268. Memorandum From Secretary of Defense McNamara to President Johnson1


  • Actions recommended for Vietnam

Evaluation of the situation. In the report of my last trip to Vietnam almost a year ago,2 I stated that the odds were about even that, even with the then-recommended deployments, we would be faced in early 1967 with a military stand-off at a much higher level of conflict and with “pacification” still stalled. I am a little less pessimistic now in one respect. We have done somewhat better militarily than I anticipated. We have by and large blunted the communist military initiative—any military victory in South Vietnam the Viet Cong may have had in mind 18 months ago has been thwarted by our emergency deployments and actions. And our program of bombing the North has exacted a price.

My concern continues, however, in other respects. This is because I see no reasonable way to bring the war to an end soon. Enemy morale has not broken—he apparently has adjusted to our stopping his drive for military victory and has adopted a strategy of keeping us busy and waiting us out (a strategy of attriting our national will). He knows that we have not been, and he believes we probably will not be, able to translate our military successes into the “end products”—broken enemy morale and political achievements by the GVN.

[Page 728]

The one thing demonstrably going for us in Vietnam over the past year has been the large number of enemy killed-in-action resulting from the big military operations. Allowing for possible exaggeration in reports, the enemy must be taking losses—deaths in and after battle—at the rate of more than 60,000 a year. The infiltration routes would seem to be one-way trails to death for the North Vietnamese. Yet there is no sign of an impending break in enemy morale and it appears that he can more than replace his losses by infiltration from North Vietnam and recruitment in South Vietnam.

Pacification is a bad disappointment. We have good grounds to be pleased by the recent elections, by Kyʼs 16 months in power, and by the faint signs of development of national political institutions and of a legitimate civil government. But none of this has translated itself into political achievements at Province level or below. Pacification has if anything gone backward. As compared with two, or four, years ago, enemy full-time regional forces and part-time guerrilla forces are larger; attacks, terrorism and sabotage have increased in scope and intensity; more railroads are closed and highways cut; the rice crop expected to come to market is smaller; we control little, if any, more of the population; the VC political infrastructure thrives in most of the country, continuing to give the enemy his enormous intelligence advantage; full security exists nowhere (not even behind the US Marinesʼ lines and in Saigon); in the countryside, the enemy almost completely controls the night.

Nor has the Rolling Thunder program of bombing the North either significantly affected infiltration or cracked the morale of Hanoi. There is agreement in the intelligence community on these facts (see the attached Appendix).3

In essence, we find ourselves—from the point of view of the important war (for the complicity of the people)—no better, and if anything worse off. This important war must be fought and won by the Vietnamese themselves. We have known this from the beginning. But the discouraging truth is that, as was the case in 1961 and 1963 and 1965, we have not found the formula, the catalyst, for training and inspiring them into effective action.

Recommended actions. In such an unpromising state of affairs, what should we do? We must continue to press the enemy militarily; we must make demonstrable progress in pacification; at the same time, we must add a new ingredient forced on us by the facts. Specifically, we must improve our position by getting ourselves into a military posture that we credibly would maintain indefinitely—a posture that makes trying [Page 729] to “wait us out” less attractive. I recommend a five-pronged course of action to achieve those ends.
Stabilize US force levels in Vietnam. It is my judgment that, barring a dramatic change in the war, we should limit the increase in US forces in SVN in 1967 to 70,000 men and we should level off at the total of 470,000 which such an increase would provide.4 It is in my view that this is enough to punish the enemy at the large-unit operations level and to keep the enemyʼs main forces from interrupting pacification. I believe also that even many more than 470,000 would not kill the enemy off in such numbers as to break their morale so long as they think they can wait us out. It is possible that such a 40 percent increase over our present level of 325,000 will break the enemyʼs morale in the short term; but if it does not, we must, I believe, be prepared for and have underway a long-term program premised on more than breaking the morale of main force units. A stabilized US force level would be part of such a long-term program. It would put us in a position where negotiations would be more likely to be productive, but if they were not we could pursue the all-important pacification task with proper attention and resources and without the spectre of apparently endless escalation of US deployments.

