The Doomsday Machine

by Daniel Ellsberg

  Their four operational missiles, at one fixed site aboveground, were thin-skinned and liquid-fueled, with highly volatile fuel that couldn’t be stored and that would take hours to load. A single U.S. missile warhead, landing several miles away, would destroy all four with near certainty. In 1961, at the high point of the Berlin crisis, in terms of actual survivable missile capability against the United States, the Soviets had no deterrent at all.

  Khrushchev had been totally bluffing about his missile production rates. He had said he was turning them out “like sausages.” That was realistic about his medium- and intermediate-range missiles within range of Europe and our overseas bases. But about ICBMs it was a flagrant lie. Moreover, it meant that he had consciously forsworn the crash effort needed to give him a credible first-strike capability in the only period when that might have been feasible.

  Our assumptions about his aims and his sense of their requirements were now put entirely in question. Or they should have been.

  My first reaction was that this startling turn of events must be made known to my colleagues at RAND as soon as possible, even though they weren’t officially authorized to see the new estimate. I flew back to Santa Monica and scheduled something that was unusual at RAND and a first for me: a Top Secret briefing. Nearly all the work at RAND except for key reports was at the Secret level. Though everyone in the building, including secretaries and maintenance staff, had to have Top Secret clearance, many employees never had occasion to use it.

  At RAND they took the regulations about classified procedures very seriously. That was never done to the same degree in the offices I frequented in Washington, where most of the documents being carried around (even in one’s briefcase going from the Pentagon to the State Department or the White House) were Top Secret. A Top Secret briefing at RAND was by invitation only, in a room with a RAND security guard at the door, checking off attendees by name on a list on a clipboard. That was something I never experienced in Washington.

  “Briefings” were the major form of oral communication of studies and results to RAND colleagues or to Air Force audiences. They were almost always accompanied by charts on a chart stand or projected on slides, with graphs or bullet points. I’d given many briefings at RAND, but never with charts. It wasn’t my style. I didn’t use the blackboards that everyone had in their offices either; I can’t think well on a vertical surface.

  But this time, when everyone had been checked off and had settled down, I started by saying, “Herman [Kahn] says you should always have charts, so for once I’ve made some.” They were on a chart stand. I’d lettered them myself, in red ink, with “Top Secret” at the top and bottom of each chart, as appropriate.

  The first chart said, “Yes, Virginia, there is a missile gap.”

  I flipped to the next one: “It is currently running about 10 x 1.”

  Then the third: “In our favor.”

  There was no response at all from the audience of about fifty department heads, top management, and key researchers filling one of our larger conference rooms at one end of the building. With puzzled looks, they waited. I explained: The latest intelligence estimate was that the Soviets had exactly four ICBMs, soft, liquid-fueled missiles at one site, Plesetsk. Currently we had about forty operational Atlas and Titan ICBMs. This was not including the intermediate-range ballistic missiles (IRBMs) we had within range of the Soviet Union, programmed to be about 120 within a year; or the Polaris sub-launched missiles that could be within range of the Soviets, almost 60 within a year. Hence, in terms just of ICBMs alone, the numbers were ten to one in our favor.

  To sum up the heated discussion that followed: no one believed me. No one. “How would they know that?” was the theme. I couldn’t tell them. They had belatedly learned, only the previous year, about the U-2 program, and then only because Khrushchev had shot down Gary Powers. Before he was captured, only a handful of RAND employees had been cleared for Talent, and those few had meticulously followed the rules and given no hint of the program to any others at RAND.

  Likewise, half a dozen or so RAND engineers had Keyhole clearance (as I learned later, when I had it). They had actually been critical to catalyzing most of the national overhead reconnaissance programs, first with planes, then balloons, the U-2, then satellites. Even if they hadn’t heard the latest results—which reflected the fact that the latest Corona passes had completed adequate coverage of all the suspected missile sites in the Soviet Union—they would have guessed immediately what the new estimate was based on. But if any of them were in my audience that day, they said nothing.

