On the History of the Arm’s Length Between Intelligence and Policymaking

Source: Marrin, S., & Davies, P. H. J. (2009). National Assessment by the National Security Council Staff 1968–80: An American Experiment in a British Style of Analysis? Intelligence and National Security, 24(5), 644–652.

XXXXXX

The relative proximity between intelligence analysts and decision-makers is one...factor that can affect the usefulness and accuracy of the intelligence analysis provided to decision-makers. In 1949, Sherman Kent – who made substantial contributions to postwar US intelligence doctrine – pointed out that ‘intelligence must be close enough to policy, plans and operations to have the greatest amount of guidance, and must not be so close that it loses its objectivity and integrity of judgment’.1

There is something of a tradeoff in terms of the degree to which intelligence analysis and intelligence analysts should incorporate policy questions into their appreciations. This tradeoff – dubbed by Marrin ‘the Proximity hypothesis’4 – is implicit in much of the literature that addresses the interaction between intelligence and decision-making.

The underlying logic of this tradeoff is as follows; when intelligence analysis is produced close to policy-making, it tends to be more useful for improving decision-making because it addresses policy-maker interests. However, proximity can also lead to incidents of distortion – or politicization – if policy-maker biases or preferences are incorporated into the analysis. Conversely, intelligence analysis that is produced distant from policymaking tends to focus less on policy-maker interests and as a result is more ‘objective’ and contains less distortion. The corresponding downside, though, is that because the analysis cannot address policy-maker interests as readily, it can also be less useful in improving policy-maker judgement.

In other words, there is a fundamental tradeoff in terms of analytic independence and usefulness that hinges on the relative proximity between intelligence analysts and decision-makers.5

….

Competing Perspectives

Over time, many participants and observers of the intelligence process have decided that they prefer one side of the tradeoff over the other. According to Arthur Hulnick, ‘some early writers believed that intelligence must distance itself from policymaking, reach independent judgments about world affairs, and avoid tailoring intelligence judgments to satisfy the ideologic drives or policy preferences of decision-makers. These writers and practitioners might be thought of as traditionalists, since their writings reflect early views about the need for a certain level of separation.’7 Hulnick then contrasted this traditionalist approach with an activist approach whose adherents ‘believed that there was a kind of symbiotic relationship between intelligence and policy and that they were and should be closely tied together’. Hulnick’s distinction between traditionalist and activist approaches provides effective labels to use for describing the tradeoff that Kent first articulated in 1949.8 These competing perspectives or camps have engaged in a running battle in the intelligence literature quite literally for decades.

Supporting Independence

Sherman Kent is widely cited as the primary proponent of analytic independence through distance from decision-makers. As Douglas Dearth has observed, Kent ‘outlines several reasons for avoiding what he calls ‘‘the disadvantage of getting intelligence too close to policy.’’ He says that ‘‘what is unquestionably gained in guidance may well be lost in the integrity and objectivity of the operation.’’’9 As senior CIA analytic methodologist Jack Davis notes, Kent’s argument for analytic independence ‘took aim at those . . . who saw no great value in an independent intelligence unit. His doctrinal standards (consisting of administrative independence and scholarly objectivity balanced by vigorous pursuit of guidance) . . . probably seemed in the ball park to most discussants’.10

However, while Kent may be identified as an early proponent of analytic independence, he was not the earliest. As Michael Warner, the current Chief Historian of the Office of the Director of National Intelligence, has pointed out, an independent analytic capability was supported by many in the national security infrastructure as early as 1945.11 For example, Warner quotes an official as saying in 1945 that ‘intelligence must be divorced from policy making’. In addition, ‘as early as 1946 General William J. Donovan said, ‘‘intelligence must be independent of the people it serves so that the material it obtains will not be slanted or distorted by the views of the people directing operations’’’.12 Or as Allen W. Dulles put it in 1947, ‘for the proper judging of the situation in any foreign country it is important that information should be processed by an agency whose duty it is to weigh facts, and to draw conclusions from those facts, without having either the facts or the conclusions warped by the inevitable and even proper prejudices of the men whose duty it is to determine policy and who, having once determined a policy, are too likely to be blind to any facts which might tend to prove the policy to be faulty. . . . All we can do is to see that we have created the best possible mechanism to get the unvarnished facts before the policy makers, and to get it there in time.’13

