The Collective Risk Social Dilemma: A Meta-analysis. [Link] with Matt Walker and Till Weber.
The collective risk social dilemma (Milinski et al., 2008) is frequently used to simulate the collective action problem posed by climate change: players repeatedly make costly contribution decisions under strategic uncertainty to reach a collective target and avoid a probabilistic loss. Numerous behavioural experiments have investigated how the costs and benefits of contributing (e.g., group size, probability of loss), features of the decision-making environment (e.g., inequality, uncertainty), and interventions (e.g., communication, intermediate milestones) affect collective success. We pool data from 42 published and unpublished studies, comprising 168 treatments and 14,077 participants, and conduct a meta-analysis to quantify the relative effects of these factors. We introduce the Collective Risk Ratio, a novel index that captures the strength of incentives for collective action generated by different experimental designs. Our treatment-level analysis shows that the Collective Risk Ratio predicts successful collective action. After controlling for the strength of incentives, we find that larger groups and uncertainty about key game parameters substantially reduce collective action. We also collect and re-analyse available participant-level data (12,413 participants) to explore within-group contribution dynamics. Early contributions strongly predict collective success and are themselves primarily influenced by the Collective Risk Ratio. The findings have timely implications for climate negotiations and underscore the important role of economic incentives in promoting collective action.
Working papers
How Do Income-based Social Norms influence Environmental Decision-making? [Link] with Smriti Sharma, Matt Walker, and Till Weber.
The prevalence of efforts to combat global warming is often underestimated, and existing research shows that generalised norm (e.g., population averages) information can increase individuals’ willingness to act. However, the relevant benchmark may depend on income. Because income shapes both the constraints of environmental action and the fairness concerns attached to it, individuals may respond differently to norm information about their income in-group and out-group. This study examines whether income-based norm information affects beliefs, incentivised environmental donations, and support for redistributive climate policies among high- and low-income UK adults. In an online experiment with 603 high-income and 606 low-income participants, they were randomly assigned to receive self-reported information about climate engagement among either their own income group or the other income group, or to a control condition without information. The results suggest that providing norm information significantly shifts beliefs about specific pro-environmental behaviours. However, we do not detect statistically significant treatment effects on environmental donations. Donation responses nevertheless vary with in-group closeness and inequality aversion, with these patterns differing across income groups. These findings suggest that income-based norm information can correct misperceptions, but its behavioural effects are conditional on the group identities and fairness concerns through which such information is interpreted.
When Later Action is Cheaper: Coordination under Collective Risk
In many collective-action problems, such as climate change mitigation, exogeneous technological progress can reduce the future cost of action, creating incentives to delay. Delay may be especially consequential under collective risk, where groups must reach a shared contribution target before a deadline to avoid a substantial risk of loss. Whether declining future costs alter contribution timing and collective success, however, remains unclear. I examine this question in a ten-round, real-time interactive online experiment using a collective-risk social dilemma (CRSD). I randomly assigned 496 participants (124 four-person groups) to conditions in which contribution costs either remained constant or declined across rounds. Declining costs significantly shifted contributions towards later rounds. Despite a unique symmetric cost-efficient equilibrium requiring the same minimum expenditure as under constant cost, collective success fell substantially. The findings suggest that lower future contribution costs can therefore be counterproductive when it makes delaying contributions more attractive.