BEFORE IT REPEATS

A Sourced Analysis of Authoritarian Escalation Against Targeted Communities in the United States

Release 1 — Public Intelligence Briefing

March 2026

Starwater Research / Starwater LLC

With full context and power for truth, while seeking freedom and safety of all — may respect prevail.

I. Foreword

You are holding this document because I believe you deserve to see the full picture.

I am writing this from inside one of the targeted communities. I have spent immense personal effort compiling what no single news cycle, no single organization, and no single academic discipline has assembled in one place: what is being done to us — right now, in the United States — who built the infrastructure, how much money funds it, which historical patterns it replicates, and what the people who faced those patterns before did about it.

This work does not claim neutrality. It claims honesty. The details change. The structure does not.

The central thesis, held as hypothesis and tested by evidence across forty research reports, is this: what is occurring in the United States is not a collection of unrelated policy disputes. It is a coordinated escalation against targeted communities — transgender and gender-nonconforming people, the broader LGBTQIA2S+ community, immigrants, journalists, political dissidents, racial minorities, disabled people, labor organizers, and others — built on infrastructure that has been funded, designed, and deployed over decades, using mechanisms that the historical record documents with precision. The targeting of different communities is treated as separate stories — trans rights on one beat, immigration on another, press freedom on a third — and that fragmentation serves the architecture of targeting, because the infrastructure is the same, the money is the same, the organizations are the same, and the legal mechanisms are the same. This work makes that visible.

A note on method: this research was conducted with AI assistance. This is stated transparently. I directed every question, shaped every analytical decision, evaluated every finding, and take responsibility for every conclusion. The AI system’s role was to process volume — forty reports, thousands of sources, dozens of cross-connections — at a scale no individual researcher could manage alone. That this scale of compilation required AI assistance is itself relevant to the analysis. If a system with substantial analytical capacity struggles to hold the full picture in one pass, that tells you something about the nature of what is being documented. The difficulty of assembling this information is part of the evidence.

This is Release 1 — a public intelligence briefing establishing the pattern, mapping the current situation, showing the cross-group connections, and providing pragmatic guidance. A comprehensive second release, expanding every section with additional depth, additional case studies, and additional community-specific analysis, is in development.

You are trusted. Trusted enough to receive the full picture and decide what to do with it. That decision is yours. This work makes sure you can make it with open eyes.

There is a window. It is still open. This document is about what you can see through it — and why it matters that you look now.

This document discusses suicide, institutional harm, family rejection, detention, conversion therapy, and targeted persecution in documented detail. Sections IV and V contain the densest clinical and institutional material. Some of this will be difficult — particularly if you or someone you love is directly affected. You have permission to stop, to skim, to read in pieces, to come back later. The information will still be here. If you are in acute distress, you may want to read Section VI (What Is Working) first — it documents active resistance, reversal data, and practical preparation. Take care of yourself first.

A Note on Terminology: LGBTQIA2S+ and LGBTQ+

This work uses “LGBTQIA2S+” — lesbian, gay, bisexual, transgender, queer, intersex, asexual, Two-Spirit, and additional identities — in its opening, closing, and highest-impact statements to honor the full breadth of the communities affected. The inclusion of “2S” recognizes Two-Spirit people, whose identities predate colonization and whose erasure is itself part of the pattern this work documents. In running text, “LGBTQ+” is used for readability, as is common in policy and research contexts. Both forms refer to the same broad community; neither is intended to exclude anyone. Where a source, organization, or piece of legislation uses a different form (such as “LGBTQI” or “LGBT”), that original usage is preserved in quotation. The choice to use the full acronym at key moments reflects a deliberate decision: when the stakes are highest, precision and inclusion matter most.

II. What This Work Claims and What It Does Not

Before presenting the evidence, this section establishes the analytical framework, credits the scholars whose work makes this analysis possible, addresses the strongest objections to this project, and defines the language discipline that governs every claim in every section that follows.

Six Intellectual Predecessors

This work builds on six scholars. Each is credited honestly, including the strongest objections their work raises against this project.

Gregory Stanton created the Ten Stages of Genocide[^2] as an early-warning framework. This work uses Stanton’s stages as diagnostic markers — instruments for recognizing escalation patterns before they reach terminal stages. Stanton himself placed the United States at Stages 3, 4, and 10 regarding Native and Black Americans. This work documents indicators consistent with Stages 3 (Discrimination), 4 (Dehumanization), and 10 (Denial through data erasure) regarding targeted communities broadly. It does not claim genocide is imminent. It claims the early-warning indicators are present, and that the entire purpose of early-warning frameworks is to identify patterns before it is too late to act on them.

Robert Paxton defined fascism with behavioral specificity[^3] precisely to prevent the word from becoming meaningless through overuse. This work does not use the word “fascism.” It uses “authoritarian escalation” and lets the documented evidence speak for itself. Paxton’s concept of “stage arrest” — most escalatory movements stall without reaching their terminal stages — is central to this work’s honest assessment. Roger Griffin argued that what is emerging may be “something worse” than fascism because it operates within democratic structures rather than overthrowing them. The documented evidence supports Griffin’s observation.

Hannah Arendt identified five mechanisms of totalitarian control[^4]: the destruction of factual reality through propaganda, the progression from isolation to loneliness, bureaucratic “rule by Nobody,” the imperial boomerang, and what Bettina Stangneth corrected in the popular understanding of the “banality of evil.” This work deploys Arendt’s mechanisms as diagnostic indicators — not as proof that the United States is experiencing totalitarianism. Dana Villa[^5] insists there is “still a world of difference” between authoritarianism and totalitarianism. This work accepts Villa’s challenge. The United States retains an independent judiciary, opposition parties, a free press, active civil society, and federalism. Arendt’s subjects had none of these. Roger Berkowitz’s formulation — identifying “intermediate stages and analogies” while insisting we “not abandon reality” — is the methodological model this work follows. The presence of Arendt’s mechanisms is diagnostic, not definitive.

Cedric Robinson established that racial hierarchy is constitutive of capitalism, not incidental to it[^6]. Ian Haney López documented how coded racial appeals fragment the cross-racial solidarity that would threaten concentrated wealth. This work extends Robinson’s framework to gender, sexuality, disability, and immigration — an extension made by scholars including Ruth Wilson Gilmore, Jodi Melamed, Dean Spade, Lisa Duggan, and Sami Schalk. The extension is the current state of the scholarship. What this work adds is empirical documentation: the One Big Beautiful Bill Act[^8] fused tax cuts with community targeting in a single legislative instrument. The structural interdependence is not theoretical. It is legislative fact.

Nancy MacLean documented the sixty-year pathway from James Buchanan’s economics to the Koch network’s institutional infrastructure[^7]. Her work was attacked for methodological overreach. This project’s financial documentation is substantially stronger because the post-2017 disclosure environment — IRS 990 filings, ProPublica investigations, the $1.65 billion Barre Seid transfer — provides data that archival research alone could not. This work distinguishes between documented coordination (Project 2025, Leo judicial appointments, ADF model legislation, the Heritage-Danube Institute agreement) and inferred alignment (corporate DEI rollbacks, broader institutional compliance). Where the evidence documents coordination, this work says so. Where it infers alignment, it marks the inferential step.

Language Discipline

This work maintains a five-category distinction that governs every claim:

“Documents show…” — Documented fact. Named sources, verifiable dates, public records.

“The pattern across [X] institutions…” — Documented pattern. Multiple independent data points converging.

“This suggests…” — Supported inference. A logical connection between documented facts, with the inferential step marked.

“The parallel to [historical case] is…” — Analogy. Illuminating, not proving equivalence.

“If current trends continue…” — Projection. A trajectory based on evidence, not a prediction.

The reader can trust that when this work states something as fact, it is documented. When it draws an inference, it says so. When it makes a comparison, it names the limits. This discipline is not a formality. It is the difference between a work that can withstand hostile scrutiny and one that cannot.

III. The Historical Record

History is not a threat. History is a gift. It is the only evidence we have of what happens next, because what happens next has happened before.

The four case studies that follow were selected because each illuminates a different mechanism of the system being documented: legal infrastructure that outlives the regime that built it, government neglect as a form of targeting, covert operations reframed as domestic security, and an interlocking legal architecture designed so that no single law can be challenged in isolation. Each case includes both the escalation arc and the resolution arc, because how something ended is as analytically important as how it began.

