Job Market Paper
When Trade Distortions Disappear: Trade Reallocation
and Social Conflict in Qing China
Zixin Wang· Latest draft: [Oct 2026]
Abstract
Trade distortions can sustain economic activity in the places that benefit from
them, raising the question of what happens when they are removed. I study the 1842 Treaty
of Nanjing, which ended the Canton System, a restriction that had concentrated China’s
foreign trade in a single port for nearly a century. The new treaty ports let trade shift to
lower-cost routes, sharply reducing traffic along the old inland routes to Canton. Using an
original county-year panel of 427 counties from 1820 to 1860, I find that counties along the
pre-1842 routes experienced a large rise in rebellion after the treaty, while the hinterlands
of the new ports show no comparable increase. This asymmetry, concentrated in transithub counties, points to an opportunity cost channel through displaced transport workers
rather than rapacity. Affected counties became substantially more susceptible to Taiping
mobilization after 1851, linking the trade shock to one of the deadliest civil conflicts in
history.
Operational AI Embedding and Green Innovation: When Rapid Adaptation Matters
with Wei Zhao, Mahak Sharma, and Hao Chen · Working paper
Abstract
Firms increasingly report using artificial intelligence (AI), but disclosure alone does not reveal whether AI is embedded in the workflows where operating decisions are monitored, coordinated, and revised. We introduce disclosed AI-enabled execution capability (AEC), a sentence-level construct that treats annual-report text as evidence of workflow embedding only when AI language is tied to concrete execution features and operational or governance processes. Using a large panel of Chinese A-share listed firms from 2020 to 2025, we examine the link between AEC and subsequent green invention patenting. Once firm fixed effects, prior innovation, and adjacent digital measures are accounted for, the average within-firm association is small, but it is markedly stronger when operating conditions require rapid adaptation. The same pattern holds under alternative timing restrictions, non-text task proxies, and an external measure of environmental-enforcement pressure. The design is observational: disclosed AEC is measured from corporate reports and does not verify system use.
Physics-informed modeling of salinity-gradient power generation and operating conditions
with Wei Zhao and Hao Chen · Working paper
Abstract
Salinity-gradient power generation converts the energy released by mixing low- and high-salinity water into hydraulic power. Because recoverable power depends on both applied pressure and the salinity difference across the membrane, searching for suitable operating conditions can require many costly transport simulations. We develop a physics-informed model that represents the feed and draw channels separately and couples them through membrane transport conditions. Validated against finite-element simulations across the main operating range and nearby boundary conditions, the model accurately reproduces the pressure-dependent power maximum and the gains in available power as the salinity difference grows. It predicts a full field in a fraction of a second, offering a large speed-up over high-fidelity simulation, and so enables rapid screening of pressure and salinity combinations before selected conditions are examined in detail.
