Long-Classified U.S. Estimates of Nuclear War Casualties During the Cold War Regularly Underestimated Deaths and Destruction

This report by an ad hoc panel of the JCS, headed by Air Force Lieutenant General Hubert R. Harmon, was the first systematic effort to estimate the destruction to the Soviet Union that a U.S. atomic attack would cause and the extent to which the destruction would be decisive to the outcome of a U.S.-Soviet war. The report drew on the strategic war plan that was in effect in early 1949, TROJAN, which targeted 75 Soviet cities with 133 atomic bombs, in the range of 20 kilotons. The planners assumed that atomic weapons would be “particularly effective against built-up populated areas, such as cities.” Urban targets were important as sites for major industrial plants, including war-supporting industry. While destruction of industrial and military capability was the main objective, widespread population losses were taken for granted.

The planners sought “Class A” damage to their urban targets, which was that level of destruction that would render a building or installation “unsuitable for its intended use until major reconstruction is accomplished.” The area where the atomic offensive would produce Class A damage was assumed to be the area which would experience an overpressure of 5 pounds per square inch (PSI), “the pressure in the blast wave in excess of atmospheric pressure.”

The report projected as many as 2.7 million Soviet fatalities and 4 million additional injuries. The estimates were based on an assumed population density of 10,000 persons per square mile. For each bomb detonated, some 20,000 persons would be killed, whether from collapsed buildings, flash burns, or radiation sickness, like the Hiroshima-Nagasaki victims. The 20,000 figure was “reasonably conservative in light of the experience at Hiroshima and the computed population densities of average Soviet cities.” Each bomb would also produce another 30,000 casualties from burns, wounds, and other causes.

The destruction produced would “vastly complicate the problems for living for the remainder of the 28,000,000 in the 77 target cities” (depending on the effectiveness of civil defense). The survivors would simultaneously have to provide first-aid, extinguish fires, restore utilities, clear rubble, and find housing.

Despite the estimated population losses and industrial damage – a 30 to 40 percent reduction in industrial capacity – the panel concluded that, even if the U.S. conducted the attack perfectly, it would not in itself "bring about capitulation, destroy the roots of Communism, or critically weaken the power of Soviet leadership to dominate the people." Moreover, Soviet military capability to take control of "selected areas" of Western Europe, the Middle East, and Far East "would not be seriously impaired," although fuel and lubricant shortages would subsequently decrease troop mobility.[17]

Even if the atomic air offensive would leave Soviet power intact, the Harmon committee concluded that “the advantage of its early use would be transcending” because it was the “only means of inflicting shock and serious damage to vital elements of the Soviet war making capacity.”

Prepared only a few months before the first Soviet atomic test in September 1949, the premise of the Harmon report was the U.S. atomic monopoly.  The strategic situation that would gradually emerge would require calculations of casualties on both sides of a nuclear exchange.

This version of the Harmon Report includes the unredacted JCS cover memo as found in the Navy’s operational archives at the Naval Heritage and History Command. It replaces the same pages that were excised in the file copy in JCS records at NARA.

To assist military and civilian planners, whether for targeting or civil defense purposes, RAND analysts used early high-speed computers to produce a highly detailed analysis of the vulnerability to nuclear attack of a large city (Moscow), a medium city (Dayton OH), and the Geneva Steel Plant in Utah. Their analysis focused on the more immediate impact of attacks on those targets, not on the “over-all consequences,” which they saw as the topic of another study.

Computing enabled the authors to divide up the three targets into parts so they could estimate the impact of a nuclear detonation on specific sections. Recognizing that cities are complex entities, they divided Moscow – the metropolitan area – into 1,400 sections based on such factors as population, number of dwellings, and square feet of industrial roof cover. The authors then aggregated the results for each section to determine the overall destruction, The point of the study was to answer such questions as how much explosive yield would be necessary to “take out” a city, how did damage vary with yield, numbers of weapons, height of burst, and bombing error, and how much could warning reduce casualties.

The analysis included high-yield weapons with a range up to incredible levels of up to 25 megatons. A ground-burst 10-megaton weapon, delivered by bomber without warning, could destroy half of metropolitan Moscow’s capital assets and two-thirds of its people – well over three million out of a population of five million. Destruction was calibrated in terms of blast effects. The remaining third would be “seriously injured” or eventually killed by fallout. A one-megaton weapon ground-burst on Dayton would cause comparable fractional losses. A smaller, 100-kiloton weapon, dropped on the Moscow city center, would also cause terrible damage, killing outright 500,000 people.

