Using a multivariate probit method to model decisions on cocoa farm rehabilitation in Ghana

This study examines cocoa farm rehabilitation decisions among smallholder farmers in the Oti Region of Ghana, where productivity has declined due to aging tree stocks and disease pressures. Using cross-sectional survey data from 241 farmers and a multivariate probit (MVP) model, the study jointly analyzes adoption of four rehabilitation strategies: coppicing, phased replanting, planting seedlings under old trees, and complete replanting. The likelihood ratio test rejects independence of decisions, indicating statistically significant correlations in unobserved factors across strategies. However, only 12% of farmers adopt all four practices, suggesting that selective combinations are more common. The results show that rehabilitation adoption is significantly associated with farmer characteristics (age, gender, education, household labor), farm size, and institutional factors including cooperative membership, access to planting materials, training, information, record keeping, and off-farm activities. Age is positively associated with coppicing, phased replanting, and under-planting, while education is positively associated with complete replanting but negatively associated with some other strategies. Gender shows negative associations with phased replanting and under-planting. Farm size, cooperative membership, access to planting materials, record keeping, training, and off-farm activities also display differentiated associations across rehabilitation options. These findings indicate that rehabilitation choices vary across strategies rather than following a uniform adoption pattern. Policy efforts that improve access to planting materials, strengthen strategy-specific extension support, and address gender and labor constraints may better align with observed rehabilitation behavior. This study contributes to the literature by jointly modelling rehabilitation strategies and providing evidence from Ghana’s cocoa sector.