The Assignment
Emadeddin Baghi was not a dissident when he started counting. He was a researcher — a journalist and scholar who had worked within the system, writing for reformist publications and engaging with the Islamic Republic’s own institutions. When the Bonyad Shahid — the Foundation of Martyrs and Veterans Affairs — needed someone to conduct a forensic census of the revolution’s dead, Baghi was the kind of insider they trusted with the task.
The Bonyad’s purpose was administrative. Families of verified revolution martyrs were entitled to substantial benefits: monthly pensions, priority housing, university admissions for their children, access to subsidized goods. The program was expensive. The government needed to know exactly how many families qualified.
To determine this, Baghi needed to answer a deceptively simple question: How many people did the Shah’s regime actually kill?
The Methodology That Makes This Airtight
The Bonyad’s registration system had a structural feature that researchers outside Iran could only dream of: a built-in incentive to overcount.
Every Iranian family that could document a death attributable to the Shah’s security forces — whether from SAVAK operations, military actions during protests, or deaths in custody — was eligible for compensation. The incentive was not trivial. Martyr families occupied a privileged tier in post-revolutionary society. There was every reason to register a death and no financial reason to withhold one.
This means Baghi’s data represents a ceiling on the actual death toll, not a floor. If the true number had been higher, more families would have come forward to claim benefits. The Bonyad, backed by the full administrative apparatus of the state, actively sought registrations. Mosques, revolutionary committees, and local officials helped families document claims.
Despite all of this — the financial incentive, the institutional support, the political motivation to validate the revolution’s founding narrative — the Bonyad could verify only 3,164 deaths across the entire sixteen-year period from 1963 to 1979.1
The Statistical Impossibility
The claim of 60,000 deaths doesn’t just conflict with Baghi’s data. It conflicts with basic organizational arithmetic.
SAVAK employed approximately 5,000 full-time agents for a population of 35 million.2 Compare this to the East German Stasi, which deployed 91,000 officers and maintained a network of 174,000 informers for a population of 17 million — one informer for every 66 citizens.3 The Stasi achieved something approaching total surveillance through sheer saturation. SAVAK, with roughly one agent per 7,000 citizens, could not.
SAVAK compensated for its limited manpower through what the historical record consistently describes as exemplary terror — torturing and publicizing the fates of specific targets to create an atmosphere of fear that outstripped its actual capacity for surveillance. The “myth” of SAVAK’s omnipresence was its primary force multiplier. The reality was a mid-sized security apparatus that relied on reputation rather than saturation.
An organization of 5,000 agents killing 60,000 people over sixteen years would require each agent to have personally killed twelve people — a rate of assassination more consistent with a war than an intelligence service. The logistical infrastructure required for mass killing on that scale — body disposal, record-keeping, operational security — would far exceed what 5,000 employees could sustain.
The number was always impossible. It just took Baghi to prove it.
The Breakdown
Baghi’s forensic audit revealed three distinct periods, each with its own character.
The 1963 Uprising: Fewer Than 100
The June 1963 protests following Khomeini’s arrest are routinely described in revolutionary literature as a “massacre of thousands.” Baghi documented 32 deaths in Tehran and fewer than 100 total. The Bonyad, which had every reason to find more, could not.
The Guerrilla War (1971-1977): 341-368
After the Siahkal incident of 1971 — when Fedayeen-e Khalq guerrillas attacked a gendarmerie post in northern Iran — SAVAK waged a decade-long counter-insurgency against two armed movements: the Marxist Fedayeen and the Islamo-Marxist MEK.
| Category | Deaths |
|---|---|
| Killed in gun battles | 177 |
| Executed (some after military tribunals) | 91 |
| Died under torture | 42 |
| Disappeared | 15 |
These were overwhelmingly armed fighters engaged in an asymmetric war. The Fedayeen conducted bank robberies, assassinations, and bombings. The MEK assassinated three American military officers: Lt. Col. Lewis Hawkins, Col. Paul Shaffer, and Lt. Col. Jack Turner. By 1977, SAVAK had effectively destroyed both organizations’ operational capacity.
The number who died under torture — 42 over six years — is damning evidence of systematic abuse. But it is not the number of a genocide. It is the number of a dirty counter-insurgency war.
The Revolution (1978-1979): 2,781
The vast majority of all casualties — nearly 88 percent — occurred in the revolution’s final year. The death toll escalated sharply from September 1978 onward, as street protests grew and the military’s willingness to fire on crowds increased.
The Tehran Coroner’s Office recorded approximately 895 revolution-era death certificates. Behesht-e Zahra Cemetery logged roughly 768 burials. These independent records bracket the same order of magnitude as Baghi’s total.
