Public Institutions Can’t Outsource Their Reasoning

Guest Blogger

For the Balkinization Symposium on the Global Political Economy of Artificial Intelligence.

Ignacio Cofone

[This essay distills part of the argument in Ignacio Cofone, Institutional Accountability and Legitimate Inference in Algorithmic Adjudication: Beyond Trustworthy AI, forthcoming in Cambridge Forum on AI Law and Governance (2026).]

AI does not relieve courts and administrative agencies of the duty to defend the reasoning behind their decisions but, on occasion, it does make that duty harder to satisfy.

For most of the 2010s, the Dutch tax authority used a self-learning algorithm to flag potential fraud in claims for childcare benefits, a means-tested subsidy that helps parents cover daycare costs. The system assigned higher risk scores to families with certain characteristics, including dual nationality. When civil servants reviewed flagged claims, they were given no information about why the system had assigned the score. More than 26,000 families were wrongly accused, many of them ordered to repay tens of thousands of euros in full, often with penalties and no installments. Some lost their homes, their jobs, or custody of their children. In January 2021, the entire Dutch cabinet resigned over the scandal.

The Dutch Data Protection Authority called the practice unlawful and discriminatory when it fined the government under the GDPR. A parliamentary inquiry found a violation of the rule of law. The Dutch high administrative court had reviewed individual cases for years without catching any of this. All these failures came back to the same institutional defect. A self-learning model was producing decisions affecting thousands of families, and the institution running it could not, on demand, reconstruct any of those decisions in terms the law could evaluate. The reasoning path leading to action against a family was opaque to everyone in a position to challenge it: the family, the civil servants reviewing the flag, and any court asked to review what the institution had done.

In Toeslagenaffaire, everything that is supposed to make a system like this accountable was in place. There was an approved algorithm and a procurement process behind it. Civil servants reviewed every flagged case. Procedures, escalation paths, and appeal rights existed on paper. None of those gave the institution the ability to answer for the outcome. The AI does not take that role on. Trustworthiness in adjudication is the institution’s work.

Courts and agencies derive their authority from the procedures they follow, such as rules about what evidence may be considered, requirements to give reasons, opportunities to contest, and standards of review. None of these properties belong to a model. The institution still has to justify that output the way law requires, defend it under cross-examination, and respond on appeal. When AI participates in a decision, the institution needs to ask whether it can keep doing those things well. When it cannot, accuracy alone will not save the decision.

The standard policy response treats this as a technical problem with a technical fix: explainable AI, meaning systems designed to give an account of how they reach their outputs. The European AI Act requires explainability, and multiple US bills propose disclosure of model logic. The premise is that if a system can describe what it does, accountability follows. But it does not follow necessarily. Such a description tells a court or an interested party how the system reaches its outputs, but not whether a decision that incorporates those outputs rests on grounds the law permits. That is a question courts and agencies have to answer.

Call the capacity to answer it traceability. Traceability is the ability to reconstruct the reasoning path from evidence to decision in terms that can be evaluated against legal standards. Explainability tells a reviewer which features the model weighted. Traceability requires that those features be legally permissible considerations, that the weight assigned to them be defensible, and that the affected party have had a real opportunity to challenge them. Many explainability tools produce counterfactual statements that show what feature mattered most by varying it (e.g., “if the defendant had lived in a different neighborhood, the score would have been lower”). That statement describes the model output but does not justify the decision; it does not, for example, tell stakeholders whether neighborhood is a permissible ground for sentencing. The model alone cannot answer the legal question. The institution has to.

Three things follow. First, a model’s lack of transparency does not relieve an institution of its traceability duty. Opacity makes traceability harder to satisfy, but the duty runs to the reasoning path the institution constructs around the model and not to the model’s internals: what the output represented, what weight the decision-maker gave it, how it was integrated with other evidence and applicable law, and how the affected party could contest each of those steps.

Second, traceability is what existing doctrine already demands once AI is involved. Due process requires that an affected party be able to identify and challenge the basis of a decision against them. Arbitrary-and-capricious review under administrative law requires that an agency consider the relevant factors and explain how they connect to the choice it made. Equality doctrines require that decisions not rest on impermissible grounds. None of these doctrines is satisfied by an explanation of how a model works. All of them require that the institution show why the resulting decision rests on grounds the law permits.

Third, this reframes what deploying an AI system commits an institution to, whether the system is built in-house or procured from a vendor. The duty to justify a decision is the institution’s regardless of what tools it uses to aid in the decision. Deploying AI does not move that duty to the model, the vendor, or the engineer. The procurement contract or internal documentation must let the institution obtain the information it needs to justify its reliance on the system and the decisions that follow. The Dutch tax authority deployed a model whose internal logic it could not interrogate, even at the level of its own civil servants. The moment a penalized family asked why, the institution had no answer.

One reply to all of this is that better accuracy and explainability will close the gap. They will not because the gap is not technical. A perfectly accurate model still tells a court only what the case is statistically, not what the law permits an institution to do with that information. Counterfactual explanations and feature attributions describe the model with more precision, but they still cannot tell a court whether a decision the institution reached on the basis of the model rests on grounds the law permits. Better model accuracy and technical description, while desirable, do not answer questions of law.

When an institution cannot account for its decision in terms the law can evaluate, it has tried to outsource its authority to the AI system it relied on. That is not authority a court or an agency has to give. In Toeslagenaffaire, when the institution was finally asked to defend its decisions in legal terms, it could not. Whether the model was accurate or sophisticated was beside the point. The same standard applies wherever AI shapes decisions about rights, from risk scores in bail hearings to generative outputs in administrative decisions.

Ignacio Cofone is Professor of Law & Regulation of AI, University of Oxford. You can reach him by e-mail at ignacio.cofone@law.ox.ac.uk. 

Balkinization Symposium on the Global Political Economy of Artificial Intelligence

JB

Trump, vaccination, and the right to kill children

Andrew Koppelman

Does liberty mean a right to kill other people’s children, slowly and painfully?  It would, if the Trump Administration had its way.  It’s pressing states to allow broad vaccine exemptions as part of its larger campaign against settled medical practice.

