Foreign Relations of the United States, 1945–1950, Emergence of the Intelligence Establishment

420. Memorandum From the Secretary of State’s Special Assistant for Research and Intelligence (Armstrong) to the Under Secretary of State (Webb)⁰

Washington, May 2, 1950.

SUBJECT

• National Intelligence Estimates

As you are aware, it has long been apparent that existing mechanisms and procedures for producing national intelligence estimates are inadequate. Although this situation was pointed out with considerable clarity and force by the Dulles–Jackson–Correa Committee report of January 1, 1949, later endorsed generally by the National Security Council in NSC 50 of July 1, 1949, it has been equally apparent that no effective steps will be taken to bring about improvement unless specific instructions are issued by the NSC.

The problem is a difficult and complicated one and has been under intensive study in R for more than six months. The results of our initial investigation are embodied in a departmental staff study dated January 3, 1950, attached as Tab A. I believe that you will find this study useful as general background on the question, especially with respect to the differences of view which have existed between CIA and the other intelligence agencies, notably State.

While the departmental study was under way we began discussions with General Magruder of the Department of Defense, who was independently working on the problem, and with the service intelligence chiefs. These discussions have led to a joint staff study and proposed NSC directive dated May 1, 1950, appended as Tab B, on which there is now State-Defense agreement at the intelligence level. We have discussed our ideas with the Intelligence advisers and are now in process of putting them before Messrs. Kennan, Nitze, and Rusk. Mr. Lay has also been kept informed of the various drafts, though he has not of course been asked to take a position.

We sincerely believe that the mechanisms and procedures set forth in the proposed directive will make possible the production of national intelligence of the high quality which the President and the National Security Council have the right to expect.

Before the study is put before the NSC, however, it will be necessary, or at least desirable, to show it to the Director of Central Intelligence, and obtain his views. There are a number of ways in which this might be done, but the Defense representatives agree with us that the approach definitely should not be at the IAC level and in fact should follow a general approval by you and Secretary Johnson or Under Secretary Early. If you are in agreement with the proposal, you and Mr. Early could appropriately lay it before Admiral Hillenkoetter on an informal basis.

It is very probable that your support and personal assistance will be necessary in any further steps that are taken. We stand ready at any time to give you further briefing and to discuss the various procedures we might use to put it into effect.

Tab A¹

Washington, January 3, 1950.

Department of State Staff Study on National Intelligence

Problem

What is the nature of national intelligence and how can its production be improved?

Facts

1. Responsibility for the production of national intelligence rests with the Central Intelligence Agency. This responsibility derives, in the first instance, from the National Security Act of 1947 which provides (Section 102 (d) (3)) that the CIA shall “correlate and evaluate intelligence relating to the national security, and provide for the appropriate dissemination of such intelligence within the Government using where appropriate existing agencies and facilities”. The Act further provides that “the departments and other agencies of the Government shall continue to collect, evaluate, correlate and disseminate departmental intelligence”.

2. Pursuant to the National Security Act, the National Security Council issued a series of intelligence directives (NSCID’s) for the guidance of CIA and the other intelligence agencies. In these directives “intelligence relating to the national security”—called “national intelligence”—is defined, and principles to govern its production and dissemination are established. The definition given is as follows:

“National intelligence is integrated departmental intelligence that covers the broad aspects of national policy and national security, is of concern to more than one Department or Agency, and transcends the exclusive competence of a single Department or Agency or the Military Establishment.” (NSCID–3)

a. The principle governing the production of national intelligence is:

“In so far as practicable, [the CIA]² shall not duplicate the intelligence activities and research of the various Departments and Agencies but shall make use of existing intelligence facilities and shall utilize departmental intelligence for such production purposes.” (NSCID–3)

b. The principle governing dissemination of national intelligence is:

“Intelligence so disseminated [i.e. to the President, the NSC, etc.] shall be officially concurred in by the Intelligence Agencies or shall carry a statement of substantially differing opinions.” (NSCID–3)

3. The National Security Council, with the assistance of the Dulles–Jackson–Correa Committee, recently re-examined, and specifically reaffirmed as sound, these provisions of the National Security Act of 1947 and the NSCID’s relating to national intelligence (NSC 50, section 1 c and 5 c). The NSC concluded, however, that the directives have not been effectively carried out, principally because there has been within the CIA a confusion between responsibility for producing national intelligence estimates and responsibility for miscellaneous research and reporting (NSC 50; section 5 c).

4. In referring to the directives relating to the production of departmental and national intelligence, the NSC agreed with the Dulles Committee that the CIA should interpret and follow the NSC directives so as to refrain as far as possible from competitive intelligence activities in the production of research intelligence estimates (NSC 50; section 5 c).

5. The NSC also reaffirmed the principle that the CIA should draw upon the specialized intelligence production of the agencies in order to prepare coordinated national intelligence estimates, and declared that a procedure should be adopted which would permit the Director of Central Intelligence to take full advantage of departmental opinion while retaining sole responsibility for the final national intelligence product. (NSC 50; section 5 c).

Discussion

An objective appraisal of the problem requires a fuller analysis of the various factors which have contributed to its existence. Such an analysis logically calls for consideration of possibly divergent views as to the nature of national intelligence, as to the meaning of the directives relating thereto, and to the utility of the directives in practice. The discussion which follows, therefore, seeks: first to place in focus the major unresolved issues in the broad field of national intelligence production; second, to consider the concept of national intelligence as it bears upon such issues; third, to interpret the directives in terms of the theory of national intelligence; and fourth, to test the theory against practical realities.

