In 1825, Chief Justice John Marshall articulated what might be called the civics-book model of American government: “the legislature makes, the executive executes, and the judiciary construes the law.”Footnote 1 Marshall’s simple formula seems to emerge straightforwardly from the structure of the United States Constitution, which begins each of its first three articles with a “vesting clause” granting a particular kind of governmental power to a specific and distinctively constituted governmental institution: “All legislative Powers herein granted shall be vested in a Congress of the United States,”Footnote 2 “[t]he executive Power shall be vested in a President of the United States of America,”Footnote 3 and “[t]he judicial Power of the United States, shall be vested in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish.”Footnote 4 What could be more “legislative” than the making of laws, what could be more “executive” than their execution or implementation, and what could be more “judicial” than their construction or interpretation? Wasn’t Marshall just stating the obvious?
There are two problems with Marshall’s account of the civics-book model of American government. The first is that it does not remotely describe the actual American government. Put aside the extent to which courts make law rather than interpret itFootnote 5 and the myriad ways in which Congress tries to control the implementation of law. The most notorious departure from the civics-book model is the emergence of what for lack of a better term is often called “the administrative state.”Footnote 6 By any plausible metric, far more of the legal norms that govern people’s conduct come from executive action than from legislative action.Footnote 7 As a matter of sheer numbers, as of 2012 the Code of Federal Regulations, which collects the binding rules promulgated by federal administrative agencies, had four times as many pages as the United States Code, which codifies the statutes enacted by Congress.Footnote 8 That does not even count the binding norms that result from administrative agency adjudication, which account for the vast bulk of, for example, federal labor law. And in terms of importance, Congress has turned over to administrative agencies, with only minimal statutory direction, primary responsibility over everything from environmental protection to financial stability.Footnote 9 To a large extent, in the modern American government the executive rather than the legislature makes the law.
The second problem with the civics-book model of American government is theoretical rather than descriptive: The model depends on a distinction among legislative, executive, and judicial powers, but it is not at all clear what kinds of activities actually count as legislating, executing, and judging. Chief Justice Marshall was fully aware of this problem. In the same passage in which he announced the civics-book model, Chief Justice Marshall went on to say: “but the maker of the law may commit something to the discretion of the other departments, and the precise boundary of this power is a subject of delicate and difficult inquiry.”Footnote 10 Four decades earlier, James Madison had remarked:
Experience has instructed us that no skill in the science of government has yet been able to discriminate and define, with sufficient certainty, its three great provinces—the legislative, executive, and judiciary …. Questions daily occur in the course of practice, which prove the obscurity which reigns in these subjects, and which puzzle the greatest adepts in political science.Footnote 11
So the Constitution asks us to draw non-obvious distinctions among the legislative, executive, and judicial powers, but it provides no definitions for those terms. It assumes that if we think about them hard enough, we will figure it out, at least in a wide range of cases. And thus was framed the problem of congressional subdelegation of legislative authority.
For more than two centuries, courts, Members of Congress, and scholars have wrestled with the problem of ascertaining how much and what kind of authority Congress can vest in executive (and judicial) actors without crossing a constitutional line.Footnote 12 To be sure, no congressional statute has been found by the Supreme Court to be unconstitutional on subdelegation grounds since 1935, but any thoughts (or wishes) that the issue had been laid to rest by the New Deal are hard to maintain in the face of five Justices in 2019 expressing interest in invigorating some kind of subdelegation doctrine.Footnote 13
A description or an analysis of the subdelegation doctrine is the work of a lengthy article (or two, or three, or four),Footnote 14 so for present purposes I have to take much for granted. In particular, I take for granted that the question of the constitutionality—as opposed to the policy wisdom, or political legitimacy, or aesthetics—of subdelegation must be resolved by reference to the Constitution’s original meaning. A description of “original meaning” would also be the work of several articlesFootnote 15; suffice it for now to say that the original meaning of a conceptFootnote 16 is the set of criteria for ascertaining referents of that conceptFootnote 17 that were intended by the concept’s author,Footnote 18 which, in the case of the United States Constitution (which has a hypothetical legal author and multiple actual authorsFootnote 19), operationalizes to the meaning that would be intended by a hypothetical reasonable observer at the time of the concept’s authorship.Footnote 20 The particular thoughts in the heads of concrete historical actors can be evidence of original meaning, but those thoughts do not constitute that meaning. Concepts can include referents—past, present, and future—not within the specific contemplation of the concept’s author but which factually satisfy the criteria for inclusion in the concept. In other words, the ultimate question is: What would a reasonable, informed, eighteenth-century observer (who may or may not correspond to any actual, concrete eighteenth-century observer) think about the criteria for including or excluding within the concepts employed in the Constitution various things and relations in the world?Footnote 21 Of course, not everyone is intellectually interested in ascertaining the original meaning of the Constitution—to which one can only invoke the great philosopher Ricky Nelson.Footnote 22
My goal in this article is not to explore, in the abstract, how the original meaning of the Constitution handles the problem of subdelegation of legislative power. I have done that elsewhere, at much too great a length to summarize here. Rather, my focus is on one particular rationale for an actor to confer discretion on another actor: the supposed expertise of the delegee. My aim here is to discuss the extent to which the original meaning of the Constitution contemplates reliance by Congress on the expertise of others and the permissible boundaries of that reliance. In order to get at that question, I will need to say something about the Constitution’s broader approach to subdelegation, but that discussion will necessarily be a bit sketchy.
Exploration of expertise and subdelegation begins with something that seems, at first glance, to have little to do with the Constitution, expertise, administrative agencies, or modern government: the eighteenth-century private law of fiduciary duty. I beg indulgence; the relevance of the following tale will become clear in due course.
