Applying the Network Stochastic Frontier Approach to study the efficiency of Taiwan's banking

傳統研究銀行業經營績效時，需考量將存款視為投入或產出，認定方法的不同將導致估計出的效率水準差異極大，本研究應用兩階段的銀行生產模型，第一階段將存款視為中間產出，第二階段與勞動和資本一起瑝作投入，生產另外三種產出。這種作法使得存款既保有產出的特性，又是銀行生產過程的要素投入，不必再主觀的去認定存款對銀行究竟是投入或產出。本研究以1997-2011年的50家台灣銀行業資料，採用隨機邊界分析法估計銀行廠商的技術效率水準與成本效率水準及兩個階段間要素投入量的分配比例。實證結果發現銀行在第一階段生產存款的過程中，要素的投入配置比例較低，大部分的要素是在第二階段生產放款、投資與非利息收入時被使用。另外也發現台灣整體銀行業生產存款的過程存在技術無效率，而生產放款、投資與非利息收入等時的成本效率值甚高。本研究也發現生產存款的過程中，泛公營銀行的技術效率高於民營銀行，這點與過去的研究比較不同。In previous studies about bank productivity and efficiency, researchers have to indentify deposit as either input or output, as a result, applying different analysis framework produce diverging outcome. In this study we consider bank production model as two-stage. Bank produces deposit as intermediate product which is one of input factor to produce various output in second stage. We don’t have to identify deposit as input or output subjectively by applying this analysis framework. We apply SFA model to estimate technical efficiency and cost efficiency and factor distribution parameter of Taiwan banking industry by 50 Taiwan banks panel data in 1997-2011. Empirical result shows that bank uses little input factor in first stage, in addition, technical inefficiency commonly exists in bank first stage production in contrast with that high cost efficiency in second stage. We also find that pan-publicly banks have both high technical efficiency and cost efficiency in comparison with private banks