It is just before 13.00. Hungry, you rush to the UvA canteen after a long class. Same steps and same classmates as many other days, but today, something has changed. Today on the menu of each stand in the canteen, there are traffic light labels, one label per meal option. Dishes with beef and lamb are red. A lot of cheese, and again the label is red. With fish and chicken, the labels turn yellow. Salads and other vegan dishes are green. Below the colors, you can read the grams of CO2 associated with each option. Now that you have all this information, will you opt for a more sustainable meal?
A few years ago, university students in Cambridge and Bonn faced exactly this choice: traffic light labels indicating the carbon impact of every meal suddenly appeared in their canteens. Lohmann et al. (2022) estimate that, in Cambridge, the labels reduced average CO2 emissions per meal by 4.3%. Schulze-Tilling (2024) shows that in Bonn the labels reduced by between 2 and 7 percentage points the probability that students buy a meat-based meal (which was, on average, among the most polluting options).
Can carbon labels be part of the solution to fight global warming? The results from the university canteen seem encouraging. Sure, a few percentage points reduction in CO2 emissions from food will not stop climate change. But introducing labels is cheap, so it might be worth the effort. Yet, university students are young, highly educated, and generally concerned about climate change. Can carbon labels push the general population towards greener diets as well?
In a recent working paper (Imai et al., 2023), we investigate this question. We move from the idea that carbon labels work because people would like to reduce their footprint, but they don't know how much each of their actions pollutes. The labels correct consumers' misperceptions about CO2 emissions and allow them to make more sustainable decisions. This knowledge-to-action mechanism is behind many policy interventions. For example, the European Commission stated that to facilitate the green transition, it "aims to ensure [...] that consumers have better information to be able to make an informed choice."
Our research project answers the question in two steps. First, we identified the products for which correcting consumers' misperceptions should generate significant changes in behavior. These are the products for which a) many people think the emissions are lower than they are, and b) the people who make this mistake care about climate change. It turns out that people are worst at guessing the emissions generated by more polluting products. Hence, we predicted that one of the most promising targets of information is beef – one of the most polluting foods. Meanwhile, we predicted no consumer reaction to information about a similar product like poultry meat. Chickens are about ten times less polluting than cows – mainly because cows produce large amounts of methane, a potent greenhouse gas. These predictions rely on data from a representative sample of US residents.
In the second step, we tested if these predictions were correct. To do so, we ran an experiment with another large sample of US consumers. In this experiment, we asked participants how much they were willing to pay for a shipment of meat to their home address. The shipment was real, and people who wanted to receive it needed to pay for it. We divided our participants into four groups: chicken-with information, chicken-without information, beef-with information, and beef-without information. The information told the participants about the average emissions associated with the production and distribution of one pound (0.45 kg) of each type of meat. Participants engaged with the information: If they learnt how, they started thinking that poultry meat was also very bad for the environment. Yet, the information doesn't reduce the amount the participants want to pay for the meat. This null effect of information was expected for chicken. It is contrary to our predictions for beef.
Our results are bad news for carbon labels. Beef is one of the products for which we should see the largest effects of labels, yet we can't detect any. Our sample is representative of the US populations (according to age and gender); we can exclude that the null result is driven by inattention, confusion, lack of trust in the scientists, participants that don't eat meat, or meat being a product whose consumption is hard to cut. We also don't find any self-reported reduction in meat consumption 2 weeks after the intervention. Overall, our results indicate that increasing the general population's knowledge about food carbon footprint does not result in more climate-friendly diets.
Yet, not all hope is lost. Our intervention targets the misperception channel, but there is another reason why carbon labels might push people to make greener choices. Schulze-Tilling (2024) shows that students' food choices change if they are forced to think about CO2 emissions without receiving any new information about them. In her experiment, this salience-of-emissions mechanism is responsible for most of the effects of carbon labels. This salience channel is muted in our experiment as we mention CO2 emissions and their consequences to every participant, including those that don't receive the information about the exact emissions associated with the meat they can buy. The salience of environmental impact is also important for other decisions. Fang et al. (2023) show that shower meters that give real-time feedback about water temperature and usage reduce energy consumption. This study took place in students' dorms in Bonn. However, the salience channel has yet to be tested with the general population.
Taking stock, carbon labels could have pushed you to choose a more sustainable meal today. Mostly because they would have made you think – and maybe feel uneasy - about the emissions you were about to cause. Yet, if this is the reason, reading this article before going to the canteen might be as good as seeing a traffic light label next to your favorite food.
Sources:
Fang, X., Goette, L., Rockenbach, B., Sutter, M., Tiefenbeck, V., Schoeb, S., & Staake, T. 2023. Complementarities in behavioral interventions: Evidence from a field experiment on resource conservation. Journal of Public Economics, 228, 105028.
Lohmann, Paul, Elisabeth Gsottbauer, Anya Doherty, and Andreas Kontoleon. 2022. “Do Carbon Footprint Labels Promote Climatarian Diets? Evidence from a Large-Scale Field Experiment.” Journal of Environmental Economics and Management, 114, 102693.
Imai, Taisuke, Davide D. Pace, Peter Schwardmann, and Joël J. van der Weele. 2022. “Correcting Consumer Misperceptions About CO 2 Emissions.” CESifo working paper 10138.
Schulze -Tilling, Anna. 2023. “Changing consumption behavior with carbon labels: Causal evidence on behavioral channels and effectiveness.” Mimeo, University of Bonn.