Since al Qaeda bombed the U.S. embassies in Kenya and Tanzania in 1998 and then attacked the World Trade Center three years later, the United States has dedicated billions of dollars and thousands of lives to addressing the threat of terrorism. Over time, policymakers converged on economic development as a key to ending terrorism, in the belief that poorer people are more susceptible to the appeals of violent groups or more likely to perpetrate violence themselves. If economic development aid raised incomes, the thinking went, support for militant groups would diminish.
This logic has taken hold at the highest levels of American policymaking. In 2010, U.S. President Barack Obama argued in favor of sending more development aid to poor countries, because “extremely poor societies” are “optimal breeding grounds for disease, terrorism, and conflict.” The same year, Secretary of State Hillary Clinton concurred, declaring economic development an “integral part of America’s national security policy.”
Yet there is no evidence that economic development changes attitudes toward violent militant groups, or even that it is the poor whose attitudes are problematic. A number of scholars, including Claude Berrebi, Alberto Abadie, and Alan Kreuger and Jitka Malečková, have found that people who join terrorist groups are predominantly from middle-class or wealthy families. Public opinion scholarship, such as that of Najeeb M. Shafiq and Abdulkader Sinno, and Mark Tessler and Michael Robbins, suggests that differences in income and education do not explain variation in support for suicide bombing and other forms of violence. According to Oeindrila Dube and Juan Vargas, job loss appears to correlate with greater violence in Colombia. And Effi Benmelech, Berrebi, and Esteban Klor have found that poor economic conditions enable Palestinian groups to recruit higher-quality operatives. But in another study, Eli Berman explained that regions with higher unemployment in Afghanistan, Iraq, and the Philippines are actually less violent. In Iraq, moreover, there is no evidence that large-scale development programs impact violence, although small-scale programs administered with deep knowledge of the local context do. Even then, the mechanisms that link small-scale aid programs to diminished violence remain unknown.
Closing this gap in understanding about the relationship between poverty and terrorism could not be more pressing. The United States and its allies have already directed billions of dollars in development aid to Pakistan and Afghanistan in the last ten years, with no demonstrable impact on the spread of Islamic militancy. With scarce resources to spend, they should be more careful about how they invest them.
The stakes are particularly high in Pakistan. The country provides haven for Islamist terrorists that operate in India and Afghanistan and is itself the victim of a militant insurgency that has killed or injured some 35,000 Pakistanis since 2004. Currently, programs meant to address the problem of homegrown Pakistani militancy by alleviating poverty dominate the Western aid agenda. The 2009 U.S. Kerry-Lugar-Berman bill, for example, proposed spending $7.5 billion on economic development in Pakistan, with the express aim of “combating militant extremism.” To test the assumption that poor people are more likely to become radicalized, we fielded a 6,000-person, nationally representative survey of Pakistanis in the four provinces of Punjab, Balochistan, Sindh, and Khyber Pakhtunkhwa (formerly North-West Frontier Province) in the spring of 2009.
The survey measured attitudes toward four important militant groups: al Qaeda; the Afghan Taliban; the so-called Kashmiri groups, which include Lashkar-e-Taiba, Jaish-e-Mohammed, Hizbul Mujahideen, among others; and sectarian groups such as Lashkar-e-Jhangvi and Sipah-e-Sahaba. The survey was much larger than any previous effort and, for the first time, included rural Pakistan. Previous studies had been undermined by low response rates, perhaps because they asked Pakistanis directly about their support for militant groups. Instead, we measured attitudes toward the groups using an indirect questioning technique called an “endorsement” experiment. We presented respondents with a set of four policy issues, including World Health Organization’s administration of polio vaccinations and the redefinition of the Durand Line separating Pakistan from Afghanistan, and asked how much they supported each. Some respondents were told that one of the four militant groups supported the policy. Comparing the support for each policy of those who were told a militant group supported the policy with those who were not gives the measure of support for the group.
The data revealed four findings that undermine common wisdom about support for militancy in Pakistan. First, survey participants were generally negatively inclined toward all four militant organizations. Contrary to some popular accounts, Pakistanis do not have a taste for militants. Moreover, they appear to differentiate between groups in subtle ways. Pakistanis were far more likely to believe that the Kashmiri groups provide public goods -- schools, health clinics, and the like -- than they were to associate other organizations with such positive activities. They were also much more likely to say that the Kashmiri groups are fighting for good things, such as justice and democracy than they were about the others.
Second, Pakistanis living in violent parts of the country, in Khyber- Pakhtunkhwa in particular, strongly disliked these groups. This is likely because they pay a disproportionately high price for militant violence, regardless of their views about the groups’ goals. Those from comparatively peaceful areas do not bear the full costs of militant action.
Third, poor Pakistanis nationwide disliked the militant groups about two times more than middle class Pakistanis, who were mildly positive toward the groups. We suspect that this is because much of Pakistan’s militant violence is concentrated in poorer areas and in the bazaars and mosques where less affluent people sell goods, shop, and pray. In addition to being in more physical danger than the rich, the poor are at more of an economic risk from attacks, and income losses are more consequential for them. Wealthier people often have servants run errands to bazaars, and when they do personal shopping they are likelier to do so in upscale stores in their own neighborhoods, which are safer.
Finally, this dislike is strongest among poor urban residents. The negative relationship between poverty and support for militancy is three times stronger in urban Pakistan than in the country as a whole. This finding reinforces the idea that the dislike of the groups is driven by greater exposure to their attacks, which are concentrated in urban areas.
Overall, the findings suggest that arguments tying support for militancy to individuals’ socioeconomic status -- and the policy recommendations that often flow from this assumption -- require substantial revision.
Most governments, including that of the United States, are still reeling from the global recession and looking to make budget cuts where possible. Development assistance aimed at alleviating poverty should not be stopped; countries such as Pakistan have legitimate development needs pertaining to education, health care, and economic growth to support its massive youth bulge. But expecting those programs to reduce militancy is misguided. There are many good reasons to offer development assistance, but counter-radicalization, counterinsurgency, and counterterrorism are not among them.