THE INFLUENCE FROM FINANCIAL PERFORMANCE ON EARNING PER SHARE (EPS) WITH SOUNDNESS LEVEL OF BANKS AS INTERVENING VARIABLE AT BUKU IV BANKS CATEGORY IN INDONESIA

The purpose of this research was to analyze those influence of financial ratios towards earnings per share with soundness level of banks as an intervening variable based on risk-based bank rating method at BUKU IV Banks in Indonesia for the period of 2015-2019. The six banks are PT Bank Mandiri Tbk, PT Bank Rakyat Indonesia Tbk, PT Bank Negara Indonesia Tbk, PT Bank Central Asia Tbk, PT Bank CIMB Niaga Tbk and PT Bank Danamon Indonesia Tbk. This research used secondary data with quantitative approach. Data analysis technique used panel data regression by purposive sampling. The results showed that NPL & GCG had significant negative influence on soundness level of banks and NPL had significant negative impact on EPS. GCG had significant positive influence against EPS, and ROA & CAR did not have significant positive impact towards soundness level of banks & EPS, while soundness level of banks were not found to have significant negative affect to EPS, and soundness level of banks did not mediate NPL, GCG, ROA and CAR in affecting EPS.


INTRODUCTION
The economic crisis which occurred in Asian countries including Indonesia has caused banking sector to experience downturn and difficult times that are marked by the depreciation of rupiah exchange rate against USD dollar, the tight liquidity which resulted in bank having difficulties in obtaining fresh fund and the increasing number of non-performing loans. With this condition, it has resulted in a crisis of customer's faith and the poor Interbank Money Market.
The case is not about the weakness of a bank's supervisory function, but a bank is a business entity which is fully obliged to maintain and protect national banking from various risks and supervision is much necessary. In the beginning, in the Bank Indonesia's regulations and Wilopo (2011) showed that GCG had significant influence on company value. Meanwhile, Sakai and Asaoka in Tjondro and Wilopo (2011) had observed that GCG had significant positive influence towards PER.
Based on the expansion from the background and the tendency of influence between bank's performance and the soundness level of banks towards Earning Per Share (EPS), therefore, the researcher conducted this research which was focused on the six Commercial Banks at Bank BUKU IV (PT. Bank Mandiri Tbk, PT Bank Rakyat Indonesia Tbk, PT Bank Negara Indonesia Tbk, PT Bank Central Asia Tbk, PT Bank CIMB Niaga Tbk and PT Bank Danamon Indonesia Tbk) with core capital of at least 30 trillion rupiah.

THEORETICAL REVIEW Risk Profile
Risk Profile is an assessment from internal risk factors and implemented of risk management quality in bank operationals with 8 (eight) risks namely credit risk, operational, market, legal, liquidity, strategic, compliance and reputational (Financial Services Authority Regulation Number 18/POJK.03/2016). There are several risks, but only use the credit risk in soundness level of calculation which measured by Non-Performing Loan (NPL) ratio.

Good Corporate Governance (GCG)
Good Corporate Governance (GCG) is an assessment factor of internal management performance with a self-assessment system by calculating the GCG component. Based on Bank Indonesia Regulation No.13/1/PBI/2011 and Bank Indonesia Circular Letter (SEBI) Number 15/15/DPNP of 2013, banks carry out their own GCG assessment based on a self-assessment system on a regular basis by searching for annual reports published for the determination of GCG assessment. GCG assessment consists of three aspects, which are Governance Structure, Governance Process and Governance Outcomes.

Rentability
Rentability (Earnings) that could be carried out by using Return on Assets (ROА) and Earning Per Share (EPS) ratios. Return on Assets (ROA) is a ratio to measure about bank's capability to generate profits and manage the bank's overall business efficiency level. Meanwhile, Earnings Per Share (EPS) used to compare the available or generated income (net income) to the number of outstanding shares at the end of period.

