Landlord and tenant — Rents Acts — Protected tenant — Statutory successor — Same-sex partner — Paragraph 2 of Schedule 1 to Rent Act 1977 — Whether same-sex partner entitled to succeed to statutory tenancy — Whether Fitzpatrick v Sterling Housing Association Ltd surviving Human Rights Act 1998 — Whether Article 14 read with Article 8 of European Convention on Human Rights engaged — Whether para 2 of Schedule 1 to 1977 Act can be interpreted to apply to same-sex couples
In 1983, WJ was granted a residential tenancy of a flat protected by the Rent Act 1977. Until his death in January 2001, he shared the flat with the respondent tenant, his same-sex partner. Following WJ’s death, the appellant landlord brought proceedings for possession. The county court judge held that the respondent did not succeed to the tenancy as the surviving spouse within the meaning of para 2 of Schedule 1 to the 1977 Act, but that he did become entitled to an assured tenancy by succession as a member of the original tenant’s “family” under para 3 of that Schedule. The Court of Appeal allowed the respondent’s appeal, holding that he succeeded to the tenancy under para 2 of the Schedule. It accepted the respondent’s contentions that: (i) the distinction made in para 2(2) between the security of tenure available to the survivor of a heterosexual couple and that available to the survivor of a homosexual couple was discriminatory and infringed Article 14 of the European Convention on Human Rights, as read with Article 8; and (ii) para 2 should be read and given effect to under section 3 of the Human Rights Act 1998, in a way that was compliant with the Convention rights. The landlord appealed.
Held: The appeal was dismissed.
Per Lord Nicholls, Lord Steyn, Lord Rodger and Baroness Hale: Article 14 read with Article 8 of the Convention was engaged in relation to para 2 of Schedule 1 to the 1977 Act. Paragraph 2 of the schedule, construed without reference to section 3 of the Human Rights Act 1998, violated the respondent’s Convention right under Article 14, taken together with Article 8. The application of section 3 to para 2 has the effect that para 2 should be read and given effect to as though the survivor of a homosexual couple were the surviving spouse of the original tenant.
Per Lord Steyn: The Court of Appeal’s interpretation of the expression in para 2 — “as his or her wife or husband” to mean “as if they were his wife or husband” — was well within the power contained in section 3 of the 1998 Act.
Per Lord Millett, dissenting in respect of the application of section 3 of the 1998 Act: The exercise under section 3 is one of interpretation, not of legislation. Section 3 requires the court to read legislation in a way that is compatible with the Convention “so far as it is possible to do so”; that must be done by a process of interpretation alone. Paragraph 2 cannot be interpreted in a way that would remove the distinction between heterosexual and homosexual couples.
The following cases are referred to in this report.
Bellinger v Bellinger [2003] UKHL 21; [2003] 2 AC 467; [2003] 2 WLR 1174; [2003] 2 All ER 593, HL
Chapman v United Kingdom 27238/95; Coster v United Kingdom 24876/94; Beard v United Kingdom 24882/94; Lee v United Kingdom 25289/94; Smith v United Kingdom 25154/94 (2001) 33 EHRR 18
Dudgeon v United Kingdom A/45 (1982) 4 EHRR 149
Fitzpatrick v Sterling Housing Association Ltd [2001] 1 AC 27; [1999] 3 WLR 1113; [1999] 4 All ER 705; (2000) 32 HLR 178, HL
Fretté v France 36515/97 [2002] ECHR 156; (2004) 38 EHRR 21
Ghaidan v Godin-Mendoza [2002] EWCA Civ 1533; [2003] Ch 380; [2003] 2 WLR 478; [2002] 4 All ER 1162; [2003] HLR 35
International Transport Roth GmbH v Secretary of State for the Home Department [2002] EWCA Civ 158; [2003] QB 728; [2002] 3 WLR 344
Karner v Austria 40016/98 (2004) 38 EHRR 24
Litster v Forth Dry Dock and Engineering Co Ltd; sub nom Forth Estuary Engineering v Litster [1990] 1 AC 546; [1989] 2 WLR 634; [1989] 1 All ER 1134
Marleasing SA v La Comercial Internacional de Alimentación SA C106/89 [1990] ECR I-4135
Marzari v Italy 36448/97 (1999) 28 EHRR CD 175
Michaelek v Wandsworth London Borough Council; sub nom Michalak v Wandsworth London Borough Council [2002] EWCA Civ 271; [2003] 1 WLR 617; [2002] 4 All ER 1136; [2002] HLR 39, CA
Petrovic v Austria 20458/92 (2001) 33 EHRR 14
Pickstone v Freemans plc [1989] AC 66; [1988] 3 WLR 265; [1988] 2 All ER 803
Pretty v United Kingdom (2002) 35 EHRR 1
R (on the application of Anderson) v Secretary of State for the Home Department [2002] UKHL 46; [2003] 1 AC 837; [2002] 3 WLR 1800; [2002] 4 All ER 1089
R (on the application of Carson) v Secretary of State for Work and Pensions [2002] EWHC 978; [2002] 3 All ER 994, QBD
R (on the application of Carson) v Secretary of State for Work and Pensions [2003] EWCA Civ 797; [2003] 3 All ER 577, CA
R (on the application of Douglas) v North Tyneside Metropolitan Borough Council [2003] EWCA Civ 1847; [2004] 1 All ER 709
R (on the application of Erskine) v Lambeth London Borough Council [2003] EWHC 2479; [2003] NPC 118
R (on the application of Hooper) v Secretary of State for Work and Pensions [2003] EWCA Civ 813; [2003] 1 WLR 2623
R v A (No 2) [2001] UKHL 25; [2002] 1 AC 45
R v Director of Public Prosecutions, ex parte Kebilene; R v Director of Public Prosecutions, ex parte Boukemiche; R v Director of Public Prosecutions, ex parte Souidi; R v Director of Public Prosecutions, ex parte Rechachi; sub nom R v Director of Public Prosecutions, ex parte Kebelene [2000] 2 AC 326; [1999] 3 WLR 972; [1999] 4 All ER 801, HL
R v Hughes [2002] UKPC 12; [2002] 2 AC 259; [2002] 2 WLR 1058
R v Lambert [2001] UKHL 37; [2002] 2 AC 545; [2001] 3 WLR 206; [2001] 3 All ER 577, HLR v Secretary of State for the Home Department, ex parte Simms; R v Secretary of State for the Home Department, ex parte O’Brien; R v Governor of Whitemoor Prison, ex parte Main [2000] 2 AC 115; [1999] 3 WLR 328; [1999] 3 All ER 400
Rojas v Berllaque [2003] UKPC 76; [2004] 1 WLR 201
Röösli v Germany 28318/95 (1996) DR 85 149
S v United Kingdom 11716/85 (1986) 47 DR 274
S (Care order: Implementation of care plan), In re [2002] UKHL 10; [2002] 2 AC 291; [2002] 2 WLR 720; [2002] 2 All ER 192
Union Colliery Co of British Columbia Ltd v Bryden [1899] AC 580
Vasquez v The Queen [1994] 1 WLR 1304; [1994] 3 All ER 674
Walden v Liechtenstein 33916/96 unreported 16 March 2000
Wilson v First County Trust Ltd (No 2); sub nom Wilson v Secretary of State for Trade and Industry [2003] UKHL 40; [2003] 3 WLR 568; [2003] 4 All ER 97
Yearwood v The Queen [2001] UKPC 31; [2001] 5 LRC 247
This was the hearing of an appeal by the appellant, Ahmad Raja Ghaidan, from a decision of the Court of Appeal allowing an appeal of the respondent, Antonio Godin-Mendoza, from a decision of Judge Cowell, sitting in West London County Court, dismissing the appellant’s claim for possession against the respondent and declaring that he held an assured tenancy.
Monica Carrs-Frisk QC and Jonathan Small (instructed by Hugh Cartwright & Amin) appeared for the appellant; Rabinder Singh QC and Paul Staddon (instructed by Oliver Fisher) represented the respondent; Philip Sales QC and Judith Stone (instructed by the Treasury Solicitor) appeared for the First Secretary of State, the intervenor.
Giving the first opinion, Lord Nicholls said:
My lords,
[1] On the death of a protected tenant of a dwelling-house, his or her surviving spouse, if then living in the house, becomes a statutory tenant by succession. But marriage is not essential for this purpose. A person who was living with the original tenant “as his or her wife or husband” is treated as the spouse of the original tenant: see para 2(2) of Schedule 1 to the Rent Act 1977. In Fitzpatrick v Sterling Housing Association Ltd [2001] 1 AC 27, your lordships’ House decided this provision did not include persons in a same-sex relationship. The question raised by this appeal is whether this reading of para 2 can survive the coming into force of the Human Rights Act 1998. In Fitzpatrick, the original tenant had died in 1994.
[2] In the present case, the original tenant died after the Human Rights Act 1998 came into force on 2 October 2000. In April 1983, Mr Hugh Wallwyn-James was granted an oral residential tenancy of the basement flat at 17 Cresswell Gardens, London SW5. Until his death on 5 January 2001, he lived there, in a stable and monogamous homosexual relationship, with the defendant, Mr Juan Godin-Mendoza. Mr Godin-Mendoza is still living there. After the death of Mr Wallwyn-James the landlord, Mr Ahmad Ghaidan, brought proceedings in West London County Court claiming possession of the flat. Judge Cowell held that, on the death of Mr Wallwyn-James, Mr Godin-Mendoza did not succeed to the tenancy of the flat as the surviving spouse of Mr Wallwyn-James within the meaning of para 2 of Schedule 1 to the Rent Act 1977, but that he did become entitled to an assured tenancy of the flat by succession as a member of the original tenant’s “family” under para 3(1) of that Schedule.
[3] Mr Godin-Mendoza appealed, and the Court of Appeal, comprising Kennedy, Buxton and Keene LJJ, allowed the appeal: see [2002] EWCA Civ 1533; [2003] Ch 380. The court held that he was entitled to succeed to a tenancy of the flat as a statutory tenant under para 2. From that decision, Mr Ghaidan, the landlord, appealed to your lordships’ House.
[4] I must first set out the relevant statutory provisions and then explain how the Human Rights Act 1998 comes to be relevant in this case. Paragraphs 2 and 3 of Schedule 1 to the Rent Act 1977 provide:
2–(1) The surviving spouse (if any) of the original tenant, if residing in the dwelling-house immediately before the death of the original tenant, shall after the death be the statutory tenant if and so long as he or she occupies the dwelling-house as his or her residence.
(2) For the purposes of this paragraph, a person who was living with the original tenant as his or her wife or husband shall be treated as the spouse of the original tenant.
3–(1) Where paragraph 2 above does not apply, but a person who was a member of the original tenant’s family was residing with him in the dwelling-house at the time of and for the period of 2 years immediately before his death then, after his death, that person or if there is more than one such person such one of them as may be decided by agreement, or in default of agreement by the county court, shall be entitled to an assured tenancy of the dwelling-house by succession.
[5] On an ordinary reading of this language, para 2(2) draws a distinction between the position of a heterosexual couple living together in a house as husband and wife and a homosexual couple living together in a house. The survivor of a heterosexual couple may become a statutory tenant by succession, the survivor of a homosexual couple cannot. That was decided in Fitzpatrick. The survivor of a homosexual couple may, in competition with other members of the original tenant’s “family”, become entitled to an assured tenancy under para 3. But even if he does, as in the present case, this is less advantageous. Notably, so far as the present case is concerned, the rent payable under an assured tenancy is the contractual or market rent, which may be more than the fair rent payable under a statutory tenancy, and an assured tenant may be evicted for non-payment of rent without the court needing to be satisfied, as is essential in the case of a statutory tenancy, that it is reasonable to make a possession order. In these and some other respects, the succession rights granted by the statute to the survivor of a homosexual couple in respect of the house where he or she is living are less favourable than the succession rights granted to the survivor of a heterosexual couple.
[6] Mr Godin-Mendoza’s claim is that this difference in treatment infringes Article 14 of the European Convention on Human Rights (ECHR) read in conjunction with Article 8. Article 8 does not require the state to provide security of tenure for members of a deceased tenant’s family. Article 8 does not in terms give a right to be provided |page:134| with a home: see para 99 of p427 of Chapman v United Kingdom 27238/95 (2001) 33 EHRR 399. It does not “guarantee the right to have one’s housing problem solved by the authorities”: see Marzari v Italy 36448/97 (1999) 28 EHRR CD 175, at p179. But if the state makes legislative provision, it must not be discriminatory. The provision must not draw a distinction on grounds such as sex or sexual orientation without good reason. Unless justified, a distinction founded on such grounds infringes the Convention right embodied in Article 14, as read with Article 8. Mr Godin-Mendoza submits that the distinction drawn by para 2 of Schedule 1 to the Rent Act 1977 is drawn on the grounds of sexual orientation and that this difference in treatment lacks justification.
[7] That is the first step in Mr Godin-Mendoza’s claim. That step would not, of itself, improve Mr Godin-Mendoza’s status in his flat. The second step in his claim is to pray in aid the court’s duty under section 3 of the Human Rights Act 1998 to read and give effect to legislation in a way that is compliant with the Convention rights. Here, it is said, section 3 requires the court to read para 2 so that it embraces couples living together in a close and stable homosexual relationship as much as couples living together in a close and stable heterosexual relationship. So read, para 2 covers Mr Godin-Mendoza’s position. Hence, he is entitled to a declaration that, on the death of Mr Wallwyn-James, he succeeded to a statutory tenancy.
Discrimination
[8] The first of the two steps in Mr Godin-Mendoza’s argument requires him to make good the proposition that, as interpreted in Fitzpatrick, para 2 of Schedule 1 to the Rent Act 1977 infringes his Convention right under Article 14 read in conjunction with Article 8. Article 8 guarantees, among other matters, the right to respect for a person’s home. Article 14 guarantees that the rights set out in the Convention will be secured “without discrimination” on any grounds such as those stated in the non-exhaustive list in that article.
