Yani Monalisa* | Sugiarto
© 2022 IIETA. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
OPEN ACCESS
The urgency of the internal model of default credit for the rural bank in Indonesia is increasing due to the recent acceleration of the increase in Non-Performing Loans at rural banks. This research will formulate an internal model of default credit that can reduce the level of default risk in rural banks based on the explanatory sequential design of mixed research method. The findings of quantitative analysis integration by Chi-Square Analysis, Discriminant Analysis, and Logistics Regression analysis at BPR BKK Pekalongan Regency in 2021 will be explored further by qualitative research. Using nine variables that affect debtors' failure, this study finds that the interest rate is a variable that consistently affects the status of default loans using the integration of 3 analyses and qualitative analysis. The study results indicate that rural banks need to pay more attention to determining credit interest rates when prospective debtors apply for credit. The determination of interest rates is related to compensation for risks faced by rural banks in connection with asymmetric information about the debtor's ability to pay while considering the interest rate determined by the IDIC.
rural bank, an internal model of default credit, explanatory sequential design
Credit is one component that plays a vital role in supporting the spinning of the economy. The credit provided by the bank has an extensive influence in all areas of life, especially in the economic field of the ability of credit to turn the wheels of the economy [1, 2].
Public can get credit from many channels, such as bank, peer-to-peer lending [3, 4], crowdfunding [5], and other financing institutions that are informal credit [6]. At Indonesia’s rural banks, there was an increase in credit from IDR 98.2 trillion in 2018 to IDR 116.6 trillion in 2021 or 5.88% per year [3]. Unfortunately, this increase in credit volume is accompanied by the rise in default loans every year. As seen in Table 1, the average increase in default loans in rural banks from 2018 to 2021 reaches 14.73% annually. Rural banks’ Non-Performing Loan (NPL) increased from 2018 to 2021, with an average of 7.77% per year. Rural banks' increase in NPL must be wary of increasing rural banks' default risk.
Although credit distribution at rural banks is still smaller than the size of lending at Commercial Banks, however if the repayment of loans at rural banks is problematic, the impact will be on the security of third parties’ funds placed in rural banks. As with the function of a bank as a financial intermediary, rural banks will increase the interest rate on loans if it is deemed that prospective customers have a high default risk. The addition of the credit interest rate was carried out as one of the anticipations of the possibility of default credit on the debtor. This condition will certainly burden the debtor because it will increase its burden. The problem of high NPLs in rural banks cannot be denied; one of the reasons is the lack of rigorous analysis of the quality of prospective debtors when the rural bank finally decides to provide credit. In addition, other causes were also found, such as the decline in debtors’ business performance due to deteriorating economic conditions, etc. Deterioration in economic conditions is a systematic risk whose impact is generally felt by businessmen, including rural banks. However, suppose the cause of the increase in NPL is a lack of rigorous analysis of the payment capability of prospective debtors, then in that case, such risks can be mitigated. The internal model of default credit can function as a filter in granting credit approvals from prospective debtors to reduce rural bank default risk. The tools related to the rural bank internal default credit model are not yet available, even at the national rural bank level.
The urgency of the internal model of default credit for Rural banks is increasing due to the accelerated increase in Non-Performing Loans (NPL) in rural banks lately. For this reason, this study was made to produce a credit model that can reduce the level of NPL and default risk in rural banks. The research carried out resulted in several novelties as follows:
(1) As far as researchers know from accessible sources, no default credit model has been proven to reduce NPLs in rural banks effectively; (2) The research will involve the demand side (credit customers in rural banks) and the supply side (officers at rural banks, especially in the credit department) to obtain a model applied in rural bank operational activities; (3) Research related to rural banks carried out by previous researchers is carried out in a qualitative approach only or a quantitative approach, not mixed research. This research will use mixed research to sharpen the rural banks’ credit model results; (4) Research in rural banks is mainly related to quantitative research, which primarily discusses financial data and rural bank financial ratios to assess the performance of rural banks [7-10]. As far as the researcher's knowledge, based on accessible sources, no comprehensive research has been found that discusses the point of view of customer behavior of credit customers at rural banks related to internal default credit and default risk.
Table 1. Credit in rural banks in 2018-2021
|
|
|
|
|
|
In billions of Rupiah |
|
|
|
Collectibility |
2018 |
2019 |
2020 |
2021 |
CAGR |
|
1 |
Current |
91,959 |
101,379 |
102,775 |
99,464 |
2.65 |
|
2 |
Special Mention |
|
|
|
9,280 |
- |
|
3 |
Sub-Standard |
1,137 |
1,373 |
923 |
982 |
-4.77 |
|
4 |
Doubtful |
1,287 |
1,512 |
1,292 |
1.058 |
-6.32 |
|
5 |
Lost / Default |
3,837 |
4,520 |
5,780 |
5,795 |
14.73% |
|
|
Total |
98,220 |
108,784 |
110,770 |
116,579 |
5.88% |
|
|
NPL |
6,261 |
7,405 |
7,995 |
7,836 |
7.77% |
|
|
NPL Ratio |
6.37% |
6.81% |
7.22% |
6.72% |
|
CAGR=compound annual growth rate = $\left(\frac{ { value \quad ending }}{{value \quad beginning }}\right)^{1 / t}-1$
According to Law No. 10 of 1998, Banks in Indonesia are business entities that aim to collect funds from the public in the form of savings and then distribute them back to the public in the form of credit or other forms to improve the standard of living the people. One of the reasons for the concentration of bank business in lending is the nature of the bank's business as an intermediary institution between the surplus side and the deficit side. The primary source of bank funds comes from the public, so morally, they must channel it back to the community in credit [11, 12].
Implementing internal non-performing loans based on the Basel Committee on Banking Supervision (BCBS) is left entirely to each bank because the model developed must be accountable. This makes each bank have a different business model and impacts the credit risk faced by banks [13]. Credit risk is caused by the debtor's inability to repay part, or all of the loan payments provided by the bank [14]. In anticipating credit risk early, the bank determines the collectibility of each loan that has been given into five categories, namely, Collectibility 1 (Current), Collectibility 2 (In Special Mention), Collectibility 3 (Substandard), Collectibility 4 (Doubtful), and Collectibility 5 (Lost/Default) [15].
In the banking world, information asymmetry can occur between lenders (banks) and borrowers (debtors), where borrowers have more information than banks regarding sources of income and their ability to pay [16]. Thus, the relationship with customers (borrowers) in providing credit arises due to information asymmetry [17].
It is hoped that with a good relationship with customers, the bank can find out the actual condition of the borrower. For this reason, banks are required to conduct qualitative analysis based on the 5C [18] or 7P [12]. The 5C principles include: (1) Character is the nature or character of the customer; (2) Capacity, namely the analysis used to see the customer's ability to pay credit; (3) Capital is to assess the capital owned by the customer to finance credit; (4) Condition, namely the current and future general conditions of course; (5) Collateral is a guarantee given by a customer to a bank to finance the credit they propose. Meanwhile, credit analysis based on 7P Principles includes (1) Personality or personality is an assessment used to determine the personality of the prospective customer; (2) Purpose, namely the purpose of taking credit; (3) Party, meaning that in distributing credit, banks are divided into several groups; (4) Payment is a method of credit payment by customers; (5) Prospect, namely to assess future expectations, especially for the object of credit being financed; (6) profitability, meaning that loans financed by banks will provide benefits for both parties, both banks, and customers; (7) Protection, meaning the protection of the object of credit being financed.