Install a barrier. A portion of the 470,000 troops—perhaps 10,000 to 20,000—should be devoted to the construction and maintenance of an infiltration barrier. Such a barrier would lie near the 17th parallel—would run from the sea, across the neck of South Vietnam (choking off the new infiltration routes through the DMZ) and across the trails in Laos. This interdiction system (at an approximate cost of $1 billion) would comprise to the east a ground barrier of fences, wire, sensors, artillery, aircraft and mobile troops; and to the west—mainly in Laos—an interdiction zone covered by air-laid mines and bombing attacks pin-pointed by air-laid acoustic sensors.

The barrier may not be fully effective at first, but I believe that it can be made effective in time and that even the threat of its becoming effective can substantially change to our advantage the character of the war. It would hinder enemy efforts, would permit more efficient use of the limited number of friendly troops, and would be persuasive evidence both that our sole aim is to protect the South from the North and that we intend to see the job through.


Stabilize the Rolling Thunder program against the North. Attack sorties in North Vietnam have risen from about 4,000 per month at the end of last year to 6,000 per month in the first quarter of this year and 12,000 per month at present. Most of our 50 percent increase of deployed attack-capable [Page 730] aircraft has been absorbed in the attacks on North Vietnam. In North Vietnam, almost 84,000 attack sorties have been flown (about 25 percent against fixed targets), 45 percent during the past seven months.

Despite these efforts, it now appears that the North Vietnamese-Laotian road network will remain adequate to meet the requirements of the Communist forces in South Vietnam—this is so even if its capacity could be reduced by one-third and if combat activities were to be doubled. North Vietnamʼs serious need for trucks, spare parts and petroleum probably can, despite air attacks, be met by imports. The petroleum requirement for trucks involved in the infiltration movement, for example, has not been enough to present significant supply problems, and the effects of the attacks on the petroleum distribution system, while they have not yet been fully assessed, are not expected to cripple the flow of essential supplies. Furthermore, it is clear that, to bomb the North sufficiently to make a radical impact upon Hanoiʼs political, economic and social structure, would require an effort which we could make but which would not be stomached either by our own people or by world opinion; and it would involve a serious risk of drawing us into open war with China.

The North Vietnamese are paying a price. They have been forced to assign some 300,000 personnel to the lines of communication in order to maintain the critical flow of personnel and materiel to the South. Now that the lines of communication have been manned, however, it is doubtful that either a large increase or decrease in our interdiction sorties would substantially change the cost to the enemy of maintaining the roads, railroads, and waterways or affect whether they are operational. It follows that the marginal sorties—probably the marginal 1,000 or even 5,000 sorties—per month against the lines of communication no longer have a significant impact on the war. (See the attached excerpts from intelligence estimates.)

When this marginal inutility of added sorties against North Vietnam and Laos is compared with the crew and aircraft losses implicit in the activity (four men and aircraft and $20 million per 1,000 sorties), I recommend, as a minimum, against increasing the level of bombing of North Vietnam and against increasing the intensity of operations by changing the areas or kinds of targets struck.5

Under these conditions, the bombing program would continue the pressure and would remain available as a bargaining counter to get talks started (or to trade off in talks). But, as in the case of a stabilized level of US ground forces, the stabilization of Rolling Thunder would remove the prospect of ever-escalating bombing as a factor complicating our [Page 731] political posture and distracting from the main job of pacification in South Vietnam.

At the proper time, as discussed on pages 6–7 below,6 I believe we should consider terminating bombing in all of North Vietnam, or at least in the Northeast zones, for an indefinite period in connection with covert moves toward peace.


