Introduction

Rear Admiral Roscoe H. Hillenkoetter became Director of Central Intelligence on May 1, 1947. His appointment had been decided upon by the National Intelligence Authority the preceding February. The Central Intelligence Agency’s first Historian, Arthur Darling (who had access to “oral tradition” as well as documents) believed that it was the Department of State’s “turn” to nominate a candidate for the job, but that the military services were unwilling to entrust the post to a civilian at a time when the pending national security legislation indicated that the post would become more important than before. Accordingly, Admiral Leahy’s influence prevailed and a naval officer was appointed. (Darling, The Central Intelligence Agency, pages 194–195)

According to an account recorded many years later by Ludwell L. Montague (then a State Department officer detailed to CIG, later a senior CIA official, and still later a historian of the period), he was told of Hillenkoetter’s appointment when he went to the White House on February 26, 1947, to consult with the President’s Naval Aide, Admiral Foskett, on another matter:

In any case, Hillenkoetter came to office with certain advantages. He was in many respects more qualified than his predecessors in terms of intelligence experience. Both Souers and Vandenberg had served tours in intelligence before becoming Director of Central Intelligence, but Hillenkoetter had had considerably more intelligence experience than either, albeit at less senior levels. He had served three tours of duty as assistant naval attaché or naval attaché at Paris and one tour with the U.S. Embassy at Vichy in 1940–1941, where Leahy came to have a high regard for his abilities as an intelligence officer. During World War II, Hillenkoetter had served as intelligence chief on the staff of the Commander in Chief, Pacific Ocean Area, Admiral Nimitz.

[Page 747]

Hillenkoetter also came to office with the advantage of becoming the first Director of Central Intelligence to serve as a statutory official; unlike his predecessors, he would head an agency established by law and enjoying regular appropriations.

At the same time, Hillenkoetter inherited most of the controversies that General Vandenberg had begun. These might well have arisen in one form or another in any case. But Vandenberg had been so energetic and aggressive and had served so briefly that most of the uproar he provoked was at full strength when Hillenkoetter took over; for example, his designation as “executive agent” of the NIA, his substantial expansion of the CIG’s analytical and research work, and his efforts to define strategic and national policy intelligence in ways that would enhance the authority of the Director of Central Intelligence.

Because such issues involved the basic definitions and ground rules of the national intelligence structure, they often tended to obscure some of the less spectacular progress that was taking place. During Hillenkoetter’s directorship, for example, there were continued (and sometimes painful) efforts to develop a scientific intelligence capability, the Atomic Energy Commission became a member of the IAB, the program of basic intelligence known as the National Intelligence Survey came into being, the working level continued to struggle with the problem of intelligence estimates and managed to produce a number of them, and in general the daily work of correlating and analyzing information went forward.

At the senior levels, however, the main intelligence issue continued to be the line of demarcation between the Central Intelligence Group/Central Intelligence Agency and the departmental intelligence organizations. At the time Hillenkoetter took office, there were two items on the agenda of the Intelligence Advisory Board related to this fundamental question. One was a draft national intelligence directive on the coordination of intelligence production (in effect a companion piece to the earlier directive on coordination of intelligence collection) which proposed to allocate by agency the responsibility for production of finished intelligence. The other was a proposal to define and spell out the Intelligence Advisory Board’s procedures and prerogatives.

Before either of these issues had come to a head, Hillenkoetter, at his first formal meeting with the National Intelligence Authority on June 26, 1947 (Document 319), asked the NIA to withdraw the delegation of power given to his predecessor at the Authority’s last meeting in February, under which the Director of Central Intelligence functioned as the “agent” of the Secretaries of State, War, and Navy on matters within his jurisdiction. (Document 185) This was a conciliatory gesture, intended to improve relationships with the departmental (especially armed services) intelligence chiefs and to ease some of the other controversies with and within the Intelligence Advisory Board.

