Intelligence oversight in the UK

The oversight structure for the security and intelligence agencies in the UK has been constructed piecemeal since the first footings were dug when the system by which ministers authorised the interception of communications was found to be in breach of the European Convention on Human Rights (ECHR).Footnote ³ The subsequent Interception of Communications Act 1985 provided a statutory basis for the procedure, established a judicial commissioner to review the procedural legality of warrant-issue and a Tribunal to receive and investigate complaints from the public. As a result of later legislation providing statutory mandates for the Security Service (MI5) in 1989 and for the Secret Intelligence Service (SIS or MI6) and Government Communications Headquarters (GCHQ) in 1994, similar commissioners and tribunals were set up. Matters were rationalised to some extent by the Regulation of Investigatory Powers Act 2000 (RIPA). Now, there is an Interception of Communications Commissioner (ICC) appointed under RIPA s.57 who reviews the warrant issuing process, ministers' and agencies' performance in acquiring and disclosing communications data and the process by which agencies other than intelligence and military gain access to encrypted data.Footnote ⁴ RIPA s.59 provides for a single Intelligence Services Commissioner (ISCommr.) who reviews procedures regarding warrants for the ‘interference with property’, electronic surveillance, ‘covert human intelligence sources’ (CHIS) and the requirement for the disclosure of encrypted data. The ISCommr. also reviews the use of these powers by the military outside Northern Ireland.Footnote ⁵ The Commissioners' annual reports are very brief and amount to little more than a rehearsal of their powers and some information on the numbers of ‘errors’ made by intelligence, police and other agencies. Most recently the ISCommr. reviewed the allegation that GCHQ had intercepted mobile telephone calls between those responsible for the Omagh bombing in August 1998 and disseminated those to police in Northern Ireland prior to the bombing.Footnote ⁶ Sir Peter Gibson's report (at least, the published version) found that GCHQ would not have been able to do this in 1998 as described in the programme but did leave unanswered other questions about the quality of intelligence sharing prior to the attack.Footnote ⁷

RIPA ss 65–70 rationalised the previous arrangements for complaints tribunals into a single Investigatory Powers Tribunal (IPT) which is empowered to receive and investigate complaints against the intelligence agencies. It may hold hearings in public or private and take evidence that would not be admissible in court. If a complaint is upheld, the IPT has the power to award compensation, quash or cancel warrants and require the destruction of information wrongly obtained. In general, the IPT may not disclose any information without the consent of the originator.Footnote ⁸ The IPT does not publish annual reports or statistics relating to the outcome of its investigations but those of both the ISCommr. and ICC include almost identically uninformative paragraphs in relation to the IPT. Since 2000 the IPT has received over 600 complaints and, for the first time, upheld one in 2005–06.Footnote ⁹ The shock of this seems to have been too much for the reporting system since in the two subsequent reports, the ISCommr. did not say whether or not any were upheld.

The most public part of the architecture of oversight, however, has been the Intelligence and Security Committee (ISC), established by the Intelligence Services Act 1994. The reasons for the creation of the ISC include the end of the Cold War, greater government openness after Thatcher, continuing controversies about intelligence such as Northern Ireland and arms sales to Iraq and a realisation among officials that to institute a manageable reform at a time of their choosing was preferable to having it forced upon them should there be a future scandal.Footnote ¹⁰ Structurally, the most significant point is that ISC is a ‘committee of parliamentarians’, appointed by and reporting to the Prime Minister, not a select committee whose first point of reference would be members of Parliament.

Studies of ISC's first ten years came to the overall conclusion that the Committee has achieved more than might have been expected. Formally, its mandate is the same as for a select committee: to review the ‘expenditure, administration and policy’ of the intelligence agencies and it has succeeded in establishing the degree of trust with the agencies that is necessary for any meaningful oversight.Footnote ¹¹ It developed oversight from scratch, examined some operational issues, despite these not being formally part of its remit, introduced significant financial accountability,Footnote ¹² probed complex and technologically challenging areas of intelligence, produced interesting and informative reports that added to public knowledge, despite redactions,Footnote ¹³ and has established itself as a serious critic of the agencies.Footnote ¹⁴ On the negative side, Mark Phythian notes the ISC's failure to investigate various allegations relating to Libya, Sierra Leone and Northern Ireland and to accept too readily proposed redactions to its reports.Footnote ¹⁵ Similarly, it is suggested that the ISC has failed to generate public or parliamentary trust in the agencies and has generally granted the agencies a cleaner bill of health than the facts warranted.Footnote ¹⁶

