*This is the seventh post in a symposium on Jed Stiglitz’s “The Reasoning State.” For other posts in the series, click here.
Over the last month or so, it has been such a thrill to host two book symposia here at the blog—the first one on Peter Shane’s Democracy’s Chief Executive and now this one on Jed Stiglitz’s The Reasoning State. Both books advance ambitious visions of the administrative state. For Professor Shane, it’s about limiting excessive presidentialism and empowering federal agencies to fulfill their statutory mandates based on administrative processes imbued with what he dubs “democratic constitutionalism.”
For Professor Stiglitz, the administrative state is vital to the American form of government because (and when) the regulatory process requires reason giving and rational decisionmaking. To be sure, his argument is not that Congress has no ability to build the capacity and expertise to do the job. In his account, Congress probably could if it wanted. (I’d add that it could do so by leveraging federal agencies to help draft those legislative initiatives. Molly Reynolds explores congressional capacity in her contribution to this symposium.)
Nor is his argument based in a democracy theory that federal agencies are as much if not more democratically accountable than Congress (or the President). Indeed, the book concludes with a note (pp. 294-295) about how Congress should make the major policy decisions, but we can, do, and should trust agencies to regulate the “nonfundamentals.” Or as Michael Livermore perhaps best summarizes Stiglitz’s vision in his symposium contribution, “Procedurally constrained agencies delivering credibly reasonable policy choices have served the country well. They, like elections and legislatures, are now a firmly emmeshed in our constitutional order.”
There is so much to like about this book, and the other contributors have already covered so much ground. Read them all here. In my contribution, I want to focus on several themes in administrative law that intersect with the arguments and findings in The Reasoning State. These observations overlap significantly with my thoughts on Democracy’s Chief Executive.
Democratic Accountability in the Administrative State
In my contribution to the Democracy’s Chief Executive symposium, I cast doubt on the idea that federal agencies’ “democratic pedigree” is “at least as strong as that of Congress itself.” I wish I had read Professor Stiglitz’s book before writing up that review, as his analysis nicely captures some of the “disanalogies” (pp. 142-143; footnote omitted):
Though prominent accounts channel deliberative democracy through the administrative state, it is important also to take account of disanalogies. Deliberative democracy premises on the idea of citizens, that is, individuals of roughly equal status and authority forging a consensus through reasoned discussion. The bureaucratic context, by contrast, features incredible asymmetries between the parties, that is, between the official and the regulated entity or the public more generally. This is not a meeting among equals. Moreover, the objective of the administrative process is not in any realistic sense to forge a consensus. Administrative deliberation, as it is, may result in moderation and tailoring, but at least one party is almost always worse off from regulation—to a first, second, and third approximation, pareto improving choices do not exist. The hope and expectation is that a regulation enhances social welfare, of course, but as the pages of the Federal Register and Federal Reporter demonstrate, the administrative process is often fraught and antagonistic, without any illusion even from the start of consensus as an aspiration.
To be sure, Professor Stiglitz does not dismiss entirely the idea of deliberative democracy in the administrative state. In his view, agency reason-giving requirements imposed by Congress (and courts) legitimate the agency action, at least when it comes to nonfundamentals.
I’ll have more to say about the democracy theories of administrative law on a panel at the AALS annual meeting in January. But I wanted to flag here how The Reasoning State adds to this discussion. I’d also note that another AALS panelist, Anya Bernstein, explores these themes of public participation and deliberation in her contribution to this symposium.
Constraints on Bureaucracy Beyond Judicial Review
Like Democracy’s Chief Executive, The Reasoning State seems to focus on agency rulemaking as the main form of regulatory action. A fixation on rulemaking may limit the reach of a theory of the administrative state, as the vast majority of federal agency actions today happen outside of the rulemaking context—in adjudications, informal orders, enforcement decisions, subregulatory guidance, and so forth. Consider, for instance, major regulatory actions that have reached the Supreme Court in recent years: DACA/DAPA, DACA rescission, the attempt to add a citizenship question to the census, the CDC eviction moratorium, the OSHA COVID vaccine-or-test requirement for large employers, and the student loan cancellation program.
Moreover, most of these agency actions evade, or at least are insulated from, judicial review, such that the external judicial constraints on reasoning giving and rationality will be much, much weaker in these contexts. Does the “reasoning state” only extend to notice-and-comment rulemaking? How do we craft regulatory processes outside of the rulemaking context and accompanying external checks to encourage reason giving and rationality in those regulatory contexts as well?
The Reasoning State provides some answers and, more importantly, sketches out a framework that can be adapted to these contexts. But much more work needs to be done, and overarching theories of administrative governance must better take into account bureaucracy beyond rulemaking (and oftentimes beyond judicial review). I hope to contribute to that discussion in my current book project, which I preview here.
