Price and modelling of food behaviours: new models for the evaluation of price policies – PRIMOFOOD

The PRIMOFOOD project aims to innovate in modelling the effect of food prices on household purchasing behaviour. It is in line with axis 1.5 of the AAPG2019 "Food and Food Systems". The proposed innovations will allow for a better assessment of the effectiveness and distributive effects of nutritional taxation policies. The identification of price effects can be based either on the econometric analysis of existing market data or on the analysis of experimental data generated in lab. Econometric methods may have limited internal validity, while experimental methods have questionable external validity. Our project therefore proposes to address the respective weaknesses of these two methods and, beyond that, to exploit their complementarities.
WP1 will develop innovative experimental analyses of price effects. First, using innovative protocols, we will analyze the effect of large price changes, similar to the levels used in micro-simulations of pricing policies, but much larger than the changes observed in market data. The objective is to identify possible salience and reference price effects, which can induce non-linearities and discontinuities in consumer reactions to prices. We will also test the potential complementarities between pricing policies and nutritional labelling of the NutriScore type. We will also study the effects of social norm, cognitive load, and the existence of opportunity costs.
WP2 will develop a structural econometric model of quality and quantity demand under multiple constraints. Beyond the usual budget constraint, the model will include a nutritional intake constraint, with possible extension to multiple linear constraints. The objective is to reflect that the choices of some households are constrained by the need to ensure a minimum of energy intake, or for some potentially addictive goods (alcohol and sugar). This econometric modelling work poses various theoretical and practical challenges. The empirical work will use Kantar WorldPanel scanner data on the consumption of non-alcoholic drinks by French households, and will aim in particular to identify the existence of effects of sugar habituation.
WP3 will aim to show the possibility of evaluating the ex-ante evaluation of pricing policies by a micro-simulation of taxation policies based on a model that integrates the outcomes of the two work packages. We will first cross-validate the experimental and econometric methods, comparing the econometric approach with the data generated in the experiments. This will allow us to identify opportunities to improve the specifications of the econometric model, an improvement that will be implemented by a Bayesian method. Finally, we will carry out micro-simulations based on the case of the taxation of sweetened drinks, which has been in place in France since 2012. This will allow us to better characterize the effectiveness of this tax, as well as its distributive and welfare impacts on the various socio-economic segments of the population.
PRIMOFOOD provides methodological contributions to the scientific communities and to important public policy issues. The team includes researchers covering the spectrum of issues addressed, from fundamental methodological issues to public policy expertise. The project proposes an international approach, with the ambition of replicating part of the work on American data, with a view to comparing food systems. Finally, the project will produce tools to better understand the links between the long-term dynamics of price changes and the sustainability of food systems.