I. How to Tame Your Agent
Suppose that in 1788 you are an American with important business contacts overseas. You market your services to exporting firms as a business agent, or “factor,”Footnote 23 to represent American businesses seeking to sell in foreign markets. One of those businesses hires you as an agent to sell its agricultural products in Europe. What are your legal responsibilities as a factor?
Those responsibilities will be set forth most prominently in the instrument of agency by which you were hired. That instrument will be a contract of sorts, and you and your principal can work out the details of your arrangements in the contract, covering everything from your compensation to the scope of your authority (How big a deal are you allowed to arrange? Could you sell the whole company if you got a good enough offer?) to your liability if something goes wrong with a business deal that you have set up (Are you liable for negligence? Strictly liable, acting as an insurer of all transactions that you arrange? Not liable even for negligence?). The contract can be as detailed as you and your principal—and your lawyers—want to make it.
And there’s the rub. The more detailed the contract, the more expensive it will likely be to negotiate and then subsequently enforce. The range of matters that could potentially arise in a contract between a principal and a factor is staggering. Even if you and your principal think of most of those potential issues, figuring out how to frame contractual provisions that resolve them satisfactorily is no mean feat. The lawyers will be happy to try to do it—at their usual hourly fees, of course—but you and the principal may have some qualms about paying for the privilege.
The law of agency might help you out a bit. Principals and agents have been making contracts like this for a very long time (even in 1788), addressing precisely the kinds of issues that you currently face. Some resolutions of those issues may have become so standardized that they are part of the background set of legal rules of agency against which any particular contract is drafted. There might be conventions about the scope of a factor’s authority in specific circumstances that the law will deem to be part of the contract unless those conventions are specifically disavowed by the parties. For example, one recurring question is whether an overseas factor can arrange for a sale of goods on credit rather than for cash. Since there is a very good chance that you, as an agent, will be sent to Europe accompanied by a boatload of goods to sell (the principal is probably not going to send you overseas on a months-long journey, wait for your return, then send goods out on another months-long journey afterward), fixing your authority to set the terms of sale is likely to be a very important part of the agency arrangement. There might also be important recurring questions about your authority to make on-the-spot decisions based on changing circumstances. If you were sent to sell goods in France, for example, but Congress then imposes an embargo on United States trade with France, do you as the agent have authority to divert to Portugal or Spain to try to sell your goods there? Can you take less than the specified minimum price if the alternative is for the goods to rot on board the ship? The contract could try to specify outcomes for all of these (and other) contingencies, but trying to anticipate each and every specific contingency is impossible, and trying to catch most of them is probably prohibitively expensive. The law can economize on transaction costs by adopting background presumptions about these matters that reflect the most likely resolutions that most parties would reach. Any parties that do not like the background rules can contract around them in a specific agency instrument, while parties that are fine with the background rules can simply say nothing and let the law do the work for them.
Those background rules can change over time. In 1700, for example, the normal assumption would be that factors could only agree to sales for cash, unless their principals had specifically given them power to arrange for sales on credit or there was a contrary custom or usage in a particular business.Footnote 24 By 1800, so many principals had granted their agents such a power that the normal background assumption had shifted, and an 1800 factor would be presumed to be able to make sales on credit unless the agency instrument specifically negated that authority.Footnote 25 Thus, identically worded contracts from 1700 and 1800 would have materially different meanings with respect to the authority of factors, because the background assumptions that form part of the meaning of every instrument—and of every act of communication—would be different.
One crucial recurring issue in agency relationships is the agent’s power to delegate authority. Presumably, the principal selected you rather than someone else as the factor because of some characteristic that the principal found appealing. Perhaps you have (or pretend that you have) contacts abroad with potential buyers that other factors might not have. Perhaps you have a reputation for being especially good at sensing changing market conditions and negotiating the best prices for principals. Whatever the reasons, the principal chose you. But suppose that after securing the contract, you decide that you want a vacation in Bermuda—or, perhaps more responsibly, you decide (maybe even correctly) that one of your associates, or competitors, will actually do a better job than you will be likely to do. Can you send someone else in your place? What if the person that you send in your place is objectively as qualified as you, by the criteria that actually induced the principal to select you? Can you delegate your agency authority to another? And even if you carry out the task yourself, you surely are going to need some assistance. If you are not a lawyer, you might need to engage the help of a lawyer to draft contracts and arrange for title documents. You might need an interpreter. If weather conditions are going to be a major consideration in the success of your venture, you might want to consult a meteorologist (or whoever in the eighteenth century might know something about weather that you would not). And the chances are pretty high that you are not personally going to pilot the ship on a cross-Atlantic journey but will want to hire someone to do that for you (and the principal). Can you do that? Or do you personally have to perform each and every task involved in carrying out your assignment? You will surely say that there are people who have expertise in matters such as piloting and weather forecasting that would help you fulfill your primary duty.
As with everything else, the parties can specify whatever they want in the contract. If the principal does not care who actually performs the tasks so long as they get done well, he or she might authorize you to delegate whatever tasks you see fit to whomever you wish, assuming that you retain ultimate responsibility for the success or failure of the enterprise. Or the principal might more specifically authorize you to hire professionals or experts necessary to carry out the task, on the assumption that you will not personally have every skill and credential necessary to effectuate all transactions. Or the contract could calibrate the authority to delegate to any degree desired, with whatever degree of specificity the parties are willing to pay the lawyers to negotiate and draft.
But what about the background rules of agency law? What will the law fill in regarding delegation if your instrument is silent?