I. Focus of the Problem

A. General.

6. The confusion with respect to national intelligence, to which the NSC referred, stemmed from differing interpretations of the controlling directives and differing concepts of national intelligence and of coordination (Dulles report, pp. 65, 69). Such differences still exist and still seriously impede the national intelligence production effort. Probably the differing interpretations and concepts of State and CIA provide the most clear-cut expression of the issues involved.

B. CIA View.

7. CIA apparently believes that:

a. Inasmuch as the Director is, by law, solely responsible for producing national intelligence estimates, collaborative preparation of such estimates is not necessary. At the option of CIA, other agencies may be asked to assist in the preparation of estimates, or the entire production process, from research through drafting, may be carried out by CIA. This production responsibility, therefore, not only forces him to reject any thought of collective responsibility with the IAC, but also justifies him in minimizing procedures for joint or collaborative preparation of drafts.

b. “Coordination” of national intelligence estimates can be adequately accomplished by a procedure which gives the other IAC agencies the opportunity to express concurrence or dissent to finished drafts presented to them by the CIA. By this token, CIA may itself write reports on any topic, subject only to the concurrence procedure.

c. The definition of “National Intelligence” cited in paragraph 2 above should be interpreted as follows:

(1) departmental intelligence is integrated if the departments participate to the extent of expressing concurrence or dissent in the CIA product;

(2) intelligence covers “the broad aspects of national policy and national security,” and becomes national intelligence if an important aspect of national policy or security is concerned, whether or not it falls wholly within the responsibility of some one department.

(3) intelligence “transcends the exclusive competence of a single Department or Agency or the Military Establishment,” whether or not such intelligence can be produced by a single departmental agency, if it has been specifically requested by the President or an inter-departmental agency, or initiated by the CIA.³

d. A large staff⁴ is needed to ensure that the DCI has the facilities to carry out his responsibility for producing national intelligence, for without such a staff under his immediate direction he cannot:

(1) meet emergency situations.

(2) protect himself from being dependent upon Departmental priorities in the production of papers.

(3) analyze the mass of information necessary to detect possible departmental bias and thus provide the independent “check” required of him.

(4) accept the responsibility involved in expressing an independent intelligence opinion sometimes at variance with the opinion of one or more of the other IAC agencies.

e. Such a staff may independently prepare intelligence papers and conduct intelligence research operations parallel to those of the other IAC agencies in all phases of production.

C. State View.

8. State believes that the CIA views outlined in A above fail to meet the intent of the NSC, have been largely responsible for the continuing failure of the efforts to produce adequate national intelligence, have produced an unwarranted duplication of effort, and have tended to disrupt the entire intelligence production effort. Specifically, State believes that in terms of current definitions:

a. The DCI’s responsibility for “producing” national intelligence prohibits him from requesting the IAC collectively to share responsibility for the final, independent judgment required by law, but imposes upon him an absolute obligation to seek the individual and collective opinion of the departmental specialists in arriving at his judgment. “Producing” does not mean that CIA does the entire work of preparing reports from the basic raw information; rather it means that CIA welds into a unified whole the oral or written contributions of the Departments.

b. “Coordination” of national intelligence is effective only insofar as it enlists a common, working collaboration of the responsible intelligence experts throughout the agencies in presenting the most authoritative reports possible. Responsibility for coordination, therefore, cannot mean that CIA should assume responsibility for all projects which happen to be referred to it. On the contrary, coordination implies, inter alia, that CIA should serve as a clearing house for referring to the appropriate IAC member any projects which fall within that member’s field of responsibility.

c. National intelligence

(1) is “integrated” only when departmental intelligence has explicitly participated in its preparation at every stage;

(2) covers broad aspects of national policy and security only when it involves topics of wide scope which require an objective balancing and interrelating of factors over which no one agency exercises exclusive jurisdiction.

(3) transcends the competence of the agencies only if it requires a fusion of functional specialties of more than one agency.

d. The quality of the central staff rather than the size is the determining factor in accomplishing the national intelligence mission; departmental staffs can provide the elements of national intelligence; acuteness rather than exhaustive research is required to check the results.

e. Central duplication is wasteful and dangerous to the extent that certain products of CIA are distributed outside regular channels and thus add to the stream of intelligence uncoordinated views that are sometimes divergent.

II. Concept of National Intelligence

A. General.

9. The conflicts in this situation divide themselves into problems of definition and of execution; all of these in turn depend in part on the concept of national intelligence. Upon the resolution of these problems hangs in considerable measure the future shape of intelligence production in the United States.

B. Background.

10. The wartime experiment with “joint” intelligence unquestionably provided experience which helped to shape the theories under which the Central Intelligence Agency was established. During the war, “joint” intelligence was produced through a committee structure, with departmental representatives drawing from their respective departments the intelligence opinion required for joint intelligence problems. This pyramidal concept proved sound except that its peak, a committee of equals, tended to produce a watered-down product. The plan recommended after the war for central intelligence retained the pyramid but replaced the committee by a Director who was empowered to express an independent view on national intelligence. It was provided that to the maximum extent possible the Director’s view should represent a fusion of expert departmental opinion, but to avoid the watering down that would result from forced agreement the device of dissent, or written statement of serious disagreement with the Director’s conclusion, at once freed the Director from being bound by Departmental views and allowed policy officials to know when such serious doubt existed.

C. Analysis of Concept.

11. This concept finds support in the Dulles report analysis (p. 70):

“In the original Central Intelligence Group it was conceived that there would be a small organization of highly qualified individuals which would limit itself strictly to national intelligence problems and base its work primarily on the specialized reports and estimates produced by the departments rather than employ a large research and analysis organization of its own.”

and at pp. 68–69:

“The concept of national intelligence underlying the statute and the directives is that of an authoritative interpretation and appraisal that will serve as a firm guide to policy-makers and planners. A national intelligence estimate should reflect the coordination of the best intelligence opinion. It should be based on all available information and be prepared with full knowledge of our own plans and in the light of our policy requirements. The estimate should be compiled and assembled centrally by an agency whose objectivity and disinterestedness are not open to question. Its ultimate approval should rest upon the collective responsibility of the highest officials in the various intelligence agencies.”

