Analysis Of Home Loan Characteristics At Auswide Bank: An Essay.

Background of Auswide Bank

Auswide Bank founded in 2015 has been operating for more than 50 years in Australia and being considered as the 10^th largest bank in terms of assets with over $3 billion of asset valuation. The extensive range of services offered by the bank is evident in form of banking and finance products which helps the customers realise their financial dreams. The bank is further listed as ADI under ASX and supervised by “Australian Prudential Regulation Authority” (Auswidebankltd.com, 2022a). The relevant values of Auswide Bank are to put an augmented focus on values such as ‘empower’, ‘make it happen’, ‘purpose’, ‘own it’, ‘wow’, ‘ethical’ and ‘real’. Since the COVID-19 pandemic, Auswide Bank has taken significant steps in encouraging the users to use the mobile app and adhere to government protocols for using the services in rideshare vehicles, airports, public transport, hospitals and retail centres (Auswidebankltd.com, 2022c).

The empower value is related with empowering the staffs and customers for initiating a change. On a similar note, the make it happen value is associated with making the relevant decisions and add them quickly for adhering to the needs of the customers. Similarly, the value related to the purpose for Auswide Bank is about identifying the purpose and being passionate about it. The own it value can be seen with owning the outcomes, customers, decisions and actions. The value related to the wow factor is all about exceeding the expectations of the customers and celebrating the success. The ethical value relates to being committed with ethical aspects and operating in a sustainable workplace. The real value is about building honesty and open relationships for delivering the promises (Auswidebankltd.com, 2022b).

Based on the data provided for the credit profile of the customers of Auswide Bank, we will be able to determine how the factors such as gender, owning a real property, applicant income, loan amount, yearly loan repayment, value of the property, nature of applicant source of income, education, relationship status, type of accommodation, age, years employed and employer type. The main purpose of this report is to support the digital transformation strategy of Auswide Bank so that the bank is able to understand its problems, business opportunities and progress toward the goals of the decision-making. These would be conducive for the bank in attaining the economic objectives, thereby acknowledging the personal values and creating more value for the stakeholders in the business. The various excerpts of the report have been conducive in stating the analytics goals, preliminary understanding of the data, describing the variables, justifying the expectation of the relationships and capability using the CRISP-DM framework (Huber et al., 2019). Based on the aforementioned business background, the research questions pertaining to creation of value for the organisation have been illustrated below as follows:

RQ1: What are the main factors which drive home loan characteristics?

RQ2: What are the different demographic factors which affect loan amount at Auswide Bank?

RQ3: What is the relationship between income and the loan amount taken by the customers at Auswide Bank?

Values of Auswide Bank

The dataset was extracted from home loan applicants at Auswide Bank for the period 1 July 2016 to 30 June 2021. The selected variables for the study are listed as follows:

1. Gender
2. Own_realty
3. Cnt_children
4. Cnt_hh_members
5. Applicant_income
6. Loan_amount
7. Yearly_loan_rep
8. Prop_value
9. Income_type
10. Education
11. Relationship_status
12. Living_status
13. Age
14. Years_employed
15. Employer_type
16. Ext_score

The aforementioned variables directly impact the home loan characteristics. These variables are related with the research questions which are to be explored. Moreover, the analytics tools used for the study are evident in form of descriptive statistics, exploratory analysis, correlation, regression analysis and Two Sample (Independent) t-Test. The overall sample size was modified as a result of outliers in the original data. The evaluation of the box plot pertaining to the original data has been conducive in identifying the lower outlier cut-off and upper outlier cut-off respectively. Similarly, the loan amount has also taken into consideration the upper and lower cut-off limits for removing any possibility of outliers in the final data. The sample size was reduced to 200 observations as the data has been analysed using MS Excel which has a restriction of maximum number of data series per chart of 255 values (Babura et al., 2018).

Table 2: Calculating the Lower and upper outlier cut-off

(Source: As created by the author)

Based on the Box plot of the original data, we can clearly see how there exists outlier above 337500 values based on applicant's income and 1616625 value as per the loan amount. 

Figure 1: Box Plot - Original Data

(Source: As created by the author)

The box plot of the final data has been prepared by taking into consideration the lower cutoff of 22500 and upper cut-off of 337500 for applicant’s income has been helpful in removing the outliers from main variables. Similarly, the lower cut-off of 537975 and upper cut-off of 1616625 has been considered for the loan amount. 