[9] It goes without saying that Article 14 is an important article of the Convention. Discrimination is an insidious practice. Discriminatory law undermines the rule of law because it is the antithesis of fairness. It brings the law into disrepute. It breeds resentment. It fosters an inequality of outlook that is demeaning alike to those unfairly benefited and those unfairly prejudiced. Of course, all law, civil and criminal, has to draw distinctions. One type of conduct, or one factual situation, attracts one legal consequence, another type of conduct or situation attracts a different legal consequence. To be acceptable, these distinctions should have a rational and fair basis. Like cases should be treated alike, unlike cases should not be treated alike. The circumstances that justify two cases being regarded as unlike, and therefore requiring or susceptible of different treatment, are infinite. In many circumstances, opinions can differ on whether a suggested ground of distinction justifies a difference in legal treatment. But there are certain grounds of factual difference that, by common accord, are not acceptable, without more, as a basis for different legal treatment. Differences of race or sex or religion are obvious examples. Sexual orientation is another. This has been clearly recognised by the European Court of Human Rights (ECtHR): see, for instance, para 32 of Fretté v France 36515/97 (2003) 2 FLR 9, at p23. Unless some good reason can be shown, differences such as these do not justify differences in treatment. Unless good reason exists, differences in legal treatment based upon grounds such as these are properly stigmatised as discriminatory.
[10] Unlike Article 1 of the 12th Protocol, Article 14 of the Convention does not confer a free-standing right of non-discrimination. It does not confer a right of non-discrimination in respect of all laws. Article 14 is more limited in its scope. It precludes discrimination in the “enjoyment of the rights and freedoms set forth in this Convention”. The court at Strasbourg has said this means that, for Article 14 to be applicable, the facts at issue must “fall within the ambit” of one or more of the Convention rights. Article 14 comes into play whenever the subject matter of the disadvantage “constitutes one of the modalities” of the exercise of a right guaranteed or whenever the measures complained of are “linked” to the exercise of a right guaranteed: see paras 22 and 28 of Petrovic v Austria (2001) 33 EHRR 14, at pp318 and 319.
[11] These expressions are not free from difficulty. In [32]-[41] of R (on the application of Carson) v Secretary of State for Work and Pensions [2003] EWCA Civ 797; [2003] 3 All ER 577, at pp592-595, Laws LJ drew attention to some difficulties existing in this area of the Strasbourg jurisprudence. In the Court of Appeal in the present case, Buxton LJ appeared to adopt the approach, espoused in the leading text book, Grosz, Beatson and Duffy, Human Rights: The 1998 Act and the European Convention (2000), at p327 in para C14-10, that “even the most tenuous link with another provision in the Convention will suffice for article 14 to enter into play”: see [9] of Ghaidan v Godin-Mendoza [2002] EWCA Civ 1533; [2003] Ch 380, at p387. In your lordships’ House, counsel for the First Secretary of State criticised this approach. He drew attention to later authorities questioning its correctness: see [21] to [22] of R (on the application of Erskine) v Lambeth London Borough Council [2003] EWHC 2479 (Admin)*, per Mitting J (“it overstates the effect of the Strasbourg case law”) and [53] to [54] of R (on the application of Douglas) v North Tyneside Metropolitan Borough Council [2003] EWCA Civ 1847; [2004] 1 All ER 709, at p722, per Scott Baker LJ.
[12] This is not a question calling for consideration on this appeal. It is common ground between all parties, and rightly so, that para 2 of Schedule 1 to the Rent Act 1977 is a provision that falls within the “ambit” of the right to respect for a person’s home guaranteed by Article 8. It is, in other words, common ground that Article 14 is engaged in the present case. This being so, and the point not having been fully argued, I prefer to leave open the question of whether even the most tenuous link is sufficient to engage Article 14.
[13] In the present case, para 2 of Schedule 1 to the Rent Act 1977 draws a dividing line between married couples and cohabiting heterosexual couples, on the one hand, and other members of the original tenant’s family, on the other hand. What is the rationale for this distinction? The rationale seems to be that, for the purposes of security of tenure, the survivor of such couples should be regarded as having a special claim to be treated in much the same way as the original tenant. The two of them made their home together in the house in question, and their security of tenure in the house should not depend upon which of them dies first.
[14] The history of the Rent Act legislation is consistent with this appraisal. A widow, who had lived with her husband, was accorded a privileged succession position in 1920. In 1980, a widower was accorded the like protection. In 1988, para 2(2) was added, by which the survivor of a cohabiting heterosexual couple would be treated in the same way as a spouse of the original tenant.
[15] Ms Monica Carss-Frisk QC submitted that there is a relevant distinction between heterosexual partnerships and same-sex partnerships. The aim of the legislation is to provide protection for the traditional family. Same-sex partnerships cannot be equated with family in the traditional sense. Same-sex partners are unable to have children with each other, and there is a reduced likelihood of children being a part of such a household.
[16] My difficulty with this submission is that there is no reason for believing that these factual differences between heterosexual and homosexual couples have any bearing upon why succession rights have been conferred upon heterosexual couples but not homosexual couples. Protection of the traditional family unit may well be an important and legitimate aim in certain contexts. In certain contexts, this may be a cogent reason for justifying differential treatment: see para 40 of Karner v Austria 40016/98 (2003) 2 FLR 623, at p630. But it is important to identify the element of the “traditional family” that para 2, as it now stands, is seeking to protect. Marriage is not now a prerequisite to protection under para 2. The line drawn by parliament is no longer drawn by reference to the status of marriage. Nor is parenthood, or the presence of children in the home, a precondition of security of tenure for the survivor of the original tenant. Nor is procreative potential a prerequisite. The survivor is protected even if, by reasons of age or otherwise, there was never any prospect of either member of the couple having a natural child.
[17] What remains, and it is all that remains, as the essential feature under para 2 is the cohabitation of a heterosexual couple. Security of tenure for the survivor of such a couple in the house where they live is, doubtless, an important and legitimate social aim. Such a couple share their lives and make their home together. Parliament may readily take the view that the survivor of them has a special claim to security of tenure even though they are unmarried. But the reason underlying this social policy, whereby the survivor of a cohabiting heterosexual couple has particular protection, is equally applicable to the survivor of a homosexual couple. A homosexual couple, as much as a heterosexual couple, share each other’s life and make their home together. They have an equivalent relationship. There is no rational or fair ground for distinguishing the one couple from the other in this context: see the discussion in Fitzpatrick, at p44.
[18] This being so, one looks in vain to find justification for the difference in treatment of homosexual and heterosexual couples. Such a difference in treatment can be justified only if it pursues a legitimate aim and there is a reasonable relationship of proportionality between the means employed and the aim sought to be realised. Here, the difference in treatment falls at the first hurdle: the absence of a legitimate aim. None has been suggested by the First Secretary of State, and none is apparent. In so far as admissibility decisions such as S v United Kingdom 11716/85 (1986) 47 DR 274 and Röösli v Germany 28318/95 (1996) 85 DR 149 adopted a different approach from that set out above, they must now be regarded as having been superseded by the recent decision of the ECtHR in Karner.
[19] For completeness, I should add that arguments based upon the extent of the discretionary area of judgment accorded to the legislature lead nowhere in this case. As noted in [70] of Wilson v First County Trust Ltd (No 2) [2003] UKHL 40; [2003] 3 WLR 568, at p589, parliament is charged with the primary responsibility for deciding the best way of dealing with social problems. The court’s role is one of review. The court will reach a different conclusion from the legislature only when it is apparent that the legislature has attached insufficient importance to a person’s Convention rights. The readiness of the court to depart from the view of the legislature depends upon the subject matter of the legislation and of the complaint. National housing policy is a field where the court will be less ready to intervene. Parliament has to hold a fair balance between the competing interests of tenants and landlords, taking into account broad issues of social and economic policy. But even in such a field, where the alleged violation comprises differential treatment based upon grounds such as race or sex or sexual orientation, the court will scrutinise with intensity any reasons said to constitute justification. The reasons must be cogent if such differential treatment is to be justified.
[20] In the present case, the only suggested ground for according different treatment to the survivor of same-sex couples and opposite-sex couples cannot withstand scrutiny. Rather, the present state of the law as set out in para 2 of Schedule 1 to the Rent Act 1977 may properly be described as a continuing adherence to the traditional regard for the position of surviving spouses, adapted in 1988 to take account of the widespread contemporary trend for men and women to cohabit outside marriage but not adapted to recognise the comparable position of cohabiting same-sex couples. I appreciate that the primary object of introducing the regime of assured tenancies and assured shorthold tenancies in 1988 was to increase the number of properties available for renting in the private sector. But this policy objective of the Housing Act 1988 can afford no justification for amending para 2 so as to include cohabiting heterosexual partners but not cohabiting homosexual partners. This policy objective of the Act provides no reason for, on the one hand, extending to unmarried cohabiting heterosexual partners the right to succeed to a statutory tenancy but, on the other hand, withholding that right from cohabiting homosexual partners. Paragraph 2 fails to attach sufficient importance to the Convention rights of cohabiting homosexual couples.
[21] Ms Carss-Frisk advanced a further argument, based upon the decisions of the ECtHR in Walden v Liechtenstein 33916/96 unreported 16 March 2000 and Petrovic. In Walden, the Liechtenstein constitutional court held that the unequal pension treatment afforded to married and unmarried couples was unconstitutional. The constitutional court did not set aside the existing legislation, given the practical difficulties involved and given also that a comprehensive legal reform guaranteeing gender equality in social security law was in the course of preparation. New legislation was enacted and came into force seven months later. The ECtHR summarily rejected an application complaining of this unequal treatment. The constitutional court’s decision served the interests of legal certainty, and, given the brevity of the period during which the unconstitutional law remained applicable to the applicant, the continued operation of the pension provisions was proportionate. In Petrovic, the applicant was refused a grant of parental leave allowance in 1989. At that time, parental leave allowance was available only to mothers. The applicant complained that this violated Article 14 taken together with Article 8. In dismissing the application, the court noted that, as society moved towards a more equal sharing of responsibilities for the upbringing of children, contracting states have extended allowances such as parental leave to fathers. Austrian law had evolved in this way, eligibility for parental leave allowance being extended to fathers in 1990. The Austrian legislature was not to be criticised for having introduced progressive legislation in a gradual manner.
[22] Ms Carss-Frisk submitted that, similarly here, society’s attitude to cohabiting homosexual couples has evolved considerably in recent years. It was only in July 2003 that the ECtHR, in Karner, effectively overruled contrary decisions, as already mentioned. The United Kingdom government responded speedily to the decision in Karner by including in two government bills currently before parliament, the Housing Bill and the Civil Partnership Bill, provisions that, if enacted, would have the effect of confirming on the face of legislation that the survivor of a cohabiting homosexual couple is to be treated in the same way as the survivor of a cohabiting homosexual couple for the purposes of para 2. The state should not be criticised for this gradual extension of the rights of cohabiting unmarried couples, first to heterosexual couples in 1988, and now more widely. The extension of para 2 to include homosexual couples would be at the expense of landlords, and, in the interests of legal certainty, this extension should be made prospectively by legislation and not retrospectively by judicial decision. Mr Wallwyn-James died more than two years before the decision in Karner.
[23] I am unable to accept this submission. Under the Human Rights Act 1998, the compatibility of legislation with the Convention rights falls to be assessed when the issue arises for determination, not as at the date when the legislation was enacted or came into force: see para 62 of Wilson, at p587. Walden and Petrovic concerned the margin of appreciation afforded to contracting states. In the present case, the House is concerned with the interpretation and application of domestic legislation. In this context, the domestic counterpart of a state’s margin of appreciation is the discretionary area of judgment the court accords parliament when reviewing legislation pursuant to its obligations under the Human Rights Act 1998. I have already set out my reasons for holding that, in the present case, the distinction drawn in the legislation between the position of heterosexual couples and homosexual couples falls outside that discretionary area.
[24] In my view, therefore, Mr Godin-Mendoza makes good the first step in his argument: para 2 of Schedule 1 to the Rent Act 1977, construed without reference to section 3 of the Human Rights Act 1988, violates his Convention right under Article 14 taken together with Article 8.
Section 3 of the Human Rights Act 1998
[25] I turn next to the question of whether section 3 of the Human Rights Act 1998 requires the court to depart from the interpretation of para 2 enunciated in Fitzpatrick. |page:136|
[26] Section 3 is a key section in the Human Rights Act 1998. It is one of the primary means by which Convention rights are brought into the law of this country. Parliament has decreed that all legislation, existing and future, will be interpreted in a particular way. All legislation must be read and given effect to in a way that is compatible with the Convention rights “so far as it is possible to do so”. This is the intention of parliament, expressed in section 3, and the courts must give effect to this intention.
[27] Unfortunately, in making this provision for the interpretation of legislation, section 3 itself is not free from ambiguity. Section 3 is open to more than one interpretation. The difficulty lies in the word “possible”. Section 3(1), read in conjunction with section 3(2) and section 4, makes one matter clear: parliament expressly envisaged that not all legislation would be capable of being made Convention-compliant by the application of section 3. Sometimes it would be possible, sometimes not. What is not clear is the test to be applied in separating the sheep from the goats. What is the standard, or the criterion, by which “possibility” is to be judged? A comprehensive answer to this question is proving elusive. The courts, including your lordships’ House, are still cautiously feeling their way forward as experience in the application of section 3 gradually accumulates.
[28] One tenable interpretation of the word “possible” would be that section 3 is confined to requiring the courts to resolve ambiguities. Where the words under consideration fairly admit of more than one meaning, the Convention-compliant meaning is to prevail. Words should be given the meaning which best accords with the Convention rights.
[29] This interpretation of section 3 would give the section a comparatively narrow scope. This is not the view that has prevailed. It is now generally accepted that the application of section 3 does not depend upon the presence of ambiguity in the legislation being interpreted. Even if, construed according to the ordinary principles of interpretation, the meaning of the legislation admits of no doubt, section 3 may none the less require the legislation to be given a different meaning. The decision of your lordships’ House in R v A (No 2) [2001] UKHL 25; [2002] 1 AC 45 is an instance of this. The House read words into section 41 of the Youth Justice and Criminal Evidence Act 1999 so as to make that section compliant with an accused’s right to a fair trial under Article 6. The House did so even though the statutory language was not ambiguous.
[30] From this it follows that the interpretative obligation decreed by section 3 is of an unusual and far-reaching character. Section 3 may require a court to depart from the unambiguous meaning that the legislation would otherwise bear. In the ordinary course, the interpretation of legislation involves seeking the intention reasonably to be attributed to parliament in using the language in question. Section 3 may require the court to depart from this legislative intention, that is, depart from the intention of the parliament that enacted the legislation. The question of difficulty is how far, and in what circumstances, section 3 requires a court to depart from the intention of the enacting parliament. The answer to this question depends upon the intention reasonably to be attributed to parliament in enacting section 3.
