson-type reversal); small cells make percentages unstable and can re-identify individuals; a single headline ratio ignores where losses occur; self-ID non-response biases disability and LGBTQIA+ counts downward; and goals without process change produce churn - hire in, leak out.

Key points

• Diversity - the mix (a fact); equity - fair systems (a design); inclusion - the lived experience (a climate); belonging - the felt outcome.
• Equality vs equity - same treatment vs treatment calibrated to need; the ramp example is the standard exam illustration.
• Surface-level vs deep-level diversity - visible/ascribed attributes vs acquired attributes, values and cognition.
• Categorisation-Elaboration Model - diversity pays only through information elaboration; social categorisation is its failure mode.
• Critical mass / tokenism (Kanter) - below about 15% a minority is a token; behaviour shifts near 30%.
• Faultlines - correlated demographic splits fracture teams more than uncorrelated differences.
• Second-generation bias - neutral-seeming practices (referrals, relocation, late meetings) producing unequal outcomes.
• Glass ceiling, sticky floor, broken rung, leaky pipeline, glass cliff - the five structural metaphors examiners expect by name.
• POSH Act 2013 - IC mandatory at 10+ workers, chaired by a senior woman with an external member, inquiry normally in 90 days, annual report; protects women.
• RPwD Act 2016 - 21 disabilities, reasonable accommodation, 4% government reservation, equal-opportunity policy and records for all establishments.
• Transgender Persons Act 2019 - non-discrimination in employment plus a designated complaint officer and equal-opportunity policy.
• Code on Wages, 2019 - no discrimination by sex in wages or recruitment for same or similar work (replaced the former Equal Remuneration Act).
• Code on Social Security, 2020 - 26 weeks maternity benefit, nursing breaks, crèche at 50+ employees (replaced the former Maternity Benefit Act).
• OSH&WC Code, 2020 - women in all establishments and in night work with consent and prescribed safeguards.
• Companies Act s.149(1) and SEBI LODR 17(1)(a) - woman director, and independent woman director for the top 1,000 listed entities.
• BRSR - SEBI's mandated ESG disclosure for the top 1,000 listed entities; Principle 5 and Section A carry the diversity, wage-by-gender and POSH numbers.
• Reasonable accommodation - adjustment that enables a qualified person to perform the job without undue hardship to the employer; it is not a lowering of the standard.
• Structured selection beats bias training - rubric-based, job-relevant evaluation is the best-evidenced single intervention.

Formulas and examples

Example 1 - Mix indices. Workforce N = 1,200: men 720, women 456, transgender/other 24. p = 0.60, 0.38, 0.02.

• Σp² = 0.3600 + 0.1444 + 0.0004 = 0.5048 → B = 1 − 0.5048 = 0.4952; maximum for k = 3 is 1 − 1/3 = 0.6667, so B* = 0.4952/0.6667 = 0.743.
• H = 0.60(0.5108) + 0.38(0.9676) + 0.02(3.9120) = 0.3065 + 0.3677 + 0.0782 = 0.7524 nats; evenness = 0.7524/ln 3 = 0.7524/1.0986 = 0.685.
• Reading: the mix is well short of an even split, driven by the dominant category.

Example 2 - Headline pay gap vs like-for-like gap. Same firm, every band paying women exactly 2% less than men:

• Mean male = (200×6 + 300×9 + 180×14 + 40×24)/720 = (1,200 + 2,700 + 2,520 + 960)/720 = 7,380/720 = ₹10.25 lakh.
• Mean female = (300×5.88 + 120×8.82 + 30×13.72 + 6×23.52)/456 = (1,764 + 1,058.4 + 411.6 + 141.12)/456 = 3,375.12/456 = ₹7.4016 lakh.
• Unadjusted gap = (10.25 − 7.4016)/10.25 = 2.8484/10.25 = 27.8%, while the like-for-like gap is only 2%.
• Conclusion: about 26 points of the gap is structural - women are concentrated in B1 - so the fix is promotion and pipeline design, not a one-off salary correction. Reporting only the 2% figure would be misleading, and only the 27.8% figure would misdiagnose the cause.

