Afghan youths look over the city of Kabul on Jan. 30, 2018. (Rahmat Gul/AP)

Budget crunchers are looking at all aspects of the fiscal 2019 White House budget, including how foreign assistance helps support U.S. security interests. Some of the $16.8 billion in the latest U.S. Agency for International Development budget, for instance, targets the agency’s  No. 1 programming issue: reducing conflict.

But what drives support for combatants in wartime? Governments, militaries and aid agencies have implemented dozens of economic interventions designed to win “hearts and minds” and reduce violence in settings as diverse as Iraq, Yemen, Pakistan, Nigeria and Somalia. These interventions include livelihood training, employment programs, cash-for-work opportunities and, increasingly, unconditional cash transfers to specific populations.

The premise here is that poverty leaves communities and, in particular, youths vulnerable to insurgent recruitment. Employment and improved economic prospects, in theory, will help shift support away from insurgents, in favor of the government. And these programs raise the opportunity costs for participating in armed rebellion, which makes it harder for insurgents to recruit.

The United States has spent billions of dollars in Afghanistan and Iraq alone in the belief that these interventions effectively persuade individuals to resist insurgent appeals and support counterinsurgency efforts. Yet rigorous empirical evidence demonstrating the efficacy of these programs is scarce. What evidence does exist is mixed, at best.

Working with Kosuke Imai and Yang-Yang Zhou of Princeton University, we took up the challenge of experimentally evaluating the effectiveness of livelihood training and cash transfers to a vulnerable population in Kandahar, Afghanistan. Specifically, we examined the U.S.-funded Introducing New Vocational Education and Skills Training (INVEST) program. While the Mercy Corps-designed and -administered INVEST program did not set out explicitly to reduce violence, it provided a unique opportunity to look at the causal links between economic interventions and combatant support.

In the program, we randomly assigned 2,597 youths to receive three to six months of vocational training; a $75 cash transfer via cellphone; both; or none. We then measured changes in economic outcomes and support for the government and Taliban about two weeks after the conclusion of the program, and then again seven to nine months later. Here’s what we found:

1) Did the program improve economic prospects? Vocational training had a modest effect on an individual’s economic conditions, we found. By the seven- to nine-month mark, those with vocational training reported more cash earned and more days worked, compared with those without training. But cash transfers did little to increase economic activity. Those who received cash alongside the vocational training saw no additional effects of the cash beyond what the training provided when we checked back seven to nine months later.

2) What is the impact on support for combatants? Opportunity cost theory suggests that if people found their economic well-being improved, they would be less supportive of political violence. But we found surprisingly little connection between economic outcomes and changes in support for either the Afghan government or the Taliban militants. Vocational training also had no effect on relative support for these combatants, either immediately after the program or seven to nine months later.

This is a sobering finding, as these individuals benefited the most from the INVEST program. But their economic gains did not translate into greater pro-government sentiment. Cash transfers alone created a “boom/bust” cycle: Government surged after the cash was received, but then quickly faded and reversed itself after seven months. Indeed, individuals who had received cash were more supportive of the Taliban than those who did not receive cash by the second survey seven to nine months later. Moreover, these individuals were the most likely to express anger at their inability to find jobs and to believe that violence against the state was legitimate.

Here’s the good news: The group that received both vocational training and cash showed a relatively large 16.7 percent decrease in its willingness to engage in pro-Taliban actions by the seven-to-nine-month mark. Compared with the initial survey results, this group also recorded a large 20 percentage-point drop in their belief that use of violence against the state is legitimate; an increased belief in the performance of both the national and local government; and an increased belief that the government was responsive to the group’s needs.

3) What’s the reason for this pro-government effect among vocational training and cash recipients, but not the others? We think high-profile programs like INVEST use cash to signal that the government is being responsive to people’s immediate needs. From survey data, we saw that individuals spent their money on basic needs, such as housing and food. However, without additional support needed to turn the cash into a longer-term benefit, the effects don’t last.

That’s why the vocational training is important: It shows the government is also working to support people in the longer term. A lone signal generated from a single economic intervention may have little impact. Two signals, however — such as the vocational training combined with cash — may indicate to people that the government cares about their welfare and is taking steps to direct benefits their way. In this light, economic interventions become windows of opportunity for officials to update citizen beliefs about government responsiveness.

What do these findings tell us about what works? 

This is just one experiment, and we recognize that it may have its limitations. But we think policies that pair short- and long-term interventions may be the best option for shifting support for combatants. Providing young people with more discretionary money — the cash transfers — appeared to provide a short-term financial boost that helped them realize the potential of the longer-term vocational training intervention. The result was lower support for political violence.

Short-term interventions are fleeting; long-term interventions do not provide people the immediate benefits to address their current needs. Combining both signals an understanding of people’s needs. We know that people become involved in violence for both short-term (i.e., economic) and longer-term (i.e., ideology and grievances) reasons. These multidimensional interventions help address a range of reasons that individuals throw their support behind combatants in wartime.

Jason Lyall is associate professor of political science and director of the Political Violence FieldLab at Yale University. Follow him @jaylyall_red5.

Rebecca Wolfe is the director of evidence and influence at Mercy Corps, a global humanitarian and development organization, and a visiting scholar at the MacMillan Center for International and Area Studies and the Political Violence FieldLab at Yale University. Follow her @rebeccajwolfe.