In March 2021, the Court of Appeal handed down a unanimous judgment in Privacy International v Foreign Secretary ‘the Third Direction case’. The unanimous decision held permitting the Security Services (MI5) to authorise criminal conduct by agents was lawful. The authority for the policy existed by necessary implication in ss 1-2 of the Security Services Act 1989. The Court of Appeal also held that prior to the 1989 Act, legal authority for the predecessor policy was vested in the Royal Prerogative.  The decision was appealed from the Investigatory Powers Tribunal (IPT), which ruled by 3-2 that the policy was lawful. 

The issue has now been rendered somewhat academic by the Covert Human Intelligence Sources (Criminal Conduct) Act 2021 which permits a broad range of public authorities (including any of the intelligence agencies) to authorise ‘criminal conduct’ by CHIS. There is no statutory limit on the nature of criminal conduct which may be authorised.

Notwithstanding the 2021 Act, the Third Direction Case has broader implications This blog argues that the ruling construes the scope of the principle of legality too narrowly and the scope of Prerogative power too broadly.

Background and Context

The dispute pertained to a previously secret policy by the Security Service entitled ‘Guidelines on the Use of Agents who participate in Criminality – Official Guidance’. This was preceded by another document authored in 1952 by then Home Secretary Sir David Maxwell Fyfe known as the ‘Third Direction’ from which the present case takes its nickname.  It is worth pausing to consider the purpose of these policies. To safeguard national security, the Security Service needs to run undercover ‘agents’ known as ‘Covert Human Intelligence Sources’ (CHIS) in proscribed organisations. Participation in the activities of these organisations may necessitate involvement with criminal or tortious conduct on the part of the agents who need to maintain their cover. A public statement by the Intelligence and Security Committee (ISC) explains that such agents ‘provide invaluable information to assist the security and intelligence Agencies in their investigations and play a significant, often critical, role in identifying and disrupting terrorist plots.’ In this regard, the policy is something of a ‘necessary evil’.

However, beyond an assurance by the Court of Appeal that nothing in the content of the policy posed a ‘significant risk’ of any breach of Convention rights, there was no express limit on the nature of criminal conduct that could be authorised. The case arose against the backdrop of morally controversial conduct by security and law enforcement agencies, including the undercover policing scandal and the death of Northern Irish solicitor Patrick Finucane. An independent inquiry into the death of Mr Finucane by Desmond de Silva QC concluded that there were ‘positive actions’ by State employees to facilitate Mr Finucane’s death. 

Privacy International advanced three unsuccessful grounds of appeal. First, they argued that the policy could not be traced to a legitimate statutory source. Secondly, they argued that the policy was unlawful because its effect was to confer a de facto immunity from prosecution upon those agents authorised to commit criminal offences. The third ground of appeal (rendered moot) by the conclusions on the first two grounds was that the policy violated the ECHR because it was ‘not in accordance’ with domestic law.

It is convenient to deal with the more concise second ground of appeal first. The Court of Appeal flatly rejected the appellant’s claim that the policy had the effect of conferring immunity from prosecution on agents. The policy explicitly stated that the ‘authorisation’ of such conduct by the Security Services had ‘no legal effect’ (emphasis original) and conferred ‘no immunity from prosecution’. In view of this, the Court held that the policy ‘did not place the Security Service above the law’. This is clearly correct.  Policies do not have the force of law (Gillick) and clearly cannot alter either statute or common law.   As such, the Third Direction would have no impact upon either the substantive application of criminal law or prosecutorial discretion. At most an authorisation under the policy may be a relevant consideration for a prosecutor.  

The Principle of Legality

In relation to ground one the appellants claimed that the policy was contrary to the principle of legality as expressed by Lord Hoffmann in ex parte Simms. It is impossible to capture the precise scope of the ‘principle of legality’ emanating from Simms in such a short post. However, Simms is famous for Lord Hoffmann’s maxim that ‘fundamental rights’ shall not be abrogated by statutory language which is ‘general or ambiguous’. If Parliament wishes to override fundamental rights or principles it must do so either using explicit statutory language or by necessary implication. Over the past twenty or so years this principle of strict interpretation has been extended to protect the principles of tort law, criminal law, and constitutional principles from inadvertent erosion by a complacent legislature. 