III. Embodiment of the Concept in Definitions

A. Collaborative Preparation.

12. It seems clear from the foregoing that the assignment to the Director of Central Intelligence of responsibility for the production of national intelligence did not imply an obligation on his part to prepare estimate entirely from his own resources. To be sure, the NSC rejected the Dulles report’s extreme suggestion of collective responsibility, but this by no means implied an opposite extreme. Indeed, the NSC explicitly endorsed the Dulles report when it reaffirmed belief in collective construction of reports. (NSC 50, section 5 c supra), a belief more sharply stated at p. 68 of the Dulles report:

“Although the Act and Intelligence Directives give the CIA the independent right of producing national intelligence, Directive No. 1 stipulates that such intelligence shall be officially concurred in by the intelligence agencies or shall carry statement of substantial dissent. As a practical matter, such estimates can be written only with the collaboration of experts in many fields of intelligence and with the cooperation of several departments and agencies of government.”

13. It seems fair to assume that the several recognitions of “departmental intelligence” which appear in the Act and in the Intelligence Directives were included with the conscious purpose that such intelligence should play a part in the national intelligence scheme. The NSC ID’s further carefully allocate responsibility for national intelligence to the CIA and for designated fields of “dominant interest” to each departmental agency in such a way that the department fields among them cover the major functional division of intelligence;⁵ this can only mean that each department was to furnish a part of the whole and to operate, for economy’s sake if nothing more, on a system of interreliance. CIA’s responsibility for coordination, in the sense of leadership in a common effort, was thus intimately related to its responsibility for national intelligence production. The Dulles report speaks of this relationship at p. 65.

The CIA is—“given the responsibility of seeing to it that the United States has adequate central machinery for the examination and interpretation of intelligence so that the national security will not be jeopardized by failure to coordinate the best intelligence opinion in the country … This responsibility has not been adequately discharged and remedial measures are necessary. There is confusion as to the proper role of CIA in the preparation of intelligence reports and estimates. This confusion has resulted from incorrect interpretation and lack of proper implementation of the statute and the directives. The reasons for this go to the heart of the national intelligence problem …”—Dulles report, p. 65.

B. Distinction Between Departmental and National Intelligence.

14. Unless the distinction between national and departmental intelligence is vigorously maintained among the members of the intelligence team there will obviously result a serious diminution in the effectiveness of the entire structure, both in productivity and type of product. The attached ORE Status Report of 1 December 1949 lists titles of recent or impending CIA “national intelligence” reports. The following titles, among others, confirm the state of confusion which exists:

• Communism in Scandinavia
• Political orientation of the West German State
• The Succession of Power in the USSR
• Postwar developments in Latin American civil aviation
• Relation of Indonesia to the Economy of the Netherlands
• UK activities in Iran
• Great Britain and the German Problem
• Soviet objectives in Latin America
• The Suppression of Communism in Turkey
• Soviet role in the UN

15. It need hardly be argued, in terms of preceding concepts and definitions, that such reports fail to fulfill the high mission assigned the CIA in this field. It may, indeed, be pointed out that under any reasonable interpretation of the allocations established in NSCID 3, most, if not all, of these topics fall largely within the field which is recognized as the responsibility of the State Department—well outside the field defined as national intelligence. The Dulles–Jackson–Correa interpretation of the pertinent directive seems sufficient (Dulles report, pp. 67, 68):

“The significant provision of Directive No. 3 for the CIA is the definition of national intelligence, for which the Agency is given exclusive responsibility … In effect the directive interprets the vague provision of the National Security Act on “intelligence relative to the national security” to cover a particular type of intelligence reasonably distinct from departmental intelligence and conforming to admittedly broad but generally comprehensible specification.”

National Intelligence “should deal with topics of wide scope relevant to the determination of basic policy, such as the assessment of a country’s war potential, its preparedness for war, its strategic capabilities and intentions, its vulnerability to various forms of direct attack or indirect pressures. An intelligence estimate of such scope inevitably ‘transcends the exclusive competence of a single Department or Agency or the Military Establishment.’ A major objective, then, in establishing the CIA was to provide the administrative machinery for the coordination of intelligence opinion, for its assembly and review, objectively and impartially, and for its expression in the form of estimates of national scope and importance.”

IV. Practicability of the National Intelligence Plan

A. Operational Practices.

16. If then the theory of the national intelligence production plan is sound, is the plan practical in operation? The CIA evidently believes not. As was noted above, the DCI’s interpretation of sole responsibility for national intelligence has led to the establishment of a large staff to assist in preparing such intelligence. The principle that maximum use should be made of departmental facilities (NSCID 3) has been forced aside through the fact of the staff’s existence and the operational practices which the CIA considers essential to its position. The result, as put by the Dulles Report (p. 54) is that “the Central Intelligence Agency itself has become a competitive producer of intelligence on subjects of its own choosing which can by no stretch of the imagination be called national intelligence.” Whether there is need for the practices which led to this result warrants detailed discussion, and can be clarified by answering a small number of questions:

B. Problems in Execution.

17. Is Cooperative Preparation Incompatible with Sole Responsibility for Production?

The responsibility for production of national intelligence was centered in one official to guard against compromise or a possible failure to unify in one place all available intelligence. It was not so centered to relieve departmental experts from all but a review function. On the contrary, it was designed to provide a common meeting ground where the specialized departmental intelligence views could be considered in relation to a broad, interdepartmental problem, could be synthesized in the form of a single, authoritative estimate, and put forth under the responsibility of the central authority. Such an estimate had to be founded upon departmental intelligence provided especially for the occasion to ensure that the central balancing of possibly divergent views was a balancing of elements tailored to the same problem. This concept of common effort had also its practical side: Maximum utilization with minimum duplication of the intelligence production resources of the Government.

