On the morning of January 20, 2001, in the final hours before George W. Bush took the oath on the Capitol steps, William Jefferson Clinton sat in the White House signing clemency grants. The pen moved across one hundred and forty names. Among them, near the bottom of a list that aides had assembled in a scramble of late-night phone calls and faxes from defense lawyers, was a fugitive commodities trader named Marc Rich. Rich had fled to Switzerland in 1983 rather than face a sixty-five-count federal indictment for tax evasion, racketeering, and oil trades with Iran during the hostage crisis. He had spent seventeen years on the run. He had never stood trial. And in the span of a signature, the forty-second president erased the legal jeopardy of a man most of the Justice Department had never been consulted about.

The reaction was immediate and bipartisan in its disgust. Jimmy Carter, a fellow Democrat, called the Rich pardon disgraceful. Congressional committees opened investigations. Federal prosecutors in the Southern District of New York, the office that had built the case against Rich, began examining whether the grant itself broke any laws. It did not, and that was the point that should have unsettled everyone. The pardon of Marc Rich was constitutional. It was, by the plain text of the document that governs the office, entirely within the president’s authority. A man could buy nothing from the Treasury, command no army on his own say-so, and appoint no judge without the Senate. But he could, on his last morning in office, with no review, no appeal, and no obligation to explain himself, lift a federal indictment off a fugitive whose ex-wife had given more than a million dollars to Democratic causes and the Clinton library fund.

That is the strange machinery this article is about. The pardon is the one place in the Constitution where the founders built a king-shaped hole and then handed every president the key. It is the most unilateral power the office contains, more discretionary than the veto, less checkable than the executive order, and older than the party system, the cabinet as we know it, or the idea that the presidency would ever be anything other than a constrained republican magistracy. To follow the pardon from the Constitutional Convention to Clinton’s last morning is to watch a single clause do work its authors never imagined, accumulate uses they never sanctioned, and resist every attempt to fence it in. The clause has not changed since 1787. Almost everything else about how it gets used has.

The Clause Itself: Twenty-Eight Words and a Convention Fight

The grant lives in Article II, Section 2 of the Constitution, tucked between the commander-in-chief clause and the treaty power. The president, it reads, “shall have Power to grant Reprieves and Pardons for Offenses against the United States, except in Cases of Impeachment.” Twenty-eight words. Two carve-outs, and only two. The offense must be federal, which means a president cannot reach into a state court and undo a conviction for a state crime. And the power cannot be used to undo an impeachment, which keeps a president from rescuing himself or an ally from removal by Congress. Everything else is open. There is no requirement of a hearing. No requirement of remorse from the recipient. No requirement that the Justice Department recommend the grant, or even know about it. No standard of evidence, no burden of proof, no appeal. The clause does not even require that the recipient have been convicted, or charged. A president may pardon a person for conduct that has not yet produced an indictment, which is precisely what makes the so-called preemptive pardon possible.

That openness was not an accident, and it was not unexamined. The Framers argued about it. At the Philadelphia convention in the summer of 1787, the pardon power drew sustained objection, and the loudest objector was the man later canonized as the Constitution’s principal architect. James Madison worried about exactly the scenario that would haunt the office two centuries later: a president using the pardon to shield his own confederates, or worse, to cover crimes in which he himself was complicit. Edmund Randolph of Virginia moved to bar the president from pardoning cases of treason specifically, arguing that treason was the one crime where the president’s own interest might be most entangled, since a chief executive could conceivably be the head of the very plot the treason laws were written to punish. The objection had teeth. If a president organized a rebellion, the logic ran, he could simply pardon his foot soldiers and dissolve the prosecution from the top.

The convention rejected Randolph’s amendment, and the man who argued it down was George Mason’s opposite on this question and the author of the document’s most muscular defense of unified executive authority. The case that carried the day rested on a wager about crisis. In moments of insurrection, the argument went, a single decision-maker who could offer a credible, immediate promise of mercy might restore peace faster than any court or legislature could. A pardon held in reserve was a tool of pacification. James Wilson, who did much of the structural drafting, made the point that the impeachment carve-out was itself the safeguard: a president who abused the pardon to cover treason would have committed an impeachable offense in doing so, and Congress could remove him for it. The pardon could not save the president from impeachment, even if it could save his accomplices from prosecution. That was thought to be enough.

Alexander Hamilton built the public argument for it in Federalist No. 74, one of the shortest and least-read papers in the entire collection, and the one that did the most to shape how the clause would be understood. Hamilton’s defense had two prongs. The first was a theory of mercy: criminal codes, he wrote, are necessarily severe, and without an “easy access to exceptions” the administration of justice would wear “a countenance too sanguinary and cruel.” Someone had to hold the dispensing power, the authority to soften the law’s edge in the individual case, and Hamilton argued it should rest in a single person because a single conscience feels the weight of the decision in a way a committee never will. A body of men, he wrote, would shelter behind one another and grant clemency too freely or refuse it too coldly; one man, accountable and exposed, would be more scrupulous. The second prong was the crisis theory the convention had already accepted. In seasons of insurrection, Hamilton argued, “a well-timed offer of pardon to the insurgents or rebels may restore the tranquillity of the commonwealth,” and the moment for such an offer might be so fleeting that requiring legislative concurrence would let it pass. Speed and singularity were features, not bugs.

Read in 2001, or in any of the controversies since, Federalist No. 74 reads like a document that anticipated everything except the thing that actually went wrong. Hamilton foresaw the rebel-pacifying pardon, and presidents would use it exactly as he predicted. He foresaw the mercy pardon, the softening of a sentence too harsh for the crime, and that too became routine. What he did not foresee, or chose not to dwell on, was the pardon as a private favor, the grant that serves not the commonwealth and not mercy but the president’s friends, donors, and family. Madison had glimpsed that danger and lost the argument. The clause that emerged was, by design, almost ungovernable, defended on the theory that a single conscience would govern itself.

Washington Sets the Template: Mercy as Statecraft

The first president to use the power used it exactly as Hamilton had advertised. In 1794, western Pennsylvania erupted over a federal excise tax on distilled spirits, the Whiskey Rebellion, the first serious armed test of whether the new government could enforce a law on its own citizens. George Washington did something no president would do again: he personally took the field, riding at the head of a militia force of roughly thirteen thousand men assembled from four states, a number larger than most of the armies he had commanded during the Revolution. The show of force worked. The rebellion collapsed without a pitched battle. The federal government had demonstrated that it could compel obedience.

Then Washington demonstrated the other half of the equation. Of the men rounded up and brought east for trial, two were convicted of treason and sentenced to hang: John Mitchell and Philip Weigel. In 1795, Washington pardoned both. His stated reasoning tracked Hamilton’s crisis theory almost word for word. The point of the prosecutions had been to vindicate federal authority, and that had been accomplished by the convictions themselves. Executing two confused frontier farmers, one of whom witnesses described as barely competent, would gain the government nothing and cost it the goodwill of a region it needed to reconcile. Mercy, deployed after the demonstration of strength, was the tool that healed the wound the rebellion had opened. Washington had shown the country that the government could punish; the pardon showed that it would not be vindictive. The sequence, force first and clemency second, became the template for what the pardon could do at the level of statecraft. It was a deliberate act of national repair, and it established that the pardon power would be used not merely for individual mercy but as an instrument of political reconciliation. The healing pardon, the grant that closes a national rupture rather than serving a single petitioner, traces its lineage directly to those two reprieves.

Washington’s successor confronted a different kind of test and answered it with a different kind of pardon. John Adams, near the end of his single term, faced the Fries Rebellion, another tax revolt, this one in eastern Pennsylvania in 1799, where a Pennsylvania-German auctioneer named John Fries led resistance to a federal property tax. Fries was twice convicted of treason and twice sentenced to death. Adams, over the objection of his own cabinet, pardoned Fries and the other condemned men in 1800. The cabinet wanted blood; Adams calculated that hanging tax protesters would do more to inflame the country than to settle it, and he extended a general amnesty to everyone involved. The decision contributed to the fracture between Adams and the Hamiltonian wing of his own party, the same fracture that helped cost him reelection. It was an early demonstration of a pattern that would recur across two centuries: the pardon that is right on the merits is frequently the pardon that is expensive in politics, and presidents who use it for reconciliation often pay for the courage at the ballot box. Adams’ willingness to put national calm over party vengeance belongs in the same family as his refusal, the same year, to take the country into a popular war with France, a country-over-party instinct that defined his presidency even as it ended it.

