Unveiling Plant-Product Productivity via First-Order Conditions: Robust Replication of Orr (2022)

TL;DR

This paper evaluates the replicability of Orr (2022)'s method for estimating plant productivity, confirming its robustness across different samples and extended periods using reconstructed Indian machinery data from 2000-2019.

Contribution

It demonstrates the robustness of Orr (2022)'s productivity estimation method across various samples and time periods, validating its applicability beyond the original dataset.

Findings

Main productivity patterns are reproducible.

Results are robust to exclusion threshold variations.

Extended periods confirm the method's stability.

Abstract

We assess the replicability of Orr (2022)'s method for estimating within-plant productivity across product lines, which combines demand estimation with cost minimization. The original study uses input price shocks in other output markets as instrumental variables, with exclusion restrictions based on downstream purchase shares. Reconstructing the original dataset of Indian machinery producers from 2000-2007, we reproduce the main productivity patterns and demonstrate their robustness to variations in the exclusion threshold. The main results remain robust in extended samples (2010-2019, 2000-2019) when calibrating demand parameters to Orr (2022)'s 2000-2007 estimates, as estimation on these extended periods yields inadmissible demand systems.