- 1 Introduction by Jack S. Levy, Rutgers University
- 2 Review by Uri Bar-Joseph, Haifa University
- 3 Review by Philip H.J. Davies, Brunel University
- 4 Review by Michael S. Goodman, Department of War Studies, King’s College London
- 5 Review by Keren Yarhi-Milo, Princeton University
- 6 Author’s Response by Joshua Rovner, Department of Strategy and Policy, U.S. Naval War College
The U.S. intelligence failures associated with 9/11 and with Iraqi weapons of mass destruction generated renewed interest in the question of intelligence failure, the study of which had been disproportionately influenced by the study of the failures at Pearl Harbor, Barbarossa, and Yom Kipper. The Iraqi WMD case in particular focused more attention on the question of the politicization of intelligence, an age-old problem but one that had been neglected in studies of the classic cases. The subsequent scholarly literature has focused on the policy question of the proper relationship between intelligence and policy, and on the causal questions of where and when politicization is most likely to occur and the role it plays in the processes leading to intelligence failure.
H-Diplo | ISSF Roundtable, Volume III, No. 17 (2012)
Joshua Rovner. Fixing the Facts: National Security and the Politics of Intelligence. Ithaca: Cornell University Press, 2011. ISBN: 978-0-8014-4829-4 (hardcover, $35.00).
Thomas Maddux and Diane Labrosse, H-Diplo/ISSF Editors
George Fujii, H-Diplo/ISSF Web and Production Editor
Introduction by Jack S. Levy, Rutgers University
Published by H-Diplo/ISSF on 9 July 2012
The U.S. intelligence failures associated with 9/11 and with Iraqi weapons of mass destruction generated renewed interest in the question of intelligence failure, the study of which had been disproportionately influenced by the study of the failures at Pearl Harbor, Barbarossa, and Yom Kipper. The Iraqi WMD case in particular focused more attention on the question of the politicization of intelligence, an age-old problem but one that had been neglected in studies of the classic cases. The subsequent scholarly literature has focused on the policy question of the proper relationship between intelligence and policy, and on the causal questions of where and when politicization is most likely to occur and the role it plays in the processes leading to intelligence failure.
Fixing the Facts is a major contribution to the study of the politicization of intelligence in modern democracies, and thus to our understanding of the more general phenomenon of intelligence failure. Joshua Rovner identifies a variety of pathologies of intelligence-policy relations, types of politicization, and standard explanations for politicization, and he engages normative questions about the proper relationship between intelligence and secrecy in democratic states. The core of the book, however, is Rovner’s theoretical argument about the sources of politicization and his efforts to test his theory against the historical evidence. Standard explanations for politicization involve the personal proximity of intelligence officers and policy makers; organizational proximity between intelligence agencies and the political decision-making process; and the organizational dependence of intelligence agencies and analysts on political leaders for resources, autonomy, and career prospects. Rovner argues that these arguments cannot explain variations in politicization across a number of historical cases. The puzzle is why political leaders would incur the bureaucratic and political risks of politicizing intelligence when they can simply ignore intelligence and implement their desired policies.
Rovner then lays out his argument that the politicization of intelligence derives primarily from domestic politics. Political leaders’ goals of maintaining their positions of power and gaining support for their policies generate incentives for them to periodically pressure intelligence agencies and analysts to manipulate intelligence in a way that reinforces their preferred policies. Political leaders are most likely to impose this pressure after they make public policy commitments, and after a critical constituency emerges in opposition to those policies. Rovner tests his explanatory model against the leading alternatives for six historical cases, organized in terms of three paired comparisons (which incorporate variations in the dependent variable by including cases in which intelligence was ignored rather than politicized): the Johnson Administration’s responses to intelligence analyses of the domino theory in 1964 and estimates of enemy strength in Vietnam in 1967; the Nixon and Ford Administrations’ estimates of the size and aims of the Soviet strategic arsenal; and U.S. and British estimates of Iraqi capabilities and intentions in the 1998-2003 period. In this series of detailed case studies, Rovner finds that his domestic oversell model of politicization provides a much better explanation for important variations in politicization across these cases than do traditional explanations based on individual or organizational proximity or dependence.
Each of the reviewers in this Roundtable agrees that Fixing the Facts is a major contribution to our understanding of the politicization of intelligence, and that its strengths include the combination of a novel theoretical argument and a series of rich historical case studies. Yet each reviewer highlights some of the book’s limitations. I find it useful to organize these criticisms into four major categories: internal validity, external validity, conceptual issues, and normative considerations such as the proper role of secrecy in the intelligence processes of a democratic state. I focus on all but the normative issues here. Internal validity concerns the question of whether the evidence supports the theory in the particular cases Rovner examines – whether each historical case study demonstrates that politicization was a major source of intelligence failure and that it derived from the domestic sources specified by Rovner rather than from individual or organizational proximity or dependence. External validity concerns the extent to which Rovner’s findings can be generalized beyond his specific case evidence to other cases of the politicization (or neglect) of intelligence. Here it is useful to distinguish between the relevance of Rovner’s findings for other American cases and for other non-American cases, presumably in democratic regimes. Rovner does not attempt to explain the politicization of intelligence in non-democratic regimes, where the personal and organizational dependence of intelligence analysts and agencies on political leaders presumably plays a critical role.
The reviewers generally agree that Rovner’s well-documented historical analyses provide strong evidence for his theoretical arguments, but they raise questions about some of his particular interpretations. They question whether leaders’ public commitments and the rise of a critical constituency to challenge the political leader were the only or even primary factors involved in the politicization of intelligence. Keren Yarhi-Milo, for example, argues that Rovner’s domestic overall model does not provide a satisfactory explanation for the “Team B” affair in the Ford Administration. She also emphasizes the potential explanatory power of individual-level psychological variables, including motivated biases, which are neglected by Rovner. Uri Bar-Joseph also emphasizes the role of individual-level factors, which he illustrates by referring to an earlier U.S. case – President Dwight Eisenhower’s decision during the “missile gap” controversy at the end of his second term not to try to manipulate intelligence to maintain his domestic stature. Bar-Joseph emphasizes the need to supplement Rovner’s model with three additional sets of variables: individual decision-making variables such as political leaders’ operational code belief systems, leadership styles, and personalities; organizational variables (both institutional and cultural) such as the strength of the organization and its level of professionalism; and the importance of the issue. I suspect that some students of intelligence might agree with Bar-Joseph that Rovner underestimates the extent to which a healthy and professional organizational culture often frustrates attempts at manipulation by political leaders, and consequently that it sometimes deters such efforts.
Turning to the criterion of external validity, several of the reviewers accept large parts of Rovner’s analyses of his American cases but question the applicability of the model to other political systems. Philip H. J. Davies argues that “the way intelligence works in the United States is highly idiosyncratic,” and that relatively few general analyses of the U.S. system are transferable to other national systems.” Michael S. Goodman argues that the relevance of leaders’ concerns with public justifications for their actions, especially in the face of opposition to their policies, is “very much dependent on a particular system of government – the U.S. model.” Similarly, Bar-Joseph criticizes the tendency of American scholars of intelligence “to present their findings as universal.” He argues that Rovner’s model is “partially valid” for the United States but “much less so” for other democratic states.
The reviewers go on to identify specific respects in which intelligence-policy relations in other democracies differ from those in the United States in ways that are consequential for the analysis of politicization. Davis argues that Rovner (like other American analysts) fails to understand the operation of the intelligence system and its relationship to the policy process in the United Kingdom, in part because the independent Civil Service has no counterpart in the United States. Goodman makes a similar point about the consequences of political appointments of the heads of intelligences agencies in the UK but not in the United States. Bar-Joseph argues that the politicization of British intelligence regarding Iraqi WMD was a rare event in the history of British intelligence, where the more common pattern is the intervention of intelligence in the political process, not top-down politicization. He argues that in Israel “there is no record of an attempt by policymakers to politicize intelligence.”
I now turn to some conceptual issues, beginning with the definition of the central concept of politicization, about which the reviewers say surprisingly little. Rovner defines politicization as “the attempt to manipulate intelligence so that it reflects policy preferences” (5), though later (29) Rovner drops the word “attempt” and speaks only of the manipulation of intelligence. I much prefer the latter formulation. If top-down political pressure on intelligence analysts is rejected by analysts and has no causal impact on the intelligence product, I would not say that intelligence has been politicized. It is important to distinguish attempts to manipulate intelligence and those that succeed. We should also remember that successful politicization (as defined above) does not necessarily imply either that the leader is successful in diffusing opposition to his policies, or that the politicization of intelligence has a causal impact on policy. It is conceivable that in the absence of politicization, political leaders might simply have neglected intelligence and followed identical policies. These are empirical questions that raise difficult methodological issues. 
A more troubling issue is that Rovner’s fairly broad definition includes not only pressure from political leaders for intelligence to please, but also the production of estimates by intelligence agencies with the deliberate aim of undermining policy decisions (30). Each of these patterns is an important source of intelligence failure, each can poison the intelligence-policy relationship, and each can feed on the other. These patterns are sufficiently different, however, that we need different concepts to describe them. I prefer to reserve the concept of the politicization of intelligence to top-down manipulation of the intelligence product. The question is not whether the intelligence process is political, but what causal mechanisms are involved. As John Ferris argues, “Intelligence always has a political context, but that need not produce politicization….”
In his elaboration of different types of politicization, Rovner makes a useful distinction between “direct” and “indirect” politicization. The former involves coercive pressure (or rewards) or “manipulation by appointment” (31) of sympathetic or malleable intelligence directors or analysts, whereas the latter involves more tacit signals and subtle threats. Some of Rovner’s types of politicization are more problematic. One of his types of politicization includes “widely held strategic assumptions and social norms [that] restrict the bounds of acceptable analysis” (205). These “embedded assumptions” do indeed shape both the intelligence product and views of policymakers, but they are best conceived as an analytically distinct path to intelligence failure rather than as a form of politicization.
This brings me back to Yarhi-Milo’s argument for an alternative psychological explanation for the politicization of intelligence. She notes “the importance of motivated biases in explaining both why policy-makers would try to interfere, influence, or have a biased interpretation of intelligence, as well as why intelligence officials would revisit and at times radically change their estimates in the fact of a change in policy.” I agree on the causal importance of psychological variables in intelligence, even if the result is a less parsimonious explanation of politicization than Rovner advocates. There is an important analytic distinction, however, between a process in which political leaders interfere with and try to influence intelligence, and one in which political leaders have a “biased interpretation of intelligence.” Each leads to cases of policy driving intelligence rather than intelligence driving policy, and each can contribute to intelligence failure, but they do so through different causal mechanisms or paths.
The first (interference and influence) involves a political mechanism in the form of top-down pressure on intelligence analysts to produce a particular intelligence product. The second (biased interpretation) involves a psychological process, one that does not necessarily involve the manipulation of intelligence agencies. It may lead, for example, to strongly held beliefs, high levels of confidence in those beliefs, and a decision to neglect intelligence. The political and psychological mechanisms are analytically distinct, but the two can reinforce each other in practice.  Empirically distinguishing the two is another matter, and one that involves some difficult methodological issues, as Yarhi-Milo illustrates with reference to British intelligence estimates of the Nazi threat in the 1930s.