18. Can the DCI Rely upon a System of Cooperative Preparation for National Intelligence Purposes?

The DCI is empowered to secure from the departments all available intelligence and to request special intelligence studies (NSCID 3). He is also empowered to secure from the Departments all intelligence information which he requires. The tools therefore are at hand for him to carry out his missions of coordination and production. The materials available enable him to plan the national intelligence program, to spot weaknesses in departmental intelligence programs, to be alert to crisis situations, and to initiate or respond to novel projects of high national importance. There remain, even so, possible difficulties which must be examined further:

a. Can the DCI receive prompt and adequate service on a national intelligence request which conflicts in an agency with other work in progress? To be sure, without a considerable increase in departmental intelligence staffs, the DCI cannot expect to receive prompt and adequate service on the variety of miscellaneous reports currently being produced as “national intelligence.” However, with respect to problems of such scope and importance as to warrant a national intelligence estimate, it will generally be found that the relevant agencies already have work in progress on the subject. It is then simply a matter of adjusting a current agency project to the national intelligence problem rather than of initiating a totally new project. The difficulty can in part also be resolved through effective programming which would give participants notice of forthcoming burdens and permit adjustments of their schedules. Should these principles fail, the DCI’s relationship with the IAC and NSC ensures any necessary attention to a well considered project of national intelligence. There remains, however, one possible obstacle: agency limitations in personnel and equipment. To meet an emergency of this sort the NSCIDs contemplate that the DCI shall support the departments in the interest of the over-all Governmental intelligence structure (NSCID 1) and either assist the agency in securing from usual sources the necessary reinforcement or himself supply the reinforcement to the agency.

b. Is collaborative composition incompatible with the speed required in crisis situations? Collaborative composition does not necessarily imply time-consuming composition. It implies rather a flexible means of obtaining the fullest feasible application of pertinent viewpoints to a given problem. Normally a national intelligence crisis estimate (cf. Estimate of Russian intentions, March 1946) is a brief expression of conclusions evolved from a maximum pooling of evidence and ideas with a minimum of composition. Collaborative composition may thus on one occasion involve a careful, laborious process of group drafting; on another, individual drafting from agency submissions for group consideration; on another, individual drafting on the basis of group discussion. Topic and deadline in each case could determine the method.

c. Can the DCI detect possible departmental bias and thus provide the independent “check” required of him? An abiding fear of any intelligence officer is the fear that operational pressures may force intelligence to lend unjustified support to a given policy or that personal feelings may develop prejudice. The danger is inherent in any intelligence plan. Obviously if State, wedded to a particular policy, presents facts distorted by faulty preconceptions, the resulting intelligence report will be defective. Unquestionably, such a danger is an important reason why the DCI has final responsibility for the national intelligence position. Detection of bias, however, is as readily accomplished through collective as through central preparation. The caliber of the CIA staff is largely determinative in either event, but that staff, under a collective theory, has greater assistance from other knowledgeable agencies, whose cross-views may bring out hidden faults. In reality the problem of bias is a purely practical one; to achieve the perfect solution, one would require a skyscraper of staffs of equal size, each checking the findings of the one on the floor beneath; in actual practice, on the relatively high level of interpretation at which such bias becomes a problem, it is not extensive files but intelligent and inquiring minds that constitute an effective safeguard.

d. Could the DCI continue to exercise the sometimes independent judgment required of his position? Independence of judgment obviously does not deny a theory of collaborative preparation. As was indicated earlier, the fixing of responsibility for national intelligence in a single body was designed to overcome the dangers of compromise inherent in a joint decision; collaborative preparation, on the other hand, is the method of obtaining all relevant views in order that the final judgment may be based upon the best available intelligence in the various related fields. The more full the examination of all factors which condition a single problem, the sounder normally should be the conclusions derived therefrom. Reliance upon a single staff as a means for arriving at independent conclusions, checked only by review, would seem less wise than a hearing in the first instance, of the best thoughts of the specialty staffs of the departments.

e. Could a small high-level staff provide the DCI with the support he requires in fulfilling his responsibility for producing national intelligence? Both the NSC and the Dulles Report agreed that a small staff would and should be sufficient for this purpose. (NSC 50 section 6 c.) They added that central research should be confined to recognized fields of common interest. In those two thoughts are clear recognition that the duplicatory role of CIA is wasteful, that national intelligence is a special type requiring expertise rather than numbers to produce, and that intelligence research outside of the departments is useful only in fields where departmental intelligence research is totally lacking or is being carried out wastefully in more than one department. If, as has been suggested above, the national intelligence plan is free from serious practical as well as theoretical defects, it is difficult to understand by what theories a large staff is justified. It is doubtful that the US Government is prepared to support the ideal mentioned earlier, a series of departmental and independent groups each duplicating the work of the other in order to test its accuracy. At some point, there has to be acceptance of the principle of delegated responsibility.

19. Do Departmental Personnel Ceilings Warrant CIA Retention of a Parallel Intelligence Research Staff to Absorb Overflow Departmental Work?

