A friend of mine, Sarah Holewinski, recently wrote about how the United States has not made progress on protecting civilians when conducting military operations abroad. It is a compelling and disheartening list. As a scientist who has visited war zones and pored through reams of military data to find creative ways to better spare civilians, the list rings all too true.

To me, the next step is asking the question: Why? To solve a problem, we first need to understand it. Fortunately, steps have been taken to do just this. I have lead or contributed to a dozen studies on civilian protection commissioned by the U.S. Defense Department and others, including analysis of real-world civilian casualty incidents. This work has proved useful over the last decade: It enabled a reduction of civilian casualties in Afghanistan starting in 2009, informed the 2016 U.S. executive order on civilian casualties, and influenced the United Nation’s Protection of Civilians program to include policy and practice as well as compliance with international law. If we want to understand the overall lack of progress that Sarah points out, this foundation of study helps us to see the problems and the options with greater clarity.

The lack of progress on civilian protection affects more than U.S. military operations. It also impacts the conduct of U.S. partners—exemplified by the tragic and extreme toll on civilians from the U.S.-supported, Saudi-led coalition in Yemen, but repeated elsewhere with other U.S. partners. And a lack of U.S. leadership exacerbates the global problem of civilian protection. In May 2020, U.N. Secretary General António Guterres described the past year as a “year of suffering” for civilians in armed conflict and called on all States to redouble efforts to protect civilians.

U.S. government officials have spoken in the past about doing “everything possible” to protect civilians in conflict, but actions in the past have fallen far short of that aspiration. Building on my research—reading one operational report after another, studying patterns and trends in conflict after conflict—and experience working with the United States and its partners to seek improvements, I offer three reasons for the U.S. lack of progress: We don’t learn, we don’t lead, and we don’t help our partners—or hold them to a high enough standard.

We Don’t Learn

The United States has done more than any country in the world to understand how civilian harm happens. When I started my work on civilian harm in 2009, I had the privilege of working with U.S. and international forces in Afghanistan under the command of Gen. Stanley McChrystal. That experience highlighted for me what was possible. Due to McChrystal’s leadership and guided by evidence-based solutions, the civilian toll dropped by 20 percent from 2009 to 2010, showing that it is indeed possible to better protect civilians. Data also showed that these efforts did not reduce mission effectiveness nor did they endanger U.S. forces. I have continued to work on the issue of civilian protection since then, studying other operations and identifying opportunities for doing better.

Having worked on this issue for many years, I have examined well over 1,000 actual incidents of civilian harm using operational records and investigations, then analyzing patterns and trends. As a result, there is much information available about civilian harm and how it can be reduced. For example, in my work I found:

  • Not all civilian harm results from “collateral damage” where civilians are inadvertently harmed in an attack of a military target. Instead, civilian harm often occurs when civilians are misidentified as military targets themselves and are attacked in that mistaken belief;
  • There are many misconceptions about civilian harm that do not hold up to the data. For example, drones aren’t necessarily as surgical as many believe, with one study showing drones being ten times more likely to cause civilian harm than manned aircraft;
  • Available weapons, equipment, and capabilities are often ill-suited for avoiding civilian casualties, including a lack of robust tools to detect civilians and civilian casualties inside buildings, little ability to identify and deconflict damage to infrastructure, and consistent challenges in determining hostile intent. For example, using green laser dazzlers to signal to drivers to stop at checkpoints is a bad idea (since green means go), but that’s what forces are often given;
  • We often don’t know the true scale of civilian harm: the assessment process improperly relies heavily on internal military intelligence and does not give enough weight to external sources;
  • Effective command guidance matters. For example, in Afghanistan, the average number of civilians killed in errant airstrikes dropped over 40 percent after the 2010 Tactical Directive amended guidance to forces concerning potential civilian casualty risks. On the other hand, command guidance discouraging warning shots at checkpoints increased the risk to civilians because forces instead resorted to more deadly disabling or lethal fire.
  • Civilian protection is not just tactics but taking a comprehensive approach that starts with mission planning and encompasses strategic, operational, and tactical levels of war.

Overall, we have a much greater understanding of challenges and ways we can improve. But to improve, we need to put that understanding into action. This can be done in two ways: operational learning and institutional learning. The 2010 Tactical Directive is one example of operational learning, finding what needs to be improved within the context of an operation and addressing it. When this operational learning is integrated into the conduct of operations, it can be effective and rapid. For example, during the Afghanistan campaign between 2009-2012, the military command would regularly send me their data on civilian casualties so I could help monitor and give suggestions from back in Virginia. In January 2011, I noticed new spikes in civilian casualties in two areas of operations. I did some quick analysis to isolate what they were from, developed some recommendations, and sent them to an aide of the then-commander, Gen. David Petraeus. The command verified the finding and took quick action to mitigate the risks, and sure enough, the numbers went back down in a matter of months. This idea of monitoring trends and looking for quick ways to identify and respond to risks was something I later wrote into section 4 of the 2016 civilian casualty executive order.

