Stephen Peter Rosen, Winning the Next War (Cornell University Press: Ithaca, NY, 1991)

Author. ( - )
Harvard professor of national security and military affairs; taught strategy at USN War College
Director of political-military affairs for Reagan National Security Council

End of Cold War, Rise of Eastern states (China, Japan, India)

Explaining how military institutions innovate in peace, war, and technology

Innovation does
not depend on civilians
…or the enemy.


§ Military organizations resist change - resist outside intervention & suppress mavericks

• Innovation occurs slowly over time

§ Peacetime innovation – caused by ideological struggles

• Long time available --> less urgency; outside intervention effective to extent supports/protects innovators; mavericks – may increase internal resistance (ex: Billy Mitchell); new theory of victory & new paths for promotion; size of defense budget not a significant factor

§ Wartime innovation – caused by success/failure in combat

• Time constraints --> urgency; requires time – more innovation in long wars; requires feedback – combat assessment; intelligence on enemy; requires new measures of effectiveness; centralized (more direction/focus) vs. decentralized (more buy-in)

§ Technology innovation – need to manage uncertainty

• Technology push vs. requirements pull

• More than rational evaluation of relative utility: limited by intelligence about enemy’s current & future actions and limited by uncertainty of efficacy of new tech and total R&D costs

• Importance of simulation

Central Proposition
250 – The overall picture of American military research and development in the period from 1930 to 1955 is one of technological innovation largely unaffected by the activities of potential enemies, a rather self-contained process in which actions and actors within the military establishment were the main determinants of innovation.

Other Major Propositions.
20 – Control over the promotion of officers is the source of power in the military.
21 – peacetime military innovation occurs when respected senior military officers formulate a strategy for innovation, which has both intellectual and organizational components. Civilian intervention is effective to the extent that it can support or protect these officers.
35 – When military innovation is required in wartime, however, it is because an inappropriate strategic goal is being pursued, or because the relationship between military operations and that goal has been misunderstood.
57 – Peacetime military innovation in the United States has, in fact, proceeded remarkably independent of intelligence about foreign military powers. Instead, perceptions of change in the structure of the international security environment have been the source of such innovation.
128 – The intellectual and organizational changes necessary to evaluate new ways of fighting are as important as the development and production of new technologies.
147 – assessing the relative strategic effectiveness of an innovation can be important not only for the promotion of the innovation, but for the proper formulation of larger strategy.
199 – It must be emphasized that the intelligence relevant to technological innovation was not only that which concerned the technical parameters of enemy equipment. An understanding of the enemy’s organizational habits also led to the identification of vulnerabilities.
228 – When organizations were made responsible for both research and procurement, research would suffer, and innovations would be slow in coming.
243 – uncertainties about the enemy and about the costs and benefits of new technologies make it impossible to identify the single best route to innovation.
252 – Rather than money, talented military personnel, time, and information have been the key resources for innovation.
253 – intelligence about the behavior and capabilities of the enemy has been only loosely connected to American military innovation.
255 – They decide when to go to war, and determine overall level of military budgets. … Civilian political leaders, however, do not appear to have had a major role in deciding which new military capabilities to develop.
256 – Independent civilian scientists have clearly been important in stimulating technological innovation by pushing back the frontiers of knowledge … they did not play a major policy role in pushing the American military to develop capabilities it otherwise would have neglected
257 – Thinking about peacetime military innovation requires a focus on the next 20-30 years.
259 – Research and development of a range of new military technologies, and mobilization plans for their mass production, rather than large-scale procurement of any one system, would appear to be the best way of managing uncertainty. … A strategy that would prepare military innovations for this new world had to focus on the management of uncertainty, rather than on the construction of new capabilities tailored to predictions of what future wars would look like.

