On Friday, February 5, 1937, at noon, Franklin Roosevelt summoned his cabinet and the Democratic leadership of both houses of Congress to the White House for what they assumed was a routine briefing. Vice President John Nance Garner, Speaker William Bankhead, Senate Majority Leader Joseph Robinson, House Majority Leader Sam Rayburn, Senate Judiciary Chairman Henry Ashurst, and House Judiciary Chairman Hatton Sumners filed in. Attorney General Homer Cummings was already seated. The press waited in an adjoining room. Roosevelt handed each man a copy of a message he was about to deliver to Congress within the hour and a copy of a bill draft. He gave them ten minutes to read. Then he announced he was sending both to the Capitol immediately. There would be no consultation, no markup, no negotiation. The bill, which the administration insisted on calling the Judicial Procedures Reform Bill, would add up to six new justices to the Supreme Court, one for every sitting member of the bench who had passed seventy years of age and refused to retire.

Six of the nine sitting justices were over seventy. The bill could expand the high tribunal from nine to fifteen at a stroke. Sumners, walking back to the House, told Rayburn in the cloakroom: “Boys, here’s where I cash in my chips.” Garner, who would oppose the plan privately throughout the spring, was reported to have walked back to the Senate holding his nose with one hand and giving the thumbs down with the other. The reaction was not what Roosevelt expected. He thought he had a mandate. He had won 523 electoral votes to Alf Landon’s eight in November 1936, the largest electoral margin since James Monroe ran unopposed in 1820. Within six months, his own party would kill his bill on the Senate floor by a vote of seventy to twenty. He would never recover the legislative momentum of his first term. And yet, by the time the proposal died in July, the constitutional revolution it was meant to compel had already arrived through other means. The threat had done the work. That paradox is the story of the 1937 fight, and it is the cleanest single test in American political history of the difference between losing a bill and winning a war.

The Court That Would Not Bend: 1935 to 1936

The bench Roosevelt inherited in March 1933 was the same one his cousin Theodore had filled out in his second term and that William Howard Taft, William McKinley, Warren Harding, Calvin Coolidge, and Herbert Hoover had stocked between them. Of the nine sitting justices when FDR took office, four had been appointed by Republicans during the 1920s. Willis Van Devanter, appointed by Taft in 1910, was the senior conservative. James McReynolds, appointed by Wilson in 1914 as the lone Democratic exception and ironically the most reactionary member of the bench, was open about his anti-Semitic refusal to speak to Brandeis. George Sutherland came from Harding in 1922, Pierce Butler from Harding in 1923. These four men formed what Roosevelt’s allies would soon call the Four Horsemen, after the four horsemen of the apocalypse. They voted as a bloc, and the bloc voted to strike down the regulatory and welfare-state experiments that the Roosevelt administration was producing at a pace unprecedented in American legislative history. The remaining five justices, often grouped as the Three Musketeers (Louis Brandeis, Benjamin Cardozo, and Harlan Fiske Stone) plus the two moderate centrists Chief Justice Charles Evans Hughes and Owen Roberts, would join the Horsemen often enough to give the New Deal a string of crushing losses through 1935 and 1936.

The pattern began on January 7, 1935, when Panama Refining Co. v. Ryan struck down a small portion of Section 9(c) of the National Industrial Recovery Act on non-delegation grounds. The administration shrugged off the loss as technical. The shrug was premature. On May 6, 1935, the bench killed the Railroad Retirement Act of 1934 in Railroad Retirement Board v. Alton Railroad Co. by a five-to-four vote, with Roberts joining the Four Horsemen, holding that pensions for railway workers exceeded the commerce power. The decision telegraphed where the swing votes were heading on broader New Deal questions, but it was treated by the press as an isolated setback because the retirement act had been hastily drafted and was widely seen as constitutionally vulnerable.

Then came Black Monday. On May 27, 1935, the bench handed down three unanimous decisions in a single morning, each of which gutted a central element of the early New Deal. Schechter Poultry Corp. v. United States invalidated the National Industrial Recovery Act, the centerpiece of the first hundred days legislative program covered in detail by FDR’s Hundred Days bill-by-bill reconstruction. The justices ruled that Congress had unconstitutionally delegated legislative power to the executive and that the regulation of intrastate poultry slaughter exceeded the commerce power. The fact that the decision was nine to zero, with Brandeis and Cardozo joining Sutherland and McReynolds, made it impossible for Roosevelt to dismiss as ideological. Humphrey’s Executor v. United States, decided the same day, ruled that the president could not remove members of the Federal Trade Commission without cause, distinguishing the holding from Myers v. United States (1926) on the ground that quasi-legislative and quasi-judicial bodies stood apart from purely executive officers. Louisville Joint Stock Land Bank v. Radford, also released that morning, struck down the Frazier-Lemke Farm Bankruptcy Act on Fifth Amendment grounds. Three statutes died in one session.

Roosevelt held a press conference four days later, on May 31, that ran ninety minutes and included his famous complaint that the bench had returned the country to “the horse and buggy days” of constitutional interpretation. The phrase was widely reported and widely mocked. The administration’s lawyers, particularly Solicitor General Stanley Reed and Assistant Attorney General Robert Jackson, began drafting contingency strategies, but Roosevelt himself was at this point still publicly committed to working within the constitutional framework as the justices had defined it. That posture lasted seven more months.

On January 6, 1936, the bench struck again in United States v. Butler. By a six-to-three vote, with Roberts joining the Four Horsemen and Hughes adding a sixth vote, the Agricultural Adjustment Act was held unconstitutional. The majority opinion, written by Roberts, advanced what came to be called the “T-square” theory of judicial review, in which the justices simply laid the statute alongside the Constitution and noted whether it fit. Stone’s dissent, joined by Brandeis and Cardozo, mocked the theory as childish: “The only check upon our own exercise of power is our own sense of self-restraint.” The AAA’s processing tax, which had funded payments to farmers for reducing crop acreage, fell with the statute. Around four hundred million dollars in tax refunds were potentially at stake. Congress would respond by re-enacting the program in modified form as the Soil Conservation and Domestic Allotment Act of 1936, but the constitutional point had been made: the bench would strike at the heart of the New Deal whenever it could find a doctrinal toehold.

On May 18, 1936, Carter v. Carter Coal Co. killed the Bituminous Coal Conservation Act of 1935 by a five-to-four vote, again with Roberts and Hughes splitting from the liberal three on the wage-and-hour provisions. The decision narrowed the commerce power to a point that, taken at face value, made any national labor regulation constitutionally impossible. The Bituminous Coal Act had been carefully drafted to address objections raised in Schechter, and its invalidation suggested that no redrafting would suffice. Three weeks later, on June 1, 1936, Morehead v. New York ex rel. Tipaldo struck down a New York minimum wage law for women by a five-to-four vote, with Roberts joining the Horsemen and Hughes dissenting alongside the liberal three. Morehead reaffirmed Adkins v. Children’s Hospital (1923) and seemed to say that no state, let alone the federal government, could legislate a minimum wage for any class of workers. Even Republican newspapers found the holding extreme. The Republican platform that summer would call for “the protection of women and children” through state action, an implicit repudiation of Tipaldo from the justices’ own ideological camp.

By the end of June 1936, the bench had invalidated, in whole or in part, the National Industrial Recovery Act, the Agricultural Adjustment Act, the Bituminous Coal Conservation Act, the Railroad Retirement Act, the Frazier-Lemke Farm Bankruptcy Act, the Municipal Bankruptcy Act, and the New York minimum wage law. The Wagner Act, the Social Security Act, the Tennessee Valley Authority’s full operations, and the Public Utility Holding Company Act of 1935 had not yet reached the high tribunal but were widely expected to fall when they did. The administration’s legal program was, in any reasonable forecast, a dead letter the moment any of those statutes reached a final conference vote.

The 1936 Mandate Roosevelt Misread

The November 1936 election was the largest landslide in modern American political history. Roosevelt won 60.8 percent of the popular vote to Landon’s 36.5 percent. He carried forty-six of forty-eight states. Only Maine and Vermont voted Republican, giving rise to the Democratic quip that “as Maine goes, so goes Vermont.” The electoral college tally was 523 to 8. In the Senate, Democrats expanded their majority to seventy-six of ninety-six seats. In the House, the Democratic margin reached 334 of 435. No president since had assembled or has since assembled an equivalent coalition. Roosevelt read the result as a mandate not merely for continued New Deal legislation but for whatever institutional rearrangement was necessary to make the program durable. He told Cummings shortly after the election that “the people are with us. The Court is not. We will see what the people want done about that.” The reading was both correct and disastrously narrow. The voters had endorsed Roosevelt and the New Deal. They had not specifically endorsed any particular method of dealing with the judicial obstacle.

There were three obvious options on the desk in November 1936. The first was a constitutional amendment. Several proposals were in circulation in academic and progressive circles. Robert La Follette Jr. of Wisconsin had drafted language requiring a two-thirds supermajority of the justices to invalidate any federal statute. Senator George Norris of Nebraska favored an amendment giving Congress power to override judicial nullification by a two-thirds vote of both houses, an idea drawn from Canadian and Australian models. Roosevelt’s own Attorney General had at various points entertained an amendment narrowing the due process and commerce clauses. The amendment route had the advantage of legitimacy: it would change the rules openly and require the assent of three-quarters of state legislatures, satisfying both democratic and federalist principles. It had the disadvantage of time. Even a swift amendment campaign would take three to five years, and the New Deal program could not wait that long with each successive statute heading toward judicial graveyard.