The national intelligence production plan, indicated above, envisages the entire intelligence structure as a pyramid; the base, four strong intelligence agencies to whom are delegated responsibilities functionally related to the work of the departments they serve; the apex, the central authority deriving its main strength from the base and, through its national intelligence staff, seeking out and arranging the essential union of departmental ideas which might otherwise never unite. The apex is only as strong as the foundation. When it placed in one central authority responsibility for the structure, the NSC remembered peace-time departmental temptations. It imposed upon the DCI the task of keeping intelligence strong. Retention of parallel intelligence research in CIA not only would tend to destroy the symmetry of the pyramid but in fact would ignore a prime CIA responsibility—to ensure as sound as possible an investment of intelligence resources. Inability in any Department fully to perform its intelligence mission should be of concern to the DCI in terms of strengthening the weak spots, or, under his coordinating responsibility, in terms of recommendation to the NSC. Parallel reporting absorbs additional Departmental effort in concurrence or dissents and thus puts impossible strains on the weaker links.

Conclusions

I. The production of national intelligence remains at an ineffective level largely because of a continuing misunderstanding on the part of the DCI as to the legal connotations of his responsibility for producing national intelligence. This misunderstanding has caused the CIA to seek to establish a national intelligence staff which can provide, from its own resources, all the materials required for national intelligence purposes. As a result there has occurred distortion in the meaning of national intelligence which, in turn, has caused regrettable wastage, through duplication, of intelligence resources. This, in turn, has brought on confusion as to the concept of US intelligence production and has created interdepartmental friction which could have serious consequences.

II. The clarity of the directives relating to national intelligence and the recent NSC comments thereon suggest that some part of the present CIA implementation of those directives may stem from doubt as to the practicability of the plan which they outline. In fact, however, the plan is sound, both in theory and in practice. Its success depends upon a common realization of its team aspects and the role which each member must play. The CIA must itself understand and must itself establish within the IAC a real feeling of the joint mission which national intelligence represents. The DCI must likewise appreciate more keenly his broad responsibility for strengthening the entire intelligence structure.

III. The small staff recommended by the Dulles Staff, working in close collaboration with present departmental facilities, can provide the DCI the support necessary to fulfill his national intelligence production mission. Problems will remain, priorities and the like, but none sufficiently serious that they cannot be solved through the forthright leadership of the DCI.

Tab B⁶

Washington, May 1, 1950.

Joint Department of State–Department of Defense Staff Study

PRODUCTION OF NATIONAL INTELLIGENCE

Problem

1. To provide for the more effective production of national intelligence and to provide for continuous surveillance of current intelligence.

Facts Bearing on the Problem

2. Responsibility for the production of national intelligence rests with the Central Intelligence Agency. This responsibility derives, in the first instance, from the National Security Act of 1947 which provides (Section 102(d)(3)) that the CIA shall “correlate and evaluate intelligence relating to the national security, and provide for the appropriate dissemination of such intelligence within the Government using where appropriate existing agencies and facilities”.

3. Pursuant to the National Security Act, the National Security Council issued a series of intelligence directives (NSCID’s) for the guidance of CIA and the other intelligence agencies. In these directives, “intelligence relating to the national security”—called “national intelligence”—is defined, and principles to govern its production and dissemination are established.

4. The National Security Council, with the assistance of the Dulles–Jackson–Correa Committee, recently re-examined, and specifically reaffirmed as sound, those provisions of the National Security Act of 1947 and the NSCID’s relating to national intelligence (NSC 50, section 1c and 5c). The NSC concluded, however, that the directives have not been effectively carried out, principally because there has been within the CIA a confusion between responsibility for producing national estimates and responsibility for miscellaneous research and reporting (NSC 50, section 5c).

5. The NSC also reaffirmed the principle that the CIA should draw upon the specialized intelligence production of the agencies in order to prepare coordinated national intelligence estimates, and declared that a procedure should be adopted which would permit the Director of Central Intelligence to take full advantage of departmental opinion while retaining sole responsibility for the final national intelligence product. (NSC 50, section 5c).

Discussion

6. Experience since the issuance of NSC 50 has indicated that the quality of national intelligence estimates has not substantially improved and that unwarranted duplication of effort as between the various agencies has continued. This condition results in large measure from

a. Continuing disagreement between CIA, on the one hand, and the departmental agencies, on the other, as to the meaning of national intelligence;

b. Inadequacies of existing mechanisms and procedures for production of national intelligence;

c. Continuing confusion within CIA between its responsibility for producing national intelligence and its responsibility for miscellaneous research and reporting.

7. The definition of national intelligence given in NSCID–3 needs further and more precise interpretation. It has not proved adequate as a guide to operating officials, and it has been subject to varying interpretations. To resolve this ambiguity, the meaning of the term “national intelligence” should be restated by the National Security Council, both in terms of content and in terms of end use.

8. Revision of the definition, however, will not in itself insure the production of authoritative national intelligence estimates reflecting the coordination of the best intelligence opinion in the Government. This can be accomplished only through the revision of existing mechanisms and procedures, which are inadequate. They should be revised to insure that national intelligence estimates reflect in every instance an integration of the best intelligence opinion of the Government. Such an integration can be fully achieved only through a cooperative process of preparation in which departmental contributions, oral and written, are synthesized and departmental intelligence opinion and advice, explicitly brought to bear at all stages of production.

9. Continuing confusion within CIA between its responsibility for national intelligence estimates and its responsibility for miscellaneous research and reporting can only be cured by an organizational separation of the staffs carrying out the two functions. A special group should be created in CIA to handle its national intelligence functions and, in addition, the surveillance and dissemination of current intelligence required by executives responsible for the formulation and execution of national policy. A separate office should be responsible for carrying out such research and reporting functions of common concern as may be prescribed by the NSC under the provisions of Section 102(d)(4) of the National Security Act of 1947.

Conclusions

10. The National Security Council should define national intelligence in terms of both content and end use so as to provide a clear guide for operating officials.