Pursue a vigorous pacification program. As mentioned above, the pacification (Revolutionary Development) program has been and is thoroughly stalled. The large-unit operations war, which we know best how to fight and where we have had our successes, is largely irrelevant to pacification as long as we do not lose it. By and large, the people in rural areas believe that the GVN when it comes will not stay but that the VC will; that cooperation with the GVN will be punished by the VC; that the GVN is really indifferent to the peopleʼs welfare; that the low-level GVN are tools of the local rich; and that the GVN is ridden with corruption.

Success in pacification depends on the interrelated functions of providing physical security, destroying the VC apparatus, motivating the people to cooperate, and establishing responsive local government. An obviously necessary but not sufficient requirement for success of the Revolutionary Development cadre and police is vigorously conducted and adequately prolonged clearing operations by military troops, who will “stay” in the area, who behave themselves decently and who show some respect for the people.

This elemental requirement of pacification has been missing.

In almost no contested area designated for pacification in recent years have ARVN forces actually “cleared and stayed” to a point where cadre teams, if available, could have stayed overnight in hamlets and survived, let alone accomplish their mission. VC units of company and even battalion size remain in operation, and they are more than large enough to overrun anything the local security forces can put up.

Now that the threat of Communist main-force military victory has been thwarted by our emergency efforts, we must allocate far more attention and a portion of the regular military forces (at least half of the ARVN and perhaps a portion of the US forces) to the task of providing an active and permanent security screen behind which the Revolutionary Development teams and police can operate and behind which the political struggle with the VC infrastructure can take place.

The US cannot do this pacification security job for the Vietnamese. All we can do is “massage the heart.” For one reason, it is known that we do not intend to stay; if our efforts worked at all, it would merely postpone the eventual confrontation of the VC and GVN infrastructures. The [Page 732] GVN must do the job; and I am convinced that drastic reform is needed if the GVN is going to be able to do it.

The first essential reform is in the attitude of GVN officials. They are generally apathetic, and there is corruption high and low. Often appointments, promotions, and draft deferments must be bought; and kickbacks on salaries are common. Cadre at the bottom can be no better than the system above them.

The second needed reform is in the attitude and conduct of the ARVN. The image of the government cannot improve unless and until the ARVN improves markedly. They do not understand the importance (or respectability) of pacification nor the importance to pacification of proper, disciplined conduct. Promotions, assignments and awards are often not made on merit, but rather on the basis of having a diploma, friends or relatives, or because of bribery. The ARVN is weak in dedication, direction and discipline.

Not enough ARVN are devoted to area and population security, and when the ARVN does attempt to support pacification, their actions do not last long enough; their tactics are bad despite US prodding (no aggressive small-unit saturation patrolling, hamlet searches, quick-reaction contact, or offensive night ambushes); they do not make good use of intelligence; and their leadership and discipline are bad.

Furthermore, it is my conviction that a part of the problem undoubtedly lies in bad management on the American as well as the GVN side. Here split responsibility—or “no responsibility”—has resulted in too little hard pressure on the GVN to do its job and no really solid or realistic planning with respect to the whole effort. We must deal with this management problem now and deal with it effectively.

One solution would be to consolidate all US activities which are primarily part of the civilian pacification program and all persons engaged in such activities, providing a clear assignment of responsibility and a unified command under a civilian relieved of all other duties.7 Under this approach, there would be a carefully delineated division of responsibility between the civilian-in-charge and an element of COMUSMACV under a senior officer, who would give the subject of planning for and providing hamlet security the highest priority in attention and resources. Success will depend on the men selected for the jobs on both sides (they must be among the highest rank and most competent administrators in the US Government), on complete cooperation among the US elements, and on the extent to which the South Vietnamese can be shocked out of their present pattern of behavior. The first work of this reorganized [Page 733] US pacification organization should be to produce within 60 days a realistic and detailed plan for the coming year.

From the political and public-relations viewpoint, this solution is preferable—if it works. But we cannot tolerate continued failure. If it fails after a fair trial, the only alternative in my view is to place the entire pacification program—civilian and military—under General Westmoreland. This alternative would result in the establishment of a Deputy COMUSMACV for Pacification who would be in command of all pacification staffs in Saigon and of all pacification staffs and activities in the field; one person in each corps, province and district would be responsible for the US effort.

(It should be noted that progress in pacification, more than anything else, will persuade the enemy to negotiate or withdraw.)

Press for negotiations. I am not optimistic that Hanoi or the VC will respond to peace overtures now (explaining my recommendations above that we get into a level-off posture for the long pull). The ends sought by the two sides appear to be irreconcilable and the relative power balance is not in their view unfavorable to them. But three things can be done, I believe, to increase the prospects:

Take steps to increase the credibility of our peace gestures in the minds of the enemy. There is considerable evidence both in private statements by the Communists and in the reports of competent Western officials who have talked with them that charges of US bad faith are not solely propagandistic, but reflect deeply held beliefs. Analyses of Communistsʼ statements and actions indicate that they firmly believe that American leadership really does not want the fighting to stop, and that we are intent on winning a military victory in Vietnam and on maintaining our presence there through a puppet regime supported by US military bases.

As a way of projecting US bona fides, I believe that we should consider two possibilities with respect to our bombing program against the North, to be undertaken, if at all, at a time very carefully selected with a view to maximizing the chances of influencing the enemy and world opinion and to minimizing the chances that failure would strengthen the hand of the “hawks” at home: First, without fanfare, conditions, or avowal, whether the stand-down was permanent or temporary, stop bombing all of North Vietnam. It is generally thought that Hanoi will not agree to negotiations until they can claim that the bombing has stopped unconditionally. We should see what develops, retaining freedom to resume the bombing if nothing useful was forthcoming.

Alternatively, we could shift the weight-of-effort away from “Zones 6A and 6B”—zones including Hanoi and Haiphong and areas north of those two cities to the Chinese border. This alternative has some attraction in that it provides the North Vietnamese a “face saver” if only problems [Page 734] of “face” are holding up Hanoi peace gestures; it would narrow the bombing down directly to the objectionable infiltration (supporting the logic of a stop-infiltration/full-pause deal); and it would reduce the international heat on the US. Here, too, bombing of the Northeast could be resumed at any time, or “spot” attacks could be made there from time to time to keep North Vietnam off balance and to require her to pay almost the full cost by maintaining her repair crews in place. The sorties diverted from Zones 6A and 6B could be concentrated on the infiltration routes in Zones 1 and 2 (the southern end of North Vietnam, including the Mu Gia Pass), in Laos and in South Vietnam.8

To the same end of improving our credibility, we should seek ways—through words and deeds—to make believable our intention to withdraw our forces once the North Vietnamese aggression against the South stops. In particular, we should avoid any implication that we will stay in South Vietnam with bases or to guarantee any particular outcome to a solely South Vietnamese struggle.

Try to split the VC off from Hanoi. The intelligence estimate is that evidence is overwhelming that the North Vietnamese dominate and control the National Front and the Viet Cong. Nevertheless, I think we should continue and enlarge efforts to contact the VC/NLF and to probe ways to split members or sections off the VC/NLF organization.
Press contacts with North Vietnam, the Soviet Union and other parties who might contribute toward a settlement.
Develop a realistic plan providing a role for the VC in negotiations, post-war life, and government of the nation. An amnesty offer and proposals for national reconciliation would be steps in the right direction and should be parts of the plan. It is important that this plan be one which will appear reasonable, if not at first to Hanoi and the VC, at least to world opinion.

The prognosis. The prognosis is bad that the war can be brought to a satisfactory conclusion within the next two years. The large-unit operations probably will not do it; negotiations probably will not do it. While we should continue to pursue both of these routes in trying for a solution in the short run, we should recognize that success from them is a mere possibility, not a probability.

