The Fascism File — The Definitions and the Institutions

The Word and the Record: What the File Shows About the Status of American Fascism

The word has a scholarly literature, and the literature disagrees with itself, in print, about the same man. The country has a record: courts ruled and were answered, an election was held on schedule and the opposition won it, and the government’s own death list for its detention system reached its highest year. The desk files the definitions as their authors state and apply them, and the record beside them, and stops where the record stops. Every claim is attributed.

By Eunice, the desk · Collected and analyzed at 1230 ET; revised at 1300 ET, Saturday, November 7, 2026.

The File in Brief

There is no single scholarly definition of fascism, and no agreed instrument that measures it. What exists is a documented field and a documented record, and the desk files both. The field, in its own texts: Robert O. Paxton defines fascism as political behavior and, after January 6, 2021, applied his own definition to Donald Trump, in print, twice. Roger Griffin defines it as a revolutionary ideology of national rebirth and has declined to apply it. Umberto Eco listed fourteen features and wrote, in the same essay, that any one suffices for fascism to coagulate around it. Jason Stanley defines fascist politics as ten tactics and applies all ten. Stanley G. Payne does not apply the label. The constitutional record is shorter: the United States is a republic in form and a democracy in the source of its authority—the standard description is the compound of the two. The compliance record is documented in both directions: across the fifty years before 2025 the desk found no instance of a president openly refusing a final Supreme Court order, and in 2025–2026 it found written judicial findings that the executive violated lower-court orders—a judge’s catalog of 96 orders violated by ICE across 74 cases; a criminal-contempt finding reversed on appeal; the flagship individual case complied with after 83 days—alongside compliance with the Supreme Court’s two largest merits rulings against the administration in 2026. The deaths record is the government’s own: on ICE’s published detainee-death list, fiscal 2026 stands at 36 listed deaths with two more announced in the agency’s own releases, the highest fiscal year since the agency began operating, against 26 deaths across the preceding administration’s four years; deaths during enforcement operations have no official list and no agreed total, and this report prints the categories separately and sums nothing silently. The revised record prints causes and manners where the official record states them: most stated custody deaths in fiscal 2025–2026 are medical, one fiscal 2026 death carries a medical examiner’s homicide ruling, and 16 of the 38 fiscal 2026 deaths have no public cause. The institutional record as of this week: the midterm elections of November 3, 2026 were held on schedule, and on Associated Press calls the opposition party won both chambers of Congress; no opposition party has been banned, no news organization shuttered by the government, and the courts are ruling against the executive in hundreds of cases and being, in the main and eventually, obeyed, with the exceptions concentrated in immigration enforcement. The expert indices moved: V-Dem reclassified the United States from a liberal democracy to an electoral democracy; Freedom House scored it 81 of 100 and still Free; both measure democracy, not fascism, and say so. Christian nationalism measures at about one American in ten as full adherents, stable since 2022. The fifty-year frame is also documented: the Church Committee’s 1976 report on the FBI’s covert campaign, the post-2001 surveillance settlement built by presidents of both parties, and deportation totals by administration from the department’s own yearbook, in which the current removal pace sits inside the historical range while the detention population and death counts stand at records. The report files these records. It selects no framework and announces no measurement; the frameworks are printed as their authors’ positions, on the shelf below, at full strength.

The Definitions: What the Scholars’ Own Texts Say

Paxton defines fascism by what fascists do. Robert O. Paxton, in The Anatomy of Fascism (2004), argues the doctrine is comparatively incoherent and the behavior is the definition. His closing sentence: “Fascism may be defined as a form of political behavior marked by obsessive preoccupation with community decline, humiliation, or victimhood and by compensatory cults of unity, energy, and purity, in which a mass-based party of committed nationalist militants, working in uneasy but effective collaboration with traditional elites, abandons democratic liberties and pursues with redemptive violence and without ethical or legal restraints goals of internal cleansing and external expansion.” His five stages run from the creation of movements through rooting, seizure, and exercise of power to radicalization or entropy; only Fascist Italy and Nazi Germany passed through all five, and each later stage requires the earlier without the earlier guaranteeing the later. Paxton’s own application is a published record in three steps. In 2016–2017 he resisted the label—“We should hesitate before applying this most toxic of labels”—while noting Trump adopted fascist themes. On January 11, 2021, in Newsweek, he wrote that the January 6 attack “removes my objection to the fascist label” and that “the label now seems not just acceptable but necessary,” emphasizing that the circumstances were profoundly different from interwar Europe. In October 2024 he confirmed the diagnosis—“It’s the real thing. It really is”—while saying the word “generates more heat than light.”

Griffin defines fascism by its mythic core, and sets the bar at revolution. Roger Griffin, in The Nature of Fascism (1991), gives the one-sentence minimum: fascism is “a genus of political ideology whose mythic core in its various permutations is a palingenetic form of populist ultra-nationalism”—palingenesis being rebirth, the myth of a decadent nation regenerated into a new order, and ultranationalism being nationalism that subordinates individual rights and pluralism to the organic nation. His test is ideological: without the revolutionary rebirth myth aimed at a new order and a new man, a movement is in his framework authoritarian or populist, not fascist. Griffin’s own application is a non-application: he has described Trump as an “ochlocrat,” a leader who enables mob rule, practicing illiberal populism, and reads “Make America Great Again” as rebirth rhetoric without the coherent revolutionary ideology, party apparatus, or project of systemic overthrow his definition requires. The desk notes the sourcing plainly: Griffin’s position is documented in scholarly surveys of his published commentary; the desk pulled no single Griffin text devoted solely to Trump, and the same caveat attaches to Payne below.

Eco listed fourteen features and disclaimed the list as a system in the same essay. Umberto Eco, who grew up under Mussolini, published “Ur-Fascism” in The New York Review of Books on June 22, 1995. The features: the cult of tradition; the rejection of modernism; the cult of action for action’s sake; disagreement as treason; fear of difference; appeal to a frustrated middle class; obsession with a plot; the enemy as both too strong and too weak; pacifism as trafficking with the enemy; contempt for the weak; the cult of heroism linked to death; machismo and weaponry; selective populism, the People as a monolithic voice embodied by the leader; and Newspeak, an impoverished vocabulary to limit critical reasoning. Eco’s own caveat is in his text and is load-bearing for any use of the list: the features “cannot be organized into a system; many of them contradict each other, and are also typical of other kinds of despotism or fanaticism. But it is enough that one of them be present to allow fascism to coagulate around it.” By its author’s design, the checklist can register a positive on a single feature. Eco died in 2016 and made no application to Trump.