I have aimed in this essay to provide a useful introduction to the four excellent reviews of Fixing the Facts, and also to add some additional comments of my own. The bottom line is clear: the book is essential reading for all of those who are interested in the complex relationships among intelligence, politics, and policy.
Joshua Rovner is Associate Professor of Strategy and Policy at the U.S. Naval War College, adjunct professor in the School of International and Public Affairs at Columbia University, and reviews editor of The Journal of Strategic Studies. He is the author of Fixing the Facts: National Security and the Politics of Intelligence (Cornell University Press, 2011). Other recent publications include “The Heroes of COIN,” Orbis, Vol. 56, No. 2 (Spring 2012), pp. 1-18; and “Dominoes on the Durand Line? Overcoming Strategic Myths in Afghanistan and Pakistan,” with Austin Long, Foreign Policy Briefing, No. 92 (Washington, DC: Cato Institute, June 2011). The views here are his alone.
Jack S. Levy is Board of Governors’ Professor of Political Science at Rutgers University, and a Senior Associate at the Saltzman Institute of War and Peace Studies at Columbia University. He is a former President of the International Studies Association (2007-08) and of the Peace Science Society (2005-06). He is author of War in the Modern Great Power System, 1495-1975 (1983) and co-author (with William R. Thompson) of Causes of War (2010) and The Arc of War: Origins, Escalation, and Transformation (2011). His current projects include balance of power theory, preventive war, intelligence failure, political oppositions and war, the militarization of commercial rivalries, the causes of World War I, and grand strategy in the 1930s.
Philip H. J. Davies is Director of the Brunel University Centre for Intelligence and Security Studies (BCISS). He holds a BA Hons in Sociology (Trent University), MA in Social and Political Thought (York University (Canada) and a PhD in Sociology from the University of Reading. He was one of the authors of the current UK military Joint Intelligence Doctrine (JDP 2-00 Understanding and Intelligence Support to Joint Operations) and recently completed a two-volume comparative study of the US and UK national intelligence systems (Intelligence and Government in Britain and the United States, Praeger Security International, 2012).
Uri Bar-Joseph is a Professor at the International Relations Division at Haifa University, Israel. He concentrates on strategic and intelligence studies, especially focusing on the Arab-Israeli conflict and Israeli security policy. In addition to more than 30 refereed journal articles in journals such as Political Science Quarterly, The Journal of Conflict Resolution, Security Studies, Political Psychology and Armed Forces and Society, he has written five books, the most recent of which is The Angel: Ashraf Marwan, the Mossad, and the War of Yom Kippur (Hebrew, 2011).
Michael S. Goodman is a senior lecturer in the Department of War Studies, King’s College London. He has published widely in the field of intelligence history and scientific intelligence, including Spying on the Nuclear Bear: Anglo-American Intelligence and the Soviet Bomb (Stanford University Press, 2008), Spinning Intelligence: Why Intelligence Needs the Media, Why the Media Needs Intelligence (Columbia: Columbia University Press, 2009), and more recently Learning from the Secret Past: Cases in British Intelligence History (Georgetown University Press, 2011). His articles have appeared in Intelligence and National Security, the International Journal of Intelligence and Counterintelligence, CIA Studies in Intelligence, Cold War History, the Journal of Cold War Studies, the Journal of Strategic Studies, Contemporary British History, Historical Studies in the Physical Sciences, Diplomacy and Statecraft, and International History Review. He is currently on secondment to the Cabinet Office where he is writing the Official History of the Joint Intelligence Committee.
Keren Yarhi-Milo is an Assistant Professor at the Department of Politics and the Woodrow Wilson School for Public and International Affairs at Princeton University. Before joining Princeton University in 2009, Yarhi-Milo was a pre-and-post doctoral fellow at Harvard University’s Olin Institute and The Belfer Center, respectively. Her book MS, titled “Knowing Thy Adversary: Leaders, Intelligence, and Assessments of Intentions in international Politics,” is near completion and is based on her dissertation for which she received the Kenneth Waltz Award for the best dissertation in the field of International Security by the American Political Science Association in 2010. She also has an article forthcoming in International Studies Quarterly. Yarhi-Milo holds a Ph.D. in Political Science from the University of Pennsylvania, and a B.A., summa cum laude, from Columbia University.
Politicization of intelligence is a universal problem, one that has existed throughout history and throughout the world. It is true that the first instance of espionage in Biblical times — that of the raven and the dove, which Noah sent to see “if the waters were abated from off the face of the ground” — did not involve politics. But in this case Noah was the intelligence producer and consumer and the spies were not human. When the setting gets more complex politics is far more likely to enter the process. Take, for example, the most famous spy mission of the Bible. The twelve spies who were sent by Moses to survey Canaan returned after an extensive fact-finding mission and informed Moses, Aaron and the whole congregation of the Children of Israel, that the Promised Land was a “land of milk and honey.” But according to the estimation of ten of them, it was populated by too strong enemies and could not be occupied. Only two spies, Joshua and Kaleb, disagreed. The voice of the people in this case was clear and was very different from the voice of God: The Israelites did not want to embark on the dangerous occupation mission and threatened to stone the two spies that supported this option. Under the popular pressure the ten spies changed the facts, terming Canaan now as “a land that eateth up the inhabitants thereof” and, implicitly, one that is not worth the risk of a military defeat. Ultimately, the Israelites were left in the desert for forty years and only their next generation conquered Canaan.
Taking the politicization of intelligence as a universal problem that should be studied across various political cultures is important since most students of the subject are American and, though they focus on the relations between intelligence producers and consumers in what is a distinctly American context, they tend to present their findings as universal in nature. This is problematic. Just as most world states did not follow the American model of healthcare so, in many democratic states the politicization of intelligence, even under favorable conditions (as per the model that Joshua Rovner suggests), is not a common phenomenon.
At the center of Rovner’s model is the need of policymakers to garner support for a specific foreign policy line to which they are committed and which is opposed by other actors involved in the conduct of national strategy. Consequently, gaining the approval and support of intelligence organizations, which are deemed by the public as both knowledgeable and objective, becomes of critical importance for the committed policymakers. Under these circumstances policymakers are more likely to pressure the intelligence producers to change the intelligence estimate according to their policy preferences. The six case studies used by Rovner validate his model.
The model that is tested in In Fixing the Facts is indeed sensible. Nevertheless, it does suffer from certain shortcomings, the most important of which being that while the model is partially valid for the U.S. it is much less so for intelligence-policy relations in other democratic states. I will first discuss, very briefly, the cases of Britain, France and Israel in order to show that the pattern of politicization is very rare there while other patterns are more common. I will then show that even in the American context politicization, as defined by Rovner, was not necessarily policymakers’ choice even when they found themselves in situations which (according to his model) should have made this choice more likely. Finally I will suggest an alternative model that is, to my mind, more applicable to the study of the politicization.
Rovner is quite convincing when he argues that Britain’s intelligence estimate with regard to Iraq’s phantom WMD arsenal also reflected a positive response to political pressure. But a more thorough study of the history of the British intelligence community would have shown that this type of politicization was the exception and not the rule. There is no other clear record in British history in which intelligence estimates did not correspond with the government policy in highly controversial issues and political pressure was wielded vis-à-vis intelligence makers to change their estimate. In some major cases the opposite was true. In 1920, for example, military and intelligence officers conspired against Prime Minister Lloyd George since they regarded his Soviet policy as being too soft. Four years later, retired and active duty intelligence officers leaked the famous Zinoviev letter to the press in order to drive out the Labor party from power in the upcoming elections. According to some sources, in 1974 MI5 officers considered leaking to the press information that would portray Labor Prime Minister Harold Wilson as a “security risk,” although the official history of Britain’s Security Service gainsays this claim.
The case of France presents a similar pattern. Although not as much is known about the relationship between policy and intelligence in this country since the Dreyfus affair, the scarce evidence shows that the combination of an intelligence culture which is based on careerism, political expediency, lack of accountability and resort to covert action, with a political culture in which intelligence may be used for parochial interests, created unclear boundaries between intelligence and politics. The outcome resembles more the British pattern of intelligence intervention in politics rather than Rovner’s model of politicization: As Douglas Porch concluded, “Virtually every serving president of the Fifth Republic was a target of a scandal or “dirty trick” with origins deep in secret service conspiracies.”
In the case of Israel, where national intelligence estimates determine a large share of the public agenda, there is no record of an attempt by policymakers to politicize intelligence. In most cases where policy and intelligence do not correspond, Israeli prime ministers tend to ignore (but not necessarily neglect) the intelligence estimate. Thus, for example, when the Director of Military Intelligence (who in Israel is, ex officio, the national estimator), General Uri Sagie, came in 1991 to Prime Minister Yitzhak Shamir to convince him that Syria was willing to sign peace with Israel in exchange for the return of the occupied Golan Heights, Shamir, who objected in principle to any return of territory that was occupied in the 1967 war, heard him politely and ignored his estimate. Similarly, when Prime Minister Ehud Barak decided to pull the Israel Defense Force (IDF) out of the self proclaimed “security zone” in southern Lebanon, Military Intelligence (Aman) warned that such a move would enable Hezbollah to carry out terrorist attacks against Northern Israel. Barak did not resort to any kind of pressure to change this estimate that had played a major role in the public debate about the planned withdrawal, which ultimately took place in May 2000. Perhaps the only case where intelligence makers were accused of coloring their estimates according to political preferences is that of Brig, Gen. Amos Gilad, who as the head of the Research Brigade of Military Intelligence provided in 2000 an estimate regarding Yasser Arafat that served the policies of Prime Minister Ehud Barak, but had contradicted the assessment of Gilad’s own experts on this issue. But even here there is no claim that Barak had actively attempted to politicize the intelligence product. As in France and in Britain, Israeli intelligence agencies did attempt a few times to pursue, in an unauthorized manner, an independent policy. The most famous episode of this kind took place in 1954 when the Director of Military Intelligence ordered an operation in Egypt intended to cause the failure of British-Egyptian negotiations on the evacuation of the British military from the zone of Suez. It ended in a complete operational and political disaster. As far as we know, Israeli intelligence officers had never conspired against their own bosses.
These examples should suffice to show that Rovner’s model – which assumes that public commitment and the emergence of critical constituency create a situation in which, because of fear of potential costs, policymakers have a greater demand for intelligence support and hence to politicize intelligence – lacks validity in non-American contexts. But a closer look at policy-intelligence relations in the U.S. shows that even in this context the model is of limited validity. Here I will resort to two examples: the first will demonstrate how the politicization of intelligence can take place, and in a rather effective manner, but not in the method described by Rovner’s model. The second example will show that even when the conditions for politicization that are set by the model exist, politicization does not necessarily take place. This example will call into question the validity of the variables that, according to Rovner, determine the likelihood of politicization and will suggest alternative ones.
In many respects the most evident case of politicization of the intelligence process in U.S. history is that of the nomination of William Casey’s as DCI in early 1981. Casey, the last of the “cold warriors,” served as Ronald Reagan’s campaign manager in 1980 and was “the first genuine political appointee at the CIA.” He brought with him to office a new conception of the role of DCI, at the center of which stood the need to improve analytical estimates, so that they better reflected the requirements and the agenda of the new administration. During his six years’ tenure Casey succeeded, indeed, in rocking the CIA analytical boat and turn its products into those that were more applicable to policymakers’ needs but at the cost of seriously damaging their objectivity and accuracy.
