The new Western way of war in the post-Vietnam era is typified by the transfer of risk from soldiers to enemy civilians to reduce their own casualties and, by implication, the political costs stemming from the growing domestic social sensitivity to casualties.¹ Risk transfer is accomplished by using excessive lethality with relatively limited discrimination between combatants and noncombatants. Exercising greater caution to avoid civilian casualties probably increases the danger to soldiers.

An examination of Israel's mode of fighting in the Gaza Strip shows that it has apparently followed the same pattern and transferred the risk away from its own soldiers to enemy civilians by launching aggressive military campaigns. Following Operation Protective Edge (2014), the Israeli organization Breaking the Silence, which documents soldiers' testimonies, argued that the "guiding military principle of 'minimum risk to our forces, even at the cost of harming innocent civilians,' alongside efforts to deter and intimidate the Palestinians, led to massive and unprecedented harm to the population and the civilian infrastructure in the Gaza Strip."² The former head of the IDF's (Israel Defense Forces) International Law Department echoed this logic.³ Israel denied this argument and insisted that it respected the principles dictated by international law.⁴

This paper addresses this controversy, which has not been adequately discussed by students of Israel's military policies. Scholars indeed showed that Israel has experienced growing casualty sensitivity domestically, as have other industrialized democracies. The first Lebanon War (1982) marked the watershed. Sensitivity to casualties was reflected in the public discourse and protests by bereaved parents and other groups.⁵ This sensitivity reached its peak in the second Lebanon War of 2006. Then, politicians and military commanders were accused of giving exaggerated weight to the fear of casualties and therefore delaying ground operations to the last minute.⁶ At the same time, Israel created a force-casualty trade-off by becoming more sensitive to its own military losses at the expense of the losses of enemy civilians. This trade-off is reflected in the increasing use of lethality against the Gaza Strip, evident in the ratio of fatalities between Israeli soldiers and Palestinian citizens. This ratio increased from one Israeli soldier over six civilians killed in Gaza during the first Intifada (1987-93) to one Israeli soldier over 84 Palestinian civilians killed in Operation Cast Lead (2008-09).⁷

However, these studies leave us with three major gaps. First, an analysis is missing of the recent Operation Protective Edge and the extent to which it showed a continuation of the pattern evident in previous operations. Second, there is no in-depth analysis of the differences in the numbers of civilian fatalities reported by different agencies to resolve the dispute between the Israeli government and civil-rights organizations. Third, as mentioned above, I argue for a force-casualty trade-off by calculating the fatality ratio of Israeli soldiers to Gazan civilians, a ratio that might indicate the scale of the transfer of risk. Nevertheless, this analysis ignores the possibility that variations in, or the stability of, the ratio between the death of one's own soldiers and enemy civilians do not necessarily indicate changes in the scale of the transfer of risk. Rather, this ratio is an outcome that can be produced by other factors, such as changes in the opponent's capabilities, without directly affecting the risk of death for civilians.

By analyzing three campaigns Israel launched against Gaza between 2006 and 2014, I argue that Israel has followed the same pattern of many Western armies engaged in urban warfare and transferred the risk away from its own soldiers to enemy civilians by using excessive lethality with relatively limited distinction between combatants and noncombatants. This argument relies on analyzing multiple combinations of fatality ratios whose trends are implicitly agreed upon by both Israel and its critics, and on the extent to which these ratios are mirrored in practices on the ground. After providing a general background, I will present the methodology for measuring variations in the transfer of risk. The fourth and fifth sections of the paper provide the fatality ratios and the practices they represent.

BACKGROUND

The Gaza Strip has been under Israel's direct and indirect control since 1967. Hostilities began in 1987, when the first Intifada, the Palestinian uprising against Israel's rule in the West Bank and Gaza, broke out. In 1993, the PLO and Israel signed the Oslo Accords, after which Israel withdrew its forces from Gaza's urban areas in 1994, while leaving the settlements that Israel had established in Gaza in place under the protection of the IDF. In 2000, the second Intifada signified another round of hostilities between Israel and Palestinian armed militias in both Gaza and the West Bank. In 2005, Israel completed its withdrawal from Gaza by pulling out its forces and evicting the approximately 8,000 Jewish settlers. But the withdrawal did not end the hostilities. As early as 2001, local militias began firing high-trajectory weapons at Israeli civilian settlements near the Israel-Gaza border. These attacks have intensified since Israel's withdrawal; despite its redeployment, Israel has maintained indirect control by overseeing its border crossings and impeding the construction of both a seaport and an airport in Gaza. Hostilities escalated in 2007, when the Islamist movement Hamas took over the Gaza Strip after factional fighting with Fatah militias. A separate Hamas-governed mini-state was then created.