[Page 748]

The nub of the argument over Intelligence Advisory Board procedures was the enduring debate over whether the IAB was a “board of directors” with independent authority of its own or whether it was created primarily as a sounding board for the Director of Central Intelligence, who could accept or disregard its advice as he chose. At the 14th meeting of the IAB on July 17, 1947, representatives from the Central Intelligence Group, Department of State, and Naval Intelligence frankly exchanged quite different views on the subject. (Document 320)

Eventually, an agreement was reached and on September 16, the NIA, in one of its last official acts before going out of existence, approved National Intelligence Authority Directive No. 11, which provided that recommendations to the Authority by the Director of Central Intelligence would be submitted first to the IAB and that any statements of non-concurrence would accompany the recommendation onward to the NSC. The directive also provided: “Any recommendation two or more IAB members believe should be submitted to the NIA (NSC) will be submitted by the Director of Central Intelligence in such form as to set forth his recommendation and the comments of the IAB members.” (Document 325)

With the coming into effect of the National Security Act of 1947, the Intelligence Advisory Board ceased to exist, or so it would appear, since the legislation made no mention of such a body. The issue revived, however, in the context of Hillenkoetter’s proposals for changes in the intelligence structure to reflect the new statutory authority. As we have seen (the section on the National Security Act), in a memorandum of September 11, 1947, to the National Intelligence Authority (Document 220), Hillenkoetter recommended that the existing NIA directives be continued in force for 60 days while new versions were prepared; that the National Security Council establish a subcommittee consisting of the Secretaries of State and Defense “to furnish the active direction of the Central Intelligence Agency”; and “while I believe it presumptuous and awkward on my part to suggest,” that he attend all meetings of the National Security Council. He also notified the Council (and the Departmental intelligence chiefs) that he proposed to establish an Intelligence Advisory Committee “which will to all intents and purposes continue in existence the Intelligence Advisory Board established under the National Intelligence Authority.” (Document 222) He accompanied his recommendation with a proposed charter for the committee which included the provisions of NIA Directive No. 11 concerning the committee’s right to submit differing views to the National Security Council.

The National Security Council at its first meeting approved Hillenkoetter’s recommendations concerning the continuance in force of NIA Intelligence Directives for 60 days and agreed to his attendance at meetings of the Council. The Council took no action on his proposal for a [Page 749] supervising subcommittee of the NSC to oversee the CIA, but soon approved Souers’ amended version, which deleted the part of his recommendation that would have carried over the provisions of NIA Directive No. 11. (Document 325)

The disputes between the Director of Central Intelligence and the departmental intelligence heads resumed in earnest in November, when Hillenkoetter put before them the proposed revisions of the National Intelligence Authority directives (henceforth to be known as National Security Council Intelligence Directives or NSCIDs). The NSCIDs, like their predecessors, would be, under the legislation, the basic charters for the national intelligence system, and as such they offered considerable possibility for controversy. The State Department’s intelligence chief, W. Park Armstrong, for instance, thought Hillenkoetter’s drafts went “beyond the intent of Congress,” which “was not … to establish a single head for all Federal intelligence agencies.” (Document 331)

At a meeting between departmental representatives and Hillenkoetter on November 20, 1947, the discussion of the proposed directives was contentious. (Document 332) An ad hoc working group established at the meeting prepared a revised set of directives which were unacceptable to Hillenkoetter. By December 8, however, the air had cleared sufficiently so that at a follow-on meeting, Hillenkoetter and the departmental intelligence heads reached agreement on the first batch of directives. What happened in the meantime to bring about this change is not documented, but according to Darling (The Central Intelligence Agency, page 215) the precipitating event was Secretary of the Army Royall’s letter of November 26, 1947, to Souers (Document 230) concerning Souers’ amended version of Hillenkoetter’s recommendation for an Intelligence Advisory Committee. Royall objected that the new arrangement failed to follow the old in not requiring the Director of Central Intelligence to submit all proposed recommendations to the NSC through the IAC in order to obtain the latter’s views.

