What is the Basel III liquidity coverage ratio?

The short answer

The 2008 financial crisis exposed banks that were technically solvent but ran out of cash to meet obligations. Basel III, finalised in 2013 and implemented in major countries by 2015, addressed this with a new rule: banks must hold enough high-quality liquid assets to survive a 30-day stress period without external funding.

The numerator of the ratio is HQLA — assets that can be converted to cash quickly without losing value. The denominator is the projected net cash outflow over 30 days under a defined stress scenario (deposit run-offs, drawn credit lines, derivative collateral calls). The ratio must be at least 100 %.

The intent was straightforward: prevent another Bear Stearns or Lehman, where institutions failed because they could not access liquidity even when their assets were ultimately solvent. The actual effects of the rule have been more complex than the designers anticipated.

→ Background: Why do repo markets matter for financial stability?

What the data shows

The structural parameters of the LCR (BIS, Federal Reserve, OCC):

• Minimum LCR: 100 %, fully phased in by January 2019.
• Level 1 assets (no haircut): central bank reserves, U.S. Treasury securities, agency debt with explicit U.S. government guarantee.
• Level 2A assets (15 % haircut, capped at 40 % of total HQLA): agency debt, agency MBS, certain sovereign debt.
• Level 2B assets (50 % haircut, capped at 15 % of total HQLA): investment-grade non-financial corporate bonds, certain equities.
• HQLA at large U.S. banks: rose from less than 4 % of total assets in 2006 to over 10 % by 2010, with further accumulation through full LCR implementation.

The exception worth noting: U.S. implementation excludes certain assets that Basel III allows globally. Privately issued mortgage-backed securities and bonds issued by financial institutions are not eligible HQLA in the U.S. version, even though they qualify in some other jurisdictions.

→ Dataset: U.S. bank reserves

Why it happens — the macro mechanism

The LCR’s effects on the broader financial system extend well beyond its narrow purpose of bank liquidity management.

Channel 1 — the demand sink for Treasuries. By making Treasuries the dominant Level 1 asset and applying no haircut, the LCR created structural demand for U.S. government debt from large banks. This was deliberate — Treasuries were chosen as the safest collateral. But the effect has been to transform Treasuries from a pure investment asset into a regulatory good that banks must hold regardless of price.

Channel 2 — the regulatory collateral concentration. Here is the angle that distinguishes serious analysis. The LCR has effectively turned Treasuries into a single-asset regulatory standard for liquidity. When Treasury markets are stressed (as in March 2020), the entire HQLA buffer becomes harder to monetise — and the very assets banks were required to hold for safety became the source of liquidity strain. The Bank Policy Institute and others have documented how this creates an “upside-down” effect: the regulation designed to prevent runs can amplify market stress by concentrating bank holdings in one asset class whose liquidity is itself state-dependent.

Channel 3 — the unintended interaction with discount window stigma. The LCR treats borrowing from the Fed’s discount window as essentially equivalent to selling HQLA — both reduce the available HQLA buffer. Combined with the historical stigma of discount window borrowing, this means that during stress, banks are reluctant to use the central bank as a backstop precisely when the central bank wants them to. The September 2019 episode and subsequent SRF design were partly responses to this dynamic.

Synthesis by regime: in the pre-2014 regime, banks held a much smaller liquidity buffer and intermediated more aggressively in money markets. In the post-2015 LCR regime, banks hold large HQLA buffers but are less willing to lend reserves overnight, which contributed to the September 2019 distribution problem. In acute stress regimes (March 2020, March 2023), the rigidity of the LCR can become a constraint on the system’s ability to absorb shocks even when aggregate HQLA looks abundant.

The LCR transformed Treasuries from an investment asset into mandatory regulatory collateral — solving one liquidity problem while creating a new concentration risk.

→ Framework: Monetary regimes, interest rates and liquidity

What it means for different economic actors

Banks have adjusted their balance sheets significantly. The cost of holding low-yielding HQLA has become a permanent drag on profitability, and Treasury management functions have grown substantially in importance. Internal LCR optimization has become a sophisticated discipline.

Treasury markets have benefited from structurally higher demand for Treasuries from the banking sector. This has likely depressed yields somewhat relative to a counterfactual without the LCR, although isolating the effect is empirically difficult.

Money markets have lost some intermediation capacity. Banks that historically lent reserves freely overnight now optimize that lending against LCR considerations, contributing to the kind of distribution failure observed in September 2019.

A common error is to evaluate the LCR purely on its narrow objective. By that standard it has succeeded — no large bank has failed for pure liquidity reasons since implementation. But the broader effects on market functioning, central bank operations, and Treasury market dynamics are equally important and harder to assess.

Practical observation

Go deeper

Related questions

• Why do repo markets matter for financial stability?
• What happened during the September 2019 repo crisis?
• How does funding liquidity differ from market liquidity?
• What is the Fed’s Standing Repo Facility?

Frequently asked questions

Did the LCR succeed in its main objective?

By the narrow standard of preventing pure liquidity failures at large banks, yes. No major U.S. bank has failed for the kind of overnight funding seizure that brought down Bear Stearns in 2008. But the 2023 regional bank failures (SVB, Signature, First Republic) showed that the LCR’s 30-day stress assumption does not capture the speed of modern deposit runs amplified by social media. The standard was designed for a slower world than the one banks now operate in.

Why does the LCR effectively favour Treasuries over other collateral?

Treasuries receive zero haircut as Level 1 assets, while corporate bonds receive a 50 % haircut as Level 2B. Holding $1 of Treasuries provides $1 of regulatory liquidity buffer; holding $1 of corporate bonds provides only $0.50. This dramatic asymmetry creates strong incentives to concentrate HQLA portfolios in Treasuries, which has implications for both bank profitability (Treasuries yield less than corporates) and Treasury market structure (banks become major price-insensitive buyers).

How does the LCR interact with the discount window?

Borrowing from the discount window for less than 30 days does not improve a bank’s LCR, because the loan generates an outflow within the LCR’s 30-day horizon. This means banks gain no LCR benefit from accessing the central bank as a backstop, which combined with discount window stigma makes the discount window effectively useless for many institutions in stress. Bank Policy Institute analysis has called this an “upside-down” effect of the regulation that limits the central bank’s ability to function as lender of last resort.

Last updated — 12 May 2026