Following the October 7 atrocities committed by Hamas on the people of Israel, the Israeli Defense Forces (IDF) response in Gaza has drawn criticism for the resulting civilian death toll. The conflict has now claimed more than 30,000 Palestinian lives, with some estimates putting the daily death toll as exceeding that of any other major conflict of recent years. This is in contrast to statements from individuals such as Ron Dermer, Israel’s Minister of Strategic Affairs, who has described Israel’s efforts to minimize civilian casualties as effective and “unprecedented.” Some commentators agree with Dermer. For example, John Spencer, an expert on urban warfare, has claimed that the IDF “has implemented more measures to prevent civilian casualties than any other nation in history” and is “the gold standard.”
As an analyst who has helped define the field of civilian harm mitigation, worked with militaries to implement it, and helped develop U.S. policies on civilian harm as a senior advisor to the State Department, I find this statement misguided. Yes, the IDF takes a number of steps designed to protect civilians, for example, the practices of roof knocking and warning calls and texts to residents. But the gold standard for civilian harm mitigation is not a checklist of steps but rather an iterative process to learn and adapt. Israel has yet to demonstrate that it has embraced this process. More importantly, the data–not just the staggering death toll, but key attributes of the campaign–suggest Israel’s steps are not working.
To be sure, protecting civilians in Gaza is no easy task. Conducting military operations in the densely populated urban areas of Gaza while avoiding civilian harm poses immense challenges. Hamas hides within the civilian population and uses subterranean tunnels to move weapons and fighters. U.S. and coalition operations over the past twenty years show that even the most professional militaries, committed to compliance with international law, face real challenges mitigating civilian harm in conflict in similar or analogous urban settings, including in Iraq, Syria, and Afghanistan.
But after more than four months of fighting, it is important to closely examine the question, “how is the IDF doing and what can it do better?” Civilian harm mitigation is an emerging expectation in modern warfare, evidenced in part by the widespread commitment to declarations like the Political Declaration on the Use of Explosive Weapons in Populated Areas. It is also a critical question for countries providing military support to Israel to inform the nature and level of their continued support. This question should be of key interest to the Israeli government itself as public support domestically and internationally wanes due to the scale of civilian harm.
What is Civilian Harm Mitigation?
Civilian harm mitigation is a relatively recent development, and there is often confusion about what it means. The concept emerged from U.S. military operations over the past two decades, but even the United States is still in the process of institutionalizing these lessons. Among its foundations is the idea that modern militaries consider the protection of civilians in their efforts to comply with international humanitarian law such as the Geneva Conventions and their Protocols. This includes, for example, a process for estimating collateral damage, selecting weapons that will be effective against the target but reduce damage to other nonmilitary structures, weaponeering processes designed to find the optimal angle and direction of approach for an attack, and developing “No-strike lists” for protected and sensitive sites.
Systematic U.S. civilian harm mitigation efforts, for example, were introduced in Iraq to address chronic problems at checkpoints leading to the deaths of civilians. It involved tracking incidents and learning lessons from them. As forces at checkpoints learned to guard against attacks while reducing the risk to civilians, fewer civilians were killed.
A more mature and comprehensive process was developed in U.S. and international operations in Afghanistan as a response to mounting civilian harm. Starting in 2006, Afghan leaders and the international community expressed alarm over significant numbers of U.S.-caused civilian casualties in Afghanistan, and U.S. leaders saw the issue of civilian harm becoming a strategic problem detracting from the military campaign.
This realization drove a different, practical approach toward civilian harm. In late 2008, U.S. and multinational forces began tracking data about each potential and actual incident involving civilian casualties. The original intent of this tracking was to counter allegations of civilian harm. But the data eventually served another purpose. In 2009, CNA (a not-for-profit research organization) began analyses of civilian harm for the U.S. Department of Defense (DOD) using operational records, combined with information from non-governmental organizations, to discover patterns of civilian harm. This analysis yielded new insights that enabled practical, focused steps that U.S. and international forces could take to reduce the risk to civilians in operations in Afghanistan. These steps were codified in a revised International Security Assistance Force Tactical Directive, new guidance for escalation of force, and many other measures. These analytically grounded changes reduced the rate of civilian casualties per operation. Overall, civilian casualties fell by approximately 20 percent in the first year, with additional reductions over time as the process continued.