According to Darling’s account, this finally led Hillenkoetter to seek a meeting with Forrestal, who in turn summoned the Army and Navy Secretaries and intelligence chiefs to his office. At that session, Hillenkoetter presented his view of how the intelligence system should function under the National Security Act, and Forrestal then simply instructed the intelligence chiefs that they were not to interfere. (Darling, The Central Intelligence Agency, pages 215–216)

There appears to be no way of knowing whether or not this is a complete and accurate account. In any case, at the National Security Council meeting on December 12, 1947, the issue also arose in the context of an oral report that Hillenkoetter gave on “organization, activities and plans” of the Central Intelligence Agency. (Document 335) According to the spare account in the minutes, “Secretary Forrestal said, and the other [Page 750] members agreed, that there was no question as to the coordinating authority of the Director of Central Intelligence. Secretary Forrestal said that Admiral Hillenkoetter tells the Departments what he needs for collation and evaluation and the Departments are required to provide it.” (Ibid.)

Although a truce of sorts appeared to have been declared, Hillenkoetter’s problems were just beginning. A month after the NSC had seemed to give him a vote of confidence, it decided that “two or three specially qualified individuals not in the Government service should make a comprehensive, impartial, and objective survey of the organization, activities, and personnel of the Central Intelligence Agency.” (Document 336) Within another month, the Intelligence Survey Group, or Dulles committee, had come into being, consisting of Allen Dulles as chairman, Mathias F. Correa, and William H. Jackson.

There is some uncertainty about the origins of the Dulles Survey, although the common denominator in all accounts is Secretary of Defense Forrestal. Souers, writing some years later, placed the origin in discussions he had with Forrestal (presumably around the time the National Security Act came into effect) on how the NSC would discharge its responsibilities for supervising the Central Intelligence Agency. Forrestal proposed that Souers exercise oversight on behalf of the Council. When Souers demurred, the two agreed “to appoint a committee to make a survey so that the N.S.C. members would be able to carry out their responsibilities by obtaining the independent judgment of a committee whose members were reasonably qualified to advise the Council.” (Souers to Montague, October 30, 1969; Central Intelligence Agency Historical Files No. 206106, Job 84T00286R, Box 2, Folder 12; see also Souers to Montague, August 13, 1970; CIA Historical Files No. 206102)

Montague thinks that the real impetus came from Robert Blum of Forrestal’s staff, who had a watching brief over intelligence and was strongly influenced by published criticism of the intelligence machinery. (Montague, General Walter Bedell Smith as Director of Central Intelligence, pages 39–41) Darling believes that Blum and a colleague on Forrestal’s staff, John Ohly, were the architects of the investigation but also notes that Hillenkoetter himself had proposed to Forrestal in the fall of 1947 that at the start of the new system under the National Security Act of 1947, it was a good time to take stock in the intelligence field. (Darling, The Central Intelligence Agency, pages 298–301) Others may have shared Hillenkoetter’s idea.

Whatever the source of Dulles’ inquiry (and all of the above accounts may be correct), there seems to have been an uneasiness about the intelligence system. The controversies that had engrossed the intelligence agencies were not purely bureaucratic struggles. They represented strongly held differences of view in which there was often merit [Page 751] on more than one side of the case. Some of the criticism also seems to have come from within CIA, reflecting internal differences about how the agency should function. Perhaps more than anything else, whether justified or not, there was impatience and concern over a perceived lack of “intelligence success,” a sense that the U.S. intelligence system was inadequate and its accomplishments limited at best, and a feeling that the resources invested had not produced commensurate results.

In April 1948, the survey group’s executive secretary expressed some of this thinking when he complained that the “CIA has not performed well or not performed at all its two functions of coordinating government intelligence activities and of correlating the evaluation of intelligence. CIA’s mission and actual operations within both these fields are uncertain, undefined and subject to much controversy and bitterness,” and he agreed with other government agencies’ “criticisms that CIA is organized as a top heavy bureaucracy and is hampered by a predominance of military personnel in key positions.” (Document 344)

