Objective

To help a large European investment bank enhance stress-testing capabilities and meet tight regulatory deadlines by creating econometric models to project a wide range of macroeconomic and business variables and developing alternate scenarios for CCAR submission.

CRISIL's Solution

• Modelled variables including vehicle sales, export of goods, vehicle imports, non-farm payroll, aggregate hours worked (private sector) and personal consumption
• Expanded on client suggestion to use exogenous modelling approach (linear regression) by also employing endogenous modelling using vector auto regression
• Applied univariate analysis: factor selection, variable transformation, lag average, number of lags, rank transformation
• Applied multivariate analysis: multi-collinearity tests, goodness-of-fit, residual analysis – serial correlation, heteroskedasticity: multi-collinearity tests, goodness-of-fit, residual analysis – serial correlation, heteroskedasticity
• Stressed and adverse stressed scenarios developed by simulating the beta-coefficients of the model within their respective confidence levels - for 75% and 90% confidence levels, respectively
  

Client Impact

CRISIL's exogenous models from both linear regression and vector auto regression significantly enhanced client stress testing and allowed the investment bank to meet tight deadlines for required CCAR submission.

Questions

Looking for high-end research and risk services? Reach out to us at: