Chapter 9: The Presidency — Powers, Constraints, and the Expanding Executive

9.1 Two Pictures of the President

There are two pictures of the American presidency that compete for your attention.

The first is the campaign-trail picture. A candidate stands in front of an enormous flag, promising to fix immigration, end the drug crisis, restore manufacturing, lower the price of insulin, defend American allies, get tough on adversaries, balance the budget, eliminate waste, and bring the country together. Each promise is delivered as if the president alone could deliver it. A president, in this picture, is something between a king and a CEO: a person who, on Day One, can sign documents, issue orders, and reshape the country.

The second picture is the one you encounter the first time you sit in a White House office and try to do something. You discover that the federal government is a machine of two and a half million civilian employees plus another two million in uniform, that most of those employees do not work for you, that Congress controls the money you would need to do almost anything new, that the courts can pause your actions while litigation proceeds, that the agencies you nominally lead have their own statutes, their own career staff, their own institutional cultures, and their own relationships with the congressional committees that fund them. You discover, in Harry Truman's famous formulation, that you cannot just sit at your desk and shout "do this" and "do that." You will sit there, Truman said, and "nothing will happen."¹

Both pictures are partly true. The American president really is constitutionally and politically more powerful than any head of government in any other developed democracy with comparable demographic scale. But the American president is also, on most domestic-policy questions, embedded in a separation-of-powers system that requires the cooperation of Congress, the courts, the bureaucracy, the states, and the public. The campaign-trail picture overstates what the president can do alone. The "nothing will happen" picture overstates what the president cannot do alone. The truth is in the gap between them, and that gap has been moving — almost continuously since 1789, and especially since the 1930s — in the direction of more presidential power.

This chapter is the first of three on the executive branch. Chapter 10 covers the vice presidency, the Cabinet, and the Executive Office of the President — the people and institutions that make the presidency operational. Chapter 11 covers the federal bureaucracy: the agencies, the civil service, and the long-running fight over whether the people who actually carry out federal policy should be more responsive to political control or more insulated from it. The current chapter is about the office itself: the constitutional powers, the legal doctrines, the political constraints, and the long arc of expansion.

We will be honest, in this chapter, about something that is sometimes obscured in textbook treatments: both parties expand executive power when they hold the White House, and both parties criticize executive expansion when they do not. We will name the cases. Obama's DACA executive action and his pen-and-phone strategy. Trump's travel ban and his 2025 second-term executive-order pace. Biden's student-loan forgiveness attempt and his executive orders on climate, immigration, and labor. Trump's 2025 Schedule F revival, his tariff actions under emergency authority, and his expanded use of the recess-appointment doctrine. The expansion is bipartisan. The justifications are bipartisan. The pattern is structural, not merely a function of any one administration's character.

We will also be honest that some recent disputes are genuinely contested and not yet doctrinally settled. The unitary-executive theory is one of those. Trump v. United States (2024) is another. We will steel-man both sides on each, and we will not pretend the constitutional questions have been resolved when they have not.

Let us start where every honest discussion of presidential power has to start: with the actual text.

9.2 What Article II Actually Says

Article II of the Constitution is short. It runs to roughly a thousand words. By comparison, Article I — which establishes Congress — runs to about 2,300. The Founders wrote far more about the legislature, and far less about the executive, than later students of the document expect. This brevity is a feature, not an oversight. The Founders had just finished a war against an executive (King George III) whom they regarded as having claimed too much power. They had also lived under the Articles of Confederation, where there was no national executive at all, and discovered that the absence of one made the central government incapable of executing its own laws. Article II is the compromise: an executive with enough power to function, described in terms general enough to permit adaptation to circumstances the Founders could not anticipate.

We will take the principal clauses in order, and translate each.

9.2.1 The Vesting Clause

The opening sentence of Article II reads: "The executive Power shall be vested in a President of the United States of America." That is it. Eleven words. They are perhaps the most contested eleven words in the Constitution.

In plain English: there is one president, not a council; the executive power belongs to that president, not to a committee or to Congress; and the executive power is, whatever else it is, executive, not legislative or judicial. What Article I does for Congress in much greater detail ("All legislative Powers herein granted shall be vested in a Congress of the United States"), Article II does for the president in this single sentence.

The contest is over the meaning of "the executive Power." Strong unitarians — we will meet them again in §9.4 — read this clause to vest in the president, by virtue of the office, all executive power as it was understood in 1789. That includes broad authority to direct subordinate executive officers, to remove them at will, to control the bureaucracy, and to act with energy in all matters that are not specifically assigned by the Constitution to Congress or the courts. Weak unitarians read the clause more narrowly: it vests in the president the executive powers that the Constitution itself enumerates in Article II, plus those that Congress has chosen, by statute, to confer on the executive. The argument turns, in part, on the difference between the executive Power (the article suggests something pre-existing, with content that does not require enumeration) and the executive powers (a more limited reading that requires enumeration to give content).

This is not merely an academic dispute. Seila Law v. CFPB (2020), Collins v. Yellen (2021), and a string of removal-power cases in between have turned, doctrinally, on what the Vesting Clause was understood to do.

9.2.2 The Take Care Clause

Article II, Section 3 directs that the president "shall take Care that the Laws be faithfully executed." Plain English: the president has a constitutional duty to enforce the laws Congress passes, even laws he or she disagrees with, and to do so in good faith.

This clause is doing several things at once. It is a duty: the president shall take care, not may take care. It is a constraint: the president must execute the laws faithfully, which has been read to forbid the president from refusing to enforce a duly enacted statute simply because the president dislikes it. And it is also, in some readings, an empowerment: to take care that the laws are faithfully executed, the president must have at least some authority over those who do the executing.

The Take Care Clause has been at the center of every major dispute about presidential non-enforcement. When the Obama administration declined to defend the Defense of Marriage Act in court (2011), critics argued the Take Care Clause was being violated. When the Obama administration declined to deport certain categories of undocumented immigrants under DACA (2012), critics argued the same. When the Trump administration declined to defend the Affordable Care Act in court (2018) and supported a lawsuit to invalidate it, critics — including many of the same critics, now from the other side — argued the same. When the Biden administration imposed a moratorium on federal student-loan repayment that the Supreme Court eventually held to exceed statutory authority (Biden v. Nebraska, 2023), the Take Care question came up again. The pattern is clear: every modern administration has been accused, by its political opponents, of failing to take care that laws it found inconvenient were faithfully executed. Whether each accusation was correct on the merits is a separate matter; the political pattern is bipartisan.

9.2.3 The Commander-in-Chief Clause

Article II, Section 2 makes the president "Commander in Chief of the Army and Navy of the United States, and of the Militia of the several States, when called into the actual Service of the United States." Plain English: when the country is at war, or when the militia (now the National Guard) has been activated for federal duty, the president is the top military officer.

What the clause does not say — and the omission is critical — is that the president has the power to start a war. The power to declare war is given, by Article I, Section 8, to Congress. The constitutional architecture, as a matter of text, is that Congress decides when there is a war, and the president runs the war once it has been declared. We will return to the gap between this architecture and modern practice in §9.5.

