Forecasting Volatility of Returns for Corn Using GARCH Models

Abstract

The purpose of this paper is to model and forecast volatility of returns for corn futures prices using GARCH models. Non-linear models from the GARCH family, specifically TGARCH and EGARCH are employed to assess the role of asymmetries and to analyze the time varying volatility of corn futures prices. The results reveal that the corn return series react differently to good and bad news. The presence of leverage effect would imply that the negative news has bigger impact on volatility than positive news of the same magnitude. The estimated volatility models were compared using symmetric measures for their forecasting accuracy. It is found that the EGARCH model provides the best out of sample forecasts for corn among all the GARCH specifications.