The original US nuclear-war plans were prepared at a time when bombers predominated, atomic bombs were in very short supply, and there was a lack of co-ordination of plans. The 1947 plan concentrated on destroying the war-making capability of the Soviet Union through attacks on government, political and administrative centres, urban–industrial areas and fuel-supply facilities. It was to have been achieved by dropping 133 atomic bombs on seventy Soviet cities in thirty days, including eight on Moscow and seven on Leningrad. This plan was ambitious, to say the least, primarily because the USA possessed a total of only thirteen atomic bombs in 1947 and fifty in 1948.
In May 1949 a new plan, named Trojan, required 150 atomic bombs, with a first-phase attack against thirty cities during a period of fourteen days. Then, also in 1949, came another plan, named Dropshot, in which a war in 1957 would be centred on a thirty-day programme in which 300 atomic bombs and 20,000 tonnes of conventional bombs would be dropped on about 200 targets, with the atomic bombs being used against a mix of military, industrial and civil targets, including Moscow and Leningrad.
All these plans included substantial quantities of conventional bombs and were essentially continuations of Second World War strategies. Interestingly, even as early as 1949, planners were earmarking command centres for exemption from early strikes (‘withholds’ in nuclear parlance), so that the Soviet leadership could continue to exercise control.
Up to about 1956 the US authorities based their plans on intelligence which was far from complete and, as a result, they deliberately over-estimated their opponent’s strengths. In 1956, however, the Lockheed U-2 spy plane started to overfly the USSR, and this, coupled with more effective ELINT and SIGINT, meant that a more accurate picture was obtained.
By the early 1950s the production rate of atomic bombs had increased to the point where US field commanders – both air-force and navy – began to make plans for their use, each of them planning to use them in support of his battle. There was nothing in this situation to prevent several commanders from selecting the same target; indeed, a US Senate committee was told that in the Far East 155 targets had been listed by two commanders and 44 by three, while in Europe 121 airfields had been targeted by two commanders and 31 by three.
A first attempt at some form of co-ordination was made in a series of conferences held in the early 1950s, which achieved partial success. The situation came to a head, however, when the navy’s Polaris SLBMs and the air force’s Atlas ICBMs became operational in the early 1960s, and a Joint Strategic Target Planning Staff (JSTPS) was established, with an air-force lieutenant-general at its head but with a naval officer as his deputy. The first outcome of the JSTPS’ work was the Comprehensive Strategic Target List (CSTL), which identified 2,021 nuclear targets in the USSR, China and their satellites, including 121 ICBM sites, 140 air-defence bases, 200 bomber bases, 218 military and political command centres, 124 other military targets, and 131 urban centres.5 This CSTL duly became an integral part of the first Single Integrated Operational Plan (SIOP), which was produced in December 1960.
There were no alternatives in this SIOP-60, which consisted of one massive nuclear attack, and it was one of the earliest targets for reform when President Kennedy took power in January 1961. US planners were now, however, aided by virtually complete satellite coverage of the USSR, giving them information on the potential enemy never previously available (and which, among other things, made it clear that the supposed ‘missile gap’ did not exist). This resulted in SIOP-63, which consisted of five categories of counter-force option: Soviet missile sites; bomber and submarine bases; other military targets; command-and-control centres; and urban–industrial targets. This received serious criticism from three separate quarters. First, from within the USA, because of what appeared to be a first-strike strategy; second, from the USSR, which denied the possibility of controlled counter-force warfare; and, third, from NATO allies, who were very alarmed by the total absence of any urban targets (the so-called ‘no cities’ strategy).
President Kennedy’s secretary of state for defense, Robert McNamara, then developed the concept of Assured Destruction, which he described in public first as ‘one-quarter to one-third of [the Soviet Union’s] population and about two-thirds of its industrial capacity’ and later as ‘one-fifth to one-fourth of its population and one-half to two-thirds of its industrial capacity’. Whatever was said in public, however, within the US armed forces SIOP-63 was not withdrawn, and thus there appears to have been a marked divergence between the public and the internal rhetoric.
