Abstract:
We leverage quasi-experimental variation to study how group size influences free-riding behavior within a high-expense environment. When buildings lack apartment-specific heat meters, tenants use simple heuristics to split a common bill. We estimate that the staggered rollout of a corrective technology, "submetering," reduces heating expenses by 17%, on average. Machine learning techniques uncover substantial heterogeneity, consistent with strategic exit of free-riders and coordination failures in large buildings. Tenants in smaller buildings show minimal response and are surprisingly price elastic. Only a minority of households exploits the free-riding incentives. Targeted submetering policies can be much more cost-effective than universal mandates.