top of page
  • Writer's pictureWilliam Webster

Valuations: Model Governance and Fraud Detection

Introduction

Financial valuations, an integral part of the financial industry, involve assessing the worth of financial products. While numerous methods are available for valuation, selecting the most appropriate one for each product is crucial. However, even with meticulous approaches, valuations are not immune to errors, and recognizing these inaccuracies is essential to prevent significant financial repercussions.


Valuation Accuracy and Its Importance

Valuations typically integrate into financial statements with an assumption of accuracy. In most cases, these numbers are correct and are appropriately recorded in accounts. However, occasional discrepancies emerge due to various factors such as differences in valuation models, timing of market data acquisition, or minor variations in applied models. While often these are innocuous, there are situations where incorrect valuations pose significant problems, notably in instances of poor model governance or outright fraud.


Model Governance in Valuation

Model governance refers to the oversight of financial models used for valuing complex and often illiquid financial products. These models, characterized by their mathematical intricacy, require careful handling due to their sensitivity to input changes. Subject matter experts, typically with robust quantitative skills, are entrusted with this task. However, their expertise does not guarantee infallibility. It's not uncommon to find models that are more theoretical than practical, leading to valuation issues.


For example, valuing illiquid mortgage securities might involve using a swap-based yield curve supplemented by a credit spread from various entities. While seemingly credible, this approach can introduce significant basis risk and inaccuracies, particularly if the basket of credits for the credit spread is narrow. These valuation errors can have drastic consequences, especially when used in stress tests for determining capital requirements, often leading to guesswork presented as definitive numbers at high-level decision-making forums.


Mitigating Model Governance Risks

To mitigate risks associated with model governance, a thorough examination of the model and its methodology is imperative. This process involves questioning the logic and simplicity of the model, understanding the source and impact of its inputs, and assessing the scenarios used in stress testing. Such scrutiny ensures the valuation is more than a theoretical construct and is grounded in practical and logical reasoning.


Fraud in Financial Valuations

The second critical issue in financial valuations is fraud. Despite improved oversight and regulatory frameworks, the risk of mispricing to conceal fraud remains. Typically, this involves traders or investors inflating the value of assets to hide losses. These fraudulent activities can persist unnoticed for extended periods, compounding losses until they become unsustainable.


Preventing Valuation Fraud

Effective fraud prevention in valuations requires strict segregation of duties, ensuring that those involved in trading or investment are separate from the valuation process. Regular and independent sampling of trades and assets to validate their market value is also crucial. This involves marking them to market independently and ensuring that the data and models used are current and appropriate. Comparing these independent valuations with reported values can provide assurance of accuracy and detect potential manipulations.


Conclusion

In summary, the realms of model governance and fraud are two primary areas of concern in financial valuations. Both can lead to significant, sometimes irrecoverable, losses. Addressing these risks involves critical evaluation of valuation models, independent sampling, and clear segregation of roles in the valuation process. By implementing these measures, firms can safeguard against the perils of incorrect valuations and ensure financial integrity.

5 views0 comments
bottom of page