What the LCR Actually Promises—and What It Does Not
This chapter examines the Liquidity Coverage Ratio as a specific regulatory test: how much qualifying liquid-asset value a bank has against modeled net cash outflows over the next 30 days. The useful judgment is not whether the ratio sounds high, but whether the reader can identify what has been counted, which stress assumptions produced the denominator, and which forms of liquidity remain outside the test.
The LCR matters because “liquid” is not a natural property that an asset either has or lacks. It depends on whether the asset can retain value, be sold or pledged in strained markets, and meet the timing of an obligation. Regulation turns those judgments into a formula. The formula creates comparability, but it also creates boundaries: a bank can satisfy the test without having a guarantee that every commitment will be funded at every moment.
What counts as high-quality liquid assets
The LCR's numerator is the bank's stock of high-quality liquid assets, usually abbreviated HQLA. These are assets intended to remain valuable and marketable during a period of financial stress. Cash and central-bank reserves are the clearest examples. Highly rated sovereign or central-bank securities can also qualify because they are generally expected to retain market liquidity when private assets are difficult to sell.
The designation is not simply a list of safe-looking instruments. The LCR assigns different regulatory treatment to different assets. Level 1 assets receive no haircut in the BIS paper's description and can make up the full HQLA stock. Level 2 assets, such as certain high-grade public-sector or covered bonds and, under national discretion, some lower-rated corporate debt or equities, receive mandatory haircuts. A haircut means that the value recognized for regulatory coverage is less than the asset's market or accounting value. Level 2 holdings are also capped as a share of the HQLA stock, with a lower cap for the riskiest Level 2B assets.
The size of those adjustments makes the composition of the portfolio matter, not just its gross value. In the revised standards described by the BIS paper, Level 2A assets receive at least a 15% haircut, while Level 2B assets receive haircuts of 25% to 50%. Level 2 assets together cannot exceed 40% of HQLA, and Level 2B assets cannot exceed 15%. These percentages are regulatory recognition rules, not predictions that an asset will literally lose exactly that amount in the next crisis. They are a way of preventing a bank from meeting the ratio by concentrating too heavily in instruments whose liquidity is more fragile. Two banks with the same market value of securities can therefore report different recognized HQLA if their portfolios have different levels of exposure to the capped categories.
The logic is straightforward. In a calm market, a broad portfolio may appear readily saleable. In a crisis, however, market depth can disappear, prices can gap lower, and the act of selling can impose losses. A regulatory numerator that recognized every asset at full value would overstate what could be raised under pressure. Haircuts and composition caps make the numerator more conservative by asking how much reliable value remains after allowing for stress conditions.
That conservatism does not mean that HQLA are identical to cash in every situation. A Treasury security and a reserve balance may both receive favorable treatment, but they move through different settlement and funding channels. A bank may sell or pledge an asset, face a counterparty or operational constraint, or encounter a market that is functioning but less deep than the model assumes. The LCR deliberately simplifies those possibilities into regulatory weights. It is a resilience measure, not a complete map of every transaction required to turn an asset into payment capacity.
This is why the numerator should be read as recognized stressed value rather than as a promise of frictionless monetization. The Federal Reserve paper makes the broader point through its discussion of buffers: even liquidity that appears on a bank's balance sheet can be difficult to deploy if doing so brings the bank close to a binding requirement or signals weakness. The LCR's asset rules are therefore only the first filter. They tell us which resources count for a specified test; they do not settle how the bank will use those resources when several obligations compete.
The numerator also has an opportunity cost. The BIS paper reports that banks more exposed to the LCR increased their LCR primarily by increasing HQLA, rather than by making a comparable reduction in modeled net outflows. That adjustment can strengthen resilience and reassure creditors. But the paper also places it beside research suggesting that liquid assets can displace loans, reduce profitability, or change the funding model. Those costs are not proof that the LCR is ineffective. They show that recognizing more liquid assets may improve one part of a bank's balance sheet while changing what the bank can earn or lend elsewhere.
How a 30-day stress test turns flows into a ratio
The denominator is projected total net cash outflows over a 30-day stress scenario. In simplified form, the LCR is:
[ \text{LCR} = \frac{\text{stock of HQLA}}{\text{total net cash outflows over 30 days}}. ]
The denominator begins with expected cash outflows. These are not limited to customers withdrawing deposits. The regulatory calculation applies runoff assumptions to categories of liabilities and off-balance-sheet commitments, including wholesale funding that may not roll over, derivative collateral requirements, and certain credit commitments. The calculation then recognizes expected cash inflows from contractual receivables, but only up to a cap of 75% of total expected outflows. Net cash outflows are the difference between the two.