The second option was to wait. The average age on the bench was over seventy. Van Devanter was seventy-seven, Brandeis eighty, Hughes seventy-four, McReynolds seventy-five, Sutherland seventy-four, Butler seventy. Actuarially, vacancies would arrive within Roosevelt’s second term. The trouble was that conservative justices had famously refused to retire under Republican presidents and there was no reason to expect them to step aside in time. Holmes had served until age ninety. Field had served until eighty-one. McReynolds in particular had told friends he would die on the bench before letting Roosevelt name his successor, and would in fact do so, retiring only in February 1941. Waiting felt to Roosevelt like submission to the dead hand of the past.

The third option was statutory: change the size or jurisdiction of the bench by ordinary legislation. The size of the Supreme Court is not fixed in the Constitution. Congress has set it at six (1789), then five (1801), then six again (1802), then seven (1807), then nine (1837), then ten (1863), then seven (1866, although attrition rather than removals achieved the reduction), then nine again (1869, the current configuration). Eight statutory adjustments in the eighty years between the Judiciary Act of 1789 and the Judiciary Act of 1869 had set and reset the number. The 1866 reduction, designed by Radical Republicans to prevent Andrew Johnson from filling vacancies, was the most recent partisan precedent. Court expansion by ordinary statute, in other words, had a long American history, and the constitutional objection to it was political rather than textual.

Cummings, who had been quietly working on the problem since 1935, presented Roosevelt in early December 1936 with a memorandum proposing the statutory route in a form deliberately disguised as efficiency reform. The memorandum drew on a 1913 proposal by James McReynolds himself, then Wilson’s Attorney General, who had urged Congress to authorize the appointment of additional federal judges whenever sitting judges over seventy refused to retire. McReynolds had aimed his 1913 proposal at the lower federal bench, not the Supreme Court, but the principle was the same. Cummings, with characteristic legal showmanship, attached a copy of McReynolds’s 1913 letter to the memorandum, knowing that McReynolds was now seventy-five and the senior conservative on the high tribunal whose vote Roosevelt most wanted to neutralize. The historical symmetry delighted Roosevelt. It also misled him. He concluded that a proposal originally made by McReynolds himself could not credibly be attacked as a violation of judicial independence. The conclusion proved to be the largest political miscalculation of his presidency.

The Bill Drafted Behind a Closed Door

Cummings drafted the bill between mid-December 1936 and late January 1937 in near-total secrecy. The drafting circle consisted of Cummings, Solicitor General Stanley Reed, Assistant Attorney General Joseph Keenan, Reed’s deputy Alexander Holtzoff, and one or two Department of Justice lawyers including Carl McFarland. Roosevelt himself approved the text. Not consulted: Felix Frankfurter, who as the administration’s most respected outside legal advisor would normally have been brought in immediately; Senate Majority Leader Joseph Robinson, who would be expected to lead the bill on the floor and whose buy-in was essential; Senate Judiciary Chairman Henry Ashurst, through whose committee the bill would have to pass; House Judiciary Chairman Hatton Sumners, who would handle the House side; Vice President Garner, whose Senate connections Roosevelt would soon desperately need; Postmaster General James Farley, the chief political operative of the administration; and the rest of the cabinet, who would learn of the proposal only on the morning of February 5. The secrecy was not accidental. Roosevelt had decided to present the bill as a fait accompli precisely because he expected resistance from his own party’s congressional leadership and wanted to deny them the time to organize opposition before he had set the public framing.

The bill that emerged from this insulated drafting process had three principal provisions, of which the first was the political bomb and the other two were administrative cover. The first provision authorized the president to appoint a new federal judge for every sitting judge who had reached the age of seventy and refused to retire within six months, with a cap of fifty new judgeships across the federal system and a sub-cap of six on the Supreme Court. With six justices then over seventy, the cap would in practice mean six new high-tribunal seats, raising the total to fifteen. The second provision authorized circuit-court intervention in federal trial-court business when calendars became congested. The third reorganized the federal judiciary’s administrative office. Cummings calculated, accurately, that the second and third provisions would attract bipartisan support from anyone who had practiced in the federal courts during the previous decade. He miscalculated, fatally, in believing that the second and third provisions would camouflage the first.

The drafting circle also produced the public message that Roosevelt would deliver on February 5. The message, around four thousand words, framed the entire problem as one of judicial efficiency. The federal courts, the message argued, were overburdened with cases and increasingly unable to render decisions within reasonable time. Aged judges, the message argued, lost the mental quickness necessary to manage modern litigation. The Supreme Court itself was straining to keep up with its docket. The remedy was additional judges to share the workload. The word “constitution” appeared only twice in the entire message. The Court’s recent decisions striking down New Deal legislation were not mentioned. The political objective of the bill, which any informed reader would understand, was deliberately kept out of the official rationale.

Cummings, Reed, and Roosevelt all knew the efficiency framing was thin. Several of the justices, particularly Brandeis (then eighty), were among the most efficient on the bench. Hughes ran the high tribunal’s calendar with notorious precision and dispatch. The argument that adding justices would speed work rather than slow it ran counter to the experience of every appellate court that had ever expanded its membership. Hughes himself would shortly demolish the efficiency claim in a letter that became the political turning point of the entire fight. But the framing was chosen anyway, for two reasons. First, the administration believed that an honest constitutional argument would be too easily portrayed as an attack on judicial independence. Second, Roosevelt enjoyed political theater and considered the McReynolds 1913 letter a delicious gotcha that would put conservatives on the defensive. Both calculations were wrong. The deceptive framing handed opponents a moral high ground they would never relinquish. The McReynolds gotcha proved irrelevant to senators who took the principle of an independent judiciary seriously regardless of its inventor.

Jeff Shesol’s 2010 study Supreme Power: Franklin Roosevelt vs. the Supreme Court, the definitive modern reconstruction of the entire episode, treats the secrecy of the drafting as the original sin of the campaign. Roosevelt’s refusal to consult his own party’s congressional leadership poisoned the relationship with Garner, Robinson, Ashurst, Sumners, and Wheeler before the bill was even introduced. William Leuchtenburg’s earlier study, The Supreme Court Reborn: The Constitutional Revolution in the Age of Roosevelt, is gentler on the procedural failures and harder on the substantive failure to make the constitutional argument honestly. Shesol and Leuchtenburg disagree on emphasis, but both agree that the chosen frame was a strategic catastrophe. So does Marian McKenna’s exhaustive Franklin Roosevelt and the Great Constitutional War, which traces the long-term damage to the relationship between the Roosevelt White House and the Senate Democratic conference. The three most thorough modern accounts converge on the same verdict: the bill was lost before it was introduced.

February 5, 1937: The Announcement That Stunned Washington

The morning briefing of cabinet and congressional leaders was scripted to minimize friction. Roosevelt was at his most charming. He distributed the message and bill, allowed ten minutes for reading, and answered no substantive questions. Sumners’s “cash in my chips” line was the first warning. Garner’s nose-holding gesture was the second. Robinson, a fiercely loyal Arkansan who had been Senate Democratic leader since 1933 and was Roosevelt’s chief congressional ally, listened in stunned silence, then walked back to the Capitol determined to defend the bill out of personal loyalty even as he privately doubted its merits and its political prospects. Robinson’s promised reward for carrying the fight, although Roosevelt was vague about specifics, was the next Supreme Court vacancy. That promise would shape the next five months of legislative maneuvering and would haunt Roosevelt when Robinson did not live to claim it.

The press release of the message at one in the afternoon set off a forty-eight-hour cascade of editorial reaction. The New York Times ran a banner headline. The Washington Post called the bill “the boldest stroke yet of a bold and assertive Executive.” The Chicago Tribune denounced it as “the destruction of constitutional government.” The St. Louis Post-Dispatch, normally a Roosevelt ally, ran an editorial headlined “Mr. Roosevelt Asks for a Stacked Court.” The phrase “court-packing,” used pejoratively in nineteenth-century political debates and rarely heard in the twentieth, was back in circulation by Saturday morning. Within a week it had displaced “judicial reorganization” in every newspaper outside the openly pro-Roosevelt few.

The political coalition against the bill assembled with remarkable speed. Hatton Sumners, the House Judiciary Chairman whose committee would normally have first crack at the legislation, simply refused to let the bill come before his committee. Roosevelt and Cummings, anticipating possible House trouble, rerouted the bill to the Senate Judiciary Committee through a procedural device, but the public signal of Sumners’s refusal was already out. Burton Wheeler of Montana, a progressive Democrat who had been Robert La Follette’s vice-presidential running mate on the Progressive ticket in 1924 and was widely respected on the left, announced his opposition on February 8. Wheeler’s defection mattered enormously because it broke the framing that opposition would come only from conservatives. Carter Glass of Virginia, the conservative Democratic eminence grise on monetary policy, followed shortly thereafter. By the second week of February, the anti-bill coalition included progressives (Wheeler, Borah, Hiram Johnson), conservatives (Glass, Bailey, Byrd), Republicans (with a few exceptions remaining silent on the calculation that Democrats should be allowed to argue), and a steadily growing number of moderate Democrats who privately favored some judicial response but doubted this particular vehicle.

Roosevelt’s first defense was the March 4 Victory Dinner address to the Democratic National Committee at the Mayflower Hotel. The speech ran twenty-eight minutes and was carried live on national radio. Its tone was triumphant and combative. “If we would keep faith with those who had faith in us,” Roosevelt told the audience of fifteen hundred party loyalists, “if we would make democracy succeed, I say we must act now!” The speech mentioned the bench only obliquely, framing the issue as a fight between the will of the people, expressed in the 1936 landslide, and an obstructive minority defending privilege. The address worked among the partisan loyalists in the room. It did not work in the Senate, where the very Democrats Roosevelt needed to convert were watching for any sign that the president understood their concerns about judicial independence. The Victory Dinner speech contained no such signal, and several wavering Democrats moved firmly into the opposition camp within the following week.

