Reader question: A bank can look solvent on its balance sheet and still run out of usable cash. How does the Liquidity Coverage Ratio turn thousands of assets, deposits, derivatives and commitments into one test of whether the bank can survive a severe 30-day liquidity stress?
The Basel Liquidity Coverage Ratio, or LCR, is a rule-based stress algorithm. It does not ask whether every asset exceeds every liability in accounting value. It asks a narrower operational question:
Does the bank hold enough unencumbered high-quality liquid assets to cover its modeled net cash outflows over the next 30 calendar days under a prescribed stress scenario?
The core ratio is:
LCR = Stock of HQLA / Total net cash outflows over the next 30 calendar days.
Under the Basel standard, the ratio should be at least 100% on an ongoing basis, subject to supervisory use of the buffer during periods of stress.
The mathematics looks like a fraction. The real algorithm is the classification, haircut, cap, run-off, inflow and reconciliation machinery that constructs its numerator and denominator.
What this page owns — and what it does not
This page owns the public computational transformation:
bank balance-sheet and off-balance-sheet positions → HQLA stock + stressed 30-day cash flows → LCR.
It does not replace the reverse-stress-testing algorithm, which searches for scenarios that break a constraint; the collateral-optimisation algorithm, which allocates eligible assets across competing calls; or exact banking arithmetic, which controls precision and reconciliation.
This is prudential-liquidity mathematics and systems education. It is not a judgment about the safety of any particular bank and not personalized financial advice.
The ratio is a 30-day stress construction, not an ordinary cash forecast
The Basel Committee describes the LCR as a measure designed to promote short-term resilience by requiring a sufficient stock of unencumbered HQLA that can be converted into cash to meet liquidity needs during a 30-calendar-day stress scenario.
The denominator is therefore not “what management thinks customers will probably do next month.” Prescribed run-off and inflow factors intentionally impose stress assumptions. The result is a regulatory scenario measure, not a point forecast.
Step 1: build the HQLA candidate inventory
The numerator begins with assets that may qualify as high-quality liquid assets. The algorithm should not simply take every marketable security on the balance sheet.
An asset must pass the applicable Basel and jurisdictional eligibility criteria, including characteristics such as low credit and market risk, ease and certainty of valuation, low correlation with risky assets, listing or market structure where relevant, and a demonstrated ability to be monetised in stressed markets. Operational requirements also matter: HQLA must generally be under the control of the bank’s liquidity-management function and free of legal, regulatory or operational impediments to monetisation.
A Treasury security locked into a transaction that prevents its use during the stress window can be economically liquid but operationally unavailable. The LCR distinguishes those ideas.
Step 2: classify HQLA by level
The Basel HQLA framework separates qualifying assets into three practical groups:
- Level 1: highest-quality assets, generally not subject to the Basel LCR haircut and not limited as a share of HQLA;
- Level 2A: qualifying assets subject to a 15% haircut and included within the overall Level 2 cap;
- Level 2B: additional qualifying assets permitted by the relevant supervisor, subject to larger haircuts and tighter composition limits.
Under the Basel standard, Level 2 assets together may comprise no more than 40% of the HQLA stock after required adjustments, and Level 2B assets may comprise no more than 15%.
Haircuts convert market value into regulatory liquidity value
Suppose a bank holds 100 million of an eligible Level 2A security. With a 15% haircut:
LCR value = 100 × (1 − 0.15) = 85 million.
The haircut is not a forecast that the asset will lose exactly 15% in price. It is a prudential conversion factor that reduces how much liquidity value the asset contributes to the buffer.
Qualifying Level 2B assets receive larger Basel haircuts depending on asset type. For example, the Basel framework applies a 25% haircut to qualifying RMBS in the Level 2B category and 50% haircuts to certain qualifying corporate debt and common equity categories.
Composition caps create a second adjustment after haircuts
Haircutting each asset is not enough. The HQLA algorithm must also enforce the portfolio composition limits.
Consider a simplified post-haircut inventory:
- Level 1 = 60;
- Level 2A = 35;
- Level 2B = 20.