In the aftermath of this exercise, the concept of “civilian harm mitigation” became institutionalized as a part of U.S. policy and practice. Civilian harm mitigation is a practical and systematic approach to identify and reduce risks to civilians from military operations to the maximum extent feasible. It has two key elements: (1) it is a learning process—continually identifying risks to civilians and adapting to reduce them when possible and (2) it is is a comprehensive process—making adaptations not just in the trigger-pull moment of the attack but from the design and planning of the operation to execution. The process continues as the military force assesses, responds to, and learns from civilian harm when it occurs.
This comprehensive learning process is illustrated in the life cycle shown below:
Does Israel Meet the Gold Standard for Civilian Harm Mitigation?
Is Israel adhering to the gold standard for civilian harm mitigation in its operations in Gaza? Spencer, who is chair of urban warfare studies at the Modern War Institute, makes three specific points to support his “gold standard” claim. First, the IDF issued warnings in northern Gaza to notify civilians to evacuate before attacks began. Second, the IDF calls and texts civilians in proximity to specific areas before an attack. And third, the IDF employs “roof knocking,” using a small, non-destructive munition on the roof of a structure to create a noise and warn civilians to flee before attacking the structure. Spencer asserts that other militaries do not use these measures, and thus the IDF is an excellent example of civilian harm mitigation.
But how effective was the evacuation order in northern Gaza? The initial order for evacuation called for over a million people to move to southern Gaza within 24 hours. Experts in humanitarian evacuation found this order to fall significantly short of what is practically possible to do. Moreover, even in an evacuation, many people cannot or do not want to leave—we see this even in evacuation orders in the United States ahead of deadly hurricanes. The poor, the elderly, those with children or those with medical conditions can be extremely challenging to relocate. It was also not clear where civilians could safely evacuate to, particularly after the IDF struck several of the evacuation routes. They were ordered to move to southern Gaza, but that area is not immune from attacks, and both Israel and Egypt have closed their borders. An evacuation order in this context is never going to be a panacea—but one issued with so little time to react and with such poor options and no true sanctuary is of limited value in guarding civilians from military attacks.
The IDF also calls and texts areas before attacks. How effective is this precaution? There is evidence that this measure can be effective in encouraging civilians to evacuate a building. For example, in IDF operations in 2014, many buildings were struck with very few civilian deaths. However, the operations in 2014 were of a much lower intensity, both in terms of the number of strikes and size of ordinance. In the current Gaza operations, there are so many attacks with such large areas of destruction that just leaving a building would not necessarily protect civilians. We also know that communications infrastructure has been crippled from the attacks in Gaza. In this context, how many of these messages even got through? The IDF does not appear to be providing communications capabilities to bridge such gaps and support delivery of precautionary messages.
The final measure given as an example of IDF civilian harm mitigation is warning occupants with roof knocking. In 2014 operations, the effectiveness of this precaution seems to have been mixed. It may be effective when combined with other warnings such as phone calls and leaflets, as it was originally conceived to be used. But using bombs—even small ones—to communicate can be confusing to civilians, especially in an urban war zone. For example, on a multi-floor building, how many floors would hear this roof knocking, and will they know unambiguously that it is a warning to them? In interviews after previous operations in 2014, civilians said they could become confused. Was the roof knock targeting their building or was it a building adjacent to them, or was it a high yield attack elsewhere in the area? In the current, high-intensity operations in Gaza, it is easy to see how this confusion could be magnified.
Ultimately, warnings are only effective in mitigating harm when they are understandable and clear to the party in their context. This was a hard lesson that U.S. forces struggled to learn in checkpoint operations in Iraq and Afghanistan—while measures taken in checkpoints seemed clear from the perspective of the military, the perspective of a civilian in a conflict zone can be quite different. Measures the IDF used in past, lower intensity operations like calls, texts, and roof knocking could have been modestly effective in that setting but of questionable effectiveness in a higher intensity conflict setting.
