State Department cipher machines and communications security in the early Cold War, 1944–1965

ABSTRACT From 1944 the State Department attempted to improve its communications security by creating a Division of Cryptography and mechanising the encryption process. This article assesses the effectiveness of these reforms and shows that State’s new cipher equipment had cryptographic vulnerabilities. Moreover, the department was unable to maintain physical security at the Moscow embassy and through espionage and technical surveillance the KGB broke the ciphers and read American communications. The paper concludes by analysing the impact of this security failure, including the claim that intercepted messages influenced Stalin’s decision to approve the North Korean invasion of South Korea in 1950.

a mistake, however, for SIGFOY was badly designed and constructed. 18The machine made frequent encryption errors, was liable to break down and difficult to repair.It was loathed by its operators too -the American consul in Madras complained to the State Department that SIGFOY was 'fantastically inefficient' and had a 'physical process of coding and decoding [that]  could hardly be more awkward'. 19The army also experienced problems with SIGFOY and declared it obsolete in December 1945. 20The State Department followed suit and withdrew SIGFOY from service in May 1946. 21It was an ignominious start to mechanised encryption for the State Department.
The United States army and navy did have more reliable cipher systems which on the face of it were better alternatives for the State Department.The most important of these was SIGABA, a sophisticated rotor cipher machine. 22Most rotor cipher machines at this time had rotors which moved in a single, mechanical stepping motion like an odometer but SIGABA's five cipher rotors were electrically controlled and had unpredictable, irregular stepping which made hostile cryptanalysis far more difficult. 23During the Second World War, the United States employed SIGABA extensively for military and navy communications yet the Axis powers failed to break any of the machine's ciphers. 24But precisely because SIGABA was so technologically advanced and secure, the army and navy saw it as too valuable to risk in exposed diplomatic posts after the war.The services did not even want to issue SIGABAs to their own service attachés in locations where there was a danger of physical compromise such as in the Soviet Bloc. 25he army and navy instead developed two new rotor cipher machines, SIGROD and the CSP 1700, which were cryptographically nearly identical to each other and collectively known as the Combined Cipher Machine Mk II (CCM). 26These were five rotor machines that made use of the SIGABA chassis but did not have its sophisticated rotor maze.The CCM was consequently less secure, but any physical compromise of the machine would not disclose the new cryptographic technology to an enemy.In the late 1940s the US army and navy began distributing the CCM to their attachés and overseas military posts.The State Department took a similar approach.Parke and the Division of Cryptography acquired 225 cipher machines from the US navy and modified them to create a relatively simple five rotor cipher machine, the MCB, which was close to the CCM in its encryption procedures. 27In 1946, the State Department issued the MCB to many of its embassies and legislation, although the American embassy in Moscow only seems to have received the machine in January 1948. 28 few American diplomatic missions, such as those in London and Paris, also had access to the army's SIGTOT one-time teletype cipher machine and the State Department later developed its own version of SIGTOT, known as MOT.29 In theory, these devices were highly secure because their teletype key tapes were only used once and then destroyed.They were also exceptionally fast for they were capable of on-line encryption.Cipher machines normally operated off-line, with encipherment and decipherment separate processes to transmission, but with on-line encryption a plain text could be typed into the cipher machine and simultaneously encrypted and transmitted.At the receiving end messages were automatically decrypted and typed out by a teleprinter.However, despite having these advantages of speed and security, the use of SIGTOT and MOT was severely constrained because they were not suitable for network communications and required massive amounts of teletype tape that had to be securely stored and deposed of after use, a challenge for many embassies with a limited amount of space.Moreover, the MOT was prone to mechanical failure.In 1955, the State Department refused to issue it to the American embassy in Tehran because the device needed 'fairly constant preventative maintenance'.30 So in the late 1940s and early 1950s the MCB would be the mainstay of State Department machine encryption.
