multi - directional efficiency analysis: a case study of iranian banking industry

in very recent papers asmild et al. (2003) and halved et al. (2004) based upon potential improvement efficiency index (bogetoft and hougaard,1999) have suggested an alternative nonparametric frontier approach namely multi-directional inefficiency analysis (mea) to estimate the performance efficiency. mea has many advantages over data envelopment analysis (dea) as the standard nonparametric frontier method to performance evaluation. mea could provide more reliable efficiency measures because it is able to assess the relative improvement potentials for each input or output type separately. this paper provides the first performance study over banking industry applying multi-directional efficiency analysis for a dataset of iranian banking system over 2004-2007. the findings of this paper support holvad et al. (2004) in which dea models over-underestimate the input excesses. moreover,, measuring the sizes and directions of potential improvement for each input separately across the banks and years, provides useful information from managerial and policy implications point of view. one of the most interesting points of these findings perhaps is that for the iranian sampled banks, generally, labour force, on average, accounts for the largest proportion of excess costs. thus, the labour reduction has the first priority in the path of efficiency improvement. jel: c01, c02, c12, c42