Foreign Relations of the United States, 1950–1955, The Intelligence Community, 1950–1955

189. Paper Prepared in the Central Intelligence Agency¹

Washington, August 25, 1954.

THE “NET ESTIMATES” PROBLEM

Introduction. One of the most basic problems faced by intelligence agencies is that of obtaining adequate information of operational matters and of using such information to produce meaningful “net estimates” of the capabilities and intentions of other nations, taking account of our own acts and facts as they must appear to others. In general, U.S. doctrine in the military services has prescribed a sharper separation between intelligence and operations than exists in the UK and some, if not most, other nations, although in the practical operation of field staffs this separation is usually mitigated or overcome through working understandings. At the highest levels in the services, however, and even in the State Department, there is a strong tendency either to keep operational matters wholly in “operational” channels, or, if they are conveyed to intelligence offices, to impose restrictions against discussion with outside agencies. This memorandum deals with efforts made by CIA since the fall of 1950 to meet this problem, in three contexts: (1) National Intelligence Estimates handled through regular machinery; (2) specific “net estimates” or “net evaluations” handled by special machinery; (3) the Watch Committee, handling intelligence from the warning standpoint.

1. Operational Information and National Intelligence Estimates.

In the NIE field, a distinction must be made between the furnishing of operational assumptions and the furnishing of specific operational facts, especially concerning our own capabilities. The former has never presented great difficulties, and is now in satisfactory shape. The latter, however, has been troublesome on several occasions, and no satisfactory overall solution has been reached.

In one type of estimate, dealing with the consequences of possible U. S. courses of action, operational assumptions as to U.S. policy are the foundation of the estimate. The only problem is that of ensuring that the assumptions have appropriate backing, and this is now usually done through the CIA Adviser to the NSC Planning Board, who may consult the Planning Board as a whole or may deal directly with the departments most concerned. Assumptions are cleared at whatever level is necessary to assure their solidity.

In the more general type of estimate, dealing with probable developments in a given situation, it is occasionally necessary to have—in addition to the always implicit assumption of no drastic change in overall US policy—specific assumptions on such matters as US aid levels, where the US may affect the situation drastically and immediately. Since the estimate may be designed to provide the basis for policy in these very respects, it is sometimes necessary to make an arbitrary assumption for the future. For example, the currently pending estimate on developments in Taiwan, to be completed for submission to the Planning Board when the question of aid levels is discussed, will be based on the assumption that the scope and nature of US programs remains as at present. Thus, the estimate will not purport to be definitive (even within the usual limits of predictability) but will be a benchmark obviously subject to adjustment if a decision is made to alter the scope and nature of the programs.…² In any event, there is no substantial present problem in obtaining such assumptions, which are usually framed in consultation with the CIA Advisor to the NSC Planning Board, or occasionally directly by departments principally concerned.

In contrast with the relative simplicity of the policy-assumptions problem, the problem of obtaining specific own-capability facts—or even assumptions—has been complex and difficult. Early in the history of the present NIE machinery, in April 1951, an ad hoc solution was reached for one case, an estimate (NIE 27)² on the likelihood of invasion of Taiwan. For this estimate, it was obvious that the intelligence community needed to know, generally at least, the dispositions of US forces in the area, since these forces were the principal obstacle to Chinese Communist action. After some negotiation via the service intelligence heads, appropriate “assumptions” approximating the real facts were provided and used.

This specific case was not then made the basis for a general solution, although the need was discussed at the working level. No specific proposal was submitted by O/NE to the Director, or by him to other agencies.

In two major fields, experience has subsequently highlighted the vacuity of estimates prepared without clear knowledge of our own capabilities. With respect to Soviet Bloc capabilities to attack Western Europe, all estimates through 1950 had been able to proceed on the assumption of virtually no Western opposition. From 1951 onward, this assumption became increasingly less valid, and in the preparation of the estimates there were prolonged discussions leading finally to the use of a fairly meaningless formula that the Soviet Bloc could “launch” a lot of campaigns, including a full-scale offensive in Western Europe. Whether any meaningful answer could have been provided in Washington without duplicating the activities of SHAPE is doubtful, but the fact is that no machinery existed even for getting and incorporating (with proper credit) the current conclusions of SHAPE. As they finally stood the estimates were certainly not helpful to anyone on this point.

