We’ve been doing some philosophy of law in the PoT reading group lately, and a recent piece came up by Cass Sunstein on “The Rule of Law.” The abstract claims that “this [Sunstein’s] account of the rule of law conflicts with those offered by (among many others) Friedrich Hayek and Morton Horwitz, who conflate the idea with other, quite different ideas and practices.” This statement caught my attention because Sunstein’s account of the rule of law seems very sensible. It is in fact quite similar to the account given by our current reading group book author Lon Fuller (that’s how Sunstein’s piece came up), which also seems sensible. At the same time, I am a strong admirer of Hayek. So, if the sensible Sunstein/Fuller account of the rule of law conflicts with Hayek’s, what’s the conflict? What “different ideas and practices” does Hayek “conflate” the rule of law with? Essentially none, it turns out. Sunstein’s, Fuller’s, and Hayek’s conceptions of the rule of law are largely the same. Sunstein misunderstands Hayek’s argument that the rule of law requires economic freedom. At least, so I will argue in what follows.
Sunstein defines the rule of law as commitment to seven principles: (1) “clear, general, publicly accessible rules laid down in advance”; (2) no retroactive lawmaking; (3) “conformity between law on the books and law in the world”; i.e., the laws as written are actually enforced; (4) right of alleged lawbreakers to a hearing before judges with some degree of political independence; (5) separation between lawmaking/enforcement and law-interpretation; i.e., neither the lawmakers nor the police get to decide what counts as a crime, but independent judges who interpret the laws as written; (6) “no unduly rapid changes in the content of law”; (7) “no contradictions or inconsistency in the law”.
This list of requirements might seem unremarkable—indeed, obvious and boring. However, it is important in that it imposes constraints on what laws a legislature may enact. That is, it sets limits on legislative authority, since not just any “law” enacted by the recognized legislative authority will satisfy all the seven principles. The rule of law in this sense forms a moral standard by which government action is judged and limited.
Sunstein claims no originality for his conception of the rule of law, and indeed he cites Fuller—who defines the rule of law by an essentially similar list of principles—as a key influence. This is the account of the rule of law that I said seems very sensible.
What, then, does Sunstein say Hayek gets wrong about the rule of law? Sunstein’s criticism is based on chapter 15, “Economic Policy and the Rule of Law,” of Hayek’s 1960 book, The Constitution of Liberty, where Hayek argues that government “measures designed to control the access to different trades and occupations, the terms of sale, and the amounts to be produced or sold” are incompatible with the rule of law. In other words, attempts by the government to control wages and prices, production quantities, and who can take what jobs are incompatible with the rule of law. And the basic reason Hayek cites is that such controls cannot be maintained through general rules but must “involve arbitrary discrimination between persons” (336).
Sunstein objects that Hayek has no general criterion of what is arbitrary. For instance, in regulating who can take which jobs, surely it is permissible to require that taxi drivers have good enough eyesight or that medical doctors demonstrate a certain level of competence. Indeed, Hayek explicitly allows such rules to be nonarbitrary. Presumably, what would count as an arbitrary restriction would be, say, requiring that only white males can be taxi drivers or doctors. But how do we determine that the latter requirement is arbitrary, but the former are not? We might say that the former restrictions are relevant to the performance of the jobs in question, whereas the latter is not. But how do we know what is “relevant” to the performance of a given job? The standard of relevance cannot be derived from the principles of the rule of law alone. What is relevant to the performance of a given job depends on the substantive particulars of that job, as well as on prevailing societal norms and values (such as, perhaps, a commitment to social equality). Thus, nearly any requirements might be compatible with the rule of law—or not—depending on our assessment of the substantive facts about a given job and our societal values. Therefore, the rule of law per se does not require a free market in access to jobs.
This argument is entirely reasonable, but it is not an objection to Hayek. Hayek does not say that arbitrariness consists in having legal requirements that are not “relevant,” but in having legal requirements that are based on the discretionary judgments of officials rather than general rules. Thus, one way—the most direct—to have arbitrary rules about access to jobs is to have local officials simply hand out licenses at their own discretion. But a more insidious way is to have restrictions that distort the market and thus require officials to make discretionary exceptions on the basis of “local need.” Hayek is explicit about this one–two punch in which market restrictions violate the rule of law (337). First, rules may directly assign discretionary power to officials. Second, restrictions on the market may prevent it from functioning adequately, thereby creating “emergencies” that can be addressed only by discretionary authority.
This is particularly clear in the case of wage and price controls. Here Hayek points out that, first, long-term rules that fix prices directly cannot keep up with the constantly changing circumstances that determine prices in a market. Second, rules that attempt to determine prices indirectly as, say, a certain percentage over cost, “will not be the same for all sellers and, for this reason, will prevent the market from functioning” (337). This argument is obscure as stated. I think he must mean something like the following. If an average cost to produce some widget is assumed on the basis of conditions at the time of making the law, this will favor producers who have lower costs and penalize those with higher costs, however these may fluctuate, regardless of the reasons for the lower or higher costs, whether bad luck, bad planning, temporary market distortions with regard to labor or other factors of production, shortages, surpluses, and so forth. As Hayek emphasizes elsewhere (“The Use of Knowledge in Society,” sec. 4), particular economic agents are always confronted with particular challenges, not “average conditions.” On the other hand, if each producer is allowed a certain margin of profit above his own cost, no matter what it is, the leads to well-known problems of its own, beginning (but certainly not ending) with gutting the incentive to reduce costs. The result either way is discretionary intervention by government officials or else a dysfunctional market. Third, since prices different from what the market would set must necessarily result in either surpluses or shortages, some way will have to be found to deal with these, and this again is bound to involve discretionary coercion. Fourth, Hayek claims that “experience has amply confirmed” that effective price controls always come down to “the judgment of authority concerning the relative importance of particular ends” (337). I think this is probably true, but he cites no evidence for the claim.
