Expected credit loss staging and significant increase in credit risk (SICR)

ECL is usually the single largest and most volatile judgment-driven line in a bank's income statement: staging determines whether a loan carries a 12-month or lifetime loss allowance, and the difference between Stage 1 and Stage 2 for the same exposure can be an order of magnitude. Because SICR is assessed relatively (change in lifetime probability of default since origination) rather than against an absolute credit quality bar, two banks holding the same loan can reach different staging conclusions, so the assumptions and triggers used directly affect comparability, capital planning, and dividend capacity.

Is the significant-increase-in-credit-risk assessment comparing the current lifetime PD to the PD expected at origination for the same instrument, rather than to an absolute risk grade? Are quantitative triggers (the 30-day past-due backstop), qualitative triggers (watchlist, covenant breach, forbearance), and multiple forward-looking macroeconomic scenarios all being combined, rather than relying on any single indicator? For loans moving back to Stage 1, is there evidence of sustained improvement over a defined period rather than an immediate release on first sign of cure? Is Stage 3 classification supported by objective evidence of impairment (default, distress-driven concession) assessed separately from the SICR test?

1. At each reporting date, compare the current lifetime probability of default (PD) to the PD expected at origination for the same instrument, not to an absolute risk grade.
2. Apply the rebuttable presumption that risk has increased significantly once contractual payments are more than 30 days past due, but do not rely on this backstop alone.
3. Overlay qualitative triggers: watchlist status, covenant breaches, forbearance flags, sector or macro deterioration specific to the borrower.
4. Incorporate forward-looking information using multiple macroeconomic scenarios (base, upside, downside) rather than a single point forecast.
5. Move exposures from Stage 2 back to Stage 1 (cure) only after observing sustained improvement, applying the same rigor as for deterioration.
6. For Stage 3, confirm objective evidence of credit impairment (default, significant financial difficulty, concession granted due to distress) separately from the SICR test.

Measure Stage 1 ECL as the 12-month expected credit loss (portion of lifetime ECL from default events possible within 12 months); measure Stage 2 and Stage 3 ECL as lifetime expected credit loss. Use probability-weighted outcomes across at least a base, upside, and downside macroeconomic scenario rather than a single deterministic path, and discount expected shortfalls at the original effective interest rate. Incorporate collateral value (net of realization costs and haircuts) into loss-given-default. Reassess PD, LGD, and EAD parameters at least annually and whenever macro conditions or portfolio composition shift materially.

Disclose: (1) reconciliation of the loss allowance and gross carrying amount by stage, showing transfers between stages; (2) inputs, assumptions, and estimation techniques (PD, LGD, EAD models, macroeconomic scenarios and their weightings); (3) sensitivity of the ECL to reasonably possible changes in key assumptions; (4) the definition of default and SICR criteria used, including the 30-days-past-due backstop and any rebuttal; (5) collateral held and its effect on ECL; (6) write-offs and recoveries during the period.

• Relying solely on the 30-day past-due backstop as the SICR trigger while ignoring qualitative and forward-looking indicators
• failing to reassess and re-stage a modified loan once it cures, or conversely releasing it to Stage 1 too early without evidence of sustained performance
• applying a single deterministic macroeconomic scenario instead of probability-weighted scenarios
• using origination-date risk grade as a static SICR benchmark without updating it for portfolio-wide changes in origination standards
• double-counting collateral benefit already reflected in the loan'
• s contractual terms.