ANALISIS DETERMINAN EFISIENSI PERBANKAN INDONESIA TAHUN 2010-2014 : PENDEKATAN DATA ENVELOPMENT ANALYSIS (DEA) DAN STRUCTURE-CONDUCT-PERFORMANCE (SCP)

Abstract

This research aims to analyze the technical efficiency level and its determinants in Indonesian banking industry during the period of 2010 – 2014. To measure efficiency level, the Data Envelopment Analysis (DEA) is applied based on the three approaches: production, intermediation, and asset. The objects of research are 100 public banks, which consist of six groups, i.e. 4 Shareholder Banks (Bank Persero), 31 Foreign Exchange National Private-Public Banks, 21 Non-Foreign Exchange National Private-Public Banks, 24 Regional Development Banks (BPD), 10 Mixed Banks, and 10 Foreign Banks. In addition, to analyze the determinants of efficiency level in Indonesian banking industry, this research uses the Structure-Conduct-Performance (SCP) framework. In this research, Technical efficiency level is a performance indicator used as dependent variable, while independent variables consist of Structure (S) and Conduct (C) such as: Market Share (MS), Capital Adequacy Ratio (CAR), Loan to Deposit Ratio (LDR), Non-Performing Loan (NPL), and Net Interest Margin (NIM). To estimate the determinant of banking industry, the regression with Tobit model is applied. Based on the three approaches of efficiency measurement, the results show that the Shareholder Bank (Bank Persero) experiences the highest production-efficiency level. Meanwhile, Regional Development Bank (BPD) has the highest intermediation-efficiency level and the Non-Foreign Exchange National Private Public Bank has the highest assets-efficiency level. Tobit regression shows that Market Share (MS) has positive influence toward production and intermediation-efficiency level, but it has negative influence on asset-efficiency level. Similarly, CAR has positive influence on production and assets-efficiency. LDR negatively influence the production and assets efficiency. NIM has positive effect on intermediation-efficiency.