As in the case of access to job opportunities, Sunstein objects that a general law fixing prices is only “arbitrary” from the perspective that the market is the only appropriate means of setting prices, not from the perspective of the rule of law per se. Hayek’s preference for market determination of prices “is a reasonable judgment, but it is not part of the rule of law” (12). And again, this misconstrues what Hayek says. Hayek is clear that governance by general rules is nonarbitrary by definition. But in order for a system constrained by the rule of law
to function properly, it is not sufficient that the rules of law under which it operates be general rules, but their content must be such that the market will work tolerably well. The case for a free system [sc., system governed by the rule of law] is not that any system will work satisfactorily where coercion is confined by general rules, but that under it such rules can be given a form that will enable it to work. If there is to be an efficient adjustment of the different activities in the market, certain minimum requirements must be met; the more important of these are, as we have seen, the prevention of violence and fraud, the protection of property and the enforcement of contracts, and the recognition of equal rights of all individuals to produce in whatever quantities and sell at whatever prices they choose. Even when these basic conditions have been satisfied, the efficiency of the system will still depend on the particular content of the rules. But if they are not satisfied, government will have to achieve by direct orders what individual decisions guided by price movements will. (338)
Thus, Hayek’s claim is that material considerations, for example of microeconomics, show that a workable system constrained by the rule of law requires tolerably free market conditions. Sunstein is right to insist that free market principles do not follow directly from the principles of the rule of law. But Hayek doesn’t claim they do. Rather, the claim is that a system constrained by the rule of law cannot survive without reintroducing arbitrary coercion—in a sense that Sunstein himself would recognize as arbitrary—unless it includes certain kinds of rules.
Notably, in the chapter of The Constitution of Liberty immediately preceding chapter 15, Hayek lays out his own conception of the rule of law, which, although less elaborate than the Sunstein/Fuller conception, conforms well with it. Hayek stresses four major principles of the rule of law. Laws must be: (1) general, “containing no references to particular persons, places, or objects” (315); (2) known and certain; (3) applied equally to all; and (4) not applied at the discretion of the legislator or enforcer. Hayek also mentions that laws should be long-term measures and never retroactive (315). The only principles named by Sunstein that I don’t see explicit mention of by Hayek in this chapter are the third and seventh. Moreover, Hayek expressly emphasizes the role of the rule of law as “a limitation upon all legislation” that thereby constitutes “a meta-legal doctrine or a political ideal” that sets the terms of what the law ought to be (310–311). Thus, so far from conflicting with the Sunstein/Fuller conception, Hayek’s conception of the rule of law is essentially similar to it.
These matters are important because, it seems to me, Hayek’s case for “a free system” or libertarian order rests fundamentally on his conception of the rule of law as a prohibition on discretionary coercion. As is well-known, Hayek’s basis for a libertarian order differs radically from the more familiar grounding in the idea of natural rights or a “non-aggression axiom” that we find in writers like Ayn Rand and Murray Rothbard. For Hayek, the essential point always seems to be to liberate people from subjection to arbitrary authority. This is why his conception of what a libertarian order requires is so much less strict than that of the writers just mentioned. For example, he will not be that libertarian who says, “taxation is theft!” and who opposes government social welfare programs and even the very institution of government as inherently coercive. That is because he has a different conception of coercion (or rather, of the sort of coercion that is incompatible with individual liberty). It is a striking claim that the ban on arbitrary coercion encapsulated in the rule of law is sufficient to generate a tolerably free market, libertarian order. Given its potential importance, it deserves to be clearly understood.
Policy of Truth 
This post is illuminating and thought-provoking. I don’t know Sunstein or Hayek’s work well enough to comment directly on their views, but the post provokes a thought on the rule of law that I think applies to both of them, and probably to Fuller as well.
As background: I work on the implementations and operations side of health care revenue cycle management. Things in my field work like this: Hospitals contract with third-party payers, whether private commercial ones (e.g., Aetna) or governmental ones (Medicare, Medicaid) for reimbursement on medical procedures performed. The payers then promise to reimburse the hospitals for these procedures according to a very complicated set of rules. Payers often deny claims based on various loopholes in the rules, so there is an additional rule-guided appeals system in place for pursuing appeals for reimbursement.
Given the complexity of both sets of rules–collections and denials–and the labor-intensive quality of applying them to a large volume of accounts, the hospitals outsource the billing, collecting, and denials processes to companies like mine. We then specialize in pursuing the accounts receivable and denials, and collecting on whatever claims we can, to whatever extent we can. When we’re successful, we deliver part of the proceeds back to the hospitals, taking a cut in contingency fees for ourselves.
To some degree, the rules involved in the preceding system are set by private, commercial contractors. But given the degree of government involvement in health care, a large proportion of the rules are governmental, formulated and administered by the Centers for Medicare and Medicaid Services (CMS). CMS was originally enabled by the legislation that brought Medicare and Medicaid into existence in 1965, but now operates by administrative rules set internally by the agency itself. For obvious reasons, Congress has no interest in running CMS on a daily basis, or in supervising its relation to individual hospitals (much less individual patient accounts). So CMS has that job. Given CMS’s stature in the field, CMS rules have essentially become the de facto default rules for all other payers, whether commercial or governmental. An additional complication is that Medicare Advantage plans straddle the fence between private and public entities, but are governed by CMS rules. Such plans are very popular, which gives CMS rules wider application in health care than they’d otherwise have.
It’s an interesting question whether the application of these administrative rules–the CMS rules I’ve just described–counts as an instance of the “rule of law.” If so, I think the application presents a challenge to Sunstein’s, Hayek’s, and Fuller’s accounts of the rule of law.
In describing Sunstein’s account of the rule of law, you say:
You describe the list as “unremarkable,” but once we consider them in practice, e.g., from the perspective of an administrative technician tasked with applying them to real-life cases, they’re anything but that. Their apparent clarity quickly becomes a murk of obscurity and contestation even in the simplest, garden-variety case. I encounter dozens of such cases every day.
Take principle (1): laws must be “clear, general, publicly accessible rules laid down in advance.” Now try to apply this to CMS rules governing reimbursement rates for Medicare Severity Diagnostic Related Groups (MS-DRGs). Here is the link for CMS’s MS-DRG Classifications.
https://www.cms.gov/medicare/payment/prospective-payment-systems/acute-inpatient-pps/ms-drg-classifications-and-software
I’ve deliberately not explained what any of that means–and it’s the tip of the iceberg. I assume that the rules in question are obscure to most people, and in that respect inaccessible. If they are “clear and accessible,” they are clear and accessible only in a highly revisionary or specialized sense of those terms. It takes years of training to understand the rules in question, and years of continuing education classes (at $200/session) to keep on top of them. How clear is any of that to the average person? Given its unclarity, how accessible is it in any practically relevant sense? If it’s neither, are CMS rules not an instance of the rule of law? Or are we obliged to re-conceive the rule of law to accommodate cases like this? No matter what answer one gives, Sunstein’s criterion is not “unremarkable.” Ironically, in invoking “clarity,” it demands clarification.
