The primary goal of a Significant Risk Transfer (SRT) transaction is to reduce a bank's regulatory capital requirement. This is achieved by lowering the Risk-Weighted Assets (RWA) associated with a specific pool of loans. The benefit, known as capital relief, is the difference between the capital required before and after the transaction.

Before an SRT, a bank holds capital against the full RWA of the loan portfolio. After transferring the credit risk of the junior and mezzanine tranches to investors, the bank is only required to hold capital against the senior tranche it retains. Since the senior tranche is the safest, its RWA is significantly lower than that of the entire original portfolio.

The effectiveness of this relief depends on how RWA is calculated for the securitised tranches. Regulators provide a strict hierarchy of methods for banks to use.

Under the Basel framework, there are specific methods to calculate the RWA for securitisation exposures. The two most common are the Securitisation Internal Ratings-Based Approach (SEC-IRBA) and the Securitisation Standardised Approach (SEC-SA).

There's a clear pecking order. Banks with approved internal models for the underlying loan types must use SEC-IRBA. This approach allows them to use their own risk parameters (like probability of default) to calculate capital, which often results in a more risk-sensitive and lower RWA figure. Banks that cannot use SEC-IRBA must use SEC-SA, which relies on prescribed risk weights set by the regulator. A third method, the Securitisation External Ratings-Based Approach (SEC-ERBA), can be used if the tranches have an external credit rating, but it's less common in the SRT market.

The choice of method directly influences the potential RWA relief and, therefore, the economic viability of the deal. Banks generally prefer SEC-IRBA as it better reflects the underlying risk and can lead to greater capital efficiency.

The finalisation of Basel III rules, often called Basel 3.1, introduces a major constraint: the 'output floor'. This rule is designed to reduce the variability in RWA calculations between banks and limit the capital benefit of using internal models.

The output floor sets a lower limit on the RWA a bank can report. A bank's total RWA calculated using its internal models cannot be less than 72.5% of the total RWA calculated using only the standardised approaches. In essence, it creates a 'floor' below which the benefits of internal models cannot fall.

This has a direct impact on SRTs. If a bank has a portfolio of very low-risk loans, its internal models might calculate a very low RWA. However, the output floor might force the bank to report a higher RWA, reducing the pre-transaction RWA value. This shrinks the potential $RWA_{Relief}$ that can be achieved, making an SRT less attractive or even uneconomical.

This dynamic forces banks to be more selective about which portfolios they securitise. The key metric becomes the RWA density of the portfolio, which is the RWA divided by the total loan amount. Portfolios with high RWA density, such as loans to corporations or credit card receivables, remain good candidates for SRTs because there is a substantial amount of RWA to transfer. In contrast, portfolios with very low RWA density, like prime residential mortgages, may see their SRT efficiency significantly reduced by the output floor.

Ultimately, the mechanics of regulatory capital are the engine of the SRT market. The interplay between internal models, standardised rules, and overarching constraints like the output floor determines which deals are possible and how they must be structured to achieve their purpose.