During his tenure as DCI (and that of Robert Gates as Deputy Director of Intelligence) the Agency’s analytical products reflected less the available intelligence evidence and more Casey’s ideology and beliefs. Melvin Goodman, who had worked under him, described it as follows: “[w]hen a finished product of the directorate of intelligence (DI) ran up against his naturally hard line views, he immediately moved to kill the product and punish its authors.” Consequently, major intelligence estimates, ranging from the level of Soviet involvement in international terrorism and the probability of Soviet intervention in Iran after Khomeini’s death, to the likelihood of a popular revolution in Mexico and the magnitude of the Sandinista threat to American interests in Latin America, reflected less the available data and analysts’ expertise and more the political preferences and the ideological beliefs of the DCI – President Reagan’s man in the CIA.
Casey’s nomination presents an alternative to the model suggested by Rovner. Instead of politicizing inconvenient estimates, the nomination of the administration’s man as DCI enabled in-house control of the entire process, thus obviating the need to tamper with the finished product. In this sense, the nomination of Casey was like nominating Team B (or the Office of Special Plans) to head the CIA.
History also shows that under conditions which are highly favorable for the use of even genuine intelligence estimates in order to garner domestic support, other factors prevent such behavior. The launch of the Soviet Sputnik in October 1957 triggered the “missile gap” debate, the biggest public controversy of its kind during the 1950s. President Eisenhower was attacked by Democratic politicians, the military-industrial complex, and the media in general, for pursuing a policy that allowed the USSR to get ahead of the USA in the ICBM race. Senator John Kennedy famous accusation, that the US “was fiddling as Rome burned,” typified the kind of attacks to which the president was subjected. On Rovner’s model, Eisenhower was supposed to use the excellent intelligence ammunition at his disposal – the information that was gathered for years by the U2 reconnaissance flights, which indicated that the Soviets, in fact, had no such advantage –to convince the public of the sense in his policy. Politicization of the intelligence product was not even required. However, Eisenhower avoided making such a move and rejected a suggestion by his Secretary of State to disclose the intelligence foundations of his policy. In his memoirs he wrote about this episode: “I was hampered, of course, by the fact that I could not reveal secrets which in themselves would have reassured our people.”
The “missile gap” controversy illustrates that even in the USA, where politicization of intelligence seems to be the practice more than in any other democratic state, domestic politics cannot be the only explanatory variable. Here, three other factors, at three different levels of analysis, account for the outcome. The first is on the policymaker level. Eisenhower entered office with a better grasp of national security, military, and intelligence affairs than arguably any other American president. He was haunted by the surprise attacks on Pearl Harbor and Korea and feared a similar attack throughout his presidency. During WWII he learned to appreciate the critical contribution of intelligence, especially Signal Intelligence (SIGINT) and Imagery Intelligence (IMINT), as a safety belt against such a surprise attack and as a means to ensure military effectiveness. And since he attributed such a critical role to intelligence, his relationships with DCI Allen Dulles, his brother, Secretary of State John Foster Dulles, as well as other close aides, were based, first and foremost, on mutual trust. In addition, Eisenhower’s personal sense of responsibility and decency prevented him from sacrificing prime collection assets such as the U2 in order to preserve his political stature. Factored together, Eisenhower’s beliefs and his personality explain why he did not resort to using secret intelligence, or to manipulate intelligence to defend a controversial policy.
The second factor, at the organization level, involves the organization’s strength. During the 1950s, before failures such as the Bay of Pigs tarnished its public image, the CIA was at the height of its power. In terms of bureaucratic and public opinion wars, the agency was a strong actor. Its analytical section produced high quality strategic intelligence estimates that proved, in retrospect, to have been more accurate than alternative estimates that were produced by the military organs. The CIA’s analysts, particularly at the Office of National Estimates (ONE), were professional experts committed to high level of professional standards and values. So were others, including the DCI. Altogether, this healthy atmosphere created convenient conditions for the production of objective estimates. Thus, for example, when ONE completed in 1958 the “Estimate of World Situation” and DCI Dulles was unhappy with its gloomy outlook, the analysts and the DCI conducted a series of discussions at the end of which Dulles accepted the estimate of his experts. Had attempts to politicize intelligence products been made at this time (they had not been), this professionalism and healthy institutional practice could have frustrated them.
Finally, at the systemic level, the relevant factor is the level of importance of issue at stake. Obviously, the more crucial the intelligence product is, the less will policymakers tend to politicize it and the more will intelligence officers be ready to defend their estimates against an attempt to alter them according to political preferences. At the height of the Cold War, when a nuclear confrontation with the USSR was considered likely, the issue at stake was the feasibility of a Soviet nuclear attack against the U.S., or, on other words, the fate of the United States. None of the issues at stake in the cases that Rovner discusses in his study was as important as this one.
This discussion leads to three main conclusions. First, the fact that politicization of intelligence is quite common in the United States and far less so in other democratic states, indicates that this type of behavior tends to flourish in certain political and intelligence cultures. This is not the place to enter into a discussion of the uniqueness of the American political culture, which developed separately and in a different manner from the political cultures of other democratic states. Suffice it here to say that some values, such as individualism and competiveness, which are so prominent in American culture and are less salient in other political cultures, may contribute to the tendency to politicize intelligence in the U.S. Intelligence culture is defined here as “modes of thought and action derived from perceptions of national historical experience, aspirations for self-characterization, and distinctive state experiences, with respect to the role of intelligence information and analysis in shaping foreign policy.”  The most relevant cultural trait that may explain why politicization of intelligence is more common in the American intelligence culture is the aggressiveness of bureaucratic competition of this community, which is expressed in both lack of bureaucratic cooperation and a tendency towards “intelligence to please” as a means to influence policy and gain power within the policymaking machinery. Second, President’s Eisenhower conduct during the “missile gap” controversy demonstrates that despite favorable conditions for the use of intelligence (not even calling for its politicization), American policymakers may avoid such a move. This suggests that Rovner’s model should be refined. Intuitively, it is possible to suggest that the likelihood of politicization of the intelligence product when the political interest to do so exists depends on three additional clusters of factors: At the decision-making level, the policymakers’ personal straits, leadership style, and operational code; at the bureaucratic level, the intelligence organization’s bureaucratic strength, its level of professionalism, and its level of commitment to the ethics of intelligence analysis; and, at the system level, the importance of the issue at stake.
The final conclusion is that the subject of politicization of intelligence demands more research. Rovner does a good job in showing (pp.192-97) the damage that this type of behavior caused American national interest in 1968 Vietcong OB controversy, the 1969 SS-9 episode, the 1975 Team B case, and the American and British estimates of Iraq’s WMD capabilities prior to its invasion in 2003. It is quite possible that the politicization of the intelligence process with regard to the Viet Cong OB in 1967, or the Iraqi WMD capabilities in 2002-03, was paid for in thousands of American lives and many more foreign citizens’ deaths. The importance of this subject justifies further studies like Fixing the Facts.
It is essential to acknowledge at the outset of this review that Fixing the Facts is a thoughtful and valuable contribution to how we discuss and think about intelligence, or at least intelligence in the U.S. context. Having spent much of the last decade engaged in a comprehensive comparative analysis of British and American intelligence I think it is important to stress that the way intelligence works in the United States is highly idiosyncratic and very few of the conventional concerns and conceptions of intelligence treated as necessary and inherent in U.S. thought are transferable to other national systems. With this in mind I believe that the most useful contribution I can make to this symposium is to comment in some detail on the books’s comparative case study, its execution and the implications of its execution for Rovner’s wider argument.
That being said, there are aspects of Fixing the Facts’s discussion of the U.S. system with which I might take some issue. Rovner opens the volume with the assertion that “Politicization is the most significant problem in intelligence-policy relations and it deserves special attention” (p.5). That it might well have been in the post-Iraq miasma of the last decade. But politicisation did not figure in the substantial dysfunctions that contributed to America’s vulnerability to the 9/11 attacks. To be sure, the book demonstrates a better understanding of the interagency process underlying the production of U.S. National Intelligence Estimates (a machinery often as weakly understood by Americans as outside commentators, many of whom tended, erroneously, to attribute the Estimates to CIA). This gives the volume’s discussion of cases like Vietnam and the Team A/Team B experiment a strong basic foundation on which to build what is a generally well-considered analysis. However, one occasionally notes minor errors of fine detail such as the claim that National Intelligence Council replaced the Board of National Estimates (BNE) in 1973 (130). While Director of Central Intelligence William Colby established a National Intelligence Office (NIO) in 1973, manned by a suite of National Intelligence Officers, they were not constituted as the National Intelligence Council until December 1979, three years after Team A/Team B. More significantly, while the increased closeness of the National Intelligence Officers to policy-makers is acknowledged in the book there is no recognition of the fact that this transition prompted widespread concerns that such proximity would increase the risk of politicisation (an acute enough concern to prompt the angry resignation of then-BNE chair John Huizenga). Indeed, one could quite reasonably argue that the events of 2002 were a manifest demonstration that Huizenga’s concerns were entirely well-founded.
Unfortunately, this Fixing the Facts’ presentation of intelligence processes outside the author’s native political arena is far weaker. There can be no doubt that the book demonstrates a fundamental failure to understand how the UK’s intelligence system functions and its relationship to the policy processes in Whitehall and Downing Street. It is, for example, asserted that “Unlike their British counterparts, U.S. intelligence agencies are purposefully removed from the policy process” and that “This kind of separation is nonexistent in the United Kingdom where government ministers have more input in estimates” (p.140). Both assertions about the UK are profoundly inaccurate. It would be grossly inappropriate (as well as almost prohibitively difficult) for ministers of the Crown – the only constitutionally recognised policy makers in the British model of government – to have an input in the intelligence assessment process. Indeed, it would be no less a transgression for ministers to influence the policy analysis process as well – an assertion no doubt all but incomprehensible to a reader from the United States. To be sure, as Fixing the Facts states, “ministerial representatives as well as intelligence chiefs” sit on the UK’s national assessments body the Joint Intelligence Committee (not “Council” as Rovner calls it) (p.140 infra). But those officials represent their ministries and not their ministers. In the British system of government that distinction is both basic and vital.
This failure to understand the British system is common amongst U.S. observers, largely because of they do not understand the mechanisms and implications of the UK-model independent Civil Service – a model of governance shared with Canada, Australia and New Zealand and on which the League of Nations and United Nations secretariats were consciously modelled. In the study of intelligence this misunderstanding can be attributed in a large part to an unfortunate item of drafting by Michael Herman, quoted religiously by many U.S. observers as well as in this volume, to the effect that the UK’s top-level intelligence coordination and assessment body called Joint Intelligence Committee (JIC) includes “policy makers” and “policy people” (the phrasing recurs with some variation in a number of his pieces, most notably his 1996 Intelligence Power in Peace and War.) This difficulty arose largely because Herman was, understandably, trying to find a common conceptual language to talk about the U.S. and UK systems at a time when discussion of the British system was far less developed than it is today. Unfortunately the U.S. parlance adopted by Herman implies assumptions peculiar to the U.S. ‘spoils-and-churn’ system of appointed public officialdom that the UK does not share.