The Dulles survey group seems to have taken a fairly activist approach from the beginning. As noted earlier, Kennan consulted Dulles about the psychological warfare problem in April 1948, and in the following month, under the stimulus of NSC debate on covert operations, the survey group produced an “interim report” on the relationship between secret intelligence and secret operations. Later in May, Forrestal and Lovett invited Dulles to meet with them to discuss the subject and in the course of the conversation offered him the job of heading up covert operations under the emerging “autonomous” arrangements. Dulles declined. (See the section on Psychological and Political Warfare)

In another memorandum of June 4, 1948, the Dulles inquiry’s executive secretary argued that the group must decide on the type of report it wished to submit to the National Security Council, and continued:

There was also a separate inquiry into intelligence in 1948 by the Hoover Commission’s Committee on National Security Organization (also known as the Eberstadt committee). Moreover, in the aftermath of the Bogota riots of April 1948, which broke out during the Inter-American Foreign Ministers Conference, a Congressional investigation into the handling of intelligence on the disturbances led to recriminations [Page 752] between the Department of State and CIA and to widespread press allegations of an “intelligence failure.”

Although the Intelligence Advisory Committee was fairly quiescent during 1948, this seems to have been mainly because it met so few times during the year. This may have reflected in part a sense that the Dulles inquiry presaged major changes, but it is also possible that Hillenkoetter was simply trying to reduce frictions by convening the committee as little as possible. A number of unsettling questions persisted, however. One of the most serious was the vexing issue of “strategic and national policy intelligence.” The Department of State in its presentation to the Dulles survey group strongly criticized CIA for duplicating departmental intelligence analysis rather than exercising its coordinating function by using the departmental agencies as part of the national intelligence process. (Document 354) This issue was to break out again next year with renewed vigor after presentation of the Dulles report.

Although a considerable amount of source material is available on the 1947–1948 period, the same tendency toward formal bureaucratic documentation noted earlier is even more evident. At a good many key junctures, there is no written record that takes the reader behind the scenes into the policy process. Compared to 1946 and 1947, when the IAB met fairly often, the absence of IAC minutes from the fall of 1947 through the end of 1948 eliminates one of the major sources for intelligence history during this period. Some of the material connected with the Dulles inquiry gives a sense of what lies behind the more formal documentation, but in general much reading between the lines is required.

From the beginning of 1949 to the outbreak of the Korean war in June 1950, the national intelligence system was preoccupied with the Dulles Report and its aftermath, the effort to put into effect the survey group’s recommendations on the organization and management of the national intelligence effort. Although the intelligence structure continued to develop and take shape at the working level, there is an inescapable sense that at the policymaking levels this was a period of drift and uncertainty, in which there was a marked inability to resolve conflicts or to solve problems.

The Dulles Report was a long and complicated document covering a wide span of activities by the Central Intelligence Agency and the departmental intelligence elements. Overall, the Dulles survey group criticized the Central Intelligence Agency for failure to perform its coordinating functions adequately, for failing to fulfill its assigned responsibilities to produce national intelligence, and for poor management of both itself and its interdepartmental responsibilities. The report made no recommendations for changes in existing legislation and agreed that the National Security Council was the proper supervising body for the Central Intelligence Agency. But it urged much closer liaison between the [Page 753] Agency and the Secretaries of State and Defense. (An extract from the report is printed as Document 358; the entire report is in the Supplement.)

The Dulles Report was particularly critical of the Central Intelligence Agency’s management of the intelligence process. On the “most vital problem of coordination of intelligence activities,” for example, it “emphasized that coordination can most effectively be achieved by mutual agreement among the various agencies. With the right measure of leadership on the part of the Central Intelligence Agency, a major degree of coordination can be accomplished in that manner.” (Document 358) On national intelligence (intelligence estimates), the Dulles committee was similarly blunt: “With one or two significant exceptions, whose occurrence was largely fortuitous, the Central Intelligence Agency has not as yet effectively carried out this most important function.” (Ibid.)

The report went on to recommend that the Central Intelligence Agency should draw the departmental agencies more deeply into the estimative process and replace its Office of Reports and Estimates (whose performance the survey group criticized) with a small estimates division which would synthesize the estimative contributions of the other agencies. Finally, the Dulles Report proposed that the Intelligence Advisory Committee be revitalized so as to consider and pass upon all estimates and to become collectively responsible for them.

