Sometimes discrimination can be easy to spot – for example, if a hotel turns you away because you’re gay. This is called direct discrimination. This is when you’re treated differently simply because of who you are.

But there are other times when you may be treated in the same way as everybody else, but it has a different and worse effect on you because of who you are. This is also discrimination. The Equality Act 2010 calls this indirect discrimination.

Read this page to find out more about indirect discrimination.

Remember, indirect discrimination is when:

  • there’s a policy, practice or rule which applies to everybody in the same way
  • it places people who share your protected characteristic at a disadvantage
  • it places you personally at a disadvantage
  • the person applying the policy, practice or rule can’t show there’s a good enough reason for it.

What’s indirect discrimination?

The law which says you mustn’t be discriminated against is called the Equality Act 2010. Discrimination which is against the Equality Act is unlawful. This means you can take action in the civil courts.

Indirect discrimination is when there’s a practice, policy or rule which applies to everyone in the same way, but it has a worse effect on some people than others. The Equality Act says it puts you at a particular disadvantage.

What’s a practice, policy or rule?

A practice, policy or rule can be formal or informal. It can be a one-off decision or a decision to do something in the future. It includes things like arrangements, criteria, conditions, qualifications or provisions.

What’s important is that it applies to everyone in the same way, that it’s neutral.If something only applies to some people who all have the same protected characteristic, it would be direct discrimination.

The amended definitions of indirect discrimination are in the context of section 1(1) of the Race Relations Act (RRA) 1976(now section 1A) and section 1(2) of the Sex Discrimination Act (SDA) 1975. The original definitions in the SDA 1975 and the RRA 1976 contained a requirement that the employer demonstrate the practice introduced is ‘justifiable’ irrespective of the sex or the race or nationality of the person to whom it is applied. European legislation has had an impact upon these definitions via the Race Directive 2000/43/EC; the Burden of Proof Directive 1997/80/EC and the Equal Treatment Amendment Directive 2002/73/EC

In the context of the SDA 1975, the original provisions stated that the applicant had to prove firstly that the employer had applied a neutral requirement or condition in relation to which a considerably smaller proportion of women than men could comply and secondly, that it had been to the applicant’s detriment because she could not comply with it. The employer had the potential defence of showing that the requirement or condition was ‘justifiable’ despite the sex of the applicant 1.

The Burden of Proof Directive has caused the term “requirement or condition” to be replaced with the wider “provision, criterion or practice”, and it has become sufficient to show that the requirement or condition is to the detriment of the applicant and is or would be to the detriment of a considerably larger proportion of women. The former requirement to demonstrate ‘inability to comply’ has been removed.

The Equal Treatment Amendment Directive has added to this definition, which has resulted in the new Employment Equality (Sex Discrimination) Regulations 2005 (SI 2005/2467), which was in force from 1 October 2005. These regulations have removed the requirement to show ‘whether the impact of a provision, criterion or practice is to the detriment of a considerably larger proportion of women’ and replaced it with an assessment of whether the provision, criterion or practice “puts or would put women at a disadvantage”. This recognises that it is not always possible or necessary for applicants to show detailed statistics to demonstrate how they have been disadvantaged. Furthermore, the employer now has to show that the provision, criterion or practice is a ‘proportionate means of achieving a legitimate aim’, rather than ‘justifiable’ 2.

1 Sex Discrimination and Flexible Working: NLJ 28 March 2003 pp 470-471

As these changes are very new, there is a lack of case law to demonstrate its impact, which is arguably to widen the scope of the definition of indirect discrimination in the SDA 1975. Previous case law demonstrates the rigidity of the former approach. In Coylman Coyle v Georgiou EAT/535/00, the employer’s actions in demanding a return to full time working hours after maternity leave or a less attractive job was held to be indirectly discriminatory, though the employer was successful in proving that the requirement was justified, in the case of the applicant being a conveyancing solicitor.