The danger is clearest with measles. Children usually recover from it, albeit often with some permanent damage. But a 2016 California study found that when infants under a year old are infected, subacute sclerosing panencephalitis (SSPE) – in which the virus lingers in the brain and slowly destroys it – develops in about one in 600 cases, far higher than the 1-in-100,000 figure doctors long cited. Long after the initial infection, the child becomes clumsy and forgetful, then experiences headaches, fevers, and seizures, loses motor control, and finally becomes bedridden and dies. There is no cure.

Babies can't be vaccinated on the standard schedule until 12 to 15 months old — earlier doses, at six months, are only given for travel to outbreak areas and don't count toward full protection. (Under the schedule revisions the Department of Health and Human Services just proposed, some of that timeline would push even later.) Until a child is protected, the only way to guard against SSPE is to keep them away from unvaccinated people. In Germany in 2000, an unvaccinated 11-year-old boy was taken to the pediatrician with a fever. He infected six children in the waiting room, including three babies. Two of them developed SSPE.  Natalie died in 2011 and Micha in 2013.

Measles is astoundingly contagious. You can catch it by entering a room two hours after an infected person left. The country has recorded more cases since early 2025 than in the previous 25 years combined, according to Centers for Disease Control and Prevention data. More than 70 percent of confirmed cases are children. The administration's anti-vaccine rhetoric is a principal driver, and if it gets its way, the unvaccinated share of the population will keep rising. Scale up that one-in-600 figure, and that's a lot of dead children.

I elaborate in a new column at USA Today, here.

26th Amendment Circuit Split

Gerard N. Magliocca

The Fourth Circuit today held that a South Carolina law restricting "no-excuse" absentee mail-in ballots to voters 65 and older violates the 26th Amendment. This decision stands in contrast to other circuits that upheld related laws during the pandemic. The Supreme Court has never heard a 26th Amendment case. This could be the Term where that silence ends.

Technology and Constitutional Rot

JB

I have posted a draft of my latest article, Technology and Constitutional Rot, on SSRN. Here is the abstract:

Technological change affects the Constitution not only by creating new problems of constitutional interpretation, but also by changing the distribution of economic and political power. New technologies alter what governments and private actors can know, control, predict, and accomplish. When constitutional institutions fail to adapt, large shifts in power and control can lead to constitutional rot: the process by which a democratic republic becomes less democratic and less republican over time. The United States has suffered from increasing constitutional rot over the past four decades; rapid technological change has made the rot worse.

Our Second Gilded Age has many similarities to the technological and economic upheavals of the First Gilded Age, which was also a period of constitutional decay. We now live in the Algorithmic Society, where public and private decision making increasingly depend on algorithms, artificial intelligence, platforms, large-scale data collection and digital surveillance. The Algorithmic Society produces an Algorithmic State, including a National Surveillance State organized around prediction and prevention. These changes disproportionately strengthen the executive branch, weaken Congress, limit practical judicial oversight, and increase government’s dependence on powerful private firms and privately-owned technological infrastructure.

Large global technology companies increasingly exercise governing power over communications, commerce, national security, access to information, and public debate. As a result, government and technology firms engage in repeated attempts at mutual co-optation, blurring traditional distinctions between public and private power. Digital platforms reshape the public sphere through an attention economy that weakens knowledge, intensifies political polarization, and erodes trust. These changes exacerbate four central features of constitutional rot: growing wealth inequality, political polarization, loss of trust, and policy disasters.

Despite these dire circumstances, the United States has recovered from periods of constitutional rot before. The First Gilded Age gave way to the reforms of the Progressive Era and the New Deal. Today constitutional renewal involves a different set of reforms: new digital privacy laws and new antitrust, civil-rights, and due-process protections. Renewal means reconstructing and improving the knowledge institutions that underwrite our democracy. It requires structural changes that better allow Congress to check and oversee the executive branch. And it requires reforms that secure and strengthen our system of democratic representation.

Analyzing War Powers: Ways Forward (Part 3)

Stephen Griffin

 Despite the lack of constitutional amendments with respect to war powers, there are two inescapable legal changes that must be kept in mind when analyzing questions of war powers in the present.  Along the pathway pioneered by Bruce Ackerman, one might call them “amendment-level.”

The first is that the United States has global responsibilities (along with the ability to project power) it did not have in the eighteenth and nineteenth centuries, not only because the foreign policy landscape is different but because of duly ratified treaties and other agreements that are their de facto equivalent.  The second is that the War Powers Resolution (WPR) is, as a prominent scholar recently reminded us,  “binding law,” and, by its terms and purpose, remains a cutoff point for the legal relevance of historical executive practice.  In other words, although this has never been acknowledged by executive branch lawyers, it reset the table with respect not only to executive-legislative relations but with what counts as a sound argument.

There’s also a logical point that should be borne in mind – that the oft-cited distinction between offensive and defensive wars never made any sense, either in the eighteenth century or in our own.  As becomes evident if we treat war powers in the context of foreign affairs (see Part 1 of these posts), war is always a political and policy choice.  As such, it is governed by the same set of provisions in the Constitution irrespective of whether it can be classified as offensive or defensive.  Furthermore, war is a choice even if we are “suddenly attacked.”  Believe it or not, American history has featured a number of instances where U.S. military forces were egregiously attacked, yet no one jumped to the conclusion (as Hamilton once mistakenly argued) that we were automatically at war.  On top of this, it appears that the legal effect of the U.N. Charter (one of those treaties mentioned above) was to rule out recourse to offensive war as an instrument of foreign policy and also, by the way, declarations of war.  But that just moves us back to where we should be, which is whether Congress authorized the war in question.  As reviewed in post 2, where the past used declarations of war, we use AUMFs.  No harm, no foul.

