A review of econometric approaches for estimating input allocation in agriculture: a focus on strengths and weaknesses

Understanding input allocation—the process of distributing aggregated inputs across production activities (Britz, 2004)—plays a central role in agricultural economics. However, it remains a challenge, particularly in multi-output systems where firm-level input data is often aggregated. Traditional survey-based approaches are costly and often impractical, highlighting the need for more effective methods. In this context, this paper reviews four main econometric approaches to input allocation: classical regression models, random parameter methods, entropy-based techniques, and Bayesian estimation. Each method is assessed in terms of its theoretical foundations, strengths, and limitations. Particular focus is given to two recent methodologies—SAEM-based random parameter modeling and Highest Posterior Density estimation—which address key issues of econometric techniques of input allocation. By comparing these approaches, this study provides practical guidelines to help users choose the most suitable method based on relevance, simplicity, and computational strength. Finally, potential extensions, including the integration of machine learning techniques into cost allocation models, are discussed for future research purposes.