It was the second field, however, that of air defense of the Soviet Union, which seemed to General Smith even more forceful, and the experience in this field, in the spring and summer of 1952, contributed heavily to the campaign launched by him in October 1952, as discussed in the next section. What happened was simple. A pioneer national estimate on Soviet air defense capabilities ran into prolonged agency disagreements, in which it became more and more clear that any description of the effectiveness of Soviet defenses depended entirely on assumptions as to our capabilities and strategy of attack. Eventually, it was recognized that without clear guidance on these matters the estimate was, in the words of General Smith, merely an inventory of Soviet assets, and it was finally approved by the Intelligence Advisory Committee after most qualitative statements had been cut out, and after the title had been altered to, “The Scale and Nature of the Soviet Air Defense Effort 1952–54.”³

So far as National Intelligence Estimates are concerned, the situation since then has been as before—that operational information is not made available to the intelligence community on any systematic basis adequate for its employment in such estimates. With the development of increasingly close and cordial working relationships both in the IAC itself and at the National Estimates Board and Staff levels, it is safe to say that a great deal of such information is in fact fed informally into the estimates. But much operational information is still withheld wholly, and the overall situation is far from satisfactory.

2. Special Machinery for “Net Estimates”

In the summer of 1951, concern over continental defense, within the National Security Council, led to discussion of the need for an authoritative “evaluation” of the net capability of the USSR to injure the United States. Although General Smith made no affirmative effort to have sole primary responsibility for the effort—and in fact is believed to have expressed his opposition to the assignment—the Joint Chiefs of Staff were diffident, and the upshot was an NSC directive, of August 30, 1951,⁴ that DCI prepare a “summary evaluation” in collaboration with the Joint Chiefs of Staff and with the internal security committees.

The resulting exercise, handled through regular channels and without any central mechanism, was a nightmare. The Intelligence Advisory Committee speedily produced the necessary basic estimate of Soviet gross capabilities (SE–14, 18 October 1951),⁵ and the internal security committees furnished adequate contributions. However, the Joint Chiefs of Staff contribution was delayed for several months and when finally produced, in May 1952, proved to be based on the most extreme possible estimates of Soviet capabilities and on several other questionable assumptions, of a largely intelligence nature, concerning Soviet attack strategy. General Smith regarded this contribution as unsatisfactory and assigned two members of the Board of National Estimates, Mr. E. M. Hoover and Vice Admiral B. M. Bieri (former Deputy Chief of Naval Operations), to the task of producing an integrated evaluation merging all contributions, and modifying the JCS contribution. The ensuing procedure was trying to all concerned. It finally produced a report which General Smith accepted as the best obtainable in the circumstances, and which he forwarded to the NSC on 14 October 1952.⁶

In his accompanying memorandum General Smith noted the defects of the report, and gave three “primary reasons” for them. The third of these was:

“c. There is at present no machinery to plan, guide, coordinate and produce an appraisal or estimate based on the integration of national intelligence with military, political, and economic operational data dealing with our own capabilities.”

General Smith recommended that as an interim measure he be instructed to prepare terms of reference for a more adequate study of the problem, and that the Council:

“c. Concurrently, instruct the Director of Central Intelligence to examine, in collaboration with officials of other governmental bodies as needed, the adequacy of present machinery, and the character of any new machinery that may be required in order to plan, guide, coordinate, and produce for the National Security Council, upon request, evaluations in the nature of ‘Commander’s Estimates,’ of the USSR’s capabilities and intentions vis-à-vis the United States, based upon the integration of military, political, and economic operational data dealing with United States’ capabilities and intentions, and national intelligence.”