Stanley defines fascist politics as tactics, and applies his definition in full. Jason Stanley, a philosopher of propaganda, in How Fascism Works (2018), is explicit that a nation need not be a fascist state to exhibit fascist politics. His ten pillars: the mythic past; propaganda, including anti-corruption language used corruptly; anti-intellectualism; unreality, the flooding of the public sphere until shared truth collapses into loyalty; hierarchy; victimhood, the dominant group portrayed as the true victim; law and order, the in-group as naturally lawful and the out-group as inherently criminal; sexual anxiety; “Sodom and Gomorrah,” the heartland against the decadent city; and Arbeit macht frei, the out-group as lazy and public goods reserved for the in-group. Stanley’s own application is the most direct of the set: he has called Trump a fascist leader and MAGA a fascist movement in power, in his books and interviews from 2018 forward. That is his stated position, recorded here as his.

The field does not agree, and the disagreement is part of the record. Stanley G. Payne, whose A History of Fascism, 1914–1945 (1995) distinguishes fascism from other authoritarianisms by anti-liberalism, anti-conservatism, and selective anti-capitalism, a party militia, and the goal of a totalitarian new man, has not applied the label to Trump or MAGA, describing the movement as authoritarian populism lacking fascism’s ideological coherence, and warning that overuse of the term undermines precision. Ruth Ben-Ghiat and Timothy Snyder apply authoritarian and fascist frameworks affirmatively, their category of choice in much of their work being “authoritarian” or “strongman,” with fascism used for specific lineages and tactics. Two assessments by former subordinates are the most quoted non-scholar entries in the file. John Kelly, the president’s longest-serving chief of staff, told the New York Times in recorded interviews published October 22, 2024, after reading aloud a general dictionary-style definition: “Certainly the former president is in the far-right area, he’s certainly an authoritarian, admires people who are dictators—he has said that. So he certainly falls into the general definition of fascist, for sure.” The definition was the general one he read aloud, not a scholarly framework. Gen. Mark Milley is quoted in Bob Woodward’s War (2024) calling Trump “fascist to the core”—a secondhand quotation. Both statements came in a campaign’s final weeks; the Trump campaign called Kelly’s accounts fabricated.

Republic, Democracy, or Both: The Constitutional Record

The Constitution uses one of the two words, and the founding texts define it as the other word’s mechanism. The word “democracy” does not appear in the United States Constitution. The word “republican” does, at Article IV, §4: “The United States shall guarantee to every State in this Union a Republican Form of Government.” Madison, in Federalist No. 10, distinguishes a “pure democracy,” citizens administering government in person, from a republic, “a government in which the scheme of representation takes place”; in No. 39 he defines a republic as “a government which derives all its powers directly or indirectly from the great body of the people.” The structure the text creates: representatives chosen by popular election; senators originally state-chosen, directly elected since the Seventeenth Amendment; a president chosen through the Electoral College; judges tenured during good behavior; and a Bill of Rights removing specified subjects from majority vote. The descriptive consensus of the reference authorities—the National Constitution Center, the Annenberg Classroom, and the standard civics references—is that the United States is both: a republic in form, representative government under a constitution with no monarch, and a democracy in the source of its authority, the people acting through elections. The two terms answer different questions—who holds ultimate authority, and how it is exercised—and the public contest is over the emphasis, not over any text. The undemocratic features of the system are themselves constitutional: the Electoral College, under which a president can lose the popular vote, as has happened five times, most recently in 2000 and 2016; equal Senate representation regardless of population; life-tenured judges; and the Article V amendment supermajorities. These are features of the 1787–1791 design, not departures from it.

Presidents and the Courts: The Compliance Record

The structural fact governs everything below: the Supreme Court has no enforcement arm. Compliance rests on the executive’s Article II duty to take care that the laws be faithfully executed, and on political cost. The Federal Judicial Center’s own history states the norm plainly: substantial compliance, with defiance on rare occasions. Outright presidential refusal to obey a direct Supreme Court order is, in the documented record, close to nonexistent; the historical pattern is resistance, delay, and narrow reading, followed in the landmark cases by compliance. The two pre-window exceptions: in Worcester v. Georgia (1832), Georgia defied the Court’s ruling and President Jackson took no action to enforce it against the state—the famous line attributed to him is apocryphal—and Cherokee removal proceeded; and in Ex parte Merryman (1861), Chief Justice Taney, sitting as circuit judge, held Lincoln’s suspension of habeas corpus unconstitutional, and Lincoln disregarded the order in the Civil War’s opening weeks. That is the clearest case of a president ignoring an order directed at his own conduct—by a single justice on circuit, in wartime, 165 years ago. Against those stand the compliances: Truman, held in Youngstown (1952) to have seized the steel mills unlawfully, complied immediately; Nixon, ordered unanimously in 1974 to produce the tapes, complied and resigned sixteen days later; and at Little Rock in 1957 the executive sent the 101st Airborne to enforce the Court’s orders against a defiant governor.

Across the fifty-year window before 2025, the desk found no documented instance of a president openly refusing a final Supreme Court order directed at him. From Carter through Biden, administrations lost major cases and complied: the George W. Bush administration complied with Hamdan v. Rumsfeld (2006) and then sought and obtained new legislation; the Obama administration complied with the health-care ruling as decided; the Biden administration complied with the ruling striking its student-loan program and then pursued narrower alternatives through rulemaking. That sequence—comply with the order, contest the policy space around it—is the standard modern pattern, the first Trump administration’s travel-ban litigation its standard example. The desk found no court-documented case in this window of a judicial finding that a president defied an order. That is the baseline.

