We now turn to the second set of hypotheses that emerge from our model, those regarding legislative organization and congressional delegation. Our intent is to place the study of legislative organization, and in particular the committee system, within the larger framework of the U.S. governmental process and to explore how the possibility of delegating authority to the executive alters legislators' incentives, behavior, and policy choices. The key prediction analyzed here is that, since congressional policy making and delegation are substitutes, albeit imperfect ones, inefficiencies in the legislative process should lead rational legislators to rely less on committees and to shift more discretionary authority to executive agencies.

We begin our analysis by first reviewing the different theories of legislative organization, concluding that Congress's most distinctive feature – its system of strong committees – must derive not from legislators' desire for informational or distributive benefits, but from incentives created by the larger constitutional system, including separate powers. We then test this theory in three ways: (1) by relating the presence or absence of committee outliers to interbranch conflict; (2) by exploring the impact of committee–floor conflict on Congress's decision to delegate; and (3) by examining the ways in which parliamentary devices like restrictive rules and multiple referrals affect agency discretion.