The principal and agent in 1788 would find refreshingly clear and direct answers to those questions. Eighteenth-century agency law had well-developed, well known, nuanced, and strong views on the default presumptions regarding delegation of agency authority. The baseline rule (with some important qualifications to be noted shortly) was that power delegated to an agent had to be exercised personally by the agent to whom the power was delegated. That is, the agency instrument itself is a delegation of power from the principal to the agent. Any further transfer of that power by the agent would be a subdelegation, and subdelegation in the eighteenth century was presumptively unlawful.
This principle was about as clear and settled as any eighteenth-century legal maxim could be. One of the most important sources of eighteenth-century legal lore was Matthew Bacon’s A New Abridgement of the Law, first published in 1730. Founding-era and early American use of and reverence for this treatise was enormous.Footnote 26 Bacon wrote: “One who has an Authority to do an Act for another, must execute it himself, and cannot transfer it to another; for this being a Trust and Confidence reposed in the Party, cannot be assigned to a Stranger whose Ability and Integrity were not so well thought of by him for whom the Act was done; therefore an (a) Executory having Authority to sell, cannot sell by Attorney.”Footnote 27 This language persisted, and was elaborated with examples, through seven editions of the treatise, running into the middle of the nineteenth century.Footnote 28
Other private-law treatises were in complete agreement with Bacon’s account of subdelegation by agents. Samuel Livermore wrote in 1818: “An authority given to one person cannot in general be delegated by him to another; for being a personal trust and confidence it is not in its nature transmissible, and if there be such a power to one person, to exercise his judgment and discretion, he cannot say, that the trust and confidence reposed in him shall be exercised at the discretion of another person.”Footnote 29 Joseph Story’s celebrated treatise on agency law similarly explained: “one, who has a bare power or authority from another to do an act, must execute it himself, and cannot delegate his authority to another; for this being a trust or confidence reposed in him personally, it cannot be assigned to a stranger, whose ability and integrity might not be known to the principal or who, if known, might not be selected by him for such a purpose.”Footnote 30 James Kent, one of the most famous early nineteenth-century jurists and legal commentators, confirmed: “An agent, ordinarily, and without express authority, has not power to employ a sub-agent to do the business, without the knowledge or consent of his principal. The maxim is, that delegatus non potest delegare, and the agency is generally a personal trust and confidence which cannot be delegated ….”Footnote 31 I am not aware of any eighteenth- or nineteenth-century source that says otherwise about the agency law of subdelegation. “The founding-era rule against sub-delegation of delegated agency authority is as clearly established as any proposition of law can be established.”Footnote 32
So, would you as a factor in 1788 have to pilot the ship to France (and swab the decks in the bargain)? No, or at least not unless the agency instrument specifically said that you had to perform that task yourself. The hard-and-fast rule against subdelegation was subject to three crucial but limited qualifications.
First, the parties could by contract authorize any degree of subdelegation of authority that suited them. As Livermore put it in 1818, “an authority may be delegated to another, where the attorney has an express authority for that purpose.”Footnote 33 Of course, given the strong background presumption against subdelegation, “when it is intended, that an agent shall have a power to delegate his authority, it should be given to him by express terms of substitution.”Footnote 34 The parties would need to speak clearly to overcome the normal rule against subdelegation.
Second, there might be contexts in which custom or usage permits a measure of subdelegation that would normally be forbidden.Footnote 35 One would have to look to the usage or custom of the particular trade to discover whether such permission to subdelegate is available in any specific agency setting.
Finally, the rule against subdelegation only applied to those aspects of tasks that require the judgment or discretion of the agent. For ministerial or incidental tasks, the natural assumption (subject to alteration by the parties in the agency instrument) is that some kind of subdelegation is permitted.
If someone is given a power of attorney to sell property and prepares an advertisement for a newspaper, surely a subordinate can mail or phone the advertisement into the newspaper, though the attorney would no doubt have to prepare the advertisement and select the appropriate newspapers in which to place it. Whether a task involves discretion or is merely ministerial or mechanical may at times raise difficult legal questions (as anyone familiar with the law of mandamus, which turns on the same distinction, can attest), but over a wide range of cases one can normally glean the extent to which agents must exercise discretion from the terms of the instrument and the nature of the task.Footnote 36
It can sometimes be difficult to ascertain which tasks are the ones which the agent was specifically hired to perform personally and which are incidental to those primary tasks, but the context of the agency arrangement will often provide answers. In other words, “there are many cases wherein from the nature of the duty, or the circumstances under which it is to be performed, the employment of subagents is imperatively necessary, and the principal’s interests will suffer if they are not so employed. In such cases, the power to employ the necessary subagents will be implied.”Footnote 37
Joseph Story aptly summed up the eighteenth- and nineteenth-century law of agency subdelegation:
But there are cases, in which the authority [to sub-delegate] may be implied; as where it is indispensable by the laws, in order to accomplish the end; or it is the ordinary custom of trade; or it is understood by the parties to be the mode, in which the particular business would or might be done …. So, where, by the custom of trade, a ship broker, or other agent, is usually employed to procure a freight or charter party for ships, seeking a freight, the master of such a ship, who is authorized to let the ship on freight, will incidentally have the authority to employ a broker, or agent for the owner, for this purpose. And the same principle will apply to a factor, where he is, by the usage of trade, authorized to delegate to another the authority to substitute another person to dispose of the property. In short, the true doctrine, which is to be deduced from the decisions, is, (and it is entirely coincident with the dictates of natural justice,) that the authority is exclusively personal, unless, from the express language used, or from the fair presumptions, growing out of the particular transaction, or of the usage of trade, a broader power was intended to be conferred on the agent.Footnote 38
This excursion into eighteenth- and nineteenth-century agency law may (or may not) be intellectually interesting, and the potential travails of an eighteenth-century factor may (or may not) make for a good story, but what does any of this have to do with the constitutional authority of twenty-first century American administrative agencies?