9.2.4 The Treaty Power and Appointments

Article II, Section 2 also gives the president the power to "make Treaties, provided two thirds of the Senators present concur," and to "nominate, and by and with the Advice and Consent of the Senate, … appoint Ambassadors, other public Ministers and Consuls, Judges of the supreme Court, and all other Officers of the United States, whose Appointments are not herein otherwise provided for."

Plain English: the president negotiates treaties with foreign countries, but a treaty does not become binding U.S. law unless two-thirds of senators present agree. The president nominates federal judges, ambassadors, and senior executive officials, but those appointments require Senate confirmation by simple majority.

Two practical features have emerged that the Founders did not anticipate.

First, the executive agreement has largely supplanted the formal treaty for international commitments. An executive agreement is a deal the president signs with a foreign government that does not require Senate ratification. Some are based on prior congressional authorization (so-called "congressional-executive agreements," like most modern trade deals); others are pure executive agreements based on the president's own constitutional authority. The Iran Nuclear Deal of 2015 — formally the Joint Comprehensive Plan of Action — was structured as an executive agreement rather than a treaty, in part because the Obama administration knew it could not get sixty-seven votes for a treaty in the Senate. The Paris Climate Agreement (2015–2016) was similarly structured as an executive agreement. Both were entered without Senate approval; both were exited (Trump 2017–2018) by simple presidential action; the Iran deal was rejoined and then exited again across administrations. The treaty form, with its supermajority requirement, has become rare for any politically contested commitment.

Second, the Senate-confirmation slowdown has changed how appointments work. As of the 2020s, hundreds of senior executive positions either go unfilled for long stretches or are filled by acting officials who did not require Senate confirmation. The slowdown has multiple causes — the rise of the routine filibuster on judicial and some executive nominations until the 2013 nuclear-option precedent, the Trump administration's preference for acting officials, the polarization of the confirmation process — but the practical effect has been to make Senate advice-and-consent less of a real check than the constitutional design contemplated.

9.2.5 The Recess-Appointment Power

The Constitution permits the president to "fill up all Vacancies that may happen during the Recess of the Senate, by granting Commissions which shall expire at the End of their next Session" (Article II, Section 2). Plain English: when the Senate is on recess, the president can fill vacancies without Senate confirmation, and the appointment lasts until the end of the next Senate session.

This clause was designed for a Senate that met for short periods and was then absent for months at a time. In the modern Congress, where the Senate is technically in session most of the year (often via brief "pro forma" sessions held precisely to prevent recess appointments), the clause has become a matter of recurring dispute. The Supreme Court's decision in NLRB v. Noel Canning (2014) limited the recess-appointment power: the Court held that pro forma sessions count as sessions for the purposes of the clause, and that recesses must be of substantial duration (the Court suggested ten days or more) before the recess-appointment power can be exercised. The Trump administration's 2025 effort to revive aggressive recess-appointment use, including pressure on the Senate majority leader to schedule recesses long enough to permit appointments, is an active site of constitutional and political contestation as of this writing.

9.2.6 The Pardon Power

Article II, Section 2 grants the president "Power to grant Reprieves and Pardons for Offences against the United States, except in Cases of Impeachment." Plain English: the president can forgive, or reduce the punishment for, federal crimes, with one exception (impeachment).

The pardon power is, by constitutional text, almost unlimited. The president can pardon before charges are filed, after conviction, after sentencing, or after release; can pardon a single person, a class of persons, or — at the limit — anyone the president chooses, for any reason or for no stated reason; and is not subject to congressional override on individual pardons. The two textual limits are that pardons cover only federal offenses (state convictions are unreached) and that the pardon power cannot be used to undo an impeachment.

Recent practice across both parties:

• Gerald Ford pardoned Richard Nixon (1974) for any federal crimes Nixon might have committed in office, before any indictment. The pardon was extremely unpopular and is widely credited with contributing to Ford's loss in 1976; it is also widely credited, by historians and political scientists, with allowing the country to move past Watergate.
• Jimmy Carter issued a blanket pardon (1977) to Vietnam-era draft evaders, fulfilling a campaign promise.
• George H. W. Bush pardoned six Reagan-administration figures involved in the Iran-Contra affair (1992), including former Defense Secretary Caspar Weinberger before his trial.
• Bill Clinton issued 140 pardons on his last day in office (January 20, 2001), including the pardon of fugitive financier Marc Rich. The Rich pardon prompted bipartisan criticism and a federal investigation that produced no charges.
• Barack Obama issued 1,927 commutations and pardons over his two terms, the most of any modern president, focused heavily on non-violent drug offenders serving long sentences under prior mandatory-minimum regimes.
• Donald Trump (first term) pardoned former Sheriff Joe Arpaio (2017), former campaign chair Paul Manafort (2020), longtime adviser Roger Stone (2020), former National Security Advisor Michael Flynn (2020), and others connected to the Russia investigation, plus various politically connected figures.
• Joe Biden issued a categorical pardon of federal marijuana possession offenses (2022 and again in 2024), pardoned his son Hunter Biden (December 2024) on charges that had already been brought (a politically controversial act that Biden had previously said he would not take), and issued a series of preemptive pardons for members of the January 6 Select Committee and for several Biden-administration officials in his last days in office (January 2025).
• Donald Trump (second term) pardoned approximately 1,500 January 6, 2021 defendants on his first day back in office (January 20, 2025), a categorical pardon that included defendants convicted of violent offenses against police officers as well as those convicted of trespass and disorderly conduct.

Two patterns are visible. First, the use of the pardon power has become more politically charged, with more pardons of close associates, more last-day pardons, and more categorical pardons aimed at political constituencies. Second, both parties have engaged in this expansion. Critics of any of these pardons can point to defenders' pardons as precedent. Reform proposals — none of which have advanced significantly — have included a constitutional amendment requiring pardon disclosure and reasoning, a statutory ban on self-pardons (the constitutionality of a self-pardon has never been adjudicated), and waiting periods before pardons of close associates. The status quo, as of 2026, is that the pardon power remains as broad in practice as it appears on the constitutional page.

9.2.7 The Veto and Presentment

Article I, Section 7 — technically about Congress, but operationally about the presidency — provides that every bill passed by both chambers "shall, before it become a Law, be presented to the President of the United States; If he approve he shall sign it, but if not he shall return it, with his Objections," and that the bill becomes law over the president's veto only if both chambers re-pass it by two-thirds majorities.

Plain English: every bill that passes Congress goes to the president. The president can sign it (it becomes law), veto it (it returns to Congress), or do nothing for ten days (with two distinct outcomes, depending on whether Congress is in session).

There are three principal forms of veto.

1. Regular veto. The president returns the bill to Congress with a statement of objections. Congress can override by two-thirds vote in each chamber.
2. Pocket veto. If Congress adjourns within ten days of the bill being presented, and the president has not signed it, the bill does not become law and Congress cannot override (because Congress is not in session).
3. Line-item veto. A power some governors hold to veto specific provisions of a bill while signing the rest. The federal Line Item Veto Act of 1996 attempted to grant this power to the president; the Supreme Court held it unconstitutional in Clinton v. New York (1998) on the ground that it violated the Presentment Clause's requirement that the president act on bills as Congress passes them, not on edited versions.