Jefferson Turns the Pardon Into a Repudiation

If Washington made the pardon an instrument of healing, Thomas Jefferson made it an instrument of judgment, a way for one administration to render a verdict on the one before it. The vehicle was the Alien and Sedition Acts, the 1798 laws under which the Adams administration had prosecuted and jailed newspaper editors and at least one sitting congressman for the crime of criticizing the government. Matthew Lyon, a Vermont representative, had been convicted of sedition for accusing Adams of “unbounded thirst for ridiculous pomp” and other ordinary political invective; he was fined and imprisoned, and he ran his reelection campaign from a jail cell, winning. James Thomson Callender, a scurrilous pamphleteer who would later turn on Jefferson and break the Sally Hemings story, had been convicted under the same law.

Jefferson came into office in 1801 regarding the Sedition Act as not merely bad policy but unconstitutional, a violation of the First Amendment that no court had yet struck down and that had in fact expired by its own terms the day before his inauguration. He could not undo the law, but he could undo its effects. He pardoned those convicted under it and ordered fines remitted where he could. His reasoning was explicit and constitutional rather than merely merciful. He argued that he was bound to consider the law a nullity, “as absolute and as palpable as if Congress had ordered us to fall down and worship a golden image,” and that pardoning its victims was his duty as an officer sworn to the Constitution, not a discretionary act of grace. This was a new use. Washington’s pardons had served the present; Jefferson’s served the past, retroactively repudiating the judgments of a prior administration and asserting that the executive’s own reading of the Constitution could override a verdict the courts had let stand. It planted a principle that would echo through later debates about whether a pardon can function as a statement that the underlying law itself was illegitimate, a question that would resurface in the amnesty proclamations of the twentieth century. The conciliatory spirit Jefferson announced in his first inaugural, the famous insistence that “we are all Republicans, we are all Federalists,” found one of its sharpest practical expressions in the pardon power, even as the pardons themselves drew a hard line under the Federalist prosecutions he meant to repudiate.

Lincoln and the Mass Mercy of a Civil War

No president before or since used the clemency power on the scale Abraham Lincoln did, and none used it with such varied purpose. Lincoln operated the pardon on two registers at once. At the level of national policy, he issued the Proclamation of Amnesty and Reconstruction on December 8, 1863, the document that put the pardon at the center of his plan to end the war and rebuild the Union. The proclamation offered a full pardon, with restoration of property except in slaves, to any Confederate who would take an oath of future loyalty to the United States and accept emancipation. It carved out exceptions, high Confederate officials and certain categories of officers, but for the rank and file it was a standing offer of amnesty in exchange for an oath. This was the rebel-pacifying pardon Hamilton had theorized, deployed at industrial scale. Lincoln paired it with his ten percent plan, the proposal that a seceded state could begin reconstituting a loyal government once a number equal to ten percent of its 1860 voters had taken the oath. The amnesty was not an afterthought to Reconstruction; it was the mechanism. Lincoln was trying to make rejoining the Union cheaper than staying out of it, and the pardon was the price discount.

At the individual level, Lincoln was famous, and within the army infamous, for his willingness to spare condemned soldiers. The cases came across his desk constantly: deserters, sentries who fell asleep on duty, boys who broke under fire, men sentenced to be shot for cowardice. Lincoln looked for reasons to commute, and his generals complained that his leniency was undermining discipline. He developed informal categories of cases he would always pardon, including the soldier too young to be held fully responsible and the deserter who had a credible story of going home to a sick family. His son Robert Todd Lincoln, looking back, estimated that his father had granted clemency in thousands of cases over the course of the war, a figure historians treat as plausible given the volume of military justice the conflict generated. The image of Lincoln staying up late to find a way to save a frightened private became part of the mercy mythology of his presidency, and unlike most such myths it rests on a documentary record of remitted sentences. The same wartime instinct toward leniency coexisted with the hardest exercises of executive power in American history, the suspension of habeas corpus and the prosecution of a total war, and the pardon was where Lincoln let the other side of his temperament show.

What Lincoln demonstrated was the pardon’s elasticity. The same clause that could heal a single condemned teenager could also be the legal instrument of a national reconciliation program affecting hundreds of thousands. Lincoln stretched it in both directions at once, and in doing so he set up the bitter fight that followed his death, because the amnesty program he designed as a tool of reunion would, in the hands of his successor, become a weapon that Congress saw as a tool of betrayal.

Andrew Johnson and the Pardon as a Battlefield

Andrew Johnson inherited Lincoln’s amnesty machinery and turned it into the central front of the Reconstruction war between the executive and Congress. Where Lincoln had used conditional pardons as a calculated instrument inside a broader Union strategy, Johnson used them lavishly and, in the eyes of the Radical Republicans, recklessly, to restore the antebellum Southern ruling class to power as quickly as possible. His Amnesty Proclamation of May 1865 offered pardon to most former Confederates who took a loyalty oath, while reserving fourteen excepted classes, including high officials and wealthy planters whose taxable property exceeded twenty thousand dollars, who had to apply to the president personally. Johnson then proceeded to grant those individual applications by the thousands. By the time he left office he had issued pardons to a substantial majority of the excepted Confederates who sought them, and in December 1868, on Christmas Day, he issued a final universal amnesty extending unconditional pardon to essentially all remaining former Confederates, including those previously excluded.

To Johnson, this was the fulfillment of Lincoln’s policy of malice toward none. To the Radical Republicans in Congress, it was the deliberate dismantling of the war’s results, restoring to power the men who had led the rebellion and who were, in the same months, enacting the Black Codes to re-subordinate the freedmen. The pardon power became one of the flashpoints in the broader struggle that ended in Johnson’s impeachment. Congress could not directly stop the pardons, since the clause put them beyond legislative reach, but the lavishness of Johnson’s clemency fed the conviction that the president was an obstacle to be removed, not a partner to be negotiated with. The impeachment ultimately turned on the Tenure of Office Act and Johnson’s firing of Secretary of War Edwin Stanton, but the pardon controversy was part of the atmosphere of distrust that made removal thinkable. Johnson survived in the Senate by a single vote, and the episode established something important about the clause: there was no constitutional remedy for a pardon Congress hated, short of impeaching the president for the totality of his conduct, and even that mechanism failed to reach the clemency itself. The pardon was, functionally, beyond recall. The veto fights and pardon fights of the Johnson years together taught Congress that a determined president wielding his enumerated powers could grind the legislature’s program to a halt, a lesson about the limits of legislative supremacy that the Reconstruction Congresses absorbed the hard way.

Ulysses Grant, succeeding Johnson, continued the work of winding down Civil War-era prosecutions through more measured clemency, and he confronted the opposite problem in the South: not too little punishment of former Confederates but the rise of the Ku Klux Klan and the collapse of federal will to protect the freedmen. Grant’s decision to prosecute the Klan under the Enforcement Acts represented the punitive instinct that Johnson’s amnesties had abandoned, and the contrast between the two presidencies on the question of who deserved mercy and who deserved the full weight of federal law marks one of the genuine hinge points of Reconstruction. The pardon power sat at the center of that divergence, a single clause that one president used to forgive the planter class and another declined to use as he turned federal prosecutors loose on the Klan.

The Routine Century: Clemency Becomes Bureaucracy

For most of the period between Reconstruction and the Second World War, the pardon power did its quiet, unglamorous work, and the volume was enormous. This is the part of the story that the headline controversies obscure. Across American history the clemency power has been exercised somewhere in the neighborhood of twenty thousand to twenty-one thousand times, and the overwhelming majority of those grants were not Marc Rich or even Whiskey Rebels. They were ordinary acts of mercy and sentence adjustment, granted to people whose names no one remembers, processed increasingly through a bureaucratic apparatus rather than decided at the president’s elbow.

The institutionalization mattered. In 1865 the attorney general’s office began taking a formal role in processing clemency petitions, and over the following decades the function migrated into what became, in 1894, the Office of the Pardon Attorney within the Department of Justice. The pardon attorney’s office reviewed petitions, investigated cases, solicited the views of prosecutors and sentencing judges, and forwarded recommendations to the president through the attorney general. This created a two-track reality that persists to this day. On the formal track, a petitioner waited the required years after completing a sentence, filed an application, submitted to investigation, and received a recommendation that the president usually but not always followed. On the informal track, a president could ignore the entire apparatus and grant a pardon to anyone he chose, with or without the pardon attorney’s knowledge, which is exactly the track Clinton used for Marc Rich and which Justice Department veterans found so alarming precisely because it bypassed the institution built to discipline the power. The clause itself does not require the formal track. The Office of the Pardon Attorney exists by administrative choice, not constitutional command, and a president remains free to route around it whenever he likes.