While that idea is now national policy—Section 4 of the executive order is still in effect today—that policy commitment to operational learning has not been met in practice. For the campaigns in Iraq, in Syria, and in Afghanistan since 2015, the risk of civilian casualties increased over time with no mitigating actions being taken, and the calculation of this risk was only done after operations concluded. The operational learning seen earlier in Afghanistan—real-time intervention to address risks to civilians—was not repeated in later operations.

The United States also struggles with institutional learning. The various studies and reviews over the past decade have produced many recommendations, recommendations that largely have not been acted on. The result is that the same mistakes tend to occur over and over again. Because the United States has a poor institutional memory, this very fact of repeated mistakes is often not even recognized. For example, in the 2015 U.S. strike on an MSF hospital in Afghanistan, the investigation and DOD follow-on actions did not recognize that many of the features seen in that incident had occurred many times before, and thus missed an opportunity to address those long-standing problems. And while the evidence is clear that a comprehensive approach is needed for civilian protection, the United States has never taken such an approach. This is exemplified in a tragic way by operations in Raqqa, Syria, where some efforts were made to support the protection of civilians, mostly at the tactical level, but the lack of shaping and monitoring of risks led to an approach that left the city in ruins and caused human suffering and tragedy that will last far longer than the conflict itself.

Institutional memory is critical for progress. When Thomas Edison invented the light bulb, he tried thousands of materials for the filament before finding the right solution. What if Edison had not meticulously kept a record of the materials he tried? Without learning from his mistakes, it is very unlikely he would have been successful. But the U.S. military lacks such institutional memory, such systematic tracking of lessons, of successes and failures with regard to civilian protection. Without that, progress is much more difficult. And it is telling that this situation is allowed to persist, suggesting that civilian protection is not a true priority for DoD. Consider if this were the situation for aircraft carrier performance or fighter aircraft effectiveness: Would this persistent lack of progress be tolerated?

We Don’t Lead 

The United States understands the challenges of reducing and assessing civilian casualties better than any military on the planet and has developed a number of evidence-based best practices, though its overall practice has been uneven. As the international community continues to wrestle with how to reduce the civilian toll of military operations—for example, in initiatives such as the U.N. Secretary General’s Protection of Civilians initiative and Ireland’s political declaration on explosive weapons in populated areas—the United States has much to offer in terms of both conceptual and concrete approaches.

While the United States has introduced into these discussions the notion of sharing good military practices and practical measures that States can take to protect civilians during conflict, the United States can take a stronger and more forward-leaning stance internationally. For example, after the issuance of the 2016 Executive Order on civilian casualties, Sarah and I proposed to have the United States lead the development of a joint declaration on civilian casualties, getting other countries to agree to similar policy commitments and then having signatory countries work together to share best practices. Then-Deputy Secretary of State Antony Blinken and International Committee of the Red Cross President Peter Maurer both agreed to speak at a launch event for the new joint declaration at the U.N. General Assembly in September 2016. But the effort was scuttled when legal counsel in DoD and State cautioned that doing so could risk the creation of new norms that the United States would then be beholden to.

While it is right to be aware of possible effects of such initiatives, the risk—which to me still seems unclear since the effort was modeled on policy commitments that the United States had already made—should be weighed relative to the benefits. And the benefits of U.S. leadership in the overall area of civilian protection are many. U.S. leadership can promote humanitarian goals: greater involvement and leadership in these venues can help the international community to meet its aspirations to better protect civilians. It is also smart strategy: such efforts enhance the legitimacy and reputation of the United States while highlighting the differences between the United States and competitors such as Russia and China.

What are some specific areas in which the United States can show leadership? Such leadership could include:

  • strengthening transparency initiatives regarding civilian harm, setting a standard for sharing of victim demographics, locations, and whether amends were made;
  • working constructively with international groups and civil society, and encouraging other militaries to do the same;
  • developing practical and technical solutions to persistent problems such as protection of critical infrastructure and hospitals that can be used by itself and by partners; and
  • helping to develop implementation strategies for initiatives such as the U.N.’s POC program that are appropriate to different partners and military contexts.

We Don’t Help Our Partners—Or Hold Them to a High Standard

The United States works with foreign nations to meet national security, foreign policy, and economic objectives. Strengthening a foreign country’s security forces can carry risks, and over time security assistance laws and policies have aimed to mitigate them. Specifically, foreign military sales carry a requirement for end use monitoring, where the United States seeks to ensure that military equipment is safeguarded from potential espionage and not transferred to third parties without authorization. Similarly, legislation in 1997 restricted assistance to foreign security forces when evidence was found of gross violations of human rights.

