Elderlake Partners

01 · Leverage ratio

Art 429–430 CRR

Purpose & history

The measure that cannot be modelled down

Pre-2008 balance sheets ran 40–50× equity while reporting comfortable risk-based ratios — low risk weights, not low risk, did the work. The leverage ratio is the deliberate antidote: Tier 1 over a total exposure measure, no weights anywhere. Its virtue and its vice are the same property: it prices a bund and a bridge loan identically. That is why it is a backstop, calibrated (3%) to bind only when risk weights fall low enough to be suspicious — and why for low-RWA business models (Dutch mortgage banks pre-floor) it was for years the actual binding constraint.

The exposure measure is accounting-plus: derivatives enter via a modified SA-CCR (collateral recognition restricted — received collateral cannot shrink the balance sheet), SFTs add a counterparty add-on, and off-balance items convert at their credit CCFs with a 10% floor. G-SIIs carry a leverage buffer of 50% of their G-SII buffer, with its own MDA analogue.

2008
40–50× leverage with sound risk-based ratios — the exhibit for the backstop.
2010
Basel III introduces the ratio; disclosure-only at first.
2021
Binding 3% in the EU (CRR2); G-SII add-on follows.
2023–
Window-dressing response: averaging requirements for SFT exposures at large banks.
2025 ff
As the output floor lifts TREA, banks flip from leverage-bound to floor-bound — the two backstops trade places.

Requirements

LREM productionOn-balance carrying values gross of most netting; derivatives per leverage-specific SA-CCR rules (cash VM offset only under strict conditions); SFT add-ons; off-balance at CCFs floored at 10%.
Averaging & window-dressingDaily averaging of SFT components for large institutions; quarter-end optics are now themselves a supervisory topic.
Central-bank reserve treatmentExclusion is a discretion with a recalibration price — if invoked, the 3% is adjusted up; track the national position.
G-SII leverage buffer50% of the G-SII buffer, with distribution restrictions on breach (L-MDA) — a second ladder to monitor.
Dual-constraint planningEvery capital action tested against risk-based and leverage stacks simultaneously; the binding constraint labels the marginal cost of every asset.

Calculations

Building the exposure measure On-balance €240bn + derivatives (mod. SA-CCR) €6bn + SFT add-on €4bn + off-balance (CCF, floor 10%) €10bn = LREM €260bn
T1 €11bn → leverage ratio = 4.2% vs 3% requirement.
G-SII add-on G-SII buffer 1.5% → leverage buffer 0.75% → requirement 3.75%; at 4.2% the headroom is €260bn × 0.45% ≈ €1.2bn of T1 — thinner than the risk-based headroom suggests.
Which backstop binds — the density rule Leverage binds when RWA density (TREA/LREM) < leverage req. / T1 req.
Minima only: 3% / 6% → density < 50%
Realistic full T1 stack ≈ 11%: 3% / 11% → density < ≈27%
Dutch mortgage banks pre-floor: density in the low 20s → leverage-bound. The floor pushes density up through ~30% → the constraint flips to the floor. Same balance sheet, new binding rule, different marginal pricing.

Changes & sources

  • Sources: Art 429–430 CRR; leverage disclosure templates; BCBS LEV standards for lineage.
  • CRR3 left the ratio's design alone — its role changes anyway, because the floor re-sorts who it binds.
  • Watch: the US SLR recalibration debate — transatlantic daylight on how hard the backstop should bite dealers and Treasury markets.

Independent analysis by Elderlake Partners, as at August 2026. Not legal, accounting or supervisory advice.