Trump Cancels Drone Strike Civilian Casualty Report: Does It Matter?

April 2, 2019

On March 6, without explanation, President Donald Trump revoked an Obama-era rule requiring an annual public report on U.S. drone strikes, including civilian casualties. This prompted worries: What are they trying to hide? Are they planning something big? Covering up a disaster? As The Atlantic’s Conor Friedersdorf put it, “Trump just gave himself more power to kill in secret.”

These fears are overblown. Not because criticisms of the drone campaign are unwarranted, but because the statistics the Obama administration generated weren’t especially believable. (Though the reporting requirement was in place last year, the Trump administration ignored it.) Obama’s official numbers systematically underestimated civilian casualties and did not disclose the location of strikes.

The drone program, which refers to extrajudicial strikes against suspected terrorists outside of war zones — i.e., Pakistan, Yemen, and Somalia, rather than Afghanistan, Iraq, or Syria — began under George W. Bush and accelerated under Barack Obama. When, in the last year of his presidency, Obama ordered the public disclosure of information on casualties resulting from strikes taken outside traditional war zones, it was the first time the United States officially acknowledged the extent of the drone campaign. But everyone already had a good idea.

As I note in my book Drones and Terrorism, the report claimed 473 drone strikes occurred from 2009 to 2015, with an estimated civilian casualty rate of 2.63 percent to 4.30 percent, substantially lower than independent estimates. A second report covering 2016 claimed 53 strikes occurred with one civilian casualty, accounting for 0.2 percent of an estimated 431–441 people killed. Add that to the first report and, according to the Obama administration, civilians made up just 2.27 to 3.73 percent of drone strike deaths. However, independent estimates suggest a much higher civilian casualty rate, perhaps due to the way the government classifies combatants.

Measuring Drone Strikes

To determine how frequently American drones kill civilians, I used statistics from two non-governmental organizations: New America (based in the United States) and the Bureau of Investigative Journalism (BIJ, based in the United Kingdom). Both organizations use a variety of local and international sources to compile strike and casualty data, offering low and high estimates. To account for the full range of possibilities, I used whichever of those two groups’ low estimate was lowest and high estimate was highest.

Here’s what the combined stats look like for Obama’s two terms:

U.S. Drone Strikes (2009–2016)

Country Number of Drone Strikes Estimated Total Deaths Estimated Civilian Deaths Civilians as Percentage of Total Deaths  
Low High Low High Low High Low High
Pakistan 353 375 1934 3415 129 634 6.67 % 18.57 %  
Yemen 162 184 801 1393 89 170 11.11 % 12.20 %  
Somalia 31 35 264 388 0 12 0 % 3.09 %  
Total 546 594 2999 5196 218 804 7.27 % 15.47 %  


Compared to an estimated civilian casualty rate of 7.27 to 15.47 percent, the Obama administration’s claim of 2.27 to 3.73 percent appears unrealistic, and few drone researchers took it seriously. Obama officials did not explain their methodology, nor break down the data by country, but the large discrepancy likely reflects the government’s incentive to downplay harm to civilians.



When it comes to drone strikes, the U.S. government classifies all “military-aged males” as militants unless posthumous evidence explicitly proves otherwise. Essentially, this assumes that any man (or boy) in the vicinity of a known terrorist who appears to be between the ages of 14 and 60 is also a terrorist. Because official numbers count fewer people as civilians, the numerator (civilian deaths) is lower, resulting in a lower civilian casualty rate.