To end what it perceived as Israel's siege of Gaza, which Israel tightened after Hamas took control, the latter intensified its shelling of Israel's civilian population and increased the distance of its rockets. Israel launched several large-scale operations to suppress the rocket fire. Relevant to this article are the operations that involved ground incursions in which the dilemma about risk transfer is evident.

The first operation was Summer Rains, conducted June-November 2006. It was a five-month small-scale operation in which Israel bombed civilian infrastructure and made brigade-size incursions into urban areas. With the intensification of rocket attacks that threatened about a million Israelis, Israel launched Operation Cast Lead in December 2008-January 2009, the largest operation since the withdrawal of 2005. After a week-long aerial attack on Hamas installations in which Hamas refused to seek a ceasefire on terms favored by Israel, Israel launched a division-size ground invasion that lasted another two weeks. Against a similar background, Israel launched Operation Pillar of Defense in 2012, accomplished without a ground invasion. The last operation was Protective Edge in 2014. After 10 days of aerial assaults during which Hamas made several attempts to enter Israeli territory through tunnels, the government set a new goal: to destroy the tunnels by launching a ground incursion lasting nearly another three weeks. A ceasefire was achieved only after a further three weeks of IDF shelling from the air, land and sea.

In all three operations, Israel's goal remained the same: to protect its civilian population by stopping the rocket attacks, rather than the unrealistic goal of unseating the Hamas regime or seeking any territorial gains. However, in each of these operations, complementary goals were also defined. In Summer Rains, Israel reacted to the abduction of the soldier Gilad Shalit (a day before the operation began), seeking to prevent his transfer to Egypt. In Cast Lead, Israel also attempted to stop weapons smuggling from Egypt into Gaza. In Protective Edge, Israel destroyed tunnels leading from Gaza into Israel.⁸

To recall, Israel concurrently experienced growing domestic sensitivity over casualties. This affected Israel's policies in Gaza as well. Although Israel recaptured the major Palestinian cities in the West Bank during the second Intifada (mainly in 2002), it refrained from deploying its forces in the congested urban areas of the Gaza Strip. And even when Hamas intensified its firing of rockets after the withdrawal of 2005, the notion of "Gaza phobia" discouraged Israel from re-entering the Strip for a large-scale ground operation (aside from the limited ones in Summer Rains). The military preferred to delay; Israel thus accepted a six-month Egyptian-brokered ceasefire agreement with Hamas in June 2008.⁹ Against this background, it was clear that Israel could launch a risky ground incursion only when it could claim legitimacy from both the international community and its own citizenry for an aggressive attack that shifted the risk from its soldiers to enemy civilians. Presenting Hamas as escalating its attacks on Israel's civilian communities served this goal in Cast Lead¹⁰ and later in Protective Edge.

METHODOLOGY

Four categories of the number of fatalities form the basis of the analysis:

• Palestinian civilians, including noncombatants, those who were not involved in the fighting but were directly killed by Israel's attacks

• Palestinian combatants in the armed militias, although not defined as soldiers, who were directly killed by Israel's attacks

• Israeli civilians, including noncombatants, who were not involved in the fighting and were killed by shelling and rocketing from Gaza

• Israeli soldiers directly killed in ground operations, including by friendly fire. To isolate the impact of shifting risk, I excluded 21 soldiers killed by attacks on military compounds within Israeli territory. As these casualties could not be reduced by further risking Gazan civilians, I also decreased the number of Palestinian combatants killed in these attacks.

Information about these numbers comes from two sources:

(1) Formal Israeli sources. The Internet sites of the Israeli Prime Minister's Office and the Israel Ministry of Defense provide information about the deaths of Israeli soldiers and civilians, while the IDF releases numbers about Palestinian fatalities through the Israeli Ministry of Foreign Affairs.¹¹ Israel's numbers best reflect its intentions about the permissible level of risk to which it might expose Gazan civilians and its perceptions about who is a civilian and therefore protected against direct, intentional attack.