The raw numbers from this era dwarf modern practice, and the comparison is the single most counterintuitive fact about the institution’s trajectory. Grover Cleveland, William McKinley, Theodore Roosevelt, William Howard Taft, and Woodrow Wilson each granted clemency in the hundreds or thousands. The pardon was a working tool of the justice system, used routinely to correct disparities, reward rehabilitation, and clear the records of people who had served their time. The modern image of the pardon as a rare, almost exotic act, deployed mainly in scandal, is a distortion produced by the collapse in volume during the late twentieth century. For the first hundred and fifty years of the republic, the pardon was less a scandal machine than a mercy bureaucracy, grinding through thousands of unremarkable cases that never made a newspaper.

Truman, Carter, and the Amnesty of War Resisters

The twentieth century revived the mass-amnesty tradition twice, both times to settle the human aftermath of a war. Harry Truman, in December 1947, issued a proclamation pardoning roughly fifteen hundred men who had been convicted of violating the Selective Training and Service Act during the Second World War, the conscientious objectors and draft resisters who had refused to serve. Truman did not extend the amnesty to everyone; he appointed a three-member Amnesty Board to review the cases of the more than fifteen thousand men convicted of draft violations, and the board recommended pardons for only about ten percent of them, the ones whose objection appeared genuinely rooted in conscience rather than convenience or disloyalty. The selective character of the Truman amnesty is worth noting because it shows a president using the pardon as an instrument of moral sorting, distinguishing the principled objector from the ordinary evader, a distinction the clause permits him to draw on whatever basis he chooses.

Jimmy Carter went further and faster, and made the war-resister amnesty one of the first acts of his presidency rather than a late reckoning. On January 21, 1977, his first full day in office, Carter issued a proclamation granting an unconditional pardon to the men who had violated the Selective Service Act during the Vietnam era, primarily by evading the draft, a population that affected roughly ten thousand men directly and far more in its symbolic reach. Carter framed it not as forgiveness of a crime but as an act of national reconciliation after a war that had torn the country apart, the same logic Washington had applied to the Whiskey Rebels and Lincoln to the Confederates. The grant was controversial in a way Truman’s had not been. Veterans’ organizations objected that pardoning the men who fled to Canada while others died in the jungle was a slap at those who had served. Carter narrowed the scope to draft evaders and pointedly did not extend it to deserters, who faced a separate and more limited discharge-upgrade program, drawing a line that preserved some distinction between refusing to go and abandoning a unit in the field. The Carter amnesty closed the legal chapter of Vietnam-era draft resistance, but the political wound it touched never fully healed, and the episode demonstrated again that the reconciliation pardon, however justified on the merits, almost always costs the president who issues it among the constituency that paid the war’s price.

Ford and the Pardon That Defined the Modern Power

Everything before September 8, 1974, was prologue to the single act that did more than any other to shape how Americans understand the pardon. On that Sunday morning, Gerald Ford went on television and announced that he was granting Richard Nixon “a full, free, and absolute pardon” for all federal crimes Nixon “has committed or may have committed or taken part in” during his presidency. Proclamation 4311 covered the entire Watergate period in a single sweep, lifting the threat of indictment off a former president who had resigned only a month earlier and who had not been charged with anything. It was, and remains, the most consequential individual pardon in American history, and it crystallized two features of the power that had always been latent in the clause but had never before been displayed so starkly.

The first feature was the preemptive pardon. Nixon had not been indicted. He had not been tried or convicted of any crime. Ford’s pardon reached forward to cover offenses that existed, at that moment, only as the subject of a special prosecutor’s ongoing investigation. This confirmed that the clause’s silence about conviction was meaningful: a president could pardon a person for conduct that had never been adjudicated, foreclosing prosecution before it began. The legal grounding came from an 1866 Supreme Court case, Ex parte Garland, in which the Court had described the pardon power as unlimited except by the impeachment exception and had said it could be exercised at any time after the commission of the offense, “either before legal proceedings are taken, or during their pendency, or after conviction and judgment.” Ford’s lawyers relied on exactly that language. The Nixon pardon was the most prominent test of the principle, and it held. A president could erase a crime that no court had ever named.

The second feature was the political cost, and Ford paid it in full. The pardon detonated his presidency. His approval rating dropped from the seventies into the forties almost overnight. The press secretary Jerald terHorst resigned in protest. The widespread suspicion, never proven and consistently denied by Ford, was that a corrupt bargain had been struck, that Nixon had agreed to resign in exchange for a promise of immunity from his designated successor. Historians who have examined the question, including the scholarship focused specifically on the Ford controversy, generally conclude that no explicit deal existed, but that the appearance of one was unavoidable given the sequence of events, and that the appearance was politically fatal regardless of the truth. Ford himself maintained that he acted to spare the country the spectacle of a former president on trial, to let the long national nightmare end, and to free his own administration to govern rather than spend years consumed by Nixon’s legal fate. Years later, even many of his critics came around to the view that the pardon, whatever its costs, had served the country by closing the Watergate wound, and in 2001 Ford received the Profile in Courage Award for the decision. The reappraisal is real, but it took a quarter-century, and it did not save his 1976 campaign, which he lost to Carter by a margin small enough that the Nixon pardon plausibly made the difference.

The Ford pardon matters most for what it institutionalized. It established the preemptive pardon as a usable tool, demonstrated that the power could be deployed at the highest level of national politics to resolve a constitutional crisis, and embedded in the political memory the idea that a president could and would use clemency to protect the man at the top of the system rather than the forgotten private at the bottom. Lincoln had used the pardon to spare frightened teenagers; Ford used it to spare a president. The trajectory of the next several decades would run in Ford’s direction, not Lincoln’s. The decision belongs in the same conversation as the question historians still debate about what a different choice would have meant for the country and for Ford’s own political survival, and it remains the reference point against which every controversial pardon since has been measured.

Reagan, Bush, and the Pardon as a Shield

If Ford used the pardon to close a crisis, the presidents who followed used it increasingly to protect their own, and the line between reconciliation and self-interest grew harder to locate. Ronald Reagan’s clemency record was modest in volume and mostly unremarkable, but it included one grant that pointed toward the future. In 1989, on his way out of office, Reagan pardoned George Steinbrenner, the owner of the New York Yankees, who had been convicted in 1974 of making illegal corporate contributions to Nixon’s reelection campaign and of obstruction of justice for trying to cover it up. Steinbrenner was a wealthy, well-connected figure whose offense was a campaign-finance crime, and the pardon restored his civil rights and cleared his record. It was the kind of grant that, a century earlier, would have passed unremarked among thousands of routine clemencies, but in the low-volume modern era, where pardons had become rare and therefore scrutinized, it stood out as a favor to a prominent man with the right connections.

George H.W. Bush delivered the grant that, before Marc Rich, was the modern benchmark for a self-protective pardon. On Christmas Eve 1992, six weeks after losing his reelection bid and four weeks before leaving office, Bush pardoned six figures from the Iran-Contra affair, the scandal in which Reagan administration officials had secretly sold arms to Iran and diverted the proceeds to the Nicaraguan Contras in defiance of a congressional ban. The most consequential of the six was Caspar Weinberger, Reagan’s former defense secretary, who was scheduled to stand trial weeks later on charges of lying to Congress. Weinberger’s trial threatened to put Iran-Contra back in the headlines and, more dangerously for Bush, to expose what Bush himself, as Reagan’s vice president, had known and when. The independent counsel, Lawrence Walsh, reacted with open fury, declaring that the pardons completed a cover-up and noting pointedly that Bush had been a subject of the same investigation. Walsh’s charge was that Bush had pardoned witnesses whose testimony might have implicated him, a use of the power that edged toward exactly the self-shielding scenario Madison had warned against at the convention. Whether or not Bush acted to protect himself, the timing, on Christmas Eve, after an electoral defeat, on the eve of a trial that could have damaged him, established the template of the last-minute defensive pardon and demonstrated that the clause offered no recourse against it. Walsh could rage; he could not undo a single signature.

By the time Bush left office, the modern pattern was set. The pardon had become, in its most visible form, an instrument deployed at the end of a presidency, often after electoral defeat, frequently to benefit the president’s allies, and beyond the reach of any check but public opinion and the historical record. The mercy bureaucracy still ground on in the background, processing its thousands of forgotten petitions, but the public face of the power had become the eleventh-hour grant to a connected insider. Clinton would take that pattern to its logical and most notorious conclusion.

