Cotton Futures Price Variability: The Role of China's Cotton Inventory Policy

Understanding the role of China’s cotton inventory policy in affecting cotton price variability has recently become an issue of considerable importance to policy makers, producers, commodity traders, and researchers. This paper empirically examines the determinants of price variability in the U. S. cotton futures markets, with a special focus on the cotton inventory policy of China.

China, the world’s largest cotton growing country, has a price support policy for subsidizing domestic cotton farmers. Under this policy, the government of China purchases domestically produced cotton at above world market prices. Currently, China holds more than half of the World’s cotton stocks. While some participants in the global cotton industry fear that China has certain deeper motives to manipulate the global market using the massive reserve, continued stockpiling of cotton suggests a continuation of market distortion.

Our empirical analysis investigates the variability of weekly averages of cotton futures settlement prices observed between 2003 and 2012. We evaluate the relative contribution of four types of shocks to the variability of cotton futures prices. These are shocks to real economic activity (flow demand), net current production (flow supply), external market price changes (co-movement), and rational speculation (inventory demand). Data for this analysis are obtained from Commodity Research Bureau (CRB) database of futures prices and the United States Department of Agriculture (USDA).

The empirical analysis proceeds in two distinct directions. The first uses conditional heteroscedasticity (CH) models to investigate deterministic factors related to futures price variability as well as autoregressive conditional heteroscedasticity (ARCH) and generalized autoregressive heteroscedasticity (GARCH) relationships. Maximum likelihood techniques are used to estimate the CH, ARCH, and GARCH models. The analysis is repeated including and excluding China’s inventory of cotton. A second component of the empirical analysis involves the application of nonstructural vector-autoregressive (VAR) models to market-based measures of price volatility.