Paragraph 175: The Law That Outlived Everything

In 1871, the newly unified German Empire adopted Paragraph 175 of the criminal code, criminalizing sexual acts between men.[^9] The law was not a Nazi invention. It predated the Nazi regime by sixty-four years. What the Nazis did was expand it — broadening the definition in 1935 so that a glance, a touch, an alleged intention could result in prosecution. Under the expanded statute, convictions surged. Men were sent to concentration camps, where they wore pink triangles and suffered mortality rates estimated between 60 and 65 percent — among the highest of any group.

When the Nazi regime fell in 1945, the law did not fall with it. Democratic West Germany retained Paragraph 175 in its Nazi-expanded form and continued to prosecute men under it for another twenty-four years. Between 1945 and 1969, approximately 50,000 men were convicted under a democratic government using a statute that had been weaponized by a fascist one. East Germany, under Soviet influence, reformed the law by 1968. West Germany, the democracy, took until 1969 to partially reform it and until 1994 — forty-nine years after the fall of the Nazi regime — to repeal it entirely.

The total: 123 years. 140,000 convictions. A law that persisted through an empire, a republic, a dictatorship, an occupation, a partition, and a reunification.

What this teaches: Legal infrastructure outlives regimes. Once a definitional framework is embedded in a legal system — once the bureaucratic machinery exists to classify, identify, and process a population — that machinery acquires gravity independent of the ideology that created it. Paragraph 175 persisted not because every postwar German official was a Nazi, but because the law existed, the courts existed, the precedents existed, and repealing it required affirmative political will that took decades to materialize. The parallel to EO 14168’s definitional cascade through every federal and state system¹ is structural, not analogical: a single definitional change is now propagating through agencies, courts, registries, and benefit systems in the same self-reinforcing pattern.

But the law was repealed. In 2002, the German government formally pardoned all men convicted under Paragraph 175 during the Nazi era. In 2017, it extended pardons to those convicted under the democratic government as well and established a compensation fund. It took 132 years from enactment to full acknowledgment. It took organized people who refused to accept that the law’s persistence made it permanent.

ACT UP and the AIDS Crisis: When the Government Chose to Watch People Die

Between 1981 and 1996, the United States government’s response to the AIDS epidemic killed people through deliberate inaction.[^10] President Reagan did not publicly say the word “AIDS” until 1987, by which time over 20,000 Americans had died. Funding was delayed, research was deprioritized, and the communities most affected — gay men, intravenous drug users, Black and Latino communities — were treated as populations whose deaths were politically acceptable. This was not neglect born of ignorance. The administration had the epidemiological data. It had the institutional capacity to respond. It chose not to. The mechanism was the same one documented in the current healthcare erosion: knowledge of disproportionate impact, available alternatives, and ideological alignment with the targeting rhetoric.

ACT UP — the AIDS Coalition to Unleash Power — was founded in March 1987 at the Lesbian and Gay Community Services Center in New York City. Within three years, it had 140 chapters worldwide and had developed what remains the most effective model of crisis resistance in the modern democratic record: a dual strategy combining direct action with institutional mastery.

The direct action is what people remember — die-ins at the FDA, ashes thrown on the White House lawn, political funerals where the bodies of the dead were carried through the streets. But what changed policy was the Treatment and Data Committee, later the Treatment Action Group, which produced activists who understood clinical trial design, FDA regulatory processes, and pharmaceutical development pipelines with the same fluency as the scientists running the programs. They did not merely demand access to treatment. They demonstrated, with technical specificity, how the existing regulatory framework was killing people by delaying access to drugs that worked, and they proposed specific procedural reforms that the FDA ultimately adopted.

The results were measurable: the FDA overhauled its drug approval timeline. The Ryan White CARE Act[^27] provided $3.2 billion. Protease inhibitors, when they arrived in 1996, turned a death sentence into a manageable condition — for those who could access them.

What this teaches: Communities that master the systems threatening them gain leverage that moral arguments alone cannot provide. The convergence of street protest and institutional fluency — outsider pressure and insider expertise operating simultaneously — is the pattern that produces results. ACT UP did not win because it was loud. It won because it was loud and right and technically precise, all at the same time.

But the cost was real. Research on long-term ACT UP participants documented elevated PTSD rates and deep psychological costs.[^56] Forty-nine percent of survivors in one study reported they had never planned for old age because they did not expect to reach it. Spencer Cox, who helped design the clinical trials that led to protease inhibitors, died at 44 of AIDS-related complications because he had stopped taking the medications he helped bring into existence. The movement fractured along race, class, and gender lines under partial success — the same fracture pattern visible in current LGBTQ+ advocacy, where trans people of color bear the heaviest targeting while resources flow to more privileged segments. Any resistance model that does not account for what resistance costs the people doing it is incomplete.

COINTELPRO: The Government’s Playbook for Reframing Dissent as Threat

Between 1956 and 1971, the FBI operated a series of covert programs under the name COINTELPRO — Counter Intelligence Program — targeting civil rights organizations, Black liberation movements, the antiwar movement, the American Indian Movement, and others.[^11] The programs were not investigations. They were operations: infiltration, disinformation, harassment, blackmail, and, in documented cases, assassination. The stated purpose was to “disrupt, misdirect, discredit, and otherwise neutralize” organizations the FBI deemed threats to national security.

The mechanism that matters most for this analysis is the reframing of political dissent as a security threat. Martin Luther King Jr. was not surveilled because he was a criminal. He was surveilled because the civil rights movement threatened the existing power structure, and the FBI’s institutional mandate could be redirected to frame that threat as a matter of national security rather than political disagreement. The “Black Identity Extremist” designation of 2017 — manufactured from six unrelated incidents — and the TIVE (“Transgender Ideology-Inspired Violent Extremism”) proposal manufactured from a non-transgender shooter follow the same pattern: fabricate a threat category from cherry-picked or misattributed incidents, then use the category to justify surveillance and enforcement.

The Church Committee[^11] exposed COINTELPRO in 1975 and recommended a statutory FBI charter to replace executive-branch-controlled Attorney General Guidelines. That charter was never enacted. The AG Guidelines proved malleable — amended four times under the Bush administration alone. The current NSPM-7 and Bondi memo architecture[^14] exploits precisely the gap the charter would have closed. The FBI’s Security Index operated for 39 years without statutory authority. The Bondi list has no more legal basis today than the Custodial Detention Index had in 1939.

What this teaches: The shift from illegal covert operations to legally shielded operations may be the most significant structural evolution in the targeting infrastructure. COINTELPRO was exposed and condemned. Its successors are designed to be legally unchallengeable — not because the activities have changed, but because the legal architecture has been redesigned to shield them from the kind of scrutiny that the Church Committee brought.

Apartheid South Africa: The Full Arc

The South African apartheid system was not a single law. It was 148 interlocking statutes[^12] designed so that each required every other — classification leading to segregation, segregation enabling control, control justifying suppression. The Population Registration Act of 1950 classified every person by race. The Group Areas Act segregated them geographically. The Bantu Authorities Act created the legal fiction of independent homelands. And 145 additional laws wove the system together so tightly that challenging any single element left the rest intact.

The resistance that ultimately dismantled this system was not any single force. It was five forces converging simultaneously[^13]: international sanctions that made the economic cost of apartheid untenable, internal resistance (the United Democratic Front with 700+ affiliated organizations and 3 million members, plus COSATU labor strikes that shut down key industries), an economic crisis that the regime could not resolve without Black labor participation, a military stalemate in Angola and Namibia, and the end of the Cold War which eliminated the regime’s strategic value as an anti-communist ally.

No single force was sufficient. Moral arguments alone did not end apartheid. International pressure alone did not end it. Internal resistance alone did not end it. The convergence — all five forces operating at the same time, each amplifying the others — is what produced the negotiated transition that led to the 1994 elections.

What this teaches: The Convergence Model. Resistance succeeds when multiple vectors of pressure operate simultaneously. This is the most consistent finding across every historical case in this research: zero cases document successful single-vector resistance. ACT UP combined street protest with institutional mastery. Anti-apartheid forces combined international economic pressure with domestic labor action with political organizing with military resistance with geopolitical realignment. The current US situation is structurally unique in one critical respect: no external compulsion mechanism equivalent to international sanctions is viable against the world’s dominant military and economic power. What substitutes for external pressure — federalism creating alternative power centers, civil society depth, economic self-interest — is the strategic question this work identifies.

But the revolution was unfinished. South Africa dismantled formal legal apartheid but preserved the economic structures: a Gini coefficient of 0.63 (among the highest in the world), 72% of agricultural land still owned by white South Africans, 32.9% unemployment. Legal change without structural economic transformation left the conditions for regression intact. This is the central warning for any future recovery: the law can change while the machinery persists.

