16 November 2026: The bill I could not trace

By the middle of November 2026 I had acquired a peculiar new occupation.

In addition to practising law, writing, and pursuing proceedings against public authorities, I was trying to find out how much the British government had spent on its lawyers in cases brought by me. The question had an almost comic simplicity. I was a taxpayer. The government had lawyers. I had asked it to explain decisions about me. The Ministry refused. How much had the public paid for the resulting dispute?

The Government Legal Department had been able to provide a figure. On 17 September 2026 it disclosed, under my Freedom of Information request, £50,101.40, excluding VAT, in expenditure recorded against two proceedings involving the Ministry of Defence. One started in 2025 and the other followed in 2026. The schedule distinguished internal casework, client-related work and other categories. It was an accounting record of legal work, not a judicial determination that the work was necessary, reasonable or paid in precisely the same amount by its departmental client. That distinction is important. It is also why I next asked the client department for its own records.

On 28 October, in response to FOI2026/19326, the Ministry of Defence invoked section 12(1) of the Freedom of Information Act 2000 in relation to expenditure for 2025/26. Locating, retrieving and extracting the relevant information, it explained, would exceed the statutory cost limit of £600, calculated for central government by reference to 24 hours of staff time. An earlier payment process had ceased operation. The relevant items could not be retrieved through one readily searchable case-specific field. The Department would have to identify transactions manually across financial records. Its response distinguished this from some information identified for 2026/27. It was not a request for another twenty-four hours in which to answer; it was an estimate of the work required to produce an answer. [1]

A freedom-of-information refusal does not demonstrate an absence of invoice controls.

Financial records may be properly authorised while being difficult to retrieve by the search criteria used in a FOI request; accruals, casework charges, payment dates and departmental accounting categories need not coincide. Yet an ordinary taxpayer could reasonably ask a second question. If a department can instruct government lawyers on a case, what system allows it to reconcile the legal work recorded by those lawyers with its own approval, financial and performance records? More fundamentally still, how was the decision to continue commissioning that work made?

On 16 November I pursued both questions. I renewed an earlier submission to the House of Commons Public Accounts Committee about the cumulative cost of an unresolved administrative dispute. The secretariat responded that it could not indicate whether the Committee would take up the suggestion, but pointed me towards the National Audit Office, on whose technical work much parliamentary financial scrutiny depends. Later that afternoon I submitted the material to the NAO. I had not secured an investigation; I had referred evidence for consideration. I also asked my constituency MP, Andrew Rosindell, to consider Written Parliamentary Questions about the Ministry’s checks on Government Legal Department expenditure. In a further request that day, I proposed questions about the recorded consideration of administrative review, correction and value for money before and during litigation. Whether any of those questions would be tabled remained for Parliament, not me, to determine. [2]

I had started with the bill. I had arrived at the decision that generated the bill.

Before describing how that happened, it is necessary to step away from my dispute. The relevant question is not whether a particular ministry has a large legal budget, or whether a claimant dislikes being defended against. A government must sometimes contest litigation: public powers, public safety, statutory obligations and important issues of legal principle cannot always be negotiated away. The question is whether the State’s capacity to defend a decision is matched by an equally effective capacity to recognise, examine and, where appropriate, correct an administrative problem before the cost of defending it multiplies.

1. The citizen and the two ledgers

I came to this question as a commercial solicitor. Much of my professional instinct was formed around an unromantic proposition: litigation is an expensive means of managing uncertainty, and sensible clients want to know what they are buying. They ask what the dispute is really about, who has authority to settle it, what evidence would change the assessment, which contractual or regulatory mechanism can resolve it, what the next procedural step will cost, and what objective it serves. Those questions do not replace legal judgment. They make legal judgment accountable to the client’s practical interests.

The commercial lawyer’s first ledger belongs to the claimant. It contains court fees, legal advice, counsel’s fees, expert evidence, time diverted from work and the risk of paying some of the other side’s costs. The second belongs to the defendant. In a commercial dispute the defendant usually finances that ledger out of its own resources, or through insurance and other funding arrangements. Its officers and shareholders may ask whether the expense remains proportionate to the company’s interests. They may still choose to fight on principle. But the cost is visible as a commitment of an identified organisation.

A dispute with the State changes the arrangement. The citizen still encounters the first ledger directly. The public authority’s defence is supported through public finances: departmental budgets, charges for government legal services and the wider costs of justice. The claimant contributes, along with other taxpayers, to the revenue from which public institutions are funded. This is not literal double billing to the same person; tax does not confer ownership of a particular invoice, and not all taxpayers contribute equally. It is nonetheless a real asymmetry.

The individual is exposed to concentrated personal costs, while the institution draws on a collective financial structure that the individual cannot ordinarily control.

The asymmetry is not proof of wrongdoing. It is an important reason to demand clear governance. Public spending should be authorised, recorded and audited; but accuracy of accounts is not the same as value for money. A perfectly accurate bill may still represent work that could reasonably have been avoided by resolving an underlying administrative question earlier. Equally, an expensive defence may be justified because an apparently inexpensive correction would compromise a lawful public objective, affect other cases, or concede a legal principle that requires determination. Those competing possibilities must be examined, not presumed.

A preliminary cost model I prepared for parliamentary scrutiny in July 2026 illustrates the difference between estimates and actual expenditure. I estimated approximately £4,500 for the initial creation and handling of the personal record I was challenging. I separately modelled approximately £30,000 budget for an independent examination of the relevant records, coordination between institutions and possible corrective steps. Neither sum was an audited bill or a proven price for rectification. They were working assumptions, designed to ask what alternative administrative action might cost.