The second Trump administration’s record runs in both directions, and the volume is itself documented. Litigation trackers counted roughly 700 to 800 active or filed challenges to administration actions by 2026; one tracker’s mid-2026 snapshot records plaintiffs winning 262 rulings to the government’s 126, and other trackers give different splits—the desk prints the figure as one named tracker’s dated snapshot, because no rate in this file is method-free. A CNN analysis identified at least 77 federal court orders containing sharp judicial criticism of the administration, from 69 judges, including 11 Trump appointees. The Supreme Court also ruled for the administration repeatedly on its emergency docket in 2025—the White House counted 21 victories by September 2025—and then ruled against it on the merits in the two largest cases of 2026. In Learning Resources, Inc. v. Trump, on February 20, 2026, the Court held 6–3 that the International Emergency Economic Powers Act does not authorize the president to impose tariffs; the administration complied with the judgment while substituting a replacement tariff under a different statute, which the Court of International Trade then struck on May 7, 2026. In Trump v. Barbara, on June 30, 2026, the Court held that children born in the United States to unlawfully or temporarily present parents are citizens at birth under the Fourteenth Amendment, invalidating the core of the administration’s executive order. The desk records the holdings and dates as reported; it did not pull the slip opinions themselves, and says so rather than smoothing it over. In the large majority of the individual cases, the documented pattern is the standard one: appeal, seek a stay, and, when stays are denied, implement the order while appealing.

The documented exceptions are written down by judges, and they are concentrated in immigration enforcement. In March 2025 the administration deported more than 200 alleged Tren de Aragua members to El Salvador under the Alien Enemies Act while Judge James Boasberg’s order barring removal was being issued; on April 16, 2025, Boasberg found probable cause of criminal contempt through “willful disobedience of judicial orders”; on April 14, 2026, an appeals panel blocked the contempt proceedings, holding the underlying order was not clear and specific enough to support criminal contempt. Both halves are the record. In January 2026, Judge Patrick Schiltz, a George W. Bush appointee, documented in an order 96 court orders that ICE had violated across 74 cases—the most concrete judicial count in the record, because it is a judge counting his own court’s orders. By January 2026, reporting counted at least 35 show-cause orders since August 2025 requiring the administration to explain why it should not be held in contempt; the figure is press-derived and approximate. The most litigated single case is Kilmar Abrego Garcia’s. Removed to El Salvador on March 15, 2025, under a 2019 withholding order barring that removal, he was conceded by the government, in its own March 31 filing, to have been removed “because of an administrative error.” The District of Maryland ordered the government on April 4 to facilitate and effectuate his return; the Supreme Court on April 10 left the facilitation requirement in place without noted dissent; the district court found on April 11 that the government had made no meaningful effort to comply. The administration’s position through April and May was that “facilitate” did not obligate it to secure a release from a sovereign El Salvador; President Trump said on April 11, “If the Supreme Court said bring somebody back I would do that. I respect the Supreme Court.” On June 6, 2025—83 days after the removal, 57 days after the Supreme Court’s order—the government returned him to the United States, simultaneously indicting him in Tennessee on human-smuggling charges; that prosecution was later dismissed on vindictive-prosecution grounds, a district court finding the closed investigation had been reopened a week after the Supreme Court’s order, and the Justice Department appealed the dismissal in August 2026. No government body counts defied orders, and no official defiance rate exists; a Washington Post analysis relayed by secondary trackers, putting defiance at roughly one-third of cases decided against the administration, is a method-dependent press construction whose dataset the desk did not read. The comparison the sourced record supports is this: in the 1976–2024 window, judicial findings that the executive violated orders are rare to the point that the desk found none at presidential level; in 2025–2026 such findings exist in writing from multiple federal judges, alongside an appellate reversal of the major contempt proceeding, ultimate compliance in the most prominent individual case, and compliance with the Supreme Court’s two largest merits rulings; and no president in the window, including this one on the record assembled here, has openly refused a final Supreme Court merits order and simply continued the struck policy unchanged. What is new is the volume of adverse findings about compliance with lower-court orders, concentrated in immigration enforcement. That sentence is as far as the sourced record goes.

Deaths in Connection With ICE: The Government’s Own Numbers

“People killed in connection with ICE” is not one number, and this report does not print it as one. The record contains three distinct categories that public argument routinely sums. Category A is deaths in ICE custody—in a facility, in transit, or in a hospital while custody remained active—the only category with an official, statutory, name-by-name public list. Category B is deaths during enforcement operations, fatal encounters with ICE and Border Patrol agents outside detention; no official list exists, and counts are press and tracker constructions that disagree. Category C is deaths shortly after release or transfer, excluded from ICE’s count by definition; cases are documented individually and no official total exists.

Category A rests on the agency’s own list, kept under a statutory reporting requirement. Since the Department of Homeland Security Appropriations Act of 2018, ICE is required to publish reports of in-custody deaths, and its Detainee Death Reporting page lists each death by fiscal year. A study in JAMA, published April 16, 2026, assembled the full series from ICE’s own sources: 272 deaths in custody from fiscal 2004 through partial fiscal 2026. The annual totals on ICE’s own reporting, for the statutory-publication era, are these.

The administration-spanning comparison comes from the same list. Read across administrations, the list records 26 deaths in custody from January 20, 2021 to January 19, 2025, the preceding administration’s four years, and 57 named on the list from January 20, 2025 to August 23, 2026, the current one’s first nineteen months, plus at least two more announced by ICE in September 2026; independent counts running to late October 2026 reach about 60 since January 2025. The detained population rose over the same period: ICE held roughly 73,000 people in mid-January 2026, a record, on the agency’s figures as reported. Human Rights Watch, in June 2026, measured the custody mortality rate up about 138 to 140 percent year over year against a 77 percent rise in the detained population—on its measurement, deaths rose faster than the population; JAMA’s rate series points the same direction.

