Robert L. Powell, "Bob," died on December 13, 2021. Bob was one of the world’s foremost applied game theorists and made important contributions to our understanding of the causes of war and political conflict more generally.
A mathematics major at Harvey Mudd College, he completed an M. Phil in International relations at the University of Cambridge in 1982 and his PhD in Economics at UC Berkeley in 1985. Bob subsequently taught political science at the University of Michigan (1985-1987) and Harvard University (1987-1990), and then returned to Berkeley in 1990, where he was the Robson Professor of Political Science until he passed.
Bob pioneered the use of modern non-cooperative game theory (mainly developed in the 1980s) to reconsider and rebuild central arguments of international relations theory. His work consistently sought to go beyond general claims about anarchy and conflict, to more clearly identify specific strategic settings, mechanisms, and paths that might lead to organized violence in some cases but not in others. A recurrent theme is the idea that in a surprisingly diverse set of contexts, both interstate and civil conflict is driven by the anticipation of adverse shifts in relative military capability or opportunity, coupled with constraints on the parties’ ability to either regulate or commit not to take advantage of favorable shifts.
Bob’s earliest work made groundbreaking contributions to explanations for armed conflict that are based on the parties’ uncertainty about each other’s willingness or ability to use force. His first book, Nuclear Deterrence Theory: The Search for Credibility (1990), was recognized by the National Academy of Sciences in 2012 with the William and Katherine Estes Award for Behavioral Research Relevant to the Prevention of Nuclear War.
Here Bob used recently developed methods of incomplete-information game theory to reconsider Thomas Schelling’s and other classic deterrence theorists’ representations of “crisis bargaining” between nuclear-armed adversaries. He formalized Schelling’s idea of nuclear crises (like the Cuban Missile Crisis) as a “competition in risk taking,” modeling it as a kind of auction in which the players “bid” in amounts of risk of nuclear catastrophe, or with limited strikes in a “counterforce war of attrition.” The insight that successful deterrence between nuclear adversaries is fundamentally a political rather than a military problem, and that there is no military or technical way to ensure success, “manage escalation,” or maneuver the opponent to ensure your preferred outcome, is once again highly policy relevant as the US government tries to strengthen deterrence with respect to possible attacks on Taiwan and Ukraine.
Bob returned to deterrence theory on several occasions. He extended an auctions-based model of brinkmanship to questions about national missile defense and new nuclear states (International Security, 2003), and he proposed a way to analyze the interaction between relative conventional capabilities and nuclear risk (International Organization, 2015). Both are problems not directly or satisfactorily treated in the classical literature that focused on nuclear risk during the Cold War. He also studied the problem of allocating resources across possible targets of terrorist attack, pointing out that optimal policy should, roughly speaking, seek to equalize an attacker’s expected payoffs across the highest value targets, rather than minimize vulnerability net of costs (American Political Science Review, 2007a and 2007b). The general principles are relevant to allocating resources to defend critical infrastructure against state-based attacks as well, a problem of rapidly increasing policy importance.
Bob’s widely-read second book, In the Shadow of Power: States and Strategies in International Politics (1999), reconsidered three central means by which states have sought security or gain—arming, allying, and threatening to use force. Arming is a waste (less “butter”) but may be needed for deterrence of a revisionist competitor. A high cost of deterrence can then make going to war in hope of reducing the need to spend so much to deter in the future more attractive, despite the costs and risks of war. Developed using a straightforward but highly original dynamic game model, these arguments provide an alternative grounding for common realist claims about the “security dilemma,” the offense-defense balance, and the relationship between arms levels and the risk of conflict. Another chapter presents one of the first non-cooperative game theory models of alliance formation that introduces realistic frictions (like costly conflict), which Bob shows work against a universal tendency for balancing.
In the Shadow of Power also contains a chapter on an explanation for costly conflict that Bob came to see as surprisingly general, both theoretically and empirically. Namely, costly conflict can result from large and rapid shifts in the distribution of power when states exhibit limited ability to commit to future promises. This work transformed our understanding of power transitions and war in international relations and is useful for evaluating and qualifying claims about preventive war and United States and China relations today.
