Single-source intelligence as the basis for assessments

SIS and GCHQ's place in the intelligence machinery is well-known and rarely questioned. These agencies traditionally produce single source secret intelligence reports in response to intelligence requirements which are set by Whitehall customer departments and prioritised by the JIC. The Cabinet Office's Assessments Staff combines SIS and GCHQ reporting with information, intelligence and analysis from other departments, partner countries and open sources. The preparation of all-source assessments are then finalised and approved by the JIC for issue to senior officials and Ministers.

Intelligence officers rarely claim that secret intelligence should dominate an assessment. As long ago as 1947, Allen Dulles, later to become director of the CIA, said: ‘A proper analysis of the intelligence available by these overt, normal and above board means would supply us with over 80 per cent, I should estimate, of the information, required for the guidance of our national policy.’Footnote ² Unarguably, today the contribution of open sources to find intelligence product is in the 90 per cent plus range. However, an efficiency argument is not the same as an effectiveness argument.Footnote ³ Not withstanding this, official reports have repeatedly argued that in practice the JIC tends to elevate intelligence reports over other sources of information. In 1983, the Franks Report recommended that the government should review ‘arrangements for bringing to the Joint Intelligence Organisation's attention information other than intelligence reports’.Footnote ⁴ Over a decade later in 2004 the Butler Report recommended that ‘assessments should not give undue weight to intelligence reports over wider analysis of historical, psychological or geopolitical factors’.Footnote ⁵

The process of basing assessments on a small number of single source reports is one that would be rejected if used as academic social science methodology. For a source's information to be processed as an intelligence report, the individual has to find himself in the social or professional orbit of an intelligence officer, has to establish a trusting personal relationship with that officer, and ultimately has to be willing to break his own institution's rules, at some risk to his own interests, in return for some material or psychological reward. It is often forgotten just how ad hoc this process is, and how dependent it is on the ability to develop a relationship between the officer and his source. Sources of human intelligence are in many cases at least partly self-selecting, either demonstrating in early contact with intelligence officers a willingness to ‘overstep the mark’ of what they should reveal, or in the most literal cases (known in intelligence circles as ‘walk-ins’) offering their services directly. It is not an exaggeration to say that key sources of intelligence, producing numerous reports over several years, became so because they ‘got on well’ socially with their early intelligence interlocutors; while other individuals, who may not have been so easily engaged at official functions or social gatherings, but who may have had much better information to offer, were overlooked.

Clearly the effect of this stereotypical serendipitous ‘selection by cocktail party’ is reduced by careful targeting and an expansion of the methods by which officers meet potential sources. However the element of luck and the emphasis on the personal can never be wholly removed; the process inevitably introduces a high degree of bias into the sample of opinions being reported.Footnote ⁶ Even when the source is honestly representing his own views and those of his sub-sources, by virtue of his accessibility, motivation to reveal privileged information and willingness take significant risks he is likely to be an unusual individual who is not necessarily representative of his institution as a whole. Arguably, intelligence failures over WMD in Iraq were because SIS were expediently talking to those people they could gain access to and who were willing to talk, rather than an unbiased sample of people who could make a sophisticated judgement on Saddam Hussein's WMD policy.Footnote ⁷ This is not to criticise SIS; merely to point out that secret intelligence by definition relies on recording information from exceptional individuals selected through an ad hoc and highly personal process, rather than conducting a careful analysis of the depth and breadth of information available. It is precisely for this reason that intelligence officers and official reports have themselves warned against relying on it too heavily.

Of course intelligence reports routinely separate ‘facts’, as reported by the source, from the source's opinion. But the type of high-level material that makes it into JIC reporting and is passed on to Ministers is rarely a quantitative fact such as the number of weapons and their locations, or a firm plan for an invasion. Instead it is often a judgement on a leader's intention and preferences, based on a discussion between the leader and the source or sub-source, or an observation made by the source. While reports can aim to tease out a leader's motives and any hidden agenda, the problem of a single, biased, perspective on the leader's state of mind remains. Reports of this nature are also heavily influenced by the nature of the relationship between the leader and the source or sub-source. Despite source descriptions and caveats, it is difficult for an intelligence customer to judge how much weight to give a report accredited to ‘a senior adviser to the President’, without understanding in much more detail the role of that adviser and his personal and political relationship with the President in question – information not available for source protection reasons.

