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 andvegetables in the period 1999-2004 in Denmark wasinvestigated using quantile regression and two previouslyoverlooked problems were identified. First, the change in theten percent quantile samples decreased. This could have beencaused 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 wasnot caused by alterations in the distribution of covariates butby changes in the coefficients and therefore a change inbehaviour. The reason for this development is probably due tolow income groups becoming relatively more incomeconstrained since the gap to the high income group have grownconsiderably at the lower end of the distribution. The secondproblem was that the education inducing gap became larger in2004 indicating that uneducated people have not responded aswell to the health related information flow. These resultssuggest that information campaigns have not been assuccessful as previously thought; more importantly the resultsindicate that information campaigns alone will do a poor job insolving the identified problems. Other instruments targetinguneducated and low income groups more directly are needed.Instruments which make fruit and vegetables relativelycheaper would undoubtedly have an effect on low incomegroups and send a strong signal to the uneducated population.