Clinton’s Last Morning and the Marc Rich Affair

Return to the scene that opened this article, because it is the case where every thread of the institution’s evolution converges. Bill Clinton’s clemency activity on January 20, 2001, was unusually heavy even by the standards of a departing president. He issued one hundred and forty pardons and thirty-six commutations in his final hours, a burst of last-day grants that included reasonable acts of mercy, defensible commutations of disproportionate drug sentences, and a handful of grants that ranged from questionable to indefensible. His own half-brother Roger Clinton received a pardon for an old cocaine conviction. Two former business associates connected to the Whitewater investigation were among the recipients. But the grant that dominated everything was Marc Rich.

The facts of the Rich pardon assembled almost every concern the institution had ever raised. Rich was a fugitive, not a petitioner who had served his sentence and waited the required years; he had spent seventeen years evading American jurisdiction in Switzerland precisely to avoid the trial his pardon now made unnecessary. The Office of the Pardon Attorney, the bureaucratic mechanism built to discipline the power, had been bypassed almost entirely; the Justice Department’s own prosecutors, the people who had built the sixty-five-count indictment, were not meaningfully consulted, and several learned of the pardon from the news. The petition had reached Clinton through a back channel of private lawyers and political intermediaries, including Rich’s attorney Jack Quinn, a former Clinton White House counsel who knew how to reach the president directly. And the financial dimension was unavoidable: Rich’s ex-wife, Denise Rich, was a prominent Democratic donor who had given more than a million dollars to Democratic candidates and causes, including substantial contributions toward the Clinton presidential library, and who had personally lobbied for the pardon. The combination of a fugitive recipient, a bypassed Justice Department, a back-channel petition, and a major donor connection produced a scandal that consumed the early weeks of the Bush administration and prompted congressional hearings and a federal investigation that ultimately found no prosecutable crime.

That last fact is the crucial one. The investigation closed without charges because the pardon, however unseemly, was constitutional. There is no law against pardoning a fugitive. There is no law requiring a president to consult the Justice Department. There is no law against pardoning the ex-husband of a donor, even a large one, absent proof of an explicit quid pro quo that no investigator could establish. Clinton offered a labored public defense, arguing that the Rich case had merit on the underlying tax questions and that respected legal figures had vouched for the grant, but the defense persuaded almost no one, and Clinton’s own former aides and allies regarded the pardon as a self-inflicted stain on his final day. The episode demonstrated, more completely than any case before it, that the clause’s near-absolute discretion meant near-absolute immunity from consequence. A president could do something that looked exactly like the sale of a federal favor, and as long as no smoking-gun bargain could be proven, the only sanctions available were reputational. Marc Rich kept his pardon. He died in 2013, in Switzerland, a free man as far as American law was concerned.

The Numbers: A Power in Decline by Volume, Rising by Notoriety

The findable artifact at the center of this institution’s story is quantitative, and it cuts against intuition. The pardon power has not grown more active over time. It has collapsed in volume while rising in visibility, which is a large part of why each modern grant draws scrutiny that a nineteenth-century pardon never would have.

Consider the per-presidency clemency counts across the twentieth century, and the contrast becomes stark. Woodrow Wilson granted clemency more than two thousand times across eight years. Franklin Roosevelt, across his unprecedented twelve years and the demands of the Depression and the war, granted clemency well over three thousand times, the highest total of any modern president, reflecting both his long tenure and the era’s continued treatment of the pardon as a working tool. Harry Truman granted clemency roughly nineteen hundred times across not quite eight years, a rate of around two hundred and fifty per year. Then the volume begins to fall. By the end of the century, the totals had dropped by an order of magnitude. The pattern that emerged in the late twentieth century, with George H.W. Bush granting fewer than eighty clemency actions across four years and his successors issuing grants in the low hundreds across two full terms, represented a collapse from the historical norm. To put the modern restraint in perspective, presidents who served two full terms at the close of the twentieth century granted clemency at rates that would have struck Truman or Roosevelt as astonishingly stingy, a few hundred grants across eight years where their predecessors had granted thousands.

The reasons for the collapse are worth naming because they explain why the modern pardon feels so different from the historical one. The drug war and the explosion of the federal prison population created a backlog of clemency petitions that the bureaucracy could not process, even as the political risk of any grant rose sharply. The Willie Horton politics of the late 1980s, in which a single furloughed offender’s later crime sank a presidential campaign, taught every subsequent president that leniency was electoral poison and that a single recipient’s recidivism could define a presidency. Presidents responded rationally to the incentive: grant as few pardons as possible, and concentrate the political risk in a single end-of-term batch issued when no election remained to lose. The result is the modern paradox. The pardon is used less than at any point in American history, yet it generates more controversy than ever, because the grants that survive the new risk-aversion are disproportionately the favors to insiders that a president will accept the reputational hit to deliver, timed for the moment when the hit costs the least. Mercy for the anonymous many declined; favors for the connected few became the visible residue.

Here is the artifact in tabular form, a comparison of major presidential pardons across the two centuries, showing the range from statecraft through mercy through questionable personal and political benefit:

Date	President	Recipient(s)	Stated Rationale	Political Context	Contemporary Reception	Long-Term Assessment
1795	Washington	Mitchell and Weigel, Whiskey Rebels	National reconciliation after rebellion suppressed	First armed test of federal authority	Broadly accepted as statesmanlike	Founding template for the healing pardon
1800	Adams	John Fries and Pennsylvania tax rebels	Avoid inflaming the country with executions	Split with Hamiltonian wing of his party	Cabinet opposed; contributed to his defeat	Vindicated as a country-over-party choice
1801	Jefferson	Sedition Act convicts including Lyon and Callender	Law deemed unconstitutional and void	Repudiation of Federalist prosecutions	Celebrated by Republicans, resented by Federalists	Early assertion of executive constitutional judgment
1863	Lincoln	Confederate soldiers via Amnesty Proclamation	Make rejoining the Union cheaper than rebellion	Reconstruction strategy mid-war	Welcomed by moderates, doubted by Radicals	Foundational instrument of reunion policy
1865 to 1868	A. Johnson	Former Confederate officials and planters	Rapid restoration of the Southern order	Open warfare with Radical Republicans	Bitterly contested; fed impeachment sentiment	Seen as undermining the war’s results
1947	Truman	Roughly 1,500 WWII draft resisters	Selective post-war reconciliation	Cold War dawning	Largely uncontroversial	Model of the screened, principled amnesty
1974	Ford	Richard Nixon	End the national nightmare of Watergate	Weeks after Nixon’s resignation	Approval collapse; suspicion of a deal	Reappraised as courageous decades later
1977	Carter	Roughly 10,000 Vietnam draft evaders	National reconciliation after the war	First full day in office	Veterans objected sharply	Closed the Vietnam draft chapter; politically costly
1989	Reagan	George Steinbrenner	Restore civil rights of campaign-finance offender	Departure from office	Minor controversy	Early sign of the favor-to-the-connected pattern
1992	Bush Sr.	Weinberger and five Iran-Contra figures	Prevent re-litigation of a settled scandal	Christmas Eve, after electoral defeat	Independent counsel alleged a cover-up	Benchmark for the self-protective pardon
2001	Clinton	Marc Rich	Claimed merit on underlying tax questions	Final morning, donor connection	Bipartisan condemnation; investigation	The defining cautionary case of the modern power

The quantitative companion to the table is the per-presidency rate, and it tells the story the controversies hide. Track the totals from Wilson and Roosevelt in the thousands, through Truman near nineteen hundred, down to the late-century presidents in the low hundreds across full eight-year tenures, and the curve falls steeply across the twentieth century. The notoriety curve runs the opposite direction. The institution that was once a high-volume, low-drama mercy bureaucracy has become a low-volume, high-drama favor machine, and the two trends are causally linked: scarcity is what makes each surviving grant worth scrutinizing, and political risk is what filters out the merciful many while leaving the connected few.

What the Scholars See: The Clemency Power Under the Microscope

The academic literature on the pardon is smaller than the literature on the veto or the war power, but it is sharp, and the leading voices disagree in instructive ways about what the power is for and how dangerous its modern drift has become. The classic treatment is W.H. Humbert’s older study of the pardoning power, which framed the clause primarily as an instrument of justice, the safety valve Hamilton described, and which read the historical record as a broadly responsible exercise of a necessary discretion. Humbert’s account is institutional and somewhat trusting; it treats the abuses as aberrations within a system that mostly worked as designed, the mercy bureaucracy doing its quiet good.