The last finding concerned “super bombs.” While acknowledging that very high-yield weapons, below 25 megatons, would accomplish the requisite destruction, the authors brought a 1,000-megaton bomb into their analysis. If one was dropped on an oil refinery in Noginsk, 25 miles from Moscow, “we would gain as bonus damage the destruction of about 50 percent of Moscow.” Whether the authors knew it or not, the idea of idea of a 1,000-megaton bomb had already been raised in July 1954 by Edward Teller who saw it as the “primary” for a 10,000-megaton weapon. As noted, the authors did not evaluate consequences in their analysis, but members of the AEC’s General Advisory Committee argued, in scathing discussions of Teller’s ideas, that detonating such weapons would “contaminate the Earth.”

That nuclear attacks would kill huge numbers of people was not easily accepted by top officials. Gerard C. Smith, who was the secretary of state’s special assistant on atomic energy matters, played a major role in the nuclear education of Secretary of State John Foster Dulles. After Dulles made press conference statements that underestimated the fallout hazards associated with using low-yield nuclear weapons, Smith discussed with him home truths about weapons effects including the likelihood that megaton nuclear strikes on the Soviet Union would "cause tremendous destruction of life and property.” After Dulles took issue with that statement, Smith realized that his chief needed better information, but he could go no further at the moment because others in the room lacked "Q" clearances for atomic energy information. When Smith found an opportunity for a briefing, he told Dulles that "if the SAC strikes were successful … Russian casualties would be in the tens of millions." The criterion that Smith used for nuclear devastation was blast destruction measured in over-pressure (the air pressure above atmospheric air pressure at sea level): if the megatonnage of nuclear weapons dropped on Soviet cities was "divided into the total area of Russia,” it would "result in substantial over-pressure over the whole area.”

Not long after becoming Under Secretary of State in February 1957, Christian A. Herter received a briefing, involving Gerard Smith and General Herbert Loper, assistant to the secretary of defense for atomic energy, on nuclear weapons and their effects. The outline of the briefing on weapons effects detailed how a nuclear weapon exploded and what its destructive properties were. Treating blast as the "primary destruction agent" for “inanimate objects,” it mentioned thermal radiation as a cause of fires but did not treat fire as a predictable effect. On the vulnerabilities of humans, the outline cited “initial” radiation and fallout dangers, treating blast effects as relatively less important as casualty producers for those in shelters.

The briefing took account of new knowledge of blast affects associated with thermonuclear nuclear weapons. By the early 1950s Air Force target intelligence planners had developed methodologies to predict blast damage, methods that still remain in use. But with their huge yields, thermonuclear weapons produced transient winds that increased damage to structures, such as sheds. In other words, with the same overpressure, the blast wave’s longer duration produced more damage; the degree of damage could be predicted and incorporated into nuclear targeting. The briefing outline explicitly mentioned the sensitivity of a target like a bridge to “drag” also known as dynamic pressure.[18]

During the mid-1950s the Eisenhower administration organized a highly secret NSC Net Evaluation Subcommittee (NESC), charged with developing comprehensive studies of the net impact of nuclear weapons attacks on both sides of the Cold War. The NESC report for 1958 presented a devastating Soviet attack in 1961 involving the detonation on the United States of 553 nuclear weapons with a total yield exceeding 2 thousand megatons. The attack produced "widespread fires" burning out 169,000 square miles or 5.7 per cent of the U.S. land area. Moreover, a "lethal blanket of radiation" covered "at its maximum one-half the nation and persisted in small areas for over two years." Fifty million Americans were dead and nine million were sick or injured, out of a pre-attack population of 179 million. When housing and food supplies were destroyed, exposure and starvation caused more deaths.

