Background
Are you aware of all the models that banks are currently using in Switzerland? And are you familiar with the associated maintenance and validation requirements?
These days, banks (and securities dealers) in Switzerland use a variety of models for risk management, for business decisions and regulatory reporting purposes (see table below). As the Swiss regulator (FINMA) approves and monitors the use of many of these models, banks need to comply with additional supervisory parameters.
| Credit parameters | Credit risk | Market risk | Operational risk |
| PD (Logistic Regression) |
SA, IRB-F, IRB-A | VaR, IRC, Exp. Shortfall | BIA |
| LGD (Specific Models) | Economic Risk Capital | Greek Sensitivities | SA |
| EAD (SA-CCR, IMM) | IFRS 9 Impairment | FRTB | AMA |
| ALM/treasury | Liquidity risk | Value adjustments | Stress testing |
| IRRBB (EVE, NII) | Liquidity-adjusted VaR | CVA, DVA | CCAR (FED) |
| Behavioural Models | Intraday Monitoring | FVA, COLVA | Stress Test (PRA) |
| Replicating Factors | Liquidity Risk | KVA, MVA | Capital Planning |
With growing regulatory challenges and increased complexity, all these models require independent model validation. This raises questions about resources and competences, and leaves banks with a new category of risk: model risk.
Model risk is the potential for adverse consequences of decisions based on incorrect or misused model outputs and reports. The result can be financial losses, poor business and strategic decision-making, or damage to a bank’s reputation.
Independent model validation mitigates model risk for banks by identifying fundamental errors and avoiding the incorrect use of models and inaccurate outputs. It provides senior management and the Board of Directors with assurance, and fulfils the regulatory requirements regarding independent model validation.
Unlike the largest players in the Swiss market (such as FINMA-supervised category 1 and 2 banks), many Swiss banks do not have their own model validation framework and/or teams with in-house expertise to validate their models.
Although it is usually the CFO or CRO who owns the models, the ultimate responsibility remains with the Board of Directors. This is because the outputs of many models are used to fulfil regulatory capital and liquidity requirements for which the Board is ultimately responsible.
New mandatory independent model valuation for Swiss banks
Since 2011, major financial supervisors and international authorities, such as the FED, BIS, ECB and PRA, have issued diverse regulatory publications that set out principles for the effective management of banks’ model risk by validating models. Before 31 December 2018, the only recommendation in Switzerland was to set up a model validation framework by referring to the publications of foreign standard setters (see table below).
| Authority | Year | Publications on model risk | Description |
| FED | 2011 | Supervisory Guidance on Model RiskManagement (SR 11-07) | Comprehensive supervisory guidance with principles on effective model risk management |
| BIS | 2012 | The Internal Audit Function in Banks (BCBS 223) | Principles on supervisory guidance to assess effectiveness of internal audit |
| EBA | 2014 | Supervisory Review Evaluation Process (SREP) | Comprehensive assessment of banks’ strategies, processes, risks and capital planning |
| ECB | 2015 | Targeted Review Internal Models (TRIM) | Guidelines to reduce inconsistency and variability of capital required when banks use internal models |
| PRA | 2017 | Model Risk Management Principles for Stress Testing | Principles on model risk management for banks relying on internal models to perform stress tests |
FINMA-circular 2017/1 assigns responsibility for developing and operating risk monitoring systems, defining and applying principles and methods for risk analysis and assessments (e.g. validation of models) to the risk management function.
On 1 January 2019, FINMA issued its revised circular 2019/2mon the Interest Rate Risk in the Banking Book (IRRBB). The circular implements the global IRRBB principles issued by the Basel Committee for Banking Supervision (BCBS) for Swiss banks. It requires independent validation of risk measurement systems and models, for example asset and liability management tools for interest rate risks. Although FINMA allows smaller banks to apply some simplifications, independent validation is required for such systems and models for all Swiss banks at least every three years or in the event of a significant change in the data, system, models and parameters.
Our services: how we can support you?
New risk management, business and regulatory requirements raise various challenges in terms of the effectiveness of banks’ model risk management. This will require diverse expertise and more resources. To help banks develop their model risk management framework and run model validations, PwC offers ‘one-off’ validation services, benchmarking and a fully outsourced managed service.
PwC teams of model risk experts with a technical background and proven experience provide assurance to CFOs, CROs, Executive Management, Risk Committees and/or Board of Directors on the adequacy on the adequacy and efficiency of the bank’s controls designed to reduce model errors. PwC can provide such services to non-audit clients in each of the three lines of defence (LoD), depending on the requirements of the bank:
A bank’s management benefits from PwC’s network to benchmark approaches in use in countries where standard setters require more comprehensive and more frequent stress testing results from supervised banks than in Switzerland. With our global practice, PwC has a strong track record in supporting banks with ECB stress testing, the implementation of IFRS 9 and CECL, TRIM Margin of Conservatism and IRB models.
Banks have the option of outsourcing the model validation function to PwC. In such cases, independent and dedicated PwC teams take charge of the end-to-end model validation process, while banks focus their efforts on their core business and risk management. This way banks benefit from PwC’s industry-wide experience, can maintain high quality standards and gain increased confidence in model risk management. Additional factors to consider are ensuring service continuity, tracking industry developments and maintaining critical team sizes – all of which can be difficult to maintain internally for such a specialised task. It can therefore be effective and efficient to outsource this alongside the core model validation work.
