Agency policy making and political control: Modeling away the delegation problem

Abstract

In recent years, there has been a noticeable shift in some of the literature on political control of the administrative agencies, away from an emphasis on agency autonomy and toward the view that politicians can and do exert significant amounts of control over agency decisions. This shift has been particularly evident in the mainstream political science journals and includes both theoretical and empirical work. I argue here that this shift in thinking may be premature, and that these recent works do not really challenge our understanding of why widespread agency autonomy is an inherent byproduct of the delegation of policy-making authority from politicians to agencies. Much of the new theory of political control comes from positive theorists, who emphasize the ability of politicians to shape agencies' decision-making environments in ways that steer agencies' subsequent decisions toward the politicians' policy goals. Much of the new empirical work comes from quantitative empiricists whose statistical models of agency behavior find significant amounts of agency responsiveness to external political stimuli. Neither of these groups, however, demonstrate that politicians can overcome the delegation problem; rather, they model the problem away. Positive theorists do this by assuming away the importance of substantive policy foresight to politicians' efforts to influence subsequent agency decisions. Quantitative empiricists do this by choosing as their dependent variable agency implementation of established policies, rather than agency policy making itself. While both are driven to this result by their own separate methodological imperatives (and a commendable preference for scientific rigor), the result is the same. Neither group of models is sufficiently rich or generalizable to challenge the notion that agencies act with a great deal of autonomy when exercising most delegated policy-making authority.