I’ve never had a stronger sense in another person of a kindred awareness of this situation and intensity of his concern to change it. Thirty years later, McNamara revealed79 in his memoir In Retrospect that he had secretly advised President Kennedy, and after him President Johnson, that under no circumstances whatever should they ever initiate nuclear war.† He didn’t tell me that, but it was implicit in everything he had said at this lunch. There is no doubt in my mind that he did give that advice, and that it was the right advice. Yet it directly contradicted the mad “assurances” on U.S. readiness for first use he felt compelled to give repeatedly to NATO officials (including speeches I drafted for him) throughout his years in office, as the very basis for our leadership in the alliance.
McNamara’s assistant, Adam Yarmolinsky, had joined us for the last part of the lunch, without saying anything. When we left McNamara’s office Adam took me into his small, adjoining room and said that he had never seen McNamara prolong a lunch that way. He had talked more frankly with me than Yarmolinsky had ever heard him talk with anyone else. The point of Adam’s telling me this, and of my repeating it now, was to give weight to what he said next. “You must tell no one outside of this suite what Secretary McNamara has told you.”
I asked if he was referring to fears of the reaction from Congress and the JCS (I could have added, “NATO”). He said, “Exactly. This could lead to his impeachment.” I told him that I understood. But he went on to make that more explicit. “By no one,” he said, “I mean, not Harry Rowen, not anybody.” Now, that I understood. Evidently he knew that Harry was my closest friend and confidant, a cleared colleague with whom I normally would have shared even such sensitive information—though I’d been told not to tell anyone—unless I was specifically told not to. I never did tell anyone, not even Harry, what McNamara had said, though he would have found it as heartening as I did. But I did ask Adam one question: “As far as you know, is President Kennedy’s thinking on these subjects different from the secretary’s?”
Adam held up his thumb and forefinger pressed together, no space between them, and said, “Not an iota.”
I left the secretary’s suite thinking that here, in Robert McNamara, was someone whose judgment was worthy of my greatest trust. He had, as I saw it, the right perspective on the greatest dangers in the world, and the power and determination to reduce them. And he and his assistant had the street-savvy to know that if he wanted to achieve that, he had to keep his cards very close to his chest.
* * *
On July 25, 1961, President Kennedy issued a tough speech in connection with the Berlin crisis, calling up reserves for a possible confrontation over Berlin, warning the public that nuclear war was a real possibility, and calling for a national fallout-shelter program. Herman Kahn had argued that to make a credible first-strike threat, we had to be prepared to show that we would survive a retaliatory strike with our fallout shelters, or at least believe that we would. To do that you had to act as if you believed—as Kahn did—that shelters would make all the difference, and you had to encourage people to build them. I remember hearing at the time that McGeorge Bundy had said, “We’re going to do this not for the Herman Kahn reasons,” by which he meant that we weren’t making a first-strike threat or counting on the shelters to work; we were just … what else? Making a prudent effort that might help if a nuclear war happened to come about, I suppose.
But in fact there was no other reason than Kahn’s for the president to be talking about fallout shelters at that time. If a nuclear war had come about that year, it would have been only because our efforts to maintain access to Berlin had led us, the United States, to initiate the use of nuclear weapons, which almost surely meant a resort to general war. To be sure, the Kennedy administration didn’t make it explicit to the American public that the nuclear risks of his policy involved a U.S. first use or first strike, as Kahn might have done if he were writing the speech.
Nevertheless, the speech did set off a frenzy of concern about fallout shelters—and a great commercial interest in selling them for private homes. Charlie Hitch, the head of the economics department at RAND and the man who had hired me, actually built a fallout shelter in his backyard. (As I recall, it was eventually used for wine storage.) So did Willard Libby of the Atomic Energy Commission. His shelter, as it turned out, burned down in the midst of the Cuban missile crisis the next year, leading Leo Szilard to comment that this proved not only that there was a God but that He had a sense of humor. There were discussions in Life magazine of the ethics of equipping your shelter with a machine gun to repel neighbors without the foresight to have built one of their own from attempting to muscle into your fallout shelter. Some Catholic and Protestant theologians concluded that, yes, it was within Christian ethics to protect your family in that way.
