Investigatory Powers Bill

Written Evidence submitted by Tom Hickman (IPB 66)

1. This submission focuses on ensuring the effectiveness of the "double lock" (Part A) and ensuring the effectiveness of ex post oversight (Part B). (A submission addressing wider issues was made to the Draft Committee on the draft IP Bill). [1]

A. Ensuring effectiveness of Judicial Commissioner approval (the "double lock")

 

2. The so-called double lock is trumpeted as the key safeguard introduced by the Bill.

3. Much attention has been focused on the "judicial review" standard to be applied by a Judicial Commissioner ("JC"). [2] Of more importance however is what the JC is reviewing, what information needs to be supplied to the JC and what powers can be exercised without JC approval.

4. As currently drafted, the Bill has a door with a double lock but there are various side doors which public officials can go through to get the data which essentially undermines the double lock’s effectiveness.

5. In order to close these side-doors, I suggest that the Bill should be amended to address the following issues.

1. "Thematic" warrants delegate targeting decisions from Judicial Commissioners to the officers conducting an operation

6. The provisions for approval of warrants relating to a "group of persons who share a common purpose" or who "may carry on, a particular activity" (cl. 15(2) (interception) and 90(1) (EI) - so-called "thematic warrants" – remain too vague and broad with the consequence that rather than approving the scope of such warrants, JCs will effectively delegate authority to the police and intelligence services.

7. The notion of a "group’ is vague. It could also apply to persons who are jointly engaged in a demonstration (a "group") even if they have no prior association. Or to a "faith group" of persons who share a certain faith or approach to their faith. An interception could also apply to a group of persons on the basis that they "may" in the future carry out a particular activity although not carrying out such an activity: thus it could apply to a particular faith group on the basis that it may in the future carry out certain extremist activities.

8. Clause 90(1) is even broader and would also allow, for example, EI warrants to be granted in respect of equipment in a particular location or in more than one location for the purposes of single operation. The Government accepts that this could apply to "cover a large geographic area or involve the collection of a large volume of data". [3] It could apply to all mobile phones in a town, for example.

9. Such warrants give extensive discretion to executive officers. They essentially delegate the task of deciding who is a proper target for interception to such officers rather than it being determined by the Secretary of State and the JC. This is contrary to the fundamental objection against general warrants which require coercive action to be authorised by the person responsible for authorising the action. [4] Furthermore, such warrants do not have the same protections for JC approval of examination of material as required for bulk intercepted material, thus essentially allowing a form of mini-bulk interception with lesser protections for individual rights.

10. Cl. 15(2)(a) should be deleted and the powers in cl. 90(1) greatly reduced. In the case of interception, that would still allow warrants to relate to an "organisation" (Cl. 15(1)(a)) as well as allowing warrants to include more than one person targeted by a particular operation (cl. 15(2)(b)) (if such warrants could be obtained for each person individually). This latter point would address the Government’s concern about a kidnapping [5] because a warrant relating to Kidnappers 1 – 5, for example, would remain available. It is a mistake to think that interception warrants for more than one person need to be able to specify the true identity of each person, rather than identify that there is a person (true identity unknown) who is a suspect. The powers in Cl. 15(2)(a) are not needed for such a situation.

11. The Government’s other stated concern to justify thematic warrants relates to interception of users of illicit websites. There is nothing wrong in principle in intercepting all communications accessing illicit websites by analogy with all persons phoning an illicit premises, such as a known criminal base or hideout. But that does not justify the great width of the power in the current draft clauses. An amendment to Cl. 15(1)(b) to state "premises or other specific location (including website)" would meet this concern without the unnecessary width of cl. 15(2)(a); 90(1)(b)).

2. Modification powers by-pass Judicial Commissioners

12. There are very wide powers in clauses 30 and 104 for modification of judicially approved warrants. If it is intended that Judicial Commissioners will examine the detail of warrants – who they apply to and what safeguards are in place – then this scheme is subverted by allowing public officials to amend the warrants to add different persons or remove or alter safeguards without judicial approval.

13. For example, a warrant for persons X, Y and Z who are the subject of an investigation into a particular extremist faith group, could be amended to add more and more people who associate with such persons (or who are family members etc) without judicial oversight.

14. It is difficult to see how the double lock can provide public confidence if public officials can obtain data through such a side door. In the case of adding a person, a JC should be informed, for example, not only of the addition, but why it was not reasonably practicable to include the person at the outset (see cl. 27(4)). That will introduce an important discipline and protection into the double lock system that is currently absent.

3. Need to specify safeguards in warrants.

15. Cl 27 (8) prescribes that where a targeted warrant authorises interception or the examination of communications,

"the warrant must specify the addresses, numbers, apparatus, or other factors or combination of factors, that are to be used for identifying the communications." (emphasis supplied)

16. This is welcome. But there should also be a requirement for warrants to specify the restrictions on intercepting communications or retaining and examining such communications as are considered necessary and appropriate – for instance requiring destruction of communications found to be between or from family members of a target, or intercept that is found to be with the target’s lawyer or medical professional.

