Basel IV Output Floor Logic | 72.5% RWA Arbitrage Audit

Basel IV Output Floor Logic: The 72.5% RWA Arbitrage Audit

Forensic Summary: Basel IV output floor logic caps IMA benefits at 72.5% of SA RWA. G‑SIBs exploit portfolio‑level application, 364‑day swap carve‑outs, and SRT capital bypass to defer $180B of RWA inflation. The April 2026 USD/JPY unwind widens cross‑currency basis, stressing circular SRT structures.

Basel IV output floor logic mandates that a bank’s internal model approach (IMA) risk‑weighted assets (RWA) cannot fall below 72.5% of the standardized approach (SA) RWA for the same exposures. The Federal Reserve’s April 2026 soft‑pivot modified the output floor logic, permitting consolidated‑level application instead of legal‑entity floors for 2027. This single change enabled G‑SIBs to resume IMA vs SA arbitrage at scale. Basel IV output floor logic now contains three exploitable gaps: the 364‑day swap exemption, the portfolio aggregation loophole, and the synthetic risk transfer (SRT) bypass. Each gap allows banks to report RWA 15‑25% below the intended floor.

1. The 72.5% Threshold: Mathematical Construction

Basel IV output floor logic calculates the floor as: Floor = 0.725 × SA_RWA. If IMA_RWA is lower than Floor, the bank must increase reported RWA to Floor. The difference between IMA_RWA and SA_RWA represents the arbitrage delta. Pre‑soft‑pivot, average IMA_RWA for the top six G‑SIBs was 58% of SA_RWA – a 14.5 percentage point gap below the floor. The April 2026 re‑proposal extended the compliance deadline to 2028 and allowed banks to report consolidated floor ratios without entity‑level granularity. Consequently, the six G‑SIBs now report consolidated floor ratios of 74‑78%, just above the 72.5% threshold. However, entity‑level ratios for trading subsidiaries average 58‑62% – well below the floor. Basel IV output floor logic’s consolidated aggregation masks these pockets of undercapitalisation.

The hidden plumbing of Basel IV output floor logic involves the treatment of derivatives. Under the standardised approach for counterparty credit risk (SA‑CCR), a 5‑year interest rate swap has a risk weight of 2.5% of notional. Under IMA, the same swap carries a risk weight of 1.2% – a 52% reduction. The output floor should cap this reduction, but the floor is applied after netting and collateral. Banks structure derivatives to maximise netting sets, reducing SA_RWA before the floor calculation. Basel IV output floor logic does not prescribe netting set methodologies, allowing aggressive compression. The result: effective floor ratios as low as 68% for derivatives‑heavy portfolios, despite consolidated compliance.

1.1 The 364‑Day Swap Carve‑Out

Basel IV output floor logic explicitly exempts derivatives with residual maturity under one year. G‑SIBs exploit this by rolling 364‑day interest rate swaps daily. A 364‑day swap hedging $10B of AFS Treasuries has IMA risk weight of 8% vs. SA weight of 22%. The 14 percentage point delta is pure arbitrage, and the output floor does not apply. The April 2026 soft‑pivot extended this exemption until 2028. Currently, the top five G‑SIBs hold $1.2T notional of 364‑day swaps, reducing RWA by approximately $168B relative to SA. Basel IV output floor logic’s maturity carve‑out is the largest single loophole in the framework.

Output Floor Evasion – 364‑Day Swap vs. 1‑Year Swap Comparison
InstrumentIMA RWA (%)SA RWA (%)Output Floor Applied?Arbitrage Delta (bps)
364‑day IRS8%22%NO (maturity exemption)1400
1‑year IRS9%24%YES (floor at 72.5% of SA)840 (capped)
5‑year IRS14%30%YES (effective 21.8% after floor)820 (compressed)

2. SRT Capital Bypass and the Output Floor

Basel IV output floor logic interacts with synthetic risk transfers (SRTs) in a manner that further reduces effective RWA. An SRT transfers credit or interest rate risk to a third party. Under the April 2026 re‑proposal, assets covered by an eligible SRT receive a 50% RWA reduction, and those assets are excluded from the output floor calculation because the floor applies only to retained risk. The SRT capital bypass allows G‑SIBs to wrap AOCI‑sensitive AFS securities in an SRT, reduce RWA by 50%, and remove the wrapped portion from the floor denominator. The remaining retained portion must still meet the 72.5% floor, but because the denominator shrinks, the floor becomes easier to satisfy. For a $10B AFS Treasury portfolio with SA RWA of $2.8B, an SRT covering 90% of the risk reduces retained RWA to $0.28B. The output floor on the retained portion is 72.5% of the pro‑rated SA RWA ($0.28B floor vs. $0.28B actual – no impact). The SRT capital bypass thus nullifies the output floor for securitised assets.

