By the mid-1960s it was becoming clear to Marshall that RAND’s core mission of advising the Air Force on the “preferred instrumentalities and techniques” for intercontinental strategic warfare required an in-depth assessment of the relative military power of the two competitors. For nearly two decades Marshall had thought increasingly about the problem of measuring military power. But it was William Kaufmann who persuaded him to set down his thoughts on this issue.
Kaufmann had left RAND in 1961 and joined the political science department of the Massachusetts Institute of Technology. In 1966 Kaufmann was asked to run a panel at the September meeting of the American Political Science Association. This opportunity prompted him to ask Marshall to present his views on estimating military power. The result was Marshall’s classic “Problems of Estimating Military Power” (P-3417). The paper detailed the profound conceptual problems and practical difficulties confronting all attempts to define appropriate measures of the capability of one country’s military forces to deal with those of another across a range of relevant contingencies.
Marshall’s skepticism about reliance on quantitative metrics was evident in his observation in the paper’s introduction where he declared that any attempt to estimate the military power of the United States, or of any other nation, contained so many conceptual and practical challenges that “there seem to be only problems and very few, well-accepted adequate methods of making such estimates.”54 The most common approach to estimating one nation’s military power relative to another’s, he elaborated, was to count up the personnel under arms, the weapons of various types, and the major formations (divisions, fleets, air wings, etc.) on each side and compare the numbers. But such tabulations, Marshall argued, were “an evasion of the problem” because they said “nothing about the actual capabilities of the forces of one country to deal with another.”55 They ignored the influence of geography and logistics on combat outcomes as well as the effects of human error, misguided doctrine, stodgy planning, and the unpredictable behavior of governments and military organizations under the stresses of wartime.
At a practical level, a simple compilation and comparison of opponents’ forces implicitly assumed that quantitatively equal forces would give rise to equal military power. But in Marshall’s judgment the whole history of military conflict indicated that actual outcomes could diverge dramatically from comparisons of the numbers of personnel, weapons, and military formations. For example, in World War II the German attack through the Ardennes in May 1940 quickly defeated the forces of France and Low Countries and ejected the British from the continent at Dunkirk, even though the Allies had more men, divisions, and tanks than did the Germans.
In light of such disconnects between the usual quantitative comparisons of military forces and combat outcomes, Marshall concluded that the conceptual problems of constructing adequate or useful measures of military power had yet to be squarely and honestly faced. Defining adequate measures, he wrote, “looks hard, and making estimates in real situations looks even harder.”56 These conclusions reflected Marshall’s long-standing interest in military history and preference for empirical data on a diverse range of factors over simple “bean counts” of forces and weapons, or abstract theory. His reservations concerning anyone’s ability to predict probable war outcomes would receive dramatic confirmation in the ease with which US-led coalition forces ejected the Iraqis from Kuwait in 1991 despite the belief of many that coalition forces would suffer high casualties in attempting to do so.
In the second half of his paper, Marshall returned to the problem of estimating the future military forces of potential opponents. Based on his earlier collaboration with Loftus he identified the key challenge as the generation and documentation of useful hypotheses concerning the behavior of large governmental institutions and military bureaucracies.57 Rational-actor explanations of the slow, complicated process of Soviet military force developments since World War II, he wrote, “were substantially in error.”58 He warned that until it was possible to arrive at a far better understanding of the decision-making processes within typical military bureaucracies, it was doubtful that US analysts could produce even rough projections of the USSR’s military posture—or that of any other nation—more than four or five years into the future.59
As with the problems of estimating military power, those relating to estimating the future military postures of potential adversaries have not been definitively resolved in the decades since Marshall’s 1966 paper. Estimating likely combat outcomes in actual situations appears to remain beyond our powers of prediction. War continues to be an incredibly complex phenomenon characterized by massive uncertainty. Strategy, tactics, decision-making in times of immense stress, and other human factors matter greatly. Even in the short-term, combat outcomes are highly sensitive to scores of such factors, and small changes in any of them can mean the difference between victory and defeat.
