Fall, 2026
November 6 (Friday), 2026, 11:30AM-12:30PM (PST) / 1:30-2:30PM (CST) / 2:30-3:30PM (EST)
“Command and Loyalty: Political Choices of Military Elites in Civil War”
Linchuan Zhang (Emory University)
Abstract: The military is vital to the survival of authoritarian regimes, but variations in officers’ loyalty pose a fundamental challenge to dictators. I develop a framework that analyzes how dictators strategic position appointments shape officers subsequent allegiance. I theorize that officers occupying key command positions are more likely to remain loyal because of the benefits tied to their roles. Furthermore, the effect of position depends on factional affiliation: insiders whose loyalty comes from personal connections to the ruler are less influenced by appointments, while officers who are outsiders to the inner circle are more motivated to remain loyal when appointed to important positions. Empowering outsiders with important positions, however, can be a double-edged sword: once disloyal, they are more likely to defect than desert, causing greater harm to the regime. Empirically, I analyze Kuomintang (KMT) military elites during the second phase of the Chinese Civil War (1946-1950) using an original dataset of over 800 KMT generals, and find evidence consistent with the theory. This study highlights the importance of position appointments in shaping military loyalty and emphasizes the heterogeneity of individual officers’ decision-making, contributing to research on authoritarian politics, civil-military relations, and conflict processes.
Discussants: Mai Hassan (MIT) and Tyler Jost (Brown)
November 13 (Friday), 2026, 11:30AM-12:30PM (PST) / 1:30-2:30PM (CST) / 2:30-3:30PM (EST)
“Bureaucratization as a Bumpy Journey: The Rise and Fall of Imperial Chinese State”
Shuyi Yu (University of Chicago)
Abstract: Historians have long documented the rise and fall of the Imperial Chinese state. This paper introduces bureaucratization as a potential mechanism underlying these dynamics. It develops a game-theoretical model to analyze the inherent principal-agent problem in state administration, identifying two types of agents available to the ruler: elites and bureaucrats. The process of bureaucratization, driven by exogenous forces such as technological advancements, has enabled the state to monitor an increasing portion of administrative tasks and delegate them to bureaucrats rather than elites. Relational contracts with bureaucrats offer higher returns than those with elites because bureaucrats can be more effectively monitored. Interestingly, as bureaucrats gradually took on administrative duties, state capacity followed an N-shaped trajectory. The mid-phase decline in state capacity is attributed to the weakening credibility of the ruler’s contract with the elites and their reduced efforts. This decline aligns with the significant downturn experienced by the Chinese state during the late Imperial period, followed by a resurgence in modern times as bureaucratization eventually reaches its full form. Furthermore, this model may have broader applicability to other civilizations.
Discussants: Clair Yang (University of Washington) and Sean Gailmard (UC Berkeley)
November 20 (Friday), 2026, 11:30AM-12:30PM (PST) / 1:30-2:30PM (CST) / 2:30-3:30PM (EST)
“Assimilation as a Courtesy: Manchu Bureaucrats and Imperial Personnel Strategy in Qing China (1726-1820)”
Cheng Cheng (New York University)
Abstract: How do minority rulers govern majorities? A dilemma arises when small dominant groups must govern through majority-facing institutions while preserving the ethnic boundaries that sustain minority rule. I argue that rulers manage this tradeoff by cultivating partially assimilated co-ethnic elites: officials who acquire majority cultural fluency while remaining embedded in minority institutions. I evaluate this claim using an original panel of roughly 34,000 position-years and 5,200 Qing officials between 1726 and 1820, using the adoption of Chinese courtesy names by Manchu bannermen as a proxy for cultural fluency. Manchu with courtesy names received more important local post in Manchu-specific association that vanishes among Han officials once credentials are controlled. At position transitions, courtesy-name officials accumulate significantly more merit grades conditional on their prior record, suggesting that the personnel system systematically rewarded this dimension. Over the career, these officials rotated more frequently between majority-facing civil posts and boundary-reinforcing banner posts, and were more likely to be recalled from provincial postings back to central court service, consistent with an incentive structure that encouraged partial assimilation while limiting over-assimilation. These findings suggest that assimilation can function not only as a strategy of subordinate groups, as has been shown in the literature, but also as a technology of rule for dominant minorities. My project develops partial assimilation as a framework for understanding how minority regimes preserve ethnic closure while staffing majority-facing institutions.
Discussants: Peng Peng (Washington University at St. Louis) and Vasiliki Fouka (Stanford)
December 4 (Friday), 2026, 11:30AM-12:30PM (PST) / 1:30-2:30PM (CST) / 2:30-3:30PM (EST)
“Digital Panopticon, Built on Cheap Surveillance”
Alison Sile Chen-Zhao (University of California, San Diego)
Abstract: Autocrats have long relied on imprisonment and surveillance to suppress disaffected citizens. However, pervasive incarceration disrupts production and weakens the state, while traditional surveillance is labor-intensive and generates more records than bureaucrats can process. Digital surveillance, powered by AI and large-scale infrastructure, offers an alternative: it allows citizens to stay productive while enabling the regime to preempt opposition. Through formalizing state-citizen interactions under surveillance, this paper shows how digital surveillance reshapes authoritarian social control. The technology affects citizens unevenly. People who pose moderate threats to the regime are most likely to be surveilled. As the technology grows cheaper, surveillance expands to lower-threat citizens; as it grows more accurate, it expands to higher-threat ones. A further finding runs against common intuition: cheap, instead of advanced, surveillance is the key to authoritarian durability, because its constant use deepens self-censorship across the population. An ongoing data collection effort, drawing on large language models to organize hard-to-process Chinese sources, provides empirical traction on the theoretical arguments.
