Works in progress
The Strategic Calculation of Negotiation Before and After Ceasefire: An Empirical Puzzle and Theory
What drives belligerents’ decisions of when to stop fighting and when to negotiate during interstate war? This paper shows an empirical puzzle of changing patterns in wartime negotiations after World War II: states substitute post-ceasefire negotiations with pre-ceasefire negotiations. I examine the duration of negotiations before and after ceasefires with a dataset of 91 interstate wars from 1823 to 2001, featuring 35,521 conflict days. I find that, first, fewer negotiations occured after ceasefire post-World War II, compared to pre-World War II period. Second, wartime negotiations lasted longer. Using a multi-state hazard model with time-varying covariates, I show that fighting is less likely to end after negotiation begins after 1945. Why do states negotiate more during the war but avoid peaceful negotiation after ceasefire? I argue that increased domestic and international pressure for peace after World War II increased the cost of restarting war, making negotiation after ceasefire inefficient for belligerents. Thus, states have an incentive to prolong negotiations before ceasefire to gain concrete advantage in the resulting settlement. I demonstrate this strategic calculation using a case study of the 1988 Iraq-Iran ceasefire negotiation, where two sides bargained over what needed to be negotiated before implementing a UN-sponsored ceasefire.
Great Power Competition Revisited (with David A. Lake)
When and how great powers compete is a central question to International Relations. This paper develops a game-theoretic model to study the conditions under which great powers exercise influence in target countries through domestic “allies.” It explores situations in which two political groups within a third country compete for control of government policy, and great powers can potentially sway the policy outcomes by supporting groups whose policy preferences are most closely aligned with their own. The model with complete information suggests that great powers are more likely to intervene when their stakes are high in the target country, political polarization within the country is intense, and the allied faction is comparatively weaker. Under incomplete information, great powers may excessively support a faction to dissuade competition, or strategically limit their involvement to avoid escalation into a full-scale great power contest. This model implies that great powers choose to compete on a case-by-case basis and competition and conflict will not necessarily be a universal phenomenon. We illustrate the model through brief historical applications to the Soviet Union and the United States’ involvement in the Korean Peninsula in the 1950s and Cold War dynamics in Europe.
Tables Have Turned: Alternative Bargaining With Stochastic Shift of Power
I study an infinitely repeated symmetrical bargaining model where 1. Players decide the split of the prize either through peaceful negotiation or costly fighting. 2. In any period, the relative military power shifts following a Markov process and the transition matrix is symmetrical. Under complete information, this bargaining model has a generically unique perfect equilibrium payoff for each player, but the equilibrium outcome is inefficient if the probability of power transition is too high or too low. This model shows of how volatility, rather than asymmetrical information or commitment problem, could cause bargaining delay.
US-China Talent Flow In Critical Technology Sector: Evidence From LinkedIn Data (with China Data Lab)
Competition for talent in the critical technology sector is a central aspect of the US-China relationship. The United States is concerned that talent migration from top US technology firms to Chinese firms could accelerate China's growth in critical technologies, impacting both economic and security interests. This project explores where people work in China after being employed at top US tech companies and whether many of them transition to Chinese tech companies. Using LinkedIn data, we analyze 651,546 individuals who had worked at top US tech companies and had experience living in China between 1990 and 2020. Our findings cast doubts on the US’s the concern over losing talents: 93% of people who worked at US tech companies in the US remained in the US, with only 6% moving to work in China. Among those who returned to China after working at US tech firms, employment was spread across various firms rather than concentrated in large Chinese tech companies.