Just about everything.
II. A Matter of Trust
The United States Constitution is a fiduciary instrument. Guy Seidman and I—building on the pioneering work of Rob Natelson—have made the book-length argument for this claim elsewhere,Footnote 39 and I will not rehearse the lengthy argument here. Suffice it to say that viewing government in fiduciary terms was natural to the founding generation, with express language describing government officials as fiduciaries finding its way into numerous state constitutions of the era.Footnote 40 Those state constitutional provisions were declaratory rather than substantive; the fiduciary character of government infused all constitutions of the era even when they did not contain express fiduciary clauses. The United States Constitution is entirely of a piece with the fiduciary state constitutions from the prior decade. The Constitution is, in chief measure, an entrustment of authority by the principal, “We the People,”Footnote 41 to various governmental agents to handle some aspects of the principal’s affairs.Footnote 42 That kind of entrusting is exactly what fiduciary instruments do. Whether one thinks that the Constitution is most like a power of attorney,Footnote 43 is more akin to a corporate charter,Footnote 44 strongly resembles a trust,Footnote 45 or is sui generis,Footnote 46 the Constitution belongs to the family of eighteenth-century documents known as fiduciary instruments. Accordingly, if there were background legal principles governing that entire family of instruments, whatever their precise characterization, in effect in the late eighteenth century, there is good reason to think that ascertaining the meaning of the Constitution requires resort to those background interpretative principles.Footnote 47
The eighteenth-century principles regarding subdelegation of agency authority were not confined to factors. They cut across the entire family of fiduciary relationships. Their eighteenth-century application did not depend on the particular characterization of the agency instrument involved, though the qualifications to the general rule against subdelegation are somewhat dependent on context, so that their application might depend on the specific agency instrument at issue. Accordingly, the best understanding of the Constitution is that when it delegates authority to government agents, such as Congress, the president, and the federal courts, it expects those agents to exercise that power personally and not subdelegate it to other actors, unless the subdelegation is either expressly authorized by the instrument or falls within one of the recognized qualifications to the general prohibition.
In the case of Congress, to which We the People delegated authority to manage a wide range of We the People’s affairs, there is no express authority in the document to subdelegate authority. The only conceivable source would be the Necessary and Proper Clause, which gives Congress power to “make all Laws which shall be necessary and proper for carrying into Execution”Footnote 48 all constitutionally vested powers. But that clause is actually an instantiation, and incorporation, of background principles of agency law, including the agency-law principle against subdelegation.Footnote 49 Subdelegations of legislative authority to other actors is simply not a “proper” means for implementing federal powers.
Nor is there any plausible way to construe the Constitution as implicitly authorizing subdelegation of legislative authority. Even a casual look at the Constitution’s overall structure shows that it is hyperfocused on the selection procedures for the actors in whom is vested governmental power. A staggering percentage of the Constitution’s length (including the amendments) is concerned with defining the electorate, the electoral procedures for Congress and the president, and the appointment process for judicial and other executive actors. This is obviously a document that cares very deeply and profoundly about who is exercising governmental power. The idea that a document so focused on selection procedures for agents contains implicit authorization for subdelegation is, frankly, simply absurd. Such a claim does not pass a straight-face test. Nor was there a clearly established eighteenth-century custom or usage that allowed legislators to pass off their responsibilities to others. The starting point for constitutional analysis, assuming that one is interested in what the Constitution actually means,Footnote 50 is that Congress cannot subdelegate its delegated authority to others.
To be sure, one can reach the same conclusion about the constitutional rule against subdelegation without the machinery of fiduciary law. One can derive a strong constitutional rule against subdelegation of legislative authority from straightforward textual and structural analysisFootnote 51 or from any number of plausible normative add-ons to the constitutional text and structure.Footnote 52 “[T]here are few propositions of constitutional meaning as thoroughly overdetermined as the unconstitutionality of subdelegations of legislative authority.”Footnote 53 In the end, however, the most fundamental source of the principle against subdelegation is the fiduciary character of the underlying document; the textual and structural arguments gain much of their currency from the background rules of agency law that inform them. Accordingly, the private-law rules of agency regarding subdelegation describe the constitutional rule for Congress when it tries to vest authority in other actors.
III. You’ve Got to Give It To Me
The fiduciary principles regarding subdelegation go to the very core of modern administration. The administrative state is built on subdelegation. No one would care about the administrative state if it did not control people’s lives, fortunes, and sacred honors; and the executive bureaucracy has that power because Congress has given it over. To be sure, a crucial part of the story is the successful assertion, by all three departments of the national government, of centralized authority that goes far beyond any plausible constitutional account of national legislative power.Footnote 54 The administrative state then vests a large percentage of that (usurped) power in agents exercising subdelegated legislative power. I will say more about the relationship between the scope of federal power and the problem of subdelegation shortly, but for this essay I want to focus primarily on the subdelegation phenomenon at a general level. If the Constitution is a social contract of sorts, what are the terms of that contract? Do fiduciary legislators breach that (fiduciary) social contract when they empower others to act in their stead?