Causes and Manners: What the Official Record States

A death count without cause and manner leaves unlike deaths in the same digit. ICE’s announcements are the first limitation: in the current period they often describe a detainee as found unresponsive or as having passed away and state no cause. The desk therefore re-read the agency’s announcements and death reports beside coroners’ and medical examiners’ rulings where reporting located them. The name-by-name record for the two completed fiscal years is:

The medical column is ordinary lethal medicine, stated in the agencies’ and examiners’ words. It includes cardiac arrest and heart failure, pneumonia and sepsis, stroke, liver and kidney failure, chronic disease, withdrawal, and one death certificate attributing death to an allergic reaction. The fiscal 2026 vehicle entry is Jose Castro-Rivera, struck on a highway while fleeing officers during an enforcement stop; ICE carries him on its detainee-death page although the death was not in a facility. The two fiscal 2025 homicides were not killings by officers or detainees: Norlan Guzman-Fuentes and Miguel Angel Garcia-Hernandez, detainees being processed at the Dallas field office, were shot by an outside sniper who attacked the office; Garcia-Hernandez died of his wounds days later.

The one fiscal 2026 homicide ruling remains disputed and unadjudicated. Geraldo Lunas Campos, 55, died at Camp East Montana on January 3, 2026, while being restrained by contract security staff. The El Paso County medical examiner ruled the cause asphyxia due to neck and torso compression and the manner homicide—a medical classification meaning death at the hands of another, not a criminal judgment. DHS’s revised account says Lunas Campos violently resisted staff and continued trying to take his own life, and that he stopped breathing during the ensuing struggle. Fellow detainees told reporters that guards held him in a chokehold and that he said he could not breathe. A Government Accountability Office report in June 2026 described a use-of-force death at the camp in January, said the contractor had not provided reports its contract required, and found associated evidence missing or destroyed. No court has adjudicated the death.

The suicide count depends on the instrument, and the gap is a finding. A name-by-name reading of ICE’s own announcements yields five suicides in fiscal 2026 and three in fiscal 2025, including cases ICE called presumed suicide. The Associated Press, working from autopsy reports, coroners’ rulings, and police records, counted at least ten suicides from January 2025 through May 2026, seven in fiscal 2026 alone. The counts differ because ICE releases frequently state no manner and some deaths announced without a cause were later classified as suicides by coroners. This report prints both instruments and does not average them.

No comparable cause table exists for the preceding administration. ICE publishes the earlier counts, but no source located published a cause-by-cause tally at fiscal-year granularity for those years. The only peer-reviewed classification is JAMA’s pooled 2021–2026 band—65 deaths spanning both administrations and therefore not divisible between them: other or undetermined, 49.2 percent; cardiovascular, 18.5 percent; infectious, 13.8 percent; suicide, 10.8 percent; neurologic, 7.7 percent. JAMA’s prespecified classification contains no homicide category anywhere in its series; a death by force in its sources would fall inside other or undetermined. The absence of a prior-administration cause ledger is a limit of the public record, not a gap this report fills by inference.

The negative space is part of the record. Malnutrition, dehydration as a stated cause, abandonment as a stated cause, sleep deprivation, waterboarding, torture, and beating do not appear in any official stated cause located for fiscal 2025 or fiscal 2026. One adjacent record must travel with that sentence: ICE’s report on Genry Ruiz Guillen listed hypernatremia and hyperosmolality—conditions associated with severe water deficit—with rhabdomyolysis and a catatonic disorder. ICE did not state dehydration or neglect, and the desk does not rename its words.

The care record is separate from the cause record, and it is adverse in the findings located. In June 2024, independent physicians retained by the ACLU, American Oversight, and Physicians for Human Rights reviewed ICE’s own death reviews for 2017–2021 and judged 95 percent of the deaths preventable or possibly preventable with adequate care—an attributed expert judgment about care over a period spanning two administrations, not a set of cause-of-death rulings. The Associated Press’s May 2026 examination of its ten identified suicides reported repeated violations of ICE’s suicide-prevention standards, including intake screening, risk monitoring, and mental-health care; DHS responded that suicide remained rare and that staff followed strict safety protocols. A death attributed to untreated illness in a facility obligated to treat it therefore presents a documented care question. It is not, on any official cause located for these two years, a torture or beating death. Both halves stand.

The Comparative Baseline: Detention, Jails, and Prisons

The rate beside the count changes the comparison that can honestly be made. JAMA’s ICE series is stated per 100,000 detainee-years, adjusting for person-time in custody. The Bureau of Justice Statistics’ jail and prison series are stated per 100,000 people in average daily population. Those denominators are related but not identical; the populations also differ in age, health, legal status, sentence length, and turnover. The figures are therefore printed together with their constructions attached.

BJS’s comprehensive national mortality series ends in 2019; no newer nationally comparable jail or prison baseline was located. The jail and prison figures are therefore the last measured baseline, not a contemporaneous 2025–2026 measurement. State and federal prisoners are generally older and held for years, and their deaths are dominated by chronic illness accumulated over long sentences. ICE detainees are a civil population with much shorter average stays, which is why the detainee-year denominator matters. On the printed constructions, even partial fiscal 2026’s 88.9 remains below BJS’s last local-jail baseline of 167, while rising from 13.0 in fiscal 2023. The comparison is limited to what the two series measure.

The Care Owed on Paper

The written standard is distinct from delivery. ICE’s Performance-Based National Detention Standards, Standard 4.3, require covered facilities to provide a continuum of care including screening, prevention, diagnosis, and treatment; medical, dental, and mental-health intake screening no later than 12 hours after arrival; a comprehensive health assessment within 14 days; 24-hour emergency medical and mental-health services; prescribed medication and medically necessary treatment without interruption; and referral or transfer when a detainee’s needs exceed facility resources. The standards operate through ICE policy, contracts, and inspections and are not, standing alone, independently enforceable by a detainee in court.

Because immigration detention is civil, claims arise under the Fifth Amendment’s Due Process Clause rather than the Eighth Amendment’s punishment clause. The Supreme Court’s cases require nonpunitive conditions and describe civil detainees as entitled to more considerate treatment than convicted criminals; Youngberg v. Romeo supplies the professional-judgment standard. The circuits are split in application: several apply deliberate indifference, borrowed from Eighth Amendment doctrine; the Fourth Circuit has applied professional judgment in a civil-detention case; and objective-indifference variants have spread since Kingsley v. Hendrickson. That is the floor and the governing dispute as written. The death reports and reviews above are the separate record of performance.