Capital
Capital could be measured by Capital Adequacy Ratio (CАR). According to Dendawijaya (2009:121) and Irawan and Syarif (2019), Capital Adequacy Ratio (CAR) is a ratio which showed that how far all bank assets had contained risk (loans, investments, securities, claims at other banks) were also funded from bank capital itself, besides to obtain funds from the sources outside the bank, such as community funds, loans and others.

Theoretical Framework
Based on these descriptions and explanations which had been done above, this research was formed because of interdependence between variables which considered important to be studied. These research model based on theoretical framework could be seen in the image as below.

Hypothesis
The hypothesis in this research were 1) Non Performing Loans (NPL) had an influence toward Soundness Level of Banks; 2) Good Corporate Governance (GCG) had an affect to the Soundness Level of Banks ; 3) Earnings had influence towards the Soundness Level of Banks, 4) Capital Adequacy Ratio (CAR) had an affect the Soundness Level of Banks ; 5) Non Performing Loan (NPL) had influence towards Earning Per Share (EPS); 6) Good Corporate Governance (GCG) affects Earning Per Share (EPS); 7) Earnings had impact to Earning Per Share (EPS); 8) Capital Adequacy Ratio (CAR) affects those Earning Per Share (EPS); 9) Soundness Level of Banks had influence towards Earning Per Share (EPS); 10) Soundness Level of Banks had contribution in mediation Non Performing Loans (NPL) in affecting to Earning Per Share (EPS); 11) Soundness Level of Banks had contribution in mediation Good Corporate Governance (GCG) in affecting to Earning Per Share (EPS); 12) Soundness Level of Banks had contribution in mediation Earnings in affecting to to Earning Per Share (EPS); and 13) Soundness Level of Banks had contribution in mediation Capital Adequacy Ratio (CAR) in affecting to Earning Per Share (EPS).

RESEACH METHODS
This research theoretically aims to examine those influence from bank's financial performance on soundness level of banks and its impact to bank's earning per share by mathematical model approach and hypothesis test as a way of testing. These independent variables in this research were Non Performing Loans (NPL), Good Corporate Governance (GCG), Earnings (Profitability) using ROA (Return on Assets) as a parameter, and Capital Adequacy Ratio (CAR). The intervening variable in this research was the soundness level of Banks. The dependent variable in this research was Earning Per Share (EPS). These population and sample were Commercial Bank from BUKU IV Category which consisted of six banks, namely PT Bank Mandiri (Persero) Tbk, PT Bank Rakyat Indonesia (Persero) Tbk, PT Bank Negara Indonesia (Persero) Tbk, PT Bank Central Asia Tbk, PT. Bank CIMB Niaga Tbk and PT Bank Danamon Indonesia Tbk for the period of 2015-2019. The sample selection used purposive sampling method and sample selection based on predetermined criteria. This research used secondary data with quantitative approach. In accordance to research objectives that would be analyzed to those impact from independent variables towards dependent variable with the help of intervening variables, path analysis methods were used and data analysis techniques used panel data regression.

RESULT AND DISCUSSION Statistical Description of Research Variables
From these research results, it is known that average NPL ratio was 2.54%, this had indicated that NPL value was still within the specified maximum NPL limit stipulated by Bank Indonesia. The average GCG ratio was 1.7%, this means that Bank was carried out of its own GCG (self-assessment) routinely. The average ROA ratio was 2.88% with maximum ratio was 4.20% and minimum ratio was 0.47% with standard deviation of 0.93%. CAR from BUKU IV Bank showed that these average ratio was 20.80% with maximum ratio was 24.20% and minimum ratio was 16.28% with standard deviation of 1.96%, this means that CAR met the requirements which stipulated by Bank Indonesia (at least 8%). The average from BUKU IV bank's soundness level of banks ratio showed 88.34% with maximum ratio was 96.67% and minimum ratio was 76.67% by standard deviation of 4.61%. The average Earnings Per Share (EPS) ratio from BUKU IV bank showed that 535% with maximum ratio was 1.159% and a minimum ratio was 34.00% with standard deviation of 334%.

