Orders in Washington.
Choices in Chile.

Chapter 01 / 1962–September 1970

The intervention preceded the crisis

Chile’s September 1970 election was the latest contest in a long domestic argument over property, inequality and political authority. Eduardo Frei’s Christian Democrats had offered reform within a non-Marxist framework; Salvador Allende’s Popular Unity coalition proposed a parliamentary road to socialism. Conservatives defended property and resisted the left, while the coalition itself included parties with different views about revolutionary change. These were Chilean political projects with organized constituencies. Foreign money entered this contest; it did not create every conviction or divide from outside.

American covert political work had begun well before Allende’s presidency. The CIA’s later report to Congress identifies party assistance in 1962 and an expanded election program in 1964, when Frei benefited from secret support. The 1975 Senate investigation describes both campaign financing and propaganda. Organizational contacts, media channels and assets persisted into later elections. This infrastructure matters because the abrupt presidential directive of September 1970 did not reach an agency starting from nothing. It also matters that previous operations had different objectives: helping an electoral candidate and provoking a military coup were distinct policy choices.

For the 1970 presidential election, the 40 Committee authorized a campaign against Allende’s coalition rather than direct CIA support for either opposing candidate. A campaign to damage one candidate still intervened in the electoral contest. Its narrower design also distinguishes it from the large 1964 operation and from private American corporate contributions. CIA contact with business, CIA funding, and business money were overlapping but nonidentical channels. A claim about one cannot automatically establish the amount or recipient of another.

Allende won a plurality on4 September, about36 percent of the vote, rather than an outright majority. Under Chile’s constitutional process Congress would choose between the leading candidates. Calling him democratically elected is consistent with that process; saying he had already won a popular majority is not. The interval before the congressional vote gave his opponents a possible opening. By 14 September, presidential adviser Viron Vaky was assessing a plan to elect Jorge Alessandri, have him resign, and enable Frei to seek a new mandate. Vaky questioned both the risks and the necessity. The later turn to a secret military operation therefore followed a debated choice inside Washington, not a demonstrated absence of alternatives.

Sources: S01 · S02 · S03 · S04

Chapter 02 / September–October 1970

Two tracks, one presidential demand

On 15 September 1970 Richard Nixon met Henry Kissinger, CIA director Richard Helms and Attorney General John Mitchell. Helms’s handwritten notes record an urgent instruction to prevent the unacceptable outcome in Chile, with no embassy involvement and up to$10 million available if needed. William Broe’s memorandum the following day translates the directive into an organizational assignment: prevent Allende from taking power or unseat him; keep the undertaking outside normal State and Defense coordination; give Thomas Karamessines overall responsibility with a special task force. These are contemporary records of authorization, not merely later recollections.

The terms TrackI and TrackII distinguish routes, though the political pressures and contacts were not hermetically separate. TrackI worked through the embassy and the 40 Committee on political maneuvers intended to block confirmation. TrackII sought a military intervention and was deliberately concealed from Ambassador Korry and much of the normal policy machinery. This division explains why a diplomatic cable can oppose a coup while a CIA directive presses for one. The contradiction is evidence of compartmentalization and disagreement, rather than a reason to erase one set of documents.

Korry’s9 October warning was specific. He judged that a successful intervention required the complicity of Frei or Army commander René Schneider, doubted support for retired General Roberto Viaux and feared a disastrous failure. His assessment was neither political neutrality nor endorsement of Allende; he was already engaged in pressure against the incoming government. It does show that officials could recognize the difference between their desired result and available Chilean capabilities.

On 16 October CIA headquarters reaffirmed continued pursuit of a coup, preferably before Congress voted but continuing afterward if necessary. The same cable discouraged Viaux from acting alone because his available forces were judged insufficient. It sought broader coordination among conspirators and maintained interest in other officers, including Camilo Valenzuela. Reading only the discouragement clause gives a false impression of a general cancellation. Reading only the coup instruction misses a real operational distinction between groups and plans. Washington ordered pressure; Chilean officers still had to decide whether to move, whom to follow and how to overcome commanders committed to the constitutional transfer. That gap between authorization and capability became the central failure of TrackII.

Sources: S04 · S05 · S06 · S07 · S08 · S30

Chapter 03 / October 1970

A kidnapping plan ended in death

René Schneider, the Army commander-in-chief, stood in the way of officers who wanted to prevent Allende’s accession. His commitment to the constitutional process made his removal a practical requirement in several coup schemes. The CIA’s 2000 review acknowledges that American officials knew of, and accepted, the plotters’ view that his abduction was essential. Kidnapping the commander was therefore not a harmless conversation at the margin of electoral politics. It was a coercive operation intended to disable an institutional barrier to a coup.