Judicial Review and the Major Questions Doctrine
While reading The Reasoning State, I couldn’t help but think about how the Roberts Court seems to have embraced additional constraining doctrines to encourage agencies to engage in reasoned decisionmaking and to otherwise advance Professor Stiglitz’s vision of the reasoning state.
Consider a few examples. The Roberts Court requires an agency to be more deliberative when it changes an existing agency policy, including to consider reasonable alternatives and reliance interests. It discourages an agency from attempting to supplement its reasoning on judicial review. It may take a harder look at pretextual reason giving. And if an agency wants deference for an agency policy/interpretation of an existing regulation, the Court has articulated a number of important judicial guardrails, such as a reasonable explanation, a showing of official agency action, a connection of the new policy/interpretation to agency expertise, and a lack of unfair surprise or harm on regulated entities. My guess is that Professor Stiglitz would embrace these judicial innovations, when properly applied, as effective constraints on reason-lacking agency actions.
Of course, the Roberts Court’s latest and biggest administrative law innovation is the new major questions doctrine, on which Beau Baumann has provided a really helpful literature here. For those searching to understand the policy rationales for this doctrine, the Notice and Comment blog has hopefully been a useful source for diverse perspectives.
Dan Farber, for instance, has articulated how the major questions doctrine could help combat excessive presidentialism. The requirement of “clear delegations from Congress on matters that are critically important,” Professor Farber explains, is “not so much as a way of preventing Congress from giving away too much power as a way to prevent Presidents from snatching powers they were not given.”
Alternatively, I have argued that the doctrine as a policy matter can help address the difficulty of delegation and time. Simply put, textually broad statutory delegations to federal agencies can become a source of authority for agencies to take action at a later time. This later action could be wholly unanticipated by the enacting Congress and may not receive support in the current Congress. Some version of the major questions doctrine could address this temporal problem of delegation.
In The Reasoning State, Professor Stiglitz provides an information-processing policy rationale for some version of a major questions doctrine. He argues (pp. 293-294):
[O]ur broad objectives as a society should be determined through elections. This is the irreducible link that makes us democratic rather than some other form of government….
The key to understanding the proper electoral domain and the proper reasoning domain, therefore, concerns information. In the broad strokes, or in simple policy areas, such as abortion or gun control, electoral accountability functions well, and such policies ought to be left to the electoral domain. In the particularities, or in the complex policy areas, such as financial regulation or communications policy, electoral accountability does not function well, and such policies out to be reserved for the reasoning state. The public possesses the information to judge the first set of policies natively, or with modest assistance from interest groups; the public does not possess the information to judge the second set of policies, natively or with the assistance of interest groups.
In terminology resonant of the major questions doctrine, Professor Stiglitz argues (pp. 194-195) that “[t]he broad contours and thrust of major policies areas, therefore, ought to be established by Congress.” In her symposium contribution, Bijal Shah also seems sympathetic to this aspect of Professor Stiglitz’s theory, albeit recognizing that “[t]he inability of Congress to represent and act on majoritarian preferences in these contexts is a problem for another post.”
To be sure, Professor Stiglitz does not opine on the major questions doctrine as a matter of law. Nor do any of the rest of us (at least in the linked pieces above). But for those seeking to understand the policy case for some version of the major questions doctrine, The Reasoning State advances one theory that merits sustained attention.
Constraining Versus Empowering Regulatory Action
Perhaps I’m just projecting, but The Reasoning State seems to focus on process, reason giving, and judicial review mostly as means of constraining agency action. After all, the public (and courts and Congress) may be more likely to trust a bureaucracy that is constrained. But on further inspection, I think that’s a misreading of Professor Stiglitz’s vision. These features also empower federal agencies to advance effective government. And as Gillian Metzger observed in her contribution to the Democracy’s Chief Executive symposium, “Ensuring that elections translate into implemented policy is also essential for preserving democracy.” Effective policy implementation, in turn, increases trust in bureaucracy.
On a more personal note, I fondly remember our discussions at an interdisciplinary conference at Cornell years ago where Jed sketched out his trust theory and suggested a number of ways to empirically test it. It has been fun to follow the development of this project over the years. The finished product was definitely worth the wait (and the years of research on his part).
Yet, like excellent scholarship typically does, this is not truly a finished project. The book raises so many questions for further inquiry, some of which are empirical (as Kevin Stack suggests), others about evolving (or devolving) trends in bureaucracy (as David Lewis notes), and still others concern theory and paths for policy reform. I look forward to reading the future work on this subject by Professor Stiglitz and many others in the years to come.
Christopher J. Walker is a Professor of Law at the University of Michigan Law School.