These commentators' main criticism was of the Committee's poor performance in the controversy about Iraqi weapons of mass destruction (WMD). The Committee lacked ‘appetite’Footnote ¹⁷ to take on the politicisation of intelligence and ‘stumbled badly’ with a narrowly-focused inquiry compared with that conducted a year later by Robin Butler.Footnote ¹⁸ Glees et al and I concluded that the credibility of the ISC requires that this issue be re-visited, especially in the light of the emergence of the ‘Downing Street memos’ indicating Tony Blair's political commitment to the Iraqi invasion early in 2002 and the extent to which intelligence was abused by the Government in drumming up support.Footnote ¹⁹ Clearly this will not happen in the light of Brown's announcement in June 2009 of a privy council inquiry into the lead up to and conduct of the Iraq invasion, but this weakness was symptomatic of a deeper structural issue as to the Committee's role and to whom it thinks it is speaking. The ISC addresses itself to Whitehall, not the public.

A review of the history of intelligence oversight shows that concern with propriety is uppermost after scandals relating to intelligence abuse and with effectiveness after intelligence ‘failures’, but these swings of the pendulum show that the best accountability systems are those which seek both effectiveness and propriety all the time. The ISC, however, was designed and has performed primarily with effectiveness in mind; it is more of a tool in the Whitehall armoury for the management of the intelligence community than of Parliament in maintaining democratic accountability in the interests of propriety as well as effectiveness. Indeed, two early examinations of the new ISC made the point that its function should be seen more as the executive seeking to maintain control over a controversial area of governance in the context of the changing risk environment after the end of the Cold War.Footnote ²⁰ David Omand, at different times Director of GCHQ and Intelligence and Security Co-ordinator, saw the role of the ISC as ‘to oversee the efficiency and administration of the secret services’ and, noting how the Committee had extended its remit, explained how the government had used this, for example, to investigate the 2002 Bali bombing, and how the agencies also found it useful to have an authoritative group of parliamentarians who could counter the misinformation intelligence attracts in the media.Footnote ²¹ In one sense this is no surprise: the very principle at the heart of the UK parliamentary system of government is the organisation and control of parliament by the executive but, given the historical evidence of the harm that can be done to civil liberties and human rights by intelligence agencies, this is unsatisfactory.

Off-balance? The ISC and human rights

Reflecting this ‘managerialist’ emphasis, there was no substantial reference to human rights in any ISC Report of its first ten years. Then, in 2005 the ISC reported on the intelligence agencies' dealings with detainees in Afghanistan, Guantánamo and Iraq.Footnote ²² It noted that intelligence officers had expressed their concern at the US treatment of detainees in a small proportion (fewer than 15 out of 2000) interviews witnessed, criticised the lack of training of personnel in the implications of the ECHR before they were deployed and criticised the lack of follow up when concerns were expressed to US authorities. Yet ISC showed keen sensibility to the agencies' dependence on the US for intelligence and did not tackle the ‘difficult questions’ involved in the use of information possibly obtained under torture, satisfying itself by quoting at length the Foreign Secretary's utilitarian justification for this.Footnote ²³ This was an abdication of maintaining accountability – the question of abuse of prisoners by soldiers was, per se, one for the Defence Select Committee but, judging by its report, the ISC did not even inquire as to whether this was at the behest of intelligence personnel by way of preparing detainees for interrogation, as it was at Abu Ghraib.Footnote ²⁴ It was subsequently revealed in a court martial in 2007 that such techniques had been used also by some British soldiers, as revealed originally in a Red Cross report in 2004.Footnote ²⁵

