The Indicator Framework v1.0

Premise

The Trump administration has been charged by many with authoritarianism. This site does not argue the charge; it keeps the record. Each entry weighs the evidence for and against an authoritarian reading: against history, against precedent, and against what the accumulated record already shows.

This page publishes the instrument in full: how entries are produced, the twelve standing indicators, the five-zone event scale, the ten scoring principles that gate every entry, the scholarly case against an authoritarian reading, the calibration cases the framework draws on, and the corrections and versioning policy.

How entries are produced

Entries on this site are produced by an automated research-and-synthesis pipeline operating under the published rules on this page. There is no human byline, and none is claimed. Three times a day the pipeline:

1. Polls a fixed set of sources: primary government documents (Federal Register, whitehouse.gov), court-filing data (CourtListener), and a cross-spectrum press layer from Fox News to NPR to Just Security (full list, with what each provides, on the Sources page).
2. Triages new items for whether they touch any of the twelve indicators below, in either direction. A contrary-signal item (court compliance, a reversal, power-sharing) is explicitly in scope, not noise, and a run that finds none says so rather than manufacturing a story.
3. Researches each triaged story against a published watchlist and sourcing rules, then retrieves pattern context: what the accumulated record already shows on the relevant indicators, entities, and threads.
4. Synthesizes one entry per story, structured per the scoring principles below, evidence-against populated before evidence-for.
5. Gates every entry through a compliance check before publication: against-cell populated, base rate named, disconfirmer specified, zone call justified, tone checked for alarm-adjectives. A failing entry is rewritten once with the failure specified, or it does not publish.

Automated production is disclosed, not hidden, because the alternative, a fabricated byline, would be worse. The About page explains why an exhaustive, checkable methodology is treated as a stronger trust signal than an author's name; in short, the COVID Tracking Project became a citation source for federal agencies on the strength of process transparency alone, with no institutional pedigree, and this site is built on the same bet.

The twelve indicators

Each indicator is a slow-moving standing reading, not a per-article score, decomposed from the frameworks surveyed below into a checkable statement with a specified condition that raises it, a specified condition that lowers it, and a named disconfirmer. Full provenance, current reading, and changelog for each live on its own indicator page; what follows are the full definitions.

Tier 1: the electoral / rule-of-law core.

1. Electoral integrity & acceptance of results. The government accepts the possibility of losing power and commits to peaceful transfer; election administration is not manipulated to favor incumbents. Raises: refusing to accept results, casting systemic doubt on legitimacy without evidence, manipulating administration or maps. Lowers: the opposition wins real elections and takes office; routine partisan redistricting has bipartisan precedent.
2. Judicial independence & compliance with court orders. The executive complies with adverse rulings and does not pack, purge, or delegitimize courts. Raises: defying court orders, delegitimizing judges, purging or packing courts. Lowers: the executive loses in court and complies; criticizing a ruling while obeying it is normal.
3. Weaponization of state institutions against opponents/critics. Coercive and investigative state power (prosecutors, tax authority, regulators, security services, inspectors general) is applied neutrally, not selectively against named rivals, media, or critics. Raises: investigations targeting opponents, retaliatory licensing/tax action, purging watchdogs. Lowers: the action is neutrally applied, precedented, or grounded in a legally authorized theory of executive control.
4. Freedom of expression & press freedom. The government does not use state power to punish, license-threaten, or coerce media and critics. Raises: licensing/defamation threats tied to coverage, retaliation against outlets, pressure on universities. Lowers: press-access disputes have precedent; a plural, adversarial media still operates freely.

Tier 2: checks-and-balances / liberal-democracy core.

5. Executive aggrandizement vs. Congress & statutory limits. The executive operates within statutory and constitutional limits and respects the power of the purse. Raises: unilateral decree beyond statute, impounding appropriated funds, ignoring statutory notice. Lowers: EO volume alone is a weak, bipartisan signal; unitary-executive theory legitimizes much intra-executive control. Look for illegality or defiance, not assertiveness.
6. Politicization of the civil service, law enforcement & military. Professional civil service, DOJ/FBI, and the armed forces remain insulated from partisan or personal loyalty tests. Raises: loyalty purges, installing loyalists in enforcement agencies, drawing the military into domestic politics. Lowers: normal appointee turnover; the professionalized military staying out of politics.
7. Toleration or encouragement of political violence. Leaders unambiguously condemn political violence and do not pardon, praise, or align with it. Raises: pardoning or praising politically motivated violence, ties to armed groups. Lowers: clear condemnation; isolated rhetoric without material follow-through.
8. Delegitimization of opponents as illegitimate/criminal enemies. Political rivals are treated as legitimate adversaries, not criminals or traitors. Raises: branding opponents enemies or criminals combined with a coercive move to punish them. Lowers: heated rhetoric alone, without coercive follow-through, is common and low-specificity.