The empirical pattern of vetoes is straightforward: most presidents veto a small number of bills (Obama 12, Trump-1 10, Biden 13 through mid-2024), and most vetoes are sustained. Override is rare. In the modern Congress, the threshold of two-thirds in both chambers is essentially never met on a partisan bill; the president's veto pen is, in practice, an absolute negative on legislation the president's party opposes.

9.2.8 The State of the Union and Ambassador Reception

Article II, Section 3 directs that the president "shall from time to time give to the Congress Information of the State of the Union, and recommend to their Consideration such Measures as he shall judge necessary and expedient." Plain English: the president has to report to Congress periodically on how the country is doing, and can propose legislation. The same section gives the president the power to "receive Ambassadors and other public Ministers."

These clauses look minor, and in some senses they are. But the State of the Union has evolved into the principal annual presidential address, broadcast to every household with a television, and the ambassador-reception clause is the textual basis for the president's exclusive control over diplomatic recognition. When the United States recognized Israel in 1948 (Truman), opened relations with the People's Republic of China in 1979 (Carter), broke relations with Iran in 1980 (Carter), restored relations with Cuba in 2015 (Obama), or recognized Jerusalem as Israel's capital in 2017 (Trump), each of those acts was based on the ambassador-reception clause. Recognition power, as a matter of constitutional doctrine, belongs to the president alone (Zivotofsky v. Kerry, 2015).

9.3 The Long Arc of Executive Expansion

The American presidency in 2026 is more powerful, in almost every measurable respect, than the presidency the Founders described in Article II. Tracing the expansion is partly an exercise in constitutional history and partly an exercise in political economy: power flowed to the office for reasons.

Let us walk the arc, briefly.

Hamilton's "energy." Federalist No. 70 (Hamilton, 1788) argued for a "vigorous Executive," and specifically for a single executive (rather than a council) on the ground that a single executive could act with energy, secrecy, and dispatch — the qualities needed in foreign affairs and military emergencies. Hamilton's argument is the deepest theoretical grounding for the presidency's claim to robust independent authority. It is also the argument that strong unitarians return to repeatedly.

Jefferson and the Louisiana Purchase (1803). Jefferson, who had argued for a strict reading of presidential power, executed the Louisiana Purchase on terms he himself doubted he had constitutional authority to accept. The doubling of the country's territory was popular enough that no constitutional challenge took hold, but the precedent — that a popular president can stretch his powers in the face of an opportunity — was set early.

Jackson and the veto. Andrew Jackson (1829–1837) used the veto more freely than any prior president, vetoing twelve bills (more than all his predecessors combined). His veto of the recharter of the Bank of the United States (1832) was based not on a constitutional objection — the Supreme Court had already held the Bank constitutional — but on Jackson's policy preference. The precedent that the president could veto on policy grounds, and that the veto is a political weapon in addition to a constitutional check, was established.

Lincoln and the Civil War (1861–1865). Lincoln took an enormous range of executive actions in the first months of the Civil War without prior congressional authorization: he raised troops, suspended habeas corpus, ordered a naval blockade of Southern ports, expanded the army beyond statutory limits, and disbursed funds. He sought congressional ratification afterward, which Congress provided. The Lincoln precedent is that, in genuine national emergency, the president may act first and seek authorization later. It is the deepest historical foundation for unitary-executive arguments and the deepest worry of those who fear that "emergency" becomes the all-purpose justification for circumventing Article I.

Theodore Roosevelt and the "stewardship" presidency (1901–1909). TR articulated what he called the "stewardship theory": the president, as the only nationally elected official, is steward of the public welfare and may do whatever the Constitution and laws do not specifically forbid. This is the inverse of William Howard Taft's later "literalist" theory: the president may do only what the Constitution and laws specifically permit. Most subsequent presidents have, in practice, operated closer to TR's stewardship theory than to Taft's literalism, regardless of party.

FDR and the New Deal (1933–1945). Franklin Roosevelt expanded the executive branch in scale and in claimed authority more than any president before or since. The administrative state — the network of regulatory agencies that now governs much of American economic life — was largely built between 1933 and 1945. The Executive Office of the President was created in 1939 to give the president institutional capacity to manage the expanded executive. The Supreme Court, after striking down several New Deal initiatives in 1935–1936, accommodated the expansion after 1937. The wartime presidency that followed — including the Japanese-American internment (Korematsu, 1944), one of the most consequential and most criticized exercises of executive power in American history — established the modern pattern of broad presidential discretion in war.

The Cold War presidency (1945–1989). With the United States as a superpower in continuous near-conflict, the imperial presidency that Arthur Schlesinger Jr. would name in 1973 took shape. The National Security Act of 1947 created the modern national-security apparatus. The Korean War (1950–1953) was fought without a declaration of war. So, eventually, was Vietnam, on the basis of the Gulf of Tonkin Resolution (1964), which Congress later regretted but did not repeal until 1971. Presidents claimed inherent authority to act in foreign affairs that Congress only intermittently contested.

Nixon and the post-Watergate retrenchment (1969–1974). Richard Nixon's administration represented the high-water mark of imperial-presidency claims and the beginning of their reversal. Nixon was forced to resign rather than face certain impeachment and conviction over Watergate. United States v. Nixon (1974) established that executive privilege has limits. Congress passed the War Powers Resolution (1973), the Impoundment Control Act (1974), the Foreign Intelligence Surveillance Act (1978), and the Inspector General Act (1978) in an effort to claw back authority. The arc, briefly, reversed.

Reagan and the conservative re-expansion (1981–1989). Reagan's administration revived strong-executive claims under a conservative banner: the president's authority to direct the executive branch, to control regulatory agencies through OIRA review, to make signing-statement claims about how laws should be interpreted. The Office of Legal Counsel, under Theodore Olson and others, articulated what became the modern unitary-executive theory. Reagan's Federalism Executive Order (12612, 1987) was an early example of a presidential effort to discipline regulatory agencies' relationships with the states.

Post–Cold War, post-9/11. The Cold War's end did not retract presidential power. The first Gulf War (1991), the Bosnia and Kosovo interventions (1995–1999), and the post-9/11 Authorization for Use of Military Force (2001) all confirmed that presidents could act militarily on broad and sometimes vague legislative grants. The 2002 AUMF authorizing force against Iraq was the last formal AUMF Congress has passed. Both — the 2001 and 2002 AUMFs — remain operative today, almost a quarter-century after they were enacted. We will return to AUMFs in §9.5.

Obama's "pen and phone" (2009–2017). Frustrated by Republican congressional opposition after 2010, the Obama administration explicitly committed to a strategy of executive action where legislative action was blocked. DACA (2012), DAPA (2014, blocked by the courts), the Clean Power Plan (2015, ultimately stayed and replaced), and the Iran Nuclear Deal (2015, structured as an executive agreement) all reflected this strategy. Republicans denounced executive overreach; the Obama legal team replied that the president was acting within statutory authority and prosecutorial discretion.

