| Summary: | 碩士 === 國立中興大學 === 應用經濟學研究所 === 93 === ABSTRACT
To cope with the competitive trend of liberalization and globalization, the management of financial institutions in Taiwan is under a survival pressure and is thinking about how to uplifting its performance. Because Taiwanese banking management is using “Branch System”, the performance of each bank relies on the efficiency of its branches. To evaluate the efficiency of each branch, the management of the financial institutions often uses a performance evaluation approach, but the approach involves subjective factors and is not able to indicate the improvement scale and aspect needed by inefficient branches.
This study uses the efficiency evaluation method - Data Envelopment Analysis (DEA) to measure 81 branches of one bank in Taiwan, using input/output oriented mode as well as CCR and BCC models to obtain operational efficiency, pure technical efficiency and scale efficiency. An empirical result would be analyzed and suggestions would be made by this study to provide decision making references to the management of the bank.
The summary of this research’s result are as follow:
1.According to input/output variables, the operational efficiency of 81branches is all the same. However, the analysis results of technical efficiency and scale efficiency among inefficient branches aren’t quite alike. Among the 81 branches, only 8 branches are relatively efficient (the ratio is about 9.9%) and 73 branches are inefficient (the ratio is a high 90.1%).
2.By examining the relation of the location of the branches and their efficiency values, we could find the most efficient branches are located in the Taoyuan-Hsinchu-Miaoli region. The greater Taipei region ranks second. Although most branches are located in Taipei, since this is a much more competitive market, their operational efficiency isn’t as high as one would expect. Judging by the urbanization of the locations of the branches, we find that the most efficient ones are located in the cities. Therefore, the locations of the branches and their efficiency have some relation.
3.By using the DEA method to evaluate the efficiency value of each branch, the result is less different with the existing group. The main difference between DEA and performance evaluation is the distributions of weights. The weights of DEA come from linear programming, and the weights of earnings of the performance evaluation is 50%. Technically, the DEA method is used to find out the reasons for the increase of efficiency of different branches. However, economically, whether or not this increase of efficiency is worthwhile still needs further analysis.
4.According to the result of the DEA method, among the branches that have an efficiency value of 1, some of them have a poor performance in practice. This shows that the input/output operational efficiency of these branches’ productive essentials is higher than the other branches. As for the meaning of reference frequency, generally the more reference frequencies taken represents the evaluated unit is relatively more stable and efficient. According to the research of substantial evidence, this study finds that the more a branch is referred to (bench mark) merely shows other evaluated branches have similar input/output ratio, production mode or production criteria. Through the sensitivity analysis, we deleted branch D5 and C17 for they are the most referenced branches in input-orient mode (they both belong to group 3). We reran DEA afterward and found that the branches with much more reference frequencies are still centralized around D5 and C17, proving the validity of this theory.
|
|---|
