When designing the credit score table we may classify the factors into groups such as:
After deciding the different factors under each group, we need to put them in a table. This can be done with an Excel spreadsheet if the bank does not have a dedicated program such as the Fair Isaac credit scoring program. We then start fixing the weighted value for each factor, and then for each tier under each factor.
The score should be reviewed on a monthly basis by repeating the same process, starting with getting an updated delinquency report from IT, which should be sorted by each factor and if major changes have taken place, the score should be recalculated. In this way, the scores can be fine-tuned on a monthly basis upon review of the progress of the delinquency.
To do this exercise you need to decide which factors you think are best to evaluate the credit applicant for the (for example) auto loan. Divide the factors into four groups as explain above. Start assuming the weighted value for each group and then divide each factor within the group into value. Under each factor, you need to divide it into different tiers (for example, with the source of income $4,000 to $5,999 as tier 1, then $6,000 to $8,000 as tier 2). You will then allow a weighted value for each tier.
You will need to have one tier representing the highest risk as zero. The lowest risk tier is assigned the highest value that you have assumed for that factor. The total value for all tiers could be any value – there is no need to use 1 to 1000 or any other rounded figure – as the customer will be evaluated by a percentage. For example, if the total score is 1550 and the applicant got a score of 1230, the credit score will be 79 percent.
The vintage report is a report generated on a monthly or sometimes quarterly basis to analyse and measure the performance of a portfolio of credit product (or the whole credit portfolio) over different periods following the granting of credit or loans. The vintage report can help track the progress of payments of certain borrowers, group borrowers, segments, particular retail credit product portfolios, or even the whole retail credit portfolio.
The vintage report’s analytical data helps reliably score and efficiently monitor potential risk exposure.
|Dates||No of Grants|
This table shows that in one year of performance, in the month of January 21, there were 15,000 number of cases granted. In the third month from the date of granting the credit, there were 25 cases of defaults.
That is 0.00167% of the cases.
The number of default cases increased to 60 in the fourth month (or 0.004%), and reached 321 in month twelve (0.021%). If the expected loss planned is for 3% then we are in safe status as the highest is 2 percent. We can also look for certain dates. For example, in May 21, there were 42 cases of defaults in month four with a percentage of 0.0025.
This gives us a a chance to understand the defaults movements and the rate of growth or decline. Noting these movements, we can tell that we are on the right target track – or that we need to fine-tune our credit policy.