Foreign Relations of the United States, 1969–1976, Volume II, Organization and Management of U.S. Foreign Policy, 1969–1972

229. Memorandum From the President’s Deputy Assistant for National Security Affairs (Haig) to the President’s Assistant for National Security Affairs (Kissinger)¹

Washington, March 27, 1971.

SUBJECT

• Review of the Intelligence Community

Attached is the Presidential approval of yours and Shultz’s joint memorandum on intelligence reorganization. Per your instructions and with the agreement of Shultz, the PFIAB was given a copy of the basic study with the review of the various options. It has not seen this package.

As I told you earlier, knowledge of this package has leaked and the Intelligence Community, as well as the PFIAB, are buzzing with its implications. I told George Shultz that the package was returned approved and that we are holding it in strict confidence here, pending comments from the PFIAB on the study itself.

I do not believe we will ever get this study off the ground unless strong Presidential clout is put behind it. I am also somewhat fearful that some of the recommendations such as the establishment of a DDI may prove to be unworkable. For this reason, however we proceed from here we must have the President’s man in overall charge of the reorganization effort. I agree with you that Jim Schlesinger might be an ideal person to work directly for the President and you in implementing this package. If this arrangement is adopted, you should do so with your eyes open. This package is going to be the most controversial gutfight in recent bureaucratic experience. It is likely that you will have the PFIAB against you, the military, the entire Intelligence Community and a full array of Congressional opponents. The most important thing I can think of at the moment is organizing to organize.²

Attachment

Washington, undated.

Memorandum From the President’s Assistant for National Security Affairs (Kissinger) and the Director of the Office of Management and Budget (Shultz) to President Nixon³

SUBJECT

• Review of the Intelligence Community

Your expression of serious concern about the performance and cost of the intelligence community has led to a detailed review of the community’s present structure and a searching analysis of how you might improve its operation. The review shows that the overriding weakness of the community is the absence of authoritative leadership. Appearances to the contrary, the community lacks both a leader and a staff unit—intimately involved with programmatic issues but without a vested interest in any particular collection system or program—to solve recurrent problems of overlap and duplication, make efficient choices between new and competing systems, and consider the balance between various methods of collection and production. The Director of Central Intelligence (DCI), while nominal head, suffers from his many burdensome and often conflicting roles; in addition he directly controls only 15 percent of the community’s current budget of [dollar amount not declassified]. The United States Intelligence Board (USIB)—in which the DCI is assisted by the heads of the National Security Agency, the Defense Intelligence Agency, the Service intelligence chiefs, and State/INR—has also proved incapable of decisive action with respect to community matters. Even within the Department of Defense, which contains 85 percent of the resources, effective direction and control is lacking.

In the absence of leadership, the community has accumulated a host of serious problems. The National Security Council Intelligence Directives (NSCIDs) which established the division of responsibilities and govern community relations are largely obsolete. The blurring of traditional community boundaries has caused its activities to become increasingly fragmented and functionally disorganized. Under the impact of new technology, a generally unhealthy and costly competition in collection has been fostered, largely without guidance or planning. At the same time, the concerns of the intelligence consumer and the welfare of the producer have been relegated to the sidelines.

These conclusions are not new. Presidential commission, the PFIAB, special study groups, and BOB/OMB have often expressed dissatisfaction with the performance and cost of the intelligence community and recommended various piecemeal organizational reforms. In the absence of forceful and persistent leadership and under the impact of continuing technology change, these reforms have largely failed. We believe that the solution to the community’s problems and achievement of your objectives requires major changes in the organization and functioning of the community. These changes must attack three major problems: (1) the leadership of the community as a whole, (2) direction and control of Department of Defense intelligence activities, and (3) the division of functional responsibilities.

We recommend the following major changes:

• • The role of the DCI should be modified and CIA restructured so that they are separated from direct responsibility for the conduct of intelligence collection and covert action operations which, along with residual activities, would be inherited by a new agency; the DCI would acquire a strong Presidential mandate to plan, program, and review all intelligence resources and continue to produce all national intelligence required by the President, the NSC, and other national level consumers. A reorganization plan for the new agency is required.
• • A Director of Defense Intelligence should be created in order to direct and control all Defense intelligence resources; the DDI would concentrate on rationalizing Defense collection and processing capabilities. Substantive intelligence production would remain under the Defense Intelligence Agency, which would continue to report through the Joint Chiefs of Staff.
• • The NSCIDs should be revised to redraw current functional boundaries among intelligence agencies; in particular NSCID #1 should be changed to make the United States Intelligence Board strictly an advisory body to the DCI.

As the attached report indicates, there are other courses of action, both more and less radical, that you could take in light of your objectives. We believe, however, that our recommendations constitute the right steps at this time. They focus on the establishment of leadership centers both within the community as a whole and within the Department of Defense; they require minimum legislative action and cause minimum disruption of the community commensurate with your concerns; and they leave you the option of further modification later.

If you approve the recommended actions, we will prepare the necessary implementing letters and directives.⁴

Henry A. Kissinger
George P. Shultz

Attachment

Washington, March 10, 1971.

Study Prepared by the Staffs of the Office of Management and Budget and the National Security Council⁵

A REVIEW OF THE INTELLIGENCE COMMUNITY

I. Introduction: The Costs and Benefits of Intelligence

The operations of the intelligence community have produced two disturbing phenomena. The first is an impressive rise in their size and cost. The second is an apparent inability to achieve a commensurate improvement in the scope and overall quality of intelligence products.

