Abstract
Time-series data for the USA and Europe representing prices of agricultural commodities, biofuels and fossil fuels are used for a comparative analysis of long-run price relationships. There is some evidence for cointegration between ethanol and gasoline, especially for the USA, and in the case of biodiesel, stronger evidence of cointegration between biodiesel, diesel and soya oil for both the USA and the Europe. Finally, biofuel prices do not seem to influence agricultural commodity prices or fossil fuel prices.
Notes
1. Like other studies of (bio) fuel prices we use prices in nominal terms. For example biodiesel and diesel are relatively close substitutes (or used in a certain combination of shares) where use of a common deflator might be relevant but of minor importance when estimating the cointegration models in the present analysis (apart from making the interpretation of e.g. the intercept term or a deterministic time trend, if included, different).
2. If for example α1 is significant, then a shift in P2 will influence ΔP1.
3. Corresponding to the data in –.
4. From 366 observations, the number of deleted outliers is 13. Including these outliers also reveals problems of heteroscedasticity, but again deleting outliers gives in most cases satisfactory results from an LMARCH(1) test.
5. Like for the ethanol model, the normality assumption will not hold when outliers are included, but again it is a relatively small number of outliers deleted.