Trump's first term (2017–2021). The Trump administration pursued an aggressive executive-action agenda from the opposite direction: travel-ban executive orders, deregulatory actions, the family-separation policy, the use of emergency-declaration authority to redirect funding for the border wall. Democrats denounced executive overreach; the Trump legal team replied that the president was acting within statutory and constitutional authority.

Biden (2021–2025). Biden issued more executive orders in his first 100 days than any predecessor, many reversing Trump-era orders. The student-loan forgiveness initiative was eventually struck down (Biden v. Nebraska, 2023). Vaccine-mandate executive actions during the COVID-19 pandemic produced extensive litigation. Republicans denounced executive overreach; the Biden legal team replied that the president was acting within statutory authority.

Trump's second term (2025–). The Trump 2.0 administration's first weeks produced an unusually high volume of executive orders, including orders on federal workforce restructuring (the Schedule F revival), tariffs under emergency-declaration authority, agency reorganizations, and immigration enforcement. Litigation is extensive and ongoing. As of this writing in 2026, multiple suits are pending; some early rulings have constrained particular orders, others have not.

The pattern is clear. Each modern administration enters office promising to use executive power to achieve goals Congress has not authorized; each modern administration is criticized by the opposing party for doing so; each modern administration, once it leaves office, leaves precedents and institutional capacities the next administration of either party can use. Anyone who has spent a serious amount of time in Washington across multiple administrations will tell you the same thing: the structure of incentives produces this trajectory regardless of which party holds the White House.

9.4 The Unitary Executive Debate

Within constitutional law, the deepest dispute about the modern presidency goes by the name "the unitary executive theory."

The phrase has been used in two quite different senses, and a lot of public confusion comes from conflating them.

The weak unitarian thesis is: the executive branch is hierarchical, with the president at its head; the president has at least some power to direct subordinate executive officers; and Congress cannot, consistent with Article II, fully sever the president from the execution of federal law. This is something close to a consensus position among constitutional scholars. The president cannot, for example, be prevented by Congress from firing every executive officer for any reason; some core removal authority is part of the executive power.

The strong unitarian thesis goes further: the executive power vested in the president by Article II includes the power to direct or remove every officer who exercises executive power, with at most narrow exceptions. On this view, statutes that insulate executive officers from presidential removal — independent agencies whose commissioners can only be fired "for cause," for instance — raise serious constitutional questions. Strong unitarians associated with this view include Steven Calabresi (Northwestern), John Yoo (Berkeley), Saikrishna Prakash (Virginia), and a number of judges nominated to the federal bench by Republican presidents since the 1980s. The intellectual home of the strong-unitarian movement has been the Federalist Society and the Office of Legal Counsel during Republican administrations.

The anti-unitarian position is associated with scholars including Peter Strauss (Columbia), Cass Sunstein (Harvard), Lisa Heinzerling (Georgetown), and Bruce Ackerman (Yale). On this view, the Constitution's text does not require strong unitarianism; the historical practice from 1789 onward has accepted some executive officers who are not under direct presidential control; and the policy consequences of strong unitarianism — concentrating effective control of the regulatory state in the president — are inconsistent with the rule of law and with congressional Article I authority over the structure of the executive branch.

Three recent Supreme Court cases have shifted the doctrine in a strong-unitarian direction.

Seila Law v. CFPB (2020). The Consumer Financial Protection Bureau, created by the Dodd-Frank Act of 2010, was led by a single director removable by the president only for "inefficiency, neglect of duty, or malfeasance in office." The Supreme Court (5–4, with Chief Justice Roberts writing) held that this removal restriction was unconstitutional. A single-director independent agency, the Court reasoned, "concentrates power in a unilateral actor insulated from Presidential control" in a way the Constitution does not permit. The Court left in place removal restrictions for multi-member commissions (Humphrey's Executor, 1935), but the doctrinal trajectory was clear.

Collins v. Yellen (2021). The Federal Housing Finance Agency had a similar single-director structure with for-cause removal. The Court applied Seila Law and struck down the removal restriction.

Trump v. United States (2024). The Court held, 6–3, that a former president has at least presumptive immunity from criminal prosecution for "official acts" taken while in office. The reasoning relied heavily on Vesting Clause and unitary-executive arguments: a president, the majority reasoned, must be free to take vigorous official action without the prospect of post-office criminal liability. The dissent — and many constitutional scholars across the ideological spectrum — argued that the decision creates a category of presidential immunity that the Constitution does not expressly grant and that risks placing presidential conduct beyond the rule of law. We will return to this case in §9.10.

Steel-manning both sides

The strongest unitarian argument, as Yoo, Calabresi, and Prakash present it, runs like this: The constitutional text vests "the executive Power" in a president, singular. The Founders rejected proposals for a plural executive precisely because they believed that energy, accountability, and dispatch required a single chief executive. If Congress can sever pieces of executive power from presidential control by creating "independent" agencies, the unitary structure the Founders chose can be steadily eroded. Moreover, accountability requires unity: voters know who to hold responsible for executive-branch failures only if the president is genuinely in charge of the executive branch. An "independent" CFPB or Federal Reserve diffuses accountability in a way that is bad for democratic legitimacy.

The strongest anti-unitarian argument, as Strauss, Ackerman, and Sunstein present it, runs like this: The constitutional text does not say that all executive officers are removable at the president's pleasure; it says that "the executive Power" is vested in a president and that the president "shall take Care that the Laws be faithfully executed." Both phrases are compatible with a structure in which Congress, exercising its Article I power to "make all Laws which shall be necessary and proper for carrying into Execution the foregoing Powers, and all other Powers vested by this Constitution in the Government of the United States, or in any Department or Officer thereof," creates executive officers with structural protections from presidential control where Congress concludes the public interest requires it. The practice of independent agencies goes back to the Interstate Commerce Commission of 1887 and was not seriously challenged on constitutional grounds for nearly a century. Humphrey's Executor (1935), holding that for-cause removal restrictions on FTC commissioners were constitutional, was settled doctrine for eighty years. The strong-unitarian campaign to dismantle that doctrine is itself a constitutional innovation, however much it is dressed in originalist rhetoric. And the policy consequences — concentrating control of the regulatory state in the president — are not obviously consistent with the Founders' design, which, after all, was for a separation of powers, not a fusion of legislative-executive authority in the executive.

The honest position, as of 2026, is that the doctrinal trajectory has been moving in the strong-unitarian direction, that Trump v. United States further accelerated that trajectory, and that the next set of cases — likely involving the Federal Reserve, the FTC, and various inspector-general protections — will determine how far the trajectory goes. The constitutional debate is genuinely live.

9.5 War Powers

The constitutional architecture for war is straightforward in text and confused in practice.

Article I, Section 8 gives Congress the power "to declare War, grant Letters of Marque and Reprisal, and make Rules concerning Captures on Land and Water." It also gives Congress the power to raise armies, maintain a navy, and "make Rules for the Government and Regulation of the land and naval Forces."

Article II, Section 2 makes the president "Commander in Chief of the Army and Navy of the United States."

The textual division is: Congress decides whether to go to war; the president runs the war once it has been decided.

