Skip to content

What Makes An Administrative Agency “Democratic”?


Katharine Jackson (@kvj2108) is Assistant Professor of Law at the University of Dayton.

It is no question that many of the projects envisioned by LPE require an active state. But just at the time when we need our state to be at its most capable, its agencies – the tools it has at its disposal – are being undermined. And the critics aren’t beholden to any particular political ideology. Followers of LPE will be familiar with anti-administrativism’s right-wing attacks: that agencies, violating the rule of law and the separation of powers, can only be recalled from their constitutional exile either by subsuming them under an unboundunitary executive or throwing them to the dubious mercy of invasive, conservative courts.

Yet they are not left unscathed by the left. Blake Emerson, in his excellent book, critiques agencies for their lack of democratic accountability and transparency. Sabeel Rahman targets their inegalitarian and dominating implications. Many blame administrative technocracy for a variety of social and political ailments: the reification of social differences and the juridification of human nature; corruption and regulatory capture; the depoliticization of economic issues and the subsidization of globalized financial capitalism and, ultimately, the constellation of populist politics currently threatening liberal democratic states.

The solution generally offered is democratization. This is the right instinct. Yet it cannot mean opening up everything to direct citizen participation. These moves risk not just the efficiency required to implement democratic will effectively. They also risk reproducing within agencies the same social and economic inequalities that corrupt regular lawmaking. They may also rely too much on a “rational” public consensus about how policy is to made and implemented. It is unclear how our commitment to political equality is to be maintained if certain groups can self-select into newly-democratized administrative action. “One person, one vote” might be the baby thrown out with the undemocratic bathwater – leaving the residue of corporatism behind.

To be sure, Rahman’s republican conception of administration admirably takes on the dominating aspects of economic governance. But it, like most republican theories, does not make a positive case for collective democratic agency. It strives only to make that agency more palatable. It can explain why marginalized voices ought to have more say in building a Green New Deal – but not whether we should be able to have a Green New Deal to begin with. On the other hand, Blake Emerson’s deliberative democratic argument makes a positive case for the administrative state as the expression of the popular sovereign. But it relies on a consensus-oriented rationalism and leaves unspecified how public reason is to percolate into administrative action.

In any event, scholarship thus far has not reconciled the relationship between (democratized) agency policymaking and the regular lawmaking done by Congress. Indeed, most literature within political theory (Richardson’s 2003 Democratic Autonomy being a notable counterexample) takes an “agency instrumentalist” view. This is the idea that the “will of the people” is expressed in Congress and that agencies must, “like a machine without a mind,” implement the will of its sovereign principle. Nevertheless, as Emerson notes, agencies, just like courts, will inevitably make policy as they apply general rules to particular circumstances. Meanwhile, agencies’ instrumental reasoning will always confront values choices. As Rawls once noted, what amounts to “good” means is an essentially contested concept. Even capitalist firms have to decide whether to privilege short-term profits over long-term value.

To ameliorate the inexorable agency costs, theorists generally pose two different solutions: (1) a democratization of agency discretion, e.g., by making notice and comment procedures more robust; or (2) forcing Congress to elaborate their intent in fine-grained detail or undertake more robust oversight. Yet the first option opens Congressional lawmaking to relitigation behind citizens’ backs – an opportunity rarely passed up by corporate power. The second may reproduce the problem, as Congressional Representatives will inevitably rely on some “non-democratic” bureaucracy for any statutory specification.

Regardless, both moves inadvertently replicate a conceptual mistake committed by many anti-administrativists. This essay will rectify this mistake. Doing so will not only situate administration more comfortably within contemporary theories liberal constitutional democracy. It will also point to useful reforms.

Most normative scholarship on administration relies on an outmoded conception of popular sovereignty. It is one that holds that a country’s highest legitimate lawmaking power is held, literally, by a body with a will. To be sure, for much of modern Western history, sovereignty was thought to be held by a specific body: literally, the person of the monarch. Hobbes, for example, imagined a state of nature full of individual human beings forming a covenant to authorize and relinquish their natural liberties to a “Mortall God,” i.e., the modern state, represented (or re-presented) literally in the flesh of the king.

After the democratic revolutions, some theorists analytically merged the sovereign and her subjects. They imagined “the people” not only replacing the monarch as sovereign, but also governing itself as a subject. They thereby attempted to create an identity between ruler and ruled. Rousseau’s volonté générale is a model for this kind of thinking. It is a sovereign-subject hat trick: the ruled become the ruler, the democratic “people,” understood as an organic body, a “unitary macro-subject,” with a mind and will of its own. Carl Schmitt likewise endorsed this identity between governed and governor, although with some obviously alarming implications. So too does Woodrow Wilson, an early (racist) Progressive supporter of administration.

In contemporary legal politics, this unitary macro-subject is thought to be represented – however imperfectly – by Congress. As a result, administrative agencies retain legitimacy only if they remain loyal fiduciaries of Congress, itself the sovereign’s loyal fiduciary.