The surviving account differentiates groups. Viaux’s associates sought supplies and support; headquarters judged his proposed action unlikely to succeed and warned against premature movement. The 16 October cable nonetheless encouraged wider planning. Separately, the CIA supplied Valenzuela’s group with three submachine guns, ammunition and tear-gas equipment for an abduction scheme. According to the later agency review, those firearms were returned unused. These distinctions bear directly on the claim that CIA-supplied guns killed Schneider. The assembled record does not establish that specific claim, even though it firmly establishes CIA assistance to a kidnapping project.

On 22 October Schneider was shot during an attempted abduction by men associated with Viaux. He died on25 October. The CIA retrospectively said that the fatal attempt proceeded independently of its direction at that moment and that it found no evidence of an intention to kill him. Those assertions address command and intended outcome; they do not cancel the earlier encouragement of a coup or acceptance of abduction. The risks of violently seizing an armed commander were real even if the intended political mechanism depended on keeping him alive. Intent, assistance, foreseeable danger and the identity of the gunmen are separate questions.

After the shooting, CIA briefing notes still imagined that a state of emergency and Valenzuela’s control in Santiago might provide the sequence needed for a coup. The forecast failed. The 28 October review acknowledged that Schneider’s resistance and death had inhibited further action. Congress confirmed Allende on24 October; he took office on3 November. In November the CIA paid$35,000 to a surviving Viaux-group representative, a payment its 2000 report linked partly to keeping prior contacts secret. That after-the-fact payment is consequential evidence of a continuing relationship and concealment concern. Its timing does not make it an advance murder fee. The bounded conclusion is serious enough: a documented American coup effort assisted abduction planning, and the wider plot produced a fatal attack without achieving its intended transfer of power.

Sources: S02 · S07 · S09 · S10 · S24

Chapter 04 / November 1970–1972

The failure changed the method

Allende’s inauguration ended the immediate attempt to stop his accession, but not Washington’s campaign against his government. National Security Decision Memorandum93, issued9 November 1970, paired a restrained public posture with efforts to prevent his consolidation and restrict policies contrary to American interests. The distinction between public diplomacy and private pressure was explicit. A government could maintain relations while pursuing an adversarial program through other channels.

A CIA options paper prepared before the inauguration illustrates the available repertoire: exploit divisions within Popular Unity, support opposing political groups and media, maintain influence with security forces, and create uncertainty. An options paper is evidence of proposed means, not proof that every item was approved or carried out. Subsequent authorizations and reporting establish substantial actual assistance to opposition parties and media. El Mercurio received secret support. Such support affected the resources available for opposition activity without making every newspaper article, party decision or public protest a CIA instruction.

Economic pressure also requires a mechanism more precise than a slogan. Nixon’s directive was followed by policies restricting assistance and seeking leverage through finance and debt. In January 1972 Treasury Secretary John Connally urged creditor coordination to isolate Chile and maintain pressure. This was a deliberate effort to worsen the government’s bargaining position. It coexisted with domestic decisions about nationalization, compensation, wages, spending, investment and monetary expansion; fluctuations in copper earnings; and supply and production problems. A CIA assessment in August 1971 anticipated substantial growth that year while warning of shrinking reserves and emerging constraints. Even hostile official analysis did not describe a single switch that Washington could flip to create or remove every economic difficulty.

The covert campaign’s precise scale depends on time period and accounting rules. The Senate investigation and the 2000 agency report present different aggregate totals; neither is a complete public recipient-level ledger. The reliable conclusion is a sustained multimillion-dollar political operation. Counting an authorization as a payment, or combining private corporate money with all CIA project spending, would create false precision. The same care applies to strikes: the Senate found no approved direct strike-funding program, while noting support for private-sector organizations connected with strikers and at least one unauthorized transfer onward. A formal prohibition therefore does not establish that no assistance reached a strike, and evidence of one transfer does not establish that every strike was financed abroad.

Sources: S01 · S02 · S11 · S12 · S13 · S14 · S27

Chapter 05 / January–August 1973

By 1973 the paths were converging

During 1973 political conflict and economic crisis narrowed the room for accommodation. Popular Unity received 43.4 percent in the March congressional vote and retained a substantial electoral constituency; the March congressional election did not produce an opposition victory sufficient to resolve the contest by simply removing the president through Congress. Government supporters, opposition parties, business organizations, workers and the armed forces faced different incentives. In March Ambassador Nathaniel Davis reported opposition anxiety that a U.S.–Chilean settlement might ease Allende’s economic problems. These Chilean opponents wanted pressure maintained for their own political purposes, rather than waiting passively for instructions.

