Asset Quality, Earnings, and Capital, Oh My!
How will the recession impact your institution? Particularly in these volatile economic times, stress testing can help drive an answer.
Improving the accuracy of credit risk scorecards leads to better lending decisions, better portfolio management, and ultimately, better risk management. These models are a core component of assessing borrowers’ probability of default, loss given default, and expected loss. DCG's extensive validation experience — combined with our deep knowledge of cutting-edge statistical and machine learning methods — will help you maintain and improve your credit risk scorecards.
The biggest driver of inconsistent and unreliable credit model results is the variable quality of the statistical techniques used to build the model. A rigorous, precise validation by DCG's team of PhDs and statisticians can help determine whether your methodology is suitable for your goals and whether the analytical method is appropriate and reliable.
Not all credit models are statistical in nature. Even statistical models may have a qualitative component. Non-statistical methods can be effective in capturing credit risk, but can be tricky to properly validate. DCG's extensive validation experience verifies that your non-statistical models meet regulatory expectations.
Drawing from a toolkit of more than 20 different testing techniques, DCG's goal is to ensure that your models not only work today, but will perform reliably in the future. DCG's testing includes back-testing, benchmarking, cross validation, statistical tests, and proprietary techniques. Auditors and regulators praise the DCG approach for its comprehensiveness and depth.
The DCG team's risk rating and scorecard model results are reliable for conducting many other functions—such as allowance setting and bottom-up stress testing—that depend on the accurate estimation of your borrowers’ credit risk.
Helped improve institution’s risk position modeling through identification of missing contractual characteristics in loan file by performing data validation.
Identified deficiencies in assumptions development and support process for deposit behavior characteristics. Helped institution better understand alternatives to defend deposit assumptions and deposit stability. Strengthened model assumptions and re-focused management on strategic deposit initiatives.
Re-validated risk models for $2B bank after exam found third-party vendor's validation “did not provide sufficient rigor.” DCG's re-validation with effective challenge uncovered myriad input and model deficiencies that, once addressed, changed bank’s risk profile and strategic direction.