The U.S. retaliatory attack on the Sino-Soviet bloc involved 805 bombers, 5 ICBMs and 41 IRBMs. That included every city with a population of over 25,000. Moscow itself was targeted with a 100-megaton attack by missiles and bombers (see document 7). The contemplated U.S. attack completely destroyed "command facilities" in Moscow, Beijing, and Pyongyang and killed 71 million people at once; 30 days later, a total of 196 million people had died (out of a total of 952 million people in the bloc). The attack reduced the Soviet GNP by 75 per cent for the year following the attack. The total bloc GNP was decreased by about 56 percent. The U.S. attack "would virtually eliminate [the Soviet Union] as a world power."

As devastating was the picture of a devastated U.S. and Soviet Union, the report nevertheless found that at the end of the nuclear exchange, "The balance of strength would be on the side of the United States." This sort of optimistic reading would persist in NESC reports, for example, according to the 1962 report, "the net balance … would favor the US."

This NESC report was presented to Eisenhower and the National Security Council. Eisenhower was concerned about the level of destruction, seeing it as too much to strike every city with a population of over 25,000 people, and that it would be better to determine those targets whose destruction would “most economically paralyze the Russian nation.” He did not see any point to adding to the destructiveness of an attack that could produce “a 100 per cent pulverization of the Soviet Union." "There was obviously a limit – a human limit – to the devastation which human beings could endure." These considerations led Eisenhower to request a study on what types of retaliatory attacks would best deter the Soviet Union from future attacks: an attack on a military-target system or an "optimum mix" of military and urban-industrial targets. The request was NSC Action 2009, and the resulting studies would lead to the first Single Integrated Operational Plan (SIOP).

In his review of the NESC report (see Document 5), Policy Planning Staff Director Gerard C. Smith pointed out that 100 megatons in nuclear weapons were targeted on Moscow and raised questions about the morality of the over-destruction. Referring to fallout hazards, he noted that the U.S. retaliation against Soviet targets would have “serious effects on non-belligerent nations.”

At a meeting with members of the State Department Policy Planning staff, two White House science advisers, David Z. Beckler and an associate, Dr. Waters (first name unknown), discussed a range of security issues, including the effects of nuclear weapons, “clean” nuclear weapons, and the problem of warning time. On nuclear weapons effects, they reviewed the particulars of the devastating impact of a thermonuclear war on people in the Northern Hemisphere. Starting with the base-line of 10,000 megatons as the estimated explosive yield of a U.S.-Soviet nuclear war, Beckler declared that the stratospheric fallout would cause “enormous” numbers of casualties, but they “would represent only a small portion of the total casualties from all causes (blast thermal radiation, fire, and local fallout).” Beckler was speaking long before the possibility of “nuclear winter” had been recognized, although some contemporaries had considered the adverse implications of global nuclear war for the climate.

Dr. Waters opined that fears of genetic mutations from nuclear weapons radiation were “exaggerated” because “a large proportion of the potential parents would perish as a result of the massive attacks.” He also observed that radioactive iodine “one of the most harmful products of nuclear explosions has been given scant attention” compared to strontium-90. Yet it was “an extremely great hazard to children as it attacks their thyroid glands.” This point became more palpable in the years ahead, in light of the thyroid gland diseases experienced by the young Marshall Islanders who were exposed to radioactive fallout from the 1954 Castle Bravo test.

The devastation of nuclear war that U.S. nuclear targeting plans could produce was itemized in a report produced in June 1961. Then the Berlin Crisis was heating up because Khrushchev had just set his six-month deadline for a treaty with East Germany. This Joint Chiefs of Staff report on “Berlin Contingency Planning” included key details of the first SIOP, the centralized war plan covering the nuclear weapons systems of both the Air Force and the Navy.[19] The report, sent to Secretary of Defense Robert McNamara, was for National Security Adviser McGeorge Bundy to use for answering questions posed by White House adviser and former Secretary of State Dean Acheson.

According to the report, if the full force assigned to the SIOP struck the Soviet Union, 54 percent of its total population would be killed, 108 million out of 217 million people. If the smaller alert force (bombers on 15 minutes to 2-hour alert) was used, total Soviet casualties would be 37 percent, or about 80 million. Against China, the full force would kill a smaller proportion because there was a smaller urban population: 16 percent of an estimated 650 million, or about 104 million. Poland, the site of Soviet air defense installations, would suffer over 2.5 million casualties in an attack by the full force. The casualty estimates included those caused by fallout and probably other radiation effects.

