The Economics of Strategic Learning in Trade Wars: Evidence from the Trump -Xi Natural Experiment. -- Conditional Reciprocity and Behavioral Synchronization in US-China Trade Wars--

How do rival leaders learn to retaliate without trust or treaties? This paper exploits a rare “same leaders” natural experiment to trace strategic learning in bilateral economic conflict. We analyze two periods of the Trump–Xi trade war — “Trump–Xi 1.0” (2017 2020) and “Trump–Xi 2.0” (2025) — in order to isolate how experience shapes escalation in tariff retaliation. We introduce the Bilateral Learning Strength Index (BLSI). This novel metric captures two behavioral dimensions: conditional reciprocity—how predictably one side responds in kind—and behavioral synchronization—how closely rivals mirror each other’s timing and intensity across repeated trade actions. Using data on 37,098 U.S.–China trade actions, we find that escalation in Trump–Xi 2.0 is substantially more constrained. Retaliatory responses are both more disciplined and far more synchronized, with a correlation coefficient of 0.884. These results suggest that through repeated interaction, adversarial leaders converge toward implicit rules of engagement—thresholds for retaliation that stabilize conflict dynamics even in settings devoid of formal treaties or mutual trust. The framework has potential applications well beyond trade wars, including central bank coordination, oligopolistic competition, and international monetary spillovers—any environment in which actors engage repeatedly without binding agreements. The BLSI also lays the groundwork for "Quantified Conflict Studies”, which could enable strategic forecasting, AI-assisted diplomacy, and real-time monitoring of conflict behavior in trust-deficient settings. By making strategic learning empirically measurable, this paper contributes to understanding not only how conflict escalates but also how it may evolve toward patterned stability rather than chaos.