This biased sampling matters less when the source is reporting objective facts. If the source passes on the technical blueprint of a new missile system, or the location or mobile telephone number of a terrorist suspect, then the selection of the source need not affect the validity of that specific report. However, even in this case a biased sample can have some impact, because such a report, if not balanced by other reporting, may give a skewed view of the target organisation's priorities, for example by emphasising its weapons development activities and preparations for attack (capability), but not its willingness to negotiate or its reasons for exaggerating its military power (intent). Arguably, the fact that the JIC set a Category One intelligence requirement on Iraq's weapons programme and only a Category Three requirement on Iraqi politics inadvertently guaranteed that the JIC would have a skewed view of the possible explanations of Iraq's apparent preparation of WMD.Footnote ⁸

These risks are well-known. Intelligence agencies do their best to guard against them by seeking a wide range of sources and providing separate, clearly-defined comments and context from the source, the source-handler and Head Office. But the implied criticisms from the Franks and Butler reports suggest that over-dependence on single-source reporting is a persistent concern. Authors such as Richard J. AldrichFootnote ⁹ and Andrew RathmellFootnote ¹⁰ argue that with the end of the Cold War, the growth of globalisation and the increasing importance of perceived risks from networks of and/or non-state, non-hierarchical actors, then the narrowness and sample bias of single source reporting may become even more problematic. It is, for example, difficult to see how single-source reporting could add much of strategic use to clarify issues such as climate change, energy security, global financial stability or food assurance.Footnote ¹¹ On these questions it is difficult for the JIC to compete with the mass of information available from market signals (such as the oil price), academia think tanks and numerous private sector organisations. Indeed, there is already collaboration to some degree.Footnote ¹²

Other countries, notably the US, already have mechanisms to generate assessments based on a richer base of source material than secret intelligence. The National Intelligence Estimates (NIE) draw in particular on extensive analysis of open source material from the recently created Open Source Centre (OSC). However access to open and other source intelligence did not prevent the CIA overruling evidence from a range of competent departments in assessments leading up the 2003 invasion of Iraq.Footnote ¹³ It may therefore not be enough to draw on a breadth of information; the structures which exert authority over the collation and delivery to policy-makers of that information will also influence the nature of assessments. One conclusion may be that not only should the information base of assessment be broadened, but also the power base of the structures through which it is compiled and delivered should be extended. The logical extension of this argument – involvement of non-governmental (for example, academic and commercial) bodies in the compilation of assessment – may be a step too far for Whitehall, but it may be key to the delivery of balanced assessment.Footnote ¹⁴

Do the assessments matter anyway?

This article supports long-standing warnings that intelligence assessments rely too much on single-source intelligence reporting and too little on wider historical, geopolitical, economic and psychological perspectives. However, there are also other persistent problems in the preparation of assessments which, despite being well-known, have never been fixed. In 1982, The Nicoll ReportFootnote ¹⁵ assessed the JIC's failure to predict the Soviet Invasion of Czechoslovakia in 1968, the Chinese invasion of Vietnam in 1979 and the Soviet invasion of Afghanistan in 1979. The report said ‘in several of these studies, it has appeared that the JIC made up its mind very early, and either did not change it at all (as with USSR/Czechoslovakia) or only later, in the face of very strong evidence of military preparations’.Footnote ¹⁶ The report added ‘there has been a tendency to assume that factors which would weigh heavily in the United Kingdom would be equally serious constraints on countries ruled by one-party governments and heavily under the influence of a single leader’.Footnote ¹⁷

These analytical failings can be summarised as confirmation bias – the use of evidence that supports an existing hypothesis and mirroring; assuming that the target organisation has the same influences and thought processes as one's own. These heuristics are well-known in the field of psychology and have been thoroughly explored in the literature on cognitive biases and analytical failings and are widely recognised in scholarly writings.Footnote ¹⁸ What is striking is that criticisms of these failings are repeated, often unconsciously but in very similar terms,Footnote ¹⁹ in Percy Cradock's study of the JIC,Footnote ²⁰ and in the Franks and Butler Reports. Yet they show little sign of being addressed by Government either now or over the past quarter of a century.Footnote ²¹

