Proxy Variables and Drone Effectiveness

This is the third part of my discussion of President Obama’s speech defending the use of drone attacks, given last Thursday (a transcript of the whole speech is here).  I was interrupted by other interesting posts that I wanted to write.  Having discussed the moral and legal dimensions of the argument in favor of drones, now I would like to raise an issue regarding the evaluation of drone strikes’ effectiveness.

President Obama highlighted the effectiveness of the US drone attacks on al-Qa’ida and insurgents in Afghanistan as a strong argument in favor of using them (although he himself noted, “To say a military tactic is legal, or even effective, is not to say it is wise or moral in every instance”).  His summary of their effectiveness is worth quoting in full:

To begin with, our actions are effective. Don’t take my word for it. In the intelligence gathered at bin Laden’s compound, we found that he wrote, “we could lose the reserves to the enemy’s air strikes. We cannot fight air strikes with explosives.” Other communications from al Qaeda operatives confirm this as well. Dozens of highly skilled al Qaeda commanders, trainers, bomb makers, and operatives have been taken off the battlefield. Plots have been disrupted that would have targeted international aviation, U.S. transit systems, European cities and our troops in Afghanistan. Simply put, these strikes have saved lives.

That last sentence is worth repeating: “Simply put, these strikes have saved lives.”  Earlier he had put it, “In sum, we are safer because of our efforts.”

On one level, it would be foolish to dispute the efficacy of drone strikes.  Just today Pakistani sources reported that a US drone strike killed the Taliban’s #2 commander.  As President Obama euphemistically put it, “Dozens of [al-Qa’ida terrorists] have been taken off the battlefield.”  President Obama argued that sending in soldiers (“putting boots on the ground”) would kill more civilians and make more enemies, so on this logic the drones are positively effective with less risk than conventional military means.  They were designed to enable “targeted killing” and they were designed well for that goal.

But on another level, the effects of drone strikes after the initial death are not so clear.  The effects of the strike do not end there, but play out among the people (both terrorists and non-terrorists) who knew the people killed.  The effects continue to play out among the people who live near the spot of the attack, or among people who could imagine a drone flying over their own heads.  Even when the person(s) killed were the intended targets (usually the case), is it true that “these strikes have saved lives” and “we are safer” because of drone strikes?

The short answer is that we cannot know for sure.  Safety in the abstract is not measurable, and the number of lives that would have been lost if another choice had been taken is not knowable.  When social scientists and policy analysts want to measure something that is not conveniently quantifiable or not accessible to measurement, they often identify a “proxy variable” that they can measure, and use the proxy variable as a stand-in for the unmeasurable end.  This is the best way it is done, but it is important to remember that the proxy variable is only a proxy, decided by fiat to be related to some unmeasurable or inaccessible quantity.  Failing to remember the distinction between the proxy and the real leads to bad reasoning and blind spots in the analysis.

In this case, President Obama seems to reason implicitly that since al-Qa’ida forces are trying to kill civilians, killing them first before they get to where they can kill civilians results in fewer civilian deaths.  Thus the number of al-Qa’ida operatives killed by drones becomes a proxy variable for increased safety and saved lives achieved by drones.  It sounds plausible, it makes sense, but is it true?

One can think of other proxy variables which are a bit closer to the desired goal of “safety.”  The number of terrorist attacks attempted and/or committed in a given time frame within the US, for example.  Has this gone down?  I’m not so sure.  (Someone with more patience or more time than I have could comb through Wikipedia’s List of Terrorist Incidents and count incidents per year in any given geography.)  Other factors, less accessible or less quantifiable, certainly come into an assessment of increased or decreased safety.  More accurate than the number of al-Qa’ida operatives killed might be the total number of al-Qa’ida operatives at a given time, although this is inaccessible because they are unlikely to tell us if we called them up to ask.  But drone strikes do not necessarily result in a decrease of al-Qa’ida membership, as the retired Air Force general acknowledged in saying that drone strikes can contribute to “creating a recruiting poster for Al Qaeda.”  There are also the intangibles such as the level of cultural repugnance against terrorist violence among societies from which al-Qa’ida operatives are recruited and the level of anti-American sentiment which al-Qa’ida, can leverage for recruiting or material support.

And the uniqueness of the US drone strikes program has put the US in a vulnerable position to erroneous reports of drone attacks.  For example, yesterday in Somalia al-Shabab, the local branch of al-Qa’ida, claimed to have shot down a US drone.  This is no doubt a tactic to recruit additional forces, and regardless of the veracity of the claim, it will work.  As a potential example, since Iran has drones, all Tehran would need to do in order to stoke anti-American sentiment in Pakistan or Afghanistan would be to send a few drones and a few rockets over its borders eastward, and the US would be blamed for any resulting deaths.  This is a vulnerability resulting from the US use of drone attacks.

Have drone strikes and “targeted killing” saved lives and increased safety?  I’m not saying they have not.  But the proxy variable needs to be recognized for what it is.  All the evidence cited by President Obama for the effectiveness of drones demonstrates their military effectiveness.  I’m just not sure that military effectiveness is a useful proxy for the goal of internal peace and civil safety.  Rather more complex questions need to be asked before the affirmation of drone strikes’ effectiveness in attain those goals is justified.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s