[31] On this, the first point to be considered is how far, when enacting section 3, parliament had intended that the actual language of a statute, as distinct from the concept expressed in that language, should be determinative. Since section 3 relates to the “interpretation” of legislation, it is natural to focus attention initially upon the language used in the legislative provision being considered. But once it is accepted that section 3 may require legislation to bear a meaning that departs from the unambiguous meaning the legislation would otherwise bear, it becomes impossible to suppose that parliament had intended that the operation of section 3 should depend critically upon the particular form of words adopted by the parliamentary draftsman in the statutory provision under consideration. That would make the application of section 3 something of a semantic lottery. If the draftsman chose to express the concept being enacted in one form of words, section 3 would be available to achieve Convention-compliance. If he chose a different form of words, section 3 would be impotent.
[32] From this, the conclusion that seems inescapable is that the mere fact the language under consideration is inconsistent with a Convention-compliant meaning does not of itself make a Convention-compliant interpretation under section 3 impossible. Section 3 enables language to be interpreted restrictively or expansively. But section 3 goes further than this. It is also apt to require a court to read in words that change the meaning of the enacted legislation, so as to make it Convention-compliant. In other words, the intention of parliament in enacting section 3 was that, to an extent bounded only by what is “possible”, a court can modify the meaning, and, hence, the effect, of primary and secondary legislation.
[33] Parliament, however, cannot have intended that, in the discharge of this extended interpretative function, the courts should adopt a meaning inconsistent with a fundamental feature of legislation. That would be to cross the constitutional boundary section 3 seeks to demarcate and preserve. Parliament has retained the right to enact legislation in terms that are not Convention-compliant. The meaning imported by the application of section 3 must be compatible with the underlying thrust of the legislation being construed. Words implied must, in the phrase of my noble and learned friend Lord Rodger of Earlsferry, “go with the grain of the legislation”. Nor can parliament have intended that section 3 should require the courts to make decisions for which they are not equipped. There may be several ways of making a provision Convention-compliant, and the choice may involve issues calling for legislative deliberation.
[34] Both these features were present in In re S (Care order: Implementation of care plan) [2002] UKHL 10; [2002] 2 AC 291. There, the proposed “starring system” was inconsistent in an important respect with the scheme of the Children Act 1989, and the proposed system had far-reaching practical ramifications for local authorities. Again, in R (on the application of Anderson) v Secretary of State for the Home Department [2003] UKHL 46; [2003] 1 AC 837, section 29 of the Crime (Sentences) Act 1997 could not be read in a Convention-compliant way without giving the section a meaning that would be inconsistent with an important feature expressed clearly in the legislation. In Bellinger v Bellinger [2003] UKHL 21; [2003] 2 AC 467, the recognition of Mrs Bellinger as female for the purposes of section 11(c) of the Matrimonial Causes Act 1973 would have had exceedingly wide ramifications, raising issues ill-suited for determination by the courts or by court procedures.
[35] In some cases, difficult problems may arise. No difficulty arises in the present case. Paragraph 2 of Schedule 1 to the Rent Act 1977 is unambiguous. But the social policy underlying the 1988 extension of security of tenure under para 2 to the survivor of couples living together as husband and wife is equally applicable to the survivor of homosexual couples living together in a close and stable relationship. In this circumstance, I see no reason to doubt that application of section 3 to para 2 has the effect that para 2 should be read and given effect to as though the survivor of such a homosexual couple were the surviving spouse of the original tenant. Reading para 2 in this way would have the result that cohabiting heterosexual couples and cohabiting heterosexual couples would be treated alike for the purposes of succession as a statutory tenant. This would eliminate the discriminatory effect of para 2 and would do so consistently with the social policy underlying para 2. The precise form of words read in for this purpose is of no significance. It is their substantive effect that matters.
[36] For these reasons, I agree with the decision of the Court of Appeal. I would dismiss this appeal.
Giving the second opinion, Lord Steyn said:
My lords,
[37] In my view, the Court of Appeal came to the correct conclusion. I agree with the conclusions and reasons of my noble and learned friends Lord Nicholls of Birkenhead, Lord Rodger of Earlsferry and Baroness Hale of Richmond. In the light of those opinions, I will not comment on the case generally.
[38] I confine my remarks to the question of whether it is possible, under section 3(1) of the Human Rights Act 1998, to read and give |page:137| effect to para 2(2) of Schedule 1 to the Rent Act 1977 in a way that is compatible with the ECHR. In my view, the interpretation adopted by the Court of Appeal under section 3(1) was a classic illustration of the permissible use of this provision. But it became clear during oral argument, and from a subsequent study of the case law and academic discussion on the correct interpretation of section 3(1), that the role of that provision in the remedial scheme of the 1998 Act is not always correctly understood. I would therefore wish to examine the position in a general way.
[39] I attach an appendix (not reproduced here) to this opinion that lists cases where a breach of an ECHR right was found established, and the courts proceeded to consider whether to exercise their interpretative power under section 3 or to make a declaration of incompatibility under section 4. [Table not reproduced in this report]. For the first and second lists (A and B), I am indebted to the Constitutional Law Division of the Department of Constitutional Affairs, but law report references and other information have been added. The third list (C) has been prepared by Laura Johnson, my judicial assistant, under my direction. It will be noted that, in 10 cases, the courts used their interpretative power under section 3 and, in 15 cases, the courts made declarations of incompatibility under section 4. In five cases in the second group, the declarations of incompatibility were subsequently reversed on appeal: in four of those cases, it was held that no breach had been established, and, in the fifth case (R (on the application of Hooper) v Secretary of State for Work and Pensions [2003] EWCA Civ 813; [2003] 1WLR 2623), the exact basis for overturning the declaration of incompatibility may be a matter of debate. Given that, under the 1998 Act, the use of the interpretative power under section 3 is the principal remedial measure, and that the making of a declaration of incompatibility is a measure of last resort, these statistics by themselves raise a question about the proper implementation of the 1998 Act. A study of the case law reinforces the need to pose the question of whether the law has taken a wrong turning.
[40] My impression is that two factors are contributing to a misunderstanding of the remedial scheme of the 1998 Act. First, there is the constant refrain that a judicial reading down, or reading in, under section 3 would flout the will of parliament as expressed in the statute under examination. This question cannot sensibly be considered without giving full weight to the countervailing will of parliament as expressed in the 1998 Act.
[41] The second factor may be an excessive concentration on linguistic features of the particular statute. Nowhere in our legal system is a literalistic approach more inappropriate than when considering whether a breach of a Convention right may be removed by interpretation under section 3. Section 3 requires a broad approach concentrating, among other things, in a purposive way, upon the importance of the fundamental right involved.
[42] In enacting the 1998 Act, parliament legislated “to bring rights home” from the ECtHR to be determined in the courts of the United Kingdom. That is what the White Paper said: see para 2.7 of Rights Brought Home: The Human Rights Bill (1997) (Cm 3782). That is what parliament was told. The mischief to be addressed was the fact that Convention rights as set out in the ECHR, which Britain ratified in 1951, could not be vindicated in our courts. Critical to this purpose was the enactment of effective remedial provisions.
[43] The provisions adopted read as follows:
Interpretation of legislation
3.–(1) So far as it is possible to do so, primary legislation and subordinate legislation must be read and given effect in a way which is compatible with the Convention rights.
(2) This section
(a) applies to primary legislation and subordinate legislation whenever enacted;
(b) does not affect the validity, continuing operation or enforcement of any incompatible primary legislation; and
(c) does not affect the validity, continuing operation or enforcement of any incompatible subordinate legislation if (disregarding any possibility of revocation) primary legislation prevents removal of the incompatibility.
Declaration of incompatibility
4.–(1) Subsection (2) applies in any proceedings in which a court determines whether a provision of primary legislation is compatible with a Convention right.
(2) If the court is satisfied that the provision is incompatible with a Convention right, it may make a declaration of that incompatibility.
…
If parliament disagrees with an interpretation by the courts under section 3(1), it is free to override it by amending the legislation and expressly reinstating the incompatibility.
[44] It is necessary to state what section 3(1), and, in particular, the word “possible”, does not mean. First, section 3(1) applies even if there is no ambiguity in the language in the sense of it being capable of bearing two possible meanings. The word “possible” in section 3(1) is used in a different and much stronger sense. Second, section 3(1) imposes a stronger and more radical obligation than to adopt a purposive interpretation in the light of the ECHR. Third, the draftsman of the Act had before him the model of the New Zealand Bill of Rights Act 1990, which imposes a requirement that the interpretation to be adopted must be reasonable. Parliament specifically rejected the legislative model of requiring a reasonable interpretation.
[45] Instead, the draftsman had resort to the analogy of the obligation under the EEC Treaty on national courts, as far as possible, to interpret national legislation in the light of the wording and purpose of directives. In Marleasing SA v La Comercial Internacional de Alimentación SA C-106/89 [1990] ECR I-4135, at p4159, the European Court of Justice defined this obligation as follows:
It follows that, in applying national law, whether the provisions in questions were adopted before or after the directive, the national court called upon to interpret it is required to do so, as far as possible, in light of the wording and the purpose of the directive in order to achieve the result pursued by the latter and thereby comply with the third paragraph of Article 189 of the Treaty.
Given the undoubted strength of this interpretative obligation under EEC law, this is a significant signpost to the meaning of section 3(1) of the 1998 Act.
[46] Parliament had before it the mischief and objective sought to be addressed, viz the need “to bring rights home”. The lynchpin of the legislative scheme to achieve this purpose was section 3(1). Rights could be effectively brought home only if section 3(1) was the prime remedial measure, and section 4 a measure of last resort. How the system modelled on the EEC interpretative obligation would work was graphically illustrated for parliament during the progress of the bill through both Houses. The Lord Chancellor observed that “in 99% of the cases that will arise, there will be no need for judicial declarations of incompatibility”, and the Home Secretary said: “We expect that, in almost all cases, the courts will be able to interpret the legislation compatibly with the Convention”: Hansard (HL Debates) 5 February 1998, at col 840 (3rd reading) and Hansard (HC Debates) 16 February 1998, at col 778 (2nd reading). It was envisaged that the duty of the court would be to strive to find (if possible) a meaning that would best accord with Convention rights. This is the remedial scheme that parliament adopted.
[47] Three decisions of the House can be cited to illustrate the strength of the interpretative obligation under section 3(1). The first is R v A (No 2) [2001] UKHL 25; [2002] 1 AC 45, which concerned the so-called rape-shield legislation. The problem was the blanket exclusion of prior sexual history between the complainant and an accused contained in section 41(1) of the Youth Justice and Criminal Evidence Act 1999, subject to narrow specific categories in the remainder of section 41. In subsequent decisions, and in academic literature, there has been discussion about differences of emphasis in the various opinions in A. What has been largely overlooked is the unanimous conclusion of the House. The House unanimously agreed on an interpretation under section 3 that would ensure that section 41 would be compatible with the ECHR. The formulation was, by agreement, set out in [46] of my opinion in that case, as follows: |page:138|
The effect of the decision today is that under section 41(3)(c) of the 1999 Act, construed where necessary by applying the interpretive obligation under section 3 of the Human Rights Act 1998, and due regard always being paid to the importance of seeking to protect the complainant from indignity and from humiliating questions, the test of admissibility is whether the evidence (and questioning in relation to it) is nevertheless so relevant to the issue of consent that to exclude it would endanger the fairness of the trial under article 6 of the Convention. If this test is satisfied the evidence should not be excluded.
This formulation was endorsed by Lord Slynn, in [13] at p56 of his opinion, in identical wording. The other law lords sitting in the case expressly approved the formulation set out in [46] of my opinion: see Lord Hope in [110], at pp87-88; Lord Clyde in [140], at p98; and Lord Hutton in [163], at p106. In so ruling, the House rejected linguistic arguments in favour of a broader approach. In the subsequent decisions of the House in In re S (Care Order: Implementation of Case Plan) [2002] UKHL 10; [2002] 2 AC 291 and Bellinger v Bellinger [2003] UKHL 21; [2003] 2 AC 467, which touched upon the remedial structure of the 1998 Act, the decision of the House in A was not questioned. And, in the present case, nobody suggested that A involved a heterodox exercise of the power under section 3.
[48] The second and third decisions of the House are Pickstone v Freemans plc [1989] AC 66 and Litster v Forth Dry Dock & Engineering Co Ltd [1990] 1 AC 546, which involve the interpretative obligation under EEC law. Pickstone concerned section 1(2) of the Equal Pay Act 1970 (as amended by section 8 of the Sex Discrimination Act 1975, and regulation 2 of the Equal Pay (Amendment) Regulations 1983 (SI 1983/1794)), which implied into any contract without an equality clause one that modifies any term in a woman’s contract that is less favourable than a term of a similar kind in the contract of a man:
(a) where the woman is employed on like work with a man in the same employment…
(b) where the woman is employed on work rated as equivalent with that of a man in the same employment…
(c) where a woman is employed on work which, not being work in relation to which paragraph (a) or (b) above applies, is, in terms of the demands made on her (for instance under such headings as effort, skill and decision), of equal value to that of a man in the same employment.
Lord Templeman observed, at pp120-121:
In my opinion there must be implied in paragraph (c) after the word “applies” the words “as between the woman and the man with whom she claims equality”. This construction is consistent with Community law. The employers’ construction is inconsistent with Community law and creates a permitted form of discrimination without rhyme or reason.
That was the ratio decidendi of the decision. Litster concerned regulations intended to implement an EC directive, the purpose of which was to protect the workers in an undertaking when its ownership was transferred. However, the regulations protected only those who were employed “immediately before” the transfer. Having enquired into the purpose of the directive, the House of Lords interpreted the regulations by reading in additional words to protect workers not only if they were employed “immediately before” the time of transfer but also when they would have been so employed if they had not been unfairly dismissed by reason of the transfer: see Lord Keith, at p554. In both cases, the House eschewed linguistic arguments in favour of a broad approach. Pickstone and Litster involved national legislation that implemented EC directives. Marleasing extended the scope of the interpretative obligation to unimplemented directives. Pickstone and Litster reinforce the approach to section 3(1) that prevailed in the House in the rape-shield case.