Category B has no spine, and the spread among its constructions is itself the finding. A compiled list of shootings by United States immigration agents since January 20, 2025 counts at least 43 shooting incidents resulting in 11 deaths, as of late October 2026. The two names the whole country knows are in this category, not the custody category: Renée Nicole Good, 37, a United States citizen shot by an ICE officer in Minneapolis on January 7, 2026, and Alex Jeffrey Pretti, 37, a United States citizen and intensive-care nurse shot by Border Patrol personnel in Minneapolis on January 24, 2026; the department stated Pretti was armed, and video analyses reported he was holding a phone, and the desk records both, unresolved. Other trackers give different totals: Reuters documented seven fatal shootings through July 17, 2026; The Trace counted nine fatalities by October 29, 2026; broader constructions, taking in vehicle pursuits and other encounters, reach 22 to 24. Figures above roughly 70 for January 2025 through October 2026 that circulate in public argument are sums of the custody list, the operations constructions, and tracker-identified unlisted cases across both agencies: a construction, not a count, and printed here only as one.

The Institutions: Opposition, Press, Elections, Courts

The frameworks in the first section of this report were written against regimes that abolished the institutions in this one. The commission asks what the record shows about their current operation, and each item below is observable. The 2026 midterm elections were held on Tuesday, November 3, 2026. On Associated Press race calls as of November 4–7, 2026—uncertified, with counting continuing—the Democratic Party, the opposition, won control of both the House and the Senate. In the Senate, Democrats flipped seven Republican-held seats and held every contested Democratic seat, standing at at least 54 Democratic and aligned seats to at least 44 Republican, with Alaska and Mississippi uncalled. In the House, at least 222 seats were called for the Democrats against the 218 needed. Democrats also flipped five governorships. The new Congress convenes in January 2027, producing divided government for the term’s second half; Democrats did not win the two-thirds supermajorities required to override vetoes or convict on impeachment alone. No major legal challenge to the results had been reported by November 5–7. Election-administration litigation is part of the record: at least 34 federal district-court decisions had halted or blocked administration efforts touching election administration, among them the March 2025 executive order imposing documentary proof of citizenship on the federal registration form—8 of the decisions by judges appointed by Trump himself, on the Brennan Center’s catalog. The administration complied with the injunctions while appealing.

The opposition party operated openly throughout, and the press was not shuttered. Through 2025 and 2026 the Democratic Party held seats, campaigned, raised funds, won special elections, and litigated against the administration. No opposition party has been banned, deregistered, or barred from ballots in any state in the record assembled. No United States news organization has been shuttered by the government in the record, and critical coverage is continuous; it is the source of much of this file. The documented pressure episodes, each attributed: the administration barred the Associated Press from pooled events in February 2025 over its refusal to adopt “Gulf of America” in its stylebook; Judge Trevor McFadden, a Trump appointee, ruled the exclusion violated the First Amendment, and the government dropped its appeals in March 2026. A Reuters review found 75 cases in 2025–2026 in which judges found First Amendment infringements by the administration, spanning universities, law firms, and media matters. The president and his officials routinely characterize critical outlets as “fake news” and enemies, rhetoric documented in the administration’s own posts.

The expert indices moved sharply in 2025–2026, and what they measure is democracy, not fascism; the file prints them under that label and no other. Freedom House, in Freedom in the World 2026, covering 2025, scores the United States 81 of 100, down 3 points from 84, its lowest since the scale began in 2002, and down 12 since 2005; the United States remains rated Free. The V-Dem Institute, in its Democracy Report 2026, records the Liberal Democracy Index falling from 0.75 to 0.57 in one year, its rank from 20th to 51st of 179, and reclassifies the United States from a liberal democracy to an electoral democracy for the first time in over fifty years, describing the speed as unprecedented; its analysts also note the electoral mechanics remain fundamentally intact—citizens vote, and elections are clean—with the decline located in liberal checks, rights, and executive constraint. The Economist Intelligence Unit’s Democracy Index 2025 scores the United States 7.65 of 10, a flawed democracy, its classification since 2016 and its lowest United States score since the index began. These are expert-coded assessments with published methodologies and known critics, and their 2025 moves are contested, including by commentators who argue the coders’ priors drive the results. The file prints the scores, the instruments’ nature, and the contestation. None of the three claims to measure fascism, and the desk does not use them as if they did.

Christian Nationalism: The Scholars’ Measure

The field’s standard sociological treatment defines the term as a demand on the state, not as a synonym for being Christian, evangelical, or patriotic. Andrew L. Whitehead and Samuel L. Perry, in Taking America Back for God (2020), define Christian nationalism as “a cultural framework—a collection of myths, traditions, symbols, narratives, and value systems—that idealizes and advocates a fusion of Christianity with American civic life,” contending that America has been and should remain distinctively Christian “from top to bottom.” Their measure is a six-item scale from the Baylor Religion Survey, built on agreement that the federal government should declare the United States a Christian nation, advocate Christian values, allow religious symbols in public spaces, and take steps to preserve the nation’s Christian character, should not enforce strict separation of church and state, and that United States success is part of God’s plan. On 2007 and 2017 data they sorted Americans into Rejecters, about 21.5 percent, Resisters, Accommodators, and Ambassadors, about 19.8 percent at the top; roughly half the country scored as Accommodators or Ambassadors—a figure that depends heavily on where the scale’s middle is cut, which the authors themselves discuss.

The independent series that tracks the question now measures a stable minority. The Public Religion Research Institute built a separate five-item battery—that the government should declare America a Christian nation, that United States laws should be based on Christian values, that if the United States moves away from its Christian foundations “we will not have a country anymore,” that being Christian is important to being truly American, and that God intends Christians to exercise dominion. Its 2025 American Values Atlas, more than 22,000 interviews conducted February to December 2025 and reported February 17, 2026, finds Adherents at 11 percent and Sympathizers at 21 percent, together about 32 percent of Americans, against Skeptics at 37 percent and Rejecters at 27 percent. The shares have been stable since measurement began in late 2022; its chief executive stated there is no evidence of widespread national growth in the ideology over the series. By party and religion, supporters are a majority of Republicans, at 56 percent, 67 percent of white evangelical Protestants, and 54 percent of Hispanic Protestants, while 82 percent of Democrats and 87 percent of the religiously unaffiliated are Skeptics or Rejecters. The institute also records that 38 percent of Adherents and 28 percent of Sympathizers agreed that “because things have gotten so far off track, true American patriots may have to resort to violence to save the country.” The label is contested from inside: many who score as Adherents reject the term as pejorative, while a smaller, visible set of writers adopts it affirmatively—the 2022 book The Case for Christian Nationalism, by Stephen Wolfe, argues openly for a Christian prince and an explicitly Christian polity; its representativeness is precisely what the survey data measures. Several members of Congress and state officials have used the term of themselves since 2021; the desk built no verified name-by-name catalog in this pass and prints no names it did not individually source.

