Anchored to Experience: How Hazard-Proximate Climate Metrics Shape Adaptation Choices Among Ghanaian Cocoa Farmers

• In Event: Clean Energy Transitions in Developing and Rural Contexts

Abstract

Climate adaptation policy assumes that providing smallholder farmers with better climate information will drive the adoption of adaptive technologies. This paper challenges that assumption and reframes the problem as one of information design: not whether climate data reaches farmers, but whether its structure maps onto the cognitive and experiential frames through which farmers evaluate risk.
We administered a Discrete Choice Experiment (DCE) to 578 smallholder cocoa farming households in Ghana, testing willingness-to-pay (WTP) for solar-powered irrigation under experimentally varied conditions of climate information format and loan structure. The DCE design was informed by Participatory Rural Appraisal (PRA) across 12 communities, which mapped farmers' risk mental models and identified the experiential benchmarks they use to evaluate climate threats.
Affordability acts as the binding constraint on adaptation, with loan cost and payback period dominating utility shifts across all specifications. Generic seasonal rainfall forecasts had no statistically significant effect on adoption probability. However, hazard-proximate metrics, specifically, consecutive dry days benchmarked against the previous growing season, significantly shifted utility, but only when anchored to recent lived experience. Latent Class Analysis reveals that a pooled estimate would mask this effect: a dominant segment (92.5% of treated farmers) exhibited significant WTP of GH₵ 309.51 for experientially anchored advisories, while a smaller segment, disproportionately female and lower-education, systematically opted out of all adaptation packages regardless of information design or financial terms, suggesting that the binding constraints for this group operate beyond the scope of the intervention.
These findings argue for bundling experientially anchored climate advisories directly with financial enablers rather than disseminating climate information as a standalone policy instrument. For the opt-out segment, neither information redesign nor improved financial terms shifted adoption, indicating that additional structural supports, such as targeted subsidies or asset-building programs, may be prerequisites for participation. More broadly, these findings expose the limits of uniform intervention design: adaptation policies that ignore gendered and educational disparities in baseline adaptive capacity risk widening the gaps they aim to close.

Author

• Carlos Calvo Hernandez, University of Notre Dame
  Presenting Author