Spatial Externalities, Inefficiency, and Sufficient Statistics

How much economic inefficiency is generated by spatial externalities such as agglomeration or congestion? What can we learn with data and variation around an inefficient equilibrium? We express deadweight loss in an equilibrium model with spatial externalities, building on Harberger (1964). Our expressions depend on two empirical objects, an externality matrix and a Slutsky substitution matrix. This provides a basis for assessing how modeling assumptions, especially related to patterns of demand substitution and the variation used for estimation, may constrain inefficiency conclusions. We illustrate extensions to our approach with two examples: electric vehicle chargers and peak-hour traffic congestion.