Factors Affecting The Likelihood Of Loan Delinquency

The Debt Ratio

Question:

Write an essay on “ACME Bank”.
 

The importance of determining the influential variables in identifying the likelihood to customers for forfeiting loan and become a delinquency cannot be overlooked. The primary factors are selected for viewing a credit scoring variables along with likelihood for loans. The factors are included as following:

The ‘Debt Ratio’ variable helps to calculate monthly debt payments and costs of maintenance along with other allied costs, divided by gross income. This variable provides a ratio value that is able to imply if the debt mount exceeds the borrower’s monthly income, since the latter is required to exceed the former. Psychological studies indicate that in case of a greater debt ratio, the borrower has a higher tendency of committing the felony of loan forfeiture.  In other words, if the debt ratio exceeds 1, that is when the income is lesser than the debt amount, then the inclination towards delinquency rapidly increases.

This variable identifies if the requirement of the loan is more necessary to the borrower than his monthly salary. Lesser compulsion of the borrower towards the loan with respect to the monthly salary reduces the predisposition of the borrower towards acts of delinquency. In cases where the obligation proportion was significantly greater than 1, the propensity towards committing a felony was shown to have increased at an alarming rate (Sovern 2014).

A second variable is the ‘Monthly Income’, which denotes the gross income of the loan borrower.

Another variable, known as the ‘Number Of Open Credit Lines And Loans’, indicates the number of open loans that the borrower has along with lines of credit and car loan installments. It has been seen that if the debt ratio and the number of open loans of a borrower increases simultaneously, the proclivity towards committing such a crime is augmented.

A different variable, called the ‘Number Real Estate Loans Or Lines’ corresponds to the number of mortgages placed by the borrower against a real estate loan with home equity line of credit (HELOC). This variable is a tool for measuring credit scores that reveal the real estate mortgage loans with lines of credit (Garmaise 2015). It is quite akin to the ‘Number Of Open Credit Lines And Loans’ variable – while the former informs about the number of loans and credit lines, the second variable reveals the number of mortgage loans along with credit equity lines (Moulton et al. 2015).

Finally, ‘Number Of Dependents’ is a variable that represents the number of dependents that the loan borrower requires in case of a particular loan (Xiao et al. 2014). Greater the number of dependents, greater is the urgency from the borrower’s perspective with regard to repayment of the loan and resolution of debt issues. Statistical reports attest to the fact that dependents catalyze the process of repayment of debts back to the bank from the borrower.

When the numbers of people who have taken loan are notable, the borrower sincerely wants to resolve the commitment successfully. Studies that have been performed earlier reveal that the charges are in fact a momentum for the borrower, as they reminded him or her to pay the necessary debt amount back to the bank (Brown 2016).

Monthly Income

The number of people who have faced delinquency in span of the past three months or more, have been revealed by the variable named ‘Serious Dlqin 2yrs’. This variable also gives away the frequency in which the borrower needs to handle the charges against delinquency.

After detecting the various factors of tendencies related to delinquency, which has surfaced in the loan forfeiting cases, the variables from the table are selected, that can be removed from consideration, to reduce complexities. The following include the selected variables along with justifications for omitting them.

Another scoring option for credit, showing the mortgage loans of real estate with credit lines of is the ‘Number Real Estate Loans Or Lines’ (Jiminez et al. 2014). The variable is almost equivalent to the variable known as the ‘Number of Open Credit Lines and Loans’ (Emekter et al. 2015). The first variable shows the open loans as well as credit lines, whereas, the number of mortgage loans and the credit equity lines are revealed by the second variable. For this particular analysis, there must be a certain variable for presenting the number of open loans and mortgages along with the equity lines of credit. In some cases, both of the variables may have redundancy of data and some variances. It is almost sometimes indistinguishable from the variable that is mentioned above. The earlier variable shows the open advances and extensions in credit, whereas, the second variable is at jeopardy to reveal the amount of home loan taken in advances along with credit value lines. Presently in this research, for demonstrating the quantity of open advances and home loans beside extensions to value credit there should be one specific variable. Sometimes, both of the variables (Acharya et al. 2013) represent information that is repeated and inconsistencies.

