The RBA has given up the quantity lever. Now the price on the window is everything
A demand-driven liquidity regime only delivers ample reserves if banks are willing to ask for them. Price the Reserve Bank's window too far above the cash rate target and "ample" becomes a name, not a condition.
Australia's central bank, the Reserve Bank of Australia, runs the account system banks use to pay each other. The balances in those accounts are called settlement balances, and banks need enough of them each day to settle payments. Until now the Reserve Bank has largely decided how many of those balances exist. It is moving to a system where banks themselves decide how much to borrow, by pledging assets to the Reserve Bank at a price the Bank posts.
Letting banks decide how many reserves they hold is sensible, but it only works if borrowing from the Reserve Bank is cheap and ordinary. Price the window above the cash rate target and banks will avoid it exactly when it matters.
The word doing the work is not “ample”
“Ample” is a word about quantity. The regime the Reserve Bank of Australia has named after it will be settled by a price.
On 25 August the Bank’s Head of Domestic Markets, David Jacobs, set out the path in a speech called “The Road to Ample — Towards a Demand-driven Liquidity Regime”. Both halves of that title are terms of art, and they point in slightly different directions. “Ample” is the destination: a banking system holding comfortably more settlement balances — the money banks keep at the central bank in order to pay each other — than it strictly needs to get through a normal day. “Demand-driven” is the method. The Reserve Bank does not decide how many of those balances exist. It posts a price and a list of acceptable collateral, and lends whatever quantity the banks come and ask for.
That is a genuine surrender of control, and a deliberate one. The alternative is a supply-driven system, in which the central bank estimates the system’s appetite for reserves and buys assets until it thinks it has satisfied it. Guess too high and you carry a balance sheet you did not need. Guess too low and you starve the payment system on a Tuesday afternoon. Demand-driven removes the guess. But it swaps one hard problem for another: having handed the quantity lever to the banks, the Bank keeps only the price lever, and everything now turns on where that price sits.
Why the spread is the entire design
Banks will obtain reserves mainly through repurchase agreements — selling securities to the Reserve Bank and agreeing to buy them back a short time later. The rate on those operations becomes the marginal cost of settlement liquidity for the whole system. Nothing else in the framework competes with it.
Set that rate at or barely above the cash rate target — the overnight interest rate the Reserve Bank Board actually sets — and no bank has any reason to sit on a shortfall. The observed cash rate is welded to the target, because a cheap alternative is available on demand to anyone with collateral. Set it meaningfully above the target and you have invited every bank treasurer in the country to economise: hold a fatter precautionary buffer, hesitate at the margin, arrive at the window last rather than first. The quantity of reserves in the system is then not “ample” in any meaningful sense. It is whatever reluctance permits.
This is the part of the transition that deserves argument, and it is the part most easily lost in the language of plumbing. A demand-driven regime does not guarantee ample reserves. It guarantees only that reserves will be as ample as banks are willing to ask for at the posted price.
Stigma is a design flaw, not a temperament
The standard objection to relying on banks to come forward is stigma — the worry that borrowing from the central bank marks a firm as troubled. Stigma is usually discussed as if it were a psychological quirk that education might cure. It is not. It is a rational response to a price signal. If the window costs more than the market, using it says something about the user, because a solvent bank with options would not pay the premium. The premium creates the inference. Remove the premium and the inference dissolves, because everyone uses the facility all the time and using it says nothing at all.
That is why the design and the culture cannot be separated. A regime that wants routine, unremarkable borrowing has to price borrowing so it is routine and unremarkable. Frequent operations, a broad collateral list, pre-positioned collateral tested in advance rather than in an emergency, and a spread small enough that no board paper is required to justify tapping it.
The strongest case for making it expensive
The case against me is serious. A window priced at the target makes the central bank the cheapest lender in the country. The interbank market — banks lending spare balances to each other overnight — withers, because there is no reason to shop around. With it goes a price signal and a form of peer discipline: banks that other banks are wary of lending to used to find that out early and cheaply. And a cheap, permanent facility encourages institutions to hold thinner buffers of liquid assets, knowing the central bank will fill the gap.
I think that argument fails on its own terms. An interbank market that exists only because the alternative has been deliberately overpriced is not price discovery. It is a toll, and what banks are paying is not a price for credit risk — central bank repo is collateralised — but a fee for administrative discouragement. Nor should the operating framework be doing the supervisor’s job. If banks are holding too few liquid assets, that is a matter for liquidity rules, capital rules and supervision. Using the settlement system’s pricing as a substitute for prudential regulation gets you weak versions of both: a rate target that misses at quarter-ends, and risk-taking that the rules were supposed to constrain anyway.
There is a further Australian reason to go cheap. The Payments System Board is currently consulting on the role of the Reserve Bank Information and Transfer System — the country’s high-value settlement system — in supporting settlement in a tokenised ecosystem, and on the future of account-to-account payments. Both point toward liquidity needs that are faster, lumpier and harder to forecast. A regime that expects banks to pre-empt their own shortfalls, at a penalty, is exactly the wrong regime for that world.
The Reserve Bank should publish, from the first operation under the new regime, a standing monthly series: total amount bid and allotted, the number of distinct counterparties bidding, the weighted average operation rate expressed as a spread to the cash rate target, and the daily distribution of the observed cash rate around that target. None of this is commercially sensitive at an aggregate level, and all of it is the evidence on which the design will be judged.
It should also commit in advance to a trigger. If the number of bidding counterparties narrows, or the observed cash rate spends more than a handful of days above target by more than a few basis points, the spread comes down without a review process. Domestic Markets owns the operation rate; the Payments System Board owns the settlement-system changes that will shape demand for it. Both should say now what number would make them move.
What would change this view
Set out the observable test. If the Reserve Bank launches the regime with its open market operation rate at a visible spread above the cash rate target — say ten basis points or more — and then, across four consecutive quarters of published operations, the number of distinct bidding counterparties is broad and stable rather than concentrated in a handful of names, and the actual cash rate stays within a couple of basis points of target through quarter-ends and year-end, then I am wrong.
If that happens, the spread is doing useful work: it is preserving an interbank market without rationing reserves, and the stigma problem I am warning about did not materialise in Australian conditions. I would then argue the opposite case — that the Bank should hold the spread rather than compress it, and that other central banks running near-zero spreads are carrying more balance sheet than they need.
Worked from
- The Road to Ample - Towards a Demand-driven Liquidity Regime
Reserve Bank of Australia · 2026-08-25