Research Master Pre-Defense
On the Identifiction of Production Functions
-
SeriesResearch Master Defense
-
SpeakerJorge Cal Medina
-
LocationAmsterdam
-
Date and time
July 17, 2024
14:00 - 15:30
| In this paper, we examine the identification strategies employed in the control func- tion approach to estimate production coefficients. Our objective is to first, shed light on identification issues with a special emphasis on weak identification and second, to formulate new practical guidelines for practitioners to address them. We uncover significant propensity for weak identification. Particularly, we find that the choice of proxy plays an important role for the identification strength of production coefficients, tying in with previous literature. We also find that per- sistence in production input data complicates estimation, which is a well-known issue in the dynamic panel literature. Further, we find that commonly used esti- mators are not robust to weak identification. This is problematic since they deliver seemingly reasonable results while being misguided. We propose a version of those estimators based on the GMM framework to improve the reliability of their empir- ical results. The GMM framework further allows us to derive diagnostic tests to gauge the relevance of weak identification issues. |