Michael Spagat
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Michael Spagat has a total of 32 co-authors. Here are the top co-authors by number of shared publications:
- Stijn van Weezel
Nijmegen School of Management, Radboud University, Nijmegen, The Netherlands
5 shared publications Debarati Guha-Sapir
Centre for Research on Epidemiology of Disasters, Université Catholique de Louvain, Ottignies, Louvain-la-Neuve, Belgium
3 shared publicationsEran Bendavid
Primary Care and Population Health, Department of Medicine, Stanford University, Stanford, CA, USA
3 shared publicationsJon Pedersen
Oslo, Norway
3 shared publicationsKhalil Shikaki
Palestinian Center for Policy and Survey Research, Ramallah, occupied Palestinian territory
3 shared publicationsMichael Robbins
Department of Politics, Princeton University, Princeton, NJ, USA
3 shared publications
Latest Publications
- 28 October 2026
- Published by Oxford University Press (OUP) in Journal of Peace Research
The impact of drone warfare on suicide bombings in Pakistan
Drone strikes against militant groups have become a central counterterrorism tool, yet their effectiveness remains debated. We use weather-based instruments and a major drone-base closure to identify the causal effect of Central Intelligence Agency drone strikes in North Waziristan, Pakistan, on suicide bombings nationwide during 2008–2016. Our instrumental variables strategy improves on recent work by combining cloud cover interacted with precipitation—which directly impedes camera-based targeting—with a base closure providing qualitatively different variation. Focusing exclusively on suicide bombings allows us to test specific theoretical mechanisms and provide evidence no existing study offers: quantification of the total programme effect, heterogeneity between decapitation and non-decapitation strikes, and systematic analysis of attack target types. We find each drone strike causes approximately one additional suicide bombing within the subsequent month, with 18%–23% of all suicide bombings during this period causally attributable to the drone programme. Strikes eliminating militant leaders provoke reactions roughly 8 times stronger than ordinary strikes. These reaction attacks predominantly target security installations, not civilians, contradicting theories emphasizing capacity loss or undisciplined violence and instead supporting our proposed mechanism: militant groups strategically “green light” pre-prepared suicide attacks as face-saving responses to authority challenges. We formalize these predictions as four testable hypotheses—blowback effects, decapitation amplification, strategic targeting, and rapid response—all of which receive empirical support. Our findings complement recent evidence that drone strikes increase terrorism and that leadership targeting disrupts organizational control, while providing novel evidence on suicide bombing dynamics, decapitation heterogeneity, and strategic target selection. The results suggest policymakers should anticipate month-long vulnerability windows after strikes, particularly following leadership kills, and harden security targets rather than assume militants will shift to softer civilian targets.
- 28 October 2026
- Published by Oxford University Press (OUP) in Journal of Peace Research
The impact of drone warfare on suicide bombings in Pakistan
Drone strikes against militant groups have become a central counterterrorism tool, yet their effectiveness remains debated. We use weather-based instruments and a major drone-base closure to identify the causal effect of Central Intelligence Agency drone strikes in North Waziristan, Pakistan, on suicide bombings nationwide during 2008–2016. Our instrumental variables strategy improves on recent work by combining cloud cover interacted with precipitation—which directly impedes camera-based targeting—with a base closure providing qualitatively different variation. Focusing exclusively on suicide bombings allows us to test specific theoretical mechanisms and provide evidence no existing study offers: quantification of the total programme effect, heterogeneity between decapitation and non-decapitation strikes, and systematic analysis of attack target types. We find each drone strike causes approximately one additional suicide bombing within the subsequent month, with 18%–23% of all suicide bombings during this period causally attributable to the drone programme. Strikes eliminating militant leaders provoke reactions roughly 8 times stronger than ordinary strikes. These reaction attacks predominantly target security installations, not civilians, contradicting theories emphasizing capacity loss or undisciplined violence and instead supporting our proposed mechanism: militant groups strategically “green light” pre-prepared suicide attacks as face-saving responses to authority challenges. We formalize these predictions as four testable hypotheses—blowback effects, decapitation amplification, strategic targeting, and rapid response—all of which receive empirical support. Our findings complement recent evidence that drone strikes increase terrorism and that leadership targeting disrupts organizational control, while providing novel evidence on suicide bombing dynamics, decapitation heterogeneity, and strategic target selection. The results suggest policymakers should anticipate month-long vulnerability windows after strikes, particularly following leadership kills, and harden security targets rather than assume militants will shift to softer civilian targets.