Raw post-haircut HQLA would be 115. But Level 2B cannot simply contribute all 20 because the 15% cap applies, and Level 2 as a whole cannot exceed 40% of the adjusted stock.
The Basel framework specifies cap calculations after haircuts and after taking account of the unwind of certain short-term securities-financing and collateral-swap transactions maturing within the 30-day period. A production implementation should use the formal Basel cap formulas rather than a naive “multiply total HQLA by 15%” shortcut, because the denominator itself changes when an excess category is removed.
Why the 40% and 15% caps are recursive-looking
Suppose Level 2B is too large. Removing excess Level 2B reduces total HQLA. That lower total can in turn change the amount of Level 2 assets permitted under the 40% cap.
This is why the Basel application guidance expresses the caps through adjusted amounts and maximum functions rather than an informal percentage check. The implementation problem is a constrained composition calculation.
Step 3: construct stressed cash outflows
The denominator begins with contractual and contingent cash outflows expected within 30 calendar days. Each category is multiplied by its prescribed run-off or drawdown factor.
Conceptually:
Expected stressed outflow = Exposure amount × prescribed outflow factor.
Retail deposits, small-business deposits, operational deposits, non-operational wholesale funding, secured funding, derivatives, committed facilities and contingent obligations can receive different treatments.
The factor depends on economic and behavioral features, not merely accounting labels.
Stable versus less-stable retail deposits
The Basel framework distinguishes qualifying stable retail deposits from less-stable deposits. The standard’s application guidance includes 5% and 10% baseline run-off categories for these classes, with national supervisors able to apply more conservative treatments where conditions warrant.
For a stylised example:
100 million stable deposits at a 5% run-off factor produce:
5 million stressed outflow.
100 million less-stable deposits at 10% produce:
10 million stressed outflow.
The algorithm therefore needs deposit attributes such as customer type, deposit insurance status and relationship characteristics. A single “customer deposits” balance is insufficient.
Wholesale funding is more sensitive to counterparty type and purpose
Unsecured wholesale funding from financial institutions can receive much higher run-off assumptions than relationship-based operational deposits. Funding from non-financial corporates, sovereigns and other legal entities has its own categories.
The mathematical lesson is that liquidity behavior is encoded through a classification tree:
liability → customer class → operational/non-operational → maturity/withdrawal characteristics → run-off factor.
An incorrect branch in that tree can be more material than a small arithmetic error.
Derivatives create liquidity outflows even when accounting exposure is small
Derivatives can require cash or collateral during stress because market values change, collateral calls rise, downgrade clauses activate or collateral substitution becomes more demanding.
The LCR therefore contains derivative-related outflow treatments beyond simple contractual payments. A derivatives portfolio that is close to zero market value today can still generate material 30-day liquidity needs.
This is one reason liquidity risk is not the same thing as counterparty-credit exposure.
Committed facilities are conditional cash claims
A bank may have promised undrawn credit or liquidity facilities. The customer has not yet drawn the cash, so there may be no funded loan asset today. Under stress, however, some portion of the commitment can be drawn.
The LCR applies prescribed drawdown assumptions to eligible off-balance-sheet commitments. The denominator therefore reaches beyond the accounting balance sheet.
Step 4: construct permitted cash inflows
Cash inflows expected within the 30-day horizon are also classified and multiplied by prescribed inflow factors.
Examples can include contractual receipts from performing loans, maturing placements and certain securities-financing transactions.
But the LCR deliberately does not allow a bank to assume that every expected receipt will arrive and fully fund every outflow.
The 75% inflow cap creates a minimum self-funded stress requirement
The Basel LCR generally caps recognised cash inflows at 75% of total expected cash outflows, subject to specified treatments and supervisory provisions.
In its simplest form:
Recognised inflows = min(calculated inflows, 0.75 × calculated outflows).
Then:
Net cash outflows = calculated outflows − recognised inflows.
This implies that, under the baseline cap, net cash outflows cannot fall below 25% of gross outflows merely because the bank expects large contractual receipts.
A numerical inflow-cap example
Suppose gross stressed outflows are 200 and calculated inflows are 190.