The practical division, since 1945, has not been that. The United States has fought five wars long enough to be remembered by name — Korea (1950–1953), Vietnam (1964–1973), the first Gulf War (1991), Afghanistan (2001–2021), and Iraq (2003–2011 and 2014–) — plus dozens of smaller engagements. Of these, only the first Gulf War received a congressional authorization that came reasonably close to a traditional declaration of war (the Authorization for Use of Military Force Against Iraq Resolution of 1991). None received a formal declaration of war in the constitutional sense. Korea was fought on UN authority and presidential commander-in-chief power. Vietnam was fought on the basis of the Gulf of Tonkin Resolution (1964), a broad authorization Congress later regretted. The post-9/11 wars in Afghanistan and Iraq have been fought, in part, on the basis of two AUMFs:

• AUMF 2001, authorizing force against "those nations, organizations, or persons" the president determined "planned, authorized, committed, or aided" the September 11, 2001 attacks. Used as the legal basis for the Afghanistan war, the global counter-terrorism campaign, drone strikes in Yemen, Somalia, Pakistan, Libya, and Syria, and operations against the Islamic State after 2014. Still operative as of 2026.
• AUMF 2002, authorizing force against Iraq. Used as the original authority for the Iraq war, and in the 2010s reinterpreted to cover operations against the Islamic State in Iraq and Syria. Still operative as of 2026.

Congress passed the War Powers Resolution in 1973 to constrain unilateral presidential war-making. The resolution requires the president to consult with Congress before introducing forces into hostilities "in every possible instance"; to report to Congress within 48 hours when forces are introduced; and to withdraw the forces within 60 days (with a 30-day extension for safety) unless Congress has either declared war or specifically authorized the use of force.

The empirical record of the WPR is unflattering. Every president since Nixon has, in some form, claimed that the resolution unconstitutionally infringes on the president's commander-in-chief authority and has acted accordingly. Reagan's deployment to Lebanon (1982–1984), the Grenada invasion (1983), the Panama invasion (1989), the Bosnia and Kosovo interventions (1995, 1999), the Libya intervention (2011), the strikes against Syria (2017, 2018), the recent strikes in Yemen (2023–2024), and the response operations in Iraq and Syria (2024) all involved either non-compliance or contested compliance with the WPR's procedural requirements. Only twice has Congress invoked the WPR to attempt to force a withdrawal — both times via concurrent resolutions vetoed by the president (Trump's 2019 Yemen-resolution veto being the most recent example).

The honest description is that the WPR has not effectively bound any president of either party. Its consultation provisions are routinely treated as a notification courtesy rather than a constraint; its 60-day clock is routinely either ignored or argued not to apply; and Congress has rarely shown the political will to force the issue, in part because forcing the issue carries political risk for the chamber doing the forcing.

Steel-manning both sides

The strongest executive-prerogative argument runs: The world after 1945 is one of fast-moving threats — nuclear weapons, terrorism, cyber-warfare — in which the deliberative pace of Article I declaration is incompatible with effective national defense. The president, as the only official with continuous access to intelligence and the diplomatic apparatus, is best situated to make rapid decisions about military deployment. The WPR, on this view, attempts to enforce a 19th-century template on 21st-century security challenges. AUMFs, broad as they are, reflect a congressional judgment that the executive should have flexibility in the conduct of long campaigns whose details Congress cannot anticipate.

The strongest congressional-prerogative argument runs: The Founders deliberately separated the war power from the executive precisely because they had seen what happens when one person can take a country to war. A president with continuous capacity to start small wars without congressional approval, and to escalate them at will, is exactly the constitutional architecture the Founders rejected. AUMFs from 2001 and 2002 cannot reasonably be read to authorize, twenty-five years later, drone strikes against organizations that did not exist in 2001 and that no member of Congress was voting on when those AUMFs passed. The proper response is to repeal both AUMFs, replace them with narrower and time-limited authorizations, and force the executive to come back to Congress when the authorization expires.

Reform proposals have circulated for years. Senators Tim Kaine (D-VA) and Todd Young (R-IN) have repeatedly introduced legislation to repeal the 1991, 2001, and 2002 AUMFs and replace them with narrower authorities; some bipartisan support exists, but no replacement has cleared both chambers as of 2026.

9.6 Executive Orders, Proclamations, and Memoranda

Below the level of treaty-making and war-making, presidents act primarily through three kinds of unilateral instruments.

Executive orders are presidential directives to the executive branch, published in the Federal Register, that have the force of law to the extent the president is acting within constitutional or statutory authority. They are numbered sequentially; we are now in the high four-digits since the numbering began in 1907.

Presidential proclamations are formal announcements, often ceremonial, but sometimes substantively important (tariffs under emergency authority, for example, are often issued by proclamation).

Presidential memoranda are similar to executive orders in legal effect but are not numbered, are not always published in the Federal Register, and have lower public visibility. They have grown as a fraction of unilateral presidential action in recent decades, in part because they attract less attention than EOs.

The empirical pattern: every modern president issues roughly 30–60 executive orders per year, with the volume varying by administration and by year. EO totals across administrations: Reagan 381 over 8 years; G.H.W. Bush 166 over 4; Clinton 364 over 8; G.W. Bush 291 over 8; Obama 276 over 8; Trump-1 220 over 4; Biden roughly 160 over 4; Trump-2 issued over 70 in his first month back in office (January–February 2025), an unusually high pace.

Famous and consequential executive orders, across the partisan spectrum:

• EO 9981 (Truman, 1948) — desegregated the U.S. armed forces. One of the most consequential civil-rights actions of the 20th century.
• EO 10730 (Eisenhower, 1957) — federalized the Arkansas National Guard and ordered them to enforce school integration in Little Rock.
• EO 11246 (LBJ, 1965) — required federal contractors to take affirmative action to ensure non-discrimination in employment. The textual basis for federal affirmative-action requirements until 2025.
• EO 12612 (Reagan, 1987) — imposed federalism principles on agency rulemaking.
• DACA Memorandum (Obama, June 2012) — directed the Department of Homeland Security to defer deportation of certain undocumented immigrants brought to the U.S. as children. Implemented as an exercise of prosecutorial discretion rather than a formal executive order, but with similar legal effect.
• EO 13769 (Trump, 2017) — the first version of the "travel ban," restricting entry from seven Muslim-majority countries. Substantially revised after litigation; ultimately upheld in Trump v. Hawaii (2018).
• EO 13985 (Biden, 2021) — directed federal agencies to advance racial equity in their programs. Reversed by Trump in 2025.
• EO 14172 (Trump, 2025) — restoring "names that honor American greatness," renaming geographic features and federal lands.
• EO 14173 (Trump, 2025) — ending federal affirmative-action requirements, revoking EO 11246.

The reversibility question — what can a successor undo? — is critical. As a rule, an executive order can be revoked by a subsequent executive order. Most of what one president does by EO, the next president can undo by EO. The Trump-Biden-Trump sequence on environmental, immigration, and labor regulation is largely a record of mutual reversals.