The Court of Appeal considered this formulation of legality as a ‘statement of principle’ to be irrelevant. The principle of legality did not apply because the policy was lawful under sections one and two of the Security Services Act 1989. Under the principle of legality in Simms Parliament has always been able to violate fundamental rights or constitutional principles by ‘necessary implication’. The test for determining what constitutes a ‘necessary implication’ comes from the House of Lords ruling in Morgan Grenfell. According to Lord Hobhouse, determining when a statute creates a necessary implication is a question of ‘express language and logic’ as opposed to ‘interpretation’. He then expressed the somewhat tautological view that a necessary implication is one which ‘necessarily follows from the express provisions of the statute construed in their context’. This was not to be confused with a mere ‘reasonable implication’. Although sections 1 and 2 of the 1989 Act were silent as to the powers of the Security Service, the Third Direction was deemed lawful by necessary implication because the Security Service already possessed the power to authorise criminal conduct under the Royal Prerogative before the Act was enacted. In so holding, the Court of Appeal appears to have cast aside the requirement of ‘language’ and proceeded entirely according to a logic of pragmatism.

Thereafter, the Court described the appellant’s reliance on the principle of legality as ‘paradoxical’. This was because the Third Direction purported to authorise criminal conduct with the aim of ‘preventing the taking of innocent life and to inhibiting the activities of those having no regard whatsoever to any principle of legality.’  This dictum elides two senses of ‘legality’ in an unhelpful way. The principle of legality as expressed in Simms is materially different from respect for legality as an individual or organisation’s general respect for the criminal law.  Legality in the Simms sense is both a subset of the rule of law and an evolving set of standards by which the common law seeks to evaluate the legislative process. It is not about whether an individual or group respects or disregards legal norms. These are the two different sense of legality which the Court of Appeal elides. These different senses of legality are incommensurable. They should not be crudely traded off in the evaluation of whether a statute is a sufficient legal basis for a policy. 

During its analysis of the principle of legality the Court also expressly rejected the approach to legality by Professor Graham Zellick QC in his dissenting judgment in the IPT. He held that Morris v Beardmore demonstrates that the principle of legality prevented the authorisation of tortious conduct which adversely impacted common law rights ‘in the absence of express provision or necessary implication to the contrary’ by Parliament. In Morris the House of Lords held that the Road Traffic Act 1972 did not authorise a police officer to engage in trespass to breathalyse a suspect. By extension, the Security Services Act 1989 should not be capable of authorising more serious, criminal conduct by way of statutory ambiguity or silence. However, The Court of Appeal distinguished Morris because it concerned the Road Traffic Act 1972, a context ‘far removed’ from the present case. 

Instead, the Court held that the present case was analogous to Buckoke v Greater London Council. In Buckoke the Council’s policy permitting Fire Service vehicles to ignore traffic lights on the way to emergencies was deemed lawful. According to the Court, this analogy was apposite because ‘the background…also involved the potential protection of life and property’. However, the tolerance of such a minor illegality does not readily equate with the implications of the Third Direction. There is no doubt that permitting the running of a red light on the way to extinguish a fire is a pragmatic perspective for the law to take. When weighed in the balance this is a tolerable offence if the result is the preservation of life. However, national security is (at its margins) an inherently political concept. It is complicated and fluid. There can be reasonable disagreement about which groups are subversive. For this reason, it seems unwise to render the principle of legality nugatory in the face of a policy authorising a potentially broad range of criminal conduct because the legal system has overlooked minor criminality in the face of a straightforward and apolitical danger (fire) elsewhere.  

The above contextual approach also sits uneasily alongside a recent judgment of the Administrative Court. In Privacy International v Foreign Secretary and GCHQ section 5 of the Intelligence Services Act 1994 was deemed incapable of authorising ‘thematic’ warrants for large-scale computer hacking by GCHQ. The Administrative Court held that the national security context did not exempt s 5 from the strictures imposed by the principles of legality. Therefore, the Foreign Secretary could not rely upon section 5 due to the common law’s ‘longstanding aversion’ to general search warrants. The potential consequences of the Third Direction case are more offensive to legality than the creation of thematic warrants. The policy has the potential to surpass the invasion of digital privacy in its impacts. The principle of legality cannot at once be contextual and a-contextual if it is to be a workable principle of public law.

