Estimating theoretically consistent demand systems using cointegration techniques with application to Greek food data

The use of time-series techniques (cointegration and error correction models) has been offered as a potentially promising way in estimating demand systems that resolve the perennial issue of violating the theoretical postulates of homogeneity and symmetry. The empirical results of this paper, based on linearized AIDS and Greek food data, indicate that sample size, aggregation schemes and model specification seem to be important factors in the estimation of a theoretically consistent demand system. They also show that homogeneity is sensitive to sample size, while symmetry is sensitive to the aggregation scheme adopted.