“Impact Assessments should be published when decisions are made.” That is the recorded position of the Minister for Migration and Citizenship, given to a Lords committee this spring (para 313). It is the settled position of the Court of Appeal that the duty to have regard to a policy’s impacts must be performed before and at the time the policy is under consideration. This post is about the space between those two statements — and a set of dates, already in the diary, that will determine whether it closes or widens. None of this is, at bottom, about immigration: it concerns the relationship in time between ministerial decision-making and the evaluative duties Parliament and the common law attach to it.

The context is Earned Settlement: a policy materially altering the position of up to 1.6 million people — the Minister’s own figure — already part-way through published immigration routes, its retrospective application confirmed, its consultation drawing some 200,000 responses, its assessments unpublished, promised “in due course”. No court has been asked to rule; the principal rules are yet to be laid. The question is not whether an impact assessment will eventually exist — it plainly will — but whether it will arrive while still capable of doing the work the law assigns to it: a question of the assessment’s timing, not its existence. Nothing here concerns the policy’s substantive merits, and nothing asserts that anything the Government has done is unlawful. The exercise is narrower: set out the legal test and the timeline, and let the reader observe the direction of travel.

The doctrine: assessment as an antecedent, not an appendix

Three lines of authority converge on the point at which the Minister’s formulation and the court’s part company: the legal value of evaluative material is a function of when it enters the decision-making process.

The first is the public sector equality duty in s.149 of the Equality Act 2010. In Bracking v Secretary of State for Work and Pensions [2013] EWCA Civ 1345, McCombe LJ drew the principles together at [26]: the duty must be fulfilled before and at the time a policy is under consideration; regard must be exercised “in substance, with rigour and with an open mind”; and an assessment produced after the decision has in substance been taken is what the courts have called a rearguard action incapable of discharging the duty retrospectively. A late equality assessment is thus not a cure but evidence that the duty was not performed when the statute requires.

The second is the Tameside duty. In Secretary of State for Education and Science v Tameside Metropolitan Borough Council [1977] AC 1014, the House of Lords held, per Lord Diplock at 1065B, that a decision-maker must take reasonable steps to acquaint himself with the relevant information. Tameside prescribes no procedural form; but information-gathering can only perform its function if it occurs while still capable of influencing the decision. Acquisition in order to decide presupposes acquisition before decision: a minister who announces the destination and commissions the map afterwards has inverted the exercise.

The third concerns consultation. R v Brent London Borough Council, ex p Gunning (1985) 84 LGR 168 established, and the Supreme Court in R (Moseley) v Haringey London Borough Council [2014] UKSC 56 endorsed, that consultation must occur while proposals are at a formative stage, and that responses must receive conscientious consideration before decision. Both limbs are temporal. Consultation after public commitment to the outcome fails the first; responses that cannot realistically be digested before rules settle fail the second.

An assessment arriving after the policy it was meant to inform has been fixed satisfies the letter of a publication commitment while defeating its purpose.

The timeline

What follows are the argument’s factual premises: dates and public record, nothing else.

In May 2025 the Immigration White Paper announced that the standard qualifying period for indefinite leave to remain would double from five to ten years; whether the change would reach those already on qualifying routes was left open. In November 2025 the Government opened the A Fairer Pathway to Settlement consultation (CP 1448). Its foreword recorded that the ten-year baseline had already been announced — the consultation would address how settlement might then be earned — and proposed applying the changes “to everyone in the country today” without settled status; the document committed to publish a summary of responses “alongside economic and equality impact assessments … in due course”. Ten days in, a Government response to a parliamentary petition on application to those already in the UK stated “No decision has been taken”. On 16 December 2025 the Home Affairs Committee wrote requesting the equality impact assessment; the Home Secretary replied on 2 February 2026 that the impacts had been considered in developing the proposals, committing to a finalised assessment alongside the consultation response. The consultation closed on 12 February 2026 with over 200,000 responses. No impact assessment has been published; the Government’s position remains “in due course”.

On 5 March 2026 the first associated change — a heightened English-language requirement, applying to those already on a pathway — was laid in the Immigration Rules; the same Statement carried, that day, full impact assessments for two other measures. In a speech that day to the Institute for Public Policy Research, the Home Secretary said the Government was “currently reviewing the responses” while confirming “applying any rule changes to those who are in the UK today” without settled status. On 13 March 2026 the Home Affairs Committee reported, urging thorough assessment of the final policy’s impacts before implementation — getting changes right, it observed, matters more than speed. The response fell due on 13 May 2026; none has yet been provided. In March too, the Secondary Legislation Scrutiny Committee criticised assessing impacts only “iteratively” — which, it said, made the changes impossible to scrutinise fully.

On 23 June 2026 the Lords Justice and Home Affairs Committee published Settlement, Citizenship and Integration (HL Paper 13). At paragraphs 126–127 it recorded evidence invoking R (HSMP Forum Ltd) v Secretary of State for the Home Department [2008] EWHC 664 (Admin) and restated the principle of prospectivity; at paragraph 128 it recorded the Minister’s assurance of the legality of retrospective action, assessments to follow “in due course”; and at paragraphs 129–131 it concluded that retrospective change would be “manifestly unfair—and may be unlawful”, recommending against retrospective application to those already on a qualifying route. At paragraphs 315–316 it found the Home Office approach to impact assessments inconsistent, recommending publication when the decision is made but before implementation, alongside the legislation. The Government’s response falls due in late August 2026. Implementing rules are expected in autumn 2026.