A simple illustration shows why the cap changes the meaning of the ratio. Suppose a bank's modeled gross outflows are 100 units and its contractual inflows are 90. The rule recognizes no more than 75 units of inflows, leaving 25 units of net outflows. If the bank has 25 units of recognized HQLA, its ratio is 100% for this simplified scenario. Without the cap, the same contractual inflows would leave only 10 units of net outflows and make the ratio appear much higher. The example is not a forecast of any particular bank; it shows how the rule refuses to treat expected receipts as a complete substitute for a liquid-asset reserve.
The inflow cap is one of the most important parts of the design. A bank may expect money to arrive from borrowers, counterparties, or maturing assets, but during a system-wide stress those inflows may be delayed, disputed, or dependent on the same market conditions that are causing the bank's outflows. Allowing every expected receipt to offset every expected payment would make the ratio look comfortable precisely when counterparties are least reliable. Limiting recognized inflows means that the bank must hold a stock of qualifying assets even when it expects some cash to come back.
This also explains why the denominator is a scenario rather than a forecast in the ordinary business sense. The question is not “What will cash flows most likely be next month?” It is closer to “How much recognized liquid value would cover a prescribed pattern of stress-related outflows, after conservative treatment of inflows?” The answer depends on the runoff rates and other regulatory assumptions. Change the assumed behavior of depositors, wholesale creditors, derivatives counterparties, or commitment holders, and the modeled denominator changes even if the bank's assets do not.
The scenario also makes the LCR a test of interaction, not just of isolated liabilities. A bank may be able to manage a deposit outflow in one state and a wholesale-funding rollover in another, yet face a much harder problem if both occur while collateral calls rise and borrowers draw committed lines. The regulatory calculation does not reproduce every possible joint crisis. It selects a prescribed combination of outflows and inflows so that institutions can be compared and supervisors can ask whether the bank has a reserve against that combination. The result is useful precisely because it is standardized; its limits follow from the fact that a standardized scenario cannot be identical to every future shock.
The 30-day horizon is equally important. It makes the ratio focused and operational: the bank should have enough qualifying assets to navigate an intense short-term funding episode. But a 30-day test does not describe every maturity mismatch, every slow-moving credit deterioration, or every longer crisis. The BIS paper distinguishes the LCR from the Net Stable Funding Ratio, which addresses funding stability beyond a one-year horizon, and focuses its own analysis on the LCR. That distinction is a reminder that liquidity risk is multi-horizon. Passing one horizon does not eliminate pressure at another.
Nor does the LCR answer whether a corporate borrower can draw a committed line. The ratio's denominator includes some off-balance-sheet commitments through regulatory outflow assumptions, but the borrower's actual ability to draw depends on the facility's terms, financial covenants, and the bank's decision to accommodate a simultaneous wave of requests. A commitment can be legally outstanding, economically constrained, or operationally difficult to fund. The LCR provides a bank-level buffer against a modeled class of outflows; it does not convert every contingent promise into segregated cash for the beneficiary.
The difference between qualifying HQLA and deployable liquidity can be summarized as a sequence:
The diagram is a measurement path, not a claim that the final comparison guarantees the outcome of a crisis. A bank can have a compliant ratio and still face three further questions. Can its recognized assets be mobilized at the required speed and value? Can the relevant borrower or counterparty actually take the promised funding? And will creditors continue to supply or roll funding while the bank uses its room?
The applicable requirement must also be identified before the percentage is interpreted. The Basel formulation described in the BIS paper uses a 100% standard, but the papers also discuss regulatory categories and differing requirements in the United States. A ratio is not a buffer until the reader knows the floor against which it is measured. “110%” by itself is incomplete: it could represent substantial room under one requirement and far less under another. The next question is therefore not whether the bank is above zero or above a round number, but how much distance separates the observed ratio from the rule that governs that bank.
This boundary is the foundation for the rest of the series. The LCR is neither a meaningless accounting label nor a universal guarantee. It is a structured, stress-based test that disciplines the asset stock and the assumed cash-flow profile. Its value lies in making resilience visible and comparable. Its limitation is that the percentage does not, by itself, reveal how a bank will behave as it approaches the regulatory floor, which borrowers will receive liquidity, or how creditors will respond.
A careful reading therefore keeps three statements separate. First, HQLA are assets that receive favorable recognition because they are expected to retain value and marketability in stress. Second, the denominator is a modeled 30-day net-outflow requirement, not a complete forecast of all future liquidity needs. Third, compliance with the ratio is a condition for resilience, not proof that every asset, credit line, or funding relationship is immediately usable. Once those distinctions are in place, the next issue becomes visible: two compliant banks can have very different room above the floor, and that room can shape their willingness to act.