A0498
Title: Measurement matters: Challenges for robustness in two-stage efficiency frameworks
Authors: Lukas Fryd - Prague University of Economics and Business (Czech Republic) [presenting]
Ondrej Sokol - University of Economics, Prague (Czech Republic)
Abstract: Two-stage efficiency analysis, commonly used in the evaluation of agricultural policies, typically involves estimating technical efficiency in the first stage of ten through data envelopment analysis (DEA) or stochastic frontier analysis (SFA) and then examining the impact of various policy or structural variables on the estimated efficiency scores in the second stage. Using farm-level data, different approximations of production inputs and outputs are demonstrated to lead to considerable variation in estimated technical efficiency. This heterogeneity further propagates into the second stage, where the estimated determinants of inefficiency or efficiency can vary significantly even under nearly identical input-output approximations. The issue persists across different model specifications and distributional assumptions, including half-normal, truncated, and exponential inefficiency terms in SFA models as well as DEA model. This variability has profound implications: It undermines the robustness of policy recommendations derived from two-stage analyses and challenges the comparability of empirical results across studies. Moreover, it raises concerns about the potential for intentional or unintentional manipulation of factor approximations to influence analytical outcomes.