[49] A study of the case law listed in the appendix to this judgment reveals that there has sometimes been a tendency to approach the interpretative task under section 3(1) in too literal and technical a way. In practice there has been too much emphasis upon linguistic features. If the core remedial purpose of section 3(1) is not to be undermined a broader approach is required. That is, of course, not to gainsay the obvious proposition that inherent in the use of the word “possible” in section 3(1) is the idea that there is a Rubicon that courts may not cross. If it is not possible, within the meaning of section 3, to read or give effect to legislation in a way that is compatible with Convention rights, the only alternative is to exercise, where appropriate, the power to make a declaration of incompatibility. Usually, such cases should not be too difficult to identify. An obvious example is R (on the application of Anderson) v Secretary of State for the Home Department [2002] UKHL 46; [2003] 1 AC 837. The House held that the Home Secretary was not competent under Article 6 of the ECHR to decide on the tariff to be served by mandatory life sentence prisoners. The House found a section 3(1) interpretation not “possible”, and made a declaration under section 4. Interpretation could not provide a substitute scheme. Bellinger is another obvious example. As Lord Rodger observed — “in relation to the validity of marriage, Parliament regards gender as fixed and immutable”: see [83] of [2003] 2 WLR 1174, at p1195. Section 3(1) of the 1998 Act could not be used.
[50] Having had the opportunity to reconsider the matter in some depth, I am not disposed to try to formulate precise rules about where section 3 may not be used. Like the proverbial elephant, such a case ought generally to be easily identifiable. What is necessary, however, is to emphasise that interpretation under section 3(1) is the prime remedial remedy, and that resort to section 4 must always be an exceptional course. In practical effect, there is a strong rebuttable presumption in favour of an interpretation consistent with Convention rights. Perhaps the opinions delivered in the House today will serve to ensure a balanced approach along such lines.
[51] I now return to the circumstances of the case before the House. Applying section 3, the Court of Appeal interpreted “as his or her wife or husband” in the statute to mean “as if they were his wife or husband”. Although there has been some controversy about aspects of the reasoning of the Court of Appeal, I would endorse the reasoning of the Court of Appeal on the use of section 3(1) in this case. It was well within the power under this provision.
[52] I would also dismiss the appeal.
Giving the third opinion, Lord Millett said:
My lords,
[53] Paragraphs 2 and 3 of Schedule 1 to the Rent Act 1977, as amended by the Housing Act 1988, provide:
2.–(1) The surviving spouse (if any) of the original tenant, if residing in the dwelling-house immediately before the death of the original tenant, shall after the death be the statutory tenant if and so long as he or she occupies the dwelling-house as his or her residence.
(2) For the purposes of this paragraph, a person who was living with the original tenant as his or her wife or husband shall be treated as the spouse of the original tenant.…
3.–(1) Where paragraph 2 above does not apply, but a person who was a member of the original tenant’s family was residing with him [in the dwelling-house] at the time of and for the [period of 2 years] immediately before his death then, after his death, that person or if there is more than one such person such one of them as may be decided by agreement, or in default of agreement by the county court, shall be [entitled to an assured tenancy…].
[54] As my noble and learned friend Lord Nicholls has observed, and as this House decided in Fitzpatrick v Sterling Housing Association Ltd [2001] 1 AC 27, on an ordinary reading of para 2(2), the survivor of two persons of the opposite sex living together as man and wife in a dwelling-house that is subject to the Rent Acts has a statutory right to succeed to the statutory tenancy of the deceased tenant, but the survivor of two persons of the same sex living together in similar circumstances has no such right. He or she merely has a right, in competition with other members of the deceased tenant’s family, to claim an assured tenancy, but not only is an assured tenancy less advantageous than a statutory tenancy but the survivor’s entitlement, if disputed by other members of the late tenant’s family, is at the discretion of the court.
[55] I agree with all my noble and learned friends, whose speeches I have had the advantage of reading in draft, that such discriminatory treatment of homosexual couples is incompatible with their Convention rights and cannot be justified by any identifiable legitimate aim. I am, moreover, satisfied by the powerful and convincing speech of my noble and learned friend Baroness Hale that, for the reasons she gives, such treatment is not only incompatible with the Convention but is unacceptable in a modern democratic society at the beginning of the 21st century. This is not to say that it was always, or even until fairly recently, unacceptable, but times change, and, with them, society’s perceptions change also (a commonplace usually dignified by being rendered in Latin).
[56] It follows that, unless the court can apply section 3 of the Human Rights Act 1998 to extend the reach of para 2(2) to the survivor of a couple of the same sex, it must consider making a declaration of incompatibility under section 4. The making of such a declaration is in the court’s discretion (section 4 provides only that the court “may” make one), and it may be a matter for debate as to whether it would be appropriate to do so at a time when, not merely has the government announced its intention to bring forward corrective legislation in due course (as in Bellinger v Bellinger [2003] UKHL 21; [2003] 2 AC 467), but parliament is currently engaged in enacting remedial legislation. It is, however, unnecessary to enter upon this question, for there is a clear majority in favour of the view that section 3 can be applied to interpret para 2(2) in a way that renders legislative intervention unnecessary.
[57] I have the misfortune to be unable to agree with this conclusion. I have given long and anxious consideration to the question of whether, in the interests of unanimity, I should suppress my dissent, but I have come to the conclusion that I should not. The question is of great constitutional importance, for it goes to the relationship between the legislature and the judiciary, and, hence, ultimately to the supremacy of parliament. Sections 3 and 4 of the Human Rights Act 1998 were carefully crafted to preserve the existing constitutional doctrine, and any application of the ambit of section 3 beyond its proper scope subverts it. This is not to say that the doctrine of parliamentary supremacy is sacrosanct, but only that any change in a fundamental constitutional principle should be the consequence of deliberate legislative action and not judicial activism, however well meaning.
[58] Sections 3 and 4 of the Human Rights Act 1998, so far as material, provide as follows:
Interpretation of legislation
3.–(1) So far as it is possible to do so, primary legislation and subordinate legislation must be read and given effect in a way which is compatible with the Convention rights.…
Declaration of incompatibility
4.–(1) Subsection (2) applies in any proceedings in which a court determines whether a provision of primary legislation is compatible with a Convention right.
(2) If the court is satisfied that the provision is incompatible with a Convention right, it may make a declaration of that incompatibility…
…
(6) A declaration under this section (“a declaration of incompatibility”) —
(a) does not affect the validity, continuing operation or enforcement of the provision in respect of which it is given…
[59] Several points may be made at the outset. First, the requirement in section 3 is obligatory. In R v Director of Public Prosecutions, ex parte Kebilene [2000] 2 AC 326, at p373, Lord Cooke described the section as “a strong adjuration” by parliament to read and give effect to legislation in a way that is compatible with Convention rights. With respect, it is more than this. It is a command. Legislation “must” be read and given effect to in a way that is compatible with Convention rights. There is no residual discretion to disobey the obligation that the section imposes.
[60] Second, the obligation arises (or at least has significance) only where the legislation in its natural and ordinary meaning, that is to say, as construed in accordance with normal principles, is incompatible with the Convention. Ordinary principles of statutory construction include a presumption that parliament does not intend to legislate in a way that would put the United Kingdom in breach of its international obligations. This presumption will often be sufficient to enable the court to interpret the statute in a way that will make it compatible with the Convention without recourse to section 3. It is only where this is not the case that section 3 comes into play. When it does, it obliges the court to give an abnormal construction to the statutory language and one that cannot be achieved by resort to standard principles and presumptions.
[61] This is a difficult exercise, for it is one that the courts have not hitherto been accustomed to perform, and where they must accordingly establish their own ground rules for the first time. It is also dangerously seductive, for there is bound to be a temptation to apply the section beyond its proper scope and trespass upon the prerogative of parliament in what will almost invariably be a good cause.
[62] Third, there are limits to the extent to which section 3 may be applied to render existing legislation compatible with the Convention. The presence of section 4 alone shows this to be the case, for it presupposes the existence of cases where the offending legislation cannot be rendered compatible with the Convention by the application of section 3.
[63] There are two limitations to its application that are expressed in section 3 itself. In the first place, the exercise that the court is called upon to perform is still one of interpretation, not legislation: legislation must be “read and given effect to”. Section 3 is in marked contrast with the provisions in the constitutions of former colonial territories in relation to existing laws that are incompatible with constitutional rights. Such provisions commonly authorise the court to construe such laws “with such modifications, adaptations, qualifications and exceptions as may be necessary to bring them into conformity with the constitution”.
[64] This is a quasi-legislative power, not a purely interpretative one; for the court is not constrained by the language of the statute in question, which it may modify (that is, amend) in order to bring it into conformity with the constitution. In R v Hughes [2002] UKPC 12; [2002] 2 AC 259, the Privy Council deleted (that is, repealed) express words in the statute. In doing so, it exercised a legislative, not an interpretative, power. Such a power is appropriate where the constitution (particularly one based upon the separation of powers) is the supreme law, and where statutes that are inconsistent with the constitution are, to the extent of the inconsistency, automatically rendered void by the constitution. A finding of inconsistency may leave a lacuna in the statute book that in many cases must be filled without delay if chaos is to be avoided and that can be filled only by the exercise of a legislative power. But it is not appropriate in the United Kingdom, which has no written constitution and where the prevailing constitutional doctrine is based upon the supremacy of parliament rather than the separation of powers. Accordingly, section 4(6) provides that legislation that is incompatible with a Convention right is not thereby rendered void, nor is it invalidated by the making of a declaration of incompatibility. It continues in full force and effect unless and until it is repealed or amended by parliament, which can decide whether to change the law and, if so, from what date and whether retrospectively or not.
[65] In some cases (In re S (Minors)(Care Order: Implementation of Care Plan) [2002] UKHL 10; [2002] 2 AC 291 and R (on the application of Anderson) v Secretary of State for the Home Department [2002] UKHL 46; [2003] 1 AC 837 are examples), it would have been necessary to repeal the statutory scheme and substitute another. This is obviously impossible without legislation, and cannot be achieved by resort to section 3. That is not so in the present case. In other cases (Bellinger is an example), questions of social policy have arisen that ought properly to be left to parliament and not decided by the judges. I shall return to this point later.
[66] In the second place, section 3 requires the court to read legislation in a way that is compatible with the Convention only “so far as it is possible to do so”. It must therefore be possible, by a process of interpretation alone, to read the offending statute in a way that is compatible with the Convention.
[67] This does not mean that it is necessary to identify an ambiguity or absurdity in the statute (in the sense of being open to more than one interpretation) before giving it an abnormal meaning in order to bring it into conformity with a Convention right: see R v A (No 2) [2001] UKHL 25; [2002] 1 AC 45, at p67 and 87 per Lord Steyn and Lord Hope. I respectfully agree with my noble and learned friend, Lord Nicholls, that even if, construed in accordance with ordinary principles of construction, the meaning of the legislation admits of no doubt, section 3 may require it to be given a different meaning. It means only that the court must take the language of the statute as it finds it and give it a meaning that, however unnatural or unreasonable, is intellectually defensible. It can read in and read down; it can supply missing words, so long as they are consistent with the fundamental features of the legislative scheme; it can do considerable violence to the language and stretch it almost (but not quite) to breaking point. Per Lord Steyn in [44] of A, at p67: The court must
strive to find a possible interpretation compatible with Convention rights…
(Emphasis added.)
But it is not entitled to give it an impossible one, however much it would wish to do so.
[68] In my view, section 3 does not entitle the court to supply words which are inconsistent with a fundamental feature of the legislative scheme; nor to repeal, delete, or contradict the language of the offending statute. As Lord Nicholls said in Rojas Berllaque (Attorney General for Gibraltar intervening) [2003] UKPC 76; [2004] 1 WLR 201, in [24], at p208-209:
There may of course be cases where an offending law does not lend itself to a sensible interpretation which would conform to the relevant Constitution.
This is more likely to be the case in the United Kingdom, where the court’s role is exclusively interpretative than in those territories (which include Gibraltar) where it is quasi-legislative.
[69] I doubt that the principles that I have endeavoured to state would be disputed; disagreement is likely to lie in their application in a particular case. So it may be helpful if I give some examples of the way in which I see section 3 as operating.
[70] In the course of his helpful argument, counsel for the Secretary of State, who did not resist the application of section 3, acknowledged that it could not be used to read “black” as meaning “white”. That must be correct. Words cannot mean their opposite: “black” cannot mean “not black”. But they may include their opposite. In some contexts, it may be possible to read “black” as meaning “black or white”; in other contexts, it may be impossible to do so. It all depends upon whether “blackness” is the essential feature of the statutory scheme, and, although the court may look behind the words of the statute, they cannot be disregarded or given no weight, for they are the medium by which parliament expresses its intention.
[71] Again, “red, blue or green” cannot be read as meaning “red, blue, green or yellow”; the specification of three only of the four primary colours indicates a deliberate omission of the fourth (unless, of course, this can be shown to be an error). Section 3 cannot be used to supply the missing colour, for this would be not to interpret the statutory language but to contradict it.
[72] The limits on the application of section 3 may thus be, in part at least, linguistic, as in the examples I have given, but they may also be derived from a consideration of the legislative history of the offending statute. Thus, although it may be possible to read “cats” as meaning “cats or dogs” (on the footing that the essential concept is that of domestic pets generally, rather than felines particularly), it would obviously not be possible to read “Siamese cats” as meaning “Siamese cats or dogs”. The particularity of the expression “Siamese cats” would preclude its extension to other species of cat, let alone dogs. But suppose the statute merely said “cats”, and that this was the result of successive amendments to the statute as originally enacted. If this had said “Siamese cats” and had twice been amended, first to read “Siamese or Persian cats” and then to read simply “cats”, it would not, in my opinion, be possible to read the word “cats” as including “dogs”. The legislative history would demonstrate that, although parliament had successively widened the scope of the statute, it had consistently legislated in relation to felines, and had left its possible extension to other domestic pets for future consideration. Reading the word “cats” as meaning “cats or dogs” in these circumstances would be to usurp the function of parliament.
[73] In A, the offending statute had laid down an elaborate scheme to prevent the defendant to a charge of rape from adducing certain kinds of evidence at his trial. Read without qualification, this could exclude logically relevant evidence favourable to the accused and deny him a fair trial, contrary to Article 6 of the Convention. The House read the statute as subject to the implied proviso that evidence or questioning that was required to ensure a fair trial should not be treated as inadmissible. The House supplied a missing qualification that significantly limited the operation of the statute but that did not contradict any of its fundamental features. As Lord Steyn observed in [45] at p68, it would be unrealistic to suppose that parliament, if alerted to the problem, would have wished to deny an accused person the right to put forward a full and complete defence by advancing truly probative material.
[74] For my own part, I have no difficulty with the conclusion that the House reached in that case. The qualification that it supplied glossed but did not contradict anything in the relevant statute. Neither expressly nor implicitly did the statute require logically probative evidence to be excluded if its exclusion would have the effect of denying the accused a fair trial. The meaning of the statute was not ambiguous, and, in the absence of section 3, the proviso could not have been implied. But if it had been expressed, it would not have made the statute self-contradictory or produced a nonsense.
[75] Lord Hope, who had more difficulty in the application of section 3, observed (loc cit) that compatibility was to be achieved only in so far as this was possible, and that it would plainly not be possible if the legislation contained provisions that expressly contradicted the meaning that the enactment would have to be given to make it compatible. He added that the same result must follow if they did so by necessary implication, as this too was a means of identifying the plain intention of parliament. Lord Steyn said the same in [59] of Anderson, at p894:
Section 3(1) is not available where the suggested interpretation is contrary to express statutory words or is by implication necessarily contradicted by the statute.
citing in support Lord Nicholls in [41] of In re S (Minors), at pp313-314.
(Emphasis added.)
[76] I respectfully agree with this approach, although I would add a caveat. I do not understand the word “implication” as entitling the court to imply words that would render the statute incompatible with the Convention; that would be entirely contrary to the spirit of section 3. They mean only that the incompatibility need not be explicit, but, if not, it must be implicit, that is to say, manifest on the face of the statute.
[77] It is obvious that if para 2(2) of Schedule 1 to the Rent Act 1977, as amended, had referred expressly to “a person of the opposite sex” who was living with the original tenant as his or her husband or wife, it would not be possible to bring the paragraph into conformity with the Convention by resort to section 3. The question is whether the words “of the opposite sex” are implicit, for, if they are, the same result must follow. Reading the paragraph as referring to persons whether of the same or of the opposite sex would equally contradict the legislative intent in either case. I agree that the operation of section 3 does not depend critically upon the form of words found in the statute; the court is not engaged in a parlour game. But it does depend upon identifying the essential features of the legislative scheme, and these must be gathered, in part at least, from the words that parliament has chosen to use. Drawing the line between the express and the implicit would be to engage in precisely that form of semantic lottery to which the majority rightly object.
[78] In the present case, both the language of para 2(2) and its legislative history show that the essential feature of the relationship that parliament had in contemplation was an open relationship between persons of the opposite sex. I take the language first. Paragraph 2(1) provides that “the surviving spouse” of the deceased tenant shall succeed to the statutory tenancy. The word “spouse” means a party to a |page:141| lawful marriage. It may refer indifferently to a lawfully wedded husband or a lawfully wedded wife, and to this extent is not gender-specific. But it is gender-specific in relation to the other party to the relationship. Marriage is the lawful union of a man and a woman. It is a legal relationship between persons of the opposite sex. A man’s spouse must be a woman; a woman’s spouse must be a man. This is the very essence of the relationship, which need not be loving, sexual, stable, faithful, long-lasting or contented. Although it may be brought to an end as a legal relationship only by death or by an order of the court, its demise as a factual relationship will usually have ended long before that.
[79] Another basic feature of marriage is that it is an openly acknowledged relationship. From the earliest times, marriage has involved a public commitment by the parties to each other. Whether attended by elaborate ceremonial or relatively informal, and whether religious or secular, its essence consists of a public acknowledgment of mutual commitment. Even primitive societies demand this, because the relationship does not concern only the immediate parties to it. The law may enable them to dispense with formalities, but not with public commitment. In some Polynesian societies, it is said, young men and women marry by the simple process of taking a meal together in public.
[80] Paragraph 2(2) provides that a person who was living with the original tenant “as his or her wife or husband” shall be treated as “the spouse of the original tenant”. Mathematically, there are four possibilities: “his wife”, “her wife”, “her husband” and “his husband.” But two of these are nonsense. A man cannot have a husband, and a woman cannot have a wife. In order to be treated as the spouse of the original tenant, a person must have been living openly with the tenant as his wife or her husband. In any given case, of course, only one person can qualify. If the tenant was a man, that person must have been his wife or have lived with him as his wife; if a woman, he must have been her husband or lived with her as such. The paragraph is gender-specific.
[81] It seems clear that parliament contemplated an open relationship, whether legal (para 2(1)) or de facto (para 2(2)), the essential feature of which is that, unlike other relationships, it subsists and can subsist only between persons of the opposite sex. A loving relationship between persons of the same sex may share many of the features of a de facto marriage. It may, as Baroness Hale describes it, be “marriage-like”, but it is not even de facto a marriage, because it lacks the defining feature of marriage.
[82] In my opinion, the words “of the opposite sex” are unmistakably implicit. Although not expressed in terms, they are manifest on the face of the statute. The parties are not required merely to live together but to do so as husband and wife. They are not merely given the same rights as married persons but are treated as if they were married persons. If the draftsman had inserted the words “being of the opposite sex” expressly, he would have produced a comical tautology. If he had inserted the words “whether of the same or opposite sex” he would have produced a self-contradictory nonsense. Persons cannot be or be treated as married to each other or live together as husband and wife unless they are of the opposite sex. It is noticeable that, now that parliament is introducing remedial legislation, it has not sought to do anything as silly as to treat same-sex relationships as marriages, whether legal or de facto. It pays them the respect to which they are entitled by treating them as conceptually different but entitled to equality of treatment.
[83] I turn to the legislative history. As originally enacted, the Rent Act 1977 provided, as had earlier Rent Acts, that on the death of a statutory tenant without a widow but leaving a member of his family living with him at his death, that person should become a statutory tenant by succession. If there were more than one such person, then, in the absence of agreement between them, the court would have to decide which of them would become the statutory tenant. If the tenant had left a widow who was residing in the dwelling-house at his death, however, she would have a right to a statutory tenancy in priority to other members of his family.
[84] It is an important feature of this legislation that the widow succeed to the tenancy by virtue of her status, much as she would succeed to her late husband’s estate on intestacy. She merely had to produce her marriage certificate. She did not have to prove that the marriage was happy, or stable, or long-lasting, or that the parties had been faithful to each other. The marriage could have been unhappy, tempestuous, or very recent; she could have been unfaithful; her husband could have begun divorce proceedings. Provided that she was living in the dwelling-house (not necessarily with her husband) at the date of the tenant’s death and that he was still her husband, she would be entitled to become the statutory tenant. She did not have to prove that she deserved to do so. Merit did not come into it.
[85] The primacy given to the widow’s claim would not, of course, be at the expense of the landlord. She was a member of the tenant’s family, and if she was the only member of his family who was residing in the dwelling-house when he died, she would be entitled to a statutory tenancy anyway. It would only be when there was more than one member of his family who qualified that her priority would be of any significance, and that this would be at the expense of the others. Even then, hers would normally be the most deserving claim. This provides the key to an understanding of the legislative policy behind the Rent Acts. The widow was to succeed by virtue of her status alone; she was not to be required to prove that she was more deserving than, say, her mother-in-law. Parliament sought to avoid provoking bitter and unseemly family disputes over the succession. A similar policy informs the law governing intestate succession. A statutory tenancy was the creature of statute. It was merely “a status of irremovability”; it was not property. It did not form part of a deceased tenant’s estate and could not pass under his will or on his intestacy. But parliament could determine what should happen on the tenant’s death, and, in effect, it provided for it to pass to his widow in much the same way as if it were part of his estate and he had died intestate.
[86] Unfortunately, as originally enacted, the Rent Act 1977 conferred the right on the tenant’s widow, and the Court of Appeal held that this was gender-specific. Paragraph 2(1) applied only where the tenancy was in the husband’s name and he predeceased his wife. It did not apply in the converse, although much less common, case where the tenancy was in the wife’s name and she predeceased her husband. Today, of course, section 3 of the Human Rights Act 1998 would permit this to be corrected. “Widow” cannot mean “widower”, but it can mean “widow or widower”. Reading in the words “or widower” would not contradict the express terms of the statute or create a nonsense. Parliament had failed to include the less common case (perhaps because it overlooked it or thought that it had provided for it), but it had not manifested an intention to exclude it.
[87] Parliament responded promptly. In 1980, it amended the Rent Act by substituting “surviving spouse” for “widow”, producing para 2(1) in its present form. For the reasons that I have mentioned, this did not prejudice the landlord. It merely meant that, like the widow, the widower could succeed by virtue of his status without having to prove that he was more deserving than, say, his stepson.
[88] In 1988, however, parliament made two changes of greater significance. The first was the consequence of a policy decision to introduce more flexibility into the housing market by phasing out statutory tenancies. No more statutory tenancies were to be created in future; they would be replaced by assured tenancies. Paragraph 3 was amended so that, on the death of a statutory tenant leaving a member of his or her family residing in the dwelling-house, he or she would become an assured tenant, but not a statutory tenant.
[89] Parliament did not, however, alter the right of a surviving spouse to become a statutory tenant in the place of the deceased tenant. This must, I think, have been due to a reluctance to enact what might be seen as retrospective legislation. The wife or husband of a statutory tenant had a vested right to succeed to the tenancy on the tenant’s death. He or she had more than a hope of succession, for there was nothing the tenant could do to prevent it. To the extent that this might be thought to run counter to the policy to phase out statutory tenancies, it should be recalled that the surviving spouse’s claim arose only on the death of a statutory tenant. Eventually, there would be no more statutory tenancies, and para 2 would become a dead letter. The same aversion to retrospective legislation was not, however, seen as applying, or as |page:142| applying with the same force, to the claims of other members of the tenant’s family.
[90] The present case is concerned with the second change: the introduction of para 2(2). This was compelled by changes in society. Couples are increasingly living together openly as man and wife without actually marrying. It is possible that this will become the norm rather than the exception. To extend the privileges of marriage to those who choose not to marry was formerly highly controversial; it was thought by many to undermine the status of marriage. It is less controversial today. By 1988, parliament considered that it was sufficiently acceptable to enact it in legislation.
[91] By enacting para 2(2), therefore, parliament was responding to changes in society. The timing of such a response is, under our constitutional arrangements, peculiarly a matter for the legislature and not the judiciary. Parliament’s policy, however, had not changed. The survivor, whether a spouse or merely treated as a spouse, should still have the right to succeed to the statutory tenancy by virtue of his or her status. The difference was that he or she no longer had to prove that the relationship was recognised by law; it was sufficient that it existed in fact. The claimant no longer had to produce a marriage certificate; it was sufficient that he or she and the deceased tenant had lived openly together as husband and wife. This probably was seen as entrenching on the landlord’s rights, for it must have been far from clear in 1988 that the “common law wife” or husband was a member of the other party’s family. But landlords do not ask to see their tenants’ marriage certificates, and the encroachment, if any, was easily justifiable.
[92] The expression “living together as man and wife” or “as husband and wife” is in general use and well understood. It does not mean living together as lovers, whether of the same or the opposite sex. It connotes persons who have openly set up home together as man and wife. Although other factors may be significant where the question arises between the parties themselves, in a context such as the present, it must depend largely, if not exclusively, upon outward appearances. It cannot depend upon the relationship being a happy, or long-lasting or stable one. This would be contrary to parliament’s long-standing policy: the survivor must succeed by virtue of his or her status. He or she is to be treated as having been the spouse of the original tenant because that is what, to all intents and purposes and to all outward appearances, the claimant was. This is, of course, not to say that they must hold themselves out as husband and wife: couples who live together as husband and wife rarely do so. It means only that they must appear to the outside world as if they were husband and wife.
[93] There is, indeed, a paradox at the heart of modern society. For centuries, the civil and canon law, the common law of Europe as it has been called, did not require any form of religious or secular ceremony to constitute a marriage. Persons who openly set up home together and lived together as man and wife were presumed to be married, and if they had consummated the marriage they were married; marriage was by habit and repute. The combined effect of the Council of Trent and the Marriage Acts put an end to all that. But there is nothing new in treating men and women who live openly together as husband and wife as if they were married; it is a reversion to an older tradition.
[94] By 1988, therefore, Parliament had successively widened the scope of para 2(1). First applying only to the tenant’s widow, it was extended first to his or her surviving spouse and later to a person who had lived with the tenant as his or her spouse, although without actually contracting a legally binding marriage. The common feature of all these relationships is that they are open relationships between persons of the opposite sex. Persons who set up home together may be husband and wife or live together as husband and wife; they may be lovers; or brother and sister; or friends; or fellow students; or share a common economic interest; or one may be economically dependent on the other. But parliament did not extend the right to persons who set up home together; only to those who did so as husband and wife.
[95] Couples of the same sex can no more live together as husband and wife than they can live together as brother and sister. To extend the paragraph to persons who set up home as lovers would have been a major category extension. It would have been highly controversial in 1988, and was not then required by the Convention. The practice of contracting states was far from uniform, and parliament was entitled to take the view that any further extension of para (2) could wait for another day. One step at a time is a defensible legislative policy that the courts should respect. Housing Acts come before parliament with some frequency, and parliament was entitled to take the view that the question could be revisited without any great delay. It is just as important for legislatures not to proceed faster than society can accept as it is for judges, and under our constitutional arrangements the pace of change is for parliament.
[96] Parliament, as I have said, is now considering corrective legislation in the Civil Partnerships Bill currently before the House in its legislative capacity. The bill creates a new legal relationship, called a civil partnership, that the persons of the same sex may enter into by registering themselves as civil partners. It inserts the words “or surviving civil partner” after the words “surviving spouse” in para 2(1), and adds a new para (2)(b):
(b) a person who was living with the original tenant as if they were civil partners shall be treated as the civil partner of the original tenant.
[97] There will thus be four categories of relationship covered if the bill becomes law: (i) spouses, that is, married persons (necessarily being persons of the opposite sex); (ii) persons who live together as husband and wife, who are to be treated as if they were married (and who must therefore also be of the opposite sex): (iii) civil partners (who must be of the same sex), who are given the same rights as, but are not treated as if they were, married persons; and (iv) persons who live together as if they were civil partners without having registered their relationship, who are treated as if they had done so. This is a rational and sensible scheme that does not involve pretending that couples of the same sex can marry or be treated as if they had done so.
[98] Among the matters that parliament will have had to consider in debating the Civil Partnerships Bill are: (i) which statutes to amend by extending their reach to civil partners and persons living together as civil partners: (ii) whether such statutes should extend to unregistered civil partnerships in every case, or whether in some cases it would be appropriate to require the parties to register their relationship before taking the benefits of the statute: (iii) whether the bill should be retrospective to any and to what extent: and (iv) from what date should the new provisions come into force. Presumably, some time must elapse before a system of registration can be established: should unregistered civil partners have to wait until it is? These, and no doubt other matters, are questions of policy for the legislature.