Figure 2: Box Plot – Final Data

(Source: As created by the author)

As per the analysis of the descriptive statistics, we are able to identify how the applicant’s average income applying for loan is $181754.80. Additionally, the average loan amount applied by the applicants is evident with $878401.9. A similar note, the yearly total loan repayments is evident in form of $35185.07. Based on the median values, we can see the median of applicant’s income as 180000, loan amount as 808650 and yearly total loan repayments as 32647.50. The standard deviation of all the numeric variables is very low and closely clustered around the mean, thereby ensuring reliability of the dataset. Additionally, the mode is also free from any outliers. The confidence interval of 95% for all the variables indicate how the probability of 95% confidence interval ensures that most of the data will fall close to the mean (Kaur, Stoltzfus & Yellapu, 2018).

Descriptive Statistics 

Table 1: Descriptive statistics of all the Numeric variables

(Source: As created by the author)

The demographic characteristics having an impact on the loan amount is evident in form of relationship status, gender, education, living status, income type and employer type. Based on the depiction of gender, we can see that female have not only applied for more amount of loan comparison to the male part they also have higher property value and income. Such a trend is due to higher female population in Australia (Knoema.com, 2021). 

Research Questions of the Study

Figure 3: Relationship of gender on loan amount, property value and income

(Source: As created by the author)

It is evident that as per relationship status, the married individuals have applied for higher home loan in comparison to the individuals who are widow, single, separated or engaged in civil marriage. In this regard, we can discern that married individuals are more in need for home loans (Rachmansyah, Harijono & Prabowo, 2021). 

Figure 4: Relationship of relationship status on loan amount, property value and income

(Source: As created by the author)

The individuals with secondary or secondary special are higher applicants for home loan. This specifies that people with higher qualification have disposed higher tendency for taking loans (Rachmansyah, Harijono & Prabowo, 2021). 

Figure 5: Relationship of education on loan amount, property value and income

(Source: As created by the author)

Individuals who are living in House/apartment have higher tendency of applying for home loan (Rachmansyah, Harijono & Prabowo, 2021). 

Figure 6: Relationship of living status on loan amount, property value and income

(Source: As created by the author)

It is quite evident that working categories of individual have demonstrated higher disposition for applying for a home loan (Delmelle, Nilsson & Schuch, 2021). 

Figure 7: Relationship of income type on loan amount, property value and income

(Source: As created by the author)

In terms of employer type, the individuals applying for home loans belong to business entity type 3. 

Figure 8: Relationship of employer type on loan amount, property value and income

(Source: As created by the author)

It has been discerned that individual with ownership of real estate has a tendency of not only taking higher home loans but also more amount of property value as a result of higher income (Delmelle, Nilsson & Schuch, 2021). 

Figure 9: Relationship of employer type on loan amount, property value and income

(Source: As created by the author)

The research hypothesis for the study can be further formulated as follows:

H₀₁: There does not exist any direct relationship between income and the loan amount taken by the customers at Auswide Bank

H₁: There exists a direct relationship between income and the loan amount taken by the customers at Auswide Bank

H₀₂: There does not exist any relationship between characteristics such as number of children living with the loan applicant, number of people with the loan applicant, years of employment, age, income, property value and external credit rating on the loan amount taken by the customers.

H₂: There exists a direct relationship between characteristics such as number of children living with the loan applicant, number of people with the loan applicant, years of employment, age, income, property value and external credit rating on the loan amount taken by the customers.

The t-Test: Two-Sample has been used for the depicting the relationship that exists between income and the loan amount taken by the customers at Auswide Bank. The stress level of income of the applicant (M=181754.8, SD=66357.16, n=200) was hypothesised to be lesser than the stress level of amount of loan taken by the individuals (M=878401.895, SD=272901.90, n=200). Moreover, this difference was not found to be significant as, t (398) = 1.965, p =3.8907E-124 (1 tail). In this regard, we fail to reject the null hypothesis H₀₁ (Kelter, 2020)_. 

Data and Methods

Table 3: t-Test: Two-Sample Assuming Equal Variances

(Source: As created by the author)

The correlations matrix suggests that the highest correlation exists between property value and loan amount with a correlation of 0.96. Along with this, there also exists significant relationship between value of the property and loan amount. This is evident with a correlation of 0.64. 