The 75% cap is:
0.75 × 200 = 150.
Recognised inflows = 150, not 190.
Therefore:
Net cash outflows = 200 − 150 = 50.
If HQLA is 55:
LCR = 55 / 50 = 110%.
A naive implementation using all 190 of inflows would produce net outflows of only 10 and an LCR of 550%, radically overstating the regulatory liquidity position.
A compact end-to-end example
Assume that after HQLA eligibility tests, haircuts, transaction-unwind adjustments and composition caps, the bank has 120 of recognised HQLA.
Its classified 30-day stress engine produces:
- retail deposit outflows = 20;
- wholesale funding outflows = 75;
- derivative/collateral outflows = 25;
- facility drawdowns and other outflows = 40.
Total outflows = 160.
Calculated inflows = 100.
75% inflow cap = 120, so all 100 inflows are recognised.
Net cash outflows = 160 − 100 = 60.
LCR = 120 / 60 = 200%.
Now change only calculated inflows to 150. Recognised inflows are capped at 120, so net outflows become 40 and LCR becomes 300%, not 1,200%.
LCR is nonlinear because caps bind
If a bank adds one more unit of Level 2B assets while already at the 15% cap, that unit may contribute no additional HQLA. If an additional contractual inflow arrives while the 75% inflow cap already binds, that inflow may contribute no additional denominator relief.
Therefore:
marginal balance-sheet amount ≠ marginal LCR benefit.
This makes LCR optimisation a constrained, piecewise problem rather than a simple ratio-management exercise.
Inputs and outputs
A robust LCR calculation engine can require:
- legal entity and consolidation perimeter;
- asset identifiers, market values and eligibility attributes;
- encumbrance and monetisation status;
- HQLA level and haircut;
- securities-financing and collateral-swap maturity data;
- deposit/customer classification;
- funding maturity and operational status;
- derivative cash flows and collateral features;
- committed and contingent facilities;
- contractual inflows;
- currency and legal-entity transfer restrictions;
- jurisdictional rule version.
Outputs can include HQLA before and after caps, gross outflows, gross inflows, capped inflows, net outflows, LCR, composition concentrations, currency-specific diagnostics and data-quality exceptions.
Evidence polarity: what supports confidence?
Evidence for a reliable LCR result includes asset eligibility traceable to the rulebook, HQLA haircuts reproducing Basel examples, Level 2 caps enforced after the required adjustments, deposit classes tied to documented attributes, derivative and commitment outflows reconciled to source systems, inflows capped correctly, and numerator/denominator movement explainable from one reporting date to the next.
Evidence against confidence includes unexplained HQLA growth without new assets, Level 2B exceeding its permitted composition, HQLA assets that are encumbered or operationally inaccessible, deposit run-off rates determined only by product name, recognised inflows above the permitted cap, or large ratio changes caused solely by mapping-table updates.
Counterexample: more marketable securities do not always improve LCR
A bank buys a corporate bond that is liquid in normal markets. If it does not satisfy HQLA eligibility criteria, it adds assets but may add zero LCR numerator.
If it qualifies only as Level 2B and the 15% cap already binds, it can still add little or no recognised HQLA.
Counterexample: a deposit with long contractual maturity is not automatically stable
Behavioral and withdrawal characteristics matter. A deposit product that can be withdrawn early without meaningful penalty may not behave like its nominal contractual maturity suggests.
The algorithm therefore needs the regulatory definition of stability rather than a simple date comparison.
Counterexample: large expected inflows do not eliminate the need for HQLA
The 75% inflow cap deliberately prevents a bank from constructing a near-zero denominator solely from expected incoming cash. A bank must retain a meaningful HQLA buffer against gross stress outflows.
Counterexample: an HQLA asset can fail operationally
An eligible government security held in the wrong legal entity, pledged beyond the stress horizon or subject to transfer restrictions may not be available where the liquidity need occurs.
Classification truth and operational availability are separate tests.
Weak links in implementation
Security-master mapping error. Wrong issuer, rating or asset class produces the wrong HQLA level.
