The Appraisal Trap – Green Book Appraisal, Gunning Compliance and the Governance Conflicts That Create Judicial Review Exposure

Public bodies running major consultations routinely use HM Treasury’s Green Book to appraise options, and routinely face legal challenge under the Gunning principles, the court-established legal test for lawful public consultation in the UK. The two frameworks share a common goal but impose conflicting demands on timing, disclosure and decision-making. This article examines those conflicts and their governance implications.

Most practitioners who run major service change or infrastructure consultations are familiar with both frameworks independently. The Green Book sets the Treasury standard for how public sector options must be appraised: full cost-benefit analysis, risk-adjusted costs, distributional impacts, sensitivity analysis, and a documented preference for the option with the highest Net Present Social Value. The Gunning principles, confirmed by the Supreme Court in Moseley v Haringey [2014], set the legal standard for how those options must be put to the public: at a formative stage, with sufficient reasons, adequate time, and genuine consideration of responses.

The problem is that the two frameworks were designed for different purposes and pull in different directions at precisely the moments that matter most. Understanding where they conflict is not an academic exercise: it is a pre-condition for defensible consultation design.

The Timing Problem: When Has the Decision Already Been Made?

The most structurally significant conflict between the Green Book and the Gunning principles concerns timing. The Green Book requires a thorough options appraisal before a preferred option is selected. The Gunning test requires consultation to occur “at a formative stage”, meaning before the decision is made. The tension arises because a well-executed Green Book appraisal, by design, produces a clear preferred option. Publishing that appraisal at consultation stage risks looking like a concluded view presented for ratification rather than a genuine proposal open to challenge.

Courts have been alert to this. In Moseley, Lord Wilson emphasised that consultation must be more than “going through the motions.” In R (Joicey) v Northumberland CC [2014] EWHC 3657, the court found that where a decision-maker had effectively pre-committed, the consultation was unlawful regardless of its procedural form. The Heathrow judgment (Spurrier [2019]) revisited the same issue in an infrastructure context: extensive technical work preceding consultation does not itself invalidate the process, but the framing of consultation materials must leave real room for the preferred option to change.

The practical implication for Green Book users is uncomfortable: the more rigorous and complete your pre-consultation appraisal, the greater the risk that consultation appears predetermined. Organisations that complete a full Business Case before consulting, as many infrastructure and NHS bodies routinely do, must be especially careful about how the appraisal is framed in public materials. Phrases like “our preferred option” or “the recommended approach” in consultation documents have been cited in judicial review proceedings as evidence of a closed mind.

Governance risk | If your consultation document describes a preferred option derived from a completed Green Book appraisal, you need to be able to demonstrate that the appraisal assumptions were genuinely revisable in light of consultation responses, and that you revised them, or documented why you did not.

The Disclosure Dilemma: How Much Appraisal Evidence Must Be Published?

Gunning Limb 2 requires that consultees receive “sufficient reasons” to respond intelligently. The Green Book generates exactly the kind of quantitative evidence that would allow intelligent responses: cost tables, risk distributions, distributional breakdowns by income group, sensitivity ranges. The question is how much of that evidence must appear in consultation materials, and in what form.

Public bodies frequently resist full disclosure of appraisal evidence on three grounds: commercial sensitivity (particularly in infrastructure procurement), the risk of pre-empting scrutiny processes such as Health Overview and Scrutiny Committee (HOSC) review, and the concern that partial or preliminary figures will be misrepresented or taken out of context. Each of these concerns is legitimate. None of them is a complete answer to Limb 2.

West Berkshire DC v SoS CLG (2015) is the clearest authority on what insufficient disclosure looks like in practice. The court struck down a consultation that withheld the economic reasoning behind an affordable housing exemption, finding that consultees could not meaningfully respond without understanding the basis for the proposal. The same logic applies to Green Book appraisals: if cost-benefit analysis drove the preferred option, consultees are entitled to understand that analysis at a level of detail sufficient to challenge its assumptions.

The quality of disclosure matters as much as the quantity

A related and underappreciated problem is that Green Book appraisals are technical documents written for economists and senior officials. Disclosing them in full does not necessarily satisfy Limb 2 if the disclosure is not intelligible to a lay consultee. The Gunning standard is not simply “information has been made available”: it is that consultees can form and express a considered view. An Appraisal Summary Table buried in a technical appendix, without plain-language interpretation, may fail the intelligibility test even though it has technically been published.

