Before I head off on holiday, I wanted to flag a deceptively unremarkable paper just published in the BIS Working Paper series — one whose importance is almost camouflaged by its technical language and routine presentation. Yet, it touches on a topic that we hold dear: the underappreciated role intraday liquidity played in the 2008 crisis, and the extent to which the abundance of excess reserves has been quietly masking unresolved vulnerabilities in how RTGS systems manage intraday liquidity.
If you read our perpetual future Long Read, you will have picked up on these issues, not least because our standing position is that the invention of perpetual future may have inadvertently created a mechanism that can help manage upcoming intraday liquidity constraints without the need for de facto subsidization.
Thus, even though the BIS paper appears, at first glance, to be another dry contribution to the literature on payment systems, its substance is actually quite striking. Beneath the technicalities lies an implicit admission that the market can no longer postpone addressing these inherent weaknesses, and that a market-based solution will have to emerge sooner rather than later.
Here are the relevant bits, starting with a passage acknowledging that RTGS wasn’t necessarily the panacea for so-called Herstatt settlement risk that everyone initially thought it was:
The authors go on to highlight that the new scarce liquidity environment could put pressure on the standing RTGS system, especially now that stablecoins are reintroducing elements of prefunding and gross settlement (I’ve helpfully pointed out the “where we are at” point):
And here is the key pressure point as they see it:
To combat these risks, the BIS authors — Rodney Garratt, Morten Bech, Marko Nanut Petric, and Caner Ates — seek a “novel auction-based liquidity-saving mechanism” for real-time gross settlement systems (RTGS).
In essence, rather than relying on banks pre-committing a fixed amount of liquidity or purely first-come-first-served queues, they believe short auctions should be used throughout the day to find efficient ways to settle payments that reduce how much liquidity each bank needs at any given moment.
The theory goes that without a mechanism like this, banks often have to hold large amounts of idle cash or central bank balances to make sure they can cover their outgoing payments. When abundant reserves are no longer available, that becomes expensive or impractical. Hence the urgency now that reserves are getting tighter.
Crucially, the paper argues that competitive pricing — meaning letting banks bid for when and how much liquidity they need or can supply — is essential to making the proposed liquidity-saving mechanism work well because it reveals participants’ true costs and preferences and thereby aligns incentives so that liquidity can be shared efficiently.
This, however, is analogous to the “surge pricing mechanisms for intraday liquidity” that I’ve been advocating for for a very long time. But hey, thanks for the non-acknowledgement! (As you can tell, I’m not bitter at all that every time I’ve raised the issue with central bankers or payments experts — with the exception of the guys at Finteum and one chat with a BIS official — they’ve either rolled their eyes or quickly changed the subject.)
Of course, it’s also been my contention that a de facto auction-based LSM for prefunded liquidity already exists in the form of perpetual futures systems in crypto markets. The difference is that such systems usually run funding cycles every 8 hours, whereas the proposed auction-based LSM would net at 10-minute intervals, like err, Bitcoin. Other than that, traders post collateral (often stablecoins) that sit idle until needed to absorb P&L flows and funding transfers.
Those stablecoins are effectively a prefunded liquidity pool that enables continuous settlement. The funding rate, together with cross-rates between spot, perp, and funding markets, implicitly prices the opportunity cost of tying up that collateral intraday. If alternative funding mechanisms exist — for example borrowing stablecoins elsewhere, rotating collateral, or using different venues — arbitrage across those rates determines where liquidity is held and when. The result is a market-implied price of short-term, collateralized liquidity.
An auction-based LSM, as proposed in the paper, on the other hand, would be even more analogous to the revival of a Libor market for intraday liquidity.
Serious CBankers would presumably scoff at that suggestion. But I don’t personally see why that’s not the case.
Both LIBOR and the proposed LSM are mechanisms for discovering a price at which banks that are short of liquidity can induce banks with surplus liquidity to lend to them over a very short horizon. In LIBOR’s case, that horizon was overnight or term unsecured funding; in the LSM’s case, it is the next intraday settlement window. The economic question being answered is fundamentally the same: what rate clears the marginal unit of liquidity between banks right now?
In all three cases — LIBOR-era interbank markets, perp funding, and the BIS LSM — the system is discovering the marginal price at which balance-sheet capacity is willing to move across participants to keep a continuous system running.
The big unanswered question in the paper is whether an auction-based LSM would dispense liquidity on a collateralized basis (a la the new norm for daylight overdrafts) or on an uncollateralized basis. If the former, then it would be even more analogous to the perp systems already in place.
Whatever the case, the paper’s conclusion hints at the problems that will arise if these systems are not introduced soonish.
The analytical framework developed here allows for an explicit assessment of the trade-off between additional liquidity savings and increased settlement delays. By expressing both in comparable monetary terms, using an interest rate that reflects the opportunity cost of liquidity, operators can make transparent and well-informed design decisions.
Incorporating an auction-based LSM could provide payment system operators with new tools to enhance efficiency and resilience.
Related links:
The Fed’s quiet pivot to servicing intraday liquidity — and why it matters — The Blind Spot (March 25, 2025)
The crypto innovation traditional finance needs — Bloomberg (August 31, 2022)
IMPORTANT The Uber surge pricing model really is coming to liquidity markets — The Blind Spot ( November 17. 2023)
Changes to intraday funding plumbing are coming sooner than you think — The Blind Spot (Sept 30, 2022)
What’s up with Fed daylight overdrafts data? — The Blind Spot (May 24, 2022)
How RTGS killed liquidity: US tri-party repo edition — FT Alphaville (October, 2019)
How RTGS inadvertently killed system liquidity — FT Alphaville (October, 2019)