Amongst U.S. commentators perhaps Richard Neustadt truly came closest to articulating the contrasts between British Civil Service system and the U.S. model of a largely appointive and hence heavily partisan public service. The British Civil Service is a permanent cadre of officials, promoted internally through an internal process which admits of political contact only at the point of the Prime Minister formally approving the topmost appointments at the level of Permanent Undersecretary. And the Prime Minister would be seen as infringing the autonomy of the Civil Service were he or she to challenge the recommendations of its internal appointment process (as Margaret Thatcher briefly and largely unsuccessfully attempted in the 1980s). The Prime Minister really only appoints two echelons of officials on his or her own recognisance and judgement: Cabinet ministers and junior departmental ministers. This amounts to around 35 individuals, in contrast with the U.S. President who appoints some 3500 people, many of whom bring with them personal appointees eventually occupying the top four or five public service echelons beneath immediate Cabinet and sub-Cabinet appointments.
The function of the Civil Service is both to execute ministerial decisions and to advise and inform ministers to assist them in their decision-making process. The Civil Service’s ‘advice to ministers’ consists of surveying options and offering evaluations of potential costs and benefits but they are formally prohibited from ‘taking a view’ on those options. Civil Servants may manage the implementation of policy but they are not entitled to ‘make’ it. Consequently, even when offering ‘advice’ on policy the role of Civil Servants rests on a notion and degree of impartiality and independence that in the United States is reserved solely for the intelligence community. Thus the distinction between policy-makers and non-partisan advice is exactly the same as in the United States but occurs at a different level of the machinery of government. It is in this context, therefore, that Reginald Hibbert (also quoted in Fixing the Facts) could observe that the Foreign Office is ‘a huge assessment machine.’ Assessment is the essential role of the Civil Service – but not advocacy. Within this structure, the role of the JIC is to provide a common knowledge base for interdepartmental policy discussions directed towards generating Civil Service advice to ministers, and most particularly where that knowledge base depends in some significant part on information clandestinely acquired. The representatives of the policy departments on the JIC are, therefore, speaking as non-partisan analytical voices and as conduits of departmental sources of open and confidential information. This is what Herman actually meant with his notion of “government assessment.”
Thus, in the UK, political appointees are at least two stages removed from the intelligence process. Intelligence in Britain is actually more rigorously separated from policy-makers and their influence than in the United States where the immediate working-level departmental points of contact for the intelligence community are appointed, politically partisan, policy advocates. This undermines the assertion in Fixing the Facts that ‘If organizational design matters, then politicization should be more common in the United Kingdom, and it should have been much more intense in the run-up to the war in Iraq’ (140 infra). First, if organizational design matters then politicization should be less common in the UK system than in the U.S. (and there is good reason to believe this is so, some of which appears in Fixing the Facts as part of the case against the (misconstrued) organisational approach). And from that it would follow that any politicization prior to Iraq would not only be more delayed but would – more importantly – require a substantial compromise of the much more fundamental and comprehensive independence of the Civil Service and especially its interdepartmental coordinating mechanisms in the Cabinet Office.
Indeed, something that Fixing the Facts fails to adequately acknowledge is that the UK’s intelligence process as such was not significantly politicized. Much of the discussion (including in this book) of UK intelligence on Iraq has hinged to a greater or lesser degree since the invasion on a Cabinet Office publication Iraqi Weapons of Mass Destruction: the Assessment of the British Government,better known as the ‘September Dossier’. To be sure, the September Dossier was a politicised product, but was a public information document and not a formal intelligence community appreciation. While based on August and September JIC assessments, it was drafted separately from them. Nor either was it subject to the de rigueur challenge and review processes of the JIC’s substantive subcommittees or Current Intelligence Groups. The assertion in this book that that there was a hardening of the UK assessment view on Iraqi non-conventional weapons in the latter half of 2002 is accurate as far as it goes. But the acknowledged reasons for this are not properly examined. The principal independent inquiry after the Iraq war was led by retired Cabinet Secretary Robin Butler (Lord Butler of Brockwell). Lord Butler argued that the main factor driving any hardening of analytic lines in the Autumn was the JIC’s production of worst-case oriented papers for military contingency planning in August 2002. His conclusion was that strength of the JIC’s judgement calls in the 9 September assessment of Iraq – on which the September Dossier was largely drawn – arose from Assessment Staff drafters carrying the worst-case thinking of the August papers forward to the September paper. But the 9 September assessment was, as Butler also notes, scrupulously caveated to indicate the lack of direct evidence. The problem, then, was of analytical cross-contamination rather than politicization as such. Indeed, the book goes to some length to avoid reckoning directly with Butler’s case against a diagnosis of politicisation, relegating this significant item of substance to a footnote (239n7).
This is not to say that there was no compromise of analysis and assessments at some level. One of Butler’s most significant findings – which Fixing the Facts entirely overlooks – is Prime Minister Tony Blair’s refusal to convene the Cabinet Defence and Overseas Policy Committee (DOPC) and the supporting Civil Service machinery of the Overseas Policy and Defence Official Committee (OPD(O)) and the OPD(O) secretariat, at that time called the Overseas and Defence Secretariat (ODS). Prior to Prime Minister David Cameron establishing the UK National Security Council, DOPC was the counterpart to the U.S. National Security Council in London, chaired by the Prime Minister and made up mainly of the Foreign, Defence and Home Secretaries and the Chancellor of the Exchequer (Cameron’s NSC is essentially an expanded DOPC that meets weekly rather than as-required). This was, therefore, the direct equivalent of an American President going to war without convening the NSC or getting the NSC Secretariat to work through the policy issues and implications. As a result, Butler notes “excellent quality papers were written by officials but these were not discussed in Cabinet or in Cabinet Committee.” One of these papers, the British electorate would eventually learn, was a dire warning of the impending consequences of invading Iraq and the problems that pacification and reconstruction were likely to (and eventually did) entail.
Consequently, while Fixing the Facts’ overarching thesis that internal politics and the exigencies of persuasion drive politicisation may be true, it is also trivially true. That politicians exaggerate, and use the information sources available to them opportunistically and disingenuously, may be — as the old Soviet-era Russian joke runs — Pravda (truth), but it is not Izvestia (news). And that making the intelligence process public increases the vulnerability of that sector of government to such opportunism and disingenuousness is also well-trodden ground. As Anthony Glees and I observed in 2004 when the intelligence on Iraq was still in the headlines, openness is not always a virtue and secrecy not always a vice. Drawing to a close, Fixing the Facts concludes bleakly that the only real protection against the politicisation of intelligence assessment is old fashioned secrecy (203). But this achieves no more than to echo Ken Robertson’s prophetical warning a quarter of a century ago that “Bringing intelligence into current political debates has the effect of ‘politicising’ the intelligence services in a way that is undesirable … subjecting them to pressures from political actors to produce results or take certain views”. And so, in the last analysis, the fact that intelligence publicised becomes intelligence politicised is also truth but hardly news.
This is Joshua Rovner’s first book and is based upon his Ph.D. thesis. Fixing the Facts considers what is, at first glance, a relatively well-trodden subject: the politics of intelligence. In fact, the book’s title neatly summarises Rovner’s starting position – there is something wrong with intelligence, particularly in its interface with policy. Fixing the Facts therefore attempts to look at why this is the case and what might be done to redress it.
At face value the book is all about the politicisation of intelligence; in other words, how “policy makers incorporate intelligence into the decision-making process” (p.3). Rovner employs a nice mixture of theoretical and empirical approaches to the subject, and the book works well through its combination of the two. In an appendix Rovner lists a huge number of possible reasons for politicisation to occur, what he calls “pathologies of intelligence-policy relations” (p.205). Rovner’s starting assumption is that politicisation is really an inevitable by-product of the interaction: if you want policy-makers to use intelligence, and there’s no point for its existence otherwise, then that policy involvement will inevitably lead to a degree of politicisation. The key is in managing and minimising this.
To solve this perennial problem Rovner suggests a number of possible solutions. The first part of the answer lies in the question of proximity: how close should the policy-makers be to the intelligence community? This has been an enduring debate, beginning with Sherman Kent’s book Strategic Intelligence for American World Policy in 1949 and the review by Willmoore Kendall the same year. Rovner’s thesis is based upon “domestic politics” (p.11) and is predicated on two conditions: i) that policy-makers are tempted use intelligence as a public justification for their actions and it is therefore intimately connected to the fate of the policy; ii) that opposition to a policy makes it more likely that intelligence will be used in this way. Both conditions are understandable, but they are very much dependent on a particular system of government – the U.S. model.
Rovner never really defines what he means by “policy-maker.” Although this might sound like an academic and rather tedious point, it is highly relevant because the book purports to be applicable to both U.S. and UK governments; indeed, of the four detailed case-studies in the book, one refers directly to the UK. Furthermore, Rovner states in his conclusion that “this book has surveyed almost all of the prominent case of politicisation in the United States and the United Kingdom over the last forty years” (p.189), certainly a very bold claim. Whilst his case-studies and the way in which he applies his theoretical model to the US system work well, his arguments are less well developed with respect to and less well borne out by the British system.
There is, of course, a whole raft of differences between the UK and U.S. systems, but it is necessary to highlight a few salient ones, especially as they serve to undermine some of Rovner’s claims. For starters, the heads of the intelligence agencies in the UK are not political appointees, as they often are in the US. From the outset, then, this has direct connotations on the question of proximity between the two camps. Furthermore, the British system is based on consensus: politicians only receive a unified intelligence assessment; the high-level U.S. equivalent has a dissent function, which historically has been well used. The implications are clear: it is simpler in the U.S. for intelligence to be used as a justification for a policy decision when there is a greater array of judgments on offer.
Similarly, the term “policy-maker” itself is troublesome: in the U.S. context it is clear from Rovner’s book that he is referring to officials in the Administration; in the UK it can refer to elected politicians or, more often, civil servants in policy-making departments. The difference is crucial as the former decide policy and are political; whereas the latter deliver policy and are apolitical. Their interaction with the intelligence community is therefore paramount. This has contemporary salience too. Historically high-level intelligence in the United Kingdom has taken place in the Joint Intelligence Committee: a forum where intelligence officials and policy-making civil servants sign up to agree assessments. In the new National Security Council, intelligence officials rub up against senior politicians, often without any civil servants from policy departments being present. Politicisation then, is not a simple question of having two entities: ‘intelligence’ and ‘policy’.
The solution Rovner offers to dealing with the problem is secrecy. If there is more secrecy then governments will work better, decisions won’t need to be based on the public use of intelligence, assessments won’t make headline news, and policies will improve. There is nothing wrong with this, but it does beg the question ‘did policy work better in the halcyon days of secrecy’? The simple is answer is probably not. In the wake of the invasion of Iraq and the questions that have arisen over intelligence since, is it always the “powerful tool” that Rovner believes it to be (p.48)? Probably the answer is, by and large, yes, but surely not for everyone. Similarly, in a country like the UK where, Iraq notwithstanding, intelligence is far more secret, has politicisation been better controlled? The simple answer is that it’s not so simple.