The enacted footprint is concentrated at state level and in litigation; no federal confessional settlement exists in the record. The documented state items include Ten Commandments display laws, Louisiana’s 2024 classroom-display statute among them, enjoined by federal courts and litigated through 2025–2026, and Bible-infused curriculum materials approved in Texas in 2024—each a real enactment, each contested in court with mixed outcomes. At the federal level, no statute declares the United States a Christian nation; the Establishment Clause stands; and the Supreme Court’s religion jurisprudence has moved toward accommodation in funding and display cases under doctrines grounded in history and coercion tests, not in Christian nationhood. Self-described Christian nationalists hold real offices, and the movement’s organizational footprint is public and active. Its programmatic maximal goals, a confessional state and Christian-law frameworks, are not enacted federal law anywhere in the record assembled. The measurement the record supports is this: a minority ideology with majority weight inside one major party’s coalition—about one American in ten as full adherents, one in three as adherents or sympathizers, stable across the 2022–2025 measurement—with its enacted footprint at state level and in the courts. Whether that constitutes a component of fascism is an application of the first section’s frameworks, left where this report leaves all applications: with the reader.

Capitalism, Corporatism, and the Fascist Economy

The literature’s central term is the one public argument most often gets wrong, and the file corrects it from the standard references. In the scholarly and historical usage, corporatism derives from the Latin corpus, body, and means the organization of society into state-recognized functional bodies—employers’ syndicates, labor syndicates, professional guilds—through which the state manages each sector. It does not mean rule by business corporations; the term for that is corporatocracy, and conflating the two is a documented category error the standard references state explicitly. The widely circulated line that fascism should be called corporatism because it is a merger of state and corporate power, attributed to Mussolini through Gentile’s Enciclopedia Italiana entry, is an attribution the desk could not verify in Gentile’s text; it circulates in polemics from 1944 forward, and this report does not print it as a Mussolini quotation. Mussolini’s own definition is on the record, from his address to the National Council of Corporations on November 14, 1933: “Corporatism is disciplined, and therefore also a controlled economy… Corporatism surpasses socialism and surpasses liberalism, creating a new synthesis.” Fascist corporatism in practice ran through the Charter of Labour of 1927, the twenty-two corporations of 1934, and the replacement in 1939 of the elected Chamber of Deputies by the Chamber of Fasces and Corporations.

The scholarly record on the reference regimes’ economies is that private property survived and independence did not. Fascist Italy and Nazi Germany preserved private property and profit; big business was not expropriated, and in Germany major firms profited enormously. What both regimes abolished was independence: independent trade unions were dissolved, a bright-line fact with no ambiguity in the record; prices, production, and foreign exchange were increasingly state-set; and the Four-Year Plan apparatus in Germany and the IRI holding system in Italy made the state a dominant economic actor. Paxton’s formulation treats fascism as neither capitalism’s agent nor its enemy: a third force that allied with elites and business against the left while pursuing its own goals, at business’s expense when the two conflicted. Payne’s typology counts selective anti-capitalism among the fascist traits. The Marxist tradition defined fascism as, in a 1935 formulation, “the preservation by violence of the private ownership of the means of production”—that tradition’s attributed position. The test the historical record sets up is structural, with three marks: abolition of independent trade unions; compulsory state-sector bodies replacing market and parliamentary allocation; and formal subordination of firms to national plans. The contemporary United States economy exhibits none of the three. Unions remain legal and independent; no corporative chamber exists; and the 2025–2026 federal economic fights—tariffs by executive action, industrial subsidies, government equity stakes in individual firms, the 2025 Intel arrangement among them, labeled state capitalism by critics across the spectrum—occur inside a market system with private ownership intact and are litigated in open court, the tariff program itself struck down in the case recorded above. Scholars who argue the capitalism point against the label, Griffin and Payne among them, make precisely this argument: that the current economic program is recognizably anti-statist in its domestic core—deregulation, tax cuts, hostility to the administrative state—which sits awkwardly with classical fascism’s statist economics. Scholars who apply the label, Stanley among them, argue contemporary fascism can be market-fused. Both positions are in the peer-reviewed record and are printed here as positions.

The question “how fascist is capitalism” inverts in the sources, and the file prints the inversion. No major fascism scholar in the set defines capitalism itself as fascism. The live dispute is three-cornered: whether fascism is capitalism’s emergency instrument, the Marxist tradition’s position; an autonomous movement that used and then subordinated capital, Paxton’s and Griffin’s; or a politics compatible with multiple economic systems, including neoliberalism, Stanley’s. There is no measurement instrument for the question beyond the three positions, and the desk files the three.

The Fifty-Year Frame: The Hegemon at Home

The denominator comes first: the power the United States actually holds. On the Stockholm International Peace Research Institute’s April 2025 figures, for 2024, United States military expenditure was $997 billion, 37 percent of the world total, 3.2 times the second spender, China, at an estimated $314 billion, and 66 percent of all NATO spending. The United States maintains the world’s only global basing and alliance network; the precise base count is method-dependent and the desk adopts none. Domestically, the federal government fields the largest incarceration system in the democratic world—1,852,900 people incarcerated at year-end 2023, on the Bureau of Justice Statistics’ count, filed in Special Report No. 2. The outer question, how the hegemon uses its power abroad, is scoped here and not built: the documented external record of the window is a research project of its own size at this paper’s standard, and this report does not pretend to it. The domestic question is answerable from the government’s own investigated record, and the three baselines follow.