[Page 735]

The solution lies in girding, openly, for a longer war and in taking actions immediately which will in 12 to 18 months give clear evidence that the continuing costs and risks to the American people are acceptably limited, that the formula for success has been found, and that the end of the war is merely a matter of time. All of my recommendations will contribute to this strategy, but the one most difficult to implement is perhaps the most important one—enlivening the pacification program. The odds are less than even for this task, if only because we have failed consistently since 1961 to make a dent in the problem. But, because the 1967 trend of pacification will, I believe, be the main talisman of ultimate US success or failure in Vietnam, extraordinary imagination and effort should go into changing the stripes of that problem.

President Thieu and Prime Minister Ky are thinking along similar lines. They told me that they do not expect the enemy to negotiate or to modify his program in less than two years. Rather, they expect the enemy to continue to expand and to increase his activity. They expressed agreement with us that the key to success is pacification and that so far pacification has failed. They agree that we need clarification of GVN and US roles and that the bulk of the ARVN should be shifted to pacification. Ky will, between January and July 1967, shift all ARVN infantry divisions to that role. And he is giving Thang, a good Revolutionary Development director, added powers. Thieu and Ky see this as part of a two-year (1967–68) schedule, in which offensive operations against enemy main force units are continued, carried on primarily by the US and other Free World forces. At the end of the two-year period, they believe the enemy may be willing to negotiate or to retreat from his current course of action.9

Note: Neither the Secretary of State nor the JCS have yet had an opportunity to express their views on this report. Mr. Katzenbach and I have discussed many of its main conclusions and recommendations—in general, but not in all particulars, it expresses his views as well as my own.

[Page 736]


Extracts from CIA/DIA Report “An Appraisal of the Bombing of North Vietnam through 12 September 1966”10

There is no evidence yet of any shortage of POL in North Vietnam and stocks on hand, with recent imports, have been adequate to sustain necessary operations.
Air strikes against all modes of transportation in North Vietnam increased during the past month, but there is no evidence of serious transport problems in the movement of supplies to or within North Vietnam.
There is no evidence yet that the air strikes have significantly weakened popular morale.
Air strikes continue to depress economic growth and have been responsible for the abandonment of some plans for economic development, but essential economic activities continue.

Extracts from a March 16, 1966 CIA Report “An Analysis of the Rolling Thunder Air Offensive against North Vietnam”11

Although the movement of men and supplies in North Vietnam has been hampered and made somewhat more costly [by our bombing],12 the Communists have been able to increase the flow of supplies and manpower to South Vietnam.
Hanoiʼs determination [despite our bombing] to continue its policy of supporting the insurgency in the South appears as firm as ever.
Air attacks almost certainly cannot bring about a meaningful reduction in the current level at which essential supplies and men flow into South Vietnam.

Bomb Damage Assessment in the North by the Institute for Defense Analysisʼ “Summer Study Group”13

What surprised us [in our assessment of the effect of bombing North Vietnam] was the extent of agreement among various intelligence agencies on the effects of past operations and probable effects of continued [Page 737] and expanded Rolling Thunder. The conclusions of our group, to which we all subscribe, are therefore merely sharpened conclusions of numerous Intelligence summaries. They are that Rolling Thunder does not limit the present logistic flow into SVN because NVN is neither the source of supplies nor the choke-point on the supply routes from China and USSR. Although an expansion of Rolling Thunder by closing Haiphong harbor, eliminating electric power plants and totally destroying railroads, will at least indirectly impose further privations on the populace of NVN and make the logistic support of VC costlier to maintain, such expansion will not really change the basic assessment. This follows because NVN has demonstrated excellent ability to improvise transportation, and because the primitive nature of their economy is such that Rolling Thunder can affect directly only a small fraction of the population. There is very little hope that the Ho Chi Minh Government will lose control of population because of Rolling Thunder. The lessons of the Korean War are very relevant in these respects. Moreover, foreign economic aid to NVN is large compared to the damage we inflict, and growing. Probably the government of NVN has assurances that the USSR and/or China will assist the rebuilding of its economy after the war, and hence its concern about the damage being inflicted may be moderated by long-range favorable expectations. Specifically:

As of July 1966 the U.S. bombing of North Vietnam had had no measurable direct effect on Hanoiʼs ability to mount and support military operations in the South at the current level.
Since the initiation of the Rolling Thunder program the damage to facilities and equipment in North Vietnam has been more than offset by the increased flow of military and economic aid, largely from the USSR and Communist China.
The aspects of the basic situation that have enabled Hanoi to continue its support of military operations in the South and to neutralize the impact of U.S. bombing by passing the economic costs to other Communist countries are not likely to be altered by reducing the present geographic constraints, mining Haiphong and the principal harbors in North Vietnam, increasing the number of armed reconnaissance sorties and otherwise expanding the U.S. air offensive along the lines now contemplated in military recommendations and planning studies.
While conceptually it is reasonable to assume that some limit may be imposed on the scale of military activity that Hanoi can maintain in the South by continuing the Rolling Thunder program at the present, or some higher level effort, there appears to be no basis for defining that limit in concrete terms, or for concluding that the present scale of VC/NVN activities in the field have approached that limit.
The indirect effects of the bombing on the will of the North Vietnamese to continue fighting and on their leadersʼ appraisal of the prospective [Page 738] gains and costs of maintaining the present policy have not shown themselves in any tangible way. Furthermore, we have not discovered any basis for concluding that the indirect punitive effects of bombing will prove decisive in these respects.

Robert S. McNamara
  1. Source: Johnson Library, National Security File, NSC Meetings File, Manila Conference. Top Secret. Also printed in The Pentagon Papers: Gravel Edition, vol. IV, pp. 348–355. McNamara submitted the report to the President after visiting Vietnam October 10–13.
  2. November 30, 1965; for text, see Foreign Relations, 1964–1968, vol. III, pp. 591594.
  3. Printed below.
  4. Admiral Sharp has recommended a 12/31/67 strength of 570,000. However, I believe both he and General Westmoreland recognize that the danger of inflation will probably force an end 1967 deployment limit of about 470,000. [Footnote in the source text.]
  5. See footnote on page 7. [Footnote in the source text. Reference is to footnote 8 below.]
  6. Section 2e, below, entitled “Press for negotiations.”
  7. If this task is assigned to Ambassador Porter, another individual must be sent immediately to Saigon to serve as Ambassador Lodgeʼs deputy. [Footnote in the source text.]
  8. Any limitation on the bombing of North Vietnam will cause serious psychological problems among the men who are risking their lives to help achieve our political objectives; among their commanders up to and including the JCS; and among those of our people who cannot understand why we should withhold punishment from the enemy. General Westmoreland, as do the JCS, strongly believes in the military value of the bombing program. Further, Westmoreland reports that the morale of his Air Force personnel may already be showing signs of erosion—an erosion resulting from current operational restrictions. [Footnote in the source text.]
  9. Lodgeʼs report of McNamaraʼs meeting with Thieu and Ky is in telegram 8161 from Saigon, October 11. (Department of State, Central Files, POL US–VIET S) Westmorelandʼs memorandum for the record of McNamaraʼs meeting with Ky, October 11, is in the Johnson Library, Westmoreland Papers, #9 History File.
  10. Not found.
  11. Not found.
  12. All brackets are in the source text.
  13. A full copy of the report, “The Effects of U.S. Bombing on North Vietnamʼs Ability To Support Military Operations in South Vietnam and Laos: Retrospects and Prospect,” August 29, is in the Johnson Library, National Security File, Country File, Vietnam, box 192, The Effects of U.S. Bombing. Attached is a memorandum from Ginsburgh to Rostow, September 13, critiquing the five points excerpted in the Appendix printed here.