Eleanor Roosevelt’s role through February and March of 1937, which the Shesol account underscores, was significant and quietly skeptical. Her My Day column avoided the question, but in private letters and conversations she expressed reservations about both the substance and the tactics. The first lady’s stance, traced more fully in the institutional biography of first ladies as political operators, prefigured the influence she would exert on later New Deal questions. Frances Perkins, the Secretary of Labor and the only cabinet member with substantial legal training other than Cummings, was also quietly uncomfortable. She would later write in The Roosevelt I Knew that she had wished Roosevelt had taken the amendment route despite the longer timeline. Henry Stimson, Hoover’s Secretary of State and a Republican statesman Roosevelt courted intermittently, publicly opposed the bill in a Foreign Affairs essay that circulated widely on the Hill. The bipartisan establishment, by mid-March, had largely turned against the proposal.

The March 9 Fireside Chat and the Efficiency Frame’s Collapse

Roosevelt’s second major public defense was the fireside chat of Tuesday, March 9, 1937, the ninth in his series of radio addresses. The chat ran thirty-two minutes and was estimated to have reached fifty million listeners on the NBC, CBS, and Mutual networks. Its opening line was vintage Roosevelt confidence: “Last Thursday I described in detail certain economic problems which everyone admits now face the nation.” The “everyone admits” was characteristic. The chat then made what was, at last, a substantive constitutional argument, after a month of clinging publicly to the efficiency frame. Roosevelt walked listeners through the New Deal decisions of 1935 and 1936. He described the Court as having read its own economic and social predilections into the Constitution. He invoked the precedent of the Lochner era, when the bench had struck down progressive labor legislation for thirty years, and the Dred Scott decision of 1857, which he described as the most catastrophic judicial intervention in American history. He compared the present Court to a team of three horses: the executive, the legislature, and the judiciary. When one horse refused to pull in harmony with the other two, he argued, the nation could not move forward.

The horse metaphor was Roosevelt at his rhetorical best, and the constitutional argument was the one he should have made from the start. But the timing of the shift was politically devastating. By March 9, the efficiency framing had been the bill’s official rationale for five weeks. Opponents had spent five weeks ridiculing the efficiency framing as obviously pretextual. When the president finally pivoted to the honest constitutional argument, the pivot itself confirmed that the original rationale had been a smokescreen. Wheeler exploited the timing brilliantly in subsequent floor speeches. So did Glass. The president’s credibility as a straight shooter, which had been one of his most valuable political assets, took a substantial and durable hit.

The constitutional argument itself, separately considered, was also less powerful in 1937 than it would prove in retrospect. Roosevelt’s complaint that the Court read its policy preferences into the Constitution was true. But his proposed solution, adding six justices who would predictably vote his way, raised the obvious question of why the abuse of judicial review was bad when the conservative bench did it and acceptable when Roosevelt arranged for a liberal bench to do it. The principled response, which Roosevelt’s lawyers explored privately but never developed in public, would have been to propose constitutional procedural reforms (supermajority requirements for invalidation, term limits, jurisdictional adjustments) that would constrain whatever majority happened to be sitting. The expedient response Roosevelt actually chose, simply adding seats until his side won, did nothing to constrain future packing by future presidents. Republicans pointed this out repeatedly, and the point landed.

Cummings, in private correspondence with Reed in mid-March, conceded that the political situation had deteriorated since the announcement. Robinson, in private meetings with Roosevelt at the White House, urged consideration of a compromise that would add only two justices rather than six. Roosevelt, characteristically, rejected the compromise on the grounds that it would not convey the same constitutional message to the bench. The decision to refuse compromise in March was, in retrospect, the second major strategic error of the campaign, after the secrecy of the drafting. Two new justices on a closely divided Court would have been a politically deliverable victory and would have produced the same constitutional revolution that Roosevelt sought. By holding out for six, he won zero.

Hughes’s Counterstrike: The March 21 Letter to Wheeler

Charles Evans Hughes was seventy-four years old in March 1937 and at the height of his juridical reputation. He had served on the high tribunal from 1910 to 1916 (resigning to run for president against Wilson, whom he came within a few thousand California votes of beating), as Harding’s and Coolidge’s Secretary of State, as a judge on the Permanent Court of International Justice, and since 1930 as Chief Justice. He was, by common consent, the most respected lawyer in the United States. He was also a savvy political operator who understood that the bench’s strongest defense against Roosevelt’s bill was to demolish the official efficiency rationale on which Roosevelt had publicly staked the case. Hughes set out to do exactly that.

On Sunday, March 21, 1937, Hughes met privately at his home with Brandeis (the senior associate justice and a Wilson appointee whom Roosevelt could not credibly attack as a conservative obstructionist) and Van Devanter (the senior conservative). The three justices reviewed the docket statistics, the speed of the bench’s recent decisions, and the procedural reforms the Chief had implemented since 1930. They agreed on the content of a letter Hughes would send to Burton Wheeler, the leader of Democratic opposition in the Senate. Brandeis signed off on every paragraph. Van Devanter signed off on every paragraph. Hughes signed it as Chief Justice and added a note that the letter spoke for the bench, although he had not formally polled all nine justices.

The letter, delivered to Wheeler late on Sunday March 21 and read by Wheeler to the Senate Judiciary Committee on Monday March 22, ran around eight pages and was a methodical demolition of the efficiency claim. Hughes documented that the high tribunal had heard arguments in 717 cases in the 1935 term and disposed of all of them before adjournment. The docket was fully current. Petitions for certiorari were being granted or denied within weeks of filing. The bench had not been behind on its work at any point in the past five years. Adding new justices, the Chief noted, would slow rather than speed work because each new member would have to be briefed on every pending case and would participate in every conference. The bench worked best with nine, Hughes argued, because each justice could read every brief and participate fully in every conference. A bench of fifteen would inevitably balkanize, with subgroups handling subsets of cases and the collegial deliberative process degrading. The letter closed with a careful note that Hughes had not consulted all of his colleagues but believed his factual claims about the bench’s actual workload were uncontroversial among the justices.

The political effect of the letter was devastating. Hughes had been the bench’s most credible centrist, willing to vote with the liberals on Tipaldo and the Bituminous Coal Act dissents and against the executive on Schechter and Humphrey’s Executor. His letter could not be portrayed as a partisan defense by an entrenched reactionary. The fact that Brandeis had signed off on every paragraph eliminated the framing of the letter as a conservative cri de coeur. Wheeler, reading the letter aloud in committee with theatrical timing, drove the point home: the bench’s most respected justices, including its senior Wilson Democrat, had unanimously rejected the rationale on which Roosevelt was selling the bill. From Monday March 22 forward, the efficiency framing was effectively dead in any informed political conversation in Washington.

Cummings’s response was to argue, both in newspaper interviews and in private memoranda to Roosevelt, that Hughes had violated the separation of powers by intervening in a legislative debate. The argument was technically defensible. The Chief Justice had indeed waded into political waters. But the argument also conceded the central factual point of the letter: the workload claim was false. Once the workload claim was false, the bill’s official rationale was false, and the only remaining argument for the bill was the constitutional one Roosevelt had begun making on March 9. That argument had to fight on its own merits, without the protection of the efficiency framing. It would prove unable to win the fight.

The Switch in Time: West Coast Hotel v. Parrish

On Monday, March 29, 1937, eight days after Hughes’s letter, the bench handed down West Coast Hotel Co. v. Parrish, a five-to-four decision upholding a Washington State minimum wage law for women. The majority opinion, written by Hughes, distinguished Adkins v. Children’s Hospital (1923) and Morehead v. New York ex rel. Tipaldo (1936) so narrowly that it effectively reversed both. Owen Roberts, who had joined the Four Horsemen in Morehead nine months earlier to strike down a substantively identical New York minimum wage law, now joined Hughes, Brandeis, Stone, and Cardozo to uphold the Washington statute. The shift was extraordinary. A five-to-four decision against state minimum wage laws in June 1936 had become a five-to-four decision for them in March 1937, on essentially the same facts, with the only change being the position of Owen Roberts.

The phrase “a switch in time saves nine,” coined by humorist Cal Tinney shortly after the decision, captured the popular interpretation: Roberts had switched his vote to save the bench from packing. The interpretation was so politically resonant that it became the standard reading of the entire episode for decades. The trouble with the interpretation is that the actual vote in West Coast Hotel had been taken at conference on December 19, 1936, six weeks before Roosevelt announced the bill. Justice Stone was ill and not present at that conference, leaving the original vote at four-to-four. The case was carried over for re-argument with Stone present, and the final five-to-four decision in March 1937 reflected the December votes plus Stone’s vote rather than any post-announcement switch by Roberts. On the strict chronology, Roberts had switched in December 1936 between the Tipaldo decision of June 1936 and the West Coast Hotel conference of December 1936, well before the court-packing bill existed.