(2) B'Tselem, the Israeli Information Center for Human Rights in the Occupied Territories, provides information about Palestinians fatalities. B'Tselem, other human-rights organizations, the United Nations and the IDF disagree on the distinction between combatants and noncombatants. They also disagree on the number of people directly killed by the IDF rather than those who died from natural causes or hostilities between Palestinian militias. I used B'Tselem's numbers because they lie between the IDF's figures and those published by Palestinian organizations and the United Nations. B'Tselem is less institutionally biased than the government of Israel or the Palestinian NGOs, and conducted exhaustive investigations by the organization's field researchers questioning Palestinian sources and cross-referencing their information with others publicly available.12

These numbers yield four combinations of fatality ratios, each of which reflects another aspect of the fire policies and, hence, their impact on risk transfer:

• Ratio between Israeli soldiers and Palestinian civilians. As noted earlier, this ratio reflects the trade-off between force-produced fatalities and self-casualties and monitors its variations.¹³ It indicates the extent to which soldiers are exposed to risk in their activities among the enemy civilian population relative to the risk they impose on those civilians. Therefore, an increased ratio may indicate a decline in the military's willingness to sacrifice its soldiers by shifting the risk to enemy civilians, but this cannot always be the only inference. Therefore, other ratios are calculated.

• Ratio between Israeli soldiers and Palestinian combatants. This ratio helps test the validity of the ratio between soldiers and enemy civilians to isolate variations in the nature of the mission and its goals, or the rival's capabilities and tactics within the same arena. It indicates the extent to which enemy combatants challenge the IDF troops. Thus, a decline in the ratio of soldiers to enemy civilians may indicate greater risk to Israeli soldiers and hence encourage the military to shift the risk to enemy civilians, either to coerce the opponent into accepting a ceasefire,¹⁴ or to reduce the risk to its own soldiers by easing restrictions on the use of force, sparing the need to exercise more caution to avoid civilian casualties.

• Ratio between Palestinian combatants and Palestinian civilians. Such a ratio represents the extent to which the Israeli troops distinguish between combatants and noncombatants in their efforts to transfer the risk away from their own soldiers. The greater the transfer of risk, the lower the level of discrimination.

• Ratio between Israeli civilians and Palestinian civilians. Increasing the risk to one's own civilians may increase aggressiveness towards enemy civilians to coerce the adversary to accept a ceasefire. Direct rather than collateral targeting of enemy civilians may also be motivated by factors other than shifting the risk, such as deterrence, punishment or even revenge.

MEASURING THE FATALITY RATIO

Table 1 sums up the number of fatalities sorted by the different sources of information.

---

---

As expected, Israel's count of Palestinian civilian fatalities is significantly lower than B'Tselem's, and, by implication, Israel lists higher numbers of Palestinian combatants. Such discrepancies raise the question of how to apply the principle of distinction, as prescribed by international law, between civilians and combatants, forbidding indiscriminate military attacks and limiting offensives to military objectives alone.¹⁵ Application of this principle affects the body count and thus helps explain the discrepancy in the numbers provided by different sources. As both B'Tselem¹⁶ and Amnesty International¹⁷ reported, Israel justified attacks on civilian installations and the deaths that resulted by claiming that these installations were actually being used for military purposes. In short, Israel categorized noncombatants as combatants. However, Israel is formally and morally committed to the principle of distinction, and its fire policy is monitored by legal advisers. Since at least Operation Cast Lead, lawyers have been attached to divisional commands and involved in approving plans and operational decisions.¹⁸

Furthermore, the application of the principle of distinction partly affects the application of another principle, proportionality, holding that even an attack against a military target might still be prohibited if the attack would cause greater harm to civilians than the military advantage gained.¹⁹ However, it is debatable how to apply these two principles to wars between a state and a nonstate actor such as the Hamas-controlled Gaza Strip. Arguments about proportionality affect the justification for civilian casualties even when the distinction is not debated. Nevertheless, more relevant to this study are debates about the application of the principle of distinction; they are more easily quantified by calculating the ratio between enemy combatants and noncombatants.

