The Attorney-General is an ancient office that is endowed with a very wide range of discretionary powers and functions. Some of those are classed as ‘public interest functions’. These derive from statutes and the prerogative, and include granting consent for certain prosecutions, applying for orders preventing vexatious litigants from bringing proceedings, entering a nolle prosequi, intervening in family cases and appointing an advocate to the court (amicus curiae). For each of these functions, the Attorney-General acts as the ‘guardian of the public interest’ (R v Attorney General, ex p Ferrante (unreported, 1 July 1994, Popplewell J)).

In R (Campbell) v His Majesty’s Attorney-General for England and Wales [2025] EWHC 1653 (Admin), the Divisional Court held that there are certain discretionary ‘public interest functions’ of the Attorney-General that are ‘categorically immune from [judicial] review’. Lord Justice Stuart-Smith and Chamberlain J candidly acknowledged that this outcome may seem ‘anomalous…in the light of the rest of the modern law of judicial review.’ But their Lordships stated that they were bound by several earlier precedents to reach this conclusion and that ‘it is not now open to any court below the Supreme Court to decide otherwise’.

I argue that the earlier precedents binding on lower courts conflict with modern developments in public law. If the issue reaches the Supreme Court, there are good reasons for it to abandon this anomalous category of non-justiciable functions. Without judicial review, there is a risk that the Attorney-General could abuse an unbridled discretion. This could mean that powers intended to be exercised in the public interest could lawfully be exercised for private benefit or other improper purposes. Consistent with modern developments in judicial review, exercises of these powers should at least be reviewable in the courts on limited grounds such as for fraud or bad faith.

Campbell and its Progeny

Geoffrey Campbell was tragically killed in the World Trade Centre in New York City on 11 September 2001. An inquest into his death was conducted by the Senior Coroner for West London who found that Campbell died when an American Airlines aircraft was deliberately flown into the building as part of an attack by Al-Qaeda.

Matthew Campbell, Geoffrey Campbell’s brother, does not accept that conclusion. He hypothesises that the building collapse was caused by the detonation of pre-planned explosives or incendiaries. And he believes that fresh evidence not before the Senior Coroner supports his view. He wished to apply to the High Court for an order quashing the original inquisition and directing another investigation under s 13 of the Coroners Act 1988. But that very section states that an application may only be made ‘by or under the authority of the Attorney-General’. The Solicitor-General (who exercised the function of the Attorney-General under s 1(1) of the Law Officers Act 1997) refused to grant authority to quash the inquest. Matthew Campbell then sought judicial review of that refusal.

A hearing was held to determine the preliminary issue as to whether the challenged decision was justiciable. The Divisional Court (Stuart-Smith LJ and Chamberlain J) held that, on current authority, the public interest functions of the Attorney-General are immune from judicial review. The Court gave careful attention to Gouriet v Union of Post Office Workers [1978] AC 435 where the Attorney-General’s refusal of consent played an important role in determining whether Mr Gouriet could bring proceedings for an injunction in his own name. The Law Lords in their speeches emphasised the unique constitutional role of the Attorney-General to singularly represent the public interest, noting that accountability for any error lies in the political field in Parliament and not in the legal field through the courts.

Although Gouriet was about relator actions (the enforcement of public rights on behalf of others), the Court in Campbell held that two Court of Appeal decisions—both decided after the landmark decision in Council of Civil Service Unions v Minister for the Civil Service [1985] AC 374—confirmed that the scope of the immunity extends to a refusal to enforce certain legislation (R v HM Attorney General, ex p Edey (unreported, 26 February 1992, Court of Appeal)) and a decision not to consent to proceedings for contempt of court (R v Solicitor General, ex p Taylor (1996) 8 Admin LR 206). The Divisional Court held that the immunity extended to the public interest functions more broadly, and the combined effect of Gouriet, Edey and Taylor is that ‘it is not now open to any court below the Supreme Court to decide otherwise’.

Challenging the Claims for Non-justiciability

I argue that, if this issue were to reach the Supreme Court, the Court should depart from these earlier decisions and recognise at least some limited grounds on which the public interest functions of the Attorney-General can be subject to judicial review. I address six claims contained in the earlier precedents. Each one of these justifications should be rejected in light of modern developments in constitutional and administrative law.

Claim 1: not appropriate for courts to review weighing of public interest

The first claim is that weighing different aspects of the public interest is one of ‘political judgment’, which is ‘not appropriate for decision in the courts’ (Gouriet at 524C). Not only is the Attorney-General the ‘guardian of the public interest’ (Ferrante),Lord Fraser went so far to say that ‘he [the Attorney-General] alone is entitled to represent the public interest’ (Gouriet at 519A).

