本研究衡量銀行分支機構的經營效率，使用Translog隨機邊界生產函數法，以2007至2010年度合計四年為研究期間，選取國內某商業銀行175家分支機構，共700筆觀察值為研究對象，經由實證分析得到下列結論： 一、分行位於工商業發達的縣市（地區），放款總額技術效率值較高，但放款營收技術效率值卻較低，係因銀行間彼此競爭的結果。分行位於工商業不發達的縣市（地區），表現與上述相反。實際盈餘因非利息收入增加，有助技術效率值表現。 二、非工業區分行放款總額技術效率較工業區技術效率值高，係因工業區承作個人放款（房貸）少；但工業區以中小企業居多，所以放款營收及獲利表現較佳。 三、都市型分行較非都市型分行放款總額之技術效率值高，但因競爭造成利差因素，所以放款營收及實際盈餘技術效率值表現不佳。 四、營業規模愈大的分行，通常位於工商業較發達的縣市（地區）且成立年數較久，客戶資源多，雖因競爭以致放款利息收入減少，但非利息收入較高，所以整體盈餘水準較營業規模小的分行為佳。 The purpose of this study is to measure the operating efficiency of bank branches by using Translog stochastic frontier production function approach. In this study, we select a domestic commercial bank as an example, which consists of 175 branches and collects operating data from the year of 2007 to 2010.Our empirical results indicate: First, due to the competition, branches located in the industrial and commercial areas have higher technical efficiency values from the measure of loans, but lower values form the measure of revenues. However, branches located in the less developed areas, the performance is reversed. Second, due to personal mortgage, branches located in the non-industrial areas have higher technical efficiency values from the measure of loans. However, branches located in the industrial areas have higher technical efficiency values from the measure of revenues and earnings because of mostly small and medium-sized enterprise. Third, due to the interest differential, branches located in the urban areas have higher technical efficiency values from the measure of loans, but lower values form revenues and earnings. Finally, due to non-interest income, large-scale branches have higher technical efficiency values from the measure of earnings.
