The accuracy of the prediction is also shown in Table 4.10 Classification Table, which shows that in 45 cases predicted to be unable to pay the debt, the model correctly predicted 41 cases, so the accuracy rate is 91.1%. As for the 46 cases that actually paid the debt, the model incorrectly predicted 3 cases (that is, they did not pay), the accuracy rate is now 93.5%. From there, we can calculate the correct prediction rate of the entire model to be 92.3%.
4.3.3. General comments and model selection:
Both models have a good overall fit (Sig.OB = 0.00). The forecast results of both models are also consistent with the observed data (Sig.HL of both models > 0.05).
- The accuracy of the forecast results of both models is very high, more than
90%.
- Through the value of “Cox & Snell R-squared” in the results obtained from
The logistic regression estimates of the models show that the overall logistic model and the restricted logistic model explain 66.9% and 71.3% of the variation in the probability of customers repaying their debts, respectively.
Therefore, the limited logistic regression model has a higher suitability in reflecting the current situation of personal credit lending at VSB Dong Nai. The model is estimated as follows:
Log e [
] = -52.333 + 0.169 age + 0.954 career + 0.386 time
+0.116 tglamcv + 1.197Nha_o + 0.335Tncanhan– 0.088Du_no + 0.803DV
+1,640Payment
In summary, with the logistic model, to clearly distinguish whether a customer is able to repay the debt or not, the indicators to consider are:
* Working time ( = 0.386) means that under the condition that other factors remain unchanged, if working time increases by 1 unit, the customer's probability of paying debt increases by 1.47 times compared to the group with lower working time (That is, Log e (khanangtrano) = 0.386 Khanangtrano = 1.47 )
* Current job time ( =0.116) means under the condition
Other factors remaining constant, if work time increases by 1 unit, the customer's probability of paying the debt increases by 1.12 times.
* The indicator of having a house or not ( =1.197) means that under the conditions of
Other factors remaining constant, if the customer group is highly rated as living in their own home, the probability of repaying the debt increases by 3.30 times compared to the group living in rented houses.
* Personal income ( =0.335) means that under the condition that other factors remain constant, the group of customers with highly rated personal income such as income greater than 120 million has a debt repayment probability 1.39 times higher than the group with income from 36 - 120 million;
* Customers who use banking services ( = 0.803) means that, assuming other factors remain constant, if the group of customers who use banking services are highly rated, such as using savings and card services, their ability to repay debt is 2.232 times higher than groups that are rated lower, such as those who only use savings services.
* Customers with deposits at the bank ( = 1.640) means that under the condition that other factors remain unchanged, the group of customers with highly rated bank deposits such as deposits greater than 500 million will have an increased ability to repay debt.
5.15 times higher than the lower rated group such as the group that only deposited 100 – 500 million.
4.4. Synthesis of factors affecting the debt repayment ability of individual customers at VSB Dong Nai from the experimental model
The results of the logistic linear regression test showed that 9 factors out of 15 factors included in the initial model had a strong impact on the model and met the author's construction purpose. The 9 factors include: age, occupation, working time, current job time, housing, personal income, outstanding debt, use of banking services, and deposit accounts.
According to the results from the model, we see that the savings deposit variable at the bank and the housing variable have the strongest impact on the customer's debt repayment probability, consistent with practical theory and the current situation at VSB Dong Nai.
According to the selected model given above, we have a summary table.
Factors affecting customers' ability to repay debts are as follows:
Factors
Impact on dependent variable belong(Khanangtrano) | |
Age | + |
Industry | + |
time | + |
Tglamcv | + |
House | + |
Tncanhan | + |
Du_no | - |
DV | + |
Instructions | + |
Maybe you are interested!
-
Factors affecting the debt repayment ability of corporate customers at Vietnam Joint Stock Commercial Bank for Investment and Development - Long An Branch - 1 -
Factors affecting the choice of bank for deposit transactions of individual customers - 1 -
Expectation Factors Affecting Individual Customers' Decision to Choose a Savings Bank -
Factors affecting the acceptance of using e-banking services by individual customers at An Binh Commercial Joint Stock Bank, Dong Nai branch - 10 -
Research on factors affecting the decision to choose savings deposit service of individual customers at Vietnam Joint Stock Commercial Bank for Investment and Development - Thua Thien Hue branch - 17
Table 4.11: Factors affecting the debt repayment ability of individual customers at VSB Dong Nai
The Age variable has a (+ ) impact on the customer's ability to repay the debt. This means that the older the customer is, the higher the probability that the customer will be able to repay the debt.
The variable Occupation (nghenghiep) has a positive impact on the variable ability to pay.
customer debt. We can see that the direction of the variable's impact is quite reasonable.
It is clear that when a customer has a job, the customer's ability to repay the debt is higher.