11. Existing mechanisms and procedures for the production of national intelligence estimates should be revised so as to insure that such estimates represent in every instance an integration of the best intelligence opinion in the Government.

12. The staff within CIA responsible for carrying out the Agency’s national intelligence functions should be organizationally separate from the staff responsible for research and reporting activities of common concern.

Recommendations

That the National Security Council

a. Approve and issue the attached directive implementing the conclusions of this paper.

b. Direct that the strength of the National Intelligence Group, described in paragraph 5 of the directive, shall not exceed 100 officer and/or professional personnel, of whom not more than 20 may be detailed by departmental agencies. One half of the necessary personnel authorizations or spaces for officer and professional personnel shall be charged against existing CIA allotments.

c. Note that after the accumulation of operating experience, the Director of Central Intelligence, with the advice and assistance of the IAC, will recommend to the NSC such modifications in the strength and composition of the National Intelligence Group as may be necessary.

d. Note that the Director of Central Intelligence, with the advice and assistance of the IAC, will submit revisions of existing NSCID’s and DCID’s required to bring those directives into conformity with the attached directive.

Attachment⁷

NSC DIRECTIVE

Pursuant to the provisions of paragraphs (d) and (c) of Section 102 of the National Security Act of 1947 and in furtherance of paragraphs 5 and 6.a(2) of NSC 50, the National Security Council hereby authorizes and directs the following readjustment of the functions and organization of the Intelligence Advisory Committee and of the Central Intelligence Agency, in order specifically to insure more effective production of national intelligence estimates and studies, and to provide continuous surveillance of current intelligence:

General

1. National intelligence is that intelligence required in the formulation of policy at the national (i.e. supra-departmental) level. As stated in NSCID–3, such intelligence inevitably “covers the broad aspects of national policy and national security”, and normally it “transcends the exclusive competence of a single department or agency or the military establishment”. Regardless of its content, however, intelligence used in the formulation of national policy must have an interdepartmental basis, i.e., it must embody departmental intelligence resources and be subjected to adequate safeguards against bias. Strictly political or strictly military intelligence estimates, for example, should be considered as national intelligence in those cases where such estimates are required in the formulation of policy at the national level.

Because national intelligence serves as a basis for the formulation of policy at the national level, it is essential that it represent in every instance an integration of the best intelligence opinion of the departments. Such an integration can be fully achieved only through a cooperative process of preparation in which departmental contributions, oral or written, are synthesized and departmental intelligence views are explicitly brought to bear at all stages of production. Intelligence which has not been so produced should not be disseminated as national intelligence.

Intelligence Advisory Committee

2. Composition

In order to facilitate the effective functioning of the IAC, the Director of Central Intelligence, the intelligence chiefs of the Departments of State, Army, Navy and Air Force and the Federal Bureau of Investigation shall each designate a qualified individual who shall be on full time duty at IAC headquarters and who will be representative of his agency in matters covered by this directive not requiring the action of the head of his agency. The Director of the Joint Intelligence Group and the Director of Intelligence of the Atomic Energy Commission may designate such representatives if they so desire. The representative of the Director of Central Intelligence shall concurrently be head of the National Intelligence Group (hereafter described).

3. Mission

a. In addition to its existing functions, the mission of the IAC with respect to national intelligence shall be

(1) To initiate timely national intelligence requirements to cover future foreseeable needs for national estimates and studies of a continuing, recurrent, or emergency nature, such requirements to be recommended to the DCI for preparation by the National Estimates Staff (hereafter described).

(2) To advise the DCI as to the desirability and feasibility of national estimates and studies which he may initiate.

(3) To review drafts of national estimates and studies prepared by the National Estimates Staff and resubmit them for modification or clear them expeditiously for consideration by the DCI by unanimous concurrence or with record of substantive dissent by individual members. All issues under review on which dissent is pending shall be considered by the assembled IAC.

(4) To maintain contact through the Chairman, IAC, with the NSC staff and national planning agencies of the Government in order to be cognizant at all times of contemplated high level negotiations, plans or projects which should be soundly based on national intelligence estimates.

(5) To maintain close liaison with the Joint Intelligence Committee in order to coordinate the projects for estimates in the two committees and to insure the integration of appropriate parts of national estimates with joint estimates.

b. In addition to the functions outlined above, it shall be the mission of the IAC

(1) By keeping under continuing review all critical current intelligence, to be prepared to draw nationally significant conclusions therefrom, if an emergency threatens or prompt executive action is indicated. Such conclusions shall be recommended without delay to the DCI for immediate transmission to key executive officials.

(2) To recommend to the DCI the nature and specifications of periodic current intelligence summaries and reports, centrally prepared and appropriately edited, required by executive officials responsible for national policy.

(3) To recommend to the DCI the preparation of special intelligence products in recognized fields of common interest.

(4) To be responsible for the conduct of all discussions with foreign intelligence agencies respecting the formulation of combined national intelligence estimates and the preparation, in collaboration with foreign intelligence agencies, of such combined estimates.

4. Location

The permanent headquarters for deliberations of the IAC shall be conveniently located with respect to the National Intelligence Group (hereafter described) and its facilities. The housing and facilities for both the IAC and the National Intelligence Group shall be provided by the CIA.

National Intelligence Group (NIG)

5. Composition

The National Intelligence Group shall be established as an organic part of CIA. It shall consist of a National Estimates Staff and Current Intelligence Staff. A part of the professional and all clerical personnel of the Group shall be provided by CIA. A proportion of the professional personnel shall be provided by detail from departmental agencies for duty with CIA; each individual so detailed shall be acceptable to the DCI and responsible to him. The DCI shall provide the necessary personnel authorizations as well as reimbursement for civilian personnel assigned. He shall also provide spaces for assigned military personnel which will be additional to the personnel authorizations of their respective service intelligence agencies.