In one case, I reclustered B'Tselem's numbers. In Cast Lead, 248 policemen were killed by Israeli attacks, 99 during the first minutes of the military operations. Israel clarified that "an overwhelming majority of the police forces were also members of the Hamas military wing or activists of Hamas or other terrorist organisations. Therefore, considering Hamas 'police' casualties as civilians is inappropriate."²⁰ True, the UN's report of inquiry²¹ rejected Israel's position, claiming "there is insufficient information to conclude that the Gaza police as a whole had been 'incorporated' into the armed forces of the Gaza authorities …[and therefore], this attack failed to strike an acceptable balance between the direct military advantage anticipated…and the loss of civilian life."²² However, this is not a case of collateral damage resulting from an attempt to shift the risk away from its own soldiers but a direct targeting of what Israel considered armed forces. Methodologically (and, of course, without judging Israel's interpretation of the law), accuracy suggests clustering policemen and combatants together, so the number of 598 combatants in the B'Tselem row includes 248 policemen and 350 combatants.

Table 2 presents the combinations that yield the fatality ratios.

---

---

Table 3 measures the relative changes in the fatality ratio from one campaign to the next. The fatality ratio in Summer Rains receives the value of 1, and the next operations are measured relative to the previous campaign. The cumulative value measures Protective Edge relative to Summer Rains.

---

---

Remarkably, according to Table 3, the cumulative values are almost the same, regardless of the differences between the sources. As both sources agree, the fatality ratio between Israeli soldiers and Palestinian combatants steadily declined (by 70-80 percent); Israeli soldiers were gradually exposed to a higher level of risk from operation to operation. Both sources show the same relative trend, although Israel counted more Palestinian combatants than B'Tselem because it included more individuals in this category.

While in Summer Rains the campaign generally relied on aerial and artillery attacks, and the troops conducted limited incursions, in Cast Lead larger forces were deployed to Gaza for a longer period of time. More important, in Cast Lead, the government set the minimal goal of achieving a long-term ceasefire by reinforcing Israel's deterrence.²³ A flexible plan evolved from this goal that allowed a limited ground offensive and minimal close confrontations with Hamas forces, without going deeper into the urban areas and incurring the attendant casualties.²⁴

In sharp contrast, in Protective Edge, the operation was expanded to a ground invasion aimed at destroying Gaza's tunnel system. This thrust lasted for nearly three weeks. As these tunnels were deeply embedded in densely populated areas ruled by Hamas forces, Israeli troops were engaged in direct frontal assaults on them, unavoidably increasing the risk to Israeli soldiers and claiming the lives of more of them than in Cast Lead. Furthermore, relative to Cast Lead, Hamas had significantly improved its ability to cause damage; this was evident in the defense system it deployed against Israeli troops and the increased effectiveness and cohesion of its combat forces.²⁵

Because the tunnels were portrayed as a genuine security threat to the Israeli civilian communities in proximity to the Israel-Gaza border, and the public perceived the tunnel-destroying operation as effective,²⁶ Israeli society tolerated the deaths of 44 soldiers in the ground operation on Gazan soil. Similar to the U.S. case, tolerance for casualties increases with public expectations that the military operation will be a success.²⁷

Increasing risk explains the fatality ratio between Israeli soldiers and Gazan civilians. Over time, this ratio declined by more than 50 percent, although the sources disagree about the transition from Summer Rains to Cast Lead. While, from Israel's perspective, it increased its soldiers' exposure to risk to a greater extent than shifting the risk to Gaza's civilians, B'Tselem's inquiry implied the opposite. In other words, while from Israel's perspective, it proactively attacked military installations staffed by armed combatants, from B'Tselem's perspective, it attacked noncombatants. However, both sources agree that, from Cast Lead to Protective Edge, Israeli soldiers were at greater risk in the shift from a maneuverable situation with minimum encounters with Hamas combatants to an operation in a heavily armed, densely populated territory with more trained and better-equipped Hamas forces. On the surface, Israel reduced the level of risk transferred from soldiers to civilians, as the fatality ratio of Israeli soldiers and Palestinian civilians declined by about 50-60 percent, from 1:33 to 1:17 according to Israel's number, or from 1:86 to 1:32 according to B'Tselem. But the reality is more complicated.

