Join War on the Rocks and gain access to content trusted by policymakers, military leaders, and strategic thinkers worldwide.
For eight years I helped decide whether American counter-terrorism money was working.
As a programs advisor in the State Department’s Bureau of Counterterrorism from 2018 to 2025, I oversaw about 20 projects in East Africa worth roughly $80 million a year. None of that money moved without a performance monitoring plan. We established baselines before work started and agreed on indicators in advance. When a project underperformed, or the reporting came late, we withheld funding. Reporting tracked activities against those metrics across the life of the program. Independent evaluations validated what implementers said they had done.
That machinery was not the bureau’s idea. In 2014, the House Foreign Affairs Committee held a hearing titled “The State Department’s Counterterrorism Bureau: Budget, Programs, and Evaluation,” pressing the Coordinator for Counterterrorism on whether the bureau could show what its spending achieved. The following year, the Foreign Affairs Manual codified the Office of Programs and charged it with monitoring and evaluating what the bureau funded. That is the office I worked in.
Nor was it self-administered. In 2020, the State Department’s Office of Inspector General inspected the bureau’s foreign assistance and found that counter-terrorism was not complying with department standards for monitoring and evaluation. The inspection issued a numbered recommendation, published the bureau’s response, and held it open until the bureau produced documentation. Inspectors found one employee serving as grant or agreement officer representative for 86 projects worth almost $147 million, a span of control they flagged as beyond what the bureau could properly oversee. The assistance side failed the standard, too. The difference is that somebody wrote it down.
I should be straight about what all that measurement produced. Some of it is demonstrable. I kept a running record of high-profile attacks in Mogadishu when I first worked there, and it runs to 10 in 2015, 18 in 2016, and 42 in 2017. That last year included the October truck bomb that killed 587 people — a vehicle loaded in Shabelle Valley that passed several checkpoints on the Afgoye road and reached a busy intersection in the center of the city. Multi-donor support hardened the fixed checkpoints on the main approaches, and the mobile vehicle checkpoint capability we funded closed the gaps between them. In December 2019, al-Shabaab tried the same weapon again and could not get in. The truck detonated at the checkpoint on the western edge of the city. No attack in Mogadishu has since approached that scale. Separately, a countering violent extremism program encouraged or assisted more than 1,200 defectors to leave al-Shabaab between 2020 and 2024.
Defeat of al-Shabaab was never the standard. We build counter-terrorism assistance the way we build fire protection: sensors and alarms so you know when something is heating up, a response capability that can move before it spreads, and prevention that keeps the number of sparks down. The proactive unit with source handlers and surveillance teams, the watchlist tied to immigration systems, and the financial reporting center were the alarms. The investigative team, the explosive ordnance capability, and the forensic lab were the response. The defector program was prevention. Judged that way, the record is mixed, but it is legible. Mogadishu stopped burning at the scale it once had. Outside the city, we never contained al-Shabaab, and it is arguably stronger now than it was then.
I can tell you where it worked and where it did not, because the reporting existed to say so. A campaign that publishes no outcome data does not produce a disappointing answer. It produces no answer and no way to assess its effectiveness.
The kinetic component of the same policy has moved the other way.
Judging whether a kinetic strike was lawful or effective requires knowing whom it struck and how many it killed or injured. Across two combatant commands, the United States has stopped producing that information, and the mechanisms that once produced it have lapsed within the past 18 months. This is not to say the strikes are unlawful or unwise. Rather, no one outside the Pentagon is in a position to say so, and the same government demands far more of a grant recipient than it does of an airstrike.
The obvious objection is that this compares two different categories of government action. Foreign assistance is discretionary spending appropriated by Congress and properly subject to performance conditions. Military operations run under different authorities with a deliberately narrower disclosure regime, and importing grant-management expectations into targeting is not truly an apples-for-apples swap. The objection would hold if the standard I am describing came from the assistance world. It does not. In the National Defense Authorization Act for Fiscal Year 2018, Congress required the Secretary of Defense to deliver an annual accounting of civilian casualties caused by military operations. Congress wrote that requirement for military operations specifically and did not borrow it from the grant world. The department has missed it each of the last three years.
In January 2024, U.S. Africa Command reported that a strike near Caad, in central Somalia, killed six al-Shabaab members but injured or killed no civilians. By March 2025, the command had reduced the detail to a line saying the strike killed enemy combatants and injured or killed no civilians. Releases in August 2026 carry a date, an organization, and a distance from the nearest town, and nothing else. The releases across all three years close with the identical sentence that the command will not release specific details about units and assets, to ensure operations security. The caveat sits above the detailed and the empty reporting alike. Whatever changed, it was not the security constraint.
The command’s own accounting stopped, too. Africa Command began producing quarterly civilian harm assessments in April 2020, and members of Congress wrote welcoming the practice and urging it to continue. The most recent assessment on their website covers the period ending June 30, 2025. The command has left four quarters unpublished, July 2025 through June 2026, covering the largest sustained strike campaign in its history. Africa Command published the assessments in accordance with Department of Defense policy on civilian harm, and nothing in that policy obliges a combatant command to publish quarterly.
The mandatory report is separate. Section 1057 requires an annual accounting of civilian casualties be given to the congressional defense committees no later than May 1 each year. The edition covering 2025 reached Congress on Aug. 10, 2026, three months late. The 2024 edition arrived in July 2025, while the 2023 edition arrived in December 2024. The assessment was not the holdup. The report states its determinations as of Feb. 1, 2026, three months before the statutory deadline and six months before the document reached Congress.
The same provision requires that the unclassified report be made available to the public at the same time it goes to Congress, unless the secretary certifies in writing that the publication would threaten national security. The Intercept reported that the department delivered the document to members of Congress without releasing it to the public. CBS News and Stars and Stripes described obtaining the report rather than receiving it.