17. There should be a specific statutory requirement for these protections to be set out in a warrant, which would ensure that these are matters that always form part of the JC’s proportionality assessment and determination.

18. I therefore propose clause 27 be amended to read,

"the warrant must specify

(i) the addresses, numbers, apparatus, or other factors or combination of factors, that are to be used for identifying the communications; and

(ii) such additional protections and restrictions on the obtaining, retention or use of intercepted communications and secondary data as is considered necessary to ensure the proportionality of the warrant.

An additional sub-paragraph should be added to clause 101 (Requirements which must be met by EI warrants) introducing the same protection to EI warrants.

4. Full JC approval of all bulk datasets

19. The protection of the JC to approve bulk personal dataset warrants is, as currently drafted, of very little value. These can be held under "class" warrants which specify a general class of dataset (cl 177, 179). When agencies apply for such class warrants it will be said that such bulk datasets are in general terms necessary for the purposes of national security or preventing serious crime. The JC will have no sight of the individual datasets or even be told what individual datasets make up the class.

20. It will be almost impossible for the JC to do anything other than approve such a warrant. Moreover, there will be far less than complete visibility about what is included within the class and therefore what is being authorised.

21. Since datasets are retained and are not in general time sensitive, there is no reason why retention of all bulk datasets cannot be approved specifically by a JC. The double lock power is only meaningful if a JC considers whether each individual dataset is necessary. That does not mean that numerous bulk datasets cannot be approved under a single warrant: but it does mean that approvals cannot be made for bulk datasets that are never disclosed or identified to a JC.

22. Thus, a warrant could still be issued for travel information generated by airlines. However, each bulk-data set should be listed so that the JC has visibility over the amount, breadth and nature of the information being approved. Such a list might, for instance, identify bulk datasets that the JC would not have appreciated would be covered by a class warrant.

23. In short: the provision for class datasets should be deleted. Practical difficulties can be accommodated without a class approval.

5. Public rulings

24. If the double lock system is to engender genuine public confidence, JCs should be specifically empowered to make public redacted rulings on points of principle. The Government appears to have accepted that JCs are able to instruct counsel to act as amicus, which is to be welcomed.

6. Communications data

25. JC approval should be required at least for the most sensitive categories of communications data [6] including but not limited to,

a. information between lawyers and their clients and doctors and health professionals and their patients. At present judicial approval is limited to journalistic sources;

b. movement / location information (including that acquired by bulk interception)

7. Separation of JC approval role from ex post oversight

78 It is an important part of the judicial function and vital for public confidence that persons exercising judicial functions do not receive briefings from and do not meet informally or formally with those who make applications to them that they are required to approve or reject. The submissions and evidence presented to a judge or quasi judicial officer should be limited to that which is submitted within the formal confines of the judicial or quasi-judicial process.

77 It is therefore important that there is an institutional or sub-institutional separation between JCs who consider warrant applications and inspectors who engage in post-hoc investigations and monitoring.

 

79 This means that it would be inappropriate for JCs to carry out the task of inspectors. They should not be going in to the agencies and meeting with them formally or informally or visa versa. Whilst such a system has operated in relation to the Intelligence Services Commissioner hitherto, it should not be expanded under the new Act. The IP Bill therefore should draw a distinction between JCs whose job it is to approve or reject warrants, and inspectors who have an ex post reviewing function.

8. Independent Counsel

80 There will be many cases in which the role of the JC will only be truly effective if independent counsel can test and question the position of the agency seeking the warrant. The Government has accepted in principle that it will sometimes be appropriate for JCs to be aided by security cleared Counsel. [7]

81 But the Government has made no statutory provision for this, merely stating that the budget make provision for it. This is not good enough. The protection of individual rights should not be dependent on the extent of budget provision made by the Government. Provision for this should be made in the legislation, providing JCs with a clear statutory mandate which can, if necessary, be invoked to resist budgetary pressures.

B. Ensuring effectiveness of oversight

1. The IPC’s ex post oversight function

26. Submissions on the need for a separation between the oversight functions of the IPC and the role of JC’s in giving prior approval for warrants have been set out above.

27. The provisions on error reporting, although improved since the November draft Bill, still contain three fundamental defects.

 

85 First, the conditions set out for error reporting disclosures remain unsatisfactory. The focus of cl. 198(1)(b) and (2) is on the seriousness of the error defined in terms of the consequences of the error, which must have caused "significant prejudice or harm to the person concerned". What does this mean in a context when the error is covert? Does it have to be shown that the error caused some distress or financial harm to the individual concerned? There are three problems:

(a) It is very unclear what the clause contemplates or what constitutes a serious error.

(b) Even if a person has been significantly prejudiced by an error this may not be apparent to the IPC. The IPC will not be able to ask a person how they have been affected by an error before they have disclosed the error.

(c) Such a condition would mean that many serious errors are not disclosed even if there is no or no significant harm to national security or crime detection at all (because for example a person is entirely unrelated to any such concerns and/or an operation has been closed), simply because no prejudice is known or can be shown. That is unlikely to comply with the European Convention on Human Rights.

 

87 The IPC should have a discretion to disclose errors and should be able to balance national security against the seriousness of the error in deciding whether to report the error to the person or organisation concerned. The JC will consider representations from the police or agencies and is perfectly capable, and it is perfectly proper, for him to make that determination himself.

 

88 Secondly, it needs to be made clear in statute what limits there are on the error reporting power: it is not appropriate for this to be done by Code. Sub-clause (9)(b) currently provides that the limits on the power are to be prescribed in the Code: it is wrong for the executive to determine what errors should be concealed and which should be disclosed.

 

89 Finally, the definition in sub-clause (9)(a) should be expanded to include errors by telecoms companies who are responsible for a large proportion of the errors that have occurred.

 

2. The IPT

 

71 The IP Bill includes an important reform to the IPT in clause 208 providing for a right of appeal to the Court of Appeal.

 

72 The introduction of a right of appeal will bring the IPT into the civil justice system and it will no longer be a mere complaints body for surveillance and intelligence services matters (which is essentially the current model). That is a good thing.

 

73 But once it is recognised that the IPT is an independent tribunal and part of the civil justice system, and not a mere complaints process, other changes to the legislative regime under which it operates are called for:

(a) At present the IPT’s rules are made by "the Secretary of State" (RIPA s.69(1) cf. First and Upper Tribunal Rules made by the Tribunal Procedure Committee and the Special Immigration Appeals Commission rules which are made by the Lord Chancellor). It is obviously inappropriate for the IPT to determine cases pursuant to rules made by one of the parties to the complaint. Although the IPT does have power to dis-apply rules that are contrary to the Human Rights Act 1998, (a) in determining this issue it gives considerable leeway and deference to the Secretary of State, and (b) being able to disapply a rule is not the same as writing the rules; the entire structure of the IPT’s procedure and approach to complaints is determined by the Secretary of State. The power to design the IPT’s rules remains an extremely important power over the tribunal which is exercised by the Secretary of State. This is especially true because the primary legislation (RIPA s.67) gives statutory force the rules in certain respects including in restricting the jurisdiction and powers of the IPT.

(b) This gives rise to the second point. At present the IPT cannot disclose any information or documents provided by the intelligence services or public bodies without that entity’s consent (RIPA s.67(7) Rules r. 6(1), (2)). It is a fundamental constitutional principle that the courts and tribunals determine whether material can be disclosed and are not dictated to by the Government (Duncan v Camel Laird, Conway v Rimer). [8] The Court of Appeal will inherit this inability to order disclosure with out the consent of the police authority or intelligence agency before it (Senior Courts Act s.15(3)).  It is constitutionally inappropriate for the Court of Appeal to be prevented from ordering disclosure without the consent of a party to it if it considers it necessary in the interests of justice. It is no answer to say that appeals will be restricted to points of law, because sometimes facts will be relevant to explain or determine such points, or disclosure will be a necessary consequence of the court’s determination.

74 In short, if the IPT is to be treated as a genuinely independent tribunal then Parliament should remove the power to write its procedural rules from the hands of the Secretary of State and it should give it the power to decide when documents should be disclosed in the public interest (which power would be subject to an appeal to the Court of Appeal).

 

 

DR TOM HICKMAN

Reader in Public Law, University College London and barrister at Blackstone Chambers

April 2016


[1] Neither that paper nor the present should be taken to contain a comprehensive list of every concern relating to the Bill held by the author.

[2] The author’s views on this point were summarized in the Joint Committee on the Draft IP Bill HL, 3 February 2016, Paper 93, HC 651 at §426.

[3] Operational Case For Bulk Powers at 8.7.

[4] See author’s evidence to the Joint Committee on the draft IP Bill at para 14.

[5] See Government Response to Joint Committee on the draft IP Bill ("Response") at Response 32. This would also allow for warrants in the two examples given in the draft Interception Code at page 30 para. 5.12. But it would require the intelligence services to specify each unidentified person to whom the warrant related.

[6] The extent to which independent authorization should be required when JC approval is not required is currently the subject of litigation (R (Davis & Watson) v SSHD) in which the author is involved and which this note does not address.

[6]

[7] Response, op cit, 62.

[8] It is no answer that in IPT proceedings, the Government will not always have an option to concede issues to avoid tribunal ordered disclosure in surveillance-related cases in the IPT, as (a) that does not meet the objection of principle, and (b) this option is not always open in other proceedings either legally or practically.

 

Prepared 14th April 2016