  “Why would the CIA even think we should believe this?” I wasn’t supposed to know the answer to that question myself. I knew better than to jeopardize my chance of getting the clearances (as I would later do, toward the end of the year) by revealing the basis for the intelligence estimates. At least one in the audience, Amron Katz, was a reconnaissance expert who had known about the U-2 program and knew that it hadn’t discovered any ICBMs. But he had written a number of RAND memoranda conjecturing the possibilities for the Soviets to confound our reconnaissance by camouflage and concealment and distraction. He was not inclined to believe these findings without having studied the evidence in detail (even though, unknown to me and most of the rest of us except for top management, he had played an important role in the Corona program). The others of us had all spent the last several years in anxiety of a possibly imminent Soviet threat of attack with bombers and a sizable number of missiles.

  Few, probably, took seriously the Air Force estimates—all that were officially available to RAND—of hundreds to thousands of Soviet missiles in the near future. But to the extent they had heard of the more moderate CIA estimates, they regarded those as quite possibly too low. (The rumored Army and Navy “estimates” were beneath contempt.) We had all read of McNamara’s assertion that “there was no missile gap,” but hardly anyone at RAND paid any attention to that. And at most it implied that the Soviets might not have much more than the forty ICBMs we deployed in 1961. That, in combination with bomber and submarine-launched attacks, was quite enough, according to our analyses, to paralyze SAC.

  Only a few who had seen the actual NIEs in Washington—no longer available at RAND—were even aware of the official dissenting footnotes by the Army and Navy that predicted “only a few” Soviet ICBMs in 1959, 1960, and 1961. If they had seen those, they would surely have reacted the same way my Air Force colleagues in the Pentagon did, believing that the Army and Navy were taking service bias to wild, almost treacherous extremes.

  Two of the top Soviet experts at RAND were Arnold Horelick—later head of Soviet estimates at the CIA—and Myron Rush. (Rush’s claim to fame was that he had, almost alone, predicted the rise of Khrushchev to top power from studying the sequence of photographs of Kremlin officials gathered together for parades in Red Square or other formal occasions. It was an esoteric form of intelligence that gave rise to the term “Kremlinologist.”) In 1959 they had co-authored a Top Secret memorandum—uncommon, as I’ve said, at RAND—that warned with unusual urgency that the Soviets were probably conducting a crash program on ICBMs that would give them a significant first-strike capability as early as 1959 (i.e., right then). Their main basis for this was a close analysis of all Khrushchev’s statements on the subject. Their premise was that Bolsheviks did not bluff. On that assumption, the sequence of his allusions to rockets and sausage making told them that he had already arrived at the capability he had earlier predicted and now claimed.

  They were wrong. Khrushchev had been bluffing. That was what the new estimate was saying. It was correct, as Horelick and Rush themselves acknowledged,86 not much later, in a Top Secret report that was subsequently published. But many at RAND had believed their earlier memorandum, and my briefing was not enough to change that inclination.

  More important, the estimate contradicted and essentially invalidated the key RAND studies on SAC vulnerability since 1956. Those studies had explicitly assumed a range of
uncertainty about the size of the Soviet ICBM force that might play a crucial role in combination with bomber attacks. Ever since the term “missile gap” had come into widespread use after 1957, Albert Wohlstetter had deprecated that description of his key findings. He emphasized that those were premised on the possibility of clever Soviet bomber and sub-launched attacks in combination with missiles or, earlier, even without them. He preferred the term “deterrent gap.” But there was no deterrent gap either. Never had been, never would be.

  To recognize that was to face the conclusion that RAND had, in all good faith, been working obsessively and with a sense of frantic urgency on a wrong set of problems, an irrelevant pursuit in respect to national security. That is not a recognition that most humans in an institution are quick to accept. It was to take months, if not years, for RAND to accept it, if it ever did in those terms. To some degree, it’s my impression that it never recovered its former prestige or sense of mission, though both its building and its budget eventually became much larger. For some time most of my former colleagues continued their focus on the vulnerability of SAC, much the same as before, while questioning the reliability of the new estimate and its relevance to the years ahead.

