| Paper 1 examines the combination of a presumptive tax and environmental subsidy as an alternative to a direct tax on a negative externality. The government often cannot directly tax pollution effectively, either because a firm can avoid the tax or because it is difficult to measure emissions. Two general equilibrium models are constructed. A first-best model is built to establish the equivalence between the tax-subsidy combination and a direct tax. Another model is built to derive an explicit tax-subsidy solution when the government has a revenue-raising requirement. The tax and subsidy from the two models are then compared and interpreted in an environmental context. Paper 2 examines firm location decisions in the context of the environmental justice literature. Many studies argue that a correlation between waste facility location and poor or minority communities points to a disturbing trend of siting discrimination. However, these studies use current socioeconomic characteristics to assert discriminatory behavior, and thus do not accurately reflect variables that a firm takes into account when deciding where to locate. To properly analyze a firm's location decision, I examine the relationship between polluting plants established in Texas after 1975 and neighborhood characteristics at the time of siting. I use a conditional logit model and include variables such as poverty and race, as well as the cost of land and labor, and the presence of pre-existing plants. Paper 3 examines whether a plant varies the amount it pollutes according to the socioeconomic characteristics of the neighborhood in which it is located. To properly evaluate plant pollution behavior, two econometric issues are addressed. First, to control for the fact that a firm is not represented in the sample unless it is a heavy polluter, a sample selection model is used. Second, to address possible endogeneity between plant pollution and neighborhood characteristics, pollution is defined relative to a baseline. Papers 2 and 3 use census tract level data from the U.S. Census of Population and Housing matched to data from the Toxic Release Inventory, the U.S. Census of Manufactures, and the Texas Elections Database. |