What cannot easily be undone:

1. Actions vested with reliance interests. When an EO induces individuals or businesses to take significant action in reliance on the policy (DACA recipients build careers; companies invest in compliance), withdrawal raises legal and political costs. Courts have sometimes held that the rescission must follow proper administrative procedure (DHS v. Regents of UC, 2020 — the DACA-rescission case).
2. Actions incorporated into statute or rule. When a successor administration codifies an EO into a final rule under the Administrative Procedure Act, undoing it requires another rulemaking.
3. Personnel and structural changes. An EO that fires a category of officials, or restructures an agency, may be reversible in form but not in fact — fired civil servants do not necessarily come back.

In short: executive orders are powerful but generally reversible, with key exceptions. They are not, as a category, equivalent to legislation, which is why their proliferation across administrations has been a structural feature of American politics in the 21st century rather than a partisan grievance.

9.7 Signing Statements and Executive Privilege

Two further presidential instruments deserve treatment.

Signing statements are statements issued by the president when signing a bill into law, indicating the president's interpretation of the bill or the president's intention not to enforce particular provisions on constitutional grounds. The practice goes back to James Monroe, but the modern aggressive use traces to the Reagan administration. George W. Bush issued more than 750 constitutional objections in signing statements over his eight years; he used signing statements heavily on national-security legislation, asserting that he would interpret various provisions consistently with his commander-in-chief authority. Obama narrowed the practice but continued it; Trump and Biden have used signing statements at lower rates.

The constitutional question is whether a signing statement that announces non-enforcement amounts to an unconstitutional line-item veto in disguise. The Supreme Court touched the question indirectly in Hamdan v. Rumsfeld (2006), where the Court rejected the Bush administration's interpretation of the Detainee Treatment Act despite a signing statement asserting otherwise. The lesson, doctrinally, is that signing statements do not bind courts; courts read the statute Congress actually passed.

Executive privilege is the doctrine that the president may withhold certain communications and information from Congress and the courts when disclosure would impair the executive's ability to function. The doctrine has constitutional roots — the separation of powers requires that the executive be able to deliberate confidentially — but it is not absolute.

The foundational case is United States v. Nixon (1974), in which the Supreme Court unanimously held that President Nixon had to comply with a subpoena for the Watergate tapes. Executive privilege exists, the Court held, but it is not absolute and it cannot defeat a subpoena in a criminal investigation when there is a demonstrated, specific need for the evidence. Nixon resigned 16 days after the decision.

Modern privilege disputes:

• Trump v. Mazars (2020). Three congressional committees subpoenaed Trump's personal financial records from his accountants and bankers. The Supreme Court declined to apply executive privilege to non-government records but adopted a four-factor test for evaluating congressional subpoenas of presidential records. The case was remanded; the records were eventually turned over after Trump left office.
• Trump v. Vance (2020). A New York grand jury subpoenaed Trump's tax returns from his accountants. The Supreme Court rejected absolute presidential immunity from state criminal subpoenas; the records were eventually produced.
• Mazars v. Trump and follow-on litigation. Multiple post-presidency disputes about which records were subject to privilege, with mixed outcomes.

The post-Nixon, post-Mazars doctrine, in summary: executive privilege exists; it is qualified, not absolute; criminal investigations and serious congressional oversight needs can overcome it; the contours are case-by-case and political negotiation usually resolves disputes before courts have to.

9.8 The Veto and the Legislative Endgame

Section 9.2.7 introduced the veto. A few empirical and political points deserve emphasis.

First, the threat of a veto often does more work than the veto itself. Most presidents issue Statements of Administration Policy (SAPs) before votes on major bills, signaling whether the president will sign or veto. Legislative leaders then re-engineer the bill, drop provisions, or abandon the effort. The veto is, in this sense, a bargaining tool whose visible use is rare because its deterrent effect is constant.

Second, the override threshold is essentially never met on partisan bills. Two-thirds of both chambers requires either bipartisan supermajority support or a chamber composition so lopsided that override is automatic. Neither has been the modern norm. The most overridden modern presidents are those who lost their party's majority in Congress and then vetoed extensively (Andrew Johnson, with 15 overrides of 29 vetoes, holds the record). Modern overrides are rare — Carter, Reagan, Clinton, and Obama each suffered single-digit override counts; Trump-1 had one override (the National Defense Authorization Act, 2020); Biden had no overrides; Trump-2 has none as of this writing.

Third, the line-item veto question is settled at the federal level by Clinton v. New York (1998). Congress would have to pass a constitutional amendment to give the president line-item-veto power, and proposals to do so have not advanced.

9.9 Presidential Immunity: Trump v. United States (2024)

In July 2024, the Supreme Court issued Trump v. United States, a 6–3 decision (Roberts writing for the majority; Sotomayor, Kagan, and Jackson in dissent) that established a new framework for criminal immunity of former presidents.

The framework, in summary:

• A former president has absolute immunity for actions taken pursuant to "core constitutional powers" — those exclusively committed to the president by Article II.
• A former president has at least presumptive immunity for actions within the "outer perimeter" of official responsibility. Whether this presumptive immunity can be rebutted, and how, was left to the lower courts to develop.
• A former president has no immunity for unofficial acts.
• Evidence of official-act conduct cannot be used in a prosecution for unofficial-act conduct.

The case concerned the federal indictment of former President Trump on charges related to the 2020 election and the events of January 6, 2021. After the Supreme Court's decision and Trump's reelection in November 2024, the federal prosecution was dismissed.

Steel-manning both sides

The strongest pro-immunity argument, as the majority articulates it: A president who fears post-office criminal prosecution for the close-call decisions of office cannot govern with the energy and decisiveness that Article II contemplates. The risk is not that a malicious president will exploit immunity; the risk is that a public-spirited president will second-guess every decision because the next administration of the opposing party might prosecute. The Founders did not include criminal-immunity language explicitly, but the structural design — separated powers, an executive with energy — implies it. The presumption tracks the Speech or Debate Clause's protection of legislators: not because legislators are above the law, but because the alternative — making official acts subject to retaliatory prosecution — is incompatible with separated powers.

The strongest anti-immunity argument, as the dissent and many critics put it: The Constitution says the president "shall be removed from Office on Impeachment for, and Conviction of, Treason, Bribery, or other high Crimes and Misdemeanors" — and the same Article specifically contemplates that a president removed by impeachment "shall nevertheless be liable and subject to Indictment, Trial, Judgment and Punishment, according to Law" (Art. I, §3). The constitutional text contemplates post-office criminal liability; reading immunity into the Constitution strains against that text. The structural argument cuts both ways: a president who knows that no domestic court can ever hold him criminally accountable for "official" actions has fewer incentives to stay within law. The framework as articulated leaves too much undefined: what is an "official act"? Are tweets directing rally-goers official? Are conversations with advisers about staying in office official? The doctrinal vagueness permits, in practice, far broader immunity than the Court's framework expressly grants.

The honest position, as of 2026, is that the doctrine is new, that the lower courts are now working out the application of the "outer perimeter" presumption, and that future administrations of either party will be operating under a substantially different liability environment than any prior president has faced. The ultimate political and constitutional implications will play out over years.