What Happened When the Data Emerged
Baghi published his findings in the 1990s. The reaction was immediate and telling.
His research was not challenged on methodological grounds. No one presented counter-evidence showing that the Bonyad had missed thousands of eligible families. No rival census emerged with a higher count. The data was not disputed — it was suppressed.
The regime faced what might be called the Baghi Paradox: a system that had commissioned an audit of its enemy’s crimes discovered that the audit threatened its own legitimacy more than any external critic could. The number 3,164, verified by the regime’s own institution using the regime’s own records, posed an existential problem.
If the Shah killed 3,164 people over sixteen years, and the Islamic Republic killed 8,000-9,500 in its first four years alone — and 3,800-5,000+ more in the prison massacres of a single summer in 1988 — then the comparative arithmetic was devastating. The regime that justified its existence by the violence of its predecessor had, by its own researcher’s count, surpassed that violence within its first decade.
Baghi was not punished for getting the number wrong. He was punished for getting it right.
The Verification
Baghi’s findings do not exist in isolation. Multiple independent sources converge on the same order of magnitude.
| Source | Method | Finding |
|---|---|---|
| Baghi / Bonyad Shahid | Forensic audit of martyr registrations | 3,164 total (1963-1979) |
| Tehran Coroner’s Office | Death certificates | ~895 (revolution period) |
| Behesht-e Zahra Cemetery | Burial records | ~768 (revolution period) |
| Ervand Abrahamian | Academic synthesis from SAVAK files and oral histories | ~368 guerrilla deaths (1971-1977) |
| US State Department | Diplomatic cables (1976) | 3,400-3,500 political prisoners |
| ICRC / Red Cross | Physical prison inspections (1977) | ~3,200 political prisoners |
| Parviz Sabeti (SAVAK Third Directorate chief) | Post-1979 testimony | 3,200-3,700 political prisoners |
When the Shah’s own security chief, the US State Department, the ICRC after physical inspections, an independent academic historian, and the Islamic Republic’s own Martyrs Foundation all converge on the same numbers — the data is telling you something.
The outlier is Amnesty International’s 1975 report estimating “25,000 to 100,000” political prisoners.4 Amnesty acknowledged the difficulty of obtaining information in a closed society, and their estimates relied heavily on exile opposition testimony — the same sources that produced the 60,000 figure. When the ICRC was granted physical access to prisons in 1977, they found approximately 3,200 inmates.5 The “100,000” appears to have been either a cumulative total of all arrests and detentions over many years, or an uncritical repetition of opposition claims.
A prison population of 3,200 places the Shah’s Iran in the category of a repressive authoritarian state — comparable to Chile under Pinochet (3,216 killed or disappeared over 17 years). It is an order of magnitude below the totalitarian saturation of the Eastern Bloc. And it is a fraction of the Islamic Republic’s subsequent toll.
The Paradox
The Baghi Paradox is this: the most reliable data on SAVAK’s death toll was produced by the institution with the strongest motive to inflate it, and it demolished the inflation.
Every feature of Baghi’s methodology — the financial incentive structure, the institutional support, the political context — biased the audit toward overcounting. And still, the count refused to exceed 3,164.
This is the nature of forensic evidence. It does not care about founding myths or constitutional preambles or the narratives that revolutions require. It counts what it finds and stops counting when there is nothing left to find.
The Islamic Republic’s constitution still carries the number 60,000. The Bonyad’s archive still carries the number 3,164. The paradox between them is not a historical curiosity. It is the gap between a state’s mythology and its own records — a gap nineteen times wide, and still growing.
This article is a companion to The Number That Changed Everything. For an hour-by-hour reconstruction of how 88 deaths on Black Friday became 4,000, see Black Friday: 88. For the forensic record of SAVAK’s torture methods, see The Apollo.
Footnotes
-
Emadeddin Baghi, forensic audit for the Bonyad Shahid (Foundation of Martyrs and Veterans Affairs), published 1990s; corroborated by Emadbaghi.com archives ↩
-
SAVAK personnel estimates from declassified US diplomatic cables, Office of the Historian, FRUS 1969–76, Vol. XXVII ↩
-
Stasi personnel and informer network figures, Federal Commissioner for the Stasi Records (BStU), Berlin ↩
-
Amnesty International, annual report on Iran, 1975; later qualified regarding methodology limitations ↩
-
International Committee of the Red Cross, confidential prison inspection reports to the Shah’s government, 1977; US State Department cables estimating 3,400–3,500 political prisoners, 1976 ↩