The ISC's next foray into the walk on the dark side that was the ‘war on terror’, was to examine ‘extraordinary rendition’ (ER).Footnote ²⁶ In November 2002 Bisher al-Rawi, an Iraqi who had lived in UK since 1984, and Jamil el-Banna, a Jordanian-Palestinian with refugee status in the UK, were arrested at Gatwick after ‘suspicious items’ were found in their luggage. Both men were known to the Security Service as ‘close associates’ of Abu Qatada, the radical cleric. After interviews, the men were released and eventually resumed their journey to The Gambia where they were arrested on arrival. Shortly, they were handed to US officials and ‘rendered to detention’, first in Afghanistan and then in Guantánamo in February 2003. Now, several times around their departure from the UK, ‘U.S. authorities’ – presumably the CIA – were sent information relating to the arrest and release of the men including caveats that the information was not to be acted upon.

Overall, the ISC report asks us to accept that the UK agencies had only slowly realised the implications of the ER policy, at least when it moved beyond the removal of fighters from Afghanistan. ISC made no substantial criticism of agencies' behaviour. However, the time and resources spent on this inquiry were minimal and incapable of revealing impropriety on the part of the agencies had they sought to conceal it; certainly compared with the judicial inquiry conducted by Justice Dennis O'Connor into the Maher Arar case in Canada.Footnote ²⁷ However, by the traditional standards of British reluctance to criticise the US on intelligence matters, the Report's conclusion as to the impact of the CIA ignoring the caveats is stark:

Another of the cases referred to in ISC's rendition report is that of Binyam Mohamed al-Habashi, an Ethiopian national who had sought asylum in the UK in 1994. He travelled from UK to Pakistan in June 2001 and was arrested at Karachi airport in April 2002 for travelling on a false passport from Afghanistan where he had allegedly been fighting with the Taliban. Binyam Mohamed was held in Pakistan for three months, during which time he was interrogated by MI5 officers, though they deny telling him that he would be tortured, as he alleges. The ISC reported that the interview was conducted within the Service's guidance.Footnote ²⁹ In July Binyam Mohamed was subjected to ER to Morocco where he remained for eighteen months before being rendered to Kabul and then Guantánamo where he was held until returning to the UK in February 2009. He alleges that he was subject to appalling torture in Morocco, including beatings and being cut with a scalpel all over his body but the ISC refrained from giving details.Footnote ³⁰ The ISC accepted that there is a ‘reasonable probability’ that intelligence passed by the Security Service was used in his subsequent interrogation and agreed

Given what they said about the caveats in the earlier case, ISC might have wondered, given the state of knowledge of ER by July 2007, whether any assurances would have been worth the paper they were written on.

The ISC Report referred to its meetings with ‘other officials’ from MI5 (in addition to those named) thus there is no way of knowing whether these were actually those concerned in the interrogation of Binyam Mohamed. These officers did eventually appear in the High Court at the hearing of his case to obtain documents provided by the UK to the US. In the course of its judgment, the court was far more critical than the ISC. It noted that MI5 had facilitated interviews by or on behalf of the US when he was being detained unlawfully in Pakistan and when it must have known he was being detained by ‘a foreign government’ (presumed to refer to Morocco) and that this co-operation with the US ‘was far beyond that of a bystander or witness to the alleged wrongdoing.’Footnote ³²

This case and other developments forced the ISC to re-visit the issue – since publishing its Rendition report, further legal cases were pending in which entirely credible allegations of torture at the hands of the Pakistani Inter-Service Intelligence (ISI) and of UK collusion had been made.Footnote ³³ Ministers' continued repetition of the mantra that UK agencies ‘do not participate in, solicit, encourage or condone the use of torture […]’ was undermined further in March 2009 when the Attorney General announced she was calling in the Metropolitan Police to investigate possible criminality by intelligence officers involved in Binyam Mohamed's detention and torture.Footnote ³⁴ The ISC's re-investigation prompted the ‘rediscovery’ of thirteen MI5 and MI6 documents that should have been disclosed to the High Court, leading to a government apology, and ISC made recommendations to the Prime Minister as to how the guidance to intelligence officers should be amended.Footnote ³⁵ Brown told the Commons that new guidance would be published once it had been reviewed by ISC but the Foreign Secretary told the Foreign Affairs Committee in June 2009 that it was unlikely the guidance promulgated in 2004 would be published.Footnote ³⁶ Whatever the particular outcomes of the police investigation, several civil actions, and continuing inquiries by parliamentary committees for foreign affairs and human rights, they are likely to expose the continuing inadequacies of ISC modus operandi.

