Non-Parametric Tests for Firm Efficiency in Case of Errors-in-Variables

textabstractThis paper develops a novel statistic for firm efficiency called efficiency depth that allows for statistical inference in case of errors-in-variables. We derive statistical tests that require minimal statistical assumptions; neither the sample distribution nor the noise level is required. An empirical illustration for European banks illustrates that - despite the minimal assumptions- the tests can have substantial discriminating power in practical applications