Internal Models Approach (IMA) for Market Risk: A Complete Guide Under FRTB

Market risk—the potential for losses due to changes in market prices (e.g., interest rates, equity prices, foreign exchange)—is a critical concern for financial institutions. To mitigate this, regulators require banks to hold capital as a buffer against unexpected losses. The Internal Models Approach (IMA) is an advanced regulatory framework that allows banks to use their own internal risk models to calculate market risk capital requirements, offering a more tailored alternative to standardized methods. Introduced under Basel II and refined by the Fundamental Review of the Trading Book (FRTB) (Basel III), IMA is designed to better align capital with a bank's actual risk profile. This blog explores how IMA works, its key components, updates under FRTB, and its pros and cons.

Table of Contents#

  1. What is the Internal Models Approach (IMA)?
  2. How Does IMA Work?
  3. Key Components of IMA
    • 3.1 Expected Shortfall (ES)
    • 3.2 Stressed Expected Shortfall (Stressed ES)
    • 3.3 P&L Attribution Test
    • 3.4 Model Validation
  4. FRTB Updates to IMA
  5. Advantages of IMA
  6. Challenges of Implementing IMA
  7. Conclusion
  8. References

What is the Internal Models Approach (IMA)?#

The Internal Models Approach (IMA) is a regulatory framework that permits banks to use their proprietary risk models to calculate market risk capital requirements, subject to strict regulatory approval. Unlike the Standardized Approach—which uses fixed risk weights set by regulators—IMA leverages a bank's internal data, models, and risk management expertise to estimate risk more accurately.

IMA is primarily used for the trading book (assets held for short-term trading, e.g., equities, derivatives) and is governed by the Basel Committee on Banking Supervision (BCBS). To qualify for IMA, banks must demonstrate robust model governance, data quality, and risk management practices.

How Does IMA Work?#

IMA follows a structured process to ensure reliability and consistency:

  1. Data Collection & Risk Factor Identification: Banks gather historical market data (e.g., prices, volatilities) and identify key risk factors (e.g., interest rates, equity indices) relevant to their trading portfolio.
  2. Model Development: Banks develop internal models to estimate risk measures (e.g., Expected Shortfall) using statistical techniques (e.g., Monte Carlo simulations, historical simulation).
  3. Backtesting & Calibration: Models are tested against historical data to ensure accuracy. For example, predicted losses (from ES) are compared to actual losses to validate performance.
  4. Regulatory Approval: Regulators (e.g., the Federal Reserve, EBA) review the model's design, backtesting results, and governance to approve its use for capital calculations.
  5. Capital Calculation: Once approved, the bank uses the model to compute daily market risk capital, which is reported to regulators.

Key Components of IMA#

IMA relies on several core components to ensure robust risk measurement:

3.1 Expected Shortfall (ES)#

Under FRTB, IMA replaces the traditional Value-at-Risk (VaR) with Expected Shortfall (ES) as the primary risk measure. ES estimates the average loss that occurs beyond a specified confidence level (e.g., 97.5%) over a 10-day horizon. Unlike VaR, which only measures the threshold of loss, ES captures the severity of extreme losses, making it more sensitive to tail risk.

Example: A 97.5% ES of 10millionmeansthat,in2.510 million means that, in 2.5% of cases (i.e., extreme market conditions), the average loss over 10 days will be at least 10 million.

3.2 Stressed Expected Shortfall (Stressed ES)#

To account for severe market conditions, IMA requires banks to calculate Stressed ES, which uses a 12-month historical stress period (e.g., the 2008 financial crisis) to estimate ES. This ensures capital requirements reflect losses during periods of market turmoil, preventing undercapitalization in crises.

3.3 P&L Attribution Test#

The Profit & Loss (P&L) Attribution Test validates that the model accurately captures the drivers of a bank's trading P&L. Banks compare:

  • Model P&L (HPL): P&L predicted by the risk model using hypothetical positions.
  • Actual P&L (RTPL): Realized P&L from trading activities.

Under FRTB, the P&L Attribution Test uses two statistical metrics:

  • Spearman Correlation Test: Measures whether the daily ranks of the risk model P&L and desk P&L are highly correlated. A low correlation indicates the risk model may not adequately capture the desk's risk drivers.
  • Kolmogorov-Smirnov (KS) Test: Compares the statistical distributions of the two P&L series to assess whether they are sufficiently similar.

