Improving the productivity of the agricultural sector is part of one of the Sustainable Development Goals set by the United Nations. To this end, many international organizations have funded training and technology transfer programs that aim to promote productivity and income growth, fight poverty and enhance food security among smallholder farmers in developing countries. Stochastic production frontier analysis can be a useful tool when evaluating the effectiveness of these programs. However, accounting for treatment endogeneity, often intrinsic to these interventions, only recently has received any attention in the stochastic frontier literature. In this work, we extend the classical maximum likelihood estimation of stochastic production frontier models by allowing both the production frontier and inefficiency to depend on a potentially endogenous binary treatment. We use instrumental variables to define an assignment mechanism for the treatment, and we explicitly model the density of the first and second-stage composite error terms. We provide empirical evidence of the importance of controlling for endogeneity in this setting using farm-level data from a soil conservation program in El Salvador.

翻译：提高农业部门生产力是联合国设定的可持续发展目标之一。为此，许多国际组织资助了旨在促进发展中国家小农户生产力与收入增长、抗击贫困及增强粮食安全的培训和技术转让项目。随机生产前沿分析可成为评估这些项目有效性的有用工具。然而，考虑这些干预措施中通常固有的处理内生性，直到最近才在随机前沿文献中得到关注。本研究通过允许生产前沿和低效率均依赖于潜在内生二元处理，扩展了经典随机生产前沿模型的最大似然估计。我们使用工具变量定义处理的分配机制，并显式建模第一阶段与第二阶段复合误差项的密度。基于萨尔瓦多土壤保护项目的农户层面数据，我们提供了在此背景下控制内生性重要性的经验证据。