臺灣地區金控與非金控證券商之共同邊界效率分析

Abstract

資料包絡式分析法是一種常被廣泛使用來針對決策單位之績效進行評估的分析方法，可以同時處理多個投入變數及產出變數，亦可評估各決策單位在不同時期的績效表現，其應用範圍相當廣泛。而在經濟理論中，Tirole (1988, Ch. 4) 介紹過垂直效率理論，對於有母體支撐之組織結構，其效率遠優於獨立組織，本文期望透過分析金控與非金控體系下之證券公司效率比較，以符合上述之經濟理論。本研究採用資料包絡式分析法中的差額變數模型(SBM)，針對2011年至2015年24家上市櫃證券公司，分成金控與非金控體系兩群組，進行個別效率表現之分析，再利用共同邊界模型 (Metafrontier) 比較兩體系下的證券商經營效率。本研究包含兩個投出項及三個產出項，其中投入項包含:總資產、營業費用，產出項則為:經紀收入、承銷收入及自營收入。研究結果得出，在24家上市櫃證券公司中，僅4家 (3家金控、1家非金控) 證券商在資料期間內皆具有共同邊界效率。

Data Envelopment Analysis (DEA) is a popular method in analyzing decision-making unit’s efficiency. It can manage multiple inputs and outputs at the same time and evaluate the performance in different period. Tirole (1988, Ch. 4) introduced the theory of vertical control of industrial organization with the conclusion that the downstream firms with more resource and support from the upstream firms can be more efficient. In the securities industry, we expect to support such a theoretical prediction that the securities firms under financial holding companies are more efficient than independent ones. In this paper, we chose 24 securities firms, including 17 listed and 7 OTC securities firms in Taiwan market during 2011-2015, in order to compare if securities firms under financial holding companies or securities firms not under financial holding companies are more efficient. In the first step, we group these 24 securities firms into two groups, one is financial holding companies and the other one is not under financial holding companies. By using slake-base measure (SBM), the within-group efficiency scores of securities firms are then computed. In the second step, we compare the efficiency between all the 24 securities firms by metafrontier method. This research chooses assets and expense as our input variables, and brokerage income, underwriting income, and dealing income as the output variables. There are 5 securities firms among the 24 firms have metafrontier efficiency during the data period, among which three are FHC security firms and only one is an independent securities firm.