Code Warriors

by Stephen Budiansky

  But nothing could heal the gaping distrust that now yawned where secrecy and plausible deniability once held fast. On October 25, 1978, President Jimmy Carter signed into law the Foreign Intelligence Surveillance Act, which for the first time imposed statutory limits on NSA’s activities. The act required the approval of the attorney general and a new Foreign Intelligence Surveillance Court for NSA to conduct electronic surveillance against foreign targets within the United States or against U.S. citizens and permanent residents anywhere in the world; it set up a permanent system of congressional oversight of NSA activities; and it specifically outlawed the kinds of dragnet collection operations within the United States that the Shamrock program had entailed, regardless of whatever subsequent measures the agency made to minimize the reading of communications of U.S. persons that it snared in the process. The most salient passage made it a crime, punishable by fine and imprisonment, to engage in “electronic surveillance under color of law” except as specifically authorized by statute. It was the first recognition that for covert intelligence operations to be accepted as legitimate and justifiable in a democratic society, they also had to be legal.21

  —

  At the same time, however, the very act of bringing covert operations under a legal framework marked another downward step in the erosion of traditional American values: it was a formal acquiescence in the kind of moral ambiguity in foreign affairs that after decades of Cold War conflict had come to seem normal, no longer a temporary or deplorable necessity but enshrined in permanent and legally chartered institutions of government.

  At the bottom of it all was, of course, the superpower nuclear standoff, which, based as it was on the constant threat of mass annihilation of entire civilian populations, had a way of making a mockery of any traditional notions of ethics and principle in the conduct of national policy: when the alternative was cataclysm, many compromises seemed acceptable, even desirable, by comparison. The policy of détente championed by Kissinger during the Nixon and Ford administrations aimed to reduce the risk of nuclear war by stabilizing the relationship between the United States and the Soviet Union, but in seeking to lock in the status quo of the superpower rivalry, détente also further froze in place the Cold War institutions that had grown up along with it. Mutually assured nuclear destruction, espionage and surveillance, and all of the other moral compromises of the Cold War once seen as aberrations were instead becoming parts of the world order.

  Détente might reduce U.S.-Soviet tensions, but it was not therefore going to do anything about reducing the now multibillion-dollar-a-year global espionage enterprise. The efforts of each side to learn the negotiating positions of their adversary in talks that began in the mid-1970s to freeze or reduce nuclear arms and stabilize the political status quo in Europe if anything intensified the targeting of government communications links in Moscow and Washington by the rival spy agencies. The need for each side to verify compliance with arms treaties and reassure itself that in stepping away from the brink of nuclear war it had not left itself vulnerable to military surprises or technological breakthroughs in weapons systems by the other placed even more demands on intelligence.

  In Washington, a thicket of antennas began sprouting on the rooftop of the Soviet embassy on 16th Street, just a few blocks from the White House. In August 1974, acting on a warning from NSA, Kissinger informed President Ford, “It is very probable that the Soviets are intercepting out-of-city telephone conversations of key Washington officials, since such calls are usually on radio links which can be intercepted with rather simple and commercially available equipment.” A report “identified 10,000 leased government circuits terminating in the Washington area for which protection seemed prudent. About 4,000 of these circuits are now on microwaves and exploitable, and the remaining 6,000 are already on cable but must be tagged to see that they remain there.”22

  A Soviet defector, Arkady Shevchenko, subsequently confirmed that other well-situated facilities used by Soviet diplomats in the United States—an eighteen-story residential complex in Riverdale in the Bronx, located on one of the highest spots in New York City; a Long Island mansion on thirty-seven acres in Glen Cove that served as a recreational retreat; and another recreational facility on the Eastern Shore of Maryland—were used to eavesdrop on microwave towers that served not only the major long-distance telephone links to Washington and New York, but also Andrews Air Force Base, where the president’s Air Force One was based, and Norfolk, Virginia, home of the U.S. Atlantic Fleet. “It is most unlikely that these sites were selected for any other reason than microwave interception,” an AT&T analysis concluded. The Soviets’ Washington-area intercepts of unencrypted phone calls by U.S. officials and defense contractors were, according to KGB defector Vasili Mitrokhin, Moscow’s “most important source of intelligence on the foreign and defense policies of the Ford and Carter administrations.”23

