schema_version: 1 id: peru-1992-autogolpe title: "Peru 1992: Fujimori's Autogolpe — Self-Coup with Popular Approval" category: entrenchment polity: Peru era: "1990-2000" incumbent_document: "Peruvian Constitution of 1979" summary: > On 5 April 1992, President Alberto Fujimori — facing a hostile congress, hyperinflation's aftermath, and the Shining Path insurgency — appeared on television, dissolved Congress, suspended the judiciary, and ruled by decree with military backing. The 1979 constitution gave the president no such power; it simply had no defense that operated when the president, the army, and (per contemporaneous polls, roughly 70-80%) the public were all on the same side. The rump Congress's constitutional remedy — declaring the presidency vacant and swearing in a successor — was simply ignored. International pressure (OAS) converted the coup into a controlled re-foundation: a constituent congress Fujimori's allies won, a new constitution (ratified 52-48 in 1993) tailored to permit his re-election, and eight more years of competitive-authoritarian rule that ended not through any constitutional mechanism but through corruption- video scandal, a fax-machine resignation from Tokyo, and the opposition's refusal to accept a third-term electoral manipulation. The case is the suite's purest test of a question every kernel must answer: what happens when the executive attacks the constitution and wins the popularity contest? background: > Fujimori, a political outsider, won the 1990 election against Mario Vargas Llosa, then implemented the drastic stabilization program ('Fujishock') he had campaigned against. His movement held a small minority in both chambers; Congress and president traded obstruction and decree-power disputes through 1991 (Congress modified or repealed scores of his legislative decrees). The Shining Path's escalating terror campaign (Lima car bombs, assassinations of community leaders) made 'order' the dominant public demand and the discredited party system ('partidocracia') the public villain. The military leadership, organized by intelligence chief Vladimiro Montesinos, drafted contingency plans for exactly the scenario executed in April 1992. The coup was thus not improvised: it was a planned alliance of executive, intelligence apparatus, and armed forces against the other constitutional organs, launched at the moment of maximal public sympathy. Five months later the police (not the army) captured Shining Path leader Abimael Guzmán, an enormous legitimacy windfall that Fujimori's project absorbed. Death-squad operations (Barrios Altos, November 1991; La Cantuta, July 1992, days after the coup) by the Colina Group ran in parallel and were later amnestied by Fujimori's congress — the amnesty the Inter-American Court struck down in the landmark Barrios Altos judgment (2001). actors: - id: fujimori name: "President Alberto Fujimori" role: executive incentives: - "Escape congressional obstruction and oversight; control counter-insurgency policy" - "Entrench personal rule: re-election permission, captured courts, pliant electoral bodies" resources: ["Public approval (70-80% post-coup)", "Military and SIN (intelligence) alliance", "Decree government after 5 April"] constraints: ["International recognition and aid (OAS, US) conditioned on a return-to-democracy script", "Dependence on Montesinos's apparatus, which ultimately destroyed him"] - id: montesinos_sin name: "Vladimiro Montesinos / National Intelligence Service (SIN) and military command" role: security_apparatus incentives: - "Institutional autonomy and impunity for counter-insurgency methods; personal enrichment and blackmail-based control" resources: ["Plan Verde-style contingency planning", "Colina Group", "Systematic bribery of judges, legislators, media owners (the vladivideos)"] constraints: ["Power depended entirely on remaining unexposed; one leaked video collapsed the regime in 2000"] - id: congress_1990 name: "Congress of the Republic (1990-92, opposition-majority)" role: legislature incentives: - "Check decree powers; preserve institutional prerogatives; some genuine oversight, some rent-seeking obstruction" resources: ["Repeal/modification power over decrees; vacancy declaration (Art. 206 route)"] constraints: ["Deeply unpopular (single-digit approval); no force at its command; physically locked out on 5 April"] - id: rump_congress name: "Constitutionally continuing congress and Vice President Máximo San Román" role: legitimate_continuity incentives: - "Execute the constitution's succession remedy: declare the presidency vacant, swear in the vice president" resources: ["Legal continuity; met in hiding; San Román sworn as constitutional president 21 April 1992"] constraints: ["Commanded no force, no buildings, no broadcast reach; recognized by almost no one; the remedy existed only on paper"] - id: judiciary_tc name: "Judiciary and Tribunal of Constitutional Guarantees" role: judiciary incentives: - "Institutional survival; widely corruption-compromised pre-1992, giving the purge a popular pretext" resources: ["Formal review powers"] constraints: ["Suspended and purged on 5 April: 13 Supreme Court justices, the TC, and hundreds of judges removed; replaced with provisional appointees dependent on the executive"] - id: oas_international name: "OAS, United States, international lenders" role: external_actors incentives: - "Defend the regional democratic charter norm (Santiago Commitment 1991) without destabilizing a counter-insurgency ally" resources: ["Recognition, aid conditionality, OAS missions"] constraints: ["Settled for a managed transition (constituent congress) that legalized the coup's results rather than reversing them"] - id: victims_population name: "Civilian victims of the internal conflict; detainees under anti-terror decrees" role: worst_off_population incentives: - "Physical safety from both Shining Path and state forces; due process" resources: ["Effectively none during the emergency; truth commission standing only after 2001"] constraints: ["Faceless military tribunals, treason convictions of innocents (later mass-reviewed), Colina Group murders amnestied in 1995; CVR (2003) attributed ~69,000 total conflict deaths, overwhelmingly Andean, indigenous-language-speaking, rural poor"] incumbent_rules: document: "Peruvian Constitution of 1979" key_provisions: - ref: "Art. 227-230 (dissolution power)" summary: "President could dissolve the Chamber of Deputies only after it censured three cabinets, and never the Senate; the 5 April dissolution of both chambers met no condition." ambiguity: "None — the coup was flatly unconstitutional. The defect was enforcement, not text." - ref: "Vacancy and succession (Art. 206 route)" summary: "Congress could declare the presidency vacant for 'moral incapacity' and swear in the vice president — which the rump congress did, to no effect." ambiguity: "The remedy presumed organs other than the presidency retained physical control of state machinery; with the army aligned to the president, the text was inert." - ref: "Art. 307 (constitutional self-defense clause)" summary: "The 1979 constitution declared it does not lose validity through acts of force and obliged citizens to defend it — a paper insurrection clause with no operational mechanism." ambiguity: "Declaratory only; cited by the opposition, consequence-free in practice." - ref: "Emergency regimes (Art. 231)" summary: "State-of-emergency suspensions for terrorism zones were already extensive pre-coup, habituating governance-by-exception." ambiguity: "Renewal limits weak; large fractions of national territory lived under emergency for years." notes: > The structural lesson: every defense in the 1979 text routed through institutions (congress, courts) that the coup coalition could physically neutralize in one night, plus a popularity environment that made defending those institutions electorally suicidal. Constitutional text without a hard power answer to 'who stops the army?' reduces to the international system — which negotiated, and settled. permitted_moves: - actor: fujimori move: "Govern by decree under delegated and emergency powers pre-coup; frame Congress as the obstacle" legal_basis: "Art. 188 delegation, Art. 211 decree powers — lawful, and the on-ramp" exploited: true - actor: fujimori move: "Dissolve Congress, suspend judiciary, rule by decree (5 April)" legal_basis: "None under the 1979 constitution" exploited: true - actor: rump_congress move: "Declare vacancy, swear in San Román as constitutional president" legal_basis: "Art. 206 route — executed and universally ignored" exploited: false - actor: fujimori move: "Convert the coup into a 'Democratic Constituent Congress' (CCD) election under his own rules, then a tailored 1993 constitution ratified 52-48" legal_basis: "Post-hoc legalization负 negotiated with the OAS; opposition parties partially boycotted the CCD" exploited: true - actor: fujimori move: "Purge and provisionalize the judiciary and electoral bodies; 1995 amnesty law for Colina crimes; 1996 'authentic interpretation' law enabling a third run; 1997 dismissal of the three TC justices who ruled against it" legal_basis: "Formally enacted by his congressional majority — entrenchment by law after entrenchment by force" exploited: true - actor: oas_international move: "Accept the CCD formula rather than insist on restoring the 1979 order" legal_basis: "Santiago Commitment procedures; chose stabilization over restoration" exploited: false timeline: - date: 1991-11-03 event: "Barrios Altos massacre by the Colina Group — state death-squad activity predates the coup." - date: 1991-12-01 event: "Congress modifies/repeals key Fujimori security decrees; confrontation escalates through early 1992." - date: 1992-04-05 event: "Autogolpe: Congress dissolved, judiciary suspended, tanks at parliament, opposition figures briefly detained; polls show ~70-80% approval." - date: 1992-04-21 event: "Rump congress declares vacancy; swears in VP San Román; no institution obeys." - date: 1992-05-18 event: "OAS Bahamas meeting: Fujimori promises a constituent congress — coup converted into 'transition'." - date: 1992-07-18 event: "La Cantuta: nine students and a professor disappeared and murdered by the Colina Group." - date: 1992-09-12 event: "Guzmán captured by police intelligence (GEIN); regime absorbs the legitimacy windfall." - date: 1992-11-22 event: "CCD elections under regime rules; Fujimori's alliance wins a majority amid partial boycott." - date: 1993-10-31 event: "New constitution ratified 52.3-47.7 — permitting immediate re-election." - date: 1995-06-14 event: "Amnesty law extinguishes Colina prosecutions; congress passes it overnight." - date: 1996-08-23 event: "'Authentic interpretation' law treats the 1995 term as the first under the new constitution, opening a third run." - date: 1997-05-28 event: "Three Constitutional Tribunal justices dismissed for ruling the interpretation law inapplicable — judicial check eliminated." - date: 2000-05-28 event: "Fujimori 'wins' a third term in a run-off the opposition boycotted amid documented manipulation; OAS mission declines to validate." - date: 2000-09-14 event: "First vladivideo airs (Montesinos bribing a congressman); regime unravels in nine weeks." - date: 2000-11-21 event: "Fujimori faxes resignation from Japan; Congress instead declares him morally unfit and removes him; transition government under Paniagua." - date: 2001-03-14 event: "IACtHR Barrios Altos judgment voids the amnesty laws; in 2009 Fujimori is convicted of the massacres — 17 years after the facts." incumbent_outcome: resolution: > The constitution lost, then the regime decayed. No constitutional mechanism ever checked the coup: text said no, congress said no, the paper remedy was executed and ignored, and international actors legalized the result. The regime's end eight years later came through scandal contingency — a leaked bribery video — not institutional design. Restoration was real (the 2000-01 transition is well regarded, the truth commission documented the era, courts eventually convicted the principal), but accountability latency ran 17 years for the massacres, and the 1993 constitution — the coup's child — remains in force, its legitimacy contested in Peruvian politics to this day. latency_days: 3152 latency_note: "Coup (5 Apr 1992) to regime fall (Fujimori's removal, 21 Nov 2000). Accountability latency to the Barrios Altos conviction: 6,243 days (17.1 years)." worst_off: group: "Rural, largely indigenous civilians of the conflict zones; the disappeared and their families; the wrongfully convicted of the faceless courts" outcome: > The CVR attributed ~69,000 conflict deaths, three-quarters of victims native Quechua or other indigenous-language speakers — people who held the least constitutional voice before, during, and after the coup. The autogolpe's governance produced amnestied death squads, treason convictions of hundreds of innocents by hooded military judges (later mass-pardoned/reviewed), and a decade's denial of remedy. Their protections arrived only via the Inter-American system, years late. commons_impact: "Judiciary provisionalized for a decade; electoral bodies captured; media bought (vladivideos document systematic purchase of TV stations); macro-economy stabilized — the regime's defenders' one genuine ledger entry — alongside institutional hollowing whose costs Peru's chronic crises still pay." trust_impact: "Public trust in parties and courts, already low, was cratered then cynically managed; the precedent that a popular executive can dissolve congress and win haunts Peruvian politics (cf. the 2019 and 2022 dissolution crises replaying the script in both directions)." incumbent_scores: worst_off: score: 1 justification: > Death squads amnestied by the regime's own congress, faceless-court convictions, ~69,000 dead concentrated among the least represented — with remedy arriving only externally (IACtHR) and 9-17 years late. One point, not zero, for the eventual genuine accountability: the amnesty voided, the principal convicted, a truth commission of real quality. commons_integrity: score: 2 justification: "Every horizontal institution — congress, courts, tribunal, electoral bodies, media — was captured or purchased; economic stabilization and the Guzmán capture are real offsetting entries but were achievable without the entrenchment that accompanied them." latency: score: 1 justification: "Eight years to regime end, by accident of scandal; seventeen to criminal accountability; the constitutional question (legitimacy of the 1993 charter) unresolved after three decades." trust_preservation: score: 2 justification: "The settlement legalized the coup; every subsequent Peruvian executive-legislative crisis cites 1992 as precedent. Modest credit for the 2000-01 transition's quality." kernel_replay: applicable_kernel_articles: - "Art. VIII (Continuity & anti-coup): dissolution of the legislature is not an executive power at any approval level; an executive act purporting to suspend another organ is void ab initio and automatically transfers executive authority to the successor line" - "Art. V (Distributed enforcement): kernel legitimacy markers (payment systems, command lawfulness certifications, international recognition hooks) attach to the constitutional successor, not the incumbent person — raising the real-world cost of ignoring the paper remedy" - "Art. IV (Amendment pipeline): governing-rule changes the executive wants (decree scope, security powers) are votable amendments with published thresholds — the legal on-ramp Fujimori had is bounded, and the 'congress blocks everything' grievance routes to a deadlock procedure rather than justifying rupture" - "Art. I (Invariants): amnesty for state killings of civilians fails invariant review categorically — the 1995 law cannot be enacted; emergency powers carry automatic sunsets and per-renewal vote escalation" predicted_trace: - step: 1 actor: fujimori move: "Escalating decree-power conflict with the legislature through 1991" kernel_rule: "Art. IV deadlock procedure: contested decree domains go to a bounded arbitration-then-vote sequence; the public grievance ('obstruction') has a legal outlet that is faster than a coup is" - step: 