Changing the incentive to pollute: Heterogeneous effects of waste pricing policies.

The externalities of household waste generation and treatment can be internalized with policies pricing unsorted waste disposal known as Pay-As-You-Throw (PAYT). To internalize pollution externalities of municipal waste, Pay-As-You-Throw (PAYT) pricing of unsorted waste are adopted worldwide. Using machine learning methods on a unique high-dimensional data panel of Italian municipalities, this paper estimates PAYT causal effects on municipal unsorted, recycling, and total waste heterogeneously with respect to a large set of municipal characteristics. I derive a household waste generation model that predicts effect heterogeneity, and I employ an R-learning random forest estimator that accounts for possibly endogenous adoption. Results show that PAYT is on average effective, decreasing unsorted waste per capita up to -60%, an effect driven by households preferring recycling (+41%) over total waste reduction. Both adoption and compliance are found to be heterogeneous in pre-policy waste levels, and explained by complex interactions of covariates. Continuous treatment analysis shows a modest and time-varying price elasticity of waste demands. Social costs savings are positive, and predicted to increase if all municipalities were to introduce the policy.