Against them stood a later official GLD record of £50,101.40 in legal expenditure (pre-defense and pre-permission stages) across two proceedings, excluding VAT. The figures cannot be treated as an apples-to-apples comparison. But their divergence creates an intelligible research question: had the relative merits, risks and costs of correction, independent investigation and continued defence been evaluated? [3]

HM Treasury’s Managing Public Money and the public spending accountability framework place value for money alongside regularity, propriety and feasibility in the responsibilities of public financial management. The National Audit Office has separately drawn attention to weaknesses in government cost information and the difficulty of using it to understand the real cost of services. The Government Legal Department’s Annual Report and Accounts for 2025–26 says that litigation is funded primarily by hourly charges to client departments. These are established components of the system. They do not tell us which official, if any, considered the cumulative cost of my particular administrative dispute as it moved through several channels. [4]

I therefore began treating the costs not as the conclusion to an accusation, but as evidence of an institutional journey. Every charge had a procedural origin. Every procedural step answered, postponed or reclassified some earlier question. To understand the expense it was necessary to reconstruct the decisions in the order in which they arose.

2. Before the bill, there was a decision

This chapter follows a different question from Chapter Two. There I ask how a person may be expected to challenge a personal decision that was never properly communicated, and why notice, reasons and an identifiable decision matter to effective access to justice. Here I begin when the person challenges (or attempts to challenge) the position taken by the public authority. The new problem is not simply what the authority decided. It is what the authority decides to do when its conduct is questioned.

There are at least five distinct events. An official action occurs. The affected person (hopefully) receives notice, discovers a record or experiences the consequences of an undisclosed decision. She asks for reasons, correction or review. The institution must then choose a response within its actual powers. That response may settle the problem, generate a refusal, direct the person to a statutory reviewer, leave a question unanswered or become the subject of litigation. The route may be urgent, irregular or parallel; it is not a compulsory ladder along which every claimant must climb. But the institution’s choice at each junction matters.

One reason it matters is that an ordinary administrative problem and a legal claim are not interchangeable objects. A person may ask, ‘Who entered this information, and on what basis?’ The litigation lawyer may instead ask, ‘What claim has been pleaded, what relief is available, and what must the defendant admit or deny?’ Both questions are legitimate. Only the first directly tests the provenance and continued accuracy of the administrative record. A successful procedural defence does not necessarily answer it.

This distinction also prevents a mistaken assumption that every disagreement with a public authority proves a state failure. A decision may be lawful, properly reasoned and fairly upheld. A complaint may be mistaken. A particular review procedure may not have jurisdiction. Effective administration must be able to confirm correct decisions as well as correct incorrect ones. The purpose of a genuine review is not compulsory concession; it is an accountable and competent answer.

For me, the separation became visible in the correspondence before my first judicial review. In September 2025 I sought clarification of the existence, authorship, grounds and operation of a personal restriction I suspected I had in connection with Northwood Headquarters. The Government Legal Department acknowledged my pre-action letter, asked for additional time and continued procedural correspondence. What I did not obtain before issuing proceedings was a substantive answer to the core factual questions I had posed. The relevant evidence supports that narrower proposition; it would be inaccurate to say that nobody replied to the letter at all. [5]

The distinction would be revisited months later. On 14 March 2026 I sent a pre-action/pre-complaint letter concerning equality and human-rights issues, asking nine questions about the possible use of my birthplace, the records relied upon and the identity and nature of the personal restriction. The letter was sent to general email address for new peoceedings within Mod.

GLD Public law department already dealing with my administrative proceedings replied on 27 March. It argued that the correspondence did not constitute a distinct, recognisable pre-action step and appeared to overlap with the judicial review already in progress. It warned that a new claim arising from the same matters would be defended and might face an abuse-of-process objection. It did not supply the nine requested administrative explanations. That is not a non-response in the postal sense. It is a response which treated the request primarily as an issue of administrative law litigation procedure. [6]

Whether MoD staff undertook a separate substantive administrative assessment, before or after those replies, remains a question for evidence. I cannot infer from the form of a GLD letter that no internal review occurred. I can ask who had authority to carry one out, whether it did occur, what records were created and how its conclusions informed the decision to continue the legal defence.

3. The point where procedure begins to shape reality

The importance of these distinctions is easiest to see when the initial decision is hidden. A public authority may know the precise text and operational effect of a person-specific instruction while the affected person knows only that something adverse has happened. The person may encounter refusal at a gate, an unexpected condition on an application, or a record revealed months afterwards. The experienced effect is not itself the whole decision.

Discovering that an instruction exists is not necessarily the same as being told its legal basis or the route for challenging it.

In R (Anufrijeva) v Secretary of State for the Home Department [2003] UKHL 36, the House of Lords examined the legal effect of an uncommunicated decision affecting an asylum applicant’s benefits. The case should not be stretched into a rule that every internal note requires individual notice. Its constitutional importance is the connection between communication of a decision, its legal effect and the practical possibility of challenge. In a system where rights depend on knowledge, non-notification may change the entire path a person must take to obtain an answer. [7]

It is tempting to draw one idealised procedural chart: decision, complaint, review, court. Real public administration is more complicated. Some decisions are urgent and take effect before ordinary review can operate. Some are not properly notified. Some have a designated statutory appeal; others do not. Some complaints can be considered while judicial review is pending; others may be unsuitable for a particular ombudsman once court proceedings begin. There are also strict time limits. Under CPR 54.5, judicial review ordinarily requires prompt filing and, subject to exceptions, no later than three months after grounds arise. A pre-action letter or complaint does not simply stop the clock. [8]

The correct starting point is therefore a decision map, not a mandatory obstacle course. It must show both a notified decision and the distinct route of an undisclosed decision discovered through its effects. It must record when administrative consideration is actually available, and where an adequate alternative remedy exists. It must allow a protective judicial review where delay would defeat the right to challenge. And it must preserve the possibility of administrative reconsideration and mediation even after proceedings begin.

That map is the connective tissue between the legal arguments in this chapter and the financial question with which it began.

4. What a complaint is supposed to accomplish

A complaint is often described as a low-cost alternative to a lawsuit, as though its principal virtue were that it relieves the courts of work. That description reverses the order of importance.

The first purpose of a competent complaints system is administrative: to learn whether something went wrong and decide what the institution should do about it. The reduction of litigation may follow from that work, but an institution can avoid court proceedings for a time while still failing to address the underlying problem.