In the chapter on “The Direction of the Central Intelligence Agency,” the report was harsh in its criticisms. While acknowledging the difficulties of the CIA’s work, the strong pressures to show results, and the short period of time which the Agency had to “demonstrate substantial accomplishments,” the survey group nonetheless complained that the “directing staff of the Central Intelligence Agency has not demonstrated an adequate understanding of the mandate of the organization or the ability to discharge that mandate effectively,” and it proceeded to detail the CIA’s unsatisfactory working relations with other departments. Finally, on the question of whether the Director of Central Intelligence should be a military officer or a civilian, the report declared “that the Director should be civilian because we are convinced that continuity of tenure is essential and complete independence of Service ties desirable for carrying out the duties of the Director. The post cannot properly be filled as a mere tour of duty between military assignments. Unless there is such continuity of service and complete independence of action, the Director will not be able to build up the esprit de corps, the technical efficiency, the loyalty of home staff and field workers, which are essential to the success of the enterprise.” (See the Supplement)

Given the nature of the Dulles report, and in particular its strong criticism of the management and direction of the Central Intelligence Agency, Hillenkoetter’s position as Director of Central Intelligence had [Page 754] obviously become a difficult one. He was certainly aware that senior officials at the Department of State and elsewhere believed that he should be replaced. The adoption of NSC 10/2 by the NSC 6 months earlier had been, in a sense, a limited vote of no-confidence in Hillenkoetter. The Dulles Report signified more explicit vote of disapproval. Nonetheless, Hillenkoetter remained as Director of Central Intelligence for another 22 months after its submission. During this period his effectiveness in the position was reduced, and there was an impasse over the reform of the intelligence system. But, for whatever reason, Hillenkoetter remained in office.

The Central Intelligence Agency comments on the Dulles Report were submitted to the Executive Secretary of the National Security Council on February 28, 1949. (Extracts are printed as Document 371; the full text is in the Supplement) The tone was set in the first paragraph, which stated: “The observations of the Committee are, in general, accurate, and its objectives are sound; its conclusions, however, are, in many respects, faulty, and the recommendations for their attainment are, in many cases, impracticable.” (Document 371) In general, the CIA’s memorandum of comments, which rejected most of the Dulles Report’s basic findings, was as sharp in the tone of its defense as the Dulles Report had been in its criticisms. The report’s criticism of the Agency’s performance in the field of national intelligence estimates drew a particularly strong rejoinder, perhaps because the report had been highly specific in its proposals for remedies, most of which were unacceptable to the Agency.

The CIA also rejected the Dulles Report’s proposal that the Intelligence Advisory Committee assume collective responsibility for national intelligence estimates. It claimed instead that “the Central Intelligence Agency has sole responsibility for its estimates … and does not share this responsibility collectively with the members of the Intelligence Advisory Committee.” (Ibid.) The comments further charged that the “Report does not reflect a clear understanding of this vital aspect of the coordination theory. To the Committee, ‘national’ intelligence is merely ‘coordinated’ intelligence, and coordinated intelligence is joint intelligence, at the mercy of departmental bias.” (Ibid.)

At only one point in the CIA memorandum is there a glimpse of the personal feelings that must have been engendered by the situation. Concurring with the Dulles committee’s statement that continuity of service was required for the Director of Central Intelligence, the CIA memorandum elaborated on the deleterious effects on employee morale and productivity resulting from changes of directors and organization. This harm, it added with scarcely veiled sarcasm, “was particularly noticeable last Fall just before election when literally dozens of rumors were extant in Washington that one of the Committee members was to become Director of Central Intelligence as soon as the election was over.” (Ibid.)