Leuchtenburg’s account, which is the most influential modern reconstruction of this exact question, takes the December timing seriously but argues that the political pressure on the bench had been building throughout the fall of 1936 even before the specific court-packing bill was announced. Roosevelt’s landslide reelection had been a clear public verdict on the New Deal. The 1936 Democratic platform had explicitly criticized the bench. Roberts could read political tea leaves as well as anyone in Washington. The December 1936 switch was not a response to the specific February bill, but it was plausibly a response to the political climate that produced both the bill and the landslide that produced the climate. Shesol agrees on the chronology but emphasizes that the public release of the decision on March 29, 1937, in the middle of the court-packing fight, was certainly understood by everyone in Washington as a strategic capitulation by the bench, regardless of when the actual vote had occurred. The political reality and the chronological reality, on this reading, diverge.

What is clear, on either reading, is that West Coast Hotel was a doctrinal watershed. The decision repudiated the freedom-of-contract framework that had governed substantive due process since Lochner v. New York (1905). It accepted that the state’s police power extended to economic regulation of wages and hours. It revived the rational-basis standard of review for economic legislation that the Lochner-era jurisprudence had largely abandoned. From West Coast Hotel forward, no major piece of New Deal or progressive economic legislation would be invalidated on substantive due process grounds. The Lochner era, as a coherent body of constitutional doctrine, ended on March 29, 1937. The pattern of bench reversal under political pressure, examined more systematically in the broader institutional pattern of Supreme Court timing across presidencies, would shape every subsequent constitutional fight.

For Roosevelt, the West Coast Hotel decision was both vindication and threat. Vindication because the bench had moved, exactly as he wanted. Threat because the bench’s movement undercut the political case for his bill. If the high tribunal could be persuaded to read the Constitution sensibly without packing, the urgency of the legislative remedy collapsed. Senator George Norris of Nebraska, an old progressive who had been edging toward support of the bill, used West Coast Hotel as the rationale for backing off. Several other wavering Democrats followed Norris. The bill’s vote count, which had been ambiguous in early March, began to tilt against the administration by the first week in April.

Jones and Laughlin: The Wagner Act Survives

On Monday, April 12, 1937, two weeks after West Coast Hotel, the bench delivered National Labor Relations Board v. Jones and Laughlin Steel Corp., a five-to-four decision upholding the National Labor Relations Act of 1935. Hughes again wrote for the majority, joined again by Brandeis, Cardozo, Stone, and Roberts. The decision constituted the most significant expansion of the Commerce Clause in American constitutional history to that date. Hughes wrote that the Wagner Act’s regulation of industrial labor relations fell within the commerce power because labor unrest in major industries had a substantial effect on interstate commerce, even when the immediate labor disputes were intrastate. The reasoning effectively abandoned the direct-versus-indirect-effects distinction that the bench had used in Schechter and Carter Coal to confine federal regulatory authority. The Commerce Clause, after Jones and Laughlin, reached anything that substantially affected interstate commerce, whether the immediate activity itself was interstate or not.

The doctrinal expansion produced by Jones and Laughlin would not be fully tested until Wickard v. Filburn (1942), where the bench (by then thoroughly transformed by Roosevelt appointments) would extend the Commerce Clause to cover a single farmer’s wheat consumed on his own farm. But the structural turn was made in April 1937. Every subsequent expansion of federal regulatory power for the next sixty years, including the Civil Rights Act of 1964 (which was sustained under the Commerce Clause rather than the Fourteenth Amendment in Heart of Atlanta Motel v. United States), traces its constitutional foundation to Jones and Laughlin. The labor regulation, antitrust expansion, environmental regulation, occupational safety, and consumer protection statutes of the New Deal and Great Society eras would all rest, ultimately, on the doctrinal turn taken in April 1937 by Hughes, Roberts, Brandeis, Stone, and Cardozo.

Politically, Jones and Laughlin completed what West Coast Hotel had begun. The combination of the two decisions meant that the bench had now upheld the two centerpieces of the second New Deal (the minimum wage principle and the National Labor Relations Act) and had effectively confessed that its earlier resistance to New Deal legislation was no longer operative. The constitutional revolution Roosevelt had set out to compel in February was substantially complete by mid-April. Senator Wheeler, leading the opposition to the bill, made the political case explicit in a Senate floor speech on April 14: “Why pack the Court? The Court has packed itself.” The line circulated widely. Within two weeks, Senate vote-counters were reporting to Robinson and to Vice President Garner that the bill no longer had the votes for passage and could probably not be saved even with substantial amendments.

Van Devanter Retires: May 18, 1937

Congress had passed the Supreme Court Retirement Act on March 1, 1937, in the middle of the broader court-packing fight. The statute, technically titled the Sumners-McCarran Act, extended to Supreme Court justices the same retirement benefits that had been available to lower federal judges since 1869. Justices reaching age seventy with ten years of service could retire on full salary. The previous arrangement had been that justices who retired forfeited their salary, which had been a major factor in the persistence of aged justices on the bench: Holmes had served until ninety in part because he could not afford to retire. Van Devanter, who was seventy-seven and had served twenty-six years, suddenly had a financial path to retirement that did not exist before March 1937. Pierce Butler and George Sutherland faced the same arithmetic.

On Tuesday, May 18, 1937, Van Devanter wrote to Roosevelt informing him of his decision to retire effective June 1. The vacancy was the first Roosevelt would be allowed to fill in over four years in office, and its arrival coincided with the moment when his court-packing bill was visibly dying in the Senate. Van Devanter’s departure was widely interpreted as proof that the bill was no longer needed. If conservative justices were now going to step aside under the new retirement statute, Roosevelt could remake the bench through ordinary attrition, without recourse to expansion. Wheeler, Glass, and the rest of the opposition immediately deployed Van Devanter’s letter as the closing argument: the bench was correcting itself, the older justices were going home, and the constitutional emergency that Roosevelt had invoked in February had passed.

The Van Devanter vacancy created an immediate political crisis for Roosevelt over his commitment to Joe Robinson. Robinson had been promised the seat in exchange for leading the bill on the Senate floor. Robinson, sixty-five, had spent his entire adult life seeking the Supreme Court seat that Van Devanter’s resignation now made available. But Roosevelt, characteristically, hesitated. Robinson was a moderate Arkansas Democrat whose voting record on labor and civil rights questions was uncertain. The first New Deal vacancy was politically too valuable to spend on a moderate when the obvious candidate could be a movement liberal. Roosevelt delayed the nomination, asking Robinson to keep fighting for the broader court-packing bill on the Senate floor while the appointment question remained open.

Robinson, hurt but loyal, accepted the delay. He continued to lead the floor fight through late May and early June, even as the bill’s prospects became increasingly hopeless. Cabinet members and close advisors privately urged Roosevelt to make the Robinson nomination formal, both to honor the commitment and to mollify the Senate Democrats whose votes Roosevelt still needed on other parts of his legislative program. Roosevelt declined. The pattern of withholding the promised reward while extracting more loyalty service was characteristic of Roosevelt’s executive management style and was a recurring source of resentment in the Senate. The delay would matter more than anyone realized when Robinson’s body gave out under the strain six weeks later.

The Senate Judiciary Report: June 14, 1937

The Senate Judiciary Committee, chaired by Henry Ashurst, finally voted on the court-packing bill on Monday, June 14, 1937. The vote was ten to eight against. Seven Democrats joined the committee’s three Republicans to recommend that the bill be rejected. The majority report, drafted by Wheeler, ran twenty-three pages and was the most unsparing repudiation of a presidential legislative initiative ever issued by a committee of the president’s own party. Selected phrases from the report:

“It is a measure which should be so emphatically rejected that its parallel will never again be presented to the free representatives of the free people of America.” That sentence was the rhetorical climax of the report and was widely reprinted as proof that the Senate Democratic caucus had moved decisively against the administration. The report continued: “We recommend the rejection of this bill as a needless, futile and utterly dangerous abandonment of constitutional principle.” And further: “It would not banish age from the bench, nor abuse from power. It would not, by establishing precedent, increase the number of justices, but would set a precedent which any future Congress, working its will, might invoke to enlarge or diminish the court as suited the temper of the time.”

The committee report’s argumentative force came from its emphasis on precedent. Wheeler and the committee Democrats focused on the long-term institutional consequences of allowing court expansion to become a routine political weapon. Roosevelt’s six new seats today, the report argued, would become a future Republican president’s six new seats tomorrow. The bench would cease to be a counter-majoritarian institution and would become an extension of whichever party controlled the executive and the Senate at any given moment. The argument was sophisticated and was the most damaging single piece of writing produced against the bill at any point in the five-month fight.

Roosevelt’s response was political theater rather than substantive engagement. He summoned Robinson and several committee Democrats to Hyde Park in late June for what he described as an opportunity for the senators to enjoy the country air. The visit was awkward. Robinson, exhausted, stayed for two days and returned to Washington with no clear instructions on whether to push for a floor vote or to accept defeat with a face-saving substitute. The other senators sensed that Roosevelt had no plan beyond continued personal pressure. The committee report’s verdict, in any case, was now a public document, and the Senate’s calendar for July would dispose of the bill one way or another.

Joseph Robinson’s Death: July 14, 1937

On Tuesday, July 13, 1937, Joseph Robinson delivered a two-hour Senate floor speech in defense of the court-packing bill. The speech was the longest he had given in years. He was sixty-five and had been complaining of chest pains for several weeks. He had not received the Supreme Court nomination that Roosevelt had promised in February. The Senate Judiciary Committee, on which several of his closest personal friends served, had repudiated his president. The bill he was carrying had no clear path to passage. Robinson finished the speech, walked back to his apartment at the Methodist Building on Capitol Hill, dined alone, and went to bed.