Milton, Mill, and (Maybe) Aristotle on Why Hearing Horrible Speech Makes You a Better Person

Andrew Koppelman

What’s Wrong with the War Powers Debate (Part 2)

Stephen Griffin

It has been difficult for participants in the war powers debate to describe accurately the legal status quo.  Is the WPR’s constitutionality in dispute?  If it is constitutional, does it play a meaningful role?  Are AUMFs the constitutional equivalent of declarations of war?  What does the Constitution require in the context of the national security state built after WWII?  What is the legal position of the executive branch as represented in OLC opinions?  And so on.

To make progress in this fraught area, we need to think differently about the way the Constitution is enforced.  It is not only enforced through the judiciary.  The other institutions responsible for enforcing the Constitution, including political parties, do not enforce it all in the same way.  The political branches deserve that name because they do not necessarily follow a juridical model.  For example, they do not necessarily rely on or pay much attention to common law reasoning based on exemplar cases or historical practice.  So while the Constitution is enforced politically as well as legally, they do not look the same.

In this light, consider a recent attempt to make sense of the war powers debate by Jeh Johnson and Ethan Klaris, published in a 2026 issue of the Journal of National Security Law & Policy.  This article attracted attention because Johnson has practical experience.  He was General Counsel of the Department of Defense and later Secretary of Homeland Security in the Obama administration.  Johnson and Klaris interviewed several key members of Congress with respect to their views on war powers, albeit mostly Democrats.  Their subtitle states their fundamental theme: “Executive Expansion and Congressional Retreat.”  This is supposed to describe the course of war powers throughout American history.  I contend this is misleading as history, but it is actually difficult to find a steady throughline to contest their argument.  This is because they fail to set forth criteria that would enable us to determine whether, say, Congress was retreating or advancing.

This failure is part of a larger problem with the war powers debate.  Making distinctions about how and why Congress acts is long overdue.  Besides the distinction between juridical and political enforcement of the Constitution, we need also to think about the purpose of the “declare war” clause and the WPR.  Is it to make war constitutionally legitimate or, in the end, to slow down or stop contemplated military operations?  Some commentators appear to assume that if there were full compliance with the Constitution and the  WPR, this would sidetrack or terminate most proposed military adventures.  But that is a non sequitur, a literal “does not follow.”  The legal provisions respecting war are about legitimacy in the first instance, not about putting national security policy on a particular pathway. 

Similarly, consistent with the oft-repeated dictum that Congress is a “they,” not an “it,” some members of Congress, at least beginning in the Cold War and after, have always supported vigorous presidential leadership in foreign affairs, including with respect to decisions for war.  In contemporary times, some adhere to an “ideology of threat,” namely that the world is a dangerous place and must be periodically curbed through the application of overwhelming military force as determined by the president.  The late Senator Lindsey Graham was an excellent example.  Further, despite the genuine bipartisan support for the WPR, in the wake of Vietnam and with the return of Republican presidents, the parties developed an asymmetry with respect to war powers, especially in the Senate.  Democrats remained haunted by Vietnam for years, while Republicans shook the trauma off, figuring (like Nixon) that they were not ultimately responsible.

The Iran War and the War Powers Debate (Part 1)

Stephen Griffin

President Trump’s February decision to go to war with Iran has produced a negative reaction in Congress and inspired another round in the war powers debate among pundits and legal commentators.  Once again, probably more heat has been generated than light.  In particular, there is so much emphasis on whether and to what extent Congress has a role under the Constitution that the issue of why Congress has a role gets short shrift.  It is as if the commentariat is looking for Congress to simply punch a ticket rather than make a substantive decision.  One has the sense that the whole debate is occurring through the wrong end of a very long telescope.

In these posts I will apply the position I develop in my 2013 book Long Wars and the Constitution to the case of the Iran War.  I start not with any specific clause of the Constitution, but its design with respect to the conduct of diplomacy and foreign affairs.  In contrast to war initiation, when this topic arises in litigation the executive branch has plenty of precedents to draw on in support of its exercise of legal authority.  Yet it is also clear from both the text of the Constitution and historical experience that there is no such thing as conducting the nation’s business overseas without Congress.  In this respect, the historical record of joint branch involvement in foreign affairs resembles the role of Congress in regulating the president’s commander-in-chief power, including in wartime, described in a set of articles by David Barron and Marty Lederman.  In other words, a situation of shared rather than exclusive powers.

This point about the shared conduct of foreign affairs goes deeper than to say Congress is involved by constitutional design.  It is that there is no proper way of describing, understanding, or evaluating specific presidential war powers decisions and actions without considering how they figure in the ineluctably connected tasks of advancing the foreign policy and protecting the national security interests of the United States.  These responsibilities are exclusively national and certainly presidents have usually been at the forefront.  Further, while the presidential “first-mover” status that many decry is partly the result of presidential initiative, it is not best characterized by the archaic term “usurpation” as it is mainly a product of congressional delegation, whether by law or through the party-political process.  Either way, there is no inherent violation of the constitutional plan.

What difference does this approach make when questions of presidential war powers do arise?  Consider first the commonly asked question of whether we are at “war.”  “War” is arguably a vague term, with some clear cases and others on the borderline.  The concentration by pundits and members of Congress on defining “war” nonetheless permitted the Trump administration to claim briefly that the Iran War was over(!) before the sixty day mark established by the War Powers Resolution (WPR).  My approach is different.  Reduced to a bare minimum, it says: stop thinking about “war.”  At least, stop thinking about war in terms of number of ships, troops, planes and what they do or can do.

Instead think about the Trump administration’s goals and strategy with respect to the conduct of foreign affairs with Iran.  If, just for example, the goal is “regime change” and reasonable evaluations of what it would take to achieve this in Iran involve military actions or the threat of war, then we are in the kind of “declare war” and “necessary and proper” territory described in the Constitution that was the basis for the WPR.  As we are in the realm of shared power over foreign affairs and national security, Congress is required to have a role in making the relevant decisions.  What counts is the overall objectives and strategy rather than parsing exactly how many aircraft carriers, planes, and troops are needed to achieve them. 