The horrific casualty figures were the direct consequence of standard operating procedures used by war planners. According to this JCS report, SIOP-62 stipulated a 95 percent assurance factor for the destruction of targets “representing a direct threat” to the continental U.S. Typically against such military targets, ground bursts would be used, which would produce more fallout. To meet the assurance criteria, SIOP-62 posited a mass attack by an alert force of over 1,700 nuclear weapons. Some Designated Ground Zeroes (DGZs) would be struck by two or more weapons, creating more devastating weapons effects. The high degree of assurance sought was one of the bases for the “overkill” that some have associated with U.S. nuclear war plans.

The Joint Chiefs’ total estimated deaths for the full force attack were far fewer than the estimate of 600 million fatalities for the Soviet Union, its allies, and China that JCS Chairman Maxwell Taylor provided to President John F. Kennedy also in 1961, as disclosed by Daniel Ellsberg.  His revelation of this startling number is important, but the document that included them is elusive, despite searches in the records of the Joint Chiefs of Staff at the National Archives.[20]

When SIOP-62 was devised, no one in the military or the civilian sector had responsibility for determining what the appropriate numbers of weapons in the plan should be or how much destruction it required. The SIOP was simply a “capabilities plan,” where the JSTPS simply assigned to the plan all the nuclear weapons that the services had put at its disposal, based on their budgetary assessments.” It was not until the end of the Cold War that the SIOP was subject to scrutiny at the (civilian) Secretary of Defense level.[21]

These documents detail the destructiveness, in terms of fatalities, of a U.S-Soviet nuclear war in 1966, triggered under various circumstances (e.g., first strike by, or 15 minute warning for, both superpowers). Strike categories were broken down into military, military-urban industrial, or urban-industrial targets. The estimates of fatalities included not only U.S. and Soviet, but also Western European, NATO allies. Each section indicated fatalities from a first strike but also fatalities from a retaliatory strike. The estimate was further broken down into ground bursts, producing more fallout, and air bursts, producing far less. Moreover, some sheltering was assumed for both the United States and the Soviet Union, but none for Western Europe. Thus, a Soviet first strike ground burst attack killed 115 Western Europeans and 56 million Americans, while first strike air bursts killed 15 and 25 million respectively. Numbers of warheads associated with the strike and counter-strikes were not disclosed.

According to the chart, Polaris Submarine-launched ballistic missiles were held in reserve, to respond to the Soviet first strike, for later bargaining purposes, but if they had been used in the initial strike the level of Soviet fatalities would have been higher.

The second page indicated aim points, numbers of alert force warheads, and fatalities for a U.S.-NATO strike against Soviet bloc targets, under conditions of 15-minute warning of a Soviet strike or no warning of a Soviet nuclear attack. Strike option "II", which included a substantial NATO medium-range missile force, estimated better "expected kill" results than otherwise.

As the Cuban Missile Crisis was unfolding, CJCS Taylor aide William Y. Smith sent information that suggested the worst-case outcome. It was a summary of information on fatality levels from various U.S. strategic attacks prepared by Robert McNamara’s assistant for atomic energy matters. The illustrative Soviet fatality numbers are excised completely from the document. The excisions typify the Pentagon’s variable approach to releasing casualty numbers, suggesting considerable discomfort with their declassification.

The optimistic conclusion of earlier NESC reports – that the U.S. would come out ahead in terms of the “balance of strength” – disappeared in the 1963 report. Summarizing a longer report, the "Oral Presentation" prepared for an NSC meeting analyzed hypothetical nuclear war scenarios during1964-1968; in each, one of the superpowers initiated a preemptive strike and the other retaliated. The NESC's conclusions were grim: for "the years of this study … neither the US nor the USSR can emerge from a full nuclear exchange without suffering very severe damage and high casualties."

To measure the impact of the attacks, the NESC focused on fatalities: with their growing ICBM force, the Soviets would be killing more. In a 1964 preemptive attack, they launched 108 weapons on the U.S., killing 63 million people. In a 1968 attack, they struck with 409 weapons and killed 134 million. In both 1964 and 1968, a U.S. preemptive attack killed about the same numbers in the Soviet Union, about 140 million.

