Concentrating implementation: How ideological agreement between the House and Senate affects delegation to the bureaucracy

By Jordan Carr Peterson

Many scholars and observers of American legislative politics focus on the substantive public policies contained in laws Congress enacts. For instance, oceans of ink have been spilled explaining the failure of congressional Democrats to include a public option in the Affordable Care Act. This emphasis on the policy content of law is self-evidently justified as legislation is the most obvious and most central output of a legislature. In a recently published study, however, I argue that it is also important to consider the manner in which Congress designs a system of implementation for the policies contained in the laws it passes.

Developing a theory of why Congress writes some policies into law but not others is critical for understanding the legislative component of the policy process. To cease consideration of congressional enactments at the policies they prescribe, however, ignores some crucial (if nuanced) realities of policy implementation in American government. Namely, this limited focus neglects to account for how Congress structures the role of the federal bureaucracy in implementing the policies the legislature creates. It mistakenly punctuates the analysis of legislative policy-making with a period instead of a comma.

In pursuit of a more thorough examination of the legislative process, I offer a theory of how Congress designs the bureaucratic implementation of public policy that builds on earlier scholarship in political science. My argument rests on the foundational notion that in the institutional expanse of the contemporary administrative state, federal agencies predominate in carrying out the public policies that Congress writes into law.

I argue that the legislative decision to delegate implementation authority for public policies to fewer versus more administrative agencies is a meaningful choice. The more bureaucratic officials there are charged with the task of implementing a given policy, the more likely implementation efforts may be frustrated by what economists and political scientists term “agency loss.” According to this concept, there is a principal (here, Congress) that relies on an agent or set of agents (here, federal bureaucrats) to see some outcome achieved (in this instance, public policy implemented). The involvement of more agents (bureaucrats) means a greater likelihood of agency loss representing a substantial hurdle to successful policy implementation – the policy design equivalent of why coordinating group projects seems to become more difficult when more people are added (assuming no one is allowed to free ride).

Why, then, is there any variation in how Congress delegates the responsibility for implementing the policies it creates? In other words, do circumstances exist in which Congress would more strongly prefer to see the laws it enacts successfully implemented than others? I argue and find that there is an ideological explanation to these questions, rooted in the legislative branch’s bicameral design.

I contend that Congress is more likely to concentrate implementation authority in fewer administrative agencies when there is greater ideological proximity between the House and the Senate. I rely heavily on Keith Krehbiel’s theoretical framework of pivotal politics, which suggests the distribution of ideological preferences within and across institutions goes a long way in determining those institutions’ policy outputs. According to this theory, pivotal actors are those at the ideological “tipping points” within institutions – i.e., the median member of the House, which requires only a simple majority to pass bills, and the so-called “filibuster pivot” in the Senate, with its functionally supermajoritarian vote requirement – whose preferences determine the collective policy preference for the body as a whole.

I argue that concentrated delegation of implementation power is more likely given greater ideological proximity between chambers of Congress because enacted legislation will be closer to the sincere preference of the pivotal member of one chamber if that member’s ideology is more proximate to that of the pivotal member in the other chamber. More concretely expressed, it is more likely that legislation will contain conservative (or moderate, or liberal) policy prescriptions if the pivotal members in both chambers are conservative (or moderate, or liberal) than otherwise. Quite simply, intercameral ideological congruence obviates the demand for intercameral policy compromise.

Because pivotal members of Congress are ultimately more likely to prefer policy outcomes given minimal ideological distance between chambers, it stands to reason that legislators making delegation choices under such conditions would view agency loss resulting from unsatisfactory policy implementation by bureaucrats more negatively than were those policies more remote from the legislators’ preferences. Indeed, facing the enactment of laws that specify policy programs they regard as suboptimal, legislators may even positively anticipate the potential for agency loss at the hands of the bureaucracy, and thus seek to avoid concentration of implementation authority. These considerations should motivate members of Congress to concentrate authority delegated to the bureaucracy among fewer agencies – i.e., to “put all their eggs in one basket” – given greater ideological congruence between the two chambers.

To test my argument’s claims, I collected information on all legislation in David Mayhew’s dataset of important congressional enactments from 1947 to 2012. Using Poisson regression – an econometric technique that estimates countable outcomes – I created a statistical model to predict the number of federal administrative agencies to which Congress will delegate policy implementation authority given variation in ideological proximity between pivotal members of the House and Senate.

I found that moving from a relatively low to a relatively high level of intercameral ideological distance, the predicted number of agencies receiving delegated implementation authority moves from 3.96 to 6.31. Because the statistical analysis controls for a host of other political, structural, and economic factors that may bear on legislative delegation choices, the results suggest that all else equal, congressional management of designing a policy implementation scheme is strongly associated with the distribution of ideological preferences within Congress.

Criticisms of administrative governance typically contend there is a democratic accountability deficit when policy is made by bureaucrats, resulting in the divorce of policy outcomes from popular preferences. If citizens elect legislators to design policies that reflect their own preferences, then alterations to congressional enactments made by unelected bureaucrats represent a subversion of the democratic will. Contemporary policy implementation, however, necessarily relies on satisfactory performance by a cooperative (or coerced) bureaucracy. My research suggests that legislators can circumvent the problem of agency loss by distributing implementation powers across fewer bureaucratic institutions, and my analysis indicates that such concentrations of authority are more common given ideological congruence between the two chambers of Congress.

Jordan Carr Peterson  (@JordanCarrP) is assistant professor of political science at Texas Christian University.