Khrushchev’s response to Kennedy’s tough stand was to begin construction of the Wall around East Berlin on August 13. That stopped the hemorrhaging of skilled workers and their families from East Germany to the West, which had been the most urgent pressure on the Soviet regime to change the status of West Berlin. But Khrushchev didn’t withdraw his year-end deadline for turning over control of access to Berlin to the East Germans, a development, we believed, that inevitably would lead to war.
* * *
In late August 1961, I visited Strategic Air Command headquarters in Omaha to find out the SAC reaction to the cable I had drafted, with Colonel Lukeman’s help, for McNamara to send to General Thomas Power, commander in chief of SAC. The cable urged Power to find ways to adapt current planning and operations as soon as possible to the war-plan guidance to the JCS I had drafted, which wasn’t scheduled for full implementation until the next year.
I talked to Colonel Dave Liebman, now chief of war plans for SAC, whom I had known and worked with earlier when he frequented the offices in the Air Force Plans Division along with Lukeman. Liebman said that, after some initial reservations, my guidance had been received with approval. He said the attitude in Omaha, radiating from General Power, was “we can work with this.” That was good news to my ears; I wouldn’t have bet on it before the visit. (Looking back, I should have been uneasy, at best, that my guidance had gone down so well with Power.)
In the course of the conversation, he remarked how unhappy he and most of his colleagues at SAC were with President Kennedy’s lack of resolve and toughness with respect to Berlin. He mentioned that the president was perceived as being scared of the prospect of nuclear war, even though, he remarked, at the urging of his boss General Power, the JCS had assured the president that “if worst came to worst” and it was necessary to go to general war over Berlin, “a preemptive attack on the Soviet Union would result in less than ten million deaths in the U.S.”
The words “less than ten million deaths,”80 as a reassurance, came more easily from the lips of a SAC officer than from most other humans aside from Herman Kahn, but even so I was startled to hear an estimate that low. I said, “Ten million? That’s the population of metropolitan New York! One large warhead, or a couple of them, on New York or L.A. could give you that. How could Power say it would be that low?”
“Well, that’s what he believes, and that’s what the JCS told the president,” Liebman said. “They told the president that he should understand, in going into his bargaining with the Russians, that he had the capability to back up his threats to that extent if worst came to worst.”
Obviously, they weren’t including Allied casualties in Western Europe, although the Soviets had hundreds of medium- and intermediate-range missiles within range of Europe, along with medium-range bombers—more, in fact, than we had ever predicted. (It later became clear that the U.S. intelligence community had underestimated Soviet nuclear forces targeted on Europe and England by as much as they had overestimated Soviet capabilities against the United States. In particular, the Soviets had bought themselves a medium-range missile force capable of making one deep smoking hole of West Germany: a final solution to their German
problem.) What is more, they had also produced and deployed missiles and bombers that could cover all our overseas bases, which were scattered throughout Western Europe, North Africa, the U.K., and Japan. Soviet attacks on these targets, which couldn’t remotely be eliminated by our forces striking preemptively, would effectively annihilate the populations of all these areas.
Moreover, as the JCS had informed the president earlier, merely the fallout from our own attacks on the Soviet bloc would result in a hundred million dead in Western Europe, along with another hundred million deaths in other areas contiguous to the Soviet Union and China. But the JCS evidently presumed the president would be so much less concerned with allies and neutral civilian deaths resulting from U.S. preemptive escalation and Soviet retaliation that they didn’t need to mention any of this.
From the documentary record, they could have been right about that. It was typical of U.S. strategists, then and later, to leave European, North African, and Asian casualties entirely out of account in weighing the deterrent balance. And I don’t know of any instance of a president or any civilian official raising this point. In retrospect, that’s a startling commentary.
