自從績效評估方法被提出來以後,很多文章就以這個方法對商業銀行的生產力和效率進行評估。主要是通過資料包絡法(DEA)和隨機邊界法來實現(SFA)。本研究主要利用隨機邊界法,結合道格拉斯生產函數和Translog模型對2005年至2012年25家台灣商業銀行的經營績效進行評估。實證結果發現,樣本期間台灣銀行之經營績效平均值為0.870,也就是說,樣本期間台灣銀行產業尚有13%(1-0.870)的無效率可以改進。而從時間上來看,樣本台灣銀行經營效率總體呈現出上升的趨勢。而在2008年金融危機期間呈現出經營績效下降的趨勢,在危機過後重新上升,在2012年達到最高效率值0.966。
The performance evaluation method was widely used to estimate the productivity and efficiency of bank since it was proposed. Most studies used Data Envelopment Analysis (DEA) or Stochastic Frontier Analysis to achieve the estimation.Thisstudy applied the Stochastic Statistical Frontier Approach model, conbime with Cobb-Douglas production model and Translog model to estimate the 25 banks in Taiwan during 2005 to 2012. The empirical analysis result shows that the technical efficiency of Taiwan banks is 0.870, which means that there is 13% inefficiency to improve. The technical efficiency seems to rise during the period, but it fall during the 2008 Financial Crise. After the crise, the efficiency rise up to 0.966 in 2012.