The coup in Zimbabwe in November ended the second-longest interlude between coup attempts worldwide in modern history. Nearly 500 days passed between the failed putsch in Turkey in July 2016 and last month’s move against President Mugabe. The longest interlude occurred a decade earlier, when more than 600 days passed between successful coups in Fiji in December 2006 and Mauritania in August 2008.

Those two data points suggest that lulls between coup attempts have grown longer, and thus that coups have become rarer. A careful look at the complete event set bears out that impression.(1) The bar chart below plots the interludes between coup attempts worldwide since 1950. Instead of marking the location of those events in time, the vertical bars mark the time between those events in sequence, and the overlaid line summarizes change across history in the length of those bars. The labels on the horizontal axis identify the earliest and most recent events along with every hundredth event in between.

coup.waittime.bar.line-1

As the chart shows, the frequency of coup attempts remained fairly constant from the end of the colonial era circa 1960 until the collapse of the Soviet Union in 1991. Over the ensuing 35 years, the central tendency in the waiting time between coup events has grown from less than two months to almost half a year.

Those shifts in the international system have not rendered coups obsolete, but they have helped make them rarer.

The location of the inflection point in that curve is not an accident. The disintegration of the USSR ended the Cold War rivalry that drove both sides to abet or even catalyze coup attempts against governments sympathetic to their geopolitical foes. In the post–Cold War period, international norms favoring democracy and opposing usurpations of power have also strengthened, and regional associations and foreign donors are now more likely to speak and sometimes even act against military interventions in politics. Those shifts in the international system have not rendered coups obsolete, but they have helped make them rarer.

Looking ahead, we can’t foresee precisely when the next coup attempt will occur, but we can make an educated guess. If we treat the waiting times between previous coup attempts as a time series, we can use a statistical model suited to that type of data to forecast how long the current lull will last.(2) That exercise points to early May 2018, with a 50-percent prediction interval spanning from March 28 to June 11. That particular prediction may prove wrong, but the size of the expected gap between events underscores the broader trend toward longer lulls.

NOTES

  1. The data used in this analysis come from Professors Jonathan Powell and Clayton Thyne, who define coups as “overt attempts by the military or other elites within the state apparatus to unseat the sitting head of state using unconstitutional means.” Those attempts do not have to result in any deaths to count as coups, and a coup is considered successful if its perpetrators take and hold power for at least seven days. See their website for more details.

  2. We used the ‘forecast’ package in R to fit a univariate time-series model to the data on waiting times between coup events and then to generate our forecast.

Jay Ulfelder

Political science Ph.D. (Stanford 1997), research director for the US government-funded Political Instability Task Force (2001-2011), Good Judgment Project superforcaster.