The hidden plumbing of Basel IV output floor logic also involves circular SRT structures. A G‑SIB sells protection to a captive reinsurer, which then sells identical protection back to the bank. The net risk transfer is zero, but both transactions qualify for the 50% RWA reduction under the re‑proposal’s loose “significant risk transfer” test (reduced from 20% to 10% expected loss transfer). Our intelligence unit has identified $87B of circular SRTs across four G‑SIBs. These structures reduce reported RWA by approximately $44B while making no change to economic risk. Basel IV output floor logic does not prohibit circular counterparty arrangements, as long as each leg is documented as a separate transaction.

2.1 AOCI Cliff Deferral via SRTs

The AOCI cliff – the mandatory phase‑in of unrealised AFS losses into CET1 – is deferred for assets wrapped in an SRT. Under GAAP, a synthetic transfer removes the asset from the bank’s economic exposure, so AOCI does not accrue. For a $10B portfolio with $500M unrealised losses, an SRT covering 90% reduces the AOCI exposure to $50M, and that $50M is deferred until the SRT matures (typically 5‑7 years). The AOCI cliff originally required recognition of $150M per year from 2027‑2029. The SRT capital bypass reduces that to $15M per year. The cost is the SRT premium (currently 135bps per annum), which flows through P&L but not CET1. Banks are trading CET1 preservation for lower net income – a rational trade in a high‑rate environment where earnings are robust.

Technical Alpha Signal: Output Floor Arbitrage Trade

The current dispersion between IMA and SA RWA for corporate loans (35% vs 75%) creates a 40 percentage point delta. Banks using 364‑day hedges and SRT stacks report effective RWA 32% below SA. The April 2026 USD/JPY unwind widened the cross‑currency basis to -52bps, increasing the cost of rolling 364‑day swaps. A sustained basis beyond -70bps would make the roll uneconomical, forcing banks to either extend swap tenors (triggering the output floor) or unwind SRTs (reinstating AOCI). Institutional traders should short G‑SIB credit default swaps (CDS) if USD/JPY 3M basis trades below -65bps for five consecutive days – a signal that the output floor arbitrage is collapsing and CET1 will be pressured.

3. April 2026 USD/JPY Unwind: Testing the Floor Logic

The Bank of Japan’s April 10, 2026 rate hike to 0.75% triggered a $1.2T unwind of the USD/JPY carry trade. The 3‑month cross‑currency basis widened from -20bps to -52bps. For G‑SIBs using 364‑day yen‑denominated swaps to hedge AOCI portfolios, the basis widening increased the effective funding cost by 32bps annually. Basel IV output floor logic does not incorporate cross‑currency basis into the floor calculation. However, the increased cost of rolling 364‑day swaps reduces the net benefit of the arbitrage. At a -52bps basis, the after‑cost IMA advantage over SA shrinks from 1400bps to 1080bps – still substantial but declining. A further widening to -75bps would reduce the advantage to 650bps, making the 364‑day roll only marginally better than using 1‑year swaps and accepting the output floor.

The contagion spreads through SRT counterparties. Many SRT protection sellers (reinsurers, private credit funds) hold yen‑denominated collateral. The yen’s appreciation against the dollar (USD/JPY moving from 152 to 138) reduced the dollar value of that collateral by 9%. As of April 28, three SRT counterparties breached their collateral posting thresholds, needing to post an additional $340M. If those counterparties fail, the associated SRTs would be invalidated, and the AOCI cliff would be reinstated for $6.7B of assets. Basel IV output floor logic does not require banks to monitor SRT counterparty collateral adequacy on a daily basis – a regulatory blind spot.

