Comparison of Efficiency and Modelling of Islamic Banks and Conventional Banks in Indonesia

This study aims to analyze the efficiency comparison between Islamic banks and conventional banks in Indonesia in 2012-2014. The data in this study were chosen through purposive sampling from 20 Islamic banks and conventional banks in Indonesia. The method used in this study is non-parametric approach with data envelopment analysis (DEA) whereas input and output variables are treated in intermediary function. The input variables are total asset, total saving (third party fund), and price of labor while the output variables are total financing (loans) and total operational expenses. To measure the efficiency level of Islamic banks and conventional banks the independent sample t test is used.The result of the study shows that there is no significant difference of efficiency between Islamic banking and conventional banking in 2014 because of the significant value (2-tailed) only at 0.537 where P-value is higher than α=0.05 Ha is refused. There is no difference of efficiency between Islamic banks and conventional banks in efficiency scale (ES). The empirical factors that affect Islamic banks and Conventional banks efficiency are ROA, CAR, and FDR variables. On the other hand, NPF results insignificantly and affects negatively towards Islamic banks efficiency. Lastly, ROA, NPL, LDR, and CAR had significantly affected Conventional banks efficiency in Indonesia from 2012 to 2014.

efficiency analysis using non-parametric method of data envelopment analysis (DEA). The results showed merging banks experienced during the last three years the efficiency of the Year 2001Year -2003. Then the merger does not produce good efficiency for the bank. However, most banks that merged to experience efficiency. Firdaus & Hosen (2013), using two-stage method of analysis, namely the first stage using the DEA, and the second stage uses Tobit models or perform regression for factors that affect efficiency. The results show there is no optimum efficiency in Islamic Banks during the period 2010-2012.
Another study using DEA method are Fauzi (2013), Joko Sarjono (2008), Shahid, et al (2010), Vitello & Sutarno (2010) used the DEA and comparison with the ROA, Yudhisthira (2004), using the method of DEA and OLS, Ahmad & Noor (2011), using the two-stage analysis: DEA and Tobit, Shafitranata (2011), Daman Huri & Susilowati (2004), Widyastuti (2011) using DEA and regression model panel (fixed effect model), and Perwitaningtyas (2014 ) using DEA and multiple Regression. While the research results Fauzi (2013), showed a significant difference between the efficiency of CB and IB both CRS and VRS. Sarjono (2008) investigated the efficiency analysis of Islamic banks with DEA method, case studies Mumalat bank, Bank Syariah Mandiri andBank Mega Syariah Year 2005-2007, the results showed the bank muamalat got scale the highest efficiency of all three banks. Haseeb Shahid researching "efficiency comparison of Islamic and conventional banks of Pakistan". The results showed conventional banks in Pakistan more efficient technically of the Islamic Bank. But cost allocation and cost efficiency showed a healthy competition. While the value of tstatistic showed no significant difference between the average efficiency score Islamic banks and conventional banks except in 2008.
Similarly, Ahmad and Noor (2011), conducted the research with two stage analysis method: DEA and Tobit. The results showed that there were technical efficiency in the profitability of Islamic banks and demonstrate efficiency tends to favor an Islamic bank. While research Wijayanto & Sutarno about the performance efficiency of intermediation owned banks in Indonesia by using DEA. Results showed BNI, BTN, and Mandiri have relative efficiency level of 100%. BNI obtain relative efficiency level of 88.5%. While the average score of efficiencies gained by the bank approximately 97.13%.
Recently research Widayastuti (2011), using the method of DEA and panel data model, showed that the average efficiency rate of 96.15% of Bank BNI, BRI amounted to 92.56%, the bank amounted to 98.50% and 96.51% of BTN. While the results of the regression using panel data shows that only asset variables that significantly influence the performance efficiency of state-owned banks in Indonesia. While third-party funds and loans did not significantly affect the efficiency of state-owned banks.
From some research above suggest that research into the level of efficiency of banking can be done with different methods and approaches vary. But in this study researchers used intermedisasi approach in determining the input and output of banking, while the methods used to see the level of bank efficiency using Data Envelopment Analysis (DEA). Then comparative efficiency of using the One Sample Test Furthermore, to analyze the effect of efficiency with that contained in Islamic banks and conventional banks researcher using Fixed Effect Model (FEM) 2012-2014.

