Estimate: ~15% (range 10–20%)
Base rate. Forced irregular removal of authoritarian leaders runs roughly 2–4% per leader-year (general literature estimate, not verified against Archigos this session). The question’s disjunction — flee or cease to be president — covers the full removal distribution. Starting prior: ~5%.
What justifies a premium over base rate:
Three stressors are converging simultaneously. Maduro’s capture (January 2026) severed Cuba’s last major oil patron; imported oil supply dropped an estimated 75–90% within weeks, with Venezuelan supplies fully cut and Mexican supplies halted by late January. The US is running a targeted fuel blockade via EO 14380 (January 29, 2026), authorizing tariffs on any country supplying Cuba. GDP has contracted over 15% since 2020 per Cuba’s own CEEC, with three consecutive years of negative growth. Population loss stands at roughly 10% by government figures (HRW), possibly 18% per independent estimates. Infrastructure is visibly failing: 20-hour blackouts, airports out of jet fuel, schools shuttered.
Discounts: Cuba survived comparable deprivation during the Special Period (1991–95); the US has pursued regime change for 65 years without success; authoritarian regimes persist at very low equilibria.
The mechanism problem — and why it caps the estimate:
Bottom-up revolution is improbable: no organized opposition, island geography, intact security apparatus. The likeliest pathway is internal PCC substitution — Díaz-Canel is the first leader without revolutionary legitimacy, making him uniquely expendable.
However, as of February 27, 2026, there is no documented evidence of internal fracture. No public criticism from senior officials, no separation of his dual roles, no credible leaks. He was re-elected February 24 with 459/460 votes. Raúl Castro’s December 2025 postponement of the 9th Party Congress reinforced old-guard authority over Díaz-Canel, not against him.
Novel pressure vector not in earlier drafts: On February 25, the US licensed Venezuelan oil sales to Cuba’s private sector, excluding government and military — designed to empower non-state actors. This could erode state control without triggering a leadership exit, creating a “zombie state” scenario the question may not capture.
Wildcard: Cuban forces killed four passengers from a Florida-registered boat on February 26–27; US citizens reportedly among the dead. If confirmed, this could shift US posture faster than any economic variable.
Update triggers:
| Signal | Revised estimate |
|---|---|
| Senior PCC figures publicly distancing | 25% |
| Military refusing to suppress major protest | 35%+ |
| Credible reporting of succession deliberations | 20–25% |
| Alternative oil supply secured | 8% |
| US attention shifts, sanctions enforcement relaxes | 8% |
Prediction markets. Polymarket prices exit by year-end at ~55–65%; Kalshi 45–55%. The multi-outcome “Next leader out” contract has $512K+ volume on Díaz-Canel; standalone Cuba contracts remain thin ($6K–$71K) and likely diaspora-heavy. The gap between 15% and market consensus reflects disagreement about whether unprecedented simultaneous pressure crosses a threshold base rates can’t capture — not a factual dispute.
Highest-leverage unknown: Military loyalty. GAESA controls ~40% of GDP. No evidence of US contact with military leadership or internal military dissent.
cafebedouin.org 