From Climate-Finance Approval to Measurable Resilience in Latin America and the Caribbean: Dynamic Panel Evidence and Small-Sample Robustness

Multilateral climate-finance approvals may support resilience, but approval, disbursement, execution, and measurable national outcomes occur at different stages. This study examines associations between country-specific multilateral approvals and climate vulnerability in a 20-country Latin American and Caribbean panel for 1995–2024. The preferred dynamic two-way fixed-effects models use 333 observations from 18 countries during 2006–2024, common samples, Driscoll–Kraay inference, wild cluster bootstrap tests, and a half-panel jackknife. To avoid mechanical overlap, total finance is first estimated alone and then decomposed into mutually exclusive adaptation and other-finance components. Total approvals are statistically imprecise. The component model also produces imprecise estimates until controls and institutional interactions are added; in that specification, adaptation finance and its interaction are positive, but the adaptation coefficient weakens under wild cluster inference, and both terms lose precision after jackknife correction. Results are stable to alternative starting years, bandwidths, and 20-country models that omit the controls responsible for missingness. Vulnerability remains strongly persistent. The findings do not demonstrate that climate finance is ineffective; they show that annual approval commitments, especially when sparse and upstream from implementation, cannot reliably identify realized resilience effects at the national level.