This dissertation investigates whether the Environmental Protection Agency (EPA) acts free from the influence of elected representatives when making enforcement policy choices, or whether external political actors affect the implementation of enforcement policy by the Agency. The dissertation addresses the debate in the literature over whether bureaucracies are free to make policy choices according to their own preferences, or whether they remain responsive to political overseers through mechanisms outside of the budgetary process. Chapter 1 presents a theoretical model that investigates whether the president's and Congress' preferences have a direct effect on the total number of enforcement cases taken by the EPA independent of budget. The model indicates that the Agency's principal on whom enforcement imposes the highest political cost may directly influence EPA enforcement outcomes above and beyond budgetary considerations, because of the threat of that actor sanctioning the Agency administrator once he chooses how aggressively to enforce. The preferences of this principal only constrain the EPA administrator's choice over how aggressively to enforce if the political cost imposed on the principal is high enough. If the political cost imposed by enforcement is relatively low, the administrator is free to direct the EPA to enforce as he chooses. In Chapter 2, a linear approximation of the theoretical model is estimated using OLS and multinomial logit techniques. The empirical model examines whether the president or the EPA's primary Congressional oversight committees significantly influenced the total number of enforcement actions taken by the EPA or the distribution of enforcement outcomes across the three types of cases available to the Agency, independent of budget, from fiscal year 1972 to 1995. The empirical estimation indicates that the preferences of the president and Congress regarding the environment affect both the total number of enforcement cases taken by the EPA and the case mix pursued by the Agency. The empirical estimation also indicates, to the extent possible given the available data, that the presidents influence was more pervasive both in the area of total enforcement cases and case mix.
My Research