Another method that banks can use to reduce the level of non-performing loans when they want to disburse credit, especially to types of credit that have a high level of sensitivity, is by applying the principle of prudence (selectiveness in choosing debtors) through an assessment of the profile of prospective debtors [19] including the value of the collateral held precisely and accurately [14].
To suppress the occurrence of more significant credit losses, the bank must control the credit distribution process properly and correctly from an early age. There are three methods that banks can use in controlling credit quality [20], namely adjusting the credit ceiling to the debtor's risk profile, monitoring, and fostering debtors. In addition to these three methods, banks must also properly value collateral because they can be a source of liquidity when debtors fail to pay. Collateral can be used to bind the debtor's responsibility for his obligations. The weak value of collateral can be a factor in the reluctance of debtors to fulfill their obligations [21].
Several authors have examined the value of relationship loans. Banks typically produce valuable and unique information about the financial health of their customers' companies; however, most of this information cannot be easily transferred, as long-term loan relationships exist [22]. For example, companies that have better relations with banks bear lower interest rates [23, 24], enjoy better credit access and more favorable term of loans [25, 26], and have more financial certainty. Problems and thus tend to enjoy lower costs [27, 28], and have a positive (negative) stock price reaction. In addition, Gadanecz et al. [29] mentions that companies with more information asymmetry have more benefits than loans.
Several studies have concluded that the characteristics of debtors and the qualifications of credits influence the status of credit. Age, family size, and income all impact rural credit accessibility [30, 31]. There is substantial evidence that the interest rate is a positive factor [16, 32-34]. In a number of developing nations, education is one of the most significant determinants of credit accessibility. Educated household heads imply superior knowledge, farming skills, and credit market information [31, 32].
Many commercial banks are only willing to make lending decisions based on collateral because the size of a landholding is more acceptable for institutional lenders as risk management and loan security. It is believed that collateral increases the likelihood of household repayment [35, 36], which is why the majority of poor and small households cannot borrow [37]. In Pakistan, the lack of collateral prevents farmers from gaining access to rural credit [38].
The research was conducted using an explanatory sequential design, a mixed method of QUAN à qual research. The main focus is to explain the results obtained from the implementation of quantitative research by exploring specific results in more detail using qualitative research. In the first stage, quantitative data were collected, and the results were analyzed. Then proceed with the collection of qualitative data, which is used to explain the results of the quantitative analysis. The classification of mixed research methods used includes triangulation to seek convergence, proof, and correspondence of results from different methods, complementary to seek elaboration, improvement, clarification of results from one method with results from other methods, and expansion to broaden the scope of the investigation, gaining perspectives. Broader scope uses different methods for different investigation components [39].
3.1 Sampling method
This research was conducted at PT. BPR BKK Pekalongan Regency. PT. BPR BKK Pekalongan Regency is a Trading Company for the People's Credit Agency of the District Credit Agency, a Regional Owned Enterprise (BUMD) with 9 Branch Offices and 1 Head Office with the head office address Jl. Mandurorejo No. 344 Kajen, Central Java.
The research was conducted on debtors and the management of PT. BPR BKK Pekalongan Regency. Information extracted from both debtors and managers will complement each other in extracting critical variables related to default loans from debtors. With the triangulation technique, relevant data will be extracted from reliable sources to obtain reliable findings.
Against debtors, the research was conducted on members of the debtor population, namely all debtors of PT. BPR BKK Pekalongan Regency. The sample population is customers who are still debtors from 2021 until now. The sampling technique used is purposive sampling with the criteria of debtors who have complete data. From 6,766 debtors, 6,727 debtors were obtained from PT. BPR BKK Pekalongan Regency, which has complete customer data, will then be the target of research observations. The research uses the snowball sampling technique by extracting information from key personnel who have the authority to make decisions regarding credit distribution.
3.2 Data collection
Data mining was carried out using triangulation techniques using the integration of observation techniques and communication techniques in the period from June 2021 to March 2022, covering the variables in the PT. BPR BKK Pekalongan Regency and demographic data from debtors and data obtained from crucial personnel of PT. BPR BKK Pekalongan Regency has the authority to make decisions regarding credit distribution.
3.2.1 Secondary data
The data used is the data of all debtors of PT. BPR BKK Pekalongan Regency that meets the data completeness aspect includes variables in the database and demographic data from debtors. The number of debtors is 6,766 customers, and only 6,727 customer data is considered due to incomplete customer data.
3.2.2 Primary data
In addition, the researchers explored primary data by conducting in-depth interviews with several debtors as key informants. The researchers also have in-depth interviews with the BPR BKK Pekalongan Regency's director and the managers have the authority to decide on credit distribution.
3.2.3 Literature study
Researchers use a literature study to explore the factors that have a significant effect on credit status that have a solid potential to influence debtor failure.
3.3 Analysis techniques
The findings from the quantitative analysis were confirmed and corroborated with information obtained from the core personnel of PT. BPR BKK Pekalongan Regency has the authority to make decisions regarding credit distribution. The internal credit model is built by integrating Chi-Square analysis, Discriminant analysis using the stepwise method, and Logistics Regression analysis using the stepwise method. Data on the dependent variable, credit status, as many as 6.727, will be analyzed using the Logistic Regression method, which uses a scale of 0 (not default) and 1 (default). In addition, to sharpen the results of the analysis, this study also divides based explicitly on the type of credit provided, namely Working Capital Loans (n=4,959), Consumption Loans (n=1,393), and Investment Loans (n=375).
The use of analysis in general (all loans in total without differentiating the type of credit) and precisely (each type of credit) is to see which independent variables consistently affect the dependent variable.
3.3.1 Chi-square analysis
A chi-square analysis is a correlation analysis that studies the close relationship of two or more variables of concern to categorical research variables [40]. In conducting the analysis, a level of significance of 0.05 was used. The independent variable with P-value <0.05 was stated to be significantly related to the dependent variable with the following calculations (Eq. (1)).
$\chi^2=\sum\left(\frac{O-E}{E}\right)^2$ (1)
where, χ2 is the chi-square test statistic; Σ is the summation operator (it means “take the sum of”); O is the observed value; E is the expected value.
3.3.2 Discriminant analysis
Discriminant analysis is part of multivariate statistics, helpful in examining and modeling relationships between variables [39]. In this analysis, a significance level of 0.05 was used.
3.3.3 Logistics regression
Logistic regression analysis is used when the dependent variable takes a dichotomy or binary, and the independent variable is a continuous variable, categorical, or both.
The logistic regression model used is (Eq. (2)).
${\mathrm{P}\left(\pi_i=j\right)=\frac{1}{1+e^{-(z)}}}$ (2)
where, z=β0+β1 X1+β2 X2+ …. +βp Xp; β0, β1, β2 …βp are regression parameters; X1, X2, … Xp are explanatory variables;
π is the probability that credit will default.
In logistic regression, stepwise regression is used. The regression equation starts with a single independent variable and then adds or removes independent variables used in the final model. Thus, the independent variable that significantly influences the dependent variable will stay in the model [41].
3.4 Research variables
From the available debtor database, the researcher explores the possible variables to be empowered, informing the internal model of default credit of PT. BPR BKK Pekalongan Regency, which is the research objective (Table 2).