Encumbrance lag. Collateral status updates arrive after the LCR snapshot.
Deposit misclassification. Stable/less-stable status is inferred from product label instead of customer and insurance attributes.
Double counting. A cash flow appears in both contractual and contingent feeds.
Cap ordering error. HQLA composition limits are applied before required transaction-unwind adjustments.
Inflow-cap omission. All expected inflows are recognised.
Sign error. A derivative outflow arrives as a negative amount and is accidentally added as an inflow.
Currency blindness. Aggregate LCR looks strong while a significant currency has a material shortfall.
Stale rule version. National implementation or Basel FAQ changes are not incorporated.
Diagnostics: how to test the algorithm
- zero-inflow test: with inflows set to zero, net cash outflows must equal gross outflows.
- inflow-cap test: raise inflows above 75% of outflows and verify recognised inflows stop increasing under the baseline rule.
- Level-2B cap test: add excess Level 2B assets and verify recognised HQLA respects the composition cap.
- Level-2 cap test: construct a portfolio dominated by Level 2A/2B and verify the overall 40% cap.
- haircut test: reproduce 15%, 25% and 50% category examples where the Basel classification applies.
- encumbrance test: pledge an HQLA asset and verify availability changes appropriately.
- deposit migration test: move a deposit from stable to less-stable and verify the outflow increases.
- commitment test: add an undrawn facility and verify off-balance-sheet outflow treatment.
- reconciliation test: HQLA and cash-flow populations reconcile to the source balance sheet plus documented off-balance-sheet feeds.
- explainability test: every material day-on-day LCR movement is decomposable into balance, market value, classification, haircut, cap or cash-flow changes.
What would falsify confidence?
Confidence should be withdrawn if the engine permits ineligible or encumbered assets into HQLA; if Level 2 caps can be breached; if recognised inflows exceed the permitted cap without an explicitly documented rule treatment; if gross outflows do not reconcile to source liabilities and commitments; or if the same portfolio produces different ratios merely because processing order changed.
Alternatives and limits
LCR is a standardized 30-day stress metric. It does not replace contractual maturity ladders, intraday-liquidity monitoring, survival-horizon analysis, internal stress testing or the longer-horizon Net Stable Funding Ratio.
A bank can have an LCR above 100% and still face a liquidity problem outside the 30-day horizon, in a particular currency, in a trapped legal entity or under a stress more severe than the standard assumptions. Conversely, supervisors may expect banks to use HQLA buffers during genuine stress rather than mechanically preserve a ratio at the cost of destabilizing behavior.
How this connects to the surrounding knowledge estate
LCR classification relies on exact position and cash-flow data, linking directly to banking money arithmetic. HQLA availability intersects with collateral optimisation. Derivative margin liquidity connects to initial-margin algorithms. Reverse stress can search for combinations of deposit run-off, collateral calls and HQLA impairment that force the ratio through 100%.
Verification and update triggers
Preserve the Basel/jurisdictional rule version, asset-eligibility mapping, haircut table, composition-cap logic, deposit classification, outflow/inflow factors, encumbrance state and source-system snapshot. Revalidate after rule changes, new Basel FAQs, new asset classes, deposit-product redesign, collateral-system changes, legal-entity restructuring or unexplained ratio breaks.
Primary and high-quality references
- Basel Committee on Banking Supervision, Basel III: The Liquidity Coverage Ratio and liquidity risk monitoring tools, 7 January 2013; now consolidated into the Basel Framework.
- Basel Framework, LCR30 — High-quality liquid assets, including Level 2A/2B haircuts and composition caps.
- Basel Framework, LCR40 — Cash inflows and outflows, including the 30-day denominator and run-off/inflow treatment.
- Basel Framework, LCR99 — Application guidance, summarising factors and HQLA composition calculations.
- Office of the Superintendent of Financial Institutions, Liquidity Adequacy Requirements (2026), Chapter 2 — Liquidity Coverage Ratio, a current supervisory implementation cross-check.
Educational boundary: This article explains standardized bank-liquidity mathematics. It does not assess a specific institution’s safety, predict a bank run or provide personalized financial advice.