The use of Markov-Switching GARCH models in a Mexican rice spot price hedging algorithm with CME rice futures

The present paper tests the effectiveness of using symmetric Markov-Switching GARCH (MS-GARCH) and asymmetric MS-EGARCH models in a hedging decision algorithm for the Mexican spot rice price. The rationale of the simulated algorithm is to hedge the rice price with a short, one-month Chicago Mercantile (CME) rice future if the producer forecasts a high volatility or distress scenario at t + n (the models assumed a two-scenario context). Such forecasts were made with the MS, MS-GARCH, or MS-GARCH model. Also, the simulations assumed an unconditional non-switching location parameter (arithmetic mean) or a conditional parameter estimated with the impact of the CME rice future price and its speculaton ratio. With weekly simulations and t + 1 and t + 4 weeks hedging horizons, the simulations found that using Gaussian MS-GARCH in t + 1 and t+ 4 adds additional income due to hedging, but this result only holds in the short-term and due to market issues. Consequently, this work is among the first to test the benefits of the MS-GARCH model for agricultural non-commodity price hedging and for food security applications.