[99] All this will be foreclosed by the majority. By what is claimed to be a process of interpretation of an existing statute framed in gender-specific terms, and enacted at a time when homosexual relationships were not recognised by law, it is proposed to treat persons of the same sex living together as if they were living together as husband and wife and then to treat such persons as if they were lawfully married. It is to be left unclear as from what date this change in the law it to take place. If we were to decide this question, we would be usurping the function of parliament, and if we were to say that it was from the time when the ECtHR decided that such discrimination was unlawful, we would be transferring the legislative power from parliament to that court. It is, in my view, consonant with the Convention for the contracting states to take time to consider its implications and to bring their laws into conformity with it. They do not demand retrospective legislation.
[100] Worse still, in support of their conclusion that the existing discrimination is incompatible with the Convention, there is a tendency, in some of the speeches of the majority, to refer to loving, stable and long-lasting homosexual relationships. It is left wholly unclear whether qualification for the successive tenancy is confined to couples enjoying such a relationship or, consistently with the legislative policy that parliament has hitherto adopted, is dependent upon status and not merit.
[101] In my opinion, all these questions are essentially questions of social policy that should be left to parliament. For the reasons I |page:143| have endeavoured to state, it is, in my view, not open to the courts to foreclose them by adopting an interpretation of the existing legislation that it not only does not bear but that is manifestly inconsistent with it.
[102] I would allow the appeal.
Giving the fourth opinion, Lord Rodger said:
My lords,
[103] I have had the advantage of considering the speeches of my noble and learned friends, Lord Nicholls, Lord Steyn and Baroness Hale, in draft. I agree with them and would accordingly dismiss the appeal. In view of the importance of the issue, I add some observations on section 3 of the Human Rights Act 1998 (the 1998 Act).
[104] Section 3, which, as Lord Hoffmann remarked in R v Secretary of State for the Home Department, ex parte Simms [2000] 2 AC 115, at p132A, enacts the principle of legality as a rule of construction, provides:
(1) So far as it is possible to do so, primary legislation and subordinate legislation must be read and given effect in a way which is compatible with the Convention rights.
(2) This section —
(a) applies to primary legislation and subordinate legislation whenever enacted;
(b) does not affect the validity, continuing operation or enforcement of any incompatible primary legislation; and
(c) does not affect the validity, continuing operation or enforcement of any incompatible subordinate legislation if (disregarding any possibility of revocation) primary legislation prevents removal of the incompatibility.
In R v Director of Public Prosecutions, ex parte Kebilene [2000] 2 AC 326, at p373F, Lord Cooke described section 3(1) as “a strong adjuration” by parliament to read and give effect to legislation compatibly with Convention rights. Nevertheless, the opening words of subsection (1) show that there are limits to the obligation. That is reflected in subsection (2)(b) and (c) as well as in the next section, section 4, which applies in those cases where a higher court is satisfied that, despite section 3(1), a provision is to be regarded as incompatible with a Convention right. In that event, the court may make a declaration of incompatibility. Although it is therefore clear that there are limits to the obligation in section 3(1), they are not spelled out. In a number of cases, your lordships’ House has taken tentative steps towards identifying those limits. The matter calls for further consideration in this case.
[105] In addressing the question, it is useful to bear in mind section 6(1) and (2):
(1) It is unlawful for a public authority to act in a way which is incompatible with a Convention right.
(2) Subsection (1) does not apply to an act if
(a) as the result of one or more provisions of primary legislation, the authority could not have acted differently; or
(b) in the case of one or more provisions of, or made under, primary legislation which cannot be read or given effect in a way which is compatible with the Convention rights, the authority was acting so as to give effect to or enforce those provisions.
Subsection (3) goes on to define “public authority” as including a court.
[106] Inevitably, when section 3 comes to be considered by a court, the focus is on the approach that section 3(1) requires the court to adopt when reading a statutory provision that, on a conventional interpretation, would be incompatible with a Convention right. Nevertheless, the section is not aimed exclusively, or indeed mainly, at the courts. In contrast to section 4, which applies in terms only to “a court” of the level of the High Court or above, and in contrast also to section 6, which applies only to public authorities, section 3 is carefully drafted in the passive voice to avoid specifying, and so limiting, the class of persons who are to read and give effect to the legislation in accordance with it. Parliament thereby indicates that the section is of general application. It applies, of course, to the courts, but it applies also to everyone else who may have to interpret and give effect to legislation. The most obvious examples are public authorities such as organs of central and local government, but the section is not confined to them. The broad sweep of section 3(1) is indeed crucial to the working of the 1998 Act. It is the means by which parliament intends that people should be afforded the benefit of their Convention rights, “so far as it is possible”, without the need for any further intervention by parliament. In [44] of R v A (No 2) [2001] UKHL 25; [2002] 1 AC 45, at pp67-68, and in his speech today, Lord Steyn has referred to what ministers told parliament about how they anticipated the obligation in section 3(1) would work in practice. However that may be, section 3(1) requires public authorities of all kinds to read their statutory powers and duties in the light of Convention rights and, so far as possible, to give effect to them in a way that is compatible with the Convention rights of the people concerned. In practice, even before the 1998 Act came into force, many public authorities had reviewed the legislation affecting them so as to be in a position to comply with this obligation from the date of commencement. This was a wise precaution. Once the 1998 Act came into force, whenever, by virtue of section 3(1), a provision could be read in a way that was compatible with Convention rights, that was the meaning that parliament intended that it should bear. For all purposes, that meaning, and no other, is the “true” meaning of the provision in our law.
[107] The second point to notice is that, so far as possible, legislation must be “read and given effect” compatibly with Convention rights. The use of the two expressions, “read” and “given effect”, is not to be glossed over as an example of the kind of cautious tautologous drafting that used to be typical of much of the statute book. That would be to ignore the lean elegance that characterises the style of the draftsman of the 1998 Act. Rather, section 3(1) contains not one, but two, obligations: legislation is to be read in a way that is compatible with Convention rights, but it is also to be given effect in a way that is compatible with those rights. Although the obligations are complementary, they are distinct. So there may be a breach of one but not of the other. For instance, suppose that legislation within the ambit of a particular Convention right requires a local authority to provide a service to residents in its area. The proper interpretation of the duty in the legislation may be straightforward. But even if the local authority were to interpret the provision correctly and provide the appropriate service, if they were to provide the service only to those residents who support the governing political party, the local authority would be in breach of Article 14 in relation to the other article concerned and, in terms of section 3(1), would have failed to give effect to the legislation in a way that is compatible with Convention rights. So, even though the heading of section 3 is “Interpretation of legislation”, the content of the section actually goes beyond interpretation to cover the way in which legislation is given effect.
[108] Next, the 1998 Act discloses one clear limit to section 3(1). It is not concerned with provisions that, properly interpreted, impose an unavoidable obligation to act in a particular way. This can be seen from a comparison of paras (a) and (b) of section 6(2). According to para (a), section 6(1) does not apply, and a public authority will therefore have acted lawfully if, as a result of primary legislation, “the authority could not have acted differently”. An example might be a provision requiring a local authority to dismiss an application if the applicant had failed to take a particular step within seven days. Even if this were to result in the violation of a Convention right, the local authority must dismiss the application and, in doing so, they will have acted lawfully: they cannot act differently in terms of the legislation. By para (b), on the other hand, the public authority will also have acted lawfully if, in the case of one or more provisions of primary or secondary legislation “which cannot be read or given effect in a way which is compatible with the Convention rights”, the authority were acting so as to give effect to, or enforce, those provisions. Paragraph (b) echoes the language of section 3(1) and therefore deals with the (different) situation where, in terms of section 3(1), it has not proved possible to read and give effect to a provision in a way that is compatible with Convention rights. In that situation, as section 3(2)(b) provides, the validity, continuing operation and enforcement of the legislation are not affected and so it would be |page:144| lawful for a public authority to act in terms of the legislation. In that case, too, section 6(2) disapplies section 6(1).
[109] If incompatible provisions that require a public authority to act in a particular way, and leave it with no option to act differently, do not fall within the scope of section 3(1), this can only be because, by definition, it is not possible to read them or give effect to them in a way that is compatible with Convention rights. This makes sense. If a provision requires the public authority to take a particular step that is, of its very nature, incompatible with Convention rights, no process of interpretation can remove the obligation or change the nature of the step that has to be taken. Nor can the public authority give effect to the obligation by doing anything other than taking the step that the Act requires of it. In such cases, only parliament can remove the incompatibility if it decides to repeal or amend the provision. The most that a higher court can do is to make a declaration of incompatibility under section 4.
[110] What excludes such provisions from the scope of section 3(1) is not any mere matter of the linguistic form in which parliament has chosen to express the obligation. Rather, they are excluded because the entire substance of the provision, what it requires the public authority to do, is incompatible with the Convention. The only cure is to change the provision, and that is a matter for parliament and not for the courts: they, like everyone else, are bound by the provision. So, from section 6(2)(a) and (b), one can tell that, however powerful the obligation in section 3(1), it does not allow the courts to change the substance of a provision completely, to change a provision from one in which parliament says that x is to happen into one saying that x is not to happen. And, of course, in considering what constitutes the substance of the provision or provisions under consideration, it is necessary to have regard to their place in the overall scheme of the legislation as enacted by parliament. In International Transport Roth GmbH v Secretary of State for the Home Department [2002] EWCA Civ 158; [2003] QB 728, for instance, the Court of Appeal held that it was impossible for the court to use the interpretative obligation in section 3(1) in effect to recreate the fixed penalty scheme enacted by parliament so as to turn it into a scheme that was compatible with Article 6. As Simon Brown LJ observed, at p758C-D, it would have involved turning the scheme inside out — something that the court could not do. Only parliament, not the courts, could create a wholly different scheme so as to provide an acceptable alternative means of immigration control.
[111] Another illustration of this limitation on the obligation under section 3(1) is to be found in the decision of your lordships’ House in R (on the application of Anderson) v Secretary of State for the Home Department [2002] UKHL 46; [2003] 1 AC 837. Section 29 of the Crime (Sentences) Act 1997 provided that “if recommended to do so by the Parole Board, the Secretary of State may… release on licence” certain life prisoners, viz convicted murderers. The House was satisfied that it was incompatible with the Convention rights of Mr Anderson, who had been convicted of murder, for the power to release him to lie with the Home Secretary, rather than with a judicial body. Counsel for Anderson submitted accordingly that, under section 3(1) of the 1998 Act, section 29 of the Crime (Sentences) Act could be read and given effect to in a manner that would be compatible with his Convention rights. In effect, this would have amounted to reading the section in such a way as to deprive the Home Secretary of the express power to release him. The House rejected this submission since it was clear that, under section 29, the power of release and the power to determine how long a convicted murderer should remain in prison for punitive purposes were to lie with the Home Secretary and with no one else. In these circumstances, in the words of Lord Bingham in [30] of Anderson, at p883C-D:
To read section 29 as precluding participation by the Home Secretary, if it were possible to do so, would not be judicial interpretation but judicial vandalism: it would give the section an effect quite different from that which Parliament intended and would go well beyond any interpretative process sanctioned by section 3 of the 1998 Act…
The “judicial vandalism” would lie not in any linguistic changes, whether great or small, that the court might make in interpreting section 29 but in the fact that any reading of section 29 that negated the explicit power of the Secretary of State to decide on the release date for murderers would be as drastic as changing black into white. It would remove the very core and essence, the “pith and substance” of the measure that parliament had enacted — to use the familiar phrase of Lord Watson (in a different context) in Union Colliery Co of British Columbia Ltd v Bryden [1899] AC 580, at p587. Section 3(1) gives the courts no power to go that far. In these circumstances, the House made a declaration of incompatibility, which left it to the minister, and ultimately to parliament, to decide whether to remedy the incompatability by amending or repeating section 29 and, if so, how.
[112] In reaching this conclusion, Lord Bingham had regard to the well-known words of Lord Nicholls in [39] of In re S (Care Order: Implementation of Care Plan) [2002] UKHL 10; [2002] 2 AC 291, at p313, where the relevant distinction is drawn:
The Human Rights Act reserves the amendment of primary legislation to Parliament. By this means the Act seeks to preserve parliamentary sovereignty. The Act maintains the constitutional boundary. Interpretation of statutes is a matter for the courts; the enactment of statutes, and the amendment of statutes, are matters for Parliament.
Whatever can be done by way of interpretation must be done by the courts and anyone else who is affected by the legislation in question. The rest is left to parliament and amounts to amendment of the legislation. As Lord Nicholls pointed out, it is by no means easy to decide in the abstract where the boundary lies between robust interpretation and amendment, but he added in [40], at p313:
For present purposes it is sufficient to say that a meaning which departs substantially from a fundamental feature of an Act of Parliament is likely to have crossed the boundary between interpretation and amendment. This is especially so where the departure has important practical repercussions which the court is not equipped to evaluate. In such a case the overall contextual setting may leave no scope for rendering the statutory provision Convention compliant by legitimate use of the process of interpretation. The boundary line may be crossed even though a limitation on Convention rights is not stated in express terms.
[113] The problem facing the House in In re S was, in some ways, the opposite of the problem that was to come before the House in Anderson. In the earlier case, the interpretation of the Children Act 1989, which the Court of Appeal had adopted in reliance upon Articles 6 and 8 of the Convention, did not involve removing any power from a statutory body. Rather, in the view of the House, the starring system devised by the Court of Appeal involved conferring on the courts a power to supervise the way in which local authorities discharged their parental responsibilities under final care orders. This was to depart substantially from “a cardinal principle” of the Children Act, that the courts are not empowered to intervene in the way local authorities discharge their responsibilities under such orders: see [42] of Anderson, at p314. Lord Nicholls, with whom all the other members of the House agreed, went on to hold, in [43], that the innovation made by the Court of Appeal “passes well beyond the boundary of interpretation”. There was no provision in the Children Act that lent itself to the interpretation that parliament was conferring this supervisory function on the court. On the contrary, conferring such a function was inconsistent in an important respect with the scheme of the Act: “It would constitute amendment of the Children Act, not its interpretation.” In that situation, it was not possible to “read in” to the Act or any of its provisions a power to set up such a system. That would be to produce a meaning that departed substantially from a fundamental feature of the Act and so crossed the boundary between interpretation and amendment.