Government proposals for ISC reform

In 2008 the Brown Government made a number of proposals for marginal changes to the ISC: primarily, the PM would consult with the Leader of the Opposition on appointments from a list provided as for select committees. The Government has not acceded to the suggestion that the ISC Chair be appointed from the Opposition,Footnote ³⁷ as is the model in the Public Accounts Committee. This would increase the legitimacy of the Committee since its reputation has not been enhanced by a succession of ex-ministers being temporarily ‘parked’ there.Footnote ³⁸ The Committee might meet in public but only to receive briefings from agency heads or ministers because of possible security risks to lower-ranking staff. No change was envisaged to the reporting mechanism but there would be debates in both Commons, to be led by the ISC Chair rather than a minister, and Lords, led by the senior Lords Committee member.Footnote ³⁹ These were agreed after discussion incorporated with the July 2008 debate on the ISC Report for 2006–07Footnote ⁴⁰ but will make little difference.

The issue of parliamentary ownership remains the central one; on this Glees et al and Phythian disagree: the former arguing that the present status is the ‘least bad’ while the latter argues the necessity of making ISC a select committee.Footnote ⁴¹ In its present form, the ISC has succeeded in building trust so that the agencies feel free to make documents available and discuss issues without fear of material being leaked or published in such a way as to damage security. The argument against parliamentary ownership, in which the committee would make the final decision as to what to publish, is that the agencies would be far more reticent and the sum of public knowledge would be less than it is now. A select committee has powers not available to the ISC to demand ‘people and papers’ but implementing this regularly against intelligence agencies is simply not feasible in the UK system. But as long as the ISC can appear as the ‘PM's poodle’, as it did over Iraqi WMD, it will lack legitimacy. There is no reason why a parliamentary committee could not develop in time the same degree of trust over information handling as has the ISC; this is an issue where the agencies will just have to bite the bullet. For example, there is nothing to stop the ISC, as they do now, from interacting with agencies over what should be redacted from reports in what has been described as ‘quite a ping pong’Footnote ⁴² before sending reports directly to parliament.

However, the government has accepted that the ISC is under-staffed. The six current staff are acknowledged to be over-worked in terms of preparing for evidence sessions and drafting reports; they do not conduct any research.Footnote ⁴³ The government intends to reinstate the position of investigator with a potential pool of people for specific projects, which would be an improvement, but this would still mean that the Committee is dependent routinely on what they are told by agency heads and other officials. In its first report, the ISC said: ‘we hope and expect to be kept informed of information that is relevant to our remit’Footnote ⁴⁴ but some years later, Tom King, the first ISC Chair, observed more realistically ‘[…] if you don't ask for it, you don't get it.’Footnote ⁴⁵ The lack of researchers with clearance adequate to access files and personnel leaves the ISC with no way of checking on the completeness or otherwise of what they are told, as has been recently acknowledged to be the case with their first report on the July 2005 London attacks when ISC just ‘listened to’ what they were told compared with the second report when members studied the relevant surveillance tapes, videos and notes.Footnote ⁴⁶

The issue of ISC access to information is partly one of formal powers: a key difference between the commissioners and the IPT is that members of the agencies have a legal duty to provide them with information on request whereas the agency heads and, ultimately, the minister act as gatekeepers for what the ISC can receive.Footnote ⁴⁷ Governments have not routinely denied ISC access: there is just one case where the Government have refused access to a document for some (redacted) number of years.Footnote ⁴⁸ But there are also issues in relation to committee resources and tenacity: members clearly feel that the agencies are straightforward with the ISC and that, if they are misled, it is more because the government itself is unaware of relevant information or, as in the case of rendition, has been misled by another Government.Footnote ⁴⁹ However, members need to remain highly aware of their vulnerability to being misled – further inquiries into the agencies' knowledge of and/or collusion with torture indicate thisFootnote ⁵⁰ and they do not need to look any further than the difficulties faced over the years by inquiries into police and intelligence-related matters in Northern Ireland to see this danger writ large.Footnote ⁵¹

