Chinese Rural Industrial Productivity and Urban Spillovers

Chinese rural industry has grown three times faster than national GDP, surpassing agriculture in size in 1987, and now nearing half of the total Chinese economy. We use a rich, new county-level data set to explore this dramatic growth. We find that a Cobb-Douglas production function explains over 80 percent of across-county variation in 1991 rural industrial output per capita, with little role for idiosyncratic regional or provincial fixed effects. There is a very large effect on productivity from being near cities (30 to 35 percent higher productivity for a county one standard deviation above average in nearness to population centers) due to embodied technology transfer from urban residents. We find strong support for the hypothesis that saving from past agricultural income has provided start-up capital for rural enterprises. However, higher land-labor ratios lead to greater allocation of labor and capital to agriculture instead of industry, although induced inflow of migrants reduces the effect on industrial labor. Nearness to cities and more education increase capital and labor in rural industry. Substantial explanatory power (one third or more) for industrial labor and capital is attributed to provincial fixed effects, possibly reflecting local commercial and migration policies.