The Last Warrior

by Andrew F. Krepinevich

  Besides challenging Posen’s sanguinary assumptions regarding the two sides’ mobilization rates, Cohen again noted that the Soviet Union could compel its allies to mobilize, whereas the United States could not. Cohen then turned to the issue of surprise: the possibility that NATO intelligence would either fail to provide timely warning of a Warsaw Pact attack or the warnings would be ignored. Cohen recalled that the commander of the US Army in Europe first heard about the Warsaw Pact’s August 1968 invasion of Czechoslovakia, involving over 200,000 Soviet, Bulgarian, Hungarian, and Polish troops, not from his intelligence staff, but from an Associated Press dispatch. He also noted that despite a number of intelligence warning indicators, the Israelis were surprised by the Egyptian and Syrian attacks in October 1973 that initiated the Yom Kippur War.103 Cohen reiterated Roberta Wohlstetter’s warning that it is risky to assume that accurate intelligence will reach senior decision makers promptly and that they will be able to filter it from the “noise” of less relevant information. To this might be added Graham Allison’s observation in Essence of Decision that individuals and organizations can, at times, filter out useful intelligence when it does not match their preferred view of how their “rational” rivals would behave.

  In assessing a military balance such as that between NATO and the Warsaw Pact in the 1980s, Marshall’s view has long been that the first task is to compare the quantities of men and equipment (tanks, artillery, troops, etc.) each side has at its disposal. Second, an assessment needs to capture qualitative differences between the opposing men and equipment (such as the firepower, mobility, and survivability of a US main battle tank versus that of its Soviet counterpart, Soviet conscripts versus the professional soldiers of an all-volunteer army, differences in command and control, etc.). Third and most crucial, in Marshall’s judgment, a good net assessment needed to incorporate intangible, often uncertain, variables such as geography, weather, logistics, warning times, surprise, and readiness along with asymmetries between the two sides in training, tactics, military doctrine, campaign strategy, and theater objectives.

  To address the first two of these requirements, Posen used the US Army’s WEI/WUV methodology. Mearsheimer did the same, asserting that, “The best measures of relative conventional strength are those that capture the full range of the combat capabilities of the forces, including their mobility, survivability, and firepower (including the rate and lethality of their fire). The measure that best captures these capabilities in ground forces is the ‘armored division equivalent’ (ADE) score that the Pentagon uses as its basic measure of ground force capability.”104

  An Armored Division Equivalent (ADE) is based on the WUV score of a standard US armored division. If the WUV aggregate for a US armored division was 120,000 and that of a Soviet tank division was 96,000, then dividing both units by the WUV score of the US division would yield ADEs of 1.0 and 0.8 for the US and Soviet units, respectively. ADEs could then be calculated for both sides’ ground forces throughout the European theater, thereby producing a quantitative force ratio between NATO and Warsaw Pact ground forces. Using this methodology both Mearsheimer and Posen concluded that the Warsaw Pact had an ADE advantage over NATO of 1.2 to 1.105

  Setting aside the fact that the Army’s Concepts Analysis Agency that first developed the WEI/WUV scoring system discounted its utility as a valid means for assessing the relative effectiveness of opposing forces in actual combat, Posen adjusted the theaterwide force ratio by assigning a multiplier of 1.5—an increase of 50 percent—to NATO ADEs because NATO allocated “1.2 to 2 times the personnel as the Pact to generate a given unit of firepower.”106 Here Posen was endeavoring to reflect NATO’s higher investment in support forces, or logistics, than the Warsaw Pact in the theaterwide force ratio.

