Exploring 200 Years of U.S. Commodity Market Arbitrage: A Structural Time Series Model Approach

This project contains code and data to replicate "Exploring 200 Years of U.S. Commodity Market Arbitrage: A Structural Time Series Model Approach." The raw data encompass more than 570,037 monthly price observations for 72 locations and 103 goods spanning from 1700 to 1950. The cleaned data consist of 385,224 monthly observations across 70 locations and 42 goods from 1750 to 1949.<br>This paper uses a structural time series model to explore U.S. commodity market convergence, efficiency, and intertemporal smoothing from 1750-1949. I find near-continuous convergence that is largely concentrated in the frontier, broad antebellum efficiency gains, and intertemporal smoothing from the late 1800s onward among the most perishable goods. The results reveal new periods of integration across all three metrics and underscore the rapid rate of integration on the frontier.<br>