The sentence opening this post appears at paragraph 313 of the same report: the Minister told the Committee, repeatedly, that impact assessments should be published when decisions are made — before that point, in his words, the department is still assessing and analysing. Set against the doctrinal sequence — evidence, assessment, policy choice, consultation, decision — the Government has not merely followed an inverted order; it has stated the inversion as method. Everything above is fact. Only the final link — rules expected to be laid with assessments still trailing — remains a projection; “expected” is doing deliberate work.

The collision, conditionally stated

Everything here is contingent on sequencing yet to occur. The argument is not that any legal threshold has been crossed in this instance; it is that the public record is progressively assembling the factual predicates from which such arguments are ordinarily constructed.

If the implementing rules are laid before the impact assessments are published, or so soon after that they could not realistically have informed them, three questions would arise on the public record. First, pre-determination: the outcome was signalled in the consultation’s own foreword and confirmed — in the March speech while responses were under review, in the Minister’s evidence to the Lords — before any assessment existed. Ministers are entitled to be predisposed; the complaint concerns sequence, not states of mind. Second, Bracking: on that sequencing and the Minister’s account of method, the assessment would look less like the rigorous, open-minded antecedent consideration s.149 requires and more like the rearguard action incapable of curing an earlier omission; a doctrine that demands regard before and at the time of decision sits uneasily with the Minister’s description of a process in which the department is still assessing until the decision is made. At that point, the assessment risks becoming justificatory rather than deliberative. Third, Gunning: the baseline extension was never itself consulted upon, having been announced six months earlier; for what was, the concern is not that minds were closed but that sequencing progressively narrowed the practical space within which 200,000 responses — and the assessments themselves — could influence the outcome. Whether consideration in those conditions can be conscientious is the question. The Commons committee noted an uncomfortable precedent: the 2022 care-route expansion — the failure Earned Settlement answers — was itself launched without a prior impact assessment (para 46).

The hedges must be stated with equal prominence. Section 31(2A) of the Senior Courts Act 1981 obliges a court to refuse relief where it appears highly likely the outcome would not have been substantially different absent the conduct complained of; a defendant would argue the direction was so settled that an earlier assessment would have changed nothing — an irony (pre-determination invoked to answer pre-determination) that does not deprive it of force. And s.149 requires due regard, not publication; the department asserts the impacts were considered in developing the proposals — though a consideration said to have occurred but never evidenced invites, rather than repels, scrutiny. Further, an assessment that is hurried but exists before laying transforms the complaint from absence into one about quality and genuineness, a harder, more evaluative claim.

But two things follow. Section 31(2A) disciplines remedy, not the analysis of whether the duty was performed when the law requires. And the constraint binds courts, not Parliament — which matters, because an assessment has two audiences, and the due-regard answer reaches only one of them. For the decision-maker, publication is evidence of regard; for Parliament, publication is the performance — an unpublished assessment can assist no committee. Statements of Changes are laid under the disapproval procedure in s.3(2) of the Immigration Act 1971 — negative in style — and an assessment arriving after the practical point of commitment cannot realistically assist committees or Members before a change takes effect. All three committees have said as much; the Minister’s own formulation — still assessing until published — concedes the equation.

What to watch

A timing argument is falsifiable; the coming months supply the tests. The first is late August 2026, when the response to the Lords committee falls due, against a Commons response already two months overdue. The second is the expected autumn laying of the rules. The third, and decisive, checkpoint is sequence: whether the economic and equality impact assessments appear before the Statement of Changes, with enough clearance for Parliament and consultees to digest them — a sequence to which the Government is already part-committed by its February promise.

Said plainly: if the assessments precede the rules, with a genuine interval for scrutiny, and the instrument carries considered transitional provision, the questions raised here largely dissolve; the doctrine asks for timing, and timing can still be delivered. The path to doing this properly remains open; the record shows a Government that has left itself a narrowing window in which to walk it.

The constitutional point

Section 149 and Tameside share an architecture: they are ex ante constraints, designed to shape decisions before commitments harden into settled outcomes. If a department may announce a direction, confirm its application to an existing cohort, and produce the assessment only “when decisions are made”, the duties survive in form but not in function — homework marked by its own author, submitted after the deadline. On 9 July 2026 the Lords debated the Constitution Committee’s rule-of-law report and the Government reaffirmed its unequivocal commitment to the rule of law. The autumn will show what that commitment means in the tense that matters to public law: the future one.

Zonglin Lyu is Co-Founder of the Skilled Worker Justice Alliance (SWJA).

(Suggested citation: Z. Lyu, ‘Marking Their Own Homework, After the Deadline: Earned Settlement and the Timing of an Impact Assessment’, U.K. Const. L. Blog (17th July 2026) (available at https://ukconstitutionallaw.org/))