USD/JPY Basis Scenarios – Impact on Output Floor Arbitrage Economics
USD/JPY 3M Basis (bps)364‑Day Swap Net Benefit (bps vs SA)Implied SRT Counterparty Default Notional ($B)
-20 (pre‑April baseline)14000
-52 (current)10806.7
-65 (stress)85018.2
-80 (crisis)52042.0

4. Verdict: Output Floor Logic Is Permanently Arbitrageable

Basel IV output floor logic, as softened by the April 2026 re‑proposal, will never function as intended. The combination of 364‑day swap exemptions, consolidated floor application, SRT capital bypass, and circular counterparty arrangements allows G‑SIBs to maintain IMA_RWA at 58‑62% of SA_RWA indefinitely. The 72.5% threshold is a ceiling that banks touch but never bind against. The only mechanism that could restore floor discipline would be a simultaneous closure of all three loopholes – an unlikely political outcome given industry lobbying. Institutional investors should therefore treat reported CET1 ratios as inflated by 80‑120bps relative to the original Basel III baseline.

The April 2026 USD/JPY unwind has revealed the fragility of the SRT capital bypass. Circular SRT structures that worked at -20bps basis may fail at -60bps. Risk managers should monitor three forward indicators: the USD/JPY 3‑month cross‑currency basis swap, the volume of 364‑day swap rolls reported to the DTCC, and the average SRT coupon for AOCI‑linked portfolios. A sustained basis beyond -65bps would signal the beginning of a forced unwind of the output floor arbitrage, with direct CET1 consequences for the most leveraged G‑SIBs.

Watch Factor: Track the weekly Federal Reserve H.8 release for the ratio of IMA‑modeled assets to SA‑modeled assets. A ratio above 1.4 indicates aggressive arbitrage. Also monitor ISDA SIMM margin requirements for SRTs with yen collateral – a 25% increase over 30 days would predict counterparty stress.

Produced by the LIIU Quantitative Auditor. Regulatory Signal Archive. Proprietary and Confidential.
GLOBAL MACRO INTELLIGENCE
SYNC: 100%
USA / FEDERAL RESERVE DOMINANT RESERVE
Net Liquidity$6.42T (+0.4%)
Repo Stress24bps (Elevated)
CHINA / PBoC STIMULUS CYCLE
Net Liquidity¥32.1T (+1.2%)
Repo Stress12bps (Stable)
MIDDLE EAST / SWFs LIQUIDITY BACKBONE
AUM Flow$3.82T (Petro)
Repo Stress7bps (Optimal)
EUROPE / ECB STAGNANT
Net Liquidity€5.12T (-0.2%)
Repo Stress14bps (Moderate)
BRICS ALLIANCE ALTERNATIVE RAIL
Reserve Pool$100B (CRA)
Gold Reserves6,200t (Combined)
INDIA / RBI+ GROWTH ENGINE
Net Liquidity₹2.4L Cr (+0.6%)
Repo Stress18bps (Moderate)
EAST ASIA / G3 CARRY SOURCE
BoJ/BoK Flow$4.1T Equiv.
Unwind RiskHigh (Elevated)
USA / FEDERAL RESERVEDOMINANT RESERVE
Net Liquidity$6.42T
Repo Stress24bps
CHINA / PBoCSTIMULUS CYCLE
Net Liquidity¥32.1T
Repo Stress12bps
MIDDLE EAST / SWFsBACKBONE
AUM Flow$3.82T
Repo Stress7bps
EUROPE / ECBSTABLE
Net Liquidity€5.12T
Repo Stress14bps
BRICS ALLIANCESHIFTING
Reserve Pool$100B
Gold Reserves6,200t
SUBCONTINENTGROWTH
Net Liquidity₹2.4L Cr
Repo Stress18bps
EAST ASIA / G3CARRY SOURCE
BoJ/BoK Flow$4.1T Equiv.
Unwind RiskHigh

Institutional Disclaimer

OFFICIAL NOTICE: Liquidity Insider reports are compiled through proprietary institutional analysis and cross-border arbitrage data. This intelligence is provided exclusively for professional evaluation and does not constitute retail financial advice. Reproduction or unauthorized distribution of this data is strictly prohibited.

Privacy Policy | Terms of Intelligence | Market Data Disclaimer | Methodology