How much power may Congress delegate to federal agencies?

Editor’s Note: This essay first appeared in the Originalism Blog. The subject is very important, but the essay is rather technical: It discusses the extent to which Congress may delegate responsibilities to federal agencies. Some scholars have argued recently that most modern delegation is unconstitutional because it grants legislative, and often judicial, power to the executive branch. Those scholars are seeking to resuscitate the Supreme Court’s almost moribund non-delegation doctrine. However, a new article by Julian Mortenson and Nicholas Bagley, two University of Michigan professors, claims there is no non-delegation doctrine—that Congress may delegate any powers it wishes to the executive so long as it does not “alienate” those powers permanently.  As the following essay demonstrates, I think neither position is completely accurate.

*****

Julian Mortenson and Nicholas Bagley deserve credit for sharpening thought on the non-delegation principle. They certainly have sharpened mine. They also have collected useful instances of early congressional delegation, which advocates of the non-delegation principle will have to address.

My purpose here is not to rebut their paper nor to define the limits of the non-delegation doctrine. As becomes evident below, I do not think there is a single, coherent non-delegation principle. Rather, my purpose is to correct a few of M&B’s assumptions and suggest a framework for future investigation.

For those who cannot read this entire post, my “syllabus” is as follows:

The correct question is not (as M&B contend) whether legislative power can be delegated but the extent to which delegated legislative power can be further delegated. M&B cite delegations by the British Parliament and by the states to the Confederation Congress, but neither is directly relevant, at least as M&B employ them. Moreover, the Constitution does not delegate to Congress a single “legislative power,” but discrete, enumerated legislative powers. The scope of each, including the extent of delegation approved by the people who conveyed it, must be sought in its particular wording and background.  It is also fruitless to try to identify in the Constitution a non-delegation principle for congressional activities in areas outside the enumerated powers. Originalists should concentrate on defining the proper scope of enumerated powers, the amount of delegation permissible for each, and why some congressional activity—with or without delegation to the executive—is outside the scope of the Constitution’s enumeration.

Because most jurists and law professors, including originalists, have no historical training, we often commit methodological errors when attempting to recreate the original meaning of a document as old as the Constitution. One  common failing is that of anachronism. Anachronism manifests itself in constitutional writings in several different ways. Perhaps the most common is the attempt to reconstruct the ratification-era understanding or meaning from materials generated years, even decades, later. While most of M&B’s historical illustrations are reasonably timely, some are anachronistic: For example, events in 1798 or 1802—whether cited by proponents or opponents of the non-delegation doctrine—tell us little about how other people, subject to different incentives, understood constitutional language a decade or more earlier.

Another kind of anachronism is to retroject, consciously or unconsciously, later attitudes into the founding generation. For example, in my experience modern legal academics do not think much of the traditional canons of construction because, they claim, those canons are tautological, conclusory, contradictory, or meaningless. But whatever we think of them, the fact is that the generation that wrote and adopted the Constitution respected them a great deal. Some legal sources characterized them as equal in authority to acts of Parliament, and ratification-era writers referenced them repeatedly, both expressly and by implication. The framers wrote, and the ratifiers adopted, the Constitution within a culture of that respect.

But Mortenson and Bagley unconsciously reflect modern attitudes when, as Will Foster observes, they fail to even mention an important founding era canon directly on point: Delegata potestas non potest delegari (delegated power cannot be delegated)—sometimes expressed as Delegatus non potest delegare (A delegate may not delegate). This maxim is crucial to any discussion of congressional delegation because, of course, under the Constitution Congress exercises only delegated powers. As I think Gary Lawson has pointed out, the relevant issue is not whether or to what extent legislative power in general can be delegated. The relevant issue is whether or to what extent delegated legislative power can be delegated further. Put another way, to what extent may our congressional agents employ executive branch subagents to carry out legislative responsibilities?

It follows that M&B’s precedents deriving from the British Parliament are not on point, at least not in the way they use them. Although 18th century radical Whigs contended that public officials were agents or trustees for the people and derived all authority from them, that was not the reigning British political theory. The reigning political theory was that Parliament—or more precisely the King-in-Parliament—was itself sovereign and the source of all political power.

It is true that one branch of Parliament, the House of Commons, consisted of members nominally representing the people. This was precisely why they were not permitted to vote by proxy—because delegating their voting authority to others would violate the Delegata potestas maxim. By contrast, peers represented only themselves, so members of the House of Lords were permitted to vote by proxy. The third branch of Parliament was the Crown, and like the Lords, represented itself only. And when the three branches voted together, they (and not the people) were sovereign. At least at this point, therefore, parliamentary delegations are not germane to the authority of a delegated-power legislature to delegate.

M&B also note that the states delegated authority to the Confederation Congress. This case is superficially appealing because prevailing American political theory did rest on radical Whig public trust/agency premises. Indeed, several state constitutions recited so.

But the Confederation example derives, I think, from another anachronism: the common habit of thinking about the Articles of Confederation as “our first Constitution.” Actually, although the Articles comprised part of the “foederal constitution” in the broader, 18th century sense (i.e., the political system), the Articles themselves were not a constitution in our sense of the word.