The Concept of Efficiency
Efficiency is a term that is relative, that is always associated with a certain criterion. Most economists bisects an efficiency standpoint. First of positive aspects, which in the thrust of human behavior are always looking for increased value, this value will form the search market mechanism. Secondly, in terms of normative, stems from the desire to create a policy, whether the policy is better than other policies (Sari, 2010: 14).
Global Review of Islamic Economics and Business, Vol. 4, No. 2 (2016) 102-114 105 Nuryana Sari (2010 states that the concept of efficiency starting from the concept of production theory that explains the relationship between input and output. The production function illustrates the input to output in a given period. One of the models used to describe the production function is the production frontier models. This line shows the relationship of input and output in the production process. The production line of the production frontier represent the maximum output level of any inputs that represent the use of technology from a company or industry. The concept of efficiency was first proposed by Farrell as a follow-up proposed by Debreu and Koopmans. Efficiency measurement proposed by Farrell can take into account the input compound. Efficiency a company consists of two components, namely technical efficiency and allocative efficiency. Technical efficiency indicates the ability of companies to achieve maximum output from a number of inputs. While the allocative efficiency of input use show a company optimally at a certain price level input. Then the two components are combined to produce a measure of efficiency or total economic efficiency (Abidin & Endri, 2009: 22).
Basic economic theory there are two terms of efficiency, the technical efficiency and economic efficiency. Economic efficiency has a macro point of view that has more reach than technical efficiency micro-view cornered. Measurement of technical efficiency tends to be limited to technical and operational relationships in the process of converting inputs into outputs (Huda & Nasution, 2014: 10). As a result, efforts to improve the technical efficiency requires only micro policy which is internal, namely the control and optimal resource allocation. Price in economic efficiency can not be considered to be given, as they can be affected by macro policies (Amrillah, 2010: 22).
In general, banking efficiency can be divided into two kinds. The first scale efficiencies. Both X-efficiency. The concept of scale efficiencies introduced by Farrell (1957), which can be simply defined as the relationship between the average production cost per unit and volume, thus the bank is said to have economies of scale when output was accompanied by a low unit cost of production. Both X-Efficiency introduced by Leibenstein (1966), refers to the deviation from the constraints of cost savings from the lowest production costs for a given level of output. X-efficiency derived from technical efficiency and allocative efficiency (Yudhisthira, 2004: 2). However, to achieve the level of economic efficiency of a company should be technically efficient. In order to achieve the level of profit (profit) is maximized, the company should produce the maximum output with given inputs (technical efficiency) and produces output with the right combination at a certain price level (allocative efficiency) (Sari, 2010: 16).

Research Method
The method employed in this study to measure the performance of technical efficiency using Data Envelopment Analysis (DEA). DEA measuring the ratio between the input and ouput owned by banks, the approach used is the intermediation approach in measuring input and output. Samples were examined in this study were 20 Islamic banks and 20 conventional banks in the period 2012-2014. Variables that used are the input variables comprised of Total Deposits, Labor Costs, and the total assets. While the output variables consist of Total Financing, and total operating income. Then to see the comparison between Islamic banks with conventional banks Years 2012-2014, the authors use the method Independent sample t test in SPSS Software 17. Furthermore, the results are grouped in the efficiency ratio of Islamic banks and conventional banks.
Furthermore, to look at the factors that affect the performance of the efficiency of using the independent variable in this study consisted of financial ratios in Islamic banks and conventional banks such as ROA, CAR, LDR/FDR, NPL / NPF. This variable is used to measure the factors that affect the performance of banking efficiency. While the scale of banking efficiency measured with DEA method is used as the dependent variable which is considered as the scale of the overall efficiency of the bank as measured by specific inputs and outputs. In this study, using a fixed effect model (fixed effect model), because the data used is the data panel of cross section data and time series with a number of diverse individuals and only time was taken in the next 3 years.