IV. The Current Situation

This section documents what is happening now — not chronologically, but mechanism by mechanism, because the system operates as an integrated architecture, not a sequence of unrelated policy decisions. Each mechanism is documented with sources. Where the evidence is strong, this section says so. Where it is contested or thin, this section says that too.

The following subsections document specific enforcement mechanisms in clinical and institutional detail, including conversion therapy practices and surveillance infrastructure. If these descriptions are difficult, the summary in Section VII provides the same conclusions without the operational detail.

The Definitional Cascade

On January 20, 2025, Executive Order 14168² redefined sex in federal policy to exclude transgender and nonbinary people from legal recognition. This single definitional change propagated through every federal system: an OPM memo imposing a 48-hour purge deadline on agency records, BOP Program Statement 5260.01 (February 19, 2026), which does not merely ban gender-affirming care for federal inmates but mandates what the American Psychological Association, the American Medical Association, and the American Psychiatric Association define as conversion therapy: forced tapering and discontinuation of hormone treatment, confiscation of social accommodations (binders, wigs, cosmetics), and mandatory psychiatric intervention — psychotherapy and psychotropic medication — to “manage” gender dysphoria, replicating the logic of pre-1973 forced psychiatric treatment of homosexuals on a captive population of approximately 2,000 transgender federal inmates[^38] — EEOC enforcement nullification under Commissioner Andrea Lucas, VA healthcare elimination, CDC data scrubbing, State Department passport and visa changes, Social Security Administration gender marker halt, cascading effects on state birth certificate recognition and Medicare billing codes, and as of March 11, 2026, a final rule replacing “gender” with “sex” on Diversity Visa lottery entry forms — with the State Department confirming, in response to public comments, that the biological sex at birth requirement now applies to all visa applications, not just the diversity lottery — extending the binary definition across the entire immigration system. The BOP issued Program Statement 5260.01 in direct violation of a standing class-wide preliminary injunction in Kingdom v. Trump, renewed three times by Judge Lamberth and extended through May 31, 2026 — meaning the federal government mandated conversion therapy practices for a captive population while a federal court order explicitly prohibiting that denial of care remained in force. At least eight inmates have reported being denied care in violation of the injunction. At least ten ICE contracts with private detention facilities have been stripped of transgender care requirements, affecting facilities from Bluebonnet, Texas to Buffalo, New York — converting detention itself into a mechanism of forced detransition.

The pattern of this cascade suggests design rather than coincidence. A single definitional change at the top of the federal hierarchy produced convergent effects across every agency and program that touches the lives of the people being reclassified. The parallel to apartheid’s 148 interlocking statutes is structural: each element reinforces every other, and challenging any single agency’s implementation leaves the rest of the system intact.

Kansas SB 244 extended this logic retroactively — invalidating approximately 1,700 driver’s licenses and 1,800 birth certificates with no grace period and a $1,000 private civil bounty. The Kansas Department of Revenue sent invalidation letters with no grace period — licenses were voided immediately — and not all affected individuals had received notification by the time the law took effect. On March 16, 2026, a Douglas County judge denied the ACLU’s request for a temporary restraining order in Doe v. State of Kansas, leaving the law in full effect during litigation. The voting rights implications are direct: invalidated IDs cannot be used for voter identification, extending document invalidation from identity into civic participation. The Lemkin Institute identified SB 244 as the first statute combining bathroom restrictions with a private monetary bounty for enforcement. Previously, document restrictions applied only prospectively. SB 244 established a new escalation model: not just preventing future recognition but voiding completed ones. Attorney General Kris Kobach set a March 26, 2026 deadline for compiling the names of affected individuals into a law enforcement database — converting document invalidation into population registration. Penalties for non-compliance include a Class B misdemeanor (up to six months in jail, $1,000 fine), 90-day license suspension, and a mandatory minimum five-day jail sentence for driving during suspension.

Idaho HB 752, which passed the Idaho House in March 2026, establishes felony-level criminal penalties for transgender people using restrooms aligned with their gender identity: a first offense carries up to one year in prison; a second offense within five years — including prior convictions from other states — is a felony carrying up to five years. The bill applies to government buildings and private places of public accommodation, with no “duty to depart” provision — arrest is possible on presence alone. Both the Idaho Fraternal Order of Police and the Idaho Sheriffs’ Association opposed the legislation. The escalation from Kansas’s civil bounty to Idaho’s criminal penalties — misdemeanor to felony, civil fine to prison sentence, bathroom restrictions to presence-based arrest — demonstrates what Heritage Foundation President Kevin Roberts named publicly as “radical incrementalism.”[^41]

On March 10, 2026, HHS finalized a rule stripping gender-affirming care from the Affordable Care Act’s Essential Health Benefits, meaning costs no longer count toward deductibles or out-of-pocket maximums nationwide — twenty-one states have filed suit.[^42] In the same month, a federal employee healthcare directive effective Plan Year 2026 prohibited coverage of hormone therapy and gender-affirming surgeries for more than 10 million federal workers and their dependents, while mandating coverage of “faith-based counseling” — a term encompassing what every major medical association classifies as conversion therapy. On February 28, NIH terminated over 300 transgender health research grants totaling more than $350 million — including basic biological studies using mouse models — and redirected funding toward “regret studies.”[^43] The combination is architectural: strip coverage (ACA), strip federal employee benefits, defund the research that documents outcomes, and redirect scientific resources toward producing evidence for further restrictions. The evidence base that future legal challenges, future policy debates, and future medical guidelines would depend on is being eliminated while the restrictions it would evaluate are being enacted.

Erin Reed, whose monthly state-by-state risk assessment maps have become a primary tracking tool for the transgender community, elevated her national designation for the United States to “Do Not Travel” for transgender people, including foreign citizens, in March 2026 — Kansas joining only Florida and Texas at the highest individual state risk level, with fifteen states at “Worst Laws” and six at “High Risk.”[^44]

Enforcement Without Law

The pattern across the current enforcement architecture is a systematic gap between what the law requires and what the government does.[^28] Bostock v. Clayton County (2020) remains binding Supreme Court precedent holding that Title VII prohibits employment discrimination against LGBTQ+ individuals. The EEOC refuses to enforce it. No statute authorizes domestic terrorist organization designations, but the Bondi memo creates a list anyway. The mechanism is not changing the law but creating a gap between the law and its enforcement — and this mechanism has been consistent across eighty years, from the Security Index of 1939 through COINTELPRO through the present.

Deacon and Litman coined the term “legalistic noncompliance” to describe what is documented at scale: 210 violated court orders across 143 cases, 57 incidents of defying or frustrating court oversight documented by the Washington Post, 26 cases documented by Just Security. The IRS shared 47,289 taxpayer records with ICE despite lacking legal authority — a federal judge found 42,695 individual violations. The pattern is consistent: execute the action, face the court challenge, comply minimally or not at all, and meanwhile dismantle the institutional capacity to produce the data, the service, or the protection in the first place.

NSPM-7[^14] was designed with five features that resist legal challenge: an anti-litigation clause disclaiming enforceable rights, a directive form targeting agencies rather than individuals (defeating standing), reliance on existing authority, national security framing triggering judicial deference, and deliberate vagueness. No direct court challenge had been filed as of March 2026.

The denaturalization pipeline demonstrates the escalation. Historical denaturalization rates averaged approximately 11 cases per year from 1990 to 2017. A June 2025 DOJ memorandum elevated denaturalization to a top-five enforcement priority for the Civil Division. USCIS field offices were subsequently directed to supply the Office of Immigration Litigation with 100 to 200 denaturalization referrals per month — a target of 1,200 to 2,400 cases annually, representing a more than hundredfold increase over the historical baseline. The mechanism weaponizes the definitional cascade: under 8 U.S.C. § 1182(a)(6)(C)(i), any discrepancy between a visa applicant’s gender marker and their sex assigned at birth can be classified as material misrepresentation — fraud. For green card holders, that accusation can be brought up to five years after permanent residence was granted. For naturalized citizens, there is no statute of limitations on civil denaturalization. The legal architecture converts a completed immigration process into a permanently provisional status, retroactively, for anyone whose documents reflect a gender transition.

The Surveillance Architecture

In February 2025 — thirty-one days after EO 14168 redefined sex to exclude transgender people — DHS’s Office of Intelligence and Analysis, the agency responsible for domestic intelligence gathering on U.S. citizens, quietly removed sexual orientation and gender identity from the list of categories protected against intelligence targeting. The previous policy explicitly prohibited conducting intelligence activities based on a person’s sexual orientation or gender identity. The revised policy prohibits targeting based on race, ethnicity, sex, religion, country of birth, nationality, and disability — but not gender identity. On March 11, 2025, after Bloomberg Government reported the change, DHS restored sexual orientation to the protected list. Gender identity was not restored.[^54] The result: as a matter of documented federal policy, DHS can conduct intelligence activities targeting someone for being transgender but not for being gay. The protection was not forgotten — it was selectively removed.