Tier 3: corroborating / contextual indicators.

9. Civil-society space (NGOs, universities, protest). Independent associations operate without funding-based or legal retaliation. Raises: funding weaponization, punitive investigations of nonprofits or universities. Lowers: civil society remains active, funded, and litigating.
10. Corruption / self-enrichment & loyalty-over-law rewards. Office is not used for personal enrichment; positions are filled on competence, not fealty. Raises: kleptocratic patterns, systemic loyalty-over-law patronage. Lowers: ordinary ethics disputes, distinguished from systemic self-dealing.
11. Information environment / disinformation. The government does not systematically flood the information space with deliberate falsehood. Raises: coordinated state disinformation at volume. Lowers: ordinary spin; a plural media that can fact-check.
12. Selective application of law to minorities & critics. Neutral laws (immigration, terrorism, and similar statutes) are not applied selectively to punish minority groups or critics. Raises: selective-enforcement patterns targeting identity groups. Lowers: enforcement within precedent and due process.

The five-zone event scale

Every published entry carries exactly one zone reading for the event it covers. Zones are ordinal but are never rendered as a numeric score, gauge, or red-green meter. They render as a horizontal, text-labeled bar with a single tick, because a false-precision number is itself a credibility risk on a subject this contested.

Zone	Meaning
Contrary signal	Evidence cuts against the authoritarian reading: compliance with an adverse ruling, power-sharing, voluntary reversal, an opposition win acknowledged.
Within precedent	Base rates cover it; consistent with ordinary executive hardball across prior administrations of both parties.
Contested	A good-faith explanation and a pattern-consistent reading are both live; the evidence does not yet discriminate between them.
Pattern-consistent	Exceeds base rates in legality, scale, or selectivity; matches the documented mechanics of at least one indicator above.
Bright line	Crosses a defined bright line: defying court orders, election manipulation, encouraging violence, prosecuting opponents without predicate, and comparable acts.

Scoring principles

These ten principles are enforced by an automated compliance check before any entry publishes. An entry that fails is rewritten once with the specific failure named, or it is dropped and logged. It is never published ungated.

1. Against-cell first. The evidence-against column must be populated before the for column; an entry with an empty against column does not publish as written. It is rewritten, or the emptiness is itself stated and justified.
2. Base-rate test. Every for-reading states what prior administrations (Reagan, Bush, Obama, Biden, Trump's first term) did comparably, and whether this event exceeds that baseline in legality, scale, or selectivity, not mere existence.
3. Legality is not exoneration; selectivity is the tell. Unitary-executive arguments answer who may act, not whether power targets critics (Varol, "Stealth Authoritarianism").
4. Popularity is not exoneration. Winning real votes has coexisted with authoritarianization elsewhere; electoral support for a policy does not settle whether the policy is being applied selectively against critics.
5. The bundle is the signal. Pattern context on every entry must say whether other indicators are moving in the same period. A single indicator moving alone reads very differently from several moving together.
6. Elections-still-competitive is the master reversibility variable. Terminal readings are not warranted while the opposition can win and take office.
7. Every entry names its own falsifier. A "what would change this reading" section is mandatory, not optional.
8. Uncertainty is surfaced, not smoothed. Scholarly disagreement is cited by name when a zone call is close, rather than resolved silently in the writing.
9. Standing readings move conservatively. An indicator's standing reading changes only on pattern, bright-line, or contrary events whose weight warrants it, and every change is a dated changelog entry with a stated rationale.
10. Exculpatory coverage is mandatory. Contrary-signal events get full entries with the same prominence as any other. A run that finds no contrary-zone candidates triggers an explicit search for them in the next research pass.

The counter-literature

If this site cannot state the strongest good-faith case against an authoritarian reading of a given action, it has not measured that action. It has asserted a conclusion about it. The following is treated as load-bearing method on this site, not as a disclaimer appended for balance.