The Public Service Ombudsman’s Principles of Good Administration emphasise getting things right, acting fairly and proportionately, being open and accountable, putting things right and seeking continuous improvement. The Ministry of Defence publishes a general complaints route, under which an individual uncertain about the correct destination may write to its Central Correspondence Unit. Its published description envisages forwarding the complaint to the part of the Department best placed to respond. The existence of a portal does not establish that a particular grievance falls within the procedure, that national-security information must be disclosed, or that a specific remedy can be granted. It does establish that the Department recognises administrative complaint allocation as a distinct function. [9]

The first question for a complaint handler should therefore be classification, not defence.

What exactly is alleged?
Does the person object to the merits of a restriction, the accuracy of a record, the conduct of an officer, the failure to provide notice, the handling of a request, or several distinct matters?

One episode can generate several questions without making them legally identical. An independent reviewer may have power over police conduct but not over the underlying security classification. A data-protection officer may test accuracy without possessing authority to reopen the original operational decision. The proper course is to allocate each issue, identify an owner and communicate a reasoned result or a clear jurisdictional refusal.

My experience gave this distinction practical importance. In September 2025 I submitted a complaint concerning the Commanding Officer at Northwood. I suspected that he made a decision affecting my rights (it was proven later that actually he did).

In early 2026, after learning that an earlier email route had been obsolete and the complaint had not been processed as expected, I sought my MP’s assistance to place the matter in an appropriate civilian channel, ultimately making my MP a postman between me and the Ministry.

Separately I raised matters about visitor-reception arrangements, police and security conduct and the handling of complaints themselves. I was not asking one reviewer to decide everything. I was asking for the institutions to identify which questions they would own, and which other body, if any, would examine the rest. [10]

Later ministerial correspondence treated various issues as overlapping with existing judicial review proceedings. That may be defensible for some questions; courts should not be bypassed where a remedy is exclusively judicial, and institutions should not be required to conduct endless parallel investigations of the same matter. Yet the conclusion that two complaints are ‘the same issue’ is itself an act of administrative classification. One needs to know who made it, what comparison was undertaken, and whether the matters left outside the scope of litigation were separately examined. It has not been done in my case.

In my complaint to the Parliamentary and Health Service Ombudsman, now the Public Service Ombudsman, I drew a line between judicial determination of legal rights and the prior administrative function of identifying, explaining and, where appropriate, correcting institutional records. I described the risk that a failure to allocate and decide separate complaints could drive the citizen into additional FOI requests, policing processes and court proceedings. The Ombudsman later notified me that case C-2205680 had moved to a stage requiring more detailed consideration. This was a procedural update, not a finding of maladministration. [11]

The research question is not whether every civil complaint must end with the citizen obtaining her preferred outcome. It is whether there is a reliable route to an accountable administrative decision—and whether the route continues to function once litigation has become possible.

5. The independent reviewer and the multiplication of decisions

Administrative escalation can protect an individual from institutional self-review. A complaints officer may lack independence from the original decision-maker; a statutory reviewer may possess distinct powers, standards and reporting obligations. But independent review also changes the legal map. It may create a further administrative decision in its own right. A later judicial review may challenge the reviewer’s decision, not simply the original incident.

My second judicial review (JR2) illustrates this. The initial subject matter concerned Service Police conduct. The Service Police Complaints Commissioner later performed a statutory review. The decision challenged in JR2 is the SPCC’s review decision: the legal questions concern what that office was authorised and required to examine, the reasons it gave and the proper procedure it followed. The fact that the complaint originated with RAF Police does not mean that the lawfulness of every RAF Police act is automatically before the court in the same way. The object of challenge has to be identified afresh.

A potential further judicial review involving Hertfordshire’s police and crime complaints arrangements would require the same discipline. If the challenge concerns the lawfulness of a Police and Crime Commissioner’s own review decision, the proper defendant and grounds depend on that decision and the statutory capacity in which it was made. It is not necessarily a judicial review of the Chief Constable’s original operational decision. Thus the state can be challenged at several levels for different legal acts generated by one connected factual history.

Each level has its own powers, time limits, evidence, decision-making records and potential financial consequences.

This is not an argument against independent review. A competent reviewer may identify a defect, remit a matter, order or recommend remedial action within statutory powers, or narrow the dispute before a court is needed. The problem arises when the complainant must guess whether the reviewer’s decision is final, what avenue remains and whether the original administrative owner still has responsibility for unanswered questions.

The commercial sector supplies a useful comparative model. Under the Financial Conduct Authority’s DISP rules, a regulated firm generally has to communicate a final response with reasons, and in the prescribed circumstances inform the consumer of referral rights to the Financial Ombudsman Service. The Financial Ombudsman Service is distinct from the business against which the original complaint was made. Its determination has its own legal character.

The model is not identical to service-police complaints or to judicial review, but it shows the value of a structured transition in which the citizen is not required to discover the next stage by accident. [12]

The budgetary consequences are more subtle than the institutional names suggest. The SPCC is an independent statutory office under Schedule 14A to the Armed Forces Act 2006, but its operational resources are supported within the Defence funding structure. Independent decision-making does not necessarily mean a wholly separate departmental funding source.
Hertfordshire policing and the PCC/OPCC, by contrast, operate in a local policing financial structure outside the Defence departmental budget. GLD itself supplies services to government clients and charges primarily by reference to the work performed. The precise coding of a particular SPCC review, the costs of defending an SPCC decision, and any onward court fees cannot be inferred from the title of the office. They require matter-level financial evidence. [13]
This produces a distinction that runs throughout this book: institutional independence, legal responsibility and budget responsibility are three different maps. A complaint can move from the first to the second while its financial costs continue to be borne inside the same broad public spending system—or it can cross into a new one. No single budget holder necessarily sees the whole sequence. The absence of an identified whole-system owner is not yet proof of failure, but it is an important question about governance.

6. A map of the pathway—and the possible exits

The following diagram is not a declaration that every claimant must exhaust every available administrative process before issuing. It is an analytical pathway. Red represents an unnotified personal decision or a refused route; orange represents the discovery of practical consequences and the proposed reassessment of options; green represents resolution; teal indicates administrative or independent review; purple marks litigation.

