MODELLING IMPORT DEMAND SYSTEMS WITH NONSTATIONARY DATA: AN APPLICATION TO THE FRENCH IMPORTS OF VIRGIN OLIVE OIL

This paper aims to provide a flexible methodological framework to estimate import demandmodels, which explicitly considers the stochastic properties of data and the endogenous/exogenousnature of some variables. The French imports of virgin olive oil have been used as a case study withSpain, Italy and the Rest of the World as main suppliers. The methodological framework starts by thespecification a reduced-form VAR. Appropriated exogeneity tests show the exogeneity of Total RealImports, indicating the appropriateness of estimating a conditional model. Two cointegrationrelationships have been found. Several restrictions have been tested in order to identify them as AIDSequations. From structural coefficients of the restricted cointegrated vectors expenditure, own- andcross-prices elasticities are computed. Results show the leadership of Spanish exports to the Frenchmarket. Italian exports compete in the French market with the Spanish exports, being highlydependent on Spanish domestic production conditions.