The revolving utilization of unsecured lines of credit is quite illegitimate with respect to the misconduct of the borrowers. It shows the entire card balance except for the real estate and debt balance for car loan of an individual divided by the total credit left in the card. This result is can be procured from the borrowers’ monthly remunerations along with estimation of open credit lines. The revolving utilization of unsecured credit lines is pretty much unwelcomed due to the strenuous investigation of the defaulters. The interest is calculated on the basis of reducing outstanding loan amount at periodic intervals. This periodic interval is determined by the payment of cardholder. The rate of interest is generally high as compared to other banks due to the greater risk factor of the lender. The unsecured credit lines can be availed depending upon the annual salary of the individual.

The borrower’s income should be greater than his debts. It is psychologically viewed that greater the debt to income ratio, greater are the chances of being a defaulter. According to the decision tree analysis, the first step is to calculate the number of open credit lines. If the number of open credit lines is equivalent to range 1, then the customer will be categorized under a new group. Now, for debt equity ratio if the value becomes less than or equivalent to 0.718, then the customer will be categorized under a pre existing group. However, in case of debt equity ratio greater than 0.718, then revolving unsecured credit utilization will be managed accordingly. 

Number Of Open Credit Lines And Loans

The first step of this exploratory analysis of the given data is to validate whether there are any unusual pattern or not.  The detailed analysis has been shown in the appendix section. According to the analysis done, it can be said that the data set did not contain any unusual pattern. However, in case of missing data, no such aspects present in the given data set. It has seen that if the characteristics of the data set named loan-delinq-train.csv is same as loan-delinq-test.csv data set.

The next step of this study was to identify any significant relationship between the target variable SeriousDlqin2yrs with predictor variables. The researcher has performed correlation analysis to identify such relationship. The detail analysis has been shown in the appendix section. From the correlation, it can be said that very few variables have significant relation with the target variables. However, it is the fact that the customer of ACME bank is likely to forfeit the loan and consequently, the loan become delinquent.  

This section of the report has demonstrated the decision tree analysis performed using the rapid miner tool. The detailed analysis has shown in the appendix section. According to the decision tree analysis, it can be said that “open credit line and loans” is taken into account as the target variables. In case of target variables, below are the list as per decision tree analysis:

[a] Number Of Time 30-59 Days Past Due Not Worse;

[b] Number Of Times 90 Days Late;

[c] Age;

[d] Debt ratio; and

[e] Number of Time 60-89 Days Past Due Not Worse;

According to decision tree, the first checking option is “Number of open credit lines and loans”. Here, if it is fall under range 1, then the customer will be considered as good in terms of credit worthiness. However, if the value falls in range 2 category, then “Debt Ratio” needs to be checked. Again, the decision tree has shown that if the debt ratio becomes less than 0.718, then the customer will be treated as good customer with well credit score. On the other hand, if the debt ratio becomes more than 0.718; then another variable named “Revolving utilization of Unsecured lines” will be checked. Here, if the value is less than or equal to 0.003 than the customer will be treated as good customer with well credit score. Finally, the target variable will be taken into consideration. According to the target variable, that is, SeriousDlqin2yrs value, if it zero then customer will be categorized under first group or else second group.     

According to the logistic regression analysis, the data set contain four variables such as debt to equity ratio, number open credit line and loan, revolving utilization of unsecured loan and serious delinquency in 2 years period, which have significant influence to determine the credit worthiness of the individual customer of ACME bank. According to the Kernel Model, here the weights of each of the attributes are as follows:

W(debt to equity ratio) = 0.708

Number Real Estate Loans Or Lines

W(number open credit line and loan) = 0.720

W(revolving utilization of unsecured loan) = 0.264 and

W(serious delinquency in 2 years period) = 1.052

These weighted values have shown that all these factors are significant predictor of loan delinquency.

Thus, the above decision tree as well as logistic regression analysis has explored that mainly, debt equity ratio, revolving utilization of unsecured loan and serious delinquency in 2 years period are the key variables. Therefore, it can be said that the customer of ACME bank is likely to forfeit on their loan and become a loan delinquency will have true outcome. 

