During the Cold War, strategists used applied math for insights into nuclear strategy. Their techniques included game theory, macroeconomics, and systems analysis. These models, brilliant as their creators were, had flaws. They led to equations that could be studied in fascinating detail, yielding great academic “publish or perish” rewards and even some Nobel Prizes, but had little to do with realpolitik and leaders who are willing to take big risks. With today’s “three body problem,” over-relying on such math could mislead again.
Flawed Modeling Assumptions
Game theory finds ideal tactics for conflicts between two opponents. It works great for something simple like checkers. But to get far it needs assumptions that, together, weaken the approach for something as tricky as convincing an adversary to never coerce or attack with nuclear arms.
In its basic form, game theory assumes both players have the same goals and want to avoid the same downsides. It also assumes all relevant information for the game is known fully and equally by both players, who are unbiased, unhurried, and perform all calculations perfectly. It assumes the game has a clear beginning and end, and moves are made in an orderly one-goes-then-the-other-goes manner.
Plainly, none of these apply well to states or blocs in a nuclear crisis or an escalating conventional shooting war. More subtly, game theory does not take account of tacit cooperation between opponents—such as to avoid nuclear Armageddon.
Modern game theorists are starting to study games with many players, in which different players have different goals and different information. But this research needs to mature more and be validated rigorously before it can be trusted enough to guide national defense.
Macroeconomics is a tool for understanding and managing the workings of a national economy. In its traditional form, macroeconomics assumes that everybody decides to buy or sell things based only on their price, and that all decisions are made with one hundred percent efficiency using complete information.
The difficulty of controlling a real-world economy is shown by the problems of fluctuating American inflation rates, recessions, and unemployment. While the conceptual framework of macroeconomics was adapted to analyze nuclear deterrence decades ago, there are practical limits to policy guidance obtained this way.
Modern research turned to what is called behavioral economics. This approach pays attention to the emotional reasons people do things. It also considers that different people seek different real and emotional rewards. Even sophisticated actors are driven by an irrational perception of risk. This emerging discipline seems promising, but still needs testing.
Systems analysis studies a problem, such as how to win a war, by breaking the warfighting into moving parts, then analyzes how those parts interact. The goal is to create statistics-driven procedures, such as body counts, that will achieve the desired outcome—victory.
Systems analysis has flaws for national defense policymaking. To get anywhere, it needs to make very difficult choices about exactly how to measure effectiveness, how to handle incomplete or absent data, how to take account of fundamentals and intangibles such as political will, and how to remove analyst bias. The near impossibility of doing all this successfully was demonstrated by the failure of systems analysis in the Vietnam Conflict.
Modern Threat Assessment
During the Cold War, other practical drawbacks of these technical tools were masked by the fact that the only major players of the nuclear deterrence game were the US and USSR. Neither became so aggressive or desperate as to resort to a nuclear attack. But there were close calls, such as the Cuban Missile Crisis, peacefully resolved by back-channel bargaining between strong-willed leaders. Details of the next nuclear crisis, if one occurs, will be totally different and difficult to model realistically in advance.
Today, there are several significant nuclear adversaries confronting the US and its allies. The chances seem high that equations cannot capture the many psychological subtleties and unknown-unknown interactions, especially when several authoritarian regimes can form an opaque axis of expansion.
A good way to test the utility of behavioral economics and modernized game theory is to see if it can yield insights on dealing better with enemies who use nuclear threats against the United States, such as to limit American support for a beleaguered ally. The cases of Ukraine and Taiwan come to mind.
Conclusion
It is not just an academic exercise to confront and prevent the possibility that Russia, China, and North Korea may take their nuclear brinksmanship beyond mere verbal threats and saber-rattling exercises. Arms control advocates and defense policymakers need to recognize that nuclear attack is not simply a verbal bargaining chip thrown around by foreign potentates who are unserious, bluffing, or who have purely defensive goals. There is no pause button. There are no do-overs.
Nuclear deterrence cannot be reduced to a blackboard puzzle. Doing it properly needs undivided attention to the subtle nuances and fine distinctions that can make all the difference in an outcome. It calls for all-out political commitment despite many daunting complexities. Above all, effective deterrence requires deep understanding of how to make adversaries feel the raw fear generated when what they value most is at risk.
Joe Buff is a risk-mitigation actuary researching modern nuclear deterrence and arms control.
Aaron Holland is an Analyst at the National Institute for Deterrence Studies. The views expressed in this article are the authors’ own.
About the Author
Aaron Holland
Aaron Holland is an Analyst at the National Institute for Deterrence Studies.
Joe Buff
JOE BUFF, home office cell number 518-719-5284, has long experience researching practical applications in math and actuarial science, management consulting, and national defense. His award-winning publications include six best-selling submarine technothrillers (the Captain Jeffrey Fuller series), and numerous articles and op-eds about undersea warfare, geostrategy, and nuclear deterrence in The Submarine Review, American Submariner, the U.S. Naval Institute Proceedings, The Day (New London) and on Military.com’s DefenseTech blog. Joe has presented to the Naval Submarine League, the U.S. Submarine Veterans, the U.S. Submarine Museum (Groton) and the New York State Military Museum (Albany). In early 2024 Joe was a guest on ANWA’s NucleCast, hosted by Dr. Adam Lowther, and he has begun authoring and coauthoring articles for the National Institute for Deterrence Studies’s Global Security Review.
Employment History:
1997-Today: Joe Buff Inc., “a defense think tank of one.” Issued Best Practices reports On 21st Century Nuclear Deterrence – Volume I (2021, 420 pp.) and Deep Space Deployments: A Durable U.S. Triad’s Final Frontier? (2022, 115 pp.). The former’s Volume II is now in progress. Senior Fellow at The National Institute for Deterrence Studies.
1996-1997: Merrill Lynch, research associate and then director in the Insurance Strategies Group. Developed methodologies, spoke and published, and consulted with Merrill’s life insurer clients about investment risk control and optimum asset/liability portfolio positioning.
1987-1995: Towers Perrin, associate and then principal (junior partner) in this top-10 global actuarial management consulting firm’s Tillinghast division. Researched, spoke, wrote, and consulted to clients re technical methods, organizational management processes, and internal communication practices for better investment performance with proper risk control. Created and ran engagements each bringing in six figure revenues, involving Joe’s own Asset/Liability Efficient Frontier Paradigm, and Intercompany Investment Performance Survey (33 participants). Advanced the state of practice in insurance simulations modeling and ideal strategy selection.
1984-1986: Morgan Stanley, research associate and then product manager in the Fixed Income Research Department. Helped develop modernized techniques for quantifying life insurance company investment risk exposures, consulted to clients, assisted traders in major investment portfolio restructurings. Helped adapt option pricing theory to life insurance product liabilities.
1979-1984: Guardian Life, actuarial student, assistant actuary, then associate actuary and director of market planning, in individual health and then group life and health. Special R&D projects to modernize and make more accurate company actuarial reserving, pricing, and profit projection models. Helped to introduce better strategic planning and marketing research practices.
1977-1979: Trainee at pension plan consultants Kwasha Lipton, then Thacher & associates.
Education: Fellow of the Society of Actuaries 1980, MS algebraic topology MIT 1977, BA math NYU 1973. Won a third place Westinghouse Science Scholarship, then an NSF Fellowship.