Research
I study how trade and market connections shape local economic outcomes. My research combines quantitative spatial models, panel econometrics, and detailed geographic data, with a focus on China and emerging Asia.
Job market paper
The Local Multiplier Effects of Trade: A Quantitative Analysis of China
With Kensuke Suzuki, Xiaocong Xu, and Junfu Zhang.
Research question. Why do export shocks generate different effects on output and employment across local economies?
Approach. We combine evidence on local export shocks with a multi-sector quantitative spatial model of 31 Chinese provinces. The model includes inter-provincial input-output linkages, non-tradable services, and labor mobility, and is calibrated using China’s 2017 MRIO table, ICIO data, and Customs micro-data.
Contribution. The analysis connects multiplier differences to domestic production networks and the intra-regional manufacturing trade share. Provincial export-GDP multipliers range from 0.30 to 1.02 in the model’s counterfactual simulations. These are model-implied output responses; employment responses have different units and are reported separately.
My role. Quantitative model development, structural parameter recovery, calibration-data work, and counterfactual analysis.
Working papers and preprints
High-Speed Railway and Market Integration: Evidence from Vegetable Price Dispersion in China
Research question. How is passenger connectivity related to spatial price differences in wholesale commodity markets?
Data and approach. A panel of 3.99 million city-pair-month observations covers ten vegetables across 117 Chinese cities from 2014 to 2022. Wholesale prices are combined with smartphone-based passenger-mobility measures and rail connectivity information.
Evidence. HSR-connected city pairs exhibit 2–5% lower price dispersion in the reported specifications. Passenger-mobility controls and commodity comparisons help examine possible mechanisms. These conditional associations do not, by themselves, establish a causal effect of high-speed rail.
Related application. The Vegetable Price Intelligence project uses a broader historical price dataset for forecasting and risk analysis. Its forecasting performance answers a different question from this paper’s analysis of market integration.
Population Agglomeration and Spatial Reallocation in China
Research question. How have population concentration and the geographic distribution of settlement changed across China?
Data and approach. Annual WorldPop rasters for 2002–2020 are harmonized at approximately 5 km resolution for descriptive spatial analysis.
Evidence. The analysis documents increasing concentration and changes in the population-weighted geographic center. The spatial Gini rises from 0.852 to 0.868, and the population-weighted centroid shifts about 17 km southeast. These patterns describe spatial change rather than identify its causes.
Read the preprint on SSRN · Explore the companion atlas
Methods and data
- Economic models: quantitative spatial equilibrium, trade and production networks, calibration, structural inversion, and counterfactual simulation.
- Empirical methods: panel econometrics, shift-share instruments, event studies, gravity models, and geospatial measurement.
- Data: MRIO and ICIO tables, Customs micro-data, wholesale prices, retail scanner panels, passenger mobility, and population rasters.
- Tools: Python, Stata, R, MATLAB, SQL, SAS, GeoPandas, and ArcGIS.