Research

We study how households respond to being defaulted into an indexed-price electricity contract after a period under fixed prices, using high-frequency data from a major Italian utility during the 2021–2023 energy crisis. Quasi-random timing of the default generates large, unanticipated price shocks. Defaulting increases bills by up to 56 percent, while consumption declines slowly, by up to 9.8 percent after eighteen months, implying low expenditure elasticity. Welfare analysis shows that fixed-price contracts trade efficiency for protection: they distort consumption and generate deadweight losses, yet provide implicit subsidies that are regressive in absolute terms but progressive relative to income, highlighting central equity–efficiency trade-offs.

In recent years, a global movement against climate change has been rising. Existing evidence on how these political events impact real-world environmental behavior is scarce. In this paper, we employ electricity-use data covering 1.5 million Italian households over the period 2015-2019 to show that climate strikes cause a significant reduction in energy consumption, although these effects are modest in magnitude and short-lived, as is typical of salience-induced behavioral changes. A granular measure of social media attention to climate change, derived from universal-coverage Twitter data, and sentiment analysis suggest that climate protests direct public attention towards climate change topics and foster emotions that are strong motivators for action, such as anger and positive feelings of solidarity. These results suggest that episodes that heighten attention to climate change may lead to actual behavioral change, but their effect is small and short-lived. 

We randomize the installation of air purifiers across primary school classrooms to reduce children's exposure to air pollution. The intervention reduces indoor fine particulate matter (PM2.5) daily concentrations by 33% (18% during school hours) and decreases student absenteeism by 17%. We find larger effects among students with higher pre-treatment absenteeism, and when ambient pollution levels are low enough for purifiers to bring indoor concentrations below health-relevant thresholds. Treated students report fewer respiratory symptoms and exhibit greater awareness of air quality. The intervention has no effect on cognitive skills, mood, and aggressive behavior. Each avoided absence day costs approximately Euro 6.9, yielding a benefit-cost ratio of about 9.5:1, even ignoring potential benefits to academic performance.

In developing countries, community-driven development programs often partner with indigenous associations under the assumption that their members possess social capital that facilitates collective action. Yet it remains unclear whether such capital is portable across settings, or whether cooperative advantages are specific to longstanding social relationships. We study informal indigenous associations, social clubs known as grins, in urban Mali using a field experiment involving 635 groups (6,245 individuals). We implement step-level public goods games with two contribution thresholds in three environments: within existing grins, in groups composed of grin members drawn from different associations, and in groups of non-members. Existing grins exhibit contribution rates that are 3–10 percentage points higher across orders of play. They are more likely to meet contribution thresholds in initial interactions. To help distinguish membership effects from relationship-specific effects, we compare grin members and non-members when both play in stranger groups, and we separately compare grin members playing within versus outside their own associations using inverse probability weighting to adjust for observable group characteristics. We find little evidence that grin members contribute more than non-members when placed in mixed stranger groups. By contrast, cooperation remains higher when individuals interact within their own grin, consistent with relationship-specific social capital rather than portable membership effects. Higher contribution thresholds increase contributions but reduce provision success across environments. The cooperative advantages of indigenous associations appear to be tied to embedded social relationships rather than membership alone, with implications for the design of decentralized public good provision in low-capacity settings.