Long-term estimates of the future force postures and strategies of opponents face similar difficulties. Unless one has considerable insight into the histories, preferences, and incentives of the various actors able to influence a nation’s choices in preparing for or conducting war, such estimates are unlikely to be accurate. The USSR’s continued buildup of nuclear forces after the Kremlin achieved rough parity with the United States in the early 1970s remains a classic example of the difficulties of estimating adversary behavior. Marshall’s solution to these problems in the 1960s was to redouble his efforts to understand the actual behavior of Soviet organizations. Here he found an ally in the Pentagon in his former RAND colleague Ivan Selin.
From 1960 to 1965 Ivan Selin had been a research engineer at RAND. In 1963 he left RAND and joined Enthoven’s Office of Systems Analysis. There he began working on how the United States could most efficiently address the growth in Soviet strategic nuclear forces, especially in intercontinental ballistic missiles. By 1967 he realized that he needed better forecasts of future Soviet nuclear forces and how the Russians might react to improvements in US nuclear forces.60 Aware of Marshall’s continuing efforts to use insights into Soviet organizational behavior to improve such estimates, Selin asked Marshall to put together a group of specialists in management and organizational theory to see whether their perspectives could help generate better projections.61 Besides Marshall, the main participants in this exercise were James March from the University of California–Irvine; Joseph Bower and Roland Christensen from the Harvard Business School; Andris Trapans, a RAND Soviet specialist; and Graham Allison.
It took some doing to get this project under way. Selin and Marshall had agreed that the group would need access to the best available intelligence on those elements of the Soviet bureaucracy most able to influence decisions about the USSR’s nuclear forces. Obtaining the requisite security clearances proved easier said than done. In the end, the deputy defense secretary, Paul Nitze, had to prevail upon Richard Helms, the director of the Central Intelligence Agency, to grant everyone involved the access needed for the project.
During 1968 the group met four times, and in August of that year it produced a report for Selin. The document rejected the use of simple models of Soviet behavior that assumed effective central direction, complete information, and consistent goals, none of which are usually found in business firms, corporations, or other large organizations. Instead, the group characterized the decision processes of the Soviet military-industrial complex as a loosely coupled collection of intermittently disrupted problem-solving clusters, and noted that some major Soviet decisions were actually made by organizations far below the Politburo, the Main Military Council, or the Soviet General Staff. A preeminent fact about all large organizations, the report argued, is that their sheer size precludes any single central authority having either the time or the information to make all the important decisions. Overall, the Soviet military establishment’s decision-making exhibited the “bounded” rationality characteristic of any large organization, including military services.
&nb
sp; None of these conclusions should have been surprising. They certainly were not to Marshall. They reflected lines of thought that he had pursued since his collaboration with Loftus, if not earlier, as well as March’s findings from his work with Cyert and Simon. But by 1968 Marshall had compiled considerably more empirical evidence for these judgments than he had possessed when he and Loftus started SOVOY. As a result, a much stronger case could be made for embracing an organizational behavior approach to RAND’s research on both Soviet and US military decision-making, especially in the areas of resource allocation and force posture choices.
Instituting this approach within RAND is precisely what Marshall and Sidney Winter proposed to the Air Force in March 196762 and to RAND’s new president, Henry Rowen, in December of that year. In their memo to Rowen (M-8668), Marshall and Winter began by observing that the field they termed “Organizational and Management Studies” did not exist. The efficient-resource-allocation models used in game theory and employed less formally in strategic analyses, they argued, tended “to abstract completely” the constraints imposed on top-level decision makers by the “cumbersome and contrary organizations below them.” It was therefore “time to attempt the development of some intellectual tools powerful enough to close at least some of the gap between the models of systems analysis and the realities of policy making.”63 The potential benefits, they argued, included more effective policy analysis and implementation by RAND for its clients, better assistance to the Air Force in dealing with its force management problems, better theory development to clarify the relationships between rational choice and actual decision processes, and improved predictions of the behavior of governments and military organizations. This, then, was their case for establishing a RAND Department of Management Sciences.
