University of Southampton OCS (beta), School of Management PhD Conference 2011

Font Size: 
Forecasting the hedge ratio in commodities’ and agricultural futures markets: Evidence from GARCH Models

Building: Building 2
Room: Room 3043
Date: 2011-05-11 11:50 AM – 12:10 PM
Last modified: 2011-05-10


Trader use futures market for either price discovery or hedging risk. Voluminous futures’ trading is used to reduce and transfer risk in futures market by substituting for a cash market transaction.


Accurate forecasting of hedge ratio is important for investors when they are planning and making decision on investment.  Agricultural and commodities’ futures trading make up to 11 percent in futures market. However, the two markets are not well studied by prior scholars because of high price volatility of agricultural products and commodities. To our knowledge, very few studies investigate the forecasting of HR in the commodity and agricultural futures markets. In this paper we study the predictive power of six most frequently used GARCH family models, including GARCH, BEKK-GARCH, GARCH-X, BEKK-X, GJR-GARCH,  and TGARCH.

Five agricultural commodities containing coffee, wheat, soybean, live cattle and live hogs are explored. The first three products are categorized as storable goods that can be stored for years and have more stable price than non-stable demand-supply; the live cattle and live hogs are non-storable products which have physically or economically infeasible storage and demand almost equals to supply for non-storable goods. Based on 28 years data (1980-2008), in-sample estimation and non-overlapping two out-of-sample forecasts of hedge ratio and return for storable and non-storable products are executed with normal and student t distributed error. The forecasting accuracy of them is compared comprehensively under seven error evaluations.

Overall, results indicate that forecasting power of models somewhat depends on the commodity, the error distribution, and forecast horizon. However, the asymmetric GARCH models have great predictive power in HR forecasting for non-storable commodity.

Full Text: Slideshow