A Generalized Theory of Hedonic Estimation Using Experimental and Quasi-Experimental Designs

This paper develops a generalized framework for hedonic estimation that exploits experimental and quasi-experimental designs to relax many assumptions required by previous approaches. In the context of an idealized experiment, the identifiability is established of the distribution across consumers of the Marginal Willingness to Pay (MWTP) for a given product-specific amenity with few restrictions imposed on preferences, technologies, or the market. Conditions are presented for identification of alternative definitions of amenity value in the context of quasiexperiments and field experiments, and tradeoffs between modeling assumptions and data availability are discussed.