Model risk mitigation for pricing services: from the model owner’s lens

Financial markets rely heavily on quantitative models for decision-making, making effective model risk management crucial. Attika Raj, senior specialist, complex securities pricing at LSEG Data & Analytics, emphasises the importance of the first line of defence – model developers and owners – in ensuring model accuracy and reliability, particularly in pricing services

Financial markets are a complex system, in which firms rely heavily on quantitative models – risk, valuation and predictive – to make informed decisions. Model risk, the risk of errors in these models, arises from data inputs, processing logic, assumptions and output variables that constitute the model. Effective management of this risk is essential to ensure accurate and reliable outcomes.

To manage model risk, firms use various risk management frameworks, guided by regulatory standards such as the US Federal Reserve’s SR 11-7 guidelines. These guidelines emphasise the three lines of defence approach: model developers and owners (first line of defence), validators and reviewers (second line) and the auditors (third line). Each line plays a critical role in ensuring the robustness of the models.

Ensuring accuracy in pricing services

Pricing services provide valuations and prices for financial securities, often calculated using complex models. For instance, a corporate bond valuation model might use market data such as interest rate curves, security attributes (coupons, face value, maturity), issuer credit curves, trade information and other inputs to calculate bond price using such techniques as stochastic interest rate modelling or the Monte Carlo simulation.

However, these models carry risks related to input quality, assumptions, calibration processes and calculation logic. For example, the option‑adjusted spread, which is typically derived from a trade or evaluated price using a model, can vary significantly based on the volatility assumption and the model type. Given that the debt market usually quotes on a spread basis, this variability can lead to securities comparability issues and differences in portfolio valuation calculations, potentially leading to adverse outcomes.

Different asset types also introduce varied risk factors. A corporate bond valuation model primarily incorporates interest rate risk and credit spread risk factors, whereas a commercial mortgage-backed security valuation model also has commercial real estate-related risks. Therefore, different tailored risk strategies are necessary for different models to ensure effective risk management.

Pricing services outputs are widely used by financial firms for assets and liabilities valuations, by auditors for issuing qualified and unqualified opinions, and by asset managers for fund net asset value calculations. Poor-quality prices can pose risks not only to individual firms, but can also potentially lead to systemic risks across financial markets.

Regulatory implications

Financial services firms are subject to various regulatory requirements that mandate rigorous vetting of the prices used. For example, the US Securities and Exchange Commission rule 2a-5 requires investment companies to assess risks associated with fair-value determinations, establishing fair-value methodologies and overseeing pricing services. Hence, pricing services must manage their model risks proactively and transparently, providing clients with high-quality prices and effective price challenge resolutions to their clients.

Additionally, they need to prepare comprehensive documentation of their risk management frameworks, techniques and findings, making it easier for the clients to fulfil their regulatory requirements.

Ensuring sound model risk management

To ensure high-quality pricing, a pricing service must focus on the quality of its models. This is best ensured by implementing a sound model risk management framework. Model owners or developers – as the first line of defence – must ensure the models are well designed, logically sound and aligned with their intended use. This involves thorough testing and validation of design, logic and data inputs. Additionally, they must document the model’s description, logic, assumptions, procedures, test results and any limitations.

Model owners must also evaluate model assumptions, limitations and outputs rigorously, using qualitative as well as quantitative techniques – wherever applicable – such as sensitivity analysis, backtesting, outcome analysis and benchmarking to test model outputs.

Model owners should establish a well‑defined price validation and quality assurance framework, ensuring periodic assessment, testing and validation of pricing models. This framework and practice should be well articulated and executed in a timely and efficient manner. Clear communication between model owners and model reviewers, validators and auditors is essential. This allows model owners to incorporate suggestions for improvements and resolve issues promptly.

The end-goal is to mitigate the majority of the model risk at its source before involving the second and third lines of defence.

Conclusion

Model risk is a significant concern for pricing services. Poorly managed model risk can lead to adverse outcomes for firms and potentially escalate to systemic risks. The role of model developers and owners is critical in managing this risk, ensuring they deliver high‑quality and reliable prices to clients and the broader market.

 

Sponsored content