“The impact of discretionary loan loss provision of sharia financing on financial performance”

This study aims to investigate the role of discretionary loan loss provision of sharia financing on the Islamic commercial banks’ financial performance in Indonesia. Partial Least Squares-Structural Equation modeling (PLS-SEM) is used to examine the relationship between loan loss provisions and financial performance in 13 Islamic commercial banks for 4.5 years. The analysis of the outer model shows that the probability of default and loss given default are determinants of loan loss provision, while financial performance is determined by return on assets, non-performing financing, net operating margin, and operating costs on operating income. The results of this study indicate that loan loss provisions have a direct effect on financial performance. Further investigation shows that the return on sharia financing contributes to increasing the impact of loan loss provisions on financial performance (indirect influence). The findings contribute to the literature by showing that discretionary loan loss provision can occur in sharia financing. The study is very important in terms of awareness of management behavior related to financial performance. The study has implications for management policies related to the prerequisites of potential clients.


INTRODUCTION
Sharia is one of the solutions offered by Islamic banks to avoid riba transactions. At the time the contract is made, both Islamic banks (the investor and the customer) do not set specific interest rates as profits received or expenses to be paid. The statement of Financial Accounting Standards (PSAK, 2007, par. 04) defines sharia as a contract of business cooperation between two parties where the first party (the owner of the fund) provides all funds, while the second party (fund manager) acts as a manager. The profits are divided between them according to an agreement, while the financial loss is only borne by the owner of the fund (IAI, 2017). The definition implies that the Bank can act as both the owner of the fund and the fund manager. Sharia funds disbursed by fund owners (the Bank) are recognized as sharia investments when cash payments or transfers of non-cash assets to fund managers occur (PSAK, 2007, par. 12).
However, in practice, when banks act as fund owners, generally applying sharia is more of a financing nature (Suhartini & Anwar, 2016). As a financing product, sharia has a risk of loss due to loss or negligence of the fund manager (Bank Indonesia, 2013). The risk of loss is anticipated by forming a reserve for impairment of assets. This risk is in accordance with the Decree of the Board of Directors of Bank Indonesia http://dx.doi.org/10.21511/bbs.14(4).2019.04 No. 31/147/KEP/DIR dated November 12, 1998, which requires the establishment or allowance for earning assets losses. On the other hand, the Statement of Financial Accounting Standards (PSAK) number 50, 2006 (Revised 2014) also regulates the treatment of allowance for possible losses on earning assets in the form of allowance for impairment losses. Banks are required to establish reserves for impairment losses if there is objective evidence of impairment of financial assets as one or more events that have occurred after the recognition of the assets and have an impact on the estimated future cash flows. Allowance for impairment losses is treated as a deduction from sharia financing (Bank Indonesia, 2013).
Loan loss provisions are usually associated with bank management policies in anticipating investment loss risks and improving financial performance. Many studies have succeeded in explaining the relationship between loan loss provisions and financial performance ( The study is different from previous research. Its uniqueness lies in investigating the specific discretionary focus related to sharia financing and its measurement model.
The main purpose of this study is to analyze the influence of discretionary loan loss provisions for sharia products on the financial performance of Islamic commercial banks; secondly, to investigate the role of sharia income in the context of loan loss provision and financial performance.