Example 3 - Adverse impact in selection. A graduate-trainee drive: 2,000 male applicants → 240 selected; 800 female applicants → 72 selected.

• SR(men) = 240/2,000 = 12.0%; SR(women) = 72/800 = 9.0%.
• AIR = 9.0/12.0 = 0.75 < 0.80 → adverse impact flagged; the process must be shown to be job-related.
• To reach AIR = 0.80, women's selection rate must be 0.80 × 12.0% = 9.6% → 0.096 × 800 = 76.8 → 77 selections (5 more), or the screening step causing the difference must be validated or redesigned.

Example 4 - Is the representation target feasible? Current N = 1,200 with 456 women (38.0%). Plan: close next year at 1,320 with 45% women; attrition 15% of opening headcount, exits proportional to mix.

• Exits = 0.15 × 1,200 = 180, of which women = 0.38 × 180 = 68 (rounded).
• Hires = 1,320 − 1,200 + 180 = 300.
• Women needed at close = 0.45 × 1,320 = 594; women remaining = 456 − 68 = 388 → women hires needed = 594 − 388 = 206.
• Required diversity hire rate = 206/300 = 68.7% of all hires. Against a market applicant pool that is nowhere near 68.7% female, the target is not achievable in one year by hiring alone; either extend the horizon or cut female attrition.

Example 5 - Cost of the inclusion gap. Voluntary attrition is 18% for women and 13% for men; average female CTC ₹7.40 lakh; replacement cost taken at 50% of CTC.

• Excess exits = (0.18 − 0.13) × 456 = 22.8 ≈ 23 people.
• Replacement cost per exit = 0.50 × 7.40 = ₹3.70 lakh → total = 23 × 3.70 = ₹85.1 lakh a year.
• Inclusion survey for the same population: 850 valid responses of 1,200 (71% response). Favourable 544, neutral 170, unfavourable 136 → inclusion index = 544/850 = 64.0%; net favourable = (544 − 136)/850 = 48.0%. Split by gender, women 52% favourable vs men 70% - an 18 percentage-point gap that matches the attrition difference and points at manager behaviour and progression, not pay.

Where it is used

• IT services and GCCs (TCS, Infosys, Wipro, Accenture, Cognizant, Capgemini India): gender-mix targets, returnship programmes for women after career breaks, accessibility of delivery platforms, client-mandated supplier-diversity and pay-reporting clauses.
• Manufacturing, mining and steel (Tata Steel, JSW, Mahindra, Hindustan Zinc): women on shop floors and in underground mining under the OSH&WC Code's night-work provisions, hostel and sanitation infrastructure, transgender hiring in operations.
• Banking, insurance and consulting (HDFC Bank, SBI, Axis, Deloitte, EY, KPMG India): POSH governance, pay-equity audits, board composition under Companies Act s.149 and SEBI LODR, BRSR Principle 5 reporting.
• FMCG and retail (HUL, Marico, Titan, Lemon Tree Hotels, Trent): inclusive frontline hiring, employees with disabilities in store and hotel operations, inclusive marketing and product accessibility.
• Job roles: DEI lead / head of inclusion, HR business partner, talent acquisition and campus lead, compensation and benefits analyst (pay-equity work), people-analytics analyst, ESG and sustainability reporting analyst, POSH external member and IC secretary, CSR manager.
• Public sector and government: reservation administration under Articles 16(4)/16(4A), RPwD 4% reservation and equal-opportunity policies, accessibility audits under the Accessible India Campaign.

