Explaining commodity prices through asymmetric oil shocks: Evidence from nonlinear models

Linkages between oil and 25 other commodity prices are examined using annual data for 1900 to 2011. We identify long-run relationships using both linear and nonlinear ARDL models and capture short-run causalities through asymmetric Granger causality tests. Nonlinearity can't be rejected for the relationship between oil and most other commodity prices. Long-run positive impacts of oil price increases are found for 20 commodities and short-run negative impacts for 13 commodity prices. Oil prices don't have much impact on beverage or cereal prices once endogeneity is accounted for, but they have substantial impact on metal prices.