[Page 755]

The Navy and the Department of State also reacted to the Dulles Report. Admiral Inglis, Director of Naval Intelligence, reasserted in a memorandum of March 4, 1949, his long-held position that the Intelligence Advisory Committee should be explicitly reconstituted as a board of directors superior to the Director of Central Intelligence. (See the Supplement)

The Department of State recommendations emphasized other changes that reflected the Department’s major concerns and objectives in the intelligence field at the time: active IAC participation with the Director of Central Intelligence “in the continuing coordination of intelligence activities and in the production of finished estimates;” CIA utilization of “the facilities of the IAC members who should assume collective responsibility for them;” integration of secret operations and secret intelligence “in a single self-administered office within CIA;” and appointment of a civilian “of considerable stature and prominence, possessing the requisite qualifications of experience and willingness to serve,” as Director of Central Intelligence. (Document 378)

When the National Security Council met to consider the Dulles findings on April 7, 1949, it was confronted not only with a bulky report consisting of lengthy analyses and many complex recommendations, but also with a number of papers from the intelligence agencies offering comments, observations, or dissents. It was manifestly impossible for the NSC itself to sort out such a large volume of proposals and recommendations and it was decided that the Secretaries of State and Defense, as the two members of the Council most directly affected, should arrange for the preparation of a more concise document for further Council consideration. (Document 380) The two Secretaries designated General Joseph T. McNarney (an adviser to the Secretary of Defense) and Carlisle H. Humelsine (Executive Secretary of the Department of State) for this purpose.

The resulting document, formally designated as NSC 50 and more familiarly known as the McNarney Report, was submitted to the NSC on July 1, 1949. (Document 384) In effect, it reduced the Dulles Report to its recommendations (omitting most of the analysis and commentary) and commented on them, endorsing some, disagreeing in whole or in part with others, and proposing alternative or modified courses of action. For example, the McNarney Report agreed that the IAC should be more active in intelligence coordination and proposed that the Director of Central Intelligence should be formally designated the chairman of the IAC. It also endorsed the criticism of deficiencies in national intelligence, but rejected the proposal for the collective responsibility of the DCI and the IAC in the estimative process (apparently equating this with requiring unanimous agreement). Instead, the McNarney Report recommended a procedure under which all intelligence estimates would either [Page 756] be concurred in by the IAC or, in those cases where agreement could not be reached, there would be concurrent submissions of dissenting views with the estimate.

NSC 50 further proposed that CIA “should interpret and follow the NSC Intelligence Directives so as to refrain as far as possible from competitive intelligence activities in the production of research intelligence estimates,” and it concurred only in part in the Dulles committee’s criticism about military officers in key CIA positions, recommending instead that the matter be worked out by the DCI in collaboration with the Secretaries of State and Defense. On the question of the directorship, McNarney agreed on the importance of continuity, suggesting that the director should be a civilian or, if a military or foreign service officer, either retired or serving a final tour of duty. Finally, the McNarney Report gave a limited endorsement to the Dulles Report’s overall strictures on CIA’s leadership, while softening the criticism and labeling it “too sweeping.”

The aftermath of the Dulles and McNarney Reports extended over the following year and beyond. In a series of documents known as the “Four Problems,” the Department of State proposed an enhancement of the IAC through the strengthening of its Standing Committee and its secretariat; agreed ground rules governing the production of national (estimative) intelligence; and a general curtailment of research and analytical activity by CIA in favor of the Departments. In general, the thrust of the paper was to structure the estimative process along lines that would increase the participation of the Departmental intelligence components in the process and carefully define CIA’s coordinating and synthesizing role. (Document 393)

CIA rejected most of the Department of State proposals as well as the Dulles Report’s recommendations concerning the mechanism for producing estimates. In May 1950, a joint State–Defense proposal (the Webb Report or Webb–Magruder Report) on the organization of work on national estimates was presented, again proposing detailed rules governing the estimative process and defining a more limited central role for the CIA in the production of estimates. It was not forwarded to Hillenkoetter until July, however. By then the Korean war had broken out, which distracted attention from these proposals, and by October, Hillenkoetter had returned to the Navy and General Walter Bedell Smith had become Director of Central Intelligence. Most of the intelligence problems with which the government had been grappling in the period since the Dulles Report was submitted were still unresolved. But the war and the marked differences in style (and situation) between Hillenkoetter and Smith would create a new environment for dealing with them.