The State Department believed that with the MCB, SIGTOT and MOT it had eliminated the cryptographic weaknesses of the past and secured American diplomatic communications.In a Washington Post article in 1948 the department revealed that the encryption of messages to its embassies 'is now handled by machines rather than by hand, and machine codes are subject to quick changes and to complexity of uses which make it extremely difficult if not impossible for an important message in code to be read by foreign agents'. 31During Senate hearings in 1951 a State Department representative recognised that there had been considerable criticism of the department's codes in the Second World War but reassured senators that '[w]e have taken a lot of interest in this and we feel that we have a security code, as secure as there is in the world'. 32

Weaknesses in the cipher machines
Unfortunately, this confidence was misplaced for there remained significant problems with State Department communications security.The new cipher machines had inherent weaknesses and control of the American embassy in Moscow was very still poor, with the State Department unable to prevent Soviet espionage and technical surveillance.The flaws in the cipher machines became first apparent in 1951.The CIA was also an operator of SIGTOT and the agency discovered that the machine produced compromising electro-magnetic emissions when encrypting which could reveal the contents of messages. 33Alerted to the problem, the State Department in March 1951 had to warn its diplomatic missions in London, Paris, Moscow, Frankfurt and Vienna that the plain text of SIGTOT messages could inadvertently be conducted down the signal line with the encrypted text. 34he State Department instructed the posts to stop using SIGTOT wherever possible and cease operating one-time tape cipher machines in on-line mode, which considerably reduced their speed.
The United States then found that the CCM and MCB cipher machines were also less secure than originally thought.In December 1951, an investigation by American and British experts uncovered vulnerabilities in the CCM (which in NATO use had been given the code name AJAX). 35Details of this discovery are still classified but two contemporary documents of the American Armed Forces Security Agency (AFSA), which at that time was responsible for service communications security, have the titles 'Possible Cryptanalytic Compromise of the Combined Cipher Machine' and 'Brief History of the AJAX Crisis of Dec 51'. 36The British signals intelligence agency, GCHQ, advised its Canadian counterpart that the '[n]ature of insecurity of CCM is newly discovered exhaustion attack . . .US authorities agree with us on insecurity'. 37As an immediate response, the Americans had to modify the CCM and change its operating procedures, which made the device more time consuming and unwieldy to use. 38Even with these emergency countermeasures, the AFSA felt that it was urgent to replace the CCM.This all had implications for the State Department's MCB machine.The Director of the AFSA explained in June 1952 that: The State Dept. is now using a cipher machine which is essentially the cryptographic equivalent of AJAX . . .All attacks which work on AJAX . . .work on the CSP 2200 [MCB].It is considered imperative therefore that the State Dept.proceed at once to improve their cryptographic position. 39 1951-52 then, there was something of a crisis in American cryptography with the MCB, CCM and SIGTOT all found to have vulnerabilities.The State Department responded by bringing into service a successor to the MCB, the MEC.Designed by Parke and the Division of Cryptography, the MEC was a rotor cipher machine with electrical stepping for its rotors, like SIGABA, but with a different rotor control maze. 40According to an NSA history, Parke had sought to 'adopt as much of the SIGABA logic as was needed to achieve sound security' while at the same time 'avoid risking the SIGABA details in hazardous locations in which the Department of State had to use its cryptographic machines'. 41ompromise of the new MEC machine would therefore only compromise the principle of electrical stepping, not the embodiment of it found in SIGABA.