Table 2. Research variables
|
Variables |
Label |
Measurement Scale |
Definition |
|
Credit status |
Y |
Nominal |
The status of loans, divided into (1) Default and (0) Not Default |
|
Credit type |
X1 |
Nominal |
The type of loans, divided into (1) Working Capital Loans, (2) Consumption Loans, and (3) Investment Loans |
|
Occupations |
X2 |
Nominal |
The occupations of debtors, divided into (1) employee, (2) self-employment and (3) unemployment |
|
Age |
X3 |
Ratio |
Age of debtors at the time of borrowing |
|
Number of collaterals |
X4 |
Ratio |
The number of collaterals at the time of borrowing |
|
Beginning Ratio |
X5 |
Ratio |
The beginning ratio is collateral value divided by limit facility* |
|
Ending Ratio |
X6 |
Ratio |
The ending ratio is collateral value divided by debit balance** |
|
Term of loans |
X7 |
Ratio |
The duration of loan contracts. |
|
Frequency of instalments |
X8 |
Nominal |
The frequency of loan payments, divided into (1) monthly and (0) otherwise |
|
Interest rate |
X9 |
Ratio |
The interest rate varies depending on the amount of money borrowed and the initial scoring |
*The collateral value is the selling price of the property through auction after calculating the risk of the sale and is used to determine the maximum loan amount. The limit facility is the maximum amount of financing (credit value) provided by the bank.
**The Debit Balance is the principal balance of the loan ceiling that has been agreed upon in the credit agreement and will be reduced when the debtor makes regular installments. In general, banks provide loans of a maximum of 70-80% of the estimated value of the collateral.
4.1 General characteristics
In Table 3, from 6,727 debtor data at BPR BKK Pekalongan Regency, 91% of credit status is in the non-performing category, and only 9% of credit status is the default. Loans disbursed, 73.7% are working capital loans, 20.7% are consumption loans, and the rest are investment loans. This is in line with the commitment of BPR BKK Pekalongan Regency to support SMEs in obtaining credit. In addition, it can be seen from the job characteristics of the debtors in BPR BKK Pekalongan Regency that 63.6% are self-employed, 26% are employees, and the rest are homemakers or students, 10.3%. The frequency of 99.1% of installments is dominated by monthly installment payments. This is intended to reduce the risk of default.
BPR BKK Pekalongan Regency also provides a loan program without collateral to support MSME activities. It was recorded that 828 loans were disbursed without collateral, and the rest used various types of collateral, from 1 collateral – to 16 collateral (Appendix 1). Meanwhile, Table 4 shows that the age range of debtors is from 20 years to 82 years, with an average age of 44 years.
The average beginning ratio of 3.47 is greater than the average ending ratio of 6.90 times. In the ending ratio, the debtor has repaid part of the loan. The term of loans also varies from a short term of loans of 5 months to an extended term of loans of 276 months. The interest rate set at BPR BKK Pekalongan Regency is 9% - 16% per annum, but in practice, the interest rate is in the range of 6% - 21%, depending on the number of funds borrowed and the credit score of the debtor.
Table 3. General characteristics of the debtor (1) in BPR BKK Pekalongan regency
|
Variables |
Option |
Value |
|
|
Credit Status |
Not Default |
6,124 |
91.0% |
|
Default |
603 |
9.0% |
|
|
Credit Type |
Working Capital |
4,959 |
73.7% |
|
Consumption |
1,392 |
20.7% |
|
|
Investment |
375 |
5.6% |
|
|
Occupation |
Employee |
1,750 |
26.0% |
|
Self-employment |
4,281 |
63.6% |
|
|
Unemployment |
696 |
10.3% |
|
|
Frequency |
Monthly |
6,666 |
99.1% |
|
Otherwise |
61 |
0.9% |
|
|
Age |
<30 years old 30-39 years old 40-49 years old >49 years old |
580 1,751 2,390 2,006 |
8.6% 26.0% 35.5% 29.8% |
|
Number of collateral |
None Enough High |
828 5,484 415 |
12.3% 81.5% 6.2% |
|
Beginning Ratio |
Nol Less Enough Good |
1,175 222 1,414 3,916 |
17.5% 3.3% 21.0% 58.2% |
|
Ending Ratio |
Nol Less Enough Good |
1,175 461 4,067 1,024 |
17.5% 6.9% 60.5% 15.2% |
|
Term of Loans |
5-12 months 13-24 months 25-36 months >36 months |
529 1,686 2,490 2,022 |
7.9% 25.1% 37.0% 30.1% |
4.2 Quantitative findings
4.2.1 Chi-square analysis
In the chi-square analysis, three types of credit will be tested using nine independent variables (X1 to X9). Meanwhile, in a more specific test, for each type of credit (working capital, consumption, and investment), the test will use eight independent variables (X2 to X9).
The Chi-Square test results, where the dependent variable uses a scale of 0 (not default) and 1 (default), can be seen in Table 5. In the test on 3 types of credit where n=6,727 all independent variables have a P-value<0.05, except variable age (χ2=7.246; P=0.0645) and variable frequency of installments (χ2=1.864; P=0.1722). This means that all independent variables affect the dependent variable, except for the age variable and the frequency of installments variable.
The test results are more specific, where each credit is tested, working capital loans (n=4,959), age (χ2=6.147; P=0.1047) and the frequency of installments (χ2=1.424; P=0.2327) have no relationship with credit status (default and not default). In consumption loans (n=1.393), only employee variable (χ2=1.023; P=0.5996) which has no relationship with credit status, the rest all independent variables have P value<0.05, which means all independent variables except employee variable, has an effect on the dependent variable. In investment loans (n=375), only the term of loans (χ2=26,551; P=0.00) and interest rate (χ2=35,171; P=0.00), which have a relationship with credit status (default and not default).
Of all the independent variables tested, it turns out that the term of loans variable and the interest rate variable consistently become two variables that have a relationship with credit status, both tested in all types of loans and in specifically for each type of loan.
4.2.2 Discriminant analysis
For discriminant analysis, the results of data processing are presented in Table 6, where three types of credit will be tested by using nine independent variables (X1 to X9) and more specific testing on each type of credit (working capital, consumption, and investment), will also be tested, but will only use eight independent variables (X2 to X9).
As shown in Table 6, for the three types of credit, the variables included in the credit model are the type of working capital loans (X1(2)), self-employment (X2(2)), age (X3), and the number of collaterals (X4), term of loans (X7), frequency of installments (X8), and interest rate (X9). The predictive accuracy of the discriminant model for these three types of credit is 72.4% in the correct classification and 72.3% in the correct cross-validation group at the time of classification.
For a more specific analysis of working capital loans, the variables included in the internal credit model are self-employment (X2(2)), beginning ratio (X5), frequency of installments (X8), and credit interest rates (X9). The prediction accuracy of discriminant model 1 for working capital loans was 80.9% correctly classified, and the 80.9% group was cross-validated correctly at the time of classification.
For the type of consumer loans, the variables included in the internal credit model are age (X3), beginning ratio (X5), ending ratio (X6), term of loans (X7), frequency of installments (X8), and interest rates (X9). The prediction accuracy of discriminant model 3 for consumer credit is 78.5% in the correct classification and 78.3% in the correct cross-validation group at the time of classification.
The only variables included in the internal credit model for this type of investment loan are the interest rate (X9). The prediction accuracy of the discriminant model 3 for the type of credit in investment is 65.6% in the correct classification, and 65.6% for the group is cross-validated correctly at the time of classification.