DRGs determine payment rates for procedures done for every permutation in any set of diagnoses. They change every year, subject to ad hoc mid-year revisions for budgetary reasons that arise from contingent factors (e.g., a pandemic), and fluctuations in specific patterns of morbidity (an uptick in the fentanyl crisis), length of stay (too many old people got sick this year), payment (suddenly Aetna has gotten stingy), reimbursement, etc. Is that sufficient “generality laid down in advance”? If it is, then Sunstein’s criterion is basically toothless. Virtually anything qualifies. If it isn’t, then the entire Medicare system flouts the rule of law. I have to wonder if his account of the rule of law is sensitive to this dilemma.
Now take principle (2), the prohibition on retroactivity. Let me describe a purely hypothetical case loosely modeled on real-life events with which I might hypothetically be familiar, and which might vaguely resemble what I am tasked with doing tomorrow morning.
A certain MS-DRG rule applies payment rates to geometric mean length of stay (GMLOS). In other words, for any MS-DRG diagnosis, there is a corresponding GMLOS, and the Medicare payment rate is determined by adherence to that GMLOS. The hospital gets paid the optimal rate for a given DRG only if its GMLOS for the DRG is within the prescribed limits. The administrative/budgetary aim is to reduce mean length of stay across the board for as many patients as possible.
The GMLOS table is thousands of lines long in an Excel spreadsheet. It’s supposed to be updated each year. It governs payments from July 1 to June 30 of a given year. Failure to update the table at the right time means that hundreds of thousands, perhaps millions, of charges will be wrongly calculated in a given year.
Suppose (purely hypothetically, of course) that a diligent clerk discovers that the GMLOS tables for, say, five hospitals were not updated for two years in a row–for 2020 and 2021, let’s say. This means that all of the 2020 and 2021 accounts done in 2020 and 2021 were wrongly calculated at all five hospitals. As for accounts that have already reached final disposition, those errors are at this point in the past, and will (probably) be treated as a sunk cost. But if any of those 2020-2021 accounts still remain unpaid and on the books, they are being wrongly calculated right now.
Suppose that on discovering all this, the clerk is told (now, in 2023) to update the workflow so that the remaining 2020-2021 accounts will (retroactively) be properly calculated. Notice that in this case, the legislation itself is not retroactive, but the administrative means of enforcing it, is. The clerk is, at some potential inconvenience to someone at some distance from him, reaching back into those incorrect accounts and re-calculating the charges, then calmly accepting whatever ex post facto ramifications arise from his making the change, and moving on. (Like all clerks, he reasons: “The bad news is that these changes may adversely affect some other clerk. The good news is that they won’t affect me.”)
Is this a case of retroactive law in Sunstein’s sense? Technically, it isn’t, but functionally, it is. The ills that arise from retroactive law arise in this case as well, and arise through retroactivity in the application of legal rules. So even if this isn’t literally a case of “retroactive legislation,” it’s not clear that that matters. It’s a retroactively applied legal rule.
I should add that this sort of thing happens all the time in my field–probably millions of times a day, with problematic application to millions of accounts a day, involving transactions in the billions upon billions of dollars. Health care is the third biggest industry in the United States (after pharmaceuticals and banking), so this is not a fringe or exotic phenomenon. This is what the rule of law actually looks like in health care: retroactivity is par for the course. Question for Sunstein: is health care administration then law-like or lawless?
As for (3), the preceding example does double duty here. Mistakes of the preceding sort are made all the time, as are the fixes required to rectify them. Strictly speaking, all such mistakes and all such rectifications deviate from the black letter of the “law” (taking administrative rules as “law”). If so, deviations from the law are ubiquitous in health care administrative law. If the rule of law requires adherence to all CMS rules exactly as written, and requires it all at once, then the rule of law is a noumenal phantom no one ever encounters in the phenomenal world.
Cass Sunstein may have the luxury of thinking otherwise, but I don’t. He has the luxury because as an academic lawyer, his relationship to administrative law is entirely spectatorial and retrospective. He arrives at the scene as a critic surveying what others have already done. It’s different if you arrive at the scene having to get things done ab initio. The question you then face is: which rules shall I follow, which shall I violate, and which shall I ignore? No one can follow them all. The legislators and lawyers have made sure of that. Once the law takes the form of specialized volumes numbering in the tens of thousands of pages, it becomes pointless to demand adherence to all of it. To do so is to traffic in fantasy. Once you cross a certain level of complexity (long ago crossed), the idea of consistency between the letter of the law and its real-world practice becomes an absurdity, not a necessary condition of the rule of law.
Though I’ve aimed this diatribe at Sunstein, I think there is a problem here for Hayek, as well. There is no clear sense in which the administrative rules I’ve described are “arbitrary.” On the contrary, the rationales given for them are spelled out in thousands upon thousands of pages of mind-numbingly explicit memoranda, appendices, legal opinions, and explanations. In no case have I ever encountered an administrative rule whose rationale was, “I just felt like enacting this regulation today, so I did.”
It is not clear to me why “having legal requirements that are based on the discretionary judgments of officials rather than general rules” violates the rule of law. In criminal justice, prosecutors have to decide, on a discretionary basis, which crimes to prosecute and which to ignore. Police officers have a similar discretion. The same thing is true of administrative clerks, judges, and the like. Discretion is an ineliminable part of human life. Why should it be eliminated from law? I don’t mean to suggest that it’s never a problem. But it can’t always be one, either. If discretion is a problem, we need a better account of how and why it is. There is a huge difference between the discretion one encounters at the hands of CMS even at its worst, and the discretion one encounters from, say, COGAT (the agency responsible for the military occupation of Palestine) even at its best. The difference isn’t captured by Hayek’s saying that both operate “by discretion.”
I’ve gone on too long, but I have a lot to say about your long paragraph on wage and price controls. There is a sense in which the entire American health care sector operates under price controls. In one sense, it leads to a “dysfunctional market,” and in another sense, it doesn’t. A great deal turns on the criterion of functionality. But more on that some other time.
(In this comment, I write exclusively at my own initiative and responsibility, and do not intend to represent my employer, CorroHealth, LLC, or any other organization.)