When General Smith’s recommendations were forwarded by the Secretary of Defense to the Joint Chiefs of Staff for comment, the JCS responded by a sharply critical memorandum, dated November 21, 1952.⁷ There ensued negotiations, which were limited to the terms of reference and procedure for a new study, but which also gave an opportunity for General Smith to clarify his ideas to the Joint Chiefs of Staff on the overall problem. In the light of the change of administrations then in process it was finally decided to let the overall recommendation (subparagraph c. quoted above) lie over, while proceeding with a new “net evaluation” on the basis of an entirely novel procedure. This procedure, embodied in NSC 140,⁸ was approved by President Truman on January 19, 1953, and accepted by the Eisenhower Administration without change. It created a Special Evaluation Subcommittee, chaired by a direct Presidential appointee, Lt. General Idwal H. Edwards, USAF (Ret.)—who was in fact nominated by the Joint Chiefs of Staff under a gentlemen’s agreement with General Smith—with representatives of the JCS, CIA, and the internal security committees as full members.

The so-called Edwards Committee was given a very short deadline, May 15, 1953, for the preparation of a new net evaluation to serve as the basis for a policy appraisal of the whole field of continental defense. Through the able leadership of General Edwards and with a large share of credit also to Lt. General H. R. Bull, representing CIA, the deadline was met. The resulting report (NSC 140/1, May 18, 1953)⁹ although slightly marred by one misunderstanding with the JCS working level on terms of reference, was a highly valuable effort substantively. Organizationally it seemed to CIA, and, it is believed, to the NSC Secretariat and to others familiar with the earlier failure and with the general problem, to prove that net evaluations or estimates could be done, even on the most complex problems, through a process of constant interchange of intelligence and operational information (under appropriate security safeguards), and that the resulting net papers were a vast improvement on anything that could be done by intelligence and operations working at arms’ length from each other.

With this pioneer demonstration, attention turned for a time back to the more general problem. Largely as a result of CIA urging, the report of the President’s Committee on International Information Activities (Jackson Committee), published June 30, 1953,¹⁰ included as its very first recommendation the following:

“1. The necessary measures should be taken to provide net estimates of political, economic and military capabilities.” (Page 3)

While this recommendation might have provided a lever for reopening the over-all question and resurrecting General Smith’s recommendation c. of the preceding October (which was still technically unfinished business in the NSC), two factors combined to make this appear undesirable to the DCI. One was the replacement of most of the Joint Chiefs of Staff in August 1953; the other was the fact that in its policy consideration of continental defense (NSC 159 series)¹¹ the NSC referred all organization questions to the Office of Defense Mobilization for study (NSC action No. 873 d., August 1953).¹² Since it was at first thought that Office of Defense Mobilization’s study would be quickly completed, it seemed clearly wiser to await it, and to work out agreement with the Joint Chiefs of Staff on a proper procedure in the limited field of continental defense before taking up the over-all question again.

As it worked out, ODM encountered substantial delay both in preparing its recommendations and in getting them cleared by the NSC Planning Board. Only in April 1954 did ODM’s paper become available to NSC.¹³ At this point the JCS comment on the ODM proposals revealed that the “new Chiefs” were not happy about a net estimates procedure under which ultimate final responsibility was not vested solely in them, and there ensued a round of negotiations between Mr. Dulles and Admiral Radford, which clarified the issue but failed to produce agreement.¹⁴ A split paper was finally submitted to the President and the NSC on June 9, 1954,¹⁵ and the issue was resolved by the President in favor of joint responsibility in the Chairman of the Joint Chiefs of Staff and the Director of Central Intelligence, with other government agencies represented on an appropriate basis.¹⁶ The President designated Admiral Radford as Chairman for the exercise, which is now proceeding with a deadline of 1 November 1954. Rear Admiral Robbins is directing the project on Admiral Radford’s behalf, while General Bull is again representing CIA. Substantively the scope of the study has been broadened to cover specifically all key US installations overseas, thus making possible a far more refined and comprehensive view of Soviet attack strategy than was possible in the Edwards group. Procedurally, it appears at this writing that the project is being handled at the working level with a greater degree of close cooperation even than in the Edwards group.

