New developments in fruit and vegetables consumption in the period 1999-2004 in Denmark - A quantile regression approach

The development in the consumption of fruit and vegetables in the period 1999-2004 in Denmark was investigated using quantile regression and two previously overlooked problems were identified. First, the change in the ten percent quantile samples decreased. This could have been caused by changes in the distribution of covariates. Therefore, the counterfactual decomposition of Machado and Mata (2005) was used and the methodology established that the change was not caused by alterations in the distribution of covariates but by changes in the coefficients and therefore a change in behaviour. The reason for this development is probably due to low income groups becoming relatively more income constrained since the gap to the high income group have grown considerably at the lower end of the distribution. The second problem was that the education inducing gap became larger in 2004 indicating that uneducated people have not responded as well to the health related information flow. These results suggest that information campaigns have not been as successful as previously thought; more importantly the results indicate that information campaigns alone will do a poor job in solving the identified problems. Other instruments targeting uneducated and low income groups more directly are needed. Instruments which make fruit and vegetables relatively cheaper would undoubtedly have an effect on low income groups and send a strong signal to the uneducated population.