The integrated enforcement pipeline[^15] connects policy framework to operational capability: NSPM-7 provides the policy framework, the Bondi memo provides operational mandates, Joint Terrorism Task Force referrals provide the investigative mechanism, cash bounties provide lead generation, DOGE cross-agency database access provides the data backbone, Palantir’s ImmigrationOS ($30 million contract) provides data aggregation, Clearview AI and Mobile Fortify provide field identification, state registries and Medicare billing codes provide population identification, and FISA Section 702[^30] backdoor searches provide warrantless access to communications.

Each component of this pipeline has precedent. What is structurally novel is the simultaneous deployment of all components within an integrated framework — and the construction of a master database that connects them. DOGE has built an integrated immigration enforcement database aggregating data from the Social Security Administration, the Treasury Department, the Office of Personnel Management, HHS, and Veterans Affairs into a single searchable system — with Palantir’s ImmigrationOS ($30 million contract) providing the data aggregation platform.[^39] The scale of data access is not disputed — only the safeguards are. Separately, in June 2025, DOGE personnel uploaded the full SSA database to an unsecured cloud server without mandated security protocols, exposing personal data on more than 300 million Americans. In March 2026, the SSA Inspector General opened an investigation after a whistleblower alleged a former DOGE employee copied the NUMIDENT database — containing Social Security numbers, dates and places of birth, citizenship status, race, ethnicity, and parents’ names for more than 500 million living and deceased Americans — onto a personal thumb drive. The data flows are not theoretical. They are operational. FBI queries of Americans’ communications data under Section 702 surged 35% in 2025 — from 5,518 to 7,413 per month — a year after the Reforming Intelligence and Securing America Act expanded the definition of “electronic communications service provider” so broadly that any entity with access to a Wi-Fi router, a server, or a cell tower can now be compelled to turn over communications data, earning the law the characterization “Make Everyone a Spy.” Section 702 expires April 20, 2026. A bipartisan Government Surveillance Reform Act — requiring warrants for Americans’ communications and closing the data broker loophole — was introduced on March 12, 2026 by Senators Wyden and Lee, but faces long odds in an administration that has increased 702 query volume by 35%. The most likely outcome — a clean extension without reform — preserves the expanded ECSP definition and the absence of a warrant requirement (which failed by a 212-212 tie vote in the House). The result is a foreign intelligence authority, designed for overseas targets, feeding directly into a domestic enforcement pipeline targeting the same communities documented throughout this work.

The foreign-to-domestic surveillance pipeline — legal authorities created for foreign intelligence systematically migrating to domestic use through incidental collection, backdoor searches, raw intelligence sharing, and parallel construction — has been documented across 3.4 million FBI queries on Section 702 data in a single year. The DOGE database integration closes the last structural gap: it connects the surveillance inputs (702, Clearview AI, SWS25 tracking codes, state registries, Medicare billing codes) to the enforcement outputs (ICE operations, denaturalization referrals, benefit terminations) in a single integrated system.

The pipeline extends internationally. Secretary of State Marco Rubio’s February 2025 guidance to visa adjudicators on Executive Order 14201 directs consular officers to flag all suspected transgender visa applicants with a global tracking code — “SWS25” — for enforcement across the entire consular system.[^35] The guidance grants consular officers authority to deny visas based on “reasonable suspicion” of transgender identity and classifies any discrepancy between a visa applicant’s current gender marker and their sex assigned at birth as grounds for permanent ineligibility under material misrepresentation statutes. Immigration attorneys have confirmed the memo sets no limitation on implementation — consular officers could apply it to any and all visa categories, not just those covered by the originating executive order. The result is a global surveillance flag embedded in the visa processing system, operationalizing the definitional cascade documented in Section IV into the international enforcement architecture.

The Chilling Effect — How Laws Cause Harm Before They Pass

This is the strongest single finding in the research corpus.[^16]

The pattern across institutions is this: at least 42 hospitals paused gender-affirming care in states where that care remains legal under shield laws — including, in March 2026, NYU Langone’s shutdown of its Transgender Youth Health Program and Mount Sinai’s discontinuation of services for both new and existing transgender patients, both in New York, a shield-law state.[^45] Over 1,000 nonprofits rewrote their mission statements — 50% of which had zero federal funding. 65% of Fortune 500 companies exited the Human Rights Campaign’s Corporate Equality Index. Every major media outlet surrendered Pentagon press credentials rather than resist. The system does not require the government to enforce compliance directly. It delegates enforcement to the institutions themselves, through the combination of financial dependency, surveillance credibility, and selective visible punishment.

The New York hospital closures reveal a structural finding that updates the analysis: shield laws are necessary but not sufficient. Legal protection does not produce institutional compliance when the federal pressure architecture operates through funding mechanisms, regulatory threat, and institutional risk calculus that outweighs state-level protections. Erin Reed’s March 2026 risk assessment downgraded New York, Maine, Colorado, Arizona, and Pennsylvania — all states with legal protections — based on clinic closures and institutional capitulation. The pattern documented here is not the failure of shield laws. It is the emergence of a pressure architecture that operates beneath and around them.

The pre-inauguration corporate cascade of June through October 2024 is the most significant evidence point: institutions self-censored months before any government mandate was issued. The system was, in a word documented by the research, “pre-wired.” The mechanism is anticipated enforcement, not actual enforcement. This is almost impossible to challenge empirically — the compliance happened, the timing is documented, and the institutions acted before they were told to.

Arendt’s isolation-to-loneliness mechanism describes the psychological pathway: institutional isolation (hospitals cease care, employers retract policies, organizations rewrite missions) produces individual loneliness (the person has no institutional anchor), which produces either compliance or withdrawal. The chilling effect does not need to touch every person directly. It needs to remove the institutions that people rely on, and the people remove themselves.

The Human Cost of the Architecture

The mechanisms documented above — the definitional cascade, the enforcement gap, the chilling effect, the surveillance infrastructure — are not abstractions. They produce measurable harm in the bodies and minds of the people they target, and increasingly, in the people around them.

A 2024 study published in Nature Human Behaviour[^34] — peer-reviewed, with a sample of over 61,000 transgender and nonbinary youth — found that the passage of anti-trans laws in a state produced a 7% to 72% increase in suicide attempts among transgender youth under 18, depending on the type of law. CDC government data from the 2023 Youth Risk Behavior Survey documented that 25.9% of transgender high school students attempted suicide in the past year — approximately five times the rate of cisgender male students. These are not advocacy estimates. They are government surveillance data and peer-reviewed research with causal analysis.

The pathway from policy to self-destruction follows the minority stress framework documented across decades of research: external stressors (discrimination, rejection, hostile policy environment) produce internal responses (concealment, hypervigilance, negative self-concept) that compound into depression, anxiety, PTSD, substance abuse, and suicidality. The Family Acceptance Project documented that LGBTQ+ youth who experience family rejection are 8.4 times more likely to attempt suicide. Forty percent of homeless youth in the United States identify as LGBTQ+ — roughly six times their share of the general youth population — and 68% of those served by providers have a documented history of family rejection. The Supreme Court’s ruling in Mirabelli v. Bonta, permitting forced disclosure of students’ gender identity to parents, operationalizes this pathway: the state compels the outing, the family rejects, the youth has nowhere to go. The mechanism connects.

What is new — and what distinguishes the current moment from baseline minority stress — is the scale and simultaneity. A 2025 survey by the Movement Advancement Project and NORC found that 85% of transgender Americans report the current political environment has negatively impacted their mental health. Fifty-five percent of transgender respondents have taken active steps to reduce their visibility — going back into hiding. Eighty-four percent of transgender and nonbinary respondents have made major life decisions since November 2024: relocating, changing jobs, updating legal documents while they still can, crossing state lines for medical care. This is not the chronic background stress of being a minority. This is acute crisis response to an environment that is actively closing around them.

The pattern extends beyond the targeted communities. Peer-reviewed research published in Frontiers in Psychiatry in 2025 documented that 64.9% of people living in a threatened democracy exceeded the clinical threshold for political grief — a condition that mediated 41% of variance in depression, 30% in anxiety, and 36% in stress. Healthcare providers forced to stop providing care they know is effective report secondary traumatic stress and burnout. Teachers caught between forced outing mandates and their duty to protect students describe an impossible position. Parents of transgender children describe grief, fear, relocation, and the destruction of plans they had for their families’ futures.

