Yesterday, protesters in Armenia accomplished a rare thing: they compelled a sitting national leader to step down. Around the world, popular protests are common, but they usually don't swell into sustained movements—especially when met with repression—and they rarely succeed in pushing longtime leaders out of office.

So, will the departure of Serzh Sargsyan open the door to freer and fairer elections in Armenia, as protest leader Nikol Pashinyan has demanded? Koto’s Analogy Engine (AE) can quickly give us an empirical take on that question. AE makes it easy to find historical cases of leadership change under structural conditions similar to those in Armenia today, and then to see how they trended after that event. The distribution of post-event outcomes in similar cases can then be read as a forecast of sorts for Armenia.

The results should give Armenian protesters cause for optimism. Using GDP per capita (IMF), level of democracy (V-Dem), level of corruption (V-Dem), leader’s time in office (REIGN), and the occurrence of a leadership change (REIGN) as the basis for comparison, the Analogy Engine identified a number of cases from the past 30 years similar to Armenia today. The chart below shows how the level of democracy changed over the preceding and ensuing five years in the 10 most similar cases, which came from all around the globe.

  • Five years after their leadership changes, all but one of the most similar cases had experienced some improvement in its degree of democracy, and none saw a decline.
  • The gains ranged from a hair’s breadth to a substantial 0.25 on a scale from 0 to 1, with a median increase of approximately 0.1.
  • In the cases where significant gains occurred, those gains mostly came within two years of the leadership change.


In its most recent update, Varieties of Democracy (V-Dem) gave Armenia a score of approximately 0.4 on that electoral democracy index. From that starting point, an improvement in the neighborhood of 0.1 could be enough to move Armenia back to the “mostly democratic” side of the ledger. Past performance in similar cases is no guarantee of future outcomes in this one, but the historical record is certainly encouraging in this instance, and that’s not always the case.

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.