Robinson’s housekeeper found him on the bathroom floor the next morning. He had died of a heart attack during the night. He was alone. He had not received last rites. His Senate colleagues learned of the death from a Capitol Hill messenger around eight in the morning. The funeral was held in Little Rock, Arkansas, four days later, with most of the Senate attending. Vice President Garner, who had returned to Washington from a private trip to attend the funeral train, took the occasion to begin lobbying senators individually against the court-packing bill, on the explicit ground that Robinson’s death made continued legislative combat indecent. The lobbying campaign, by all accounts, was effective. Several Democrats who had been quietly committed to vote for the bill as a personal favor to Robinson now found themselves freed from any obligation.

Roosevelt, characteristically, did not attend the funeral. The decision was bureaucratically defensible (presidential travel to senatorial funerals was rare in 1937) and politically catastrophic. The Senate, then as now, ran on personal relationships and small symbolic gestures. Garner, on the funeral train back from Little Rock, methodically counted votes and reported to Roosevelt by telephone that the bill could not pass. Roosevelt refused to accept the count. He sent personal letters to several wavering senators arguing for one final effort. The letters were ignored. By July 19, even Cummings and Reed were urging Roosevelt to accept defeat and salvage what could be salvaged.

The cause of Robinson’s death (heart failure under physical and emotional strain) was beyond Roosevelt’s control. The political consequences of Robinson’s death were partly within his control and partly the result of the administration’s accumulated mistakes. Robinson was widely admired in the Senate Democratic caucus as a man who had taken personal political risk to defend a bill he himself had reservations about, in exchange for a promised reward that Roosevelt had repeatedly delayed delivering. His death came to symbolize, for Senate Democrats, the broader pattern of Roosevelt’s relationship with his own party’s congressional leadership: extracting maximum loyalty service in exchange for vague and often unfulfilled promises. The symbolic damage outlasted the immediate question of the court-packing bill and would shape Roosevelt’s legislative relationships through the rest of his second term.

July 22: The Bill Goes Down 70 to 20

On Thursday, July 22, 1937, the Senate voted seventy to twenty to recommit the court-packing bill to the Senate Judiciary Committee. Recommittal was, in practice, the same as defeat. The committee, which had voted ten to eight against the bill in June, was not going to reverse itself in July. The vote signaled the formal end of the court-packing fight as a serious legislative proposition. The vote count was the most lopsided defeat of a major presidential initiative since the rejection of the Treaty of Versailles in 1919 and 1920, which Wilson had pushed against the same kind of opposition from his own party leadership covered in the analysis of the Wilson Versailles loss. Twenty-one Democrats voted against their president. Of the twenty senators who voted for recommittal opposition (that is, against killing the bill), only one or two would still have voted for the original six-justice version had it come to a final passage vote.

Congress did pass, on August 24, 1937, the Judicial Procedures Reform Act of 1937, a face-saving measure containing the second and third provisions of the original Cummings bill: the circuit-court intervention authority and the administrative reorganization. The provisions were genuinely useful procedural reforms and remain in modified form in current federal judicial administration. But the political reality was that Roosevelt had been beaten by his own party on his signature institutional initiative, in a roll call that left no ambiguity about who had defected and why. The vote of seventy to twenty would be referenced in Senate debates for decades whenever the question of expanding the high tribunal was raised. The 1937 vote became the institutional precedent that, for almost ninety years, made statutory expansion of the bench politically unthinkable.

Roosevelt issued no public statement on July 22. He did, several weeks later, write a public letter to Senator Logan of Kentucky claiming victory in the broader struggle over judicial supremacy. The letter argued that Roosevelt had achieved his constitutional objectives through the West Coast Hotel and Jones and Laughlin decisions, the Van Devanter retirement, and the upcoming Hugo Black nomination, without needing to pass the original bill. The argument was substantively correct but tonally tin-eared. Roosevelt was claiming a strategic win while standing in the ruins of a tactical disaster he himself had engineered. The Logan letter circulated as proof, for many in Washington, that Roosevelt did not fully grasp the political damage he had done to his own coalition.

Did Roosevelt Actually Lose? The Strategic Verdict

The strategic question of whether Roosevelt won or lost the broader court fight has divided historians for seventy years. Three positions dominate the historiography. The Leuchtenburg position, developed in The Supreme Court Reborn (1995), is that Roosevelt won the war by losing the bill. The constitutional revolution he sought (overruling Lochner-era substantive due process, expanding the Commerce Clause, accepting the New Deal regulatory state) arrived through West Coast Hotel, Jones and Laughlin, Steward Machine, and Helvering v. Davis, all decided in 1937 by the bench Roosevelt had threatened to pack. The Roosevelt appointees of 1937 through 1943 (Black, Reed, Frankfurter, Douglas, Murphy, Byrnes, Jackson, Rutledge, and Stone elevated to chief) completed the revolution doctrinally. The Wagner Act, Social Security, and Fair Labor Standards Act all survived, were extended, and became permanent fixtures of American governance. By 1944, every substantive constitutional question Roosevelt had cared about in February 1937 had been resolved in his favor. On the Leuchtenburg view, the threat had done the work, and the threat would not have been credible without an actual bill on the table.

The Shesol position, developed in Supreme Power (2010), is that Roosevelt lost more than he gained. The court-packing fight cost him the unity of his party at the moment when the 1937 to 1938 economic recession would require maximum legislative effort. The Conservative Coalition of Southern Democrats and Republicans, which would block civil rights, labor, and economic legislation for the next three decades, formed in direct response to the court fight. The Senate Democratic caucus learned to defy the White House and never fully relearned the habit of automatic loyalty. The Fair Labor Standards Act of 1938 only barely passed and required compromise that would have been unnecessary in 1935 or 1936. The Reorganization Bill of 1938 was defeated outright. The wage and hour amendments Roosevelt sought to the Social Security Act were watered down. The price of getting West Coast Hotel and Jones and Laughlin, Shesol argues, was the destruction of Roosevelt’s ability to extend the New Deal into civil rights, comprehensive health care, full employment, and the other items on the unfinished progressive agenda.

The McKenna position, developed in Franklin Roosevelt and the Great Constitutional War (2002), splits the difference. McKenna agrees with Leuchtenburg that Roosevelt obtained his constitutional objectives. McKenna agrees with Shesol that the political cost was substantial and long-lasting. McKenna’s distinctive contribution is to emphasize that the political cost was not principally about the immediate 1937 to 1938 legislative season but about the long-term relationship between the executive and the legislative branches. Roosevelt’s defeat broke the spell of presidential omnipotence that the 1936 landslide had created. Subsequent presidents could not, after 1937, count on routine deference from Congress on any institutional question. The check on presidential expansion did not come from the formal constitutional structure but from the Senate Democratic caucus’s discovery, in 1937, that it could defy a popular president and survive politically. That discovery is what makes the court-packing fight a permanent landmark in the history of the office, in McKenna’s reading.

David Kennedy’s Freedom from Fear (1999), which is the most authoritative single-volume account of the entire 1929 to 1945 period, is closer to the Leuchtenburg position on the constitutional question but echoes Shesol on the political damage. Kennedy argues that the political damage was the price Roosevelt was always going to pay for the constitutional victory, and that the price was acceptable given the stakes. Smith’s FDR (2007) is more sympathetic to Roosevelt personally and emphasizes that the political damage was inflicted by Roosevelt’s own poor handling of the campaign rather than by any inherent flaw in the constitutional objective. The historiographical convergence, across all five major modern treatments, is that the constitutional revolution was real and substantial, the political damage was real and substantial, and the question of whether the trade was worth it is genuinely contested.

The honest verdict on the strategic question requires distinguishing what Roosevelt sought from what Roosevelt got. He sought, in February 1937, a six-seat expansion of the bench that would have created a 9-6 (or potentially worse for the conservatives, depending on retirements) Roosevelt majority on the high tribunal. He got, by 1943, a 9-0 Roosevelt-appointed bench through ordinary attrition combined with the retirement statute’s incentives. The 9-0 outcome was a more thorough institutional victory than the 9-6 he sought, achieved at lower constitutional cost (the bench’s size was preserved at nine, the institutional norm against packing held, the principle of judicial independence was reinforced by Roosevelt’s defeat). On those metrics, Roosevelt won. He sought, in February 1937, the rapid invalidation of Lochner-era substantive due process and the expansion of the Commerce Clause. He got both by April 1937, before he lost the legislative fight. On those metrics, Roosevelt also won. He sought, in February 1937, to demonstrate that the executive could overcome resistant judicial supremacy. He demonstrated the opposite: the legislative branch can constrain the executive even on an institutional question backed by an overwhelming electoral mandate. On that metric, Roosevelt lost.

The composite verdict is that Roosevelt won the constitutional war and lost a major political battle that mattered chiefly for what it taught Congress about its own power. Both outcomes were genuinely consequential. The constitutional revolution shaped the next sixty years of American governance. The political defeat shaped the next eight years of Roosevelt’s presidency. Both effects are visible in the long story of the second-term presidential curse pattern, of which the 1937 court fight is the cleanest single case.

Nine Appointments Without Packing: 1937 to 1943

Roosevelt named nine justices between August 1937 and January 1943, without expanding the size of the bench. The sequence: Hugo Black (August 1937, replacing Van Devanter), Stanley Reed (January 1938, replacing Sutherland, who retired after Van Devanter’s example showed the new retirement statute could be used without forfeiture), Felix Frankfurter (January 1939, replacing Cardozo, who died in July 1938), William O. Douglas (April 1939, replacing Brandeis, who retired at eighty-three in February 1939), Frank Murphy (January 1940, replacing Butler, who died in November 1939), Harlan Fiske Stone elevated to Chief Justice (July 1941, after Hughes retired in June 1941, with the seat itself going to Stone rather than to an outside appointee), James F. Byrnes (July 1941, replacing McReynolds, who finally retired in February 1941), Robert H. Jackson (July 1941, replacing Stone’s old associate justice seat), and Wiley Rutledge (February 1943, replacing Byrnes, who left the bench to direct economic mobilization).