Of all the independent variables tested, it turned out that only the interest rate variable consistently became the only variable that had a relationship with credit status, both tested in general and specifically for each loan.
4.2.3 Logistic regression analysis
In logistic regression analysis, where the dependent variable uses a binary scale, the test will be carried out with two scales, namely 0 (not default) and 1 (default). This test will generally be carried out on three types of credit and, precisely, on each type of credit (working capital, consumption, and investment).
Table 4. General characteristics of the debtor (2) in BPR BKK Pekalongan regency
|
Variables |
N |
Minimum |
Maximum |
Mean |
Median |
Mode |
Std. Deviation |
|
|
Age |
years old |
6,727 |
20 |
82 |
44 |
44 |
42 |
9.7674 |
|
Number of collaterals |
unit |
6,727 |
0 |
16 |
0.97 |
1.00 |
1.00 |
0.6023 |
|
Beginning Ratio |
times |
6,727 |
0.00 |
88.00 |
3.47 |
2.27 |
0.00 |
4.8515 |
|
Ending Ratio |
times |
6,727 |
0.00 |
703.47 |
6.90 |
3.54 |
0.00 |
17.9930 |
|
Term of loans |
months |
6,727 |
5 |
276 |
41 |
36 |
36 |
28.0731 |
|
Interest Rate |
% |
6,727 |
6.00 |
21.00 |
11.21 |
11.00 |
11.00 |
1.6821 |
Table 5. Chi-square analysis
|
Independent Variables |
Label |
All Types of Loans |
Working Capital |
Consumption |
Investment |
||||
|
Chi Square Test |
P value |
Chi Square Test |
P value |
Chi Square Test |
P value |
Chi Square Test |
P value |
||
|
Credit Type |
X1 |
43.886 |
0.0000 |
||||||
|
Occupation |
X2 |
50.759 |
0.0000 |
18.274 |
0.0001 |
1.023 |
0.5996 |
0.683 |
0.7106a |
|
Age |
X3 |
7.246 |
0.0645a |
6.147 |
0.1047a |
8.647 |
0.0344 |
7.466 |
0.0584a |
|
Number of Collaterals |
X4 |
53.894 |
0.0000 |
9.472 |
0.0088 |
19.906 |
0.0000 |
0.000 |
1.000a |
|
Beginning Ratio |
X5 |
119.791 |
0.0000 |
25.221 |
0.0000 |
98.071 |
0.0000 |
0.000 |
1.000a |
|
Ending Ratio |
X6 |
94.327 |
0.0000 |
9.752 |
0.0208 |
87.716 |
0.0000 |
0.000 |
0.6097a |
|
Term of loans |
X7 |
70.568 |
0.0000 |
107.145 |
0.0000 |
12.300 |
0.0064 |
26.551 |
0.0000 |
|
Frequency |
X8 |
1.864 |
0.1722a |
1.424 |
0.2327a |
0.000 |
0.0000 |
0.000 |
1.000a |
|
Interest Rate |
X9 |
76.627 |
0.0000 |
315.564 |
0.0000 |
7.338 |
0.0255 |
35.171 |
0.0000 |
a Statistically insignificant for P<.05
The logistic regression method generates the internal model of default credit in Table 7 and is processed using the stepwise method. It can be seen in Table 7 that for all types of credit (3 types of credit), all independent variables are included in the internal model of default credit, except for the age variable (X3), the number of collaterals (X4), and the frequency of installments (X8), with % correct prediction of 91.70% (df=14; n=6.727).
For working capital loans, logistic analysis using the stepwise method shows that the independent variables that are included in the internal model of default credit are occupation variables (X2), beginning ratio (X5), ending ratio (X6), term of loans (X7), frequency of installments (X8) and interest rates (X9), with a % correct prediction of 90.90% (df=13; n=4,959).
For consumer loans, Logistics Analysis using the stepwise method (Table 7) shows that the independent variables included in the internal model of default credit are the beginning ratio variable (X5), term of loans (X7), and interest rate (X9), with % correct prediction of 95.50% (df=8; n=1.393).
Lastly, on investment loans, Logistics analysis using the stepwise method shows that the independent variable included in the internal model of default credit is only the interest rate variable (X9), with a % correct prediction of 90.10% (df=1; n=375).
4.2.4 Model for all types of loans
In general, with data on all types of credit (n=6,727), the nine independent variables that are thought to affect credit status, only two independent variables, namely term of loans (X7) and interest rate (X9), consistently have a significant role in the internal model of default credit (Table 8).
The term of loans has a positive effect on credit status; the longer the term of loans taken by the debtor, the riskier the credit status will be. Similarly, the interest rate also has a positive relationship with credit status, where the higher the interest rate determined by the bank, the riskier the credit status [16, 42].
Table 6. Discriminant analysis
|
Independent Variables |
Label |
Estimated Coefficients |
|||
|
All Types of Loans |
Working Capital |
Consumption |
Investment |
||
|
Credit type (1) |
X1(1) |
||||
|
Credit type (2) |
X1(2) |
1.199 |
|||
|
Occupation (1) |
X2(1) |
||||
|
Occupation (2) |
X2(2) |
0.555 |
0.572 |
||
|
Age |
X3 |
0.011 |
0.028 |
||
|
Number of collaterals |
X4 |
0.225 |
|||
|
Beginning Ratio |
X5 |
-0.021 |
0.210 |
||
|
Ending Ratio |
X6 |
0.084 |
|||
|
Term of loans |
X7 |
0.014 |
0.010 |
||
|
Frequency of installments |
X8 |
1.860 |
3.726 |
||
|
Interest |
X9 |
0.557 |
0.863 |
0.239 |
0.666 |
|
(Constant) |
-10.597 |
-13.556 |
-5.047 |
-8.314 |
|
|
Original grouped cases correctly classified |
72.40% |
80.90% |
78.50% |
65.60% |
|
|
Cross validated grouped cases correctly classified |
72.30% |
80.90% |
78.30% |
65.60% |
|
Table 7. Logistic regression analysis
|
Independent Variables |
Label |
Estimated Coefficients |
|||
|
All Types of Loans |
Working Capital |
Consumption |
Investment |
||
|
Credit types |
X1 |
||||
|
Credit type (1) |
X1(1) |
1.481 |
|||
|
Credit type (2) |
X1(2) |
2.067 |
|||
|
Occupation |
X2 |
||||
|
Occupation (1) |
X2(1) |
0.174 |
0.106 |
||
|
Occupation (2) |
X2(2) |
0.589 |
0.763 |
||
|
Beginning Ratio |
X5 |
||||
|
Beginning Ratio (1) |
X5(1) |
2.444 |
1.081 |
1.622 |
|
|
Beginning Ratio (2) |
X5(2) |
3.832 |
2.00 |
3.131 |
|
|
Beginning Ratio (3) |
X5(3) |
3.382 |
1.514 |
2.987 |
|
|
Ending Ratio |
X8 |
||||
|
Ending Ratio (1) |
X6(1) |
-0.734 |
-0.665 |
||
|
Ending Ratio (2) |
X6(2) |
-0.262 |
-0.223 |
||
|
Term of loans |
X7 |
||||
|
Term of loans (1) |
X7(1) |
0.994 |
0.764 |
0.159 |
|
|
Term of loans (2) |
X7(2) |
1.782 |
1.613 |
-1.946 |
|
|
Term of loans (3) |
X7(3) |
2.190 |
1.868 |
1.146 |
|
|
Frequency |
X8 |
-1.957 |
|||
|
Interest Rate |
X9 |
||||
|
Interest Rate (1) |
X9(1) |
15.987 |
17.078 |
16.664 |
2.643 |
|
Interest Rate (2) |
X9(2) |
18.530 |
21.070 |
17.709 |
|
|
Constant |
-25.645 |
-22.825 |
-22.313 |
-3.975 |
|
|
df |
14 |
13 |
8 |
1 |
|
|
Total Observation |
6,727 |
4,959 |
1,393 |
375 |
|
|
% Correct Prediction |
91.70% |
90.90% |
95.50% |
90.10% |
|
The two variables, term of loans and interest rate, were retested to obtain a more robust internal model of default credit based on parsimony. In Table 9, the default credit model is obtained with less df (df=5) than before df=14 (Table 5), with a % correct prediction of 91.0%, not much different from the previous 91.7% (Table 8).