The Groundhog Day Tariffs and the Problem of Executive Discretion

David Super

     Amidst a flurry of wildly erroneous charts and figures President Trump announced his “Liberation Day” tariffs on April 2, 2025.  These included the boldest action any U.S. President has ever taken against the Heard and McDonald Islands.  After almost a year, the Supreme Court struck them down as unfounded in law.  President Trump’s immediate reaction was that he would reimpose broad tariffs on some other legal basis.  He immediately ordered sweeping tariffs under a legal authority valid for only 150 days as he declared that his Administration would spend that time figuring out how to make the tariffs permanent. 

     As promised, on July 24, 2026, President Trump announced a new round of tariffs against sixty countries, including almost all our significant trading partners.  (He does, however, appear to have gone soft on the Heard and McDonald Islands.)  He bases this new round on section 301 of the Trade Act, which authorizes tariffs and other measures when “an act, policy, or practice of a foreign country…is unjustifiable and burdens or restricts United States commerce”.  President Trump asserts that these countries gain an unfair advantage in trade with the United States because they do not exclude production inputs that come from forced labor, at least not to the extent that we do.

     The Administration’s announcement gave a whole new meaning to “threadbare”.  It failed to show that all these countries benefit from forced labor at all, much less that the products of forced labor give them an advantage equal to 10% or 12.5% of their trade with the U.S.  The Administration also neglected to explain what its reaction would be if our trading partners were to impose similar or greater tariffs on us because of our “unjustifiable” acts, policies, or practices of gutting organized labor or blocking decarbonization, both of which surely give us a far more dramatic cost advantage than any benefit derived from forced labor in third countries.  Unsurprisingly, these tariffs have now been challenged in court.   

     Internationally, this will do further damage to our reputation as a reliable trading partner.  Economically, it will further tangle international trade and likely raise costs here and abroad.  Legally, these endlessly repeating tariffs raise fundamental questions that go far beyond Trade Law. 

     In a complex modern economy, numerous important decisions cannot be specified by Congress in advance.  Yet Members of Congress of both parties will hesitate to delegate discretion to the executive if they expect it will often be exercised in bad faith to reverse their policy choices our to retaliate for personal grievances.  The solution to this dilemma is central to any plan to reconstruct Administrative Law after President Trump leaves office. 

     Rebuilding Administrative Law is much too ambitious a topic for a post here.  (I have a full article on that in progress.)  But President Trump’s Groundhog Day tariffs do suggest a few possible responses to the problem of wildly abusive exercises of discretion. 

     First, over the years Congress has grossly overestimated how much administrative flexibility is actually needed.  Putting a large thumb on the scales for flexibility likely seemed easy in an era when norms of good faith were widely and when a more robust set of checks and balances created a need to avoid Congress’s wrath.  As I have argued elsewhere, flexibility is not the unalloyed good it is often portrayed to be.  Instead, it has costs and benefits that should be weighed thoughtfully.

     Second, in cases such as trade policy where Congress genuinely cannot anticipate serious problems that may arise, a legislative path is still possible.  Over the years, we have established numerous mechanisms under which the President may submit specific types of requests to Congress and is assured of an expedited decision.  Among these are reports from military base closure commissions, proposed recissions of appropriated funds, and even some kinds of trade treaties.  These rules typically provide rigid limits on the scope of presidential requests that may benefit from expedited procedures and then set out distinctive procedural rules.  These include an accelerated schedule that prevents the presidential requests from being bottled up in committee, from being sabotaged with controversial amendments, or from being filibustered in the Senate.  The request still fails if a majority in either chamber votes to reject or table it, but the special rules prevent Members from hiding behind the obscurity of congressional procedure to kill the measure unaccountably. 

     Rather than allowing the President to impose tariffs unilaterally, Congress could rewrite section 301 and similar authorities to allow submission of proposals for tariffs of a specified amount on a specified range of goods and services from one or more specified countries.  Legislation with those provisions (but no others) would then be assured of expedited consideration.  Members favoring different changes in trade policy would be free to try to move legislation through the regular order, but that would not prevent the President’s proposal from coming to a timely vote in precisely the form the President submitted.  Absurd proposals like the Groundhog Day tariffs likely would have difficulty finding champions on the House or Senate floors; serious, well-supported recommendations would be assured of floor consideration.

     Finally, some courts may feel uncomfortable declaring that the Administration’s obviously concocted factual rationale for these tariffs is, well, obviously concocted.  They should not.  Remarkably, President Trump himself has made the courts’ task here considerably easier.  The Court rejected the first Trump Administration’s effort to put a question seeking citizenship information on the decennial Census after finding that its stated reason for doing so was obviously insincere.  “Our review is deferential, but we are not required to exhibit a naiveté from which ordinary citizens are free.  The reasoned explanation requirement of administrative law, after all, is meant to ensure that agencies offer genuine justifications for important decisions, reasons that can be scrutinized by courts and the interested public.  Accepting contrived reasons would defeat the purpose of the enterprise.  If judicial review is to be more than an empty ritual, it must demand something better than the explanation offered for the action taken in this case.”

     In addition, the Unified Executive Theory enables courts to weigh the totality of the Administration’s behavior when assessing its credibility.  Previously, a tendentious or even dishonest action by one agency could be attributed to an isolated bad bureaucrat and deemed largely irrelevant to what other agencies are doing.  Now, the Court has determined that the President’s broad, unfettered removal power is essential to making subordinate officials “remain accountable to the President, and the President to the people.” 

     Since the Twenty-Second Amendment limited presidents to two terms, we have been led 37% of the time by presidents who were never again going to be accountable to the people.  (President Trump has been immune from any further accountability to voters since he resumed office.)  To reap the intended benefits from concentrating accountability in the President, some of the President’s accountability must be to the other two branches of the federal government.  Now that the Unified Executive Theory empowers the President to impose strict, uniform standards of conduct, Congress and the courts can and should assume that the actions of any Executive Branch official reflect on the entire branch, at least unless prompt disciplinary measures are imposed. 

