Political Assassination Almost Never Works: Twenty-One Centuries of Evidence

What the Data Actually Shows

Economists Benjamin Jones and Benjamin Olken assembled the most comprehensive dataset on political assassination ever compiled: 298 serious attempts on national leaders between 1875 and 2004. Only 59 succeeded, roughly one in five. The rest failed because of body armor, bad aim, faulty bombs, or the kind of dumb luck that separates a head wound from a grazed ear.

Their key finding is counterintuitive. Successful assassinations of autocrats do produce measurable political change: transitions toward democracy are 13 percentage points more likely after a successful assassination than after a failed attempt. But this is not because assassination is an effective tool. It is because autocracies built around a single personality are inherently fragile, and removing the personality exposes the fragility. The assassination does not build democracy. It creates a vacuum, and sometimes democracy fills it. Sometimes something worse does.

In democracies, successful assassinations produce almost no measurable institutional change. The system absorbs the shock. Andrew Johnson replaces Abraham Lincoln. Lyndon Johnson replaces John F. Kennedy. The machinery of government continues, usually in a direction the assassin did not want.

The Backfire Pattern

The most striking feature of assassination history is not that it fails. It is that it consistently produces the opposite of its intended effect.

Julius Caesar, 44 BC. The conspirators killed Caesar to preserve the Roman Republic. The assassination triggered civil wars that destroyed the Republic entirely and produced the Empire, the exact form of one-man rule the assassins had tried to prevent. Within two decades, Caesar’s adopted heir Augustus held more power than Caesar ever had.

Abraham Lincoln, 1865. John Wilkes Booth shot Lincoln to avenge the Confederacy and punish what he saw as tyranny. Lincoln had been pushing for a lenient Reconstruction. His successor, Andrew Johnson, initially took a harder rhetorical line against the South before pivoting to a policy that was simultaneously more lenient toward former Confederates and more hostile to Black civil rights. The result satisfied nobody. Historians have debated for a century and a half whether Reconstruction would have gone differently under Lincoln, but the consensus is that Booth’s bullet did not help the Confederate cause. It eliminated the one leader who had the political authority to manage a more controlled transition.

Archduke Franz Ferdinand, 1914. Princip and the Black Hand wanted South Slav liberation from Austria-Hungary. The assassination set off a chain of alliance activations, ultimatums, and mobilizations that none of the conspirators had anticipated. Austria-Hungary did collapse, but so did the Russian, Ottoman, and German empires, and the resulting settlement at Versailles created conditions so unstable that Europe spent the next two decades sliding toward an even larger catastrophe.

Zhang Zuolin, 1928. Officers of Japan’s Kwantung Army assassinated the Manchurian warlord with a railway bomb, hoping to create chaos that would justify Japanese military intervention in Manchuria. The assassination did lead to the conquest of Manchuria, but it also strengthened the Japanese military’s grip on domestic politics, weakened Japan’s civilian democracy, and set Japan on the path toward the war with the United States that would end in total defeat and two atomic bombs.

Modern Targeted KillingThe military practice of eliminating specific adversary leaders or commanders through surveillance-guided strikes, as opposed to broader military campaigns.: The Same Pattern at Industrial Scale

If individual assassination has a poor track record, what about systematic, state-sponsored targeted killing? Israel has conducted the most extensive campaign of leadership decapitation in modern history, and the results are instructive.

In 1995, Mossad assassinated Fathi Shiqaqi, the founder of Palestinian Islamic Jihad, in Malta. Three decades later, the organization is larger and more dangerous than when Shiqaqi led it. In 1997, a botched attempt to poison Hamas operative Khaled Meshaal in Jordan resulted in the operatives’ capture and Meshaal’s transformation into, as former intelligence officials have described him, a Palestinian hero who rose to Hamas’s top leadership. In 2004, Israel killed Sheikh Ahmed Yassin and Abdel Aziz al-Rantisi, Hamas’s two most senior leaders, within weeks of each other. Hamas responded by increasing its attacks, and the organization that was supposed to be decapitated went on to win the 2006 Palestinian legislative elections.

Former Mossad director Zvi Zamir described targeted killing as “a measure of last resort,” a tactical response rather than a strategic solution. Mossad veterans themselves have acknowledged the core problem: every person eliminated, no matter how senior, gets replaced. As the authors of a comprehensive study of Israel’s assassination campaigns concluded after reviewing half a century of evidence: “it is demonstrable that targeted killings are not the answer.”