Table 4: Correlation Matrix

(Source: As created by the author)

Table 5: Regression analysis

(Source: As created by the author)

The regression analysis has been used for testing H₂. Based on the high R-square of 0.924 we are able to discern that the independent variables such as Cnt_Children, Cnt_Hh_Members, Years_Employed, Age, Applicant_Income, Prop_Value And Ext_Score is able to influence the dependent variable by up to 92.43%. However, as the p value of all the variables except property value is greater than 0.05, we are unable to reject the null hypothesis. Therefore, we can say that the property value as a direct impact with the loan amount.

Based on the findings of t-test, correlation and regression analysis for the numeric variables, we can discern that there exists a statistically significant relationship only between property value and loan amount. However, the use of exploratory analysis for the categorical variables really suggests how the demographic factors such as relationship status, gender, education, living status, income type and employer type have a direct impact on the amount of loan applied by the individuals.

The main limitation of the study can be seen with only consideration of Auswide Bank which prevents generalisation of the analytical results. In addition to this cause of the other variables which can be taken into consideration can be seen with previous credit history of the customers with other banks pertaining to home loan applications. Some of the main organisational issues which can prevent Auswide Bank in obtaining such data can be related with confidentiality clause of other banks related to customer internal report. These issues can be addressed by designing an industry leading method for computing customer credit rating using the latest AI technology (Chacko, Antonidoss & Sebastain, 2020).

The explicit value which can be drawn from RQ1 can be seen with factors such as gender, ownership of real property, children, income, the lonely payments, property value, income type, education, relationship status, living status, age, years employed, employer type and region rating affecting home loan characteristics. Additionally, the various demographic factors such as relationship status, gender, education, living status, income type and employer type directly adds value of the business when it comes to depicting the demographic factors affecting loan amount at Auswide Bank. The use of t-test is unable to find a statistically significant relationship between income and loan amount applied by the individuals. However, it is important for the bank to consider total property value before disbursing the final loan amount to the customers.

References

Auswidebankltd.com. (2022a). Auswide Bank - Small things. Big difference. Retrieved 31 January 2022, from

https://www.auswidebankltd.com.au/about-us/about-auswide-bank/

Auswidebankltd.com. (2022b). Auswide Bank - Small things. Big difference. Retrieved 31 January 2022, from

https://www.auswidebank.com.au/about-us/careers/our-values/

Auswidebankltd.com. (2022c). Important Information | Auswide Bank (2022). Retrieved 31 January 2022, from https://www.auswidebank.com.au/campaign/important-information/

Babura, B.I., Adam, M.B., Samad, A.R.A., Fitrianto, A. & Yusif, B., (2018), November. Analysis and assessment of boxplot characters for extreme data. In Journal of Physics: Conference Series (Vol. 1132, No. 1, p. 012078). IOP Publishing. <DOI:10.1088/1742-6596/1132/1/012078>

Chacko, A., Antonidoss, A. & Sebastain, A., (2020). Optimized algorithm for Credit Scoring. International Journal, 9(1.3). < https://doi.org/10.30534/ijatcse/2020/5691.32020>

Delmelle, E.C., Nilsson, I. & Schuch, J.C., (2021). Who's Moving In? A Longitudinal Analysis of Home Purchase Loan Borrowers in New Transit Neighborhoods. Geographical Analysis, 53(2), pp.237-258. <

https://doi.org/10.1111/gean.12234>

Huber, S., Wiemer, H., Schneider, D. & Ihlenfeldt, S., (2019). DMME: Data mining methodology for engineering applications–a holistic extension to the CRISP-DM model. Procedia Cirp, 79, pp.403-408.

<https://doi.org/10.1016/j.procir.2019.02.106>

Kaur, P., Stoltzfus, J. & Yellapu, V., (2018). Descriptive statistics. International Journal of Academic Medicine, 4(1), p.60. <DOI: 10.4103/IJAM.IJAM_7_18>

Kelter, R., (2020). Analysis of Bayesian posterior significance and effect size indices for the two-sample t-test to support reproducible medical research. BMC Medical Research Methodology, 20(1), pp.1-18. <DOI:10.1186/s12874-020-00968-2>

Knoema.com. (2021). Australia Male to female ratio, 1950-2021 - knoema.com. Retrieved 31 January 2022, from https://knoema.com/atlas/Australia/topics/Demographics/Population/Male-to-female-ratio

Rachmansyah, Y., RA, A.D., Harijono, H. & Prabowo, R., (2021). The Determinants of Home Mortgage Default Probability: The Effect of Loan and Borrower’s Characteristics. <DOI:10.4108/eai.4-11-2020.2304567>