[114] Again, it is important to notice that the problem identified by the House did not derive from any perceived difficulty in finding language to frame a power to require a report on the progress of the local authority. Rather, the problem was that however the courts might frame the power, they would be introducing something that was not to be found in the Children Act, and, more particularly, something that was actually inconsistent with one of its cardinal principles. If such a change to the Act was to be made, parliament would have to make it. |page:145|
[115] In the second passage from his speech in In re S, which I have quoted in [112] above, Lord Nicholls made the further point that a departure from a fundamental feature of an Act of parliament may be more readily treated as crossing the boundary into the realm of amendment where it has important practical repercussions that the court is not equipped to evaluate. It appears to me that difficult questions may also arise where, even if the proposed interpretation does not run counter to any underlying principle of the legislation, it would involve reading into the statute powers or duties with far-reaching practical repercussions of that kind. In effect, these powers or duties, if sufficiently far-reaching, would be beyond the scope of the legislation enacted by parliament. If that is right, the answer to such questions cannot be clear-cut and will involve matters of degree that cannot be determined in the abstract but only by considering the particular legislation in issue. In any given case, however, there may come a point where, standing back, the only proper conclusion is that the scale of what is proposed would go beyond any implication that could possibly be derived from reading the existing legislation in a way that was compatible with the Convention right in question. In that event, the boundary line will have been crossed and only parliament could effect the necessary change.
[116] Although he was disagreeing with the other members of the House on the interpretation point, the approach of Lord Hope in [109] of A, at p87, is similar to the reasoning of Lord Nicholls in In re S. In Lord Hope’s view, “the entire structure of section 41” of the Youth Justice and Criminal Evidence Act 1999 contradicted the idea of reading into it a new provision entitling the court to give leave for evidence to be led of the complainant’s previous sexual behaviour with the accused whenever this was required to ensure a fair trial. It seemed to him that “it would not be possible” to read in such a provision “without contradicting the plain intention of Parliament” in section 41(2) to forbid the exercise of such a discretion unless the court were satisfied as to the matters identified by that subsection. In his view, parliament had taken a deliberate decision not to follow examples to be found elsewhere of provisions giving the court an overriding discretion to admit such evidence. In the phraseology of Lord Nicholls in In re S, for Lord Hope this was a “cardinal principle” of section 41 and it was not open to the courts to read the section in such a way as to depart substantially from it.
[117] It was in this context that Lord Hope expressed the view, in [108] of A, at p87, that it will not be possible to achieve compatibility with Convention rights by using section 3(1) “if the legislation contains provisions which expressly contradict the meaning which the enactment would have to be given to make it compatible” or, indeed, if the legislation contains provisions that do so by necessary implication. Lord Hope repeated this observation in [79] of R v Lambert [2001] UKHL 37; [2002] 2 AC 545, at p585, and, for the reasons that I have already given, I agree with it. But this is not to say that where a provision can be read compatibly with the Convention without contradicting any principle that it enshrines or the principles of the legislation as a whole, such an interpretation is not possible simply because it may involve reading into the provision words that go further than the specific words used by the draftsman.
[118] When parliament provided that, “so far as it is possible to do so”, legislation must be read and given effect to compatibly with Convention rights, it was referring, at the least, to the broadest powers of interpreting legislation that the courts had exercised before 1998. In particular, parliament will have been aware of what the courts had done in order to meet their obligation to interpret domestic legislation “so far as possible, in the light of the wording and the purpose of the [Community] directive in order to achieve the result pursued by the latter…”: see para 8 of Marleasing SA v La Comercial Internacional de Alimentación SA C-106/89 [1990] ECR I-4135, at p4159 (emphasis added). Both Pickstone v Freemans plc [1989] AC 66 and Litster v Forth Dry Dock & Engineering Co Ltd [1990] 1 AC 546 show how, long before 1998, this House had found it possible to read words into domestic regulations so as to give them a construction that accorded with the provisions of the underlying Community directive. As Lord Oliver of Aylmerton noted in Litster, at p577A-B, Pickstone had established that:
the greater flexibility available to the court in applying a purposive construction to legislation designed to give effect to the United Kingdom’s Treaty obligations to the Community enables the court, where necessary, to supply by implication words appropriate to comply with those obligations…
Lord Oliver was satisfied that the implication that he judged appropriate in that case was entirely consistent with the general scheme of the domestic regulations and was necessary if they were effectively to fulfil their purpose of giving effect to the provisions of the directive.
[119] Your lordships are also familiar with the exercise that has to be carried out under some Caribbean constitutions to bring existing laws into conformity with the rights guaranteed by the constitution. In [85] and [86] of Lambert, at p587, Lord Hope cites two examples: Vasquez v The Queen [1994] 1 WLR 1304 and Yearwood v The Queen [2001] UKPC 31; [2001] 5 LRC 247. Both of them show how far the Privy Council has been prepared to go in substituting very different words for the words of the relevant provision in order to bring it into conformity with the relevant rights guaranteed by the constitution. Such cases are instructive in suggesting that, where the court finds it possible to read a provision in a way that is compatible with Convention rights, such a reading may involve a considerable departure from the actual words.
[120] In other respects, however, the Privy Council decisions may not provide a sure guide to the approach to be adopted under section 3(1). They are all concerned with constitutions that are the supreme law, with which other laws must conform on pain of invalidity. Clearly, that applies irrespective of whether the effect of the constitution is to make the whole, or only part, of a law invalid and also irrespective of the legislature’s intention in enacting the law. The typical constitution, or constitutional order-in-council, contains a provision to the effect that existing laws are to be “construed with such modifications, adaptations, qualifications and exceptions as may be necessary to bring them into conformity with” the constitution. I have taken these words from section 134(1) of the Constitution of Belize, which was under consideration in Vasquez. The language of such provisions may be thought to go even further than the language of section 3(1) of the 1998 Act, especially in saying that existing laws are to be construed with such modifications etc “as may be necessary” to bring them into conformity with the constitution. In particular, since the constitution is supreme, the necessary modifications to a law may well involve making what, in the present context, would properly be regarded as amendments to the legislation. I refer in particular to the judgment of Lord Nicholls in para 24 of Rojas v Berllaque (Attorney-General for Gibraltar intervening) [2004] 1 WLR 201, at pp208-209, where, in relation to a similar “far-reaching obligation on courts” in Gibraltar, he said:
The court is enjoined, without any qualification, to construe the offending legislation with whatever modifications are necessary to bring it into conformity with the Constitution.
He added:
There may of course be cases where an offending law does not lend itself to a sensible interpretation which would conform to the relevant Constitution.
By contrast, the 1998 Act deliberately maintains the sovereignty of parliament and section 3(1) is framed accordingly. For that reason, the Privy Council authorities should be treated with some caution since they are the product of constitutional systems that differ from that of the United Kingdom in this important respect.
[121] For present purposes, it is sufficient to notice that cases such as Pickstone and Litster suggest that, in terms of section 3(1) of the 1998 Act, it is possible for the courts to supply by implication words that are appropriate to ensure that legislation is read in a way that is compatible with Convention rights. When the court spells out the words that are to be implied, it may look as if it is “amending” the legislation, but that is not the case. If the court implies words that are consistent with the scheme of the legislation but necessary to make it compatible with |page:146| Convention rights, it is simply performing the duty that Parliament has imposed upon it and others. It is reading the legislation in a way that draws out the full implications of its terms and of the Convention rights. And, by its very nature, an implication will go with the grain of the legislation. By contrast, using a Convention right to read in words that are inconsistent with the scheme of the legislation or with its essential principles as disclosed by its provisions does not involve any form of interpretation, by implication or otherwise. It falls on the wrong side of the boundary between interpretation and amendment of the statute.
[122] When Housman addressed the meeting of the Classical Association in Cambridge in 1921, he reminded it that the key to the sound emendation of a corrupt text does not lie in altering the text by changing one letter rather than by supplying half a dozen words. The key is that the emendation must start from a careful consideration of the writer’s thought. Similarly, the key to what it is possible for the courts to imply into legislation without crossing the border from interpretation to amendment does not lie in the number of words that have to be read in. The key lies in a careful consideration of the essential principles and scope of the legislation being interpreted. If the insertion of one word contradicts those principles or goes beyond the scope of the legislation, it amounts to impermissible amendment. On the other hand, if the implication of a dozen words leaves the essential principles and scope of the legislation intact but allows it to be read in a way that is compatible with Convention rights, the implication is a legitimate exercise of the powers conferred by section 3(1). Of course, the greater the extent of the proposed implication, the greater the need to make sure that the court is not going beyond the scheme of the legislation and embarking upon amendment. Nevertheless, what matters is not the number of words but their effect. For this reason, in the Community law context, judges have rightly been concerned with the effect of any proposed implication, but have been relaxed about its exact form: see, for example, Lord Keith and Lord Oliver in Pickstone, at p112D and p126A-B.
[123] Attaching decisive importance to the precise adjustments required to the language of any particular provision would reduce the exercise envisaged by section 3(1) to a game where the outcome would depend in part upon the particular turn of phrase chosen by the draftsman and in part upon the skill of the court in devising brief formulae to make the provision compatible with Convention rights. The statute book is the work of many different hands in different parliaments over hundreds of years and, even today, two different draftsmen might choose different language to express the same proposition. In enacting section 3(1), it cannot have been the intention of parliament to place those asserting their rights at the mercy of the linguistic choices of the individual who happened to draft the provision in question. What matters is not so much the particular phraseology chosen by the draftsman as the substance of the measure that parliament has enacted in those words. Equally, it cannot have been the intention of parliament to place a premium on the skill of those called on to think up a neat way round the draftsman’s language. Parliament was not out to devise an entertaining parlour game for lawyers, but, so far as possible, to make legislation operate compatibly with Convention rights. This means concentrating on matters of substance, rather than on matters of mere language.
[124] Sometimes it may be possible to isolate a particular phrase that causes the difficulty and to read in words that modify it so as to remove the incompatibility. Or else the court may read in words that qualify the provision as a whole. At other times, the appropriate solution may be to read down the provision so that it falls to be given effect in a way that is compatible with the Convention rights in question. In other cases, the easiest solution may be to put the offending part of the provision into different words that convey the meaning that will be compatible with those rights. The preferred technique will depend upon the particular provision and also, in reality, upon the person doing the interpreting. This does not matter since they are simply different means of achieving the same substantive result. However, precisely because section 3(1) is to be operated by many others besides the courts, and because it is concerned with interpreting and not with amending the offending provision, it respectfully seems to me that it would be going too far to insist that those using the section to interpret legislation should match the standards to be expected of a parliamentary draftsman amending the provision: cf Lambert in [80], at p585, per Lord Hope. It is enough that the interpretation placed on the provision should be clear, however it may be expressed and whatever the precise means adopted to achieve it.
[125] My lords, in the light of that discussion, I can deal fairly briefly with the particular provisions with which the House is concerned in the present case: paras 2 and 3 of Schedule 1 to the Rent Act 1977. The House considered them in detail in Fitzpatrick v Sterling Housing Association Ltd [2001] 1 AC 27, which concerned the claim by the survivor of a long-term homosexual relationship to be treated as “the surviving spouse” of the original tenant in terms of para 2. The House unanimously rejected that claim, but, by a majority, held that he fell to be considered as “a member of the original tenant’s family” in terms of para 3. So he was entitled to an assured tenancy, rather than a statutory tenancy.
[126] In reaching the conclusion that the appellant should be regarded as a member of the original tenant’s family, the majority identified the characteristics of such a person for the purposes of the Rent Act. Lord Nicholls put the matter in this way in Fitzpatrick, at p44A-D:
The question calling for decision in the present case is a question of statutory interpretation. It is whether a same sex partner is capable of being a member of the other partner’s family for the purposes of the Rent Act legislation. I am in no doubt that this question should be answered affirmatively. A man and woman living together in a stable and permanent sexual relationship are capable of being members of a family for this purpose. Once this is accepted, there can be no rational or other basis on which the like conclusion can be withheld from a similarly stable and permanent sexual relationship between two men or between two women. Where a relationship of this character exists, it cannot make sense to say that, although a heterosexual partnership can give rise to membership of a family for Rent Act purposes, a homosexual partnership cannot. Where sexual partners are involved, whether heterosexual or homosexual, there is scope for the intimate mutual love and affection and long-term commitment that typically characterise the relationship of husband and wife. This love and affection and commitment can exist in same sex relationships as in heterosexual relationships. In sexual terms a homosexual relationship is different from a heterosexual relationship, but I am unable to see that the difference is material for present purposes. As already emphasised, the concept underlying membership of a family for present purposes is the sharing of lives together in a single family unit living in one house.
Lord Slynn and Lord Clyde reasoned to similar effect: see Fitzpatrick, at pp38B-39C and pp51H-52E respectively.
[127] Of course, some homosexual relationships of the type described by Lord Nicholls will have lasted longer and will have been happier than others. The same goes for the equivalent long-term heterosexual relationships. What these passages make clear, however, is that a long-term homosexual relationship is to be treated as being the same as a long-term heterosexual relationship in all respects save in sexual terms. There is no material difference between them, so far as membership of the original tenant’s family is concerned, for the purposes of the Rent Act. As Lord Nicholls pointed out in Fitzpatrick, at p42A-B, the wife is a member of her husband’s family and the husband is a member of his wife’s family. Therefore, the only reason why the House held that a homosexual partner could not be regarded as “the surviving spouse” of the original tenant in terms of para 2(1) was that the extended definition of “spouse” in para 2(2) (“a person who was living with the original tenant as his or her wife or husband”) was framed in a way that connoted a relationship between two persons of opposite sexes. That must be the starting point of any consideration of the matter in the present case.