This was also the approach followed by the Employment Appeal Tribunal (EAT) in Fox v Betesh Fox EAT/0363/01, in which again the requirement set by the employer for the applicant to return to full time working hours after maternity leave was held to be justified, in this case the applicant being a legal secretary. ‘Justifiability’ appears to have been influenced by a critical evaluation of whether there was a real need for the condition to be imposed, the impact upon the women, and whether in balancing both factors, the employer’s need outweighed the detriment to the women 3.

The approach to date favours the use of statistic information by the applicant in proving a disadvantage, which could now be affected by the new regulations. In London Underground v Edwards (No. 2) 4, the Court of Appeal found that although 20 out of 21 female operatives could comply with a new shift pattern, this was a ‘considerably smaller’ proportion of women than men, as all men could comply with the new pattern.

When statistics could not be provided, this was previously judged against the applicant, as shown in Carshalton College v Morris EAT/0673/01, where the applicant, a hairdressing tutor, had failed to provide statistical evidence to demonstrate that a considerably smaller proportion of child caring mothers could comply with the employer’s requirement to work the occasional evening as well as part time hours. This case would probably have been decided differently under the amended provisions of SDA 1975, as the only requirement would be to show she had suffered a detriment, and furthermore, may not have had to produce statistical information.

It is a feature of direct discrimination that actual or hypothetical comparators can be identified. Direct discrimination applied where a person is treated less favourably on the grounds of race or sex than a person who is not of the same race or sex. In Shamoon v Chief Constable of the Royal Ulster Constabulary 5, the House of Lords stated that the situation of an actual comparator should be in all material respects the same as the applicant’s, apart from his/her race/sex. The situation of a hypothetical comparator has to be drawn from the available evidence 6.

3 Allonby v Accrington & Rossendale College and others CA [2001] IRIR 364
4 [1998] IRIR 364 CA
5 [2003] The Times March 4
6 Different Treatment and Disparate Impact: Discrimination’s double whammy: NLJ 28 March 2003 pp 467-468

Claims based upon indirect discrimination appear to have been based upon more technical grounds, which were defined in Jones v Chief Adjudication Officer 7 as identification of the criterion, identification of the relevant population, or the ‘pool’ of those satisfying all the other relevant criteria, division of the ‘pool’ into two groups of those who satisfy the criterion and those who do not, statistical prediction of the proportion of each group which consist of women, the actual male/female balances in the groups, comparison of actual and predicted gender balances and if the women are under-represented in the first group and over-represented in the second the criterion can be proved to be discriminatory. Previously, if it was hard to judge the impact of a criterion, the courts required supporting statistical evidence from the applicant.

The requirement for production of statistical information means that the correct ‘pool’ for comparison must be chosen, as shown in London Underground v Edwards 8, in which the correct pool was held by the EAT to be all employees to whom the new arrangement applied, rather than all single parents affected by the new arrangements as argued by the applicant. If the wrong pool was chosen then the applicant would have to request an adjournment and have a costs order against her before collecting new statistics.

In relation to the ‘pool’, it has been held that job offers made to family and friends or to a particular person without considering any others, will not be indirectly discriminatory on racial or gender grounds, as the employer has not considered any pool of candidates 9. It is notable that no statistics were adduced.

This is a surprising decision as it appears to be contrary to the guidance provided in the Codes of Practice 10. This conflicts with R v Secretary of State for Education, ex parte Schaffer 11, where the actions of the employer in allocating grants to previously married lone parents was regarded as building an act of discrimination into the pool 12.

In Chief Constable of Avon and Somerset Constabulary v Chew 13 there was an absence of reliable statistics to show who in the relevant pool was able to comply with the condition, a shift pattern, and the tribunal was forced to adopt a different approach and evaluate how many officers could comply with the condition, before deciding that the condition was discriminatory as the majority of people who were unable to comply had child care responsibilities 14.