- Open Access
- 11 June 2026
- Published by Elsevier in The Lancet Global Health
- Vol. 14 (8) , 103992
Response to DellaPergola and Zlochin
- 11 June 2026
- Published by Elsevier in The Lancet Global Health
- Vol. 14 (8) , 103992
Open AccessResponse to DellaPergola and Zlochin
- Top Cited PapersOpen Access
- 18 February 2026
- Published by Elsevier in The Lancet Global Health
- Vol. 14 (4) , e552-e559
Violent and non-violent death tolls for the Gaza conflict: new primary evidence from a population-representative field survey
- 18 February 2026
- Published by Elsevier in The Lancet Global Health
- Vol. 14 (4) , e552-e559
Top Cited PapersOpen AccessViolent and non-violent death tolls for the Gaza conflict: new primary evidence from a population-representative field survey
- Top Cited PapersOpen Access
Violent and Nonviolent Death Tolls for the Gaza War: New Primary Evidence
High-quality war mortality estimates, such as those that were produced for Kosovo, Iraq, and Darfur play a critical role in illuminating the human cost of war. During the tumult of war they are, however, challenging to obtain due to the conflicts themselves. The Gaza Ministry of Health (GMoH) has provided regular updates to their post-October-7 violent death tally for the Gaza Strip. However, GMoH reports have attracted both criticism and support. Here, we present results from a large-scale household survey, the Gaza Mortality Survey (GMS), which provides independent estimates of war-related deaths between October 7, 2023 and January 5, 2025. Our findings suggest that violent mortality has significantly exceeded official figures. Our central estimate for the extent of GMoH undercount closely matches a separate estimate made using capture-recapture methods. We also find that nonviolent excess deaths, often overlooked in conflict assessments, also represent a substantial burden. These results underscore the feasibility of mortality surveillance methodologies in some highly challenging war-affected settings and provide a crucial empirical foundation for assessing the true human cost of the war.
Top Cited PapersOpen AccessViolent and Nonviolent Death Tolls for the Gaza War: New Primary Evidence
High-quality war mortality estimates, such as those that were produced for Kosovo, Iraq, and Darfur play a critical role in illuminating the human cost of war. During the tumult of war they are, however, challenging to obtain due to the conflicts themselves. The Gaza Ministry of Health (GMoH) has provided regular updates to their post-October-7 violent death tally for the Gaza Strip. However, GMoH reports have attracted both criticism and support. Here, we present results from a large-scale household survey, the Gaza Mortality Survey (GMS), which provides independent estimates of war-related deaths between October 7, 2023 and January 5, 2025. Our findings suggest that violent mortality has significantly exceeded official figures. Our central estimate for the extent of GMoH undercount closely matches a separate estimate made using capture-recapture methods. We also find that nonviolent excess deaths, often overlooked in conflict assessments, also represent a substantial burden. These results underscore the feasibility of mortality surveillance methodologies in some highly challenging war-affected settings and provide a crucial empirical foundation for assessing the true human cost of the war.
- Open Access
- 27 February 2024
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 19 (2) , e0297895
The violent death toll from the Iraq War: 2003–2023
From the beginning of the Iraq war, in March of 2003, to the present day, controversy has swirled around the death toll of the war. This paper narrows down the range of uncertainty for the numbers and trends in violent deaths in the war. I assemble and appraise all primary sources that cover the period from March of 2003 onwards—six sample surveys plus a casualty recording project (Iraq Body Count [IBC]). Data permitting, I present cumulative monthly figures with, for the surveys, 95% bootstrapped uncertainty intervals. The analysis uncovers a core of high-quality mainstream sources that are highly consistent with each another. In addition, there are three outlier surveys that are compromised by serious flaws and produce estimates far outside the mainstream. Discarding the outlying and flawed surveys reveals a clear picture of the violent death toll from the Iraq war. IBC figures, extended to include combatants, occupy a central position within the mainstream range of estimates. The strong consistency across the high-quality sources provides a rare validation of three war-death-measurement methodologies—household-based surveys, sibling-based surveys, and casualty recording. Methodological success notwithstanding, we must transcend the numbers to truly comprehend the human costs of the war.
- 27 February 2024
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 19 (2) , e0297895
Open AccessThe violent death toll from the Iraq War: 2003–2023
From the beginning of the Iraq war, in March of 2003, to the present day, controversy has swirled around the death toll of the war. This paper narrows down the range of uncertainty for the numbers and trends in violent deaths in the war. I assemble and appraise all primary sources that cover the period from March of 2003 onwards—six sample surveys plus a casualty recording project (Iraq Body Count [IBC]). Data permitting, I present cumulative monthly figures with, for the surveys, 95% bootstrapped uncertainty intervals. The analysis uncovers a core of high-quality mainstream sources that are highly consistent with each another. In addition, there are three outlier surveys that are compromised by serious flaws and produce estimates far outside the mainstream. Discarding the outlying and flawed surveys reveals a clear picture of the violent death toll from the Iraq war. IBC figures, extended to include combatants, occupy a central position within the mainstream range of estimates. The strong consistency across the high-quality sources provides a rare validation of three war-death-measurement methodologies—household-based surveys, sibling-based surveys, and casualty recording. Methodological success notwithstanding, we must transcend the numbers to truly comprehend the human costs of the war.
