Fact-check
Did the U.S. Stop Reporting Civilian Casualties After 14,000?
Four distinct accountability failures — EO revoked, statute missed, ROE loosened, war-zone labels quietly expanded — collapsed into one claim that doesn't hold up on closer inspection.
By The Crosscheck Desk · 2026-08-25
The US simply stopped reporting civilian casualties after around 14,000 in the anti-ISIS campaign
The Trump administration had a trend of lying about military costs (on civilian casualties)
Trump revoked Obama's civilian casualty executive order and missed the first statutory reporting deadline
Four Mechanisms, Not One
The widely shared formulation — that the U.S. "stopped reporting civilian casualties" — collapses at least four legally and operationally distinct phenomena. Treating them as one distorts the accountability picture in both directions: it overstates what was formally ended, and understates the non-repealed mechanisms that remained in force but were consistently undermined in practice.
1 — EO 13732 (Obama, 2016) → Revoked by EO 13862 (Trump, March 2019)
Executive Order 13732, signed by President Obama on July 1, 2016, required the Director of National Intelligence to release, by May 1 each year, an unclassified summary of U.S. strikes against terrorist targets outside areas of active hostilities, with assessments of combatant and civilian deaths across all agencies — including the CIA. President Trump revoked Section 3 of that order via Executive Order 13862 on March 6, 2019.[1] What was lost: the government-wide, agency-spanning count, including CIA operations. What remained: the narrower DoD-only statutory report below.
2 — NDAA FY2018 §1057 Statutory Report (Still in Force)
Section 1057 of the National Defense Authorization Act for Fiscal Year 2018 requires the Secretary of Defense to submit an annual report on civilian casualties from U.S. military operations to Congress, with a public unclassified version.[2] An executive order cannot repeal a statute. This requirement survived Trump's 2019 EO and remains law. The Trump administration missed its first deadline under this provision — May 1, 2018 — and later requested an extension.[3] Subsequent reports were filed but drew consistent criticism for severe undercounting relative to independent estimates.[4] Non-compliance with a standing statute is a sharper story than repeal; do not flatten one into the other.
3 — Rules of Engagement: Targeting Standards and Approval Levels
Obama's 2013 Presidential Policy Guidance (PPG) required "near certainty" of no civilian harm and "near certainty" of target identification before authorizing strikes outside active hostilities. Trump's 2017 changes retained the "near certainty" civilian-harm standard for women and children but lowered the target-identification standard to "reasonable certainty" for adult civilian males — a lower bar — and pushed strike-approval authority to lower command levels.[5] Defense Secretary Mattis announced a shift from "attrition" to "annihilation" in the ISIS fight in 2017, publicly stating that "civilian casualties are a fact of life."[6] These changes affected how many strikes were authorized, not what got reported after the fact.
4 — Designating New "Areas of Active Hostilities"
This is the least-discussed mechanism, and arguably the most consequential for the specific reporting regime. The Obama-era PPG applied its strictest standards — including "near certainty" and detailed post-strike reporting — to operations outside areas of active hostilities. In March 2017, Trump designated parts of Somalia and Yemen as "areas of active hostilities,"[7] which moved those operations into the war-zone rules category and out of the stricter regime, without repealing or revoking anything. No announcement was required; the mechanism worked silently. Biden later reversed these designations in 2022.[8]
The claim that the U.S. "simply stopped reporting" most accurately describes Mechanism 1 — the EO 13732 Section 3 reporting that was formally revoked in 2019. Mechanisms 2, 3, and 4 operated simultaneously, affecting volume of harm, quality of documentation, and scope of reporting, but none constituted a complete stop.
Tracing the "14,000" Figure
The specific figure "around 14,000" civilian casualties in the anti-ISIS campaign does not clearly correspond to any official count or any single widely-cited independent tally. Here is what the best-documented numbers actually show, and where "14,000" may have come from.
Official coalition count (OIR/CENTCOM): The U.S.-led coalition (Operation Inherent Resolve) confirmed approximately 1,417–1,475 civilian deaths across Iraq and Syria for the entire campaign through 2019.[9] That is an order of magnitude below 14,000.
Airwars "Fair/Confirmed" estimate: The monitoring organization Airwars tallied a range of 8,264–13,360 civilian deaths likely attributable to coalition strikes — based on cross-verified reporting from multiple credible local and international sources.[10] The upper bound of this range (13,360) rounds to "about 14,000" and is the most plausible source of the figure in circulation. This is the most likely origin of the 14,000 number. If that is the source, two points of conflation occur: Airwars' upper estimate includes a wide uncertainty band, and it covers the entire campaign from 2014 through the near-end of the caliphate, not a specific point at which reporting "stopped."
