U.S. Air Force Strategic Deterrence Analytic Capabilities: An Assessment of Tools, Methods, and Approaches for the 21st Century Security Environment (2014)

Summary

The United States developed and used nuclear weapons in the Second World War and, since the surrender of Japan, has maintained a nuclear capability to deter and influence the behavior of adversaries and assure allies. Over time, geopolitical developments have transformed what started as a bipolar world order after that war, which involved the United States and the Soviet Union and their respective allies, into the current multinodal global reality, in which nonstate and state actors play an important role. Since the early 1960s, the U.S. strategic nuclear posture has been composed of a triad of nuclear-certified long-range bombers, intercontinental ballistic missiles, and submarine-launched ballistic missiles. Also, since the early 1970s, U.S. nuclear forces have been subject to strategic arms control agreements. The large numbers and diversified nature of the U.S. nonstrategic (tactical) nuclear forces, which cannot be ignored as part of the overall nuclear deterrent, have decreased substantially since the Cold War. While there is domestic consensus today on the need to maintain an effective deterrent, there is no consensus on precisely what that requires, especially in a changing geopolitical environment and with continued reductions in nuclear arms. This places a premium on having the best possible analytic tools, methods, and approaches for understanding how nuclear deterrence and assurance work, how they might fail, and how failure can be averted by U.S. nuclear forces.

In a 2013 speech following negotiations for the New Strategic Arms Reduction Treaty (entry into force: February 5, 2011), President Obama took a further step and announced that the United States had “determined that we can ensure the security of America and our allies, and maintain a strong and credible strategic deter-

rent, while reducing our deployed strategic nuclear weapons by up to one-third…” and that he intended to “seek negotiated cuts with Russia to move beyond Cold War nuclear postures.”¹ President Obama’s announcement carried with it a series of complex conceptual and analytic challenges. For example, if nuclear weapons are to take a lesser role in U.S. security strategy, what role should that be? In which circumstances is it reasonable and credible to pose a nuclear threat? At lower levels of deployed nuclear weapons, which systems and postures are essential for maintaining a strong deterrent to attack by both known and unforeseen adversaries? What should be cut, and how is this to be done without causing harm to strategic stability in multiple areas of the globe? Which nuclear capabilities, if any, are needed to assure allies of U.S. commitment to their security? Each question is made all the more challenging when considered in light of U.S. fiscal austerity, global power shifts, and other changes currently under way in the international environment.

STUDY APPROACH AND CAVEATS

While this study of the Committee on U.S. Air Force Strategic Deterrence Military Capabilities in the 21st Century Security Environment was mutually requested by the Deputy Assistant Secretary of the Air Force for Science, Technology, and Engineering and the Assistant Chief of Staff of the Air Force for Strategic Deterrence and Nuclear Integration, the results are intended to inform the Air Force research enterprise as a whole, as well as the larger audience of stakeholders involved in issues of deterrence and assurance generally and nuclear deterrence and assurance in particular. During this study of analytic tools, methods, and approaches for strategic deterrence and assurance of adversaries and allies, it became apparent that no single tool, method, or approach could address the array of deterrence and assurance challenges the Air Force and the nation will face in coming years. It also became evident that there is a critical deficit in the Air Force capacity to sustain high-quality analysis in support of its newly broadened nuclear deterrence and assurance responsibilities. Namely, the Air Force lacks a means for organizing and ensuring the training necessary to build a cadre of methodologically savvy analysts conversant in nuclear deterrence and assurance issues. Simply put, regardless of the analytic tools it possesses, the Air Force has too few people with the personal experience and rigorous analytic training required to generate the analyses necessary to determine the nuclear force structures and postures most likely to be effective deterrents. Before discussing the specific items in the terms of reference and where

________________

¹ Executive Office of the President, Transcript of remarks by President Obama at the Brandenburg Gate, Berlin, Germany, June 14, 2013, http://www.whitehouse.gov/photos-and-video/video/2013/06/19/president-obama-speaks-people-berlin#transcript.

and how this study addresses them, the study’s orientation to the issues and, thus, what is included and what is left out of this report are explained.

First, much of Cold War era deterrence theory and analysis assume a causal or nearly causal relationship between possessing massive physical power and being able to deter unfavorable actions. One of the results of this assumption has been that, until relatively recently, higher priority was given to developing tools, methods, and approaches for estimating physical effects of weapons than to the human perceptual aspects of deterrence and assurance.² Well-founded understanding of adversary and ally perceptions, motivations, and decision processes is a critical precondition for producing the types of analyses needed to support planning for nuclear capabilities relevant to assuring multiple actors across a variety of international circumstances. As a consequence, this study focuses on tools, methods, and approaches for understanding human behavior and does not address assessments of physical effects and capabilities.

Second, there are literally scores of analytic tools, methods, and approaches.³ It would be neither reasonable nor useful to conduct a comprehensive review of all of them. Instead, the study leveraged the substantial expertise of the committee membership, previous reviews, and numerous briefings and discussions in workshops and committee meetings to identify a set of appropriate tools, methods, and approaches and assess their general applicability to deterrence and assurance issues, as well as the type of analytic role (e.g., data generation, decision support) for which each tool, method, and approach is best suited. Relatedly, this report does not suggest either a single or a set of silver bullets for addressing the range of issues confronting the Air Force, and nowhere does the report imply or state that computers or checklists might replace the human intellect.

Third, the report is not limited to nuclear deterrence or assurance. Of note, there do not appear to be standard definitions of basic deterrence-related concepts within the U.S. national security community. Theoretically, if not doctrinally, assurance of allies, together with deterrence of adversaries from nuclear use, or deterrence of other activities by way of nuclear threats are at the far ends of a spectrum of influence activities that concern the U.S. defense establishment. Moreover, some argue that attempting to consider nuclear deterrence in isolation from

________________

² Official recognition of the importance to deterrence and assurance of understanding human decision processing and perceptions is illustrated in Joint Publication 1-02, which states that deterrence is “the prevention from action by fear of the consequences … a state of mind brought about by the existence of a credible threat of unacceptable counteraction;” and the 2006 Deterrence Operations Joint Operating Concept, where “deterrence” equates to “decisive influence over [adversaries’] decision making” by increasing the costs associated with taking an action and decreasing the rewards and costs of restraint.

³ National Research Council, 2013, U.S. Air Force Strategic Deterrence Capabilities in the 21st Century Security Environment: A Workshop Summary, The National Academies Press, Washington, D.C.

TABLE S-1 Items in the Terms of Reference (TOR) and Corresponding Recommendations

other deterrence considerations—to the degree that was possible during the Cold War—is increasingly difficult and likely to be shortsighted in the current security environment.

METHODOLOGY FOR RESPONDING TO THE TERMS OF REFERENCE

The recommendations discussed in this section are organized as they relate to the five items in the terms of reference (TOR). Briefly, these are (1) to identify key issues in 21st century deterrence and assurance analysis; (2) describe and assess analytic tools, methods, and approaches; (3) discuss how the Air Force could respond to deterrence and assurance needs, including suggesting an analytic framework; (4) suggest how the Air Force might evaluate and validate new tools, methods, and approaches; and (5) recommend specific classes of tools, methods, and approaches. All of the TOR are listed in the left-hand column of Table S-1.⁴ How and where each item of the TOR is addressed in the report are described in

________________

⁴ Appendix A also provides the TOR.

^a Chapter 4 provides suggestions for Air Force organizations that would have roles in implementing the report’s recommendations.

^b Air Force, 2013, Flight Plan for the Air Force Nuclear Enterprise, Washington, D.C. The TOR are contained in Appendix A.

^c B. Bragg, ed., 2011, Concepts and Analysis of Nuclear Strategy Framework Report, prepared by NSI, Inc., for the Strategic Multilayer Assessment Office, Department of Defense, http://nsiteam.com/publications/.

the right-hand column of Table S-1 and discussed in more detail in the next section. Supplemented by the discussions and examples provided in Chapters 2 and 3 and supporting appendixes, the individual recommendations should be read as aspects of an overarching theme of the report, which is the need for the Air Force to refocus and sustain its intellectual capital in the areas of deterrence and assurance in general and political understanding of nuclear issues in particular. Table S-2 provides a complete list of report observations, findings, and recommendations mapped against the TOR.