Some of the legislative grants of authority in the modern administrative state seem obviously unconstitutional at first glance. There appears to be no way to justify giving an agency official power to grant broadcast licenses “if public convenience, interest, or necessity will be served thereby;”Footnote 55 to fix prices which “in his judgment will be generally fair and equitable;”Footnote 56 to approve a corporate financial structure if it “does not unduly or unnecessarily complicate the structure, or unfairly or inequitably distribute voting power among security holders;”Footnote 57 or to “purchase … troubled assets from any financial institution, on such terms and conditions as are determined by the Secretary,”Footnote 58 where the term “troubled assets” means “any other financial instrument that the Secretary … determines the purchase of which is necessary to promote financial stability.”Footnote 59 And what about a criminal statute that addresses the question of retroactive application by proclaiming: “The Attorney General shall have the authority to specify the applicability of the requirements of this subchapter to sex offenders convicted before the enactment of this chapter ….”Footnote 60 If these statutes are not unconstitutional subdelegations of legislative power, one might well ask, then what grant of authority to executive (or judicial) actors could possibly be unconstitutional?
As it happens, all of these congressional grants of authority to executive agents have either been specifically upheld by the Supreme CourtFootnote 61 or (in the case of the Troubled Assets Relief Program) would so obviously be upheld by the Court that there was no point in forcing a challenge. These statutes are treated by the case law as a fortiori approval of any and every grant of authority from Congress to executive (or judicial) agents. As Justice Scalia aptly put it, “What legislated standard, one must wonder, can possibly be too vague to survive judicial scrutiny, when we have repeatedly upheld, in various contexts, a ‘public interest’ standard?”Footnote 62
The magic word that has long sought to stop deeper inquiry into subdelegation is “expertise.” “C’mon, man: Do you really want (or expect) Congress to understand the chemical properties of particulate matter, to figure out the hydrodynamics of groundwater flow, to design appropriate criteria for drug trials, and so on? These grandstanding hacks can’t even pass a budget. You need people who know what they are doing—the people who have graduate degrees from Ivy League schools—to make the rules for a complex society. You need experts.”
Claims of “expertise” have been the lifeblood of the administrative state for more than a century. As Reuel Schiller put it in an elegant and readable study of the role of expertise in the development of the New Deal:
The notion that expertise should be used to formulate public policy was hardly an invention of the New Deal. Indeed, Progressive reformers of the 1910s and 1920s proffered expertise as the solution to a host of problems disturbing the social order at the beginning of the twentieth century. They believed that the scientific method could be applied to these problems. Technocratic experts, having arrived at a solution to a given problem, should be allowed to implement it.Footnote 63
These ideas bled naturally into the 1930s, when much of the modern administrative state was constructed. Felix Frankfurter expressed the Progressive vision well when he wrote:
[I]n the modern world the simple virtues of honesty and public devotion are not enough. Alone they will not unravel the tangled skein of social-economic complexities. They cannot even analyze the issues to which answers must be found …. Compelled to grapple with a world more and more dominated by technological forces, government must have at its disposal the resources of training and capacity equipped to understand and deal with the complicated issues to which these technological forces give rise.Footnote 64
James Landis, perhaps the most important figure in the design of the New Deal, opined that public policy “could most adequately be developed by men bred to the facts,”Footnote 65 because “[w]ith the rise of regulation, the need for expertness becomes dominant.”Footnote 66 In New Deal thinking, “expert” and “government” go together like “Brady” and “Belichick.”Footnote 67
“Hardly a relic of the New Deal-era, this [reverence for expertise] remains the dominant view among scholars today … [and] so deeply pervades the scholarly literature that it is difficult to isolate highly relevant pieces.”Footnote 68 Indeed, a simple WESTLAW search in the law review database for “expert!/s admin!” yields the dreaded “10,000” response. Invocation of expertise is more than just “a part of the narrative explaining legislative delegations to administrative agencies.”Footnote 69 It is absolutely central to the modern justification for the whole apparatus of the administrative state. The Supreme Court declared as much, with admirable clarity, in 1989 when it upheld an open-ended grant to the United States Sentencing Commission to establish sentencing ranges for federal crimes with the pithy observation: “in our increasingly complex society, replete with ever changing and more technical problems, Congress simply cannot do its job absent an ability to delegate power under broad general directives.”Footnote 70
Here is the question of the day: Could this actually be right as a matter of original meaning? After all, one of the accepted eighteenth-century grounds for subdelegation of agency authority was, in the words of Justice Story, “where it is indispensable … in order to accomplish the end.” If making use of expertise is truly “indispensable” in the modern world, does the Constitution perhaps accommodate that usage? Could all of those New Deal cases upholding grants of authority to agencies to pursue the public interest actually be rightly decided?
The argument becomes even stronger if one thinks (1) that the Constitution most resembles a trust instrument from within the family of fiduciary instruments and (2) the Constitution’s norms for subdelegation are the norms for trust instruments in effect at the present moment rather than the specific norms that were in place in 1788. In other words, one can be an “originalist” with respect to the notion that the Constitution is a fiduciary trust instrument but think that the specific content of trust law is meant to evolve over time.Footnote 71 Modern trust law, especially over the past quarter century, has expanded the range of permissible subdelegations of authority by trust managers, in recognition of the vast expertise in investment strategy that is available to those managers.Footnote 72 While earlier versions of the Restatement of Trusts recited the law’s longstanding skepticism towards subdelegation,Footnote 73 the Third Restatement of Trusts, as currently modified, provides:
(1) A trustee has a duty to perform the responsibilities of the trusteeship personally, except as a prudent person of comparable skill might delegate those responsibilities to others.
(2) In deciding whether, to whom, and in what manner to delegate fiduciary authority in the administration of a trust, and thereafter in supervising or monitoring agents, the trustee has a duty to exercise fiduciary discretion and to act as a prudent person of comparable skill would act in similar circumstances.Footnote 74
As investment practices and finance theory get more complex, it becomes more plausible to think that even a skilled trustee might reasonably determine that there is someone else better situated to manage a particular portfolio (or portion thereof).