It begins with the central insight of Chapter Two: notification cannot be assumed. A person may discover consequences before discovering the decision itself. A complaint or request for reasons may follow. An internal response can correct, maintain or refuse to examine the position. Where a suitable statutory or independent review exists, the next challenge may properly concern that review decision. Where there is no adequate remedy—or where urgency and limitation require protection—judicial review may arise earlier.

Mediation and negotiation are drawn alongside the route, not as a compulsory box through which every dispute must pass. They may be useful before the initial administrative response, after an adverse decision, during a statutory review if permitted by that scheme, at pre-action stage or during litigation. They do not suspend time limits without the relevant legal arrangements, and they cannot replace a public authority’s legal obligations or a court’s determination of an issue that genuinely requires adjudication.


FIGURE 1 — The administrative-to-litigation pathway, with continuing opportunities for mediation

7. The decision to defend is itself a decision

Once litigation appears likely, an institution’s lawyers acquire necessary responsibilities. They must identify a proper respondent, assess pleadings and evidence, advise on duties of candour or disclosure, protect privileged material, preserve lawful confidentiality and ensure compliance with procedural deadlines. Their advice may be essential to the integrity of the State’s position.

But there is a difference between advising on legal defensibility and deciding whether the original administrative problem can and should still be addressed by the institution with power to do so.

The Judicial Review Pre-Action Protocol is an unusually revealing document because it contains objectives that are often overshadowed by the courtroom end of the process. Paragraph 3 calls for identification of issues and sharing relevant information, informed decisions about how to proceed, attempts to settle or reduce issues, avoidance of unnecessary expense and efficient proceedings when litigation cannot be avoided. Paragraph 5 reflects the principle that judicial review is generally inappropriate where an adequate alternative remedy exists. The protocol also recognises urgency and the need not to allow pre-action exchanges to defeat a claimant’s time limits. Paragraphs 9–12 invite consideration of non-court methods and the possible costs consequences of refusing ADR without adequate justification. Paragraphs 20 onwards address the defendant’s response to the proposed claim. [14]

These provisions are not a universal statutory command to perform a formal cost-benefit analysis before defending every judicial review. Nor do they grant the claimant power to dictate an administrative investigation. They nevertheless express a legal culture in which litigation is not the only recognised route and in which the parties are expected to consider the scope and necessity of proceedings. The unresolved practical question is who within a government department is responsible for acting on that expectation beyond the exchange of correspondence.

In commercial practice the answer is often expressed through an identified client representative. A solicitor advises; an authorised client decides on settlement objectives, commercial risk, budgets and acceptable alternatives. Government is more complicated (or reckless). The legal client may be a minister or department, while the record-holder, original officer, complaints team, accounting officer and litigation team occupy distinct roles. Some information may properly be restricted for national-security reasons. The case may carry an important precedent risk. The need for ministerial judgment may be real. None of these features makes the original administrative function disappear.

I have come to think of this as the missing distinction between the litigation file and the problem file. The litigation file asks how the defendant should answer the claim. The problem file asks what happened, whether the record is correct, who has authority to reconsider the decision and which human consequences remain unaddressed. The same evidence may be relevant to both; different legal protections and decision-makers may apply. Yet if the litigation file becomes the only active file, an institution can spend substantial resources defending a position without necessarily giving a citizen a conclusive answer to the original administrative question.

That is a hypothesis to be tested, not a finding about the internal conduct of the Ministry of Defence in my cases. My March 2026 correspondence demonstrates how a request for administrative explanations can receive a reply framed predominantly around the procedural status of an existing lawsuit. It does not reveal the full decision-making process behind the reply. The obvious next evidence is a policy, delegation, minute or record showing who considered alternative administrative action and when. A document recording that correction was considered and rejected for sound reasons would be valuable evidence. So would a document showing that the issue was referred to the proper official and resolved. The purpose of the inquiry is to find the decision, not to presume its answer.

8. The commercial lawyer crosses into public law

The comparison with commercial disputes is not an argument that government should operate as a private company. Public bodies are constrained by statute, equality obligations, public-law principles and duties to persons other than the immediate complainant. A commercial party can often settle on terms that would be unsuitable for a public authority, especially where legality, precedent or sensitive powers are involved. Still, regulated commercial and contractual systems have developed mechanisms worth studying: explicit decision ownership, escalation levels, independent dispute resolution, and mediation designed to preserve a working relationship or avoid the costs of litigation.

Ohpen Operations UK Ltd v Invesco Fund Managers Ltd [2019] EWHC 2246 (TCC) provides a particularly concrete example. The parties had agreed to a tiered dispute-resolution clause.

A commercial dispute concerning an IT platform moved through attempted negotiation and higher internal levels, with provision for mediation under the CEDR model procedure. Mrs Justice O’Farrell held the relevant mediation obligation enforceable and stayed proceedings to enable mediation after pleadings had clarified the issues. She did not conclude that every dispute must settle. The order recognised the usefulness of a defined opportunity for structured negotiation even after proceedings were issued. [15]

For a commercial lawyer, the striking feature is the architecture. The agreement identified who was to speak to whom, what kind of escalation was required, and what would happen before the parties used the court as the principal forum. Those arrangements were not proof of goodwill. They were process design: organisations recognised that a dispute could be serious without being irreversibly judicial.

Public law cannot copy a contract’s consent-based machinery wholesale. Administrative legality cannot be outsourced to a mediator. But a public department may still establish internal procedures ensuring that the person with power to correct a record is consulted before lawyers are instructed to contest all aspects of a claim. A complaint can be partly well-founded and partly unfounded. Reasons can be supplied without conceding the lawfulness of a decision. Some issues can be narrowed while others properly proceed to court.

The Court of Appeal’s decision in Churchill v Merthyr Tydfil County Borough Council [2023] EWCA Civ 1416 makes the comparison particularly interesting. The underlying dispute was a nuisance claim involving land owned by a council, and the proposed non-court route was the council’s own complaints procedure. The Court of Appeal held that a court may, in appropriate circumstances, stay proceedings or direct parties to engage in non-court dispute resolution, provided the order is proportionate and does not impair the essence of the right to a judicial hearing. The Court did not itself order a stay in that case. It identified factors including fairness, effectiveness, relative resources, urgency, cost and the risk of prejudicing legal time limits. [16]

One could read Churchill as a tool for institutions to force citizens into complaints systems. That reading would be incomplete. The case demands scrutiny of the quality of the proposed procedure: independence, clarity, timing, ability to produce meaningful redress and the power imbalance between the parties. An internal complaint without a deadline, an independent decision-maker or a realistic route to correction may be a poor substitute for a court.