The Royal Prerogative 

It may be, however, that the nature of national security requires a certain degree of trust in the Security Services to exercise such a broad power with appropriate restraint.  But even if one can accept the pragmatic reasons for jettisoning a more robust vision of the principle of legality for the public interest, the Court’s assessment of the scope of the Royal Prerogative raises broader concerns.  Although it was now clear that all the powers of the Security Service (whether express or implied) were entirely statutory powers, because as a matter of necessary implication’ under sections 1 and 2 of the 1989 Act the Security Service retained ‘the (essential) power to instruct agents to participate in criminality’. In reaching the conclusion, the Court affirmed the dictum of Laws LJ in R (A) v Director of Establishments of the Security Service that ‘[a]ll the functions of the Security Service are and have been since the coming into force of the Security Service Act 1989, statutory functions’ (emphasis original). Before 1989, however, the power to authorise criminal conduct was vested in the Royal Prerogative. 

Despite being described as a ‘residue’ of discretionary power (Dicey) and a ‘relic of a past age’ (Burmah Oil) the precise nature and scope of the prerogative powers more generally remains contentious. The recent landmark judgments in Miller (No. 1) and Miller (No. 2) placed significant constraints on the Royal Prerogative. These included preventing its use to subject fundamental rights to unilateral executive repeal and ensuring that the Government could not use Prerogative powers to prorogue Parliament to frustrate its constitutional role. However, the present case illustrates the Janus-faced nature of legal control to which the Prerogative is subject. It is well-known that the Government’s power to deploy the armed forces in combat overseas, arguably the ultimate use of force, is a Prerogative power. However, what is perhaps less well-known is the range of interference with domestic legal rights which can occur and have occurred in the name of national security.  These include trespass (De Keyser’s Royal Hotel) and the destruction of private property (Burmah Oil). Although both rulings occurred within an explicit paradigm of war, the Court of Appeal held that lawful exercise of the Prerogative was possible in wider circumstances. This was because threats to national security ‘are, no doubt, not to be equated precisely with threats arising in time of war’ (emphasis added). The interests of national security are significantly more protean than the paradigm of war, albeit there is clearly some overlap. In fact, Northumbria Police Authority meant that ‘the Secretary of State had authority at all times (not just in times of war or actual emergency) to maintain the Queen’s peace and to keep law and order’.  Action under the Prerogative to this end was lawful unless it was ‘contrary to statute’.

The Third Direction Case takes a broad view of the Security Service’s statutory powers. It also emphasises that the interests of national security requires a pragmatic approach to statutory interpretation. At the same time, it is somewhat ambivalent as to the relationship between statute and Prerogative in the broader national security context.  In Miller (No. 1) the majority accepted the proposition in De Keyser’s Royal Hotel that an Act of Parliament in a cognate area deprives the executive of recourse to the Prerogative. Although De Keyser’s Royal Hotel was cited, this part of the ruling was not explicitly reaffirmed as a general proposition. Given the breadth the Court accorded to the 1989 Act, and the enactment of the new Covert Human Intelligence Sources Act 2021, the proposition from De Keyser’s Royal Hotel most likely still holds in the present context. However, the ruling leaves a glimmer of doubt as to whether all exercises of the Prerogative in the name of national security affecting domestic legal rights have truly been extinguished by statute. An appeal would be an opportunity for the UK Supreme Court to offer a welcome clarification.

My thanks to Professor Alison L. Young and Professor Michael Gordon for helpful comments on a previous draft. All errors and omissions are mine alone. 

Hayley J. Hooper is an Associate Professor of Law at the University of Oxford

(Suggested citation: H. J. Hooper, ‘The Principle of Legality and Prerogative Power after the Third Direction Case’, U.K. Const. L. Blog (26th April 2021) (available at https://ukconstitutionallaw.org/)