When ancestral land was the characteristic trust asset, trust administration required of the trustee relatively little expertise or authority. Trustees were mostly stakeholders, and the family lived on the estate and managed its affairs. Today, by contrast, financial instruments have become the typical assets of the trust, and these assets require active fiduciary administration. Managing a portfolio of marketable securities is as demanding a specialty as stomach surgery or nuclear engineering ….Footnote 75
The modern Restatement accommodates this development, and it even recognizes that it might be obligatory for a faithful agent to subdelegate authority in certain circumstances. A comment to the current section on delegation notes: “A trustee's discretionary authority in matters of delegation may be abused by imprudent failure to delegate as well as by making an imprudent decision to delegate.”Footnote 76
If that is an accurate statement of the modern law of trusts, and if Congress can be analogized to a trustee, and if the content of the Constitution’s rules on subdelegation change with times and circumstances, then there might be a plausible case that Congress not only may subdelegate authority but must do so. If all that holds true, then those scholars who argue that the modern administrative state is constitutionally required Footnote 77 might have a stronger point than some of us have credited.Footnote 78
As it happens, two of the premises in the above argument are not correct. First, the Constitution is not best analogized to a trust. A trust is a particular kind of arrangement involving management of a body of assets. In the founding era, a trust might even have had a narrower meaning as referring principally to arrangements for the management of land.Footnote 79 Government involves much more than asset management. It involves regulatory matters over many aspects of the principal’s life, not just devising wise investment strategies for maintenance of capital or the payment of debts. A power of attorney, which is a more flexible and diverse kind of instrument, far better describes the operations of government than does a trust.Footnote 80 It is true that the eighteenth century (and earlier times) were full of expressions describing government officials as holding their offices “in trust,” while almost nobody spoke of holding government office under a power of attorney.Footnote 81 The Constitution even speaks expressly of an “Office of … Trust.”Footnote 82 But references to “trust” as a concept are not the same things as references to a “trust” as a particular kind of legal arrangement. The holder of a power of attorney holds their power as a matter of trust, but that does not abolish the legal distinction between a power of attorney and a trust. To say that government officials hold their offices “in trust” says that they are fiduciaries and that the offices (unlike offices of profit) are not for sale, but it does not say what form that fiduciary obligation takes.
Second, the Constitution’s norms of fiduciary conduct do not change over time. The things and relations in the world to which the Constitution’s terms and underlying concepts apply change over time, as facts and circumstances change, but the criteria for identifying something as within those terms do not change. This all follows from the nature of concepts and the nature of meaning, though to explicate the point further would require a full description of a theory of meaning, which is a project for another day.Footnote 83
But even so, the question remains whether sound norms of fiduciary conduct permit, or even require, Congress to subdelegate certain tasks to experts. That is the question to which all of the machinery in this essay thus far has been leading.
No one doubts for a moment that Congress is able, and perhaps even required in the responsible exercise of its authority, to draw upon advice and knowledge from others. If administrative agencies simply made proposals or suggestions to Congress, no one would have any constitutional objection. Surely it is “necessary and proper for carrying into Execution” governmental powers for Congress to create mechanisms to generate information and make recommendations (though, as I will explain shortly, it matters very much how those mechanisms are constructed and employed). The real issue concerns actual transfer of decisional authority to agents to make law through regulations or binding adjudications.Footnote 84 Does that kind of formal transfer of decisional authority ever fall within the bounds of permissible subdelegation under eighteenth-century fiduciary norms?
The (too) easy answer is “no.” This is not a case in which, for example, a lay fiduciary agent needs to have documents properly filed and so engages a lawyer to navigate the local recording laws. That would clearly be an instance in which some measure of subdelegation “is indispensable … in order to accomplish the end.” Congress is fully capable of passing laws; it does not need to subdelegate that authority to others. If Congress wants help drafting those laws, it can obtain such help, but it does not need to have some technical expert enact them.
The argument from necessity really involves the claim that Congress cannot pass enough laws without subdelegating an army of “junior varsity Congress[es]”Footnote 85 to pick up the slack. When the Supreme Court said that “Congress simply cannot do its job absent an ability to delegate power under broad general directives,” the “job” that it had in mind was not passing laws but passing enough laws to satisfy the public’s (or the Court’s?) demand for them. Could Congress micromanage the entire country without subdelegation? Surely not. If Congress’s “job” is to micromanage the entire country, then of course it is “indispensable” to fulfillment of that job to subdelegate a good portion of the task. On the other hand, if Congress’s “job” is to legislate within the confines of its granted authority as laid out in the Constitution, the need for subdelegation dissipates.
Accordingly, one very important conclusion, which I have not adequately emphasized in prior work, is that one cannot really discuss the problem of subdelegation without also discussing the principle of enumerated federal powers. If Congress really does have a general legislative authority to promote the “general welfare,” as any number of modern scholars argue,Footnote 86 there is a quite plausible case to be made that a power to subdelegate follows naturally from that authority, though there would be empirical questions whether (1) promoting the general welfare calls for extensive rather than minimalist federal interventionFootnote 87 and (2) Congress could carry out that general-welfare-promoting function without formally transferring lawmaking authority to others and simply using procedural mechanisms to streamline its own lawmaking ability and enhance its output.Footnote 88 But those would be the proper terms of debate: What, exactly, is Congress’s constitutional “job,” and can it perform that “job” without subdelegation? Thus, Justice Gorsuch may have been too hasty to assume that “[t]he separation of powers does not prohibit any particular policy outcome, let alone dictate any conclusion about the proper size and scope of government.”Footnote 89 It may not be possible to analyze the problem of subdelegation without also considering “the proper size and scope of government.”
