One intellectual means of assessing requirements is to look at the likelihood of an event and its relative importance to national security concerns. Of great concern will be high-likelihood and high-importance events. It should be easier to assess importance (which should be based on known or stated national interests) than it is to assess likelihood (which is itself an intelligence judgment or estimate). (Likelihood, however, is not a prediction. See the discussion in chap. 6.) For example, during the cold war a Soviet nuclear attack would have been judged a high-importance but low-likelihood event. Italian government instability would have been judged a high-likelihood but low-importance event. Of the two, the Soviet issue would rank higher as a priority or intelligence concern because of its potential effect, even though an attack seemed possible in only a few instances and an Italian government fell several dozen times.
In both Panel A and Panel B of Figure 4-1, the issues that fall closer to the upper right reflect more important intelligence requirements. However, there may not be startling clarity as to likelihood or there may be a debate as to issues’ relative importance.
The hidden factor that drives priorities is resources. It is impossible to cover everything. The United States, for example, has long had interests in every part of the globe, although some are more significant and more central than others. For decades, the U.S. intelligence community has used a variety of processes to set priorities. The most recent example is the National Intelligence Priorities Framework (NIPF), which supports a national security policy directive (NSPD) signed by President George W. Bush in February 2003.
According to congressional testimony by director of central intelligence (DCI) George J. Tenet, the NIPF provides for semiannual reviews of intelligence priorities by the president and the NSC. Tenet described it as being more flexible and more precise than any previous intelligence priority system. The NIPF is connected directly to analytic and collection resources to ensure that the most urgent needs are being covered and that gaps can be identified quickly. The system is also used for planning in the five-year budget cycle. Other testimony revealed that each topic in the NIPF has an intelligence topic manager who helps determine collection requirements. The NIPF appears to be a more pervasive system in terms of overall intelligence community functions and a more flexible system than has been used in the past.
All priority systems must address the issue of
priority creep.
Issues can and do move up and down in a priority system. This is actually a positive occurrence as it shows that the priority system is dynamic and responsive to changes in the international situation. The problem is that issues cannot receive significant attention until after they have begun moving up to the higher priority tiers, at which point they must compete with the issues already in that bracket. Priority creep can become a problem as analysts or policy makers seek higher priority for certain issues. Priority creep is further exacerbated by the difficulty encountered in returning issues to lower priority status once they have become less urgent. Neither the intelligence analysts working on that issue nor the policy makers whom they support are eager to admit that the issue is no longer as important. After all, it is their issue. This underscores the problem with any intelligence requirements system. Such a system is, of necessity, static between reviews. Even if the requirements are reviewed and re-ranked periodically, such as the six-month review in the NIPF, they remain snapshots in time. Policy makers or intelligence officials must decide on the requirements and resources to be applied to them. However, the nature of international relations is such that unexpected issues inevitably crop up with little or no warning. These are sometimes referred to as
ad hoes.
When an issue like this arises, some policy makers and intelligence officers exert pressure to give the new issue a priority high enough to compete with other high-priority issues. Some resistance is felt, usually from those whose access to intelligence resources is threatened. Not every ad hoc merits higher priorities. (Some intelligence analysts speak of the
tyranny of the ad hoes.)
Moreover, a system that constantly responds to each ad hoc soon has little control over priorities and quickly breaks down. Thus, the system that preserves a modicum of flexibility or a modest reserve capability is more responsive to the realities of intelligence requirements. Policy makers often have little time or inclination to conduct periodic reviews of intelligence priorities, even as often as annually. As a result, static, potentially outdated requirements and the necessity to make requirements decisions can be problems for the intelligence community. This was apparently the problem with the priority tier system used under President Bill Clinton. Once the system was introduced near the middle of his presidency, Clinton was not interested in visiting the relative rankings again. Without his input the priorities could not be changed, resulting in a set of priorities that were increasingly divorced from international realities and that came to be dominated by issues pushed to higher priorities by their intelligence managers.
Moreover, if a requirement cannot be met with current collection systems, developing the technical systems or the human sources will take time. Thus, uncertainty about requirements or lower priorities for some of them will affect the development of collection capabilities.
COLLECTION
Collection derives directly from requirements. Not every issue requires the same types of collection support. The requirements depend on the nature of the issue and on the types of collection that are available. For example, concerns over possible threat from cyber attacks likely derive little useful intelligence from imagery as the locus of the threat cannot be captured in a photo. Much better intelligence might be derived from signals intelligence, which can reveal capabilities or intentions. Collection is also the first—and perhaps the most important—facet of intelligence where budgets and resources come into play in precise terms (as opposed to broader discussions when priorities are at issue). Technical collection is extremely expensive and, because different types of systems offer different benefits and capabilities, the administration and Congress must make difficult budget choices. Also, the needs of agencie vary, further complicating the choices.
How much information should be collected? Or, put another way, does more collection mean better intelligence? The answer to these questions is ambiguous. On the one hand, the more information that is collected, the more likely it will include the required intelligence. On the other hand, not everything that is collected is of equal value. Analysts must wade through the material—to process and exploit it—to find the intelligence that is really needed. This is often referred to as the “wheat versus chaff problem.” In other words, increased collection also increases the task of finding the truly important intelligence.