  Cohen questioned Posen’s use of the amount of invested resources as the proper metric for gauging relative military effectiveness. He noted that NATO logistics forces were larger, in part, because of the greater emphasis given by the alliance’s governments to providing their troops with creature comforts, which do not exhibit a high correlation to overall military effectiveness. He also noted that while Warsaw Pact forces operated with a common set of equipment and spare parts established by the Soviet Union, the major NATO militaries typically fielded their own major items of equipment, greatly complicating the challenge of providing logistics support while causing considerable duplication of effort and outright waste, further diminishing the value of investment levels as a surrogate for military effectiveness.107

  Finally, as Marshall appreciated, there were major asymmetries in the two sides’ military objectives and the doctrines by which they hoped to achieve their objectives in time of war. These differences exerted a substantial influence on their respective logistics systems. The Soviets planned on achieving their objectives in a short campaign; thus their need for spare and replacement parts and major repair facilities was far less than those needed for the NATO forces, whose major weapon systems were generally more sophisticated than those of the Warsaw Pact forces and thus more difficult—and expensive—to maintain in an operational state. NATO forces also had to be prepared to defend against an attack for as long as it might take to prevail, and could not risk assuming the war would be short. Consequently Cohen concluded “The merits of NATO’s logistical and organizational practices require discussion, not an assumption that the West’s practices are invariably correct, and certainly not an increase of NATO firepower scores by fifty percent.”108

  Remarkably, the Optimists accorded relatively little attention to the role of air power in their assessments, even though air forces had played a major role in conventional warfare at least since the beginning of World War II. Cohen noted that Mearsheimer “simply excludes air power from his analysis.”109 Posen also omitted air forces, asserting that NATO’s were superior to the Warsaw Pact’s. His justification was that because he had assessed NATO ground forces as “fully competitive” with the Warsaw Pact’s, adding aircraft to the assessment would only reinforce his judgment that the European conventional balance did not favor the Soviets.110

  Cohen responded that a true net assessment of the balance would reject the Optimists’ notion that air forces are just another input into the combat power layered atop ground forces. In particular he cited Soviet air doctrine regarding the “initial air operation” at the beginning of a war.111 In fact, the Soviets had concluded by the mid-1980s that NATO air forces constituted the greatest threat to the success of a Warsaw Pact ground offensive in a conventional war in Central Europe.112 Hence the Soviets developed the theater, or strategic, air operation, which called for massive attacks not on NATO’s front-line ground forces but on its air forces. They planned to employ tactical ballistic missiles to attack NATO air bases within minutes of the onset of war, so as to pin down the allies’ aircraft until Warsaw Pact attack aircraft arrived. These attacks would also be supported by chemical weapons attacks and Soviet special forces raids to finish the job of destroying most of NATO’s air forces on the ground. Although NATO fielded a network of surface-to-air missile batteries, there were concerns that in the chaos of war, these forces were as likely to engage NATO aircraft as they were the enemy’s. Given that NATO commanders viewed Soviet success in the initial air operation as a “recipe for disaster” in terms of their ability to blunt a Warsaw Pact offensive, Cohen’s criticism of the Optimists’ discounting of this aspect of the European balance was certainly justified.113

  Coming back to theater-wide ADE ratios, a debate ensued over what kind of a force ratio advantage was required for either side to achieve its war objectives. While Epstein and Posen relied to a significant extent on models designed to simulate the dynamic interactions between NATO and Warsaw Pact forces, Mearsheimer eschewed models and focused on the ratio of the forces measured in terms of ADEs. He concluded that while “it is difficult to say precisely what ratio represents a decisive advantage for an attacker . . . the defender would probably be in serious trouble if the overall bal
ance of ADEs in a theater of the size and geography of Central Europe favored the attacker by 2:1.” Mearsheimer went on to assert that as a rule of thumb: “A theater-wide balance of 3:1 or more would almost surely mean rapid defeat for the defender. The attacker could then easily concentrate forces along various breakthrough sectors and simply steamroll over the defender’s forces, regardless of how well the defender fought.”114 Yet as Marshall knew, historical data from some 571 land battles reviewed by the military analyst Trevor Dupuy indicated that force ratios “do not make any significant difference” in which side ultimately won.115

  Mearsheimer’s colleagues Joshua Epstein and Barry Posen also rejected his use of the 3:1 rule. Unlike Mearsheimer, however, they employed models as part of their methodology. As Epstein explained: “For purposes of theater-level net assessment, the most revealing way to evaluate the conventional balance is through dynamic analysis. In a thorough dynamic analysis, an explicit mathematical model—a formal idealization—of warfare is used to simulate the mutual attrition of engaged forces, the flow of reinforcing units into the battle, and the movement of battlefronts over time” [emphasis in the original].116

