Time Series Analysis of Grain Futures Prices: Comparison of Short Term Forecasting

• Main Authors: Chin-Dee Wong, 翁靖迪
• Other Authors: Yu-Hui Chen
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/36706069247035802220

碩士 === 國立臺灣大學 === 農業經濟學研究所 === 96 === The purpose of this thesis aims to establish short-term forecasting models for the futures prices of grains. ARMA-GARCH, level VAR, and differenced VAR model models are chosen here to analyze the dynamic interactions among wheat, soybeans and Corn traded in Chicago Board of Trade and the spot price of crude oil in the western Texas. Then, these interesting relations are applied to predict grain prices 3-month in advance. Judged purely by model forecastability, the empirical results have shown that the ARMA-GARCH model performs better than other two VAR models in the grain futures prices considered. The impulse response analysis and the forecast error variance decomposition further indicate that oil price directly impacts the futures prices of corn, and oil and corn prices later push the wheat and soybeans prices. In short, these grains are closely related with the rising oil prices which makes those grain futures prices go up.