The more recent and more skeptical synthesis comes from the work that treats the pardon power comprehensively as a constitutional problem rather than a settled good, the scholarship associated with the argument that the clause’s near-absolute character has always carried within it the seeds of exactly the abuses that materialized in the Bush and Clinton grants. This body of analysis, the most thorough modern legal-historical treatment of the power, emphasizes that the Framers’ safeguard, the assumption that a single conscience would discipline itself and that impeachment would catch the worst cases, has proven inadequate. The impeachment carve-out stops a president from pardoning himself out of removal, but it does nothing to stop the self-protective pardon of accomplices, the donor-adjacent pardon, or the family pardon, none of which is impeachable on its own and all of which the clause permits. The comprehensive treatment of the pardon power lands on a structural conclusion: the safeguard the convention relied on was a character assumption, and character assumptions are the weakest possible constraint on a constitutional power.

A separate strand of scholarship focuses on the modern cases specifically. The work on the Ford pardon dissects the Nixon decision and the deal-or-no-deal question, generally concluding that no explicit bargain was struck but that Ford’s failure to demand a statement of contrition from Nixon, and his decision to act so quickly and unilaterally, guaranteed the appearance of corruption that destroyed his standing. The analysis of the Clinton pardons treats the Rich affair as the case where every latent danger of the power became manifest at once, the fugitive recipient, the bypassed Justice Department, the donor connection, the back-channel petition, and reads it as the clearest demonstration that the clause’s discretion is functionally unreviewable. The scholarship on the modern cases, taken together, names a transformation that the older institutional literature did not anticipate: the migration of the power’s center of gravity from mercy for the powerless to favors for the powerful.

Where do these accounts disagree? Humbert and the older institutional tradition trust the system and treat abuses as exceptions; the comprehensive modern legal-historical treatment distrusts the structure and treats the abuses as the predictable output of an under-constrained clause; the case-specific scholarship on Ford and on Clinton splits the difference, finding the individual grants explicable on their own terms while conceding that the cumulative effect has been a normalization of self-serving clemency that the Framers neither intended nor foresaw. The evidence, weighed honestly, favors the structural pessimists on the trajectory and the institutional optimists on the volume. The optimists are right that the overwhelming majority of the twenty thousand-plus grants across American history were ordinary mercy, processed responsibly through the pardon attorney’s office. The pessimists are right that the modern grants that survive the new political risk-aversion are disproportionately the self-interested ones, and that the structure offers no remedy. Both can be true at once, because they are describing different parts of the same distribution: a long tail of quiet mercy and a short, loud head of connected favors.

The Named Claim: The Clemency Ratchet

The thread that ties two centuries of pardons together is best understood as a ratchet, and naming it makes the pattern visible. Call it the clemency ratchet: every controversial pardon that survives expands the norm-range of what subsequent presidents can do, while every attempt to restrict the power requires political capital that almost no one is willing to spend on a problem that only manifests at the very end of a presidency. The ratchet turns in one direction. It does not turn back.

Watch the mechanism operate across the cases. Ford’s preemptive pardon of an uncharged Nixon proved that the power could reach forward to cover crimes never adjudicated, and once that was established it could not be un-established; the preemptive pardon became a permanent feature of the toolkit. Bush’s Christmas Eve grants to Iran-Contra figures proved that a defeated president could use his final weeks to foreclose prosecutions that threatened his own circle, and the demonstration that this drew no legal consequence lowered the bar for the next president who contemplated it. Clinton’s Rich pardon proved that even a fugitive with a donor connection could be cleared on the last morning without the Justice Department’s involvement, and the failure of the subsequent investigation to produce charges confirmed that the only price was reputational. Each grant that survived without legal sanction became a precedent, not in the formal sense of binding law, but in the practical sense of a thing that had been done and could therefore be done again. The norm-range widened with each cycle.

The restriction side of the ratchet barely exists, which is what makes the mechanism so durable. Constraining the pardon power would require a constitutional amendment, since the clause itself cannot be narrowed by statute, and amending the Constitution to fence in a power that causes trouble only at the margins, only at the end of presidencies, and only in a handful of high-profile cases per generation has never commanded anything close to the political energy required. Congress can investigate, as it did after the Rich pardon, and it can shame, but it cannot reach the grant itself. The courts have repeatedly affirmed the power’s near-absolute character, from Ex parte Garland in 1866 through the modern reaffirmations. The Office of the Pardon Attorney can be bypassed at will. The ratchet has no pawl to stop it from turning, no institutional brake that engages automatically when the power is abused. The only constraint that operates at all is the one Hamilton banked on and that has proven the weakest, the conscience of the individual president, supplemented by the reputational cost of a bad grant. For a president on his way out the door, who will never face the voters again, the reputational cost is a price many have shown themselves willing to pay.

The clemency ratchet is the institutional-biography version of the series’ larger argument about the modern presidency. The office was built for a republic of constrained magistrates, and the pardon was one of the places where the Framers deliberately built in a zone of monarchical discretion, justified by a crisis theory and an assumption about character. That zone of discretion, like the emergency powers forged in the Civil War, the Depression, the World Wars, and the Cold War, has not contracted with the conditions that justified it. It has expanded through use, normalized through repetition, and proven immune to the checks the founders thought sufficient. The pardon did not need a war to ratchet outward. It only needed presidents willing to test the edges of a power that, by design, has almost no edges, and the steady accumulation of those tests across two centuries. The same expansion-and-never-return pattern that defines the executive order and the imperial presidency more broadly operates here in miniature, through the normalization of increasingly personal and political grants rather than through any formal change to the twenty-eight words themselves.

The Complication: One Clause, Incommensurable Uses

The honest difficulty with judging the pardon power is that its uses do not sit on a single scale, and any attempt to grade the institution by one metric collapses distinctions that matter enormously. Washington’s pardon of two confused Whiskey Rebels and Lincoln’s amnesty for hundreds of thousands of Confederates were acts of statecraft in the highest sense, deployed to heal national ruptures that no court could close. Truman’s screened amnesty for principled conscientious objectors was an act of moral discrimination, sorting the genuine objector from the mere evader. Ford’s pardon of Nixon was a wager that closing the Watergate wound was worth more to the country than the satisfaction of a trial, a wager that cost Ford his presidency and that took twenty-five years to be vindicated. These belong to one family of uses, the pardon as an instrument of the commonwealth’s repair, exactly the use Hamilton defended in Federalist No. 74.

Bush’s Iran-Contra pardons and Clinton’s Rich pardon belong to a different family, and the difference is not merely one of degree. These grants served not the commonwealth and not mercy but the president’s own circle, his political allies, his donors, and in the Rich case a fugitive whose connection to the president ran through money. To average these against Washington and Lincoln, to produce a single verdict on whether the pardon power is good or bad, is to commit a category error. The power is a tool, and tools are judged by their uses, and the uses span from the noblest acts of national reconciliation to the most squalid favors a departing officer can dispense. The clause makes no distinction among them. The same twenty-eight words authorized the Whiskey Rebel reprieve and the Marc Rich grant, and the Constitution offers no internal standard for telling them apart.

This is the genuine complication, and it resists the tidy resolution a single thesis would prefer. One could argue that the noble uses justify the power despite the abuses, that a constitutional mechanism capable of ending Watergate or reuniting a war-torn country is worth the occasional Marc Rich. One could argue the opposite, that any power so easily turned to private favor is a structural defect regardless of its better uses, a loaded weapon left lying around for the rare president who will misuse it. The evidence does not adjudicate between these positions, because they rest on different values rather than different facts. What the evidence does establish is the trajectory: the noble uses cluster in the earlier history and in moments of genuine national crisis, while the self-interested uses cluster in the modern low-volume era and at the ends of presidencies. The clemency ratchet has not made the noble uses impossible, but it has made the squalid ones more normal, and the institution that once tilted toward mercy now tilts, in its most visible exercises, toward favor. Judging the power means holding both families of cases in view at once and refusing to let either erase the other.

What the Courts Built Around a Clause They Could Not Limit

The judicial history of the pardon is short, but the handful of cases that exist did more to define the power’s edges than any statute, and they almost uniformly confirmed how few edges there are. The foundational decision is Ex parte Garland in 1866, decided in the bitter aftermath of the Civil War, in which a former Confederate sought to practice law before the Supreme Court after receiving a presidential pardon. The Court used the occasion to describe the power in the broadest possible terms, holding that it is unlimited except for the impeachment carve-out, that it reaches every federal offense, and that it may be exercised at any time after the commission of the offense, before charges, during a prosecution, or after conviction. The Court went further, declaring that a full pardon, when accepted, “releases the punishment and blots out of existence the guilt, so that in the eye of the law the offender is as innocent as if he had never committed the offense.” That sweeping language about blotting out guilt would be qualified by later cases, but the central holding, that the power is near-absolute and reaches uncharged conduct, has never been disturbed. Ford’s lawyers relied directly on Garland when they advised that the president could pardon a Nixon who had been indicted for nothing.

