A lot of great points from the other forecasters, such as @DKC mentioning Cameroon, here.
A long-term base rate visualization is at VOA News. I haven't crunched the numbers, but from looking it, maybe 30%-35% of 6-month periods since the turn of the 21st century have had successful coups.
Good article here from WarOnTheRocks discussing various factors involved, which may boost that base rate, for example:
Moreover, the recent withdrawal of French and European troops, who had been at the forefront of counter-terrorism efforts in the region, coincides with the emergence of private military companies. These companies, operating with less transparency and accountability, may exacerbate the existing security vacuum. Simultaneously, the “reckless exploitation of natural resources” by external actors and corrupt local leaders threatens the region’s stability. This volatile combination is likely to foster increased instability, potentially leading to more coup attempts. Furthermore, such circumstances could inadvertently strengthen populist support for regimes that emerge in the aftermath of these coups.
There are also a ton of uncertainties (since it's not like we know what plots are underway). Also, it's notable that the resolution criteria require a successful coup. This is a strong argument for going with the long-term base rate; however, I want to nudge it upward due to factors such as the recent withdrawal of French and European troops.
@ctsats That's perfect, thanks. I'm glad you did that work. Yeah, I think this is one where the base rate has to hold a lot of weight. One thing I'm curious about, if you start it with 2000 as your first year, what answer do you come up with for the base rate?
Using the same (simple...) Poisson distribution fit with only the last 24 data points (2000-2023), the results come up somewhat lower: ~1.0 successful attempts per year, with a probability of ~65% for at least one successful coup on a calendar year (12 months), i.e. closer to your guess above of 30-35% per 6-month period.
(A reminder that the respective numbers for the 1960-2023 data are 1.65 and 81%).
A lot of great points from the other forecasters, such as @DKC mentioning Cameroon, here.
A long-term base rate visualization is at VOA News. I haven't crunched the numbers, but from looking it, maybe 30%-35% of 6-month periods since the turn of the 21st century have had successful coups.
Good article here from WarOnTheRocks discussing various factors involved, which may boost that base rate, for example:
There are also a ton of uncertainties (since it's not like we know what plots are underway). Also, it's notable that the resolution criteria require a successful coup. This is a strong argument for going with the long-term base rate; however, I want to nudge it upward due to factors such as the recent withdrawal of French and European troops.
@johnnycaffeine I have crunched the VoA numbers here, just in case you are interested: https://www.randforecastinginitiative.org/comments/139379
FWIW, they led me to an initial forecast identical to yours here :)
Using the same (simple...) Poisson distribution fit with only the last 24 data points (2000-2023), the results come up somewhat lower: ~1.0 successful attempts per year, with a probability of ~65% for at least one successful coup on a calendar year (12 months), i.e. closer to your guess above of 30-35% per 6-month period.
(A reminder that the respective numbers for the 1960-2023 data are 1.65 and 81%).