The 1927 Arcos affair is an interesting test for another of Rovner’s points: that politicisation is almost always bad (p.192). Here was a law enforcement raid on a Soviet trade delegation that resulted in a brief lull in Anglo-Soviet diplomatic relations. To publicly justify the raid, the Prime Minister read out a series of intercepted Soviet cyphers in Parliament. Thus, the lack of secrecy led the Russians to alter their codes and the British lost the ability to read them. Politicisation was therefore detrimental, just as it had been three years earlier with the government’s publication of the Zinoviev letter. Yet, there are other examples where politicisation can be useful: the 1971 expulsion of Soviet intelligence officials from the UK, Operation Foot, was a success and dented Russian intelligence capabilities for years to come. To see the use/non-use of secrecy as the precursor for success is therefore too one-dimensional.
There are several other aspects that warrant further consideration: methodologically it is far easier to spot politicisation than it is to spot the successful use of intelligence, but does this bias our impression of the severity of politicisation? There is a difference between the various types of intelligence product – strategic or tactical assessments; and predictive versus explanatory intelligence – what are the implications for politicisation? What about examples of good policy that are, admittedly, based on politicisation? This leads to a final question about context: is evidence of politicisation more important than the final policy?
Fixing the Facts is a fantastic addition to the growing literature on the intelligence-policy relationship. It advances the field in our theoretical understanding of politicisation, and the very useful combination with empirical case-studies serves to emphasise Rovner’s conclusions. Its timing is apt with the increasingly public use made of intelligence in the twenty-first century. The case-studies he presents are well researched, useful, and necessary to add flesh to his theoretical sections. Rovner certainly advances the field and his prophetic words need to be understood by anyone thinking about using intelligence publicly to justify policy. What the preceding few paragraphs imply is not that the book is wrong or bad, just that its theoretical examination is not universally applicable to all governments and systems, and that it would have been useful to emphasise from the outset that it was predicated on the American system of government. A final word, and one that Rovner would surely support, belongs to a former British Prime Minister who said something along the lines of “the point of a secret service is that it is secret.” Quite.
The troubled nature of the relationship between policy-makers and intelligence analysts, and the existence of political manipulation of intelligence products, is unfortunately something very familiar to scholars of intelligence and those who work in the intelligence community and government. The Bush administration’s push for war against Iraq in 2002 and 2003 in this regard presents nothing but the most recent and vivid example of a blatant case of politicization. Despite the familiarity of the subject matter, Joshua Rovner has nevertheless provided us with Fixing the Facts, an important and refreshing book that will be useful to all students of international security, especially those interested in foreign policy and intelligence analysis.
Fixing the Facts greatly contributes to our understanding of this phenomenon because it is the first study that not only provides a solid theoretical framework for understanding the political rationale of politicization, but also is a rigorous and rich historical work highlighting different instances and types of politicization.It covers several case-studies that include the Johnson Administration’s manipulation of Vietnam War estimates, the Nixon administration’s pressure to change the National Intelligence Estimates (NIEs) so to justify antiballistic missile programs, Team B’s ideologically motivated reading of Soviet capabilities and intentions during the final months of the Ford administration, and the British and American politicization of intelligence leading up to the recent Iraq War. It is the most systematic treatment of politicization of intelligence thus far. And it is a very strong one. As such, it significantly advances the literature on intelligence-policy relations, which still has much ground to make up relative to the civil-military relations literature.
In the rest of this review, I first summarize the book’s argument. Second, I highlight several points pertaining to its theoretical framework and methodology that merit special attention in any study of such phenomenon as politicization. Finally, I focus on some gaps in the empirics, which I believe, deserve more scrutiny.
To begin, Rovner’s claim is that bureaucratic factors and individual and organizational proximity between policy-makers and intelligence officials cannot adequately explain instances of politicization. Rather, politicization can be best understood through the prism of domestic politics. Describing an oversell model of politicization, he writes, “…domestic political pressures create incentives for policymakers to oversell the amount and quality of information on security threats, regardless of the nature of personal relationships or organizational designs”(11). In brief, he contends that when a leader makes a public commitment to a policy that is opposed by a ‘critical constituency,’ then politicization of intelligence is needed to signal to audiences that the defense establishment supports the leader’s policy. Thus, a leader’s public commitment and the presence of a critical constituency are the theory’s two independent variables affecting both the presence and degree of politicization (e.g., indirect versus direct politicization). (36)
The straightforwardness of the two independent variables in this ‘oversell model’ indeed makes the theoretical framework parsimonious and elegant. However, as I note later in this review, some of the case studies Rovner discusses – the Team B episode in particular – are not explained very well by these two independent variables. At times, the relationship between the independent and dependent variables appears to be over-simplified. Indeed, it is not entirely clear why we should expect ‘political cost’ to be the only intervening variable affecting the level and magnitude of politicization. It seems to me that while Rovner is correct to focus on the consumer side of the politicization equation (i.e., the policy-maker), the measures needed to manipulate intelligence also highly depend on factors pertaining to the intelligence community. For example, if the intelligence community is already divided regarding a question that is pertaining, say, to the adversary’s intentions, policy-makers may only need to use indirect measures or subtle signals in order to politicize intentions, even in the face of a strong public commitment. Similarly, the focus on public commitments seems to be too narrow, not just because policy-makers are at times forced to publicly commit to policies that are not consistent with their own policy preferences, but also because at times there are [reputational, political, and strategic] costs inherent in manipulating intelligence that affect both the willingness and ability of policy-makers to engage in politicization. The question then becomes whether and to what extent there are additional intervening variables or causal mechanisms that influence patterns of politicization found in the empirical chapters.
Another point that is related to the book’s main claim is the ability of the study to rule out alternative rationales for politicization. To be fair, the study identifies two main competing explanations for the occurrence of politicization: individual/organizational proximity and organizational design. Rovner nicely shows that politicization occurred in the face of a close relationship between the president and the Director of Central Intelligence [DCI], as well as in cases where the DCI interacted infrequently with the president. Rovner is also able to effectively argue with the help of the matching cases of politicization of intelligence in Britain and the United States in the period leading to the war in Iraq that the design of intelligence organizations does not affect whether politicization occurs. Those who believe in the importance of these variables may counter, however, that these factors would not prevent politicization from happening, but that they would affect how often politicization occurs as well as its pervasiveness. Of course, examining this proposition is outside the scope of this book: such an exercise would require a more thorough analysis of the frequency and magnitude of politicization across time and country. Nevertheless, this study is not able at this point at least to completely rule out the significance of these variables in affecting the overall likelihood or frequency of politicization.
Another alternative mechanism that Fixing the Facts surprisingly does not discuss at length pertains to a psychological explanation for politicization. Indeed, scholars of foreign policy decision-making and even those writing on the specific topic of intelligence failures have long noted the importance of motivated biases in explaining both why policy-makers would try to interfere, influence, or have a biased interpretation of intelligence, as well as why intelligence officials would revisit and at times radically change their estimates in the face of a change in policy. According to this alternative interpretation, policy-makers’ commitment to a particular policy course would drive them to seek information that supports their policy not because of domestic political concerns, but rather due to their need to avoid stress and value trade-offs. Similarly, one can make the argument that intelligence officials would change estimates on the basis of a similar need to avoid a value trade-off that would arise if and when their estimates undermine their country’s policy. The psychological mechanism is inherently and significantly different from the domestic politics mechanism that Rovner suggests. Nevertheless, to empirically distinguish between the two is not an easy task. Despite relatively good access to primary documents and intelligence reports for scholars, it is still not entirely clear what to make of the fact that the British intelligence estimates of the Nazi threat became more optimistic around the same time that the Chamberlain cabinet saw war with Germany as inevitable. One could claim that this is a case of politicization for domestic political reasons, or an example of a motivated bias that affected British intelligence’s reading of the balance of power in the wake of a policy shift, or merely a case of intelligence learning and updating that simply happened to coincide with Chamberlain’s changing views and policy.
To be sure, Fixing the Facts, does advance in some instances such strong evidence in support of the domestic political explanation for politicization that engaging in the plausibility of alternative explanations seems unnecessary. But in some other instances, in my view, alternative explanations could have been discussed in a more straightforward manner. For example, it is not always clear what role motivated bias played in producing the estimates on Iraq, or how a reading of the evidence through the prism of motivated bias would even look in this case. The evidence for the Nixon administration’s manipulation of the Soviet estimates is strong, but the claim could have been more convincing had the author engaged more significantly with alternative explanations as to why the Soviet estimates changed (as he nicely does in the Johnson Administration’s case) and how long after the occurrence of politicization the estimates remained biased. If Rovner’s explanation for politicization were correct, then we should perhaps see the intelligence community reverting back to its original conclusions following a change in political leadership, when a leader’s commitment to a policy diminishes, or when a critical constituency that opposes that policy becomes less popular. Looking at this time dimension could have served as another way of testing the theory’s causal mechanism.
Moving to other cases, in his discussion of the exercise in competitive analysis in 1976, also known as Team B episode, Rovner does offer a plausible alternative explanation for the Ford administration’s indirect manipulation of intelligence. But it is also a case where the evidence does not fully support the book’s theoretical claim. As Rovner claims, the Team B episode shows the limits of the ‘oversell model.’ A different reading of this case might suggest that it does not support the book’s theory at all. Indeed, such a high value on both of the theory’s independent variables should have resulted in the direct manipulation of intelligence. Instead, one might question whether this was even a real attempt to indirectly politicize intelligence and the extent to which intelligence estimates were even really politicized as a result of this exercise. Though I believe that Rovner’s treatment of this case is solid and his conclusions are correct, this case study unfortunately raises considerable ambiguity about the scope conditions of the theory. In the Team B case, it is the role of the president’s private preferences that appear to matter, but as I noted earlier in the review, one can think of additional variables that can significantly affect the occurrence and type of politicization. Further, the Team B episode also highlights the problem of how to code politicization. President Ford could have easily come up with a much more sophisticated way to manipulate intelligence had he wished to do so. Instead, his choice to put in charge a group of a right-wing ideologically motivated people showed very poor judgment on his part. Perhaps it was an attempt to appease domestic opposition to détente. Whatever the reason behind Team B, the extent to which and in what way exactly this exercise actually influenced the intelligence community’s conclusions about Soviet capabilities and intentions in the NIEs is debatable. The National Intelligence Officer on Soviet strategic weapons and the principle drafter of NIE 11-3/8-76, Howard Stoertz, for instance, claimed that the Team B report had only a subtle effect on that and subsequent NIEs, as it merely made the intelligence community more prudent in its assessments of Soviet capabilities and intentions that were not based on hard evidence. Further, subsequent NIEs on the Soviet threat were perhaps more hawkish in nature, but the Team B exercise was not the most obvious reason for that change.
