Labelling food for a demand-driven transition towards sustainable consumption
The food sector is responsible for a substantial share of the anthropogenic impact on our living environment. To counteract this, the focus so far has mainly been on efficiency improvements, including waste- and input reductions. Besides this, however, an even greater leverage would lie in consumers’ dietary patterns. Affluent food consumption patterns are characterized by increasingly inferior nutritional qualities and adverse environmental impacts. Therefore, the use of sustainability labels on packaging has seen unprecedented growth in recent decades, which has resulted in a confusing multitude of labels. Consumers make very little use of them in their food choices. Besides harmonising the current multitude of sustainability labels, additional measures are needed to actually bring about a change in food consumption behaviour. This dissertation explored the potential of an overall environmental impact score to guide behaviour, as well as various limiting factors and complementing interventions that may promote further behavioural change.
Chapter 1 provides a general introduction, elaborating on factors influencing food choices, possible interventions to adjust these, and a general theoretical framework on the effect of labels on food choices. Next, chapter 2 provides a description of recent developments and challenges in environmental impact labelling of food.
Chapter 3 comprises a first empirical study. This chapter starts from the problem that consumers today can take into account different, seemingly overlapping, aspects in order to account for the environmental impact of their food choices. To this end, a discrete choice experiment is applied on vegetable consumption. In addition, the study tries to reinforce proenvironmental preferences by evoking a (non-)sustainable self-view. We found a strong preference for good Eco-Scores, about as strong as for lower prices and local origin. Furthermore, an aversion of strictly seasonal consumption was observed, and organic labels appeared to be least important. Finally, we find that too much confidence in a sustainable self-view leads to less sustainable preferences than when one has some doubt in this view.
Chapter 4 examines how the joint display of Eco-Scores and NutriScores influences food choices in an E-grocery. In addition, general and specific consumption recommendations are tested. Product images were manipulated in order to evaluate their inhibiting role on the abovementioned interventions. Jointly displaying Eco-Scores and NutriScores led to improvements in nutritional values, but not in environmental impacts. Similar improvements were attributed to the general recommendation, while the specific recommendation led to both nutritional and environmental improvement. The product images did not interfere significantly with any of the above effects. This study highlights the challenge of competing interests when making food choices, as well as the potential of alternative, less costly interventions to achieve behavioural change.
Then, part of Chapter 4 is replicated in Chapter 5. Again, the joint display of Eco-Score and Nutri-Score is tested in an E-grocery. Several supporting digital functionalities are added, such as product recommendation agents, personalised feedback on environmental impact and nutritional values of choices, and a social norm on these values. As in Chapter 4, the results again indicate that displaying Eco-Score and Nutri-Score initially only improves nutritional values and not environmental impacts. However, in combination with an appropriate recommender system, environmental improvements could also be attributed to the scores. This underlines the importance of an enabling food environment for shifting behaviour, in addition to a harmonised labelling system.
Chapter 6 aims to extrapolate the previous chapters to physical purchasing behaviour, in a mock-up fishmonger store. Besides the manipulation of an Enviroscore, information on the meaning and use of the score was also manipulated. This study also evaluates the difference in effects between filleted packed fish versus fresh unprocessed fish. Finally, it investigates to what extent observed effects could be considered reflective. Displaying an Enviroscore alone, here without a Nutri-Score, increased the choice for environmentally friendly alternatives by 15%. This effect did not differ between the two processing types of fish. Displaying information on the score, however, completely cancelled out this effect. Moreover, effects of the score on food choices were mostly automatic and unmotivated. This study confirms the potential of an environmental score, when not displayed together with other scores and provides some first formal insights in the type of observed effects.
Chapter 7 brings together all research chapters and provides some policy recommendations on sustainability labels and behavioural change interventions more generally. Overall, this PhD concludes with a relatively promising potential of environmental impact scores to guide food choices which is, however, conditional on several factors. As such, competition with other food attributes remains, as well as the need for more physical facilitation of healthy sustainable food choices. In addition, mere unmotivated responses to environmental scores require a visual aggregation of different sustainability dimensions in order allow for evaluations at a glance. Therefore, sustainability labels should not be seen as silver bullet, but rather as part of a broader policy approach. Consumers’ behavioural change is rather unlikely to occur in a system that does not move along. An actual food system transformation requires a shift in mindset of all actors throughout the value chain.