  Likewise, the Air Force, and especially SAC, was reluctant and slow to accept the new figures, despite the fact that they seemed to support what SAC and the JCS thought was a desirably tough U.S. position on the Berlin crisis. Both RAND and the Air Force expected the Soviets to build up their missile force. But that buildup, which did begin in 1963–64 (particularly after Khrushchev was replaced by Brezhnev), could never promise the Soviets the strategic advantages it might have offered in 1958–62.

  Meanwhile the Berlin crisis itself still appeared very serious. The president’s attempt to mobilize public opinion for a confrontation precisely by raising the serious possibility of nuclear war had backfired. His decision to encourage a major private fallout-shelter program was a misjudgment, mobilizing instead great controversy. The Russians continued to affirm their determination to sign a peace treaty and turn over access control of Berlin to the East Germans.

  Flying back to Washington in late September following my abortive attempt to reorient thinking at RAND, and with the Berlin game and Abe Chayes’s conclusion still fresh in my mind, I had one immediate concern: how could this new estimate be used to change our prospects in Berlin?

  West Berlin remained deep within Soviet-controlled territory. The erection of the Berlin Wall had commenced, with Kennedy’s acceptance (even relief). From Khrushchev’s point of view, that was a solution to his immediate problem: the exodus of emigrants from East Germany through Berlin. It even turned out to be an adequate solution to his longer-term problem of stabilizing the regime in East Germany, and thus strengthening the Soviets’ position in Eastern Europe. But that wasn’t immediately seen or accepted by Khrushchev, and still less by the West. Khrushchev’s ultimatum about giving control of access to the East Germans by the end of the year was still standing, as were his warnings against our trying to maintain our access by any military means.

  Now suddenly both these threats appeared to have been based on an immense, years-long bluff about his strategic “parity” with the United States. Recently discovered documents from the Soviet archives show that he was at this time bluffing his own Warsaw Pact allies87—as well as ours, in NATO—about this parity to reassure them about his management of the crisis and the risks of his apparently provocative diplomacy.

  So why not let him know, privately, that his bluff had been discovered and that he should withdraw his ultimatum and his threats? I set out to draft proposals along those lines.

  CHAPTER 11

  A Tale of Two Speeches

  For years the specter of a “missile gap” had been haunting my colleagues at RAND and in the Defense Department. The revelation that this had been illusory cast a new perspective on everything. It might have occasioned a complete reassessment of our own plans for a massive buildup of strategic weapons, thus averting an otherwise inevitable and disastrous arms race. It did not; no one known to me considered that for a moment. But in the short run, it offered other opportunities, particularly regarding the problem of Berlin.

  My first thought was for President Kennedy to convey this new understanding of the situation directly to Premier Khrushchev. He could do this through private secret channels to minimize Khrushchev’s humiliation and reluctance to back down. I wrote two memos, basically for Kennedy’s eyes. To get them to Kennedy, I gave them to Carl Kaysen, who was working for McGeorge Bundy on nuclear issues and whom I’d dealt with on the delegation problem earlier in the spring.

  One of the memos I gave him,88 dated October 9, was a recommended set of talking points for Kennedy to address to Khrushchev or to some representative of his. The other memo, “A Proposal for Educating Khrushchev,” was to explain the first to the president—to make the purpose of the message to Khrushchev explicit.

  The idea was to make it clear to Khrushchev that we knew exactly what he had, and I proposed to tell him not only the number—four ICBMs—but also the precise coordinates of the base at Plesetsk. For completeness we could include the coordinates of the Tyuratam test site, where they had a couple of test missiles. The implied message was: “You can drop all this bullshit you’ve been putting out about ‘parity’ and ‘superiority.’ We know what you’ve got and where it is. You’ve got hardly anything, and what little you have is vulnerable. So stop talking about giving us trouble on Berlin. You know, and we know, you are in no position to do that.” Those were not the words, but that was the meaning I wanted to convey.