Later in the conversation in Liebman’s office, we were discussing the CIA’s latest estimate of current Soviet missiles, issued in June: 50 to 100 ICBMs as of mid-1961.81 The USAF assistant chief of staff for intelligence did not concur; he believed there were “at least” 120 ICBMs, maybe more, with 300 expected by mid-1962. Likewise, the director of intelligence in the State Department thought there were 75 to 125 missiles at present, but “possibly” 200, with 150 to 300 in a year.
There was also a dissenting footnote from the intelligence branches of the Army and Navy maintaining that the Soviets had deployed only “a few” missiles from mid-1960 to mid-1961. When I read the June 7 National Intelligence Estimate (NIE) in the Pentagon, it was the first time I’d seen this outlier estimate in print (along with a page of arguments justifying it). By the decision of the Eisenhower administration, contractor corporations like RAND had been cut off from receiving NIEs in 1958. From that time on, RAND employees had been exposed only to Air Force intelligence estimates for Soviet offensive forces. We knew these tended to be higher than CIA estimates. I’d heard bitter rumors about Army and Navy heresy on the missile gap from Air Force officers, who regarded it as virtually treasonous. They saw the Army and Navy as willing to jeopardize national security by espousing fairy tales, with no other reason than to minimize the Air Force budget requests for missiles. This was the first time I’d actually seen in official writing what they were talking about.
Aside from the presumptively service-biased Army and Navy branches, then the intelligence community put the lowest estimate of Soviet ICBMs in June as slightly higher than the forty U.S. Atlas and Titan missiles operational at the time, or perhaps twice that, but possibly ranging upwards of three to five times as many as ours. The key projection was when the Soviets would or might have 300 or more. That was generally agreed to be the number that would allow coverage by missiles alone of all SAC bases in the United States and abroad as well as our soft ICBM sites. General Thomas Power had testified to Congress that the Soviets might have had that critical, dangerous arsenal in 1960, as Herman Kahn had been predicting. The CIA, as of June, put that as an upper level in mid-1963. The State Department had it as an upper level in mid-1962, and the Air Force definitely estimated about 300 by mid-1962, with about 550 in mid-1963 and over 1,000 in 1965.
Those latter estimates underlay the Air Force pressure for huge increases in force size. McNamara was just then confronting the question of the prospective scale of a force of Minuteman ICBMs, solid-fuel (quickly launched) missiles in hardened silos, a kind of deployment that didn’t exist yet on either side. McNamara couldn’t admit inside the Pentagon that he was even considering a number as low as one thousand, which was his private target. General Power, with LeMay’s support, was asking for ten thousand. McNamara told the president82 that we really didn’t need more than four hundred, but that a thousand was the lowest figure he could get through Congress and “not get murdered.”
The last year in which U.S. retaliation to a Soviet first strike would depend entirely on “soft” air bases and missile sites, subject to destruction by two hundred to three hundred Soviet ICBMs, was 1962. After that, thousands of Soviet ICBMs would be needed for high assurance of destroying in a first strike the large number of hardened ICBM silos the United States was programming (aside from U.S. submarine-launched Polaris missiles). In other words, 1962 was the last year the Soviets could hope to achieve a disarming first-strike capability with moderate to high confidence.
However, when I asked Liebman why the Air Force was continuing to contradict the lower CIA estimate, he gave a more concrete reason. “We just don’t believe it. There is too much evidence that they have more than that.” Then he said, “Do you know what the old man [General Power, his boss at SAC] thinks they have?”
I waited to be told.
“One thousand. He’s sure they have a thousand. Right now.”
I thought for a moment, then asked, “How many missiles does he think he knows the exact location of?” I was asking how many ICBMs SAC thought it could target now, out of the thousand that Power believed they had.
“About two hundred.”
“Two hundred,” I repeated. I remember pausing for a moment before saying, “So that leaves about eight hundred ICBMs whose location we don’t know well enough to target?”
He nodded.
I said, “How does that fit with the estimate that we would have less than ten million casualties following a U.S. first strike?”