The 2008 IISS Strategic Survey states: ‘The problem [of Iraqi WMD intelligence failures] was not so much one of intelligence analysis as of the inability of the UK's analytical community to put themselves into the minds of those whose behaviour they were analysing’.Footnote ²² In his book Know Your Enemy examining the JIC's performance between the Second World War and 1968, former JIC Chairman Percy Cradock highlighted this as a persistent shortcoming in the committee's otherwise creditable performance.Footnote ²³ In the same vein, no effort was made by the JIC or policy makers to think radically about the pre-invasion failure of the UN inspectors to find militarily significant quantities of WMD. At no point did anyone consider the possibility that, to contradict former US Defense Secretary Donald Rumsfeld, absence of evidence might in fact indicate evidence of absence. Following Butler's recommendation, some progress has been made towards establishing a UK analytical community. A Professional Head of Intelligence Analysis (head of profession) has been appointed in the Cabinet Office to recommend on professional training, standards and career structure. However, the position commands no resources, has no line-management responsibilities and can only advise and seek to exercise influence. An additional move has been to organise courses in intelligence analysis at King's College, London, to provide consistent methodology and standards. In spite of these steps, there is so far little evidence within the Assessments Staff that a permanent cadre of analysts is becoming established.Footnote ²⁴

One might reasonably conclude that a problem has been identified for twenty-five years by influential Whitehall insiders and detailed in official reports, but never been solved even after the politically-charged WMD intelligence failures, could it be that the problem simply isn't that important to Government? The memoirs of senior politicians rarely mention a JIC assessment as having been influential in their decision making. This may simply be to protect the secrecy of JIC material rather than a fair reflection of the facts. However, private conversations with some Cabinet Office officials reflect scepticism as to whether JIC papers are regularly read by senior Ministers. Indeed, it is argued that Ministers prefer the drama and clarity of a single-source report to the careful nuances of a JIC assessment,Footnote ²⁵ despite the risks inherent in over-reliance on such single-source reporting. Based on this anecdotal evidence, it is possible that when we assess the role of the JIC we overestimate the extent to which decisions are based on its assessments, and underestimate the extent to which Ministers are influenced more by their own political agendas, personal contacts and personal assessments of single-source reporting? It would be a naïve politician who gave more weight to a JIC assessment than to a barrage of tabloid headlines; the assumption of the high significance of that assessment in decision-making may be equally naïve.

SIS as global policeman

The use of single-source reporting to inform high-level policy assessments is increasingly being questioned and experiences competition from academia, the media and non-Governmental organisations. However, the intelligence community has another, more tactical set of roles which are unlikely to be challenged by overt or non-state organisations: the ability to collect and analyse intelligence on terrorist networks from their own sources; incorporation of the same from foreign liaison services, the facilitation of disruptive action by UK and foreign police and security services.

The type of intelligence that is useful in disrupting terrorist networks is often low-level, factual and fragmented.Footnote ²⁶ A typical investigation may encompass a foreign telephone number found in a diary by UK police during a forensic search, a subscriber check conducted by a foreign liaison service, a travel record from a second foreign liaison service, a credit card transaction from a third foreign liaison service, an entry record from the UK immigration service and the observations of a surveillance operation by the UK Security Service. The key to making sense of this is high-level knowledge management. SIS and GCHQ analysts increasingly spend their time preparing i2Footnote ²⁷ charts of terrorist and criminal networks operating abroad, in co-operation with their Security Service counterparts who collate the UK picture. These network analysis charts identify telecommunication or direct connections between suspect individuals, and record a network's use of specific telephone numbers, credit cards, bank accounts and methods of transport. This intelligence is shared between UK and foreign security agencies and used to support law enforcement operations, but is rarely significant enough in itself to be issued to Whitehall policy departments or incorporated into an intelligence assessment. Like a police investigation, it serves to inform operational steps taken towards an executive action – arrest, interdiction, disruption – and only emerges in the non-intelligence official sphere if submitted as evidence in court.

