Ranking, Not Point Estimates: Risk Tiering and the Case for AI-Assisted Loss Data in Caribbean SIDS Credit (World Journal of Entrepreneurship Management and Sustainable Development)

Summary

• Official lending rates across the Caribbean range from 5% to 15%, with public debt burdens exceeding 100% of GDP in several states.p.2
• Disaster losses in small states average 13% of GDP and impact 10% of the population, while chronic climate impacts consume 2% of aggregate capital stock annually.p.2
• First-order stochastic dominance holds for 100% of top-tier versus bottom-tier territory comparisons, confirming clear separation between high- and low-risk groups.p.8
• A single 1988 report of US$7 billion in losses from Hurricane Gilbert in Jamaica represents 64.8% of total recorded losses in the Caribbean regression sample.p.12

Key insights

1. Territory risk tiering is identifiable and stable across damage-function calibrations, whereas point loss magnitudes remain imprecise under sparse records.p.1
2. Official lending rates in Caribbean SIDS range between 5% and 15%, driven by high hazard exposure and public debt exceeding 100% of GDP.p.2
3. Small states experience average disaster losses of 13% of GDP, compared with roughly 1% in larger economies.p.2
4. First-order stochastic dominance holds for 100% of top-versus-bottom risk tier pairs, but for only 20.6% of within-tier pairs.p.8
5. The Caribbean wind-loss regression coefficient of 0.0502 is half the US estimate of 0.0992 and lacks statistical precision under wild cluster bootstrap testing.p.10
6. Excluding a single unverified US$7 billion record for Hurricane Gilbert reduces the Caribbean wind-loss coefficient from 0.0502 to 0.0376.p.12
7. AI-derived damage mapping from pre- and post-disaster satellite imagery can rapidly bridge the institutional loss data gap in Caribbean SIDS.p.13