TABLE S-2 Complete List of Observations, Findings, and Recommendations

Key Issues in 21st Century Deterrence and Assurance Analysis

Item 1 of the terms of reference was addressed by extensive committee debate and by input from subject matter expert speakers in a variety of disciplines, ranging from the new and eclectic, such as neurodeterrence, which combines advances in neurobiology and study of deterrence and threat behaviors, to more familiar political and technical experts with decades of experience in arms control and management of the nuclear enterprise. The “broad analytic issues and factors” gleaned from these sessions appear throughout Chapter 2, which suggests and discusses three broad categories into which recommended themes fall (see Table S-3). Chapter 2 also lays out “stressful questions” associated with peer, near-peer, regional, and nonstate challenges, as well as important deterrence and assurance issues like nuclear command and control, force modernization, air and missile defense, and geostrategic and technological changes not directly addressed in this study.

Description and Assessment of Analytic Tools, Methods, and Approaches

The first component of Item 2 of the terms of reference—assessing tools, methods and approaches—was addressed in light of the issues identified in Chapter 2 as critical to 21st century deterrence and assurance analysis. A summary of reviewed methods and tools appears in Chapter 3 with further illustration in Appendixes D and E. Given the time limitation of this study, the second element of this item was not addressed. Understanding the psychological mechanisms that govern what deters and what assures are preconditions for assessment of the attributes of various nuclear systems, technological capabilities, postures, and concepts of operation of U.S. nuclear forces. Consequently, improving the Air Force’s capacity to account for and use the types of actor- and decision-unit-specific information needed to tailor deterrence and assurance messages and activities is a necessary requirement.

TABLE S-3 Focus Issues

Doing so will allow the Air Force to better calculate the specific regional capabilities it will need to provide to allow maximum flexibility to identify and influence activities likely to be most effective in present conditions and those it may face in the future. In addition, improved understanding of the human factors involved in deterrence and assurance situations may facilitate earlier recognition of potential deterrence or assurance failures.

The Air Force needs to plan now to contribute the capabilities required to deter and assure decades into the future. Further, the Air Force would be the obvious advocate for a U.S.-government-wide program to develop systematic, multidisciplinary generalized leadership and decision-making constructs and models to improve the robustness of that planning by anticipating the range of potential behaviors, consequences, and situations that may be faced. This will also provide a baseline set of regional deterrence and assurance environments that could help analysts assert how current and future leadership changes might affect the deterrence and assurance environment. Finally, the Air Force would ideally explore the notion of “deep uncertainty” in planning support analyses in order to expand analysts’ awareness of future uncertainties and the types of circumstances most prone to significant unintended consequences. ⁵

________________

⁵ Deep uncertainty refers to “materially important uncertainties that cannot be adequately treated as simple random processes and that cannot realistically be resolved at the time they come into play” (Paul K. Davis and James P. Kahan, 2007, Theory and Methods for Supporting High Level Military Decisionmaking, RAND Corp., Santa Monica, Calif., http://www.rand.org/pubs/technical_reports/TR422.html, p. 6).

Suggested Analytic Framework for Air Force Deterrence and Assurance Needs

A high-level deterrence and assurance task framework is presented in Chapter 4. Awareness of the web of complexities involved in managing the multitude of overlapping deterrence and assurance issues led the formulation of Recommendation 1. Namely, that the Air Force design and pursue a coordinated deterrence and assurance analysis program—something that does not currently exist—to guide its efforts. Recommendations 2, 3, and 4 refer to some of the attributes that a deterrence and assurance analysis program should have. In addition, the program might include tracks to refine and apply the psychologically based concepts at the heart of deterrence and assurance; to encourage practitioner-academic dialog to facilitate practitioner access to academic strategic studies on the one hand and educate academics on operational priorities and constraints in the military setting on the other; to institutionalize and integrate ongoing efforts across the Air Force, including the Air Force Global Strike Command, the Strategic Command (STRATCOM), and elsewhere; and, finally, to train a cadre of deterrence and assurance analysts conversant in multiple analytic methods and approaches.⁶

Such a program would benefit the Air Force directly by providing a guide for developing the types of robust analyses currently lacking, but necessary to underpin and defend Air Force capabilities. The recommended deterrence and assurance analysis program would also provide the means for coordinating and monitoring analytic projects across the Air Force, limiting both (1) costs associated with duplicate independent efforts and (2) overreliance on traditional deterrence metrics—for example, damage expectancies and comparative weapons counting, which are less relevant in the current security environment.

