The Cost Kept Off the Books

The $884 billion base-budget line is the paper's own estimate. Its footnote 23 derives it by assuming that 50, 40, 25 and 20 percent of overseas contingency funding in different year bands raised the base budget. Those percentages are the author's assumptions and do not appear in any government account. The paper excludes state and local spending, allied spending, interest after fiscal year 2023 and some aid.

DoD's Section 1090 report to Congress, dated April 2024, gives cumulative war-related costs for fiscal years 2001 through 2023 of $881.9 billion for Afghanistan, $799.3 billion for Iraq and Syria and $48.7 billion for Ukraine, a total of $1,729,916 million, or $1.73 trillion. Its table footnote states that "These amounts exclude non-Department of Defense classified programs." Crawford's paper cites the department's March 2021 figure of $1.596 trillion through fiscal year 2020 and agrees that "over a trillion dollars" is correct for DoD appropriations alone.

The Center for Strategic and International Studies reviewed war-cost reporting in January 2019 and analyzed an earlier Crawford estimate. Its author, Anthony Cordesman, wrote that "some outside efforts to estimate the total cost of U.S. conflicts may sharply overestimate total costs." He wrote that the overseas contingency line "is also somewhat inflated since it includes some costs that involved low level operations and because the Congress allowed some non-war related costs to be included to avoid the limits on regular defense spending imposed by the Budget Control Act." The same report called Crawford's work "one of the best such estimates" and said the government is "failing to meet its responsibility to report meaningfully on America's wars."

The Costs of War project describes itself as research on "the ongoing consequences of the United States post-9/11 wars." Its method statements sit in each paper, and its 2025 budget page restates the $8 trillion estimate through fiscal year 2022, of which "about $2.3 trillion were spent on the budget category for post-9/11 military operations." The project's separate figure for future veterans' care comes from Linda Bilmes, who projected in August 2021 that care for post-9/11 veterans would cost $2.2 to $2.5 trillion between 2001 and 2050. That projection is double her 2011 and 2013 projections and predates the PACT Act.

GAO's report GAO-17-68 found that DoD's own non-war share of OCO appropriations rose from about 4 percent in fiscal year 2010 to 12 percent in fiscal year 2015. GAO's January 2018 follow-up letter records that "On occasion, Congress has directed that O&M OCO appropriations be used to support base requirements." It cites $9.1 billion of operation-and-maintenance OCO funds authorized for base requirements in the fiscal year 2016 defense authorization and $1 billion directed to readiness in fiscal year 2015.

CBO states that the OCO label originally followed historical convention for the first operations after the attacks of September 2001. It also records that DoD's fiscal year 2019 request planned to "increase the base budget to include most of that enduring funding in future years." Crawford notes that the Biden administration eliminated the OCO category.

DoD concurred with the first recommendation and partially concurred with the second. GAO's January 2018 letter states that DoD "has not taken any steps to implement them." GAO's January 2017 report and 2018 letter record no DoD publication of the enduring-cost estimate.

The Medicine report documents the earlier cases of the same sequence for military exposures. Chapter 12 of that report records GAO's finding that flawed exposure models undermined federal conclusions about Gulf War chemical-agent exposure (MED-12.5) and the 2014 advisory committee's findings on Gulf War illness (MED-12.6). Burn pits and the PACT Act are not covered there.

According to the National Academies release, the committee also reported that most pollutant levels at Joint Base Balad "were not higher than levels measured at other polluted sites worldwide," and that research on firefighters and incinerator workers "has not indicated increased risk" for most outcomes. An Air Force article in 2011 summarized the report as finding no evidence of harm. The release states that data gaps prevented firm conclusions.

Section 1120 applies "notwithstanding that there is no record of evidence of such disease during the period of such service." VA describes the law as "perhaps the largest health care and benefit expansion in VA history." Section 808 requires burn-pit transparency reports, and section 805 creates the Toxic Exposures Fund.

CBO's June 2022 estimate of the bill as amended by the Senate committee put total direct spending at $667.0 billion in outlays for 2022 through 2031. That figure includes $390.3 billion of current-law discretionary spending reclassified as direct spending and $113.0 billion of discretionary spending classified as direct, so part of it is not new spending. Compensation accounts for $151.7 billion of the total. CBO scored a committee version, and no score of the enacted text was located.

CBO's February 2026 baseline projects Toxic Exposures Fund budget authority of $52.7 billion in 2026 and outlays of $746.2 billion for 2027 through 2036. The fiscal year 2027 budget request lists $175.5 billion for VA medical care, of which $52.1 billion is mandatory funding from the Fund.

VA states that 55.8 percent of 2024 decedents had VA contact in 2024 or the prior five years, and 44.2 percent did not. The Limitations section, page 105, states: "Of note, there is not a single comprehensive roster of the Veteran population and source records may be incomplete or contain errors." The report states that it is not an evaluation of risk factors or of the effect of VA suicide-prevention programs. It gives a suicide rate of 80.1 per 100,000 for veterans with a recent traumatic brain injury diagnosis and 38.0 per 100,000 for those without.

The Costs of War project published a separate estimate in 2021 of 30,177 suicides among post-9/11 active-duty personnel and veterans, against 7,057 killed in war operations. Its abstract says "at least four times as many" died of suicide as in combat. The estimate combines government data, secondary literature and interviews, and VA's tables are not split by service era.