These early advocates for the independence of intelligence from decisionmaking were concerned that intelligence provided within policy departments would be biased by the preferences of those within the department. According to Warner, one solution was to create an intelligence capability separate from any policy department: ‘in essence, the idea was that the President and his key advisors needed a control variable against which to test the intelligence and policy advice coming from the departments. Only a free-standing intelligence agency, they felt, could provide such a perspective. Objectivity was valued, but independence from ‘‘policy’’ was the real desideratum.’14 Or, as Roger Hilsman, a one-time director of the State Department’s Bureau of Intelligence and Research (INR), observed in 1953, many believed that ‘the danger is bias, and the cure is an intelligence organization which is independent and which never suggests, or even thinks of, policy’.15 This approach – the traditional approach – predominated in the early Cold War years to the extent that, according to Thomas Hughes – another former director of State Department’s INR – it became intelligence  ‘theology’.16

But the downside of the traditional approach was also recognized. Former National Intelligence Officer Hans Heymann counters that ‘enforcing this kind of rigorous separation’ between intelligence analysis and decision-making in the context of National Intelligence Estimates impose[d] a splendid isolation upon intelligence that ensured its eventual policy irrelevance. The vanishing applause for its product coming from the policy side prompted intelligence to reexamine its assumptions, and a new, unconventional wisdom came to be heard. Its message was that our faith in the arm’s-length relationship was misplaced, that no such relationship really ever existed, and that close ties between intelligence and policy are not only inevitable but also essential if the policymakers’ needs are to be served.17

Supporting Integration

Parallel to the traditionalists, an activist approach developed whose proponents advocated a closer relationship between intelligence and decision-making. As Jack Davis notes, this led to a doctrinal split ‘between champions of the detached and close-support approaches to the producer– consumer relationship’.18 According to Davis, in 1949 Yale University professor Willmoore Kendall argued that ‘Kent’s endorsement of the traditional separations of intelligence from domestic affairs [is] self-defeating, if the goal of the intelligence unit is to bring to bear the knowledge in which foreign policy decisions are to be made’ because some of that knowledge involves alternative policies which, according to Kent’s view, intelligence analysts should not address.19 So, instead, Kendall preferred to see a closer relationship between intelligence and decisionmaking in which ‘the intelligence function [helps] the policymakers ‘‘influence’’ the course of events by helping them understand the operative factors on which the US can have an impact’.20

Roger Hilsman was himself another early advocate for a closer relationship between intelligence analysis and decision-making. In 1953, Hilsman argued that ‘a more effective integration of knowledge and action’ – or intelligence analysis and decision-making – will require intelligence analysts to become more policy-oriented.21 Hilsman concludes that the ‘division of labor (between intelligence and decision-making) seems to be both arbitrary and awkward’ because the warning function – which is ‘merely the act of recognizing a problem and making a preliminary analysis of it’ – is performed by both intelligence analysts and decision-makers, and ‘both organizations thus do the same kind of work, and need the same information, knowledge, method, and skills’.22 Hilsman concludes that in order for intelligence to be ‘useful and significant’ it should work much more closely with decision-makers.23

Former Director of Central Intelligence (DCI) – and current Secretary of Defense – Robert Gates is a more recent advocate of greater proximity between intelligence analysts and decision-makers. In 1989, Gates argued that ‘contrary to the views of those who are apprehensive over a close relationship between policymakers and intelligence, it is not close enough. More interaction, feedback, and direction as to strategy, priorities, and requirements are critical to better performance.’ 24 Gates later said, that the Intelligence Community has to be right next to the policymaker, that [the officer] has to be at his elbow – that he has to understand what is on his mind. He has to understand what his future concerns are. He has to understand what his agenda is. He has to understand some of the initiatives that he is thinking about taking. He has to be willing to ask the policymaker what he’s working on, or what came out of his last conversation with a world leader so that the intelligence can be made relevant. 25

Perhaps best summarizing the activist approach, former chairman of the President’s Foreign Intelligence Advisory Board Anne Armstrong has observed that ‘ultimately the intelligence officials must be guided by the needs of the policy, rather than by the search for knowledge for its own sake’. 26 Accordingly, even though a ‘strategy of narrowing the distance between intelligence and policy holds certain risks for the integrity of the former . . . the risks must be run’. 27

The underlying rationale for some who advocate greater proximity between analysts and decision-makers has to do with their conceptualization of the decision-making process in which intelligence analysis and decisionmaking are not separate functions, but flip sides of the same coin. Thomas Hughes articulated this view when he said that ‘despite the theological clarifications of their separate essences . . . the typical policy-maker and the typical intelligence officer are inevitably a little of both. In practice their roles tend to merge’.28 In 1985 Hans Heymann described Hughes’ approach as ‘a new way of thinking about intelligence and policy . . . in which the two communities were seen as awkwardly entangled and intertwined in what might be described as a competitive and often conflictual symbiotic relationship’.29

But just as the traditionalists had to contend with those who argued that their analytic products were not relevant to decision-makers, the activists were challenged by the accusation that closer proximity risks politicization. The relationship between proximity and politicization has been in the literature for decades. For example, in 1964 Klaus Knorr observed that ‘the closer intelligence is produced, in the bureaucratic context, to the consumer in government, the more it tends to acquire a bias favoring actions preferred by each consumer group’.30 This linkage between proximity and politicization has provided the cautionary note that prevented practitioners from fully embracing the activist approach.

It can, therefore, be seen that the importance of relative proximity has been embedded in the intelligence literature for decades. Some authors have argued for closing the distance between intelligence and decision-making, whilst others have defended the traditionalist approach.

A middle ground does exist between traditionalists and activists, however. Kent himself was not as uncompromising an advocate for independence as many in the literature describe him. As Joshua Rovner put it, ‘Kent himself was well aware of the dangers of overly independent analysts. While championing analytical objectivity, he was very clear that intelligence needs to seek policy guidance.’31 Rovner then points out that Kent articulated both sides of the tradeoff, and while popularly identified with the role of independence, he also observed that ‘of the two dangers – that of intelligence being too far from the users and that of being too close – the greater danger is the one of being too far’.32 Kent also said that the ‘the single most important principle of successful intelligence [is] closeness of intelligence producers to intelligence users or consumers’.33

So even though Kent is identified in the literature as an advocate for independence, a closer look at his writings reveals him to be someone who saw the tradeoff between the two values and wanted to straddle the line between them. Similarly, Percy Cradock – a former chairman of the British Joint Intelligence Committee – has observed that ‘Ideally, intelligence and policy should be close but distinct. . . . The best arrangement is intelligence and policy in separate but adjoining rooms, with communicating doors and thin partition walls, as in cheap hotels.’34

Identifying organizational structures that are able to do this, however, can be quite difficult

XXX

Citations:

1 Sherman Kent, Strategic Intelligence for American World Policy (Hamden, CT: Archon Books 1965 [1949]) p.180.

4 Stephen Marrin, Intelligence Analysis and Decisionmaking: Proximity Matters (PhD dissertation, University of Virginia, May 2009).

5 For a valuable treatment of this tradeoff and its implications, see Chapter 4 (‘Incorruptibility or Influence? Costs and Benefits of Politicization’) in Richard Betts, Enemies of Intelligence: Knowledge and Power in American National Security (New York: Columbia University Press 2007) pp.66–103

7 Arthur S. Hulnick, ‘The Intelligence Producer-Policy Consumer Linkage: A Theoretical Approach’, Intelligence and National Security 1/2 (May 1986) pp.212–14.

8 The late Michael Handel – an intelligence scholar, professor at the Naval War College and contemporary of Hulnick writing at about the same time – came up with different labels for the same tradeoff; with the ‘professional approach’ as an equivalent for Hulnick’s traditional approach balanced against the ‘realistic approach’ as an equivalent to Hulnick’s activists. See Michael Handel, ‘The Politics of Intelligence’, Intelligence and National Security 2/4 (October 1987) pp.5–38 (pp.11–12).

9 Douglas H. Dearth, ‘The Politics of Intelligence’ in Douglas H. Dearth and Royal Thomas Goodden (eds.) Strategic Intelligence: Theory and Application, 2nd ed. (US Army War College’s Centre for Strategic Leadership/Defense Intelligence Agency’s Joint Military Intelligence Training Center 1995) pp.97–121 (p.98).

