There has been a considerable amount of research on banking that addresses questions of efficiency. The banking industry poses a complex set of issues with respect to the environment in which it operates as well as the nature of its production process. Many questions remain unanswered in the literature. Does the size of banks affect levels of X-efficiency, and why? Are banks becoming more or less efficient over time? What explains the varying levels of the X-efficiencies across banks, and what role does productivity change play?;This dissertation examines these issues empirically using data on the Korean banking sector, with a stochastic frontier approach. A major innovation of this dissertation concerns the treatment of the error terms in the stochastic frontier models in order to correct for size-related heteroscedasticity, an issue that is often ignored in the literature. We first estimate the levels of X-(in)efficiency for individual banks, employing cost and profit functions. These estimates are then used to identify other bank characteristics that affect the levels of efficiency. The third section examines productivity changes and their decomposition.;The overall results suggest that large nationwide banks in Korea are the most efficient, as is indicated by the significant relationship between a bank's market share and its cost efficiency and one form of its profit efficiency. All banks in the sample display much higher levels of cost efficiency than profit efficiency. The overall higher levels of efficiency for large nationwide banks do not appear to lead to higher rates of innovation, as expressed through productivity change. The rate of productivity change is the highest for the small nationwide banks, the group that are the worst performers in the efficiency estimations. The results also indicate that the rate of change in total factor productivity (TFP) is inversely related to the number of years that a bank has been in operation, while bank levels of efficiency are positively related to their age.
展开▼