Trump’s Drone Campaign

Under the Trump administration, the drone program has continued, but its focus has shifted away from Pakistan to Yemen and Somalia. In contrast to claims that Trump substantially increased drone strikes, the data shows consistency across administrations. Here are the casualty figures, using the same non-governmental sources as above:

U.S. Drone Strikes (2017–March 2019)

Country Number of Drone Strikes Estimated Total Deaths Estimated Civilian Deaths Civilians as Percentage of Total Deaths  
Low High Low High Low High Low High
Pakistan 6 13 16 68 0 4 0 % 5.88 %  
Yemen 91 166 213 349 27 55 12.68 % 15.76 %  
Somalia 98 103 664 949 0 30 0 % 3.16 %  
Total 195 282 893 1366 27 89 3.02 % 6.52 %  


The shift away from Pakistan to Yemen and Somalia was already evident in 2016, as the United States conducted only three strikes in Pakistan that year, according to my estimates. Under Trump, American drones fired somewhere between 42 and 77 strikes per year in Yemen, similar to Obama’s final year, with similar civilian casualty rates. Trump’s average number of strikes per year in Somalia is triple 2016, indicating an escalation. This accounts for the lower total civilian casualty rate, since the militants the United States targets in Somalia operate in more easily identifiable training camps away from civilian areas compared to targets in Yemen.

But the escalation was already underway, with Obama authorizing almost as many strikes in Somalia in 2016 as in his first seven years combined. It’s more a change in the location of terrorist and counter-terrorist activity than Trump altering America’s drone strategy.

This contrasts with media reports that Trump is ordering substantially more drone strikes than Obama. For example, the BBC claims that “there have been 2,243 drone strikes in the first two years of the Trump presidency, compared with 1,878 in Mr. Obama’s eight years in office,” citing BIJ data.

Here’s why it’s a mistake: The Bureau of Investigative Journalism started tracking drone strikes in Afghanistan in 2015. As an official war zone, Afghanistan is typically classified separately from the drone campaign, because the U.S. military operates openly and has more strategic options. The vast majority of the 2,243 strikes attributed to Trump took place there.

In August 2017, Trump gave a speech announcing an escalation in Afghanistan. It’s hardly a secret. And it’s unclear whether the increase in drone strikes represents an increase in overall military activity or just greater reliance on drones relative to manned aircraft or ground operations.

An apples-to-apples comparison of drone strikes in Pakistan, Yemen, and Somalia shows more continuity between administrations than disruption. The BBC’s claim is misleading and merits a correction. Worries about Trump’s management of the drone program compared to Obama’s derive, at least in part, from inaccurate information.

Canceling the Report

Citizens of democracies have a right to know when and where their government is conducting violence in their name, so reversing the reporting requirement isn’t a positive development. Though Obama’s stats were imperfect, the required disclosure moved in the direction of transparency.

However, while Trump can reverse a previous president’s executive order, the National Defense Authorization Act — which he cannot overturn unilaterally — requires disclosure to Congress, where the data can inform various oversight committees. The public can still look to New America, the Bureau of Investigative Journalism, and other non-governmental researchers for statistics on drone strikes and civilian deaths. And when the United States makes a tragic error — such as killing 12 people on their way to a wedding in Yemen in 2014 — it makes the news anyway. It’s unrealistic to think Trump could hide them.

Losing the Obama-era report isn’t the big loss for transparency that critics are making it out to be. And yet, there’s little reason for the government not to be more transparent.

Civilian casualties are the primary cost of drone strikes, morally and strategically. Killing suspected terrorists, especially leaders, is the benefit. To evaluate the costs and benefits of the drone campaign — and the broader war on terror — Americans need a full accounting. Though Obama may not have intended it, his executive order revealed the discrepancy between the government and independent civilian casualty numbers. Greater disclosure would move towards a clearer mutual understanding of drone strikes and the somewhat ambiguous designation of “enemy combatant,” help independent researchers identify weaknesses in America’s strategy, and counter inaccurate narratives about differences between Trump and Obama.

While fears that the Trump administration will now kill more civilians in secret are overblown, official numbers would still be useful. At the very least, they tell us something about who the United States believes it’s killing.



Nicholas Grossman is Teaching Assistant Professor of International Relations at the University of Illinois, Senior Editor of Arc Digital, and the author of Drones and Terrorism. Follow him on Twitter @ngrossman81.

Image: U.S. Air Force photo by Senior Airman Cory D. Payne