This argument is unpersuasive for three reasons. First, it rests on the assumption that judicial review of a public interest function necessarily requires the court to balance the relevant factors themselves. But the role of courts is far more limited. Modern judicial review operates to ensure that the decision-maker only considers relevant factors, including all mandatorily relevant factors, and not any irrelevant factors. It has repeatedly been emphasised that the weight to be given—including to give no weight at all—is a matter for the decision-maker, subject only to Wednesbury irrationality (Tesco Stores Ltd v Environment Secretary [1995] 1 WLR 759 at 770, 780).

Secondly, the courts have rejected the idea that the executive can be final arbiter of the public interest in other contexts. Before 1968, the doctrine of ‘Crown immunity’ recognised that the objection by a minister to the production of documents ‘on the ground that this would be injurious to the public interest’ was ‘conclusive’ (Duncan v Cammell, Laird & Co Ltd [1942] AC 624 at 642C). That doctrine was comprehensively rejected in Conway v Rimmer [1968] AC 910, which held that the assessment of the competing factors in public interest immunity is a matter for the courts. Therefore, the notion of courts reviewing decisions made as to the public interest is not as foreign as Lord Fraser suggests.

Thirdly, it is difficult to see why the public interest would operate as a complete bar to reviewability when the invocation of national security and foreign relations no longer operates to exclude judicial review. In these fields, the courts calibrate the intensity of their review based on their own institutional limitations with acute sensitivity to the legal framework and the constitutional role of the decision-maker, rather than exclude review completely (see this discussion following R (Duke of Sussex) v Home Secretary [2025] 4 WLR 66). There is no reason why courts could not appropriately tailor judicial review of public interest functions with reference to the institutional expertise of the Attorney-General. 

Claim 2: the Attorney-General has a special or unique position

A recurring justification in the authorities is the constitutional significance of the Attorney-General. Lord Justice Stuart-Smith (Butterfield J agreeing) said that the Attorney-General holds a ‘unique constitutional position’ (Taylor). Some of the earlier cases say that the Attorney-General is performing a ‘judicial act’ or exercises ‘high judicial functions’ (Ex p Newton (1855) 24 LJQB 247; R v Comptroller General (1899) 1 QB 909). This likely means that the functions were part of the administration of justice, rather than the Attorney-General sitting as a court. This reasoning was framed somewhat differently in Taylor. In that case, the Divisional Court held that there was no jurisdiction to judicially review the decision of the Solicitor-General (as the delegate of the Attorney-General) to refuse consent to proceedings for contempt of court under the strict liability rule in s 7 of the Contempt of Court Act 1981. The Divisional Court held that the ‘constitutional position of the Attorney-General’s vis-à-vis the Court’ is different to that of other decision-makers.

This argument does not hold water when one looks at how public law treats other law officers whose relationship is materially similar to the Attorney-General, all of whom could be said to be involved in a broad sense with the administration of justice. Take, for instance, the Director of Public Prosecutions. To prosecute certain offences, the consent of the Director is required. And a relevant factor in deciding whether to grant consent is the public interest. The Attorney-General and Director are in a very similar position: both decision-makers have discretion to grant consent to institute criminal proceedings; both decision-makers consider the public interest. The position of the Director vis-à-vis the court is not materially different to the position of the Attorney-General in Taylor. Yet, the decision of the Director to grant consent can be subject to judicial review on the limited grounds of ‘dishonesty or mala fides or an exceptional circumstance’ (R v Director of Public Prosecutions, ex p Kebeline [2002] 2 AC 326 at 371F).

One of the public interest functions is the power of the Attorney-General to enter a nolle prosequi, which has the effect of staying criminal proceedings. There is obiter that ‘the Attorney-General alone has power to enter a nolle prosequi, and that power is not subject to any control’ (R v Comptroller General of Patents, Designs and Trade Marks (1899) 1 QB 909 at 913–4). The Director of the Serious Fraud Office has a similar power to discontinue a criminal investigation. Even when this involves weighing the public interest in pursuing an important investigation into alleged bribery against the public interest in protecting the lives of British citizens, a discontinuance decision by the Director can be subject to judicial review in ‘highly exceptional cases’ (R (Corner House Research) v Director of the Serious Fraud Office [2009] AC 756 at [30]). The Director’s discretion, Lord Bingham explained, is ‘not unfettered’. The Director must ‘promote the statutory purpose’, ‘direct himself correctly in law’, ‘act lawfully’, ‘exercise an objective judgment on the relevant material available to him’ and ‘exercise his powers in good faith, uninfluenced by any ulterior motive, predilection or prejudice’ (at [32]).