While SIS's independent human sources can certainly contribute to this type of intelligence picture, SIS's main contribution is access to foreign Governments' information, via their liaison relationships with foreign intelligence and security services. Typically the information being shared is not particularly sensitive and is often simply taken from official records: purchased travel tickets, electoral role records, family connections and similar data. SIS's second contribution is to collate this information from a wide range of governments and agencies, and use it persuade foreign security or police agencies to take action against the individuals and networks concerned that are considered a threat to the UK or its interests. The main skill here is not the traditional one of running covert agents, but the quasi-diplomatic one of managing relationships with foreign services and making a case for joint action based on common interests.

SIS as global capacity builder

SIS's ability to tackle global terrorist networks and insurgencies therefore depends on building competent and co-operative liaison services in countries that are important recruiting grounds or operating bases for terrorists. While SIS has for many years provided training to liaison services as part of the quid pro quo of a bilateral relationship, this effort is now being stepped up, more emphasis is placed upon achieving sustainable increases in capability rather than simply seeking to generate goodwill. This effort has required SIS officers, in addition to their traditional skills of managing one-to-one relationships, to become experts in project management and capacity building.Footnote ²⁸

However the work that SIS does to strengthen a liaison service's capabilities should only be part of the story. There is a risk that key partner countries will end up with highly-effective security and intelligence services, but absent of comparably strong oversight and accountability mechanisms.Footnote ²⁹ This has a direct effect on SIS's ability to co-operate with their partner services; British courts will not allow SIS to pass intelligence leads to services unless they can give credible guarantees that sharing intelligence will not lead to human rights abuses, and evidence gained from services suspected of using torture will not be acceptable by the UK agencies or admissible in a UK court.

More importantly, inadvertently creating a powerful security sector without effective oversight risks undermining the UK government's wider interests in democratisation and good governance. This is ultimately counter-productive.Footnote ³⁰ Such a realisation is beginning to lead to a convergence of views between Security Sector Reform (SSR) theorists – who see SSR as a democratising and developmental activity – and counter-terrorist/counter-insurgency theoristsFootnote ³¹ who have begun to argue that it is necessary to out-govern not out-fight the enemy. However, this article argues below, the convergence in theory has not yet led to a convergence in institutional arrangements.

A second important gap resides in the notion that effective counter-terrorism and counter-insurgency in the UK's partner countries depends upon far more than the existence of an effective state-level security or intelligence agency. RAND, in a report for the US Department of Defense, argues: ‘Of indigenous security services, none is more critical than police. By maintaining neighborhood safety and enforcing the rule of law with minimal necessary force, police can be both more effective and more legitimate than combat troops, especially if combined with fair, efficient, and transparent justice and penal systems.’ Furthermore, RAND continues: ‘Even if they are not enamored with the state itself, Muslims who are concerned about family security, as most surely are, may accept and even co-operate with well-trained, well-led, well-behaved, and even-handed police.’Footnote ³²

The report also gives some rules of counter insurgency (COIN):Footnote ³³

• ‘There is no substitute for legitimate, effective local government.

• Ordinary people, the prize of insurgency and COIN, respond mainly to everyday public safety and service, or lack thereof.

• Securing the population, as opposed to “killing the enemy”, is the principal role of military operations.

• Superior firepower can fail if seen as illegitimate, and injudicious force can fan insurgency and popular support for it.

• All instruments and measures of COIN – political, economic, intelligence, police, and military – must be integrated into a coherent campaign strategy.

• Foreign military intervention cannot save corrupt or incompetent local regimes and can trigger patriotic resistance.

• Isolating insurgents from the population is more efficacious than killing them.’

SIS is effective at training other countries' security and intelligence agencies in agent running and intelligence analysis. It does not, and does not claim to, have the skills to train foreign police services in community-based policing, foreign armies in proportionate and legitimate use of force, or local authorities in the responsive delivery of services. But these are all essential tools in counter-terrorism and counter-insurgency. In particular, the day-to-day behaviour of police is crucial. An ineffective and tainted police force can foster radicalisation,Footnote ³⁴ whereas an effective police force, enjoying a positive and mutually-trusting relationship with the public, can spot radicalisation at local level and work to nip it in the bud long before a security service operating at state level becomes aware of the problem. Where military forces are operating as quasi police forces the danger of heavy-handed counter-productive action is even greater.Footnote ³⁵ Furthermore, SIS claims no skill in helping countries with de-radicalisation programmes, such as those that appear to be becoming successful in Saudi ArabiaFootnote ³⁶ and Indonesia.Footnote ³⁷

The British Government has a number of structural problems which hamper efforts to fill these gaps. In particular, the Constabulary structure of British policing means that the government finds it difficult to deploy serving police officers to advise overseas forces. Chief Constables have sole control over the officers in their Constabularies and have little incentive to send their best people overseas, given that their performance is judged, using Home Office police performance indicators, solely on the delivery of security and justice in their own Constabularies.