  Kissinger lodged a protest about the Soviet antennas with Ambassador Dobrynin, but the fact was that the United States was over a barrel in pressing the matter. By January 1977 the White House national security adviser, Brent Scowcroft, was able to report that “government communications in the Washington area have been rerouted from microwave to cable, and government communications in New York and San Francisco are in the process of being moved to cable,” but cautioned that implementing more robust protections—such as providing private U.S. telephone companies with NSA-developed technology to bulk-scramble all telephone calls—posed the classic signals intelligence dilemma: “The main problem, from a foreign intelligence perspective, in moving ahead with communications protection is that it may stimulate the Soviets to take even greater protective measures for their own telecommunications and thereby deny us a valuable and possibly irreplaceable source of information.”24

  A plan to jam the Soviets’ listening posts was also considered, but rejected for the same reason. Aside from the questionable effectiveness and cost—NSA estimated an initial outlay of $1.8 million just to target the Soviet embassy, the Soviet School in Washington, and the San Francisco consulate—a National Security Council memo noted that “NSA has consistently opposed initiating jamming operations, because they believe the U.S. will be a net loser in a jamming war.” The United States might have been more subtle than the Soviets in its use of embassies abroad as surveillance posts, but the fact was that since at least the mid-1960s the U.S. embassy in Moscow was regularly intercepting the conversations of top Soviet officials as they spoke on their car phones while being chauffeured about the Russian capital in their ZIL limousines. That operation, code-named Gamma Guppy, came to an abrupt end in 1971 after investigative reporter Jack Anderson revealed its existence in a column under the headline “CIA Eavesdrops on Kremlin Chiefs.” According to a subsequent briefing by U.S. counterintelligence officials, the Soviets changed their car phone communications immediately afterward and a CIA agent who worked as a mechanic on the limousines “was never heard from again and presumed killed.”25

  But Gamma Guppy was just one of many similar embassy-based operations that exploited insecure Soviet communications throughout the latter decades of the Cold War. In a rare admission of a U.S. SIGINT success during this era, Vice Admiral John “Mike” McConnell told a seminar at Harvard’s Kennedy School of Government shortly after retiring as NSA director in 1996, “In the mid-1970s, NSA had access to just about everything the Russian leadership said to themselves and about one another….We knew [Soviet leader Leonid] Brezhnev’s waist size, his headaches, his wife’s problems, his kids’ problems, his intentions on the Politburo with regard to positions, his opinion on American leadership, his attitude on negotiating positions.” The information more than once allowed Kissinger to outmaneuver Soviet negotiators in the Strategic Arms Limitations Talks. “That’s the sort of thing that pays NSA’s wages for a year,” a senior U.S. official who saw the intercepts told David Kahn.26

  The reliance on “close access” methods to do what conventional cryptanalysis could not intens
ified the rivalry between NSA and CIA, both of which laid claim to operate embassy listening posts and more intrusive operations involving taps, bugs, and a wave of new devices that could pick up plaintext directly from electronic equipment such as fax machines, copiers, electric typewriters, and computer terminals. The two agencies were “competing for targets, locations,” said Admiral Bobby Inman, who became NSA director in July 1977; there was also the usual gamesmanship in which neither shared its results with the other until they had already extracted and reported what they could.27

  CIA undeniably possessed the expertise in surreptitious entry, and its Staff D had a long history of carrying out breakins in foreign embassies to plant bugs or copy cryptographic material, but Inman saw that not only was there considerable duplication between the two programs but also a pressing need to modernize and miniaturize electronic taps, an area where NSA’s technological leadership had not been taken advantage of. CIA for its part saw NSA’s moves to take over the program as part of its relentless empire building: “NSA keeps picking, nibbling and lobbying away at CIA SIGINT activities,” a CIA memo complained, even though CIA’s SIGINT programs “contribute (directly or indirectly) to about 40 percent of NSA’s serialized reporting output with an Agency SIGINT budget about one-thirtieth the size of NSA’s.”28