Sharia in Islamic banks
Islamic banks have a variety of products based on sharia financial services. Financial services provided by Islamic banks, as well as conventional banks, cannot be separated from profit motivation. In sharia, there are two forms of contracts that generate benefits for Islamic banks, namely, a sales contract and a contract of cooperation. The sale and purchase contract creates a certain advantage for Islamic banks. The sale and purchase contract of Islamic banks offers products such as murabahah, salam, istisna, and ijarah. While in the contract of cooperation, Islamic banks offer mudaraba and musyaraka. Both of these contracts bring in uncertain profits. The sharia profit obtained by Islamic banks generally comes from contractual agreements (contracts) between banks (such as shahibul mal) and mudarib.
The development of sharia financing is slower than other financing in Islamic banks in Indonesia (Gilang, 2013;Amelia & Hardini, 2017;Ningsih, 2017). This can be due to the non-performing financial contribution (NPF), which always increases (Gilang, 2013). While Amelia and Hardini (2017) explained that sharia financing can reduce the capital adequacy ratio. One of the causes of the contract failure is the burden upon cap-ital of Islamic banks. However, this problem can be solved by improving the overall determinants of financial performance, such as return on assets (ROA), return on equity (ROA), and financing-to-deposit ratio (FDR) (Ningsih, 2017). Ajekwe et al. (2017) began their research by defining loan loss provisions as a non-cash expense for banks to take into account the possibility of future losses from bad loans. The bank assumes that a certain percentage of the loan will fail or become slow to pay. This method guarantees the solvency and capitalization of the bank if and when a failure occurs (Ajekwe et al., 2017). In terms of accounting for loan loss provisions on Islamic bank products, it can affect the bank's income and capital adequacy. When a bank enters into a sharia cooperation contract, it faces the risk that the customer may fail and return not full assets. The bank assumes that losses that actually occur are recognized as expenses, providing for reserve losses.

Loan loss provision
Loan loss provisions in Islamic banks refer to the deduction of the bank's net income to provide a failure of a financing contract or non-performing financing. In accounting, the practice of loan loss provisions is reasonable and generally permitted by banking regulations. In the development of further practice, management uses this method to manage compa-

Hypotheses
Based on previous research, this study has developed the following hypotheses:

Data
This study uses cross-sectional data for 4.5 years, from of January 2014 to June 2018. The object of this study is financial reports of 13 Islamic com-mercial banks registered with Bank Indonesia and the Financial Services Authority (OJK). This study observes the financial performance of Islamic Commercial Banks and Sharia Banking Statistics published by each bank and OJK.

Variable measurement
The variables in this study include three latent variables, namely, Loan Loss Provision (dependent), Return on Sharia (intervening) and financial performance (independent). Each of these variables is measured using the following indicators (see Table 1).

EMPIRICAL RESULTS AND ANALYSIS
Partial Least Square-Structural Equation Modeling (PLS-SEM) is used to test the effect of the loan loss provision on the financial performance of Islamic commercial Banks through Return on Sharia. Previously, the path model in the form of a diagram will be used to determine the direction of the relationship between the latent variables (LLP, Return on Sharia, financial performance). The path diagram at the beginning of the analysis is shown in Figure 1.
PLS-SEM analysis produces statistical values grouped into two models, namely, the outer model and inner model.   The value of cross loading shows the connection between indicators with latent variables. The association between latent variables and forming indicators must be higher than the correlation between indicators with other latent variables (Aminah, Suharsono, & Ahmad, 2016). Cross loading value in Table 3 shows that each indicator value on certain latent variables is greater in the value of other indicators. This points out that the indicators are set to form latent variables. To strengthen these results, the Fornell-Larcker criteria were tested (see Table 4). Table 4 shows the correlation values between the same latent variables, which have greater Fornell-Larcker criterion values than between different latent variables. The test results concluded that latent variables have valid criteria.