     In the context of the Unified Executive Theory, an administration’s propensity to falsify or mislead may properly be taken into account in disbelieving subsequent assertions.  This should be all the more true for facially improbable claims such as those supporting the Groundhog Day tariffs.  And this should be particularly true on issues, such as tariffs, where the Administration has made clear that it is determined to impose its will despite opposition from the coordinate branch with primary constitutional responsibility. 

     Our legal system places such a high premium on assessing character for truthfulness that it forces criminal defendants to forego taking the stand in their own defense if they wish to avoid having their prior unrelated, misconduct introduced in evidence against them.  Our law believes it is so important that the trier of fact be aware of prior bad acts possibly showing a character prone to dishonesty that we are willing to accept the risk that the jury will conclude that the defendant committed the acts charged on the basis of otherwise impermissible character reasoning.  If character for truthfulness is so important for a private individual seeking to stave off long incarceration or even death, surely it matters even more for a President facing only a policy loss to other elected representatives of the people and now equipped to enforce the highest standards of probity within his administration. 

     @DavidASuper.bsky.social @DavidASuper1

Presidential Accountability and the Twenty-Second Amendment: A Little Math

David Super

     Note:  Having just spent the morning writing a short essay about the Twenty-Second Amendment, I was dismayed to find that Jeffrey Gordon had just posted a much more thoughtful piece on the same subject.  Accordingly, I have cut away the parts of mine that seem duplicative of his.  But I did approach the problem quantitatively and so am posting a stripped down version to make that analysis available. 

*     *     *

     In Slaughter v. FTC, the Supreme Court relied on extensive Originalist arguments to strike down most congressional limitations on the President’s power to remove executive officials.  The Court declared that “Subordinates who exercise the President’s power are subject to removal by him. Then, and only then, can they remain accountable to the President, and the President to the people.”  In Seila Law LLC v. Consumer Financial Protection Bureau, the Court declared that the growth of federal powers “only sharpens our duty to ensure that the Executive Branch is overseen by a President accountable to the people.”  On numerous other occasions, the Court has insisted on deferring to the President specifically because he is politically accountable.  But is he?

     That was a reasonable, if imperfect, approximation of the President’s position for much of this country’s history.  Between the Civil War and the end of Harry S. Truman’s presidency, the current president had at least some plausible chance of facing the voters again ninety percent of the time.  In other words, only about ten percent of the time were we governed by a president who had announced that they would not run again or who had been defeated for renomination or re-election.  Just once during this period did a president serve for more than a year after having been eliminated from further electoral accountability.  (In this calculation, I am ignoring pledges not to run again made prior to the president’s initial election as those pledges were sometimes reconsidered.) 

     The Twenty-Second Amendment, limiting presidents to two terms, changed all that.  Beginning with Dwight D. Eisenhower, the first president subject to the Amendment, and running through the expected end of President Trump’s second term, the sitting president has been a lame duck 37% of the time.  Presidents Eisenhower, Nixon, Reagan, Clinton, George W. Bush, and Obama became constitutionally incapable of accountability to the voters the moment they were re-elected, slightly less than half-way through their presidencies.  President Trump resumed office in 2025 as a lame duck. 

     Whatever one thinks about Originalism in other settings, it is clearly inappropriate where the underlying document has changed in a clearly relevant way.  Little practical insight can be had by considering the original public meaning of the Eighteenth Amendment (Prohibition) or the federal courts’ jurisdiction over cases between “between a State and Citizens of another State” because those provisions have been overridden.  So, too, doctrines of sweeping deference to the President based on the Framers’ confidence in his electoral accountability should be revised to account for the more than one-third of our civic life – including the next two-and-a-half years – when that is no longer true. 

     @DavidASuper.bsky.social @DavidASuper1

Trump v. Slaughter and the Twenty-Second Amendment

Guest Blogger

Jeffrey N. Gordon 

            Supreme Court decisions of the recent term threaten to undermine the regulatory system that has supported US economic growth and prosperity for nearly 100 years.  Many have objected to the Court’s failure to give weight to the risk of such costly destabilization.  The Court’s response, presumably, is “it’s the Constitution, not us.” The flaw is that the Court has ignored a critical part of the Constitution, the Twenty-Second Amendment, which enacts a 20^th century model of presidential empowerment that calls for sustaining the “for cause” removal protection at issue in in Trump v. Slaughter and Trump vs. Cook.

            In a series of cases that culminated in this term’s decisions, the Supreme Court has transposed the Framers’ desire in 1789 to create an “energetic” president into a decision rule that Congressional fashioning of regulatory agencies headed by commissioners with “for cause” protection was an impermissible encroachment on “executive power.”   In a prior decision (2020) that overturned the structure of the Consumer Financial Protection Bureau, Seila Law v. CFPB, the Chief Justice explained the resulting constitutional structure this way: “divide power everywhere except for the Presidency, and render the President directly accountable to the people through regular elections.”

            This intellectual scaffolding, built around the unique electoral accountability of the president, collapses in the face of the Twenty Second Amendment, approved in 1947, ratified in 1951, which not only imposes term limits but also embraces prior efforts to limit presidential power.    

            The most immediate point of failure is the “lame duck” problem.  After adoption of the Twenty-Second Amendment, electoral accountability cannot, as a mechanical matter, restrain a second term president. A direct consequence of the Twenty-Second Amendment is to change the Constitutional framework.

            But the Twenty-Second Amendment’s rebuke of the current Court goes deeper.   The Amendment was enacted because the Framers of 1947 explicitly rejected  electoral accountability as a sufficient check on presidential power. Working in the decade following the creation of the New Deal, these Framers saw a president who had become vastly powerful.  They believed that the president could use these powers for patronage and favor-granting to engineer renomination and pave the way to his reelection.  Thus the sharp limit on re-election came from  their conviction that “electoral accountability” was an insufficient check.   Presidential authority needed to be subject to a hard stop even though depriving the People of the right to choose to re-elect a president repeatedly counted against democracy.

