Identifying direct discrimination in “proxy cases” after R (on the application of Coll) v Secretary of State for Justice


Dee Masters considers the recent Supreme Court decision of R (on the application of Coll) v Secretary of State for Justice [2017] UKSC 40 which was handed down on 24 May 2017.  A copy of the judgment is available here.


Coll is compulsory reading for discrimination lawyers. Lady Hale, who provided the leading judgment, examines the extent to which costs can justify discrimination, the nature of the justification defence where segregated services are provided and the interplay between the justification defence and breach of the public sector equality duty.  This blog will focus on one important aspect of the case namely the identification of direct discrimination in “proxy cases”.  These are cases where the reason for the detrimental treatment is not expressly a protected characteristic, but the claimant wishes to argue that there is still direct discrimination.  The classic example of a proxy case is where an employer has a policy of “neutrality” in the workplace and a court has to decide whether this is really a proxy for religion or particular religious communities.[1]


The facts in Coll are simple.  Ms Coll was sentenced to life imprisonment for murder.  She was eventually released on licence which involved a requirement that she reside permanently in Approved Premises (AP) in Bedford for nine months.  Approved Premises used to be known as probation hostels and bail hostels.   All Approved Premises are currently single-sex.  There are 94 for men all over England and Wales, with a number in London.  There are 6 for women.  Critically, there are none in London which is where Ms Coll had lived for most of her life and where her children/grandchildren continued to live.   The limited number and geographical distribution of women’s Approved Premises means that women are far more likely than men to be placed in Approved Premises away from their homes and family.  This is what happened to Ms Coll.  She was then further disadvantaged because once she was permitted to live away from Approved Premises in Bedford, she was forced to live in rented accommodation nearby as she had been unable to search for accommodation in London because of restrictions on her travel.

Direct sex discrimination

Ms Coll’s primary argument was that she had been the victim of direct sex discrimination.  Her case was succinctly summarised by Lady Hale in the following terms:

“26. … Being required to live in an AP a long way away from home is a detriment.  A woman is much more likely to suffer this detriment than a man, because of the geographical distribution of the small number of APs available for women.  This is treating her less favourably than a man because of her sex.”

The Supreme Court readily accepted that Ms Coll had been the victim of direct sex discrimination because:

“31. …  all the women who would be required to live in an AP when released on licence suffered the much greater risk than the men that they would be sent to an AP far from their homes and families.”

In other words, the detrimental treatment (greater risk of being far away) applied to Ms Coll because she was a woman and this risk was then realised when she was placed in Bedford.

Lady Hale elaborated on her analysis by stating that the comparative exercise was between a man being required to live in AP when released on licence and a woman in the same position:

“32. … the question of comparing like with like must always be treated with great care – men and women are different from one another in many ways, but that does not mean that the relevant circumstances cannot be the same for the purpose of deciding whether one has been treated less favourably than the other.  Usually, those circumstances will be something other than the personal characteristics of the men and women concerned, something extrinsic rather than intrinsic to them.  In this case, the material circumstances are that they are offenders being released on licence on condition that they live in an AP.  Those circumstances are the same for men and women.  But the risk of being placed far from home is much greater for the women than for the men.”

Once the comparative exercise was formulated in these terms, it was clear that s.13 of the Equality Act 2010 was engaged (“A person (A) discriminates against another (B) if, because of a protected characteristic, A treats B less favourably than A treats or would treat others”). 

This comparative exercise is critical to understanding why Ms Coll was the victim of direct as opposed to indirect discrimination even though the detrimental treatment was the greater risk of a negative occurrence (which is then realised for the particular claimant) which is more usually seen in the context of indirect discrimination claims.[2]

To elaborate, imagine a scenario in which an employer has a policy of requiring sales people to work full-time.  This is a requirement which is applied to both men and women but it places women at a greater risk of detriment (i.e. an inability to comply with the requirement would make employment unsustainable) since women are more likely to be primary caregivers to young children.  A female employee requests part time hours and is turned down which forces her to resign. She could never bring a direct sex discrimination claim because a man who asked for part time hours would also always be turned down.  Instead, she could bring an indirect sex discrimination claim on the basis that the requirement to work full-time is a PCP which places women at a particular disadvantage and she was placed at that disadvantage.  This type of classic indirect sex discrimination claim is radically different to Coll in which a male prisoner released on licence would not be exposed to the same risk of being placed away from his family; the detriment is gender specific. 