The underlying mechanism in this preventive-war example applies to a remarkably wide range of circumstances in international relations, including conflict related to first-strike or offensive advantages in military technology, and conflict related to strategic territory (see in particular “War as a Commitment Problem,” International Organization, 2006). In more recent work, Bob demonstrated why commitment problems are also foundational for understanding domestic political conflict in weakly institutionalized settings. An initial contribution in this regard was “The Inefficient Use of Power: Costly Conflict with Complete Information” (American Political Science Review, 2004). Employing his typical parsimonious approach to model construction, he demonstrated that leading theories on topics such as civil war and democratization in fact posit a mechanism that is strategically identical to the “international relations” mechanism of shifting power and commitment problems. Contributions such as these help to break down the traditional barrier between international and domestic approaches to war.
In the last decade, Bob broke new ground by theorizing attributes specific to domestic politics. Rather than taking shifts in the distribution of power, as given in “Monopolizing Violence and Consolidating Power” (Quarterly Journal of Economics, 2013), he considered how governments can attempt to strategically achieve a monopoly of violence when facing armed challengers. Actions such as taking over the army or police forces, creating a militia, and arresting opposition leaders create conditions for conflict by endogenously shifting the distribution of power away from societal actors. However, governments are often willing to tolerate conflict to achieve the “contingent spoils” that accrue from consolidating a monopoly of violence.
In a final unpublished paper, “Power Sharing with Weak Institutions,” Bob analyzed another problem specific to domestic politics. Rulers can try to alleviate commitment problems by offering power-sharing deals. Specifically, the government can offer to opposition groups permanent control over an asset that produces spoils. Such deals enable members of the opposition to enjoy spoils even at times that they cannot coercively mobilize against the regime.
Throughout, Bob was a generous teacher, mentor, colleague, and friend. Bob loved Berkeley, even as he had a highly developed appreciation for the inanities of the institution. He worked assiduously to improve the university, chairing the faculty committees on Privilege and Tenure and Research and eventually chairing the Berkeley Division of the UC Academic Senate. His service to the campus was recognized in 2018 with the Berkeley Faculty Service Award. His Game Theory in the Social Sciences course was hugely popular among undergraduate students in the political science and economics departments. Bob was able to explain complex ideas simply and intuitively, and he made the game theory material accessible to all students, regardless of their math backgrounds. While teaching this course, it was always extremely important to Bob that the students have compelling and interesting stories to accompany the math problems. One year, he had to rewrite the midterm in a hurry because someone had pulled the fire alarm. Despite the urging from his graduate student instructors to skip the elaborate set ups and just write some math problems, Bob stayed up all night constructing detailed contexts for the game theory problems. This illustrated a central idea in his teaching and mentoring: Bob viewed game theory as an important tool that can be used to understand substantive social science questions. He always wanted to make sure his students didn’t lose sight of the empirical connections of models.
Bob was always extremely thoughtful and reflective. Never quick to criticize, he nonetheless was profoundly insightful in his comments in any academic setting. On a tribute page (https://padlet.com/embed/bndw57ns0aolqsf0) hosted by the political science department at Berkeley, one former student remarked that he used to attend seminars by outside speakers just to hear what Bob would say during the Q&A period. We can all attest to how his comments on our individual works improved our research. Although much of Bob’s research was done alone, he was also a tremendous collaborator, as evidenced in a co-edited volume on strategic choice that sought to build a framework based on game theoretic insights that could be used to structure all of international relations theory. Bob was the driving intellectual force in this effort, but he also carefully nurtured the other contributors to absorb his ideas and work to build an unusually coherent volume.
Bob was a wonderful colleague and friend. He loved animals and would be delighted when someone brought a pet to the office. A former graduate student recalls the first time she met Bob: she walked into the main political science department office and found him sitting on the floor enthusiastically petting a dog. Bob was a coffee lover, and he knew all the best coffee shops in Berkeley. An avid body surfer, he swam a mile or more nearly every day at the Berkeley pool and was overjoyed when various periods of remission during his illness allowed him to return to the water. He loved taking his annual vacations to Maui and often longingly looked up the latest weather and surf conditions from his office. Bob was also a generous host, well known for his perfectly prepared dinners and relaxed evenings of conversation at his home.
Bob suffered from metastasized melanoma for several years. As his health deteriorated, he declined to take medical leave in Fall 2021, choosing to teach once again his renowned undergraduate course on game theory (https://www.robertpowellberkeley.com/lecture-notes.html). As this was a popular course with both economics and political science students, Bob did not want to disrupt the students’ progress nor impose on his colleagues to substitute for him on short notice. Under California’s “End of Life Option Act,” and only days after his final lecture, Bob chose to die peacefully at home in his favorite spot watching the sun set over San Francisco Bay. He will be missed as a teacher, colleague, and friend. ■