During the past decade alone, the cost of the intelligence community has [less than 1 line of source text not declassified]. At the same time, spectacular increases in collection activities have occurred. Where satellite photography is concerned, the increases have led to greatly improved knowledge about the military capabilities of potential enemies. But expanded collection by means other than photography has not brought about a similar reduction in our uncertainty about the intentions, doctrines, and political processes of foreign powers. Instead, the growth in raw intelligence—and here satellite photography must be included—has come to serve as a proxy for improved analysis, inference, and estimation.

The following report seeks to identify the causes of these two phenomena and the areas in which constructive change can take place. Its principal conclusion is that while a number of specific measures may help to bring about a closer relationship between cost and effectiveness, the main hope for doing so lies in a fundamental reform of the intelligence community’s decisionmaking bodies and procedures.

This conclusion is advanced in full recognition that reorganization will, at best, only create the conditions in which wise and imaginative leadership can flourish. In the absence of reorganization, however, the habits of intelligence community will remain as difficult to control as was the performance of the Department of Defense prior to the Defense Reorganization Act of 1958.

II. Cost Trends

To understand the phenomenon of increasing costs, it is necessary to consider briefly the organizational history of the intelligence community. The National Security Act of 1947 and the National Security Council Intelligence Directives (NSCIDs) of the late 1940s and early 1950s established the basic division of responsibilities among agencies and departments. This division had its origins in traditional distinctions between military and non-military intelligence, between tactical and national intelligence, and between communications (COMINT) and non-communications (or agent) intelligence. Thus, CIA was directed to employ clandestine agents to collect “non-military” intelligence and produce “national” intelligence. The Department of State was made responsible for the overt collection of “non-military” intelligence. The National Security Agency (NSA) was established to manage COMINT collection. The Military Services were instructed to collect “military” intelligence as well as maintain tactical intelligence capabilities for use in wartime. All were permitted to produce “departmental” intelligence to meet their separate needs. While not ideal, this division of functions and responsibilities worked reasonably well into the mid-1950s.

Since that time, these traditional distinctions and the organizational arrangements which accompanied them have become increasingly obsolescent. The line between “military” and “non-military” has faded; scientific and technical intelligence with both civilian and military applications has become a principal area of endeavor for almost all intelligence organizations. Similarly, under the old distinctions, the national leadership—namely the President and the NSC—concerned itself with “national” intelligence, while presumably only battlefield commanders cared about tactical intelligence. But a rapidly advancing technology which has revolutionized the collection, processing, and communication of intelligence data casts doubt on the validity of the distinctions.

Simultaneously, technological advances have created new collection possibilities which do not fit conveniently within a structure based on traditional distinctions and were not covered in the original directives. Satellite photography, telemetry intercept, electronic intelligence (ELINT), acoustic detection, and radar have become some of the most important and vital methods of intelligence collection not currently covered by any uniform national policy.

The breakdown of the old distinctions and the appearance of new collection methods has been a simultaneous process raising a host of questions about intelligence organization. Is ELINT related to COMINT, is it technical or military in nature, is it of primary interest to tactical or national consumers? [2 lines of source text not declassified] Is telemetry more similar to COMINT or to ELINT; who should analyze it? Who should be responsible for satellite photography? On the more mundane, but nonetheless critical level, questions arise about the organizational responsibilities for such topics as Sihanoukville supply infiltration, VC/NVA order of battle, and missile deployments in the Suez Canal area. Are these military or non-military issues? Is the intelligence about them tactical or national? Who should be responsible for collection and what collection resources should be tasked?

In the absence of an authoritative governing body to resolve these issues, the community has resorted to a series of compromise solutions that adversely affect its performance and cost. In general, these compromises have favored multiple and diffuse collection programs and the neglect of difficult and searching analytical approaches. The most serious of the resulting problems are outlined below in brief form, and discussed in more detail in the appendices.

1. The distribution of intelligence functions has become increasingly fragmented and disorganized.

• • The old distinctions among national, departmental, and tactical intelligence are out of date. Today, CIA is as likely to produce intelligence relevant to, say, NVA/VC order of battle as DIA or MACV, just as MACV produces many reports that are of interest to the national leadership.
• • Similarly, the relatively neat ordering of collection functions that existed after World War II has broken down. CIA now engages in a wide range of collection activities—aircraft and satellite photography, ELINT, COMINT, radar, telemetry as well as clandestine, and overt agent collection. NSA has added telemetry and ELINT to its COMINT capabilities. The Services now have a full panoply of sensors to perform a variety of functions—tactical intelligence, surveillance, early warning, and so on.