Why Political Assassination Fails: Three Mechanisms

The pattern is too consistent to be coincidence. Three structural mechanisms explain why assassination almost always fails to achieve its intended goals.

The Martyrdom Effect. Dead leaders are more useful to movements than living ones. A living leader can make mistakes, compromise, disappoint. A dead leader becomes a symbol, frozen at the moment of maximum utility. Caesar dead was more powerful than Caesar alive. Lincoln’s assassination transformed him from a controversial wartime president into a national saint. Every assassinated leader since has followed the same trajectory: death confers a moral authority that life never could.

The Hydra Problem. Organizations that depend on a single individual are rare and getting rarer. Most political movements, insurgencies, and governments have redundant leadership structures, formal or informal succession plans, and ideological foundations that do not depend on one person’s charisma. A 2015 West Point study analyzing 758 political assassinations between 1946 and 2013 found that more than half of perpetrators had prior criminal experience, suggesting these were carefully planned operations, not impulsive acts. Yet despite this level of planning, the study found that assassinations generally intensify state fragmentation, undermine democratic institutions, and decrease political participation. The assassins plan meticulously and still make things worse.

The Complexity Trap. Assassins operate with a model of the world that is far simpler than the world itself. Princip assumed that removing an Archduke would weaken an empire. He did not model the alliance system, the mobilization timetables, the domestic politics of six great powers, or the role of military planning in constraining political decisions. Booth assumed that removing Lincoln would benefit the South. He did not model Reconstruction politics, congressional dynamics, or the difference between Lincoln’s pragmatism and his successor’s rigidity. Assassins consistently overestimate the importance of the individual target and underestimate the complexity of the system the target operates within.

The Exceptions That Prove the Rule

If assassination almost never works, does it ever work? The Jones-Olken data suggests a narrow window: when an autocrat rules through personal authority rather than institutional structures, and when the system that replaces him happens to be more open. But “happens to” is doing enormous work in that sentence. The assassin cannot control what fills the vacuum. Sometimes it is democracy. Sometimes it is a worse autocrat. Sometimes it is civil war.

The assassination of Rafael Trujillo in the Dominican Republic in 1961 is often cited as a case where assassination worked. Trujillo’s three-decade dictatorship ended, and the country eventually transitioned to democracy. But the transition took years, included a civil war and a U.S. military intervention, and the democracy that emerged was fragile. Whether the same outcome would have occurred without the assassination, perhaps through the regime’s natural decay, is unknowable.

This is the fundamental problem with the “assassination works” argument: you cannot run the counterfactualA historical or logical scenario that asks 'what if?' by imagining how events would have unfolded differently under different conditions. Historians use counterfactuals to explore the weight of specific decisions or events, though they cannot be proven.. You can observe what happened after an assassination, but you cannot observe what would have happened without it. The Jones-Olken study addresses this by using failed attempts as a control group, which is methodologically clever but still imperfect. The assassin who misses and the assassin who hits are operating in the same political context, but the contexts are not identical, and small differences can compound.

The Pattern Recognition

A national leader has been assassinated in nearly two of every three years since 1950, according to the Jones-Olken dataset. The frequency has not decreased. What has changed is the sophistication: from Princip’s pistol to Mossad’s operations to what the Institute for Security and Development Policy described as an AI-assisted remote-controlled weapon used to kill Iranian nuclear scientist Mohsen Fakhrizadeh in 2020. The technology improves. The strategic logic does not.

The pattern across twenty-one centuries of recorded assassination history is consistent enough to qualify as something close to a historical law: political assassination almost never achieves its stated objectives, frequently produces the opposite of the intended outcome, and consistently underestimates the complexity of the systems it targets. The assassin sees a chess piece to be removed. History sees a system that reorganizes around the removal, usually in ways nobody predicted.

This is not an argument for the insignificance of individuals in history. Leaders clearly matter, as the Jones-Olken economic research on natural deaths of leaders demonstrates. It is an argument that removing a leader by violence introduces so much chaos, generates so much martyrdom, and triggers so many unintended consequences that the net effect is almost always negative for the assassin’s cause. The bullet, as a tool of political change, has a track record roughly comparable to economic sanctions: satisfying to deploy, easy to justify, and almost never effective.

The Jones-Olken Dataset: What 298 Assassination Attempts Tell Us

The most rigorous quantitative study of political assassination was published in 2009 by Benjamin Jones (Northwestern/Kellogg) and Benjamin Olken (MIT) in the American Economic Journal: Macroeconomics. Their dataset covers every serious assassination attempt on a national leader from 1875 to 2004: 298 attempts, of which 59 succeeded (a success rate of roughly 20%).

