Daniella Lock: The Third Direction Case Part Two: The Doctrine of Necessary Implication and Uncertainty in National Security Law

Part One of this post presented the background to the ‘Third Direction’ case, which concerns a recently disclosed Government policy to authorise agents of the Security Service (MI5) to participate in criminal conduct, and will  will soon be heard in the Court of Appeal. Part One argued that from its the outset the case illustrates significant obstacles to the rule of law which commonly arise in national security litigation, and which are contributing to a broadening gap of uncertainty between formal, public, national security law and the way that such law operates in practice. Part Two examines the IPT judgment on the Third Direction published in December, which includes a majority ruling (per Lord Justice Singh, Lord Boyd and Sir Richard McLaughlin) and two dissents (per Charles Flint QC and Professor Graham Zellick QC). The post argues that the findings by the majority, in particular its application of the doctrine of necessary implication, principally serves to enhance uncertainty with regards to the operation of national security law. The majority ruling raises a number of questions regarding the nature of the powers found to be lawful, the role of law in national security policy, and the precise role being played by oversight bodies in this area.

The judgment examines four principal issues related to whether the policy: 1. Has a lawful basis in statute or common law; 2. Amounts to an unlawful de facto power to dispense with the criminal law; 3. Whether the secret nature of the policy, in the past and in its current form, means it is unlawful under domestic principles of public law; 4. Whether the policy, the practices authorised under the policy, and its oversight, are compatible with the UK upholding its obligations under the ECHR. The majority found in favour of the Government on all four issues, concluding that the policy was lawful. The judgment represents the first time an IPT judgment has publicly included a dissent, and will be the first IPT judgment to be appealed following the coming into force of s67A of Regulation of Investigatory Powers Act 2000 (RIPA) inserted by the Investigatory Powers Act 2016 (IPA).

The Doctrine of Necessary Implication

The majority found that a lawful basis for a policy permitting the commission of criminal offences was necessarily implied by s 1(2) of Security Services Act 1989 (SSA) read with s 2(1) of the SSA. S 1 (2) SSA states the ‘function’ of the Security Service shall be ‘the protection of national security and, in particular, its protection against threats from espionage, terrorism and sabotage, from the activities of agents of foreign powers and from actions intended to overthrow or undermine parliamentary democracy by political, industrial or violent mean’. S2 (1) SSA refers to the need for the ‘efficiency’ of the Service, which the Director-General is responsible for ensuring.  

In coming to this conclusion, the majority first emphasised that it is ‘well-established in public law that a public authority has not only those powers conferred upon it by statute but also implied powers’ [55]. Second, the majority reasoned that the ability of the service to infiltrate proscribed organisations was both essential and dependent on the policy being found to be lawful. The majority reasoned that prior to the passing of the SSA, the Security Service must have been engaged in the ‘running of agents, including the running of agents who are embedding in an illegal or criminal organisation’ [60]. Furthermore the ‘express terms’ of the SSA ‘make it clear’ that it was intended to continue the existence of the Security Service and its operation [60]. On this basis, the majority concluded that it was ‘impossible’ to ‘accept that Parliament intended in enacting the 1989 Act to bring to an end some of the core activities which the Security Service must have been conducting at that time, in particular the context of the “Troubles” in Northern Ireland’ [60]. The majority further noted the reference to ‘efficiency’ in s2 (1) SSA and states that it could ‘hardly be said to be an efficient exercise’ of the performance of either the Director-General’s or the Security Service’s functions if they ‘could not carry on doing an essential part of their core activities’ [61].  In concluding there must be an implied power for the policy in the SSA, the majority asserted that the closed evidence it had examined and the ‘events of recent years, for example in Manchester and London in 2017’ serve to ‘underline the need for such intelligence gathering and other activities in order to protect the public from serious terrorist threats’ [63].  


Both dissents in the judgment depart from the majority ruling in its application of the doctrine of necessary implication. The dissent of Charles Flint QC stated thathe was unable to find a necessary implication from the words of the 1989 Act that the commission of a tort or crime against a person could be authorised by the Director-General of the Service. Flint noted that since the Interception of Communications Act 1985 ‘exceptional powers’ have been granted using ‘substantially the same drafting technique’ whereby such powers are drafted in a manner which requires a warrant or authorisation which establishes a lawful means for the power to be exercised. He highlighted that this is the general scheme used in the SSA, Intelligence Services Act 1994 (ISA), RIPA for example at section 21(2)) and the IPA (for example at section 6(3)) [129]. Flint stated that an implied power which ‘authorises conduct contrary to the general criminal and civil law’ but leaves the person engaging in such conduct ‘liable to criminal prosecution’ would be ‘extraordinary’ [129].