Soon after the report was completed, the National Security Council received a briefing by the NESC's director, General Leon Johnson, on 12 September 1963. McGeorge Bundy had already reported to President Kennedy on the basic conclusions as well as questions that he could raise. According to the record of the meeting, the judgment that the United States would not suffer less damage than the Soviet Union brought President Kennedy to the conclusion that the superpowers were in a "nuclear stalemate" and that preemption, which Bundy had equated to a first strike, no longer offered an advantage. While NESC director Johnson believed that in a stalemated situation "nuclear war is impossible if rational men control governments," Rusk was skeptical, arguing that "if both sides believed that neither side would use nuclear weapons, one side or the other would be tempted to act in a way which would push the other side beyond its tolerance level."

The trend identified in the 1963 NESC report (see Document 12) – growing numbers of U.S. deaths from a Soviet strike, whether retaliatory or otherwise – continued in this 1967 report on the foreign policy implications of major changes in the U.S.-Soviet strategic balance. With the growing numbers of Soviet ICBMs the U.S. had become even more vulnerable to destructive attack. Damage limiting capabilities were in decline. According to the estimate, in 1964, the Soviets could kill 48 million Americans in a preemptive attack; by 1968, they would be able to kill 91 million (see table at PDF page 47, with estimated numbers of weapons unspecified). By contrast, with U.S. strategic forces already built up by the early 1960s, the trend in Soviet fatalities was constant during the decade. In a retaliatory attack on Soviet cities in 1964, some 77 million would be killed. Under the same circumstances, 81 million would be killed in 1967 (see table at PDF page 47). With the MIRV program underway, doubling numbers of U.S. warheads during the 1970s, the authors of the study were confident that U.S. deterrence would remain “more than adequate.” Nevertheless, given Soviet capabilities “mutual deterrence” was at hand. The foreign policy implications of changes in the strategic balance were uncertain.

Pentagon systems analyst Ivan Selin used estimates of fatalities to make a point about levels of investments in U.S. strategic forces. Selin’s memorandum, which Deputy Secretary of Defense David Packard sent to Henry Kissinger, demonstrates that the U.S. could not undertake a major program to limit damage without inviting comparable Soviet reactions. Selin summarized the basic findings of a report prepared to answer National Security Study Memorandum 3 on the “security and foreign policy implications of a wide range of alternative budget levels and strategies for strategic and general purpose forces.”

Drawing on the NSSM 3 report, Selin’s memo identified several damage-limiting force structures involving heavier expenditures on ICBMs and ABM systems that, if used in conflict, could greatly reduce U.S. casualties in the event of a U.S. first strike, or even if the Soviets struck first. If the U.S. invested in more defenses and more counterforce, if the U.S. struck first, the Soviet response would kill fewer than 30 million Americans. If, however, the U.S. stayed at current force levels, the Programmed Force, 90 million Americans could be killed by the Soviet retaliation. If the Soviets struck first, however, U.S. deaths would be high, but fewer than 80 million with the damage-limiting forces, but around 140 million with the Programmed force.

If the Soviets built up their forces in response to U.S. damage limiting programs, then they could kill many more Americans through a first strike or through retaliation. The fatalities would be at the level of 140-145 million in the event of a Soviet first strike. Responding to the Soviet buildup by increasing ABM and air defenses could result in annual budgets in the $20 billion range (current dollars), which would also contribute to a destabilizing U.S.-Soviet arms race.

This massive study, which supported further review of SALT options during the following months, included a survey of U.S. and Soviet strategic forces, arms control implications of weapons systems, U.S. strategic and arms control objectives, and verification issues. All of that information was the basis for alternative “packages” for negotiating positions. A crucial point was what force levels among the options would provide a retaliatory capability that would be destructive enough to deter, with force effectiveness measured in part by numbers of casualties.

One table, part of a discussion of current strategic force capabilities [see PDF pages 17-19], found that through detonating 1,000 one-megaton weapons on Soviet cities, U.S. strategic forces could kill 43 percent of the Soviet population in a retaliatory strike. “The capability to kill 30% of the Soviet people corresponds to the destruction of the 150 largest Soviet cities and more than 50% of the Soviet industrial capacity.” If the Soviets struck first, the U.S. retaliatory strike would kill 120 million Soviets. The U.S. would already have lost 100 million. In a U.S. first strike, 100 million Soviets would be killed, while the Soviets would kill 110 million in the U.S. Consistent with the logic of strategic stability thinking, the ability by each side to achieve such results would constitute a stable situation because each would suffer comparably. However, should one side or the other acquire a strategic advantage, e.g., means to reduce population losses, the results could be destabilizing, especially in a crisis, when one side might have an incentive to preempt.