Chapter 1 – Thinking about Innovation
2 – Almost everything we know in theory about large bureaucracies suggests not only that they are hard to change, but that they are designed not to change.
4/5 – No good explanation of bureaucratic innovation exists. There are only contradictory results from different studies. … different kinds of innovation occur for different reasons in the same organization, and different organizations will handle innovation very differently … it is unlikely that explanations of innovation will have universal applicability
9 – defeat in war is neither necessary nor sufficient to produce innovation
10 – Richard Neustadt has listed five condition that must prevail if a president’s order is to be readily obeyed by his bureaucratic subordinates. The president himself must be clearly involved in the decision, and his order must be unambiguous. His order must be widely publicized, and “the men who receive it [must have] control over everything needed to carry it out,” and they must have no doubt of his “authority to issue” the order.
11 – civilian command to carry out military innovation is by its nature extremely difficult to enforce
22 – The problem of wartime innovation thus becomes one of the extent to which organizational learning can take place under the unique conditions of war … old and innovative methods can be tested in combat and compared.
35 – the definition of the strategic goal, the relationship of military operations to that goal, and indicators of how well operations are proceeding can be thought of as a strategic measure of effectiveness for the military organization.
35/6 – Inappropriate strategic measures of effectiveness may lead an organization to mistakenly increase its efforts, in a vicious cycle, at a time when increasing the effort put into old methods only draws the organization deeper into failure.
28 – Clausewitz may have been more correct then Sun Tzu. Intelligence relevant to innovation very likely will not be available in wartime, and wartime innovation is likely to be limited in its impact. … the longer the war lasts, the greater the impact of wartime innovation. … the more hierarchical and centralized the organization, the greater the impact of the innovation
39 – Decentralization would seem to favor innovation in those circumstances in which the operating units can collect al the relevant data themselves and can execute the innovation without the need for organizational changes elsewhere in their service.
44 – Economic analyses of military research and development have implicitly highlighted the role of intelligence in innovation. … both perceived future value and the relative urgency with which a weapon is needed … are heavily influenced by intelligence about the enemy
45 – Knowledge about enemy military technology is one kind of intelligence relevant to our choices concerning military R&D, but intelligence about how the enemy will conduct himself is also important.
46 – there does not seem to be any such unambiguous measure of the output of military technological innovation
51 – A strategy for choosing new military technologies … has to take account of an environment in which it is extremely difficult to make any conclusive analysis of the prospective cost and utility of alternative research and development programs.

Chapter 2 – The Shape of Wars to Come: Analyzing the Need for Peacetime Innovation
58 – Because of the time necessary for young officers to be promoted to senior rank, the practical side of the innovation typically took a generation to accomplish.
60 – farsighted peacetime military innovation has been possible … when military budgets were tight … and even when the military bureaucracy had swollen in size
75 – military planners were driven to consider the need for innovation by broad structural changes in the security environment in which their organizations would have to fight for the foreseeable future, not by specific capabilities or intentions of potential adversaries … simulations were thus useful in thinking through the shape of potential innovations … Once the need for an innovation was perceived, and the character of the new capabilities conceptualized … the behavior of the organization, as it prepared for and conducted war, had to change if the innovation was to be actualized.

Chapter 3 – Making Things Happen: The Politics of Peacetime Innovation
76 – innovations occurred when senior military officers were convinced that structural changes in the security environment had created the need … they did so by creating new career paths along which younger officers specializing in the new tasks could be promoted.
93 – Rethinking doctrine is a much more difficult task than simply learning how to better apply a given concept to a particular circumstance
105 – Peacetime innovations are possible, but the process is long.

Chapter 4 – The British Army and the Tank, 1914-1918
109 – Combat provides, rather than speculative arguments, clear justification for jettisoning old ways of fighting and adopting new ones.
114 – Learning from experience begins by defining a problem to be solved or identifying a measure of effectiveness with which alternatives could be judged.
128 – The delay in the successful implementation of this wartime innovation reflected not a failure to try the new technology but a failure, or rather a slowness, of organizational learning. This slowness, in turn, was directly related to the problems of defining a new strategic measure of effectiveness, of utilizing available information to evaluate the innovation, and the absence of tight central controls to ensure the implementation.