There was a long pause. Liebman narrowed his eyes and scrunched his mouth. Then he said, “You know, that’s a very interesting question. I don’t think I’ve ever heard it raised before.” He thought for a while more, then said, “There’s someone I’d like to hear you put that question to.”
He took me down into the underground bowels of SAC headquarters and introduced me to the chief of the Air Estimates Division in SAC Intelligence, Colonel George J. Keegan Jr. Liebman described him to me as a “real intellectual,” and I’d already heard of him in the Pentagon as the “father of the missile gap.” (He was one of several rivals for that honor. In the late 1970s he was a fervid proponent of a “death-beam gap”83: a race for a “directed energy, charged particle beam” in which he claimed the Soviets were ahead.)
Liebman told Keegan, who was accompanied in his dimly lit office underground by a couple of other colonels, that I had just raised an interesting question, which he asked me to repeat. I did. Keegan didn’t answer it. Instead he reacted almost exactly the way Liebman had. He looked at me expressionlessly and said, “That is an interesting question. Hmmm …”
After a short silence I said, “You know, if you’re trying to encourage the president to take a strong stand with the Russians over Berlin, it might not serve your purpose to tell him he’s facing a thousand Soviet missiles.”
Keegan sat up sharply at this and seemed shocked and incredulous. “You’re not suggesting, are you, that we should fudge our estimates?”
He looked piercingly at me, and I looked piercingly back at him, searching his face for irony and not finding any. He seemed totally unselfconscious of the widespread reputation of Air Force estimators—SAC’s above all—for inflating their numbers. But this was not, it seemed, the moment to share a smile about this.
I said, “Certainly not. Of course not.” (Fudge? I wanted to say, “Heavens, no!”)
“But …” I went on carefully, “if there should be a range of uncertainty, it might not be best from every point of view to emphasize only the upper end of that range.”
Shortly, Liebman led me away.
* * *
In September there was a political-military simulation game on Berlin run by Tom Schelling, my former Harvard mentor on bargaining theory, who was now running a number of such games for the Pentagon. This one involved quite high-
level participants, some of them current officials and some retired officials, both military and civilian. It had participants as high ranking as General Maxwell Taylor, for example, who was shortly to be the chairman of the Joint Chiefs, and later ambassador to Vietnam. At this time he was still a top-level military advisor to Kennedy in the White House.
The game was run as a command post sort of exercise where we would sit around a table and we would get simulated cables—sent by the game controller, Schelling—as if we were getting them in the war room from Germany and from various military posts, a flow of messages supposedly in real time. One of them, I remember, was that students were demonstrating at the Free University in Berlin against our military moves. (Just over a year later, Walt Rostow and I would be reading a real cable with the same substance in an actual crisis over missiles in Cuba; see chapter 12.)
The 1961 game went through various phases of probes to Berlin by us and Soviet repulses of various kinds. I remember very little of it except that, as usual in these games, it was quite hard to get the Blue Team, the U.S. side—on which I was—to decide to use nuclear weapons, though that was the basis of our actual planning. That was so obviously fraught with total disaster that it didn’t look as though anybody on our side could imagine doing that very quickly. That wasn’t, it seemed to me, because it was “only a game.” As in other simulations Schelling had designed, the participants got very caught up in it, and there was a realistic sense of urgency and tension.
Well, fine. You wouldn’t want to discover that you were working alongside a lot of officials who thought nothing of starting a nuclear war. Only, that implied that our actual contingency plans for forcing access to Berlin had a large element of unreality. Or to put it another way, it meant that they were bluffs. Or that they should be bluffs: because the reluctance of the gamers to initiate nuclear war—as the plans called for if the Soviets used anything like their full available forces in East Germany to block access—seemed sane, far saner than carrying out the plans. But there was also the possibility—all too likely—that they would turn out not to be bluffs because some U.S. units, outnumbered in the field, would use their nuclear weapons without orders, to defend or avenge themselves, against the wishes of U.S. and NATO high command. There were no locks on those tactical weapons (any more than on SAC’s strategic weapons) to prevent them from doing so.
The Doomsday Machine Page 19