Rowen, however, was not sufficiently convinced to create a new department to develop organizational theory. Although Marshall and Schlesinger devoted considerable time to explaining to Rowen their concerns about the overly quantitative focus of RAND’s strategic research and the need to develop new analytic methods to address the limitations of systems analysis, the new RAND president chose a different path. In 1968 he created a new position, director of strategic studies, and appointed Schlesinger to it. In this role Schlesinger was to consolidate and sharpen RAND’s continuing research on US and Soviet intercontinental nuclear forces.
One result of Schlesinger’s appointment seems to have been that Marshall was able to spend more of his time working on Soviet organizational behavior. In July 1968 he produced a short piece for Selin on the problems posed by the discrepancies between the observed aspects of Soviet military programs and the behavior suggested by the rational-actor model. As an alternative he offered the hypothesis that the design of Soviet weapons mainly took place several levels down within the USSR’s political and military hierarchy,64 and that the top people only decided on the major outlines of Soviet nuclear forces through their control of budgets.
Marshall stayed busy during this time in other ways as well. Starting in 1967, he and Schlesinger had begun serving as consultants to the Bureau of the Budget (renamed the Office of Management and Budget in 1970). Marshall continued to advise the CIA on its national intelligence programs. Reflecting the high demand for his talents, Marshall also undertook a comparison of US and Soviet research and development programs, systems, and technologies for N. F. “Fred” Wikner, who at the time was the special assistant for threat assessment to John Foster, director of the Pentagon’s Office of Defense Research and Engineering (DDR&E).
Although the study for Wikner (WN-7630-DDRE) was not published until 1971, it merits mention because of its call for broad net assessments to inform US decision makers on where the United States stood relative to the Soviets in key areas of US-Soviet military competition. Only such detailed analyses, Marshall suggested, would enable the Defense Department’s senior managers to do a better job of balancing risks and identifying areas of comparative advantage that the US military could exploit through the development of specific technologies and capabilities.65 It was a position he would have a much greater ability to advance during the next stage of his career.
After Richard Nixon won the November 1968 presidential election Schlesinger took a leave of absence to work on the new administration’s transition team. In January 1969 he left RAND permanently to be Nixon’s assistant director of the Bureau of the Budget, with primary responsibility for the defense portfolio. His departure left RAND’s director of strategic studies position vacant. At Rowen’s urging Marshall accepted the job.
In his new position Marshall began reviewing all the research being conducted throughout RAND on US and Soviet strategic nuclear forces. He soon concluded that they lacked coherence. While the studies addressed all sorts of nuclear and related issues, they rarely focused on the central task of finding ways in which the United States could be more effective in its nuclear competition with the USSR.
Reflecting on his two decades at RAND, Marshall was struck by the fact that there had been no US-Soviet general nuclear war during those years. Yet most analyses of the nuclear competition, at RAND and elsewhere, had focused on exploring the consequences of the so-called arsenal exchanges in the event that deterrence failed. The dominant US scenarios were either a nuclear war that began with an all-out Soviet nuclear attack on the United States or a NATO–Warsaw Pact conflict in Europe that escalated to general nuclear war. But with the Cold War entering its third decade, Marshall realized that nuclear deterrence had not failed. True, the United States and the Soviet Union had come perilously close to the nuclear abyss during the Cuban missile crisis in October 1962. But they had stepped back from the brink. Marshall also suspected that the United States’ possession of a superior nuclear posture early in the Cold War had produced strategic effects and payoffs that went beyond deterring a Soviet nuclear attack.