Chapter 5 – New Blood for the Submarine Force
130 – Though the technological capabilities necessary for certain innovations had emerged in the interwar period, concepts of operation that might have exploited the technologies did not emerge because they ran counter to basic principles of American foreign policy.
138 – individual learning and organizational learning are two different things … Organizational failure did not lead immediately to organizational learning and innovation.
139 – The pattern in both world wars was for significant turnover among the officers with field commands, but relative stability in the high command. Because operational doctrine is usually made by the high command, turnover in the officer corps in war need not be associated with changes in operational behavior.
141 – What was unique about the submarine force was that the autonomy of command at the lower levels meant that turnover at that level did lead directly to fundamental operational changes.
142 – Submarine skippers were simply told to sink ships, any kind of ships, or they would be relieved of command. … This yielded an increase in merchant ship sinkings. Innovation was decentralized. Younger officers made a new kind of war on their own.

Chapter 6 – The United States Strategic Bombing Force, 1941-1945
149 – The ability to direct strategic bombers against appropriate targets is no less a military capability than the ability to aim artillery accurately.
149 – a military catastrophe might force organizational responses even without a well-developed intellectual construct that measures relative strategic effectiveness
157 – In the absence of a strategic measure of effectiveness, the makers of AWPD-1 did not recommend on the basis of an analytical effort that compared the impact of bombing military forces, various industrial complexes, transportation systems, or producers of primary products. The emphases of the first target plan appear in large measure to have resulted from the accidental and selective availability of intelligence about Germany. (That and many hours or intensive mathematical Operations Analysis)
157/8 – Prewar civilian data about selected industries, rather than a systematic understanding of German industries, their relation to the overall German economy, and the relation of the economy to the German capacity to make war, appear to have had considerable impact on the target planning.
159 – initial British and American efforts to perform a new function, strategic target planning, were not successful because they did not define the way in which strategic bombing was to affect the outcome of the war … neither had a strategic measure of effectiveness that defined how military operations were related to strategic objectives and indicated how well those operations were progressing
169 – Wartime innovation by the OSS did produce a clear analytical mechanism for selecting targets for the strategic bombers. … But the absence of good data, either because of mirror imaging (projecting one’s own ideas onto the enemy), deceptive enemy practices, or the inappropriateness of collection techniques, meant that the selection of targets tended not to be guided by realities about the German economy.
175 – The enemy is an active force that reacts, but not always in the most likely ways, not always, even, in the ways most advantageous to himself.
176 – Germany could have made the American use of drop tanks in Europe in 1944 impossible by sending up German fighters to challenge the escort fighters over the English Channel.
177 – the hardware for drop tanks had existed before the war but had been rejected because it seemed inappropriate given observations of the enemy. Initial wartime learning, in other words, inhibited innovation. … Drop tanks worked because German fighters were pulled back from France and the Netherlands in order to protect central Germany.
178 – An innovation implemented despite what was known of enemy countermeasures led to changes in enemy behavior that made the innovation successful.
179-81 – New strategic measures of effectiveness played a significant role in three of these innovations but were decisive in only one. … The structural conditions related to intelligence and command relations do not seem to be decisive in explaining the initiation or effectiveness of wartime innovations. … Learning from wartime experience how to perform an entirely new military function was in all cases extremely difficult. … The need for time will make even the most ardent supporter of innovation skeptical about innovation in wartime. Solutions that come after the enemy has triumphed are of no value at all.