Conversely, a properly designed process may provide something litigation cannot easily achieve: a direct explanation, an operational remedy or the prompt correction of an institutional record.

The lesson is not ‘always mediate’. It is that a choice between genuine alternatives should be a choice made on reasons. A defensive reflex is not an alternative-analysis mechanism.

9. Mediation before litigation, during litigation—and sometimes afterwards

In England and Wales, mediation is not confined to commercial contracts. Paragraph 10 of the Judicial Review Pre-Action Protocol expressly lists discussion, complaints and review procedures, ombudsmen and mediation as possible non-court methods. The general Practice Direction on Pre-Action Conduct and Protocols likewise encourages ADR and, crucially, continuing consideration of settlement after proceedings have been commenced. Civil Procedure Rules amended following Churchill expressly acknowledge the court’s capacity, subject to justice and proportionality, to order engagement with ADR. These provisions require context-sensitive judgment, not a compulsory settlement. [17]

I distinguish mediation from administrative review because their powers are different. A reviewer determines a complaint according to a defined mandate and may uphold, reject, remit or require some form of action. A mediator facilitates negotiation and does not ordinarily adjudicate legality. A successful mediation may produce an agreed administrative examination, an undertaking to provide reasons, a timetable for correction, a narrowing of legal issues or an agreed stay of proceedings. Such arrangements remain subject to the authority’s lawful powers and cannot displace statutory safeguards. Some public-law cases are poorly suited to confidential compromise because they require a binding public determination or affect third parties.

The empirical work of Varda Bondy, Linda Mulcahy, Margaret Doyle and Val Reid for the Public Law Project in 2009, Mediation and Judicial Review, set out to assess both the value and the limits of mediation in this setting. A later practitioners’ handbook by Bondy and Doyle examined situations in which ordinary communication and negotiation had become stuck. The methodological lesson is valuable: claims about mediation saving money cannot simply be assumed; the right comparison concerns the particular dispute, the costs and delays of the proposed process, the remedies it may produce and the constitutional role of the court. [18]

To that I would add a question from my own practice: who attends the mediation with actual authority to change the administrative position? If the only attendees have authority to discuss legal costs or compromise pleadings, the original problem may remain untouched. Mediation is not a correction mechanism unless it can connect to the person or unit empowered to inspect and amend the relevant record. A departmental lawyer may advise on risk and draft an agreement; an operational decision-maker may need to authorise a new decision; a data controller may have to decide on rectification; a statutory reviewer may be unable to negotiate away findings made under legislation. The parties must know which is which.

Mediation can also matter at more than one point. Before claim, it may resolve a factual disagreement without judicial proceedings. After a refusal of an administrative complaint, it may help isolate the matters requiring independent determination. During a judicial review, it may support a consent order, withdrawal, reassessment or remittal—without compromising the court’s constitutional function. After judgment, agreement may still assist with implementation, sequencing of corrective action or residual practical consequences. These are opportunities, not stages every person must exhaust. Limitation, urgency and interim relief continue to matter.

It is also important not to romanticise mediation. Equality of participation can be compromised where one party holds the evidence, controls the relevant records and can afford delay while the other cannot. Confidentiality can protect candid discussion but should not be used to conceal information that the law requires to be made available or to prevent scrutiny of systemic failings. The material obtained through without-prejudice negotiation may be protected; such correspondence must not simply be reproduced in a public book as though it were ordinary disclosed evidence. The legitimacy of mediation in public administration depends in part on how it addresses those limits.

The procedural comparison with small money claims is instructive but not determinative. Under Practice Direction 51ZE, eligible defended money claims in the small claims track are referred to an HMCTS mediation service. Participation in the prescribed process is required for qualifying cases; agreement is not. A public-law claim raising the legality of an official restriction is not the same type of claim, and its appropriate dispute-resolution route cannot be inferred from that pilot. Yet the policy demonstrates that the justice system is increasingly willing to build early non-court engagement into its process architecture. [19]

10. Who can afford to bring the question?

The imbalance in litigation costs becomes more pronounced when funding is considered.

Commercial litigation can sometimes be financed through conditional fee agreements, damages-based agreements where lawful, before-the-event or after-the-event insurance, or investment by third-party funders. Such arrangements are not automatically available; they depend on prospects, recoverability, risk and commercial economics. They may be expensive, unsuitable or unavailable for a particular client. The Civil Justice Council’s June 2025 review of litigation funding shows how extensively the civil justice system now analyses these mechanisms as part of access to justice. [20]

The economic logic is immediately apparent. A funder who pays for the costs of a claim generally expects an identifiable route to recovery. A damages claim with substantial, collectible monetary relief can be modelled. A judicial review seeking reasons, quashing of a decision, recognition of procedural error or correction of a record may have enormous personal and constitutional significance without providing a funder with any obvious shareable monetary return. Public-law remedies need not be measured in damages, but funding arrangements often are. That creates a structural tension between the value of a right and the market price of enforcing it.

It would be inaccurate to say that legal aid does not exist for administrative law. Official guidance expressly identifies challenges to the way government has made decisions, and discrimination claims, among the kinds of problems that may fall within the civil legal-aid scheme. Eligibility requires consideration of subject-matter scope, merits and means; particular routes have their own rules. Ministry of Justice statistics for April–June 2026 record 976 civil representation certificates granted in matters classified as judicial review, although contemporaneous data were affected by the recovery of a processing backlog. The scheme is real. Its practical availability to a particular litigant is a different question. [21]

One person may fall outside means eligibility while lacking the liquid resources to pay privately.