Precautions Not Taken
In addition to the above measures being insufficient, there are many steps intrinsic to civilian harm mitigation that the IDF apparently could have taken but did not. One notable example is a failure to implement civilian harm mitigation practices into campaign design. While it is true that Hamas is using civilians as human shields, the IDF appears not to be doing enough to shape the operating environment to limit harm to civilians. This was a key lesson that the United States learned in Raqqa, where the campaign design for counter-ISIS operations created a siege situation that led to more civilian deaths and the destruction of civilian property. That campaign design itself created risk to civilians. The same thing is happening in Gaza now.
The IDF also appears to have adopted a high threshold for acceptable civilian loss in Gaza. This leads to the IDF making targeting decisions that create significant numbers of civilian casualties such as the Oct. 31 strike on the refugee complex at Jabalya. The level of civilian harm that resulted from that strike is something we have not seen U.S. forces doing deliberately in recent operations. There were a few high-casualty incidents in Mosul and Raqqa but they were unintentional—for example, a case where the U.S. was baited by ISIS to attack a building with civilians hidden in the basement, or an attack on an IED factory where the secondary explosions were so intense that many civilians in the vicinity were harmed. In contrast, the IDF seems willing to conduct attacks knowing that scores or even hundreds of civilians will be harmed. It appears that Israel’s risk tolerance for civilian harm compared to expected operational benefits is significantly different than in the past. U.S. officials who met with Israeli counterparts to discuss the IDF process for calculating how many civilians are considered acceptable collateral damage, “said that Israel’s bar is far higher than the United States’ would be.”
The IDF’s frequent use of high yield weapons in Gaza also stands in contrast with best practices for civilian harm mitigation. In U.S. operations, aircraft would often have a loadout with a range of different weapons and decisions to attack would consider the minimum destruction that was required to achieve the mission objective.
Collecting and leveraging data on civilian harm is also a core element of civilian harm mitigation, recording and tracking data on civilian harm, as well as engaging with other entities to make sure the information used for assessments and learning is the best quality possible. The United States has improved considerably in this area. A 2018 U.S. assessment showed that DOD internal data on civilian harm was severely limited, and it needed to also consider outside information for the most accurate assessments. Since then, DOD has worked to refine this information, including working closely with civil society groups (such as Airwars) in the form of interactions with outsiders that can be vital to military learning. This tracking and engagement to improve available data on civilian harm appears absent in Israeli operations in Gaza, representing yet another area where the IDF falls well short of the gold standard for civilian harm mitigation.
In short, what we are witnessing in Gaza is the IDF employing several precautionary measures that they have used before, with no tangible evidence of data collection, analysis, and adaptation for the current environment. The cut-and-paste approach to precautionary measures we see by the IDF falls far short of the gold standard for civilian harm mitigation. It is telling when commentaries lauding the IDF can cite only to those measures.
What Does the Data Say?
In civilian harm mitigation, it is useful to rely on metrics to capture the aggregate risk to civilians posed by military operations. One primary metric experts like myself have used in the past is the civilian casualty rate—the percent of attacks that result in harm to civilians. I am not able to determine this figure since the IDF has not released an assessment of the number of incidents causing civilian harm. Another is the average magnitude of civilian harm per incident. Together, these two metrics capture the overall risk to civilians from military actions and the scale of this risk. However, these metrics require information that the IDF has not publicly reported. Given this limitation, there is a metric that can be used in the absence of this information and represents an aggregate of the two other metrics: the average number of civilians killed (or harmed) per 100 attacks.
The number of casualties has been reported through the Hamas-run Gaza Health Ministry. While some have accused the reporting as being biased, experts have found their reporting in this operation as well as past operations to be credible, with no signs of over-reporting. And President Joe Biden saw fit to cite the aggregate statistic of 30,000 casualties in the State of the Union Address, which close observers note presumably comes from this same source. Practically, these numbers could very well underestimate the full scale of deaths in Gaza due to the widespread destruction and bodies still buried under the rubble.
Concerns about credibility may reflect confusion about what the reported casualty numbers are and what they are not: they reflect total deaths but do not include affiliation or cause. This means we need to do some work to estimate civilian deaths from the war in Gaza. Specifically, the Health Ministry numbers contain three different types of deaths: civilians killed by operations, Hamas fighters (since they are non-state actors and do not wear uniforms), and normal deaths in the population. We need to separate out the second and third types to reach the actual number of civilian deaths from the conflict.
