Statistical Inference in Conditional Nonparametric Frontier Models

Relatore:  Cinzia Daraio - Universita' di Bologna
  venerdì 24 aprile 2009 alle ore 12.00 Biblioteca DSE, Palazzina 32 - Ex Caserma Passalacqua
In recent years the interest of researchers and practitioners on the explanation of efficiency
differential has grown both on the methodological side and on the application to economic
problems. Nonparametric conditional efficiency measures (Daraio and Simar, 2005, 2007) represent
a general and appealing approach to handle external-environmental factors avoiding unrealistic
assumptions such as the separability condition between the input-output space and the space of
external-environmental factors. This paper proposes a consistent bootstrap approach to correctly
mimicking the DGP and describes how to estimate the bias, the standard deviation and the
confidence intervals of conditional nonparametric measures of efficiency. We describe how to
compute confidence intervals on the ratios of conditional on unconditional efficiency measures to
improve the detection of the impact of external factors on the production process. Finally, we
propose a bootstrap based procedure to test for the impact of the environmental factors on the
performance of the production process. Some test statistics are proposed, compared and illustrated
using simulated datasets as well as real data.

Joint paper with L. Badin & L. Simar


Referente
Angelo Zago

Referente esterno
Data pubblicazione
10 aprile 2009

Condividi