Giving Intelligence Analysts Their Voices Back: The Case for Analyst Perspectives
“Speak truth to power” — the sacred and deep cultural oath of the CIA and broader intelligence community — reminds intelligence analysts of their responsibility to provide unvarnished, non-politicized assessments to policymakers. And yet, these days speak truth to power may ring hollow for many analysts, not because organizations and leaders in the U.S. system politicize intelligence, but rather because the onerous coordination and review process has too often resulted in watered-down and less impactful analysis. Unlike in the academic community where a diversity of views based on the same evidence is encouraged, the intelligence community actively admonishes against it in favor of a review process meant to build a consensus, even if one does not exist. The end product of this bureaucratic process, known as “the corporate product,” purposefully masks analytic disagreement in favor of presenting a united “analytic line” to the policymaker.
Indeed, the corporate product has been around for ages. However, it was reinvigorated in the 1980s by then-Deputy Director of the CIA Robert M. Gates, who in 1992 as Director of Central Intelligence told a CIA audience that “they must discard the academic mindset that says their work is their own, and they must take into account the views of others during the coordination process.” Scholar Richard Betts would later accurately observe that “intelligence products are supposed to represent the best judgments of whole organizations, not single authors.” Now considered standard practice throughout the intelligence community, the corporate product is designed to avoid “confusing” policymakers with too many disparate views. Moreover, rigorous peer and management review is believed to strengthen analysis. So why change anything?
Whether we realize it or not, many policymakers are formidable analysts themselves. In many cases, they have their own deep expertise. Moreover, they have access to much of the same raw reporting and finished analysis offered by intelligence agencies as well as policy materials that are not always shared with intelligence analysts. Therefore, policymakers are typically perceptive enough to understand that different viewpoints indeed exist, and that some are more strongly supported by the available information than others.
The coordination and review process for corporate products can be considered a peer-review process, but in reality, it is a compromise process. Corporate products must maintain unity of voice to hold an analytic line. To be sure, this process has improved marginally over the last few years, with greater numbers of differing perspectives incorporated into finished corporate products. However, it is incumbent upon individual analysts to battle against others — in many cases against their own management — to get a dissent included in the final publication. As a result, most would-be dissenters simply quit, beaten down by the bureaucracy. Even more insidiously, in my experience the coordination process is commonly employed as the modus operandi to stall and ultimately eliminate dissenters from having a voice in the final product. Analysts who disagree with colleagues may feel shamed into adjusting their views to preserve their professional reputations, a phenomenon dubbed by expert Jack Davis as “tribal think,” a subset of groupthink.
Unfortunately, the coordination system has become less peer-review and more bureaucratic box-checking, meant to ensure that all organizations and analysts with a stake in the analysis (and even those who do not) have the opportunity to weigh in. And despite noble external proposals to delineate proper roles and responsibilities, the management review process in practice remains a cumbersome and messy affair. Analysts are forced through a gauntlet of multiple review layers and word-smithing that sometimes changes the original intent or message of their analysis. If considered an intelligence community-wide product forged among multiple agencies, such as a National Intelligence Estimate, then there are additional layers of coordination and review required.
In my view, this coordination process has resulted in a steady, nibbling away of candid analysis in favor of more generic assessments aimed to appease multiple constituencies. According to former CIA analyst and intelligence scholar John A. Gentry, although “published intelligence products usually are better than their original drafts” they “sometimes bear only passing resemblance to analysts’ original drafts and their personal judgments.” Policymakers are then frustrated when they receive less hard-hitting and less thought-provoking analysis. For example, according to an account by former National Security Council (NSC) staff member Will Inboden, intelligence assessments at times were so generic that it rendered them practically useless and laughable. Inboden noted that “on the NSC staff we used to craft parodies of intelligence analyses that began ‘We judge with strong confidence that China is large, important nation in East Asia…’” Unfortunately, similar refrains are common throughout the policymaking community.
To address this problem, a range of potential solutions exists, but I recommend here that the intelligence community offer analysts a channel to convey their individual insights directly to policymakers, outside of the corporate product context. An “analyst perspective” would provide policymakers with well-reasoned alternative views, otherwise known in the community as “dissents,” to the official analytic line. Analyst perspectives would undergo no coordination or review, and in essence, represent the considered personal views of one or several analysts. Such a modification to the process would better support policymakers and give analysts their voices back, allowing them to once again speak truth to power.
Interestingly, a useful analog to analyst perspectives exists in the foreign service component of the State Department. Known as a “dissent channel cable,” foreign service officers are authorized to individually or as a group author a dissent cable on current U.S. policy that is sent directly to the secretary and other senior leaders. Dissent cables are only to be authored when there is an “inability to resolve concrete differences of opinion on substantive foreign policy issues [that] has prevented such views from being reported.” For obvious reasons, the dissent channel cable is considered an extreme route, only to be authored if there is no other recourse. This is not to say that it is never employed. One recent example occurred when 51 diplomats got together to pen a dissent on Syria policy. Like dissent cables, an analyst perspective would enable those working most closely on the issues to craft their own arguments to present to policymakers.
