Identifying factor productivity from micro-data: The case of EU agriculture

We examine the plausibility of four established and innovative identification strategies for agricultural production functions using farm-level panel datasets from five EU countries. Newly suggested proxy and dynamic panel approaches provide attractive conceptual improvements over received Within and duality models. Even so, empirical implementation of such advancements does not always live up to expectations. This is particularly true for the dynamic panel estimator, which mostly failed to identify reasonable elasticities for the (quasi-) fixed factors. Less demanding proxy approaches represent an interesting alternative for agricultural applications. In our EU sample, high production elasticities for materials prevail. Hence, improving the availability of working capital is the most promising way to increase agricultural productivity.