The 2024 civilian casualty report noted two civilians killed and two civilians injured. The 2025 edition recorded 153 civilians killed and 243 civilians injured, all in Yemen.
The maritime campaign arrives with different architecture — different command, different appropriations, and a different congressional posture. It does not have a different category, though. In 2025, the government designated the cartels as foreign terrorist organizations, and that designation is the stated basis for striking the vessels. Congress placed both campaigns under the same civilian casualty reporting requirement, and the 2025 report covers them together. The government has treated counter-narcotics as counter-terrorism when it wants the authority to kill, and the reporting obligation necessarily follows.
The same report recorded no civilian casualties from the campaign of strikes on vessels in the Caribbean and eastern Pacific, which has killed no fewer than 221 people across at least 66 strikes. Representative Sara Jacobs, who sits on the House Armed Services Committee, said after receiving a briefing that defense officials told her they do not need to positively identify the people aboard the small craft transiting the Caribbean and eastern Pacific in order to strike them, and that they need only show a connection to one of the cartels the government designated as foreign terrorist organizations in 2025, or to an affiliate.
Those two facts do not reconcile. A campaign does not produce zero civilian casualties and an admitted inability to identify the dead unless the government makes that classification by designation rather than by assessment. If a vessel’s designation makes everyone aboard a combatant, the zero is not a conclusion about whom the campaign killed. It describes the category the government placed them in before the strike.
A strike on September 2 shows what that standard cannot cover. The 11 people aboard came from coastal towns in Venezuela’s Paria peninsula that local media outlets identify as the country’s main hub for human trafficking, where boats are known to carry mixed cargo of drugs, weapons, and people. Asked in a classified briefing whether any of the dead could have been trafficking victims, the Joint Staff said they could be. Rear Adm. Brian Bennett of the Joint Staff said they could be. A trafficking victim is not a member of a designated organization.
This pattern extends beyond the Pentagon. Federal law at 22 U.S.C. 2656f requires the Secretary of State to transmit an annual terrorism report to Congress by April 30, covering the calendar year before submission. The department released the most recent edition, which covers 2024, on May 1, 2026. A submission in 2026 should cover 2025, so releasing the 2024 edition then does not satisfy the requirement late. It satisfies that of a different year.
The first Lead Inspector General’s report on Operation Southern Spear, delivered to Congress in May, records what Southern Command would not disclose. The operation’s mission statement is classified. The list of organizations designated for targeting is not publicly releasable. Neither is the accounting of strikes conducted, nor the description of how the command decides between a lethal strike and a non-lethal interdiction. When asked how it measures whether the operation is working, Southern Command told the inspector general that it would not publicly release its measures of effectiveness.
Denied a strike count by the command, the inspector general built one from the New York Times and ABC News.
That leaves one measurable question: whether the campaign reduced the flow of narcotics into the United States. The only assessment arrived in the form of a leaked Drug Enforcement Administration analysis finding no change in cocaine supply or price. The same inspector general report records the Southern Command commander testifying in March that cartels had adapted by shifting routes and moving to air traffic and shipping containers.
My entire annual portfolio ran about $80 million. Southern Command reported obligating $527.9 million on Southern Spear in a single quarter, and $647 million between September 2025 and March 2026. The campaign produces a one-sentence casualty count from Southern Command and nothing from Africa Command.
Detailed disclosure of battle damage assessments has its own risks. It can reveal sources, methods, and patterns of life. Al-Shabaab and the Islamic State read press releases and put out their own. If a government names a man it killed, the group need only produce a recording of him discussing later events to counter the claim. A government whose claims are repeatedly debunked loses credibility. Civilian deaths in airstrikes can radicalize individuals with no prior affiliation with the targeted terrorist group, an observation published recently by Viktor Marsai of the Soufan Center.
None of this explains the current posture. The 2024 and 2025 releases published counts and named individuals while withholding every unit and asset involved. The quarterly civilian harm reports ran for 22 consecutive editions without an adverse operational security disclosure incident anyone has cited.
Nigeria published what Washington would not. Africa Command’s May 16, 2026, release reported an initial assessment that American and Nigerian forces killed multiple terrorists but gave no number. The May 18 release said analysts were completing their assessments, though the command never published the findings. Abuja was specific. On May 19, the Nigerian Defence Headquarters said its assessments indicated the operation had eliminated 175 Islamic State fighters, named three more commanders killed, and listed what the strikes destroyed. Nobody has verified that count. But the only public accounting of an American operation came from a partner government, which faced the same adversary and the same propaganda risk, and reported anyway.
A return to regular publication could reassure Congress and the public that their trust, in their wallet and in their conscience, is well placed. The quarterly civilian harm reports were voluntary, and Africa Command could restore them. The Secretary of Defense could publish the Section 1057 report on the statutory schedule and explain plainly how a campaign that killed at least 221 people produced zero civilian casualties. The Secretary of State could transmit the annual Country Reports on Terrorism for the year it is due.
None of this has happened on its own. Officials under this administration and the last have missed the terrorism report deadline without consequence. Congress could condition appropriation on the reporting it already requires.
Counter-terrorism programs that cannot show their work lose their funding. Strike campaigns in Africa and the Western Hemisphere are no longer showing their work or their impact, and the U.S. taxpayer continues to fund them.
Jim Loucks served as a programs advisor in the U.S. Department of State’s Bureau of Counterterrorism from 2018 to 2025, managing counter-terrorism assistance programs in East Africa. A former Idaho State Police officer, he has worked as a police advisor and security force assistance consultant in Somalia, Afghanistan, and Iraq, and now writes on counter-terrorism and national security at the CT Conundrum.
Image: Staff Sgt. Abigail Duell via DVIDS
