On the Politicization of Intelligence


The American intelligence community is once again under scrutiny, this time for the perceived “politicization” of intelligence reporting by senior officials. Fourteen years after the 9/11 attacks called into question the ability of the U.S. intelligence community to identify and respond to the al-Qaeda threat, allegations have surfaced that senior Central Command (CENTCOM) officials altered intelligence reporting on the Islamic State in Iraq and the Levant (ISIL).

According to The Daily Beast, some 50 CENTCOM intelligence analysts (including DIA analysts) formally complained that senior officials had changed their assessments on the Islamic State and the Syrian branch of al-Qaeda. The New York Times raised the issue to national attention and former Director of the Defense Intelligence Agency Lt. Gen. Michael Flynn observed that politicization is “here” and “it’s dangerous.”

A more accurate observation, perhaps, is that politicization is still here and still dangerous. In this case, the issue seems to involve pressure from CENTCOM’s intelligence chain of command rather than direct influence from policymakers, as far as we know. In September, The Daily Beast confirmed that the Pentagon inspector general was investigating the allegations, citing additional unnamed sources claiming the allegations involved senior CENTCOM officials pressuring or attempting to alter the work of analysts working within the command.

The CENTCOM incident harkens back to similar allegations about Iraq’s weapons of mass destruction programs before the 2003 invasion. Yet this case is a different twist on the politicization pathology in American intelligence.

In his landmark study Enemies of Intelligence: Knowledge and Power in American National Security, Richard K. Betts characterized 2002 and 2003 as the nadir of the history of U.S. intelligence. In one extreme case, Betts said that policymakers attempted to remove the national intelligence officer for Latin America and a State Department expert on biological weapons because their assessments on Cuba and Syria did not conform to administration views. In another case, senior administration officials released information on Iraqi weapons programs, omitting any reference to dissenting views and then prohibiting intelligence agencies from articulating dissent after the publication. A final example of politicization concerned the creation of a separate analytic effort within the Office of the Secretary of Defense to produce assessments later provided directly to White House officials without intelligence community coordination.

Debate continues on the issue of pre-2003 politicization as a result of different definitions of what constituted “politicization” of Iraq WMD intelligence and even more significant politicization of the investigations into pre-war intelligence itself. The December 2004 Senate report on pre-war Iraqi intelligence found no specific cases in which analysts were pressured to change reporting on Iraq’s weapons programs. For some, the absence of specific cases where analysts were “pressured” or coerced into changing conclusions or assessments means there was no politicization.

But the 2004 report was compiled and published by a Republican-controlled Senate. Senate Democrats believed that the report did not address fundamental issues about the administration’s skewing of intelligence and that politicization occurred in a form other than outright pressure on analysts. After intense debate among Republicans and Democrats about what would be included in the 2004 report, a political compromise was reached to delay deeper investigation into issues related to how administration officials presented pre-war intelligence or issues related to the creation of separate intelligence activities inside the Defense Department.

In the June 2008 press release for a “Phase II” report on pre-war intelligence, then-Senator John D. Rockefeller IV argued a form of politicization occurred during the selling of the war to the American public. Arguing that the George W. Bush administration “repeatedly presented intelligence as fact when in reality it was unsubstantiated, contradicted, or even non-existent,” Rockefeller concluded that “the American people were led to believe that the threat from Iraq was much greater than actually existed.” The report found nothing unlawful about the pre-war intelligence, but raised important questions about intelligence-policy relationships.

An important takeaway from investigations into pre-Iraq weapons intelligence and the current CENTCOM allegations is that allegations of intelligence politicization are almost always more intense and political than the actual politicization itself. To borrow from Betts, in cases as politically charged as the selling of the invasion of Iraq to the American people and anything related to the current administration’s ISIL strategy, “politicization is a fighting word” — and charges of intelligence politicization are themselves seized upon to further partisan agendas.

What, then, is the “politicization of intelligence?” In his bestselling textbook Intelligence: From Secrets to Policy, Mark Lowenthal locates politicization problems as a “line separating policy and intelligence” that is “best thought of as a semipermeable membrane” in which “policy makers are free to offer assessments,” while “intelligence officers are not allowed to make policy recommendations.”

Politicization issues are different from other pathologies besetting intelligence, particularly the many forms of intelligence bias and myriad cognitive biases that can dilute, distort, or influence intelligence collection, analysis, and reporting. For example, mirror imaging bias occurs when analysts project U.S. attitudes, beliefs, or perceptions onto their targets, wrongly believing that foreign behavior, intentions, or mindsets mirror U.S. ones. Anchoring bias occurs when our perception or analysis of an issue or decision is overly influenced by one piece of information, often the first information we encounter on a topic. Confirmation bias involves the tendency to search for or inflate the value of information that supports or reinforces preconceptions or beliefs, causing analysts to ignore or dismiss contrary evidence.

Such biases are largely unconscious; they act upon individuals at the cognitive level. Other biases are social in nature, acting across groups or within coordination processes. The most infamous of these is groupthink, in which members of a group unconsciously allow a desire for conformity to undermine their assessments, which results in the muting of dissent or alternative views. Individual and social biases are a problem for any decision-making or analytic process.

Politicization issues are different from biases in both their origins and their resolution. Lowenthal identified several types of politicization:

  • Analysts deliberately alter assessments to support policy outcomes;
  • Policymakers use processes and procedural rules to influence the output intelligence workflows;
  • Repeated, negative reactions to conclusions/assessments cause analysts to alter future reporting to avoid further negative reactions; and
  • Policymakers “cherry pick” or subjectively choose among competing assessments to further their policy agendas (and mute dissenting views).

Former Director of the CIA and Secretary of Defense Robert Gates addressed politicization with CIA analysts in 1992 soon after his confirmation, which was tainted by charges that Gates had been involved in politicizing intelligence on the Soviet threat during the Reagan administration. He identified politicization as changes or shifts in emphasis on specific issues or by senior officials’ questioning during routine coordination processes. This seems similar to the CENTCOM incident. He also pointed to the expansion or limitation of routes to express alternative views, another allegation in the CENTCOM incident.

To counter politicization, Gates encouraged better communication across the intelligence process and tasked managers with challenging assessments and creating a culture of openness and dissent. He also encouraged analysts to challenge managers, reduce levels of review and coordination to avoid misunderstanding, and resist the tendency to simplify issues or assessments to produce “consensus” reports. Gates expands on Lowenthal’s “semipermeable membrane” construct regarding the line separating analysts from policymakers to include:

  • Agency products edited to emphasize parochial views vs. those including dissent;
  • Anticipating preferences in formatting that highlight analytic insights vs. consumer predilections (“pandering” to policymakers);
  • Objective, critical reporting based on all available information vs. selective use of information; and
  • Expediency to “answer the mail” vs. deliberate analysis with all available sources synthesized into a coherent report.

Overall, Gates directed analysts to “ensure that underlying policy goals are not distorting our analysis.”

Betts offers a more nuanced view. “Depending on the definition of the term,” Betts argues, “politicization is to some degree inevitable, and, in some forms, necessary.” Assessing both the costs and benefits of politicization, Betts observes that few people are willing to admit “that there could be anything less than evil about politicization.” He writes that although politicization is “a fighting word,” the strict definition of “politicize” is not “ispo facto pejorative” because it literally means “to give a political tone or character” in order “to bring within the realm of politics.”

To be clear, Betts agrees that a standard — an “irrevocable norm” — exists regarding the question or type of politicization: “Policy interests, preferences, or decisions must never determine intelligence judgments.” But in some cases, politicization is benign or even beneficial. A beneficial form, for example, involves objective, insightful analysis that is formatted and conveyed in a way that prevents key information from being overlooked or ignored in the massive amounts of data policymakers must digest every day.

The very relationship between intelligence analysis and policymaking requires that assessments and reporting must be crafted and conveyed to policymakers in the time, format, and medium amenable to their needs. In addition, many policymakers are pressed for time and are unable to read the deluge of written products they receive, requiring oral briefs and presentations that are more susceptible to problems of interpretation.

We will not fully understand the nature of the CENTCOM allegations until the office of the Pentagon inspector general completes its work. Yet the fact that Secretary of Defense Ash Carter felt compelled to reiterate the need for “unvarnished, transparent” intelligence analysis on the eve of 9/11 remembrances should raise questions the 9/11 and Iraq Weapons of Mass Destruction Commissions didn’t address: the efficacy of larger relationships and processes between the intelligence and policy communities. War on the Rocks senior editor Mark Stout recently addressed these issues concerning Vietnam War estimates.

In the CENTCOM case, we might also consider why reporting primarily intended to support warfighters in the command’s area of interest should be scrutinized for alignment with strategic narratives or assessments of concern primarily to policymakers outside the command. Perhaps it is time to rethink processes and policies guiding the coordination of intelligence for internal command consumption to limit the inevitable impulse to question how reports will be received externally. Regardless of how the CENTCOM incident plays out, heightened tensions in all theaters and increased domestic political interest in foreign policy creates the potential for subsequent problems with intelligence “narratives.”

There is no easy answer to the larger question of strengthening the intelligence–policy relationship and limiting the negative dimensions of politicization. Concerning CENTCOM’s ISIL reporting, the fact that the Pentagon’s inspector general is investigating based on a complaint by analysts is actually a sign that processes to safeguard the intelligence process are working. If true, the CENTCOM incident informs our understanding of intelligence–policy pathologies and adds a case in which processes intended to preserve the integrity of the overall intelligence process worked to correct a problem. If false, it adds an example of whistleblowers using one of their outlets for redress and deserves further attention to ensure those making formal complaints remain immune from retaliation or punishment.


Robert R. Tomes, PhD is President of the MapStory Foundation and an adjunct professor of security policy studies at Georgetown University. His publications include U.S. Defense Strategy from Vietnam to Operation Iraqi Freedom: Military Innovation and the New American Way of War, 1973-2003 (Routledge, 2007), which analyzes the Cold War offset strategy as a case study in military innovation. A former intelligence community senior manager, he served as a co-chair of an interagency working group on alternative analysis.