Oversight of the bureaucracy is arguably Congress's second-most important function, since lawmaking powers are undermined if Congress is unable or unwilling to monitor the execution of its legislation. Yet despite the theoretical importance of oversight, little is known about (1) Congress's capacity for overseeing the executive, (2) its interest in doing so, or (3) the effectiveness of this oversight. To address this gap, this dissertation probes these three subjects.

First, I posit that the existence of a bifurcated congressional principal (i.e., committees and subcommittees hold oversight hearings, but only Congress may sanction recalcitrant bureaus) limits which issues are taken up in hearings. This theory is tested with a new bureaucratic infractions database, allowing for the examination of which potential oversight topics result in hearings and which are ignored. As the theory predicts, a set of political variables is closely correlated with oversight activity; bureau-chamber convergence, bureau-committee divergence and committee-chamber convergence (in conjunction with increased gridlock) are all associated with decreased oversight.

Second, I examine representatives' interests in pursuing oversight. Leveraging new subcommittee transfers and subcommittee membership data, I show, respectively, that members of Congress tend to disfavor oversight subcommittee seats and that less powerful members are overrepresented on these bodies. Furthermore, House authorization and oversight subcommittees comprised of members that have lower seniority, less favorable floor rules applied to their bills and fewer legislative successes all tend to conduct more frequent oversight. Taken together, these findings suggest that representatives view oversight as a second-best alternative to legislation, pursued by those with less legislative influence.

Third, I employ genetic matching to examine the consequences of oversight, revealing that bureaucratic infractions covered in oversight hearings are 21.6% less likely to reoccur, compared to similar infractions that committees ignore. In addition, Congress may use the budget to punish overseen bureaus.

In brief, congressional organization and members' motivations are shown to reduce Congress's ability to conduct effective oversight. These limitations notwithstanding, oversight hearings can be effective, in some cases pushing bureaus to alter their behavior.