Latent Class Modelling for a Robust Assessment of Productivity: Application to French Grazing Livestock Farms[Modélisation des classes latentes pour une évaluation robuste de la productivité : Application aux exploitations françaises d'élevage herbager]

Our objective is to extend the latent class stochastic frontier (LCSFM) model to compute productivity change, using the robust transitive productivity Fare-Primont index. The application is to three types of grazing livestock farms in France over the period 2002-2016. The LCSFM identified two classes of farms, intensive farms and extensive farms. Results indicate that productivity change and its components show only small differences between the LCSFM and the pooled model that does not account for heterogeneity. Differences across classes exist, but depend on farm type.