An interesting phenomenon, found at least in the U.S. intelligence community, is that different analytical groups may prefer different types of intelligence. For example, the Central Intelligence Agency (CIA) may put greater store in clandestine human intelligence (espionage), in part because it is a product of CIA activities. Meanwhile, other all-source analysts may place greater emphasis on signals intelligence.
PROCESSING AND EXPLOITATION
Intelligence collected by technical means (imagery, signals, test data, and so on) does not arrive in ready-to-use form. It must be processed from complex digital signals into images or intercepts, and these must then be exploited—analyzed if they are images; perhaps decoded, and probably translated, if they are signals. Processing and exploitation are key steps in converting technically collected information into intelligence.
In the United States, collection far outruns processing and exploitation. Much more intelligence is collected than can ever be processed and exploited. Furthermore, technical collection systems have found greater favor in the executive branch and Congress than the systems and personnel requirements for processing and exploitation. One reason for this appeal is emotional. A similar circumstance, for example, exists in formation of the defense budget. Les Aspin. chairman of the House Armed Services Committee (1985-1993) and later the secretary of defense (1993-1994), once observed that both Congress and the executive branch were more interested in procurement (buying new weapons) than operations and maintenance (keeping already purchased systems functioning). Buying new systems was more attractive to decision makers in both branches and, more important, to defense contractors. Operations and maintenance, although important, are less exciting and less glamorous. Collection is akin to procurement and is much more appealing than processing and exploitation.
Collection advocates argue, usually successfully, that collection is the bedrock of intelligence, that without it the entire enterprise has little meaning. Collection also has support from the companies (prime contractors and their numerous subcontractors) who build the technical collection systems and who lobby for follow-on systems. Processing and exploitation are in-house intelligence community activities. Although these downstream
activities
(the steps that follow collection) are also dependent on technology, the technology is not in the same league, in terms of contractor profit, as collection systems.
The large and still growing disparity between collection and processing and exploitation results in a great amount of collected material never being used. It simply dies on the cutting-room floor. Advocates of processing and exploitation therefore argue that the image or signal that is not processed and not exploited is identical to the one that is not collected—it has no effect at all.
No proper ratio exists between collection and processing and exploitation. In part, the ratio depends on the issue, available resources, and policy makers’ demands. But many who are familiar with the U.S. intelligence community believe that the relationship between these two phases has been and remains badly out of balance. The congressional committees that oversee intelligence have increasingly expressed concern about this imbalance, urging the intelligence community to put more money into processing and exploitation. This is often referred to as the TPEDs (pronounced tee-peds) problem. TPEDs refers to tasking, processing, exploitation, and dissemination. Tasking is the assigning of collectors to specific tasks. Of the four parts of TPEDs, tasking and dissemination are the least problematic for the intelligence community or for Congress. The processing and exploitation gap is of highest concern to Congress.
ANALYSIS AND PRODUCTION
Major, often daily, tension is evident between current intelligence and long-term intelligence. Current intelligence focuses on issues that are at the forefront of the policy makers’ agenda and are receiving their immediate attention. Long-term intelligence deals with trends and issues that may not be an immediate concern but are important and may come to the forefront, especially if they do not receive some current attention. The skills for preparing the two types of intelligence are not identical, and neither are the intelligence products that can or should be used to disseminate them to policy makers. But a subtle relationship exists between current and long-term intelligence. Like collection versus process and exploitation, a proper balance—not necessarily 50-50—should be the goal.
The U.S. system of competitive analysis—that is, having the same issue addressed by several different analytical groups—entails some analytical costs. Although the goal is to bring disparate points of view to bear on an issue, intelligence community products written within this system run the risk of succumbing to groupthink, with lowest common denominator language resulting from intellectual compromises. Alternatively, agencies can indulge in endless and—at least to the policy consumers—meaningless
footnote wars,
the only goal of which is to maintain a separate point of view regardless of the salience of the issue at stake. In the aftermath of critiques about intelligence performance on 9/11 and Iraqi weapons of mass destruction (WMD), the intelligence community has begun to put greater emphasis on collaboration, which usually means greater sharing among analysts both of their sources and their analyses. This new emphasis raises additional concerns as well, the most obvious of which is the potential for greater groupthink. (See chap. 6 for a fuller discussion.)
Analysts should have a key role in helping determine collection priorities. Although the United States has instituted a series of offices and programs to improve the relationship between analysts and the collection systems on which they are dependent, the connection between the two has never been particularly strong or responsive. Improving this relationship has been one of the goals of the NIPF. The ideal state is one in which there is analytically driven collection, that is, collectors act in response to analytic needs and not more independently or opportunistically.
The training and the mind-sets of analysts are important. Analysts must often deal with intelligence that is contradictory, both internally and when viewed in light of their strongly held professional beliefs and perhaps their own past work. The way in which analysts deal with these contradictions depends on their training and the nature of the broader analytical system, including the review process.
Finally, analysts are not intellectual ciphers. They are likely to have ambitions and want their issues to receive a certain degree of high-level attention. This is not meant to suggest that they will resort to intellectually dishonest means to gain attention, but that possibility must be kept in mind by their superiors within the intelligence community and by policy makers.