Séminaire Nancy – Esther Devilliers (INRAE)

Détails de l'événement :

“Accounting for Latent Cropping Management Practices Choices in Crop Production Models: a random parameter Hidden Markov Model Approach” with Obafémi Philippe Koutchadé, Alain Carpentier

Abstract: In this article, we account for cropping management practices (CMPs) in economists’ production functions to evaluate pesticide uses responsiveness to price changes in a context of heterogenous technology. CMPs being latent in most economists’ data sets, we consider a hidden Markov model to describe the dynamics of farmer’s CMP choice. We also account for farmers’ unobserved heterogeneity by considering a random parameter model for our production function. We use the SAEM-MCMC algorithm to estimate our mixed hidden Markov model with heterogeneous transition probabilities. An illustration on French winter wheat producers of La Marne area allows us to distinguish between high-yielding, intermediate and low-input practices.