The fatality estimates associated with various SALT options were based on immediate or “prompt” fatalities due to blast effects only. For the retaliatory strike, the analysts used a “conservative estimate” of 10 pounds per square inch (PSI) to measure the overpressure needed to destroy Soviet cities with airburst weapons. For the war-fighting calculations, they were “less conservative and assumed that an 8 psi overpressure is sufficient for severe damage to cities of either side.

The authors acknowledged that reliance on the blast pressure criterion “tends to underestimate the resulting fatalities.” Much death and damage “would be expected to result from other effects such as direct nuclear and thermal radiation, fire storms, fallout, epidemics and starvation.” [See PDF pages 297 and 299].

The NSC’s Verification Panel, which was reviewing SALT options, briefly discussed fatality estimates. When Deputy Secretary of State David Packard observed that an unspecified estimate was inconsistent with one developed by SIOP planners, ACDA official Spurgeon Keeny suggested that the computers used for SIOP calculations “underestimated deaths” [perhaps because of the emphasis on blast effects]. Packard acknowledged the problem but suggested that the estimate that he was criticizing did not consider “all factors.” Deputy Secretary of State Elliot Richardson took the discussion to another level by pointing to a “political-psychological” issue: “the imbalance between ability to inflict fatalities and reluctance to accept or cause large numbers of deaths.” This dilemma was a basic one that civilian or military officials involved in nuclear war planning necessarily face.[22]

Suggesting that Soviet civil defenses were not as ineffective as often thought, Leon Sloss argued that they could make a difference in casualties caused by a U.S. attack. Soviet losses were minimized if there was time to evacuate cities, but millions could be saved if bomb shelters were used. If “Assured Destruction” was measured simply in terms of industrial devastation, however, civil defense was irrelevant. A U.S. retaliatory strike could produce 70 percent damage to Soviet industry.

One major initiative taken late in the Nixon administration was to develop plans for the limited use of nuclear weapons in crises with the Soviet Union and other adversaries. The point was to avoid major strategic attacks and massive devastation on each other’s territory. While the National Security Council Staff was developing guidance for President Nixon to sign, in late November-early December 1973 senior administration officials participated in a crisis “simulation” in which Blue and Red Teams developed political and military responses involving limited use of nuclear weapons.

In the Scylla scenario, Iran and Iraq went to war with each other during 1976 over the latter’s invasion of Kuwait, bringing in Washington and Moscow as the latter sent troops to Iran to prevent Iraq’s defeat. Because the “Blue Team” concluded that conventional responses could not protect Iran, the U.S. initiated nuclear attacks on Soviet forces in Iran with the purpose of avoiding strategic nuclear attacks that could escalate the conflict to all-out war. The projected scenarios involved the use of dozens of nuclear weapons on Soviet forces, with the latter, also refraining from strategic attacks, responding with nuclear strikes on U.S. aircraft carrier task groups, and even U.S. naval facilities in Guam. In the ensuing onslaught, thousands of Soviet and U.S. military personnel were killed and an estimated 500,000? Iranians (civilian casualties in Guam were not mentioned). Both Moscow and Washington agreed to negotiations when they had decided that they had not “lost” the conflict.

The presidential nuclear use decision is recorded on page 74. This would have been the first use of nuclear weapons in combat since 1945, but the decision and its implications were not debated, probably because the point of the exercise was to play out a nuclear use decision. Significantly, the after-action critique involved some retrospection on that very point. Some Blue Team players wondered whether nuclear use was “warranted” (PDF p. 254) and observed that when the U.S. made its first use, “we zipped up the ladder awfully fast” (PDF p. 259). Moreover “in reality” not as many weapons should have been used against Soviet troops. As for the Soviets, one of the Red Team players suggested that their use of nuclear weapons might have been a mistake because the Soviet side would have looked better otherwise and also would have been in a more politically advantageous position [PDF p. 262].