Table I illustrates how almost all major components of the intelligence community are involved in each of its various collection and production functions.⁶

2. The community’s activities are dominated by collection competition and have become unproductively duplicative.

• • [2½ lines of source text not declassified] Despite past massive increases in the collection of photography, COMINT, ELINT, radar and other sensor data, sizeable additional collection capabilities are planned to become operational this calendar year: [4 lines of source text not declassified]. Other new collection systems are scheduled to be started in 1972.
• • The blurring of traditional boundaries has encouraged community members to engage in a competitive struggle for survival and dominance, primarily through new technology, which has resulted in the redundant acquisition of data at virtually all levels—tactical, theater command, and national.
• • Gross redundancies in collection capabilities have become commonplace as exemplified by aircraft in both CIA and Defense which collect photography, and by aircraft which compete with satellites in the collection of ELINT.
• • Collection capabilities remain in operation beyond their useful lives. As older systems lose their attractiveness at the national level, they are taken over at the command or tactical level where they duplicate higher level activities or collect data of little value.
• •Simultaneously, compartmentalization within various security systems has served to hide or obscure competitive capabilities from evaluation, comparison, and tradeoff analysis.

3. The community’s growth is largely unplanned and unguided.

• • Serious forward planning is often lacking as decisions are made about the allocation of resources.
• • The consumer frequently fails to specify his product needs for the producer; the producer, uncertain about eventual demands, encourages the collector to provide data without selectivity or priority; and the collector emphasizes quantity rather than quality.

4. The community’s activities have become exceedingly expensive.

• • The fragmentation of intelligence functions and the competitive drive for improved collection technology are important reasons why the cost of intelligence [1 line of source text not declassified].
• • A significant part of this cost growth is attributable to the acquisition of expensive new systems without simultaneous reductions in obsolescent collection programs.
• • In the absence of planning and guidance, internally generated values predominate in the community’s institutions. These values favor increasingly sophisticated and expensive collection technologies at the expense of analytical capabilities.
• • Few interagency comparisons are contemplated. Potential tradeoffs between PHOTINT and SIGINT, between PHOTINT and HUMINT, and between data collection and analysis are neglected.
• • While the budgetary process might be used to curb some of the more obvious excesses, it cannot substitute for centralized management of the community.

III. Questions About the Product

In a world of perfect information, there would be no uncertainties about the present and future intentions, capabilities, and activities of foreign powers. Information, however, is bound to be imperfect for the most part. Consequently, the intelligence community can at best reduce the uncertainties and construct plausible hypotheses about these factors on the basis of what continues to be partial and often conflicting evidence.

Despite the richness of the data made available by modern methods of collection, and the rising costs of their acquisition, it is not at all clear that our hypotheses about foreign intentions, capabilities, and activities have improved commensurately in scope and quality. Nor can it be asserted with confidence that the intelligence community has shown much initiative in developing the full range of possible explanations in light of available data. Among the more recent results of this failure to acknowledge uncertainty and entertain new ideas in the face of it, has been a propensity to overlook such unpleasant possibilities as a large-scale exploitation of Sihanoukville by the NVA to transship supplies, a continuation of the SS–9 buildup and its possible MIRVing, or Soviet willingness to invade Czechoslovakia and put forces into the Middle East.

Difficulties of this kind with the intelligence product are all the more disturbing because the need to explore and test a number of hypotheses will, if anything, expand as the Soviets project their military power and come to play a more direct global role. Yet there is no evidence that the intelligence community, given its present structure, will come to grips with this class of problems.

The community’s heavy emphasis on collection is itself detrimental to correcting product problems. Because each organization sees the maintenance and expansion of its collection capabilities as the principal route to survival and strength with the community, there is a strong presumption in today’s intelligence set-up that additional data collection rather than improved analysis, will provide the answer to particular intelligence problems. It has become commonplace to translate product criticism into demands for enlarged collection efforts. Seldom does anyone ask if a further reduction in uncertainty, however small, is worth its cost.

The inevitable result is that production remains the stepchild of the community. It is a profession that lacks strong military and civilian career incentives, even within CIA. The analysts, with a heavy burden of responsibility, find themselves swamped with data. The consumers, at the same time, treat their product as a free good, so that demand exceeds supply, priorities are not established, the system becomes overloaded and the quality of the output suffers. As if this were not enough, production, instead of guiding collection, is itself guided by collectors and the impetus of technology. Since the military are the principal collectors, they are more likely to focus on the needs and interests of their own Services than on the issues of concern to the national leadership, and they continue the wasteful practice of counterpart targeting. Under such difficult conditions, it is not surprising that hypotheses tend to harden into dogma, that their sensitivity to changed conditions is not articulated, and that new data are not sought to test them.

IV. Organizational Dilemmas

Questions about cost and product might exist even if the intelligence community possessed strong leadership. It is noteworthy, however, that they have arisen under conditions the most marked of which is a lack of institutions governing the community with the authority and responsibility to resolve issues without excessive compromise, allocate resources according to criteria of effectiveness, and consider the relationship between cost and substantive output from a national perspective.

This lack of governing institutions stems fundamentally from the failure of the National Security Act of 1947 to anticipate the “constitutional” needs of a modern and technologically complex intelligence community. The primary intent of the Act, understandably, was to prevent a recurrence of the intelligence confusions and delays that occurred prior to Pearl Harbor. These problems were seen as having resulted from defects in the central processing, production, and dissemination of intelligence. The critical need, accordingly, was to create an organization which would have access to all intelligence and report its estimates to the national leadership.

In 1947, the size and cost of individual programs were relatively small, and the scope and nature of the management problems associated with today’s community were not anticipated. Consequently the issue of how to plan and rationalize the collection of intelligence did not seem of great moment, and the Act did not explicitly provide for a mechanism to perform these functions or evaluate the scope and quality of its product.

There is another reason why the 1947 Act did so little to provide strong leadership for the community: powerful interests in the Military Services and elsewhere opposed (and continue to oppose) more centralized management of intelligence activities. Partly, this opposition arises from the belief of the Services that direct control over intelligence programs is essential if they are to conduct successful military operations; partly, it results from bureaucratic concerns. The Services are reluctant to accept assurance that information from systems not controlled by them will be available as and when they require it.

Despite such opposition, the National Security Act of 1947 did stipulate that the CIA would coordinate the “intelligence activities” of the Government under the direction of the National Security Council. However, the Act also made clear provision for the continuation of “departmental intelligence.” Since then, three Presidents have exhorted the Director of Central Intelligence (DCI) to play the role of community leader and coordinator, but his authority over the community has remained minimal. While the DCI has been the catalyst in coordinating substantive intelligence production, he has made little use of such authority as he possesses to manage the resources of the community.

Realistically, it is clear that the DCI, as his office is now constituted, cannot be expected to perform effectively the community-wide leadership role because:

• • As an agency head he bears a number of weighty operational and advisory responsibilities which limit the effort he can devote to community-wide management.
• • He bears a particularly heavy burden for the planning and conduct of covert actions.
• • His multiple roles as community leader, agency head, and intelligence adviser to the President, and to a number of sensitive executive committees, are mutually conflicting.
• • He is a competitor for resources within the community owing to his responsibilities as Director of CIA, which has large collection programs of its own; thus he cannot be wholly objective in providing guidance for community-wide collection.
• • He controls only [number not declassified] percent of the community’s resources and must therefore rely on persuasion to influence his colleagues regarding the allocation and management of the other [number not declassified] percent, which is appropriated to the Department of Defense. Since Defense is legally responsible for these very large resources, it feels that it cannot be bound by outside advice on how they should be used.
• • The DCI is outranked by other departmental heads who report directly to the President and are his immediate supervisors on the National Security Council.

In spite of these handicaps, the DCI has established several institutional devices to assist him in leading the community. They are the National Intelligence Program Evaluation Staff (NIPE) and the National Intelligence Resources Board (NIRB). However, the principal agencies have largely ignored or resisted the efforts of management by these bodies. As a consequence, the NIPE and the NIRB have concentrated on developing improved data about intelligence programs and better mechanisms for coordination. Because of their work, both institutions could prove useful to a strong community leader; however, their contribution to the efforts of the currently constituted DCI is small.

In the absence of an effective institutional framework within which one official could be held responsible and accountable for the performance and cost of the intelligence community, the United States Intelligence Board (USIB), originally established to advise the DCI, has become a sort of governing body for the community. However, the USIB has proved generally ineffective as a management mechanism for several reasons:

• • It is a committee of equals who must form coalitions to make decisions.
• • It is dominated by collectors and producers who avoid raising critical questions about the collection programs operated by their colleagues.
• • As a result, USIB’s collection requirements—which are an aggregate of all requests, new and old—mean all things to all agencies, thus leaving them free to pursue their own interests.
• • Since policy-level consumers are not represented on the Board, they are unable to give guidance as to priority needs.

Even within the Department of Defense, there is no centralized management of intelligence resources and activities. Although the Assistant Secretary for Administration has been given a responsibility in this area, together with a small staff for resource analysis, his efforts to master the Defense intelligence complex have proved of little avail for several reasons. First, not all Defense programs come under his purview, and this limits his ability to do cross-program analysis. Second, he remains responsible for his functions as Assistant Secretary for Administration.

Below the level of review provided by an Assistant Secretary, management leadership is still absent. The Directors of DIA and NSA are themselves unable to control the activities of the components supposedly subordinate to them but operated by the Military Services. Because of a history of compromises and “treaties,” the Director of the National Reconnaissance Office (NRO) is similarly unable to control a large part of his program which is run by the Deputy Director for Science and Technology (DD/S&T) in CIA.

This lack of lower-level leadership shows up in the following ways:

• • The current failure of NSA adequately to direct Service cryptologic activities, organize them into a coherent system, or manage ELINT activities.
• • Large-scale Service-controlled tactical intelligence assets, inflated by the war and partly duplicating both national and allied capabilities, but programmed and operated outside of the community.
• • A host of unresolved problems concerning organization and the allocation of resources within both General Defense Intelligence Program (GDIP) and non-GDIP activities, including: duplication in the collection of ELINT between NRO and SAC; internally overlapping activities among varous mapping, charting, and geodesy agencies, and the several investigative services; and inadequate supervision and control of counterintelligence activities.

It follows from this analysis that the President’s objectives can be achieved only if reform addresses four organizational issues:

• • The leadership of the intelligence community as a whole.
• • The direction and control of Defense intelligence activities.
• • The division of functions among the major intelligence agencies.
• • The structuring, staffing, and funding of the processes by which our raw intelligence data are analyzed and interpreted.