AP Mosul investigation (all-cause deaths): An Associated Press investigation published in December 2017 found that more than 9,000 civilians died in the final battle for Mosul — across all causes, including ISIS violence, Iraqi and coalition forces, artillery, and building collapse.[11] This is a different figure from a single battle, covering deaths by all parties, and should not be cited as the total coalition civilian death toll for the campaign. Coalition responsibility for Mosul deaths, per the AP, was at least 3,200 — still far below 14,000.
Total locally alleged deaths (unverified): Airwars logged a total of 19,233–29,841 deaths from all locally reported incidents.[10] This is the widest net and includes unverified claims. It is far above 14,000.
The Casualty Counting Table
Figures in this space are routinely conflated because they measure different things with different methodologies. The differences between rows below can exceed an order of magnitude.
| Figure | Source | Range | What It Counts | Methodology |
|---|---|---|---|---|
| Coalition confirmed civilian deaths (OIR campaign total) | CENTCOM / OIR[9] | ~1,417–1,475 | Civilian deaths from coalition strikes that CENTCOM accepted responsibility for | Internal military investigation; does not include site visits or witness interviews |
| Airwars "Fair/Confirmed" (coalition strikes) | Airwars[10] | 8,264–13,360 | Civilian deaths from coalition strikes assessed as "fair" or "confirmed" (multiple credible sources, nearby confirmed action) | Cross-referencing local and international reports, UN data, CENTCOM data; excludes single-source "weak" claims |
| Airwars "All allegations" (unverified) | Airwars[10] | 19,233–29,841 | All locally reported alleged civilian deaths from coalition actions, including unverified single-source claims | Compilation of all reported incidents; includes contested and discounted |
| Mosul civilian deaths (all causes) | AP investigation[11] | 9,000–11,000 | All civilian deaths in the battle for Mosul — coalition, Iraqi forces, ISIS, building collapse, IEDs | Mosul morgue records, Amnesty International, Iraq Body Count, UN data; covers Oct 2016–Jul 2017 |
| Mosul deaths attributed to coalition/Iraqi forces | AP investigation[11] | ~3,200 | Coalition and Iraqi military responsibility within the Mosul battle | Subset of AP Mosul total; coalition confirmed only 326 |
| Obama-era drone strikes civilian deaths (outside war zones only) | Obama admin EO 13732 report[12] | 64–116 | Civilian deaths from U.S. strikes outside active hostilities, 2009–2015 | Government assessment; critics and independent monitors placed figure far higher |
Note on DoD undercounting: The gap between official and independent figures is not unique to the ISIS campaign. DoD's 2020 Section 1057 report listed 23 civilian deaths across all active theaters; the UN Afghanistan mission alone documented 89 international-force civilian deaths in Afghanistan for the same year.[4] Analysts at Just Security identified an absence of site visits or witness interviews as the structural cause: DoD relies on operational planning data and intelligence sources, which cannot verify who occupied a building at the moment of a strike.[13]
The Steelman: Urban Warfare, Not Just Policy
The strongest case for the other side
Civilian casualties rose sharply in 2017. The correlation with the Trump inauguration is real. But so is another factor: 2017 was the year the anti-ISIS campaign moved into the densest urban cores — the residential neighborhoods of western Mosul beginning in February 2017, and then Raqqa from June through October 2017. ISIS deliberately embedded fighters among civilians, used human shields, and mined civilian infrastructure. Urban siege warfare against a force with no separation from the civilian population raises civilian casualties regardless of who is president, what the rules of engagement say, or what the reporting requirements are.[14]
The honest analytical finding is that the evidence cannot cleanly separate the contribution of policy change from phase change. The loosened ROE and lower approval thresholds likely contributed to increased strike rates. The shift to dense urban combat almost certainly drove higher per-strike civilian exposure. Both were happening simultaneously and their effects compound. Studies of the Raqqa campaign specifically found structural building collapse — from artillery, not precision air strikes — accounting for a substantial share of deaths.[15]
The score: The phase-change argument earns a score of 4 on the adversarial scale — it directly addresses the core causal claim with evidence. The correct verdict is not "Trump policy caused the civilian death spike" but rather "the Trump-era loosened ROE and reporting rollbacks compounded a casualty increase that was partly structural to the campaign's urban phase." That is both a more honest finding and a sharper one.
The Obama Baseline
Any comparative claim about Trump-era transparency requires an honest accounting of what the baseline looked like. It was not clean.