Unitary executive theory. Article II's Vesting Clause and Take Care Clause create, on this reading, a hierarchical, unified executive under direct presidential control. So removal of subordinate officials, direction of enforcement discretion, and control over the administrative state are constitutional design, not usurpation. The most rigorous academic defense is Calabresi & Yoo, The Unitary Executive (Yale, 2008), tracing the theory through every administration from Washington to G.W. Bush; the Supreme Court has increasingly embraced versions of it (Seila Law, Collins). This site applies the theory's own limit: even its proponents distinguish control of execution from weaponizing enforcement against opponents. The theory legitimizes who decides, not retaliatory selective targeting.

The descriptive-realist defense. Posner & Vermeule, The Executive Unbound (2010) argue that in the modern administrative and emergency state, legal ("Madisonian") checks are largely illusory, and the executive is in practice constrained by politics (elections, public opinion, party), not law. A president operating near the limits of law is, on this view, the normal condition of the modern administrative state; the operative question is whether political constraints still bind, not whether legal ones do.

Base rates. Executive-power expansion is bipartisan and long-run. Executive-order counts alone are a weak signal: Obama signed 277 across two terms, G.W. Bush 291, Biden 162, Trump's first term 220, the second term 268 by mid-2026, a high pace, but not categorically outside the historical range. Reagan fired roughly fifteen to sixteen inspectors general on taking office in 1981, prompting comparable bipartisan alarm; the 2025 firings are precedented in kind. What distinguishes 2025 is one aggravating fact: a federal judge found the dismissals skipped the statutory thirty-day notice and case-specific rationale required by law, the nuance this framework's for/against columns exist to carry: precedented in kind, aggravated in legality.

The direct scholarly rebuttal. Kortukov & Waller, "Authoritarianism, Reform, or Capture?" (American Affairs, 2025) is the most substantive academic pushback on the competitive-authoritarianism label. Genuine competitive authoritarianism, they argue, requires a systematically rigged electoral playing field, an opposition that cannot win absent government error, opposition figures harassed or jailed, and a controlled media ecosystem. None of that, they argue, describes the United States: the opposition remains well-financed and viable, media remains plural, and the 2024 election showed no significant malfeasance. They invoke Lincoln's Civil War suspensions and FDR's court-packing attempt as precedents of executive-dominant episodes the republic survived, while conceding real inter-branch contestation, a hostile assault on the federal bureaucracy, and unprecedented techno-corporate influence over policy. Their frame is state reform versus state capture, not regime transformation. Their disconfirming checklist (opposition viability, media plurality, ongoing contestation, courts still ruling against the executive, functioning federalism, a military that stays out of politics) supplies the specific disconfirming observations this framework checks for on every indicator.

Calibration cases

Five historical trajectories inform how this framework weights indicators and reads timelines.

Hungary (Orbán, 2010–). A 2010 constitutional supermajority was followed by judicial reshaping, media-council capture, and an electoral-map rewrite, each individually defensible as lawful reform. What discriminated backsliding from ordinary reform was the combination, sustained together: courts, media, and the electoral map moving in the same period. Observers flagged the 2011 constitutional changes early but underestimated how durable capture would prove once a two-thirds majority removed normal veto points; the U.S. lacks any comparable single-party constitutional supermajority, a structural difference this framework weighs under the "bundle is the signal" principle.

Turkey (Erdoğan, 2007–). An early reformist, pro-EU phase gave way to incremental erosion across election cycles: a judiciary weaponized against candidates, media and security forces used to make opposition wins nearly impossible, with the facade of competitive elections persisting long after the field had tilted. A 2016 post-coup purge of roughly four thousand judges marked a sharp acceleration exploiting a crisis pretext. Lesson: the mere continuation of elections is not reassuring on its own; cumulative tilt, and the exploitation of a crisis for acceleration, are what discriminate.

Venezuela (Chávez, 1999–). A 1999 "Emergency Judicial Commission" purged judges under a modernization and anti-corruption pretext, institutional capture disguised as reform, while Chávez won real, large electoral majorities through the mid-2000s even as the field tilted. Clearest lesson: genuine electoral popularity is not exculpatory, and judicial capture dressed as reform is an early, historically under-weighted signal. This is the direct basis for scoring principle 4.

Poland (PiS, 2015–2023): the reversibility case. Capture of the Constitutional Tribunal, forced judicial retirements, and public-media capture proceeded through means that were, at points, unlawful. But elections stayed competitive throughout, and in 2023 the opposition won and took power: the strongest empirical case that backsliding can be reversed at the ballot box if elections remain real, and the direct basis for scoring principle 6. It also shows that restoration, once won, is slow and legally fraught.