            The Framers of the Twenty-Second Amendment thus implicitly and explicitly embraced the limitations on presidential power in the set-up of the “independent” agencies that proliferated in the New Deal.  In the living memory of these Framers was the struggle over Congressional effort to provide “for cause” protection for agency commissioners, validated in Humphrey’s Executor v. US, in May 1935.  This case was understood to be pivotal, since the Senate’s insistence on such protection had been a major barrier to legislative conferral of extensive new powers on the Federal Reserve Board.  The Banking Act of 1935 was passed in August, three months later.

            As put by one of the principal architects of the Twenty-Second Amendment, Rep. Earl C. Michener, chairman of the House Judiciary Committee, in the speech that introduced the operative resolution in the House: “Over a period of 12 years a President can secure a firm control over the enormous and powerful machinery of Federal administration, including the so-called independent agencies.”

            In the Supreme Court’s Slaughter world this argument makes no sense, for the president can immediately remove (or credibly threaten to remove) the commissioners of “independent” agencies and thus gain immediate control.  The president need not wait out the commissioners’ terms; the accession to control is immediate upon taking office. 

            The Twenty-Second Amendment embraces a different conception: that presidential power is in part a function of tenure and that by limiting tenure, we constrain presidential power.   The goal of the Twenty Second Amendment is to limit presidential power even as the scope of government expanded.  Its immediate mechanism is term limits but those term limits are meant to gain force through interaction with pre-existing constraints on presidential power, including the protection of “for cause” removal for the commissioners of independent agencies.  Thus a set-up with lengthy staggered terms for commissioners with “for removal” protection is woven into the Twenty Second Amendment’s conception of legitimate limits on presidential power.       

            In light of the Twenty-Second Amendment, there are two distinct routes by which the Court should redress its constitutional mistake.  The first is to accept the rewriting of presidential power embodied in the Amendment, which directly countenances limitations like staggered terms for commissioners protected  by “for cause” removal. 

            The second is simply to acknowledge and address the accountability gap that is an immediate mechanical consequence of the Amendment, in a manner consistent with the goal of ordered liberty.  In other domains, this so-called “final period problem” leads to special governance rights.  For example, in a merger transaction, target shareholders always get a vote because officers and directors will be “term-limited” if the transaction goes forward.  In the setting before the Court, “for cause” protection of agency commissioner seems an appropriate way to protect ordered liberty in light of the change in the Constitutional scheme.  Even though the mechanical electoral accountability problem does not arise in a president’s first term, there needs to be consistent rule, here, “for cause” protection.

            Slaughter relies on a vision of the presidency conjured, speculatively, from 1789. As  enactment of the Twenty-Second Amendment shows, this is not “the end of history” for the Constitutional conception of presidential empowerment. We have a 20^th century conception created in light of a dramatically changed government, not inferred through Constitutional “moment” but enacted by Constitutional amendment.  And this revised conception of presidential empowerment and its limits should bind the US Supreme Court. 

Jeffrey N. Gordon is Richard Paul Richman Professor of Law at Columbia Law School. You can reach him by e-mail at jgordon@law.columbia.edu.

Justified and Unjustified Loss of Trust (in institutions)

Stephen Griffin

You may have heard of the “Yale Report” – in April Yale University released a thoroughly researched response to public concern with universities.  Authored by a panel of distinguished scholars, it is titled “Report of the Yale Committee on Trust in Higher Education.”  The Committee felt there was not much trust to report.  It cited public opinion polls to the effect that public confidence in higher education had declined twenty points over the last ten years.  The Committee usefully defined the trust relationship by saying “[t]rust is earned by doing what you say you’re going to do – and, ideally, doing it well.”

The Yale Report is part of an upsurge of concern with trust.  In particular, it seems ever since the beginning of Trump’s second term, we are awash with claims of a loss of trust in institutions.  Besides the Yale Report, examples are easy to find.  An asserted loss of trust in the public health establishment because of the Covid-19 pandemic.  A loss of trust in the federal government, including the Supreme Court.  Leading media organizations like CBS news try to ride with the times by hiring new managers who assert the need to change because of a loss of trust.

These claims are not simply offered as observations about a lamentable state of affairs.  They are being used in a systematic way to build a case for the reform of government and other key institutions in American society.  Like any foray into the sphere of public argument, such assertions can be critiqued for deficiencies of logic and factual validity.  I am in the process of writing a short paper which poses a different question: is it possible to critique such assertions normatively?  Indeed, what would a normative critique look like?

Dobbs and Now B.P.J.

Guest Blogger

Reva Siegel & Mary Ziegler

Alliance Defending Freedom (ADF), the conservative Christian legal organization that litigated Dobbs, is now trying to do for the law of equal protection what it has done for the law of due process. Consider the cases it has brought to the Court in West Virginia v. B.P.J. and United States v. Skrmetti.

In B.P.J., the Court upheld, against equal-protection challenge, laws prohibiting transgender athletes from participating on girls teams. B.P.J. depicts its holding as a minor exception to the Court’s sex-discrimination jurisprudence, one that reflects the unique context of sports. But there is a problem with this account. B.P.J. represents the third minor exception that the Court has made in its equal protection sex discrimination cases in the past year. In cases involving transgender individuals—United States v. Skrmetti, Orr v. Trump, and B.P.J.—the Court keeps introducing new exceptions to intermediate scrutiny that have no obvious connection to one another.

Our forthcoming Article in the Texas Law Review identifies what unites these sex discrimination cases: they all employ the term biological sex which the Supreme Court had not employed in equal protection law before. Biological sex seemingly refers to science—yet it can also refer to nature, religion, and gender roles, as documents like the Nashville Statement of the Council of Biblical Manhood and Womanhood (2017) of the Southern Baptist Convention demonstrate. Our Article, however, focuses on the meaning of biological sex in equal protection law.

