Capital Stress Test Model Validation

In today's environment, accurately measuring and effectively managing risk to capital is a vital need. A sweeping regulatory response, including DFAST and CCAR stress testing requirements, has created a new business risk assessment and regulatory climate for boards and managers.

Benefits of MountainView's Validation

Capital stress test (CST) models include internally-developed credit loss, PPNR, and capital reporting models, as well as vendor-based aggregator models supporting both DFAST and CCAR applications. To validate CST models, MountainView employs an end-to-end review process that is thorough, model-centric and entirely consistent with current supervisory expectations. Our systematic approach to model validation ensures that all dimensions of your CST model development, implementation, and use (conceptual design, technical and statistical correctness, outcomes accuracy and controls) are effectively challenged and affirmed, thereby empowering you to meet your DFAST and CCAR compliance and capital management requirements.

Capital Stress Test Model Validation Methodology

MountainView will perform a comprehensive validation of all models that support your capital stress testing regimen. Our CST model validation process is comprised of four primary components including model conceptual assessment, model technical validation, model forecast validation and model governance review.

The first validation component, the model conceptual assessment, evaluates the conceptual capability of each CST model to meet current regulatory mandates and relevant business needs. We evaluate the overall modeling solution as to whether it is an appropriate tool for gathering relevant risk positions, identifying risk correlates, assigning risk metrics and parameters, and converting them to stressed outcomes.

The second component, the model technical validation, evaluates the CST model’s internal construction and potential to produce accurate, reasonable stress estimates and outcomes. When completing this component of our CST model validation process, we determine if the overall level of technical and statistical adequacy of the CST model is adequate compared to leading industry practice and current regulatory requirements.

The third component, the model forecast validation, assesses the precision and reasonableness of stress estimates and outcomes predicted by the CST model given the manner in which your balance sheet and macroeconomic factors are defined for each stress scenario. Our model forecast validation is designed to deliver diagnostic verification of CST model forecasts relative to absolute standards and plausibly defined reasonableness tests. Specifically, this component of our CST model validation process deliverable determines if the stress test forecasts of risk exposure direction and magnitude relative to underlying macro conditions and modeled risk drivers are plausible.

The final validation component, the model governance review, assesses the governance solution supporting the CST model’s development, implementation and use. Specifically, this component evaluates the model’s control environment, development and implementation documentation, and framework of applicable internal policies and procedures.

Capital Stress Test Model Validation Report

MountainView’s CST model validation report deliverable will facilitate your full compliance with current DFAST and CCAR model validation requirements. The final report set deliverable includes an executive summary which presents a CST model risk rating and general assessment of the model’s development and/or implementation quality. This overall model risk rating assessment gauges the statistical robustness, technical correctness and suitability of the CST model’s development, implementation and use relative to leading practice and current supervisory expectations. The executive summary also includes a prioritized list of validation observations and recommendations to guide your remediation and model enhancement efforts post-validation. Following the executive summary, we include validation report component narratives which present detailed and comprehensive results of our assessment of model conceptual soundness, model technical correctness, model forecast accuracy and model governance.