The composition of the bench by the start of 1943 was: Stone (chief), Black, Reed, Frankfurter, Douglas, Murphy, Roberts (Hoover appointee, the last remaining non-Roosevelt justice), Jackson, and Rutledge. Eight of nine were Roosevelt appointees. Roberts would retire in 1945, giving Truman the opportunity to fill the last non-Roosevelt seat. By the time Roosevelt died in April 1945, every active justice except Roberts owed his seat to him, and the bench’s voting patterns had reflected that reality for years. The Wickard v. Filburn decision of 1942, which extended the Commerce Clause to encompass purely intrastate non-commercial activity, would not have been imaginable in 1936 and would not have been possible without the institutional revolution Roosevelt’s appointees completed.

The political path by which Roosevelt achieved the 9-0 outcome (rather than the 9-6 outcome he had sought through court-packing) was the retirement statute combined with mortality and his own electoral durability. Van Devanter retired in 1937 because the statute made it financially possible. Sutherland followed in 1938 for the same reason. Cardozo died on the bench in 1938. Brandeis retired at eighty-three in 1939, also enabled by the retirement statute. Butler died in 1939. Hughes retired in 1941. McReynolds, the most stubborn of the holdouts, finally retired in 1941 at seventy-eight. The retirement statute, passed in March 1937 in the middle of the court-packing fight, turned out to be more important than the court-packing bill itself. It removed the financial barrier to retirement and allowed conservative justices to step aside without sacrificing income. The statute was passed almost unanimously and provoked no constitutional objection. The packing bill provoked maximum constitutional objection and would have been less effective even if it had passed.

The contrast between the popular bipartisan retirement statute and the disastrous packing bill is one of the cleanest available case studies in legislative strategy. The retirement statute addressed the underlying problem (entrenched aged justices) with a politically neutral mechanism (financial incentive to leave). The packing bill addressed the same underlying problem with a politically maximalist mechanism (executive appointment of replacements without retirement). The neutral mechanism produced the same outcome (replacement of the aged conservatives) at a fraction of the political cost. Roosevelt could have had every replacement he eventually got, on roughly the same timeline, by signing the retirement statute and then waiting. He chose the maximalist path, lost the legislative fight, and received the neutral mechanism’s benefits anyway through the retirement statute’s slow action. The lesson, repeatedly drawn by subsequent administrations contemplating institutional reform, is that incentive structures often outperform direct power grabs.

The Verdict: Tactical Defeat, Strategic Victory, Lasting Institutional Cost

The verdict on the entire 1937 episode, drawing on the convergent evidence in Shesol, Leuchtenburg, McKenna, Kennedy, and Smith, breaks into three claims that all need to be held simultaneously. First, Roosevelt suffered the most consequential tactical defeat of his presidency. The court-packing bill failed in the Senate by a margin of seventy to twenty, with twenty-one Democratic defections. The political capital expended on the bill exceeded any single executive request of the New Deal era, and the capital was not recovered. Roosevelt would never again propose an institutional reorganization of comparable scope. The Reorganization Bill of 1938, the second-term legislative initiative most closely modeled on the executive-strengthening logic of the 1937 effort, would also be defeated by the same Conservative Coalition that formed during the court fight.

Second, Roosevelt achieved his substantive constitutional objectives. The Lochner era ended in March 1937 with West Coast Hotel. The Commerce Clause was reinvented in April 1937 with Jones and Laughlin. The Social Security Act survived the same year in Steward Machine and Helvering v. Davis. The Wagner Act survived in Jones and Laughlin and Associated Press v. NLRB. By the end of 1937, the bench had upheld every major statute of the New Deal that came before it. By 1943, the bench was nine-eighths a Roosevelt creation, and its voting patterns reflected the institutional transformation that Roosevelt had set out to compel in February 1937. The threat had done the work.

Third, the institutional cost of the fight extended beyond Roosevelt’s presidency and remains observable in the structure of American governance. The Conservative Coalition that formed during the court fight blocked civil rights legislation through the 1940s and 1950s, blocked the expansion of the Fair Labor Standards Act, blocked the full extension of Social Security to agricultural and domestic workers (the same workers excluded by the original 1935 statute through the segregationist compromises that made the bill possible), and shaped the Truman, Eisenhower, and Kennedy legislative agendas. The Senate Democratic caucus learned, in 1937, that it could defy a popular president on institutional questions and survive electorally. That lesson has shaped every subsequent presidency. No president since Roosevelt has proposed a substantial expansion of the high tribunal. The institutional norm against packing held until 2020s academic and activist proposals to revisit it, and even those proposals have not produced legislative action.

The composite verdict, then, is that Roosevelt’s court-packing campaign was a tactical defeat of historic proportions, a strategic victory of historic proportions, and an institutional precedent that has constrained the office of the presidency for nearly ninety years. All three claims are true. The first two are easy to accept; the third requires acknowledging that Roosevelt’s defeat, viewed across nine decades, may have been the most institutionally valuable single legislative defeat in American history. The norm against packing held precisely because Roosevelt failed.

Legacy: The Norm That Held and the House Thesis Test

The series’s house thesis (the modern presidency was forged in four crises, every emergency power created in those crises outlived the emergency, and every president since inherits an office designed for conditions that no longer exist) finds in the 1937 court-packing fight a precise case study and a precise counter-case. The court fight is a case study in the sense that the constitutional revolution of 1937 expanded federal regulatory power for the Depression emergency and the expansion outlasted the emergency: the Commerce Clause doctrine of Jones and Laughlin still governs federal regulatory authority in the 2020s, eighty-six years later, in conditions utterly unlike the conditions of 1937. The expanded federal regulatory state of the late New Deal is precisely the kind of emergency-power-outlives-emergency outcome the house thesis predicts.

The court fight is a counter-case in the sense that Roosevelt’s specific institutional grab (six new seats on the bench) was rejected by his own party, and the institutional norm against packing that emerged from his defeat has held against every subsequent president for almost nine decades. The house thesis is generally a story of executive aggrandizement that succeeds and persists. The 1937 court fight is a rare and important counter-case in which an executive aggrandizement was attempted, was defeated by the constraining branch (Congress, in the form of the Senate Democratic caucus), and the constraint became durable institutional precedent. The thesis does not predict that every executive grab will succeed. It predicts that the surviving grabs will outlive their original justifications, which is exactly what happened to the Commerce Clause doctrine and the regulatory state.

The deeper lesson of the 1937 fight, threaded across the series, is that institutional constraints on the presidency operate through politics rather than through formal text. The Constitution is silent on the size of the bench. Statutes had set it at six, then five, then six again, then seven, then nine, then ten, then nine again. Nothing in the constitutional text prohibited Roosevelt’s proposal. The constraint that defeated him was political and norm-based: senators of his own party concluded that allowing court expansion would set a precedent that future presidents would abuse, and they were willing to pay the political price of defying a popular president on a single institutional question to preserve the precedent against packing. The norm survived because senators chose to defend it. Compare this to the durability of Roosevelt’s other institutional innovations, examined in detail in the third-term decision article and the Japanese internment article: the two-term tradition was an informal norm that Roosevelt broke without effective political constraint and required a formal amendment (the Twenty-Second) to restore. Internment was an executive action that ran without effective political constraint and required a Supreme Court decision (Korematsu, never formally overruled) and a 1988 congressional apology to address. Court-packing was a legislative proposal that the Senate of his own party blocked through the simple act of voting against it.

The differential outcomes (legislative effort blocked, executive action and norm-breaking succeed) illuminate where the modern presidency’s actual constraints lie. They are not in the formal separation-of-powers structure that high school civics textbooks describe. They are in the political coalitions, partisan dynamics, and institutional norms that the formal structure presupposes but does not create. When those informal supports work (as in 1937 against court-packing), the presidency is constrained. When they do not work (as in 1940 on the two-term tradition or 1942 on internment), the presidency runs unconstrained. The lesson is uncomfortable for anyone who hopes that the constitutional text alone will hold a determined executive in check. The text, by itself, holds nothing. Politics holds, or it does not.

Roosevelt himself, in private correspondence with Cummings several years after the 1937 defeat, described the campaign as the best lesson he had received in the limits of presidential power. He never again proposed an institutional reorganization of comparable scope. The Reorganization Bill of 1938 was a substantial executive-branch initiative, but it was a comparatively narrow proposal to give the president more control over the existing federal bureaucracy, not to reshape an entire coordinate branch. After 1938, Roosevelt focused executive expansion on foreign policy and wartime mobilization, where the constitutional text gave the office more room and where political constraints were weaker. The court-packing defeat redirected the trajectory of his presidency, and through that redirection it shaped the institutional choices of every subsequent president from Truman forward.

A final point about the framing of the 1937 episode in popular memory deserves emphasis. Standard textbook treatments often describe the fight as a straightforward defeat for an overreaching president, with the bench’s independence vindicated and the constitutional order preserved. That framing is partial at best and misleading at worst. The bench whose independence was supposedly vindicated had, within weeks of the announcement, executed the most consequential ideological reversal in its history, on substantive due process and on the Commerce Clause, in apparent response to the political pressure the bill represented. The constitutional order that was supposedly preserved was simultaneously transformed beyond recognition by the same justices the bill had targeted. The president who supposedly lost obtained, through ordinary appointments enabled by the retirement statute he signed in March 1937, every institutional outcome the packing bill would have delivered, on roughly the same timeline. The honest reading is that the bench and the president fought to a draw on the legislative question and that the bench conceded on the underlying constitutional questions while preserving its institutional independence as a face-saving consolation. The popular narrative of a clean defeat for executive overreach obscures the actual transaction, in which both sides won something and both sides lost something, and the constitutional revolution proceeded regardless.

