GC&CS, and by extension Baldwin’s cabinet, had long had access to all Russian traffic, including diplomatic and intelligence material. GC&CS also monitored communications between Comintern and the British Communist Party from, for example, an unmarked intercept station located in south London targeting a transmitter based in Wimbledon.70 After the Arcos fiasco, however, the Russians realised that their codes had been broken. Predictably, they replaced the system with a better, seemingly impenetrable, encryption scheme, known as the one-time pad.71 The one-time pad system, as the name implies, used a new cipher for each message, creating huge problems for the codebreaker. As a result, intelligence dried up. Over the next few years GC&CS had access to few Soviet diplomatic messages, and the only high-grade Soviet traffic available was that of Comintern.72
Sinclair, his deputy Stewart Menzies, and Alastair Denniston, the operational head of GC&CS, were all furious about Baldwin’s use of signals intelligence in Parliament. They pointed out the inestimable value of the source, and rightly predicted the detrimental consequences of parliamentary revelation. It was to no avail. Given the Soviet threat, the suspicions about the USSR’s connections to the left wing of the Labour Party, and, most importantly, the political reputations now at stake, intelligence had become temporarily expendable.73 Sinclair lamented how decrypts had been read ‘as a measure of desperation to bolster up a case vital to Government’. He bemoaned the lack of coordination within the intelligence services, and once again pressed for all three agencies to be united under one roof.74 Denniston also attacked Baldwin for having ‘found it necessary to compromise our work beyond question’.75
Baldwin should not shoulder all the blame. As we have seen, in the immediate aftermath of the First World War, intelligence chiefs had grown so alarmed about Soviet subversion that they too had advocated publishing intelligence in order to make the public aware of the scale of the threat. In fact, even Sinclair had changed his tune. Just seven years before, confronted with fears of the Bolshevik Revolution stretching its tentacles into Britain, he had argued that ‘Even if the publication of the telegrams was to result in not another message being decoded, then the present situation would fully justify it.’76 Moreover, Baldwin had tried to learn the lessons of the secret past. He consulted cabinet minutes from 1923, when the government had faced a similar issue. Amazingly, they recommended full disclosure: ‘The advantages of basing the published British case on actual extracts from the despatches which had passed between the Soviet government and its agents, outweighed the disadvantages of the possible disclosure of the secret source from which these despatches had been maintained.’77 The whole decade of the 1920s proved a learning experience for both the intelligence agencies and the politicians, who still lacked a central intelligence machine to guide them in the use and practice of this important weapon of statecraft.
The Arcos story – told and retold – eventually became the symbol of security failure for secret services around the world. New recruits to GC&CS were warned that senior politicians, including the prime minister himself, could be horrendously indiscreet when the political stakes were high. The episode strained the trust between the secret world and Number 10. It demonstrated that intelligence officers needed to be careful about how their product was used.78 The lesson would prove crucial during the Second World War regarding the use of Ultra decrypts. Menzies, the wartime head of MI6, never forgot that nobody in government understood the importance of protecting the source as much as the intelligence professionals.79 The lessons were not lost on a future prime minister, either. With the Arcos fiasco perhaps in mind, Churchill would personally insist that the circle with whom he shared his Bletchley Park material was limited to only half a dozen of his closest ministers.80 Even in the 1970s, when cabinet ministers were being indoctrinated into the arcane mysteries of ‘sigint’, they were visited by a mysterious figure from the Cabinet Office whom they called ‘the Man from UNCLE’. This official, in fact the Cabinet Office Intelligence Adviser, proceeded to recount the story of Baldwin’s blunder – by then politely referred to as a ‘mistake’ – to reinforce the importance of never breathing a word about sigint.81
By the summer of 1929, Ramsay MacDonald was back inside Number 10. This time, however, he headed a national coalition government with Stanley Baldwin, who chose to occupy the traditional residence of the chancellor of the exchequer next door at Number 11. He became increasingly influential as MacDonald aged, withered, and became an ever more marginalised figure. Although not prime minister between 1931 and 1935, Baldwin, as lord president of the council, in practice wielded as much power as, if not more than, MacDonald. Churchill referred to him as ‘the virtual prime minister’.82
Throughout the 1920s, the intelligence community had focused on Russia, communism, and what can perhaps be seen as a First Cold War, often fought out on the fringes of empire in locations as distant as Hong Kong and Shanghai. By contrast, the cabinet ignored Ireland during the second half of the decade. When MacDonald returned to power, though, attention turned briefly to the rise of the Irish leader, Éamon de Valera, who had been involved in the Easter Rising, had fought in the War of Independence, and had opposed the settlement that created the Irish Free State. Since then he had led the nationalist party Fianna Fáil into the Dáil, and went on to win the 1932 election. The government sensed trouble, and sought to bring de Valera down. In the absence of effective intelligence-gathering machinery in Ireland, MacDonald was informed by biased, outdated and alarmist intelligence which insisted on portraying de Valera as a violent IRA man rather than a democratic statesman. MacDonald bought this line, and believed de Valera was ‘undoubtedly a complete prisoner to the Irish Republican Army’. This flawed intelligence also confirmed the prejudices of Baldwin, who had increasing sway over MacDonald and was a diehard unionist with bitter contempt for Irish nationalism. Although London was jittery about overreacting to inflammatory intelligence, MacDonald and particularly Baldwin embarked on a campaign of economic sanctions to undermine de Valera. Based on an exaggerated threat of Irish subversion, the strategy was misjudged, and merely allowed de Valera to gain ‘immense political mileage’.83
For now, though, Ireland was