And on June 17, 2025, SAMHSA — the federal agency responsible for mental health services — eliminated the 988 Suicide and Crisis Lifeline’s LGBTQIA2S+ specialty service.[^34] The service had been receiving approximately 60,000 contacts per month, with a 46% increase in use over the prior eighteen months. The government removed the crisis support for the population whose crisis it is actively producing. Crisis support still exists — the Trevor Project (1-866-488-7386), the Crisis Text Line (text START to 741741), Trans Lifeline (1-877-565-8860), and PFLAG (pflag.org) remain operational — but the federal infrastructure designed to serve this population was deliberately dismantled. This is the architecture operating as designed: increase the pressure, remove the support, and let the human consequences accumulate in ways that — because the data collection has also been defunded — can no longer be measured.

The people affected by what this document describes are not a narrow minority. They are your neighbors on Medicaid, your coworkers whose immigration status is now searchable in a federal database, your parents whose Medicare is being restructured, your children in schools where teachers are now required to report their identities to the state. More than half the country — an estimated 58% of the US population[^40] — is directly touched by these mechanisms. Not as a theoretical exercise, but through care they are losing, benefits being cut, documents being invalidated, or communities being placed under surveillance. The architecture targets specific communities first. It reaches everyone.

The Judicial Permission Structure

In June 2025, the Supreme Court decided United States v. Skrmetti[^17], applying rational basis review — the most deferential standard available — to state laws restricting transgender healthcare. Within two weeks, the Court issued GVR (Grant, Vacate, Remand) orders in four related cases, extending the framework to adult healthcare and identity documents. On November 6, 2025, the Court stayed the district court injunction in Orr v. Trump in a 6-3 decision, allowing the federal passport policy — which requires sex markers to reflect birth sex only — to take effect nationwide. Justice Jackson’s 10-page dissent, joined by Sotomayor and Kagan, was issued the same day a parallel case, Schlacter v. State Department, was also stayed. Multiple states have since cascaded the federal policy into state-level identity document restrictions — Kansas, Florida, Indiana, and Texas have moved to invalidate existing gender-marker changes on state IDs, extending the passport framework to driver’s licenses and birth certificates. The 4th Circuit subsequently extended the Skrmetti analysis to adult care bans. On March 2, 2026, the Court issued a 6-3 emergency order in Mirabelli v. Bonta — a shadow docket decision with no full briefing and no oral argument — reinstating Judge Benitez’s class-wide permanent injunction requiring California schools statewide to disclose students’ gender identity to parents. Justice Barrett’s concurrence, joined by Roberts and Kavanaugh, grounded the holding in parental rights as “deeply rooted” in American tradition — language that imports substantive due process doctrine into forced outing. Justice Kagan’s dissent, joined by Jackson, called it a “rushed and radical” resolution of “novel legal questions.” The majority explicitly invoked Mahmoud v. Taylor, the 2025 ruling that parents could opt children out of LGBTQ-inclusive storybooks, as precedent — citing the earlier case about classroom materials to justify compelled outing of minors. California filed an emergency motion on remand to the 9th Circuit on March 6, seeking an abuse exception to the injunction, but the class-wide injunction remains in force. The cascading structure is visible: each ruling bootstraps the next, each expansion cites the previous restriction as settled law.

The 6-3 supermajority that decided Skrmetti is itself a product of the infrastructure documented in this work. All six conservative justices are connected to the Federalist Society. The Leo network[^19] spent $504 million across nine core organizations to build the judicial pipeline that produced them. ADF[^18] operates the Blackstone Legal Fellowship that trains the next generation of judges — Justice Barrett was a paid Blackstone speaker five times; Judge Kacsmaryk hosted Blackstone interns and subsequently received the forum-shopped mifepristone case. The same dark money that funds the targeting legislation produced the judiciary that rules on it. This is not inference. It is documented through IRS 990 filings, public financial disclosures, and Heritage Foundation’s own reporting.

At the district court level, the administration’s loss rate is 70–80%. District courts have issued 18+ preliminary injunctions. But the structural tension is between district court receptivity and Supreme Court hostility. Trump v. CASA eliminated nationwide injunctions, meaning each district court victory has limited geographic reach. The gap between what district courts are willing to protect and what the Supreme Court permits is the defining structural tension of this moment.

Government Data Erasure

Within the first two weeks of the new administration, over 8,000 web pages were removed from federal agency websites.[^33] The OPM memo of January 29, 2025, imposed 48-hour deadlines for record purges. CDC, Census, DOJ, VA, and HHS data related to LGBTQ+ populations, hate crimes, domestic terrorism research, and police accountability were removed, defunded, or ceased. As of March 2026, 360 federal data collections have been scrubbed of Sexual Orientation and Gender Identity (SOGI) data fields across CDC, Census, and NIH — 83% through backdoor “non-substantive change” requests that bypass the public comment process required for significant methodological changes.[^46] The NIH’s termination of over 300 transgender health research grants ($350+ million) does not merely defund current studies — it eliminates the pipeline of evidence that future courts, future legislators, and future medical guidelines would require to evaluate the very policies now being enacted. The government is dismantling the instruments of measurement while expanding the mechanisms being measured.

The mechanism is not primarily the destruction of archives — though the FBI destroyed 330,000 pages of its Sex Deviates Program files in the late 1970s, and the CIA destroyed MKULTRA records. The current mechanism is the defunding of ongoing data collection: stopping the measurement so that future harm cannot be documented. The removal of hate crime tracking from DOJ, domestic terrorism research from NIJ, and LGBTQ+ health data from CDC eliminates the ability to document targeting while leaving the surveillance infrastructure that monitors the targeted population intact. The result is asymmetric visibility: the government can see the population, but the population — and the public — cannot see what is being done to it.

The most targeted instance of this erasure is in immigration detention. Congress mandated in FY2021 that ICE publish biweekly statistics on transgender people in custody.[^36] ICE complied through January 17, 2025 — its final report showed 47 transgender people in detention, an acknowledged undercount, with at least 700 book-ins documented between October 2020 and January 2025. Starting February 4, 2025 — the first report after the second inauguration — ICE omitted all transgender population statistics. It replaced “LGBTQI” self-identification language with coded bureaucratic phrasing. As of this writing, ICE has not reported transgender detention data for over thirteen months, in direct violation of a congressional mandate. The population is still being detained. It is no longer being counted. The Vera Institute — the organization that first identified the omission — documented that this is not a technical lapse. It is a policy decision to make a detained population invisible while the surveillance and enforcement infrastructure that targets them continues to expand.

V. The Infrastructure: Who Builds It, Who Funds It, Who Benefits

The targeting of different communities is treated as separate stories. It is not. The infrastructure is shared, the funding is shared, the organizations are shared, and the legislative mechanisms are shared. This section documents that infrastructure with dollar amounts, organizational names, and sourced connections.

The Financial Architecture

The unified dark money pipeline[^20] that funds the targeting infrastructure operates through five principal nodes: DonorsTrust ($1.36 billion in total giving), the National Christian Foundation ($6.03 billion in grants, $25 billion total since 1982), the Bradley Foundation ($987 million), the Koch network ($125 billion revenue empire), and the Leonard Leo network ($1.65 billion Marble Freedom Trust from a single Barre Seid donation documented through IRS filings[^19]). These five nodes fund the same organizations: Heritage Foundation, Alliance Defending Freedom, ALEC, the Federalist Society, Family Research Council, America First Legal, State Policy Network, and Consumers’ Research. The National Christian Foundation alone has donated $56.1 million to 23 organizations designated as hate groups by the Southern Poverty Law Center. $89 million went to ADF since 2016.

The same donors fund both economic deregulation and community targeting. This is not coincidental alignment. Project 2025[^21] is a published coordination document — a single blueprint containing both radical economic policy and systematic social targeting, produced by the Heritage Foundation with $50.7 million from Leo network advisory groups, now 53% implemented with 70% of the cabinet drawn from its contributor pool. The OBBBA[^8] is the legislative proof: a single reconciliation bill fusing permanent tax cuts with Planned Parenthood defunding, Medicaid gender-affirming care bans, SNAP benefit cuts ($187 billion), and border security funding — signed into law on July 4, 2025, after passing the House 218-214 and the Senate 51-50 with Vice President Vance casting the tiebreaker. Economic and social targeting in the same instrument, voted on as a package, signed into law as a package, by the narrowest possible margins. The structural interdependence that Cedric Robinson theorized is, in this bill, legislative fact.