The crisis also involved rival claims to constitutional authority. Governing parties differed over whether to consolidate negotiated reforms or accelerate revolutionary change; opponents moved from resisting individual policies toward denying the government’s legitimacy. The failure of talks between the Christian Democrats and Popular Unity closed one route to compromise. On 22 August a majority in the Chamber of Deputies accused the government of grave breaches of the constitutional and legal order. That resolution records an important opposition claim and a step in the breakdown of civilian agreement; it should not be silently converted into a judicial finding or an instruction making military overthrow lawful. The next day Prats resigned and Allende appointed Pinochet. These were dated political and command changes that a coup-centered account has to explain, not background noise around a foreign operation.

The 29 June military rising, later known as the Tanquetazo, failed. Its failure did not mean the armed forces were permanently constitutionalist, just as the failure of TrackII had not settled the future. It exposed a continuing military option. General Carlos Prats remained an important defender of the constitutional chain of command. His departure on 23 August and replacement by Augusto Pinochet altered the institutional setting. Explaining the eventual coup requires these shifts in authority and willingness inside Chile, alongside external pressure.

CIA and embassy records show disagreement at the edge between supporting opposition and encouraging military intervention. A26 June station message argued that private-sector organizations could apply strong pressure against Allende and deserved assistance despite concerns about their relationship to military action. An August report recorded Davis’s opposition to encouraging the armed forces, alongside private-sector contacts’ pleas for help. Its author concluded that American personnel in Santiago did not possess a clear solution and could not confidently predict whether the Army would move. This is difficult to reconcile with an account of a fully controlled operation proceeding from a settled American timetable.

The 25 August Colby memorandum is particularly revealing. A new funding authorization had been made conditional on the ambassador’s approval for private-sector and contingency spending. The station sought to increase opposition pressure, while Davis declined endorsement because he feared a de facto American commitment to a coup. Colby distinguished the proposed civilian aid from direct work with the armed forces while acknowledging that increased pressure could produce a coup. Those categories were operationally and politically porous. The document supports a serious causal argument about destabilization and encouragement; it does not, by itself, establish that the proposed payment was made or that CIA officers issued the final military orders. The resulting picture contains both intervention and limits: American officials had created and sustained pressure channels, yet disagreed about how far to push them as Chilean actors approached a military decision.

Sources: S02 · S15 · S16 · S17 · S22 · S23 · S31 · S32

Chapter 06 / September 1973–2026

The coup was followed by a struggle over the record

On 11 September 1973 Chile’s armed forces overthrew Allende. The assault on La Moneda, his final radio addresses and his death belong to a documented military seizure of power. Chile’s National Archive describes the bombardment and his suicide inside the palace. The fact that Allende took his own life does not make the coup nonviolent, legitimate or unconnected with his death; it distinguishes the physical manner of death from a separate allegation that CIA personnel assassinated him. The dictatorship that followed dismantled political freedoms and committed systematic abuses. Neither the government’s policy failures nor fear of Marxism removes responsibility for those acts.

The contemporary American response was favorable and deliberately careful in public. At the 12 September Washington Special Actions Group meeting, officials discussed how to signal goodwill to the junta while avoiding a conspicuous embrace and accusations of American responsibility. A situation report to Nixon that day still contained uncertainty about fighting and Allende’s fate. These records document knowledge, priorities and post-coup support. They do not supply a missing pre-coup operational command chain.

On 16 September Nixon and Kissinger discussed helping create favorable conditions while denying that they had themselves done the coup. That distinction cannot simply be accepted because participants made it: they had incentives to protect themselves and a government they welcomed. But it is consistent with the difference between the extensive documented pressure campaign and the incomplete evidence for direct American control of the September operation. Another memorandum that day relayed a much broader claim that no overthrow program had been supported. The 1970 orders contradict such a sweeping formulation. Official reassurance must be tested against other official records.

The 1975 Senate inquiry made covert political activity part of the public record, while reporting that it had found no evidence of direct U.S. involvement in the 1973 coup. The 2000 intelligence review admitted the 1970 kidnapping-related activity, military contacts, political subsidies and relationships with later human-rights abusers, while denying operational instigation of the 1973 coup. Those findings are bounded institutional conclusions. Later document releases, including presidential briefings in 2023, show how much remained concealed and how uncertain contemporary forecasts could be. Scholarship likewise differs over American causal weight and Chilean political responsibility. The strongest supported synthesis is that Washington worked to prevent and weaken an Allende government and helped sustain the environment in which his enemies acted. Proving that it commanded the decisive coup would require a more specific authenticated chain of orders, assistance and execution than the assembled record provides.

Sources: S01 · S02 · S18 · S19 · S20 · S21 · S25 · S26 · S27 · S28 · S29