The study’s methodological innovation was treating failed assassination attempts as a natural control group. Once a weapon is discharged, survival depends substantially on chance: the path of a bullet, the timing of an explosion, small shifts in a leader’s position. By comparing political outcomes after successful and failed attempts, Jones and Olken could isolate the causal effect of assassination from the political conditions that produced the attempt.

Their central findings:

• Successful assassinations of autocrats produce transitions toward democracy at a rate 13 percentage points higher than failed attempts on autocrats.
• Successful assassinations are 19 percentage points more likely to result in institutional leadership changes versus failed attempts.
• These effects persist a decade or more.
• In democracies, successful assassinations produce no measurable institutional change.
• Small-scale conflicts intensify following successful assassinations, while large-scale conflicts may end sooner.

The peak risk period for leaders was the 1910s, when assassination risk approached 1% per year. By the 2000s, individual risk had dropped below 0.3% per year, though absolute numbers of assassination events have increased due to the larger number of nation-states.

The West Point Dataset: 758 Attacks, 1946-2013

Arie Perliger’s study for the Combating Terrorism Center at West Point compiled 758 political assassination attacks by 920 perpetrators between 1946 and 2013, resulting in 954 deaths. This dataset is broader than Jones-Olken because it includes assassinations of non-head-of-state political figures.

Key findings from the CTC study:

• Members of parliament constituted 21% of targets. Opposition leaders: 18%. Heads of state: 17%. Ministers: 14%. Diplomats: 10%.
• 51.3% of assassins had prior criminal experience, indicating these are typically planned operations by experienced operatives.
• Political assassinations correlate with restricted political competition combined with high polarization, lack of consensual political ethos, and ethnically heterogeneous populations.
• Temporal concentration: South Asia saw 76% of its assassinations since the mid-1980s; Eastern Europe saw 85% after 1995.

On impact, the CTC study found that assassinations generally intensify state fragmentation prospects, undermine democratic institutions, decrease political participation, and disproportionately strengthen executive power. Assassinations of heads of state correlate with increased domestic violence and reduced democracy. Assassinations of opposition leaders show limited systemic impact but increased unrest.

Historical Case Analysis: The Backfire Mechanism

The quantitative findings are supported by case-level analysis across twenty-one centuries of recorded assassination history.

Julius Caesar (44 BC): The Liberatores killed Caesar to preserve the Republic. The assassination precipitated civil wars (44-31 BC) that ended the Republic and produced the PrincipateThe constitutional framework of the early Roman Empire (27 BC–284 AD), where emperors held supreme power while preserving the outward forms of republican institutions like the Senate. under Augustus, concentrating power to a degree Caesar had not achieved. The conspirators’ explicit goal, preserving republican institutions, was not merely unachieved; it was reversed.

Abraham Lincoln (1865): John Wilkes Booth’s stated motive was to avenge the Confederacy. Lincoln had been developing plans for a lenient Reconstruction. His death elevated Andrew Johnson, who lacked Lincoln’s political skill and congressional relationships. The resulting Reconstruction, shaped by conflict between Johnson and the Radical Republicans, produced outcomes that satisfied neither the former Confederacy nor the advocates of Black civil rights. Historians debate counterfactualsA historical or logical scenario that asks 'what if?' by imagining how events would have unfolded differently under different conditions. Historians use counterfactuals to explore the weight of specific decisions or events, though they cannot be proven., but the consensus is that the assassination did not advance any cause Booth supported.

Archduke Franz Ferdinand (1914): The Black Hand sought South Slav liberation from Austria-Hungary. The assassination triggered a cascade of alliance activations and mobilization timetables that the conspirators had not modeled. Austria-Hungary did eventually collapse, but so did the Russian, Ottoman, and German empires. The resulting Versailles settlement created conditions that led directly to a second, larger war. The Kingdom of Serbs, Croats, and Slovenes (later Yugoslavia) was formed, but spent much of the twentieth century under authoritarian rule and eventually dissolved in a series of wars in the 1990s.

Zhang Zuolin (1928): Officers of Japan’s Kwantung Army assassinated the Manchurian warlord to create a pretext for military intervention. The operation succeeded tactically: Japan seized Manchuria. But the assassination strengthened the Japanese military at the expense of civilian democracy, accelerated the trajectory toward war with the United States, and contributed to Japan’s total defeat in 1945.