The Obama administration's official count of drone-strike civilian deaths outside war zones — 64 to 116 for the period 2009–2015 — was disputed by independent monitors at the time of its release.[12] The Bureau of Investigative Journalism estimated far higher figures for the same period. The CIA's signature strike program, under which missiles were fired based on behavioral patterns rather than identified individuals, operated for years without any public acknowledgment. Obama issued EO 13732 in July 2016 — the final year of his presidency — precisely because of sustained criticism from human rights organizations and congressional pressure for greater accountability.[16] The 2016 order was a meaningful step, but it came after seven years during which the administration's own reporting was contested as undercounting.
The Trump administration's subsequent rollback was a regression from an already incomplete baseline. Framing the Obama-era record as transparent and the Trump record as opaque misses the continuity of institutional resistance to civilian casualty accounting across both administrations, while obscuring what was new under Trump: the formal revocation of even the incomplete EO requirement, the designation of expanded active-hostilities zones, and the statutory reports that arrived late and were consistently thinner than independent findings.
What "Lying" Actually Covers
The claim uses the word "lying" — and the broader framing holds that the government had "a trend of lying about military costs." For civilian casualty reporting specifically, the record supports three distinct accusations with different evidentiary weights.
The record clearly supports the first two accusations. The third — affirmative false statements — carries a higher evidentiary bar and the documentary record does not clearly meet it for civilian casualty counts specifically. (A separate question, outside scope here, is whether Trump administration officials made false statements about other military costs; see the Crosscheck pieces on munitions spending[17] and Iran war cost estimates.[18])
The appropriate framing is: the Trump administration systematically reduced its transparency obligations regarding civilian harm, and the figures it did publish were consistently far below what independent monitors found — a pattern that the administration attributed to methodology differences and critics attribute to institutional incentives to minimize harm acknowledgment.
Claim Verdicts
What to Watch
Forward Look
- NDAA §1057 report, Trump second term (2025–): The second Trump administration began in January 2025. The next Section 1057 reports are required annually; watch whether they are filed on time, whether they cover CIA-adjacent operations, and how their figures compare to Airwars' contemporaneous monitoring.
- PPG/ROE status: Biden restored some of the Obama-era PPG standards and reversed the Somalia/Yemen "active hostilities" designations in 2022. The second Trump administration's ROE posture in Somalia and Yemen bears watching for re-designation.
- Airwars monitoring as a comparison baseline: Airwars publishes monthly estimates. A widening gap between its figures and DoD §1057 reports across the second Trump term would replicate and confirm the first-term pattern. A narrowing gap would be evidence that methodology or commitment has changed.
- Congressional oversight: Several NDAA provisions have successively tightened §1057 requirements. If the second Trump administration consistently files late or thin reports, watch for enforcement actions or legislative escalation.
What's the claim?
People online say the U.S. government "stopped reporting civilian deaths after around 14,000" in its war against ISIS. They also say the government was "lying about military costs."
The short answer
These claims are misleading. The government did reduce how much it reported. But it did not simply stop at a certain number. And the "14,000" figure does not match any official or independent count.
Four things got mixed into one
There were actually four separate things that happened with casualty reporting. Treating them as one distorts what occurred.
1. A reporting rule was revoked. President Obama signed Executive Order 13732 in July 2016. It required the government to release a yearly report on strikes and civilian deaths from operations outside active war zones. President Trump signed Executive Order 13862 on March 6, 2019. It cancelled that reporting requirement.
2. A law stayed in force. Congress passed a law in 2018 — NDAA Section 1057 — requiring the Pentagon to report civilian deaths to Congress every year. An executive order cannot cancel a law. The Trump administration missed its first deadline under this law (May 1, 2018) and later asked for an extension. It eventually filed the required reports, but they were late and listed far fewer deaths than independent monitors found.
3. Rules for targeting changed. Obama's policy required "near certainty" that no civilians would be harmed before a strike. Trump's 2017 changes kept this requirement for women and children but lowered it to "reasonable certainty" for identifying adult male targets — a lower standard. Approval for strikes was also pushed to lower military ranks. Defense Secretary Mattis announced a shift to "annihilation" tactics and said civilian casualties were "a fact of life."
4. Areas were re-labeled. Obama's stricter rules only applied outside "areas of active hostilities." In 2017, the Trump administration labeled parts of Somalia and Yemen as "areas of active hostilities." This moved those operations out of the strict reporting rules without changing any law or announcing a new policy. Fewer people noticed this change than the others. President Biden reversed these designations in 2022.
Where did "14,000" come from?
