本研究是運用資料包絡分析法（Data Envelopment Analysis, DEA）及Malmquist生產力指數，對國內某證券商33家分支機構，以91年第1季至92年第4季共8季資料，進行經營效率的實證分析。 就技術效率而言，33家分支機構整體技術效率，即CCR效率平均值為0.851，各季效率差異不大。就純技術效率言，整體平均值為0.943，平均值呈逐步上升的趨勢，顯示經營及管理決策改善，使效率改善。就規模效率而言，整體平均值為0.903，91年平均規模效率呈現倒V型，92年平均規模效率則呈現N型。就跨期效率指標，Malmquist生產力指數而言，22家分支機構在追趕效率上相對有效率，表示有66%分支機構因管理與決策適當而使效率改善；但以Malmquist指數分析，各分支機構的Malmquist指數平均值均小於1，表示技術退步。針對無效率區處，建議管理者在投入與產出數量上，提出建議做為證券商分支機構改善經營效率的參考。

This research includes discuss up to 33 branches of a securities firm as analysis sample. We evaluate these branches' performance efficiency with Data Envelopment Analysis and Malmquist production index. This study focus on how the branch's excecution ability and customer services affect on performance efficiency and provid suggestions for improvement for managers. The results show there is no significant difference on Technical efficiency across seasons. For the pure technical efficiency, we find upward tendency from 2002 to 2003, which means managerial decision-making have been improving the efficiency. Generally speaking, the type of scale efficiency in 2002 is approaching V type, but the type of scale efficency in 2003 is approaching N type. As for Malmquist index, the mean of index across 33 branches is smaller than one which indicating technology has regressed. We also suggest manager to improve these inefficiency branches through changing inputs and outputs according to the Slack Vraiable Analysis.