Panel Data Regression Model Selection
Based on the output results from the F Cross-section Probability Value from models I & II, there were differences. The Cross-section F probability value from model I was 0.0014 while the Cross-section F probability value from model II showed zero where both probability values were less than 0.05 alpha so that it could be concluded that the null hypothesis from the Chow test results (model from the common effect were more precise than the fixed effect model) was rejected. The conclusion means that the fixed effect model which more appropriate to explain those impact from independent variables towards the dependent variable compared to CE model. Since the FE is the most appropriate model, the next step is to perform the Hausman test.  (2020) The results from Hausman Model I and Model II tests had different values, however the two random cross-section probability values for model I were greater than alpha 0.05, so the null hypothesis from the Hausman test (RE model was more precise than FE model) for model I accepted. Meanwhile, these results from model II test showed these probability value from random cross-section 1, meaning that RE model was not precise so the FE model was chosen to explain these model II equation. From these model selection result, it could be seen that the model which used for path analysis was RE for model I and FE for model II. In the future, RE model I and FE model II will be used to examine those research hypothesis as well as explain those impact from independent variable (exogenous) towards dependent variable (endogenous).

Multicollinearity and Normality Assumption Test
Based on these data output from multicollinearity test, it could be explained that the correlation between independent variable of NPL and GCG did not contain a value that close to number 1, even it tends to be small below 0.4, only CAR with value of 0.6. This proved that the correlation between independent variables was weak so it could be said that there were no multicollinearity symptom. Which means that multicollinearity assumption was fulfilled.  (2020) Based on output results, it showed that the Jaque-Bera Test statistics for models I and II both exceeded 0.05 and this showed that these results from residual normality test in both models were declared normally distributed so the normality assumption was fulfilled.

Model Feasibility Test
Based on the results of panel data regression, random effect for model I and fixed effect for model II, the feasibility test from model was measured from the F test smaller than the error rate (alpha) 0.05 so both models were declared fit to use and to explain those impact from independent variable towards dependent variable. This means that these path analysis model formed could be explain those influence from independent variable towards dependent variable.

Hypothesis Examination
These regression equation from path analysis model could be formulated as follows: TK= -0,3700NPL -0,5272GCG + 0,1299ROA + 0,1530CAR + ε1 EPS= -0,2420NPL + 0,03745GCG + 0,2880ROA + 0,0740CAR -0,1710TK + ε2  (2020) Based on these results from direct influence test, it could be interpreted as follows: 1) The t-statistic probability value was 0.0350, which is smaller than alpha 0.05, so it could be explained that first hypothesis was accepted. Meaning that NPL had significant impact towards Soundness Level of Banks. 2) The t-statistic probability value was 0.0001 (less than 0.05) this was indicated that these second hypothesis was accepted, meaning that GCG had significant influence towards Soundness Level of Banks.
3) The t-statistic probability value was 0.5230 (greater than 0.05) which indicated that the third hypothesis was rejected, meaning that ROA had no impact on Soundness Level of Banks. 4) The t-statistic probability value was 0.1979, which greater than alpha 0.05, so it could be explained that the fourth hypothesis was rejected. Meaning that CAR had no affect on Soundness Level of Banks. 5) The t-statistic probability value was 0.0328 which smaller than alpha 0.05, so it could be said that the fifth hypothesis is accepted, which means that NPL had significant influence over EPS. 6) T-statistic probability value was 0.0008 which smaller than alpha 0.05, so it could be explained that the sixth hypothesis was accepted. This means that GCG had significant influence towards EPS. 7) The t-statistic probability value was 0.0530 which greater than alpha 0.05 so it could be explained that the seventh hypothesis was rejected, meaning that ROA had no effect on EPS. 8) The t-statistic probability value was 0.4217 which greater than alpha 0.05, so it could be explained that these eighth hypothesis was rejected, which means that CAR had no influence towards EPS. 9) The t-statistic probability value was 0.1134 which greater than alpha 0.05, so it could be explained that these ninth hypothesis was rejected, meaning that Soundness Level of Banks had no effect towards EPS.