Frequently Asked Questions

Q: What was Franklin Roosevelt’s court-packing plan?

The court-packing plan, formally titled the Judicial Procedures Reform Bill, was a piece of legislation Roosevelt sent to Congress on February 5, 1937, that would have authorized the president to appoint one new federal judge for every sitting judge over the age of seventy who refused to retire within six months. The cap on new Supreme Court appointments was six, which would have raised the size of the high tribunal from nine justices to fifteen if all six of the sitting justices over seventy stayed on. The bill also contained two less controversial provisions: authority for circuit-court intervention in congested federal trial courts, and administrative reorganization of the federal judiciary. The first provision was the political centerpiece and the reason the bill became known as the court-packing plan. The administration officially framed the entire package as a judicial efficiency reform, although the underlying political objective of producing a bench friendlier to New Deal legislation was understood by every informed observer in Washington from the day the bill was introduced.

Q: Why did Roosevelt propose adding justices to the Supreme Court?

The proposal responded to two years of accumulating decisions in which the bench had struck down major New Deal legislation, including the National Industrial Recovery Act (Schechter Poultry, 1935), the Agricultural Adjustment Act (Butler, 1936), the Bituminous Coal Conservation Act (Carter v. Carter Coal, 1936), and state minimum wage laws (Morehead, 1936). Roosevelt and his Attorney General Homer Cummings concluded that the existing bench, composed largely of Republican appointees from the 1920s, would continue to block the New Deal program indefinitely unless its composition was changed. A constitutional amendment would take three to five years to ratify and might fail. Waiting for vacancies through retirement or death was unreliable because the conservative justices were widely expected to refuse to retire. Statutory expansion of the bench, the third option, had historical precedent (the size of the high tribunal had been changed seven times before 1869) and could be accomplished by ordinary legislation. Roosevelt chose the statutory route, in part because Cummings argued from the McReynolds 1913 precedent that the proposal could not be credibly attacked as a violation of judicial independence.

Q: How many justices would have been added if the bill had passed?

The bill would have authorized up to six new Supreme Court seats, raising the total from nine to fifteen, but only if the existing justices over seventy refused to retire. In February 1937, six of the nine sitting justices were over seventy: Brandeis (80), Hughes (74), Van Devanter (77), McReynolds (75), Sutherland (74), and Butler (70). All six could theoretically have triggered a new appointment by remaining on the bench. In practice, the bill’s structure was understood by everyone as a mechanism to add six liberal justices who would consistently vote to uphold New Deal legislation. The size of fifteen would have been roughly double the bench’s current size in 1937 and would have been the largest single statutory expansion of the high tribunal in American history. The previous largest expansion had been the addition of a tenth seat in 1863, repealed by Congress in 1866 to prevent Andrew Johnson from filling vacancies. The 1869 statute restored the size to nine, where it has remained.

Q: Did the court-packing bill ever pass?

No. The Senate voted seventy to twenty on July 22, 1937, to recommit the bill to the Senate Judiciary Committee, which had previously voted ten to eight to recommend rejection. Recommittal in this context was functionally equivalent to defeat. The bill never came to a final passage vote. A face-saving substitute, the Judicial Procedures Reform Act of 1937, passed Congress on August 24, 1937, and contained the two non-controversial provisions of the original bill (the circuit-court intervention authority and the administrative reorganization) without the Supreme Court expansion. The substitute statute is still in effect in modified form. The defeat of the court-packing bill is one of the most lopsided rejections of a major presidential initiative in American history and the most lopsided rejection of a presidential bill by a Senate of the president’s own party.

Q: What was the “switch in time” and did Justice Roberts actually switch his vote?

“A switch in time saves nine” was a phrase coined by humorist Cal Tinney in early 1937 to describe Justice Owen Roberts’s apparent shift from joining the conservative bloc in striking down a New York minimum wage law in June 1936 (Morehead v. New York ex rel. Tipaldo) to joining the liberal bloc in upholding a Washington state minimum wage law in March 1937 (West Coast Hotel v. Parrish). On its face, the switch looked like a strategic capitulation by the bench to relieve pressure from Roosevelt’s bill. The actual conference vote in West Coast Hotel, however, was taken in December 1936, six weeks before Roosevelt announced the court-packing bill. On a strict chronology, Roberts switched before the bill existed, presumably in response to the broader political climate of the 1936 election. The release of the decision on March 29, 1937, in the middle of the court-packing fight, was nonetheless understood by virtually everyone in Washington as a strategic move by the bench. Historians remain divided on whether Roberts’s switch was a response to the specific bill or a response to the broader political environment that produced the bill.

Q: What was West Coast Hotel v. Parrish?

West Coast Hotel Co. v. Parrish, decided March 29, 1937, was the Supreme Court decision that ended the Lochner era of substantive due process. The case challenged a Washington state minimum wage law for women on the grounds that it violated freedom of contract under the Fourteenth Amendment. By a five-to-four vote, with Justice Roberts joining Chief Justice Hughes, Justice Brandeis, Justice Stone, and Justice Cardozo in the majority, the bench upheld the statute. The majority opinion, written by Hughes, distinguished Adkins v. Children’s Hospital (1923) and Morehead v. New York ex rel. Tipaldo (1936) so narrowly that the two decisions were effectively reversed. After West Coast Hotel, state legislatures could regulate wages and hours without fear of constitutional challenge under the due process clause. The decision is one of the four or five most consequential rulings of the twentieth century in American constitutional history and is widely credited with completing the doctrinal transition from Lochner-era jurisprudence to the modern regulatory state.

Q: Why did Chief Justice Hughes oppose the court-packing plan?

Hughes opposed the bill for a combination of principled and institutional reasons. He believed that judicial independence required the bench to be insulated from direct executive efforts to influence its composition, regardless of the popularity of the executive or the unpopularity of recent decisions. He believed that expanding the bench would damage its collegial deliberative process by making it impossible for each justice to participate fully in every case. And he believed the official efficiency rationale was demonstrably false, which gave him a politically usable angle of attack. His letter to Senator Wheeler on March 21, 1937, focused on the efficiency claim because that was where the administration’s public case was weakest. The letter documented that the high tribunal was current on its docket, that adding justices would slow rather than speed work, and that the bench’s collegial functioning depended on its current size. The letter is widely considered the single most influential intervention by a sitting Supreme Court justice in a contemporaneous legislative debate in American history.

Q: What did Hughes’s letter to Senator Wheeler actually say?

The letter, delivered on March 21 and read by Wheeler to the Senate Judiciary Committee on March 22, 1937, ran around eight pages and made three principal arguments. First, the high tribunal was fully current on its docket: it had disposed of all 717 cases argued in the 1935 term before adjournment, and petitions for certiorari were being granted or denied within weeks of filing. Second, adding justices would slow rather than speed the bench’s work because each new member would have to participate in conferences on every pending case, fragmenting the deliberative process. Third, the bench worked best at its current size because each justice could read every brief and join every conference fully. The letter was signed by Hughes as Chief Justice but had been reviewed and approved in advance by Brandeis (the senior associate justice and a Wilson Democrat) and Van Devanter (the senior conservative). The collaboration across ideological lines made the letter impossible for the administration to dismiss as a partisan attack. The letter’s factual claims about the docket were verifiable from public records and were not seriously contested by the administration.

Q: Who led the Senate fight against Roosevelt?

The Senate floor leader of the opposition was Burton Wheeler of Montana, a progressive Democrat who had been Robert La Follette’s running mate on the 1924 Progressive ticket. Wheeler’s leadership was crucial because his progressive credentials made it impossible to frame the opposition as a conservative defense of judicial reaction. Other key opponents included Henry Ashurst of Arizona (chairman of the Senate Judiciary Committee and nominally a Roosevelt ally, who slow-walked the bill through his committee), Carter Glass of Virginia (the conservative Democratic eminence), Tom Connally of Texas (a New Deal supporter who broke with the administration on this issue), Royal Copeland of New York (a moderate Democrat), Edward Burke of Nebraska, and the Senate Republicans led informally by Charles McNary of Oregon, who tactically kept Republican opposition quiet to allow the Democratic civil war to play out. Vice President John Nance Garner, while not formally a senator, played a substantial behind-the-scenes role lobbying senators individually against the bill, particularly after Joseph Robinson’s death in July.

Q: What was Joseph Robinson’s role and why did his death matter?

Joseph Robinson was the Senate Majority Leader and the leader of the floor fight for the court-packing bill. He had been promised the next Supreme Court vacancy in exchange for leading the fight, although Roosevelt was vague about the timing and never made the promise formally binding. When Van Devanter retired in May 1937, the obvious move was to give Robinson the seat, both to honor the commitment and to demonstrate to the Senate that the new vacancy made the broader court-packing bill unnecessary. Roosevelt delayed, hoping to extract additional loyalty service before making the appointment. Robinson, sixty-five and exhausted from carrying the bill against opposition that included many of his closest Senate friends, suffered a heart attack on the night of July 13 to 14, 1937, and was found dead in his Methodist Building apartment. His death removed the bill’s most effective Senate advocate and produced a wave of sympathy for Senate defectors who had been quietly committed to vote with Robinson out of personal loyalty. Within a week of his death, vote counts confirmed the bill could not pass. The seventy to twenty recommittal vote on July 22 came eight days after Robinson’s funeral.

