Understanding the heterogeneous nature of the demand for soft drinks in Mexico: why social determinants also matter.

Background. Soft drink consumption is a risk factor for obesity and non-communicable chronic diseases, and policies to reduce it have been proposed around the world, including taxation. Little is known about the role of other social and economic factors on the demand of such goods. In addition, heterogeneity of the demand due to different levels of consumption has been rarely explored. The aim of this study is to analyse the heterogeneous nature of the demand for soft drinks to understand the role of economic and social factors (provision of safe water /local food market conditions) and draw recommendations for the design of obesity prevention. Methods. Population, cross-sectional analysis of household data from the Mexican Family Life Survey, grouped into three consumption groups (low/medium/high consumers, defined by the proportion of total household expenditure devoted to soft drink purchases) and three economic poverty groups (defined by extreme and moderate income poverty lines). Multivariate probit regressions were applied to explore factors associated to the probability to be a consumer, and simultaneous multivariate quantile regressions were used to model the quantity purchased of soft drinks. Heckman’s procedure was used to control for identification bias. Results. The adjusted probability that a household becomes a consumer is significantly higher with male, educated heads of households and higher household income. Living in localities where access to safe water for drinking and cooking needs is not universal significantly increases the probability to consume soft drinks while living in localities with convenience stores and supermarkets (local food market condition) significantly decreases it, especially in households facing extreme poverty. Demand from low-consumption households is price-inelastic (-0.97) compared with high-consumers (-1.2). Yet when the population is grouped by poverty, households in extreme poverty have a higher significant price-elasticity (-1.5) than those above moderate poverty ...