National Estimates Staff (NES)

6. Mission

The National Estimates Staff shall:

a. Break down intelligence requisitions initiated by the IAC into component requisitions upon the several departmental agencies and CIA for their appropriate contributions of intelligence material, giving clear specifications as to the nature and scope of material required and deadline dates for completion.

b. By evaluation, analysis and synthesis of intelligence contributed in appropriate form and content by the several departmental intelligence agencies and offices of CIA, prepare drafts of national estimates and studies.

c. Present draft estimates and studies for consideration of the members of IAC prior to formal review by that body. Representatives of the National Estimates Staff normally shall be present when the drafts are considered by the IAC and shall be authorized to furnish orally additional reasoning or factual knowledge in support of statements or conclusions incorporated in the drafts.

d. From critical information received from the Current Intelligence Staff or any other source indicating the need for timely executive action on any scale, prepare crisis estimates for prompt submission to the DCI and the IAC.

e. Periodically report to the DCI and the IAC on the adequacy and completeness of available information required in the preparation of national estimates.

f. Collaborate with the Current Intelligence Staff in order to insure that estimates and studies reflect the most recent, evaluated, current information, and contra-wise, that the reasoning of current intelligence publications is not inconsistent with well-confirmed long-range intelligence.

Current Intelligence Staff (CIS)

7. Mission

The Current Intelligence Staff shall be organized and staffed to effect

a. The prompt and systematic receipt, surveillance, and appropriate situation room display of all radio, wire and otherwise rapidly-dispatched current information received in all agencies and departments of the Government which may be of significance in the preparation of current intelligence required by executives responsible for the formulation or implementation of national policy.

b. The conduct of a 24-hour watch on incoming information with provisions for immediate alert of the DCI, NIG and the members of the IAC upon the indication of critical or emergency situations which might require the preparation of crisis estimates or timely executive action on any scale.

c. The identification and preparation of “hot” information obtained by offices within CIA or from other sources for dissemination by the DCI if in his opinion it is of sufficient importance and urgency to be furnished without delay to the President, the members of the National Security Council and the IAC.

d. The preparation and dissemination, under the sole responsibility of the DCI and with speed appropriate to the apparent significance of the information, of periodic current intelligence summaries patterned to the needs of the executives responsible for the formulation of implementation of national policy. To the extent practicable, the material in such summaries which is of predominant interest to a member agency shall be coordinated with that agency.

e. The operation of an intelligence situation room designed to serve the needs of the above-mentioned executive officials and their principal assistants, the IAC and the National Estimates Staff.

Missions of Departmental Agencies and Central Intelligence Agency
(with respect to national intelligence)

8. Departmental Agencies

The several departmental agencies of the IAC shall

a. Furnish upon requisition of the National Estimates Staff, within the deadline dates prescribed, appropriate intelligence in form and content as required by the NES to provide for the national intelligence projects initiated by the IAC.

b. Be prepared to make oral presentations in the NES in support of their respective intelligence contributions at their own request or when requested by the NES.

c. Furnish to the CIS in the most expeditious manner possible all radio, wire and rapidly-dispatched current information of possible significance to the current intelligence mission of the CIA. In this category of information should be included the intelligence content or implications of operational messages.

d. In the event that intelligence studies in subjects of primary interest are prepared departmentally which have the characteristics of national intelligence and are intended for dissemination above departmental level, they shall be subjected to the reviewing procedures of the NES and the IAC in order to obtain the imprimatur of “national intelligence”.

e. In the choice of personnel to be detailed for duty with the National Intelligence Group, nominate individuals of highest professional qualifications.

9. Central Intelligence Agency

a. All offices of CIA having intelligence resources shall contribute to the requirements of the NES and the CIS in accordance with the same principles as Federal agencies outside of CIA.

b. Intelligence offices of CIA, other than the National Intelligence Group, shall be restricted to the production of intelligence prescribed in paragraphs (4) and (5) of Section 102d of the National Security Act of 1947.

c. Intelligence studies prepared under Paragraph 9.b., above, if of a character to be employed as national intelligence, must have passed through the reviewing procedures of the NES and the IAC before being designated “national intelligence”.

10. Responsibility of Director, Central Intelligence Agency and Heads of other Federal Intelligence Agencies

a. By virtue of the mission assigned in the National Security Act of 1947 to the DCI to produce intelligence related to the national security (defined by the National Security Council as “national intelligence”) employing as far as practicable the resource of existing agencies, and his mission to coordinate intelligence activities to this end, the DCI is primarily responsible for the implementation of this Directive. His available means are the coordinated and integrated resources of departmental intelligence agencies of the Government and of CIA itself. The DCI is responsible for mobilizing these resources in the production of national intelligence and for the final approval and dissemination of national intelligence estimates and studies.

b. The other members of the IAC are responsible in no less degree than the DCI, however, in their obligation to support with their full resources the mechanism set up in this Directive to accomplish national intelligence objectives. They are individually responsible for making available the talent and resources of their respective agencies in all phases of the production of national intelligence.

c. “National intelligence” hereafter is limited to intelligence estimates and studies which have passed through the NES and the reviewing procedures of the IAC and have been disseminated under the final responsibility of the DCI; provided, however, that such national intelligence shall not be disseminated without a statement indicating concurrence of the members of the Intelligence Advisory Committee or, when there is a non-concurrence, a statement of substantially differing opinion. In the event any IAC member fails to act in the preparation of, or indicate his approval or dissent with respect to, a national estimate or study, the DCI may proceed with its dissemination in accordance with his responsibility.