As with dissent cables, reasonable limitations would have to be in place on analyst perspectives to make this work. First, analysts should only submit an analyst perspective when there is no other viable recourse from the corporate analytic line. Analyst perspectives would not be designed to replace the corporate product, but only supplement them. Second, to discourage frequent and unnecessary use of the channel, analysts would have to author analyst perspectives by name and might be limited in the number of dissents they can issue per year. Finally, perspectives would not be op-eds that enable analysts to take positions on U.S. policy. Just like corporate products, analyst perspectives would have to strictly follow Office of the Director of National Intelligence (ODNI) analytic tradecraft standards and steer clear of policy prescription.
According to multiple studies on the subject, because policymakers already have well-formulated views of the issues in their portfolio, they typically seek either confirmation or a new way to interpret the data. Lowest common denominator assessments may confirm what policymakers already know because the coordination process tends to yield safer, more conservative analysis. However, it seems unlikely that these same assessments could help spur policymaker reinterpretation of the data. With analyst perspectives, policymakers would receive more stimulating alternative views less encumbered by the need to compromise on the analysis. Moreover, unlike the dissents of today which typically run only for a short paragraph, analyst perspectives would run the length of full articles and therefore allow analysts to more fully flesh out alternative argumentation.
Analyst perspectives would also expand policymaker confidence in intelligence assessments that arrive on their desks. Intelligence managers tend to think that agency seals atop corporate products add an authoritativeness to their assessments. Policymakers, however, operate in highly personalized environments, and depersonalized intelligence assessments — totally devoid of authorship on the byline, which is common among most intelligence products — have, in my experience, at times resulted in policymakers desiring to know the source of the analysis to evaluate its credibility. Analyst perspectives would occasionally present this personalized effect, thereby enhancing relationships and critical dialogues between intelligence analysts and their policy clients.
On the analyst side of the equation, analyst perspectives would serve as a useful check on corporate products, constantly testing key assumptions and interpretation of the evidence. Indeed, the ODNI’s own analytic tradecraft guidance states that “analysts should be alert to influence by existing analytic positions or judgments and must consider alternative perspectives and contrary information.” The guidance goes on to state that “analytic products should identify and assess plausible alternative hypotheses.” The problem, however, is that the intelligence community’s laser-like focus on the corporate product often comes at the expense of drawing out differences in assumptions and evaluation of the data.
Another benefit of analyst perspectives is that they would impress upon policy customers the diversity of views in the intelligence community. Corporate products run the risk of providing a false narrative of consensus, making it easier for policymakers to later point to “intelligence failure” when policies go wrong. The use of analyst perspectives, however, would demonstrate that intelligence analysis is more of an art than an exact science, and that the world is far more nuanced and complicated than corporate products might suggest.
It is important to highlight that analyst perspectives would not be synonymous with “red cell” assessments. These are meant to be deliberately alternative, creative, hypothetical, and provocative, and for that reason, are given less credence by their home agencies. By contrast, perspectives would have more legitimacy within their respective organizations because line analysts themselves would author them to present alternative argumentations of the same evidence, not to be intentionally hypothetical or provocative.
A reasonable counter-argument to analyst perspectives is that they would allow policymakers to cherry-pick assessments that support their policy views. Indeed, policymakers would play favorites, but this position ignores the reality of the situation today, which is to say that policymakers already engage in this type of behavior on a regular basis. Therefore, analyst perspectives would simply represent another respected view that provides policymakers with the opportunity to compare notes, which, at the end of the day, is what they are looking for anyway.
Another critique of analyst perspectives is that they could unnecessarily empower ideological analysts to produce reports to achieve the spotlight with policymakers. Constraints on analyst perspectives, such as per year limitations and byline requirements, could effectively limit abuse of this new process. Regardless, daily intelligence briefers would continue to perform their natural filtering function, in effect limiting the volume of assessments reaching a policymaker’s desk. Furthermore, the intelligence community analytic ombudsman would continue to investigate lack of objectivity claims against analyst perspectives to prevent politicization in this new process.
Derek Grossman is a senior project associate at the nonprofit, nonpartisan RAND Corporation. He previously served at the Defense Intelligence Agency as the daily intelligence briefer to the Assistant Secretary of Defense for Asian and Pacific Security Affairs at the Pentagon. Grossman was the winner of the 2014 “Galileo Competition” which fosters new and innovative ideas in the intelligence community. The views expressed in this article are those of the author and do not reflect the official policy or position of the Defense Intelligence Agency, Department of Defense, or the U.S. government.