We develop a nonparametric test of productive efficiency that accounts for the possibility of errors-in-variables. The test allows for statistical inference based on the extreme value distribution of the L?? norm. In contrast to the test proposed by Varian, H (1985): 'Nonparametric Analysis of Optimising Behaviour with Measurement Error, Journal of Econometrics 30, 445-458, our test can be computed using simple enumeration algorithms or linear programming. An empirical application for the Dutch electricity sector illustrates the proposed test procedure.

data envelopment analysis (DEA), errors-in-variables, extreme value theory, hypothesis testing, nonparametric production analysis
Semiparametric and Nonparametric Methods (jel C14), Corporate Finance and Governance (jel G3), Business Administration and Business Economics; Marketing; Accounting (jel M)
Erasmus Research Institute of Management
ERIM Report Series Research in Management
Copyright 2001, T. Kuosmanen, G.T. Post, This report in the ERIM Report Series Research in Management is intended as a means to communicate the results of recent research to academic colleagues and other interested parties. All reports are considered as preliminary and subject to possibly major revisions. This applies equally to opinions expressed, theories developed, and data used. Therefore, comments and suggestions are welcome and should be directed to the authors.
Erasmus Research Institute of Management

Kuosmanen, T, & Post, G.T. (2001). Testing for Productive Efficiency with Errors-in-Variables: with an application to the Dutch electricity sesctor (No. ERS-2001-22-F&A). ERIM Report Series Research in Management. Erasmus Research Institute of Management. Retrieved from http://hdl.handle.net/1765/88