LikeLiked by 1 person
Hi Irfan,
Most of your discussion here depends on details I have no independent knowledge of, so it’s hard for me to comment. Nevertheless…
Many of your points seem to be of the form, “Sunstein says the rule of law requires laws to be clear/ prospective/ enforced as written, but here are some laws and they aren’t clear/ prospective/ enforced as written.” But Sunstein—or any of these Rule of Law guys—isn’t saying these are attributes laws have to have or else they aren’t laws. This might be nonobvious inasmuch as the natural law tradition we’ve talked about in the group (as opposed to the positive law tradition) seems ready to say precisely that “laws” that don’t fulfill certain moral criteria aren’t laws. However, Fuller doesn’t talk that way. For example, Fuller cites Hitler as making retroactive laws (54–55), but he doesn’t say these weren’t laws.
Anyway, “the Rule of Law” isn’t a descriptive doctrine, but a political ideal that lays down certain principles a legal system should have in order to create a “government of laws, and not of men.” This phrase is apparently due to John Adams, though the ideal is old (Adams cites Aristotle and Livy). Adams associates the rule of law with liberty (Hayek is not alone), though Aristotle (from my extremely brief review) seems to associate it only with good government. That seems to be Sunstein’s point, actually. Sunstein seeks to detach the virtues of the rule of law from broader claims about the promotion of liberty, equality, democracy, and so forth.
Proponents of the rule of law like to make lists of the principles to which laws should conform in order generate a system characterized by “the Rule of Law.” Fuller, Hayek, and Sunstein are all cases in point. To the extent that a given legal system conforms to the principles, it implements the Rule of Law. To the extent that it doesn’t, it doesn’t. But this is a matter of more or less, obviously, and these authors don’t deny that the laws that fail to conform to their principles aren’t laws, except in outrageous cases.
It turns out there’s a pretty good SEP article on the rule of law that explains all this and more, and that treats both Hayek and Fuller as major exponents of the rule of law idea.
Concerning Hayek and arbitrarity, being arbitrary for him is not a matter of whether a rationale is given for a law. A law that was promulgated totally on whim would still be nonarbitrary for Hayek, I think, as long as it was prospective, clear, promulgated publicly, etc. etc. “Arbitrary” for Hayek really seems to mean, in this context, not being at the whim of officials on the spot.
For Hayek, the core idea of the rule of law is predictability. He says somewhere in The Constitution of Liberty that the ideal is for the laws of the legal system to be like the laws of nature: knowable so that you can calculate the consequences of your actions. Hayek’s concern with arbitrarity is as a major defeater of predictability.
LikeLike
Hayek likens proper laws of the state to laws of nature at CoL p. 221.
LikeLike
I think I may have more of a case against Sunstein than against Hayek. My argument against Sunstein is that he has greatly underestimated the distance between what he regards as essential to “the rule of law” and the kind of flexibility that is required to get things done within an administrative regime of the kind that liberals like him want. My point is not just that there are a few laws here and there that contradict Sunstein’s conception of the rule of law. It’s that once you descend from law review abstractions to the operational details of administrative life, the entire CMS administrative regime violates his conception of the rule of law. That’s a huge swatch of legal life and a big chunk of change.
If I’m right, then liberals like Sunstein have to choose. Do they want the rule of law as he conceives it, or do they want a CMS-style administrative regime in, say, health care? If he wants the former, then the more assiduously he wants it, the less of a CMS-style regime he can consistently have. If so, then “Medicare for All” becomes a non-starter. And to the extent that the rest of the administrative state is like CMS, then that becomes a non-starter, too. I don’t have any experience in the non-health-care parts of the administrative state, but I suspect that similar issues arise there as arise for health care.
Put another way: either Sunstein has to find a way of saying that administrative operations are not an instance of the rule of law at all, or he has to modify his conception of the rule of law, or he has to jettison the daily operations of the administrative state. But he can’t have it all ways at once. He can’t say: “I want the rule of law as described in the abstract, and I want an administrative state that lives up to the rule of law, and I want CMS to continue with business as usual, and I want Medicare for All.” That won’t fly.
There just seems an element of a priorism in Sunstein’s views. He has this normative ideal of the rule of law, but no sense of whether it coheres with the actual practice of the administrative state he otherwise defends. If it doesn’t cohere, I suppose he could demand that the practices be changed. But the practitioners could equally demand that Sunstein’s conception of the rule of law change. In the absence of any realistic sense of the practical demands of running an administrative state , Sunstein is not in a position to take a hard line there.
On Hayek, I guess I would ask: predictability to whom? The law is more predictable to some classes of people than others. As an analogy, just think about your medical bills. Are they “predictable”? Well, they’re predictable to me, but that’s because I deal with other peoples’ medical bills all day. Normal people find medical billing incomprehensible and find their own bills unpredictable. (“How did I get charged $4,000 for a vial of antibiotic?”) It’s worth noting that medical billing is in one sense supposed to be rule-driven (hence totally predictable, however complex). Yet the decision to write off a bill is literally a matter of whim. A manager just decides, “I’m sick of this bill (or batch of bills)” and it gets written off. I don’t know how Hayek’s views apply there, assuming that the billing process is a matter of law.
In law enforcement, the standard for a “reasonable suspicion” detention is: an individualized articulable suspicion that the target of suspicion either had, was, or was imminently about to commit a specific legal infraction, where the suspected behavior would be regarded as such by a trained law enforcement officer. As long as the officer operates in good faith, the standard is met. But “suspicion” is highly discretionary. Is it a “whim”? No. But can two people look at the same evidence and come to contrary positions consistent with the standard? Yes. The standard is frequently indeterminate between contrary outcomes. I’m guessing that doesn’t count as “arbitrary” for Hayek, but it seems a tough call to me.
LikeLike
As you can imagine, I have no desire to defend the administrative state. However, I’m having trouble seeing how an administrative state bureaucracy must inevitably run afoul of Sunstein’s Rule of Law. Consider the IRS and the federal income tax. The tax code is notoriously complex, and no doubt it is full of a lot of bad rules. But I don’t think it strongly violates the rule of law. There are pretty clear rules about what you have to do to pay your taxes and how to calculate what you have to pay. It’s so cut-and-dried that you can buy software for your PC that will do your taxes for you. The rules are byzantine, to be sure, which might seem to be a problem vis-a-vis the rule of law. But I don’t think it really is. In a modern society, the law will necessarily get technical and become the province of a professional class in many areas. And so individuals and businesses will need to consult these professionals—as one might (and I certainly do) hire a CPA to do one’s taxes. But what is important with regard to the rule of law is that, however complex and technical, the law is clear and determinate, thus allowing the said professional class to exist. I feel like (but maybe I’m delusional) the income tax code is like that. There are clear rules, even if I personally don’t know them, which, if followed, mean I don’t have to worry that there’ll be a knock on my door by people who have come to arrest me. That is basically what we are talking about with the rule of law.