The ADF Litigation Loop

Alliance Defending Freedom[^18] is the single most connected entity in the targeting infrastructure. It drafts model legislation (130+ bills in 34 states, 30+ enacted into law in 2022). States enact the legislation. ADF defends it in court. The captured judiciary upholds it (Skrmetti). The precedent enables new legislation. The cycle restarts.

The same dark money — 43% of ADF’s $110.7 million revenue anonymized through donor-advised funds — funds the legislation, the litigation, and the Blackstone judicial pipeline that produced the judges ruling on ADF-sponsored cases. This is a self-reinforcing loop: money produces legislation, legislation produces litigation, litigation produces precedent, precedent enables more legislation, and the judges deciding the cases were produced by the same pipeline that funded the bills.

Heritage Foundation President Kevin Roberts stated publicly in February 2026 that the Foundation’s strategy is to “outlaw” transgender identity through “radical incrementalism” — including extending bans to adult healthcare.[^31] This is not inferred coordination. It is a named admission of the strategy this work documents through financial and organizational evidence.

International Coordination

The coordination is formalized, not informal.[^22] The Heritage Foundation and the Danube Institute have a signed cooperation agreement. CPAC Hungary is held annually with Hungarian government funding. Documented payments flow from the Hungarian government to American conservatives: Rod Dreher ($105,000 per year), Christopher Rufo ($35,000), Gladden Pappin (appointed head of a Hungarian government institute). Italian Prime Minister Giorgia Meloni confirmed publicly that these networks exist.

The funding flows bidirectionally.[^23] Russian government sources are the single largest national funder of anti-LGBTQ movements in Europe ($186.7 million documented 2009–2018, $211.9 million 2019–2023). American Christian Right groups spent $54 million internationally from 28 organizations. ADF International operates in 104 countries with 1,500 cases. The network is not hub-and-spoke. It is a mutually reinforcing ecosystem where money, strategy, and legal frameworks circulate across borders.

The infrastructure extends beyond legal and financial mechanisms into cultural institutions. In March 2026, Representative Mary Miller (R-IL) attached an amendment to HR 1329, the bipartisan Smithsonian American Women’s History Museum authorization, stating: “The Museum may not identify, present, describe, or otherwise depict any biological male as female.” The amendment passed committee on a party-line vote after Democratic cosponsors withdrew. The bill would erase transgender women — including historically documented figures such as Sylvia Rivera, Marsha P. Johnson, and Lynn Conway — from a national museum. This is not a policy restriction. It is cultural erasure through federal institutional control: the government determining which Americans qualify for inclusion in the national historical record.[^47]

The economic costs of the targeting infrastructure are themselves documented evidence of its ideological rather than rational basis. A SPARTA analysis of the transgender military ban calculated $18 billion in accumulated operational investment loss, $1 billion in initial replacement costs, and the separation of 15,000 transgender service members — 73% of whom are senior enlisted with 12 to 21 years of experience. Healthcare costs for transgender service members averaged approximately $650 per year per person, totaling roughly $8 million — negligible against any rational cost-benefit analysis. The ban is not an efficiency measure. It is a targeting mechanism wearing an efficiency costume.[^48]

What this means: The infrastructure targeting communities in the United States is not a domestic phenomenon. It is part of a documented international architecture. The same organizations operating in the US operate in Europe, Africa, Latin America, and at the United Nations. The strategies tested in one country are exported to others. When Hungary restricts LGBTQ+ rights, the Heritage Foundation studies the results. When ADF wins a case in Belize, the precedent informs litigation in Alabama. The wall being built around one community in one country is connected to walls being built everywhere.

The window described at the beginning of this document is not a metaphor. It is the distance between where the infrastructure is now and where the historical record shows it goes. That distance is measurable, and it is narrowing.

If you have read through Sections IV and V, you have just absorbed a significant amount of difficult material. The documentation of harm is dense by design — because precision is what separates this work from speculation. But documentation is itself an act of resistance: naming the architecture precisely is the first step toward dismantling it. What follows is the evidence that dismantling is possible — and already underway.

VI. What Is Working, What Has Worked, and What You Can Do

Hopelessness is a tool of the people building the wall. This section does not hand them that tool.

What follows is documented evidence of what is currently working, what has worked historically, and specific guidance for practical preparation. The historical record is unambiguous: people who navigated deteriorating situations successfully acted while they still had documents, resources, and freedom of movement. Preparation is not panic. Preparation is what the people who survived did.

What Is Currently Working

Up to eighteen states plus the District of Columbia have enacted shield laws or equivalent executive protections[^29] protecting transgender people from out-of-state legal actions, up from fourteen states before 2025 — eight states expanded protections in a single year. Massachusetts enacted “Shield Act 2.0” banning interstate data sharing for enforcement of anti-transgender laws. These states, heavily overlapping with the State Attorney General coalition (19 states plus DC filing Oregon v. Kennedy), are building a parallel legal infrastructure.

The legal resistance core — ACLU, Lambda Legal, GLAD Law, NCLR, Transgender Law Center — has shifted strategy post-Skrmetti from federal equal protection arguments to state constitutional claims, ADA disability rights frameworks, and First Amendment challenges. Rhode Island Latino Arts v. NEA and Schiff v. OPM successfully challenged “gender ideology” restrictions as viewpoint discrimination. New Mexico fully offset lost ACA subsidies at the state level. California’s Attorney General sued a hospital for ceasing care in a shield-law state. Florida’s organized advocacy stopped every anti-LGBTQ bill in the 2025 legislative session — but the pattern across 2025–2026 shows this success was not durable and demonstrates tactical adaptation by the targeting infrastructure: when direct anti-LGBTQ bills failed, the mechanism shifted. Florida HB 1001, a sweeping anti-DEI bill that passed both chambers (77-37 House, 25-11 Senate)[^55] and awaits the governor’s signature, bans local government DEI funding and programs — including Pride festival funding, LGBTQ+ community organizations, and minority business programs. It is not explicitly anti-trans. It does not need to be. The bill’s vague language — officials face removal from office for violations — destroys the local support infrastructure that targeted communities depend on without naming them, achieving through a facially neutral mechanism what direct legislation could not. Companion bills SB 1010 (criminalizing doctors who provide gender-affirming care) and HB 1639 (the “Trans Erasure” bill) continue the direct attacks simultaneously.

The international counter-force is emerging as a structural factor. On February 13, 2026, the European Parliament voted 340-141 for “full recognition of trans women as women” in United Nations negotiations — with the center-right European People’s Party largely supporting the resolution. This placed the European Union on a direct collision course with the Trump administration at the UN Commission on the Status of Women in March 2026. The vote is significant not as a symbolic gesture but as an institutional counter-position from a major democratic bloc, operating through multilateral frameworks the current US administration has systematically undermined.[^49]

Domestic electoral counter-signals are documented. In the North Carolina primary of March 4, 2026, three Democratic incumbents who had broken with their party on anti-trans votes were defeated by overwhelming margins: Nasif Majeed (who cast the sole Democratic vote to override a trans bill veto) lost 69-27; Michael Wray (who appealed to Republican voters) lost 64-36; Carla Cunningham (who cast the sole Democratic vote supporting ICE cooperation) lost 70-22. The pattern documents that Democratic primary voters are punishing anti-trans defection at the ballot box — a counter-signal to the infrastructure’s strategy of peeling off moderate Democratic support.[^50]

Oregon expanded its shield law (HB 4088) specifically to resist federal subpoenas of hospital records — a direct response to the evidence-gathering mechanisms documented in the surveillance section. Montana courts blocked nearly the entire anti-trans legislative agenda on state constitutional grounds, demonstrating that state-level judicial resistance remains viable where state constitutions provide independent protections.[^51]

District courts remain the primary check: an administration loss rate of 70–80%, 18+ preliminary injunctions, active judicial resistance from judges including Lamberth and Bates in D.C. District Court. These victories are geographically limited by the Supreme Court’s elimination of nationwide injunctions, but they demonstrate that the legal infrastructure, while constrained, is not collapsed.

The infrastructure’s strategic adaptation includes a new vector: ballot measures. As of March 2026, four states are advancing anti-trans ballot measures for the 2026 midterm elections. Maine has confirmed a measure. Colorado certified two initiatives in March 2026 — Initiative 109 (mandating sports team participation by assigned sex at birth) and Initiative 110 (restricting gender-affirming surgeries for minors and stripping state and federal funding from providers). The Colorado measures were organized by Protect Kids Colorado, a far-right evangelical group whose co-founder called being transgender “the transgender plague,” with substantial organizational support from the Colorado Catholic Conference, which used designated parish signature collection weekends at holiday masses to gather signatures.[^52] This vector bypasses the legislative process entirely, placing community targeting directly before voters — and introduces a mechanism that, if successful, is harder to reverse than legislation because it carries the democratic legitimacy of a popular vote.

