Output Floor: How Basel IV Limits Internal Model Advantages

The output floor is a regulatory constraint introduced under Basel IV that limits how much capital relief a bank can claim by using sophisticated internal risk models. Specifically, a bank’s total capital requirement under internal ratings-based (IRB) models cannot fall below 72.5 percent of what it would be under the standardised approach—a simpler, regulator-defined formula. By capping the advantage of complex models, the output floor addresses a two-decade gap between theory and practice: banks with the most advanced risk-measurement systems were often the ones with the lowest measured capital needs, which regulators found at odds with financial stability.

Why the Output Floor Exists

For nearly two decades before Basel IV, banking regulators had tolerated a paradox: the banks with the most advanced risk-management systems—the ones most sophisticated enough to model credit, market, and operational risk—often ended up with the lowest measured capital needs. A global bank using IRB models for credit risk and advanced measurement approach (AMA) models for operational risk could reduce its capital requirement to a fraction of what the standardised approach would demand.

Regulators accepted this because the idea was sound: better risk measurement should allow lower capital to cover the same level of protection. But two problems emerged:

1. Model gaming — Banks optimized models and definitions to minimize measured risk in ways that didn’t align with actual loss experience. A slight change to a default-probability curve or a loss-given-default assumption could reduce the capital requirement millions of dollars, and regulators struggled to distinguish true refinement from hidden arbitrage.

2. Procyclicality — During boom times, internal models showed very low risk because recent loss history was benign. This allowed banks to reduce capital precisely when they were taking on the most tail risk. When downturns arrived, models suddenly flagged high risk, forcing rapid capital reduction—exactly the opposite of what financial stability needs.

The 2008 financial crisis crystallized the regulators’ concern. Banks like Lehman and Bear Stearns had sophisticated risk models; the models did not prevent disaster. By the time Basel IV was finalized (2017–2020), the global consensus was clear: internal models alone could not be trusted as the sole basis for capital requirements.

How the 72.5 Percent Floor Works

Under Basel IV, a bank’s total risk-weighted assets (RWA) and resulting Tier 1 capital requirement are calculated on two tracks:

Track 1: Fully modeled approach — Use internal ratings-based models for credit risk, advanced measurement approach for operational risk, and standardised rules for market risk. Calculate total RWA and capital requirement.

Track 2: Standardised approach — Use regulatory default formulas for all risks. Calculate total RWA and capital requirement.

The bank must then hold capital equal to whichever is higher: the fully modeled Track 1 result, or 72.5 percent of the standardised Track 2 result.

Practical example:

• Standardised approach produces RWA of €1,000 million, requiring €85 million Tier 1 capital (at 8.5 percent).
• IRB models produce RWA of €500 million, requiring €42.5 million.
• Output floor = 72.5% × €85 million = €61.6 million.
• Bank must hold at least €61.6 million (the floor), not the €42.5 million its models suggest.

In this case, the bank gets a capital relief of 27.5 percent below the standardised approach—but no more.

Impact on Bank Capital Adequacy

The output floor has forced many large banks to recalculate capital requirements upward. The magnitude of the increase depends on how aggressive their internal models were:

• Banks that conservatively calibrated models saw little change.
• Banks that had heavily optimized models for minimum capital saw material increases—sometimes 10–20 percent more RWA than models alone would predict.

This reallocation serves two goals: it reduces the incentive for models to be gamed, and it converges measured risk across banks using different approaches, making the banking system more comparable and less vulnerable to model-based arbitrage.

The Standardised Approach as a Backstop

The output floor elevates the standardised approach from a default for unsophisticated banks to a binding constraint for all. This creates competition: if a bank’s elaborate IRB model only yields a 27.5 percent discount to the standardised floor, the bank may decide the cost of maintaining the model does not justify the marginal capital relief. Some banks have begun migrating back to the standardised approach, especially in less complex portfolios.

Regulators view this as a feature, not a bug. Simpler, more transparent standardised rules are easier for regulators to supervise and for investors to understand.

Transition and Phasing

Basel IV did not require all banks to comply with the output floor on January 1, 2023. Many jurisdictions, particularly the US, negotiated extended transition periods:

• January 2023 — Hard deadline for most non-US banks and some US firms.
• 2025 onwards — Phased increases in the US floor (starting at lower percentages, stepping up year-by-year).
• EU — Varied by member state; some adopted January 2023, others negotiated delays.

The transition reflected political reality: an abrupt 20–30 percent increase in capital requirements would have disrupted lending markets. Gradual increases allowed banks time to raise capital or optimize portfolios.

Interaction with Stress Testing

In the US, the output floor complicates the relationship between annual stress tests and capital adequacy. A bank’s “binding constraint” might be the stress-test requirement, the output-floor requirement, or a minimum regulatory ratio like the leverage ratio. Banks now monitor all three, and the floor can unexpectedly become the tightest constraint if internal models improve (paradoxically, better models don’t lower capital if the floor is binding).

Remaining Controversy

Critics of the output floor argue that:

• It weakens the incentive for banks to invest in better risk measurement, because the benefit is capped.
• It imposes a one-size-fits-all ceiling even on banks whose models have proven predictive and conservative.
• It may push risk into less-regulated institutions if banks reduce lending due to higher capital requirements.

Supporters counter that:

• The output floor is necessary to preserve confidence in Basel standards after repeated instances of model-based arbitrage.
• It shifts focus from regulatory capital optimization back to genuine risk management.
• Phased transition allows markets to absorb the change without shock.

Globally, the Basel Committee continues to refine the floor in light of implementation experience.

See also