Even when the Israeli government deliberately increases the risk to its own soldiers, it cannot symmetrically shift it to enemy civilians to reduce its own casualties. Israel could gain some benefits from escalating the aerial and artillery attacks without deploying ground forces or by reducing their exposure to risk. For example, by using aerial bombardments, Israeli forces could have flattened the villages from which they suspected the tunnels were being built and then sent in ground forces; nevertheless, doing so could have severely weakened Israel's legitimacy internationally. Alternatively, the government could have minimized the soldiers' risk by refraining from conducting a ground operation. However, in Cast Lead, it could not achieve the goal of a long-term ceasefire on its terms, based on deterrence, without a ground thrust.²⁸ In Protective Edge, it could not destroy the tunnels by using aerial attacks alone. The military needed boots on the ground to detect, and then destroy, the tunnels.²⁹

In other words, Israel increased the number of its own soldiers killed without a real ability to reduce it, either by increasing the number of Gaza combatants killed (due to their improved performance) or by increasing the number of Gaza civilians killed (due to legal restrictions). Up to this point, the fatality ratio of Israeli soldiers and Palestinian civilians does not indicate any transfer of risk. Indeed, it points to the opposite conclusion. However, this ratio is not the whole story. To test this inference further, we should look at the fatality ratio between Palestinian combatants and Palestinian civilians. To recall, such a ratio represents the extent to which the troops distinguish between combatants and noncombatants in their efforts to transfer the risk away from their own soldiers.

Again, we see agreement between the sources. Cumulatively, both agree that from Summer Rains to Protective Edge, Israel nearly doubled the number of Gaza civilians the troops killed for every Gaza combatant killed. This ratio climbed from 0.4 Gaza civilians to 0.8 according to Israel's figures, or from 1 to 1.9 according to B'Tselem's data.

However, the trend towards risk transfer is not refuted. Judged by its own numbers, the Israeli military increasingly shifted the level of risk from its own soldiers to Gaza civilians (by 90 percent even relative to the lower number, 80 percent, indicated by B'Tselem). Even if the extent of this shift could be slightly debated in the comparison between Summer Rains and Cast Lead, the figures clearly show that this level increased from Cast Lead to Protective Edge. Even if Israel was willing to sacrifice more soldiers to destroy the perceived threat of the Hamas tunnels, it still reduced the risk to which its soldiers were exposed by transferring part of that risk to Gaza's civilians. The number of Israeli fatalities could have been higher had Israel not increased the risk to Gaza's civilians.

Still, we might argue that, rather than targeting civilians to shift the risk away from its own soldiers, different motivations guided the IDF. To test this option, we look at the fatality ratio between Israeli and Palestinian civilians. Both sources are in agreement about the overall trend: for every Israeli civilian killed in these operations, Israel increasingly claimed the deaths of Gazan civilians by about 120 percent. While there is agreement about the increase from Summer Rains to Cast Lead of 70-140 percent, B'Tselem's figures suggest a slight decrease between Cast Lead and Protective Edge. Remarkably, Israel's figures suggest a moderate increase. Clearly, Hamas's increased firing of rockets at the Israeli population claimed a limited number of Israeli civilian casualties, mainly due to the effective performance of the Iron Dome anti-missile shield in Protective Edge³⁰ and the passive defense system created by the infrastructure of shelters and protected areas in earlier operations.³¹

It follows that Israel gradually claimed a disproportionate sacrifice from Gaza's civilians relative to that borne by its own civilians. However, risk transfer remains the major explanation for the increasing aggressiveness aimed at partly reducing the soldiers' risk, for three reasons. First, democracies bear costs other than death. The disruption of the lives of millions of Israelis, even without a significant death toll, could have had political costs. Therefore, the aerial bombing of suspected Hamas installations could have helped press Hamas to accept a ceasefire as a means of shortening the war. When the suffering of a democracy's own civilians is intertwined with casualty sensitivity and a perceived fruitless ground operation (which failed to stop the rocketing from Gaza), pressure mounts to expand the ground operation despite the causalities it might incur, ultimately encouraging the intensification of risk transfer to counter such pressure.

Second, it was the tunnel-destroying operation in Protective Edge that helped justify the Israelis' sacrifice.³² Nevertheless, extensive coverage of the damage and deaths on the Palestinian side played a role in justifying the sacrifice to the Israeli public and, hence, relaxed pressure on the leadership to escalate the ground operation.³³

Third, relatedly, and because of international monitoring, the explanation that Israel intentionally targeted Gaza civilians to achieve deterrence or for revenge and punishment should be ruled out. Furthermore, as reports from Breaking the Silence note,³⁴ there were some indications of "battlefield frenzy." Nonetheless, since most of the Gaza civilians were killed by artillery or aerial attacks, and since there was no indication of a formal policy aimed at intentionally targeting civilians, misconduct on the ground cannot account for the rise in civilian deaths.