- Open Access
- 8 February 2021
- Published by Oxford University Press (OUP) in Journal of Peace Research
- Vol. 58 (5) , 900-914
Civilian targeting in African conflicts: A poor actor’s game that spreads through space
Armed conflict actors frequently target civilian populations. Thus, an improved understanding of such behaviour could pave the way to reducing it. We use the Civilian Targeting Index (CTI) and a broad array of geo-referenced data to investigate the spatio-temporal and economic dynamics of civilian targeting by conflict actors in Africa. Two main insights are generated. First, the civilian targeting behaviour of African non-state conflict actors is strongly influenced by the behaviour of other proximate actors. In particular, non-state actors tend to increase their civilian targeting after nearby non-state actors have done so. Possible mechanisms to explain such spatial spillovers include emulation and retaliation. Second, a negative relationship between economic activity and civilian targeting exists and applies to both state and non-state actors. In addition, CTIs of non-state actors tend to increase with population density, the geographical spread of their conflict activity and conflict duration. State actors have higher average CTIs than non-state actors do, but the gap between the two actor types tends to close during long-duration conflicts.
- 8 February 2021
- Published by Oxford University Press (OUP) in Journal of Peace Research
- Vol. 58 (5) , 900-914
Open AccessCivilian targeting in African conflicts: A poor actor’s game that spreads through space
Armed conflict actors frequently target civilian populations. Thus, an improved understanding of such behaviour could pave the way to reducing it. We use the Civilian Targeting Index (CTI) and a broad array of geo-referenced data to investigate the spatio-temporal and economic dynamics of civilian targeting by conflict actors in Africa. Two main insights are generated. First, the civilian targeting behaviour of African non-state conflict actors is strongly influenced by the behaviour of other proximate actors. In particular, non-state actors tend to increase their civilian targeting after nearby non-state actors have done so. Possible mechanisms to explain such spatial spillovers include emulation and retaliation. Second, a negative relationship between economic activity and civilian targeting exists and applies to both state and non-state actors. In addition, CTIs of non-state actors tend to increase with population density, the geographical spread of their conflict activity and conflict duration. State actors have higher average CTIs than non-state actors do, but the gap between the two actor types tends to close during long-duration conflicts.
- Open Access
- 30 March 2020
- Published by Springer Nature in Population and Environment
- Vol. 42 (1) , 79-94
Excess deaths and Hurricane María
We clarify the distinction between direct and indirect effects of disasters such as Hurricane María and use data from the Puerto Rico Vital Statistics System to estimate monthly excess deaths in the immediate aftermath of the hurricane which struck the island in September of 2017. We use a Bayesian linear regression model fitted to monthly data for 2010–2016 to predict monthly death tallies for all months in 2017, finding large deviations of actual numbers above predicted ones in September and October of 2017 but much weaker evidence of excess mortality in November and December of 2017. These deviations translate into 910 excess deaths with a 95% uncertainty interval of 440 to 1390. We also find little evidence of big pre-hurricane mortality spikes in 2017, suggesting that such large spikes do not just happen randomly and, therefore, the post-hurricane mortality spike can reasonably be attributed to the hurricane.
- 30 March 2020
- Published by Springer Nature in Population and Environment
- Vol. 42 (1) , 79-94
Open AccessExcess deaths and Hurricane María
We clarify the distinction between direct and indirect effects of disasters such as Hurricane María and use data from the Puerto Rico Vital Statistics System to estimate monthly excess deaths in the immediate aftermath of the hurricane which struck the island in September of 2017. We use a Bayesian linear regression model fitted to monthly data for 2010–2016 to predict monthly death tallies for all months in 2017, finding large deviations of actual numbers above predicted ones in September and October of 2017 but much weaker evidence of excess mortality in November and December of 2017. These deviations translate into 910 excess deaths with a 95% uncertainty interval of 440 to 1390. We also find little evidence of big pre-hurricane mortality spikes in 2017, suggesting that such large spikes do not just happen randomly and, therefore, the post-hurricane mortality spike can reasonably be attributed to the hurricane.
- Open Access
- 15 November 2019
- Published by Springer Nature
- p. 129-142
The Decline of War Since 1950: New Evidence
For the past 70 years, there has been a downward trend in war sizes, but the idea of an enduring ‘long peace’ remains controversial. Some recent contributions suggest that observed war patterns, including the long peace, could have resulted from a long-standing and unchanging war-generating process, an idea rooted in Lewis F Richardson’s pioneering work on war. Focusing on the hypothesis that the war sizes after the Second World War are generated by the same mechanism that generated war sizes before the Second World War, recent work failed to reject this ‘no-change’ hypothesis. In this chapter, we transform the war-size data into units of battle deaths per 100,000 of world population rather than absolute battle deaths – units appropriate for investigating the probability that a random person will die in a war. This change tilts the evidence towards rejecting no-change hypotheses. We also show that sliding the candidate break point slightly forward in time, to 1950 rather than 1945, leads us further down the path toward formal rejection of a large number of no-change hypotheses. We expand the range of wars considered to include not just inter-state wars, as is commonly done, but also intra-state wars. Now we do formally reject many versions of the no-change hypothesis. Finally, we show that our results do not depend on the choice of war dataset.
- 15 November 2019
- Published by Springer Nature
- p. 129-142
Open AccessThe Decline of War Since 1950: New Evidence
For the past 70 years, there has been a downward trend in war sizes, but the idea of an enduring ‘long peace’ remains controversial. Some recent contributions suggest that observed war patterns, including the long peace, could have resulted from a long-standing and unchanging war-generating process, an idea rooted in Lewis F Richardson’s pioneering work on war. Focusing on the hypothesis that the war sizes after the Second World War are generated by the same mechanism that generated war sizes before the Second World War, recent work failed to reject this ‘no-change’ hypothesis. In this chapter, we transform the war-size data into units of battle deaths per 100,000 of world population rather than absolute battle deaths – units appropriate for investigating the probability that a random person will die in a war. This change tilts the evidence towards rejecting no-change hypotheses. We also show that sliding the candidate break point slightly forward in time, to 1950 rather than 1945, leads us further down the path toward formal rejection of a large number of no-change hypotheses. We expand the range of wars considered to include not just inter-state wars, as is commonly done, but also intra-state wars. Now we do formally reject many versions of the no-change hypothesis. Finally, we show that our results do not depend on the choice of war dataset.
- Open Access
- 17 October 2018
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 13 (10) , e0204639
Fundamental patterns and predictions of event size distributions in modern wars and terrorist campaigns
It is still unknown whether there is some deep structure to modern wars and terrorist campaigns that could, for example, enable reliable prediction of future patterns of violent events. Recent war research focuses on size distributions of violent events, with size defined by the number of people killed in each event. Event size distributions within previously available datasets, for both armed conflicts and for global terrorism as a whole, exhibit extraordinary regularities that transcend specifics of time and place. These distributions have been well modelled by a narrow range of power laws that are, in turn, supported by some theories of violent group dynamics. We show that the predicted event-size patterns emerge broadly in a mass of new event data covering all conflicts in the world from 1989 to 2016. Moreover, there are similar regularities in the events generated by individual terrorist organizations, 1998—2016. The existence of such robust empirical patterns hints at the predictability of size distributions of violent events in future wars. We pursue this prospect using split-sample techniques that help us to make useful out-of-sample predictions. Power-law-based prediction systems outperform lognormal-based systems. We conclude that there is indeed evidence from the existing data that fundamental patterns do exist, and that these can allow prediction of size distribution of events in modern wars and terrorist campaigns.
- 17 October 2018
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 13 (10) , e0204639
Open AccessFundamental patterns and predictions of event size distributions in modern wars and terrorist campaigns
It is still unknown whether there is some deep structure to modern wars and terrorist campaigns that could, for example, enable reliable prediction of future patterns of violent events. Recent war research focuses on size distributions of violent events, with size defined by the number of people killed in each event. Event size distributions within previously available datasets, for both armed conflicts and for global terrorism as a whole, exhibit extraordinary regularities that transcend specifics of time and place. These distributions have been well modelled by a narrow range of power laws that are, in turn, supported by some theories of violent group dynamics. We show that the predicted event-size patterns emerge broadly in a mass of new event data covering all conflicts in the world from 1989 to 2016. Moreover, there are similar regularities in the events generated by individual terrorist organizations, 1998—2016. The existence of such robust empirical patterns hints at the predictability of size distributions of violent events in future wars. We pursue this prospect using split-sample techniques that help us to make useful out-of-sample predictions. Power-law-based prediction systems outperform lognormal-based systems. We conclude that there is indeed evidence from the existing data that fundamental patterns do exist, and that these can allow prediction of size distribution of events in modern wars and terrorist campaigns.