Through the 1950s the State Department issued the MEC to its diplomatic posts.It was installed in the American embassy in London in January 1953 and then over the next two years rolled out to missions in Western Europe, the Middle East, the Far East and Latin America. 42It took longer to equip the more exposed posts behind the Iron Curtain and Moscow, Bucharest, Budapest, Prague and Warsaw only received the MEC in February 1959, replacing their MCB machines. 43The MCB was discontinued worldwide in July 1959. 44In purely cryptographic terms, the MEC was more secure than the MCB but it was difficult to use.Rather than build wholly new code machines, the State Department appears to have created MECs by modifying existing cipher equipment and the machinery was starting to show its age.A retired State Department cipher clerk later characterised the MEC as 'cumbersome, noisy, slow and prone to failure if not properly maintained'. 45

Embassy attacks, espionage and bugging
Alongside these cryptographic challenges, the State Department had to protect the physical security of its cipher machines against a variety of overt and covert attacks.When the North Korean army invaded South Korea in June 1950 the American embassy in Seoul was quickly overrun but the marine guards were able to destroy the embassy's code machines with thermite just before evacuating. 46In Taiwan, American security was less successful. 47After a court martial acquitted an American soldier in May 1957 of killing a Taiwanese civilian, an angry mob stormed the US embassy in Taipei and managed to break into the code room.The room was protected by a thick steel door, but its walls were only made of chicken wire, studs and plaster and the rioters smashed through using sledgehammers.They ransacked the room and the cryptographic safe and while the embassy's two MCB cipher machines were not damaged, 100 rotors were stolen.Many of these were later found strewn over the embassy grounds by the departing protestors.The State Department assumed that the cipher systems had been compromised and immediately ordered all Far Eastern posts to temporarily switch to manual one-time code pads for encryption.The next day couriers with replacement cryptographic material (most likely new rotors and keys) were sent out from Washington.
The Taipei embassy attack was dramatic and disruptive but the most serious breaches of physical security in this period came from Soviet espionage.In the late 1940s and early 1950s the KGB was able to suborn in succession three US army cipher staff stationed at the American embassy in Moscow and gain intelligence on the crypto systems. 48Its first informant was Sergeant James McMillin, a military cipher clerk who defected to the Soviet Union in May 1948 in order to marry his Russian girlfriend, Galina Biconish. 49Although this appears to have been a genuine relationship, the KGB had encouraged Biconish to target McMillin and facilitated his defection. 50In return, McMillin provided the KGB with documents and explained how the embassy cipher machines worked, describing the installation of the rotors and their position changes according to the key list. 51McMillin also told the Soviets about the personalities and the daily routines of the other cipher clerks, which could have helped future KGB recruitment.
The information from McMillin was valuable, but according to former KGB officer Sergey Kondrashev, who was involved in the entrapment operation, it was insufficient to break the American ciphers. 52Moreover, the US authorities moved quickly to change all their codes after McMillin's defection.They assumed that he had compromised the CCM system in Moscow which was used by most American military attachés and major army installations in Europe and the Middle East. 53The State Department was also affected, for since March 1948 State and the army had operated a combined cryptocentre in the embassy, sharing personnel and facilities.Replacement rotors and keys were immediately sent out.The Washington Post reported that the State Department had been forced to spend $80,000 on changing its cipher machines all around the world as a result of McMillin's defection. 54owever, although the State Department and US armed forces took swift cryptographic countermeasures, human security in Moscow remained a weak point and a few months later, in early 1949, the KGB was able to recruit another military cipher clerk in the American embassy. 55This agent, who was codenamed 'Jack' by the KGB and never identified by American counter-intelligence, was also involved in a relationship with a Russian woman.But unlike McMillin, Jack was willing to stay in place and sell the Soviets information about the cipher systems.He had a series of secret meetings with a KGB signals intelligence expert who asked detailed technical questions and gave him tasks to perform on the cipher machine.Jack also provided a broken rotor and a key schedule.By the time he returned to the United States in late 1949 the KGB had paid him $100,000 for his work, a considerable sum by contemporary standards.
Kondrashev claimed that Jack's assistance enabled the KGB to build a copy of the cipher machine used by the military attaché (most probably the SIGROD version of the CCM) and read the traffic from the embassy for as long as the keys supplied by Jack remained valid. 56hrough this break into the machine, the Soviets were also able to decrypt military traffic between Washington and other posts abroad.Some supporting evidence for Kondrashev's claims comes from the testimony of Stig Wennerstrom, the Swedish air attaché in Moscow between 1949 and 1951.Wennerstrom secretly worked as a spy for Soviet military intelligence, the GRU, but on occasion, he also passed on information to the American air attaché. 57The GRU discovered his double dealing when it decrypted a telegram from the American embassy in Moscow that mentioned him by name.Wennerstrom later testified to a Senate inquiry that: In a radio report, they [the embassy] used my name as well as a cipher that the Soviets were able to break . . . .Radio messages from the radio station in the American embassy were monitored and the ciphered material was assembled and efforts made to decode it, which was possible at times . . .I know that on certain occasions also the Soviets had been able to obtain the key to codes from the American Embassy. 58ndrashev did not disclose whether the intelligence from Jack also enabled the KGB to read the ciphers of the State Department's MCB machine although it was similar to CCM and had the same cryptographic weaknesses.The breakthrough against the MCB possibly came later, in 1953, when the KGB recruited yet another source in the embassy, US army Sergeant Dayle Smith. 59Smith was a cipher machine mechanic assigned to the military attaché, but he also had access to the State Department crypto equipment.In exchange for money, he gave the Soviets sensitive technical information.According to Yuriy Nosenko, a KGB defector to the United States in the early 1960s, Smith described to the Soviets the operation of the State Department cipher machines and supplied the daily settings.Nosenko emphasised to his CIA debriefers the importance of Smith, saying that 'Thanks to his help they [the KGB] were able to read your State Department codes'. 60It does seem plausible that by the mid-1950s, the KGB could read at least some State Department traffic for intercepted American embassy telegrams were distributed to the Soviet leadership during the 1956 Suez Crisis. 61et there was another factor at play: Soviet technical surveillance of the embassy.The American embassy in Moscow was poorly protected against bugging and the KGB was highly proficient at planting covert surveillance devices.In May 1953, the United States moved the embassy chancery to another building across town and set up a new joint State Department-service attaché code room.But when a Regional Security Officer inspected the code room in October 1954 he found that it was not sound proofed and had French doors opening out onto a balcony above a street. 62The doors were sometimes propped open in the summer to provide ventilation.The appalled security officer warned in his report that the Moscow embassy code room 'cannot be equalled in insecurity by any Communications Center in a United States installation in Europe'. 63et the problem of lax security was actually even worse than he realised.Before the Americans took occupancy of their new chancery Soviet workmen had remodelled the building and State Department security failed to properly supervise the works. 64The Soviets covered the building with tarpaulins, blocking view of the renovations from outside, and insisted on prior appointment for any site inspections.When Americans did come on site they were only allowed to do visual inspections and were always accompanied by the Soviet architect.All visits by US personnel were prohibited for two to three weeks in early March 1953.While American visitors were kept away, the KGB embedded over 50 microphones in the chancery building, concealing them in walls behind radiators in all the key rooms, including the embassy code room.Later in the 1950s, the Americans established separate State Department and service attaché code rooms but both of these rooms already had microphones in them. 65he microphones allowed the KGB to do more than just eavesdrop on the conversations of cipher clerks, for as the Americans slowly discovered in the 1950s, the sounds produced by cipher machines could expose the content of messages and reveal the cryptographic system.A later NSA report on the Moscow embassy bugging explained that: . . .teleprinters and cryptographic equipments do not run silently; they emit acoustical and electromagnetic energy . . .The sounds emitted by the machinery involve minute differences in amplitude, frequency, and timing as different letters and characters are processed by the equipment.Relatively insensitive pick-up devices can detect these differences if they are placed at fairly close range to the equipment as was the case in each of the Moscow code rooms.When these sounds are recorded and analyzed the plain text of messages can be reconstructed. 66e NSA believed that by exploiting the cipher machines' acoustic emissions '[i]t was technically feasible for the Soviets to have recovered the plain text of the messages encrypted and decrypted by machines in these coderooms'. 67Furthermore: . . . the sounds made by these equipments reflect the internal mechanical workings of the cipher machine; these sounds can be recorded, analyzed, and the crypto system itself could have been reconstructed.If this did occur, not only the messages processed in Moscow but also those messages processed by other posts using the same machines and the same cryptomaterial could have been lost' 68 In 1953 then, poor State Department security would have enabled the Soviets to mount a two-sided attack on the MCB, obtaining information from Dayle Smith while monitoring and analysing the machines' compromising acoustic emissions.