4.2.5 Model for working capital loans
Furthermore, the internal model of default credit was made more specific, namely based on the type of use: working capital loans (n=4.959), consumption loans (n=1.393), and investment loans (n=375). In working capital loans (n=4,959 customers), the independent variables, beginning ratio (X5) and interest rates (X9) are two variables that affect the status of credit in Chi-Square analysis, discriminant analysis using the stepwise method, and logistic regression with stepwise method (Table 10).
The beginning ratio, where the collateral value is compared to the ceiling value, shows a negative relationship, where the higher the beginning ratio, the less risky the credit status is. Collateral should be based on the potential creditworthiness of the borrower [43] so that the beginning ratio can be used as early detection for potential debtors who apply for working capital loans.
These two variables were re-examined to obtain a more robust internal model of default credit based on parsimony. In Table 11, the internal model of default credit has less df (df=5) than before df=13 (Table 7), with a % correct prediction of 90.2%, not much different from the previous 90.9% (Table 10).
Table 8. All types of loans in chi-square, discriminant, and logistic regression model
|
Independent Variables |
Label |
Chi-Square |
Discriminant |
Logistic |
|
Credit Types |
V |
V |
||
|
Credit type (1) |
X1(1) |
|||
|
Credit type (2) |
X1(2) |
V |
||
|
Occupation |
V |
V |
||
|
Occupation (1) |
X2(1) |
|||
|
Occupation (2) |
X2(2) |
V |
||
|
Age |
X3 |
V |
||
|
Number of collaterals |
X4 |
V |
V |
|
|
Beginning Ratio |
X5 |
V |
V |
|
|
Ending Ratio |
X6 |
V |
V |
|
|
Term of loans |
X7 |
V |
V |
V |
|
Frequency of installments |
X8 |
V |
||
|
Interest Rate |
X9 |
V |
V |
V |
|
Original grouped cases correctly classified |
72.40% |
|||
|
Cross validated grouped cases correctly classified |
72.30% |
|||
|
% Correct prediction |
91.70% |
|||
Table 9. The internal model of default credit for all type loan
|
Independent Variables |
Label |
Estimated Coefficient |
df |
p-value |
|
Term of loans |
X7 |
3 |
.000 |
|
|
Term of loans (1) |
X7(1) |
0.3791 |
||
|
Term of loans (2) |
X7(2) |
1.1019 |
||
|
Term of loans (3) |
X7(3) |
0.9699 |
||
|
Interest Rate |
X9 |
2 |
.000 |
|
|
Interest Rate (1) |
X9(1) |
18.5883 |
||
|
Interest Rate (2) |
X9(2) |
19.2470 |
||
|
Constant |
-21.8427 |
|||
|
Degree of freedom |
5 |
|||
|
Total Observation |
6727 |
|||
|
% Correct Prediction |
91.0% |
Table 10. Working capital loans in chi-square, discriminant, and logistic regression model
|
Independent Variables |
Label |
Chi-Square |
Discriminant |
Logistic |
|
Occupation |
V |
V |
||
|
Occupation (1) |
X2(1) |
|||
|
Occupation (2) |
X2(2) |
V |
||
|
Age |
X3 |
|||
|
Number of collaterals |
X4 |
V |
||
|
Beginning Ratio |
X5 |
V |
V |
V |
|
Ending Ratio |
X6 |
V |
V |
|
|
Term of loans |
X7 |
V |
V |
|
|
Frequency of Instalments |
X8 |
V |
V |
|
|
Interest Rate |
X9 |
V |
V |
V |
|
Original grouped cases correctly classified |
80.90% |
|||
|
Cross validated grouped cases correctly classified |
80.90% |
|||
|
% Correct prediction |
90.90% |
|||
4.2.6 Model for consumption loans
In consumer loans (n=1,393 customers), the independent variables beginning ratio (X5), term of loans (X7), and interest rates (X9) are three variables that affect the status of the type of credit, both in Chi-Square analysis, discriminant analysis with the stepwise method and logistic regression with stepwise method (Table 12).
These three variables were re-examined to obtain a more robust internal model of default credit based on parsimony. In Table 13, the default credit model is obtained with the same amount of df (df=8) and the same % correct prediction, which is 95.5%. Therefore, the default credit before (Table 7) and after (Table 13) models can be used.
4.2.7 Model for investment loans
While for investment loans (n=375 customers), the independent variable credit interest rate (X9) is the only variable that affects the status of the type of credit, both in Chi-Square analysis, discriminant analysis using the stepwise method, and logistic regression using the stepwise method (Table 14).
Since only one independent variable affects the dependent variable, credit status, the resulting model can be used as internal capital for default loans, with a % correct prediction of 90.10% (Table 7).