Again, the EPA is not high on my list of most cherished institutions, but my impression (again, maybe I’m deluded) is that they operate by publishing clear rules of what individuals and businesses can’t (or sometimes must) do. It’s possible to know with a reasonable degree of certainty what these rules are and calculate one’s actions accordingly. I just don’t see how agencies like this must necessarily violate the principles of the Rule of Law.
Contrast the behavior of the FDA in the drug approval process. This is something I know a little about because my wife works in biotech. Getting a new drug approved (or an existing drug approved for a new use) is a matter of convincing officials at the FDA that the drug is safe and effective for whatever it is to be prescribed for. This can be a matter of satisfying arbitrary-seeming demands for more data, analyses, revisions, and so forth, with short notice and short deadlines. Recently, an FDA potentate made a site visit to my wife’s company, insisting on 7:00 a.m. meetings throughout the week. This means people are flying in from the west coast and Europe to be on hand for this (to the company) important site visit. The official arrived several days after she said she would and held no meetings that week, but basically just stayed in her hotel room, and then held meetings on a couple of days the following week—meaning those who had flown in had to stay—and then left. And the company’s response to this shit basically has to be, “thank you sir may I have another.”
I realize that this is a very mild example of the arrogance of power and that in other countries it is much worse and much more pervasive. Nevertheless, this kind of thing is the target the rule of law tries to eliminate. Does the FDA have to be the way it is; i.e., an agency that personally oversees drug development with a whip in its hand? I doubt it, but that seems like a large topic for another time.
Getting back to CMS, the problem I have is that it’s not really clear to me that the medical payment system as you have described it seriously violates the rule of law principle, much less that it has to. It seems like when a hospital performs a procedure—a colonoscopy or gastric bypass or whatever—it should be able to calculate what it will be paid for that work, given the patient’s insurance situation. Are you saying that is not true? I have had all too many hospital procedures in the past several years, and that is the way it has worked so far as I can see. And if it doesn’t, what is the reason it has to be that way? Where is the arrogance of power in CMS?
LikeLike
I think it’s too strong to say that an administrative state has to run afoul of the rule of law, but I think there is a good reason to see a strong, predictable tendency there. It is not obvious that human life, as such, can comprehensively be brought under the control of general, legalistic rules. The problem is mitigated if we confine the scope of such rules to a defined, limited part of life. But the administrative state doesn’t do that. Little by little, it tries to legislate for life as a whole. The more it does that, the more it tries to do something either impossible or extremely implausible: bring human action under the literal, supervisory control of legal rules, where the rules aim to prescribe or at least constrain action down to the details (to avoid the specter of “arbitrary discretion”). What starts out as a mere practical problem eventually becomes a series of dilemmas and then a paradox: how to combine the generality and predictability of law with the flexibility and retroactivity required of administrative decision-making at the fine-grained level?
Retroactivity is required to correct inevitable mistakes, mistakes so numerous and ubiquitous that their frequency and nature have to be concealed to protect the legitimacy of the enterprise. Just as no corporation would permit total transparency, neither does any government agency. Opacity is there to conceal retroactivity.
The irony here is that I am far more sympathetic to the administrative state than you are. But I am skeptical that its operations can be reconciled with the rule of law as Sunstein defends it. Either it can’t, or large swatches of it can’t (large enough to constitute a problem for Sunstein), or the rule of law has to be re-conceived to allow the administrative state to do what corporations do without inviting any comparable scrutiny. In theory, corporations operate within the scope of the law, but they don’t take direction from law as far as operational matters are concerned. The problem I’m describing arises because government administrative agencies are often trying to do the work of private corporations, but unlike corporations, are obliged to do them through the literal prescriptions of the law. The law literally directs their operations. Corporations do not pretend that their operations proceed by anything comparable to the “rule of law.” Government agencies do. But it strikes me as a pretense.
I wasn’t sure whether you meant the contrast between the FDA on the one hand, and EPA and IRS on the other to imply that the FDA runs afoul of the rule of law whereas the EPA and IRS operate consistently with it. The behavior you describe in the case of the FDA happens at the EPA and IRS all the time. The most obvious example is audits. By their nature, audits have to involve an element of surprise. Audits work best precisely to the extent that they are vague, arbitrary, and indeed retroactive. There has to be a sense in which the audit process is never closed, or at least a threat of indefinite, arbitrary process. Otherwise, the targets would simply figure out how to game the system and outwit the auditors. But this is just a way of saying that audits work best to the extent that they flout the rule of law in Sunstein’s sense. If they adhered to the rule of law, they wouldn’t work.
I worked in the Regulatory Compliance division of a large agricultural research firm subject to EPA and Dept of Agriculture regulations. Our products included pesticides and herbicides subject to the Federal Insecticide, Fungicide, and Rodenticide Act of 1947 (FIFRA). What we faced in the way of regulatory compliance was indistinguishable from your description of your wife’s routine vis-a-vis the FDA. That was just business as usual, including the 7 am surprise site visits. If you meant that example to be non-adherence to the rule of law, the same thing is true of the EPA.
Something similar is true of IRS regulations. Here is a source I don’t often cite–National Review.
https://www.nationalreview.com/2023/01/an-irs-that-is-above-the-law-threatens-taxpayers-security/
The author is correct. The enforcement of tax law affords the target of an audit fewer protections that would be afforded the target of a murder or rape investigation. I’m not sure the whole article is visible, but the author’s point is that the audits are not circumscribed by the Fourth Amendment’s constraints on searches or seizures. You can be audited for any reason or none.
Again, here is an article suggesting that audit targets will now be selected by AI.
https://abovethelaw.com/2023/09/the-irs-will-use-artificial-intelligence-to-help-pick-which-large-law-firms-to-audit/
You couldn’t conduct traffic stops or stop and frisk people by feeding information into some equivalent of ChatGPT and asking it to spit out targets. Or rather, you couldn’t do so consistently with any plausible conception of the rule of law. But you can enforce tax law that way. Revenue collection matters more than crime reduction. The more something matters, the more administrative flexibility is required to make decisions about it. And the less subject to the rule of law it will be.
