What we do: 

​

We are a consulting team specialized in credit risk modelling.

​

In the context of active credit portfolio management, we help our clients develop or validate:

• Counterparty, facility and collateral risk models (parameters: PD, EAD/CCF/LEQ, LGD and Dilution), 

• Portfolio loss distribution models (parameter: assets’ correlation structure), and

• Securitization models (tranches’ risk parameters: PD and LGD).

​​

These models are used for the following purposes:

• Economic: structuring and pricing of credit products,

• Regulatory: IRB and ICAAP stress tests, and

• Accounting: IFRS 9 and CECL provisioning.

 

Our expertise covers:

• Regulatory and accounting requirements, 

• Data (internal/external) strategy and management, and

• Model development or validation. 

​​

Innovation: 
We owe to our clients to bring value (i.e. better model performance at a lower cost) through innovative solutions. For that reason, we champion the use of:

• Interpretable AI/ML solutions (e.g.: XGBoost, SVM, and Random Forest): initially as challenger models, but now as an integral part of the risk quantification/qualification frameworks. Interpretable, notably via Shapley Values and Importance of Features, 

• Alternative data: external (open or not) – sourced from open depositories (e.g.: OECD, ECB, FRED, etc.), data consortiums (GCD) or market places (S&P CIQ) to name a few, and 

• Simplicity: for equivalent performances, we always favour simplicity – as it often comes with lower costs, and easier review/maintenance of models. 

 

We have a solid track record on numerous wholesale and retail asset classes. 

 

From this initial specialization, we also support our clients in the modelling of ESG risks – with a particular capacity regarding the modelling of climate-related and environmental risks (with our appropriation of the PACTA and CLIMADA models).​

​

Last, but not least, we are cost aware.

We make sure that our daily rates are competitive; and that the frameworks and models we put in place are cost-effective over their life cycles – in terms of: cost of data, time required from analysts to use our models, and cost of implementation, review and maintenance. 

​

Here are some of our recent projects:

• [CIB] Development of an IFRS 9 LGD model –PIT forward looking– covering Large-Corporate unsecured exposures (first of its kind among French banks)

• ​[Banking group] IRB-repair of Financial Institutions rating system (covered counterparties: Banks, Insurers, Mutual and Alternative-Investment Funds, Independent Brokers and Asset Managers - differentiation and quantification of default risk (PD) - incl. ESG risk drivers' coverage and sensitivity analyses)

• [CIB] Building of an historical Cost-of-Risk database (15+ years, incl. 2008-9, proforma of current organization)

• [Banking group] Remediation of the IAA model covering ABCP conduits (securitization) – including the calibration of the correlation parameter (using the Vasicek framework)

• [Banking group, MRM] Independent-validation of Personal-Finance IRB LGD models 

• [Banking group] Backtesting of Factoring rating system (including the dilution parameter)

• [CIB] IFRS 9 Significant Increase in Credit Risk ('SICR') model backtesting  

[…]

 

If you have any questions regarding our value proposal, please contact us at: contact@neonrisk.com