V. Specific Organizational Issues

The effectiveness and efficiency of the intelligence community depend on a number of organizational variables. Among the most important of these variables are:

• • The power over resources available to the leader of the community. How much power the leader can exercise, particularly over collection programs, will determine the size of the economies that can be achieved within the community.
• • The size and functions of the staff provided to the leader of the community. The effectiveness of a national intelligence leader will depend not only on his power over resources, but also on how well informed he is about issues and options within the community, which, in turn, is a function of his immediate staff. Among the potential functions for such a staff are:
  • —The planning, programming, and budgeting of resources.
  • —Control over resources once allocated.
  • —Supervision of R&D.
  • —Inspection of ongoing programs.
  • —Production and dissemination of national estimates.
  • —Net assessments of U.S., allied, and opposing capabilities and doctrines.
• • The future role of the United States Intelligence Board (USIB). As matters now stand, the USIB is both a parliament and a confederate head of the community. If more authoritative leadership is established, the USIB could become simply an obstruction unless its role is specifically redefined. Since the leader of the community, however powerful, will need close and continuing relationships with producers and collectors as well as consumers, one possibility would be to reconstitute the USIB so as to formalize these relationships on an advisory basis. In any case the future role of USIB should be addressed as part of a comprehensive review of new institutional arrangements for the functioning of a reorganized intelligence community.
• • Future Defense Department control over the resources under its jurisdiction. Even without changes in the community as a whole, major improvements in effectiveness and efficiency could be achieved if Defense were to master its own massive intelligence operations. However, a number of community-wide issues would still remain; and substantially firmer Defense management of its intelligence resources could prejudice the ability of a future leader of the community to exercise his own authority.
• • The jurisdiction of either a national leader or a Defense leader over the Military Services. The three Military Services are estimated [less than 1 line of source text not declassified] on intelligence activities apart from their support of the national agencies. Yet these activities, which partly duplicate national intelligence programs, are reviewed in isolation from them. If the Services retain control over the assets for this “tactical” intelligence, they can probably weaken efforts to improve the efficiency of the community. At the same time, there is little question about their need to have access to the output of specified assets in both peace and war. How to combine overall resource management and control with this access is an issue that will require resolution.
• • The future functional boundaries of the major intelligence agencies. Collection and production activities do not now tend to be consolidated by type in particular functional agencies. Important economies can probably be achieved by rationalizing these activities. However, it should be noted that economy and organizational tidiness, without concomitant strengthening of the community leadership, might be achieved at the cost of creating even more powerful vested interests and losing diverse and usefully competitive approaches to collection problems.
• • The number and location of national analytical and estimating centers. The national estimating machinery no doubt will have to be preserved under the leader of the community in order to continue production of national estimates and inputs to the NSSM process. The continuation of DIA and the State Department’s Bureau of Intelligence Research (INR) as producers is essential as well. Beyond that, improvement in the intelligence product will probably depend to a large extent on increasing the competition in the interpretation of evidence and the development of hypotheses about foreign intentions, capabilities, and strategies. This may require not only the strengthening of existing organizations, but perhaps the addition of new estimating centers. In addition, some entirely new organizational units may be needed to perform currently neglected intelligence analysis functions, for example, to conduct research on improved intelligence analysis methods and techniques.
• • The role of the independent review mechanisms. Because of the secrecy surrounding the operations of the intelligence community, the need for strong independent review mechanisms within the Executive Branch remains particularly important. Since the President’s Foreign Intelligence Advisory Board (PFIAB), the “40” Committee, the Office of Science and Technology (OST), and the Office of Management and Budget (OMB) already exist to perform this function, the only issues are how they can be strengthened, to what extent they need larger and more permanent staffs, and whether new review boards should be created, especially to evaluate the analytical and estimating activities of the community.

Subsequent sections do not address all of these issues; nor do they exhaust the list of organizational possibilities. Only the most salient options are presented with respect to the leadership of the community, the Department of Defense, and functional reorganization. Each is described in schematic form.

VI. Leadership of the Community

The effectiveness of a new leader of the community will depend critically on his ability to control intelligence resources and make his decisions stick. Basically, there are three different roles he can play in this respect, each with different organizational implications. They are:

• • As legal or direct controller of all or most intelligence resources.
• • As de facto manager of most resources even though they are not appropriated to him.
• • As coordinator of resources that are appropriated elsewhere, as now.

Although each of the three basic approaches could be institutionalized in a number of different ways, the principal options that accord with these roles are listed below.

A Director of National Intelligence (Option #1), with the bulk of the [dollar amount not declassified] intelligence budget appropriated to his office. That office would control all the major collection assets and research and development activities, which are the most costly programs of the community and are most likely to yield large long-term savings. The Director would also operate the Government’s principal production and national estimating center and retain the CIA’s present responsibility for covert action programs. Defense and State would retain production groups, both to serve their own leadership and to provide competing centers in the analysis of intelligence inputs to the national intelligence process. The Defense Department would maintain budgetary and operational control over only the selected “tactical” collection and processing assets necessary for direct support of military forces, although these assets should be subject to the DNI’s review.

This option affords a number of advantages:

• • It pinpoints responsibility; the President knows who is in charge.
• • It permits major economies through rationalization of the community’s functions and through the elimination of duplicative and redundant capabilities.
• • It establishes a management system which can deal comprehensively with the implications of evolving technology and make efficient choices between competing collection systems.
• • It brings producers and collectors closer together and increases the probability that collectors will become more responsive to producer needs.
• • It allows the Director to evaluate fully the contribution each component makes to the final product, enabling ready identification of low performance elements and permitting subsequent adjustments to their mission.
• • It provides one responsible point in the community to which high-level consumers can express their changing needs.
• • It facilitates the timely selection and coordination of the intelligence assets necessary to provide intelligence support to the President in periods of crisis.