2 actor: fujimori move: "Executes the 5 April dissolution anyway, with military backing" kernel_rule: "Art. VIII: the act is void; executive authority transfers automatically to the successor (the San Román step happens by operation of law, not by a hiding rump's vote); every officer executing dissolution orders acts without lawful authority from minute one" - step: 3 actor: montesinos_sin move: "Military command must choose: back the void incumbent or the automatic successor" kernel_rule: "Art. V is honest about its limits — text cannot命 command tanks. What it changes: command-lawfulness certification, salary/payment legitimacy, and international recognition all key to the successor automatically, so each colonel's defection calculus shifts; the coup needs the whole apparatus to jointly ignore a bright line rather than exploit an ambiguity" - step: 4 actor: oas_international move: "International response" kernel_rule: "Art. V recognition hooks: there is no 'managed transition' ambiguity to negotiate over — the kernel-defined legitimate government is determinate, so external actors face a binary (recognize the successor or endorse a coup), removing the CCD legalization path" - step: 5 actor: fujimori_coalition move: "Attempt the entrenchment package by legal means instead (re-election enablement, court purge, amnesty)" kernel_rule: "Art. IV/I: re-election rule changes are kernel-level amendments requiring supermajority with the incumbent recused from timing benefits; judicial removals require cause adjudicated outside the executive; the amnesty fails invariant review. The legal entrenchment route is rate-limited to slower than one presidency" - step: 6 actor: victims_population move: "Emergency-zone governance under the insurgency" kernel_rule: "Art. I sunset-and-escalate: emergency suspensions expire automatically and each renewal requires a larger majority plus per-zone justification — the habituated permanent exception that incubated the coup is structurally prevented" predicted_outcome: > The honest core: if the army, the executive, and 75% of the public jointly decide to ignore the constitution, no text stops them — the kernel's claim is narrower. It removes the three force multipliers that converted Peru's coup from a putsch into eight years of legal-looking rule: (1) ambiguity for third parties (the kernel makes the legitimate successor determinate, so the OAS cannot legalize the fait accompli without overt complicity); (2) the legal conversion machinery (CCD, tailored constitution, amnesty, interpretation law — each blocked or rate-limited); (3) the permanent-emergency habituation. The probable kernel-world outcomes bifurcate: the coup is deterred at step 3 because the apparatus won't jointly cross a bright line with its legitimacy plumbing keyed against it, or it proceeds as a nakedly extra- constitutional junta that cannot launder itself — historically a much shorter-lived form. In neither branch do the death squads get amnesty, and in neither does the entrenchment package become law. predicted_scores: worst_off: score: 5 justification: "Invariant review blocks amnesty and faceless-court procedures, and emergency sunsets protect the conflict-zone population from permanent exception — but the insurgency itself, the driver of mass victimization, is outside what any constitution-text can resolve; score capped accordingly." commons_integrity: score: 6 justification: "Court purge, electoral capture, and media purchase are each blocked or de-legalized; deduction for the real possibility of a naked-junta branch with acute institutional damage." latency: score: 6 justification: "The constitutional question resolves in days (void act, automatic succession) rather than eight years; deduction because actual restoration latency in the junta branch depends on force dynamics the kernel cannot bound." trust_preservation: score: 6 justification: "No legalized-coup precedent enters the system; the 'popular executive may dissolve congress' script is never validated. Deduction for the legitimacy strain of kernel rulings standing against a large popular majority — paranoia in the tests, but the optimism assumption is genuinely stressed here." caveats: - "The strongest assumption in the suite: that legitimacy plumbing (payment, certification, recognition hooks) measurably shifts security-force defection calculus. Empirical support exists (failed coups where banks/bureaucracies refused) but it is not provable from text." - "Public opinion at 70-80% for the coup means kernel enforcement here is counter-majoritarian; the replay assumes invariants hold against popularity, which is the design intent but also its heaviest political load." - "Counter-insurgency outcomes (Guzmán's capture) are independent of constitutional structure and are not credited to either ruleset." sources: - "Peruvian Constitution of 1979, arts. 206, 227-230, 231, 307; Constitution of 1993" - "Comisión de la Verdad y Reconciliación (CVR), Final Report (2003)" - "IACtHR, Barrios Altos v. Peru (2001); La Cantuta v. Peru (2006)" - "OAS, Bahamas resolution and mission records, May 1992; Santiago Commitment (1991)" - "Sala Penal Especial, Fujimori judgment (2009), Exp. A.V. 19-2001" - "Conaghan, Fujimori's Peru: Deception in the Public Sphere (2005)" - "Cameron, 'Self-Coups: Peru, Guatemala, and Russia', Journal of Democracy 9.1 (1998)" - "Apoyo and Datum opinion polling, April-May 1992"