The variable Working time (thoigianct) also has a positive impact on the debt repayment ability variable. This means that when customers have seniority in their work, the bank will more easily make a decision to grant credit. The same is true for the variable Working time (thoigiancv) .
The Housing variable ( Nha_o ) and the Personal Income variable ( Tncanhan ) also have a positive impact on the customer's debt repayment ability variable. This is completely reasonable with its practicality and the results drawn from the model.
The variable Outstanding Debt ( Du_no ) has a negative impact on the customer's ability to repay debt. The analysis results for this variable are also completely reasonable because customers who still have debts do not easily want the bank to highly evaluate their ability to repay debt.
The variable of using other services ( DV ) and the variable of Deposit balance ( Tiengui ) have a positive impact on the ability to repay debts of customers. The bank will give priority to granting credit and appreciate these loans when customers also use the bank's services and have deposit accounts. This means that customers have financial resources without difficulties and the credits can be recovered more.
However, according to the results from the model, we see that the impact of independent variables on dependent variables is completely opposite to the research results of Kleimeier and Thanh (2006). The reason for this big difference is due to the difference in research samples, research time and research subjects of the two models.
Comparison table of the differences between the two models:
Kleimeier and Thanh (2006) | Author | |
Research sample | Retail banks Vietnam | 91 individual customers at VSB Dong Nai |
Research time | Around 2006 | 2011-2012 |
Research object | Vietnam Bank | VSB Dong Nai |
With such a large difference, the research results of Kleimeier and Thanh (2006) cannot be suitable for the current situation and circumstances of VSB Dong Nai. Therefore, the results of the model as studied above assessing the debt repayment probability of individual customers at VSB Dong Nai are relatively suitable for the current situation of VSB Dong Nai.
CHAPTER V: CONCLUSION
Credit risk is no longer a new issue but is always of constant concern due to its significant impact on the Bank's business operations. Therefore, building a model to assess the customer's ability to repay debt is extremely urgent for banks in general and Viet Thai Joint Venture Bank - Dong Nai branch in particular. Through the system of scoring and rating individual customer credit and the logistic method, it is partly shown the influence of the above factors on the customers' ability to repay debt.
Through the research process, the thesis has obtained the following results:
- Summary of research results in the world on the influence of
Factors affecting the customer's ability to repay.
- Identify factors affecting the debt repayment probability of individual customers at Viet Thai Joint Venture Bank - Dong Nai branch including 15 factors: Age, Education level, Occupation, Working time, Current working time, Housing status, Family structure, Number of dependents, Annual personal income, Annual family income, Debt repayment status with VSB, Late interest payment status, Current total outstanding debt in VND, Use of other services of VSB and Average savings deposit balance.
- Build a logistic regression model of factors affecting
Debt repayment probability of individual customers at Viet Thai Joint Venture Bank
– Dong Nai branch.
- Assess the level of influence of each factor on the debt repayment probability of individual customers at Viet Thai Joint Venture Bank - Dong Nai branch.
However, the thesis still has certain limitations:
- Limited research subjects (the total sample consists of only 91 individual customers borrowing capital at VSB Dong Nai)
- The model only evaluates based on credit scoring techniques, not customer behavior scoring techniques. Behavioral assessment is also very important because it reflects the way, attitude, honesty as well as cooperation in repaying bank debts.
- The thesis mainly relies on information from the Credit Information Center (CIC), while this information source is not updated regularly.
Further research directions:
- Expanding research subjects
- Further study of behavioral factors affecting debt repayment ability
- In addition to information from CIC, information from other credit institutions can be consulted so that customers' financial information is more complete and clear.
Some recommendations from the model study:
To accurately and transparently assess the debt repayment ability of individual customers at VSB Dong Nai, the thesis proposes the following solutions:
- Improve the appraisal level of CBTD, especially the appraisal of customer's character because this has a great influence on the customer's willingness to repay the loan.
- To serve the assessment work, CBTD must regularly update information on technical economics, forecast information on industry development, market prices, etc.
- Improve the efficiency of collecting and using information in credit activities. Full and accurate information about customers and the market plays a very important role in ensuring loan quality and limiting risks. This helps CBTD grasp information about businesses more accurately, assess the production and business situation of borrowing businesses and have a basis for re-evaluating customers' collateral. In addition to CIC information, information can be collected from the following sources:
+ From customer partners
+ From banks with which customers have relationships. Enhance cooperation
cooperation between banks in information sharing.
+ From customer management agency
- To ensure that customers' repayment probability is highly appreciated, VSB Dong Nai needs to strengthen supervision of loan use, to avoid cases where customers use loan capital for the wrong purpose, do not repay the debt but use it for other purposes, and when the debt is due, they cannot repay. In particular, it is necessary to:
+ Regular and surprise checks
+ Change the inspection content, not only checking the loan purpose but also other factors such as collateral, legality, reputation...