The recursive impact in the multivariate probit model: An application on farmers’ decisions for opting risk management strategies

Abstract

This study investigates the determinants of farmers’ risk management decisions in Khyber-Pakhtunkhwa, Pakistan, using a recursive multivariate probit (RMVP) model. Employing data from 382 farmers collected through a multistage sampling process, the study compares the RMVP with the traditional multivariate probit (MVP) model, demonstrating the superior performance of the RMVP in capturing complex decision-making patterns. Our rigorous statistical analysis demonstrates the significant impact of endogenous covariates on farmers’ risk management choices, revealing complementarity or substitutability among strategies. The study contributes to the literature by providing empirical evidence on the effectiveness of the RMVP model for understanding smallholder farmers’ risk management behavior and offering insights for policymakers to support resilient agricultural systems.