Rather than address that question here, however, I want to focus on a different aspect of Congress’s fiduciary responsibility that has application regardless of the correct outcome of the constitutional subdelegation debate: Congress’s obligation to determine that it is truly, in fact, relying on expertise when it defers to others. This is obviously relevant if Congress is permitted/obliged to subdelegate some measure of authority; a responsible fiduciary will only subdelegate to persons who it has good reason to believe can and will genuinely bring expertise to bear on a problem for which expertise is needed. Authorization to subdelegate is not authorization to shirk. It requires careful attention on the part of the agent to the task of subdelegation. The agent must determine what does and does not truly need to be subdelegated, and the agent must subdelegate only to those who there is good reason to think will do a good job. The responsibility to think carefully about the expertise of those on whom an agent relies is also relevant even under a strict regime of non-subdelegation. Congress is permitted to call upon others for advice even when it cannot give those others formal decision-making authority, but it must have good reasons for thinking that the advice is worth taking before taking it. Whether Congress is relying on agencies for advice or for lawmaking, that kind of reliance is only responsible—meaning consistent with the fiduciary social contract under which Congress acquires its power—if it represents a genuine reliance on expertise in circumstances where reliance on that kind of expertise is appropriate.
What would a responsible fiduciary approach to expertise—whether for purposes of advice or subdelegation—look like in the modern administrative state? The answer requires a careful examination of the idea of expertise and how it can be applied, and misapplied, in modern governance.
“Expertise” as a concept is the subject of an entire sub-specialty, and the tools of psychology, sociology, economics, technology, and philosophy necessary to grasp that sub-specialty are far beyond my pay grade. There is an expansive literature on expertise, which I have neither the time nor (for lack of a better term) expertise to engage with here.Footnote 90 The general problem posed and faced by this literature is: How do you know which experts to trust and when to trust them without yourself being an expert in the relevant field?Footnote 91 The questions that must be asked of Congress in the context of the Constitution are not different in kind from the sorts of questions that arise in everyday life when people must decide whether to trust self-described experts, so a firm grasp on this literature is probably a precondition to serious exploration of the underlying constitutional problem. Accordingly, my goal here is merely to outline the kinds of questions that need to be addressed by a fiduciarily responsible theory of subdelegation in the context of the administrative state rather than to address these questions in a comprehensive form.
It is important to note beforehand, however, that the Founders were keenly aware of these problems even if they did not have the vocabulary of the modern social sciences in which to frame them. The members of the House of Representatives serve two-year terms,Footnote 92 while Senators serve six-year terms.Footnote 93 By founding-era standards, these were very long terms for legislators. State legislatures pre-1788 typically had one-year terms for members of the lower chamber, and some states had legislative terms as short as six months.Footnote 94 In The Federalist, Madison defended the House’s two-year term at least partly on cognitive grounds:
No man can be a competent legislator who does not add to an upright intention and a sound judgment a certain degree of knowledge of the subjects on which he is to legislate. A part of this knowledge may be acquired by means of information which lie within the compass of men in private as well as public stations. Another part can only be attained, or at least thoroughly attained, by actual experience in the station which requires the use of it. The period of service, ought, therefore, in all such cases, to bear some proportion to the extent of practical knowledge requisite to the due performance of the service. The period of legislative service established in most of the States for the more numerous branch is, as we have seen, one year. The question then may be put into this simple form: does the period of two years bear no greater proportion to the knowledge requisite for federal legislation than one year does to the knowledge requisite for State legislation? The very statement of the question, in this form, suggests the answer that ought to be given to it.Footnote 95
The ensuing discussion of why and how federal legislatures need more opportunities to acquire knowledge than do their state counterparts shows that Madison expected a lot from federal representatives. The same applies to the Senate, whose six-year term will supposedly provide even more opportunities for acquisition of knowledge.Footnote 96 The Constitution’s terms of office for legislators thus suggest (and I deliberately use no term stronger than “suggest”) that the constitutional standards for fiduciary subdelegationFootnote 97 are quite high. The Constitution gives representatives a moderately lengthy term of office. In return, they need to gain and apply knowledge to the tasks that they take on.
We can now sketch what fiduciary subdelegation must involve. First, and most obviously, the agent must ascertain that the task is one for which the agent’s own knowledge (and capacity to learn) is inadequate. This is not as simple as it sounds. How do you know that a task exceeds your cognitive abilities if you do not have, or possibly even do not understand, the kind of information and skills needed to perform the task? Notwithstanding the very real philosophical and cognitive problems raised by this question, ordinary experience suggests that people can recognize instances where their own knowledge comes to an end (as I have in this essay with respect to the concept of expertise). It seems pretty clear that members of Congress can figure out that they do not know enough about, for example, the dispersion patterns of certain kinds of airborne particles to be able to legislate intelligently using only their own knowledge. A thoughtful exercise of subdelegation thus requires serious self-reflection, coupled with an underlying normative baseline of knowledge that every member of Congress can be expected to possess or acquire.