The figure does not match any one official or independent count. Here is what the real numbers look like:
- The U.S.-led coalition officially confirmed about 1,417 to 1,475 civilian deaths from its strikes in Iraq and Syria.
- Airwars, an independent monitoring group, estimated 8,264 to 13,360 civilian deaths from coalition strikes — using cross-checked reporting from multiple sources.
- An Associated Press investigation found more than 9,000 civilians died in the battle for Mosul alone — but that counts deaths from all sides, including ISIS violence and building collapse, not just coalition strikes.
The "14,000" number is closest to the upper end of Airwars' range (13,360), rounded up. If that is where the number came from, two things got confused: Airwars' upper estimate includes a wide range of uncertainty, and it covers the whole campaign, not a specific cutoff point where reporting stopped.
Why did deaths rise in 2017?
Civilian deaths went up sharply in 2017. That happened during the Trump administration. But it also happened to be when the campaign moved into the most crowded urban areas — the neighborhoods of Mosul starting in February 2017, and then Raqqa from June through October 2017. ISIS fighters hid among civilians and used them as human shields.
Urban warfare raises civilian deaths no matter who is president or what the rules say. Studies found that building collapse from artillery caused a large share of deaths in Raqqa. The evidence cannot clearly separate how much was caused by policy changes versus the nature of the fighting. Both likely contributed.
Was the Obama era better?
The Obama administration's official count of civilian deaths from drone strikes outside war zones was 64 to 116 for 2009 to 2015. Independent monitors placed the number much higher. The CIA's practice of striking based on behavioral patterns — without identifying specific people — ran for years without public acknowledgment. Obama signed EO 13732 in the last year of his presidency, after years of criticism about secrecy. Trump's rollback was a step backward from an already incomplete baseline.
Was the government "lying"?
The word "lying" covers three different accusations. The record supports some more than others.
- Reduced or ended reporting: Supported. The EO reporting was cancelled. The required NDAA reports were filed late and covered less than the EO did.
- Counting too low: Partly supported. Pentagon figures ran five to ten times lower than independent monitors across multiple countries and years. The Pentagon does not conduct site visits or interview witnesses, which leads to lower counts. The Pentagon says its methods are different, not wrong.
- Stating known false numbers: Not clearly supported. No documented case of officials stating a number they knew to be false. Mattis's "fact of life" comment was a policy statement, not a false claim about a count.
Reduced reporting and undercounting are real and significant. But they are not the same as lying.
Verdicts
Claim: "The US stopped reporting civilian casualties after around 14,000."
Rating: Misleading. Reporting was rolled back and degraded, not stopped at a casualty threshold. The "14,000" figure does not match any single official or verified count.
Claim: "A trend of lying about military costs" (on civilian casualties).
Rating: Misleading. The documented record shows reduced disclosure and systematic undercounting. It does not clearly support affirmative false statements about casualty numbers.
Underlying record: Trump formally revoked Obama's civilian casualty EO and missed the first statutory reporting deadline.
Rating: Mostly True. EO 13862 did revoke Section 3 of EO 13732. The first NDAA report deadline was missed. The statutory reporting continued, but late and thin.
Sources
- Trump Revokes Obama Executive Order on Counterterrorism Strike Casualty Reporting
- Congressional Action on Civilian Harm Resulting from U.S. Military Operations: Part II
- Trump Administration Stalls on Required Civilian Casualty Reports
- A Legacy of Unrecognized Harm: DoD's 2020 Civilian Casualties Report
- "Reasonable Certainty" vs "Near Certainty" in Military Targeting — What the Law Requires
- US Using Annihilation Tactics in Final Stages of ISIS Fight in Mosul, Mattis Says
- Trump's first year in numbers: Strikes triple in Yemen and Somalia
- Targeting Yemen and the Repeal of Obama Constraints
- Coalition airstrikes against ISIS resulted in more than 1,400 civilian deaths, according to data made public for first time
- U.S.-led Coalition in Iraq & Syria
- Liberating Mosul from ISIS left more than 9,000 civilians dead: report
- Obama administration discloses number of civilian deaths caused by drones
- The Fatal Flaw in DOD's Latest Civilian Casualties Report
- Urban Warfare Project Case Study #2: Battle of Mosul
- Civilian Casualties: Lessons from the Battle for Raqqa
- Trump Inherited the Drone War but Ditched Accountability
- Trump, Truth Social, and Munitions Spending — Fact-Check
- Hegseth, Murray, and Iran War Cost Estimates — Fact-Check
- Reflecting on the Civilian Casualty Executive Order: What Was Lost and What Can Now Be Gained
- Executive Order 13862 — Revocation of Reporting Requirement
