I’m not sure what the editors of the Harvard Law Review thought I would write when they invited me to author the comment about Trump v. Slaughter for the annual Supreme Court issue. Nor did I know what I would write. As Felix Frankfurter and Henry Hart put it when the Harvard Law Review asked them — 91 years ago — to write about Humphrey’s Executor v. United States, the scope of the President’s removal power is “a debate which had been waged intermittently since early in the first session of the First Congress and to which had contributed virtually all the great figures of American political history.” If that weren’t intimidating enough, the academy in recent years has exploded with scholarship debating this constitutional question, especially as it became more and more clear with each new case that Humphrey’s Executor was not long for this world.

With many sincere thanks to a lot of whip-smart scholars who talked through the issue with me, Bethany Renee Murray for excellent research, and the team of world-class editors at the Harvard Law Review, I decided to focus my comment on a question that hasn’t been debated for nearly 240 years: “severability stacking.”

In separation-of-powers cases, the Supreme Court’s severability analysis essentially asks whether Congress would’ve wanted no agency at all or an agency without the unconstitutional restriction on the President’s statutory power. Because the answer almost always is that Congress would have preferred the agency, the Court severs the restriction, but does not eliminate the agency or invalidate any delegated authority.

My sense is that the Court is probably right that if Congress were given the binary question of “no agency at all” (or no delegated authority at all) or “agency without provision X,” Congress would choose to have the agency without provision X.  But what if a bunch of statutory restrictions are severed? For example, if Congress had to choose between an agency subject to unlimited presidential removal or no agency, Congress probably would choose to have an agency.  But what about an agency subject to unlimited presidential removal and no legislative veto and centralized White House control of agency discretion and a less independent civil service (assuming Slaughter has implications for at least some senior civil servants) and presidential immunity for official acts and so on? 

The answer from Congress to each question individually probably is “we would keep the agency.”  But what if Congress were presented with all the questions at once?  In other words, if Congress knew from the get-go what would happen to the administrative state, would Congress really have created the agencies as they are?  At some point when enough severability questions are stacked, doesn’t the likely answer become “no”? But if we take the logic of modern severability seriously and think through what Congress likely would have wanted, what does that mean for the administrative state?

There are at least four options. The Court could (1) retreat from Slaughter, which isn’t going to happen and shouldn’t (though Slaughter should be kept within its logical bounds); (2) do nothing (which would result in a very statutorily powerful presidency); (3) address severability stacking directly (something almost certainly beyond the capacity of judges and likely doctrinally wrong anyway); or (4) enforce other restraints on the presidency that partially offset the post-Slaughter statutory windfall. Which path to take, it seems to me, should be the big question going forward.

Here is a link to The American Presidency Post-Slaughter. And here is the abstract:

In the wake of Trump v. Slaughter, the American presidency finds itself much more constitutionally powerful. Since 1935, Humphrey’s Executor v. United States has anchored agency independence by allowing Congress to prevent the President from firing agency officials based on policy disagreement. After Slaughter, that anchor is gone. Slaughter, moreover, may portend even greater change. 

Slaughter, however, also raises an important statutory question. The Court decided Humphrey’s Executor more than 90 years ago. Would Congress have delegated so much policymaking authority to agencies since then had it foreseen later judicial decisions that increase presidential control over that authority? The Court’s severability doctrine asks whether Congress would have preferred an agency subject to greater presidential control over no agency at all, and almost invariably concludes that Congress would have preferred the former. But even on the Court’s own terms, that is the wrong question. The right one is whether Congress would have preferred an agency with less authority over one subject to the combined loss of each restriction on presidential control. If Congress knew, for example, not only that the President would be able to fire agency heads, but also could direct discretion formally delegated to agencies, ignore legislative vetoes, and receive broad immunity, would Congress have enacted the same delegations? The Court has never addressed the aggregate effect of its holdings. Slaughter, however, forces this issue to the forefront.