[128] For the reasons that Lord Nicholls has given in his speech, I am satisfied that treating the survivors of long-term homosexual partnerships less favourably than the survivors of long-term heterosexual partnerships for purposes of the Rent Act 1977 violates their right under Article 14 in relation to Article 8(1) of the Convention. Nor, in respectful disagreement with Lord Millett on this particular point, can I discern any principle underlying the Act as a whole, or Schedule 1 in particular, that requires that only the survivor of a long-term heterosexual relationship should be treated as a statutory tenant. All that seems to have happened is that, when Schedule 1 was amended in 1988, parliament chose to extend the concept of “spouse” to someone who had lived with the original tenant in a long-term heterosexual relationship, but did not go any further. As was recognised in Fitzpatrick, society has moved on since 1988. In this particular context, even if there once was, there is no longer any reason in principle for not including within the concept of “spouse” someone who had lived with the original tenant in an equivalent long-term, but homosexual relationship. To interpret para 2 so as to include such a person would, of course, involve extending the reach of para 2(2), but it would not contradict any cardinal principle of the Rent Act. On the contrary, it would simply be a modest development of the extension of the concept of “spouse” that parliament itself made when it enacted para 2(2) in 1988. The position might well have been different if parliament had not enacted para 2(2) and had continued to confine the right to succeed to the husband or wife of the original tenant. But that bridge was crossed in 1988. So the fact that the partners in a homosexual relationship are not, and, indeed, cannot be, married is not to be regarded as a critical factor limiting the way in which para 2(2) may be interpreted under section 3(1) of the 1998 Act. Nor is there any reason to fear that the proposed interpretation would entail far-reaching practical repercussions that the House is not in a position to evaluate. Certainly, counsel for the Secretary of State, who made submissions in favour of interpreting para 2(2) in this way, did not foresee any such problems.
[129] Accordingly, in reliance upon section 3(1) of the 1998 Act, I would interpret para 2(2) as providing that, for the purposes of para 2, a person, whether of the same or of the opposite sex, who was living with the original tenant in a long-term relationship shall be treated as the spouse of the original tenant. By this means, it is possible to read and give effect to para 2 in a way that is compatible with the respondent’s Article 8(1) and Article 14 Convention rights.
Giving the fifth and final opinion, Baroness Hale said:
My lords,
[130] It is not so very long ago in this country that people might be refused access to a so-called “public” bar because of their sex or the colour of their skin; that a woman might automatically be paid three-quarters of what a man was paid for doing exactly the same job; that a landlady offering rooms to let might lawfully put a “no blacks” notice in her window. We now realise that this was wrong. It was wrong because the sex or colour of the person was simply irrelevant to the choice that was being made: to whether he or she would be a fit and proper person to have a drink with others in a bar, to how well she might do the job, to how good a tenant or lodger he might be. It was wrong because it depended upon stereotypical assumptions about what a woman or a black person might be like, assumptions that had nothing to do with the qualities of the individual involved: even if there were any reason to believe that more women than men made bad customers, this was no justification for discriminating against all women. It was wrong because it was based upon an irrelevant characteristic, which the woman or the black person did not choose and could do nothing about.
[131] When this country legislated to ban both race and sex discrimination, there were some who thought such matters trivial, but of course they were not trivial to the people concerned. Still less trivial are the rights and freedoms set out in the European Convention. The state’s duty under Article 14, to secure that those rights and freedoms are enjoyed without discrimination based upon such suspect grounds, is fundamental to the scheme of the Convention as a whole. It would be a poor human rights instrument indeed if it obliged the state to respect the homes or private lives of one group of people but not the homes or private lives of another.
[132] Such a guarantee of equal treatment is also essential to democracy. Democracy is founded on the principle that each individual has equal value. Treating some as automatically having less value than others not only causes pain and distress to that person but also violates his or her dignity as a human being. The essence of the Convention, as has often been said, is respect for human dignity and human freedom: see para 65 of Pretty v United Kingdom (2002) 35 EHRR 1, at p37. Second, such treatment is damaging to society as a whole. Wrongly to assume that some people have talent and others do not is a huge waste of human resources. It also damages social cohesion, creating not only an underclass, but an underclass with a rational grievance. Third, it is the reverse of the rational behaviour we now expect of government and the state. Power must not be exercised arbitrarily. If distinctions are to be drawn, particularly upon a group basis, it is an important discipline to look for a rational basis for those distinctions. Finally, it is a purpose of all human rights instruments to secure the protection of the essential rights of members of minority groups, even when they are unpopular with the majority. Democracy values everyone equally, even if the majority does not.
[133] It is common ground that five questions arise in an Article 14 inquiry, based upon the approach of Brooke LJ in Michalak v Wandsworth London Borough Council [2002] EWCA Civ 271; [2003] 1 WLR 617, in [20], at p625, as amplified in R (on the application of Carson) v Secretary of State for Work and Pensions [2002] EWHC 978 (Admin)*, at [52], and [2003] EWCA Civ 797; [2003] 3 All ER 577. The original four questions were:
(i) Do the facts fall within the ambit of one or more of the Convention rights?(ii) Was there a difference in treatment in respect of that right between the complainant and others put forward for comparison?(iii) Were those others in an analogous situation?
(iv) Was the difference in treatment objectively justifiable? Ie, did it have a legitimate aim and bear a reasonable relationship of proportionality to that aim?
—————————————————————————————————-
* Editor’s note: Reported at [2002] 3 All ER 994
—————————————————————————————————-
[134] The additional question is whether the difference in treatment is based upon one or more of the grounds proscribed — whether expressly or by inference — in Article 14. The appellant argued that that question should be asked after question (iv), the respondent that it should be asked after question (ii). In my view, the Michalak questions are a useful tool of analysis, but there is a considerable overlap between them: in particular between whether the situations to be compared were truly analogous, whether the difference in treatment was based upon a proscribed ground and whether it had an objective justification. If the situations were not truly analogous it may be easier to conclude that the difference was based upon something other than a proscribed ground. The reasons why their situations are analogous but their treatment different will be relevant to whether the treatment is objectively justified. A rigidly formulaic approach is to be avoided.
[135] It is common ground that one of the Convention rights is engaged here. Everyone has the right to respect for his or her home. This does not mean that the state — or anyone else — has to supply everyone with a home. Nor does it mean that the state has to grant everyone a secure right to live in their home. But if it does grant that right to some, it must not withhold it from others in the same or an analogous situation. It must grant that right equally, unless the difference in treatment can be objectively justified. There is no need for us to express a view on the degree to which a Convention right must be engaged in order to bring Article 14 into play. On any view, that threshold is crossed here.
[136] It is also common ground that there is a difference in treatment in respect of that right between the respondent and the survivor of an opposite-sex relationship. It is also common ground that sexual orientation is one of the grounds covered by Article 14 on which, like race and sex, a difference in treatment is particularly suspect. For the reasons given earlier, the grounds put forward to justify it require careful scrutiny.
[137] The parties differ upon whether the survivors of unmarried heterosexual and homosexual couples are indeed in an analogous |page:148| situation and, therefore, upon whether the basis of the difference in treatment is sexual orientation or something else. But it is impossible to see what else the difference can be based upon if not the difference in sexual orientation. Everything that has been suggested to make a difference between the appellant and other surviving partners comes down to the fact that he was of the same sex as the deceased tenant. It is the decisive factor.
[138] We are not here concerned with a difference in treatment between married and unmarried couples. The ECtHR accepts that the protection of the “traditional family” is in principle a legitimate aim: see para 40 of Karner v Austria (2003) 14 BHRC 674. The traditional family is constituted by marriage. The Convention itself, in Article 12, singles out the married family for special protection by guaranteeing to everyone the right to marry and found a family. Had para 2 of Schedule 1 to the Rent Act 1977 stopped at protecting the surviving spouse, it might have been easier to say that a homosexual couple were not in an analogous situation. But it did not. It extended the protection to survivors of a relationship that was not marriage but was sufficiently like marriage to qualify for the same protection. It has therefore to be asked whether opposite- and same-sex survivors are in an analogous situation for this purpose.
[139] There are several modern statutes that extend a particular benefit or a particular burden, granted to, or imposed upon, the parties to a marriage, to people who are or were living together “as husband and wife”: see, for example, section 62(1) of the Family Law Act 1996 and section 137(1) of the Social Security Contributions and Benefits Act 1992. Working out whether a particular couple are or were in such a relationship is not always easy. It is a matter of judgment, in which several factors are taken into account. Holding themselves out as married is one of these, and if a heterosexual couple do so, it is likely that they will be held to be living together as such. But it is not a prerequisite in the other private and public law contexts, and I see no reason why it should be in this one. What matters most is the essential quality of the relationship, its marriage-like intimacy, stability, and social and financial interdependence. Homosexual relationships can have exactly the same qualities of intimacy, stability and interdependence that heterosexual relationships do.
[140] It has not been suggested to us that the nature of the sexual intimacies each enjoys is a relevant difference. Nor can the possibility of holding oneself out as a legally married couple be a relevant difference here. Homosexuals cannot hold themselves out as legally married, but they can if they wish present themselves to the world as if they were married. Many now go through ceremonies of commitment that have the same social and emotional purpose as wedding ceremonies — to declare the strength and permanence of their commitment to one another, their families and friends. If the Civil Partnership Bill now before parliament becomes law, an equivalent status will be available to them.
[141] The relevant difference that has been urged upon us is that a heterosexual couple may have children together whereas a homosexual couple cannot. But this too cannot be a relevant difference in determining whether a relationship can be considered marriage-like for the purpose of the Rent Act. First, the capacity to bear or beget children has never been a prerequisite of a valid marriage in English law. Henry VIII would not otherwise have had the problems he had. Even the capacity to consummate the marriage matters only if one of the parties thinks that it matters: if they are both content, the marriage is valid. A marriage, let alone a relationship analogous to marriage, can exist without either the presence or the possibility of children from that relationship. Second, however, the presence of children is a relevant factor in deciding whether a relationship is marriage-like, but if the couple are bringing up children together it is unlikely to matter whether they are the biological children of both parties. Both married and unmarried couples, both homosexual and heterosexual, may bring up children together. One or both may have children from another relationship: this is not at all uncommon in lesbian relationships, and the court may grant them a shared residence order so that they may share parental responsibility. A lesbian couple may have children by donor insemination who are brought up as the children of them both: it is not uncommon for each of them to bear a child in this way. A gay or lesbian couple may foster other people’s children. When the relevant sections of the Adoption and Children Act 2002 are brought into force, they will be able to adopt: this means that they will indeed have a child together in the eyes of the law. Third, however, there is absolutely no reason to think that the protection given by the Rent Act to the surviving partner’s home was given for the sake of the couple’s children. Statutes usually make it plain if they wish to protect minor children. These days, the succession is likely to take place after any children have grown up and left home. Children, whether adult or minor, who are still living in the home may succeed as members of the family under para 3 of the Schedule. It is the long-standing social and economic interdependence, which may or may not be the product of having brought up children together, that qualifies for the protection of the Act. In the days when the tenant was likely to be a man with a dependent wife, it was understandable that preference was given to the widow over anyone else in the family. But, in 1980, that preference was extended to widowers, whether or not they were dependent upon the deceased wife. In 1988, it was extended to the survivor of unmarried marriage-like relationships, again irrespective of sex or financial dependence.
[142] Homosexual couples can have exactly the same sort of interdependent couple relationship as heterosexuals can. Sexual “orientation” defines the sort of person with whom one wishes to have sexual relations. It requires another person to express itself. Some people, whether heterosexual or homosexual, may be satisfied with casual or transient relationships. But most human beings eventually want more than that. They want love. And with love they often want not only the warmth but also the sense of belonging to one another that is the essence of being a couple. And many couples also come to want the stability and permanence that go with sharing a home and a life together, with or without the children who for many people go to make a family. In this, people of homosexual orientation are no different from people of heterosexual orientation.
[143] It follows that a homosexual couple whose relationship is marriage-like in the same ways that an unmarried heterosexual couple’s relationship is marriage-like are indeed in an analogous situation. Any difference in treatment is based upon their sexual orientation. It requires an objective justification if it is to comply with Article 14. Whatever the scope for a “discretionary area of judgment” in these cases may be, there has to be a legitimate aim before a difference in treatment can be justified. But what could be the legitimate aim of singling out heterosexual couples for more favourable treatment than homosexual couples? It cannot be the protection of the traditional family. The traditional family is not protected by granting it a benefit that is denied to people who cannot, or will not, become a traditional family. What is really meant by the “protection” of the traditional family is the encouragement of people to form traditional families and the discouragement of people from forming others. There are many reasons why it might be legitimate to encourage people to marry and to discourage them from living together without marrying. These reasons might have justified the Act in stopping short at marriage. Once it went beyond marriage to unmarried relationships, the aim would have to be encouraging one sort of unmarried relationship and discouraging another. The Act does distinguish between unmarried but marriage-like relationships and more transient liaisons. It is easy to see how that might pursue a legitimate aim and easier still to see how it might justify singling out the survivor for preferential succession rights. But, as Buxton LJ pointed out in [21] of Ghaidan v Godin-Mendoza [2002] EWCA Civ 1533; [2003] Ch 380, at p391, it is difficult to see how heterosexuals will be encouraged to form and maintain such marriage-like relationships by the knowledge that the equivalent benefit is being denied to homosexuals. The distinction between heterosexual and homosexual couples might be aimed at discouraging homosexual relationships generally. But that cannot now be regarded as a legitimate aim. It is inconsistent with the right to respect for private life accorded to “everyone”, including homosexuals, by Article 8 since Dudgeon v United Kingdom (1981) 4 EHRR 149. If it is not legitimate to discourage homosexual relationships, it cannot be legitimate to discourage stable, committed, marriage-like homosexual relationships of the sort that qualify the survivor to succeed to the home. Society wants its intimate relationships, particularly, but not only, if there are children involved, to be stable, responsible and secure. It is the transient, irresponsible and insecure relationships that cause us so much concern.
[144] I have used the term “marriage-like” to describe the sort of relationship that meets the statutory test of living together “as husband and wife”. Once upon a time, it might have been difficult to apply those words to a same-sex relationship because, both in law and in reality, the roles of the husband and wife were so different and those differences were defined by their genders. That is no longer the case. The law now differentiates between husband and wife in only a very few and unimportant respects. Husbands and wives decide for themselves who will go out to work and who will do the homework and childcare. Mostly, each does some of each. The roles are interchangeable. There is thus no difficulty in applying the term “marriage-like” to same-sex relationships. With the greatest of respect to my noble and learned friend, Lord Millett, I also see no difficulty in applying the term “as husband and wife” to persons of the same sex living together in such a relationship. As Mr Philip Sales QC, for the Secretary of State, said in argument, this is not even a marginal case. It is well within the bounds of what is possible under section 3(1) of the Human Rights Act 1998. If it is possible so to interpret the term in order to make it compliant with Convention rights, it is our duty under section 3(1) so to do.
[145] Hence, I agree that this appeal should be dismissed for the reasons given by my noble and learned friend, Lord Nicholls. I also agree with the opinions of my noble and learned friends, Lord Steyn and Lord Rodger, on the scope and application of section 3 of the Human Rights Act 1998.
Appeal dismissed.