Chapter 7 – What is the Enemy Building?
187 – many of the key decisions about new land warfare systems in the interwar period and about the first generation of long-range offensive nuclear weapons appear to have been made with surprisingly little intelligence input.
191 – Electronics warfare is a field in which intelligence is particularly important to innovation.
206 – Knowledge of Soviet air defenses was necessary to evaluate the relative capabilities of new manned bombers, cruise missiles, and ballistic missiles to penetrate to their targets.
209 – in 1950 the Weapon System Evaluation Group found itself severely constrained in its ability to evaluate new weapons technology because of a lack of intelligence data.
213 – RAND Memorandum on 8 Feb 1954 (Bruno Augentein): “the prosecution of our own ballistic missile program is a race against time.”
214 – it is reasonable to infer that at the beginning of 1954 the Americans responsible for development of strategic offensive weapons had reason to believe that the Soviets were working seriously on ballistic missiles, and that their programs were almost certainly as competent as those in the United States.
219 – Robert McNamara (1961): “the blunt fact remains that if we had more accurate information about planned Soviet strategic forces, we simply would not have needed to build as large a nuclear arsenal as we have today.”
220 – in several significant areas American military development was not shaped by intelligence about the enemy’s technology developments.

Chapter 8 – Strategies for Managing Uncertainty
221 – research and development strategies in the real world tended to become strategies for managing uncertainties
222 – Invention was the creation of a new idea for a weapon, innovation was the choice of which new ideas to develop. Knorr and Morgenstern argued, vaguely, that the military had done neither very well, while scientists had.
224 – Vannevar Bush (April 1942): “The planning of military strategy needs to be carried on with a full grasp of the implications of new weapons, and also of the probable future trends of development.”
228 – research and procurement are incompatible
230 – Introducing civilian scientists into the picture did not remove or neutralize the pressures created by procurement needs.
231 – The Air Force resisted the ICBM because it was “committed to manned aircraft, and particularly to manned bombers, and refused to change” … The bureaucratic imperative to preserve existing missions and ways of operating tends to crush the impulse to make technological innovations.
235 – Demonstrable results were the critical factor, not the origin of the technology. … Uncertainties about the value and costs of technological research are a universal problem, not confined to military or civilian researchers and managers.
241/2 – the conventional picture of services dominated by the pilots of manned aircraft who would have strangled unmanned missiles programs were it not for the intervention of civilian scientists … Hap Arnold led the move away from manned aircraft … in September 1946, LeMay … also emphasized the need to substitute missiles for manned bombers … Guided missiles carrying nuclear warheads, if feasible, could replace manned bombers in attacking the most difficult targets in the initial phase of war. The scarcity of nuclear weapons made necessary a commitment to both manned and unmanned strategic systems.
243 – If civilian scientists did not have a clear and inherent advantage in the management of military technological innovation, what alternative strategies were there? … build anything that looked promising, and then try it out … While feasible if a nation is rich enough, the strategy has less appeal when hard choices have to be made. … A strategy for military technological innovation that seeks as much flexibility as it can buy might be better than one trying to buy the one weapon that would perform the best if it could be built to specifications at the expected cost and if it eventually turned out to be the weapon which was actually needed.
244 – Type I flexibility—weapons that will be useful in every possible contingency
Type II flexibility—manage uncertainty by buying prototypes and defer large-scale production
246/7 – The possession of scientific knowledge and engineering techniques prove more valuable in meeting the exigencies of a long war than a large stockpile of obsolescent missiles
247 – the critical even in the development of ICBMs was the development of the hydrogen bomb
248 – It was not a Soviet threat, or a civilian scientific intervention in the context of fixed technological possibilities that pushed the innovation of the ICBM, but a new and unforeseen technological innovation created by civilian physicists.

Chapter 9 – Conclusion (see Major Propositions)
260 – Four concrete implications for managing uncertainty:
1) Strategic cultures are not like strategic plans. They are the result of political and cultural history and tend to be relatively stable over time. The study of these cultures would be inexpensive and could reduce our uncertainties about how these countries could use … power
2) language training!
3) construct a wide range of scenarios and conduct imaginative conflict simulations in order to explore the shape of potential wars.
4) Years may not be available when a new threat clearly emerges. A range of new capabilities needs to be thought through to provide alternatives from which policy makers can choose.

Community content is available under CC-BY-SA unless otherwise noted.