Protecting against diversion and preventing assistance to security forces that do not hold to basic international standards of human rights are necessary but not sufficient in themselves to improve outcomes for foreign partners, to promote U.S. national interests, and to have security assistance efforts align with U.S. values. Recent operations by the Saudi-led coalition in Yemen highlight the negative outcomes that can result from security assistance. The United States has provided equipment, training, and operational support to the coalition, including with existing human rights and end-use monitoring requirements, and yet the effects of this campaign have been disastrous. From my report detailing U.S. support to the Saudi-led coalition:

Saudi-led coalition airstrikes have killed thousands of civilians while crippling hospitals and critical infrastructure, contributing to a humanitarian crisis in Yemen of a magnitude not seen since World War II, including widespread famine and a cholera epidemic. Meanwhile, in the midst of this crisis, the terrorist group Al-Qaeda in the Arabian Peninsula (AQAP) has solidified its grasp on territory and resources. With civilian casualties from US-provided bombs generating increasing hostility in Yemen toward the US, the conflict has also increased AQAP’s ability to act against the US and its interests.

This is just one example of many where partner forces receiving U.S. training and equipment harm civilians during the course of their military operations. There are compelling reasons to do more to promote the protection of civilians with partners. In addition to the suffering of the civilian population, such harm undermines the partner’s strategic success while creating grievances that can prolong conflict. Civilian harm also serves as a symptom of targeting deficiencies where legitimate threats go unaddressed. This harm is also costly to the United States, causing reputational damage and strengthening anti-U.S. sentiment, weakening norms and expectations for the protection of civilians in armed conflict, and creating resistance to efforts to support foreign partners.

What more can we do? There are a number of steps we can take to strengthen the civilian protection ability of partners and promote a more responsible security assistance policy. They include:

  • Take a comprehensive approach. Right now, the United States addresses only some components of civilian protection and treats them in isolation, instead of taking a comprehensive approach that includes strategic to tactical levels;
  • Stop making decisions in the dark. Too often, decisions about providing arms or assistance are made without a deep, contextualized understanding of the civilian protection challenges that are present. This includes “operational” end use monitoring—not current inspections intended to prevent diversion (referred to as “end use monitoring”) but a new requirement to assess the operational impact of assistance, including the impact on civilians. Such an understanding is needed to both characterize risks and to identify evidence-based solutions that can help mitigate these risks;
  • Use focused conditionality. Are there significant risks with U.S.-provided assistance? We can identify mitigating steps that can be taken to address those risks, and then introduce conditionality to help convince partners of the importance of those steps being prioritized. This is far more productive than a message from the United States saying “try harder” to protect civilians but lacking specificity on what actions should be taken;
  • Push for partners to sign onto international agreements (for example, the Explosive Weapons in Populated Areas political declaration, or a new version of a joint declaration on civilian casualties) in order to raise the collective bar on civilian protection.

We can also work to cultivate regional leaders in civilian protection. Such prospective regional leaders would combine professional military operations with a political will to improve and be an example to others. Cultivating regional civilian protection leaders has a number of advantages: It helps expand the influence of the United States, it gives the issue of civilian protection a local voice in the region, and over time it will reduce the required level of effort for the United States in this area.


Looking back at the last four years, we see an administration that, too often, did not prioritize the issue of civilian protection in practice. This includes the acceptance of a sharp increase in the risk to civilians in recent operations in Iraq, Syria, and Afghanistan; the relaxation of standards for counterterrorism operations; and the certification of the sufficiency of measures to safeguard Saudi arms sales. These and other actions all increased the risks to civilians in armed conflict, and innocent lives were lost as a result.

There are already signs that the new administration will seek to improve the protection of civilians in conflict. But the United States should do more than just reverse the actions and inactions of the past four years. We understand the problems we face, with the foundation of a dozen studies and many specific, actionable recommendations. We need to act based on that understanding. Too many times I have seen yet another airstrike kill innocent civilians, and am haunted by the fact that it happened again. Mistakes happen in war, but tolerating repetition of the same mistakes is careless and callous.

So, how do we do better? How do we start to make progress on Sarah’s list? We need a few things: strong leadership by informed and determined leaders, resourcing of efforts to address identified lessons, and establishing clear responsibilities for ensuring that institutional changes are made. These responsibilities need to be clearly established, properly resourced, and sustained over time. We have the foundation to both improve civilian protection in U.S. operations and contribute to improving the protection of civilians in conflict globally. Let’s hope that in four years we can look back and see progress. If we do, lives will be saved.

All views expressed are the author’s’ own. 

Image: A picture taken on October 21, 2017 shows a general view of heavily damaged buildings in Raqa, after a Kurdish-led force expelled the Islamic State group from the northern Syrian city. Photo: BULENT KILIC/AFP via Getty Images