Figure 2. Path Analysis Chart
There were five exogeneous variables, where four of them only act as exogeneous variables (stated as independent variables), while another one has double roles as endogent variables as well (referred as intervening variables). Based on these indirect effect test result, it was shown that: 1) The direct effect from NPL on EPS was -0.2420 while these indirect effect was 0.0630.
Because these direct effect was greater than indirect effect (regardless of the negative sign), Soundness Level of Banks had not functioned well as an intervening variable.
Meaning that the tenth hypothesis of this research was rejected, so Soundness Level of Banks did not have contribution in mediation NPL in affecting EPS. 2) The direct effect from GCG towards EPS was 0.3745 while the indirect effect had 0.0899.
Because these direct effect was greater than indirect effect one, so Soundness Level of Banks had not functioned well as an intervening variable. Meaning that the eleventh

1) Test results from first hypothesis showed that NPL had significant influence towards
Soundness Level of Banks with negative regression coefficient, so NPL had significant negative influence towards Soundness Level of Banks. This means that an increase in NPL ratio will decrease the value of Soundness Level of Banks and conversely. 2) Test results from second hypothesis showed that GCG had significant influence towards Soundness Level of Banks with negative regression coefficient, so GCG had significant negative impact towards Soundness Level of Banks. This means that an increased or decreased in the value of GCG had none influence to Soundness Level of Banks. 3) Test result from third hypothesis was indicated that ROA had no significant influence on Soundness Level of Banks with positive regression coefficient, so ROA had no significant positive influence on Soundness Level of Banks. This means that an increased or decreased in ROA value did not have positive impact towards Soundness Level of Banks. 4) Test result from fourth hypothesis shows that CAR had none significant positive influence towards Soundness Level of Banks. This means that an increased or decreased in CAR value did not have positive impact on Soundness Level of Banks. 5) Test results from fifth hypothesis showed that NPL had significant impact on EPS with negative regression coefficient, so NPL had significant negative influence on EPS. 6) Test results from sixth hypothesis showed that GCG had significant impact on EPS with positive regression coefficient, so GCG had significant positive effect towards EPS. Meaning that increased in the value of GCG will result in increase on value of EPS and conversely. 7) Test result from seventh hypothesis showed that ROA had no significant impact towards EPS with positive regression coefficient (table 4.12), so ROA had no significant positive impact towards EPS. Meaning that an increased or decreased in ROA value did not have positive impact towards EPS. 8) Test results from eighth hypothesis showed that CAR did not have significant impact to EPS with positive regression coefficient, so CAR did not have significant positive impact towards EPS. Meaning that an increase or decrease in CAR value did not have positive impact against EPS. 9) Test results from ninth hypothesis showed that Soundness Level of Banks had no significant influence towards EPS with negative regression coefficient, so Soundness Level of Banks did not have significant negative impact to EPS. 10) Test results from tenth hypothesis showed that Soundness Level of Banks did not mediate NPL in affecting EPS. So there had none influence from Soundness Level of Banks in mediation that impact between NPL and EPS. This occurs because these direct relationship between NPL was significant and greater than indirect effect one, whereas Soundness Level of Banks as intervening factor, so the effect from NPL on EPS did not require the role of Soundness Level of Banks as intervening factor. 11) The test results from eleventh hypothesis shows that Soundness Level of Banks did not mediate GCG in affecting EPS. So there had no influence of Soundness Level of Banks in mediation on the effect of GCG and EPS. This happened because these direct relationship of GCG was significant and greater than indirect effect one, where Soundness Level of Banks as intervening factor, so the effect of GCG on EPS did not require the role of Soundness Level of Banks as intervening factor. 12) Test results from twelfth hypothesis showed that Soundness Level of Banks did not mediate ROA in affecting EPS. So there had no effect from Soundness Level of Banks in mediation that affect between ROA and EPS. This happened because direct relationship between ROA was significant and greater than indirect effect one, whereas Soundness Level of Banks as intervening factor, so those impact from ROA on EPS did not require Soundness Level of Banks as intervening factor. 13) Test results from thirteenth hypothesis showed that Soundness Level of Banks did not mediate CAR in affecting EPS. So there has no effect from Soundness Level of Banks in mediation that impact between CAR and EPS. This happened because these direct relationship of CAR was significant and greater than indirect effect one, whereas Soundness Level of Banks as intervening factor, so the effect from CAR on EPS did not require Soundness Level of Banks as intervening factor.

