2. Data and Methodology
2.1. Methodology
The efficiency estimates, which can be measured by applying frontier analysis, reflects the
degree of proximity of the firms/banks to a best-practice frontier. Frontier analysis provides an overall
and objective numerical efficiency values and ranking of the firms (Berger and Humphrey, 1997).
Among different type of estimation methodologies – non parametric or parametric techniques2, the
efficiency measures in this study have been estimated by using Stochastic Frontier Approach (SFA
hereafter), one of the most widely applied parametric technique. Even if Data Envelopment Analysis
(DEA) as one of the non-parametric approaches , requires fewer assumptions, less data and less
sample, it does not allow for a random error and measurement error in the construction of the frontier,
which may lead to severe problems in shaping and positioning the frontier. Furthermore, conventional
test of hypothesis associated with the existence of inefficiency and the structure of the production
technology can not be conducted with DEA (Coelli et al., 2005). Therefore, due to the drawbacks
associated with DEA, this study employs the parametric stochastic frontier approach (SFA) to
establish the cost and profit efficiency frontiers of the banks.