The Hurricane Maria death toll
Puerto Rico's official Hurricane Maria toll stood at 64 for eleven months. Independent researchers estimated thousands. The government fought the release of death records, lost, commissioned its own study, and revised the figure to 2,975.
Why this verdict
Confirmed by the Puerto Rican government's own commissioned study and by its subsequent revision: an official toll of 64 was replaced with an estimate of 2,975 after mortality records the government had resisted releasing were obtained by litigation.
- Period
- Region
- Reading time
- Sources
01The claim
That the government of Puerto Rico and federal authorities systematically undercounted deaths caused by Hurricane Maria in September 2017, and withheld the mortality data that would have shown it, minimising the scale of the disaster and the inadequacy of the response.
02Where it came from
The certified toll rose slowly and stopped at 64. The figure counted only deaths a medical examiner attributed directly to the storm — drownings, structural collapse — and not the deaths that follow when an island loses electricity for months: dialysis interrupted, oxygen concentrators without power, hospitals running on failing generators, medication unrefrigerated. Journalists and demographers in Puerto Rico and the mainland began comparing monthly death registrations against prior years and finding a large gap almost immediately.
03What the record shows
✓ CONFIRMED — It happened. The documents exist.
The excess-mortality method is the standard tool for exactly this problem and it does not depend on anyone's certification decisions: count deaths in the period, compare with the historical baseline, and the difference is attributable to the event.
Several groups ran it. Researchers from Penn State and the University of Utah, working from registry data, put excess deaths above a thousand within months. A Harvard-led household survey published in the New England Journal of Medicine in 2018 produced a central estimate above four thousand with a very wide confidence interval — a figure widely reported without the interval, which was a reporting failure rather than a research one.
The concealment is documented separately from the counting. The Centro de Periodismo Investigativo and CNN sued the Puerto Rican government for the death registry data, which it had declined to release; the courts ordered disclosure. That litigation is what makes this a cover-up entry rather than a statistics entry.
The government then commissioned George Washington University's Milken Institute School of Public Health to do the work properly. Its August 2018 report estimated 2,975 excess deaths over the six months following the storm, and found that the mortality risk was highest among the poorest and the elderly. Puerto Rico adopted that figure as the official toll.
04What it cost
The revision changed the political record of the response — 64 deaths and 2,975 deaths describe different events — and the delay meant the disaster was assessed by Congress and the public against a number that was wrong by a factor of forty-six. The GWU report also produced the more durable finding: that the deaths were concentrated among people who could not privately replace the services the state had failed to restore. Puerto Rico subsequently revised its death-certification protocols for disasters, which is the concrete fix, and the case is now a standard citation for using excess mortality rather than certification counts.