  Inman’s predecessor, Lew Allen, had worked out a tentative agreement to improve cooperation in these “Special Collection” programs, but in practice every disagreement had to be referred to the CIA director for a decision, and Inman made it a priority upon assuming the directorship to settle the matter once and for all; with the implicit backing of the White House and the House Appropriations Committee, and based on the dictum that, in his words, “if you bring money, people will cooperate,” Inman secured NSA control of funding for the entire program, after which a definitive deal was swiftly concluded between the two agencies. The CIA-NSA “peace treaty,” as it was called, set up a new joint facility in College Park, Maryland, to oversee and process material from the Special Collection program, with the two agencies rotating the chairmanship every two years. “It led to modernization of that entire collection process,” said Inman, “and yielded very high quality return.” There were still a few details to iron out, but during the month and a half from mid-February though March 1981 when Inman was acting as both NSA director and CIA deputy director, he resolved the matter by sending memos back and forth to himself approving his solutions. (“Never had the two agencies worked so well together,” he observed.)29

  Up until the very end of the Cold War the U.S.-Soviet battle of the bugs never let up. A KGB officer under diplomatic cover at the Washington embassy who was a member of the academic Washington Operations Research Council noticed that the conference room used for the group’s meetings was in the offices of System Planning Corporation, an Arlington, Virginia, company that did contract work for the Pentagon. In September 1980 he succeeded in slipping a bug under the table after one meeting. Over the next year, until the device’s battery went dead, the KGB harvested a wealth of classified information from briefings that took place in the room, including reports on U.S. nuclear deployment plans in Europe, military modernization, and SALT II negotiating positions.30

  In February 1984, tipped off by the discovery of an extremely sophisticated bug in a piece of electronic equipment at another country’s Moscow embassy, NSA began quietly swapping out eleven tons of equipment from the U.S. embassy—typewriters, teleprinters, copiers, desktop computers, and cryptographic devices, literally anything that plugged into a wall socket. Back at Fort Meade, the agency’s director of communications security, Walter Deeley, offered a $5,000 bonus to the member of his team who could find the first bug in the returned gear. After making detailed X-rays of an IBM Selectric typewriter, a technician spotted an otherwise innocuous metal bar that ran the length of the machine. It turned out to contain an ingenious magnetic detector that measured the movement of the two arms that rotated the typeface “golf ball” at each keystroke. That was sufficient to reveal what letter had been typed; the data was captured in a small electronic memory, then periodically radioed by a miniaturized VHF burst transmitter that sent a signal too brief to be detected by the standard radio spectrum analyzers employed in periodic antibug sweeps by embassy security teams. The bugs were eventually found in sixteen typewriters, and had been in place for as long as eight years. An earlier routine inspection of the typewriters in 1978 had failed to find the well-hidden devices.31

  This was no doubt the tip of the iceberg of efforts on both sides to exploit the rich possibilities to read messages at their source, without the annoying complications of cryptanalysis, that were offered by the growing use of electronic office equipment everywhere. Although KGB defectors reported that the Soviets obtained considerable information about U.S. military aircraft and other weapons technology projects through their bugging and microwave intercept operations in the United States, particularly those targeting defense contractors on Long Island, and in California, the great irony of electronic surveillance operations as the Cold War entered its final chapter of U.S.-Soviet confrontation in the late 1970s and 1980s was that “the general effect of this intelligence was probably benign—to limit the natural predispositions of [the KGB] to conspiracy theories about American policy,” in the words of Mitrokhin and the intelligence historian Christopher Andrew. In 1979, when a Washington political flap arose over accusations by several senators that the Soviets had moved a “combat brigade” to Cuba (in fact it was a unit that had been there ever since the end of the Cuban Missile Crisis in 1962), the KGB’s “intercepts of Pentagon telephone discussions and other communications enabled the Washington residency to reassure Moscow that the United States had no plans for military intervention,” according to Andrew and Mitrokhin.32