Baseline one is secret: the FBI’s covert campaign, documented by the Senate in the Bureau’s own files. The Church Committee’s final report of April 26, 1976 records that COINTELPRO, run from 1956 to 1971, was designed to “disrupt” groups and “neutralize” individuals deemed threats to domestic security—covert action against Americans engaged in lawful political activity, from the Communist Party and the civil-rights movement to the antiwar movement and elected officials. The documented tactics include infiltration, anonymous letters to spouses to destroy marriages, procuring IRS investigations to deter protest leaders, and falsely labeling members of violent groups as informants to expose them to physical attack. The campaign against Martin Luther King Jr. included wiretaps, hotel-room bugs, and an anonymous letter urging suicide. The Committee found the targets ranged far beyond persons who could properly be characterized as enemies of freedom. The reforms that followed were the Foreign Intelligence Surveillance Act of 1978, permanent intelligence committees, and the executive ban on assassinations. COINTELPRO was secret—exposed by a burglary in 1971, not by oversight—and its exposure produced that settlement. The signature domestic-enforcement technologies of 2025–2026 are conducted openly, announced by the government, and litigated in real time. Whether openness is better or worse is argument; that the mode differs from the 1956–1971 baseline is record.

Baseline two is the post-2001 settlement, bipartisan in construction and continuation. The USA PATRIOT Act’s Section 215 was interpreted to authorize bulk collection of all Americans’ telephone metadata, under a secret court order to Verizon disclosed on June 5, 2013, by Edward Snowden, compelling daily production of records on millions of customers with no suspicion requirement. The FISA Amendments Act of 2008, at Section 702, authorized programs collecting internet communications from nine major United States technology companies, targeting non-United States persons abroad while incidentally collecting Americans’ communications at scale; Section 702 has been repeatedly reauthorized, and the desk located the statute’s sunset timeline running into April 2026 without verifying the 2026 legislative outcome, and prints the gap as a gap. The USA FREEDOM Act of 2015 ended Section 215 bulk collection as such. The same settlement produced warrantless wiretapping under the President’s Surveillance Program, disclosed in 2005; the torture program documented in the Senate Intelligence Committee’s 2014 summary; Guantánamo; and the 2001 military-force authorization’s two-decade career across four presidents. The largest documented domestic-surveillance expansion of the last fifty years was built before, and inherited by, the current administration; it is the standing baseline of standard operating procedure for state power at home.

Baseline three is the deportation record itself, from the department’s own yearbook, and the current regime’s distinguishing features are not its removal totals. On the department’s Yearbook Table 39—a removal being a compulsory movement under a removal order, a return a confirmed departure without one, an expulsion the Title 42 category of 2020 to 2023, and all broad totals counts of events, not of unique people—the Obama administration’s fiscal 2009–2016 totals are 3,062,466 removals and 2,186,427 returns, a combined 5,248,893; the label applied to that record was coined by an advocate, Janet Murguía of the National Council of La Raza, on March 4, 2014: “this president has been the deporter-in-chief.” The first Trump administration’s removals and returns sum to about 1,795,000, about 2,002,000 with fiscal 2020’s Title 42 expulsions. The Biden administration’s fiscal 2021–2024 totals are 2,032,300 removals and returns plus 2,754,120 expulsions, about 4,786,000 on the broadest count; the Migration Policy Institute’s characterization is “nearly 4.4 million repatriations” through February 2024. For the current term, the Institute estimated about 980,000 deportations from January 20, 2025 through July 31, 2026—a pace below the Obama-era removal peak on the order-based measure, with the detention population and the death counts at records. The distinguishing features of the current regime in the documented record are the abolition of the enforcement-priority tiers that had structured ICE since 2014, by Executive Order 14159 of January 20, 2025—with the nuance that ICE’s fiscal 2026 budget justification reintroduced priority language nearly identical to the rescinded 2021 memo, on the Heritage Foundation’s analysis—; the scale of third-country removals, 27,943 from January 20, 2025 to August 5, 2026 on ICE data obtained under the Freedom of Information Act, as filed in Special Report No. 2; interior enforcement at a visibility not seen in the earlier regimes; and the detention-death series filed above. Behind the window stand the deep-history anchors, also filed in No. 2: the Palmer Raids of 1919–1920, which deported 556 people after mass arrests; the Mexican Repatriation of the 1930s, which removed hundreds of thousands, up to 60 percent of them United States citizens on the standard scholarly estimate; and Operation Wetback, whose official figure above one million is contradicted by the scholarly consensus of about 300,000. Mass expulsion with weak process is a recurring American practice across a century, under presidents of both parties and none. That is the baseline fact any departure claim must clear.

Filed together, the three baselines run under administrations of both parties, and the file states the continuity without drawing from it. The covert-action baseline is a bureau program of 1956–1971, exposed in 1976. The surveillance baseline was built after 2001 and continued by presidents of both parties, its founding emergency declaration renewed annually by Obama, Biden, and Trump alike. The deportation baseline’s largest combined total in the window belongs to a Democratic administration, its second-largest broad count to another, and the Republican administrations in the series sit below both on the order-based measure—while the current administration’s detention population and custody deaths stand at records on the same government’s figures. A reader running the first section’s frameworks against this record has both parties’ administrations on the page, in the government’s own numbers. The desk files the continuity and draws no conclusion from it.

January 6: The Same Instrument, Applied to the Other Famous Count

The January 6 death ledger is complete only when it carries every side and separates the day from its aftermath. The District of Columbia medical examiner supplied the causes for the deaths of the day. Attendee, officer, and defendant in the table are positions in the event and its prosecution, not findings about any death.

The only homicide on January 6 was Babbitt’s, and the official reviews cleared the officer who fired. Capitol Police Lt. Michael Byrd shot Babbitt as she climbed through the broken window of the Speaker’s Lobby door. The Justice Department declined charges after finding no evidence that Byrd did not reasonably believe deadly force necessary in self-defense or defense of Congress; the Capitol Police found the shooting lawful and within department policy. Boyland’s examiner ruling also has a printed dispute beside it: a pathologist retained by her family concluded compressional asphyxia, while District police internal affairs cleared an officer whose baton strikes near Boyland became part of the family’s use-of-force claim. No court has adjudicated Boyland’s cause.