To come back to CMS: I think you’re missing a crucial part of the medical billing process. It may not have arisen in your experience, but it constitutes a gigantic part of the process as a whole.
Suppose I have a Medicare Advantage plan. I have a contract with Medicare Advantage to pay my medical expenses under the contracted conditions. My hospital likewise contracts with Medicare Advantage for reimbursement of the procedures it performs. So far, it looks like pricing should be a rule-bound, even algorithmic and predictable affairs.
Suppose I go in for the procedure and run up $10,000 of charges. Suppose that Medicare Advantage promises to pay the hospital 80% of my charges for a certain set of procedures associated with a DRG, as long as they are “medically necessary.” They get the bill for $10,000 and decide that (a) the coding was done improperly and (b) the procedure was unnecessary. So they deny payment: 0% of incurred charges are payable. Denial (a) is known as a technical denial, denial (b) as a clinical denial. The hospital now begins the appeals process to recoup on this denial in the hopes of collecting as much of the original $10,000 amount as is compatible with the expense of the appeals process itself.
This is just a political process. Medicare’s incentive is to delay and wear the hospital down. The hospital’s incentive is to glut Medicare with so many appeals that it will grant some without really looking hard at them, on the grounds that scrutinizing them is more trouble than it’s worth. Meanwhile, Medicare has the option of ex post facto recoupment. Suppose that it agrees to pay out, but retrospectively (via an audit process) comes to believe that it has overpaid. Then it issues a recoupment order, and the hospital is required to repay what had been paid. Here is an account of the recoupment process:
Click to access overpaymentbrochure508-09.pdf
In other words, that is a brochure codifying the legal protocols governing an explicitly retroactive legal process. It says: “We agreed to pay you X, but in retrospect, we feel we shouldn’t have, so we want the money back. We are the government, so we get do-overs. You don’t, obviously. CMS could hardly function if it treated retroactivity as a reciprocal, bilateral matter.”
I guess someone could say that the Medicare recoupment process is not literally retroactive because CMS tells you, prospectively, that it intends retroactively to recoup payments. But let me set that aside.
My point is not that the appeals process is a literal, all-out free for all. My point is that it’s extremely doubtful that it satisfies Sunstein’s conditions for the rule of law. It is impossible to know how the appeals process will turn out. The central legal term, “medical necessity,” is a meaningless moving target. An enormous amount of discretion is applied at every stage. Most of the process takes place in ways that are inaccessible to outside scrutiny, so that the potential for divergence between legal rules and actual practice is extremely high. For that reason, the idea that the process is constrained by rules is just a convenient fiction. No party applying “the rules” can know, even in principle, the content of the sum total of rules that apply to their decisions. It’s not humanly possible.
Obviously, my experience is limited. I’ve only seen what I’ve seen. But I think Sunstein’s view is extremely ingenuous. Without making any claims about necessity, it strikes me as highly likely that the sorts of considerations I’ve described in the CMS case will apply in analogous cases across the board within an administrative state. They don’t have to apply universally. They just have to arise often enough to constitute a constant, recurring problem. A view like Sunstein’s fails to acknowledge the trade-off between the normative ideal of the rule of law, and the efficiency demanded as a matter of course of any administrative body. An administrative entity gets rewarded by delivering the goods. But that requires administrative flexibility and a certain degree of arbitrary enforcement power. Both of those things stand in tension with the rule of law a la Sunstein.
LikeLiked by 2 people
Sorry to take so long to get to this. I think you make a good case here. The basic argument: There’s a difference between, on the one hand, merely establishing and enforcing the equal, impersonal rules of the societal game, and on the other hand, being an agency tasked with producing some substantial deliverable (such as gov’t revenue or “safe and effective” new pharmaceuticals). The latter is the sort of thing that is done in the private sector by corporations, and there are well-known good reasons why corporations don’t operate by impersonal rules. Such an agency “requires administrative flexibility and a certain degree of arbitrary enforcement power” that is incompatible with the ideal of the rule of law.
I’m left with two further thoughts. First, I wonder the problem might be mitigated depending on how an agency construes its mission. I recall Tyler Cowen somewhere saying something like, that once the FDA frames its mission as being to insure the safety of all new drugs and to protect Americans’ trust in it (the FDA), as opposed to establishing rules for new drug approvals and working to make those the best possible rules, it’s sunk. That’s vague, but I hope you get the idea. He’s saying the FDA is taking too much burden on itself. It is inflating its mission beyond what it needs to be and creating worse outcomes by so doing. This wouldn’t completely eliminate the problem, of course (and with an agency like the IRS, the problem may be intrinsic to its job), but it might help.
Second, I continue to be struck by the fact that the behavior of the administrative state in the US (and certain other countries) is pretty good, in spite of the potential and indeed incentive for abuse. There are many many places where the administrative powers of gov’t are much abused. What makes (and maintains) the difference is the $64K question.
LikeLike
I can’t speak to the second, comparative issue you bring up, but I think that Cowen’s claim is implausible. More precisely, the distinction he’s drawing, between a minimalist and a more maximal conception of an agency’s mission, is artificial and impractical.
The clearest rationale for establishing the “best possible rules” for drug approvals is to ensure, insofar as possible, the safety and efficacy of new drugs coming onto the market. Rules only get a rationale from the end they serve, and the clearest rationale for the best possible rules is their conducivity to an end that has some claim to being, approximating, or striving to approximate, some optimal state of affairs. Ratchet down the end, and you end up having to re-describe the rules. In that case, you’re no longer talking about the “best possible rules,” but about “rules that are better than the absence of such rules would have been.” That starts to sound like a synonym for mediocrity. It would be very difficult to generate a budget or hire talent for such an agency. It would also be very demoralizing to work for. The drive to optimize is inherent in a successful enterprise. Subtract it from the enterprise, and you’re putting it on track to extinction. But leave it in, and you’re giving the agency an end that gives it a built-in impulse to mission creep.