Failing either test results in the trading desk being excluded from IMA and required to use the standardized approach instead. The tests are conducted quarterly using the most recent 250 trading days of data.

3.4 Model Validation#

Regulators mandate independent model validation to ensure models are reliable. This includes:

  • Backtesting: Comparing model predictions to actual outcomes to verify accuracy.
  • Benchmarking: Testing the model against alternative approaches (e.g., standardized methods).
  • Ongoing Monitoring: Regular reviews to adapt to changing market conditions or portfolio composition.

FRTB Updates to IMA#

The Fundamental Review of the Trading Book (FRTB), finalized by BCBS in January 2016 (BCBS 352), introduced critical updates to IMA to address weaknesses exposed during the 2008 financial crisis. Implementation has been repeatedly delayed across major jurisdictions; the EU and UK now target January 2027 for most FRTB provisions, with the IMA component specifically delayed to January 2028 to allow time for international coordination.

Key FRTB updates to IMA include:

  • Trading Book/Banking Book Boundary: Stricter rules for classifying assets as trading book (eligible for IMA) vs. banking book (not eligible), reducing regulatory arbitrage.
  • Expected Shortfall Replacing VaR: FRTB replaced Value-at-Risk (VaR) with Expected Shortfall (ES) as the primary risk measure to better capture tail risk.
  • Risk Factor Eligibility Test (RFET): Risk factors must pass both a quantitative test (requiring at least 24 verifiable prices over the preceding 12 months with no 90-day gap exceeding one month) and qualitative data standards to be included in the ES model. Risk factors that fail the RFET are classified as Non-Modellable Risk Factors (NMRFs).
  • Non-Modellable Risk Factors (NMRFs): Risk factors with sufficient data to be included in the ES model but still have limited liquidity are assigned longer liquidity horizons, increasing capital requirements. Under recent UK PRA proposals (June 2026), NMRFs are further divided into Type 1 (passing qualitative but not quantitative standards, included in ES with a capital add-on) and Type 2 (failing both tests, fully excluded from ES).
  • Liquidity Horizons: FRTB assigns specific liquidity horizons (ranging from 10 to 120 days) to different asset classes, reflecting the time needed to exit positions without significant market impact. Less liquid instruments receive longer horizons, increasing capital requirements.
  • P&L Attribution Test (PLAT): Banks must demonstrate that their risk model P&L closely aligns with actual desk P&L using two statistical tests: the Spearman correlation test (measuring rank correlation between risk and desk P&L) and the Kolmogorov-Smirnov test (comparing the statistical distributions of the two P&L series). Failing either test results in the desk being excluded from IMA.
  • Backtesting Requirements: Banks must backtest their models by comparing predicted losses to actual outcomes, with specific thresholds for exception rates that trigger supervisory action.
  • Stricter Approval Criteria: Regulators now require banks to demonstrate ES is "reliable" through rigorous backtesting, stress testing, and ongoing model validation.

Advantages of IMA#

  • Risk Sensitivity: IMA tailors capital requirements to a bank's specific portfolio, reflecting unique risk exposures better than one-size-fits-all standardized methods.
  • Alignment with Internal Risk Management: Banks already use internal models for risk management; IMA aligns regulatory capital with these practices, reducing redundancy.
  • Potential for Lower Capital: Accurate models may result in lower capital requirements compared to the standardized approach, improving profitability.

Challenges of Implementing IMA#

  • Complexity & Cost: Developing, validating, and maintaining IMA models requires significant investment in data infrastructure, technology, and expertise.
  • Regulatory Scrutiny: Approval is rigorous; banks must navigate strict documentation, backtesting, and governance requirements.
  • Model Risk: Inaccurate models (e.g., due to poor data or flawed assumptions) can lead to undercapitalization, exposing banks to losses.
  • NMRFs: Illiquid assets with limited data require manual capital add-ons, increasing operational burden.

Conclusion#

The Internal Models Approach (IMA) is a cornerstone of modern market risk regulation, enabling banks to use advanced models to calculate capital requirements more accurately. While FRTB has tightened standards—emphasizing ES, stress testing, and stricter governance—IMA remains a powerful tool for banks with sophisticated risk management capabilities. By balancing flexibility with regulatory rigor, IMA helps ensure financial stability while allowing banks to align capital with their unique risk profiles.

References#