On predetermining outcomes

Aleardo Zanghellini *

In For Women Scotland Ltd v The Scottish Ministers, the UK Supreme Court held that ‘sex’ and related concepts in the Equality Act 2010 refer to biological sex. They do not include the sex of a trans person as recorded in their gender recognition certificate (GRC), which, where sought and validly obtained, is otherwise supposed to change a trans person’s legal sex for all legal purposes (subject to certain exceptions). Much of the Court’s argument hinges on the need for equality law to provide the organisations and decisionmakers who must apply it with antecedently clear guidance. But the Court’s vision of equality law as a jurisdictional domain requiring categorical rule application doesn’t sit well with the fairness-based rationale of equality law. The practice of equality is not – cannot be – all about rigidly predetermining outcomes in the interest of efficiency and predictability. Meanwhile, the Court itself predetermines the outcome in its own judgment by not even acknowledging that there are alternatives to categorical rule application when it comes to operationalising provisions in a statute.

The judgment makes no difference to trans people’s entitlement to invoke sex-based protections when they have experienced discrimination on the ground of the sex they identify with (§§248-261). But it does make a difference in respect of trans people’s entitlement to benefit from the forms of sex-based preferential treatment permitted in the Act, and from access to single-sex or separate-sex services congruent with the sex they identify with.

The Court argues that decisionmakers and service providers must be able to rely on a univocal, coherent meaning of ‘sex’ within the Equality Act, if they are to operate in accordance with its requirements (§§151-154). This argument is based on rule of law values of certainty and reliance. But governing by predictable rules whose meaning is antecedently clear is only one way for the law to perform its action-guiding function. Depending on the jurisdictional domain, the value of reliance fostered by categorical rules free of interpretive ambiguities may appropriately give way to considerations of fairness. Fairness is at home with the use of more open-ended legal standards, which enable forms of flexible decision-making that take into account all relevant circumstances.

Tolerating ambiguity when it comes to sex-related terms under the Act, so as to enable decisionmakers to assign to them either a trans-inclusive or exclusively biological meaning depending on context, may be appropriate where one size does not fit all, and where ensuring a fair solution seems to have primacy over concerns with certainty and predictability. The Equality Act itself sometimes explicitly subordinates concerns with reliance and predictability to ones of a different order, which are served by allowing decision-makers discretion in the application of the law. This is apparent where the Act provides that under certain circumstances (calling for assessments based on standards of reasonableness/proportionality and effectiveness/practicality) services may be provided separately on the ground of sex, or to one sex only. The inescapability of these kinds of particularistic judgements in the domain of anti-discrimination law is not terribly surprising: the practice of equality, after all, is a demanding, messy business.

If I am right – if anti-discrimination law by its nature makes more room than most other jurisdictional domains for decision-makers’ discretionary assessments – appeals to the avoidance of confusion in the interest of certainty and categorical rule-application have limited purchase in this context, and the Court inflates their importance. If ‘legislative reform might engage explicitly with different gender conceptual lines, rather than treating inconsistency as a source merely of unwelcome confusion’, the case for the courts to do so is even stronger.

Categorical rule application on the basis of rigidly defined legal terms in the interests of predictability generates, by necessity, suboptimal results. On the other hand, particularistic, flexible decision-making, if well-executed, is associated with the competing value of fairness. Considering that the key concern of anti-discrimination law, as a jurisdictional domain, is with fostering fair treatment, it is perplexing for the Court to not even acknowledge the tension between predictability and fairness. The Court does recognise that sometimes the same word can bear different meanings within the same Act, but obscures the different reasons why this may be so, all but reducing these situations to cases of drafting errors (§176). Indeed, the Court does not even treat a fairness-driven particularistic approach to the question of the meaning of ‘sex’ in the Equality Act as a legitimate alternative to categorical rule application. At the very outset it frames its task restrictively, as one of determining ‘if those words can bear a coherent and predictable meaning’ (§2).

In defending, in the interest of predictability, a single meaning of ‘sex’ for the purposes of the Act, and in picking the biological understanding as the best candidate for that meaning, the Court expends much ink to argue that a trans-inclusive understanding would be over-inclusive. It would preclude us from differentiating between the experiences and needs of cis and trans men, and between those of cis and trans women (§172). But, apart from the fact that some (not all) of the Court’s arguments here are overstated, suboptimality in categorical rule application is a two-way street. It can go the way of generating over-inclusive outcomes or under-inclusive ones. Where is the Court’s discussion of the perils of under-inclusion involved in adopting a purely biological understanding of ‘sex’?

The Court appears interested in under-inclusion only when drawing attention to it may seem to provide support for adopting the biological meaning of ‘sex’. Thus, in three paragraphs (§§228, 235, 242) it argues that a non-biological understanding of sex for the purposes of the Act would, irrationally, result in treating differently GRC-holders from non-holders. But the distinction in legal treatment between GRC-holders and non-holders is drawn in the Gender Recognition Act 2004 itself, which made the policy judgement that obtaining a GRC is a proxy, however imperfect, for the seriousness and stability of a trans person’s gender identity, justifying treating those who hold it differently from those who don’t. Like all clear-cut legal rules, this policy choice is suboptimal, but that does not make it inherently irrational; and so, neither would it be inherently irrational for relevant sex-based provisions in the Equality Act to assume the understanding of legal sex set out in the 2004 Act. Suboptimality by under-inclusion is not the same as irrationality; and under-inclusion admits of degrees. Under-inclusion that is partially trans-exclusionary (as resulting from understanding ‘women’ and ‘men’ in the Equality Act as including relevant GRC-holders) seems preferable to under-inclusion that is completely trans-exclusionary (as resulting from the Court’s preferred understanding of ‘sex’ as biological).

To be clear, this isn’t a case of rogue judges manipulating legal reasoning for nefarious ends. It is not the rogue judge, but the sincere one, ‘bound, not merely by doctrine but also by the limits of his or her political and legal imagination’, that produces the kind of ‘brittle, oppressive determinacy’ we see at play in the judgment.