Inner model
Testing of the endogenous (Financial Performance) and exogenous (Loan Loss Provision) variable partial relationship is done by inner model analysis. These tests include an accuracy test of the model, a significance test and the relevance of the path coefficients on the structural model. Tables 6  and 7 show the statistical values used to analyze the inner model. The goodness of fit test aims to determine whether the model is fit so that it can be used to assess relationships between variables (Schumacker & Lomax, 2010). Some criteria are used to determine the model accuracy, for example, Chi-square, RMSEA, GFI, AGFI, and TLI. Table 6 shows the statistical values of the test results of the criteria for testing the model accuracy (Goodness of Fit) and the size of the critical value. Analysis of the comparison between the test results and the size of the critical value of each test criterion shows that all requirements conclude the model is fit.
Testing of predictive accuracy is indicated by the statistical value of the coefficient of determination (R 2 ) on the path model that is equal to 0.780. These results explain that financial performance can be predicted by loan loss provisions, both directly and indirectly, through Return on Sharia of 78%. While the remaining 22% is likely to be explained by other variables not found in the path model.
The partial analysis to determine the relationship between the variables of Loan Loss Provision, fi-nancial performance, and Return on Sharia uses the significance test and the path coefficients on the structural model. Table 7 shows the statistical values of estimated coefficients and their significance level.
Testing of the first hypothesis is shown by the estimated coefficient value of path ROS ← LLP with a positive sign of 0.748 and a significance level of 0.0154 (C.R = 5.89) smaller than α = 0.05. The test results concluded that the higher the loan loss provision, the higher the sharia income, and vice versa, the lower the loan loss provision, the lower the sharia income.
The direct effect of loan loss provisions on financial performance is indicated by the magnitude of the estimated coefficient of path FP ← LLP with a positive sign of 0.249 and a significance level of 0.0251 (C.R = 4.45) smaller than α = 0.05. This result concludes that the second hypothesis is accepted. Thus it can be stated that the higher the loan loss provision, the higher the financial performance, and vice versa, the lower the loan loss provision, the lower the financial performance.
The test results of the third hypothesis conclude that the higher the Return on Sharia, the higher the financial performance, and vice versa, the lower the Return on Sharia, the lower the financial performance. This conclusion is supported statistically as indicated by the magnitude of the estimated coefficient value of the path of FP ← PM with a positive sign of 0.498 and a significance level of 0.0314 (C.R = 3.83) smaller than α = 0.05.
The indirect effect of loan loss provisions on financial performance through Return on Sharia is shown by the estimated coefficient value of path PF ← ROM ← LLP with a positive sign of 0.373 and a significance level of 0.014 (C.R = 2.195). This result concludes that the fourth hypothesis which states that loan loss provisions have an indirect effect on financial performance through return on sharia, is accepted.

DISCUSSION
The The outer model also examines the determinants of the Sharia commercial banks' financial performance, which includes four manifest variables, namely: return on assets (ROA), non-performing financing (NPF), net operating margin (NOM), and the ratio of operating expenses to operating revenue (OER). This study concludes that the four variables have a strong correlation with the variable financial performance as the latent variable. Testing of these financial ratios is in line with research conducted by Zulfikar and Sasongko (2016). In different ways, these financial ratios explain financial performance affecting the competitiveness of Islamic banking services in Indonesia (Zulfikar & Sasongko, 2016).
The results of the analysis of the inner model concluded that the proposed hypotheses (H1 s.d. H4) in this study were significantly supported by statistical values both partially and throughout the model. Each exogenous variable (loan loss provision and return on sharia) explains the positive relationship to endogenous variables (financial performance).
Testing of the first hypothesis proves that loan loss provisions have a significant positive effect on return on sharia. Proper management of loan loss provisions will increase return on sharia. This study agrees with Liu (2017) that loan loss provisions are part of the anticipatory steps taken by banks to reduce financing risk. Knowledge of probability of default and loss given default as a determinant of discretionary loan loss provision is an early warning for bank managers to reduce the failure to obtain income derived from sharia financing. Determination of probability of default and loss given default is an effective way for bank management to increase financing income (Misankova et al., 2015).
This study proves that loan loss provisions affect financial performance. The statement is supported by the result of path coefficient test with positive and significant directions. The discretionary loan loss provision has a positive effect on financial performance as measured by return on assets (ROA), non-performing financing (NPF), net operating margin (NOM), and the ratio of operating expenses to operating revenue (OER). The discretionary loan loss provision effectively demonstrates the ability of banks to increase sharia investment returns (Meutia, 2017). Sharia investment returns will have a positive direct impact on increasing ROA, NOM, OER and decreasing NPF (Aminah et al., 2016).
This study also proves that return on sharia is a mediating variable of the indirect effect of loan loss provisions on financial performance. Even the influence of loan loss provisions on financial performance is stronger if it is through return on sharia compared to its direct influence. This finding strengthens previous research that income derived from investing in bank financing is the main focus of management in managing loan loss provisions (