Creation of a DNI has at least five potential disadvantages:

• • It gives still further responsibilities to the DCI. A major criticism of the present confederate organization is that the DCI is overloaded and cannot be expected to perform well the many functions now assigned to him. As noted, these include substantive advice to the President and to several high-level committees, day-to-day management of a large operating program, appearing as a witness before Congress, and running numerous sensitive collection and covert action projects. It should be noted, however, that with adequate staff and competent deputies, the Director should be able to delegate responsibilities and ease his task. Also, under this option, the DCI’s power would be commensurate with his present responsibilities.
• • This option could generate substantial resistance from the Secretary of Defense and the Joint Chiefs over the transfer of intelligence functions to a new agency. It would also necessitate fundamental changes in the National Security Act which might cause major congressional resistance and open debate on a range of sensitive national security issues.
• • Even if all U. S. Government intelligence assets were transferred to the Director, there would remain the serious and continuing problem of finding ways to meet the intelligence needs of Defense without, at the same time, causing the Services to reconstitute their own intelligence activities, even at the expense of other programs.
• • There could be adverse reaction from the news media and the public to a consolidation of such sensitive activities under the control of one man, even though so many of them already are controlled, in principle, by the Secretary of Defense.
• • It is possible that this option will continue the present dominant influence of collectors relative to producers and consumers in the intelligence process.

A Director of Central Intelligence (Option #2), with a strong Presidential mandate and a substantial staff. NSA, NRO, and DIA would remain under present jurisdiction. The CIA would be divided—one part supplying the DCI staff and intelligence production component, the other part, principally current CIA collection organization, comprising a new agency under a separate director. The DCI would have senior status within the Government and would serve as principal intelligence adviser to the NSC. He would produce all National Intelligence Estimates and other national intelligence required by top level national decisionmakers, and would control the necessary production assets, including NPIC. This would include continued management of a national intelligence process that involved the participation, and inputs from, other intelligence production organizations.

Under Presidential directive, the DCI would review and make recommendations to the President on the Intelligence plans, programs, and budgets of his own office, a reconstituted CIA, and the Department of Defense. He would also present a consolidated intelligence budget for review by the OMB. By this means the Director would be able to guide resource allocation and influence community organization.

Although Option #1 offers the greatest promise of achieving the President’s objectives, this option has advantages over it and over the present situation in the following respects:

• • The DCI would be freed from the day-to-day management tasks incumbent upon the head of a large operating agency with major collection and covert action responsibilities. This would enable him to devote most of his attention to substantive intelligence matters, the tasking of collectors, and community resource management issues as they relate to his production activities.
• • This option eliminates the present situation in which the DCI serves as both advocate for agency programs and judge in community-wide matters, a role which diminishes the community’s willingness to accept his guidance as impartial.
• • The reforms could be accomplished, without major legislation, by a reorganization plan and Presidential directives to the DCI, the Secretary of Defense, and the head of CIA.
• • This option would offer improvements in efficiency and effectiveness without the major disruptions in the community required under option one.
• • It would enhance the stature of the community leader while avoiding the potentially dangerous concentration of power inherent in option one.

Option #2 has several potential disadvantages:

• • Responsibility for the community as a whole would be more diffuse than under option one.
• • The ability of the DCI to supervise the detailed activities of the operating parts of the community would be weaker.
• • The new DCI, compared to the DNI under option one, would have to rely on persuasion and the process of budgetary review rather than directive authority in order to eliminate redundant and duplicative activities, resolve trade-off issues, and reduce overhead.
• • He would lack the ability to mobilize, deploy, and target collection assets in a time of crisis, unless given specific Presidential authority.

A Coordinator of National Intelligence (Option #3), who, under Presidential mandate, would act as White House or NSC overseer of the Intelligence Community, directing particular attention to:

• • Intelligence resource and management issues.
• • Representing the concerns and needs of national policy level consumers.
• • Evaluating the suitability of intelligence output in light of consumer demand.

Under this arrangement, CIA, Defense, and State intelligence responsibilities would remain essentially unchanged. The Coordinator would express the views and concerns of the President and the National Security Council on product needs and quality; he would provide guidance on present and future collection priorities; he would critique and evaluate the current performance of the community, identifying gaps and oversights; and he would conduct studies of specific intelligence community activities as required. But he would not be responsible for the actual production of intelligence. Nor would he have any direct control over resources.

This option offers two advantages:

• • The creation of this position would provide a means for more direct representation of Presidential interest in the Intelligence Community. Consumer representation in the intelligence process would be enhanced.
• • No legislation would be required, and the President would be spared a number of bureaucratic battles.

The option has several marked disadvantages:

• • There is the potential for unproductive competition between the Coordinator and the White House staff.
• • Achievement of the President’s management and resource control objectives is unlikely.

VII. Department of Defense Leadership

Although the President has indicated his desire to institute community-wide reform, changes within the Department of Defense alone could improve the allocation and management of resources and reduce the overall size of the intelligence budget. Provided that care is taken in making them, these reforms need not be incompatible with subsequent decisions about the governance of the community as a whole.