In any case, even if you set the preceding argument aside, rules have to be enforced. A regulatory agency that depends more on voluntary self-enforcement than on coercive imposition has lost its rationale as a regulatory agency. Once it admits that its regulations are largely self-enforcing, it has admitted that regulations can be self-enforcing, in which case, it raises the question: why then do we need a regulatory agency? Why not just leave the regulatory process to voluntary self-enforcement? The raison d’etre of a regulatory agency is that some things can’t be left to voluntary self-enforcement. They have to be imposed, because it’s intolerable to permit the thing they’re trying to prevent from happening.
If so, cultivating citizens’ trust in the agency becomes a built-in imperative. If an agency is going to use force, it needs legitimacy. In particular, it needs to have the sort of legitimacy that confers the authority to use force against a set of institutions that can’t police themselves. Given that description, a regulatory agency stands above the regulated institutions, and to keep itself in that superordinate position, has to make (and make good on) some claim to being morally superior to them. You can’t just assert that you occupy a position of moral superiority. You have to demonstrate it somehow in an ongoing, counterfactually stable way. Otherwise, you lose it. But that’s just what the cultivation of trust is. Take it out of the equation, and a regulatory agency starts to lose its efficacy. It just looks like a big bully wielding a big stick.
So the cultivation of trust will have to be an ongoing, standing enterprise on par with everything else a regulatory agency does. The department within it responsible for the cultivation and reproduction of trust-in-the-agency (“PR”) may well be as big as the department responsible for its ostensible first-order mission. The two things are complementary, and will likely be seen that way.
The idea of an FDA (or any regulatory agency) that somehow virtuously abstains from selling itself as a trust-inducing enterprise strikes me as fantasy. Even ordinary police departments and Departments of Health do that kind of thing. From roughly 1999-2005, I worked for the National Assessment for Educational Progress (NAEP), a program of the purely statistical arm of the US Dept of Education. We didn’t regulate anything, but we still needed to work hard to maintain the trust of the American people. Doing so qualified as a standing project of the agency itself, alongside data collection, data analysis, and all the rest. We had to work, night and day, to persuade our fellow citizens that our data collection activities had no nefarious political purpose beyond the uncontroversial one of improving of educational outcomes. That “uncontroversial” goal sounded deeply sinister to a lot of people. And it would be disingenuous of me to pretend that NAEP was a purely neutral enterprise with no political agenda whatsoever. The agenda was centrist, but it was still an agenda.
I wasn’t in the PR part of the agency, but there were people who did nothing but work at proving–to Congress, to the press, to our governing board–that we could indeed be trusted.
Click to access 2010468.pdf
Actually, as I think about this, I find myself wondering whether they have any openings, and if so, what they pay.
A federal agency, operating in a more competitive environment with higher stakes, is likely to do things in a much more intense way than, say, a local police department or Dept of Health, and could hardly be constrained from doing so. So I don’t think Cowen’s gambit offers a way out.
Now, if you’ll excuse me, I have to dust off my resume. It’s time for some mission inflation.
LikeLike
🙂 It could well be less stressful and better paying than what you have now.
I get your point about the “Cowen gambit,” but I’m not sure I described it very well. Of course, that’s hardly surprising. It’s something he said once or twice in 2020 (I think) on the MR blog. I was fuzzy about his meaning at the time, but it has stuck with me for some reason. I tried to find it before posting my earlier comment, to understand it better, but I couldn’t find it.
Here’s another stab. As I recall, he was contrasting the FDA with the FAA. They are both trying to protect public safety. But if a plane crashes, the FAA doesn’t think it’s their personal fault, whereas if a drug turns out to have some disastrous consequences, the FDA does seem to think they’re personally responsible. It’s this framing (I think) that he was saying is a big reason why the FDA is such a terrible agency.
Cowen didn’t say much more, but I imagine he was thinking about the opportunity cost of all the potentially life-saving and -bettering drugs that are delayed or never introduced at all because of a regime that emphasizes drug safety above all else. The Thalidomide disaster apparently was instrumental in motivating this attitude on the part of the FDA—and agencies such as the EMA (European). The FDA wasn’t the regulator that approved Thalidomide, but the disaster nevertheless apparently rocked the FDA and had much to do with its current procedures. This is a classic case of seen versus unseen harm due to regulatory failure. When side effects of an approved drug cause harm, that is very visible and causes newspaper headlines. But the many thousands of Americans who suffer and/or die every year for lack of treatments held up by the FDA are relatively invisible and make no headlines, and certainly no headlines linking them to the FDA. Yet the FDA is responsible for these costs also. The FDA may have more Americans’ blood on its hands than all other federal agencies combined.
So, the point would be that conceiving the goal of the FDA’s rules merely to be ensuring safety and effectiveness of new drugs is too limited. (I know, I know, I’m the one who said that!) They should also think of themselves as responsible for the efficacy of their approval process.
However, if this is the point, then we’ve strayed pretty far from the original topic. This has gotten to be less about whether or how far an administrative agency can conform to the rule of law ideal than about how to think about the mission of such agencies. If what I’m saying bears on the rule of law question, it would be in that an agency might be less dictatorial if it sees itself as establishing rules for and facilitating a valuable public purpose carried on by others, not running it and personally responsible for all outcomes.
LikeLike
A very quick response to your December 28 comment. I looked for the Tyler Cowen post, too, but though I found related ones, I didn’t find the particular one you had in mind. I didn’t really have time for a systematic search, however.
I guess I remain to be convinced that Cowen’s distinction between the FAA and FDA really exists (I’m agnostic there), but even if it does, it seems to me that it can be explained by differences between air crashes and failures of drug safety and efficacy. The costs of an air crash are likely internalized by the airline. Air crashes involve costs that an airline cannot easily externalize: lost aircraft, lost personnel, reputational loss caused by the visually spectacular nature of the event. So the FAA has less to do there, and can perhaps afford the relatively hands off attitude you ascribe to Cowen. Voluntary self-policing works in a case like this.
But the costs of clinical failure in pharmaceutical use are very different. Thalidomide-like cases are very rare. Most cases of liability-worthy failure, I suspect, involve a subtle mix of failures of safety and efficacy that blur the distinction between those two things, are hard to detect, and even when detected, involve populations that can’t impose liability on a pharmaceutical giant. So the FDA has more of a mission to bring about. So I think I disagree with this: “When side effects of an approved drug cause harm, that is very visible and causes newspaper headlines.” Maybe sometimes, but not all of the time. I think what you say is more true of cases of outright causation of harm (of the Thalidomide variety) than cases of harm caused by inefficacy.