  Epstein, like Marshall, expressed skepticism about the Pentagon’s models for conducting dynamic assessments. He observed that the British polymath and engineer “Frederick William Lanchester’s mutual attrition equations are the core of virtually all dynamic models used by the Pentagon, its contractors, and prominent independent analysts.”117 Lanchester’s so-called N-Square law, which he derived from gunfire duels between opposing lines of World War I era capital ships, postulated that outcomes would be in response to the casualties incurred over time.118 Yet the evidence shows that battles have been resolved largely on the basis of other considerations: “No matter how casualties are measured, battles have been given up as lost when casualties ranged from insignificant to overwhelming.”119 As with force ratios, the results of historical battles do not support casualty rates as a driver of outcomes. Epstein deemed this form of modeling “absurd” for several reasons and had developed his Adaptive Dynamic Model to help overcome these and other shortcomings of the Lanchester theory.

  Epstein’s Adaptive Dynamic Model, however, had its owns problems, as two other members of St. Andrew’s Prep, Jim Roche and Barry Watts, pointed out in a 1990 article published in the Journal of Strategic Studies. In it they concluded that underlying Epstein’s model were two assumptions. First, that changes in combat outputs are directly proportional to changes in inputs. In other words, the relation between inputs and outputs in war is linear. Second, Epstein assumed that the overall dynamics of a battle or campaign is the sum of the dynamics of its various components, which is to say that warfare does not include emergent phenomena. Roche and Watts found that “even the most cursory inspection of the equations underlying Epstein’s Adaptive Dynamic Model reveals his ‘calculus of conventional war’ embraces both these assumptions. . . . ”120 They argued that in reality war is fundamentally and relentlessly a nonlinear enterprise.* To back this up, they looked to Marshall’s emphasis on history and data, reviewing the disconnects between inputs and outputs in the strategic bombing of Germany during 1943–1945 and the US B-25 raid on the Japanese home islands led by Colonel Jimmy Doolittle in April 1942. The latter was particularly striking in that the immediate damage the Doolittle raid had inflicted on Japanese targets was negligible, but the fact that the sacred soil of home islands had been attacked led the Japanese to make subsequent strategic decisions that culminated in their defeat at the June 1942 Battle of Midway, where the Imperial Navy lost four fleet carriers in exchange for a single American flat top.121

  Epstein’s model attempted to capture the dynamics of a hypothetical situation in which NATO forces were attempting to trade space for time so as to improve their position. Yet the model failed to account for what Andrew Krepinevich experienced when, as an Army staff officer, he was temporarily assigned to NATO’s Northern Army Group in 1977 for a field exercise.† As the exercise progressed he found NATO’s ground forces trading space for time—all of them, that is, except for the three West German corps, which were interspersed between the other national corps along the alliance’s “layer cake”‡ defense. Owing to political considerations, the Germans felt they could not embrace a military doctrine that called for ceding territory—German cities, towns, and fellow citizens—to the Soviets in the hope that it would eventually be retaken. Consequently, not long into the war game NATO’s front line showed a series of undulations comprising large salients where the allied corps had conducted a fighting withdrawal, trading space for time, save for the three German corps, which had held fast to protect their land and countrymen.