The second dissent was given by Professor Graham Zellick QC. Zellick argued there was no evidence of any power to authorise agents to participate in criminality prior to the SSA [144] – [145], just as there was not for lawful interference with property until it was expressly introduced by s3 SSA.  Zellick contended that s1 of the SSA is not the source of any of the Service’s powers but ‘sets our functions or purposes’ and ‘might be called an objects clause’ which in fact ‘defines the limits or scope of the Service’s activities’, rendering it therefore a ‘limiting provision’ [151]. Zellick further described the intention attributed by the majority to Parliament as ‘fanciful’ [161] and detailed instances both prior to and following the SSA where the prospect of passing legislation relating to agent participation in criminality was explicitly discussed by MI5, government departments, legal advisers, ministers and the Law Officers but put to one side due to fears it would not be supported by Parliament [163]. According to Zellick, the majority’s use of necessary implication did not satisfy the requirements of leading authorities regarding the doctrine of necessary implication which require it to be only narrowly constructed ([170] – [173]). He stated that while the power to participate in criminality ‘may well be sensible and desirable, even essential’ Parliament would likely be ‘astonished to be told that it had conferred this power in 1989’ ([174]).

Uncertainty in National Security Law

The dissents provided against the majority application of the doctrine of necessary implication are powerful and determining the proper application of the doctrine of necessary implication in this context will be a key issue in proceedings brought before the Court of Appeal. Despite the fact that the case represents an attempt to clarify an area of national security law, it is clear that if the majority view is upheld, the case will create further uncertainty in the legal regime underpinning the Government’s national security activity – including in relation to the areas set out below.

Uncertainty surrounding the nature of national security powers

The reasoning employed by the majority raises the question as to what other forms of criminal participation may now be considered lawful beyond membership of proscribed organisations. This is linked to what may be aptly described as ‘backwards reasoning’ on the part of the majority in the IPT ruling. The reasoning is backwards insofar as the necessity of a specific power – for agents to join proscribed organisations – is used as a starting point to infer the necessity of a broader set of powers – a general power for agents to participate in criminality. This is in order to determine the lawfulness of that broader set of powers. This contrasts with the more common approach to judicial reasoning required by the principle of legality, for example as is endorsed in ex p Simms, which is to only infer the existence of specific powers (where such powers are invasive state powers) that are clear and unambiguous. Even if the case that there might be a need for specific powers is found to be clearly made out, that is not usually taken to determine the lawfulness of a broader set of powers of which they are part. This is because even if that broad set is desirable, Parliament will ordinarily wish to legislate including safeguards and limitations to protect the public interest and appropriately constrain the powers of the intelligence agencies.  Indeed, this is reflected in comparable recent legislation passed in Canada. The legislation refers to criminal conduct which may be engaged in by the Canadian Security Intelligence Service and sets out clear limits on the kind of criminal conduct that may be engaged in (see Section 20 (18)).

Questions regarding the nature of the powers to participate in criminality are particularly significant as the majority ruling notably avoided any analysis of whether or not action which potentially breaches ECHR rights was authorised by the policy. This is aside from stating that there was ‘nothing inherent’ in the policy which creates a ‘significant risk’ of a breach of article 3 [100]. In avoiding ruling on such matters the majority stated that a ‘fundamental difficulty’ is that the ‘question of whether there has been a breach of a Convention right is usually one to be determined after the event on the concrete facts of a particular case rather than in the abstract’ [101]. It is not clear what this means in light of the IPT presumably having examined evidence pertaining to the application of the policy in practice in closed proceedings. Ultimately the Tribunal ruled that the parties to the case did not have standing to determine the breach of Convention rights [107]. Again the reasoning for this is not entirely clear given that parties to the case included the Pat Finucane Centre which represents families affected by the conflict in Northern Ireland and therefore potentially at risk of having been directly affected by the Government’s policy (‘potentially at risk’ of ECHR violations is the basis for standing in relation to surveillance cases considered by the IPT in the case of Human Rights Watch and ors v Secretary of State for Foreign and Commonwealth Office and ors.).