After the MEC replaced the MCB at the American embassy in Moscow in February 1959 the Soviet technical surveillance team had to master the new cipher machine.Nikolai Andreev, a KGB officer involved in the surveillance operation, later recalled that the 'main difficulty was to find the "weaknesses" of the electromechanical encoder at the US Embassy in Moscow; to determine which parts of their [the Americans] machine generate spurious emissions'. 69This seems to have been done fairly quickly, for by 1959 the KGB could again read encrypted messages from the embassy. 70Andreev and his team received the Lenin Prize for this feat. 71he American authorities were unaware of Andreev's breakthrough, but by the early 1960s, there was growing alarm in Washington about State Department communications security.In April 1960, the NSA warned Parke that the department's cipher machines did not meet acceptable security standards and should be replaced at the earliest possible date. 72It is not clear what the NSA's concerns were but the State Department itself was exercised about the possible dangers posed by bugging and technical surveillance.John Hanes, the head of the department's Bureau of Security and Consular Affairs, lamented in a memorandum in 1960 that because of major advances in surveillance technology and computers 'we are under constant apprehension that the enemy . . .may be overhearing our conversations and reading our encrypted messages'. 73On 28 September 1960, the State Department suddenly ordered its embassies in Moscow, Warsaw, Prague, Bucharest and Sofia to encrypt with manual one-time pads all future telegrams classified as top secret and secret. 74This indicated a considerable loss of faith in the MEC which had only been installed in the East European embassies the previous year.The State Department might have been reacting to comments made by the Soviet leader Nikita Khrushchev at a meeting with the Iranian chargé d'affaires in August, which suggested that Khrushchev knew the contents of cables sent to the American embassy in Moscow. 75everting back to manual encryption in the Soviet Bloc embassies could only be a temporary expedient and the State Department hurriedly sought other ways to mitigate the threat from technical surveillance.One approach was to install shielding around cipher machines. 76In October 1960, the service attaché cipher machine in the Moscow embassy was placed in a special sound proofed box. 77oing a step further, an acoustically shielded room was installed in the embassy in December 1962 to house the State Department code machines. 78This protection would have prevented Andreev's KGB surveillance team from detecting and reading any more compromising emissions from the embassy's cipher machines.But shielded communications rooms were expensive and difficult to install and by 1964 they had only been put in eight American diplomatic posts. 79he other approach taken by the State Department was to ask the NSA to create a new cipher machine that could replace the MEC and be as free as possible from compromising emissions.Originally, the State Department had wanted the MEC's replacement, the KW-1, to be available by the start of 1963 but in November 1960, it advised the NSA that it needed the machine at least a year earlier and urged that 'the greatest priority possible . . .be accorded to the development of replacement equipment capable of resisting all known penetration methods'. 80Unfortunately, this new completion date could not be met and the KW-1 was only ready for production in autumn 1962. 81he State Department then encountered resistance from an economy minded Congress which wanted to slash a $4.86 million budget request for 400 KW-1s and associated teletypewriters down to just $1.5 million. 82In early October 1962 department officials had to plead their case before the Senate Appropriations Committee and publicly reveal the deterioration in State's communications security.They admitted that because of threats posed by recent technological advances '[i]t has become highly important . . .that outmoded communication equipment at overseas posts be replaced at the earliest possible date.The equipment now in use is many years old and is subject to possible compromise by foreign governments'. 83The State Department also justified the expenditure on operational grounds, arguing that its current cipher machines were 'worn out and obsolete' and 'subject to frequent breakdowns and time-consuming repairs'. 84e CIA takes over Three weeks later, the Cuban Missile Crisis transformed the situation.It is well known that during the crisis President John Kennedy and Khrushchev found it difficult to quickly communicate with each other and that this experience inspired the creation in 1963 of the Hotline, a dedicated teleprinter link between Moscow and Washington.But the crisis also hammered home to Kennedy the need to generally modernise and improve the State Department's telecommunications system, for State proved unable to cope with the high volume of urgent traffic between Washington and embassies in October 1962.For example, a crucial 12 part message from Kennedy to other heads of government was sent through State Department channels but failed to reach many posts or arrived missing three parts, resulting in several American allies being caught unawares when Kennedy gave his televised speech on 22 October announcing a blockade of Cuba. 85The fault mostly lay with the State Department's antiquated and fragmented telecommunications network, but time-consuming encryption and decryption operations aggravated delays.On 24 October, the Executive Committee of the National Security Council broke off its discussion of the on-going crisis to consider the communications issue. 