Another may meet financial tests but struggle to find a provider with suitable expertise and capacity. Some proceedings may involve several overlapping legal categories, only parts of which are within the scope of available funding. Claimants may represent themselves while facing a departmental defendant instructed through an established government litigation service. Pro bono and specialist public-law organisations can make a profound difference, but they cannot be treated as a dependable replacement for an appropriately functioning public funding system.

The funding question is not a digression. It changes the administrative calculus. A public authority may possess information the claimant must pay to seek; the claimant may have to issue a protective claim because time is running out; each subsequent dispute about reasons, evidence or procedural routing consumes more resources. The authority faces costs too, but the citizen may confront an immediate choice between financial risk and leaving the question unanswered. In R (UNISON) v Lord Chancellor [2017] UKSC 51, the Supreme Court’s account of access to justice explains why legal rights have limited practical significance if the person entitled to enforce them cannot realistically reach a court. [22]

This is one reason I resist presenting the issue as a simple contest between a private individual’s purse and a government department’s. The deeper problem is whether the design of the administrative system pushes the cost of establishing what happened onto the person least equipped to bear it. If the State can provide a timely, reasoned, authoritative administrative determination, it may reduce costs for both sides. If it cannot, even a legally arguable claim can become financially destructive before the merits have been tested.

11. One problem, several public budgets

The cost of an administrative dispute does not always stay with the institution that created the original decision. A complaint may pass to another unit or a statutory reviewer. A related information-rights request may involve a data-protection function. An ombudsman may assess complaint handling. Litigation may require departmental officers, government lawyers, external counsel, the court and administrative staff. Some of these costs sit within one broad governmental system; others belong to distinct budgets. The person at the centre experiences one continuing problem while the public financial records may describe several unrelated activities.

That observation must be handled carefully. A separate budget is not itself a defect.

Parliamentary appropriation, financial delegations and statutory independence exist for legitimate reasons. The independence of a police-complaints reviewer should not be weakened simply to make expenditure easier to consolidate. Financial oversight should be sufficiently coordinated to make the whole sequence intelligible without giving the original institution control over the reviewer’s decision.

My Service Police Complaints Commissioner matter is a particularly useful case study because legal independence and funding do not map neatly onto one another. The Armed Forces Act 2006 provides for the statutory office and the Secretary of State’s capacity to supply support.

The SPCC’s published reports describe its operating arrangements within the Defence funding environment. A judicial review of a SPCC decision therefore raises two distinct questions: who bears legal responsibility for the review decision, and where are the costs of the review and any subsequent litigation accounted for? One cannot answer the second solely by naming the defendant in the first. [23]

My other matter involving the transition to a Hertfordshire PCC review would be different again.

The office belongs to local policing governance rather than the Defence budget. Yet a complaint involving related official descriptions, police conduct or information exchange may remain connected to the same personal history. As a result, a public system can produce several lawfully distinct expense streams while no one official has responsibility to ask whether the underlying administrative issue has been comprehensively investigated and, if necessary, repaired.

Here the commercial analogy returns, but with a limit. Large companies maintain separate subsidiaries, business units and insurance arrangements. A dispute can be escalated between contract managers, compliance departments, external advisers, regulators and an ombudsman. Senior management may still demand a consolidated picture of economic exposure. The State can need a similar perspective without erasing statutory separation. What is missing is not necessarily a single controller of every decision; it may be a reliable map of costs, decisions and outstanding questions, preserved across institutional boundaries.

This is why my request to the NAO was framed as a question about governance rather than an allegation of improper invoicing. It is possible that MoD’s financial controls operate correctly on every charge that passes through them. It is also possible that no routine process aggregates the costs of the connected complaint, disclosure, investigation and litigation pathways. Both propositions can be true. If so, the problem is not a fraudulent or inaccurate ledger. It is an incomplete field of vision. As it should not take more than 24 hours for the department to answer a simple question of how much taxpayers' money the Ministry has spent on two litigation.

12. When the court asks who made the decision

A published judicial decision provides an independent example of why the attribution of procedural choices matters. In FRY, BNM1 and LRB v Secretary of State for Defence [2026] EWHC 2571 (Admin), Mr Justice Chamberlain examined repeated requests by MoD for additional time in separate Afghan relocation judicial reviews. The court criticised aspects of late and insufficiently supported extension applications and ordered specified costs consequences, including indemnity assessment for identified delay-related costs. It also required that certain future contested applications identify the MoD or GLD official who decided to seek an extension. [24]

The decision did not find that all MoD litigation expenditure was unnecessary. It did not determine the merits of my cases, obviously, and it cannot be used as proof that the same failures occurred in them. Its significance is narrower and more illuminating. When a procedural choice imposes costs on litigants and the judicial system, a court may require better evidence of the decision and its responsible author. The fact that a department’s lawyers are busy is not automatically a sufficient answer to a failure to comply with a court order.

That approach prompts a governance question one stage earlier. If an official must sometimes be identifiable for a procedural application made in litigation, can the institution identify who made the administrative judgment that litigation should be defended or prolonged, instead of reviewing or correcting an underlying issue? This is not a proposal to expose privileged advice or to make named individuals personally liable merely because a defence was unsuccessful. It is a question of traceability, delegation and reason-giving within public administration.

There is an important precedent within central government itself. In March 2020, Sir John Manzoni, then Cabinet Office Permanent Secretary and Accounting Officer, advised that further negotiation should be attempted in litigation concerning a former special adviser, drawing on legal and financial analysis and the prospective cost of continuing the dispute.

When the Prime Minister preferred to continue, Manzoni requested a formal ministerial direction. The correspondence was subsequently published. This does not establish a requirement for a ministerial direction in every civil case. It demonstrates that the financial and public-interest merits of continued government litigation can become an explicit, documented management decision. [25]

The difference between these two examples is important. FRY concerns judicial supervision of procedural conduct; the ministerial direction concerns responsibility for spending decisions within government. Neither substitutes for the merits of an individual administrative complaint.

Together they show that there are already legal and financial languages for asking who made a decision, why it was made and what consequences it carried.

