A Bayesian zero-one inflated beta modeling of dairy farmers’ decision to sell nothing or some output through contract farming

Abstract Smallholder farmers’ preferences for participation in contract farming may take the form of proportional data––whereby farmers only sell some proportions or fractions their output to contractors. We analyze determinants zero (potential corner solution) and amounts milk that through farming, using dairy data from Zambia. Bayesian linear, linear probability, hurdle models are compared with a zero-one inflated beta regression. Monte Carlo simulations show alternative biased. Meanwhile, empirical findings suggest gender marital status household head, size, delayed payment significantly drive sales farming. Additionally, experience selling collection centers, total livestock units, access marketing information, buyer's price, among others, tend affect zero-inflated outcomes. recommend regression model also provide strategies overcome farmer-engagement barriers