The government has developed some wider expertise in helping to develop accountable and effective security agencies, under the general heading of security sector reform.Footnote ³⁸ However this has primarily been seen as a development activity, led by the Department for International Development (DFID). Under the Development Act,Footnote ³⁹ DFID's sole mandate is to reduce poverty in developing countries, an objective enshrined in legislation as the sole legitimate use of development funds – DFID has no mandate to pursue the UK national interest or to support the counter-terrorist activities of other government departments. This creates a ‘governance gap’ in many of our key counter-terrorist partner countries. Because many of them are middle or high income countries they do not (and cannot) benefit from DFID programmes on good governance and reform. Put simply, poor countries receive assistance on good governance; but most high-priority CT countries are not poor.

The variance between departmental mandates and agendas creates a difficult balancing act for DFID. They are willing to be helpful to the Foreign and Commonwealth Office (FCO) and other departments by offering advice, and they argue that their development activities will ultimately reduce radicalisation as a side outcome to the primary purpose of poverty reduction. However, by law they cannot allocate development resources to these activities if UK national security is the primary aim. The FCO has begun to take an interest in filling this governance gap, using security sector reform techniques to take forward the PREVENT strand of the Government's counter-terrorist strategy.Footnote ⁴⁰ This may involve making more use of existing cross-departmental resources such as the Security Sector Development Advisory Team (SSDAT), which specialises in security sector reform and which to date has been utilised more by DFID and the MOD than by the FCO.Footnote ⁴¹ But such work is in its infancy.

The result of these structural problems and differences in mandates can sometimes lead to an incoherent set of activities in key partner countries. SIS can be found pursuing capacity building with its partner agencies, the FCO beginning to sponsor security sector reform with a counter-radicalisation purpose, DFID absent or pursuing its development mandate, the Ministry of Defence, via their Defence Attachés, pursuing defence diplomacy and offering military training and equipment; and the Home Office and police, who could offer policing advice, largely absent. The Government created a Conflict Prevention PoolFootnote ⁴² and a Stabilisation Aid Fund to deal with some of these problems by creating single funds drawn from MOD, FCO and DFID resources, but in practice this is managed by relatively junior officials and does not have the authority to direct departments to co-operate when this goes against departmental interests.Footnote ⁴³ Work paid for through these funding mechanisms has so far taken little account of the UK government's Counter-Terrorism objectives; furthermore the UK intelligence machinery is not involved in deciding how these funds are spent.

While there are some examples of good cross-departmental co-operation in Embassies, this is mostly on an ad hoc basis; dependent on the characters of the individuals involved. Few Embassies have a single, agreed, cross-departmental strategy for activities in their host country; the UK government had no single Iraq Country Strategy until 2007, four years after the invasion and expenditure of tens of millions of pounds on security, development and other interventions. There is no single list of priority countries where HMG wishes security sector reform and other good governance activities to take place, in the interests of both development and counter-terrorism.

It is not the job of SIS to provide this co-ordination and prioritisation, but it would be strongly in its interests for it to happen. At present there is no natural home for senior executive-level co-ordination across departmental activities, and one needs to be either created or developed from existing bodies. The Joint Intelligence Committee could be a natural starting point. It includes the key Whitehall policy departments and (unlike the conflict prevention pool) the intelligence and security agencies. Its intelligence requirements already carry significant authority. Could it radically change into a Joint Action Committee, setting action requirements for key terrorist countries and providing high-level, authoritative strategy? To do so would require a fundamental rethink of the purpose and mandate of the JIC; but, in a world of perceived broadening security risks, competing departmental agendas and the pressing need to synergise efforts across government, such a rethink may be overdue.