These two examples demonstrate that the public interest functions of the Attorney-General are not entirely unique. Other executive officers have identical or substantially similar functions that place them in the same position of the Attorney-General vis-à-vis the courts and in relation to the administration of justice. The availability of review for decisions taken by the DPP and Director of the SFO, albeit on limited grounds, only serves to highlight how anomalous is the absolute non-reviewability of the Attorney-General’s public interest functions.

Claim 3: the Attorney-General is accountable only to Parliament

Lord Fraser stated in Gouriet (542C) that ‘If the Attorney-General were to commit a serious error of judgment by withholding consent to relator proceedings in a case where he ought to have given it, the remedy must in my opinion lie in the political field by enforcing his responsibility to Parliament and not in the legal field through the courts’. Similarly, the Earl of Halsbury LC said that the Attorney-General could be ‘the subject of adverse comment’ in Parliament but that the courts have no jurisdiction to intervene in the exercise of his or her public interest functions (London County Council v Attorney General [1902] AC 165 at 168–9).

This claim does not accord with the modern understanding of public law. Almost any decision of the executive—whether taken by the Attorney-General or another minister—will have a political hue. The system of responsible government makes clear that all ministers are politically answerable and accountable to Parliament. But this is no justification for why ministers cannot also be legally answerable to the courts: ‘[A]lthough the courts cannot decide political questions, the fact that a legal dispute concerns the conduct of politicians, or arises from a matter of political controversy, has never been sufficient reason for the courts to refuse to consider it’ (R (Miller) v Prime Minister [2020] AC 373 at [31]). The question of whether the Attorney-General exceeded his or her legal powers is a question that is within the constitutional function of the courts to decide.

A potential rebuttal to claim 2 is that the Attorney-General is distinguishable to the DPP and Director of the SFO because he or she is a minister and is therefore subject to more direct forms of accountability by virtue of being a member of Parliament. The existence of political accountability, whether alone or cumulatively, has not been regarded as a proper basis to insulate ministers from judicial review.

Indeed, there are at least some decisions of the Attorney-General that can be the subject of both political scrutiny in Parliament and judicial review in the courts. R (Evans) v Attorney General [2015] AC 1787 offers one example. That case concerned a certificate made by the Attorney-General to override the decision of a tribunal to order disclosure (Freedom of Information Act 2000, s 53(2)). It was a decision that, like many of the public interest functions, required the Attorney-General to engage in an evaluative judgment that weighed competing public interest factors. And parliamentary oversight was facilitated by s 53(3)(a) of the Act itself which provides that, as soon as a certificate is given, a copy must be laid before each House of Parliament. Yet, the Supreme Court held that the decision to issue a certificate was subject to judicial review. This reinforces that the Attorney-General’s accountability to Parliament has not been a bar to judicial review of his or her decisions, including those that involve weighing public interest considerations.

Claim 4: any excess of power is not a jurisdictional error

It has been claimed that an excess of power by the Attorney-General ‘would not go to his jurisdiction; it would go, I think, to the conduct of his office’ (London County Council v Attorney General [1902] AC 165 at 168–9). This passage was from a time when administrative law recognised a distinction between a jurisdictional error of law (which was reviewable) and a non-jurisdictional error of law (which was not, unless the error of law was on the face of the record). This passage suggests that any error by the Attorney-General would be treated as falling in the latter category.

This distinction was however  ‘rendered obsolete’ following Anisminic and, since then, administrative law recognises that ‘a misdirection in law in making the decision therefore rendered the decision ultra vires’ (R v Hull University Visitor, ex p Page [1993] AC 682 at 701). There is no basis to maintain this obsolete distinction solely for the Attorney-General.

Claim 5: allowing judicial review is not ‘wise’ nor consistent with the constitution

Lord Wilberforce stated that maintaining the ‘exclusive right of the Attorney-General to represent the public interest’ is ‘wise’ (Gouriet at 481A). His Lordship expressed his agreement with the words of Lord Westbury LC in an earlier decision that the ‘constitution of the country has wisely intrusted the privilege with a public officer, and has not allowed it to be usurped by a private individual’ (Stockport District Waterworks Co v Manchester Corporation (1862) 9 Jur NS 266 at 267).

This view of judicial review is inconsistent with the modern understanding of the law. It is misleading to frame judicial review as a ‘usurpation’ by a private individual of the public interest. The balancing exercise between competing public interest factors remains—even on review—with the Attorney-General. It is not for a court to substitute its own views. The role of judges is limited to ensuring that exercises of public power meet basic standards of lawfulness and rationality. Broader societal concerns about judicial review amounting to such an usurpation often ignore that courts can expedite hearings and turn around decisions quickly where urgency is brought to their attention. And courts are obliged to refuse relief if it is highly likely that the outcome would not have been substantially different if the error had not occurred (see Senior Courts Act 1981, s 31(2A)).