Second, one must determine that the gap in one’s knowledge can be filled by experts. That is not always going to be so. The fact that one cannot, on one’s own, know what needs to be known does not mean that other people with different training or experiences will know it either. A gap in knowledge, without more, does not lead to subdelegation or deference to experts. The agent must have good reason to think that expertise is, even in principle, relevant to the task. If the task is to figure out the dispersion properties of certain airborne particles, perhaps that is something to which expertise can contribute. If the task is to figure out whether to impose economic costs on some people in order to confer some measure of health benefits on others, it is hardly obvious that expertise of any kind is going to be relevant to that ultimate task.Footnote 98 Again, this poses the philosophical/cognitive problem of how one can know whether expertise will be helpful without having or understanding the expertise in question. I know of no good answer to this conundrum. Probably all that one can ask of an agent is to have considered the problem and to do their best with what they have.
Third, even if expertise is potentially helpful in principle, one must have reason to think that it will, in the particular case at hand, prove helpful in practice. This is a much more complex problem than it might seem at first glance. Even apart from any concerns about the character or quality of self-described experts, which I will discuss shortly, the ubiquitous but oft-ignored problem of second-best hangs over this process.
Assume that the agent identifies a gap in knowledge relevant to a particular task of the agent. Assume further that the agent reasonably concludes both that expertise can potentially bring additional knowledge to bear on the problem and that the agent has identified good candidates for those who might possess that knowledge.Footnote 99 Isn’t it obvious that people who know more will make better decisions—with “better” here meaning nothing more (or less) than “more in accordance with facts of reality” —than people who know less? Unfortunately, it is not at all obvious, and in many cases the opposite will be so. If the additional knowledge brought to bear by the expert is the total sum of knowledge in the universe relevant to the problem at hand¸ then of course the expert is capable of generating a better decision. (Whether the expert actually generates a better decision on these assumptions depends upon how the expert chooses to employ, or not employ, the relevant knowledge.) But how often will an exercise of supposed expertise be based on the theoretically optimal evidence set for a particular problem? Isn’t it more likely (though how would anyone know this?) that expertise will often generate a marginal increment of the relevant knowledge, yielding a larger and more robust but still incomplete evidence set? If that is so, then the question is whether the larger evidence set represented by expertise is a better basis for fact-based decision-making than the smaller evidence set represented by the lay decision-maker. The answer to that question depends on something very hard to know, and perhaps even unknowable in principle: What is the shape of the path from human knowledge and experience to the metaphysically correct state of affairs? If that path is constantly upward sloping, then perhaps one could reason that expert-based decisions will consistently be better than lay decisions. There may be many fields and contexts in which something like a constantly upward sloping assumption makes sense, though I suspect that such contexts are likely to be identified post-hoc by success rates rather than in advance by theorizing. In any event, if there are contexts where the assumption does not hold, it is quite possible that decisions based on smaller evidence sets will be better than decisions based on larger ones.Footnote 100 That will be true whenever the larger evidence set represented by expertise is on a dip in the curve relative to the evidence set represented by the lay person. More knowledge, in other words, is surely a necessary step toward achieving the ideal evidence set, but at any given point short of the ideal, it is not at all clear that knowledge X + N is a better basis for decision-making than knowledge X. Maybe it is, but one would have to know the ideal evidence set and the shape of the path to that ideal evidence set in order to make that judgment.
The real question is whether there is any good way to identify instances in which more knowledge is a way station toward better decision-making but not an at-the-present ground for better decision-making. In any such instance, no deference to expertise—much less formal subdelegation to expertise—is warranted. A careful agent will have to think through these considerations before deciding to subdelegate or defer.
Fourth, even if all of the theoretical conditions for expertise to contribute to a problem are satisfied, the agent must still have good reasons to think (1) that the agent has selected the right experts in principle and (2) that those agents will employ their expertise in practice in the appropriate fashion.
The problem of how to pick the right experts is well trod, largely because it occurs routinely in trials in which lay jurors (or judges) must choose which conflicting expert testimony to credit. This is a big problem.Footnote 101 Legislators must face the same problem when deciding which experts to credit and how much, even after they have decided that it makes sense to credit experts at all. How does a responsible fiduciary choose the right experts?
Sometimes the decision process can be shaped by results. Past performance is no guarantee of future success, but if someone has successfully designed a hundred bridges that are still standing decades later, there is good reason to think that they know something valuable about bridge design. In certain contexts, there are identifiable markers of success or failure, and one can judge experts, at least provisionally, by those markers. In other contexts, however, the markers themselves, and the relationship between expert action and those markers, is more ambiguous. In those contexts, one needs some kind of proxy for results, and it is hard to find a good proxy for results.
Reliance on credentials is dubious in many—not all, but many—contexts, again for reasons that are well trod from discussions of expert testimony in litigation.Footnote 102 Credentials work best when they are validated by results, but in that case one does not need the proxy of credentials, as one can rely directly on the results. Where credentials are not validated by results, it is very hard to see their cognitive value as grounds for subdelegation.
Scott Brewer, in the context of jurors evaluating expert testimony, has aptly identified by analogy what a responsible member of Congress must deal with before deferring judgment to a supposed expert:
When a nonexpert … must decide whether to consult a putative scientific expert … that nonexpert faces four “selection problems”: (1) determining which of the intellectual enterprises that might yield expert testimony is a science; (2) determining who is a scientist capable of using her science in a manner that satisfies the standard of epistemic appraisal and the attendant level of confidence that the practical reasoner has established; (3) determining which of the intellectual enterprises that might yield expert testimony is a science that is rationally pertinent to the case … ; (4) in cases in which there is significant doubt occasioned by task (3), determining who is capable of answering (3) in a way that can identify an expert scientific discipline capable of satisfying the chosen standard of epistemic appraisal and the attendant level of confidence.Footnote 103
Members of Congress who defer without thinking through the appropriate issues to the best of their abilities fail as fiduciaries. They do not keep their end of the fiduciary social contract.