Proponents of this policy of Mutually Assured Destruction (MAD) argued that the one way to make nuclear war impossible was to make it clear that any nuclear attack would be answered by a total attack on an enemy’s population, together with its industrial and agricultural base. While such a concept might have been valid in the early days of the nuclear confrontation, it rapidly lost its credibility when it became clear that, even after such a strike, the USSR would still have sufficient weapons to make a response-in-kind on US cities.
The strategy of ‘flexible response,’ introduced in 1967, required facing an opponent with a credible reaction which would to inflict losses out-weighing any potential gain. The deterrent power of such a strategy depended on the capacity of the proposed response to inflict unacceptable losses on the opponent, while its credibility depended upon its ability to minimize the risks of higher-order losses on the responder’s own country in subsequent rounds. This posed something of a dilemma, in that the deterrent power of the response was enhanced by escalation to a higher level, while credibility tended in the other direction, since a lower-level response carried no inherent escalatory risks. The plans implement this strategy were promulgated in a revised version of the SIOP which became effective on 1 January 1976.
It became customary for all incoming presidents to initiate a review of the strategic nuclear-war plans, and that carried out by President Jimmy Carter in 1977–9, was expected to result in major changes. In the event, however, it led only to a refinement of the previous plan, together with the introduction of rather more political sophistication. Thus, for example, targets were selected in the Far East, not so much for their immediate relevance to the superpower conflict, but because their destruction would make the USSR more vulnerable to attack by the People’s Republic of China.
A further review was conducted when the Reagan administration came to power in 1981. This resulted in a new version of the SIOP, which included some 40,000 potential targets, divided into Soviet nuclear forces; conventional military forces; military and political leadership command posts and communications systems; and economic and industrial targets, both war-supporting and those which would contribute to post-war economic recovery. The plan allocated these targets to a number of discrete packages, of differing size and characteristics, to provide the National Command Authority (the president and his immediate advisers) with an almost limitless range of options.
The new plan also included particular categories of target for other plans, for possible implementation on receipt of an unequivocal warning of a Soviet attack. These included a pre-emptive strike, launch-on-warning and launch-under-attack. The plan also included a number of ‘withholds’, but stipulated that a reserve of weapons must be retained for possible use against those ‘withholds’ if the developing scenario so dictated.
The real calculations of strategic nuclear war – known as ‘dynamic’ assessments – were extremely detailed and were far more complex than the static measurements. One such ‘dynamic’ calculation in the period following the Soviets’ fielding of the SS-18 resulted in an assessment that, under certain conditions, a Soviet counter-force first strike would appear to be a possibility. In this assessment it was calculated that the USSR, which normally had only about 10 per cent of its SSBN force at sea, would gradually, and covertly, increase that number, and, if the US command decided to ride out the attack, the Soviets would then destroy approximately 45 per cent of the US strategic forces. As a result, the ensuing US counter-military retaliatory strike on the Soviets (who would be on full alert) would leave the Soviets with 75 per cent of what had been left after their first strike. This meant that the USSR would retain not only a reserve capable of carrying out either an urban–industrial strike on US cities or an attack on US ‘other military targets’, but also a reserve for use against another opponent (the so-called ‘nth-country reserve’) – an outcome which would have been distinctly favourable to the Soviets. If the USA managed to launch all its ICBMs under attack, however, the damage ratio more or less reversed: 40 per cent damage to remaining Soviet forces versus 25 per cent damage to US strategic forces.
Thus, argued the US planners, a credible US launch-on-warning/launch-under-attack capability was a mandatory element of an effective deterrent. These results posed a problem encountered in numerous US war games: that neither side could enhance stability by pursuing its own best interests of a secure deterrent potential, but, conversely, neither side could unilaterally lower its deterrent. For the US to do the latter gambled on a US judgement of how the Soviets treated ‘uncertainty’ and what their perceptions of relative advantage might have been.
This whole area highlighted the decision to launch as one of the major problems associated with missiles. Launching bombers for possible nuclear missions was relatively easy, since crews were under firm instructions that they had to receive a positive (and encoded) order from the ground to continue before reaching specified waypoints, otherwise a return to base was mandatory. Missiles, on the other hand, received only one order – to take off; there was then no turning back. Thus the decision to launch the missile was much harder to make.