That, too (harm caused by inefficacy), is a harm, sometimes, a very great one. If I market a drug that does not overtly harm you (or doesn’t do so in any spectacularly obvious way), but is hyped as curing or ameliorating your medical condition yet does nothing for you, or does something subtly harmful, I’ve harmed you. If on every occasion of the drug’s failure, you complain that the drug is “not working,” and my response is, “Yes it is; the drug has proven efficacy. Clearly there is something wrong with you, not the drug,” then I’ve compounded the initial harm. It’s not clear whether to describe this case as one where the drug is harmful or merely ineffective, but the overall effect is long-term harm. Had the marketing hype been omitted, you might otherwise have pursued other options, but the scientific-sounding hype convinced you to pursue this option. Your failure to stick with the option then gets re-described as moral weakness or irrationality. This is an “unseen harm” that fans of Bastiat tend to miss when it comes to goods like pharmaceuticals and their regulation.
My wife was in this position not with respect to a drug, but with respect to a medical device–her spinal stimulator intended for chronic back pain. It was sold to her as a miracle device, and as someone who suffered terrible chronic pain, she was only too happy to seize on the miracle. But it was not at all the miracle device it was touted as being. In retrospect, I regret that I was as motivated by wishful thinking about it as she was. In this case, the failure of the device played some very-hard-to-measure role in her death. She took her life over generalized despair over her condition, itself exacerbated by mental health and substance problems; the failure of the stimulator (if it was that) played some peripheral role. But it arguably played a harmful, largely unseen role.
I think this pattern plays out elsewhere, and to the extent that it does, and is recognized as doing so, the FDA will always have a motive for expanding its mission to cover cases like this, however attenuated the causality in question. And parts of the causality clearly not within the purview of the FDA will likely be farmed out to other agencies. We are currently suffering what’s being called a crisis of “deaths of despair,” and there is enormous pressure on government to do something about it.
https://www.newyorker.com/magazine/2020/03/23/why-americans-are-dying-from-despair
The FDA is in the forefront of dealing with this deaths of despair crisis, but pressure is building to outsource the non-FDA parts of it to other agencies.
Having said all that, I have very mixed feelings about the FDA, so I agree with you “many thousands of Americans who suffer and/or die every year for lack of treatments held up by the FDA are relatively invisible and make no headlines, and certainly no headlines linking them to the FDA.” It’s difficult to know how to capture and respond to every such cost, visible and invisible. But this would be a problem for an entity tasked with quality assurance over drug safety and efficacy, whether governmental or otherwise. The problem of setting priorities and dealing with the unintended consequences is inherent in the task itself. No institutional set up can make it go away.
The distinction you draw, between “how far an administrative agency can conform to the rule of law ideal [and] how to think about the mission of such agencies,” involves considerations that are so closely related that though we can distinguish them for conceptual purposes, in practice, virtually every discussion of the one will involve discussion of the other. The issue first arose in US history during the presidential “administration” of George Washington: when Washington was first elected, no provision had been made for the creation of any administrative infrastructure beyond the bare-bones described in the Constitution. And that’s pretty uninformative. The Constitution vests the executive power in the president, but nothing in Article II of the Constitution says anything about the president’s authority to set up the infrastructure of an executive administration– a Cabinet, for instance. We’re told that he is to receive foreign “ministers,” but nothing is said about his authority to have ministers of his own. Nothing is said about his authority to issue Executive Orders, or whether he can do so at all.
On an absurdly strict construction of the Constitution, the President can do all and only those specifically enumerated tasks listed in Article II. Anything else violates the Constitution, hence the rule of law. Taken that way, we would not only never have had cabinet-level positions in the White House, but never had an air force. Neither is mentioned in the Constitution.
On a more sensible view, “the executive power” is a definite but unenumerated power that has to be worked out as we go along, but worked out with some ideal in mind of what executive power ought to be. If so, we have to think about the idealized mission of an executive power as we go along, and reconcile it with our conception of the rule of law. If this is right, there is no good way to make a very sharp distinction between the two inquiries. you mention. Any conception of the rule of law that conflicts with our conception of the ideal mission of government, or a particular agency, will reasonably be judged inadequate; it’ll be seen as lacking a justifiable raison d’etre.
I’m left with the thought that mission creep is inherent in human institutions, and can only be combatted in a given case by counterposing one sort of mission creep with the more powerful mission creep of some competing institution. That poses a perpetual problem for the rule of law.
That was not a “very quick response,” alas. My comments seem to suffer from mission creep, along with so much else in the world.
LikeLike
It occurs to me that my off-the-cuff description of Washington’s presidency is both over simplified and imprecise, but the general point is correct: there is a sense in which the Constitution does, and one in which it doesn’t, make provision for an administrative bureaucracy of the kind necessary to run a federal government. The size and scope of the institution we now call “the cabinet” was left indeterminate. Washington filled the indeterminacy by adopting a basically Hamiltonian conception of government that isn’t specified by the Constitution, and could arguably be thought to violate the rule of law, taking the Constitution absolutely literally.
LikeLike
You’ve got that right.
You’ve got a lot of interesting thoughts here, but I guess I’m going to move on. There are some other posts I want to make before the all-too-brief winter break is over. I would like to begin posting more in general. It’s been nearly impossible for me, however, during the semesters. We’ll see if I can do any better in the coming spring. After that, I’m getting out of academia.
On the point about the supposed crisis of “deaths of despair,” you’ve mentioned that before. I know Angus Deaton won the Nobel Prize and all, but that doesn’t mean there’s really any such phenomenon as he claims. See the excellent Matt Yglesias on this.
LikeLike
When I wrote my comment, I had not yet finished chapter 4 of Fuller’s book, and then forgot to do so. I just went back and re-read it this evening, and saw that Fuller makes much the same point I make in my comments above, though in more general terms, and without all of the tedious detail I give. It’s on pp. 170-77, under the heading, “Legal Morality and the Allocation of Economic Resources”:
The whole section is worth reading, and I think, largely correct. It now occurs to me that my criticisms apply much more to Sunstein than to Fuller, and that Sunstein would do well to respond to criticisms of the sort Fuller makes in this part of his book. I have mixed feelings about Fuller’s overall theory, but find him insightful on the challenges of integrating a regulatory-administrative state with the intuitive conception of the rule of law. Managerial functions in business do not function by anything that resembles “the rule of law.” Insofar as government decides to engage in business management, it can’t help departing from that ideal as well.
LikeLike