Mike Lewis on the Lack of an IHL Transparency Requirement for Drone Strikes

There Is No Transparency Requirement for Drone Operations

Critics of drones have frequently characterized them as being illegal weapons, or at least weapons that were being used in an illegal manner. Much of the initial opposition was driven by concerns about civilian casualties which led to claims that drones failed to comply with IHL’s requirements of distinction and proportionality. But as Laurie Blank points out here (http://thehill.com/blogs/pundits-blog/defense/207352-drones-transparency-and-legitimacy) critics of drones have shifted the argument away from this more familiar legal terrain and now assert a legal requirement for transparency about the conduct of drone operations that is not found anywhere in IHL. While it must be conceded that transparency for its own sake is not a requirement under IHL, this call for transparency is cast as necessary to verify compliance with the distinction and proportionality requirements that are unquestionably part of IHL. It is argued that when civilian casualty figures vary so widely, how can there be any trust that the drone program complies with these legal requirements without more transparency into how such strikes are conducted. And to the extent that civilian casualty figures have shown an improvement, it is argued that this may be due to a number of other factors unrelated to drone policy and therefore transparency is still required to insure that drone policy is legal (http://www.huffingtonpost.com/2014/05/27/obama-drone-speech_n_5397904.html).

At the same time, the Obama administration has been unwilling, justifiably I believe, to provide transparency into how drone strikes are conducted and how targeting lists are generated. A federal judge has required the administration to declassify memos that provided the legal basis for conducting drone strikes against American citizens abroad. But even when those memos are released there will be many unanswered questions about how/whether drone strike policy has adjusted to reduce civilian casualties and insure IHL compliance. Is there any way around the impasse between the critics who demand true transparency about the drone program and the government that essentially says “trust us, we are compliant”?

One possible solution to this impasse arrived in the form of a massive and detailed data set released by the Bureau of Investigative Journalism (TBIJ) last week (http://www.thebureauinvestigates.com/2014/05/23/get-the-data-what-the-drones-strike/). This data set purportedly provides details of 383 drone strikes conducted by the US in Pakistan between June 2004 and the end of 2013. It gives the date and location of each strike and the type of target (domestic building, other structure, vehicles, etc.). It also provides ranges for the minimum and maximum numbers of killed and wounded in the strikes as well as ranges for the numbers of civilians and children killed in the strikes. Although some of this information was previously available, much of the detail is new. Information provided by this report clearly illustrates some policy changes and strongly hints at others. Granted this is only one source of information. But when it is considered that this data is provided by the TBIJ, which strongly opposes drones and is among their most vocal critics, here is what it tells us about the drone program in Pakistan.

The varying frequency of strikes across administrations has been well documented and frequently discussed. There were 51 drone strikes in Pakistan during the Bush administration, 45 of which occurred in the last two years of his presidency. According to the data set those strikes killed 167 civilians, over a third of all civilian casualties reported by the study. In fact a single drone strike against a madrassa in Chingai in 2006, before the drone program really got underway, was reportedly responsible for over 80 civilian deaths, almost 20% of all civilian casualties reported in the ten year study. It should be noted that there remain conflicting reports of Pakistani participation in the Chingai strike and some eyewitness statements that helicopter gunships took part in the attack.

The first year of the Obama administration saw 52 strikes that killed 100 civilians. Because of the statistical distortion caused by the Chingai attack, the first year of the Obama administration showed a marked improvement in the percentage of civilians killed versus insurgents over the Bush years. With Chingai removed however, the numbers look very similar (approximately 25% of those killed by drone strikes were civilians), strongly supporting the idea that the procedures being used for drone attacks remained roughly the same through the change in administration. But that continuity didn’t last and this is where the new information from the data set is helpful.

In 2009 Generals Petraeus and McChrystal emphasized the need to take extraordinary measures to reduce civilian casualties in Afghanistan. To that end, Gen. McChrystal formed a mixed military/civilian team tasked with conducting root cause analysis on every civilian casualty in Afghanistan and developing procedures to prevent such casualties in the future. This team began providing recommendations to change protocols for all types of activity from night raids to close air support to drone strikes, in 2010. And the data set pinpoints for us when at least some of those recommendations were implemented.

As of September 14, 2010 there had been a total of 174 drone strikes in Pakistan. Of those, only 14 (~8%) had targeted vehicles. Over the next 18 months there were another 145 drones strikes of which at least 64 (~44%) targeted vehicles. This indicates a clear policy change to target vehicles with much greater frequency for the purposes of reducing civilian casualties. The vehicle strikes killed 249 people of whom only 19 were civilians. This shift to vehicle strikes reduced the civilian casualty percentage from ~25% to less than 10%.

A second set of policy changes can be inferred from the data set. Even with the increased emphasis on vehicle strikes during 2011 overall civilian casualties from drone strikes were still at roughly 14%. This represented a marked improvement from 2009, but was still far from perfect. In 2012 that number decreased to less than 7% even though there were still 50 drone strikes that year which killed 200 people overall. In 2013 the minimum number of civilians killed as assessed by TBIJ reached 0 (the maximum is assessed at 4). This is out of over 100 people killed (nearly 200 is the maximum assessment) in over 25 strikes. The data set does not give us any further indications what other changes have occurred in targeting, intelligence analysis or strike procedures, but the continued sharp decreases in civilian casualty rates are not an accident.

Like the shift in emphasis to vehicle strikes, other policy changes have continued to greatly decrease civilian casualties caused by drones, to the point that no one can seriously question whether they comply with the principles of distinction and proportionality. And if there can be no question that drone strikes comply with the jus in bello requirements of IHL, then there can be no legal basis for requiring transparency into drone operations.