NUMBERS REFLECT POLICIES

These numbers reflect modifications in fire policies from operation to operation. An analysis of policies, moreover, may help reveal which source presents the reality more accurately. In Summer Rains, soldiers were risked to a limited degree. Israel used aerial attacks and artillery fire to reduce the need to resort to costly ground operations. Indeed, as a thorough inquiry by Human Rights Watch³⁵ found, the rise in Palestinian civilian deaths occurred following two changes Israel made in its artillery practices after Hamas took over Gaza. First, the IDF greatly increased the number of artillery shells fired. Second, it increased its artillery attacks in the immediate vicinity of civilian residences by narrowing the minimum distance required between a potential target and the nearest homes or populated areas. Thus, Israel could target nearby rocket-launching sites and Hamas installations. Here we see a typical practice of transferring risk: increasing the use of stand-off weapons to do the job without ground incursions, at the cost of more collateral deaths among noncombatants.

In Cast Lead, Israel used the previous practices but also deployed new ones. Among them were these:

• A kind of rolling fire-induced smokescreen preceded the advancing ground units to protect them.³⁶

• Cluster bombs or shells were used that endangered civilians.³⁷

• There was unprecedented use of inaccurate mortar shells. In one case, the army fired a mortar shell in the Jabalya refugee camp, killing over 40 civilians.³⁸

• Sterile zones were created by warning civilians to leave their homes. Then, the rules of engagement instructed the soldiers that those who stayed behind could not be considered "innocents" and that, in case of any doubt whatsoever, the soldiers were to shoot.³⁹ At the same time, Israel accused Hamas of using civilians as human shields by encouraging them to protect military sites and firing from schools and medical facilities.⁴⁰

• The rules of engagement also prescribed that, in a moment of doubt about how to treat a suspected combatant, "If you are not sure, shoot. If there is doubt, then there is no doubt."⁴¹ Division commanders were allowed to strike any house suspected of being booby-trapped.⁴²

• As described above, Israel targeted buildings that usually served civilian purposes, claiming that these were actually being used for military purposes or were homes of Hamas leaders. Apparently, unlike the practices described above, this is not a typical case of risk transfer. However, the greater the sensitivity to Israel's own casualties among the soldiers, and the greater the suffering of its own civilians who were being attacked by rockets fired from Gaza, the stronger the propensity was to escalate aggressiveness to shorten the war. The longer the operation, the stronger the pressure was to achieve a decisive victory. In this case, prolonging the operation could intensify pressure to expand the ground operation to overthrow the Hamas regime,⁴³ with the casualties that might be expected. To recall, escalated air attacks in the Kosovo War were triggered by similar motivations.

Consequently, most of the civilian fatalities among women and children occurred during the ground offensive or as a result of it, by means of aerial bombing that provided cover for the ground units.⁴⁴ It follows that, in the debate between Israel and B'Tselem over the number of Palestinian civilians killed in Cast Lead, B'Tselem is more credible. The practices described here are reflected in the organization's in-depth inquiry through which it investigated the cause of death of every single Palestinian. Accepting B'Tselem's figures provides stronger support for the argument that, from Summer Rains to Cast Lead, Israel shifted the risk from its own soldiers to Gaza's civilians, as the fatality ratio between Israeli soldiers and Palestinian civilians increased from 1:71 to 1:86. Similarly, while in Summer Rains, for every Gaza combatant killed, Israeli troops killed one Gaza civilian, this ratio increased to 1:1.3 in Cast Lead. Moreover, testimonies from soldiers further validated the link between risk-transfer practices and casualty sensitivity.⁴⁵ To recall, however, the argument about risk transfer is supported even without choosing between sources.

In Protective Edge, the IDF managed its fire policies more cautiously to avoid allegations such as those documented in the UN report following Cast Lead, known as the Goldstone Report.⁴⁶ At the same time, the military perfected its risk-transfer methods. Preferring to avoid a ground operation, Israel accepted an Egyptian-brokered ceasefire after a week of aerial attacks, even at the cost of not dealing with the problem of the tunnels. Conflicting requirements — concern about international investigations versus domestic pressure to increase aggressiveness to deal with casualty sensitivity and the need to demonstrate a decisive victory — resulted in hesitation.⁴⁷ Since Hamas refused to accept the terms of the ceasefire, the ground operation was launched.

