01 · Buffers, CCyB & MDA

Art 128–142 CRD

Purpose & history

Capital you are allowed to lose

Minima are cliff edges: breach them and the supervisory machinery of last resort starts. Buffers exist to put loss-absorbing distance above the cliff — capital a bank may consume in stress with consequences that are commercial (distribution caps via the MDA) rather than existential. Each layer answers a different question. The CCoB (2.5%, fixed) is the all-weather cushion. The CCyB is the macroprudential experiment: built when credit runs hot, released when stress arrives — the only buffer designed to go to zero on purpose, and the reason several authorities (the Netherlands at 2% among them) now hold a "positive neutral" rate in calm times: you cannot release what you never built. G-SII/O-SII buffers price systemic footprint (the higher of the two applies); the SyRB covers structural risks the others miss and stacks on top — in the Netherlands used sectorally against RRE concentration.

The unresolved debate is usability: in the pandemic, banks proved deeply reluctant to dip into buffers even when invited — MDA stigma and market expectations made nominally usable capital behave like a minimum. The macroprudential review keeps circling the obvious conclusion: more of the stack should be releasable (CCyB-like), less of it fixed.

2010

Basel III invents the buffer concept: CCoB + CCyB, MDA as enforcement.

2014–16

EU implementation; systemic buffers (G-SII/O-SII/SyRB) layer in via CRD.

2020

Pandemic: buffers released — and banks barely use the room. The usability problem becomes empirical.

2021–

Positive-neutral CCyB wave: NL, and a growing list, set non-zero rates in normal conditions.

2025

Single-stack floor: every buffer % now multiplies floored TREA — the layers rise with the floor.

Requirements

Geographic CCyB engineRelevant credit exposures mapped by obligor country; each country's announced rate (and 12-month lead times) tracked; the institution-specific rate recomputed quarterly as the exposure-weighted average.

Stack arithmeticCBR assembled per level of application: CCoB + institution-specific CCyB + max(G-SII, O-SII) + SyRB (additive, incl. sectoral); combined with P2R-CET1 into the MDA trigger point.

MDA readinessDistributable-items and maximum-distributable-amount computation ready daily in stress; board distribution policy pre-wired to buffer positions.

Capital planning above the CBRManagement buffer sized over CBR + P2G; stress paths showing which layer absorbs first and where distributions stop.

DisclosureCountercyclical buffer geographic breakdown and buffer requirements in Pillar 3; the market reads distance-to-MDA before it reads anything else.

Calculations

Institution-specific CCyB — exposure-weighted Relevant exposures: NL 60% (rate 2.0%) · DE 30% (0.75%) · US 10% (0%)
CCyB = 0.6×2.0 + 0.3×0.75 + 0.1×0 = 1.425%

The full CBR CCoB 2.5 + CCyB 1.425 + max(G-SII —, O-SII 2.0) + SyRB 0 = CBR = 5.925%
MDA trigger = 4.5 + P2R-CET1 1.125 + 5.925 = 11.55% CET1

What release is worth Stress: NL cuts the CCyB to 0 → institution rate falls 1.2 points → on TREA €72bn, ≈ €0.86bn of CET1 headroom appears overnight, with no market signal of weakness — the entire argument for building the buffer in good times.

Floor interaction Same 5.925% CBR on TREA €60bn → €3.6bn; on floored €72.5bn → €4.3bn. The buffer percentages never moved; the floor moved what they multiply.

Changes & sources

Topic	State
Positive-neutral CCyB	Spreading practice, not EU law — rates are national decisions with 12-month implementation lags
SyRB	Sectoral use (RRE) established; overlap control with Art 458 measures
Usability	Macroprudential review's standing item: shift weight from fixed to releasable

Sources: Art 128–142 CRD; ESRB rate database for CCyB rates and reciprocity; EBA MDA Q&As.
Watch: any move to make part of the CCoB releasable — the single biggest potential redesign of the stack.

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