A fair amount of the argument in favor of the unitary executive model of presidential power is rooted in an account of the Founding. I am skeptical of important parts of that account. But the point I want to raise here is about the impact, on theories of presidential power, of a later event in the making of constitutional law. That later event is the adoption of the Twenty-Second Amendment, under which a twice-elected president cannot run for re-election. The point I want to raise is about what an enactment like the Twenty-Second Amendment can tell us, not about any original vision of the presidency, but about the vision of the presidency that prevailed in 1951, when the Amendment was adopted. The further question I want to raise goes like this: if the Constitution’s provisions about the presidency were partly written in the middle of the twentieth century, what role should the prevailing understanding of the presidency as it existed at that time play in an overall theory of presidential power? The chief bearing of this question, for present purposes, is that it might induce skepticism about an important aspect of unitary executive theory: that is, the proposition that the president must be able to remove executive officers at will, and that Congress lacks the power to insulate executive-branch officials—or maybe even executive-branch personnel generally—from the presidential ax.
Unitary executive theory is heavily structural. To be sure, it comes with a (contestable) reading
of the text of Article II and various claims about the way the Founding
generation understood executive power.
But in the form that has been most potent within the judiciary, unitary
executive theory depends for its force on an idea about democratic
accountability: that the personnel of the executive branch must be accountable
to the president, who in turn is accountable to the electorate. For the president to be held responsible for
the actions of executive officials, the thinking goes, he must be able to fire
executive-branch personnel who fail to carry out their duties in the way he
thinks proper. The public’s assurance
that the president will for the most part exercise his authority over the
executive branch prudently, and in the public interest, lies in the public’s
ability to hold the president accountable at the polls.
To be sure, a modicum of reality-checking would reveal
that democratic elections are blunt mechanisms for holding executive-branch
personnel to account. Presidential
elections turn on multiple factors, some of which would be beyond the control
of presidents even if presidents were able to direct all executive-branch
action with impeccable precision. Given
the size of the electorate, the infrequency of elections, and the dizzying
array of reasons why people vote the way they do, the chances that any
particular decision a president makes about firing or retaining a subordinate
will play a significant role in a re-election campaign are rather small. Still, the basic idea that elections are a
mechanism of accountability has something to it. And one of the chief reasons that presidents
(and many other officials) are thought trustworthy to exercise the powers of
office is that they can be turned out of office if they use those powers
poorly.
The Twenty-Second Amendment puts a major dent in that
idea. It means that the President of the
United States will frequently be electorally unaccountable. If you cannot run for re-election, the voters
cannot police your behavior by threatening to refuse to return you to
office. To be sure, presidents
ineligible to run for re-election might have other incentives to stay popular
with the electorate: popularity might help them get cooperation from Congress,
and an unpopular president might damage his party’s electoral fortunes even
after he is out of the game, and so on.
But it would be hard to say that the elimination of the straightforward
electoral check makes no significant difference in the degree to which a
president is democratically accountable.
And once we recognize that the Twenty-Second Amendment makes presidents
less democratically accountable than they previously were, we should ask what
assumptions about the presidency might have helped the Americans who supported
that Amendment reach the view, all things considered, that it was safe to vest
the considerable powers of the presidency in someone who be electorally
unaccountable much of the time.
I make no strong claim about the answer. But it is worth noting that the presidency the
drafters and ratifiers of the Twenty-Second Amendment knew was not the
presidency of unitary executive theory. He
was enormously powerful: the Leader of the Free World, with access to a (small
but terrifying) stock of nuclear weaponry.
But he did not have unlimited freedom to direct the bureaucracy, or
choose the personnel, of the modern administrative state. A dozen years before Congress proposed the
Twenty-Second Amendment, the Supreme Court decided Humphrey’s Executor v. United States, holding that Congress had the
authority to constrain the president’s latitude to dismiss administrative
commissioners and, more broadly, federal officials whose duties could be
described as other than “purely executive.”
One year before proposing the Twenty-Second Amendment, Congress passed
the Administrative Procedure Act, whose rules channel and constrain the
exercises of bureaucratic power, including a great deal of power exercised by
people appointed by or answerable to the president. In short, the presidency that the framers and
ratifiers of the Twenty-Second Amendment contemplated—and whose electoral
accountability they were willing to compromise—was a president whose freedom of
action was importantly constrainable by Congress, including through the
mechanism of insulating federal officers from at-will removal.
I’m not claiming that the Twenty-Second Amendment
would not have been adopted if the constitutional law of the mid-twentieth
century had conformed to the unitary-executive model that later gained so many
adherents. How post-war Americans might
or might not have thought differently about the presidency if this or that
feature of the office had been different then is a question on which proof is
essentially impossible. In the age of
totalitarian dictators, maybe the fear of a president-for-life would have been
strong enough to inspire a constitutional term limit, and the concomitant loss
of electoral accountability during a second term, even if the president had had
the power that unitary executive theory would give him. Or maybe, in that same age of totalitarian
dictators, the successful movement for a constitutional amendment limiting
presidential terms would also have been a successful movement for a
constitutional amendment disestablishing a unitary-executive model of the
presidency, had that model been operative at the time. It’s impossible to know.
But two propositions seem clear. First, a theory of presidential power that
rests in any significant part on ideas about what the Constitution-makers
presumed about the office needs to take account of the presumptions that
operated for the Constitution-makers of 1947-51 and not just those of 1787-88. Second, the presidency of 1947-51 was not the
presidency of unitary-executive theory.
Indeed, the Constitution-makers of 1947-51 restructured the office in a
way that has meant that a great deal of the time, the president is not an
officer for whom the mechanism of democratic accountability functions in the
way that the theory standardly presumes.
(Or more precisely, the president is an officer for whom that mechanism,
which never functions more than bluntly in the real world, does not even function
as a matter of ideal theory.) Perhaps
that matters.