Credit risk modeling
Financial institutions have been taking on more complex types of credit risk than ever before resulting in the need to be able to manage credit risk more effectively. Credit risk modeling enables these institutions to estimate how much credit is ‘at risk’ in the event of default or due to changes in credit risk factors. By doing so, it enables them to price the credit risks they face more effectively and also helps them to calculate how much capital they need to set aside to protect against such risks.
Traditionally, Banks make credit decisions on an individual basis, by utilising debt servicing ratios, financial statement ratios, credit bureau history and external credit ratings. However, more sophisticated Banks now utilise models that can credit score a loan application based upon the application details and comparing these details with previous loan applications and the subsequent performance of these similar applications.
In addition to credit scoring, credit risk models can also assess the loan application on a portfolio basis. Whilst a credit score might be low, the model could determine that the granting of the loan would not adversly affect the portfolio as the new loan is not highly correlated to the existing portfolio.
More recently, the banking industry has shifted from analysing individual loans to analysing what the impact of adding the loan would be on the existing portfolio. This method particularly addresses the risk of concentration of exposure to any one particular industry/location/security type/loan size/loan type etc,.
Another benefit of risk models are the ability to adequately “price for risk” and provide optimal allocation of capital. These models calculate the probability of default (PD) of loan applications based upon the performance of similar loans within the portfolio. Loan applications with higher PD’s obviously would attract a higher interest margin to offset the higher likelihood that the borrower might default.
In my view, the key component of risk models is the correct capturing and ongoing recording of data. Banks are now capturing hugh amounts of data, not only on application of the loan, but each month the bank records “snapshots” of every exposure that records the balance of the loan and importantly any non payment of accruals.
Banks now run large teams of risk managers that focus on the performance of these models and importantly validating the results. The movement in the loss distributions over various risk grades is closely monitored – what way is the distribution moving and what “tweaks” need to be made to refine the credit decisioning on new applications.
Related posts
(No Ratings Yet)