EFFICIENCY VS. HETEROGENEOUS TECHNOLOGIES IN THETURKISH BANKING: 1990-2000

Mahmoud A. EL-GAMAL (Rice University)
Hulusi İNANOĞLU (Rice University)

In this paper, we model unobserved heterogeneity in banking technologies as a mixture model and investigate the efficiencies of 53 Turkish banks using likelihood based stochastic frontier analysis for the 1990-2000 period. The measure of "inefficiency" may be corrupted due to unobserved heterogeneity in banking technology. Therefore we estimate a mixture model using the EC (Estimation-Classification) estimator, and its companion EC-algorithm, which were introduced in (El-Gamal and Grether, 1995) to obtain data-driven classification and estimation of banking technologies.

In contrast to previous efficiency studies which tend to draw conclusions on the estimates of pooled or arbitrarily classified banks, the data-driven EC classification points mainly to distinctions between small and foreign banks vs. large and domestic ones, respectively. We then investigate the inefficiency scores with respect to the two different cost frontiers recovered by EC. Surprisingly, we do not find evidence of heterogeneity between state and private banks. Moreover, contrary to common wisdom, we find that state banks are not significantly less efficient than private banks.