兩階段DEA模型在台灣產險公司之應用

杜勻

指導教授 ： 紀美秀

國立中正大學/理學院/數學系應用數學研究所/碩士(2016年)

摘要

自Cooper 等(1978) 建立了CCR 模型後，不但奠定了DEA 方
法的基礎，也讓DEA 方法廣泛的應用在各公司單位的績效評估
上。本研究透過Kao(2008) 的2001 - 2002 年的產險公司資料分別利
用CCR、Centralized、Chen 與Despotis 四個兩階段模型評估其經營績
效，並從結果了解到四個模型在公司效率高低的評估上有相似的判斷力，
同時我們也發現CCR 模型的兩階段效率可作為其他三個模型的兩階段效
率的上界，藉由這個上界為基準點來了解公司的兩階段是否有達到應有的
表現。最後我們整理出2013 - 2014 年的產險資料，並透過四個模型的評估
結果與2001 - 2002 年的模型比較結果對照。
關鍵字:資料包絡法

關鍵字

資料包絡法

並列摘要

Since Cooper et al.(1978) built the CCR model, the foundation of
DEA have been established and applied broadly to evaluate the company
relative eciency. According to Kao and Hwang's(2008) data, about 24
non-life insurance companies in Taiwan, we employ four di erent two-stage
DEA models including CCR , Centralized , Chen(2009) and Despotis(2012)
to evaluate their relative eciency. From the outcome, we have found these
four models have similar ability for judging the companies. Moreover, the
eciency of each stage of CCR model o ers the upper bound for the e-
ciency for each stage of the other models. We may also apply the upper
bound to realize the achievement of each stage. In addition, we make use of
these four models to evaluate the eciency of the non-life insurance com-
panies' data between 2013 and 2014.
Keywords:Data envelopment analysis(DEA)

並列關鍵字

DEA

參考文獻

velopment analysis : An application to non-life insurance companies in Tai-

visional eciency and frontier projection under general network structures,

[4] Seiford, L.M.,Zhu, J., (1999), Pro tability and Marketability of the Top 55

U.S. Commercial Banks, Management Science, 45: 1270-1288.

mance of Canadian Life and Health Insurance Companies, Mathematical and