References:

Acharya, V.V., Almeida, H. and Campello, M., 2013. Aggregate risk and the choice between cash and lines of credit. The Journal of Finance, 68(5), pp.2059-2116.

Berrospide, J.M., Meisenzahl, R.R. and Sullivan, B.D., 2012. Credit line use and availability in the financial crisis: the importance of hedging.

Brown, S.R., 2016. The Influence of Homebuyer Education on Default and Foreclosure Risk: A Natural Experiment. Journal of Policy Analysis and Management, 35(1), pp.145-172.

Emekter, R., Tu, Y., Jirasakuldech, B. and Lu, M., 2015. Evaluating credit risk and loan performance in online Peer-to-Peer (P2P) lending. Applied Economics, 47(1), pp.54-70.

Garmaise, M.J., 2015. Borrower misreporting and loan performance. The Journal of Finance, 70(1), pp.449-484.

Goldbach, S. and Nitsch, V., 2015. Cutting the credit line: Evidence from Germany (No. 25/2015). Discussion Paper, Deutsche Bundesbank.

Hazen, B.T., Boone, C.A., Ezell, J.D. and Jones-Farmer, L.A., 2014. Data quality for data science, predictive analytics, and big data in supply chain management: An introduction to the problem and suggestions for research and applications. International Journal of Production Economics, 154, pp.72-80.

Jian Xiao, J. and Yao, R., 2014. Consumer debt delinquency by family lifecycle categories. International Journal of Bank Marketing, 32(1), pp.43-59.

Jiminez, G., Mian, A.R., Peydró, J.L. and Saurina Salas, J., 2014. The real effects of the bank lending channel. Available at SSRN 1674828.

Kaisler, S., Armour, F., Espinosa, J.A. and Money, W., 2013, January. Big data: Issues and challenges moving forward. In System Sciences (HICSS), 2013 46th Hawaii International Conference on (pp. 995-1004). IEEE.

Kitchin, R., 2014. The real-time city? Big data and smart urbanism.GeoJournal, 79(1), pp.1-14.

Koh, E.B., Oh, J. and Im, C., 2014. A study on security threats and dynamic access control technology for BYOD, smart-work environment. InProceedings of the International MultiConference of Engineers and Computer Scientists 2014 (pp. 12-14).

Kornacker, M., Behm, A., Bittorf, V., Bobrovytsky, T., Ching, C., Choi, A., Erickson, J., Grund, M., Hecht, D., Jacobs, M. and Joshi, I., 2015. Impala: A Modern, Open-Source SQL Engine for Hadoop. In CIDR.

Liu, H. and Motoda, H., 2012. Feature selection for knowledge discovery and data mining (Vol. 454). Springer Science & Business Media.

Lohr, S., 2012. The age of big data. New York Times, 11.

Mayer-Schönberger, V. and Cukier, K., 2013. Big data: A revolution that will transform how we live, work, and think. Houghton Mifflin Harcourt.

Moulton, S., Collins, J.M., Loibl, C. and Samek, A., 2015. Effects of monitoring on mortgage delinquency: Evidence from a randomized field study. Journal of Policy Analysis and Management, 34(1), pp.184-207.

Sagiroglu, S. and Sinanc, D., 2013, May. Big data: A review. In Collaboration Technologies and Systems (CTS), 2013 International Conference on (pp. 42-47). IEEE.

Shoenbill, K., Fost, N., Tachinardi, U. and Mendonca, E.A., 2014. Genetic data and electronic health records: a discussion of ethical, logistical and technological considerations. Journal of the American Medical Informatics Association, 21(1), pp.171-180.

Sovern, J., 2014. Fixing consumer protection laws so borrowers understand their payment obligations. Journal of Consumer Affairs, 48(1), pp.17-33.

Varian, H.R., 2014. Big data: New tricks for econometrics. The Journal of Economic Perspectives, 28(2), pp.3-27.

Vayena, E., Salathé, M., Madoff, L.C. and Brownstein, J.S., 2015. Ethical challenges of big data in public health. PLoS Comput Biol, 11(2), p.e1003904.