This paper presents an empirical technique for valuing large changes in nonmarketed local attributes (e.g., climate amenities) without data describing prices of locally traded commodities like housing. A model of endogenous sorting is used to identify individuals' indirect utility functions, from which the value of the change in the local attribute is recovered while accounting for equilibrium impacts on markets for labor and locally traded commodities. Annual amenity costs of Brazilian climate change are estimated to be between $1.6 and $8.1 billion for a moderate climate change scenario, depending upon the role of migration costs. Copyright Blackwell Publishing, Inc. 2007