The effectiveness of drone strikes

Should the U.S. use an expanded campaign of drone strikes in Iraq to help push back the Islamic State in Iraq and Syria (ISIS)? Among the many issues this question raises is the practical one: Are they effective in reducing attacks?

On the surface it would seem that they are, in that they enable the removal of potentially high-value targets with minimal risk to our troops. Looking a little deeper, though, reveals some problems. Many are discussed in a recent report from the Stimson Center, which suggests that heavy use of drone strikes "risks increasing instability and escalating conflicts."

I earlier discussed in this space one of the costs of drone strikes the report mentions: blowback. In a nutshell, the argument I made was that a targeted killing may have a localized effect on the terrorist group, reducing its capability to strike (that's good) while at the same time spurring others to join the group (that's bad). The bad effect can happen even when the targeting is perfect, if the person killed had ties to the community (or the strike itself violated national sovereignty, as the Stimson report notes).

Given inevitable imperfections in targeting despite every effort to the contrary, the likelihood of blowback can be high. The effect of this blowback can easily trump whatever localized reduction in capability the killing yields due to the presence of networks connecting individuals. Both the negative effects of blowback and the positive effects of hearts-and-minds approaches build on each other across networks via a process of positive feedback, enhancing their effects, whereas the reduction in capability does not get so multiplied. Thus, in the presence of significant blowback, drone strikes are unlikely to be effective.

However, even though actual drone strikes may not have a net positive effect in many cases, the threat of drone strikes can still be useful. To see how this can be, we must consider organizational and personnel issues within the terrorist group. Margaret Foster and I discuss some of these issues here. The argument relating to drone strikes goes like this. Terrorist groups can choose to organize in different ways; we'll call these centralized and decentralized. A centralized organization places strict control of sub-commanders in the hands of the group's leaders, and does not devolve a great deal of power to the sub-commanders. This prevents sub-commanders from developing independent bases of power. This is preferred by the leaders, all else equal, for two reasons. One, it reduces threats to the leaders' power from powerful sub-commanders. Two, with reasonable frequency, sub-commanders act out of turn and offend the numerous external backers that many groups rely upon for resources. When these backers come calling for blood, the group's leaders want in many cases to be able to give them what they want, which is much easier to do when the commanders are not independently strong.

However, this centralized structure is very susceptible to drone strikes. Commanders in a decentralized organization are more typically enmeshed in the local community, which will provide aid if called upon. This leads to a higher civilian casualty rate, more blowback and less effective drone strikes, which the government will have less incentive to attempt. In contrast, without such protections commanders will get picked off more easily with less blowback, and government will have more incentive to make frequent use of drone strikes, costing the group operationally.

The threat of drone strikes thus puts terrorist groups between a rock and a hard place. Centralizing entails large operational costs but maintains control and satisfies external backers. Decentralizing minimizes operational costs but reduces control and risks the loss of external backers. The greater the government's remote strike capability, the more likely groups will choose decentralization to avoid drone strikes and in so doing diminish their ability to function in other ways.

So we need the threat of drone strikes (to force groups to decentralize) without many actual strikes (to limit blowback). Maintaining — and making public — a strong remote strike capability while using it only in the surest of cases would help to ensure the credibility of the threat without encouraging detrimental blowback, enhancing the effectiveness of the drone strike program. It would also aid in minimizing issues of sovereignty that arise with actual strikes. Drone strikes can therefore by a useful weapon even when rarely employed.

Siegel is an associate professor of political science at Duke University.