Attachment⁸

BRIEFING FOR MR. WEBB ON NATIONAL INTELLIGENCE
(May 4, 1950)

Introductory

The “national intelligence problem” can be very simply defined: It is the problem of finding a way of insuring that the President, the NSC and other officials concerned with national policy have the benefit of the full intelligence resources of the Government. The problem of national intelligence is thus by no means the total intelligence problem; it does not, for example, directly involve the multifarious problems of collection of information nor a number of other knotty problems in the intelligence production field. But it is in a very real sense the capstone of the intelligence business. No matter how good our collection may be, unless we have a sound mechanism for producing national intelligence, a mechanism that is not only adequate but respected by the producers and users of intelligence alike, the intelligence program of the Government has largely failed.

Now, tapping the full intelligence resources of the Government, and by this I mean intelligence talent as well as intelligence information, necessarily involves an interdepartmental process and interdepartmental cooperation. And this is something relatively new in the intelligence field.

History

Prior to World War II, the U.S. did not have any mechanism for the interdepartmental evaluation of intelligence and the production of interdepartmental estimates. (In some measure, the lack was responsible for Pearl Harbor, and doubtless the Congress had this disaster very much in mind at the time it passed legislation creating the CIA.)

Joint intelligence estimates had their beginning during the war in connection with planning by the Joint Chiefs of Staff. Evidently such interdepartmental planning had to be based on interdepartmental intelligence estimates. Such estimates were put out by a Joint Intelligence Committee composed of the intelligence chiefs of the three services and representatives from State, OSS, and OWI. This body functioned satisfactorily and its estimates undoubtedly had an important influence on the planning and conduct of the war. It still exists today, minus the civilian element, as an adjunct of the Joint Chiefs.

The JIC arrangement had one major flaw, or potential flaw: As a committee of equals, it tended to produce watered-down estimates which were the lowest common denominator of the views of the members. This difficulty was avoided in the National Security Act of 1947 and in subsequent NSC directives which created the Central Intelligence Agency with a Director empowered to express an independent view.

To assist the Director in discharging this responsibility, it was originally expected that the CIA would have a small organization of highly qualified individuals which would limit itself strictly to national intelligence problems and base its work primarily on the specialized reports and estimates produced by the departments rather than employ a large research and analysis organization of its own. (Parenthetically, I may say that this is the concept which we still think is the correct one and which we seek to establish now.)

In practice, however, CIA followed the opposite course: It developed over the course of time a very large research organization of its own which not only sought to produce national estimates from the ground up but also—through sheer weight of numbers and zeal to write—began to produce reports on all manner of subjects which, as the Dulles Committee put it, could by no stretch of the imagination be called national intelligence. Gradually the process of producing a national intelligence estimate came to be something like this. CIA conceives a topic on which it decides to produce a “national estimate”, or the topic may be given to it by higher authority. CIA then proceeds, normally from its own resources, to produce a draft estimate, which is then circulated to the other agencies for concurrence, comment or dissent. This process does not insure, in fact in our view it discourages, full participation by the other agencies in the estimate. The recipient of a CIA national intelligence estimate, concurred in by State, Army, Navy, and Air, can only be sure that what he has before him is not violently opposed by the other agencies. He can be reasonably sure that it does not represent the best intelligence estimate of which the Government’s intelligence system is capable. Finally, he can be positive that the system which produced it is the most wasteful that could be devised, since it involves, in theory, the duplication in CIA of the research and analysis facilities of each of the other agencies.

To correct this situation we are proposing the following:

1. To cut the national intelligence function clearly out of the CIA Office of Reports and Estimates and to establish a separate staff within CIA to take care of it. The ORE would henceforth be limited to those research and reporting functions of common concern which can best be performed centrally.

2. To limit this staff, to be called the National Intelligence Group, to a maximum of 100 professional personnel. This we believe is a generous staff but not so large that it would be tempted to produce national estimates from its own resources.

3. To require the NIG to break down requests for estimates forwarded to it, into requisitions for intelligence on the appropriate departments. Such departmental contributions, oral and written, to be synthesized by the NIG into draft estimates.

4. To give to the IAC added responsibilities in the production of national intelligence, notably

a. The formulation and recommendation to the DCI of the subjects on which national intelligence estimates are needed.

b. The detailed review of the draft estimates prepared by the NIG prior to recommending them to the DCI for dissemination.

0. Source: National Archives and Records Administration, RG 59, Records of the Department of State, Records of the Bureau of Intelligence and Research: Lot 58 D 776, National Intelligence Problem. Secret. Drafted by W.C. Trueheart.
1. Secret.
2. These and following brackets are in the source text.
3. It might here be pointed out that unquestionably many of these and other requests relate to a popular but erroneous impression that CIA possesses information not held by the other IAC agencies and can therefore produce unique reports. In fact, all CIA information is distributed to the appropriate IAC agencies. The latter have more information on their own fields than does the CIA. [Footnote in the source text.]
4. According to the Dulles report, this staff numbered some 600 persons as of 1 January 1949. [Footnote in the source text.]
5. Political, Cultural, Sociological Intelligence—State; Military Intelligence—Army; Naval Intelligence—Navy; Air Intelligence—Air; Economic, Scientific and Technological Intelligence—Each agency in accordance with its respective needs. [Footnote in the source text.]
6. Secret. Darling reports that General Magruder was the main drafter of this study (The Central Intelligence Agency, p. 393). This view is supported by Howe’s comments in Document 410, which apparently refer to an earlier draft. It was also known as the Webb Study.
7. Secret.
8. Secret. This was not an enclosure to Armstrong’s memorandum to Webb, which is dated 2 days earlier, but was attached to the source text. It may have been the text or set of notes for a follow-on oral briefing.