  Kaysen read these and offered to discuss them with me. He was driving from the White House over to a meeting elsewhere in Washington and asked me to come with him in the car. He said, “Look, Dan. You’ve got to take into account the nature of the channel here,” using information-theory terms about the president, his boss. “Kennedy will simply never … he will not talk like this.” It wasn’t clear to me whether he was saying this critically about Kennedy or whether he agreed with the president’s style. He just repeated, “It’s unthinkable. Kennedy would not talk this way to Khrushchev.”

  It still seemed important to me, though, to get it through to the Soviets somehow that they should not commit themselves on this issue, as if they thought we believed their claims of strategic superiority or even equality. “We know these claims are not true, so don’t dig yourself in on threats that you are not going to be ready to carry out.” That was the minimum message I wanted Khrushchev to hear. “Commitments to threats like these carry real risks, and things can get out of hand.” But I got it from Kaysen that Kennedy wouldn’t say that directly to Khrushchev, either face-to-face with his representative or in a private message.

  A day or two after that, I was back at the Pentagon in the office of Adam Yarmolinsky, the assistant to the secretary of defense. I was still a RAND consultant, paid by RAND on an open-ended Air Force contract, but I was spending more than half the year in Washington working on papers and staff work in the Pentagon and State Department. Adam told me that he was working on a draft speech for Kennedy to give at a war college. Several agencies had been asked to send proposed drafts to the White House, and he was drafting one for McNamara to send. He asked me to look at it and add anything I thought should be included. So here was another chance to get the president to send my message—this way publicly.

  I put in many of the same themes I had included my memos, adjusting the tone to a public speech, not addressed personally to Khrushchev. I wrote them out longhand on notepad paper and gave them to Adam, who exclaimed, “This is good!” He put them all into his draft for McNamara. A little later he told me, “McNamara likes it. He’s sent it to the White House.”

  Days later I read the speech Kennedy made, which confirmed what Kaysen had said to me about Kennedy’s more conciliatory style. He hadn’t used anything I said. So I gave up on Kennedy as the channel for my message.

  But then I dr
opped in on my friend Timothy Stanley, Paul Nitze’s special assistant in ISA—where I hung out most of the time in Washington—for whom I’d worked on the war plans earlier in the year and who had checked out my memo on the LST at Iwakuni. Stanley had a little cubbyhole office across from the entry to the assistant secretary of defense’s office. (Three years later, I would inhabit this same office when I became special assistant to Nitze’s successor, John McNaughton.) Stanley said that he was drafting a speech for Roswell Gilpatric.

  I gave Tim my original handwritten notes and said, “Look, I wrote this for Kennedy and he didn’t use it. You can use it if you want for your speech.” He read what I’d written. It wasn’t a whole speech, just several pages of key points, including this statement: “Our forces are so deployed and protected that a sneak attack could not effectively disarm us.” Right after reading that, Tim looked up and read the next paragraph aloud:

  The destructive power which the United States could bring to bear even after a Soviet surprise attack upon our forces would be as great as—perhaps greater than—the total undamaged force which the enemy can threaten to launch against the United States in a first strike. In short, we have a second-strike capability which is at least as extensive as what the Soviets can deliver by striking first. Therefore we are confident that the Soviets will not provoke a major nuclear conflict.

  He asked with surprise, “Is that true?” I said, “Trust me, Tim, it’s true. That’s the way it is.” This particular calculation was pretty simple on the basis of what I’d just learned about the Soviet arsenal. Four ICBMs! One hundred and fifty-odd strategic bombers!

  Although the new NIE had not made any comparison of U.S. and Soviet forces, before or after an attack by either, it was easy for me to be confident of my “net assessment” of a “nuclear exchange” (in Pentagon language), startling as it was to anyone who had spent years hearing about the missile gap or reading RAND classified reports on SAC vulnerability. I was sure it would stand up in the course of bureaucratic reviews of the speech.