  Back in 1945, William F. Clarke of GC&CS, contemplating the future of signals intelligence in the postwar world, somewhat idealistically proposed (“this is probably a counsel of perfection,” he admitted) that the one step that “would contribute more to a permanent peace than any other” would be the abolishment of all code and cipher communications by international agreement.33 Three decades later, having fought the code wars to a near draw, the Cold War superpowers were finding in that visionary notion a germ of truth.

  —

  At age forty-six the youngest director in NSA’s history, Inman brought an energy and determination to take charge of the agency unseen since Canine’s tenure as its first chief. Growing up in a small town in East Texas, the son of a gas station owner, Inman had graduated from high school at fifteen and from the University of Texas with a degree in history at nineteen; after joining the Naval Reserve during the Korean War, he had rocketed up through the ranks of naval intelligence despite never having attended the Naval Academy or held a seagoing command, two normally fatal failings when it came to being considered for promotion to admiral. With finely tuned political instincts he sought to raise the agency’s profile in Washington, personally working his connections on Capitol Hill and in the White House and giving the first-ever on-the-record newspaper interviews by an NSA director, aided by a dazzling command of whatever subject on which he was holding forth. “Nearly everyone who knows him mentions a piercing intellect, honesty, unusual memory for details and prodigious capacity for work,” reported the Washington Post’s Bob Woodward. Another observer of the Washington intelligence scene described the indefatigable intellectual energy he brought to the job: “If Inman had a hearing at nine o’clock in the morning, he’d be up at four prepping for it. He’d read the answers to maybe a hundred hypothetical questions. He’d essentially memorize the answers. Then he’d go before the committee and take whatever they threw at him, without referring to a note.”34

  Like Canine, the new director instituted a system of rotation for the agency’s top managers. Unlike Canine, his aim was not to pursue some abstract management theory but to break up the clique of civilian czars who had grown accustomed to controlling the orga
nization and to start edging out the World War II generation of cryptologists who still held most of the top positions. As Inman warned one of his successors, left to its own inclinations the NSA staff would always try to make sure the director never actually knew anything or changed anything: “They want to treat you like Pharaoh, to carry you around on a sedan chair and let you have the occasional lunch with a visiting foreign delegation, but to keep you away from anything else that goes on at NSA.” He established a review committee to identify the most talented leaders at the GS-14 and GS-15 levels, and promptly assigned all eighty-one of them to new jobs; two years later he shifted them again.35

  Shaking up NSA’s business-as-usual management and cultivating a new generation of civilian leaders was one thing; making progress against the still-impenetrable high-level Soviet ciphers was another. But Inman said that “the most thoughtful analysis” of the problem he read while coming in as director was a study of the state of cryptanalysis done in 1976 by William Perry, a PhD mathematician who headed the Pentagon’s R&D programs in the Carter administration and later served as secretary of defense during Bill Clinton’s presidency. Earlier outside assessments had affirmed the 1958 Baker Panel’s fundamental pessimism about the prospects of breaking any high-level Soviet systems through cryptanalytic means alone. Richard M. Bissell, a former CIA official asked to review NSA’s programs in 1965, found that while advances had been made in the diagnosis and recovery of the one Soviet cipher machine that had yielded some results owing to the discovery of bust messages, it was still never more than a small percentage of traffic that could be read at any given time. “A massive intellectual effort was required over a number of years” even to get this far, Bissell noted, adding, “Above all it must be emphasized that the sample of intercepted traffic which turns out to be decipherable is determined by the incidence of [operator] carelessness and of machine malfunctions so that the selection is one over which we have no control.” Given the continuing edge codemakers were gaining over codebreakers due to advances in technology, “timely intelligence from high-grade systems on the World War II scale would appear to be out of the question,” Bissell concluded.36