Looking ahead, Marshall was convinced that the military competition with the Soviets would be an extended one that would continue for the foreseeable future. Consequently, any and all future choices that the United States made regarding its nuclear posture and capabilities should be seen as a series of moves aimed at improving the US position vis-à-vis that of the USSR. How was this to be accomplished? Marshall’s answer: by emphasizing areas of US competitive advantage, such as advanced technology; by exploiting known Soviet weaknesses and propensities; and by making strategic choices that posed major difficulties for Soviet decision makers. Of course, to achieve these goals required a coherent strategy. That strategy, in turn, had to meet two requirements: (1) careful net assessments to monitor how the United States was doing in the key areas of the protracted peacetime competition with the Soviet Union; and (2) greater understanding of Soviet organizational behavior, particularly of the enduring organizational and resource constraints on Soviet decision-making. Looking back after the Cold War had ended, Marshall would recall thinking that this approach represented a “terrific new way” of looking at the United States’ rivalry with the USSR.
In 1969 Marshall began spelling out this new perspective in a paper he titled the “Long-Term Competition with the Soviets: A Framework for Strategic Analysis” (R-862-PR). Although RAND did not publish R-862-PR until 1972, Marshall finished drafting it in 1970, around the same time he was being drawn into addressing Richard Nixon and Henry Kissinger’s concerns over the flow and quality of the foreign intelligence coming into the White House. R-862-PR echoed the data-driven understanding of Soviet nuclear force development that he and Loftus had pursued from the mid-1950s to the early 1960s, when they began trying to understand the behavior of Soviet organizations. More important, however, Marshall’s long-term-competition framework contained the seeds of major themes that he would later stress as the director of net assessment. During every annual Pentagon budget cycle, the military services competed for shares of the defense budget and sought to fund the forces, acquisition programs, and capabilities they considered most vital. This practice inevitably led the services
to downplay US strengths and exaggerate Soviet capabilities. Marshall’s paper on the long-term competition with the Soviets argued instead for identifying strategic choices that capitalized on areas of US comparative advantage and exploited Soviet weaknesses and behavioral propensities so as to impose disproportionate costs on the USSR.
Marshall’s long-term competition framework was even more of a watershed in influencing US Cold War strategy than was his seminal paper on the problems of estimating military power. He was—and remains—a pragmatic strategist who has consistently looked further into the future and focused more on the first-order problems of long-term competition in peacetime than those around him. And starting with a phone call in the fall of 1969 from President Nixon’s national security adviser, Henry Kissinger, Marshall would soon bring his strategic acumen and insight to Washington, DC.
4
THE BIRTH OF NET ASSESSMENT 1969–1973
We’re here to inform, not to please.
—ANDREW MARSHALL
From 1953 to 1965 the National Security Council had a secretive subcommittee that conducted annual “net evaluations” of the impact of a Soviet nuclear attack on the United States. This subcommittee was disestablished in 1965. Within a few years, however, calls began to surface for reestablishing a net assessment function within the NSC. By 1969 the need for better, more precise comparisons of where the United States stood relative to the USSR in various areas of competition was becoming increasingly apparent and ultimately resulted in Andrew Marshall’s establishing a net assessment capability in the Pentagon.
This piece of Cold War history began in January 1953. On his last day in office, President Truman established the Special Evaluation Subcommittee (SESC), which was renamed the Net Evaluation Subcommittee, or NESC, in 1955.1 The SESC’s purpose was to prepare annual reports on the net effects—overall damage, human losses, and politico-military outcomes—of a Soviet nuclear attack on the United States. The SESC’s first report in June 1953, for example, assumed that if Soviet leaders decided to initiate a nuclear attack, it would start with strikes by Soviet Long Range Aviation on US bomber bases in the United States, Europe, and the Far East along with the “heaviest possible attack” on major population, industrial, and control centers in the continental United States.2 The report estimated that, in the period from 1953 to mid-1955, 24 to 30 percent of US nuclear-capable bombers would be lost in the opening Soviet attack, the US population would suffer over 10 million casualties (50 percent of which might be fatal), and there would be initial paralysis of all industry within the areas attacked.3 Nonetheless, the SESC also concluded that even after such a Soviet attack the United States would be capable of mounting a “powerful initial retaliatory atomic air attack, continuation of [the] air offensive, and [the] successful prosecution of the war [emphasis in the original].”4
The Last Warrior Page 11