A. Result
Results of measurement Data Envelopment Analysis (DEA) with the approach of CRS and VRS Islamic banks 2012-2014 indicates the following results:  Furthermore, to analyze the factors that affect the performance efficiency Fator shown in the results of the regression Fixed Effect Model (FEM) as measured using 8.0 eviews shown in Table 1.3 as follows:

B. Discussion
The results of measurement of the efficiency of Islamic banks in 2012 showed the CRS approach shows only five Islamic banks are relatively efficient perfect (BMS, BDS, BIIS BSS and BAS). While in the year 2013 the number of Islamic banks are relatively efficient perfect increased by 6 banks. Then approach to CRS in 2014 the number of Islamic banks were efficient as many as 10 Islamic banks. However, the average efficiency of CRS approach to Islamic banks just as much as 5 Islamic banks were efficient in Year 2012-2014. The average efficiency of the approach is more dominant CRS Islamic banks experienced by sharia business unit than commercial banks sharia.
Nevertheless, every year the number of Islamic banks increased efficiency very rapidly, from 5 banks were streamlined into 10 banks were efficient at CRS approach. Then the VRS approach to Islamic banks are eight Islamic banks were efficient in Year 2012. In this VRS approach other variables held constant and does not affect other variables. In the year 2013 increased by 11 Islamic banks are relatively efficient. Then the VRS approach to Islamic banks in 2014 to approach as many as 12 banks belonging perfectly efficient. While the average Islamic banks were efficient in VRS approach is as much as 10 Islamic banks are relatively efficient.
Judging from the overall efficiency of Islamic banks in the year 2012-2014, it gained an average efficiency of Islamic banks in the approach scale efficiency (SE) as many as nine Islamic banks are relatively efficient perfect. Average experience this is the perfect efficiency of Islamic banks, while Islamic business units are also still relatively good efficiency even under Islamic banks. This is because the Islamic banks more easily compete with other banks, compared Islamic business units are still located on the premises of conventional banks.
Then the development of the efficiency of conventional banks are quite varied in 2012-2014. Judging from the CRS approach in 2012 the number of conventional banks are efficient as conventional bank 7, higher than Islamic banks. Later in the year 2013 the number of efficiency of conventional banks decreased to 3 conventional banks. While in 2014 the number of efficiency of conventional banks with an efficient approach to CRS experience as much as five banks. When viewed from an average of just 4 CRS efficient conventional banks. Sedankan the VRS approach to conventional bank in 2012 as many as 10 banks. In the year 2013 as many as nine conventional banks efficiently. In 2014 the number of conventional banks efficiently increased by 10 banks. Then the average of conventional banks in the VRS approach only as many as 10 conventional banks are perfectly efficient. Judging from the overall efficiency of the Year 2012-2014 on SE approach conventional banks an efficient amount is as much as seven conventional banks efficiently. This number is still below the average efficiency of Islamic banks. This is because conventional banks are still not optimal in managing third party funds and credit and too much cost of funds for the banking operations.
The estimation results ROA significant negative effect on the efficiency of Islamic banks. These results are consistent with research Rahman and Rosman (2013), Endri (2011), Ismail, Majid, and Rahim (2013), Ulfa (2014), Firdaus &Hosen (2013), andHasan (2005). Hasan (2005) mentions ROA has a strong correlation between the efficiency of the bank with the profitability of Islamic banks. While Ismail, Majid and Rahim (2013) reported similar results ROA positive effect on the efficiency of Islamic banks are scale efficiency (SE). So with the above results it can be said even though the variables return on asset negative effect, the higher the ROA ratio of Islamic banks more efficient.
And then the regression results for conventional bank ROA positive and significant impact on the efficiency of conventional banks in 2012-2014. The analysis results ROA at conventional banks according to the research Soetanto & Ricky (2011), which showed ROA significant effect on the efficiency of conventional banks in Indonesia, due to the amount of assets owned by a conventional bank is greater than the Islamic banks.
The estimation results NPF negative but insignificant effect on the efficiency of Islamic banks in the year 2012-2014. This means that NPF did not affect the efficiency of Islamic banks in Indonesia. This is in contrast to the initial hypothesis that NPF requires significant negative effect on the efficiency of Islamic banks. And then results for conventional banks to nonpeforming loans (NPLs) So accept the hypothesis that the NPL a significant negative effect on the efficiency of conventional banks. Means the ratio of NPL affect the efficiency of conventional banks in the year 2012-2014. This is due to the NPL ratio in conventional banks are relatively smaller and the burden of risk borne by conventional banks lower for larger banks will issue a guarantee fee, supervision to financing problems.
While the ratio of Islamic banks FDR showed Financing Deposit Ratio (FDR) positive and significant impact on the efficiency of Islamic banks. Pembiaayaan channeled Islamic banks is high and rising in the year 2012-2014, resulting in more optimal FDR variable and affect the efficiency of Islamic banks are positive and significant. And then results of the analysis Loan Deposit Ratio (LDR) a significant negative effect on the efficiency of conventional banks in Indonesia. Due to the negative and significant potential improvement can be seen from the average DEA conventional banks are still not optimal in providing financing to third parties. But the results of LDR analysis found a significant effect on the efficiency of conventional banks in Indonesia 2012-2014.
The empirical results of this study indicate CAR variables significantly influence the efficiency of Islamic banks in 2012-2014. Receiving initial hypothesis, the higher the ratio of a bank's CAR at the more efficient the bank anyway. Due to banks that have large capital to be able to maintain the stability of performance as a whole. The results are consistent with research Ulfa (2014), Endri (2011). The results of the analysis of the variable CAR to conventional bank in Indonesia in 2012-2014. Variable CAR positive and significant impact on the efficiency of conventional banks in Indonesia. These findings are consistent with results of previous studies conducted by Perwaningtyas (2014). However, most studies assume CAR variable has no effect or not related to the banking efficiency. As Soetanto & Rizky (2011), Masita (2013, shows that the health of banks is not having an effect on the efficiency that shows the Capital Adequacy Ratio (CAR), which is owned by the bank is only used by the bank to meet the policy rule of minimum capital so it does not affect the efficiency of the bank.

