No Annex of Statistical Information, the report's statistical backbone in years past, has been published with it.

And the report's own pages cite terrorism-incident rankings from the Global Terrorism Trends and Analysis Center, or GTTAC - the same data program whose funding the department allowed to lapse in March.

The report is not optional. Federal law - 22 U.S.C. § 2656f, on the books since 1987 - directs the secretary of state to transmit to Congress each year "a full and complete report" on terrorism worldwide, including, "to the extent practicable, complete statistical information" on those killed, injured, or kidnapped "by each terrorist group during the preceding calendar year."

For two decades, professionals met that statistical requirement: the National Counterterrorism Center from 2004 to 2011, the University of Maryland's Global Terrorism Database from 2012 to 2017, and GTTAC's Record of Incident Database - GRID - from 2018 on. GRID was purpose-built for the statutory task. It applied the government's own definition of terrorism, named its sources, published its codebook, and put its methodology through peer review this year in the journal Democracy and Security.

In March, the funding stopped. The team behind GRID kept collecting for a time on its own; the public record now ends in mid-May. Ambassador Adam Blackwell, a GTTAC principal, told me that appeals to senior officials at the State Department, in Congress, and at the White House all failed, and that he finds the timing difficult to credit: the United States is engaged in active campaigns against Hamas, Ansar Allah, Hezbollah, and the IRGC, and has newly designated cartels and gangs from across the hemisphere - Venezuelan, Salvadoran, Mexican, and Haitian - as foreign terrorist organizations. This is the moment Washington chose to stop counting.

Note the statutory asymmetry. The report is mandated unconditionally; only the statistics carry the qualifier "to the extent practicable." Defund the sole mechanism that made the statistics practicable, and the qualifier becomes an exit. The 2024 report's own compliance chapter makes the gap explicit: it states that the statistical requirement "is satisfied through the inclusion of a statistical annex to the report" - an annex that has not been published. The 2025 edition will be worse still, because the incidents of 2025 are not being systematically collected by anyone charged with the task.

The reflex answer - use the alternative datasets - does not survive contact with the data. I spent my doctorate inside those datasets, and their custodians have been admirably candid about the limits. By its own founding director's account, the Global Terrorism Database's source-validity rule excluded 28 percent of otherwise-eligible cases worldwide in an early-2014 sample - and 76 percent in Syria, precisely where the fighting was.