This creates a design challenge that many organisations handle badly. The default approach is to publish the full appraisal as an appendix and provide a brief non-technical summary in the main consultation document. Courts have been willing to look through this structure where the summary omits material assumptions or where the appraisal and the summary tell different stories. Where sensitivity analysis shows that a competing option becomes preferable under plausible alternative assumptions, that finding must be accessible to consultees, not left buried in appendices.

The Quantification Gap: What Happens to Evidence the Green Book Cannot Capture?

The Green Book is an optimisation framework. It aims to identify the option that maximises net present social value, with distributional considerations and unmonetised factors treated as secondary to the core cost-benefit calculation. Consultation is a deliberative process. It surfaces values, concerns and local knowledge that do not naturally reduce to monetary values, and the Green Book methodology is structurally ill-equipped to incorporate.

This tension is most visible in the treatment of equity and place. Green Book distributional analysis (Chapter 7) shows how costs and benefits fall across income groups or regions, but it works with aggregate effects. Consultation frequently surfaces specific, local, relational concerns: the effect of a road closure on an informal care network, the cumulative impact of service withdrawal on a community already facing multiple deprivations, or the disproportionate burden on groups whose costs are real but not easily monetised. These concerns are legitimate inputs to a decision. They are also largely invisible to the appraisal framework.

The legal question is what Limb 4 requires when consultation responses raise concerns that the Green Book methodology has not captured. “Conscientious consideration” does not mean acting on every response. It does mean genuine engagement with the substance of what has been raised. Where a decision-maker acknowledges a concern but dismisses it on the grounds that “it falls outside the scope of our appraisal framework,” that is a Limb 4 failure; the framework cannot be used as a shield against engagement.

The Cost-Benefit Override

A sharper version of this problem arises when consultation responses strongly favour an option that the Green Book appraisal ranks poorly. Public bodies sometimes assume that a robust economic case provides legal cover for disregarding contrary consultation views. It does not. The Gunning test requires that responses are conscientiously taken into account, not that they prevail, but the decision record must demonstrate genuine engagement with why the economically suboptimal option was rejected. A decision that simply restates the appraisal result without addressing the substantive concerns raised in consultation is vulnerable.

This is a particular risk in NHS service change consultation, where economic efficiency arguments (cost per QALY, cost per episode) are well-developed but where consultation responses frequently address access, trust, community identity and perceived abandonment, none of which register in the appraisal. HOSC scrutiny powers exist precisely because Parliament recognised that the economic case is not the whole case.

The Iterative Appraisal Problem: Which Version of the Evidence Was Consulted On?

Green Book appraisals are not static. They are revised as new data becomes available, as procurement costs crystallise, as risk registers are updated, and, critically, as consultation responses raise issues that require the assumptions to be revisited. This creates a version control problem with direct legal implications.

If an organisation consults on an appraisal in January, receives substantial responses about cost assumptions in February, revises the appraisal significantly in March, and makes its decision in April on the basis of the revised figures, there is a real question about whether consultees were given sufficient information to respond intelligently to the appraisal that actually drove the decision. The consultation was formally compliant, but the material put to the public was not the material on which the decision rested.

This problem has arisen in infrastructure DCO consultations and in NHS reconfigurations where clinical evidence reviews continued in parallel with consultation. The safe approach, and one that is rarely taken, is to treat any material revision to appraisal assumptions after consultation opens as a trigger to consider whether further consultation is required on the revised evidence. Courts have not yet established a bright line here, but the logic of Limb 2 points strongly in this direction.

There is also a subtler version of this problem. Where sensitivity analysis produces switching values (that is, where an alternative option becomes preferable if a key assumption changes by a specified amount), consultees have a legitimate interest in knowing those thresholds. If the appraisal says “Option A is preferred unless construction costs increase by more than 15%, in which case Option B becomes preferable,” that switching value is material to intelligent consideration. It is also the kind of information that is routinely omitted from non-technical consultation summaries.

---

---

How tCI Can Help

Quality Assurance
Independent review at critical stages of your engagement design. Our seven stage QA process ensures your focus group approach meets legal and good practice standards, assessing format choice, facilitation design and evidence documentation.

Early Assurance
Snapshot review during planning to sense check your engagement framework before fieldwork begins. Strengthens your approach and helps avoid costly missteps.

Bespoke Training Workshops
Sector tailored sessions that work with real projects, not hypothetical scenarios. Teams learn to design defensible focus groups, choose appropriate formats and document evidence properly. Half day or full day delivery for health, local government, planning and public service teams.

Contact tCI: hello@consultationinstitute.org