Case selection and phases

The unit of analysis is a case family. A family may contain several phases where consent, the stated basis for force, the scope of an operation or the form of U.S. involvement changes. Findings apply to the conduct and phases described in each dossier.

The inventory covers direct military operations and selected cases of covert action, operational assistance and political or economic intervention screened for a possible connection to force. The selection defines the atlas’s coverage.

Scale, gravity and victims

The atlas examines the character and scale of uses of armed force affecting a state’s territorial integrity or political independence and asks whether the acts are of sufficient gravity. It considers the intensity, duration and geographic reach of the force, alongside the consequences for people. Article 2 of Resolution 3314 makes sufficient gravity relevant to an aggression determination. Article 8 bis uses character, gravity and scale to assess whether the Charter violation is manifest for the individual crime.

Unlawful armed invasion and occupation are clear examples of aggression. The atlas also treats unlawful armed force of sufficient gravity to impose regime change for ideological or strategic reasons alone as aggression against political independence. The assessment identifies the force used and examines the claimed legal justification for the conduct in each phase.

Drawing on Tom Dannenbaum’s account of aggressive war as a wrong to the people exposed to unjustified killing, the atlas puts victims at the center of its assessment of gravity. It considers deaths, injuries, displacement and damage to homes and essential services where the record supports them. Each assessment relates the harm to the actors, conduct and period covered by the evidence.

State-level classifications

  • Prima facie U.S. aggression. The research record supports a provisional finding for specified U.S. conduct and warrants formal investigation.
  • Prima facie assistance to aggression. The evidence supports a provisional finding that the United States aided or assisted another state’s aggression and warrants formal investigation.
  • Insufficient evidence. The reviewed material falls short of establishing either finding. The case may still warrant further inquiry.

The classifications express scholarly assessments. The dossiers identify the claims, evidence and legal issues that could affect a finding.

For assistance to aggression, the Atlas draws on Article 16 of the International Law Commission’s Articles on State Responsibility. It examines the underlying wrongful act, the assisting state’s knowledge of the circumstances, whether its aid was given to facilitate the act and contributed to it, and whether the act would be wrongful if the assisting state committed it.

Documentary evidence

The project gives priority to official and primary records, including declassified documents, presidential and congressional records, military histories, judicial materials and international documents. It uses independent research and monitoring to cross-check claims and document consequences.

The dossiers distinguish documented conduct, assertions by participants and the author’s legal inferences. They identify material disputes and gaps where these bear on the result.

Mortality and other harm

The cases distinguish direct deaths, excess mortality and wider conflict deaths. Each measure is tied to its source, population, dates and attribution. Other consequences, such as injury, displacement and destruction, appear in the case evidence.

The Atlas estimates total fatalities for cases assessed as prima facie U.S. aggression or assistance to aggression. Where a case-specific study estimates both direct and indirect deaths, the Atlas uses that figure for the population and period the study covers. Where an armed conflict or sustained campaign has a direct-death estimate only, the Atlas models 4 indirect deaths per direct death. A direct toll of 1,000 therefore contributes 5,000 estimated total deaths, including 4,000 modeled indirect deaths.

The Geneva Declaration’s 2008 report uses 4:1 as a conservative assumption across conflicts. A 2015 paper circulated to the UN Security Council cites that benchmark. Case-specific ratios may be higher or lower. Indirect deaths can result from hunger, disease, displacement or disruption of health care and essential services.

The Atlas uses a sourced point estimate where available and the lower figure where the record gives a range or lower bound. It counts overlapping estimates once. Isolated incidents, custodial deaths from all causes and figures with unclear mortality coverage can contribute as partial counts. A sampled set of attacks does not establish a campaign-wide baseline. The Timor-Leste study’s mortality estimate includes modeled killings and disappearances and excess deaths from hunger and illness.

Some figures cover all sides of a conflict or extend beyond the phase assessed in the legal finding. The estimate describes the scale of the selected cases. Attribution to specific actors is reported in the case evidence. Filters select case families and preserve the full period of each selected mortality figure.

Browse mortality evidence

Economic costs

The taxpayer subtotal covers cases assessed as prima facie U.S. aggression or assistance to aggression. It combines documented federal obligations, reported spending and estimates of incremental costs or grant-funded military support. Dossiers distinguish figures based on budgets, spending, obligations and estimated resource costs. Proposed spending, loan face values, third-party spending and overlapping figures are excluded.

The Atlas marks the economic cost to victims as unquantified. Dossiers document damage to homes and infrastructure and disruption to livelihoods and services where sources are available.

The Atlas converts reported U.S. dollars to 2025 purchasing power with the U.S. city average CPI-U (2025 annual index 321.943). It uses the annual amounts when available. It gives a range when a multiyear total can only be placed between the earliest and latest price years. It states any other price-year proxy beside the estimate. Ranges can reflect spending dates or source uncertainty. The taxpayer subtotal sums the lower and upper amounts separately. Case filters preserve the period covered by the source figure. Missing figures leave a gap in coverage.

Browse cost evidence

Map, filters and counters

The map links case families to the countries and territories on the affected side of the conduct assessed. A dossier covering multiple affected locations may appear on more than one present-day map unit. The dossiers describe the historical geography and the roles of the states involved.

Interventions are counted once per case family. The administration counter counts distinct administrations represented in selected cases of prima facie U.S. aggression or assistance to aggression. Nonconsecutive administrations count separately. The country counter counts distinct mapped countries and territories. On the map, bright red marks prima facie U.S. aggression, muted red marks prima facie assistance to aggression, and deep blue marks insufficient evidence. When selected cases in the same place have different findings, aggression takes precedence over assistance, and assistance over insufficient evidence.

Period filters match a case’s documented research window. Intervention categories and presidential filters index conduct described in the case family. A match can include phases outside the filter window or a different phase from the one on which the legal finding rests.