After the early 1970s, the trail begins to thin for declassified government estimates of nuclear war casualties. A report from late in the Gerald R. Ford administration typifies the secrecy surrounding such estimates. A comparison of U.S. and Soviet strategic capabilities developed by the Joint Chiefs of Staff drew on SIOP targeting and a postulated Soviet war plan, the Red Integrated Single Operational Plan (RISOP). The estimate was based on a Soviet preemptive attack with the U.S. retaliating. The model developed ended up with fatality estimates that were consistent with previous trends. The Interagency Security Classification Appeals Panel (ISCAP) decision on this report withheld the numbers of fatalities for various attack levels but the analysis will be familiar: “Expected fatalities have almost doubled since 1966 while the totals in the USSR have remained relatively constant.” According to the Joint Chiefs, the greater numbers of U.S. deaths was due to the increase in Soviet megatonnage but also to the way that the RISOP used the weapons, plainly against urban-industrial centers.

While not including fatality estimates, this report was relevant to thinking about that problem by reviewing measurements of the lethality and destructiveness of strategic arsenals. For example, on the Soviet side, Moscow had an estimated advantage over the U.S. in the “lethal area for soft targets” and numbers of delivery vehicles while the U.S. had advantages in numbers of reentry vehicles and capabilities against hard targets (the “K factor”). With respect to measuring lethality, the report found that both “the US and the USSR … possess the potential to devastate more than 29,000 square nautical miles of the opponent's country at this level of vulnerability.”

A discussion at the close of the report was pertinent to casualty estimates. During the 1960s and early 1970s estimates of both fatalities and industrial devastation served as “dynamic measures” of capabilities for assuring deterrence. According to the report such estimating had dropped away in part because of substantial changes in the “data base used to determine weapons allocations and the understanding of weapons effectiveness.” Also relevant was that the “newly evolved concept of strategic deterrence has shifted away from emphasis on fatalities as a measure of effectiveness.” These were oblique references to the concept of “assured destruction,” which was becoming subject to the criticism that it was immoral to postulate high levels of fatalities as a measure of deterrence.

From the Jimmy Carter administration some casualty estimates developed in response to Presidential Review Memoranda have been declassified. For example, one of the PRM-10 reports indicated that under several strategic targeting scenarios, there could be no “winners” – the Soviets would lose at least 113 million people while the United States would lose at least 140 million [See PDF page 63 of linked report].

PRM 32 on civil defense, signed by Carter in September 1977, produced an interagency study that also involved the use of fatality and casualty estimates. According to national security official Samuel P. Huntington, the study took a calm view of Moscow’s civil defense efforts, but if information in the report was leaked (which was likely) it would support those who already held alarmist views. For example, the Soviets had blast shelters for the “top leadership,” which included about 110,000 people.

This brief summary of the interagency report included information on estimated Soviet casualties caused by a “US attack with 2,200 weapons directed at economic targets.” Installations deemed necessary for Soviet post-war economic recovery – industrial, energy, and transportation facilities – had become an explicit target category late in the Nixon administration and remained on the target list early in the Carter administration. According to the estimate, with two hours of notice only, an attack could kill up to 70 million people with another 50 million injured. The upshot was that “The Soviets cannot prevent massive industrial damage, a breakdown of their economy, and the destruction of their most significant material accomplishments.”

Even though the Soviets were putting significant resources into civil defense, the report found that it would not increase the danger of conflict: the civil defense program “will not embolden [the Soviets] deliberately to expose the USSR to a higher risk of nuclear attack.”

A few details of casualty estimates can be found in the minutes of a December 1981 meeting of the Reagan administration’s National Security Council, during a discussion of civil defense. A military official disingenuously told President Reagan that, if the Soviets attacked the United States, the U.S. would lose 150 million people while the Soviets would lose only 15 million from a US retaliatory attack (“less” than from World War II or the purges) because urban dwellers would have been evacuated. The discussion of Soviet efforts to protect population led a senior FEMA official, General Bennett Lewis, to claim that in five years the U.S. could evacuate enough people to protect 80 percent of the population, which meant a huge estimated loss of 20 percent. Reagan was not told, however, how much greater the Soviets losses could have been without evacuation or in the event of a surprise U.S. attack. Eventual declassifications may disclose when and whether senior officials later shared different numbers with Reagan.