13. The administrative reassessment gate

I therefore propose an analytical device, not yet a general rule of law: an Administrative Reassessment Gate. Its purpose would be to ensure that, when litigation is foreseeable or already under way, the institution still has an identified route to consider whether the underlying administrative question can be resolved or narrowed without surrendering legitimate legal positions.

The gate should not be owned solely by the litigation team. It requires an administrative decision owner capable of instructing relevant officers, verifying records and authorising lawful corrections. Lawyers should advise on legal risk, remedies, privilege and the proper scope of any proceedings. Finance staff or an appropriate delegated officer should assess material costs and alternatives. Where a statutory reviewer is independent, the process must not trespass on that reviewer’s decision-making powers. A mediator may facilitate a solution where the parties and the subject matter permit it. Each participant has a distinct function.

A credible gate would ask, and record answers to, six questions.

  1. What is the exact decision, omission or record in dispute?
  2. Has the affected person been given the notice, reasons and access routes required by law?
  3. What facts have been verified, and which remain contested?
  4. What administrative action, if any, is within the institution’s powers to take now?
  5. Which matters require independent review or a court’s ruling?
  6. What are the foreseeable costs, risks and consequences of each lawful course, and when will the choice be reconsidered?

The point is not to force a favourable outcome for the complainant. A documented assessment might properly conclude that no correction is warranted, that national-security material must remain protected, that the claimant is seeking a remedy the institution cannot provide, or that the court must resolve a difficult legal question. Such an assessment would make the choice to defend more intelligible, not less legitimate. The significant failing would be a system incapable of identifying who made the choice or what alternatives, if any, were considered.

The gate would need safeguards. It must not become a bureaucratic requirement that delays urgent judicial review or allows limitation periods to expire. It must not require disclosure of privileged legal advice or operationally sensitive information to persons not entitled to receive it. It should operate proportionately: a minor, plainly unfounded complaint should not trigger a full public expenditure inquiry. It should distinguish settlement of a legal claim from correction of an administrative record; the former does not automatically achieve the latter. Finally, it must recognise that a claimant retains the right to decline mediation and seek adjudication where the law permits.

I have called the wider research programme Administrative Injury Law. That proposed framework asks whether the law’s existing remedies, taken together, adequately identify, investigate and repair serious cumulative harm caused through public administration. The reassessment gate is one potential prevention mechanism within that theory. It aims to make an existing administrative responsibility visible before a potentially remediable problem becomes further fragmented between courts, complaints offices and financial systems. It remains a reform hypothesis, to be tested against statutory schemes, public-law doctrine and evidence of how institutions actually operate.

14. The question behind the bill

By the afternoon of 16 November, my papers were with the National Audit Office for consideration. The House of Commons Public Accounts Committee had not committed to an inquiry; my MP had been asked to consider parliamentary questions, not instructed to table them. Nothing in those acts established that MoD had breached a financial duty (yet). The courts had not determined the underlying disputed issues. The bills would continue to demand careful interpretation.

But I had located a question that could not be answered by the sum of the invoices alone.

I had originally wanted to know what it would cost a public authority to examine and, if appropriate, correct a personal administrative record. I eventually found myself asking who authorised a sequence of legal and administrative responses, whether the original question was independently re-examined, which reviewer could decide which part, and how the public bodies accounting for those steps understood the whole.

To a commercial lawyer, who I was originally before I was forced into a domain of public law, this is recognisable as the problem of decision ownership. A client may be legally entitled to fight, and a solicitor may be technically entitled to advance a defence. Neither proposition answers the separate question whether the organisation has made a considered choice about the problem it is trying to solve. In public law the stakes are more complicated, because the institution must protect public interests, statutory powers and people beyond the parties to the proceedings. That complexity makes the choice harder. It does not make the choice unworthy of examination.

I do not know whether an internal MoD minute exists recording a reasoned assessment of administrative correction as an alternative to continued litigation in my cases. I do not know whether its financial controls aggregate the expenditure of the connected complaint and review routes. Those are questions for evidence, not facts to be supplied by inference. And I do not know which of the legal challenges will succeed. The inquiry can remain worthwhile without pretending to know their outcomes.

What I do know is that a democratic system cannot measure its accountability only by its ability to defend its decisions in court. The lawfulness of a defence, the competence of a complaints system, the independence of a reviewer, the accessibility of legal aid and the reliability of financial records are related but separate conditions.

A person’s life does not fragment into those categories when something happens to her. Public administration does.

The State may have good reasons to defend a decision. The public is entitled to ask whether somebody considered the reasons for reconsidering it first—and whether that question remains alive after the lawyers have been instructed.
That is the decision before the defence.

