This article proposes two novel estimation and inference approaches for production frontiers based on extreme quantiles of feasible outputs. The first approach linearly combines two extreme quantiles to reduce the estimation bias, and uses a subsampling method to construct point estimates and confidence intervals. The second approach can accommodate any finite number of extreme quantile estimates by way of the Approximate Bayesian Computation method. The point estimators and confidence intervals are then obtained through the Markov Chain Monte Carlo algorithm. The estimations and inferences of both approaches are justified asymptotically. Their finite sample performances are illustrated through simulations and an empirical application.
Fixed-k asymptotics, extreme value theory
YANG, Thomas Tao and ZHANG, Yichong.
Simulation-based Estimation and Inference of Production Frontiers. (2017). 1-59. Research Collection School Of Economics.
Available at: http://ink.library.smu.edu.sg/soe_research/2119
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.