CONCLUSION AND SUGGESTION Conclusion
Based on these research results and those analysis from discussions which have been presented, it could be concluded from these thirteen hypothesis test results as described as follows: 1) NPL had significant negative impact on Soundness Level of Banks; 2) GCG had significant negative impact towards Soundness Level of Banks; 3) ROA had no significant positive effect on Soundness Level of Banks; 4) CAR had no significant positive impact towards Soundness Level of Banks; 5) NPL had significant negative effect on EPS; 6) GCG had significant positive effect on EPS; 7) ROA had no significant positive impact on EPS; 8) CAR did not have significant positive impact against EPS; 9) Soundness Level of Banks did not have significant negative effect on EPS; 10) Soundness Level of Banks did not mediate the NPL in affecting towards EPS; 11) Soundness Level of Banks did not mediate GCG in affecting towards EPS; 12) Soundness Level of Banks did not mediate ROA in affecting towards EPS; and 13) Soundness Level of Banks did not mediate CAR in affecting towards EPS.

Suggestion
To contribute towards these research results, several recommendations could be presented as follows: 1) To anticipate the tendency of increasing Non Performing Loans (NPL) or the risk of nonperforming loans (default), banks should be more prudent and selective in providing and expanding credit in addition to taking steps for the non-performing loans settlement in order to reduce the cost of provision for non-performing loans, which in the end would lead to improvement in financial performance and soundness level of banks. 2) In implementing and evaluating the components of Good Corporate Governance (GCG), banks are obliged to complete the GCG Self-Assessment Working Paper by seeking an annual report that is published in order to determine the value of the composite rating from each criteria by comparing it with those indicators that has been stipulated by Bank Indonesia. Banks which successfully execute and implement the good corporate governance (GCG) properly would undoubtedly give a positive response towards consumer and investor, therefore the bank would obtain customers and increase the bank stock prices more easily. The better the implementation and the management of GCG at a bank, the better the bank's financial performance because the bank's goal could be achieved. 3) To increase profitability, banks should demonstrate the management capability of a bank in managing the level of overall bank performance effectiveness in order to obtain profits by utilizing its assets. Bank Indonesia has an assessment indicator which has been stipulated that good ROA ratio is above 1.5%. The greater the ROA of bank, the better the bank's position in terms of asset usage. The bigger the level of ROA ratio would indicate that the bank has a large amount of profit, which would lead to larger amount of dividends being distributed to the shareholders. 4) In accordance with the regulation of Bank Indonesia which indicate the minimum CAR ratio level is 8% as the stipulated provisions mean that banks have to always strive to maintain the minimum CAR ratio to be the same or larger. If the CAR decrease, then the profitability would also decrease which leads to lower profitability. This issue could occur since bank has not been maximized in providing loans, ultimately resulting and driven the bank capital into a negative spread and the imbalance of increasing in assets with the addition of capital. The low CAR ratio level could lead to a decline in public's confidence which could conclusively lower profitability. When the value of Capital Adequacy Ratio (CAR) is relatively close to its minimum, then bank would need to have additional capital injection. Several ways to obtain the capital injection to increase the equity value of the bank which could be implemented through Right Issue and Initial Public Offering (IPO). 5) To increase the soundness level of banks, the assessment could be implemented using the calculation of financial ratios against those aspects and conditions of Financial Performance of a bank using RGEC method (Risk Profile, Good Corporate Governance, Earnings and Capital). In principle, the better the financial performance of a bank, the more profits and benefits it would give to shareholders, which could trigger an increase in stock prices. 6) In facing intense global competition, every bank, especially for those that have gone public, would definitely be required to have a good and reliable financial performance. The financial performance is not only in managerial terms, but also in financial terms. Bank financial performance is often used as the basis for consideration for investors in decisionmaking. One of them could be seen from Earning Per Share (EPS) factor.