We show that biological sex is a movement signature—the fingerprint of advocates who are supplying the Supreme Court with resources for the stealth overruling of United States v. Virginia (1996), Justice Ginsburg’s opinion for the Court explaining the Constitution’s guarantees against sex discrimination (which judges apply in cases of discrimination on the basis of sexual orientation and gender identity, as well).

The movement context we bring to B.P.J. clarifies the Court’s path in the last year. Reading together the “biological sex” cases weakening Virginia—Skrmetti, Orr, and B.P.J.—we can see the Court beginning to constitutionalize the claims of a movement seeking the stealth overruling of Virginia on its thirtieth anniversary. Biological sex directs judges to defer to the judgments of the political branches rather than to scrutinize sex-based state action for sex-role stereotyping—a code associating sex difference and judicial deference that revives in contemporary idiom the very traditions of reasoning from the body that United States v. Virginia rejected. 

In B.P.J., the Court asserted its ruling was narrow—that sports is an exception that implicates physical differences in a way that employment does not. But B.P.J. also sought to normalize this exception, stating without any evidence other than the suggestion of the advocates that biological sex was the ordinary meaning of sex in Title IX and the Court’s sex-discrimination cases of the 1970s. This is wrong in several ways. Title IX prohibits “discrimination . . . on the basis of sex.” The relevant unit of meaning is not the word “sex” but the phrase “on the basis of sex.” The statute addressed beliefs about gender roles (not reproductive organs)—Congress was concerned about stereotypes leading to girl’s and women’s sports receiving virtually no funding. The term biological sex appears nowhere in the text or legislative history of Title IX (the first appearance of the term came in 2016, when conservative lawmakers proposed amending the statute to define sex as biological sex).

In deciding B.P.J., the Court ignored the text and history of Title IX and its proponents concern with sex-role stereotyping. It then employed a weakened form of intermediate scrutiny, identified as such by citations to Skrmetti and Burger Court decisions like Michael M. v. Superior Court that treat intermediate scrutiny as more akin to rational basis. Most revealingly, the Court twice quoted the very language ADF has cherrypicked from Virginia and reiterates throughout its campaign to legitimate judicial deference: that “[p]hysical differences between men and women” are “enduring.” For the Court to quote this language from Virginia is like a movement wink-and-nod, bound to energize those convinced that physical difference itself justifies the elimination of intermediate scrutiny of any kind.

Will B.P.J. be strictly limited to the sports context? It is hard to trust the Court’s assurances about B.P.J’s limits if the Court is already citing Skrmetti outside the specific factual context that decision purported to address.

The Court has not yet revised Virginia into the difference-deference-doctrine that ADF and Heritage seek, but conservatives on the Court are collaborating with movement actors to weaken constitutional guarantees of equal citizenship. It remains to be seen how far this transformation will go.

Very far is the position for which Heritage advocates in a report published immediately after B.P.J. in which Heritage attacks gender-egalitarian funding of sports as “feminist social engineering” and argues for revising Title IX to respect natural sex differences in sports, citing women’s interest in yoga and cheerleading rather than competitive sports.

Reva Siegel is Nicholas deB. Katzenbach Professor, Yale Law School. You can reach her by e-mail at reva.siegel@yale.edu.  Mary Ziegler is Martin Luther King Jr. Professor of Law at the University of California at Davis. You can reach her by e-mail at mziegler@ucdavis.edu.

From Off the Wall to On the Wall: How Social Power Shapes Legal Reason

JB

I've posted my latest article, From Off the Wall to On the Wall: How Social Power Shapes Legal Reason, to SSRN. Here is the abstract:

Throughout the history of the Constitution legal arguments have moved from being "off the wall"-- crazy, incompetent, and beyond the boundaries of professional reason, to "on the wall"-- plausible claims about which reasonable minds can differ. Some of these claims have eventually become part of the law, and a few have become orthodox, so that denying them would be off the wall.

This article explains how and why this happens. Arguments move along a spectrum of plausibility depending on who occupies important nodes of power and influence in society. These include political leaders, political parties, social movements, the media, elite lawyers, and legal intellectuals. When powerful actors invest in a fringe claim, they compel the legal profession to debate it. Arguments move on the wall not by general acceptance but by compelled engagement.

Several examples show how this dynamic works in practice. Arguments for same sex marriage only slowly gained traction because gays and lesbians had little social status and had to build support over decades. The constitutional challenge to Obamacare's individual mandate moved on the wall in only a few years because the Republican Party and conservative media were unified behind it. The equal protection argument in Bush v. Gore took only a month because control of the government was at stake. The Trump administration's recent challenge to birthright citizenship had the support of a sitting president, the MAGA social movement, the Republican Party, and the party's affiliated media and institutions. Conservative legal intellectuals quickly devised arguments designed to appeal to a Supreme Court six of whose members were conservatives appointed by Republican presidents.

Underlying this phenomenon is the dual nature of legal consciousness, which combines acceptance of shared professional norms with constitutional protestantism, the belief that individuals can interpret the Constitution for themselves. The phenomenon also exemplifies law's dual character as fixed in the short run and mutable in the long run. Legal education cultivates this duality, training lawyers both to treat existing law as authoritative and to creatively deconstruct it.

The theory recasts jurisprudential debates over legal indeterminacy from the nature of language to sociology and culture. Legal materials are flexible but not infinitely malleable; they are tilted towards entrenched interests until sustained social power changes them. Following these changes, constitutional culture frequently rewrites its own history. Formerly off-the-wall claims that win out are later heralded as always having been correct, while formerly accepted decisions, such as Plessy v. Ferguson, are denounced as “wrong the day they were decided.” This triumphalist narrative obscures the intense social and political labor required to shift the boundaries of professional reason.

Fiscal Floundering

David Super

     Our nation continues its transition from congressional to presidential fiscal control.  Retainers of the ancièn regime (“Members of Congress” to you) fumble around looking for something to do as the Administration becomes increasingly assertive.  No single recent event moved me to write, but putting several pieces together reveals a picture that is worth noting.