A half-century later the Court added a subtler and more consequential gloss. In Burdick v. United States in 1915, a newspaper editor was offered a pardon to remove his Fifth Amendment privilege so he could be compelled to testify about his sources, and he refused to accept it. The Court held that a pardon must be accepted to take effect, that an unaccepted pardon is a nullity, and, most strikingly, that the acceptance of a pardon carries “a confession of guilt.” This is the doctrine that complicates the cheerful Garland language about innocence. A pardon does not declare the recipient innocent; the Court in Burdick suggested that taking one is closer to an admission. The acceptance requirement also means a recipient can decline, which matters in cases where a person would rather clear their name at trial than accept the implication of guilt that the pardon’s acceptance carries. The tension between Garland’s blotting out of guilt and Burdick’s confession of guilt has never been fully resolved, and it sits at the heart of every debate about what a pardon actually means for the person who receives it.

Two further cases mapped the power’s reach into adjacent territory. In Ex parte Grossman in 1925, the Court held that the pardon power extends to criminal contempt of court, rejecting the argument that allowing presidents to pardon contempt would let the executive nullify the judiciary’s authority to enforce its own orders. The opinion, by Chief Justice and former president William Howard Taft, who knew the clemency power from the inside, reasoned that the founders had borrowed the pardon from English practice where it had always reached contempt, and that the occasional friction with the courts was a price the Constitution accepted. And in the conditional-pardon line of cases, the Court confirmed that the greater power to pardon includes the lesser power to attach conditions, so that a president may commute a sentence or grant clemency on terms, as Lincoln did when he conditioned Confederate amnesty on a loyalty oath. The cumulative effect of this small body of law was to ratify, again and again, the near-absolute character the text suggests. The courts did not fence the power in. They confirmed that it could not easily be fenced, and then they got out of its way.

The Verdict

The pardon power is the most unilateral authority the Constitution gives a president, and after two centuries of use the case for keeping it intact rests almost entirely on its rare moments of greatness while the case against it rests on the regularity of its abuse. The verdict this article reaches is twofold and refuses to be flattened into a single judgment. On the design, the Framers got the core decision right and the safeguard wrong. They were correct that a republic needs a dispensing power, a place where the rigidity of the criminal law can be bent in the individual case and where a single decision-maker can offer credible mercy fast enough to matter in a crisis. The Whiskey Rebellion reprieves, the Sedition Act pardons, the Lincoln amnesty, and the Vietnam-era reconciliations vindicate that judgment. A government without such a valve would be more brittle and more cruel. But the Framers were wrong to think that the impeachment carve-out plus the conscience of the officeholder would suffice to police the power’s abuse. Madison saw the danger and lost the argument, and the modern record of self-protective and favor-driven pardons proves Madison was right about the danger even as Hamilton was right about the need.

On the trajectory, the verdict is that the clemency ratchet is real, that it turns in one direction, and that the institution has drifted measurably from mercy toward favor in its most visible modern exercises without any corresponding drift in the text that authorizes it. The drift is not a story of bad men, though some of the grants were bad; it is a story of a power that was always near-absolute being used, across enough cycles, to discover exactly how near-absolute it was, with each discovery becoming a usable precedent for the next president. The pardon did not need to be amended to become more dangerous. It only needed to be used by presidents willing to find its limits, and the structure ensured that what one found, the next could use.

The defensible bottom line is that the pardon power should be neither abolished nor left entirely as it is. Abolition would sacrifice the genuine and irreplaceable good of the mercy valve and the crisis-pacifying reprieve to prevent abuses that, while real, are rare and concentrated at the predictable margin of presidential terms. But pretending the abuses are not real, or that the structure adequately polices them, ignores two centuries of evidence culminating in the Rich affair. The clause that has not changed since 1787 has done both more good and more harm than its authors imagined, and an honest accounting has to hold both ledgers open. The pardon is a king’s power inside a republican office, and the republic has been living with the consequences, good and bad, for two hundred and thirty years.

Legacy and Implication: A King-Shaped Hole That Never Closed

The presidential pardon is the clearest surviving example of something the Constitution did on purpose and then could not undo: it built a zone of monarchical discretion into a republican executive, defended the choice with a theory of crisis and an assumption about character, and created no mechanism to contract the zone if the theory or the assumption failed. Every other major executive power has accumulated some countervailing structure over two centuries. The war power generated the War Powers Resolution, however weak. The appointment power runs through Senate confirmation. The treaty power requires two-thirds of the Senate. The executive order is subject to judicial review and statutory override, the subject of the institutional history of the executive order that runs from Washington to Clinton and that shows how even the most expansive ordinance power eventually drew limits from the courts and Congress. The veto, traced from Jackson’s transformative 1832 Bank Veto Message through its modern strategic use, is checked by the override and was even briefly experimented upon with the line-item veto before the Supreme Court struck that down. The pardon alone has acquired no countervailing structure that reaches the grant itself. It remains as near-absolute in 2001 as it was in 1795, the one place where the king-shaped hole the Framers cut has never been patched.

That permanence is the institution’s deepest legacy, and it connects the pardon to the series’ larger argument about an office that keeps the powers crises gave it long after the crises pass. The pardon’s expansive zone of discretion was justified by the crisis theory, the well-timed offer of mercy that quiets a rebellion, and that justification was honored in the early cases. But the discretion outlived its crisis justification, just as the emergency powers of the modern presidency outlived the emergencies that produced them. By the late twentieth century, the crisis-pacifying reprieve had become the rare exception, and the ordinary modern use, when the power surfaced at all, was the end-of-term favor. The discretion remained; the justification had quietly evaporated. The president on his last morning signing a fugitive’s pardon was exercising the same authority Washington used to heal the Whiskey Rebellion, but the healing purpose was gone, and only the unchecked discretion remained. That is the imperial-presidency pattern in its purest form: a power justified by an emergency, retained after the emergency, and used for purposes the emergency never contemplated.

The implication for how Americans should think about the office is that some of its most dangerous features are not the product of modern overreach but of original design, built in at the founding and defended by the founders themselves. The pardon was not seized by an imperial president from a weakened Congress; it was handed over freely at Philadelphia, over Madison’s objection, on Hamilton’s argument. The lesson is that the constraints on executive power that Americans rely on, the checks and balances of civic mythology, were never complete, and that the founders knowingly left at least one zone where a president would answer to nothing but his conscience and the verdict of history. The pardon is where that verdict gets written, case by case, grant by grant, from Mitchell and Weigel on the Pennsylvania frontier to Marc Rich in his Swiss exile. Two centuries on, the clause stands exactly as written, the uses stretch from the noblest statecraft to the squalidest favor, and the only check that ever fully engages is the one this article participates in: the long, slow, unforgiving judgment of the historical record. That judgment is the pawl the Framers forgot to build, and it is the only thing that ever makes the clemency ratchet feel, even briefly, like it might hold still.

Frequently Asked Questions

Q: What does Article II Section 2 actually say about the pardon power?

Article II, Section 2 of the Constitution grants the president the “Power to grant Reprieves and Pardons for Offenses against the United States, except in Cases of Impeachment.” That is the entire textual basis, just twenty-eight words. It contains only two limits. First, the offense must be federal, meaning the president cannot pardon state crimes, which remain with governors and state boards. Second, the power cannot undo an impeachment, so a president cannot use clemency to rescue an official from congressional removal. Beyond those two carve-outs, the clause sets no conditions. It requires no conviction, no charge, no hearing, no Justice Department recommendation, no standard of evidence, and no explanation. This near-total absence of internal limits is what makes the pardon the most unilateral power in the Constitution and what has allowed it to be used across two centuries for purposes ranging from national reconciliation to personal favor.

Q: Did the Founders argue about the pardon power before they approved it?