9.10 Impeachment and Removal

Article II, Section 4 provides: "The President, Vice President and all civil Officers of the United States, shall be removed from Office on Impeachment for, and Conviction of, Treason, Bribery, or other high Crimes and Misdemeanors."

The mechanics: The House impeaches by simple majority. The Senate tries the case (the Chief Justice presiding when the president is on trial), and conviction requires two-thirds. A convicted official is removed; the Senate may, by separate vote, also bar the official from future federal office.

The empirical record: Three presidents have been impeached; one (Nixon) resigned to avoid certain impeachment. Four presidential impeachments total in 235 years.

• Andrew Johnson (1868). Impeached for violating the Tenure of Office Act. Acquitted by one vote in the Senate. The episode left the Tenure of Office Act effectively unenforceable; it was repealed in 1887.
• Richard Nixon (1974). Resigned before House impeachment vote. Pardoned by Ford one month later.
• Bill Clinton (1998). Impeached for perjury and obstruction of justice arising from the Lewinsky investigation. Acquitted in the Senate (45–55 on perjury, 50–50 on obstruction).
• Donald Trump (December 2019, first impeachment). Impeached for abuse of power and obstruction of Congress arising from the Ukraine matter. Acquitted (48–52 on abuse, 47–53 on obstruction).
• Donald Trump (January 2021, second impeachment). Impeached for incitement of insurrection arising from January 6. Acquitted (57–43, ten short of the 67 needed).

No president has been convicted and removed. The two-thirds Senate threshold, in the era of party-line voting, makes conviction effectively impossible without a major bipartisan rupture.

The 25th Amendment (1967) provides an alternative path. Section 4 allows the vice president and a majority of Cabinet officers (or a body Congress designates) to declare the president unable to discharge the duties of office; the vice president becomes acting president; the president can contest the declaration; if the vice president and Cabinet majority persist, Congress decides by two-thirds vote of both chambers within 21 days. The 25th has never been invoked under Section 4, although discussions reportedly occurred during Reagan's second term, in connection with Trump after January 6, and reportedly during the final months of Biden's term in connection with concerns about the president's capacity. Section 4 is operationally a high bar; the political cost of invoking it would be severe.

The honest summary is that removal is, by design, very difficult. The Founders wanted impeachment to be available but rare, and the system they designed has produced exactly that.

9.11 The President as Party Leader

A constitutional power that does not appear anywhere in Article II, but that has become central to the modern presidency, is the president's role as leader of the president's political party.

The modern American party, particularly since the early 2000s, has organized itself around the president (when the party holds the White House) and around the most likely presidential nominee or recent former president (when it does not). The formal mechanisms are several:

• Endorsements in primaries. Presidents endorse candidates in House and Senate primaries; those endorsements move primary electorates substantially. Trump's endorsement record in Republican primaries (2018–2024) is the most studied recent example — endorsements correlate strongly with primary victories, although causation versus selection is genuinely contested in political-science research.
• Fundraising for the party. Sitting presidents headline major fundraisers; the resulting committees fund congressional and gubernatorial races. Both parties' nominees raise the largest single dollars at the presidential level and redirect to down-ballot races.
• Control of the national committee. The chair of the Republican National Committee (RNC) and the Democratic National Committee (DNC) is, in practice, selected by or in close consultation with the sitting president (or, for the out-party, the most recent nominee). The 2025 RNC operates as an arm of the Trump administration's political operation.
• Discipline of dissenters. Presidents have substantial leverage over members of their own party in Congress: through committee assignments (although those are formally controlled by congressional leadership), through campaign funding, through endorsement choices, and through public statements that move primary voters. Both parties have seen high-profile examples of incumbents losing primaries after presidential opposition (Republican incumbents 2018 and 2022; Democratic incumbents at lower rates but including some 2024 examples).

The "Trump Republican Party" is the most studied contemporary example, but the pattern is structural. Obama was the leader of the Democratic Party from 2009 to 2017; Biden inherited the role; the role passes to the most likely 2028 nominee through the primary. The Hatch Act limits some of the presidency's overt political activity (federal employees acting in official capacity cannot engage in partisan political activity), but the boundary between official duties and political leadership has been thin for a long time.

9.12 The President as Communicator

Theodore Roosevelt, in 1909, used a phrase that has stuck: the presidency, he said, was a "bully pulpit" — by which he meant a wonderful platform from which to speak to the country.

Subsequent presidents have used the platform in ways TR could not have imagined.

• FDR's fireside chats (1933–1944): radio addresses, conversational in tone, that reached households directly. The chats normalized the idea that the president speaks personally to the public.
• Eisenhower's televised press conferences (1955–1961): the first president to allow regular TV coverage of press interactions.
• Kennedy's television presidency (1961–1963): the first president whose public image was substantially shaped by television.
• Reagan's "Great Communicator" persona (1981–1989): polished, scripted, unusually effective at the prime-time televised address.
• Clinton's town-hall style (1993–2001): less formal, more interactive.
• Obama's online presidency (2009–2017): the first president whose communication strategy treated social media (initially) and YouTube/web video as primary channels alongside traditional networks.
• Trump's social-media presidency (2017–2021, 2025–): the first president to communicate substantially through unmediated short-form posts on social media. The volume and tone of presidential posting has reshaped both the news cycle and the institution of the press conference.
• Biden's lower-frequency presidency (2021–2025): fewer press conferences and fewer extended public communications than recent predecessors, partly by choice and partly under pressure as questions about cognitive capacity intensified.

The constant across these styles is the steadily increasing direct reach. A modern president can speak directly to a hundred million Americans without the mediation of a press corps. The political economy of that capacity is contested: it allows presidents to drive narratives without correction, and it allows the public to hear from the president without filter. Whether that is, on balance, good for democratic deliberation is one of the open questions of modern political science.

9.13 Modeling the Modern Presidency

Two theoretical frameworks dominate the political-science treatment of the modern presidency.

Richard Neustadt's "power to persuade" thesis (1960). In Presidential Power, Neustadt argued that the formal powers of the presidency are insufficient for effective governance, and that the working power of a president is the power to persuade — to convince other actors (Congress, the bureaucracy, the public, foreign leaders) that what the president wants is what they should want too. Neustadt's claim was empirical: Truman's experience as president, and the experience of every postwar president Neustadt studied, showed that a president who relied on command without persuasion failed. Persuasion required reputation among Washington insiders ("professional reputation") and standing with the public ("public prestige"); presidents who lost both lost the practical capacity to govern.

Stephen Skowronek's "political time" framework (1993). In The Politics Presidents Make, Skowronek argued that presidents inherit a political-regime context — a dominant ideology, a coalition of interests, a set of policy commitments — and that what a president can accomplish depends on whether the regime is robust or vulnerable, and whether the president is affiliated with or opposed to it. Skowronek identified four positions: reconstructive (opposed to a vulnerable regime, building a new one — Lincoln, FDR, Reagan), articulation (affiliated with a robust regime — TR, LBJ), preemptive (opposed to a robust regime — Eisenhower, Clinton), and disjunctive (affiliated with a vulnerable regime — Hoover, Carter). The framework offers a way to compare presidents across very different historical periods.

