When Social Comparisons Backfire: Evidence from Carbon Footprint Feedback

Do social comparisons spur climate action, or do they deepen divides? Suppose you learned that your carbon footprint was higher than your neighbor's: would you try to lower it or on the contrary, would you give up?

In numerous settings, comparative feedback, which consists in informing households how their resource use stacks up against similar neighbors, tends to push high consumers to use less and low consumers to stay the same or use more, a pattern known as “regression toward the mean”.

In our study, based on a nationally representative experiment, we analyse the effect of comparative feedback about individuals’ carbon footprints on costly mitigation behavior (1).

We document two main facts: first, optimistic bias is widespread. Most individuals underestimate their relative emissions, meaning that they emit comparatively more than they think they do. Second, instead of the usual “regression toward the mean”, comparative feedback produces here a “divergence from the mean”: low emitters increase their mitigation efforts, while high emitters do not change their behavior.

Putting comparative feedback aside, we also find that mitigation behavior is unrelated to relative emissions or socio-demographic characteristics and is driven primarily by pro-environmental preferences.

The usefulness of comparative feedback

Social norm interventions (2) have become a central tool for policymakers seeking to foster pro-environmental behavior. These interventions typically rely on descriptive norms (information about what most people do) and/or injunctive norms (information about what most people approve of).

Descriptive norms can be sharpened through comparative feedback, which informs individuals how their own behavior compares to that of others.

Comparative feedback has been widely used to encourage reductions in household electricity consumption (3), water use (4), and to improve waste sorting (5). In these settings, households are informed about how their resource consumption compares to that of similar households. A common finding is that high consumers increase their mitigation efforts and low consumers reduce or do not change them - a pattern often described as “regression toward the mean.”

The originality of our method

In our paper, we study whether receiving comparative feedback about carbon footprint affects individuals’ efforts to fight climate change. We propose that a key source of heterogeneity in responses to comparative feedback lies in individuals’ prior beliefs about their relative contributions to climate change. 

To investigate this mechanism, we conduct a demographically representative survey experiment with 1,825 adults in Germany. We calculate respondents’ carbon footprints, covering all major sources of greenhouse gas emissions (excluding flights), and position each respondent within the national distribution of carbon footprints. 

We then ask participants to estimate how their carbon footprint compares to other people's. Many got it wrong: most assuming they were greener than they actually are, others thinking they emit more than they actually do. We then reveal to a randomly selected group where they truly stand, while the rest received no such information. This allows us to measure how learning the truth changes behavior.

To capture behavioral change, we ask participants to spend up to 100 euros buying carbon allowances from the EU's emissions trading system (6). In other words, they have the opportunity to spend money to offsett their emissions. For a random subset of participants, this is a real financial decision with real consequences for both their wallet and the climate.

A widespread and problematic underestimation of emissions

We find that beliefs about relative carbon footprints are distorted. Nearly two-thirds of respondents display optimistic bias while roughly one-third exhibit pessimistic bias. In other words, 2 people out of 3 underestimate their carbon footprints compared to others’. On average, comparative feedback has no significant effect on mitigation behavior. 

However, this average hides important differences depending on whether people are too optimistic or too pessimistic about their relative carbon footprint, and whether the news they receive is better or worse than expected.

When feedback delivers positive information, it strengthens mitigation efforts. This occurs among low emitters who learn that their carbon footprints are below average, with the effect concentrated among those with the lowest carbon footprint. A similar reinforcement pattern emerges among respondents with pessimistic bias who discover they overestimated their relative emissions. The effect is strongest for low emitters who hold pessimistic bias, as the feedback provides two aligned positive signals. 

In contrast, when feedback conveys negative information, it does not affect mitigation efforts. If anything, the estimates suggest the possibility of a backfiring response.

Key findings and contributions

First, we find no average effect of comparative feedback on mitigation. This contrasts with contexts such as household electricity use, where comparative feedback also signals potential cost savings. In our context, mitigation involves a personal sacrifice without direct financial return, which may explain why our results differ from previous studies. 

Second, our study highlights the importance of individuals’ beliefs. Although beliefs have only recently received attention as a source of heterogeneity in economic behavior, they are central to understanding how individuals process information, in our case, comparative feedback. Ignoring belief heterogeneity can obscure meaningful effects that remain hidden behind null average treatment effects.

Third, we contribute to the literature on belief correction. Studies where feedback relates to domains relevant to ego, such as intelligence and beauty (7), show that positive information leads to belief updates, while negative information tends to be disregarded. Our results mirror this pattern: people respond to positive feedback about their relative carbon footprint but do not react to negative information. To our knowledge, this is the first evidence of such asymmetry in climate change mitigation.

Finally, we add to the research on climate-related misperceptions. Prior work shows that people underestimate others’ awareness of climate change issues (8), support for policies (9), and mitigation efforts (10). We show a complementary bias: individuals systematically underestimate their own carbon emissions.

What are the implications and solutions for the implementation of climate policies?

Our results carry several implications for the design of mitigation policies that rely on social norms interventions.

Comparative feedback appears most effective when directed at individuals who already engage in mitigation efforts, as it reinforces their behavior. By contrast, it is unlikely to motivate - and may even discourage - those with relatively high emissions. For these individuals, alternative approaches such as public recognition (11) may prove more effective.

Moreover, the effectiveness of comparative feedback may be enhanced when combined with injunctive norms, as demonstrated in energy conservation contexts (12). 
 

References

(1) https://www.sciencedirect.com/science/article/abs/pii/S009506962500097X

(2) https://www.pnas.org/doi/10.1073/pnas.2426768122

(3) https://www.sciencedirect.com/science/article/abs/pii/S0047272711000478

(4) https://www.aeaweb.org/articles?id=10.1257/aer.101.3.318

https://direct.mit.edu/rest/article-abstract/95/1/64/58053/Using-Nonpecuniary-Strategies-to-Influence

(5) https://www.sciencedirect.com/science/article/pii/S0095069625000087

https://www.sciencedirect.com/science/article/pii/S0305750X26000586

(6) https://climate.ec.europa.eu/eu-action/carbon-markets/eu-emissions-trading-system-eu-ets_en

(7) https://www.aeaweb.org/articles?id=10.1257/aer.20180728

(8) https://www.nature.com/articles/nclimate1743

(9) https://www.nature.com/articles/s41467-022-32412-y

(10) https://www.nature.com/articles/s41558-024-01925-3

(11) https://www.sciencedirect.com/science/article/abs/pii/S0095069614000072

(12) https://www.jstor.org/stable/40064634

 Photo by Etienne Girardet via Unsplash