GAO's report opens: "There is growing evidence that using certain weapons that produce high levels of what is called blast overpressure (the sudden onset of a pressure wave) can affect service members' brain health." The Defense Health Agency recorded 541,032 first-time traumatic brain injury diagnoses from 2000 through the fourth quarter of 2025, of which 442,538, or 81.8 percent, were mild. The agency counts each service member once, at the highest severity, to avoid over counting, and counts diagnoses made anywhere U.S. forces are located, including the continental United States. It states that most mild cases return to duty in 10 to 14 days.

The figures below are the Costs of War project's estimates, drawn from several sources, and the project states the ranges and the limits of each. The project's pages give different totals on different dates.

The project's pages also give U.S. military deaths as 7,052, 7,053 and 7,057. The Crawford and Lutz paper states that its counts are approximations from several sources and that the categories are disputed. It states that sources "tend to be conservative, only counting deaths which researchers have verified through two or more independent sources."

Savell's 2023 paper estimates 3.6 to 3.8 million indirect deaths. The figure comes from applying a ratio to the direct-death count. The paper states that "a reasonable, conservative average estimate for any contemporary conflict is a ratio of four indirect deaths for every one direct death," a ratio the paper takes from the Geneva Declaration Secretariat's 2008 report. The paper states that the precise figure is unknown. The indirect-death total is a modeled estimate and no body has counted those deaths.

Vine and colleagues estimated in 2020 that at least 37 million people were displaced by the eight most violent post-9/11 wars, and possibly 48 to 59 million. The authors write that "37 million is a very conservative estimate." The project's August 2021 update gives 38 million.

DoD reported to Congress that its 2022 operations resulted in "no civilian casualties." The report for 2023 states one civilian killed and one injured, and two ex gratia payments, against none in 2022. DoD counts a casualty where it is "more likely than not" that civilian casualties occurred, and only those attributed to U.S.-operated weapons. The 2023 report says some reports could not be assessed "due to insufficient information provided or because related reports are still pending review." The statutory deadline for the 2022 report was 1 May, and the document was generated on 12 March 2024.

The DoD Inspector General reviewed how CENTCOM documented and assessed civilian-casualty allegations in a report dated 18 May 2021 and publicly released in December 2021. It found that pre-strike targeting followed doctrine and that documentation, reporting and assessment teams could be improved. Most of its findings and recommendations are redacted in the public version.

GAO reported in February 2024 that DoD lacks comprehensive records of alleged law-of-war violations. It obtained records of 47 alleged violations by U.S. personnel in AFRICOM and CENTCOM from January 2012 to December 2022, and not all entities could provide complete records. DoD concurred with both recommendations. In March 2024 GAO found that the Action Plan captured 61 of 68 recommendations from nine earlier studies and that DoD had not addressed two key challenges, including what constitutes improvement. DoD did not concur with GAO's recommendation to set performance goals and measures.

The Inspector General's May 2026 report credits DoD with progress: a Steering Committee, new hires, a DoD Instruction in December 2023 and a Civilian Protection Center of Excellence by February 2025. It found that components did not fully implement any Action Plan objective by the end of fiscal year 2025, ended funding for the data management platform, stopped Steering Committee meetings and lost personnel. DoD partially agreed with the recommendations. The Under Secretary of War for Policy "did not agree with the need to develop and implement a plan to ensure compliance with statutory requirements," and both recommendations remained unresolved.

DoD's published position is that "The U.S. military is steadfastly committed to limiting harm to civilians."

The source of the records is the FOIA suit Khan v. Department of Defense, filed in the Southern District of New York on 13 June 2018 (18-cv-5334), over requests for CENTCOM records on civilian-casualty allegations in Operation Inherent Resolve. The counts are the Times's, and the finding is graded as reported. A military spokesman, Captain Urban, told the Times: "An honest mistake ... is not, in and of itself, a cause for disciplinary actions as set forth in the law of armed conflict." He also said investigators in hostile territory may be unable to visit blast sites. The register holds no DoD document that reproduces the Times figures.

Battlefield survival improved. A 2019 analysis in JAMA Surgery of 56,763 battle casualties found that the case-fatality rate fell in Afghanistan from 20.0 to 8.6 percent and in Iraq from 20.4 to 10.1 percent between the early and later stages of the wars. Tourniquets, transfusion and transport times accounted for 44.2 percent of the estimated reduction in mortality. The paper does not address veterans' costs.

The Costs of War suicide paper links better survival to higher lifelong disability and care costs, writing that medical advances let service members survive and "return to the frontlines in multiple deployments." That link is the project's reading. Bilmes's projection states that more than 40 percent of post-9/11 veterans were already approved for lifetime disability benefits.

Brown University's project, CBO, GAO, the DoD Inspector General and the Institute of Medicine produced the figures in this chapter on spending, enduring costs, exposure data and civilian harm. DoD and VA publish war-cost, suicide and civilian-casualty reports, and those reports state their own exclusions and limits.

No government total exists for the wars. DoD held an enduring-cost estimate it did not report. VA states there is no complete roster of veterans, and DoD's plan states there is no comprehensive civilian-harm database.