  The result of the battlefield decisions Krepinevich anticipated was nothing like what might be expected from Epstein and Posen’s models, which implicitly assumed that NATO’s members would act “rationally” to optimize their military effectiveness, when in fact bureaucratic and organizational forces were very much involved in the outcome—a conclusion Marshall had reached thirty years earlier in reviewing his RAND colleagues’ projections as to where the Soviets would locate the bases for their bomber forces. Roche and Watts captured the point well in pointing out that “Human involvement alone argues that combat interactions and processes cannot be universally linear, that effects can be all out of proportion to their causes. As a result, the measures and analyses by which we attempt to deal with such complex interactions . . . cannot be . . . adequately captured by explicit, linear mathematical formulas any more than can chaotic dynamic systems.”122 In summing up their thoughts, Watts and Roche brought their mentor Marshall explicitly into the discussion, noting his “concern . . . that as the stockpile of quantitative measures and models has grown, it has become easier to deal with whatever analytic problems happen to occur simply by pulling old measures and models off the shelf and applying them with little or no thought as to their applicability or appropriateness.”123

  While Marshall applauded the efforts of such scholars as Epstein, Mearsheimer, and Posen to assess the European balance, he was not impressed with their results. He had decided long ago, in writing his master’s thesis on Klein’s economic model, that analytic measures and models, no matter how sophisticated they might appear to be, or how much effort and talent had been committed in their construction, did not overcome the simple fact that, as he had written in 1966, “The conceptual problems in constructing an adequate or useful measure of military power have not yet been faced. Defining an adequate measure looks hard, and making estimates in real situations looks even harder.”124 Epstein, Mearsheimer, and Posen had each come up against exactly the sorts of difficulties Marshall and members of St. Andrew’s Prep had become attuned to, and their attempts at net assessments fell short of ONA’s Cold War balances.

  In little more than a year after this debate, the Soviet Union’s position in Eastern Europe began gradually to collapse, bringing with it an end to the four-decade Cold War. In a December 1988 speech to the United Nations, Mikhail Gorbachev announced his decision to reduce the USSR’s armed forces by 500,000 personnel within two years, withdraw and disband six tank divisions then stationed in Eastern Europe, and cut Soviet forces there by 50,000 personnel and 5,000 tanks.125 The following November the Berlin Wall came down, and in December 1991 the Soviet Union itself collapsed. These developments were part of a rapid succession of world-changing events that obviated the original impetus for competitive strategies against the USSR and left the United States standing alone as the world’s preeminent power.

  Although the economic and political disintegration of the Soviet Union into fifteen separate countries brought the United States a respite from years of rivalry with a major adversary, it did not bring an end to such competitions. New security challenges would emerge in the years following the Cold War. Remarkably, they would be along the lines of what had been predicted by Marshall years earlier.

  Marshall was by no means egotistical enough to suggest, a
s some did after the fact, that he had foreseen the USSR’s demise in 1991. To be sure, he had felt for some years that the Soviet economy was headed for bankruptcy, but he did not predict precisely when or how the Soviet Union would unravel. When the end finally came in the same year as the 1991 Persian Gulf War, what surprised him was its rapidity and completeness.

  Granted, many elements contributed to the USSR’s collapse. One was surely the spreading loss of confidence among Soviet elites in the future that Soviet experts, such as James Billington and Peter Reddaway, began to sense in the early 1980s—a spiritual malaise that the 1986 nuclear disaster at Chernobyl only reinforced.126 Still, Gorbachev was probably right to complain that US policies, especially Reagan’s SDI initiative, sought to “exhaust the Soviet Union economically.”127

  The Reagan defense buildup did indeed put pressure on the Soviets, but to begin to understand how much one needs to appreciate the magnitude of military burden that the Soviet’s own choices had imposed on the USSR’s economy. Marshall’s insights into the true burden enabled him to provide Weinberger, Iklé, and many others with a more accurate and nuanced assessment of how the long-term competition with the Soviets was going and whether deterrence was likely to hold.

  Looking ahead, the 1980s proved to be a period in which Marshall’s office made great strides in understanding Soviet assessments and anticipating the likelihood of a disruptive shift in the military competition based on the marriage of precision munitions with wide-area surveillance and automated command and control. In many respects these insights flowed directly from Marshall’s willingness to relentlessly challenge the Pentagon’s bureaucracy and its conventional, but often self-serving, wisdom. Combined with his own lifelong quest to better understand how the world really worked, these intellectual inclinations are what made him and ONA so valuable during the latter decades of the Cold War—and what would make these same proclivities so vitally important during the tumultuous security transition that followed its end.