Steel-manning both

Neustadt's strength: the framework forces attention to what presidents actually do day-to-day, which is bargain, persuade, threaten, plead, reward, and trade. It captures the texture of the office in a way more formal frameworks miss.

Skowronek's strength: the framework explains why some presidents who appeared successful (LBJ on Vietnam) became failures, and why some who appeared weak (Eisenhower) accomplished more than contemporaries credited. The structural variables — regime context — sometimes matter more than individual presidential skill.

Neustadt's limitation: the bargaining model fits the mid-20th-century presidency better than it fits the polarized 21st-century context, where persuasion across party lines has become rare and presidential power increasingly flows through unilateral action and partisan loyalty rather than persuasion.

Skowronek's limitation: the regime-cycle framework can become deterministic if applied mechanically. Presidents have agency; structural position matters but does not dictate outcomes.

The honest pedagogy is to use both. Neustadt for the texture of presidential decision-making; Skowronek for the structural context that shapes what decisions are available.

9.14 Presidential Popularity

Political-science research on presidential approval has identified several recurring patterns.

The honeymoon. Most newly elected presidents enjoy high approval in their first months, often higher than any other point of their tenure. The post-election effect typically dissipates by the end of the first year as the president takes politically costly positions.

The rally-round-the-flag effect. Major foreign-policy crises often produce sharp short-term increases in presidential approval, regardless of whether the policy response succeeds. The first Gulf War boosted George H. W. Bush above 80 percent; 9/11 boosted George W. Bush above 90 percent; the Russian invasion of Ukraine produced a smaller bump for Biden in early 2022. The bumps generally fade within months.

Durable polarization since approximately 2009. Beginning roughly with Obama's first term, presidential approval ratings have shown an unprecedented partisan gap: out-party identifiers approve of the president at unusually low rates, and the gap has not closed during foreign-policy rallies. Both Obama and Trump-1 saw approval gaps between Democratic and Republican identifiers exceeding 70 percentage points throughout their terms — a gap unprecedented in the polling era. Biden showed similar patterns. Trump-2's early-2025 approval shows the same structural gap.

The 40 percent floor. Both Obama and Trump-1, and now Biden and (so far) Trump-2, have shown a remarkable feature: a "floor" of approximately 40 percent approval that has been hard for political opponents to push below regardless of events. This reflects the durable partisan attachment of approximately that share of the country to its party's incumbent president. It is structurally different from the pre-2009 era, when several presidents (Carter, G.W. Bush in his second term) saw approval drop into the high 20s or low 30s.

Biden's specific trajectory. Biden took office in January 2021 with approval near 55 percent. By autumn 2021 — after the chaotic withdrawal from Afghanistan and the rise of inflation — approval dropped below 45 percent and remained below 50 percent for the remainder of his term. By 2024, with concerns about cognitive capacity intensifying, approval dropped further; Biden chose not to seek reelection in July 2024.

Trump's approval, second term. Trump returned to office in January 2025 with the typical honeymoon-related uplift; by mid-2025 his approval had settled into the high 40s to low 50s, similar to but slightly higher than his first-term levels. As of early 2026 the durable-polarization pattern continues to hold.

The political-science honest summary: presidential approval is a useful indicator, but it must be read against the durable-polarization context. A 47 percent approval rating in 2026 means something quite different from a 47 percent rating in 1976.

9.15 The ACA Anchor Returns

Throughout this textbook, we are using the Affordable Care Act of 2010 as a running example of how the institutions interact. Chapter 7 set up the legislative history. Chapter 8 detailed the procedural mechanics. The presidency chapter has its own contributions to make.

The ACA was Obama's signature domestic-policy achievement. Its design relied on executive implementation: the Department of Health and Human Services issued thousands of pages of regulations to implement the statute; the IRS issued rules on the individual-mandate tax penalty; the Department of Labor issued rules on employer responsibilities. Each of those implementing actions was, in part, a presidential exercise of Take Care Clause authority — the president, through subordinate executive officers, executing the law Congress passed.

When the Trump-1 administration came in (2017), it could not repeal the ACA (Congress narrowly failed to do so in July 2017, with Senator John McCain casting the decisive vote against repeal). But it could shape implementation through executive action: it issued executive orders narrowing enforcement of the individual mandate, expanded short-term-health-plan options that were not subject to ACA requirements, and took various administrative actions to weaken the law.

The Biden administration (2021–2025) reversed many of those executive actions, expanded ACA enrollment subsidies through the American Rescue Plan and the Inflation Reduction Act, and made record enrollment a priority. The Trump-2 administration's early 2025 actions have included executive orders to scale back several Biden-era ACA implementation choices, with extensive litigation under way.

The ACA story illustrates the modern dynamic. A statute Congress passed in 2010 has been implemented through three Republican-Democrat-Republican administrations with substantially different executive priorities. The statute has not changed in any major respect; its operation has changed substantially through executive action. This is, increasingly, how American policy works: legislation establishes the framework; presidents shape what the framework does in practice.

9.16 The Honest Synthesis

Where does this leave us?

The American presidency in 2026 is more powerful than the office described in Article II — substantially more powerful in war-making, regulatory direction, and unilateral domestic action than the Founders contemplated. The expansion has been bipartisan; both parties have used the powers when they held the office and criticized them when they did not. The expansion has been driven by structural forces — congressional delegation, the rise of the administrative state, the post-1945 security environment, party polarization, media affordances — that no single administration created and no single administration can reverse alone.

The constitutional architecture that constrains the office is real but contested. Some constraints have eroded (the WPR has not bound any modern president; the recess-appointment clause has been repeatedly fought over; for-cause removal protections for executive officers have been narrowed by Seila Law and Collins). Some constraints have intensified (judicial scrutiny of major-questions agency action, post-2022 West Virginia v. EPA; congressional appropriations leverage when divided government persists). The doctrinal landscape after Trump v. United States on official-acts immunity is genuinely new and not yet fully worked out.

The political constraints are real but variable. A unified-government president can do more than a divided-government president; both Trump in 2017 and Biden in 2021 found that unified government still imposed significant constraints (the Senate filibuster being the most consistent). A first-term president generally has more political room than a second-term president; Obama in 2009–2010 had more freedom than in 2015–2016. A president with high approval has more leverage than a president with low approval.

What this means for an informed citizen: the campaign-trail picture of presidential power is misleading in the direction of overstatement. Presidential candidates promise more than the office can in fact deliver, and they routinely fail to deliver. But the "powerless presidency" picture — the version that emphasizes only constraints — is also misleading, in the direction of understatement. Modern presidents have, and use, enormous unilateral capacity. The honest picture is the one in this chapter: a powerful office with real constraints, an expansion trajectory that is structural and bipartisan, and a set of doctrinal questions that are still being worked out, in real time, in the courts.

In Chapter 10, we turn to the people and institutions that make the presidency operational: the vice president, the Cabinet, and the Executive Office of the President. In Chapter 11, we look at the bureaucracy the president nominally leads — and the long-running fight over how responsive to political control that bureaucracy should be. The presidency is not an island; it is the most visible point in a much larger executive system. The chapters that follow will give you the rest of the picture.