本研究以選定26家上市櫃公開發行鋼鐵產業為例,利用資料包絡分析法的四種DEA模型進行產業經營績效實證分析,然後根據執行結果,進行第一階段樣本廠商的之相對效率分析、差額變數分析,與敏感度分析。第二階段將第一階段所得到之效率值加入公司財報中與償債因子相關的投入變數,來觀察對於公司表現之績效是否有顯著或非顯著之影響。 實證結果顯示,經過各模式鋼鐵公司績效分析之後發現:(1) 在BCC模式下,整體效率值僅次於Super SBM,且逐年上升。(2)在SBM模式下,整體效率值為所有模式最低,且逐年上升。(3)在AR模式下,整體效率值為所有模式次低, 且2004年至2005年時呈現上升現象,2005年至2006年則呈現下降現象。(4)Super SBM模式下,各年皆只有1家公司績效排名第1。整體效率值為所有模式最高,且逐年上升。另直接將風險視為投入變數,經風險因子調整後的鋼鐵公司之效率值呈現上升,因此是否考慮償債因子將影響公司之績效表現。 This paper examines the relationship of performance and risk on Taiwanese steel industry by means of DEA models to evaluate efficiency measures . First, that make use of four types of DEA models to create the efficiency measures of companys for analysis of relative efficiency, slacks-based measure and sensitivity. After that put to use the risk factor of working capital as relative input variable into these models to implememt and observe the variation of efficiency measures. Using annual data from 2004 to 2006, the results indicate the efficiency measures with risk factor are more efficient than the efficiency measures without allowing for risk factor. We find that advised risk factor is essential to improve the measurement of performance of business and could more accurately support the evaluation of lending risk . The risk factor is significant to affect the efficiency measures.