Conference Agenda

We present an economic experiment where two groups interact in a linear repeated public goods game. In one treatment, groups give to a public good. In the other treatment, groups take from a public good. A large literature has documented that this difference in framing can lead to substantial differences in cooperation in pay-off equivalent social dilemmas. Here, we study the existence and persistence of such a give-take framing effect when groups, not individuals, decide. We find large differences in cooperation in the first supergame of the experiment, but no difference in cooperation in the second supergame. Analyzing the group's decision-making process and their chats, we find that the results can best be explained by groups strongly reciprocating (treatment specific) initial defection in the first supergame, and (irrespective of treatment) adopting a payoff-maximing treshold strategy of initial trust building and last-round defection in the second supergame.

Information provision is a common policy response to negative externalities with low mone-

tary implementation costs, but a large literature shows that it can impose psychological “moral

tax” costs (e.g., guilt, shame, or social pressure), leading some individuals to strategically avoid

information to preserve a moral self-image. Existing work shows that accounting for this in-

formation (dis)utility can substantially change welfare conclusions in cost–benefit analyses of

information-provision programs. We know much less, however, about how the general pub-

lic evaluates such preferences—an issue central to a complete welfare analysis of information

provision.

We conduct a preregistered experiment with almost 4,000 impartial spectators from the U.S.

general population. Spectators decide whether to intervene by providing information about a

negative externality to an uninformed decision-maker, knowing (i) whether the decision-maker

wants to be informed or prefers to remain uninformed, (ii) whether the action imposes a negative

externality on a third party, and (iii) whether information would change the decision-maker’s

choice in a pro-social direction.

We find that 50% of spectators intervene even when the decision-maker prefers to remain

uninformed and information does not affect behavior. We classify spectators’ motives and

find substantial heterogeneity: the largest group consists of “moralists” who provide information

even when doing so is neither consequential nor desired, while roughly one-third are

welfare-motivated types who intervene primarily when doing so affects outcomes in a pro-social

direction and/or when the decision-maker wants information. These results suggest that information

policies may enjoy broad moral legitimacy even when they conflict with decision-makers’

preferences and have limited effects on reducing externalities. Furthermore, they highlight a

fundamental tension between welfarist and deontological perspectives in evaluating information-

based interventions.

We explore the interplay between two key individual drivers of green consumption: intrinsic moral concerns for the environment and reputational concerns for social image. Our microeconomic behavioral model characterizes choices among lifestyles differing in environmental impacts (brown/green) and conspicuousness (positional/discreet), depending on how strongly one values each of these motives.

We show that image concerns can substitute for environmental concerns in driving green consumption across a limited but central range of preferences, in particular through the purchase of green positional goods.

Such conspicuous conservation can green individual consumption (\textit{reconciling Eco and Ego}), especially among image-sensitive consumers, but it yields environmental benefits only under specific economic conditions. Indeed, the environmental impact of a lifestyle depends critically on its \textit{relative impact intensity}, i.e., the pollution per dollar spent on this lifestyle, more than on the pollution per unit of the representative good of the lifestyle, driving volume effects and behavioral rebound effects, which both reduce the environmental benefits of green lifestyles. Knowing the collective distribution of preferences may help design targeted policies, as those preferences strongly determine policy effectiveness. Our findings are especially relevant for policies that aim to foster greener consumption choices in different economic contexts (e.g., green nudging, environmental taxes with higher rates on positional goods...).

Many environmental policies rely on information campaigns to promote voluntary participation in socially desirable activities. This paper studies the effectiveness of such campaigns in the context of pharmaceutical waste recycling in France, where returning unused medicines is voluntary and managed through a delegated eco-organization. We develop a simple theoretical framework in which individuals decide whether to recycle by comparing an individual-specific valuation of recycling with a perceived cost that includes time, effort, and informational frictions. Using repeated cross-sectional survey data, we exploit a standardized informational script delivered during interviews to all respondents who report not recycling medicines and examine subsequent stated intentions to change behavior. We find that the informational message is associated with higher intended recycling only among individuals with high valuation of recycling, while individuals with low valuation remain largely unresponsive. These findings highlight intrinsic limits of information-based policy instruments and help explain why uniform campaigns implemented through delegated governance structures often generate modest aggregate effects, underscoring the role of behavioral heterogeneity in environmental policy design.