As contemporaneous dictionaries define the term, a “confederation” was a treaty, alliance, or league among sovereigns. The Articles confirm this characterization: “a firm league of friendship.” In other words, the Confederation Congress was a treaty organization—akin to NATO’s North Atlantic Council. It was not a true government. This was a reason the 1783 Peace of Paris did not refer to our country as solely “the United States.” Rather, the treaty provided that “His Britannic Majesty acknowledges the said United States, viz., New Hampshire, Massachusetts Bay, Rhode Island and Providence Plantations, Connecticut, New York, New Jersey, Pennsylvania, Maryland, Virginia, North Carolina, South Carolina and Georgia, to be free sovereign and independent states, that he treats with them as such . . . . ”

The states’ decisions to confederate were consistent with the announcement in the Declaration of Independence that the former colonies were now “free and independent states” with “full power to . . . contract alliances . . . and do all other acts and things which independent states may of right do.” There was nothing new in this: Since the days of ancient Greece, sovereign states were understood to have power to make treaties, and the power to make treaties was understood to include authority to enter into confederations and leagues—a matter discussed extensively during the 1787-90 ratification debates.

Thus, the states’ delegation of authority to the Continental and Confederation Congresses—including what M&B would refer to as “alienation” of all authority to make treaties—was not a delegation of the legislative power. It was an exercise of that treaty power, customarily thought of as executive rather than legislative in nature. The states’ pre-constitutional delegation of authority to Congress is therefore no more relevant to post-constitutional congressional delegation than acts of the British Parliament.

Of course, one might point out that the Articles still were a delegation of delegated power, in seeming violation of Delegata potestas. But there was no violation. The Delegata potestas canon, like other canons of construction, applied in absence of evidence of the parties’ agreement to the contrary. If a principal consented, either expressly or by implication, to the agent’s employment of subagents, then the maxim was inapplicable.

Moreover, consent could be inferred from the nature of the commission. As noted above, the treaty power always had included authority to confederate.

Incidentally, the Articles were not the only document though which states, after Independence and before the Constitution, employed their “foreign affairs” prerogatives to delegate authority to joint assemblies. Between 1776 and 1787, the states continued colonial practice by holding eleven formal conventions among themselves. (This is one reason I am amused by the frequent claim that the Constitutional Convention is the “only precedent” for a general amendments convention under Article V.) Most of these conclaves, including the Constitutional Convention, were empowered only to make proposals. But in at least one instance, the 1780 Philadelphia Price Convention, the states pledged to be bound by any agreement the convention might reach.

Might it also be said that in granting Congress legislative power the people implicitly authorized delegation to executive agencies? Could M&B’s British and Confederation examples be used to show such implicit consent? Not in the way they use those examples.

It is fundamental that the Constitution does not delegate to Congress “the legislative power.” Rather, it delegates about thirty discrete legislative powers—seventeen (clarified by the Necessary and Proper Clause) in Article I, Section 8, and others scattered throughout the document. Specific words and phrases define the scope of each of these powers. The framers were careful drafters, and most of these words and phrases were borrowed from contemporaneous legal or governmental practice. For example, to “regulate Commerce” was an Anglo-American legal phrase embracing the law merchant, protective tariffs, governance and facilitation of navigation, and certain related subjects. Similarly, the framers borrowed the phrase “establish Post Offices and post Roads” from British postal statutes. It encompassed building post offices and intercity roads, designating postal routes and tolls, hiring employees, enacting criminal penalties for misuse of the postal service—everything necessary and customary to develop a working postal system and an intercity highway system.

The extent to which each enumerated power authorized Congress to delegate to executive sub-agents depended on the constitutional words describing the power. The relevant history shows that the phrase “establish Post Offices and post Roads,” probably did not allow delegation to the executive to fix routes—unlike in Britain, where Parliament used exactly the same language to describe its delegation of postal functions to the executive. On the other hand, the enumerated power to “provide and maintain a Navy” contemplates Congress as playing principally a funding role, delegating extensively to executive-branch officers.

It follows that the search for a single “non-delegation” principle applicable to all congressional powers is a futile one. Instead, the scope of permissible delegation of any particular congressional power must be sought in the meaning of the words describing that power.

A remaining difficulty is that today many of Congress’s statutory programs are not within the scope of the Constitution’s enumeration. Of course, we are all familiar with professorial and juristic attempts to square those programs with the Constitution’s text—specimens of the sophistry and pretextual arguments rejected by 18th century jurisprudence and widely mocked during the ratification debates. One sign of the weakness of such arguments is that they offer no way for identifying principled originalist delegation standards for the laws they seek to justify.

Consider, example, the law upheld in Hodel v. Indiana regulating surface mining so as to preserve farm land. The court held (in the teeth of extensive founding-era evidence to the contrary) that the law was a proper exercise of the Commerce Power. Given that holding, how does one determine the extent to which Congress may delegate surface mining rules to an executive agency? Nothing in the pre-constitutional history of “regulating Commerce” informs us of this, because nothing in the pre-constitutional history of “regulating Commerce” pertains to regulating land use.

In such cases, identifying a non-delegation principle rooted in the Constitution is, in my view, a hopeless task. Originalists would better spend their time defining the scope of the enumerated powers, the amount of permissible delegation within each, and why certain federal activities are outside their scope.