There are flaws in this string of identities. “The people” cannot rule. The entire population of a diverse community of hundreds of millions cannot be present within governing institutions. Nor would that population likely ever find a unanimous general will, no matter how constrained and qualified their public reasoning – nevermind one accurately reproduced on Capitol Hill. Indeed, when anyone ever presents her will as “the will of the (real?) people,” it’s time to pay attention – because we’re entering an authoritarian danger-zone. Power, as Lefort argues, must be an empty place if it’s not to erase dissent from existence.

Instead, it’s better to understand the popular sovereign not as a proper nouna body with a will that holds legitimate power. Instead, it’s better understood as a regulative ideal and standard of critique: what kinds of procedures will let all those individual people feel like they are co-equal participants in politics? Meanwhile, democratic lawmaking does not amount to taking dictation from pre-exising popular opinion. It is simply the output of decision-making procedures that, in Stephen Holmes’ words, “actually creates a framework in which the nation can for the first time have a will.”

Under this procedural understanding of popular sovereignty, we cannot assume beforehand that agency rulemaking is a usurpation of sovereign power. Because sovereignty is held neither by Congress or the President as the “voice of the people.”

This point is hammered down by the recent constructivist turn in theories of democratic representation. Political theorists like Jane MansbridgeMichael SawardLisa DischAndrew Rehfeld and Nadia Urbinati have argued convincingly that representatives do not, however ham-handedly, transcribe voters’ pre-existing “wills” into law. Instead, it is representatives (both elected and unelected) who construct their own constituencies and give shape to voter opinions. As evidenced by our unstable and irrational responses to political science opinion surveys, we do not, as an empirical matter, enter politics with a fully-formed set of exogenous preferences. Rather, through the educative function of electoral and grass-roots politics, we come to have preferences. Campaigns stimulate and frame public debate. They motivate citizens to form judgments between policy options. Representation includes top-down processes as candidates and organizational leaders make claims. Citizens respond in a bottom-up way, offering critiques or affirmations. The vote gives people a target for their judgments. It is also the only way to treat people as moral equals when part of the population thinks their rights require one thing, and another part thinks they require something else. The ballot is not a translation app, but a dispute resolution mechanism. 

Importantly, these theorists demonstrate that the norms that liberal constitutional democracy is supposed to instantiate – the right to equal participation in governance – does not imply any particular kind of collective decision-making procedure. Representation is not some watered-down, pragmatic alternative to normatively superior, but infeasible, direct participatory democracy. It is instead “an intrinsically modern way of intertwining participation, political judgment, and the constitution of demoi capable of self rule.”

Theorists of institutional design may therefore indulge their creativity. Because the yardstick is not “transcribe the will of the people” by having them, literally, transcribe their will into the statute books.  It is instead: “procedures that allow us to understand ourselves as contributing equally in policymaking,” or what theorists sometimes call “democratic autonomy.”  There is no good reason why administrative agencies cannot be included in the matrix of representative democracy.

What are the implications for agency design? First, it is well past time to relinquish an “idolatrous” commitment the separation of powers. It was always, in Madison’s words, an “auxiliary” device, an “invention of prudence,” that served a bigger goal: non-domination and efficiency. Moreover, the balkanization of function is logically impossible. Both courts and the executive will inevitably find themselves making contentious policy choices. The executive and the legislature will adjudicate social conflicts. Congress and the Courts will consider the implementation of public ends.

Instead, it may make more sense to divide responsibilities according to the contentiousness of the issues involved – regardless whether we characterize those issues as “means” or “ends.” When citizens are mobilized around an issue, resolving their conflict in a forum that allows each a (more) direct voice in a single national conversation may best serve democratic autonomy. Once litigated through national elections, they should not be retried in a different forum. On the other hand, decentralizing and democratizing local policy-making, so long as its effects are concentrated to an identifiable sub-national constituency, may also serve it. Meanwhile, excluding certain interests from one agency (e.g., the CFPB) when they are already (more than adequately) represented elsewhere (e.g., the FSOC) can ensure participation rights remain equal. Moreover, allowing payday lenders to write consumer financial protection rules amounts to religitation.

Nor should we wring our hands raw over the threat of technocracy. Congressional politics, like agency politics, likewise finds itself grappling with “the battle of the experts.” They negotiate the constraints of international financial markets and capital flows, budgets, and economic growth. Experts can be, and routinely are, incorporated into normal electoral politics. If anything, their claims to expertise can provide, as Emerson notes, a “constitutive tension,” as citizens (and their representatives) challenge received truths from would-be Guardians. One need only observe the fruitful work on MMT and heterodox Keynesianism – indeed, Keynesianism itself – to know that technique is fruitfully politicized. When it is, citizens deserve to weigh in formally by exercising their judgment as to whether one expert better represents their views than another. Meanwhile, insulating agency experts from lobbying pressure when there is no dispute about the quality of their knowledge can serve democratic autonomy. It gives government a fighting chance to actually put laws into effect – presumably the point of collective action. To illustrate, we might permit representative politics within a highly politicized decision to increase the deficit to stave off depression. The choice involves both moral and technical conflicts. On the other hand, we might insulate CDC doctors and IRS accountants unless and until a significant opposition is mobilized.

I only gesture at a few ideas here. But it should be clear that relinquishing the hold of “popular sovereignty” over our imaginations can only liberate us as we imagine an inclusive, egalitarian, and healthy future.