Intelligence networks: yet greater challenges for intelligence oversight

But even if the ISC were to receive significantly greater resources and autonomous powers, it would still face the problem of how to oversee networks. The general ‘democratisation’ of intelligence since 1970s, of which the UK has been a part, has been predicated on the accountability of state intelligence agencies and has relied on statutory mandates, written guidelines, internal authority relations, audit trails for information and expenditure and reporting requirements. States remain significant in global intelligence but are far less able to act unilaterally than they were and, in some places, are less important than non-state actors. Much intelligence activity takes place between organisations that constitute ‘security networks’Footnote ⁵² These are mechanisms for the transfer of information and other resources involved in governance characterised by interdependence, flexibility and complementarityFootnote ⁵³ and raise complex issues of both direction/control and oversight/accountability. The former includes processes of activation, facilitation, brokering and mediationFootnote ⁵⁴ but here we concentrate on oversight.

Intelligence networks exist spatially at various ‘levels’ – local, regional, national and transnational – and within and between different sectors. Beyond the state the two main ones are the corporate, which is relatively well defined in terms of either ‘in-house’ security or contracted security companies, and a far more nebulous ‘community’ sector.Footnote ⁵⁵ Here, ‘actors’ vary widely in how formally they are organised and extent to which their activities are actually legal. They range from ‘gated communities’ and criminal organisations to separatist or national liberation movements.Footnote ⁵⁶ For example, we might note the significance of ‘Awakening’ groups in reducing levels of violence in Anbar Province, Iraq during 2007–08.Footnote ⁵⁷

For reasons of space, discussion here concentrates on national and transnational state networks. Oversight institutions are set up within national systems and may only have access to one (or a few) member(s) of state networks which leaves some very large gaps in intelligence oversight. British pragmatism and a wish to establish review that would just meet the minimalist standards of the ECHR has left the UK with an unsatisfactory ‘compartmentalised’ system of review that fails to close the gaps. For example, the ISC tried for years to obtain access to the confidential annexes of the Interception Commissioner's reports but the government refused them. Similarly, there appears to be no contact between the IPT and the ISC whereas, elsewhere, review committees have noted how they have learnt from the process of investigating complaints. It would not be appropriate to have ISC parliamentarians dealing with individual complaints but there are opportunities for mutual education. At a time when governments seek greater integration between national security operations, it is crucial that review mechanisms develop similarlyFootnote ⁵⁸ lest the democratic deficit grows ever wider.

It is worth consulting recent Canadian experience in this respect. Justice O'Connor's enquiry into the rendition of Maher Arar by the US to Syria in 2002 where he was detained and subject to torture over a period of 12 months found that RCMP officers passed a whole investigative file over to the US without the information being screened for relevance, reliability or personal information and included information that was ‘inaccurate and imprecise’, overstating any involvement Arar might have had in terrorist activity.Footnote ⁵⁹ The Commission was also charged with recommending how review of the RCMP's activities might be better conducted. After considering Canadian and foreign experience, it proposed a structure that would have the separate agency review bodies co-operating in cases where the agencies themselves co-operated operationally and shared information. The review bodies would deploy ‘statutory pathways’ that would enable them to exchange information as necessary to carry out their duties and an Integrated National Security Review Coordinating Committee, comprising of the chairs of the individual review bodies and an independent chair, would have the mandate to ensure that the gateways worked, to avoid duplicative reviews, to provide a single mechanism for the receipt of public complaints and to report on the impact of national security practices on rights and freedoms.Footnote ⁶⁰

Although Parliament is significantly absent from O'Connor's proposal since its place within intelligence accountability was beyond his mandate, and acknowledging that no country can simply ‘import’ reforms from another, the principles developed by O'Connor require serious attention. ‘Statutory pathways’ already exist in the UK, for example, between the Independent Police Complaints Commission and the parliamentary ombudsman. The ISC has grasped this nettle in some respects by extending its review beyond the three central agencies to include the Defence Intelligence Staff (DIS) and Serious Organised Crimes Agency (SOCA) and, more recently, has had informal discussions with the commissioners but too many gaps remain.Footnote ⁶¹