What the Historical Record Teaches About Resistance

Every historical case in this work’s corpus confirms a single pattern: successful resistance requires convergence. Multiple vectors of pressure operating simultaneously.

ACT UP’s dual strategy — street protest combined with institutional mastery — forced FDA regulatory reform and the CARE Act. The anti-apartheid movement required five simultaneous forces. No single-vector resistance has produced lasting results in any case documented across forty research reports.

The treatment literacy model — communities mastering the systems that threaten them — is directly applicable. The legal resistance organizations currently carrying the federal challenge possess the same kind of institutional fluency that ACT UP’s Treatment and Data Committee developed. The question is whether that fluency can be combined with the economic pressure, coalition breadth, and sustained political engagement that convergence requires.

One finding from the apartheid case study deserves specific attention: the United Democratic Front’s consumer boycotts achieved 100% compliance, effectively inverting the chilling effect against the regime’s economic base. The mechanism is directional, not fixed. The same institutional compliance that currently serves the targeting infrastructure can, under the right conditions, be redirected.

Pragmatic Preparation

Having a plan is power. Not needing to use it is also an outcome. The following guidance is priority-ordered by time sensitivity.

Documents: Obtain or update birth certificates in accessible states before legislative sessions close windows. Secure medical records before provider closures — multiple copies, stored separately. Begin the apostille process (2–4 months). Be aware that submitting a passport name-change application may trigger gender marker reversion under current federal policy.

Healthcare Continuity: If you or someone you depend on receives care that is under legislative threat, request your complete medical records now — not summaries, complete files — from every provider. Store copies separately from your home. Identify providers in shield-law states. If you take medications that require prescriptions, ask about 90-day supplies and understand your insurance’s mail-order options. If you rely on ACA marketplace subsidies, know your state’s options before open enrollment closes. The preparation is the same whether you end up needing it or not.

Digital Security: Use Signal for sensitive communications. Use a VPN consistently. Opt out of data broker databases (this is a multi-step process requiring applications to each broker). Review organizational security — a 5-layer threat model begins with individual device security and extends through organizational communications to public-facing digital presence.

Financial Preparation: Understand FATCA reporting requirements for any international accounts. Research banking options in destination countries. Document relocation cost estimates by destination. Address student debt strategies before departure if relevant.

Community and Emergency Readiness: Know your people. Have a contact list that does not depend on a single app or a single device. If you have children, make sure custody documents are current and accessible. If you are in a state that has criminalized aspects of your identity or your family’s medical decisions, know which neighboring states have shield laws and what crossing into them requires. Have a bag packed. Have a plan for your pets. These are the things people who navigated deteriorating situations tell you they wish they had done sooner.

Credentials: Begin credential evaluation for any country of potential relocation. Healthcare, legal, engineering, and teaching credentials all have country-specific portability requirements that can take months to navigate.

Exit Planning: This work’s research assessed eight countries in detail: the Netherlands, Uruguay, Spain, Germany, Finland, Portugal, Chile (subsequently downgraded due to political shifts), and Canada. Each has distinct advantages and limitations across legal protections, healthcare access, cost of living, community infrastructure, and residency pathways. Digital nomad visas, skilled worker programs, and asylum pathways are all documented with specific requirements.

Staying Informed: The organizations tracking this in real time are your primary sources: Trans Legislation Tracker (translegislation.com) monitors every bill in every state. Erin in the Morning (erininthemorning.com) publishes daily reporting and a monthly state-by-state risk map. The ACLU Litigation Tracker follows all active federal lawsuits. Movement Advancement Project (lgbtmap.org) maps shield laws, policy landscapes, and state-by-state protections. For legal help: Lambda Legal’s Help Desk, Transgender Law Center, GLAD Law, and the Sylvia Rivera Law Project provide direct assistance. For crisis support: the Trevor Project serves LGBTQIA2S+ youth, and PFLAG supports families. For immigration-specific guidance: the National Immigration Law Center. Know who your state legislators are and how they voted. Know your state attorney general’s position. The people who changed outcomes in every historical case in this work were the ones who knew what was happening while it was still happening. Updates and additional resources are maintained at starwater.xyz.

The choice between staying and leaving is yours, and both have worked historically. People who stayed and fought changed the system from within — ACT UP, the anti-apartheid movement, the civil rights movement. People who left preserved their lives and their communities and returned when conditions changed — or built new lives elsewhere. The historical record does not privilege one choice over the other. It privileges being prepared to make the choice from a position of information and agency rather than crisis and desperation.

VII. Honest Assessment

The evidence has been presented. This section draws the conclusion.

Seven independent democracy assessments[^24] converge on the same assessment: the United States is experiencing democratic backsliding at a scale and pace not previously recorded in its history. Freedom House scores have declined from 93 to 81 — the 2026 report, published in March, dropped the United States three points in a single year, to its lowest score in 54 years, ranking alongside South Africa. V-Dem’s 2026 Democracy Report (released March 17, 2026) reclassified the United States from “Liberal Democracy” to “Electoral Democracy” — the first such downgrade in over fifty years — and dropped the country from 20th to 51st globally in its Liberal Democracy Index, a 24% score loss in a single year, the steepest single-year decline V-Dem has recorded for the United States. Freedom of expression in the US is now at its lowest level since the end of World War II. Founding director Staffan Lindberg warned that if current trends continue, the United States faces further reclassification to “electoral autocracy.” Autocracies now outnumber democracies globally 91 to 88 for the first time in twenty years. Bright Line Watch expert ratings dropped from 67 to 55 — the steepest single decline since the survey began in 2017 — placing the United States between Israel and Mexico and “closer to a mixed or illiberal democracy than a full democracy.” The Economist Intelligence Unit’s global democracy average of 5.17 is the lowest since the index began. The US Corruption Perceptions Index score of 64 is its lowest ever, tied with the Bahamas. The Century Foundation’s US Democracy Meter — a 100-point assessment using 23 indicators — scored the United States at 57/100 in 2025, down from 79/100 the prior year: a 28% single-year decline, the steepest measured by any assessment tool identified in this research. Protect Democracy’s Authoritarian Threat Index — a composite metric tracking executive overreach, institutional erosion, and democratic norms — rated the United States at 3.4 out of 5 as of March 16, 2026, classified as “Severe Threat.” 72% of the world’s population now lives under autocratic rule. The United States is not an anomaly in these numbers. It is at the center of the worst decade for global democracy in a generation.

The Lemkin Institute for Genocide Prevention — a 501(c)(3) named after Raphael Lemkin, who coined the word “genocide” — issued its third Red Flag Alert on the United States on March 11, 2026.[^32] The Institute assessed that the United States is in the early to middle stages of a genocidal process against transgender people — an independent finding consistent with this work’s documentation of Stages 1 through 4. The Institute’s framework identifies a two-phase structure: the current phase of preventing identity expression and destroying institutions that reproduce identity, followed by a projected phase of containment through criminalization. Gregory Stanton, creator of the Ten Stages of Genocide framework, stated in connection with the alert that the United States is undergoing “a hardening of the categories” driven by ideology with historical precedent. This is institutional validation from the organization whose namesake created the legal framework for recognizing genocide — and whose analytical conclusion aligns with what the evidence assembled in this work independently documents.

The international response corroborates the assessment. As of March 2026, at least seven allied nations — Germany, Denmark, Finland, the Netherlands, Ireland, Portugal, and Canada — have issued government travel advisories with transgender-specific warnings about the United States, with the United Kingdom updating its guidance in more general terms.[^37] These are not advocacy organizations. They are foreign ministries advising their citizens that a NATO ally may deny them entry, detain them, or revoke their documents based on gender identity. The last time allied governments issued identity-based travel warnings about the United States at this scale, the context was Japanese internment. The diplomatic record is itself evidence.

What the Recovery Data Shows

Fifty-two percent of autocratization episodes have been reversed.[^25] Seventy-three percent in the last three decades. This means that reversal is the more likely outcome — but not the guaranteed one. Forty-eight percent of episodes were not reversed. And the Carnegie Endowment’s 2025 finding that “few recovering backsliders regain prior democratic quality” means that even the successful reversals result in diminished democracies.