Table 11. The internal model of default credit for working capital loans
|
Independent Variables |
Label |
Estimated Coefficient |
df |
p-value |
|
Beginning Ratio |
X5 |
3 |
.003 |
|
|
Beginning Ratio (1) |
X5(1) |
-1.0903 |
||
|
Beginning Ratio (2) |
X5(2) |
-0.4506 |
||
|
Beginning Ratio (3) |
X5(3) |
0.3294 |
||
|
Interest Rate |
X9 |
2 |
.000 |
|
|
Interest Rate (1) |
X9(1) |
18.1970 |
||
|
Interest Rate (2) |
X9(2) |
20.7186 |
||
|
Constant |
-20.5676 |
|||
|
Degree of freedom |
5 |
|||
|
Total Observation |
4959 |
|||
|
% Correct Prediction |
90.2% |
Table 12. Consumption loans in chi-square, discriminant, and logistic regression model
|
Independent Variables |
Label |
Chi-Square |
Discriminant |
Logistic |
|
Occupation |
||||
|
Occupation (1) |
X2(1) |
|||
|
Occupation (2) |
X2(2) |
|||
|
Age |
X3 |
V |
V |
|
|
Number of collaterals |
X4 |
V |
||
|
Beginning Ratio |
X5 |
V |
V |
V |
|
Ending Ratio |
X6 |
V |
V |
|
|
Term of loans |
X7 |
V |
V |
V |
|
Frequency of Instalments |
X8 |
|||
|
Interest Rate |
X9 |
V |
V |
V |
|
Original grouped cases correctly classified |
78.50% |
|||
|
Cross validated grouped cases correctly classified |
78.30% |
|||
|
% Correct prediction |
95.50% |
|||
Table 13. The internal model of default credit for consumption loans
|
Independent Variables |
Label |
Estimated Coefficient |
df |
p-value |
|
Beginning Ratio |
X7 |
3 |
.000 |
|
|
Beginning Ratio (1) |
X7(1) |
-2.9873 |
||
|
Beginning Ratio (2) |
X7(2) |
-1.3648 |
||
|
Beginning Ratio (3) |
X7(3) |
0.1439 |
||
|
Term of loans |
X7 |
3 |
.008 |
|
|
Term of loans (1) |
X7(1) |
0.1587 |
||
|
Term of loans (2) |
X7(2) |
-1.9462 |
||
|
Term of loans (3) |
X7(3) |
1.1458 |
||
|
Interest Rate |
X9 |
2 |
.006 |
|
|
Interest Rate (1) |
X9(1) |
16.6644 |
||
|
Interest Rate (2) |
X9(2) |
17.7085 |
||
|
Constant |
-19.3254 |
|||
|
Degree of freedom |
8 |
|||
|
Total Observation |
1393 |
|||
|
% Correct Prediction |
95.5% |
4.3 Qualitative findings
The use of explanatory sequential design, a mixed research method of QUAN à qual based on the results of quantitative analysis from the three analyses above, produces the following information. It was found that the independent variables of the term of loans and interest rate are variables that consistently have a significant role on the credit status of the debtor and therefore are essential variables for the development of the default credit model. Loans disbursed by rural banks have proven to positively impact micro and small enterprises (SME – Small Medium Enterprise) [44]. Therefore, rural banks need to pay more attention to the interest rates provided. This is due to the positive relationship between interest rates and credit status; The higher the interest rate, the higher the chance the loan will become the default. From the in-depth interviews with several debtors who were used as key informants, credit with low-interest rates has affected its business development. With the higher interest rates charged by rural banks, the burden on debtors increases, making the number of profits they get eroded. The interest expense on loans is a concern for debtors, especially the passion for running their business. Some of the main inputs that can be considered for lowering credit interest rates are suitable alternatives that can reduce the number of interest rates but still ensure the security of rural bank performance as offered by non-rural bank financial institutions. These results imply the need to expand the cooperation scheme between the people's business credit program and rural banks [45].
The finding in-depth interviews with the managers of PT. BPR BKK Pekalongan Regency, the high-interest rate charged by rural banks to debtors is also associated with asymmetric information related to the ability of prospective debtors to pay when they apply for loans. Although rural banks have tried to dig up valuable and unique information about the financial health of their prospective debtors when prospective debtors apply for loans; however, much of this information cannot be easily verified and generally only gradually comes to light with the existence of a long-term loan relationship. The inability of the rural bank to dig up the complete data creates information asymmetry, which is generally compensated by the imposition of higher credit interest rates. It is revealed that the lender-borrower relationship plays an important role in minimizing asymmetric information and impacts determining credit interest rates and debtor credit status [46]. BPR BKK Pekalongan Regency created the Penta Planner program, where five parties support each other to create justice for the people's economy. These five parties are farmers (or parties who need working capital credit), rural banks (credit providers), credit guarantee institutions (covering collateral that cannot be provided by debtors and providing guarantees in case of default), technical assistants (increasing the productivity of farmers and ensuring the quality of the harvest according to standards) and off-takers (buyers of farmers' crops at a mutually agreed price). This Penta Planner idea existed in early 2019 but could only be realized in the second half of 2021, starting with porang farmers. Pouring is a type of tuber plant usually processed into rice, shirataki, a mixture of cake products to cosmetics [47]. In addition, in 2018 alone, pouring exports could reach 254 tons, sent to Japan, Vietnam, China, Australia, and other countries. However, unfortunately, the selling price of porang at the farmer level is only Rp. 2,500 per 4 kg of porang (equivalent to \$0.04 per kg at an exchange rate of IDR 15,000/US \$1) [48]. With the low selling price of porang and limited working capital, the porang farmers cannot live in prosperity. For this reason, BPR BKK Pekalongan Regency believes that the Penta Planner program can improve the welfare of porang farmers and other farmers using loans without collateral, low-interest rates, and leading to low NPLs for rural banks. For BPR BKK Pekalongan debtors who have not been included in the Penta Planner program, the government continues to provide support, such as the banking credit structuring policy, for debtors affected by the COVID-19 pandemic [49].
Table 14. Investment loans in chi-square, discriminant, and logistic regression model
|
Independent Variables |
Label |
Chi-Square |
Discriminant |
Logistic |
|
Occupation |
||||
|
Occupation (1) |
X2(1) |
|||
|
Occupation (2) |
X2(2) |
|||
|
Age |
X3 |
|||
|
Number of collaterals |
X4 |
|||
|
Beginning Ratio |
X5 |
|||
|
Ending Ratio |
X6 |
|||
|
Term of loans |
X7 |
V |
||
|
Frequency of Instalments |
X8 |
|||
|
Interest Rate |
X9 |
V |
V |
V |
|
Original grouped cases correctly classified |
65.60% |
|||
|
Cross validated grouped cases correctly classified |
65.60% |
|||
|
% Correct prediction |
90.10% |
|||
From the integration of Chi-Square Analysis, Discriminant Analysis, and Logistics Regression analysis, it was found that only the interest rate variable affected all types of credit. By using nine variables that are suspected of influencing debtor failure, the results of the integration of Chi-Square Analysis, Discriminant Analysis, and Logistics Regression analysis at BPR BKK Pekalongan Regency in 2021 found that in Working Capital loans, the variables that affect the status of default loans are credit interest rates and beginning ratio. Meanwhile, in consumer credit, the variables that affect the status of default loans are credit interest rates, beginning ratios, and term of loans. In investment loans, only the interest rate variable affects the status of default loans at rural banks.
Explanatory sequential design findings reveal that credit interest rates positively influence default credit status, which means that the higher the credit interest rate borne by the debtor, the higher the chance that their credit will fail. With the higher interest rates charged by rural banks, the burden on debtors increases, making the number of profits they get eroded, and the debtor's ability to pay for the credit they get from rural banks decreases.
From the management of the rural bank, the determination of the credit interest rate is related to the compensation for the risks faced by the rural bank in connection with the asymmetric information of the debtor's ability to pay while also paying attention to the interest rate set by the IDIC.
The small sample size for this study, which depends only on one rural bank, may be a weakness of its application. Therefore, the generalizability of the findings may be limited due to the small sample size. On the other hand, large sample sizes may be used in future studies. Furthermore, the addition of independent variables, such as gender, education level, marital status, number of family members, monthly income, installment amount, etc., can provide a better picture of potential debtor failure in Indonesia. The Penta Planner program, which is run at BPR BKK Pekalongan Regency, can be investigated further to determine the parties' benefits in it.
Universitas Kristen Maranatha research grant supports this work. The authors also acknowledge Universitas Prasetiya Mulya and BPR BKK Pekalongan Regency for supporting this work.
