A Framework for Reform of the U.S. Intelligence Community: Chapter 4

A Framework for Reform of the
U.S. Intelligence Community

Chapter 4: Elements of Institutional Reform

Several key reforms need to be made. The most important relate to leadership and integrity issues, but some organizational changes are also desirable. They should improve efficiency, while lending institutional support to improvement and maintenance of high performance standards.

Strengthen the DCI

The Director of Central Intelligence has a number of importance powers and responsibilities, but he does not control the Intelligence Community in any meaningful way. This has been the fact for years. Former CIA officer Victor Marchetti described DCI Richard Helm's reaction to recognition of his limited powers:

In the wake of [previously described bureaucratic] defeats, Helms gave up on attempts at managing the intelligence community. At one point, months later, he observed to his staff that while he, the DCI, was theoretically responsible for 100 percent of the nation's intelligence activities, he in fact controlled less than 15 percent of the community's assets--and most of the other 85 percent belonged to the Secretary of Defense and the Joint Chiefs of Staff. Under such circumstances, Helms concluded, it was unrealistic for any DCI to think that he could have a significant influence on U.S intelligence resource-decisions or the shaping of the intelligence community (62).

This lack of centralized control has had the predictable effect of leading to inter-agency rivalries, needless functional duplication, and a complicated structure that leads to consumer confusion about just what the Intelligence Community is doing and what it thinks. An organizationally stronger DCI would be able to resolve some of these problems. Legislative reforms should include:

  • Increasing the portion of the Intelligence Community directly under the DCI's control, and simultaneously giving him more control over the budget of the areas he operationally manages.

  • Unification of national-level collection and analytic agencies under the DCI to give him maximum control over the assets. This control should ensure their most efficient, coordinated efforts, and give the DCI access to the best judgments possible in fulfillment of his responsibilities as statutory advisor to the National Security Council. This unification would require amendment of Public Law 102-496, which gives the Secretary of Defense authority over some "national" intelligence capabilities, imagery, and the National Reconnaissance Office (NRO) (63).

    Unification of the collection agencies, in particular, could help rationalize the "system" of collection programs run by NSA, NRO, the military, and CIA. Even with the coordinating committees maintained at the the community management level, the organizational structure is Byzantine, disjointed, and not really controlled by anyone. Turf motives are strong. Some agencies have overlapping functions; inefficiencies thus are greater than necessary. Former SSCI staffer Angelo Codevilla described the management of a "system" of a large number of semi-autonomous programs:

    The Intelligence Community Staff contains committees to coordinate the various agencies' plans with regard to technical collection, and the Congressional intelligence committees also sometimes prod the system toward coherence. Yet coherence is elusive, because coordination is ex post facto to budgetary planning. Each program is driven by distinct purposes, and each has its own constituencies (64).

    As with any organization, a real head is needed to assure control, efficiency, and accountability. The logical head is the DCI, whether he retains that title or is called Director of National Intelligence--as former Senator David Boren and others have proposed. There may be real financial savings associated with elimination of redundant activities and better managing the remaining programs.

    DCIs need to be stronger leaders as well. DCIs William H. Webster (1987-91) and R. James Woolsey (1993-95) were very weak, and Robert M. Gates (1991-1993) was a major creator of the Intelligence Community's problems. Leadership skills and presidential wisdom in picking a good DCI cannot be legislated, but Congress and the president can make clear that they expect strong leadership and management from directors. The DCI must establish clear ethical and performance standards, and hold all Intelligence Community personnel under his control to those standards. This leadership is essential.

    Collection and Analysis: Complements or Adversaries?

    The reorganizational debates have at times focused on whether analysis should be co-organized with collection agencies. This issue particularly has concerned analysis and human intelligence (HUMINT) collection. Opponents of location within the same agency point to the dirtying of analysis through co-habitation, of opportunities for politicization, of the need for clear organizational boundaries. However, there are major advantages to close coordination between collection and analytic organizations that I believe overwhelm the negative aspects of such collaboration:

  • Better collection guidance. Analysts who closely coordinate with collectors--who have such coordination as a performance directive--can guide collectors toward areas of current as well as long-term interest. If analytic organizations monitor consumer needs and anticipate them--the old "over-the-horizon" mission--they can push HUMINT collectors, especially, to begin the time-consuming process of developing assets that may provide critical warning intelligence and eventually information to meet current intelligence requirements. Without analytic guidance about possible long-term needs, collectors may not know how to best allocate their limited resources. Similarly, analyst guidance can help designers of technical systems make equipment best suited to meet policymakers' informational needs, not engineers' fancies.

  • Better quality collection. Analysts can help "reports officers"--the DO officers who guide field "case officers" in targeting, grade their reports, and interface with analysts to ensure an internal firewall of security for operators and their assets--to assess the quality of collection. They can offer advice for improvements in collection. Analyst guidance would reduce the much (and perhaps too much) criticized tendency of case officers to collect on issues easily collected on in order to improve production numbers. Clear quality determinants of performance--inherent rigor, completeness, and so on, as well as clear instructions from reports officers to case officers about consumer priorities as identified by analysts--would help make human collection more valuable and cost-effective.

  • Better analysis. Better collection begets better analysis. In a similar vein, the ability of analysts to talk with DO reports officers can gain them insights into the quality of reports and the reliability of assets. They can get oral expressions of confidence, concerns, or suspicions that reports officers will not commit to paper in their final disseminated reports. These affect the weight analysts place on such reporting. Thus, the value to analysts--and to reports officers and ultimately to consumers--of a good DI/DO working relationship can be very great indeed.

    Good leadership can overcome or minimize the few disadvantages of proximity. Clearly objective analysis, ensured by good leadership, will reduce policymakers' concerns about "slant." An accurate representation of the relationship between analysts and "spies" will boost understanding of the linkages between the two functions--and their significant differences. There is little purpose served by pretending no relationship. The fiction of "separateness" intended by the Carter Administration's renaming the DI the "National Foreign Assessment Center" fooled only people ignorant of intelligence.

    Unite NSA and CIA

    The National Security Agency and CIA should be unified under one head--the DCI. There are some reasons as well to integrate the organizations more formally into a single agency. These have performance as well as efficiency elements.

    First, NSA serves national intelligence needs--just as CIA does. There is no reason for NSA to be formally part of the Department of Defense and to have as its head a serving military flag officer. NSA ought to be headed by a civilian and be part of the Executive Office of the President, just as CIA long has been. Such a reorganization would in no way diminish NSA's role in serving the intelligence needs of combat commanders. Military personnel could continue to work in many NSA positions, just as they now also serve in CIA and with the NIC.

    Signals intelligence (SIGINT) and related technical collection is a significant complement to the HUMINT effort housed mainly at CIA. National-oriented collection is resource constrained and needs better coordination. Reorganization of the Intelligence Community could help accomplish this by integrating NSA's SIGINT and other capabilities better with the DO's HUMINT collection. These include, for example, joint targeting and better use of one agency's assets to pinpoint areas for exploitation by the other's capabilities. The current organizational fragmentation and bureaucratic competition work to reduce mutually beneficial joint work. Analysts working directly with both collection services could help improve the efficiency of overall collection. For example, co-location of NSA's collection with CIA's analytical strengths would foster better NSA targeting and better NSA understanding of future targeting needs--and thus improve its technical development. This coordination could become still more important if technological changes--possibly including failure to employ clipper chips, growing use of encrypted Internet communications, and more widespread use of fiber optics, for example--inhibit NSA's ability to collect information and ultimately threaten its viability as a collection agency.

    Linking CIA's analyst corps more closely to NSA would improve CIA's understanding of SIGINT--and thus also its ability to judge the value and reliability of NSA's information. It would improve the requirements process, which is now so formalistic and complex that many CIA analysts do not bother to task NSA (65).

    Unification of the DI and NSA's more modest analytic staff would tend to improve the analysis of SIGINT, while freeing NSA analysts of the needless fetters that now restrict them to assessment of world events based only on NSA's collection--that is, which prohibits them from all-source analysis. It also would allow assessment of the mass of data that NSA has collected but long has been unable to adequately process.

    Unification would tend to reduce, if not eliminate, the petty bureaucratic squabbling between NSA and CIA's Directorate of Science and Technology, in particular, which have some overlapping responsibilities. This is a serious problem that warrants high-level evaluation in a classified context.

    Reform Military Intelligence

    The complex of Defense Department intelligence organizations needs substantial additional reorganization and reform. Such reform would be complicated by the Byzantine bureaucracy that is the Department of Defense, however. The end of the Cold War and some vigorous leadership has prompted a number of consolidations and closing of intelligence units focused mainly at Warsaw Pact forces, but more can be done to refine the military's intelligence effort. Ignoring in this section NSA and the NRO, the major reforms should rationalize the work of DIA--the intelligence arm of the Joint Staff--and the four service intelligence operations.

    Each service has an intelligence headquarters, analytic functions, its own printing shop(s), schools, doctrine, personnel policies, and service-specific duties to support assigned missions. Moreover, the continued independence of the services in training, logistical, and other support functions, combined with bureaucratic turf interests to defend, means that the Joint Staff has little direct authority over the services' intelligence (or any other) functions.

    Thus, there is little prospect for real efficiency gains in military departmental intelligence until there is substantial reform of the Defense Department itself. In particular, there must be a sharp diminution of service power before parochial interests can be subsumed to the greater good of the military as a whole--and the nation. This probably means adoption of a general staff-type structure--something that now seems to be a remote prospect.

    Barring such reform, the Secretary of Defense could mandate somewhat greater powers for DIA, or legislation could give it more authority. Reasonable objectives here would include placement of certain analytic functions entirely in DIA's hands, for example, while reducing the size and scope of service organizations. It also could mean better coordination of DIA's work with that of the intelligence organizations of the unified and specified commands, and more specialization of functions. In the absence of legislation (and probably with it if the Goldwater-Nichols Act experience is a guide) there will be service resistance to diminution of the services' budgets, staffs, and responsibilities.

    Consolidation efforts might mean, for example, elimination of the Marine Corps' modest intelligence function in favor of a modified joint Navy-Marine or an Army-Marine capability. The former makes some bureaucratic sense, given history and the organization of the Department of the Navy, while the latter makes functional sense in that Marine and Army units often are deployed together or placed in similar situations--such as recent joint deployments to Haiti and Somalia, for example. The Army now handles much training for the Marines in skill areas like armor, airborne, artillery, and civil affairs, for example. It could do more. The Marines are virtually certain to oppose such consolidation for turf reasons and because it would tend to reduce further the credibility of their claims that the Marine Corps adds unique value as a second American army.

    The military intelligence analytic effort is large indeed, yet its contributions are arguably modestly cost effective. According to a Georgetown University study in early 1995, DIA and the services had some 13,000 people conducting analysis, versus 1,500 analysts at CIA (66). (Note the discrepancy between this figure and those on page 6--another indication of recurrent misperceptions and/or differing definitions of "analysts.")

    These people often reprocess reports received from DIA, CIA, and NSA for the periodic briefings they give their commanders and senior staff officers. Because packaging is important to commanders, military intelligence organizations often spend a good deal of their time making visually appealing viewgraphs and maps, and writing for in-house journals specially tailored to meet narrow commanders' or senior staff officers' desires. Because the content and graphical requirements are specific to each G-2, J-2, or commander, they must be redone at each headquarters. More reasonable presentational requirements alone would make substantial savings possible, and there are additional opportunities for consolidation of functions.

    Despite the large number of military intelligence personnel, there is continuing need for intelligence personnel to brief commanders, collect local HUMINT in tactical situations, and disseminate information to tactical unit commanders who need it. Arguably, these tactical intelligence duties are under staffed, and such personnel perform more poorly than they should. This reflects heavily the services' personnel policies that require officers regularly to rotate--meaning that they cannot become intelligence experts. There seems little high-level interest in remedying these problems. Indeed, they are a consequence of consciously chosen career development policies.

    At the other extreme, defense-related technical collection efforts are large and expensive. Some critics claim they are much more expensive than necessary. Yet while their technical powers are often very impressive, the services and CIA have had trouble figuring ways to get "national-level" technical collection to users in the field. Even senior, priority customers like Desert Storm commander General H. Norman Schwarzkopf complained about the speed of getting such information to him (67). Lower ranking officers with smaller support staffs have sometimes more scathing critiques. A common one is that such capable intelligence is useless to them because it is so highly classified and thus protected that it is unavailable in timely, convenient venues.

    All of these problems present major obstacles to accomplishment of the military intelligence portion of the "revolution in military affairs" that Andrew Marshall of the Secretary of Defense's Office of Net Assessments, and others, have postulated (68). Even if the formidable hardware and computer software problems associated with creation of a much more complex tactical military intelligence support system are solved, the system seems unlikely to be able to address anything but the conceptually simplest of military situations--the presence of battlefield threats. Complex questions of motive, intent, and most situations below high intensity conventional combat will be difficult to address because the Defense Department and the Intelligence Community have not invested in the knowledge and the human effort needed to create wisdom about such matters.

    Thus, there are major procedural, doctrinal, training, organizational, and procurement issues facing Department of Defense intelligence. One sure way not to resolve the issues is to add another star to the shoulders of the head of DIA--as Robert Gates and others have suggested as part of their effort to create a "Director of Military Intelligence" to consolidate military intelligence. A four-star J-2 (or head of DIA) would rank the three-starred J-3, or operations officer, who normally is the ranking staff officer. The suggestion thus would upset nearly 80 years of military tradition, rendering its acceptance by service personnel unlikely. The DMI idea is sound, but linking it to the idea of a promotion for the director of DIA would not be helpful.

    Revamp the National Intelligence Council

    The National Intelligence Council for years has been a weak body that neither advises the DCI extensively nor performs very many essential coordination functions. While DCI William Casey used the NIOs a good deal, William Webster paid little attention to the NIC in 1987-91. Deputy DCI Richard Kerr reduced its staffing levels and influence, evidently to increase the relative power of the DI, immediately after he became DDCI in 1989 (69). The NIOs' roles as drafter/coordinators of National Intelligence Estimates (NIEs) also have been modest and they have little independent authority over judgments or analytical resources--despite the publication in recent years of a small series of NIC analyses under its own letterhead.

    The NIC or a successor organization should be expanded to include a substantial staff of respected experts--from within and on temporary assignment from outside the Intelligence Community--and it should have expanded NIE drafting authority. A more capable NIC would serve several valuable functions. It could be:

  • A more intellectually potent body of advisors immediately accessible to the DCI. This is important if the DCI is to serve as a true intelligence adviser to presidents and the National Command Authority consistent with his/her status as statutory adviser to the National Security Council. In-house expertise is important because the DI, although also subordinate to the DCI, regards the NIC as a bureaucratic rival and has its own distinct agendas. The DI for years has scrupulously reviewed the written material its subordinate offices prepare for both the NIC and the DCI and monitors contacts of all sorts between NIC and DI people very carefully.

  • Capable of conducting independent research and concluding respected judgments that would make it a genuine center of independent assessments. This would serve as a check on the propensity of agencies to generate official analytic lines and to hold them in inter-agency debates for bureaucratic rather that substantive reasons. This has been a recurrent problem. The Intelligence Community may need restoration of a variant of the Board of National Estimates--the 1960s vintage predecessor of the NIC that was replaced because it developed problems as well.

  • A better clearinghouse of ideas and an inter-agency coordinator for the DCI or the head of national intelligence, whatever new name may be chosen for the position.

    At the same time, at least two current practices should be eliminated. The NIC in recent years has been to some extent a dumping ground for failed senior executives. George Kolt, for example, became NIO for Russia and Eurasia after he finally was terminated as head of the DI's Office of Soviet Analysis in 1991--a rare firing for cause. He is a good example of the failure of DCIs to ensure that good leaders and people with sound judgment occupy positions of authority. Indeed, he is a good example of the many judgmental errors Robert Gates made in selecting senior officers who continue to cause CIA major problems (70).

    Second, as is by now well known, NIEs are too often lowest-common-denominator documents that are wishy-washy or confusing or both. They too often contain collections of contradictory "footnotes" of dissenting opinions from members of the National Foreign Intelligence Board (NFIB)--the committee of agency heads that approves NIEs. Part of the problem is the extensive sub-division of the Defense Department's intelligence interests to service representatives who have parochial service interests at heart. This sub-division should be eliminated in favor of a unified Defense Department position on NIEs. There is no reason why as many as six Defense Department organizations--DIA, NSA, Army, Navy, Air Force, and Marines--should have voices on estimates at the final coordination level of the NFIB.

    The handling of NIEs is itself a major issue that warrants additional, special attention. NIEs have not consistently provided what consumers consider to be adequate intelligence service. The NIC is well aware of this perception and has periodically tried to improve performance. None of the efforts seems to have worked, and some have been silly. In the late 1980s, for example, then-NIC Vice Chairman Dave Gries ordered NIOs to include summaries of the Key Judgments of NIEs--in essence summaries of summaries for busy policymakers. Gries mandated that the summaries be no more than four sentences in length. The dictum was arbitrary, especially for complex subjects, and led the NIOs to use semi-colons extensively in effective circumvention of the spirit if not the letter of Gries' dictum. Some "sentences" contained two semi-colons and were the length of normal paragraphs. One reasonably can wonder what kind of policies are likely to be made by a senior government official who wants four-sentence synopses of NIEs because he will not devote time to read a two-page summary of an already watered-down, lowest-common-denominator paper. It is an unflattering commentary on the consumer community the NIC tries to serve.

    Establish Genuine Intellectual Competition

    It is essential that reform of organizations and ethics acts to assure objectivity of analysis. While this can be accomplished for short periods within single agencies through the strength of individual leaders, a better solution is creation of multiple analytic organizations that genuinely compete intellectually and create an institutional framework of checks and balances to ensure that bias is not systematized in any agency--or in the Intelligence Community as a whole. This was one of the intents embodied in the creation of DIA, of course. Few intelligence professionals believe DIA has been fully successful on the intellectual level, although it and CIA certainly have engaged in bureaucratic battles. Multiple analytic organizations entail some costs, but analysts and their support equipment are inexpensive compared to technical collection hardware. Leaked budget data indicate that DIA's total budget--not just the part that performs functions similar to those of CIA--is about $600 million annually. The DI has a quite small part of CIA's reported $3 billion budget, and INR's budget is tiny (71).

    I believe that one of the key institutional changes that led to the serious denigration of CIA's analytic performance in the 1980s was the elimination of key competing centers of analysis within the DI. This happened in two stages within four months of each other. In October, 1981, the Directorate of Intelligence was reorganized along geographic lines. This meant that primary responsibility for regions of the world was placed in the hands of a single office director. He oversaw the formulation of assessments within his area of responsibility and could overrule the judgments of his political scientists, economists, military analysts, and so on.

    The office directors could and often did enforce their analytic preferences through control of their offices' "career services" or personnel management systems. Those subordinate managers and analysts who cooperated received promotions and other preferential treatment, while those who resisted often received internally well-known punishments. This created internally well-known intellectual orthodoxies. It was another part of the transformation of much of the analyst corps into what the Office of Global Issues economist in 1990 called "scared rabbits."

    This conduct contrasted markedly with previous practice, which held that the heads of the political, economic, and military offices all had to concur on assessments of major issues before a DI position was established. This assured a thorough review of issues and a check on bias, albeit at a slight cost in terms of narrowly defined "efficiency."

    In January 1982, new Deputy Director for Intelligence Robert Gates persuaded DCI William Casey to remove the Senior Review Panel (SRP) from the DI and place it in support of the National Intelligence Council (72). The SRP was a distinguished, senior group of retired military officers, diplomats, academics, and scientists who offered independent assessments of the DI's work on part-time bases. Gates obviously did not want independent review of the work he would direct, control, and approve. Gates "replaced" the SRP with the Product Evaluation Staff that reported directly to him alone. It quickly showed itself to be little more than a support center that kept production statistics, however. It had no role in ensuring product quality at any level of review. Gates himself assumed that role. He announced that he was to be the final arbiter of "quality:"

    I am the final approving official for all of CIA's daily production of current intelligence that goes to the president and other senior government officials. The director [sic] of Central Intelligence sees it at the same time as the policy-maker. I also approve all longer-range assessments. (73)

    There was no role for an activist Product Review Staff that truly reviewed product quality and would lay blame for a failure on the doorstep of the boss.

    Note, in addition, that Gates effectively absolved William Casey of direct responsibility for the growing politicization of analysis then afflicting the DI. This effectively and accurately contradicts the now-conventional wisdom that Casey directly politicized analysis to suit his ideological beliefs. Casey was remiss in not effectively monitoring Gates' actions, however. He did not notice or countermand the growing DI practice of studying his public and private comments for policy views or geo-political beliefs to pander to.

    The loss of these two checks on objectivity arguably is largely responsible for the deterioration of the quality of CIA analysis. The loss is thus also a major cause of CIA's diminished standing with Congress, policymakers, and with the American citizenry.

    There are a number of ways that checks and balances on analytic quality can be ensured. They must be systematized, however. Establishment of competing centers within DI offices is one way. These would cost money but might be well worth the cost. Restoration of the pre-1981 organizational structure would help in some ways, but would restore some of the inefficiencies that prompted the restructuring in the first place.

    Another is an internal DI equivalent of the NIC Analytic Group--which provides analytic support to the NIOs and is a smaller version of the staff of the Office of National Estimates of the 1960s and 1970s. Along with cutting NIO positions, then-DDCI Dick Kerr reduced the size of the NIC/AG early in 1989 after his promotion from DDI--another move NIOs generally believed at the time was a move to advance the DI's interests over those of the NIC (74). A variant of this would be to return the Senior Review Panel to the DI or, better yet, to place it in the Office of the DCI with responsibility for reviewing DI products and mandating changes where problems were found. Just the existence of such an oversight body would help prevent the sometimes arrogant injection of personal beliefs by reviewers, particularly the office directors, that blew up into the "politicization" issue at Gates' 1991 DCI confirmation hearings.

    Former NIC Chairman Joseph Nye tried another approach in mandating his "Parallel Estimates Project" that was designed as a latter day Team A/Team B exercise. Nye envisioned outside organizations making assessments on selected topics while the Intelligence Community internally conducted its own study of the subject. One such project was a Hudson Institute prognosis for Europe through the year 2005. But Nye departed for the Defense Department shortly after the initiation of this project and his successor--CIA careerist Christine Williams--let it die by not appointing people to do the NIC half of the exercise. It was an unwanted challenge that offered potential embarrassment but little gain. The Nye approach, while potentially useful, was fatally flawed in that it had no institutional permanence.

    Still another way would be to separate the now combined functions of managers who both judge the worth of analysts and make promotion and assignment decisions through control of the career services. This dual authority has made it easy for managers in politicized offices like the Office of Soviet Analysis and the Office of European Analysis in the late 1980s to use bureaucratic incentives to push judgmental preferences. Creation of separate product review and management chains of authority would reduce sharply the danger of sanction- or incentive-generated bias. The alternative review chains could be groups of peers, other managers, or combinations of the two.

    This approach would not eliminate legitimate managerial prerogatives, however. An analogy is the organization in some corporate research centers where researchers--the brains who produce the ideas that translate into new products and profits, command high salaries, and receive peer review--are supported by sometimes less well paid managers who keep their hands off laboratory work but assure overall efficiency and accountability of the research effort.

    It matters little what the actual organization, so long as the reviewers have independence, resources and authority to replicate analyses, power to interview analysts, powers to intervene with management at analysts' requests, and power to appeal to senior management those cases of questionable office management conduct. There must be independence of competing analytic activities to ensure corporate analytic integrity and to bring out the very substantial best the Intelligence Community has to offer.

    It is critical to distinguish competing analytic functions--a critical part of intelligence--from organizational competetion using analytic judgments as weapons. The latter is little more than turf fighting. Unfortunately, such bureaucratic struggles have been the norm for part of Intelligence Community dealings for years--especially in the Soviet military area between CIA on one side and the DIA/military agency complex on the other (76).

    Some significant reform in this area is essential if the culture of the DI and the Intelligence Community as a whole is to change and if CIA analysis is to regain the confidence of policymakers. Identifying mechanisms that can restore this confidence should be a key objective of the reform effort.

    State/INR Performs Well

    The State Department's Bureau of Intelligence and Research (INR) serves a valuable role as a producer of competitive, alternative analyses, and may be the most cost-effective agency in the entire intelligence structure. It should be left alone. INR is small and inexpensive, yet its judgments generally are well respected within the Intelligence Community and with policymakers (77). Secretaries of State have legitimate need for an internal, controlled, and responsive analytic unit. It thus would be foolish to eliminate INR in the name of modest cost savings. State would just create a replacement under another moniker.

    INR's punchy writing style--particularly that of its daily Morning Summary--gain it the advantage of readability that regularly bests CIA's National Intelligence Daily in reader surveys, despite CIA's chronically massive efforts to make the NID user-friendly. (As an aside, it bears reflection that consumers--smart, sophisticated, and senior ones--judge an agency's worth heavily by the ease with summary, newspaper-format publications can be read. How much does this say about the consumer community? And how far does this go toward explaining why CIA spends so much time and effort trying to coddle stylistically demanding consumers? In some instances the medium has indeed become the message.)

    INR has problems, but these are internal ones that merit the attention of the Secretary of State, not lawmakers. Perhaps chief among them is the relative bureaucratic status of a tour of duty in INR. The bureau's personnel include both Civil Service personnel and Foreign Service Officers (FSOs), who have different career track incentive systems. While INR's personnel generally are quite good, State could move to enhance the promotional worth of an INR tour for FSOs, for example, to get good people to do intelligence/policy support work for the department and the White House. It also could move to equalize the status of FSOs and generally less well regarded Civil Service personnel.

    Indeed, the Commission and Congress might want to assess the extent to which INR's performance is enhanced by the fairly easy transitions of FSOs between policymaking tours in regional bureaus, policy implementing tours at embassies, and intelligence/policy support work at INR. There may be lessons for other analytic agencies whose analysts are to some extent accurately accused of being isolated from their policymaking consumers. The inaccurate corollary is that it is analysts' fault that they are so isolated. In fact, in my experience, the main reason is that analysts' managers insert themselves between analysts and policymakers to ensure control. Analysts would be happy to have closer contact with their clients.

    Reinvigorate Inspectors General

    Legislative action probably is needed to improve the performance of intelligence agency inspector general staffs. These should be tasked explicitly with performance-related oversight duties--not just investigation of criminal activities or adjudication of minor squabbles. The lack of authority of inspectors general is a problem not restricted to intelligence agencies, but the closure of intelligence from normal public scrutiny places a special responsibility on IGs that they have not always satisfied.

    Legislation to modify the responsibilities of CIA's statutory IG is needed to redress the chronic failures of CIA's Office of the Inspector General (OIG) to perform the role of guardian of the performance of the Agency. The enactment of statutory IG legislation that took effect in 1990 did virtually nothing except generate another Senate confirmation hearing and vote. The OIG continued during the tenure of Frederick P. Hitz to rubber-stamp poor managerial decisions and conduct, and to ignore the growing institutional malaise.

    The institutional reason, regardless of Hitz's personal preferences, is that the Inspector General has no real authority other than that granted to him by the DCI. To make matters worse, the OIG traditionally has been staffed by older managers with vested interests in the status quo. Even those relative few who are energetic know that they have little real power--making a spirited objection to problems futile but career-damaging. And, because many OIG personnel are on rotations from home offices that may have problems themselves, "troublemakers" on the OIG staff are subject to retribution by vindictive office chiefs. It is a formula for OIG impotence.

    Reform of CIA's Office of the Inspector General should cover several areas:

  • The Inspector General should receive instructions to expand the scope of his concerns to address the performance of CIA as a whole. In the past, the office has taken a legalistic approach that largely ignored performance of the organization as a whole.

  • The Inspector General should have enhanced powers to mandate remedial action, to impose sanctions unilaterally on perpetrators of inappropriate actions, or to go to directly--that is, to bypass the DCI--to either the president or the Congressional oversight committees with notification of problems. This unsual power would give OIG real authority to enforce accountability. It would be especially important if, as happened in the last decade, senior management itself was the Agency's key problem.

  • There should be a separate "career service" for OIG personnel so that they can have satisfactory careers in the inspections area. OIG people would therefore not be held hostage by potential retribution from line office chief heads of the career services to which they belong. They also would not be ignored as detailees "out of sight and out of mind. (78)"

    Revamp Security Procedures

    The entire security apparatus warrants review and reform. Heading the list of priorities is the abysmal working relationship CIA long has had with the Federal Bureau of Investigation. Mutual dislike and turf fighting have been legendary. Both parties undoubtedly are partly to blame; the FBI also has poor relations with the Secret service, I understand. The fall-out of the Ames affair and Presidential Decision Directive 24 issued in mid-1994 to better coordinate counter-intelligence may help, but remedies certainly are cultural--because the problems are largely cultural--and may require legislation as well as determined senior leadership and effective Congressional oversight.

    Within CIA, the Office of Security (OS) merits substantial reform and a purge of senior managers. OS has relied excessively on the polygraph, despite widely known and well publicized failures to identify problems. OS and many CIA people are well aware of other problems that are not yet public knowledge.

    Yet, in response to criticism of the polygraph as a result of its repeated failure to catch Aldrich Ames, OS apparently has responded by using the "box" still more. In early 1995, OS reportedly was disqualifying some 80 percent of applicants due to failure to pass the polygraph. These vetos come at the end of processing--after much time, effort, and money has been expended. According to a CIA person, the Agency was having trouble filling new Career Trainee (CT) classes due to polygraph disqualifications. Senior OS managers evidently are having a hard time learning from their mistakes and moving to more productive security approaches.

    Overzealous use of the polygraph has several negative effects. It undoubtedly rejects many perfectly loyal citizens who do not do well at polygraph testing--that is, who are innocently nervous during the exam. It rejects people whose breadth of experience makes simple "yes" or "no" answers difficult. Yet it is these very people whose foreign area expertise is much needed. The result is that young, inexperienced and poorly traveled people often pass; they have done and know too little to make them suspicious. And, knowing that security processing is a time consuming, troublesome and/or random part of the application process, many good prospects simply do not apply. The lack of regular injections of such experienced, mature, mid-career people arguably has been a major contributor to the establishment of intellectual orthodoxy in some offices and the allegations of poor performance that recurrently has dogged CIA analysis in recent years. Columnist Jack Anderson claims documentary evidence that CIA purposefully recruited obedient conformists (79). The Congress and the Commission should investigate this matter--from the standpoints of security and the Agency's de facto mission.

    In addition, overzealous CIA application of the polygraph simply makes no sense for overall national security when State and Defense do not require polygraph tests for all employees. Foreign governments interested in acquiring U.S. secrets including intelligence judgments and policies--as opposed to data on agents and operations--would be far better advised to try to breach security at those agencies or Congress. Even better might be the White House where, for example, former Clinton campaign operatives turned senior government policymakers saw little reason for most of 1993 even to bother applying for security clearances.

    It makes little sense to add more bolts to the front door--to the point where access by the home owner is negatively affected--when numerous back doors are left unlocked or wide open. Recall that Israeli agent Jonathan Pollard was a U.S. Navy employee and the Soviets managed to recruit numerous NSA officers over an extended period; these people were under Defense Department security rules. Foreign governments, especially nominal friends and major aid recipients interested in internal U.S. Government deliberations about policies toward them, might even put penetration of CIA fairly low on their priority lists. Again, much better choices are the White House, Congress, and State, which have relatively easy security standards and a relatively large flow of political appointments greased by modest campaign contributions. Turnover and therefore access is further increased by the willingness of many people to take temporarily small salaries in return for a coveted resume line. It is a wonderful opportunity to slip agents into positions with access to sensitive information.

    CIA's security apparatus also needs reform because it has become a mechanism for ensuring internal orthodoxy--not just security. In the 1980s, OS used its powers to turn on internal critics of Robert Gates and his proteges. At least one law suit is in process over this issue. The Office of Medical Services (OMS) collaborated in the effort, by evaluating internal critics for alleged psychological problems. Unable or unwilling to handle all the "investigations" internally, OMS contracted with a psychiatrist in Reston, Virginia and a psychologist in Washington, D.C., at least, to handle the work. The magnitude of the misuse of authority is unclear to me, but it has begun to leak into the public domain (80). This issue warrants serious investigation and significant remedial action.

    What Place the Intelligence Budget?

    The perceived poor performance and lack of accountability of intelligence agencies in recent years has prompted additional calls for publication of the Intelligence Community's annual appropriations. This long has been a staple of civil libertarians' demands, but recently has gained support among mainstream observers and the Congress as well. Intelligence agencies have resisted such publication on the grounds--which have some merit--that publication of budget figures would provide adversary nations and their intelligence agencies with information about resource allocations and thus a glimpse at sacrosanct "sources and methods."

    However reasonable the security argument, CIA especially has lost enough credibility that some publication of intelligence budget figures seems needed to reassure the citizenry and its Congressional representatives that the organizations are performing adequately and in an accountable manner. Moreover, from a practical perspective, there simply is such dissatisfaction with intelligence that aggregate figures at least are bound to continue to leak. Figure 2 is my summation of figures that appeared in but one newspaper article in mid-1994. Furthermore, senior Congressional and intelligence officers regularly quote percentage figures on staffing cuts, quantities of resources devoted to the Soviet Union or open-source collection, for example. These individually innocuous figures allow moderately astute citizen observers, and certainly curious foreign intelligence services, to put together the tidbits into a fairly accurate mosaic picture of the U.S. intelligence effort.

    There also are some additional advantages of release of budget figures that can help clear up serious misperceptions about intelligence and its performance held by Americans who do not study intelligence very much. These can help debunk the now common variants of public moans to the effect that "we spend $30 billion a year on CIA and all we get is a bumbling drunk like Aldrich Ames."

    Figure 2

    Intelligence Community Budget -- 1994

    Budget Share

    Agency ($ billion) (percent)

    CIA 3 11 State/INR --- --- NRO 7 25

    Defense NSA 4 14 DIA 0.6 2 TIARA 12 43 ---------------------------------------------------------------------------- TOTAL 28 100

    Note that the sum of line entries do not total $28 billion or 100 percent. "DMA" refers to the Defense Mapping Agency.

    Source: Walter Pincus, "White House Labors to Redefine Role of Intelligence Community," The Washington Post, 13 June 1994, p. A8. These figures were leaked to the press. Actual figures are classified.

    Chief among the misperceptions is the notion that CIA takes the lion's share of intelligence funding. In fact, as all accounts indicate, Defense Department agencies use the bulk of the money, and Tactical Intelligence and Related Activities (TIARA) takes the biggest share of that. TIARA funding and the monies provided under the General Defense Intelligence Program (GDIP) go for intelligence support of the war-fighters. These programs include expensive items like aircraft and radars. Much of the effort is as essential to conducting combat operations as guns, bullets, and food for soldiers. It is inherently part of the Defense establishment and should be accounted for as such--as well as funded by the Defense Department.

    A new accounting scheme should separate defense intelligence--that narrowly used by uniformed personnel--from national intelligence which supports the president, Executive Branch civilian agencies, and the Congress as well as some strategic intelligence needs of the Defense Department. This latter account is where the citizenry generally believes "spying" to be conducted. It amounts to much less than the $28 billion annual figure cited above by The Washington Post. The release of some figures by no means requires detailed listings of all operational accounts. To the extent that newly categorized national programs serve defense needs on a significant or recurring basis--as with imagery and other technical collection--the NRO and NSA, presumed now to be part of the Executive Office of the President and subordinate directly to the DCI alone, should "bill" the Defense Department in the sense that an appropriation for this activity should be placed in the Defense Department's intelligence budget (81).

    The separation of aggregate figures would serve a valuable governance function. It could help resource allocators in Congress (including the Congressional Budget Office and the General Accounting Office), the Office of Management and Budget, for example--as well as the Defense Department--better understand how much the government is spending on what--and thus help improve the efficiency of resource decisionmaking. Even the reporting of aggregate figures would help the citizenry and the press keep tabs on government decisionmakers without compromising sensitive information. HPSCI chairman Larry Combest has made the point similarly:

    If a project is done specifically for the military, I don't know if we should have that lumped into our [intelligence] budget (82).

    He properly suggests that the military weigh the merits of tactical intelligence systems versus weapons versus other defense activities and make decisions within the context of the defense budget--independent of the needs for national level intelligence systems and programs that serve non-military consumers of intelligence information and judgments.

    Exempt Intelligence From Affirmative Action

    The Congress should enact legislation to exempt the Intelligence Community--and particularly agencies' subcomponents that operate abroad--from affirmative action policies and laws. I do not suggest elimination of fair employment treatment or equal opportunity. However, the now major preoccupation with sexual and racial percentages has cast a pall throughout CIA and reduced sharply the once-strong sense of fraternity and unity of purpose. In my experience, it is now becoming a problem in the military as well. Just as in society as a whole, promotions are seen in the context of quotas, leading to suspicions about the fairness of promotions and reverse discrimination.

    Age and sex discrimination suits by CIA personnel are now with the Equal Employment Opportunity Commission or recently have been settled. An age grievance in the DS&T had origins in the dismissal of older white men to make managerial room for young blacks. Recall that former station chief Jeanine Brookner won a rare $410,000 settlement in late 1994 and another suit by women DO officers was resolved by a cash settlement and retroactive promotions in March 1995. Unhappiness with management remains pervasive, suggesting that the Agency will continue to be torn by allegations of discrimination unless there is significant remedial action--in both leadership and appeals arenas.

    However well motivated, affirmative action has become a national curse--creating competing groups that alternately feel discriminated against and want remedial discrimination at someone else's expense--that is particularly ill-suited to international intelligence work. It may make little operational sense to send a woman, for example, to try to recruit agents in an Islamic state that prohibits women from traveling alone or freely. (Or to establish a station of blacks in Scandinavia. Or to send six foot six inch blond weight lifters to southeast Asia.) We can assuage some moral national sense by sending women into such operational environments anyway. But foreigners are under no such real or imagined obligation to accept our moral codes; they will behave as their cultures tell them to behave. If women case officers are likely to be ineffective in such operational environments, the CIA effort will be damaged with uncertain impact on national security; the women's presence there also would be a waste of government resources. Moreover, poor performance--in terms of the usual criteria of recruitments made and reports submitted--is likely to show up at promotion time. The result may be low aggregate promotion figures for women--ostensibly evidence of actually non-existent sex bias.

    The point is simply to recognize that efficient field work requires operations in accordance with foreign cultural norms. We foolishly spit into the wind--as sailors might say--by demanding that CIA operations managers make assignments in accordance with domestic American standards of "fairness."

    This is not to say that CIA managers are fair--or that there is not a need for substantial improvement. The needed improvements are for all aspects of personnel management, however. I vividly recall a senior DO officer--a white male--who visibly shook at the mere discussion of crossing the executive who headed DO operations in Europe in the late 1980s. White men are treated poorly, too. They just do not have sex or race so often to hang claims of bias upon. Reforms need to be in at least three areas: selection of better managers; improved internal grievance mechanisms; and, a more effective Office of the Inspector General. There can be no healing of the breaches caused by affirmative action until the Intelligence Community, and CIA in particularly, accept a policy of genuine equality of opportunity consistent with mission requirements and the operational environment. The latter may be important for the DO and the DS&T, but is virtually irrelevant for the primarily U.S.-based Directorate of Intelligence.

    The Congress thus should give the agencies authority to act as common sense and operational efficiency dictate--and to keep social engineering out of U.S. intelligence. HPSCI Chairman Larry Combest put the point well, saying that intelligence:

    ... [P]robably has more of a need for multiracial individuals because we're dealing all over the world with people... We need to show balance, but the priority has to be on national defense and security. (83)

    Combest added:

    I think we had better be very careful not to use [the intelligence agencies] as areas of social experimentation. That should not be the goal. (84)

    Improve Congressional Oversight

    Congress should move to strengthen its understanding and oversight of intelligence. The Intelligence Community's problems developed over an extended time, and the citizenry and members of the Intelligence Community itself can understandably conclude that the lack of serious Congressional concern about the problems until 1994 indicates that Congress conducted its oversight responsibilities poorly. Better oversight does not mean micromanagement--at least of a healthy intelligence community--but it does mean better understanding of the performance and problems of intelligence. My own experience in dealing with the SSCI in 1991 in connection with the Gates confirmation hearings was that staffers, although inherently capable and apparently well-intentioned, simply did not understand the problems of CIA. Members were in worse shape. A Congressional staffer quipped in 1992, "'Oversight' has two meanings."(85) I believe his comment applied accurately to both oversight committees for years, and that, while there has been improvement since 1991, both have yet to reach optimal levels of understanding and staff expertise.

    Former SSCI Chairman Dennis DeConcini recognized the point in a May 1993 speech. Responding to comments by Robert Gates about the lack of expertise on the oversight committees, DeConcini said:

    CIA has not always played it straight with the oversight committees. Now part of that can be our fault. Mr. Gates pointed out there are very few members on the oversight and appropriations committees who appear to devote a lot of time to the process, and there is not a broad knowledge by many--I think he said he could count them on one hand. Well there is some truth to that ... (86)

    I therefore agree with Gates' proposal that Congress strengthen the oversight committees by removing term limits (87). It should also take steps to improve staff expertise. Several actions could help improve the performance of staffers:

  • Selection of more personnel with intelligence backgrounds. There is no substitute for experience.

  • Rotations for Congressional staffers to intelligence agencies, and possibly, vice verse. Care should be taken to ensure that possible intelligence agency personnel on rotation to the committees are not simply agents on information collection and recruitment missions.

  • Staffers' dealings with offices of agency inspectors general should be strengthened. This could help improve staff understanding, and also help oversee (and encourage) the internal overseers of agencies. In CIA's case, as mentioned, the Office of the Inspector General has been a major problem in recent years.

    Presidential Leadership Critical And finally, the firm involvement of the president in intelligence matters is essential to the coherent conduct of the agencies (as well as his own conduct of foreign policy). President Clinton's evident disinterest in intelligence matters and his difficulty in securing a willing replacement for DCI James Woolsey in early 1995 bode poorly for the success of the Aspin Commission and of any Clinton Administration DCI (89). Any DCI will be weakened by the absence of strong, public presidential commitment to reform. Intelligence professionals are adept at nothing if not obfuscation and bureaucratic infighting. Still, new DCI John Deutch has a good reputation for management in some quarters and has access to a large and growing body of evidence about the Intelligence Community's problems. I hope he will act strongly on his own initiative--independently but in concert with the Congress and the Commission. Certainly he promised the SSCI at his confirmation hearing that he would do so. Congress and the knowledgeable citizenry can, in many ways, help focus the president's attention on this essential component of America's foreign policy establishment; they can remind Deutch of his promises as well.

    62. Victor Marchetti and John D. Marks, The CIA and the Cult of Intelligence (New York: Alfred A. Knopf, 1974), p. 99.

    63. See L. Britt Snider, "Assessing the 1992 Intelligence Reorganization Legislation," National Security Law Report, Volume 15, Number 5, May 1993, p. 1.

    64. Angelo Codevilla, Informing Statecraft: Intelligence for a New Century, (New York: The Free Press, 1992), p. 117.

    65. This is a serious problem. NSA long has required written "requirements"--requests for information before the agency will gather data. These must be submitted in specific ways through specified channels. Because CIA analysts usually believe that actual responses may or not come, they are reluctant to spend the time learning how to work the system. This lack of "user-friendliness" contrasts markedly with the willingness of many parts of the DO to tailor collection to support on-going or project-specific analyst needs.

    66. Walter Pincus, "Military Espionage Cuts Eyed," The Washington Post, March 17, 1995, p. A1.

    67. See: Molly Moore, "War Intelligence Flawed," The Washington Post, June 13, 1991, p. A1; George Lardner, "In a Changing World, CIA Reorganizing to Do More With Less, The Washington Post, July 5, 1991, p. A9; and, U.S. House of Representatives, Committee on Armed Services, 103rd Congress, 1st Session, "Intelligence Successes and Failures in Operations Desert Shield/Desert Storm," August 1993.

    68. See, for example, James R. FitzSimonds, The Revolution in Military Affairs: Challenges for Defense Intelligence, (Washington D.C.: Consortium for the Study of Intelligence, 1995).

    69. Shortly after Kerr became DDCI in early 1989, then-NIC Chairman Fritz Ermarth held several NIC staff meetings (that I also attended) during which NIOs expressed fears about Kerr's designs to reduce the role of the NIC, and thus their responsibilities, in favor of the DI. Shortly thereafter, the several NIO positions were abolished. At the same time, the already small NIC Analytic Group, then a small group of roughly five senior analysts, was also reduced in strength. These people were largely refugees; several had run-ins with DI management, despite or perhaps because of their consid erable abilities, and were in bureaucratic exile under the personnel system of the Office of the DCI.

    70. George Kolt entered CIA directly at Gates' behest. In the early 1980s, Kolt was a serving U.S. Air Force colonel on rotational assignment to the NIC as an assistant NIO. NIC chairman Gates made him a full NIO--a position normally reserved for major generals. When Kolt was passed over for promotion to brigadier general, he retired and immediately became DDI Robert Gates' director of the Office of European Analysis.

    71. Walter Pincus, "White House labors to Redefine Role of Intelligence Community," The Washington Post, June 13, 1994, p. A8.

    72. Robert M. Gates, speech to DI personnel entitled, "Remarks to DDI Analysts and Managers," January 1981, p. 9.

    73. Robert M. Gates, "Is the CIA's Analysis Any Good?" The Washington Post, December 12, 1984, p. A25.

    74. Then-chairman of the NIC Fritz Ermarth addressed the issue repeatedly in NIC staff meetings in early 1989. I attended some of the sessions.

    75. This was the name given to a 1970s-vintage exercise in which CIA and outside experts separately examined some Soviet military issues.

    76. It is important to distinguish institutional turf battles from those of smaller offices and between individuals. Often the relations between individuals of different agencies are far better than those between their parent organizations. At the individual level, there often is useful sharing of views and mutual help.

    77. Several National Security Council senior staff officers have told me they prefer INR's Morning Summary to CIA's current intelligence publications.

    78. See also Gentry, Lost Promise, pp. 193-94 for a discussion of possible reforms of OIG.

    79. Jack Anderson and Michael Binstein, "CIA has an Image to Lose," The Washington Post, May 15, 1995, p. D20. Major parts of the article are wildly inaccurate and some are simply nonsense, but the authors claim a 28-page manual for recruiting called the "Personality Assessment System" dates from the 1970s.

    80. See Walter Pincus, "Intelligence Panel to Study Discrimination Charges," The Washington Post, July 25, 1994, p. A11. I know several people who experienced such treatment.

    81. The idea of one government agency "billing" another is not remarkable or new. Indeed, the Defense Department itself has pressed some of its subordinate agencies, particularly the arsenals and repair facilities, to conduct near private-sector contracting within the department. The goal was a good one worth emulating: to foster efficiency and sound decisionmaking. Similarly, some Pentagon offices distribute funds to the unified commands, which then implement their programs.

    82. Walter Pincus, "A 21st Century Intelligence Test," The Washington Post, February 20, 1995, p. A27.

    83. Ibid.

    84. Ibid.

    85. For observations on Congressional oversight, see Gentry, Lost Promise, chapter 5, pp. 123-42.

    86. Remarks of Senator Dennis DeConcini to the American Bar Association Standing Committee on Law and National Security, reprinted in "DeConcini Responds to Gates Remarks on Oversight," National Security Law Report, Volume 15, Number 5, May 1993, p. 1.

    87. Robert M. Gates, "A Leaner, Keener CIA," The Washington Post, January 30, 1995, p. A15.

    88. Gentry, Lost Promise, p. 203, and Gates, "A Leaner, Keener CIA."

    89. Multiple published reports claimed in late 1994 and early 1995 that choices such as Deputy Defense Secretary John Deutch and retired Admiral William Crowe rejected feelers. Clinton himself said, in naming the 17-member Aspin Commission, that its work was his "highest personal priority"--a claim hard to believe. See Walter Pincus, "President Launches 13-Month Review of Post-Cold War Intelligence Needs," The Washington Post, February 3, 1995, p. A20.

    Back to Table of Contents