During the course of the negotiations on the particular continental defense problem, the over-all issue was discussed, and the Director, of Central Intelligence proposed that the device of joint responsibility in DCI and the Chairman of the JCS, the one as intelligence adviser, the other as military adviser, to the President and the NSC, be extended generally to cover problems for which a specific net evaluation, or net estimate, procedure is appropriate. (Such problems, generally speaking, would be those of substantial scope, involving a complicated study of the interplay of US action and Soviet counteraction. In situations where action on both sides would be simple and predictable, the use of joint machinery might not be warranted.) For the present, this suggestion is not being pressed, since the success of the new continental defense project appears to be important in any decision.

3. The Watch Committee and Operational Information

For that part of the intelligence community that concentrates on providing warning of hostile action, knowledge of US or allied operations may be even more vital than for more long-range intelligence efforts. Without such knowledge false warnings may be given, available intelligence may be seriously misconstrued (in either direction), and intelligence effort may not be focused properly at points and areas of tension.

The 1948–54 history of this problem is covered in detail in reports prepared by the Office of Current Intelligence, the CIA component particularly charged with the watch function and with support of the Watch Committee. The essence of the story is that there has been no remotely adequate procedure for keeping the Watch Committee informed of operational matters that could have a bearing on its activities. However, as a result of the work of an ad hoc Committee designated by the IAC in 1953 to review the whole Watch Committee process, the Director of Central Intelligence, in July 1954, submitted to the NSC, and the President approved in principle, a directive that operational information necessary to the Watch Committee’s mission be furnished to it, under appropriate security safeguards.¹⁷ At the date of writing, the exact wording of this directive remains to be ironed out with the FBI Director. In CIA’s judgment the substance of this directive will provide a broad and adequate basis for the proper functioning of the Watch Committee in this respect, though no doubt particular problems of detail and interpretation will arise. The directive provides that in cases where an agency believes that overriding security considerations preclude release of information (conceded to be significant to the Watch Committee), the matter shall be referred initially to DCI, thereafter to the President if DCI and the referring agency are unable to resolve it. Thus, the Director should be in a position to iron out difficulties as they arise.

It is significant that this directive has had the full support of the JCS, in principle and in its general breadth.¹⁸ This JCS attitude may be a significant indication of the possibilities for further improvement in the fields of national estimates and of special net evaluation machinery, discussed in the earlier sections of this memorandum.

1. Source: National Archives, RG 263, HS/HC: HRP 82–2/00022, Box 1, HS/HC 111, Misc. Documents. Secret. The original went to General Clark; copies were sent to Assistant Director for National Estimates (Kent), Deputy Director for Intelligence (Amory), and William Bundy.
2. Ellipsis in the original.
2. For NIE 27, see Foreign Relations, 1951, vol. VII, Part 2, pp. 1623–1624.
3. See footnote 2, Document 120.
4. NSC Action No. 543, Document 86.
5. Not printed. (National Archives, RG 263, Soviet NIEs, 1950–1955, #20, Box 1) The actual date of SE–14 is October 23, 1951.
6. Document 131.
7. Not found, but See Document 137 and footnote 3 thereto.
8. For NSC 140, see Foreign Relations, 1952–1954, vol. II, Part 1, pp. 205–208.
9. Text is ibid, pp. 328–349. See also Document 149.
10. See Document 151.
11. See Foreign Relations, 1952–1954, vol. II, Part 1, pp. 465–489.
12. See footnote 2, Document 173.
13. Not found.
14. See Documents 176 and 177.
15. Not found.
16. See Document 182.
17. President Eisenhower approved at the 209th meeting of the NSC; See Document 187.
18. See Document 186.