SELECTED AUTHORITIES, RESEARCH AND DOCUMENTARY RECORD

[1] MoD, FOI2026/19326, response of 28 October 2026; GLD, expenditure disclosure of 17 September 2026. The 24-hour calculation concerns the FOIA search and retrieval cost ceiling, not an extension of the statutory response deadline. FOIA section 12: https://www.legislation.gov.uk/ukpga/2000/36/section/12.
[2] Author’s request to PAC, 16 November 2026, and PAC secretariat reply the same day; author’s submission to NAO, 16 November 2026; email to Andrew Rosindell MP on GLD expenditure and separate proposed Written Parliamentary Questions concerning administrative alternatives and value for money, 16 November 2026. Parliamentary receipt or tabling is not presumed.
[3] Author’s Joint Parliamentary Briefing, 24 July 2026 (working cost model); GLD expenditure schedule, 17 September 2026. The £4,500 and £30,000 figures were estimates; £50,101.40 was recorded GLD expenditure excluding VAT.
[4] HM Treasury, Managing Public Money: https://www.gov.uk/government/publications/managing-public-money; HM Treasury, Public Spending: The Accountability Framework: https://www.gov.uk/government/publications/public-spending-the-accountability-framework; NAO, Improving Government’s Productivity Through Better Cost Information (2025): https://www.nao.org.uk/reports/improving-governments-productivity-through-better-cost-information/; GLD, Annual Report and Accounts 2025–26: https://www.gov.uk/government/publications/government-legal-department-annual-report-and-accounts-2025-26/government-legal-department-annual-report-and-accounts-2025-26.
[5] Author’s JR pre-action correspondence, 20 September–7 October 2025, including GLD acknowledgement and request for additional time, 26 and 29 September 2025. See also R (White) v Secretary of State for Defence, AC-2025-LON-003621 (continuing proceedings).
[6] Author’s Pre-Action / Pre-Complaint Letter — Equality and Human Rights, 14 March 2026; GLD response, 27 March 2026; related Equality Act proceedings N00RM961. These documents are used to characterise correspondence, not establish the merits of any unadjudicated allegation.
[7] R (Anufrijeva) v Secretary of State for the Home Department [2003] UKHL 36: https://www.bailii.org/uk/cases/UKHL/2003/36.html.
[8] Civil Procedure Rules, Part 54, particularly rule 54.5: https://www.justice.gov.uk/courts/procedure-rules/civil/rules/part54; R (Burkett) v Hammersmith and Fulham LBC [2002] UKHL 23: https://www.bailii.org/uk/cases/UKHL/2002/23.html.
[9] Public Service Ombudsman, Principles of Good Administration: https://www.ombudsman.org.uk/publications/principles-good-administration-0; Ministry of Defence, Complaints Procedure: https://www.gov.uk/government/organisations/ministry-of-defence/about/complaints-procedure.
[10] Author’s complaint and forwarding correspondence, September 2025–April 2026, including 30 September 2025 complaint and MP-forwarded request of 2 February 2026; ministerial correspondence concerning the scope of separate complaints. Some procedural characterisations remain contested.
[11] Author’s PHSO/PSO complaint, C-2205680, 26 June 2026, and subsequent updates; Ombudsman procedural correspondence, 17 July and 20 October 2026. Progression to further consideration is not a finding on merits.
[12] FCA Handbook, DISP 1.6.2R, complaints final-response obligations: https://handbook.fca.org.uk/handbook/DISP/1/6.html; Financial Ombudsman Service: https://www.financial-ombudsman.org.uk/.
[13] Armed Forces Act 2006, Schedule 14A: https://www.legislation.gov.uk/ukpga/2006/52/schedule/14A; SPCC Annual Report 2024: https://www.gov.uk/government/publications/service-police-complaints-commissioner-spcc-annual-report-2024.
[14] Pre-Action Protocol for Judicial Review, especially paragraphs 3, 5, 9–12 and 20–23: https://www.justice.gov.uk/courts/procedure-rules/civil/protocol/prot_jrv.
[15] Ohpen Operations UK Ltd v Invesco Fund Managers Ltd [2019] EWHC 2246 (TCC): https://www.bailii.org/ew/cases/EWHC/TCC/2019/2246.html.
[16] Churchill v Merthyr Tydfil County Borough Council [2023] EWCA Civ 1416, particularly paragraphs 60–66 and 74: https://www.bailii.org/ew/cases/EWCA/Civ/2023/1416.html.
[17] Judicial Review Pre-Action Protocol (note [14]); Practice Direction — Pre-Action Conduct and Protocols, paragraphs 8–12: https://www.justice.gov.uk/courts/procedure-rules/civil/rules/pd_pre-action_conduct; CPR 3.1(2)(o): https://www.justice.gov.uk/courts/procedure-rules/civil/rules/part03.
[18] Varda Bondy, Linda Mulcahy, Margaret Doyle and Val Reid, Mediation and Judicial Review: An Empirical Research Study (Public Law Project, 2009): https://publiclawproject.org.uk/resources/mediation-and-judicial-review/; Varda Bondy and Margaret Doyle, Mediation in Judicial Review: A Practitioners’ Handbook (Public Law Project, 2011): https://publiclawproject.org.uk/resources/mediation-in-judicial-review-a-practitioners-handbook/.
[19] CPR Practice Direction 51ZE, Small Claims Track Automatic Referral to Mediation Pilot Scheme: https://www.justice.gov.uk/courts/procedure-rules/civil/rules/part51/practice-direction-51ze-small-claims-track-automatic-referral-to-mediation-pilot-scheme.
[20] Civil Justice Council, Review of Litigation Funding — Final Report (2 June 2025): https://www.judiciary.uk/related-offices-and-bodies/advisory-bodies/cjc/current-work/third-party-funding/.
[21] Legal Aid Agency, Work out who qualifies for civil legal aid: https://www.gov.uk/guidance/work-out-who-qualifies-for-civil-legal-aid; GOV.UK, What you can get legal aid for: https://www.gov.uk/legal-aid; Ministry of Justice, Legal Aid Statistics, April–June 2026: https://www.gov.uk/government/statistics/legal-aid-statistics-quarterly-april-to-june-2026/legal-aid-statistics-england-and-wales-bulletin-apr-to-jun-2026.
[22] R (UNISON) v Lord Chancellor [2017] UKSC 51: https://www.bailii.org/uk/cases/UKSC/2017/51.html.
[23] SPCC statutory and annual-report sources (note [13]); relevant Hertfordshire PCC/Chief Constable financial statements to be examined for any comparison of specific cost centres.
[24] FRY, BNM1 and LRB v Secretary of State for Defence [2026] EWHC 2571 (Admin), judgment of Chamberlain J dated 9 October 2026: https://www.judiciary.uk/wp-content/uploads/2026/10/R-v-FRY-and-others.pdf.
[25] Cabinet Office, Ministerial Direction Regarding Legal Proceedings, letters dated March 2020 and published in December 2020: https://www.gov.uk/government/publications/ministerial-direction-regarding-legal-proceedings.

EDITORIAL AND LEGAL STATUS

This chapter presents legal analysis and a longitudinal documentary case study. Contested assertions about the author’s proceedings are attributed to correspondence and have not been treated as judicial findings. Figures from the author’s July 2026 model are estimates; GLD figures are recorded expenditure rather than a finding of necessary or improper cost. No parliamentary question is described as tabled without an official reference, and the NAO submission is not described as an accepted investigation. The proposed Administrative Reassessment Gate is a reform hypothesis, not an enacted legal obligation. Before external publication, recheck the latest procedural status, relevant reporting restrictions, confidentiality, personal data and without-prejudice or privileged material.

Made on
Tilda