Interview questions

• Q: What is the difference between diversity, equity and inclusion? A: Diversity is the mix of people, equity is the fairness of systems that decide pay and progression, and inclusion is whether that mix is respected and heard. Diversity is counted, equity is designed, inclusion is experienced.
• Q: Equality or equity - which should an employer aim for? A: Equity, because identical treatment of people facing different barriers reproduces the existing gap; equity adjusts inputs such as accessibility, flexibility and sponsorship so that outcomes are not predicted by identity.
• Q: Does diversity automatically improve performance? A: No. Under the Categorisation-Elaboration Model it helps only when differences are used to elaborate task-relevant information; without inclusion it produces in-group/out-group splits and conflict. The published performance evidence is correlational.
• Q: How many people must a workplace employ before a POSH Internal Committee is compulsory? A: Ten or more workers; it must be presided over by a senior woman employee and include at least one external member from an NGO or with relevant legal knowledge, with the inquiry normally completed within 90 days.
• Q: Which law governs equal pay for women in India today? A: The Code on Wages, 2019, in force from 21 November 2025, which subsumed the former Equal Remuneration Act and prohibits discrimination on ground of sex in wages and in recruitment for the same or similar work.
• Q: What does the RPwD Act 2016 require of a private employer? A: Non-discrimination, reasonable accommodation, an equal-opportunity policy and records of employees with disabilities; the 4% reservation applies to government establishments, not private ones.
• Q: Can a private Indian company set a hiring quota for women? A: It can set internal aspirational goals and require diverse candidate slates, but a rigid quota that excludes other candidates risks challenge; reservation as a legal instrument applies to public employment under Article 16(4).
• Q: A manager says his team's 25% pay gap just reflects seniority. How would you test that? A: Compute the gap within each grade, role and location, then regress ln(pay) on grade, role, experience and performance and read the gender coefficient; if the within-grade gap is small the problem is distribution across grades, which is a promotion problem.
• Q: What is the four-fifths rule and when is it used? A: If a group's selection rate is below 80% of the highest group's rate, adverse impact is flagged and the selection step must be shown to be job-related. It is a US EEOC rule, used in India mainly as a diagnostic and by employers serving US clients.
• Q: Why can an unconscious-bias workshop alone fail? A: Awareness effects decay within weeks, do not change the decision process, and can trigger backlash; structured interviews, calibrated ratings, transparency and accountability change outcomes.
• Q: POSH complaints doubled after a new policy. Is that bad? A: Not necessarily - a rise usually signals that people now trust the channel. Judge it against resolution times, repeat-offender data and survey measures of psychological safety.
• Q: How would you support a transgender employee's onboarding? A: Update records and systems to the chosen name and gender with consent, provide gender-neutral facilities, brief the team on respectful conduct without disclosing details, and route any discrimination to the complaint officer required under the Transgender Persons Act, 2019.

Common mistakes

• Treating diversity and inclusion as one idea, or defining equity as equality - the single most penalised error in exams.
• Quoting the Equal Remuneration Act 1976, the Maternity Benefit Act 1961 or the Factories Act 1948 as live law; those subjects are now in the Code on Wages, the Code on Social Security and the OSH&WC Code, in force from 21 November 2025.
• Saying the POSH Act protects all employees - it protects women; protection for men and transgender employees comes from internal policy and the Transgender Persons Act, 2019.
• Confusing constitutional reservation (public employment, Article 16) with private-sector DEI goals, and calling private targets "reservation".
• Reporting one aggregate pay gap and drawing a conclusion from it, instead of splitting by grade - the structural and like-for-like gaps answer different questions.
• Quoting a consultancy correlation as proof that diversity causes profit, with no mechanism and no caveat.
• Setting representation targets without checking the arithmetic of attrition, growth and available supply - the target then needs an impossible hiring mix.
• Collecting self-identified disability, caste or sexual-orientation data without voluntary consent, purpose limitation or small-cell suppression.

Related topics

Recruitment and selection · Compensation and benefits · Labour laws and the labour codes · HR analytics · Organisation development: diagnosis and interventions · Strategic HRM and HR business partnering