NEW YORK • If the attempted coup in Turkey came as a surprise, there was a good reason: The event went against decades of research on how, when and why coups happen.
Last Friday night's uprising appeared to diverge wildly from the usual patterns. And political scientists who study coups say Turkey should have been at little risk.
NOT THE KIND OF COUNTRY AT RISK
Coups are usually driven not solely by individual plotters, but also by structural factors. Political scientists, by tracking factors like economic trends, political freedoms and public health, have identified several predictive patterns.
Mr Jay Ulfelder, who works in the area of political forecasting, has developed a mathematical model that synthesises this data to predict a country's level of risk.
Turkey, according to Mr Ulfelder's research, done in conjunction with the Early Warning Project, was a "very unlikely" candidate for a coup. It had only a 2.5 per cent probability of an attempted coup, based on 2016 data.
That placed it 56th out of 160 countries, between Laos and Iran, and was within a range considered stable. At-risk countries tend to have high rates of infant mortality, a common measurement of poverty, and poorly performing economies. Turkey's economy has been growing and its infant mortality has been rapidly declining.
Another crucial factor is what experts call elite fragmentation. If divisions open up among powerful elites - elected officials, business leaders, generals, judges and so on - their competition for resources and control can culminate in a coup.
There is, as yet, no sign of such a split in Turkey. The growing economy gives elites reason to maintain the status quo. Turkey also lacks the kind of social polarisation that disaffected elites often exploit to push forward a coup.
NOT WHAT A COUP LOOKS LIKE
Successful coups tend to be waged as "coordination games", Professor Naunihal Singh at the Air War College wrote in the book Seizing Power, which examines why coups succeed or fail.
Coups work, according to this theory, when leaders convince other officers and soldiers that their success is assured. That makes joining an act of perceived self-interest.
Plotters usually accomplish this with a predictable set of steps. A massive show of force demonstrates that the weight of the military is behind the coup.
A public statement by one or more high-level public officials shows that there is elite support. And the plotters usually establish tight control over the media and the flow of public information.
This time, Turkey's dissident officers tried to take only some of these steps, and succeeded in none of them. Rebels deployed tanks and air power in a show of force, but this was not enough to intimidate the rest of the military, which eventually overcame them.
Most notably, there was no public face of the coup to demonstrate elite support or issue a clear plan. The insurrectionists also tried but failed to control communication with the public. Mr Erdogan was able to use the FaceTime smartphone app to call a TV station, a bizarre scenario that risked the appearance of weakness but also sapped the plotters' momentum and allowed him to call for public protests.
Pro-government protests - and the lack of visible pro-coup crowds - may have also been key to thwarting the plot. Successful coups will often exploit or even coordinate with pre-existing movements, using this show of popular support to rally elites. Turkey's 1997 coup leaders, for example, worked with civil society groups and others who opposed the government. Leaders of the weekend coup attempt appeared to lack any allies.
NEW YORK TIMES