86Kennedy was briefed that: . . .one problem is that all of the Latin American enciphered telegraph facilities are made up of what we call offline facilities, so that it takes a considerable amount of time to get an enciphered message from the transmitter to the receiver, and they can't be use for two-way telegraph conferences. 87e president was exasperated and he set up a National Security Council sub-committee headed by William Orrick, the Under Secretary of State for Administration in the State Department, with orders to solve the immediate communications problem in Latin America and to draw up plans for an integrated national communications system that could bring together the networks of the State Department, CIA and military. 88ortunately, there was on hand an advanced, on-line cipher machine which could be quickly issued to the embassies in Latin America.This was the KW-26, a completely electronic cipher machine which the NSA had developed and brought into service for the American armed forces in 1957. 89The KW-26 was a generation ahead of the State Department's MEC and represented a major technological leap forward.It encrypted messages with an internal electronic key generator and since it had no rotors or other mechanical moving parts, it produced fewer compromising emissions.For reasons that are unclear, the State Department had not adopted the KW-26, but the CIA had used the machine in its overseas stations since 1958. 90The agency had spare machines available and it was directed to deploy 89 KW-26s to Latin American and other diplomatic posts by the end of November 1962. 91he KW-26 also featured prominently in the Orrick Committee's deliberations on how to set up a national communications system.The aim was to create joint State Department-CIA code rooms in embassies with common use of on-line KW-26 machines.But this posed a problem for the CIA as the KW-26 automatically decrypted and typed out incoming messages and anyone in the code room would be able to see the clear text, whether they were State or CIA. 92The agency argued that it had to maintain the privacy of its communications because otherwise they could reveal secret intelligence sources and clandestine techniques.This meant that CIA personnel would have to staff the consolidated communications centres.Although Orrick was sceptical at first the CIA's stance was 'not negotiable' and it was incorporated into the committee's planning. 93Where possible, embassy code rooms would be co-located in a single communications centre with the CIA operating the cryptographic equipment and teletype terminals. 94These consolidated centres would be equipped with KW-26 on-line cipher machines.Embassies would also have the new KW-1 cipher machine so the State Department could protect its privacy and encrypt any highly sensitive telegrams before they were passed on to CIA for on-line encryption and transmission.A few embassies and most consular establishments would have just the KW-1, presumably because of their low volume of traffic.
This plan was largely implemented in 1963.The CIA transferred one of its leading communications specialists, John Coffey, to the State Department for two years to be Deputy Assistant Secretary of State for Communications and carry out the reforms. 95In March 1963, a new State Department Office of Communications was formed with a Division of Communications Security. 96State's cryptographic staff were reassigned to the new organisation, although Lee Parke chose to resign in February 1964. 97A Diplomatic Telecommunications Service was also established with combined State-CIA communication centres in many American embassies under joint State-CIA control. 98The necessary funding was found to equip these centres with more KW-26s and provide fast, secure, on-line encryption.The KW-1, however, turned out to be a dud despite its years of development.The machine performed badly in production testing in 1963, corrupting many letters in messages. 99The State Department had to abandon the device and instead purchase a basic one-time tape cipher machine, the HW-28, to give diplomatic posts an off-line encryption capability. 100The HW-28s were issued to posts in 1965 and the State Department's vulnerable MEC cipher machines were finally withdrawn from service across the network. 101he Cuban Missile Crisis thus brought about major and long overdue improvements in American diplomatic cryptography.It also led to much greater CIA involvement in State Department communications.A retired State Department official, Nuel Pazdral, later recalled that after the missile crisis: . . .new encrypting gear was budgeted.The Central Intelligence Agency took over a large part of this because they had these machines.It was called a KW26 . . .It was pretty much space age gear for its day and the Central Intelligence Agency was the only agency that had many of these machines.So they then moved into our communication picture in the State Department for the first time.And ever since then we have had the system which we had then, which is basically that we sort of massage the data and then pass it along to Agency folks who send it onward. 102ctor Marchetti, a former CIA officer, described a similar relationship in 1974, writing that: . . . it should be understood that CIA communication clerks handle nearly all classified cables between American embassies and Washington -for both the CIA and the State Department.To have a separate code room for each agency in every embassy would be a wasteful procedure, so a senior CIA communications expert is regularly assigned to the administrative part of the State Department in order to oversee the CIA's communicators who work under State cover. 103 effect, between 1962 and 1965 the State Department had to cede much of its communications security role to the CIA.