7 [1990] IRIR 533, CA
8 [1995] ICR 574
9 Coker v Lord Chancellor [2001] EWCA Civ 1756
10 Commission for Racial Equality Code of Practice for the elimination of racial discrimination and the promotion of equality of opportunity in employment, para 1.10a and Equal Opportunities Commission Code of Practice for the elimination of discrimination on the grounds of sex and marriage and the promotion of equal opportunity in employment, para 19(c)
11 [1987] IRIR 53
12 Out of Proportion: Solicitors Journal 3 May 2002 pp 403- 405
13 [2001] All ER (D) 101 (Sep)
14 Sex, Lies and Employment Tribunals: NLJ 5 April 2002 pp 514-516

There also has to be a significant disproportion between the two groups mentioned. In Staffordshire County Council v Black 15, the EAT decided that there was not a considerable difference between the percentages of the two groups at 87.5 percent of females to 97 percent of males.  The test appears to be set in R v Secretary of State, ex parte Unison 16, in which it was held that a difference of less than 5% is not regarded as ‘considerably smaller’.

The Race Directive led to the enactment of the Race Relations (Amendment) Regulations 2003, in force on 19 July 2003, which has provided for a wider definition of indirect discrimination, namely that it will occur where an apparently neutral “provision, criterion or practice” equally applied (a) puts or would put persons of the same race or national or ethnic origins at a particular disadvantage compared with others; (b) puts that person at that disadvantage; and (c) which the employer cannot show to be a “proportionate means” of achieving a legitimate aim. The new regulations also allow proof of indirect discrimination with or without statistic evidence 17.

The difficulty of producing statistical evidence is acknowledged by the introduction of the new regulations. Previous case law usually required such evidence about the restriction of the pool into racial groups with the necessary qualities, although if the applicant had done his or her best to provide some supportive evidence, then statistical evidence would not be demanded 18. This was acknowledged in Greater Manchester Police Authority v Lea 19, where the only evidence that could be obtained was that in relation to percentages of men and women not having occupational pensions. In Harvest Town Circle Ltd v Rutherford 20, the EAT decided that the figures in relation to statutory dismissal of employees over 65 years old was only available at the relevant government department.

The problems in collating statistical evidence are particularly evident in relation to race, because of the uneven distribution of ethnic minorities in the population. In Tower Hamlets LBC v Qayyum 21, the EAT considered that the appropriate pool was all appropriately qualified local authority employees of the Bengali race, rather than all Bengali residents in the United Kingdom.

As regards burden of proof, the point is that this remains with the applicant and only shifts to the respondent after the applicant has shown certain criteria 22, which are: less favourable treatment than an actual or hypothetical comparator, circumstances of unfavourable treatment to allow an inference to be drawn that it was on the basis of the complainant’s sex (or race) 23. This appears to open up the use of comparators in the area of indirect discrimination, though the impact is awaited.

15 [1995] IRIR 234
16 [1996] IRIR 438
17 Racing Ahead: NLJ: 28 March 2003 pp 465-466
18 Jones v University of Manchester [1993] IRIR 218 CA
19 [1990] IRIR 372 EAT
20 [2002] ICR 123 EAT
21 [1987] ICR 729
22 Pratt v Sanderson International Europe [2003] EAT 0529/02, 22 October 2003
23 Discrimination Law Update: Legal Action: February 2004 pp 24-29

In conclusion, it remains to be seen whether the emphasis upon statistics in indirect discrimination claims will become lessened, some authorities suggest that the new regulations will not affect how things have proceed so far, as tribunals have the discretion whether to consider other evidence in its absence. The use of comparators has been poor to date, though future case law may demonstrate its use as suggested in Pratt. On the whole, the fact the burden of proof remains upon the applicant until certain criteria are proved shows that it will still be difficult to achieve success in such claims.

Who can be indirectly discriminated against?