Q: How many justices did Roosevelt eventually appoint to the Supreme Court?

Roosevelt named nine justices between August 1937 and February 1943: Hugo Black (1937), Stanley Reed (1938), Felix Frankfurter (1939), William O. Douglas (1939), Frank Murphy (1940), Harlan Fiske Stone elevated to Chief Justice (1941), James F. Byrnes (1941), Robert H. Jackson (1941), and Wiley Rutledge (1943). By the end of 1943, eight of the nine sitting justices were Roosevelt appointees, with Owen Roberts (a Hoover appointee) the last remaining non-Roosevelt justice. By the time Roosevelt died in April 1945, his appointees had been a majority of the bench for six years and a near-unanimous bloc on most major constitutional questions for at least three. The total of nine appointments is the largest by any president except Washington (who appointed all of the original justices and several replacements). Roosevelt achieved his institutional objective of remaking the bench through ordinary attrition rather than through the packing bill he had set out to pass in 1937.

Q: Did Roosevelt actually lose the broader fight even though the bill failed?

Historians have debated this question for seventy years. The strongest argument that Roosevelt won is that he achieved every substantive constitutional objective he had set out to compel: the end of the Lochner era (West Coast Hotel, March 1937), the expansion of the Commerce Clause (Jones and Laughlin, April 1937), the survival of Social Security (Steward Machine and Helvering v. Davis, 1937), the survival of the Wagner Act (Jones and Laughlin), and the eventual remaking of the bench through nine appointments. The strongest argument that Roosevelt lost is that the court fight produced the Conservative Coalition of Southern Democrats and Republicans that blocked civil rights, labor expansion, and economic legislation for the next three decades, and broke the spell of presidential supremacy that the 1936 landslide had created. The honest verdict is that both are true. Roosevelt won the constitutional war and lost a major political battle. Whether the trade was worth it depends on what one weights more heavily, the constitutional revolution or the political and legislative damage that followed.

Q: What was “Black Monday” in 1935?

Black Monday was May 27, 1935, the day the Supreme Court handed down three unanimous decisions striking down major elements of the early New Deal in a single morning. Schechter Poultry Corp. v. United States invalidated the National Industrial Recovery Act on non-delegation grounds and as exceeding the Commerce Clause. Humphrey’s Executor v. United States ruled that the president could not remove members of independent regulatory commissions without cause. Louisville Joint Stock Land Bank v. Radford struck down the Frazier-Lemke Farm Bankruptcy Act on Fifth Amendment takings grounds. The three decisions together represented the most consequential single-day defeat of a sitting administration’s legislative program in Supreme Court history. Roosevelt’s “horse and buggy” press conference four days later, in which he criticized the bench’s reading of the Commerce Clause as antiquated, marked the beginning of his open political conflict with the high tribunal. The events of Black Monday set in motion the chain of executive frustration that culminated in the court-packing proposal of February 1937.

Q: What was Schechter Poultry and why did it matter so much?

Schechter Poultry Corp. v. United States, decided May 27, 1935, was the Supreme Court decision that invalidated the National Industrial Recovery Act, the centerpiece of Roosevelt’s first hundred days legislative program. The case involved a Brooklyn kosher poultry slaughterhouse charged with violating NIRA’s “Live Poultry Code,” which set wages, hours, and trade practices. The bench unanimously held that NIRA unconstitutionally delegated legislative power to the executive (the non-delegation doctrine) and exceeded the Commerce Clause by reaching purely intrastate activity. The unanimity of the decision, with Brandeis and Cardozo joining the conservative bloc, made it impossible for Roosevelt to dismiss as ideological. The decision killed the centerpiece of the New Deal regulatory program and signaled that other major statutes (the Wagner Act, the Social Security Act, the Public Utility Holding Company Act) would likely face the same fate. Schechter is the single most important pre-1937 New Deal defeat and the immediate provocation for the constitutional confrontation that produced the court-packing plan.

Q: Why didn’t Roosevelt pursue a constitutional amendment instead?

Roosevelt and Cummings considered the amendment route and rejected it for three reasons. First, the time required to ratify an amendment (typically three to seven years) was longer than Roosevelt could politically wait while the New Deal program was being dismantled one statute at a time. Second, the substantive content of any plausible amendment was contested: should it require a supermajority of justices to invalidate a federal statute, or empower Congress to override invalidation, or narrow the Commerce Clause and due process protections, or address some other specific aspect of judicial review? No amendment had clear consensus support, and any specific proposal would face attack from multiple directions. Third, the political optics of an amendment fight were unattractive: Roosevelt would have to fight Republicans, conservative Democrats, and state legislatures simultaneously, and any failure would be a public defeat. The packing bill, by contrast, could pass through ordinary legislation if Roosevelt could hold his own party in line. The calculation turned out to be wrong about Roosevelt’s ability to hold his own party. With hindsight, an amendment campaign might have produced a more durable institutional victory, but the choice in 1936 was not unreasonable on the information then available.

Q: How did the court-packing fight affect the rest of the New Deal?

The fight’s political damage to Roosevelt’s legislative coalition was substantial and durable. The Conservative Coalition of Southern Democrats and Republicans that formed during the court fight blocked or weakened multiple Roosevelt initiatives in the second term: the Reorganization Bill of 1938 was defeated outright, the Fair Labor Standards Act of 1938 passed only after substantial compromise (including the exclusion of agricultural and domestic workers, the same exclusion the original Social Security Act of 1935 had used), the wage-and-hour amendments Roosevelt sought were watered down, the executive expansion of regulatory authority was constrained, and the civil rights agenda was effectively off the legislative table for the rest of Roosevelt’s presidency. The 1937 to 1938 recession, which began in August 1937, made the political situation worse and limited Roosevelt’s room to push back against the Conservative Coalition. By 1939 and 1940, with European war looming, Roosevelt pivoted away from domestic reform toward foreign policy and military preparedness, where the political environment was more favorable. The court fight did not end the New Deal, but it ended the New Deal as a legislative reform program. The administrative state Roosevelt had built remained in place, but new legislative additions were limited and incremental from 1938 forward.

Q: Has any president since Roosevelt seriously tried to pack the Supreme Court?

No president has formally proposed an expansion of the high tribunal since Roosevelt’s defeat in 1937. The norm against packing has held for nearly nine decades, despite multiple periods of political tension between the executive and the bench (most notably the Earl Warren Court during the Eisenhower and Kennedy years, the Burger Court rulings during the Nixon administration, and the Roberts Court rulings during the Obama and Biden administrations). Academic and activist proposals to expand the bench have surfaced periodically, including in the 2020s in response to the 6-3 conservative majority created by appointments under the Trump administration. The Presidential Commission on the Supreme Court of the United States, established by President Biden in 2021, examined court expansion as one option but did not formally recommend it. No legislation has been introduced with substantial chance of passage. The 1937 precedent, in which a popular president with overwhelming congressional majorities was defeated by his own party on this question, continues to function as a powerful institutional warning against statutory expansion.

Q: What did historians like Leuchtenburg and Shesol say about the constitutional revolution?

William Leuchtenburg, in The Supreme Court Reborn: The Constitutional Revolution in the Age of Roosevelt (1995), argues that the constitutional revolution of 1937 was real, substantial, and largely successful from Roosevelt’s perspective, regardless of the failure of the specific bill. The bench moved on substantive due process, on the Commerce Clause, and on the constitutional acceptability of the regulatory state. Roosevelt’s nine appointments completed the transformation. Leuchtenburg treats the legislative defeat as a tactical setback in a strategic victory. Jeff Shesol, in Supreme Power: Franklin Roosevelt vs. the Supreme Court (2010), accepts Leuchtenburg’s constitutional analysis but emphasizes the political damage to Roosevelt’s coalition and the broader institutional cost. Marian McKenna, in Franklin Roosevelt and the Great Constitutional War (2002), argues that the political cost was greater than Leuchtenburg allows and that the long-term damage to executive-legislative relations was the most important outcome of the fight. David Kennedy, in Freedom from Fear (1999), and Jean Edward Smith, in FDR (2007), are closer to Leuchtenburg on the constitutional question but agree with Shesol and McKenna on the political damage. The historiographical convergence is that the constitutional revolution was real and the political cost was real, with disagreement about the relative weight to assign each.

Q: What is the lasting legacy of the 1937 court fight?

The lasting legacy operates on three levels. Constitutionally, the doctrinal transformation of 1937 (the end of Lochner-era substantive due process, the expansion of the Commerce Clause, the acceptance of the regulatory state) remains the foundation of modern American constitutional law. Every subsequent expansion of federal regulatory authority (the Civil Rights Act of 1964, environmental regulation, consumer protection, occupational safety) rests on the doctrinal moves made by the bench in West Coast Hotel and Jones and Laughlin. Institutionally, the norm against court-packing has held for nearly ninety years and continues to constrain executive ambitions toward the high tribunal. Politically, the Conservative Coalition that formed during the fight shaped the next three decades of congressional politics and contributed to the political coalition realignment that eventually produced the Reagan revolution of 1980. The composite legacy is that Roosevelt’s most consequential single legislative defeat may have been his most institutionally durable contribution to American governance, both because the constitutional revolution he sought was achieved without the packing bill and because the failure of the packing bill produced an enduring norm against the kind of institutional restructuring he had attempted.