B. Recommendations
After conducting an analysis of this study, Source: DEA excel the authors have some advice to be reviewed in this study, as follows: 1. The sample in this study is limited to 20 Islamic banks and 20 conventional banks. While the variables used only 5 types, input variables include total assets, total deposits, and labor costs. While the output variables include the total financing and total operating income. So for researchers to develop the variables used in this study, as well as other approaches besides the intermediation approach undertaken. 2. Comparison of the efficiency of Islamic banks and conventional banks in this study used a different test. For further research can be carried out studies into the factors that affect the macro or micro level of efficiency in the banking system. 3. Should the approach model Data Envelopment Analysis (DEA) can be used in research IB, Islamic business units, and conventional banks, to obtain relevant information on the performance of the institution. 4. Later variable factors that affect the technical efficiency of banks can be deepened back in future studies 5. For academics and researchers can use a variety of variables in measuring technical efficiency in the institution or company.
Global Review of Islamic Economics and Business, Vol. 4, No. 2 (2016) 102-114 111 6. Availability of data affecting the efficiency of banks can be trusted, so that it can obtain diverse information. 7. For the Indonesian Bank and the financial services authorities should discuss and formulate a policy that is carried out in measuring financial performance. 8. This study was limited in using a regression model that related to internal factors banking, due to the availability of external data is still limited and the period used comparatively brief for 3 years.