Air Force decision makers, analysts, and, most critically, consumers of deterrence and assurance analyses may also consider adopting the habit of considering the limitations and appropriate uses of any analytic tool, method, and approach, along with the results such use might generate. Like hammers, saws, and other carpenter’s tools, analytic methods have appropriate and inappropriate uses according to the nature of the task to which they are put. As with using hand tools to build a table, well-executed analysis of issues of human perception and behavior require integrated use of multiple methods and tools to produce robust and defensible results. Using tools properly and in tandem can significantly improve analytic breadth, accuracy, and insight. As such, the approach to deterrence and assurance analysis adopted by the Air Force would ideally include as a specific goal

________________

⁶ Officials at STRATCOM discussed their general awareness of efforts to reinvigorate deterrence thought and develop deterrence experts, including educational courses offered at the Air Force Nuclear Weapons Center, establishment of a Deterrence and Assurance Working Group, and a proposed nuclear fellows program.

the careful integration of analytic techniques. Combining methods in a planned and coordinated way can also help refine analyses over time.⁷

Methods for Air Force Evaluation and Validation of Tools, Methods, and Approaches

Rather than propose a static set of deterrence and assurance tools, the report identifies factors that might be used to guide a multiyear, multimethod research agenda. ⁸ Moreover, the task of providing a framework for “validating” tools became both less relevant and exponentially more complex with the report’s focus on tools, methods, and approaches tied to better understanding of the impact of perceptual factors as opposed to capability factors on deterrence and assurance.⁹ Where human behavior is the subject of concern, there are two types of validity that must be tested: internal and external. Internal validity refers to the internal logic of the model and the degree of confidence that it actually taps into and explains the underlying construct that the researcher intends—for example, the psychological mechanisms that account for decisions to forego benefits in light of costs and thus be deterred from taking an action. Implicit in this is that the model is a comprehensive representation of that construct. External validity refers to the degree to which a model or tool is applicable beyond the particular circumstance for which it was built—for example, whether a model explaining Russian decision making would also apply to China. The means of validity testing, or validation, however, vary according to the specific tool, method, or approach used. Thus, while

________________

⁷ Integrating analytic methods need not be a costly or onerous undertaking in order to produce valuable results. In many cases, the output of one approach fits perfectly into or can help frame the required input of another. For example, social network analysis can identify key decision makers who should be subject to leadership profiling and other decision analyses. Hypotheses regarding the strategic interactions of regional adversaries derived from game theoretic analysis and case studies can be further tested in series of human (war) games, and so on. Table 3-1 in Chapter 3 illustrates the general mixes of the methods reviewed for this study. For a thorough discussion of integrating multiple analytic techniques for deterrence analysis, see Office of the Secretary of Defense Multilayer Analysis Deterrence Subgroup A Report: Deterrence-Supporting Approaches and Comparative Analysis and Integration Recommendations, June 30, 2009.

⁸ It should be noted, however, that a nuclear force posture comparison modeling project undertaken by STRATCOM (J5) for the 2010 Nuclear Posture Review may serve as the core of a development project in this area.

⁹ It is important to note that, although the words used are often the same, tools, methods, and approaches for issues of human behavior are subject to different notions of what constitutes a model and tool “validation” than is typical in engineering and other scientific disciplines. These concepts can also be different from what analysts often mean by “validation.” For example, How well has the model performed in the past? or How much confidence should I have in what it tells me? are often what analysts (as opposed to model builders) mean when they refer to “validating” an analytic model, framework, or tool.

no general framework for validation is suggested in this report, where appropriate, these issues are treated in the reviews of methods presented in Chapter 3 and Appendixes D and E.

Recommending Specific Classes of Tools, Methods, and Approaches

As noted above, the number and variety of analytic tools, methods, and approaches is enormous. Each of those reviewed for this report is relatively mature and accessible to the Air Force, if not directly to analysts, then via experts and companies that can easily be found to apply them. The complexity of planning and analysis for nuclear deterrence and assurance that will confront current and subsequent generations is likely to continue to increase exponentially. Paradoxically, rapid advances in communications technologies means that conveying deterrence and assurance messages will become increasingly difficult to control as counter-communications are easier to issue and perceived U.S. intentions become subject to literally global interpretation. The relative lack of exposure of many of today’s analysts to nuclear-related issues may make it premature for the Air Force to consider significant investment in classes of tools, methods, and approaches and certainly, in particular, in tools needed to conduct deterrence analyses now and into the future. Instead, the Air Force would do well to focus on its people first. This will ensure that Air Force personnel are able to provide the most credible and analytically based perspectives in both Air Force and joint decision fora, and that the Air Force is able to provide leaders with informed and reliable reviews and critiques of alternative force structures, sizing, and deployment options.