L_sQ_mClaim_RGB

CheQ5 - Model Validation for Banks

abstract1_5393465_RGB

Model validation must meet stringent regulatory provisions

Sound pricing and risk models play a critical role in the management and mitigation of model risk. Risk models specifically determine capital requirements which directly influence a bank’s profitability. These developments have come to regulatory attention as sound modeling approaches are essential to ensuring the long term viability of banking institutions. Banks must now conduct periodic reviews of each model to determine whether it is working as intended and if the existing validation activities are sufficient. Thorough model validation can only be achieved through active management of model risk which is most effectively done by independent model developers.

Regulatory approved Intelligent Technology

swissQuant Group is the world leader in model development and validation. We have the requisite experience, knowledge, skills, and a high level of technical expertise to validate complex models, both in structure and application. Our Intelligent Technology encompasses highly flexible modules for data clearing, proxy generation and validation, and data interfaces. It also encompasses sophisticated database designs and data models able to process large data sets with low retrieval time.

buildings1_16471217_RGB

Ask the experts in derivatives risk management and pricing

Model risk management examines the processes and activities intended to verify that models are performing in line with their design objectives and business applications. It also identifies potential limitations and assumptions and assesses their possible impact on overall risk profiles. For this reason, it is best performed by confirmed experts with appropriate experience and influence.

Pre-emptive Model Risk Mitigation

With CheQ5, swissQuant Group has developed a structured auditing process consisting of five interrelated checkpoints to improve internal model design and development. Years of risk and pricing model development and validation have resulted in extensive model validation libraries able to validate model assumptions even for large panel data. Our modular libraries for integrated model consistency testing, high-dimensional statistical tests, and visual inspection of high-dimensional data are readily available allowing for extremely fast implementation and a considerable reduction of validation project time.


Proven Intelligent Technology

  • Construction of risk and pricing models

  • Development and implementation of hedging frameworks

  • Implementation of large-scale risks systems (Big Data)

  • Regulatory approved model validation and auditing frameworks

Model validation to the point

CheQPoint-1

Data and Implementation Quality

The first check involves investigating input data for differences to alternative data sources and evaluating proxies where data is missing. Model implementation is reviewed; good theory must be accurately translated into self-explaining code.


CheQPoint-2

Conceptual Soundness and Assumptions

In a second check, model assumptions are explicitly and implicitly reviewed by reverting to our extensive software library of statistical analyses and stylized facts. Possible complexity reductions are reviewed and applied methodologies are tested and compared to their empirical model behavior.


CheQPoint-3

Back Testing and Stress Testing

Model robustness and accuracy are exposed to various extreme but plausible scenarios. Changes in inputs and parameters on model outputs are gauged to determine expected ranges. Corresponding tail risks are highlighted.


CheQPoint-4

Model in Model Consistency

This check validates the consistency of model assumptions throughout all model stages to highlight risk exposures. The interplay of all submodels is back tested and stress tested in historical, as well as artificial but plausible environments. Model interdependencies and convergence tests are investigated to ensure econometric test reliability.


CheQPoint-5

Reporting

The final check evaluates the predictive power of different models and their performance. Observed deviations are assessed and corrective actions suggested. Our automatized frameworks allow for fast representation of the final results.

References

swissQuant Group has concluded many successful validations and implementations of pricing and large-scale risk systems. We are one of the largest independent Quant teams in Europe dedicated to delivering lasting client value as demonstrated by the following selected reference projects:
  • Improved margining system of a top-tier European stock exchange clearing house through superior methodology and full model validation.

  • FX option model development for the largest U.S. clearing house.

  • Design of large-scale portfolio risk systems with hundreds of thousands of portfolios of all asset classes for the world’s leading wealth managers.

  • Validation of applied valuation and risk models and methods of new structured products for a provider of fair value pricing.