Transnational intelligence networking has received much more scholarly attention in the wake of the Rendition scandal.Footnote ⁶² Other than O'Connor's Commission, the most extensive inquiry has been that on behalf of the Council of Europe by Senator Dick Marty whose two reports provide a graphic account of the rendition process.Footnote ⁶³ But the problem is much clearer than the solution, for example, national reviewers can establish minimum legal standards for information exchange and ask that intelligence-sharing agreements be made by governments in writing and be subject to inspection but, as we know, the major strength of intelligence networks from the point of view of practitioners is their very informality. Exchange of information between people who know and trust each other is the lifeblood of transnational intelligence co-operation. Under pressure, formal policies may simply be disregarded, as Maher Arar and others discovered to their cost.

So, how on earth are individual and organisational actors working within these networks to be overseen so that they are effective and behave properly (not just legally)? In the same way that intelligence networks operate in three sectors, so do potential reviewers. Within the state sector, the Canadian Commission has shown one way in which non-parliamentary external reviewers can be linked; this can be built on to incorporate or connect with internal or ministerial processes such as inspectors general, legislators in review committees and judges, especially those involved in special courts or as commissioners. Equally, in the UK, for example, the fact that different parliamentary committees may touch on intelligence matters, as they have regarding Iraq and rendition, must not inhibit the ISC from taking a holistic view of intelligence. These national officials must also lift their eyes above the horizon and respond to both national and transnational constituencies.Footnote ⁶⁴ An example of just this kind of thinking can be seen in the recent initiative of the Belgian Standing Intelligence Agencies Review Committee: ‘With increased globalisation in mind, the Standing Committee I wishes to meet the expectations of a broader public.’Footnote ⁶⁵ Accordingly, it has translated its Activity Reports for 2006 and 2007 into English. Members of national review agencies meet at the biennial International Intelligence Review Agencies Conferences.Footnote ⁶⁶ These are (more or less) closed – they may well be of value for participants but they should consider having an additional day to which NGOs, researchers, media etc would be invited for exchanges of ideas and to establish a review network.Footnote ⁶⁷

With no researchers of their own, the ISC must maximise the use of findings by academics, scholars, journalists, NGOs and activists who can produce highly relevant knowledge even in the unpromising context of highly secret intelligence collaboration. For example, significant in mapping rendition was the work of Amnesty International,Footnote ⁶⁸ investigative journalist Stephen Grey,Footnote ⁶⁹ the Council of EuropeFootnote ⁷⁰ and the All-Parliamentary Group on Extraordinary Rendition.Footnote ⁷¹ A similar example can be found in the exposure of the US operated ECHELON system for the interception of global communications involving journalists, activists and eventually the European ParliamentFootnote ⁷² and Aldrich notes that the press got further in uncovering the Niger ‘yellow cake’ saga than official inquiries.Footnote ⁷³

Thorsten Benner et al identify a number of mechanisms with potential for network accountabilityFootnote ⁷⁴ and we should consider to what extent, if any, these can be applied to intelligence.Footnote ⁷⁵ They include professional/peer evaluation, for example, via codes of conduct. Of course, professionalism is also important in establishing networks in the first place, as in the case of transnational policing.Footnote ⁷⁶ Second, they identify ‘public reputational’, for example, naming and shaming (but they note this will not work with rogue actors of which intelligence has a few!) Also, even legitimate intelligence actors will be very hesitant to denounce others because it may involve breaching secrecy and will threaten the possibility of future co-operation – ‘you never know today on whom you may depend in future […]’ Third, the need to account for funds in contractual relationships has some potential but is fraught with difficulty since even legal intelligence organisations use ‘grey’ financial channels as part of covert operations. Finally, there is the law, although Benner et al note this is relatively little used in networks. This was especially true of intelligence that operated out with law but the role of law is more significant than it used to be. For example, establishing legal mandates, standards and controls has been a major aspect of the widespread democratisation of intelligence in former authoritarian regimesFootnote ⁷⁷ and Aldrich shows how law has had more impact since 9/11 than ‘toothless’ political mechanisms.Footnote ⁷⁸ Yet questions remain as to the effectiveness of law over the generality of intelligence co-operation.