The IDF does not regularly report the number of airstrikes in Gaza. For this calculation to be meaningful, we need to use data for both airstrikes and casualties that was current as of the IDF’s last report, so that they reflect the same time period. The last IDF report on airstrikes was on Feb 20. As of this date, the IDF reported 29,000 airstrikes in Gaza. At that date, the number of reported civilian deaths in Gaza was roughly 29,200. The number of Hamas fighters killed in action has been claimed by the IDF and Hamas, both without evidence. The IDF claims 12,000 fighters killed, while Hamas claims 6,000 fighters killed. We can look at those numbers given the demographic data from the health ministry: out of 29,200 deaths, about 9,700 were men, and the rest were women and children. If the IDF claim is true, this is equivalent to every man being killed in Gaza being a Hamas fighter and several thousand women and/or children being Hamas fighters. This is extremely unlikely, but to be extremely conservative I will use the IDF-reported Hamas casualty numbers, 12,000, to give a lower bound on civilian deaths in Gaza. In addition, we need to subtract deaths that would normally happen for this population: 4,900 deaths per year on average, equivalent to 1,838 deaths in a 4.5 month period. This gives us a minimum number of civilian war deaths of approximately 15,700. Based on 29,000 airstrikes, this leads to an average of 54 civilians killed per 100 attacks.
How does this compare to other operations? A roughly comparable operation where we have similar data is urban operations in Raqqa, Syria, against the Islamic State. In the Raqqa operation, according to DOD reporting, there were 178 civilian deaths and 10,663 airstrikes—an average of 1.7 civilian deaths per 100 attacks. This number for Raqqa was not considered to be good—never mind a gold standard. In fact, there was so much concern about the levels of civilian harm and destruction in Raqqa that DOD conducted an independent assessment of civilian harm (disclaimer: I was a member of the team that authored the report). The findings of this assessment contributed to the Civilian Harm Mitigation and Response Action Plan (CHMR-AP) directed by U.S. Secretary of Defense Lloyd Austin. In the report, we also mention that this DoD estimate is likely low. For example, our study team considered the Airwars estimate of 744 civilian deaths to be more realistic. This yields a higher average of 7.0 civilian deaths per 100 attacks.
Despite the alarm over the high rate of civilian deaths in Raqqa, one finds the minimum equivalent in Gaza—54 civilians killed in 100 attacks—is eight times greater than the Airwars-based estimate and 32 times greater than the DOD estimate. And recall that 54 is a lower bound for the Gaza ratio; it is likely far higher than this. Just as miners in California could see the appearance of iron pyrite—fool’s gold—and think they had struck the real thing, it is possible to look at the IDF’s precautionary measures and at first glance think they are practicing civilian harm mitigation. But whether evaluating the IDF’s performance on its process or its results, it fails to qualify as a gold standard.
What Next
The IDF has an opportunity to reexamine its overall approach in Gaza. Possible steps include recording and tracking civilian harm, making comprehensive assessments of civilian harm, with attention to the causes of harm and the effectiveness of precautionary measures, and considering additional steps they can take to reduce risk to civilians while maintaining effectiveness. This should also include greater transparency regarding civilian harm incidents. The IDF should seek dialogue with countries that are actively learning and striving to better mitigate harm to civilians, including the United States and the Netherlands, and with civil society organizations that may be able to support more accurate and neutral assessments, such as Airwars.
Countries that provide military support to Israel can also help encourage and support efforts to better protect civilians. This can include providing expert advisory support on civilian harm mitigation, providing capabilities and information designed to strengthen mitigation efforts, and tailored conditionality—a combination of proactive and reactive measures to promote Israel’s will and capability for civilian harm mitigation, both in Gaza and in the longer term.
A key lesson over the past two decades is that it is possible for governments and militaries to do more to protect civilians in war. That is what civilian harm mitigation is: a comprehensive, learning approach to identifying and mitigating risks to civilians in ways that both reinforce effectiveness and provide better outcomes to civilians in war zones. Civilians in Gaza and Israeli citizens deserve more from their government and their forces than rhetoric about being the gold standard. They deserve the commitment of a government and professional military to do what they can—and to ask for support—to genuinely protect civilians in war.
Editor’s Note: This analysis represents the author’s own views and not the opinions of CNA or any of its sponsors.