Appendix 1. Chi-Square
|
INDEPENDENT VARIABLES |
CREDIT STATUS |
TOTAL KREDIT |
CHI-SQUARE TEST |
P value |
||||
|
NOT DF = 6,124 |
DEFAULT = 603 |
n = |
6,727 |
|||||
|
CREDIT TYPES |
43.886 |
0.0000 |
||||||
|
Working Capital Loan |
1,331 |
21.7% |
62 |
1.0% |
1,393 |
20.7% |
||
|
Consumption Loan |
338 |
5.5% |
37 |
0.6% |
375 |
5.6% |
||
|
Investment Loan |
4,455 |
72.7% |
504 |
8.2% |
4,959 |
73.7% |
||
|
OCCUPATIONS |
50.759 |
0.0000 |
||||||
|
Employee |
1,652 |
27.0% |
98 |
1.6% |
1,750 |
26.0% |
||
|
Self-employment |
3,817 |
62.3% |
464 |
7.6% |
4,281 |
63.6% |
||
|
Unemployment |
655 |
10.7% |
41 |
0.7% |
696 |
10.3% |
||
|
AGE |
7.246 |
0.0645a |
||||||
|
<30 years old |
539 |
8.8% |
41 |
0.7% |
580 |
8.6% |
||
|
30-39 years old |
1,610 |
26.3% |
141 |
2.3% |
1,751 |
26.0% |
||
|
40-49 years old |
2,170 |
35.4% |
220 |
3.6% |
2,390 |
35.5% |
||
|
>49 years old |
1,805 |
29.5% |
201 |
3.3% |
2,006 |
29.8% |
||
|
NUMBER OF COLLATERALS |
53.894 |
0.0000 |
||||||
|
None |
810 |
13.2% |
18 |
0.3% |
828 |
12.3% |
||
|
Enough |
4,936 |
80.6% |
548 |
8.9% |
5,484 |
81.5% |
||
|
High |
378 |
6.2% |
37 |
0.6% |
415 |
6.2% |
||
|
BEGINNING RATIO |
119.791 |
0.0000 |
||||||
|
Nol |
1,156 |
18.9% |
19 |
0.3% |
1,175 |
17.5% |
||
|
Less |
208 |
3.4% |
14 |
0.2% |
222 |
3.3% |
||
|
Enough |
1,221 |
19.9% |
193 |
3.2% |
1,414 |
21.0% |
||
|
Good |
3,539 |
57.8% |
377 |
6.2% |
3,916 |
58.2% |
||
|
ENDING RATIO |
94.327 |
0.0000 |
||||||
|
Nol |
1,156 |
18.9% |
19 |
0.3% |
1,175 |
17.5% |
||
|
Less |
413 |
6.7% |
48 |
0.8% |
461 |
6.9% |
||
|
Enough |
3,642 |
59.5% |
425 |
6.9% |
4,067 |
60.5% |
||
|
Good |
913 |
14.9% |
111 |
1.8% |
1,024 |
15.2% |
||
|
TERM OF LOANS |
70.568 |
0.0000 |
||||||
|
5-12 months |
509 |
8.3% |
20 |
0.3% |
529 |
7.9% |
||
|
13-24 months |
1,599 |
26.1% |
87 |
1.4% |
1,686 |
25.1% |
||
|
25-36 months |
2,223 |
36.3% |
267 |
4.4% |
2,490 |
37.0% |
||
|
>36 months |
1,793 |
29.3% |
229 |
3.7% |
2,022 |
30.1% |
||
|
FREQUENCY OF INSTALMENTS |
1.864 |
0.1722a |
||||||
|
Monthly |
6,072 |
99.2% |
594 |
9.7% |
6,666 |
99.1% |
||
|
Otherwise |
52 |
0.8% |
9 |
0.1% |
61 |
0.9% |
||
|
INTEREST RATE |
76.627 |
0.0000 |
||||||
|
Low |
205 |
3.3% |
- |
0.0% |
205 |
3.0% |
||
|
Enough |
5,181 |
84.6% |
463 |
7.6% |
5,644 |
83.9% |
||
|
High |
738 |
12.1% |
140 |
2.3% |
878 |
13.1% |
||
|
Data are presented as number (percentage) |
||||||||
|
a Statistically insignificant for P < .05 |
||||||||
[1] Chosyali, A., Sartono, T. (2019). Optimization improving credit quality in order to overcome problem credit. Law Reform, 15(1): 98-112. https://doi.org/10.14710/lr.v15i1.23357
[2] Morina, F., Özen, E. (2020). Does the commercial bank's loans affect economic growth? Empirical evidence for the real sector economy in Kosovo (2005-2018). Planning, 15(8): 1205-1222. https://doi.org/10.18280/ijsdp.150807
[3] Keuangan, O.J. (2021). Booklet Perbankan Indonesia 2021. Jakarta.
[4] McNulty, J.E., Murdock, M., Richie, N. (2013). Are commercial bank lending propensities useful in understanding small firm finance? Journal of Economics and Finance, 37(4): 511-527. https://doi.org/10.1007/s12197-011-9191-x
[5] Pratiwi, P.Y. (2021). The impact of joint liability group lending on lowering the risk of farmer and agriculture crowdfunding in Indonesia. International Journal of Rural Management, 09730052211049595. https://doi.org/10.1177/09730052211049595
[6] Hoffmann, V., Rao, V., Surendra, V., Datta, U. (2021). Relief from usury: Impact of a self-help group lending program in rural India. Journal of Development Economics, 148: 102567. https://doi.org/10.1016/j.jdeveco.2020.102567
[7] Chou, T.K., Buchdadi, A.D. (2016). Bank performance and its underlying factors: A study of rural banks in Indonesia. Accounting and Finance Research, 5(3): 55-63. https://doi.org/10.5430/afr.v5n3p55
[8] Afriyeni, A., Fernos, J. (2018). Analysis of the determinants of the profitability performance of conventional rural banks (BPR) in West Sumatra. Jurnal Benefita, 3(3): 325-335. https://doi.org/10.22216/jbe.v3i3.3623
[9] Setia Murningsih, M.F., Purwanto, B. (2020). Factors influencing Indonesian rural banks’ credit disbursement. Jurnal Keuangan dan Perbankan, 24(2): 189-201. https://doi.org/10.26905/jkdp.v24i2.3778
[10] Sofyan, M. (2021). The performance of BPR and BPRS at the time the COVID-19 pandemic. 2nd Semin. Nas. ADPI Mengabdi Untuk Negeri Pengabdi. Masy. di Era New Norm., 2(2): 6-12.
[11] Siamat, D. (2003). Management of Commercial Banks.
[12] Kasmir, Banks and Other Financial Institutions. Jakarta: Jakarta: Rajawali Pers, 2018.
[13] Jayadev, M. (2006). Internal credit rating practices of Indian banks. Economic and Political Weekly, 41(11): 1069-1078.
[14] Maryandi, M.S., Yaya, R., Supriyono, E. (2015). Analysis of the influence of bank internal factors on NPL and application of the loan loss ratio-based model. JBTI: Jurnal Bisnis: Teori dan Implementasi, 6(2): 244-259. https://doi.org/10.18196/jbti.v6i2.2530
[15] Ikatan Bankir Indonesia, Banking Credit Business. Jakarta: Gramedia Pustaka Utama, 2018.
[16] Saputro, A.R., Sarumpaet, S., Prasetyo, TJ. (2019). Analysis of the effect of credit growth, types of credit, bank loan interest rates and inflation on credit problem. Ekspansi: Jurnal Ekonomi, Keuangan, Perbankan, dan Akuntansi, 11(1): 1-12. https://doi.org/10.35313/ekspansi.v11i1.1325
[17] Sharpe, S.A. (1990). Asymmetric information, bank lending, and implicit contracts: A stylized model of customer relationships. The Journal of Finance, 45(4): 1069-1087. https://doi.org/10.1111/j.1540-6261.1990.tb02427.x
[18] Manurung, E.T., Manurung, E.M. (2019). A new approach of bank credit assessment for SMEs. Academy of Accounting and Financial Studies Journal, 23(3): 1-13.
