ABSTRACT
This study generalizes the test performed by Simar and Zelenyuk (2007) to examine differences in the technical efficiency among groups within an industry (where ). For this purpose, the groups are divided into pairs and each group is compared with all other groups. The groups can then be classified into three cohorts: those performing better, equally and worse, relative to the benchmark group. For illustration purposes, annual data for Vietnamese banks covering the period 2005‒2012 is used.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 Our DEA model is specified by the conditions of output orientation and variable returns to scale.
2 For details on subsampling, see Simar and Zelenyuk (Citation2007).
3 The algorithm of the subsampling bootstrap process for mean DEA efficiency scores is the same with equal weights.
4 The MATLAB codes used by Simar and Zelenyuk (Citation2007) for an industry with two groups were modified for an -group industry.
5 At any significance level, if the lower bound is smaller than unity, while the upper bound is greater, the efficiency of the groups being compared will be equal.