Similarly, Ms Colls’ claim would not make any sense as an indirect sex discrimination claim because once a neutral PCP is formulated it becomes plainly and unarguably justifiable.  This point was made by Lady Hale:

“43. Ms Rose accepts that it is difficult to analyse this case in terms of indirect discrimination.  The whole point of indirect discrimination is that a PCP is applied equally to, in this case, men and women, whereas the complaint here is of unequal provision. “Shoe-horning” the complaint into indirect discrimination by identifying the PCP as the requirement to live in an AP on release does not really work, because what has to be justified is the PCP, and such a PCP is readily justifiable by the aims of protecting the public, reducing reoffending and assisting the offender’s rehabilitation.  In my view, no such shoe-horning is required.  Conduct cannot at one and the same time be both direct and indirect discrimination.  The finding that this is direct discrimination, albeit potentially justifiable, rules out a finding of indirect discrimination.”

The “exact correspondence” rule is only applicable in “proxy cases”

The Supreme Court also addressed the extent to which there needs to be “exact correspondence” between the advantaged / disadvantaged group and the protected characteristic.  The Secretary of State argued that there could not be direct sex discrimination because other women could be placed reasonably near to their homes.[3]  Lady Hale reject that argument.  The Supreme Court concluded that there is only a requirement for “exact correspondence” where the reason for the detrimental treatment is not the protected characteristic but the court must determine whether the reason for the detrimental treatment is actually a proxy for the protected characteristic. 

Lady Hale stated that:

“29. … as Ms Rose correctly points out, the “exact correspondence” test is only relevant where the actual criterion used by the alleged discriminator is not a protected characteristic but something else.  In Patmalniece it was not having the right to reside in the United Kingdom; in Preddy v Bull, it was not being married.  The question is whether some other criterion is in reality a proxy for the protected characteristic.  The best-known example is James v Eastleigh Borough Council [1990] 2 AC 751, where people who had reached the state retirement ages for men and women meant that a 61-year-old woman got in free whereas her 61-year-old husband did not.  This was held to be direct discrimination on grounds of sex.  In this case, there is no doubt what the criterion is.  It is sex, which is itself a protected characteristic.” (Emphasis added)

Lady Hale only discusses the “exact correspondence” test briefly in Coll.  However, earlier Supreme Court decisions have made it plain that “exact correspondence” will arise where there is a 100% overlap between the protected characteristic / absence of the protected characteristic within the advantaged group / disadvantaged.  This is illustrated by the following cases: 

  • In Patmalniece v Secretary of State for Work and Pensions (AIRE intervening) [2011] 1 WLR 783 the Latvian claimant, who was an asylum seeker, was denied UK state pension credit.  The Supreme Court concluded that there was no direct discrimination since there was no “exact correspondence” between nationality (UK national vs non-UK national) in both the advantaged group (i.e. those can satisfy the test of habitual residence) and the disadvantaged group (i.e. those who cannot satisfy the test of habitual residence).[4]  That is, UK nationals might not be able to satisfy the habitual residence test if, for example, they have been absent from the UK for a long period of time.[5]
  • In Preddy v Bull [2013] 1 WLR 3741, a hotel would not allow unmarried couples (homosexual or heterosexual) to share a double bed.  The claimants, a homosexual couple in a civil partnership, were denied a double bedroom.  The majority of the Supreme Court concluded that they were the victims of direct discrimination on the grounds of sexual orientation.  Lady Hale stated that direct discrimination will arise where there is an exact correlation between the advantaged and disadvantaged groups on the basis of sexuality.[6]  Since only heterosexual couples could enter into a legal relationship which satisfied the hotel’s rule on double rooms, there was direct discrimination.[7]
  • Although it is not mentioned in Coll, the same analysis of the “exact correspondence” test appears in Onu v Taiwo [2016] ICR 756 where Lady Hale provided the leading judgment on behalf of the Supreme Court.  The claimants had been mistreated because they were vulnerable immigrants.  The Supreme Court rejected the argument that this was direct race discrimination since the claimants’ vulnerable immigration status was not in dissociable from their nationality.  In reaching that conclusion, the Supreme Court concluded that the “exact correspondence” test was not met as between nationality (non-British) and vulnerable immigration status.

Lady Hale stated:

“24. In these cases, Ms Taiwo and Ms Onu had limited leave to enter on domestic workers’ visas.  It was the terms of those visas which made them particularly vulnerable to the mistreatment which they suffered.  At the relevant time, such visas were granted to workers who had already been working abroad for a year; typically they would be granted for a year, though renewable; and the employee would have to seek the approval of the immigration authorities for any change of employer while here.  In practice, therefore, such workers were usually dependent upon their current employers for their continued right to live and work in this country. 