Moreover, there are good reasons why limited grounds for judicial review should be available for the public interest functions. Consider this hypothetical example: The Attorney-General enters a nolle prosequi for an accused who is on trial. This decision immediately halts the trial and ends all criminal proceedings. It is later discovered that the Attorney-General took a bribe from an associate of the accused. The question would be how leaving exclusive and unfettered power to the Attorney-General does not go beyond the constitutional pale? Why must the court dismiss the case in the face of what is a clear affront to the rule of law? I argue it would be wise to allow judicial review of the public interest functions in these circumstances. Of course, courts should recognise that these are very broad powers with little statutory prescription, and their exercise involves consideration of sensitive policy and public interest factors. And, where a decision involves a ‘pre-eminently political judgment’, the courts should accord the view of the Attorney-General ‘great weight’ (Rehman v Home Secretary [2003] 1 AC 153 at [31]). It is right that courts should be very slow to interfere. But there still reaches a point where bribes, fraud, dishonesty, corruption and bad faith should attract judicial intervention (in much the same way as was recognised in Kebeline and Corner House).

A potential counterargument could be that there are existing checks to address possible abuses of the public interest functions. For example, if the Attorney-General has given consent to bring charges when such a decision is irrational, a court can permanently stay prosecutions to prevent an abuse of process. I accept that there are some public interest functions (such as granting consent) where error can be appropriately addressed by means other than judicial review. In those circumstances, the existence of a convenient and effective alternative remedy operates as a discretionary bar to the grant of relief in judicial review (see Sharma v Brown-Antoine [2007] 1 WLR 780, [14(4)]). But that could not be said for all public interest functions. The facts in cases like Taylor and Campbell are such that judicial review was the last resort to check on the legality of the decision. Allowing judicial review ‘ensure[s] that the rule of law is respected where no other procedure is suitable to achieve that objective’ (R (Glencore Energy UK Ltd) v HMRC [2017] EWCA Civ 1716 at [55]).

Claim 6: Parliament intended non-justiciability

Finally, it has been observed that ‘Parliament must be taken to know the law as stated in Gouriet… and if it had intended the Attorney-General’s discretion to be reviewable by this court in this instance, in my view it would have said so’ (Taylor).

The assumption that Parliament must have necessarily adopted the rule of non-justiciability is inconsistent with repeated statements that ‘[u]nless there is the clearest provision to the contrary, Parliament must be presumed not to legislate contrary to the rule of law’ (R v Home Secretary, ex p Pierson [1998] AC 539 at 591) and the ‘constitutional right of access to the courts is inherent in the rule of law’ (R (UNISON) v Lord Chancellor [2020] AC 869, [66]). Contrary to Taylor, the presumption cuts the other way: Parliament’s silence should not be assumed that they intended for non-justiciability. Under the principle of the rule of law, ‘Parliament cannot entrust a statutory decision-making process to a particular body, but then leave it free to disregard the essential requirements laid down by the rule of law for such a process to be effective’ (R (Privacy International) v Investigatory Powers Tribunal [2020] AC 491 at [123]). It is absurd to assume that Parliament intended that were the Attorney-General to act irrationally, in bad faith or corruptly in exercising a power then a person with sufficient interest would have no recourse to challenge that decision in the courts. The better view is that Parliament can be presumed to intend that any power it grants is always exercised rationally, in good faith and for a proper purpose.

Conclusion

The Attorney-General’s public interest functions are of very wide import. They range from ‘gatekeeper’ functions when giving consent or authority for certain kinds of proceedings to ‘protective’ functions when appealing an unduly lenient sentence (Criminal Justice Act 1988, s 35) or intervening in family cases (Family Law Act 1986, s 59). The precedents holding that the exercise of all these functions by the Attorney-General is non-justiciable simpliciter is anomalous, especially when compared to the reviewability, however limited, of the decisions of other similarly situated executive officers and in light of the modern developments in public law.

Whilst the Divisional Court in Campbell was only deciding the issue based on binding authorities, the matter has not been tested in the House of Lords or Supreme Court since Gouriet. If the issue does reach the Supreme Court, there are good reasons not to slam shut the courthouse door. It is fundamental to the rule of law that decisions of the executive are reviewable by the courts in proceedings brought by a party with a sufficient interest. To hold otherwise is to stand the constitutional principle of the rule of law on its head.

Dane Luo is the Farthing Scholar in Administrative Law at Pembroke College, University of Oxford

(Suggested citation: D. Luo, ‘The Anomalous Islands of Public Interest Functions Immune from Judicial Review’, U.K. Const. L. Blog (14th October 2025) (available at https://ukconstitutionallaw.org/))