Conclusion
Reviewing the period between 1944 and 1965 it is hard to escape the conclusion that the State Department failed to adequately protect its communications security.The Division of Cryptography did mechanise the encryption process by introducing the MCB, MOT and MEC cipher machines and although SIGFOY and the KW-1 were fiascos, State Department cryptography was undoubtedly far stronger in the early Cold War than it had been in 1939.The level of encryption and communications security was probably good enough to protect American diplomatic cables against attack from many foreign intelligence agencies.Crucially though, it was insufficient defence against the main enemy, the Soviet Union, and the KGB was able to read State Department and military attaché traffic at critical moments in the early Cold War.Despite all the high hopes around the creation of the Division of Cryptography and mechanisation, by the early 1960s the State Department was unable to provide fast, secure encryption and it had to be rescued by the CIA, just as it had been rescued by the American army and navy in 1944.
Two factors stand out in this failure.Firstly, the State Department did not keep up with the rapid advances in cipher machine technology from the mid-1950s.The Division of Cryptography may have had insufficient personnel and resources to do this; it only had 31 Cryptography Staff in 1961 and appears to have struggled to get funding from Congress. 104Yet even so, it is hard to understand why State did not adopt the KW-26 in the 1950s, like the CIA.Secondly, the State Department did not maintain physical security in the American embassy in Moscow in the late 1940s and early 1950s, and this gave opportunities for the KGB to install an extensive technical surveillance system and suborn cipher personnel.In this aspect, the State Department performed worse than the foreign services of some of America's allies.The British embassy in Moscow was also bugged by the Soviets but the Foreign Office managed to prevent its cipher staff from being recruited by the KGB. 105Generally, there does not seem to have been sustained attention and commitment to all aspects of communication security in the State Department, in spite of the best efforts of Parke and the Division of Cryptography.
What remains to be determined is whether State's flawed communications security gave the Soviets an advantage in the Cold War.Russian sources are largely quiet about the duration and impact of the KGB's break into the Moscow embassy's cables but we do have contemporary American damage assessments.Acting on information from the KGB defector Nosenko, the State Department discovered the bugging system in the embassy in 1964 and the President's Foreign Intelligence Advisory Board (PFIAB) commissioned the United States Intelligence Board (USIB) to lead a multi-agency investigation. 106The NSA's technical assessment was that the microphones could have enabled the Soviets to read all of the State Department' telegrams with Moscow from 1953 until October 1960, when the embassy switched to one-time pad manual encryption for top secret and secret messages. 107Once the shielded room was installed in the embassy in December 1962 all State Department traffic should have been protected.State Department cables with other East European posts were also likely 'to have been compromised during 1952-1959 through cryptomachine information derived at Moscow'. 108The military attachés were similarly affected.Attaché traffic processed in Moscow might have been read up to October 1960, when the cipher machine was placed in a sound proofed box.Since the microphones could have disclosed the workings of the CCM's cipher system, between 1952 and 1954 'all traffic in the entire attache net which included Moscow could have been compromised'. 109hese were severe, prolonged breaches of American communications security yet in its damage assessment the State Department downplayed the likely diplomatic impact. 110The department reviewed telegraph traffic with the Moscow embassy during several Cold War crises and periods of superpower negotiation and concluded that while the Soviets may have gained some small tactical advantages from reading these telegrams, in most cases there was no evidence that damage had been done to American foreign policy interests.Compromised messages during the Korean War and the 1958-59 Berlin Crisis might even have benefitted the United States by revealing American resolve.Some members of the PFIAB criticised this assessment seeing it as the State Department 'seeking to sweep unpleasant matters under the rug' and the final USIB report seems to have tempered State's optimism. 