     The congressional appropriations process is rapidly falling apart.  Lacking the filibuster, House minorities have virtually no independent power when the majority is united.  The House Republican Caucus is vehemently opposed to negotiating with Democrats so House Republican appropriators write extreme bills that can keep their Members on board but have no chance of getting the Senate Democratic votes needed to overcome the filibuster.  This makes House appropriators of both parties increasingly ineffective. 

     For the past several years, then, appropriations bills were effectively negotiated in the Senate.  This may be much more difficult this year because the Senate Appropriations Chair is one of the most endangered Republicans and can neither afford to alienate swing voters nor depress turnout in the MAGA base. 

     President Trump also threw a monkey wrench into the process of writing appropriations bills with his demand for a 44% increase in war spending in a single year.  This half-trillion-dollar increase is far more than the total combined spending of welfare, Supplemental Security Income, nutrition assistance, foster care, child care subsidies, the Earned Income Credit, the Child Tax Credit, and unemployment compensation for next year. 

     To accommodate anything resembling this spending surge, appropriators must either further increase the deficit or wreck the rest of the budget.  One might think that House Republicans would opt for deficit spending, with almost all having thrown fiscal caution to the wind in enacting last summer’s upper-income tax cut legislation.  For the moment, at least, an assortment of fair-weather fiscal hawks are insisting that the defense increase be offset with cuts in domestic spending.  (Reportedly some have already told Republican leaders what favors for their districts would suffice to induce another round of fiscal amnesia.)  But any appropriations bills that transferred anything remotely resembling this amount of money from domestic to defense spending could never get the Democratic votes necessary to surmount a Senate filibuster.  That leaves both House and Senate Republican appropriators stuck.  With many people suspecting that President Trump will veto any appropriations bill that does not contain the voter-suppressing SAVE Act, few have much incentive to take bold, risky decisions to move appropriations. 

     Speaker Mike Johnson has tried to break this impasse by proposing that they move a large part of the defense increase through special majoritarian “reconciliation” procedures.  This is getting to be quite a habit.  Republicans invoked reconciliation to pass their upper-income tax cuts last summer without Democratic votes and then did so again this spring to provide opulent funding with no accountability to the Department of Homeland Security and its immigration enforcement agencies.  Why not go for a hat trick?

     Procedurally, Republicans can indeed return to that well a third time in one Congress.  (The Congressional Budget Act limits how often the majority may invoke reconciliation, but they can make this one work.)  Politically, however, forcing Republicans to cast vote after vote to fund a wildly unpopular, failed war is not very attractive.  Moreover, reconciliation rules give Senate Democrats two separate opportunities to force numerous votes on politically uncomfortable amendments.  Practically, reconciliation consumes a lot of floor time, which is scarce with nervous Members eager to go home to shore up their re-election prospects.  And institutionally, the habit of addressing more and more difficult spending questions through hyperpartisan reconciliation sends the appropriators, particularly the genuinely bipartisan Senate Appropriations Committee, steadily toward irrelevance. 

     House Republicans have been looking desperately for offsets to make a military reconciliation bill look less like the budget-buster that it would be.  Reportedly they tried claiming that an increase in state anti-fraud administrative money would yield savings from fewer fraudulent payments in anti-poverty programs; following longstanding scorekeeping principles, the Congressional Budget Office refused to attribute any savings to this scheme as these programs already have extremely low fraud rates.  At present, Republicans lack the votes even to start the reconciliation process in either chamber.  

     Unable to move either regular appropriations bills or a partial substitute via reconciliation, attention is turning to the “continuing resolution” that will have to be enacted to prevent a partial government shutdown on October 1.  These bills typically freeze funding at the prior year’s nominal-dollar level, which stresses agencies due to inflation.  In crafting continuing resolutions, the parties typically negotiate a modest number of “anomalies”, deviations from the freeze level to reflect unusual funding needs (such as major disasters or the decennial Census). 

     The Administration just submitted its list of proposed anomalies to Congress.  It is instructive.  Not surprisingly, it is trying to smuggle in as much of its military spending request as possible into the anomalies process.  It also seeks retroactive congressional approval of impoundments of domestic programs’ funds that it has already implemented.  It is urgently seeking a $30 million increase in funding for the Department of Justice to pursue “fraud”; since returning to office, the Trump Administration has used vague, unfounded accusations of fraud to justify harassing and defunding blue states that it particularly detests (while ignoring far more serious integrity problems in red states).  The Administration’s request accelerates the availability of disaster response funds, reportedly at the insistence of a North Carolina Republican threatening to withhold his vote from fiscal legislation.  And it seeks an additional $113 million for the National Park System to cover the costs of the President’s 250^th anniversary celebration.  I certainly hope nobody tells the waste-fraud-and-abuse people what the attendance at that celebration was.  (I would not be surprised if there was an Arc d’Trump hiding in that budget request as well.)

     Implicitly signaling that bipartisan agreements on appropriations will be increasingly rare, some senators in both parties are working on legislation to regularize impasses.  Proposals’ details vary, but most would automatically activate a continuing resolution with less money than programs need to continue providing current services.  Making this the default will give those seeking budget cuts even less reason to negotiate appropriations bills and forcing programs’ proponents to make concessions to avoid automatic cuts.  This will set domestic programs’ funding on course to ratchet down over time.  Ending individually negotiated continuing resolutions also will eliminate one of the few vehicles still available for restraining the Administration’s worst excesses, such as its forthcoming rule politicizing the federal grants process.

     Government shutdowns are costly and messy, but they are also our primary means of getting the electorate to adjudicate the parties’ genuine disagreements.  Just as the disappearance of trials deprives civil litigants of benchmarks for valuing their cases, so, too, the absence of verdicts from the electorate on government shutdown would leave the parties with little guidance on how they should limit their demands.  This would make partisan divisions even sharper and more intractable. 

     To end on a positive note, I would encourage those interested in an example of Congress acting fairly forcefully to restrain the Administration’s impoundments of appropriated funds to read my recent op-ed with Cerin Lindgrensavage.

     @DavidASuper.bsky.social @DavidASuper1