10Jack Davis, ‘The Kent–Kendall Debate of 1949’, Studies in Intelligence 36/5 (1992) pp.91– 103 (p.96). National Assessment by the NSC Staff 1968–80 647

11See Michael Warner, ‘Intelligence Transformation and Intelligence Liaison’, SAIS Review 24/1 (Winter/Spring 2004) pp.85–6.

12Major Gen. William J. Donovan, 1 May 1946, Vital Speeches 12/14, p.248, as quoted in Amos Kovacs, ‘The Nonuse of Intelligence’, International Journal of Intelligence and Counterintelligence 10/4 (Winter 1997–1998), pp.383–417 (p.406). Also previously quoted in Roger Hilsman, ‘Intelligence and Policy-Making in Foreign Affairs’, World Politics (October 1952) pp.1–45 (p.4).

13Allen W. Dulles, Memo Respecting Section 202 (CIA) of the Bill to Provide for a National Defense Establishment, dated 25 April 1947, as published in Hearings on S. 758, p.525, as quoted in Hilsman, ‘Intelligence and Policy-Making in Foreign Affairs’, p.4.

14Michael Warner, ‘Intelligence Transformation and Intelligence Liaison’, pp.85–6.

15Hilsman, ‘Intelligence and Policy-Making in Foreign Affairs’, p.4. 648 Intelligence and National Security

16Thomas L. Hughes, ‘The Fate of Facts in a World of Men: Foreign Policy and IntelligenceMaking’, Headline Series #233 (New York: The Foreign Policy Association 1976) p.5.

17Hans Heymann, ‘Intelligence/Policy Relationships’, in Alfred C. Maurer, Marion D. Tunstall and James M. Keagle (eds.) Intelligence: Policy and Process (Boulder, CO/London: Westview Press 1985) pp.57–66 (pp.57–8).

18Davis, ‘The Kent–Kendall Debate of 1949’, pp.96–7.

19Ibid., p.96. For more, see Willmoore Kendall, ‘The Function of Intelligence’, World Politics 1/4 (July 1949) pp.542–52. While Davis has framed the conflicting approaches taken by Kent and Kendall as a debate, he was not the first to highlight the differences in their views. At least one other author prior to Davis made the same observation. See Anne Armstrong, ‘Bridging the Gap: Intelligence and Policy’, The Washington Quarterly (Winter 1989) pp.23–34 (27–8).

20Davis, ‘The Kent–Kendall Debate of 1949’, p.95.

21Hilsman, ‘Intelligence and Policy-Making in Foreign Affairs’, p.45.

22Ibid., p.33.

23Ibid., p.44.

24Robert M. Gates, ‘An Opportunity Unfulfilled: The Use and Perceptions of Intelligence at the White House’, The Washington Quarterly (Winter 1989) pp.35–44 (p.40).

25H. Bradford Westerfield, ‘Inside Ivory Bunkers: CIA Analysts Resist Managers’ ‘‘Pandering’’ – Part II, International Journal of Intelligence and Counterintelligence 10/1 (Spring 1997) pp.19–54 (p.27), quoting Gates from US Senate Select Committee on Intelligence Hearings. Nomination of Robert Gates to be Director of Central Intelligence (September– October 1991) Vol. I, pp.569–70.

26Anne Armstrong, ‘Bridging the Gap: Intelligence and Policy’, The Washington Quarterly 12/1 (Winter 1989) pp.23–34 (p.30).

27Ibid., p.33.

28Hughes, ‘The Fate of Facts in a World of Men’, p.7.

29Heymann, ‘Intelligence/Policy Relationships’, p.58.

30Klaus Knorr, Foreign Intelligence and the Social Sciences, Research Monograph No.17 (Princeton, NJ: Center of International Studies, Woodrow Wilson School of Public and International Affairs, Princeton University, 1 June 1964) p.42.

31Joshua Rovner, ‘Pathologies of Intelligence–Policy Relations’, Unpublished paper, 24 June 2005, 5http://web.mit.edu/polisci/students/jrovner/pathologies_of_intelligence-policy_ relations.pdf4(accessed 11 July 2006) p.6.

32Kent, Strategic Intelligence for American World Policy, p.195, as quoted in Rovner ‘Pathologies of Intelligence-Policy Relations’, p.6.