Recovering cropping management practices specific production functions: clustering and latent approaches

Résumé En

Reducing the use of pesticides and more generally of chemical inputs is a topical issue for governments. Economists generally advocate taxation for reducing polluting input uses. While econometric models tend to show that the pesticides price elasticity is low, these models mostly consider short-term adjustments. Mid-term adjustments of variable input uses are expected to be larger as reducing such uses require farmers to change their cropping management practices (CMPs). CMP is a notion closely related to the economists' production functions and used by agricultural scientists for characterizing crop production technologies. Yet, data lacking on farmers' CMPs prevents direct empirical analyses of CMPs' performances and adoption processes. The main objective of this paper is to propose original approaches for identifying farmers' CMPs in farm accountancy panel datasets with cost accounting. We consider that each CMP is characterized by a specic production function and propose approaches for identifying farmers' CMPs and the related production functions either sequentially or simultaneously. We demonstrate the relevance of our approaches through an empirical application based on a French arable crop farm accountancy unbalanced panel dataset covering the 1998-2014 period. Albeit preliminary, our empirical results demonstrate that our approaches perform relatively well. Indeed, they enable us to identify two wheat CMPs used by farmers: a low input CMP and a high yielding CMP.