A nurse weighs a child at a U.S.-funded clinic on Sept. 7, 2011 in Farza, Afghanistan. (John Moore/Getty Images)

Over the past decade, U.S. military expeditions have increasingly used development aid to undermine popular support for insurgents and extremist groups, and pacify turbulent areas. But does this tool work in conflict zones like Afghanistan?

My new research published in the American Political Science Review provides evidence that aid spending by the U.S. military in Afghanistan in many cases produced violent backlash from insurgents against troops and civilians.

Aid distributed in districts under the control of coalition forces reduced violence. But aid distributed in contested districts of Afghanistan actually increased the level of violence carried out by insurgents against both civilians and the U.S. military.

To test how aid projects affect insurgent violence, I examined 5,936 projects in Afghanistan during 2008-2010 under the Commander’s Emergency Response Program (CERP), which provided development aid funds to U.S. military commanders to use in their local area of operations. I also looked at more than 46,000 security incidents reported by an independent nongovernmental organization network and 30,000 incidents reported by the U.S. military.

From 2012 to 2013, I spent seven months in Afghanistan, carrying out dozens of interviews about the patterns of violence in the country.

The U.S. military has aimed to win hearts and minds in the past

In past conflicts, notably during the Vietnam War, the U.S. military carried out nonmilitary activities aimed at winning the support of an occupied population, such as training self-defense groups or offering medical care. In the early 2000s, this tactic was extended to development aid projects.

A decade before Gen. James Mattis became Donald Trump’s pick for defense secretary, he huddled with Gen. David Petraeus at the U.S. Army Combined Arms Center at Fort Leavenworth, Kan., to hammer out a new counterinsurgency (COIN) doctrine for the U.S. military.

By 2006, U.S. interventions had produced nominal victories against the Taliban in Afghanistan and against Saddam Hussein’s regime in Iraq. But the strongest military on the planet found itself mired in extended nation-building projects — amid rising levels of armed insurgency.

In response, Petraeus, Mattis and their staffs developed Field Manual 3-24, which formalized a COIN approach that emphasized winning the hearts and minds of civilians, rather than simply defeating enemies on the battlefield. CERP was an important part of this new approach.

Here’s the logic: Local communities are uncertain whether to cooperate with coalition forces and the new government. Aid projects help to win their trust — and build up a reciprocal relationship that boosts the legitimacy of NATO coalition forces and the host government.

With this theory of change in mind, the U.S. Congress appropriated several billion dollars to CERP projects, including $2.6 billion in Afghanistan from 2004 to 2011.

But an armed opposition typically recognizes that aid projects stand to benefit pro-government forces. This leaves insurgents with a strong incentive to sabotage these initiatives and punish civilians who may be collaborating. This tends to be especially true in contested areas — where coalition forces do not have full military control — because foreign forces are stretched further and the population is more likely to be pliable.

Here’s the new evidence of backlash

With much of the funds directed to contested areas, the net result in Afghanistan was that areas that received CERP projects saw a significant increase in violence, at least in the short term. I estimate that the close to $100 million in completed projects spent in contested districts during 2008-2010 resulted in 33 additional bombings against civilians and troops, 20 additional insurgent actions against U.S. troops and 57 additional improvised explosive device (IED) incidents against U.S. forces.

In contrast, aid-related spending in controlled areas probably prevented six civilian bombings, 18 enemy actions and 18 IED events from 2008-2010.

Congress has approved more than $5 billion in CERP funds to date in Iraq, Afghanistan and the Philippines, with minimal oversight and unclear strategic guidance for operators. In Afghanistan, the CERP program peak of $1 billion in a single year corresponded with the start of President Obama’s “civilian surge” in 2010. Thousands of CERP projects were completed during the decade, including canal digging, humanitarian aid and security infrastructure in government buildings.

Commanders weren’t given clear guidance or held accountable for their aid spending

Interviews with commanders in the field have confirmed that these projects were undertaken with little guidance or oversight. Commanders were encouraged by superiors and incentivized through promotion grading to spend as much money as they could, with little to no ability to measure success. Military operators, reflecting on CERP, reported that in general they viewed this funding as a potentially useful tool, but one that has not yet been harnessed well.

The bottom line from my evaluation is that “hearts and minds spending” is indeed capable of reducing violence against both civilians and U.S. forces — but only when it targets areas already under the control of friendly military forces.

Otherwise, these projects are more likely to generate increased levels of violence by insurgents. My research findings also detail how aid projects can easily be swamped by large-scale strategic shifts, like changes in supply routes.

The architects of the COIN strategy, including Mattis, Petraeus and Lt. Gen. Michael Flynn, are poised for a return to political power in the incoming administration. My findings suggest that some important revisions to COIN doctrine would be warranted.

My research also suggests that counterinsurgency doctrine needs to be updated to reflect a better understanding of how aid can be deployed in a stabilization mission. This includes taking a closer look at top-level documents such as FM 3-24 and congressional authorizations, but also issuing practical guidance for commanders in the field.

The evidence indicates that aid can be an effective tool for stabilization and force protection when pro-government control is already established. In other areas, it may instead backfire.

Renard Sexton is a doctoral candidate in the politics department at New York University.