Evaluating the Health Impacts of Food and Beverage Taxes

Before I wrote the original article that was published in Obesity Reviews, most of my work in this area, had been ‘modelling’. Modelling uses economic data to estimate the potential impact of price changes (from taxes or subsidies) on consumption and what that in turn would mean for people’s health (see Adam Briggs’ article for an example). Whilst I think these studies are important, I began to realise that the evidence these studies generated was only ever going to be part of the evidence jigsaw that policy makers sought. Policy makers appeared to be much more interested in ‘realworld’ evidence (i.e. studies of actual taxes), the problem of course was that very few if any countries had introduced such taxes with the explicit aim of improving health. It felt like catch-22.

But in late 2012 that all changed, when Denmark introduced a new tax on saturated fat, in part to address high levels of cardiovascular disease in Denmark. Other countries followed suit, with France taxing sweetened beverages and Hungary introducing taxes on some products high in salt and sugar in 2013. Studying these ‘experiments’ could provide much richer evidence for policy makers – not just concerning the effect of the tax on consumption of food and health, but other factors: how does industry respond?; are there other economic effects?; what do the public think?

However rigorous scientific evaluation is not easy to do. There are a number of challenges. For example people often hope to link taxes and subsidies directly to changes in health. Sometimes this may be possible, but often it won’t.

For example Cliona Ni Mhurchu’s work from New Zealand suggests that a subsidy on fruit and vegetables at 20% would lead to a 12% and 18% increase in consumption of fruit and vegetables respectively. That should contribute to reductions in cancer and heart disease, but those reductions won’t happen immediately. There are also many other things changing, treatments are improving and smoking is going down. These, and other things, will affect cancer and heart disease, which makes it very hard to point to changes in disease and confidently say those changes are due to a new subsidy or tax. Instead in our review, we suggest the evaluation should focus on changes in consumption. When changes in consumption have been shown to relatively large, it may be appropriate to look at the effects on some measures of health.

We also highlight the importance of not just focusing on benefits, but also considering potential harms that might occur from taxes or subsidies. If a tax on salt leads to people purchasing fewer salty snacks, then they might compensate by purchasing more sugar-sweetened confectionery. They may also respond to a tax by spending less money on fruit and vegetables in order to absorb the increased cost of the diet. It is important to evaluate the effect of taxes and subsidies across the whole diet rather than just the targeted product. Ultimately a full understanding of the effect of taxes and subsidies will come from a range of different approaches: evaluation of real policies, as well as modelling studies (like those that Adam Briggs and Cliona Ni Mhurchu discuss).