Killing a leader doesn't always work
May 17th 2007
From The Economist print edition

FRANZ FERDINAND found alive—first world war a mistake! Whoever wrote that graffiti must have shared the view that assassinations can change history. A new academic study* tries to analyse whether that impression is correct. It looks at 298 assassination attempts between 1875 and 2004, from the murder of Spain's Canovas in 1897 to that of Pakistan's General Zia in 1988.

The bad news for would-be Lee Harvey Oswalds is that the success rate is pretty poor. Only 59 of the 298 attempts resulted in the target's death, a hit rate of one in five. If you are planning to knock off a leader, it is far better to use a gun. Assassination attempts using firearms had a kill rate of around 30%, whereas those using bombs or other explosive devices worked only 7% of the time (but often harmed bystanders).

The good news for political leaders is that their chance of surviving office is improving. While the annual number of assassination attempts has been increasing (currently around one every two years), there are a lot more countries (and thus more leaders) than there were in the early 20th century. In the 1910s, a given leader had a 1% chance of being killed in any given year; today, the probability is around 0.3%.

Killing leaders does make a difference, but only in certain circumstances. Democracies seem to be able to cope with the loss of a president or prime minister. But in autocracies, a successful assassination was 13 percentage points more likely than a failed attempt to result in a transition to democracy. The “beneficial” impact of an autocrat's removal was still observable ten years later. However, given the low success rate of assassins, the gamble might not be worth it; autocrats who survived an attack tended to tighten their grip.

When it comes to wars, the effects are more subtle. Assassinations tend to hasten the end of intense wars (those with more than 1,000 battle deaths) but, prior to the second world war, made small-scale conflicts more intense. As for initiating conflicts, the academics find assassinations are irrelevant. So even if Franz Ferdinand had dodged the bullet, the first world war might have happened anyway.