7.3.1. Food labelling

The modelled intervention consists of statutory policy changes, requiring all manufacturers or retailers to provide information on the nutritional composition of foods sold in stores and supermarkets. Such information should be clearly visible, as well as easy to understand and intuitive. Currently, almost all the OECD countries require some sort of labelling on processed foods, which may be attached to the front of the package (i.e. front-of-pack labelling - FoP), but is more often only attached to the back (i.e. back-of-pack labelling - BoP) of the packaged foods. While BoP labels tend only to inform about the nutrient content of foods, and sometimes may be difficult to read and understand, FoP labels are usually more visually appealing and intuitive. Available evidence concludes that FoP labels, such as the Nutri-Score labels in France, or warning labels in Chile, are more effective than BoP labels [Cecchini and Warin, 2016 [8]][Campos, Doxey and Hammond, 2011 [6]] Nevertheless, currently only four OECD countries implement mandatory FoP labels with virtually all the other OECD countries having mandatory guidelines only for BoP.

In the simulation, intuitive FoP labels are assumed to decrease average daily calorie intake by 1.16%. This effectiveness parameter is primarily based on an update to a recent meta-analysis [Cecchini and Warin, 2016 [8]] concluding that the implementation of food labelling for pre-packaged processed foods from industrial production can reduce average calorie intake by 2.21%. The analysis assumes that this effect represents a lower bound of the effects of FoP labels, as they are generally found to be the most effective label type [Cecchini and Warin, 2016 [8]][Campos, Doxey and Hammond, 2011 [6]]. Furthermore, account is taken of the fact that only a fraction of all calories come from the consumption of processed foods sold in stores and supermarkets. More specifically, evidence from the United States [Lin, Guthrie and Fraz{\~{a}}o, 1999 [38]] and Canada [Nardocci et al., 2019 [41]] suggests that only 80% of the total calorie intake comes from food that would be covered by this intervention (e.g. food sold in supermarkets or grocery stores). Furthermore, the policy is assumed to affect only the consumption of processed foods, and therefore will not change the consumption of foods such as as fruit, vegetables, legumes and meat. Based on [Monteiro et al., 2018 [40]], processed and highly processed foods account for about 65-66% of all calories consumed. The resulting change in calorie intake is converted into changes in body-mass index (BMI), by using the methodology developed by [Hall et al., 2011 [29]] and assuming an achievement of the maximum reduction after 100 days followed by a constant BMI thereafter [Strychar, 2006 [56]].

The intervention is assumed to target the whole population aged at least one year old as children are likely to be affected through the food bought by their parents. However, following previous evidence [Grunert et al., 2010 [25]], only 15% of the target population is assumed to be affected by the intervention as not everyone is likely to read and consider the information on the nutrition label.

The cost of setting up a new food labelling programme includes expenses on policy administration, planning and enforcement in the form of food inspections. This cost estimates vary between USD PPP 1.15 and USD PPP 1.30 annually per capita across the countries included in the analysis. Conversely, the analysis does not account for the additional costs associated with designing and printing nutrition labels or for the potential cost associated with the reformulation of certain foods, likely to be borne by the private sector. For example, graphic design and prepress for new labels can cost around USD 4 000 per item. In addition, there may be associated printing material costs (up to USD 6 100 per product) and labour costs (about USD 5 800 per product).