Air quality regulation and the reduction of toxic and greenhouse gases

The focus of this dissertation is to examine the effect of county-level air quality regulatory status on polluting behavior across counties. Ozone is regulated subject to the National Ambient Air Quality Standards (NAAQS) of the Clean Air Act. When a county is out of compliance (or out of attainment) for the ozone standard, the county implements a strict plan for reducing the concentrations of precursors to ozone which are volatile organic compounds (VOCs) and nitrogen oxides (NOx). I use county-level attainment status for 1-hour ozone as a proxy for air quality regulatory regime. Regulation of ozone creates a tighter regulatory climate that could spill over and lead to reduced emissions of a large range of pollutants (both regulated and unregulated), primarily those tracked by the EPA's Toxics Release Inventory. From estimation using panel data in a fixed-effects framework, the results provide support for the existence of spillovers as evidenced by the reduction of non-VOC emissions associated with non-attainment status of 1-hour ozone and by the reduction of unregulated industrial carbon dioxide emissions.;I also use county-level measures of pro-environment voting from the U.S. House of Representatives as a proxy for regional heterogeneity in preferences of citizens for more or less regulation in order to estimate their effect on toxic air emissions at a local level. Even though constructing county-level voting scores from congressional district scores requires a degree of approximation in counties that lie partially in multiple districts, the fact that county lines do not change with the decennial Census allows for measures of emissions activity in specific locations over time when using panel data spanning more than ten years. From estimation using panel data in a fixed-effects framework, the results suggest that allowing for regional heterogeneity in preferences at the county level can explain within-state variation in toxic emissions where state-level aggregates fail to identify such a relationship.