Israel’s Targeted KillingThe military practice of eliminating specific adversary leaders or commanders through surveillance-guided strikes, as opposed to broader military campaigns. Campaign: Systematic Evidence

Israel has conducted the most sustained modern campaign of leadership decapitation targetingThe military strategy of systematically killing or capturing an adversary's senior leaders to destroy its command structure and render it unable to function., providing a large-N case study on whether systematic assassination produces better results than ad hoc attempts.

The evidence, summarized by former intelligence correspondents Yossi Melman and Dan Raviv, is that “after half a century, it is demonstrable that targeted killings are not the answer to the Israeli-Palestinian conflict.” Specific cases illustrate the pattern:

• Fathi Shiqaqi (Palestinian Islamic Jihad founder, killed 1995 in Malta): the organization subsequently grew larger and more dangerous.
• Khaled Meshaal (botched poisoning, Jordan, 1997): the operation’s failure transformed Meshaal into a Palestinian hero who rose to Hamas’s top echelon.
• Sheikh Ahmed Yassin and Abdel Aziz al-Rantisi (Hamas leaders, both killed 2004): Hamas increased its attacks and won the 2006 Palestinian legislative elections.
• Imad Mughniyeh (Hezbollah military chief, killed 2008 in Damascus): Hezbollah remains fully operational.

Former Mossad director Zvi Zamir characterized targeted killing as “a measure of last resort,” a tactical rather than strategic tool. The Institute for Security and Development Policy’s assessment is that targeted killings create “a cycle of violence, where groups seek revenge, potentially escalating conflicts,” and that collateral damage “can also create backlash against the state conducting the assassination, potentially fueling further violence and radicalization.”

Three Structural Mechanisms of Failure

1. The Martyrdom Amplification Effect. Assassination transforms leaders from fallible politicians into unassailable symbols. Caesar’s death created a cult that his adoptive heir weaponized. Lincoln’s assassination made him the moral center of American identity in ways his presidency alone would not have. Modern movements exploit this consciously: the killed leader’s image becomes recruitment material, and the act of assassination itself becomes evidence for the movement’s narrative of persecution.

2. Organizational Resilience (The Hydra Problem). Organizations with decentralized leadership structures, ideological rather than charismatic foundations, and formal or informal succession planning absorb leadership losses. Hamas has prepared for decapitation for decades: its cell-based structure, ideological training systems, and multiple simultaneous leadership tracks mean that eliminating one leader produces a replacement within the existing pipeline. The CTC study’s finding that assassinations disproportionately strengthen executive power suggests a related mechanism: the surviving institution centralizes authority as a defensive response, becoming harder to disrupt through subsequent attempts.

3. Complexity Underestimation. Assassins model the target, not the system. Princip modeled Habsburg vulnerability but not the alliance system. Booth modeled Lincoln’s role but not congressional Reconstruction politics. The Kwantung Army modeled Manchurian power dynamics but not the long-term consequences for Japanese civilian governance. This is a specific instance of a general problem in strategic thinking: interventions in complex systems produce non-linear effects that scale far beyond the intervention’s intended scope.

When Does Assassination Produce Its Intended Result?

The Jones-Olken data identifies a narrow set of conditions: personalist autocracies where power is concentrated in an individual rather than distributed across institutions. Remove the individual, and the institution built around them collapses. But the assassin cannot control what replaces it. The 13 percentage-point increase in democratic transitions is an average: it means some assassinations of autocrats produce democracy, and some produce chaos, civil war, or a new autocrat. The assassin is playing a lottery with someone else’s country.

The Dominican Republic under Trujillo (assassinated 1961) eventually transitioned to democracy, but only after years of instability, a civil war, and U.S. military intervention in 1965. Whether this counts as “the assassination working” depends on what you count as the intended goal and how many years of civil conflict you are willing to accept as transition costs.

Conclusion: A Near-Universal Pattern

Across 298 documented attempts since 1875, 758 political assassinations since 1946, and case studies stretching back to the Roman Republic, the pattern is consistent: political assassination almost never achieves its stated objectives and frequently produces the opposite of the intended outcome. The technology evolves, from a pistol in Sarajevo to what has been described as an AI-assisted weapon in Tehran, but the strategic logic remains flawed for the same structural reasons. The assassin models the target. History models the system. The system, being complex, reorganizes in ways nobody predicted.

This is not a moral argument. It is an empirical one. Assassination fails not because it is wrong (though it is) but because political systems are more complex than the people who try to change them by removing one piece. The bullet is to political change what economic sanctions are to economic change: intuitively appealing, politically satisfying, and almost always strategically futile.