Title of article
Frontier estimation in the presence of measurement error with unknown variance
Author/Authors
Kneip، نويسنده , , Alois and Simar، نويسنده , , Léopold and Van Keilegom، نويسنده , , Ingrid، نويسنده ,
Pages
15
From page
379
To page
393
Abstract
Frontier estimation appears in productivity analysis. Firm’s performance is measured by the distance between its output and an optimal production frontier. Frontier estimation becomes difficult if outputs are measured with noise and most approaches rely on restrictive parametric assumptions. This paper contributes to nonparametric approaches, with unknown frontier and unknown variance of a normally distributed error. We propose a nonparametric method identifying and estimating both quantities simultaneously. Consistency and rate of convergence of our estimators are established, and simulations verify the performance of the estimators for small samples. We illustrate our method with data on American electricity companies.
Keywords
Nonparametric estimation , Penalized likelihood , Deconvolution , Stochastic frontier estimation
Journal title
Astroparticle Physics
Record number
2042228
Link To Document