Finally, the Team B episode also leads one to wonder about additional cases where we would have expected to observe politicization but have not found evidence for it. Of course, the known universe of cases of politicization is bound to be misleading. After all, severe cases of politicization might leave little paper trails behind because they would prove too embarrassing or costly for either policy-makers or intelligence officials to admit to having done so even after the fact. Nevertheless, it is interesting to note that one possible case to examine would be why President Ronald Reagan and his closest advisors did not attempt to manipulate political intelligence on the Soviet threat during the second half of the 1980s, and of Gorbachev’s intentions in particular. The apparent lack of politicization is puzzling given Reagan’s commitment to deepening cooperation with the Soviet leader and his desire to engage in controversial arms control treaties. While Reagan’s commitment was public and firm, and a critical constituency of conservative hawks challenging Reagan’s decision certainly existed (his own Secretary of Defense, Casper Weinberger was even one of them), the NIEs on the Soviet Union continued to contradict Reagan’s public rhetoric on Gorbachev and the Soviet threat. George Shultz and other senior members in Reagan’s administration were increasingly unsatisfied with the NIEs’ portrayal of Gorbachev’s intentions. At numerous times they voiced their frustrations before Robert Gates, then the Deputy Director of the CIA. Yet neither Reagan nor his advisors attempted to manipulate intelligence directly or indirectly during this period. If anything, CIA officials later complained that Gates was manipulating intelligence in the opposite direction, thereby forcing the community to present more hawkish interpretation of Soviet capabilities and intentions. Why? Why did Ronald Reagan not wish to signal to his opponents that the intelligence establishment was behind him and was supporting his arms control policy and overall accommodating stance towards the Soviet Union and its leadership? What should we make of cases where senior officials in the intelligence community were manipulating intelligence on major policy issues in a manner that contradicted the president’s policy and beliefs? One can think of several explanations for the absence of politicization by senior decision-makers during this time period, but I am not convinced (yet) that an explanation based on domestic political concerns would be the most powerful one in this case.
In sum, the points I raised would have perhaps made the book’s claims more convincing to skeptics. However, it is important to note that the book offers a very strong argument about the political roots of politicization. The observable implications of the theory are supported by the empirical record as Rovner, in many instances, is able to offer both correlational and process-tracing evidence that confirms his argument. Anyone who writes about intelligence knows how rare it is to find that kind of evidence. For all these reasons, Fixing the Facts is a must-read for anyone who is interested in understanding policy-intelligence relations. It is a great example of how political scientists should approach the study of intelligence. The outcome is a fantastic book that is a pleasure to read.
Most intelligence scholars focus on how intelligence agencies do business. Historians and political scientists have written countless studies about how intelligence works and why it fails. Far less is known about how policymakers use it. This is both surprising and problematic because even the best intelligence will be irrelevant if it does not have a receptive audience. Intelligence is less important than intelligence-policy relations.
Fixing the Facts examines the interaction between political leaders and the intelligence officials who serve them. Ideally, leaders will accept intelligence at face value without swallowing it whole, and let intelligence inform the policy debate without letting it dictate policy decisions. Healthy intelligence-policy relations are difficult to sustain, however, and in some cases leaders ignore intelligence or try to shape it to their own satisfaction. Both are serious problems. Ignoring intelligence removes an important check on leaders’ preexisting beliefs and preferences, making them vulnerable to strategic tunnel vision. Politicizing intelligence skews the quality of estimates by forcing intelligence to deliver certain conclusions even when the underlying data is partial or ambiguous. It also inhibits reassessment because altering such conclusions would constitute a very public admission of failure. Finally, episodes of politicization damage intelligence-policy relations for years after the fact.
The book provides a taxonomy of the problems that plague intelligence-policy relations and a theory of politicization. It tests the theory in six cases arranged in three paired comparisons. The first pair explains why the Johnson administration ignored intelligence on Vietnam in 1964 but politicized it in 1967. The second pair explains why both President Nixon and President Ford politicized estimates on Soviet nuclear forces, even though they came into office with very different views about intelligence. The final two cases compare the experiences of the United States and the United Kingdom before the war in Iraq in 2003. Despite the fact that the organizational structure of intelligence and policy was very different in Washington and London, the pattern of politicization was almost exactly the same.
None of the reviewers disagree with my conclusions about Vietnam, the Nixon administration, or American intelligence before the war in Iraq. Keren Yarhi-Milo offers some criticism of my treatment of the Ford administration, though it is not entirely clear that we differ on the important aspects of that case. Michael Goodman and Philip Davies both criticize my description of the British system. In addition to these issues, which I discuss below, the reviewers raise four important questions.
What is politicization?
The word politicization is pervasive in the literature on intelligence, but it is rarely defined with any precision. Instead, politicization seems to include any overlap between the worlds of policy, politics, and intelligence. When I began writing Fixing the Facts, I sought a clearer definition in order to get analytical purchase on the issue of intelligence-policy relations. My definition of politicization is straightforward: the manipulation of intelligence to reflect policy preferences (p. 29). By focusing on how policymakers respond to new estimates, the book creates a framework for comparative studies by asking why these leaders alternately accept, ignore, or politicize intelligence.
Two of the participants in the roundtable imply that my definition is too narrow. Bar-Joseph discusses several cases in which intelligence officers “conspired against their bosses,” suggesting that these cases somehow undermine my theory of politicization. They do not. The issue of intelligence conspiracies is certainly important, and Bar-Joseph has written a terrific book on the topic. But it is beyond the scope of Fixing the Facts, which asks how policymakers use or misuse intelligence, not why intelligence officials conspire against their civilian masters.
Michael Goodman also muddies the water when he asserts that my starting point is the assumption that politicization is inevitable whenever policymakers and intelligence come together. I make no such claim. Instead, I explicitly write that policymakers can accept, ignore, or politicize estimates (pp 18-19). Sometimes policymakers accept intelligence and sometimes they ignore it. Politicization is not inevitable. Goodman also argues that cases like the Arcos affair are important tests of my arguments about politicization. But the Arcos affair had nothing to do with politicization. In that case, British officials publicized information they had gathered about the Soviet Union. They did not try to manipulate analysts’ views about what that information meant. Similarly, the expulsion of Soviet intelligence officials in the Cold War may or may not have been justified, but it is not an example of politicization. Intelligence conspiracies, intelligence revelations, and expulsions of intelligence officials: none are cases of politicization. The fact that perceptive scholars like Bar-Joseph and Goodman describe them as such probably says something about the concept itself. For whatever reason, even longtime observers tend to cry politicization whenever the world of intelligence merges with the world of policy. There are many ways in which intelligence interacts with policy, of course, but to call them politicization makes the term so broad that it becomes meaningless. All of this highlights need to be precise about operational definitions. A better understanding of intelligence-policy relations depends on it.
Does the theory travel?
Fixing the Facts develops a theory of politicization based on domestic politics. Briefly, it argues that leaders are likely to manipulate intelligence estimates when they publicly commit to controversial policies. In these cases they will be tempted to use intelligence to overcome domestic skepticism. Intelligence is a uniquely powerful public relations vehicle because intelligence agencies control secret information, and policymakers can exploit the aura of secrecy in order to win over critics. But if intelligence does not seem to support policymakers’ claims, then they will have large incentives to pressure intelligence officials to change their conclusions.
Bar-Joseph argues that there are few cases of politicization outside of the United States, which suggests that my theory of politicization stops at the water’s edge. His comments on the Israeli experience are particularly important. In Israel, Bar-Joseph writes, “national intelligence estimates receive a large share of the public agenda… (but) there is no record of an attempt by policymakers to politicize intelligence.” I do not know enough about the history of Israeli intelligence-policy relations to evaluate this claim, and I do not know if enough has been declassified to make a reasonable judgment. But if it is true then it does present an important challenge to Fixing the Facts, and it is extremely important to learn why policymakers avoided the obvious temptation to meddle in estimates in order to win public debates. If the theory cannot describe an important case, we should find out why. More work on the Israeli case might produce deeper insights into the causes of politicization. It might also lead to useful suggestions about how to mitigate the problem.
Bar-Joseph agrees with my conclusion about politicization in the British case, but argues that it this problem has been historically rare. This fact hardly undermines my theory, and the historical absence of politicization in the United Kingdom is not hard to understand. As I argue in the book, politicization is unlikely when governments maintain a strict secrecy regime. On the other hand, politicians are much more likely to meddle in the content and tone of estimates if they believe that intelligence will play a prominent role in the public debate. This was not the case in the United Kingdom until very recently. Indeed, the British government did not even acknowledge the existence of its intelligence agencies until after the Cold War. It is not surprising that politicization did not occur there.
A longstanding view is that close proximity to the policy process will inevitably lead to politicization, whereas political preferences are less likely to influence estimates if intelligence agencies are far removed from the policy fray. Fixing the Facts tests this claim by comparing the pattern of intelligence-policy relations in the United States and the United Kingdom in the period before the war in Iraq. Because the British intelligence community is closer to policymakers than its U.S. counterpart, politicization should have been much more intense throughout the prewar period. But a close analysis of intelligence-policy relations before the war shows that the timing and nature of politicization was virtually identical in each case. Both governments pressured intelligence officials to exaggerate the Iraqi threat, and estimates became much more ominous in both countries in the second half of 2002. Moreover, both President Bush and Prime Minister Blair exploited the persuasive power of secret intelligence in order to win over skeptical domestic audiences in the crucial months before the war. Organizational design had little to do with the outcome. Politics drove politicization.
Philip Davies and Michael Goodman argue that I misrepresent the role of intelligence in the British government. According to Davies, this undermines my claim that the theory explains intelligence-policy relations outside the United States. The book, he writes, “demonstrates a fundamental failure to understand how the UK’s intelligence system functions and its relationship to the policy processes in Whitehall and Downing Street.” He specifically criticizes my contention that British intelligence is organizationally closer to the policy process than in the United States. Davies argues that while British policymakers routinely interact with intelligence officials in the Joint Intelligence Committee, they do so as “non-partisan analytical voices and…conduits of departmental sources of open and confidential information.” Meanwhile, American policymakers, who interact with intelligence officials, are “appointed, politically partisan, policy advocates.” Davies and Goodman also argue that the independence of the civil service makes it a buffer between government ministers and intelligence officials. These officials can safely interact with intelligence officers because they are in the business of assessments rather than policy advocacy. Thus intelligence is actually closer to policy in the United States than in Great Britain.
This is unconvincing. As Davies has explained elsewhere, the British model of intelligence-policy relations is deliberately intrusive. It operates on the basis of a pull architecture, in which policymakers choose the targets and priorities of intelligence agencies rather than allowing agencies to act independently and push their views upward. Such a model depends on close and continuing contact between the policymaking and intelligence communities. If this is correct, then British intelligence is indeed closer to policy than in the United States, and a strong civil service ethic is the only thing that prevents the process from becoming corrupted. Davies also argues that U.S. intelligence officers are more likely to interact with partisan political appointees than in the United Kingdom because there are so many appointees in Washington. But many of these appointees – 3500 by his count – take jobs that have nothing to do with national security or intelligence. Davies also ignores the fact that there are many more intelligence analysts in the United States. There are said to be more than 100,000 employees in the intelligence community, most of whom are unlikely to have a political partisan as their first point of contact. On the contrary, they usually interact with career State Department or military officers rather than political staffers.
