By Andreas Raggl, Torsten Wegner and Tamar Joulia-Paris

Models are pivotal to the key processes of effective asset management. Since publishing our last article, “Why Model Risk Management matters in Asset Management” at the end of 2017, the risk of model failure has become an ever-greater focus for both market participants and policy makers. This paper provides our views on the latest trends in model risk, its management and related regulations.


Next generation models in Asset Management

Technological advances, Big Data, and AI-based advanced analytics such as natural language processing, machine learning or deep learning are pushing the limits of traditional analytics which were heavily reliant on times series and regression analysis. There is little doubt that these developments will lead to a vastly improved performance of next generation of models which will eventually take automated or autonomous business and risk decision-making to a new level. These types of high-performing new analytics are now increasingly used across investment firms to enhance client segmentation and profiling, investment or product advisory services, or to further improve asset allocation, portfolio optimization, and trade execution.

However, because of the specific fiduciary duty of asset managers to their client base, senior management and risk managers need model-based results to be clearly understood by their customers before they can deploy them in their investment processes. Outperformance on its own is insufficient. The industry is therefore implementing several initiatives, collectively known as Explainable Artificial Intelligence (XAI), designed to solve the problem of interpreting these techniques in a transparent manner thereby overcoming the “black box” nature of many advanced analytics techniques.

In this context, having model validation frameworks on hand that have been specifically developed for advanced data and analytics is critical for asset managers who want to aggressively adopt next generation models.


Regulatory trends in Asset Management 

Both in North America and in Europe, regulatory oversight is becoming more stringent – notably in models used for liquidity risk measurement, for counterparty credit risk management, stress-testing, or cash flow forecasting.

For those asset managers who are part of banking groups with a US footprint, the regulatory demands placed on Model Risk Management (MRM) are growing. Asset managers must find the right balance between meeting the regulatory requirements imposed by their banking parent with benefiting from the added value of a tailored MRM framework. In the US, for instance, we expect the SR 11-7 regulation for Model Risk Management to act as the reference point for liquidity models to comply with Act 40 Modernization which comes into force on December 1, 2018.

Regulatory scrutiny in Europe is also growing. Mirroring the issues faced by the US bank-owned asset managers, their European counterparts will also have to manage increased model-related scrutiny due to the heightened supervisory requirements issued by ECB in its latest Guide to Internal Models. These more stringent regulations are a direct outcome of the TRIM (Target Review of Internal Models) supervisory program which provides comprehensive guidance on the governance, validation and audit of internal models to manage their risks. In addition, MiFID II imposes requirements on funds around risk reporting, which should be viewed in tandem with the EBA consultation paper on capital and liquidity management for investment firms.

All of the discussed regulations require measurement and independent validation of the underlying models and tools used which means a robust MRM framework is crucial. These regulations are being introduced in the context of several wider industry trends. These include the growing use of diverse advanced analytics techniques; migration towards illiquid assets and infrastructure finance investments; compliance with the new “Simple, Transparent and Standard” securitisations standards; and stricter requirements for stress testing. We believe this multiplicity of factors may act to further propel the sophistication and complexity of modelling, increasing the need for appropriate model validation and MRM.


MRM trends in the European investment industry

Investment firms are transforming their risk management functions to embrace digitisation, manage emerging risks, reduce costs, and prepare themselves for the next downturn.

Across the industry, we see several key trends emerging in MRM:

  • Model risk is considered as an enterprise risk in its own right and, as a consequence, is finding its way into dedicated model risk policies, into the risk reporting to senior stakeholders, and into risk appetite statements. However, very few institutions hold a capital or liquidity buffer for model risk yet;
  • MRM is becoming increasingly established as a new risk function, reporting to chief risk officers, alongside investment risk and operational risk;
  • In 2018, MRM functions primarily focus on governance and model risk oversight, organization and policies, and setting up or completing model inventories. They are also increasingly carrying out more formalized model validation activities.


MRM challenges for investment firms

  • The governance and control of the model lifecycle remains strongly focused on model validation, addressing deficiencies in model coverage, model documentation and ongoing monitoring. Managing model risks consistently at both an individual fund and investment firm level and, where relevant, aligning them with the overarching bank or insurance group practices, is still work in progress for many.
  • Model risk managers are striving to establishing more complete model inventories using consistent model definitions. These inventories aim to capture all models, at fund and investment firm level, whatever their purpose, and whether developed internally or by external vendors. They also include next generation models, financial planning tools, as well as behaviour, climate and cyber risk models. Given the size and complexity of investment firm this is often a non-trivial effort.
  • Staffing MRM and Model Validation functions remain challenging, as significantly more internal resource is directed towards model development. To complete more validation work while contributing to overall cost targets, partnerships with external model validation service providers are increasingly being used to enhance the efficiency of the function.

The number and diversity of models used in asset management is expanding year on year, as traditional and advanced analytics are now put into practice across all areas of organizations. Whether models are developed internally or procured from third party providers, leading asset managers are taking the necessary steps to establish a robust and modern model risk management framework.

Because of the fiduciary duty of investment firms, this transformation is crucial for the industry to enable them to leverage the value of the new data and analytics that will significantly improve their investment performance and operational efficiency.

To tackle these increasing challenges, the framework introduced in the whitepaper “Why Model Risk Management matters for asset managers” could provide useful guidance.

© RiskDynamics | Part of McKinsey & Company