Uncertainty as to the role of law in national security protection

That such a legal basis for such a broad set of powers could be inferred to exist from a general statutory provision also raises questions as to the role of more specific legal provisions in national security law. If the Government can have national security powers read into statutory provisions which are not explicit about providing powers at all, let alone a broad set of powers, this begs the question as to the purpose of provisions which are more specific in setting out powers to govern activities of the security services. Lawmakers themselves may also question what is the point of scrutinising specific and sometimes complex statutory schemes for powers, such as those referred to by the dissenters and contained in SSA and RIPA, when the Government is able to operate on the basis of broad policies it has written up in-house and given a legal basis by general statutory provisions.

Uncertainty regarding the need to disclose national security policies

The majority ruling also sends mixed signals regarding the need for Governments to disclose their national security policies. In deciding whether the policy was consistent with the principle of legality prior to its disclosure, the majority ruling stated it saw ‘no practical purpose would be served by this Tribunal granting any remedies to the Claimants in this regard’ [90]. In avoiding a finding of historic illegality of the policy, the Tribunal departed from its previous approach to similarly broad, invasive and previously secret policies (for example in Liberty v GCHQ (No 2)  and Privacy and Greennet.). It also contrasts with a recent ruling made by the Federal Court of Canada which found extensive unlawfulness with regards to participation in criminality by the Canadian Security Intelligence Service prior to the passing of legislation which explicitly regulated such conduct. Importantly, the IPT ruling indicates to the Government that it will not necessarily be penalised when it keeps secret policies governing a significant aspect of its activity, in a manner which enables it to essentially operate outside of the law, leaving it unclear whether the Government needs to disclose such policies at all.

Uncertainty as to the purpose of IPCO/ISC

Other questions raised by the majority ruling in the judgment include the precise role of the bodies responsible for providing oversight in national security law. On the question as to whether there were adequate safeguards against the abuse of powers provided by the Government’s policy, the claimants argued that as the Intelligence and Security Committee (ISC) and the Investigatory Powers Commissioner’s Office (IPCO) had been told by the Government that their role was not to consider the lawfulness of such a policy, it was not subject to proper oversight [94]. The majority argued in response that ‘it was not necessary’ to ask the ISC or IPCO (the ruling says ‘IPT’ but this is presumably a typo) to ‘provide any endorsement of the policy’ as this was ‘not their function’ [95]. The majority adds that such ‘questions of law’ are ‘ultimately ones for courts and tribunals to determine’ [95].  

While no one would disagree that courts make determinative assessments of law, the idea that the function of oversight bodies does not include making assessments regarding the lawfulness of activities of the Security Service is not consistent with the image often presented of the role of oversight by the UK Government, as well as by the IPT itself. Indeed in requesting that Parliament pass RIPA, the UK Government (in the form of Jack Straw, Home Secretary at the time) expressly stated that the role of the commissioners would be to ‘reassure Parliament and the public’ that investigatory powers used by the Security Service and law enforcement ‘are being used properly’ (see para 777), and former judges are required to be appointed in both bodies to make such assessments (see RIPA, s59 (5), now repealed, and s227 (2) of the IPA).  Conclusions on the lawfulness of Security Service activities by the ISC and IPCO have also been quoted by the IPT. This includes when reasoning that the Government’s powers are lawful, precisely by virtue of these bodies existing to determine ‘compliance’ of the security services with their ‘obligations’, which must include their legal ones – thus satisfying ECHR requirements that sufficient safeguards exist to ensure such powers are not subject to abuse. This was a key factor in the IPT finding that hacking powers were lawful in Privacy and Greennet  [77]. In light of this, the majority’s commentary on the ISC and IPCO exposes a lack of coherence in the system of national security oversight.

Concluding remarks

It is true that national security litigation is always underpinned by some degree of uncertainty, due to the presence of secret proceedings and intentionally broad legal definitions meant to enable the security services to respond to unpredictable national security-related threats. However, as this post has highlighted, the majority ruling in the IPT in this case raises questions pertaining to the fundamental functioning of the regime underpinning Government activity in the domain of national security.   The result is to enhance uncertainty in this area of executive activity and thus undermine the rule of law in its already fragile state as far as national security powers are concerned.

Daniella Lock, @DaniellaLock, PhD student and Teaching Fellow, University College London

The author would like to give warm thanks to Professor Michael Gordon and Professor Alison Young for their helpful comments on an earlier draft of this piece. Any remaining errors are the author’s own.

(Suggested citation: D. Lock, ‘The Third Direction Case Part Two: The Doctrine of Necessary Implication and Uncertainty in National Security Law’, U.K. Const. L. Blog (24th July 2020) (available at https://ukconstitutionallaw.org/))