過去在基層金融環境中扮演重要角色的信用合作社，隨著金融市場之開放而面臨了強大的競爭。本篇研究為國內首次使用資料包絡分析法，以較客觀的角度評量估用合作社之經營績效，並修正Korobow & Stuhr (1985)之加權效率指標，以比較資料包絡分析法與參照我國金融監理制以中之檢查資料評等系統所設計之財務比率評等模型，最後以回歸分析對可能影響信用合作社經營績效之因素進行研究，俾提供主管機關對信用合作社進行管理與協助，使國家經濟得以在安定中永續發展。 研究結果顯示，資料包絡分析法與財務比率評等模型之績效評估結果不盡相同。在模型對優良信用點作社的判別上，型一誤差之控制以財務比率評等模型較佳，而型二誤差的控制則屬資料包絡分析法之BCC模式較優。因此本篇研究建議金融機構之主管機關，似可以資料包絡分析法建立金融機構績效評估模型，俾與檢查資料評等系統相互配合。另外，實證結果顯示信用合作社的經營績效逐年成長，且隨賺取之利息差額擴增而提升；而當營業所在區域競爭愈激烈、信用合作社營業據點愈多、放款品質愈差，則會使信用合作之經營績效降低。

　　　　　In this study, we conduct performance evaluation of 72 credit unions in Taiwan. We use traditional Financial Rating Model (FRM) and Data Envelopment Analysis (DEA) for assessing financial performance and technical efficiency of credit unions. We then employ modified Korobow & Stuhr's Weighted Efficiency Index (1985) to compare the results and examine factors affecting the performance of credit unions. We find that 2 to 14 of the 72 credit unions are identified as relatively efficient under the DEA each year, and only 2 to 4 of them are ranked as 'Grade A' performers under FRM. COMPARING THE weighted Efficiency Indices of the two models in classifying better performers, we find that FRM is better that DEA ON type I error control, but DEA's Type II error control is superior to the former. It suggests that the government policy making and evaluation on credit unions will be more comprehensive by compensating the current evaluation system with DEA. The result of multiple liner regression indicates that the net interest earned has a significantly positive impact on the efficiency, which has also improved over time. Besides, competition intensity, branch units, and loan overdue have significantly negative impact on the efficiency performance of credit unions.