Within the Department of Defense, there has never been an individual with formal responsibility for management of all DoD intelligence activities. The Deputy Secretary of Defense historically has been charged with this task, but he has very little staff to assist him and can devote only a modest amount of time to the complex intelligence issues that arise within his domain. Consequently, if the problems of Defense intelligence are to be resolved in a fashion satisfactory to the President, it will be necessary either to create a Director of Defense Intelligence (DDI) with specific responsibility for the Department’s collection assets, or provide the Deputy Secretary with major staff support in the form of an Assistant Secretary of Defense for Intelligence.

Neither of these posts would be incompatible with options two and three relating to community-wide leadership reform. However, the DDI concept conflicts with option one, in which the bulk of U.S. intelligence resources would be appropriated to a Director of National Intelligence.

A Director of Defense Intelligence would have the authority and responsibility to direct and control all Defense intelligence activities. He would allocate all the Defense intelligence resources, including those for tactical intelligence, the funds for the NRP, and budgets for other national programs under departmental jurisdiction. He would report to and represent the Secretary of Defense in all matters relating to the management of intelligence resources; review the need for, and conduct of, sensitive intelligence collection and operations; review all Defense intelligence “requirements” with resource implications in order to evaluate need and determine priorities; serve as the principal Defense representative on the USIB; and monitor other DoD programs which have clear implications for the collection of intelligence. Under this option the DDI would be able to reorder completely the Defense intelligence collection structure as deemed appropriate.

The DIA would be involved in collection management only if so directed by the DDI, and would concentrate on the production of finished intelligence for the Secretary of Defense and other national consumers.

It is important that the Director of Defense Intelligence be responsive to tasking by the community leader, who would be the principal substantive intelligence official of the Government. Both the community leader and the DDI should receive authoritative guidance about national consumer interests. This could be provided by a Council of Intelligence constituted within the NSC and with the Assistant to the President for National Security Affairs, the Secretary of State, and the Secretary of Defense as its members. The restructuring of USIB and revision of NSCIDs can help in establishing the appropriate DCI/DDI relationship:

The post of DDI has great prospective advantages:

• • It would provide for the concentration of resource management authority in one individual, which would allow authoritative comparisons and decisions about competing collection programs.
• • It would provide for the centralization of direction and control over all Defense intelligence activities, including conduct of sensitive intelligence collection operations.

But there are possible drawbacks as well, in that the position would:

• • Concentrate great power at a single point in Defense. This could possibly diminish the community leader’s access to information, as well as his ability to task collection systems in support of national intelligence production, and design balanced collection programs, in support of his production responsibilities.
• • Superimpose a large staff over those of other major intelligence managers within Defense (the Directors of DIA, NSA, and NRO), although a reduction in various coordination staffs should be possible at the same time.

An Assistant Secretary of Defense for Intelligence (ASD/I) who would act as the principal staff assistant to the Secretary of Defense. His responsibilities would be similar to those of the DDI, except that he would not exercise direct control over Defense intelligence collection programs, and would not be a member of USIB unless the Board were reconstituted to advise the DCI on the allocation of collection resources.

This option has a number of advantages:

• • It allows for effective cross-program analysis within Defense.
• • It avoids the concentration of power inherent in the DDI option, if that is considered a danger.
• • Compared to the DDI, an ASD/I would be more likely to respond to the needs of the present DCI or the community-wide leader established under either option two or three.

The post has a number of potential weaknesses in that, compared with the DDI, it would probably:

• • Lack both the strong mandate provided to the DDI and direct authority over Defense intelligence activities, including those carried out by the program managers.
• • Make the ASD/I vulnerable to “end runs” by major components within the Defense intelligence community who might wish to appeal directly to the Deputy Secretary of Defense.

VIII. Changing Functional Boundaries and Costs

During the past two years, the budget of the Intelligence Community has been reduced, measured in constant and current dollars, as shown in the following chart:

[chart not declassified]

To achieve further economies, particularly without major reorganization, will be difficult for several reasons.

• • Savings that we foresee as immediately feasible are likely to be counterbalanced to a considerable degree by further pay and price increases.
• • With the heavy R&D costs for proposed new systems, such as the nearly real-time photo satellite, [less than 1 line of source text not declassified] there already is built into the budget a strong upward bias which may prove difficult to control, particularly considering the intense interest in high-technology and expensive new systems for SALT and other purposes.
• • The U.S. withdrawal from Southeast Asia will permit reductions in SIGINT and HUMINT resources, but they will only partially offset the above cost increases.
• • Some of the largest savings can only result from shifting and consolidating current activities in such a way as to redraw the functional boundaries of the major intelligence organizations.

Despite these difficulties, it is the case that functional boundaries can be withdrawn without a major reorganization of Defense intelligence or the community as a whole. We should stress, however, that actions of this character will still leave a number of community-wide issues unresolved and at the same time arouse all the opposition of the military Services and the Joint Chiefs of Staff. Moreover, with the rapid evolution of technology, further changes in boundaries—and comparable upheavals—will probably have to follow in the future.

With all these cautions, there are a number of specific functional actions that can be taken at the present time. Among the most important are the establishment of NSA as a truly national cryptological service with authority over all signal intelligence, and the consolidation of a number of activities now operated separately by the military Services. The effect of these changes should be to achieve economies of scale, eliminate excessive duplication, and promote competition among like activities so as to weed out the less productive programs.

