Document Type



Administrative Law | Constitutional Law


This Article argues that recent United States Supreme Court decisions invalidating agency policymaking rely on a normatively unattractive and empirically mistaken notion of democratic popular sovereignty. Namely, they rely upon a transmission belt model that runs like this: democracy is vindicated by first translating and aggregating voter preferences through elections. Then, the popular will is transposed by members of Congress into the statute books. Finally, the popular will (now codified), is applied mechanically by administrative agencies who should merely “fill in the details” using their neutral, technical expertise. So long as statutes lay down sufficiently “intelligible principle[s]” that permit their application without significant discretionary remainder, regulation will carry democratic legitimacy because they who will the ends—the people—will the means.

Although the transmission belt model comes in procedural, deliberative, and populist varieties, each occludes the inescapable facts of political intermediation and social conflict. No matter how participatory and rational, governance will involve some officials somewhere making decisions that some citizens will dislike. Even after citizen preferences are massaged by public reason or netted out through compromise, it is impossible to speak of a single, identifiable popular will that is capable of translation into legislative codes and regulations. Collectively, voters do not behave in a way that makes them a plausible principle to a government agent.

Given these empirical problems, the transmission belt model carries several unsavory undemocratic implications. For example, it can lead to a Schmittian repression of social difference, should a demagogic leader claim to speak with the (fictional) voice of the (fictional) people. It may lead to juristocracy, as judges and technocrats claim unique insight into the substantive contents of public reason.

The Article suggests instead that models of democratic political representation serve as better criteria to assess agency legitimacy. Theories of representation recognize and take advantage of the institutional mediation of democratic input. They protect and accommodate social conflict. They recognize, unlike pluralist, deliberative, and transmission belt theories of democracy, that there will always be a gap between ruler and ruled and that there will always be officials making decisions that people may not like.

Drawing from theories of political representation within political science and political theory, the Article argues that administrative agencies themselves can serve as democratic representatives because they (1) incorporate citizen presence in their decision-making while (2) preserving the normative priority of the citizen. The Article then suggests that the oft-maligned trustee model of representation, although often associated with the paternalism of Edmund Burke and Federalist 10, can serve as an appropriate evaluative standard. Agencies’ historical commitment to an inclusive notion of the public good, as well as their dedication to the public interest (qua beneficiary) over the often partial and self-serving commands of elected officials and powerful lobbyists (qua authorizers), make the trustee model an attractive starting point. With some democratic modifications that account for deliberation and debate about the meaning of the public good, the trustee model shows why agency decision-making has strong democratic credentials.

The Article concludes by offering some modifications to the legal doctrine used by courts to assess the legitimacy of agency action, to interpret agencies’ organic statutes, and to shape the presidential removal power. If we demand that administration provide good trustee representation, we will not expect it to mechanically carry out the orders of elected representatives. Rather, we will give them sufficient autonomy to carry out their authorized mandates diligently, loyally, and in good faith.