[19] Ghulam, Y., Dhruva, K., Naseem, S., Hill, S. (2018). The interaction of borrower and loan characteristics in predicting risks of subprime automobile loans. Risks, 6(3): 101. https://doi.org/10.3390/risks6030101
[20] Hasibuan, M.S.P. (2017). Fundamentals of Banking, 11th ed. Jakarta: Jakarta: Bumi Aksara.
[21] Waweru, N., Kalani, V.M. (2008). Commercial banking crises in Kenya: Causes and remedies. Global Journal of Finance and Banking Issues, 3(3): 23-43.
[22] Petersen, M.A., Rajan, R.G. (1994). The benefits of lending relationships: Evidence from small business data. The Journal of Finance, 49(1): 3-37. https://doi.org/10.1111/j.1540-6261.1994.tb04418.x
[23] Berger, A.N., Udell, G.F. (1995). Relationship lending and lines of credit in small firm finance. Journal of Business, 68(3): 351-381.
[24] Chen, Y., Huang, R.J., Tsai, J., Tzeng, L.Y. (2015). Soft information and small business lending. Journal of Financial Services Research, 47(1): 115-133. https://doi.org/10.1007/s10693-013-0187-x
[25] Hill, T.L., Scott, J. (2015). Knows me and my business: the association between preference for relational governance and owners' choice of banks. Journal of Small Business Management, 53: 174-192. https://doi.org/10.1111/jsbm.12193
[26] Berger, A.N., Goulding, W., Rice, T. (2014). Do small businesses still prefer community banks? Journal of Banking & Finance, 44: 264-278. https://doi.org/10.1016/j.jbankfin.2014.03.016
[27] Elsas, R., Krahnen, J.P. (1998). Is relationship lending special? Evidence from credit-file data in Germany. Journal of Banking & Finance, 22(10-11): 1283-1316. https://doi.org/10.1016/S0378-4266(98)00063-6
[28] Fiordelisi, F., Monferrà, S., Sampagnaro, G. (2014). Relationship lending and credit quality. Journal of Financial Services Research, 46(3): 295-315. https://doi.org/10.1007/s10693-013-0176-0
[29] Gadanecz, B., Kara, A., Molyneux, P. (2012). Asymmetric information among lending syndicate members and the value of repeat lending. Journal of International Financial Markets, Institutions and Money, 22(4): 913-935. https://doi.org/10.1016/j.intfin.2012.04.007
[30] Bilau, J., St-Pierre, J. (2018). Microcredit repayment in a European context: Evidence from Portugal. The Quarterly Review of Economics and Finance, 68: 85-96. https://doi.org/10.1016/j.qref.2017.11.002
[31] Baba, H., Abdallah, A.H., Hudu, Z. (2015). Factors influencing agricultural credit demand in Northern Ghana. African Journal of Agricultural Research, 10(7): 645-652. https://doi.org/10.5897/AJAR2014
[32] Kosgey, Y.K.K. (2013). Agricultural credit access by grain growers in Uasin-Gishu County, Kenya. IOSR Journal of Economics and Finance, 2(3): 36-52. https://doi.org/10.9790/5933-0233652
[33] Ugwumba, C.O.A., Omojola, J.T. (2013). Credit access and productivity growth among subsistence food crop farmers in Ikole Local Government Area of Ekiti State, Nigeria. Journal of Agricultural and Biological Science, 8(4): 351-356.
[34] Duniya, K.P., Adinah, I.I. (2015). Probit analysis of cotton farmers' accessibility to credit in Northern Guinea Savannah of Nigeria. Asian Journal of Agricultural Extension, Economics and Sociology, 4(4): 296-301. https://doi.org/10.9734/ajaees/2015/13538
[35] Hussain, A., Thapa, G.B. (2012). Smallholders’ access to agricultural credit in Pakistan. Food Security, 4(1): 73-85. https://doi.org/10.1007/s12571-012-0167-2
[36] Saqib, L., Farooq, M.A., Zafar, A.M. (2016). Customer perception regarding Sharī ‘ah compliance of Islamic banking sector of Pakistan. Journal of Islamic Accounting and Business Research, 7(4): 282-303. https://doi.org/10.1108/JIABR-08-2013-0031
[37] Khoi, P.D., Gan, C., Nartea, G.V., Cohen, D.A. (2013). Formal and informal rural credit in the Mekong River Delta of Vietnam: Interaction and accessibility. Journal of Asian Economics, 26: 1-13. https://doi.org/10.1016/j.asieco.2013.02.003
[38] Ahmad, N. (2011). Impact of institutional credit on agricultural output. Theoretical and Applied Economics, 18: 10-563.
[39] Sugiarto, S. (2022). Business Research Methodology, Edisi 2. Yogyakarta: Penerbit Andi.
[40] Gani, I., Amalia, S. (2018). Data analysis tools: Statistical applications for economic and social research, Edisi Revi. Penerbit Andi.
[41] Sugiarto, S., Setio, H. (2021). Applied Statistics for Business and Economics, 1st ed. Yogyakarta: Penerbit Andi.
[42] Salifu, A.T., Tofik-Abu, Z., Rahman, M.A., Sualihu, M.A. (2018). Determinants of loan repayment performance of small and medium enterprises (SMEs) in Ghana: The case of Asante Akyem Rural Bank. Journal of African Business, 19(2): 279-296. https://doi.org/10.1080/15228916.2018.1440460
[43] Chaudhary, M.A., Ishfaq, M. (2003). Credit worthiness of rural borrowers of Pakistan. The Journal of Socio-Economics, 32(6): 675-684. https://doi.org/10.1016/j.socec.2003.10.005
[44] Anwar, M., Nidar, S.R., Komara, R., Layyinaturrobaniyah, L. (2019). Rural bank efficiency and loans for micro and small businesses: Evidence from West Java Indonesia. International Journal of Emerging Markets, 15(3): 587-610. https://doi.org/10.1108/IJOEM-11-2017-0494
[45] Rudi Purwono, D., Nugroho, R.Y.Y., Mubin, M.K. (2019). Response on new credit program in Indonesia: An asymmetric information perspective. Journal of Asian Finance Economics and Business, 6(2): 33-44. https://doi.org/10.13106/jafeb.2019.vol6.no2.33
[46] Kong, R., Turvey, C., Xu, X., Liu, F. (2014). Borrower attitudes, lender attitudes and agricultural lending in rural China. International Journal of Bank Marketing, 32(2): 104-129. https://doi.org/10.1108/IJBM-08-2013-0087
[47] Azanella, L.A., Aida, N.R. (2021). A What is a plant of Porang and the benefits of it? Kompas.com. https://www.kompas.com/tren/read/2021/08/20/114600165/apa-itu-tanaman-porang-dan-apa-manfaatnya-?page=all.
[48] Tim Redaksi CNBC, Get to knon Porang plants: The type, the benefit, the price, and the cultivation, cnbcindonesia.com, accessed on October 12, 2021.
[49] Sytra Disemandi, H., Ismail Shaleh, A. (2020). Banking credit restructuring policy on the impact of COVID-19 spread in Indonesia. Journal Inovasi Ekonomi, 5(2): 63-70. http://ejournal.umm.ac.id/index.php/jiko.