The following table of possible savings, while only an estimate, indicates what economies might be feasible as a result of redrawing functional boundaries, consolidating activities, and eliminating duplication:

	Estimated Savings (In millions of dollars)
	FY 1972	FY 1975
Unify Defense mapping, charting, and geodesy activities	[*]	[*]
Form a single Defense Investigative Service	[*]	[*]
Establish a Defense Scientific and Technical Intelligence Center	[*]	[*]
Create a National Cryptologic Service	[*]	[*]
Assign NRP aircraft and satellite responsibilities to a single Defense manager	[*]	[*]
[1 row not declassified]	[*]	[*]
[1 row not declassified]	[*]	[*]
Subject Defense tactical intelligence to consolidated review	[*]	[*]
Assign ELINT authority to NSA	[*]	[*]
[* entry in table not declassified]

A major issue arises in connection with changes of such scope and magnitude. It is whether we should attempt to make the reforms now, or await more general reorganization and allow the head of the community to exercise his judgment and authority in instituting them. Our current judgment is that reductions of this magnitude should be attempted only after a reorganization has significantly improved the capabilities of the community to direct, control, and monitor program changes. We also believe that the economies should be effected over a period of years. Without these two conditions, the reductions could prove illusory or transient, and a heavy price in disruption and lowered morale might follow.

It should be noted that the anticipated savings come primarily from collection activities; major analytical and estimating capabilities are not affected. Their improvement is the subject of the next section.

IX. Toward Improvements in the Product

Much of the emphasis by the intelligence community and the bulk of its resources go to the high technology necessary to overcome barriers to information in the USSR and China. Yet this stress on the technology of collection—admittedly important—comes at a time when improved analysis is even more important.

Because of the keener competition from the Soviets, and the narrowing gap in relative resources devoted to defense, the U.S. must refine its evaluation of foreign capabilities, intentions, activities, and doctrines rather than assume that it has the resources to insure against all possibilities. The community must also improve its current political estimates and find ways of becoming more responsive to national consumers and their concerns.

Important improvements in performance may be feasible without major reorganization. But preliminary investigation suggests that higher quality is much more likely to come about within the framework of a coherently organized community which is focused on improving output rather than input. Indeed, it seems a fair assumption that the President would be willing to rebate some of the potential savings from the community if he had any hope of improved performance as a consequence. As of now, however, he has no such assurance and may reasonably argue that, for current performance, he should at least obtain the benefit of lower costs.

Even if we knew how to measure the benefits of intelligence, it would be difficult to relate specific changes in programs to improvements in performance. Nonetheless, experienced observers believe that the following steps—all of them comparatively inexpensive—should increase the usefulness of the product to the national leadership:

• • Major consumer representation to and within the intelligence community, perhaps through a restructured USIB, a high-level consumer council, or other institutionalized ways of communicating consumer needs, priorities, and evaluations to intelligence producers.
• • Assessment of the intelligence product through quality control and product evaluation sections within the production organizations themselves.
• • Upgrading existing analytical centers to increase the competition of ideas, including a DIA with improved organization and staffing as a major competitor to CIA in the area of military intelligence.
• • Periodic reviews by outsiders of intelligence products of the main working hypotheses within the community, and of analytical methods being used.
• • A net assessment group established at the national level which, along with the NSSM process, will keep questioning the community and challenging it to refine and support its hypotheses.
• • Stronger incentives to attract good analysts, better career opportunities to hold them as analysts instead of forcing them to become supervisors in order to achieve promotion, and a more effective use of personnel already trained and experienced in intelligence.
• • Increased resources and improved organizational arrangements within the intelligence community for research on improved methods of analysis and estimation.

It is probably premature to recommend the detailed measures necessary to improve the quality and scope of the intelligence product. In the near future, this issue should be considered at greater length by the leadership of a reorganized community. Indeed, the leadership should be specifically charged with the task of product improvement as a matter of the highest priority. What steps will prove feasible will depend on the particular type of reorganization selected, and, in the present circumstances, it may be well to be guided in the choice by considerations of economy in the use of resources. But it should be stressed, in conclusion, that improvement of the product at current budget levels is simply another way of achieving the efficiency that is so desperately needed within the intelligence community as it is presently constituted.

1. Source: National Archives, Nixon Presidential Materials, NSC Files, Subject Files, Box 332, Intelligence Reorganization, Vol. I. Top Secret; Byeman Comint.
2. Kissinger wrote below this paragraph: “I wouldn’t have told Shultz it was adopted. I prefer to sit on it. I have no intention to bleed over it.”
3. Top Secret; Handle via Byeman-Comint control systems jointly. A note on the memorandum indicates that the President saw it. The memorandum is undated but another copy, attached to Smith’s June 18 memorandum to Kissinger, is dated March 22. (National Archives, Nixon Presidential Materials, NSC Files, Subject Files, Box 332, Intelligence Reorganization, Vol. I)
4. The President initialed his approval.
5. Top Secret; Handle via Byeman-Comint control systems jointly. The study, known both as the “OMB/NSC Study” and the “NSC/OMB Study” and sometimes referred to as the “Schlesinger Report,” was prepared jointly by OMB and NSC with James Schlesinger, OMB Assistant Director, playing a key role.
6. Not printed.