Sicknick’s record contains two statements by the same examiner, and both are required. Chief Medical Examiner Francisco Diaz ruled the manner natural and told the Washington Post that the autopsy found no evidence of an allergic reaction to chemical irritants and no evidence of internal or external injuries. Diaz also said Sicknick had engaged the rioters and that “all that transpired played a role in his condition.” No one was charged with killing him; two men were charged with assaulting him with chemical spray. The District’s Police and Firefighters’ Retirement and Relief Board separately found Jeffrey Smith’s January 6 injuries the sole and direct cause of his later suicide—the one formal causal adjudication attached to an aftermath death in the ledger.

The phrasing in public argument and the examiner’s record are different objects. Attorney General Merrick Garland’s January 2025 formulation referred to “the five officers who lost their lives in the line of duty”—Sicknick and the four later suicides. The examiner recorded no officer homicide and no officer death on January 6 itself. Line-of-duty recognition by the Capitol Police, the Justice Department, and benefit boards is an institutional determination printed in the research record; it is not a medical-examiner finding that rioters killed five officers. Any total drawn from the table must say whether it is counting deaths of the day, the next day, or the aftermath.

No corresponding national ledger exists for civil unrest as a whole. No federal body publishes deaths and injuries from riots by side or in total; the FBI’s LEOKA series counts officers feloniously killed in all circumstances—73 in 2021 and 64 in 2024, for example—without a riot subset for killings. Protester-side totals exist only as constructions: Forbes and Associated Press reporting counted about 19 deaths in the opening weeks of the 2020 unrest; the Guardian and ACLED counted at least 25 across 2020; an academic dataset counted 31 associated homicides; and an open-compiled multi-year list reaches 42, each under a different inclusion standard. The absence of a common instrument is part of the fifty-year record.

The Frameworks Applied: Where the Authors Land

This section does the only application work an objective file may do: it records where the frameworks’ own authors and other named scholars land, and the documented facts each side cites. It renders no measurement of its own. Under Paxton’s framework, per Paxton, the label applies, on the ground he names: open encouragement of civic violence to overturn an election. The facts his side cites: January 6, 2021; the 2020 election-overturn effort, documented in the January 6 committee record and the Georgia indictment record; the mass pardons and commutations of January 20, 2025, about 1,500 January 6 defendants, including men convicted of assaulting police; and the president’s “poisoning the blood” rhetoric of late 2023 and after, in his own words. The facts the other side cites within the same framework: Paxton’s stage model requires elite collaboration abandoning democratic liberties and redemptive violence without legal restraint as governing practice; the 2026 midterms proceeded and removed the president’s party from Congress; and the courts operate against him, as filed above. Under Griffin’s framework, per Griffin, the label does not apply: no coherent revolutionary ideology, no new-man project, no party militia, no project to replace the constitutional order. The facts his side cites: the movement’s program is restorative—“again”—rather than futural-revolutionary; its economic core is deregulatory; its organization is a personalist electoral coalition inside an existing party. The facts the other side cites: Griffin himself calls the MAGA slogan palingenetic in form, and critics inside fascism studies argue the rebirth myth need not be coherent to function, since Griffin’s framework runs on mythic core. Under Stanley’s pillars, Stanley applies all ten to the current United States in his own writing; his framework expressly contemplates fascist politics inside a still-functioning democracy, so the institutional record filed above does not, within his framework, refute the application. His critics, Griffin and Payne among them, argue the pillars describe illiberal populism generally and thereby dilute the category.

The comparative floor under the whole argument is a matter of record, and both sides of the argument stand on it. The reference regimes’ documented record includes, for Nazi Germany, the industrialized murder of approximately six million Jews and millions of others, a world war killing tens of millions, and the abolition of elections, parties, unions, and courts as independent bodies within months of taking power; and for Fascist Italy, the abolition of parliament, the murder of Giacomo Matteotti, the racial laws, and wars of conquest. No sourced element of the 2025–2026 United States record assembled in this file is of that category or scale, and the abolition list stands unexecuted in the institutional record filed above. That floor underlies two different uses, and the file prints both and takes neither: the argument that the label, applied to the present, fails by orders of magnitude; and the position of the scholars recorded above that the label, applied now, describes a politics and a trajectory—in Paxton’s terms, the territory of his first three stages—not a completed regime, which none of the four framework authors in this file claims.

The Cocked Hat

Three bearings never give a point; they give a small triangle, and this paper prints the triangle. The first side is the definition itself: the field’s frameworks disagree about the threshold, and their authors, applying their own instruments to the same man, reach opposite conclusions—Paxton and Stanley apply the label, Griffin and Payne withhold it, and Eco’s list, by its author’s design, cannot settle the question it is most often used to settle. No agreed metric exists, and the democracy indices that moved sharply this year measure a different quantity than the one in this report’s headline. The second side is the compliance record: the fifty-year baseline contains no judicial finding of presidential defiance, and the 2025–2026 record contains such findings in writing from multiple judges, an appellate reversal of the largest contempt proceeding, compliance in the flagship case after 83 days, and compliance with the Supreme Court’s two largest merits rulings—and no official scorekeeper exists by whom that record could be summed into a rate. The third side is the deaths record: the custody list is the government’s own and its current fiscal year is its highest, and the operations category beside it has no official list at all, so the one total the public argument most wants does not exist in any source and is not manufactured here. Around those three open points, the bearings that are single-valued are these: the midterms were held on schedule four days ago and the opposition won both chambers on the Associated Press’s calls, uncertified as this report files; no opposition party is banned and no news organization has been shuttered by the government in the record; the detained population and the custody death count stand at records on the government’s own figures while removal totals sit inside the historical range on the same government’s yearbook; and Christian nationalism measures at one American in ten. The log proceeds.

The Companion Essay

The argument this file does not make is published by this paper’s publisher on his own program, The Chris Abraham Show, in his own voice, as “The Worst Hitler Ever.” The essay takes the position the record above is most often argued from: that the president is the worst fascist ever—that he cannot do the thing the label names, that the villainy is performed rather than executed, and not jackbooting enough to register on the frameworks’ own scales. This paper prints no part of it and adopts none of it; the link stands so the reader can follow the argument the desk declined to carry, from the record above to the case the record supports in another man’s telling.