111The board noted that while there was no indication of Soviet action detrimental to the United States based on the compromised information, 'it must be concluded that due to the extensive period of penetration . . . the cumulative effect had resulted in serious damage to the U.S'. 112 It does appear that the intercepted State Department communications were a prized source of intelligence for the Soviet leadership.Nosenko told the CIA that the KGB regularly sent Khrushchev and the Central Committee reports on its intercepts of American embassy telegrams. 113According to Nosenko, when the source was lost, the Central Committee exerted considerable pressure on the KGB 'for continuation of the information formerly obtained in this manner' and this led the KGB Chairman to demand 'maximum effort from all KGB officers concerned with the recruitment of [American] code clerks'. 114Moscow embassy traffic published in the Foreign Relations of the United States document series contains material that would surely have been of interest to Khrushchev, such as two State Department telegrams to Moscow in May 1956 which reported that the CIA had covertly obtained a copy of his 'Secret Speech' and was considering the best way to leak it to a global audience. 115Further research into State Department and military attaché correspondence is needed to fully understand the scale and type of intelligence the Soviets may have collected between 1953 and 1962 from the bugging system.
It is also possible that an earlier breach of Moscow embassy security, the recruitment of the cipher clerk Jack by the KGB in 1949, had a major impact on superpower relations and the Cold War.The information and cipher keys Jack provided enabled the Soviets to read American military communications between Washington and other posts apart from Moscow. 116After the end of the Cold War, Kondrashev, Jack's KGB handler, claimed that some of this traffic had helped convince Soviet leader Josef Stalin to approve North Korea's invasion of South Korea in 1950. 117Viktor Abakumov, the Soviet Minister for State Security, would personally give Stalin transcripts of the deciphered American military messages and Abakumov confided to Kondrashev that these intercepts had been 'especially persuasive for Stalin'. 118They gave the Soviet leader the impression there was little danger of direct conflict with the United States if the Soviets supported the North Korean venture.According to Abakumov, that 'was why Stalin finally removed his objection to the North Koreans' long standing plans to invade'. 119his was a big claim to make and it could just have been boastful hyperbole from Kondrashev.However, in 2020 the historian Sergey Radchenko reported the discovery of a Chinese document that gave more credence to Kondrashev's story. 120The document was a record of talks in 1956 between the Chinese leader Mao Zedong and Anastas Mikoyan, First Deputy Chairman of the Soviet Council of Ministers.At the meeting Mao asked Mikoyan why Stalin had decided to give the go head to Kim Il Sung's invasion.Mikoyan replied that it was because Soviet intelligence had intercepted telegrams from General Douglas MacArthur in Japan to the United States which indicated that the Americans would not get involved if a war broke out in Korea.There does seem here a possible connection with the KGB operation in Moscow, but it is not clear whether MacArthur's messages would have been encrypted by the same type of cipher machine, the SIGROD version of the CCM, which Jack appears to have compromised.MacArthur's Far East Command had started using SIGROD in January 1948, but it also operated SIGABA in its major headquarters and it is likely that any cables from MacArthur would have been encrypted by the more secure SIGABA, which as far as we know the Soviets never broke. 121ore research is needed to elucidate this topic and for now any connection can only be speculative.Nonetheless, the Mikoyan document does raise the intriguing possibility that the Korean War, one of the bloodiest wars of the Cold War, was partly caused by poor State Department embassy security.