As the analysis above confirmed, both sources might agree that the fatality ratio between Israeli soldiers and Palestinian civilians declined from Cast Lead to Protective Edge but was balanced out by increasingly indiscriminate attacks. Indeed, the numbers reflect the descriptions provided by Israeli civil-rights organizations and the United Nations about Israel's fire policies. The practices implemented in Protective Edge included:

(1) Using explosive weapons and imprecise artillery with wide-area effects in densely populated areas to better protect the soldiers.⁴⁸ During the operation, the IDF changed policies, giving junior officers greater discretion over the amount of firepower to use and the acceptable degree of collateral damage that might be caused by reducing the civilian safety ranges.⁴⁹

(2) Creating a "sterile combat zone" by warning civilians to leave their neighborhoods before Israeli troops entered, and then allowing the working assumption that residents had abandoned the neighborhoods, thus making anyone in the area a legitimate target.⁵⁰ As one soldier testified, the military eased the rules of engagement to protect the soldiers by saying that, when in these zones, anyone "you see in the neighborhoods you're in, anything within a reasonable distance, say between zero and 200 meters — is dead on the spot… [because] he isn't supposed to be there.⁵¹" However, in many cases, warnings were absent or ineffective.⁵² It is safe to assume that, relative to Cast Lead, Israel expanded the use of this method, as when its troops dealt with tunnels embedded in densely populated areas.

At the same time, the UN commissions of inquiry documented calls by the spokespersons of Hamas to the people in Gaza to adopt the practice of shielding their homes from attack by going up on their roofs after receiving a warning from the IDF to leave, thereby encouraging the use of human shields.⁵³ Israel indeed cited instances of such calls,⁵⁴ and the use of schools as rocket-launching sites.⁵⁵

(3) Using artillery and air attacks in very close proximity to civilians to support rescue operations for soldiers, disregarding the necessary safety ranges from the civilian population. The most famous case was the rescue operation in the city of Rafah during the last days of Protective Edge. The firing was indiscriminate,⁵⁶ and as the UN commissions of inquiry that investigated the event indicated: "[T]he protection of IDF soldiers was a major consideration for the IDF, overruling, and at times eliminating, any concern for the impact of its conduct on civilians… [W]hen soldiers' lives were at stake or there was a risk of capture, the IDF disregarded basic principles on the conduct of hostilities, namely the principles of distinction, proportionality and precautions."⁵⁷

A similar picture emerged from the report published by B'Tselem,⁵⁸ which investigated 70 incidents in which the military launched strikes from the air, sea and land against residential buildings. It estimated that more than 70 percent of the people killed in these attacks were under 18, over 60 or women. Similar to the logic mentioned with regard to Cast Lead, Israel intensified its aerial attacks during the final week of fighting, lifting some of its self-imposed restrictions. One example is the destruction of luxury towers in the heart of Gaza City. This action was among the factors compelling Hamas to accept a ceasefire arrangement.⁵⁹ Failure could have sparked an expanded ground operation, with the risks to Israeli soldiers of an attempt to reoccupy Gaza and topple the Hamas regime, as some senior politicians, including the foreign minister, urged the government to do.⁶⁰

Protective Edge presented the IDF with more risks and challenges, because it required the Israeli army to deal with military targets embedded among civilians. However, the IDF shifted part of the risk to Gaza's civilians as its own figures about the fatality ratio attest.

CONCLUSIONS

This paper has addressed the controversy between Israel and civil-rights organizations and the United Nations as to whether the IDF has consistently shifted the risk from its own soldiers to enemy civilians in the repeated rounds of hostilities in the Gaza Strip. I expanded the comparative outlook by including the recent Operation Protective Edge. I also elaborated on previous studies by delving into the discrepancies in the fatality numbers provided by different sources, using more fatality ratios to enlarge the analysis of different aspects of fire policies and then correlating them with practices on the ground.