Recovery timelines across five cases: Spain consolidated in approximately 7 years. South Korea has experienced cyclical democratic recovery over 17 years, most recently achieving a martial law declaration to life sentence in 14 months — the most encouraging case in the contemporary record. Argentina required 22 years. Chile still operates with elements of a Pinochet-era constitution after 35 years of democracy. South Africa achieved formal democracy but preserved apartheid’s economic structures.

Levitsky and Ziblatt[^26], whose 2018 work documented democratic erosion, assessed in December 2025 that the United States has entered “competitive authoritarianism.” Ruth Ben-Ghiat identified three features of the current moment that are unprecedented in her comparative study of authoritarian leaders: the speed of institutional damage, the scale of simultaneous multi-domain action, and the use of economic leverage to enforce compliance across the private sector.

What This Work Documents

Stages 1 through 4 of the Five-Stage Escalation Model are comprehensively documented across multiple independent sources: rhetorical dehumanization through the manufactured concept of “gender ideology” deployed across every federal agency and over 740 state bills under consideration across 42 states[^53] (125 passed in 2025 alone — the highest annual total on record, a 668% increase since 2021); legal reclassification through EO 14168 and its cascade; institutional capture through Project 2025’s 53% implementation rate and 70% cabinet penetration; and surveillance infrastructure through the integrated enforcement pipeline connecting NSPM-7 to the DOGE master database to Palantir to Clearview AI to FISA 702.

Overlap-adjusted population analysis — using Census, CMS, Gallup, Williams Institute, and USCIS data with conservative overlap reductions[^40] — estimates that at least 200 million Americans, approximately 58% of the population, are directly or structurally affected by the mechanisms documented in this briefing through healthcare cuts, immigration enforcement, document invalidation, benefit reductions, or surveillance infrastructure. The democratic erosion documented by seven independent indices affects everyone. This is not a minority issue. The targeting of minorities is the entry point for an architecture that reaches the majority.

Stage 5 — population-level targeting — is where this work draws the line between what is documented and what is projected. The infrastructure for Stage 5 is being built. The healthcare erosion (a projected 4.8 to 14 million newly uninsured as ACA enhanced subsidies lapse — the lower figure for 2026, the higher through 2034 if subsidies are not restored — and 5.3 million through Medicaid work requirements effective 2027), the mandated conversion therapy practices in federal prisons (issued in defiance of a standing court order), the retroactive document invalidation, the construction of a master surveillance database integrating foreign intelligence collection with domestic enforcement — these are real and documented harms happening to real people right now. But this work does not claim that population-level targeting at the scale of the historical case studies is currently operational. It claims that the infrastructure is being constructed, that the historical record shows what that infrastructure has been used for before, and that the purpose of documenting the pattern is to act before it reaches the endpoint — not after.

The approximately ten-year threshold for irreversibility, documented across multiple recovery cases, is presented as plausible but speculative. What is not speculative is that every recovery case in the corpus required free elections as a necessary precondition. No democratic recovery has been achieved without them.

What Changes the Outcome

The same historical record that documents escalation documents reversal. The same research that maps the infrastructure maps the resistance. The convergence model — drawn from every case study in this work — identifies what is required: institutional mastery and legal resistance, economic pressure that makes targeting more costly than accommodation, coalition breadth across targeted communities, and sustained political engagement that does not rely on a single electoral outcome or a single legal victory.

South Korea’s 14-month trajectory from martial law to conviction. Brazil’s former president convicted and banned for 27 years while deforestation fell 62%. ACT UP’s transformation of the FDA in three years. The repeal of Paragraph 175 after 123 years. The fall of apartheid when five forces converged. These are not inspirational anecdotes. They are documented evidence that organized people, applying sustained multi-vector pressure, have reversed these patterns before.

You now have what I set out to provide: the full picture, assembled in one place, grounded in evidence, connected across communities, tested against the historical record, and honest about both the trajectory and the capacity for change.

My position — stated here because honesty requires it — is that the evidence supports action, that the window for action is real but not unlimited, and that the people in every targeted community deserve to make their decisions with open eyes.

The window is still open. What you do with what you’ve seen through it is yours to decide.

With full context and power for truth, while seeking freedom and safety of all — may respect prevail.

A Note on Reading This After March 2026

This document is a snapshot. It reflects the evidence available through mid-March 2026. By the time you read it, court rulings will have been issued, bills will have passed or died, indices will have been updated, and developments will have occurred that this work could not anticipate.

The factual claims will age. The structural analysis will not. The mechanisms documented here — definitional cascades, chilling effects, judicial permission structures, dark money pipelines, surveillance architecture — do not depend on any single bill, any single ruling, or any single administration. They are architectural. They predate this moment and they will outlast it.

Check starwater.xyz for updates and corrections. Verify the endnotes against current data. Use this document as a framework for understanding the architecture, not as a real-time report.

And if, when you read this, the trajectory has changed — if the window has widened rather than narrowed — that will be the most important evidence of all: that people who organized, who applied sustained pressure across every vector documented here, changed the outcome. Again.

Monitoring as of March 2026

The pace of developments during the production of this work is itself evidence of the system being documented.

• 4th Circuit Extension (Fall 2025): The 4th Circuit Court of Appeals extended the Skrmetti rational basis framework to adult healthcare.
• SAVE America Act (H.R. 22): Senate vote proceeding week of March 16, 2026. Requires documentary proof of US citizenship to register for federal elections. 53 Republican votes but 60 needed for cloture; Senate Democrats opposed. Expected to fail at the filibuster threshold. Senator Cornyn broke with prior position to support rule changes (March 11), but leadership lacks the votes.
• Ninth Circuit M.H. v. Hamso reconsideration — pending.
• FISA Section 702 — expires April 20, 2026. House GOP leadership pushing clean reauthorization (18+ months, no reforms) week of March 19. Rep. Jim Jordan, previously a leading reform advocate, announced support for clean extension. Government Surveillance Reform Act (bipartisan warrant requirement, Wyden-Lee-Davidson-Lofgren, March 12) being sidelined. If clean extension passes, the surveillance architecture documented in this report continues without the warrant requirement that failed by a 212-212 tie in 2024.
• V-Dem Democracy Report 2026 — released March 18, 2026. US dropped from 20th to 51st globally, losing 24% of its score in one year. Reclassified from “Liberal Democracy” to “Electoral Democracy” — first time in over 50 years. Freedom of expression at lowest level since end of World War II. Autocracies outnumber democracies 91-88 globally for the first time in 20 years. Founding director warns US reclassification likely in 2027 report.
• Colorado ballot measures — Initiatives 109 and 110 certified for November 2026. Sports ban and minor care restrictions with funding strip. Parish-organized signature collection by Catholic archdiocese.
• Smithsonian Women’s History Museum — HR 1329 amendment banning depiction of trans women as female, passed committee March 2026. Cultural erasure through federal institution control.
• AYAGDOS study — Researchers behind the “Rapid Onset Gender Dysphoria” pseudoscience (including Lisa Littman and J. Michael Bailey) recruiting transgender young people for new study. Community organizations issuing participation warnings.
• Kansas AG database — Kobach compiling law enforcement database of affected transgender individuals, deadline March 26, 2026. Population registration mechanism.
• Idaho HB 752 — Felony bathroom ban passed House, advancing to Senate. First statute with felony penalties (5 years) for restroom use.

A Note on Sources

This briefing uses compound endnotes: each numbered reference aggregates the primary sources supporting the claims in its section, rather than citing a single source per footnote. The 40 endnotes below contain approximately 100 distinct, publicly verifiable sources — government documents, court filings, academic papers, tracking organizations, and investigative journalism — each cited by name, author, date, and URL where available.

These endnotes are drawn from a research corpus of approximately 4,350 source URLs compiled across the project’s research phase, organized into forty thematic research reports covering executive orders, legislative tracking, judicial developments, surveillance infrastructure, historical case studies, funding networks, international parallels, and practical preparation. The complete source archive, with Wayback Machine preservation links, publishes alongside this work at starwater.xyz for independent verification.

The source hierarchy privileges government documents and court filings above all other sources, followed by established tracking organizations (ACLU, Trans Legislation Tracker, Freedom House, V-Dem, Committee to Protect Journalists, Reporters Without Borders, Movement Advancement Project), named journalists at established outlets, academic papers, and advocacy organization reports. Where sources conflict, the conflict is noted. Where claims rest on a single source, this is stated. Where claims are unverified, this is stated.

This is political speech, civic analysis, and inquiry into matters of urgent public concern. It is protected under the First Amendment. It is offered freely as a public service.

Release 2 — the complete work, with expanded historical case studies, additional community-specific analysis, deeper solutions, and a comprehensive source archive — is in development.

Notes