This paper applies a structural approach to the empirical investigation of futures pricing and the estimation of the minimum variance hedge ratio. Currently, the approach is ad hoc, and lacks any a priori basis of model choice. This study suggests a dynamic structural framework to integrate established theoretical interpretations of the relationship between the spot and the futures price and the hedging decision. Economic tools, specifically structural vector autoregression modelling (VAR) and the Johansen methodology, are suggested to improve the power of the analysis, with the objective of deriving a long term structurally stable optimal hedge ratio. This study finds that the VAR approach provides a theoretically sound option to hedge ratio analysis, comparing favourably to the simple OLS based approaches, and the more complex multivariate GARCH models. In applying estimated hedge ratios from the VAR methodology, we find it outperforms all other models when tested in the context of a mean-variance framework.