India (Modi, 2014–). V-Dem reclassified India from electoral democracy to electoral autocracy around 2019–21, with erosion concentrated in media ownership and journalist arrests, civil-society funding restrictions (FCRA cancellations of NGOs), and selective legal targeting of a religious minority via sedition, terrorism, and hate-speech statutes, while national elections remained large-scale and the ruling party lost repeatedly at the state level. India is the clearest illustration of Varol's stealth-authoritarianism mechanism: neutral-sounding laws applied selectively. It is also a reminder that index downgrades are judgment calls, not facts: the "electoral autocracy" label itself remains contested among India scholars given the country's genuine electoral competitiveness.

Cross-case synthesis (Carnegie). A comparative study finds erosion varies along three axes: which institution is targeted first, how fast decline moves, how severe and institutionalized it becomes. It places the current U.S. profile as distinctive on all three: intra-executive dominance and delegitimizing courts rather than packing them, circumventing Congress rather than co-opting it with a legislative supermajority, and weaponizing federal funding, all at unusual speed but comparatively lower severity and, crucially, "non-institutionalized" (norms tested, laws not yet rewritten). That leaves recovery pathways Hungary and Turkey had already lost by a comparable stage (Carnegie Endowment, "U.S. Democratic Backsliding in Comparative Perspective").

The expert-bias caveat

A 2025 study in PS: Political Science & Politics, "Expert Bias and Democratic Erosion," tests whether the left-leaning composition of the expert pools feeding instruments like Bright Line Watch and V-Dem inflates backsliding assessments, directly relevant to a site built substantially on those instruments' indicator language, and taken seriously rather than set aside.

What this site does about it: every "raises" condition requires a specific, checkable action, not an expert sentiment score, and every "lowers" condition and disconfirmer is drawn from the counter-literature above, not omitted. Standing readings default to "contested" whenever a credentialed, opposing framework offers a live alternative reading of the same facts (see the initial calibration rationale on the weaponization and civil-service indicators). Zone calls cite named scholarly disagreement rather than resolving it silently. None of this eliminates possible pool bias in the underlying scholarship this framework draws its language from. That limitation is disclosed here, not left for a reader to discover.

Scope: foreign policy

This instrument measures domestic democratic erosion. Foreign-policy aggressiveness (a hawkish posture toward an adversary, an alliance dispute, a territorial claim against another state) is not itself an indicator here and is not scored as authoritarianism. Two bounded doors let foreign affairs in:

1. War powers → the congress indicator. Unauthorized force is scored as executive aggrandizement against a constitutional limit, on the same base-rate test as domestic unilateral action. Presidents of both parties have used force without a declaration of war for eighty years: Korea, Kosovo, Libya, the Syria strikes, Soleimani. So unauthorized force alone is precedented, not automatically pattern-consistent. What discriminates is Congress's response: notification honored, authorization sought, or a War Powers Resolution vote that succeeds are contrary/mixed signals; force continued over a congressional vote moves toward pattern. Votes are logged as events either way.
2. Rhetoric toward allies/territories → the rhetoric layer. Statements disparaging a treaty ally or claiming another state's territory are logged under the information-indicator rhetoric layer (§4.5), on the same recurrence and target-specificity tests as any other official rhetoric, not a standalone indicator.

Watched, not scored. The comparative literature's alliance-drift correlate, Orbán's and Erdoğan's realignment away from Western alliance structures alongside domestic backsliding (scholarship.md, Part V), is background context for the alliance-conduct thread, never a basis for a reading change.

This boundary cuts both ways: without it, the site would either launder ordinary hawkish foreign policy into a score it wasn't built to carry, or go silent on the one place foreign conduct bears directly on the separation of powers it exists to measure.

Versioning & corrections policy

This is version 1.0 of the Indicator Framework, published 2026-07-10. Changes to indicator definitions, the zone scale, or the scoring principles will be versioned (1.1, 2.0, and so on per the scope of the change) and dated on this page; nothing here is edited silently. Changes to a specific indicator's standing reading are logged on that indicator's own page as a dated, append-only changelog entry with a stated rationale. Readings are never quietly moved.

Factual errors in a published entry are corrected in place, with the correction dated and left visible: a strikethrough plus an amendment note, never a silent edit. This is the same policy published in full, with the underlying rationale, on the Sources & Pipeline page; it is stated here as well because a methodology page that describes its own error-correction process is itself part of the instrument, not an appendix to it.