Several years ago I wrote a book with Adrian Vermeule called The Executive Unbound. The book was widely interpreted to argue something like: “the president has unlimited power and that is a good thing, too.” Last week, I published an op-ed in the New York Times (written with my colleagues Daniel Hemel and Jonathan Masur) that argued that a court might block Trump’s plan to build a wall on the Mexican border if he cannot persuade the judge that the wall passes a cost-benefit analysis. The op-ed emerged in part from as yet unpublished work with Masur that argues that judges should require regulatory agencies to use cost-benefit analysis. Are these two positions consistent? The executive must be either bound or unbound; which is it?

The popular interpretation of The Executive Unbound was always a caricature. The Executive Unbound compared two systems of government: the “liberal legalism” model in which Congress takes the lead in making policy within a system of separation of powers; and the “presidential primacy” model in which most policymaking takes place in the executive branch, by regulatory agencies under the broad supervision of the president. Our argument was that while the liberal legalism model remains the official story, the presidential primacy model more accurately depicts the workings of the government.

This confusion between ideology and reality is important. Liberal legalists have constantly complained about the wholesale shift of power to the president from Congress and the courts. And yet this shift has occurred partly because those same liberal legalists have supported consistent incremental shifts in this direction, based on the needs of governance as they developed over many decades. Just think how recently liberals supported Obama’s immigration orders (after Congress gridlocked on immigration), Obama’s climate regulation orders (after Congress gridlocked on climate change), and Obama’s recess appointments (after Congress gridlocked on his nominations). We argued that the long-term trajectory in favor of presidential power—more generally, the centralization of power within the national government in the office of the presidency—reflects economic, political, and technological trends that are (more or less) inevitable, and in the face of which the separation of powers system gridlocks. To that extent, if you want government to govern, you need to learn to live with presidential primacy.

While the title of the book and some passages may have led readers to think that the book imagines that the president is subject to literally no constraints from Congress and the courts, that was never the argument. The book would make little sense if it was. One chapter discusses presidential power during emergencies; the argument that the president’s freedom of action expands during emergencies would make little sense if the president’s freedom of action was already at a maximum during normal times. The book observes that most of the expansion of executive power took place through voluntary delegations by Congress, which, at least formally, retains the power to withdraw those delegations or impose additional constraints on them. And the book describes a historical process that is still going on; it does not claim that we have reached an endpoint of maximal presidential power, whatever that would mean.

My relatively new belief that courts should review regulators’ cost-benefit analysis is a change of mind: I used to think otherwise, and I explain my reasons in the as yet unposted paper with Masur. But I’ve always been a fan of cost-benefit analysis, and it’s only recently that I’ve worked out my thoughts about how it should be institutionalized in government. If my thoughts about the relationship between courts, regulators, and the executive are evolving at the margin, that’s just because I know more today than I did seven years ago. That said, the role of the judiciary in ensuring that agencies perform cost-benefit analysis must remain limited, for reasons that Masur and I explore.

Finally, as to Mr. Trump. The Executive Unbound acknowledged that you don’t always get the president you want. You also don’t always get the Congress you want. If the choice is between action and gridlock, I still prefer action (in part because gridlock will eventually lead to action of a bad kind).

What will happen if Trump causes the devastation that is so widely predicted? It depends on what exactly he does. If he establishes a dictatorship and starts World War III, all bets are off. A more realistic worst-case prediction at this point is something like a Nixon presidency, except without the foreign policy accomplishments, ending in impeachment, resignation, or simply an electoral loss. If so, then expect Congress, presumably now populated with Democrats, to react as it did after Nixon’s resignation—by enacting new statutes that attempt to rein in the presidency. And expect those new statutes to erode in short order just as the 1970s statutes did.