Something can be indirect discrimination if it has a worse effect on you because of your:

  • age
  • disability
  • gender reassignment
  • marriage or civil partnership
  • pregnancy and maternity
  • race
  • religion or belief
  • sex
  • sexual orientation

The Equality Act calls these protected characteristics.


A health club only accepts customers who are on the electoral register. This applies to all customers in the same way. But Gypsies and Travellers are less likely to be on the electoral register and therefore they’ll find it more difficult to join.

This could be indirect discrimination against Gypsies and Travellers because of the protected characteristic of race. The rule seems fair, but it has a worse effect on this particular group of people.

Who does the practice, policy or rule affect?

The practice, policy or rule applies to a certain group of people in the same way regardless of who they are. For example, you could have a rule that all employees in a shop must work on Saturdays. This is a neutral rule which affects everyone in the shop. This group is called the pool for comparison.

Within this group, some people with a particular protected characteristic may be put at a particular disadvantage by the rule. For example, if you’re Jewish and observe the Sabbath, you can’t work on Saturdays. It doesn’t matter that there aren’t any other Jewish people who work in the same shop. It can still be indirect discrimination if something would normally disadvantage people sharing your characteristic.

Are you disadvantaged?

You can only challenge a practice, policy or rule which you think is indirectly discriminatory if it affects you personally.


There’s a clause in your contract which says you may have to travel around the UK at short notice. It’s difficult for you to do this because you’re a woman with young children. This clause therefore places you at a particular disadvantage. It also places women generally at a disadvantage, as they’re more likely to be the carers of children.

You could challenge the clause because it affects you personally, even if you’ve not been asked to travel at short notice yet.

Are other people disadvantaged?

You need to show that people who share your protected characteristic are also disadvantaged, but that the other people in the group to whom the practice, policy or rule applies aren’t.


Your optician has a rule which allows payment for glasses by instalments for those in work. This rule applies to all their customers regardless of their protected characteristics. This is the pool for comparison. You’re not working because you’re a pensioner and so are not allowed to pay by instalments. The rule therefore places you at a disadvantage. It also disadvantages other pensioners who want to buy glasses by instalments.

Being a pensioner falls under the protected characteristic of age. This could be indirect discrimination, as one group of people who share the protected characteristic of age are particularly disadvantaged compared to another group.

How do you show disadvantage?

Sometimes, you can rely on common knowledge to show how some people are disadvantaged. For example, it’s common knowledge that Sunday is a day of worship for Christians. So a rule forcing staff to work on Sundays would put Christians at a particular disadvantage.

Sometimes, you may need to use statistics to show how people are disadvantaged by the policy, practice or rule. At other times, it may be necessary to use the help of an expert.

When might it not be indirect discrimination?

Indirect discrimination can sometimes be lawful. The Equality Actsays it’s not indirect discrimination if the person applying the practice, policy or rule, can show there’s a good enough reason for it. They would need to be able to prove this in court, if necessary. This is known in legal terms as objective justification.

Under s.19 of the Equality Act 2010, indirect discrimination occurs where person A applies to person B a provision, criterion or practice that is discriminatory in relation to a relevant protected characteristic of B’s.

A provision, criterion or practice is discriminatory in relation to a characteristic, if:

  • A applies (or would apply) it to people who do not share the protected characteristic with B (eg people who are not of the same racial group as B);
  • it puts (or would put) people who do share the protected characteristic at a particular disadvantage when compared with others who do not share it;
  • it puts (or would put) B at that disadvantage; and
  • A cannot justify the provision, criterion or practice by showing that it is a proportionate way of achieving a legitimate aim.

The courts must balance the employer’s need to apply the provision, criterion or practice, with its discriminatory effect on the employee.

The relevant protected characteristics are age, disability, gender reassignment, marriage and civil partnership, race, religion or belief, sex and sexual orientation. Protection from indirect discrimination does not apply in relation to pregnancy and maternity.