Nonlinear Granger causality between grains and livestock

Abstract

Linear and nonlinear Granger causality between three grains: corn, soybean, wheat and two livestock commodities: live cattle and lean hogs, was verified. Weak evidence of linear causal relationships was found, supporting the results published in other studies. However, strong nonlinear causal relationships between grain and livestock returns were found, which had not yet been documented in the literature on this subject. The revealed relationships have different patterns and features, and in some cases, they arise from second moment dependencies, but nonlinearities of a different type were also found. Most of the discovered nonlinear relationships are bidirectional.

Keywords

• agricultural futures contracts
• grain and livestock prices
• multivariate garch model
• nonlinear causality tests