Industrialization, urbanization and land use in China

Abstract

Rapid industrial development and urbanization transfer more and more land away from agricultural production and affect the patterns of land use intensity. This paper analyzes the determinants of land use by modeling arable land and sown area separately. An inverse U-shaped relationship between land use intensity and industrialization is explored both theoretically and empirically. The findings highlight the conflict between the two policy goals of industrialization and grain self-sufficiency in the end. Several policy recommendations are offered to reconcile the conflict.

Keywords:

• Industrialization
• urbanization
• land use and China
• JEL classifications: O13
• Q15
• R14

Acknowledgements

The authors gratefully acknowledge helpful comments from Shenggen Fan, Peter Hazell, Scott Rozelle, and participants in seminars held in Chinese Academy of Agricultural Sciences, Cornell University, IFPRI, and World Bank.

Notes

 Arable land is land that can be farmed even if it is not. Sown area or cropping area is equal to the product of arable land area and multiple-cropping index. It has been noted from the recent agricultural census that the previous official arable land area has been under-reported (p. 368, SSB, 1997; Ash & Edmonds, 1998; Smil, 1999). The debate in the literature focuses on the magnitude of under-reporting. In spite of the shortcomings of the official statistics, they are the only source for land stock at the provincial level readily available and consistently compiled over a long period. The trends of land use may not be severely affected by this problem considering that most under-reporting of arable land occurs in hilly and mountainous regions (Ash & Edmonds, 1998). Considering that the ratio of under-reporting is rather constant over time, by taking advantage of the availability of a panel data set, we can use regional dummies to largely reduce the systematic measurement errors in our econometric analysis.

 Because most of China has already been heavily populated, there are few opportunities to claim marginal land. Arable land may also be lost due to environmental changes caused by soil erosion and salinity (Huang & Rozelle, 1995; Ash & Edmonds, 1998). Because environmental changes are mostly related to population growth, an increase in agricultural inputs, and development of rural enterprises, they can partly be captured by the population and industrialization variables in the model. The data needed for a more complete analysis were not available.

 Although a rapid change in grain trade positions often has an important impact on the tightness of land use policy, other factors may also affect land policy.

 For one rationale for treating area allocations and yield separately in estimating an agricultural supply function, see McGuirk & Mundlak (1991).

 For simplicity, the subscript for each observation is omitted. In fact, the levels of output price, wage rate, and technology development may vary across regions.

 Under the assumption of constant returns to scale for production, the cost neutrality technology in equation (4) is equivalent to a profit neutral or Hicks neutral technical change (Chambers, 1988).

 Other functional forms, such as an inverse function, were also tried and the results are similar.