Yes, and the argument was pointed. At the 1787 Constitutional Convention, James Madison worried that a president might abuse the pardon to shield his own confederates or to cover crimes in which he was complicit. Edmund Randolph of Virginia moved to bar the president from pardoning treason specifically, reasoning that treason was the crime where a president’s self-interest might be most dangerous, since a chief executive could be the head of the very plot the treason laws punished. The convention rejected Randolph’s amendment. The defenders argued that in moments of insurrection a single decision-maker who could offer immediate, credible mercy might restore peace faster than any court or legislature, and that the impeachment carve-out was a sufficient safeguard against the worst abuse. Madison lost the argument, but the modern record of self-protective pardons suggests his warning was prescient.

Q: What did Hamilton say about the pardon in the Federalist Papers?

Alexander Hamilton defended the pardon power in Federalist No. 74, one of the shortest essays in the collection. His argument had two parts. First, he made a case for mercy: criminal codes are necessarily harsh, and without easy access to exceptions the law would wear a countenance too sanguinary and cruel. Someone had to hold the power to soften the law in individual cases, and Hamilton argued it should be one person rather than a body, because a single conscience feels the weight of the decision in a way a committee never will. Second, he made a case for crisis management: in seasons of rebellion, a well-timed offer of pardon might restore tranquility to the commonwealth, and the moment might be too fleeting to allow for legislative deliberation. Hamilton foresaw the rebel-pacifying pardon and the mercy pardon. He did not dwell on the danger of the private favor.

Q: Why did Washington pardon the Whiskey Rebels?

George Washington pardoned John Mitchell and Philip Weigel, the two men convicted of treason during the 1794 Whiskey Rebellion, in 1795. His reasoning followed Hamilton’s crisis theory closely. Washington had already demonstrated federal authority by personally leading roughly thirteen thousand militia to suppress the tax revolt in western Pennsylvania, and the convictions had vindicated the government’s power to punish. Executing two confused frontier farmers, one described as barely competent, would have gained the government nothing and cost it the goodwill of a region it needed to reconcile. The pardon, deployed after the demonstration of strength, was the tool that healed the wound the rebellion had opened. This established the template of force first and clemency second, and it made clear from the very beginning that the pardon would serve not only individual mercy but national reconciliation, a use that recurs throughout American history.

Q: How did Jefferson use the pardon power differently from Washington?

Thomas Jefferson used the pardon to render a verdict on the previous administration rather than to manage a present crisis. When he took office in 1801, he regarded the Alien and Sedition Acts, under which the Adams administration had jailed newspaper editors and a sitting congressman for criticizing the government, as unconstitutional. He could not repeal a law that had already expired, but he could undo its effects. He pardoned those convicted under the Sedition Act, including Matthew Lyon and James Callender, and remitted fines where he could. Jefferson framed this not as discretionary mercy but as constitutional duty, arguing he was bound to treat the law as a nullity. This created a new use of the power, a retroactive repudiation of a prior administration’s judgments and an assertion that the executive’s own reading of the Constitution could override verdicts the courts had let stand.

Q: How many pardons did Lincoln grant during the Civil War?

Abraham Lincoln granted clemency in what his son Robert Todd Lincoln later estimated as thousands of cases over the course of the Civil War, a figure historians regard as plausible given the volume of military justice the conflict produced. Lincoln operated on two levels. At the policy level, his December 1863 Amnesty Proclamation offered conditional pardons to Confederate soldiers who would take a loyalty oath and accept emancipation, the centerpiece of his plan to make rejoining the Union cheaper than rebellion. At the individual level, he was famous for sparing condemned soldiers, including deserters, sentries who slept on duty, and boys who broke under fire. He developed informal categories of cases he would always commute, and his generals complained that his leniency undermined discipline. The mercy mythology of his presidency, unlike many such legends, rests on a documentary record of remitted sentences.

Q: Why was Andrew Johnson’s use of pardons so controversial?

Andrew Johnson used the pardon power lavishly after the Civil War to restore the antebellum Southern ruling class to power quickly, which put him in open conflict with the Radical Republicans in Congress. His May 1865 Amnesty Proclamation pardoned most former Confederates who took a loyalty oath, while requiring high officials and wealthy planters to apply to him personally, and he then granted those applications by the thousands. On Christmas Day 1868 he issued a final universal amnesty covering essentially all remaining former Confederates. To the Radical Republicans, this deliberately dismantled the war’s results, returning to power the men who had led the rebellion even as the South enacted the Black Codes. The pardon controversy fed the atmosphere of distrust that produced Johnson’s impeachment, though the impeachment itself turned on the Tenure of Office Act. He survived removal by a single Senate vote.

Q: Why did Ford pardon Nixon, and was there a deal?

Gerald Ford pardoned Richard Nixon on September 8, 1974, granting a full pardon for all federal crimes Nixon committed or may have committed during his presidency, just a month after Nixon resigned and before any charges were filed. Ford said he acted to spare the country the spectacle of a former president on trial, to end the long national nightmare of Watergate, and to free his administration to govern. The widespread suspicion was that a corrupt bargain had been struck, that Nixon resigned in exchange for a promise of immunity. Historians who have examined the question generally conclude that no explicit deal existed, but that the appearance of one was unavoidable and politically fatal regardless of the truth. Ford’s approval collapsed from the seventies into the forties, and the pardon plausibly cost him the close 1976 election. Decades later, the decision was widely reappraised as courageous.

Q: What was the Marc Rich pardon and why was it so scandalous?

Marc Rich was a commodities trader who fled to Switzerland in 1983 rather than face a sixty-five-count federal indictment for tax evasion, racketeering, and oil trades with Iran during the hostage crisis. He spent seventeen years as a fugitive and never stood trial. On his final morning in office, January 20, 2001, Bill Clinton pardoned him. The scandal assembled nearly every concern the institution had ever raised. Rich was a fugitive, not a petitioner who had served his sentence. The Office of the Pardon Attorney was bypassed and the prosecutors who built the case were not consulted. The petition reached Clinton through a back channel of private lawyers. And Rich’s ex-wife, Denise Rich, was a major Democratic donor who had given more than a million dollars to Democratic causes and the Clinton library and had personally lobbied for the grant. A federal investigation found no prosecutable crime, because the pardon, however unseemly, was constitutional.

Q: Can a president pardon someone who has not been charged or convicted?

Yes. The Constitution’s pardon clause does not require a conviction or even a charge, and this silence is meaningful. The Supreme Court confirmed in the 1866 case Ex parte Garland that the power can be exercised at any time after an offense is committed, whether before legal proceedings begin, while they are pending, or after conviction and judgment. This is the legal basis for the preemptive pardon, the grant that forecloses prosecution before it starts. Ford’s 1974 pardon of Nixon was the most prominent test of the principle, since Nixon had not been indicted, and it held. A president can therefore pardon a person for conduct that has never been adjudicated, lifting the threat of prosecution off someone who has not been accused in any court. This is one of the features that makes the power so expansive and so resistant to external check.

Q: Can a president pardon state crimes?

No. The pardon power reaches only “Offenses against the United States,” which means federal crimes. A president cannot pardon someone convicted of a state crime such as most murders, robberies, or assaults, which are prosecuted under state law. Clemency for state offenses rests with state governors and, in some states, state pardon or parole boards, each operating under that state’s own constitution and statutes. This federalism limit means that a person convicted of both federal and state charges arising from the same conduct could receive a presidential pardon for the federal counts while remaining fully liable on the state counts. The division reflects the constitutional structure: the federal government and the states operate as separate sovereigns with separate criminal codes, and the president’s clemency authority extends only as far as federal jurisdiction does.

Q: What is the Office of the Pardon Attorney and can the president ignore it?

The Office of the Pardon Attorney is a unit within the Department of Justice, formally established in 1894, that processes clemency petitions. It reviews applications, investigates cases, solicits the views of prosecutors and sentencing judges, and forwards recommendations to the president through the attorney general. For most ordinary clemency grants, a petitioner waits the required years after completing a sentence, files an application, and submits to this review. But the office exists by administrative choice, not constitutional command, and a president can ignore it entirely. He can grant a pardon to anyone he chooses without the office’s knowledge or recommendation. Clinton used exactly this informal track for Marc Rich, which is why Justice Department veterans found it so alarming. The pardon attorney’s office is meant to discipline the power, but it has no authority the president cannot bypass at will.

Q: How many pardons have been granted in American history?

Across American history, the clemency power has been exercised roughly twenty thousand to twenty-one thousand times, a figure that includes full pardons, commutations of sentence, and remissions of fines. The overwhelming majority of these grants were ordinary acts of mercy and sentence adjustment given to people whose names no one remembers, processed through the bureaucratic clemency apparatus rather than decided at the president’s elbow. This volume is the part of the story the headline controversies obscure. The Marc Rich and Whiskey Rebellion grants are vanishingly rare exceptions in a vast record of routine clemency. The modern image of the pardon as an exotic act deployed mainly in scandal is a distortion produced by the late-twentieth-century collapse in volume; for the first century and a half of the republic, the pardon was a working mercy bureaucracy grinding through thousands of unremarkable cases.

Q: Why do modern presidents grant so many fewer pardons than earlier ones?

The collapse in pardon volume is one of the most counterintuitive facts about the institution. Presidents like Wilson and Franklin Roosevelt granted clemency in the thousands, and Truman granted it roughly nineteen hundred times across nearly eight years. By the late twentieth century, totals had dropped by an order of magnitude, with some two-term presidents issuing only a few hundred grants. Several forces drove the decline. The drug war and the explosion of the federal prison population created a petition backlog the bureaucracy could not process. And the politics of crime turned sharply punitive after the late 1980s, when a single furloughed offender’s later crime helped sink a presidential campaign, teaching every subsequent president that leniency was electoral poison. Presidents responded rationally by granting as few pardons as possible and concentrating the political risk in a single end-of-term batch.

Q: What was the difference between the Truman and Carter war-resister amnesties?

Both presidents used the pardon to settle the human aftermath of a war, but they differed in scope and method. Harry Truman, in December 1947, pardoned roughly fifteen hundred men convicted of draft violations during the Second World War, but he did so selectively, appointing a three-member Amnesty Board that reviewed more than fifteen thousand cases and recommended pardons for only about ten percent, the men whose objection appeared genuinely rooted in conscience. Jimmy Carter, on his first full day in office in January 1977, issued an unconditional blanket pardon to Vietnam-era draft evaders, affecting roughly ten thousand men directly. Carter’s grant was broader and faster, framed as national reconciliation, and far more controversial. Veterans objected that pardoning the men who fled while others died was an insult to those who served. Carter narrowed it to draft evaders and pointedly excluded deserters, preserving a distinction between refusing to go and abandoning a unit in the field.

Q: What were the Iran-Contra pardons and why were they controversial?

On Christmas Eve 1992, six weeks after losing his reelection bid, George H.W. Bush pardoned six figures from the Iran-Contra affair, the scandal in which Reagan administration officials secretly sold arms to Iran and diverted the proceeds to the Nicaraguan Contras in defiance of a congressional ban. The most consequential recipient was former defense secretary Caspar Weinberger, who was scheduled to stand trial within weeks on charges of lying to Congress. The independent counsel, Lawrence Walsh, reacted with fury, declaring that the pardons completed a cover-up and noting that Bush himself had been a subject of the investigation. Walsh’s charge was that Bush had pardoned witnesses whose testimony might have implicated him, edging toward the self-shielding scenario Madison warned against. The timing, on Christmas Eve after an electoral defeat on the eve of a trial that could have damaged him, established the template of the last-minute defensive pardon.

Q: Is there any way to overturn or block a presidential pardon?

Functionally, no. Once a president signs a pardon, there is no constitutional mechanism to reverse it. Congress cannot undo a pardon, because the clause places it beyond legislative reach; the Andrew Johnson era proved this when the Radical Republicans, who despised his Confederate pardons, could not stop them. The courts have repeatedly affirmed the power’s near-absolute character, from Ex parte Garland in 1866 onward, and have not recognized any authority to void a completed grant. A president cannot use the pardon to escape his own impeachment, but that carve-out does not reach the grants themselves. The only remedies are political and reputational: Congress can investigate and shame, as it did after the Marc Rich pardon, and the historical record can render its judgment. This absence of any institutional brake is precisely what makes the power so dangerous in the rare cases where it is abused.

Q: What is the clemency ratchet?

The clemency ratchet is the pattern this article names to describe how the pardon power expands over time. Every controversial pardon that survives without legal consequence widens the norm-range of what subsequent presidents can do, while every attempt to restrict the power requires political capital almost no one is willing to spend on a problem that surfaces only at the end of presidencies. The ratchet turns in one direction and does not turn back. Ford’s preemptive pardon of an uncharged Nixon proved the power could reach crimes never adjudicated; Bush’s Iran-Contra grants proved a defeated president could foreclose prosecutions threatening his own circle; Clinton’s Rich pardon proved even a fugitive with a donor connection could be cleared on the last morning. Each grant that drew no legal sanction became a usable precedent. There is no institutional brake, no pawl, that stops the mechanism from turning, only the conscience of the president and the slow judgment of history.

Q: Does the pardon power support the idea of an imperial presidency?

Yes, in a specific and revealing way. The pardon is the clearest example of a power that was justified by crisis, retained after the crisis passed, and used for purposes the original justification never contemplated, which is the defining pattern of the imperial presidency. The Framers defended the broad pardon power on the theory that a well-timed offer of mercy could quiet a rebellion, and the early cases honored that purpose. But the discretion outlived its crisis justification. By the late twentieth century, the rebellion-pacifying reprieve had become a rare exception, and the ordinary modern use was the end-of-term favor to an insider. The discretion remained while the justification evaporated. Unlike the war power or the treaty power, the pardon has acquired no countervailing structure that reaches the grant itself, making it the one place where the king-shaped zone of discretion the Founders built has never been fenced in.

Q: How do historians and legal scholars disagree about the pardon power?

The leading scholars split along a structural fault line. The older institutional tradition, exemplified by W.H. Humbert’s classic study, frames the pardon primarily as an instrument of justice and reads the historical record as a broadly responsible exercise of necessary discretion, treating abuses as aberrations within a system that mostly worked. The more skeptical modern legal-historical treatment, the comprehensive analysis of the power as a constitutional problem, argues that the clause’s near-absolute character always carried the seeds of the abuses that materialized in the Bush and Clinton grants, and that the Framers’ character-based safeguard was the weakest possible constraint. Case-specific scholarship on the Ford and Clinton pardons splits the difference, finding individual grants explicable while conceding a cumulative normalization of self-serving clemency. The evidence favors the optimists on volume, since most grants were ordinary mercy, and the pessimists on trajectory, since the surviving modern grants tilt toward favor.

Q: Why is the pardon called the most unilateral power in the Constitution?

The pardon earns that description because it is subject to fewer checks than any other major presidential authority. The president cannot spend money without Congress, cannot appoint judges or senior officials without Senate confirmation, cannot make treaties without two-thirds of the Senate, cannot command war indefinitely without appropriations, and cannot issue executive orders immune from judicial review or statutory override. The pardon answers to none of these. It requires no legislative approval, no judicial sign-off, no agency recommendation, and no explanation. The only two limits are that the offense must be federal and that the power cannot undo an impeachment. Within those bounds the president acts alone, with no obligation to consult anyone, and his decision is final and irreversible. This concentration of unreviewable discretion in a single person, exercised without process, is what makes the pardon the closest thing to a monarchical prerogative that survives in the American constitutional system.

Q: Did the pardon power ever face a serious attempt at reform?

No sustained reform has ever come close to succeeding, and the reason illuminates why the clemency ratchet is so durable. Constraining the pardon would require a constitutional amendment, because the clause cannot be narrowed by ordinary statute. Amending the Constitution to fence in a power that causes trouble only at the margins, only at the ends of presidencies, and only in a handful of high-profile cases per generation has never commanded the political energy required. After particularly egregious grants, such as the Marc Rich pardon, Congress has held hearings and proposed measures, but these efforts consistently fail to advance because the abuse, however galling, is too rare and too concentrated to mobilize the supermajorities an amendment demands. The Office of the Pardon Attorney provides a voluntary procedural check, but a president can bypass it at will. The structural reality is that the power has no automatic brake, and building one has never been worth the cost to anyone with the power to do it.

Q: What is the lasting lesson of the pardon power’s history?

The deepest lesson is that some of the presidency’s most dangerous features are not modern overreach but original design, built in at the founding and defended by the founders themselves. The pardon was not seized by an imperial president from a weakened Congress; it was handed over freely at Philadelphia, over Madison’s objection and on Hamilton’s argument. This means the checks and balances of American civic mythology were never complete. The Framers knowingly left at least one zone where a president would answer to nothing but his conscience and the verdict of history. The pardon proves that a republic can deliberately retain a king’s power and live with the consequences, good and bad, for centuries. The same clause produced the Whiskey Rebellion reprieve and the Marc Rich grant, the Lincoln amnesty and the Iran-Contra cover-up. The only check that ever fully engages is the long, slow, unforgiving judgment of the historical record, which is the pawl the Framers forgot to build.