As the cases showed, the ratio between the deaths of Israeli soldiers and Gaza civilians gradually declined because Israel increased the soldiers' exposure to risk, moving from limited incursions to a large-scale operation and then to an operation designed to destroy tunnels. A drop in the ratio between the deaths of one's soldiers and enemy civilians may indicate an increasing willingness to sacrifice soldiers without shifting the risk to enemy civilians. To test this hypothesis and assess variations in the level of risk, the fatality ratio between the combatants of both sides was calculated, confirming the increased risk imposed on Israeli soldiers, represented by the declining ratio between Israeli and Palestinian combatants. This ratio helps test the validity of the ratio between the deaths of soldiers and enemy civilians as a means of identifying variations in the nature of the mission or the rival's capabilities and tactics within the same arena. As the method demonstrates, the fatality ratios of soldiers to enemy civilians and soldiers to enemy combatants do not indicate the transfer of risk, but exactly the opposite.

Nevertheless, the picture would be incomplete if we stopped here. To test this inference further, I calculated the fatality ratio between Palestinian combatants and Palestinian civilians to evaluate the extent to which the Israeli troops distinguished between combatants and noncombatants in their efforts (or the lack of them) to transfer the risk away from their own soldiers. Indeed, these calculations demonstrated risk transfer, because over time, Israel nearly doubled the number of Gaza civilians the troops killed for every Gaza combatant. However, variations in the fatality ratio of Palestinian combatants to Palestinian civilians can also be explained by factors other than risk transfer, such as more intentional targeting of enemy civilians to achieve deterrence or exact revenge. To determine the role of the military's motivation for targeting Gaza civilians for reasons other than shifting risk, I also considered the fatality ratio between Israeli and Palestinian civilians. It showed that, for every Israeli civilian killed in these operations, Israel more than doubled the death toll of Gaza civilians relative to its own civilians' sacrifice. However, and as supported by an analysis of operational practices, the goal of shortening the war seems to have guided Israel's fire policies, another facet of risk transfer.

To validate what the numbers tell us, analyzing the risk-transfer practices behind the numbers provided further validation of the conclusion that Israel shifted the risk away from its soldiers to enemy civilians. As its troops became exposed to more danger, Israel perfected and expanded the use of risk-transfer practices. It started by intensifying the use of artillery in the immediate vicinity of civilian residences and culminated by expanding the "sterile combat zones" in which anyone located in the area was a legitimate target. Fatality ratios may overlook less visible processes unless they are supported by revealing the practices that generated the numbers.

No less important, by evaluating the significance of the differing figures provided by various, often rival, agencies, I focused on the overall trends on which different sources may sometimes implicitly agree. Remarkably, I found overall agreement among Israel's official figures and those provided by B'Tselem and the United Nations, although they had disagreements as well. Agencies do not debate methodological issues such as how to count bodies and categorize them, but legal issues with political implications — usually, how to apply international law regarding the distinction between combatants and civilians. Israel faces pressure from the international community to respect the immunity of noncombatants, and these demands are increasing by ongoing UN investigations and the prospect of war-crimes charges in the International Court of Justice. Against this background, implicit consensus is of extreme significance.

In this case, judged by its own numbers, the Israeli military increasingly shifted the level of risk from its own soldiers to Gaza civilians. Israel's figures confirm that its soldiers were gradually exposed to more danger, surpassing the level of risk imposed on enemy combatants. To mitigate this risk, Israel transferred at least part of it to Gaza's civilians by minimizing the distinction between enemy combatants and noncombatants. Israel's numbers also agree that it increasingly claimed the lives of Gaza civilians in a disproportionate manner relative to its own civilians.

Israel can be used as a typical case study of a democracy whose society is increasingly less tolerant of casualties among its soldiers battling a nondemocratic entity. Given the question of how democracies strike a balance between conflicting imperatives when fighting nondemocratic entities, Israel's dilemmas are comparable to those of other democracies fighting in Iraq, Afghanistan and other arenas. This is particularly true, given that I ruled out the possibility that motivations other than targeting civilians to shift the risk away from its own soldiers guided the IDF.

Therefore, this paper also has policy implications. Establishing a more accurate assessment of how democracies balance conflicting imperatives may contribute to the global monitoring of this conduct. This study's empirical point of departure is that Israel denies its risk-transfer tendencies, but the findings show a different picture. However, this issue is not unique to Israel. Indeed, when General Martin Dempsey, the chairman of the U.S. Joint Chiefs of Staff, declared that in Protective Edge, Israel went to "extraordinary lengths" to limit civilian casualties, the Pentagon sent a team to see what lessons could be learned from the operation.⁶¹ Such an action underscores that the question of how to monitor the extent to which democracies balance casualty sensitivity with respect to the immunity of enemy civilians is increasingly crucial.