The impact of child obesity news on UK household food expenditure

The United Kingdom (UK) has one of the highest obesity levels in the world (Mazzocchi, Traill and Shogren, 2009). As indicated by the National Health Service (2010), 25% of adults and 17% of children are obese in the UK. This last statistic represents an increase of four points in comparison to 1995. The Government Office for Science (2010) estimated that by 2050, half of the UK population would be obese, with a consequent direct annual cost of £ 1 0 billion and an indirect annual cost of £50 billion at today's prices. Governments have the role of ensuring that households have the most complete information possible about their food choices (Mazzocchi, Traill and Shogren, 2009). With this objective, the UK government has conducted information campaigns such as nutritional food labelling and the 'Change 4 Life' campaign, in order to increase nutritional awareness. Despite government efforts, obesity has been steadily increasing in the UK. This research aims to contribute to the debate on how health-related information impacts household food expenditure and whether this impact varies across income groups and household composition. This study specifically measures the impact of child obesity news on household food expenditure in the UK. To this end, the study calculated a set of elasticities for different income groups (high vs. low) and family composition (families with and without children). This set of elasticities gives us a measure of responsiveness, to change in terms of price, income and news. This study uses an augmented two-stage budgeting demand system. Demand systems combine price, income and news index data into a well-supported economic framework. The empirical analysis includes testing for homogeneity, symmetry, concavity and the time series properties of the data and the residuals. In the UK, no recent study has measured the impact of news on household food expenditure. Moreover, few empirical demand studies use structural approaches that are consistent with the time series properties of the data