26. Clearly, however, there are many non-British nationals living and working here who do not share this vulnerability.  No doubt, if these employers had employed British nationals to work for them in their homes, they would not have treated them so badly.  They would probably not have been given the opportunity to do so.  But equally, if they had employed non-British nationals who had the right to live and work here, they would not have treated them so badly.  The reason why these employees were treated so badly was their particular vulnerability arising, at least, in part, from their particular immigration status.  As Mr Rahman pointed out, on behalf of Mr and Mrs Akwiwu, it had nothing to do with the fact that they were Nigerians.  The employers too were non-nationals, but they were not vulnerable in the same way.”

Each of these Supreme Court cases where the “exact correspondence” test has been deployed are cases where the reason for the detrimental treatment is not expressly a protected characteristic i.e. test of habitual residence not satisfied, unmarried couples cannot have a double bed, vulnerable immigration status will lead to exploitation.  In order to determine whether these reasons are actually proxies for the protected characteristics, the “exact correspondence” test was then utilised.

Bearing in mind the Supreme Court’s analysis of the “exact correspondence” test in Coll, Patmalniece, Preddy and Onu, there is one passage in Coll which, at first blush, seems anomalous. 

The relevant passage is underlined below:

“30. Further, it cannot be a requirement of direct discrimination that all the people who share a particular protected characteristic must suffer the less favourable treatment complained of. It is not necessary to show, for example, that an employer always discriminates against women: it is enough to show that he did so in this case.  In the Birmingham case, some of the girls achieved a high enough mark to gain a place at a selective school.  What all the girls suffered from was the risk that if they did not get a high mark, they would not get a place – just as, in the recent case of Essop v Home Office (Border Agency) [2017] UKSC 17; [2017] 1 WLR 1343, all the BME candidates suffered from the greater risk of failing the core skills assessment required for promotion, but of course some of them passed it.  In the Birmingham case, some of the girls did of course achieve a high enough mark to get a place.  But there were some who achieved a mark which would have been high enough had they been boys but was not high enough because they were girls.  That is direct discrimination on grounds of sex.”

If the “exact correspondence” test dictates that direct discrimination will only exist in proxy cases where all of the protected class are in the disadvantaged group, the highlighted passage looks wrong.  But, it is important to remember that in Coll and R v Birmingham City Council, ex parte EOC [1989] 1 AC 115 (referred to as the Birmingham Case in the extract above), the detriment appears to have been the risk of suffering a disadvantage which was then realised for the claimants.  Seen in this context, Lady Hale’s analysis makes perfect sense.  She is simply making the obvious point that whilst all of the protected group must be disadvantaged (e.g. risk of being placed in AP away from home) not all of the protected group need experience the ultimate consequence of that disadvantage.  So, in Coll, some women could well have been placed in an AP that was near their families but this did not negate the fact that all women ran the risk that they might not, whereas none of the men were exposed in this way. 

Whilst not expressly addressed in Coll, it is also important to be clear that in non-proxy cases, there is no requirement for an exact delineation between the advantaged group/disadvantaged group and the protected characteristic.  For example, in a case where a male manager was prejudiced against women with caring responsibilities, it would be direct sex discrimination for him to deny the claimant a promotion because she was on maternity leave even if there were other pregnant women who he had not disadvantaged in the same way.  It is enough that the reason for the claimant’s detrimental treatment was her maternity leave. 

Similarly, in cases where purportedly non-discriminatory criteria are being applied, a claimant can still argue that the process is sham without needing to show “exact correspondence”.  For example, a line manager who is prejudiced against women might appear to apply objective and fair criteria when choosing who to make redundant within a particular pool, but if he actually manipulates the scoring so that one woman is ultimately dismissed, the process is a sham and the real reason for dismissal is sex even if other women escape dismissal.  In this scenario there would be no scope for the “exact correspondence” test as the alleged proxy has been shown to be a sham.


Identifying direct discrimination cases can be very difficult.  Coll is a welcome addition to the jurisprudence as it clarifies the circumstances in which the “exact correspondence” test applies, making it plain that it only applies to “proxy cases”. 


31 May 2017

[1] There are a number of blogs from Schona Jolly QC of Cloisters on the extent to which rules concerning neutrality in the work place amount to direct or indirect religious discrimination.  See, for example,  A more detailed article by Schona Jolly QC on this topic will appear in [2017] E.H.R.L.R., Issue 3.

[2] A recent example of an indirect discrimination case where the disadvantage was the risk of detrimental treatment is Essop v Home Office (Border Agency) [2017] 1 WLR 1343.

[3] Paragraph 28.

[4] Lord Hope, paragraphs 28-19.

[5] Lord Hope, paragraph 27.

[6] Paragraphs 18-22.

[7] Paragraphs 29-30.