糖心TV

Job Market Paper

Estimating Multi-Product Production Functions: What Can We Learn Without Demand Assumptions? [paperLink opens in a new window] [presentationLink opens in a new window]

Abstract: I prove that, when the demand side is unrestricted, production functions for multi-product firms are unidentified, except in population if the conditional time-series variance of inputs is unbounded. This result calls into question the robustness of existing demand-side-based estimators. I develop a novel identification strategy that does not rely on demand-side assumptions. Instead, by imposing the weaker assumption that the productivity distribution is in stationary equilibrium, I show that the production function parameters are set-identified. This approach avoids the need for instruments and is easy to scale with the number of outputs, providing a widely applicable method for estimation and conduct testing.

Work in Progress

• "Testing for Non-Jointness and Homogeneity with Non-Separable Production Frontiers"
• "Partition Dependent Expected Utility" (with ) [paperLink opens in a new window]