How deleterious is the impact of air strikes on popular opinion in Afghanistan? | Commentary by Greg Sanders

Flikr photo by  TailspinT used under a Creative Commons license. According to the United Nations Assistance Mission to Afghanistan, coalition air strikes killed 552 civilians in 2008. That figure represents 64 percent of civilian casualties caused by coalition forces and 25 percent of all civilian combat fatalities. Although coalition forces are aware of the problem – and released more restrictive guidelines in August – a November 3 bombing undermined many of the resulting gains.   Such deaths have both a human and a political toll.  At a recent CSIS event, ABC Director of Polling, Gary Langer, described their impact on popular support for the coalition. The ABC/BBC/ARD poll found that local experience with U.S. bombing and artillery fire was strongly correlated with both negative attitudes towards the U.S., towards the Afghan central and provincial governments, and regarding Afghanistan’s direction. This dissatisfaction can lead to radicalization: some 44 percent of Afghans who experienced local bombing believe that attacks against coalition forces could be justified.  Air strikes are not the only factor; they are also highly correlated with perceptions of bad security. To help differentiate between these two factors, I asked Mr. Langer to compare attitudes about bombings among those who believed security was good, and those who thought it was bad. He was kind enough to provide the results:  We know that air strikes have a negative effect on Afghan perceptions. But what is the magnitude of that effect? Perceptions of security also have a deleterious effect and can easily act as a confounding variable. These factors are particularly intertwined because – as Dr. Cordesman emphasized at the event – air strikes in Afghanistan are typically in support of U.S. ground forces. The U.N. Mission report found that “Air-strikes called in by ground forces engaging insurgent fighters represent a particular threat to civilians, as often compounds with an enemy presence also house civilians unable to leave when fighting breaks out.” Unlike in Pakistan, these are not predominantly Predator UAV strikes that occur well away from coalition ground forces. As a key side note, none of the sources mentioned differentiate between low-altitude close-air-support from coalition attack aircraft or helicopter gunships and higher altitude precision strikes from aircraft firing missiles or dropping bombs  Afghans aware of local coalition bombing are consistently more negative about the situation in their country regardless of their opinion of local security. This is particularly true for the central and provincial governments where the percentage decline in positive opinion is roughly the same regardless of the security situation. Of greater interest is the fact that bombing in areas where security is bad seems to generate both the greatest resentment towards the U.S. and pessimism regarding Afghanistan’s direction. Those two popularity ratings decline by at least a third as opposed to situations where security alone is bad. This suggests two likely explanations: 1) Goodwill from Afghans in areas where security is good partially mitigates the  blowback from air strikes. 2) Air strikes in areas with bad security have worse intelligence, cause more  casualties, and thus generate greater blowback. In terms of region, U.S. strikes in areas with poorly rated security are most widely noted in the South of the country. Afghans who rate their security poorly but do not experience bombing are most common in the Central and Eastern regions of the country. That said, while 31 percent of those in the East are in bad security areas and avoid bombing, 42 percent experience both negatives.   When focusing on the differences between the ratings within a single category, air strikes in areas with good security have the greatest impact on the rating of the provincial governments and are equivalent to the penalty for being a poor security area.  The effect on other ratings is more moderate and surprisingly absent in the U.S. case. Since the question addressed national and not local performance this suggests that the 38 percent U.S. approval rating already accounts for feelings regarding bombings. Additional research: • ABC News: 2009 Poll of Afghanistan • UNAMI: Annual Report on Protection of Civilians in Armed Conflict, 2008 (Pg. 16-17) • Full results on security and bombing Special thanks to Gary Langer for providing the requested data runs.