An Equality Impact Assessment (EqIA) is the document that turns “we thought about this” into “we can show what we thought, on what evidence, and why we decided as we did.” For a trans-inclusion policy, that isn’t optional paperwork — it’s the mechanism that makes the policy defensible when it’s challenged, whether the challenge comes from a trans employee who feels the policy is inadequate, a colleague who feels it goes too far, a journalist, a regulator, or an employment tribunal. A policy without an EqIA behind it is a position. A policy with one is a documented decision. Under scrutiny, only the second survives.
That gap isn’t hypothetical. Beyond Compliance’s research found that 50.7% of organisations have no formal policy on trans or nonbinary inclusion at all, and — among those that do — 36.4% give managers no guidance on how to apply it. A further 26.9% have no set schedule for reviewing the policy they do have, and 20.9% don’t know when it was last reviewed. Put together, roughly half of organisations have nothing written down, and a large share of the rest have a document nobody has tested, applied, or revisited. An EqIA is the process that closes that gap — not by producing the “right” answer, but by producing a defensible one.
Why this matters now
The law an EqIA sits underneath hasn’t changed shape, but a recent judgment has sharpened what the assessment needs to cover. The Equality Act 2010 already requires public bodies to have due regard to eliminating discrimination and advancing equality of opportunity under the Public Sector Equality Duty (s.149) — and gender reassignment (s.7) has always sat alongside sex (s.11) as a distinct protected characteristic. What the Supreme Court’s 2025 ruling in For Women Scotland Ltd v The Scottish Ministers ([2025] UKSC 16) clarified is that “sex” in the Act means biological sex, not sex as modified by a Gender Recognition Certificate — which affects how single-sex exceptions and the PSED operate. Gender reassignment protection under section 7 is unaffected and remains a protected characteristic in its own right.
The practical consequence for EqIA practice: a policy that touches sex-based provision now needs its assessment to consider sex (biological) and gender reassignment as two separate lines of analysis, rather than one blended judgement call. That’s a change in what the EqIA has to show its working on — not a change in whether one is needed. It was needed before. It is needed now. It will be needed under whatever the law looks like next.
Question 1: What is this policy trying to achieve, and who decided that?
Every defensible EqIA starts by writing the aim down in a sentence a stranger could read and understand — not the assumed aim, the stated one. What outcome is the policy meant to produce, and for whom? Vague aims (“supporting all our people”) don’t survive scrutiny; specific aims (providing a name and pronoun change route within a set number of working days) do, because they can be tested against what actually happens. This is also where authority sits: who has the mandate to set this aim, and is that recorded? A policy that emerged from an informal conversation and was never signed off at the right level is vulnerable regardless of its content.
Question 2: Who is affected, and what evidence says so?
An EqIA has to name the people affected and show where that came from — headcount estimates, staff feedback, incident data, benchmarking against sector research. The absence of evidence is itself relevant: a decision made without evidence-gathering is procedurally flawed, whatever the decision turns out to be. For a trans-inclusion policy, that means considering impact on trans and nonbinary staff or service users specifically — not folding them into a generic “diversity” impact statement — and, where the policy touches sex-based provision, assessing sex and gender reassignment as distinct protected characteristics rather than one combined analysis.
Question 3: If there’s a restriction, is it proportionate — and what else was considered?
Any policy that restricts access, participation, or provision on the basis of a protected characteristic has to clear a proportionality test, not just a good-intentions test. What alternative approaches were weighed, and why were they rejected? Cost alone is not a sufficient justification by itself, and a blanket rule applied without case-by-case consideration is a common failure point precisely because it looks efficient rather than because it is defensible. If the policy leaves anyone with no provision at all, that is very unlikely to be proportionate — and an EqIA needs to show this was actually tested, not assumed away.
Question 4: How will this be reviewed, and who owns it?
An EqIA is not a one-off certificate; it is dated evidence that expires. It needs a named owner, a review date, and a trigger for early review if something changes — new guidance, a new judgment, a complaint, a restructure. This is where the numbers above bite hardest: 26.9% of organisations have no set review schedule at all, and 20.9% don’t even know when their policy was last looked at. Both are governance failures independent of what the policy says, because an undated, unowned decision cannot be defended as current.
Building the evidence trail
Answer these four questions properly and you have, in effect, an evidence trail: a claim (the aim), evidence (who’s affected and how you know), an authority (the legal basis and who signed off), and a documented action — reviewed on a set date, by a named owner. That’s what a review seal looks like in practice: not a rubber stamp, but a dated, attributable statement that this decision was current as of a specific point, and will be looked at again by a specific point. Organisations that can produce that trail are not necessarily the ones with the most generous policy. They are the ones whose policy — generous or cautious — can be shown to be a considered, evidenced, proportionate, and current decision. That is what “defensible” means in practice, and it is what an EqIA is for.