FRTB IMA: Quantifying Capital Attribution

In the Fundamental Review of the Trading Book internal model approach (FRTB IMA), the primary challenge is delivering transparent and explainable capital figures, rather than computational capacity.

Currently, maintaining IMA approval requires financial institutions to continuously establish desk eligibility through detailed profit and loss attribution (PLA), rigorous backtesting procedures and robust model governance frameworks. 

Despite significant investments in risk engines and data pipelines, capital attribution remains largely unstandardized and fragmented across multiple systems.

This fragmentation creates explainability gaps, limiting the ability of risk and capital management teams to trace capital movements from aggregate risk measures down to underlying market drivers and individual trades, thereby hindering both regulatory compliance and strategic business utility.

To address this industry-wide challenge, we present a five-layered framework for FRTB IMA capital attribution that systematically decomposes capital movements from aggregate regulatory figures to granular trade-level drivers, addressing critical explainability gaps in risk management.

The framework’s interconnected hierarchy begins with the risk-theoretical layer, which decomposes total IMA capital into regulatory-aligned risk measures to establish exactly what has changed at the macro level.

It then progresses into the structural decomposition layer, which identifies capital concentrations within specific risk buckets, followed by the allocation layer, which translates these buckets into actionable, trade-level insights for front-office optimization.

The framework deepens further through the tail scenario attribution layer, which isolates the historical stress events driving tail losses, and culminates in the implementation integrity layer, which reconciles any process or control breaks.

Together, these layers enable stakeholders across the organization—from regulators and senior management to risk teams and front-office traders—to understand capital movements clearly and consistently at every level of abstraction.

Central to the operationalization of the five-layered framework is the introduction of the capital attribution score (CAS), a quantitative metric that measures the proportion of total desk capital successfully explained at each respective layer.

By mathematically defining the ratio of attributed capital to total desk capital, the CAS provides a clear indicator of attribution completeness, where a perfect score signifies full reconciliation and any deficit highlights residual, unexplained capital.

The primary advantage of this scoring mechanism is its facilitation of effective residual governance. It explicitly quantifies non-reconciling portions of capital, enabling institutions to establish targeted controls, assign clear ownership and maintain a rigorous audit trail for supervisory review.

Besides, the CAS transforms abstract explainability requirements into a measurable, governable standard, empowering banks to optimize capital allocation, enhance model validation processes and satisfy stringent regulatory expectations while supporting strategic business decisions across trading desks and business lines.

In short, the CAS quantifies the completeness of capital attribution at each layer, enabling targeted governance and auditability. It helps banks quantify and govern these residuals transparently, rather than leaving them as unexplained plug figures in capital reports.

Overall, the five-layered framework provides a standardized, auditable and actionable solution that enables financial institutions to meet regulatory expectations, optimizing front-office capital allocation and strengthening model governance.