Davies’s treatment of the Iraq case is puzzling. He acknowledges that the critical intelligence estimates on Iraqi capabilities were politicized, but contents that “the UK’s intelligence process as such was not significantly politicized.” To make this claim he notes that the public dossier on Iraq’s chemical, biological, and nuclear weapons programs was “a public information document and not a formal intelligence community appreciation.” This is a distinction without a difference. Despite the fact that 10 Downing Street was ultimately responsible for the dossier, the document was based on intelligence information and drafted in the JIC. Prime Minister Blair’s introduction was unambiguous:
The document published today is based, in large part, on the work of the Joint Intelligence Committee (JIC). The JIC is at the heart of the British intelligence machinery. It is chaired by the Cabinet Office and made up of the heads of the UK’s three Intelligence and Security Agencies, the Chief of Defence Intelligence, and senior officials from key government departments. For over 60 years the JIC has provided regular assessments to successive Prime Ministers and senior colleagues on a wide range of foreign policy and international security issues. Its work, like the material it analyses, is largely secret.
The first draft of the dossier was finished in June 2002, and each subsequent version painted a more ominous picture of Iraqi capabilities. This shift was largely due to the influence of the Prime Minister’s communications staff, which worked closely with the JIC on editing the dossier. Communications staffers sought to make it vivid and compelling because they knew the dossier was going to be released to the public, and they urged intelligence officials to draw worst-case scenarios from partial and ambiguous information. This was a remarkable intervention, and Blair acknowledged that it was “unprecedented for the Government to publish this kind of document. But in light of the debate about Iraq and Weapons of Mass Destruction (WMD), I wanted to share with the British public the reasons why I believe this issue to be a current and serious threat to the UK national interest.” The Prime Minister faced sustained public opposition and deep doubts from his own party, and took extraordinary steps to justify a war of regime change. The most important step was manipulating intelligence in order to oversell the threat posed by Iraq.
The postwar Butler Report came to a different conclusion, and Davies complains that my book “goes to some length to avoid reckoning directly with Butler’s case against a diagnosis of politicisation.” I don’t know what to make of this criticism, given that chapter 7 is basically an extended refutation of Butler’s finding. Moreover, I note in the book that the Butler report contradicts itself by recommending that the JIC chief be someone “who is demonstrably beyond influence” (p. 239n). If political pressure was not important then this recommendation should not have been necessary. Indeed, the entrenched civil service norms that Davies discusses should have been enough to prevent the politicization of intelligence.
In any case, while official inquiries like the Butler Report often yield very useful information, their conclusions are often dubious. Scholars may appreciate the work of commissioners without accepting their analyses.
Are there better explanations?
Some of the participants in this roundtable criticize Fixing the Facts for ignoring other explanations for the causes of politicization. Bar-Joseph argues that one alternative path to politicization is appointing like-minded intelligence chiefs, and he reminds us that President Reagan chose his own campaign manager as Director of Central Intelligence. I agree that this is an important form of politicization, which I describe as “manipulation-by-appointment” in the book (pp. 30-31), but this is not inconsistent with my theory. Indeed, if policymakers expect to use intelligence to justify their policies, they will be more likely to appoint friendly or malleable intelligence chiefs. It is also worth noting that policymakers occasionally appoint intelligence officials for the opposite reason. President Kennedy, for example, replaced Allen Dulles with the conservative John McCone partly as a sop to the right. The fact that McCone was unlikely to share his views made him politically useful.
Bar-Joseph also suggests that personality characteristics may condition policymakers’ receptivity to new intelligence. President Eisenhower, he notes, took intelligence seriously because of his memories of Pearl Harbor and his fear of suffering another surprise attack during his presidency. His sense of “responsibility and decency” later caused him to withhold intelligence about Soviet weaknesses that might have helped Vice President Nixon fend off attacks from Democrats during the 1960 presidential campaign. The notion that personal traits might make politicization more likely is certainly plausible, and I test it in the chapters on Presidents Johnson, Nixon, and Ford. Johnson’s unique traits cannot explain why he first ignored and later politicized intelligence on the Vietnam War. His personality was a constant, but his behavior was not. Nixon’s disdain for intelligence – which partly stemmed from his own suspicion that the CIA had conspired against him in 1960 – should have led him to ignore intelligence rather than manipulate its findings. Yet he ended up politicizing intelligence on the Soviet Union. Ford was more amenable to intelligence, yet he also ended up politicizing the Soviet estimate. The chapter on Ford also considers whether the personality characteristics of top intelligence officials make intelligence more or less likely, but the results are indeterminate. In sum, while it is possible that personality traits affect the nature of intelligence-policy relations, domestic politics provide a much more comprehensive explanation about the causes of politicization.
A third alternative is that politicization is less likely when intelligence agencies are held in high esteem. According to Bar-Joseph, attempts to manipulate intelligence will probably fail if policymakers come up against a corps of analysts confident in their capabilities and committed to professional norms of independence and objectivity. In these cases policymakers might be deterred from even trying to politicize intelligence. Fixing the Facts does test a variant of this claim, which I call the organizational dependence hypothesis (pp. 10-11). The argument here is that policymakers will target intelligence agencies that require patronage or top cover. Agencies may be structurally vulnerable (e.g. military intelligence organizations are beholden to their broader services) or politically vulnerable (e.g. agencies that have been publicly embarrassed might look for policy protection in order to avoid the budgetary axe). Presumably policymakers are able to spot the difference and will only try to politicize agencies with bureaucratic reasons to bow to pressure.
The chapter on the Ford Administration and the Team B affair offers evidence that cuts in the other direction. Hardline critics of the intelligence community had been arguing for years that national estimates of the Soviet Union were far too sanguine, and they demanded that an external “B” team be given the opportunity to review classified intelligence and write a parallel estimate. Some of Ford’s advisors warned the president that the proposed exercise was a thinly veiled attempt to force the intelligence community to shift to the right. Ford chose not to launch the Team B exercise in 1975, however, when the intelligence community was at its most vulnerable point and reeling from dual congressional investigations. He reversed course the next year, however, when his own political future was placed at risk from hawkish Republicans. The intelligence community had survived its moment of maximum weakness, but now the president’s fortunes were in danger because of shifting political currents. It was only at this point that he authorized Team B, which suggests that domestic politics were much more important than bureaucratic factors. Nonetheless, this case does not rule out Bar-Joseph’s hypothesis, and it is worth exploring further. It is also worth considering the opposite: perhaps policymakers are more inclined to politicize strong and prestigious intelligence agencies because their views will be especially persuasive in public debates.
Finally, Bar-Joseph criticizes the book for ignoring episodes in which national security was on the line. He argues that in these cases intelligence officials will ruthlessly defend their integrity and policymakers will be unlikely to challenge them in the first place. (This logic might explain Bar-Joseph’s contention that Israel, a small state surrounded by hostile rivals, has been historically spared from politicization.) He specifically argues that it misses “the height of the Cold War, when a nuclear confrontation with the USSR was considered likely, the issue at stake was the feasibility of a Soviet nuclear attack on the U.S., or, in other words, the fate of the United States.” While it is true that the book does not explore intelligence-policy relations during the crises over Berlin and Cuba, it is not clear that policymakers were responsible for those crises in their dealings with the intelligence community at the time. Before the Bay of Pigs debacle, for example, policymakers did not try very hard to determine whether intelligence analysts thought that the island was ripe for counter-revolution or about the broader consequences for the Cold War. Moreover, the cases I explore had enormous implications for U.S. national security. The first two cases consider the decision to escalate the war in Vietnam and then to extend a bloody strategy of attrition. The chapter on the Nixon administration is about how the intelligence community interpreted dramatic technological breakthroughs in the Soviet nuclear arsenal. Indeed, critics of the intelligence community worried that the Soviet Union had finally gained the ability to target U.S. nuclear forces, which meant that it might have a real incentive to strike first. The chapter on the Ford administration was also about a battle between intelligence analysts and outside critics who thought that the decade-long Soviet nuclear buildup put the United States in grave danger. And the case on the war in Iraq explores intelligence on the eve of the most consequential and costly foreign policy decision since Vietnam. Perhaps intelligence-policy relations are healthier in the face of an existential threat – and even here the historical record is mixed – but the book shows that politicization is possible even when the stakes are very high.
Yarhi-Milo believes that psychological pressures may lead both policymakers and intelligence officials to manipulate estimates. She suggests that motivated biases and the desire to avoid value trade-offs might compel leaders to meddle in intelligence. But as I discuss in the book, politicization is unlikely on these cases. Policymakers can find psychological comfort either by interpreting intelligence in ways that fit their preexisting beliefs, or they can simply ignore it (pp. 26-29). More plausible is her suggestion that intelligence officials might alter findings that seem to throw cold water on existing policies, especially if those policies have already required significant sacrifice. We might also speculate that intelligence leaders who personally believe in the wisdom of policy decisions might be psychologically incapable of tolerating bad news from their analysts. It is worth noting, however, that some U.S. intelligence agencies submitted very pessimistic analyses well into the wars in Vietnam, Iraq, and Afghanistan. I suspect it is easier for intelligence to overcome motivated bias than policymakers, who are ultimately responsible for the consequences of policy decisions.
Yarhi-Milo also criticizes my discussion of the Team B affair. I agree that the Team B case does not fit the pattern of other cases, and I conclude in the book that my theory does not fully explain why President Ford authorized the exercise. Yarhi-Milo suggests that it does not even qualify as a partial explanation, given that high values on the independent variables in the model predict direct politicization. Yet her criticism is confusing, in part because she concludes that my treatment of the case is “solid and (the) conclusions are correct.” She also agrees with my observation that for peculiar reasons, the president’s private beliefs were more important than his public statements. I argue in the book that these beliefs moderated the political pressures to politicize intelligence. Rather than manipulate the estimate directly, Ford allowed a group of known hawks to participate in the NIE review process, fully aware that this might cause the Soviet estimate to become more ominous. Yarhi-Milo suggests that something else might have been at work, but she does not say what. Nonetheless, I agree that there is probably not a single cause for politicization in this case. More attention to the Team B episode might produce some interesting findings about how policymakers reconcile their public and private views, and how that process affects their views of intelligence.
What comes next for intelligence-policy scholarship?
The reviewers make a number of useful suggestions about intelligence-policy scholarship that go beyond their criticisms of the book. Yarhi-Milo seems to agree that proximity between intelligence and policy does not provide a strong causal mechanism for politicization, but she wonders if it creates permissive conditions that make it more likely. If this is correct, then politicization should occur more often when intelligence is very close to policy, even if proximity itself is not the cause. The opportunities for politicization certainly increase the more that intelligence officials are in touch with their policy counterparts. This is also consistent with the proximity hypothesis, which suggests that politicization may occur subconsciously in these cases as policy biases seep into analysis through a kind of osmosis (pp. 7-8). I am skeptical of these ideas, but I concede that the book does not test them systematically. Doing so, as Yarhi-Milo rightly notes, will require long-term studies measuring the frequency of politicization within and among countries with different traditions of intelligence-policy relations. Such a project is fraught with methodological pitfalls, not the least of which is the problem of coding proximity in a way that makes longitudinal analysis possible. Nonetheless it is surely worth the effort, given the centrality of the proximity hypothesis on the debates about intelligence reform. We should at least determine if proximity is correlated with politicization, whether or not it is the cause.
We also need to know more about the aftermath of intelligence-policy breakdowns. Most of the book explores why relations go bad. It says far less about what happens later. I have found evidence that episodes of politicization poison intelligence-policy relations over the long-term, hardening mutual mistrust for years after the fact. But the long-term is not the same as forever, and it is important to understand the conditions that make rapprochement possible, as well as the ways in which intelligence and policy officials overcome past disputes. In addition, we can do more to investigate the effect of politicization on future estimates. The book notes, somewhat ironically, that intelligence will continue to incorporate policy preferences despite the fact that intelligence officials mistrust policymakers. The reason is that reversing course later would constitute either an admission of past errors or an admission of caving to pressure. Yarhi-Milo, however, suggests that the effects of politicization do not necessarily linger in future estimates. Testing these rival predictions should be possible given the growing number of declassified past estimates. Such a test might also reveal something interesting about organizational learning within intelligence agencies.
Michael Goodman adds his own follow-up questions. Are scholars more likely to spot politicization when policy has failed? Are they inclined to overlook it if policy succeeds? Is politicization more likely on strategic or tactical assessments? Is it more likely when intelligence is explaining the past or predicting the future? I do discuss some of these issues in the book. For instance, I suggest that politicization is more likely on strategic estimates than other kinds of intelligence, both because the stakes are higher and because they deal with inherently ambiguous problems that are open to rival interpretations. Nonetheless these are all good questions; undoubtedly there are more. I am gratified that Fixing the Facts has provoked serious questions about intelligence-policy relations, and hopeful that scholars of intelligence and foreign policy will provide useful answers.
© Copyright 2012-2015 The Authors.
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 United States License.
 Roberta Wohlstetter, Pearl Harbor (Stanford, CA: Stanford University Press, 1962); Barton Whaley, Codeword Barbarossa (Cambridge, MA: MIT Press, 1973); Avi Shlaim, “Failures in National Intelligence Estimates: The Case of the Yom Kippur War,” World Politics, 28, 2 (April 1976): 348-80.
 Richard K. Betts, Enemies of Intelligence: Knowledge & Power in American National Security (New York: Columbia University Press, 2007); Richard H. Immerman, “Intelligence and Strategy: Historicizing Psychology, Policy, and Politics,” Diplomatic History 32 (January 2008): 1-23; Paul R. Pillar, Intelligence and U.S. Foreign Policy: Iraq, 9/11, and Misguided Reform (New York: Columbia University Press, 2011); Robert Jervis, Why Intelligence Fails: Lessons from the Iranian Revolution and the Iraq War (Ithaca, NY: Cornell University Press, 2010).
 Bar-Joseph develops this argument more fully in Intelligence Intervention in the Politics of Democratic States: The United States, Israel, and Britain (University Park, PA: Penn State Press, 1995).
 Rovner argues that “leaders who set out to politicize intelligence usually succeed” (19), though this may be due in part to the fact that leaders only set out to politicize intelligence when they are fairly confident of succeeding. The question of causal impact raises difficult issues in counterfactual methodology. See Jack S. Levy, “Counterfactuals and Case Studies,” in Janet Box-Steffensmeier, Henry Brady, and David Collier (eds.), Oxford Handbook of Political Methodology (New York: Oxford University Press, 2008), 627-44.
 Another possibility is to simply distinguish between top-down politicization and bottom-up politicization.
 John R. Ferris, “‘Now that the Milk is Spilt’: Appeasement and the Archive on Intelligence,” Diplomacy & Statecraft, 19, 3 (September, 2008), pp. 527-65, at p. 545.
 To this I would add the establishment of a new office or agency for the purposes of bypassing normal intelligence channels. The establishment of the Office of Special Plans in the Pentagon after the 9/11 attacks is an example. Uri Bar-Joseph and Jack S. Levy, “Conscious Action and Intelligence Failure,” Political Science Quarterly, 124, 3 (Fall 2009): 461-88, at p. 475.
 The concept of motivated biases refers to the subconscious tendency to see what you want to see. It is driven by people’s emotional needs, by their need to maintain self-esteem, and by their interests – diplomatic, political, organizational, or personal. See Jervis, Why Intelligence Fails, chap. 4.
 These distinct mechanisms are illustrated by alternative interpretations of the U.S. intelligence failure regarding Iraqi WMD. Rovner (chap. 7) and Pillar (Intelligence and U.S. Foreign Policy, chap. 6) emphasize the role of politicization. Jervis (Why Intelligence Fails, chap. 3) emphasizes the role of individual psychological biases.
 For an argument that although Chamberlain might have been driven by motivated biases, there is very little evidence of successful pressure on intelligence agencies to distort the intelligence product, see Norrin M. Ripsman and Jack S. Levy, “Playing It Straight or Politicized Process? British Military Intelligence and the Nazi Threat, 1933-39” (Concordia University and Rutgers University, unpublished paper, 2012).
 Genesis: 8, 8.
 Numbers: 13,32
 Uri Bar-Joseph, Intelligence Intervention in the Politics of Democratic States: The United States, Israel, and Britain (University Park: Penn State Press, 1995), 257-97.
 Gill Bennett, A most extraordinary and mysterious business: the Zinoviev letter of 1924 (London: Foreign & Commonwealth Office, General Services Command ,1999)
 Peter Wright with Paul Greengrass, Spycatcher : The Candid Autobiography of a Senior Intelligence Officer (New York: Dell, 1987); Christopher M. Andrew, The Defence of the Realm : The Authorized History of MI5 (New York : Allen Lane, 2009), 627-43.
 Douglas Porch, “Classical Legacies of French Intelligence”, in: Thomas C. Bruneau and Steven C. Boraz, Reforming Intelligence: Obstacles to Democratic Control and Effectiveness (University of Texas Press, 2007), 121-146, 132.
 Uri Sagie, Lights within the Fog (Hebrew, Tel Aviv: Yedioth Ahronoth, 1998), 141; 154-59.
 Daniel Sobelman, New Rules of the Game: Israel and the Hizbollah after the Withdrawal from Lebanon (Tel Aviv University: Joffee Center for Strategic Studies, 2004), 30-38.
 Akiva Eldar, “A Thesis in Retrospect”, Haaretz, 29/10/2011.
 Bar-Joseph, 149-253.
 Melvin A. Goodman, Failure of Intelligence: The Decline and Fall of the CIA (Lanham: Rowman & Littlefield, 2008), 300.
 Bob Woodward, VEIL: The Secret Wars of the CIA 1981-1987. (New York: Simon and Schuster, 1987), 76.
 Goodman, 303.
 Dwight D. Eisenhower, Waging Peace, 1956-1961; The White House years (New York: Doubleday, 1965), 225.
 Christopher Andrew, For the President’s Eyes Only: Secret Intelligence and the American Presidency from Washington to Bush (New York: HarperPerrenial, 1995), 199.
 Willard C. Matthias, America’s Strategic Blunders: Intelligence Analysis and National Security Policy, 1936-1991 (University Park: Penn State Press, 2001), 136-38.
 Uri Bar-Joseph and Rose McDermott, “The Intelligence Analysis Crisis” in: Loch K. Johnson (ed.) The Oxford Handbook of National Security Intelligence (New York: Oxford UP, 2010),.
 Ibid., 361, 365-66.
 Philip H.J. Davies, Intelligence and Government in Britain and the United States, 2 volumes (Praeger Security International, April 2012 forthcoming)
 This was as much true for journalists as academics, see for example Seymour Hersh “The Stovepipe: How the Intelligence System Got Off Course,” The New Yorker, October 27, 2003 or Richard L. Russell, “A Weak Pillar for American National Security: The CIA’s Dismal Performance Against WMD Threats,” Intelligence and National Security 20/3 (September 2005) p.468.
 Acting Chairman, National Intelligence Council (1980) ‘Annual Report for FY 1980’ NFAC #6849-80, 15 October 1980, CIA-RDP86B00985R000300060006-8, NARA.
 William Colby and Peter Forbath, Honorable Men: My Life in the CIA. New York: Simon & Schuster, 1978, 352; for Huizenga’s own objections to the NIO formula see O/NE Director John Huizenga in a responding memo the following day, “The Use of NIEs,” 14 June 1973, CIA-RDP86B00269R001300030018-5, CREST, NARA
 (Cambridge: Cambridge University Press, 1996) p. 274.
 Richard E Neustadt. “White House and Whitehall,” The Public Interest 2 (Winter 1966) 55-69.
 Reginald Hibbert “Intelligence and Policy,” Intelligence and National Security Vol.5 No.1 (January, 1990) p. 113.
 This is made entirely explicit both by Lord Butler (Review of Intelligence on Weapons of Mass Destruction (London: TSO, 2004) p.7) and the Intelligence and Security Committee (Annual Report 2003-2004 (London: TSO, 2004) p .9).
 Herman Intelligence Power in Peace and War p. 275.
 (London: Cabinet Office, 2002).
 Butler Review of Intelligence on Weapons of Mass Destruction pp. 71-73.
 Butler Review of Intelligence on Weapons of Mass Destruction p. 147.
 Michael Smith, “Secret Papers Show Blair Was Warned of Iraq Chaos”Sunday Telegraph, 18 September 2004.
 Anthony Glees and Philip H.J. Davies Spinning the Spies: Intelligence, Open Government and the Hutton Inquiry (London: Social Affairs Unit, p. 12.
 K.G. Robertson ‘The Politics of Secret Intelligence: British and American Attitudes’ K.G. Robertson ed. British and American Approaches to Intelligence (London: Macmillan, 1987) p. 251.
 W.Kendall, ‘The Function of Intelligence’, World Politics 1:4 (1949).
 The author does discuss in brief (and eventually rules out) psychological variables in his analysis of President Johnson’s willingness to manipulate intelligence on the Vietnam War.
 Uri Bar-Joseph, Intelligence Intervention in the Politics of Democratic States: The United States, Israel, and Britain (University Park, PA: Penn State Press, 1997).
 Philip H. J. Davies, MI6 and the Machinery of Spying (London: Frank Cass, 2004), pp. 13-16.
 Iraq’s Weapons of Mass Destruction: The Assessment of the British Government, September 24, 2002, p. 3; http://image.guardian.co.uk/sys-files/Politics/documents/2002/09/24/dossier.pdf.
 Quoted in Blair’s forward to Iraq’s Weapons of Mass Destruction:The Assessment of the British Government, September 24, 2002, p. 3; http://www.gwu.edu/~nsarchiv/NSAEBB/NSAEBB254/doc05.pdf
 The 9/11 Commission was similarly lionized even though its analysis was flimsy and flawed. See Joshua Rovner and Austin Long, “The Perils of Shallow Theory: Intelligence Reform and the 9/11 Commission,” International Journal of Intelligence and Counterintelligence, Vol. 18, No. 4 (Winter 2005-2006), pp. 609-637.
 Elsewhere Bar-Joseph argues that intelligence officers with certain personality traits might be less vulnerable to psychological biases that inhibit analysis, which implies that some are also more adept at navigating relations with policymakers. See Uri Bar-Joseph and Rose McDermott, “Change the Analyst and Not the System: A Different Approach to Intelligence Reform,” Foreign Policy Analysis, Vol. 4, No. 2 (April 2008), pp. 127–145.