Towards a Demand-driven Liquidity Regime

Speech

Notes

The Road to Ample – Towards a Demand-driven Liquidity Regime David Jacobs * Head of Domestic Markets RBA 25 August 2026
– Sydney
Audio 43.3MB Q&A Transcript Watch video: Speech delivered by David Jacobs, Head of Domestic Markets, Reserve Bank of Australia, Sydney Introduction – destination ample Thank you for joining me at the Reserve Bank. I would like to acknowledge the Gadigal People of the Eora
Nation, the traditional owners of the land we meet on. I pay my respects to their Elders past and present
and extend that respect to all First Nations peoples present. Today I’m focusing on the RBA’s transition to a new regime for implementing monetary policy. To
keep the cash rate near the Monetary Policy Board’s target, the RBA sets terms on which banks can
access Exchange Settlement (ES) balances. These are a form of money that banks use to settle payments
with each other and are also known as reserves. Different regimes can achieve the cash rate target, but
each approach has different implications for the liquidity available to the banking system. The regime we’re leaving is that of high, supply-driven reserves that arose from our policy responses
to the pandemic. The RBA’s bond purchases and term funding operations injected nearly
$500 billion of reserves into the banking system (Graph 1). Banks had less need to borrow from
one another or make use of our regular liquidity operations. The rate paid on reserves – the ES
rate, tied to the Board’s cash rate target – became the anchor for short-term interest rates.
However, reserves have since fallen by more than half as the assets the RBA acquired under pandemic-era
policies have matured, and that will continue. Graph 1 The regime we’re moving to is one of ample reserves – which in the Australian context means a demand-driven framework. This is not about reaching a particular level of reserves. Rather, it
is a system that meets and adapts to banks’ demand for reserves, and in that way steers short-term
interest rates. The Board chose this approach in 2024, rather than returning to the pre-pandemic framework of scarce
reserves. Experience had highlighted the benefits of readily available reserves for monetary policy
transmission, banking system resilience and payments efficiency. 1 Under the new framework, we supply as much reserves as are demanded through our regular full-allotment
open market operations (OMOs) which offer reserves at a modest spread over the ES rate. But we don’t
attempt to push more reserves into the system beyond that which is demanded at that prevailing price. 2 In 2025, we
announced the detailed design, including the spread of the OMO rate over the ES rate. 3 As we move from a system in which the RBA determines the quantity of reserves to one in which the banking
system does, a key question is how that can occur smoothly and adapt as market conditions evolve. That
will be key to maintaining effective control of the cash rate, transmission of monetary policy and
well-functioning markets. In that context, I’ll address three issues today: How close are we to a demand-driven environment for reserves, and what are we learning as we get
closer? How will our operational framework help with a smooth transition? I’ll outline the different
roles of OMOs, other liquidity operations, and why we’re announcing today that one facility
– open repo – is no longer needed. How should market participants prepare for a world in which liquidity management once again becomes
more active for everyone? I want to stress that none of these issues bears on the stance of monetary policy itself. The cash rate
target continues to be set in the normal way by the Monetary Policy Board, and as I said we can achieve
that target under either regime. We are not yet at ample reserves It is important to know if we’re already in a demand-driven environment or if we’re still
transitioning towards that. As reserves decline, two key things will signal we are close to ample: Short-term rates will rise and increasingly be guided by the cost of obtaining reserves (the OMO
lending rate) rather than the return on holding them (the ES deposit rate). Rates will nonetheless
remain consistent with the cash rate target because both those rates are set at a small spread to the
target. 4 The banking system will determine the quantity of reserves via OMOs. As market rates rise, it will
become more attractive to use OMOs, and they will become the key way in which the system overall gets
the reserves it needs. Currently, these signals suggest that reserves are still above underlying demand. First, reserves can still be sourced somewhat more cheaply in short-term repo markets than from the RBA
(Graph 2). Money market rates across instruments and maturities have moved well above the ES rate,
but we still hear in liaison that liquidity is abundant and placing downward pressure on market
rates. 5 Graph 2 And second, use of our full-allotment OMOs is still modest. We do see a wide range of counterparties
participating regularly, and total usage has increased recently. But the banking system has essentially
allowed reserves to decline rather than trying to offset this (Graph 3). Graph 3 Taken together, those indicators suggest that we’re not yet in an environment where the level of
reserves is driven by demand. Demand for reserves is evolving and uncertain While we haven’t reached ample reserves, we’re learning more about underlying demand as we get
closer. Banks’ expectations for the reserves they will hold at ample have fallen quite a bit over time. To
gauge how demand is evolving, we regularly survey banks to get a sense of their preferred reserve
holdings in the ample regime. In 2024, banks’ responses implied that ample reserves would be around
$ 100–200 billion (Graph 4). But that has since declined to around
$ 70–100 billion – well below the current level of reserves. This
helps explain why the fall in reserves has so far seen only limited reaction in either market prices or
OMO usage. 6 Graph 4 The decline in banks’ expected reserve holdings is, if anything, a welcome development, given that
one factor is likely to be comfort with the RBA’s liquidity framework. As I’ve said, ample is
not about targeting a particular quantity of reserves; it’s about making them available in whatever
size the system needs. What we’ve seen is that, as the supply of reserves has fallen, banks have been comfortable holding
high-quality liquid securities instead (Graph 5). Banks primarily hold reserves to meet payments,
manage liquidity risks and satisfy regulatory requirements, given they’re the most liquid asset
available. But banks have been willing to hold more securities because they are confident they can
readily turn them into cash, either in private markets or at the RBA’s liquidity facilities. It
helps that OMO use has remained routine through the period of excess reserves. And we’ve emphasised
that our facilities are available to be used and banks should be ready to access them when needed. 7 Graph 5 Some of the decline may also reflect changes in incentives. Last year, we increased the cost of obtaining
reserves through OMOs to encourage greater activity in private markets under the ample framework. Though
the survey suggests this has been a less important driver of demand. More generally, these developments underscore that we need to be ready for a range of scenarios for
reserves demand. International experience also illustrates that point – underlying demand for reserves has been
difficult to predict and very different across jurisdictions (Graph 6). Australia is towards the
lower end of the range for where reserves appear to be settling. There are various reasons for such
differences, including the inherent liquidity risks in the financial system and central banks’
operational frameworks. 8 And our shorter experience with excess reserves may also
have helped to avoid some of the reliance on large reserve buffers seen elsewhere. 9 Graph 6 And looking ahead, reserves demand will continue to evolve. Demand is influenced by various factors
– payments innovation, liquidity regulation and the shape of the yield curve – all of which
could shift in coming years and possibly in different directions. 10 So, when we might reach ample reserves ultimately remains highly uncertain. It could be some years away,
particularly if banks continue adjusting to a lower-reserves environment. But it could also be earlier if
banks increase their demand or there are market frictions. In any event, our framework is designed for uncertainty, so that reserves adjust to whatever level of
demand or timing eventually emerges. A flexible liquidity framework How is our framework designed to smoothly meet the system’s demand for reserves as we get closer? Let’s step through what will happen. As our bond holdings continue to mature, that by itself will keep reducing system reserves at a rate of
about $ 20–40 billion a year. 11 But as reserves reach underlying demand,
banks will replace those drained reserves with reserves obtained from OMOs. That’ll keep overall
reserves steady in the face of further bond maturities. This graph illustrates that simple transition (Graph 7). For now, I’m just showing reserves
supplied through bond holdings and OMOs; I’ll come to other liquidity operations that could also
contribute in a moment. Graph 7 In essence, the level of reserves will become determined by demand from the banking system. Rather than
earlier decisions about our asset holdings – in the form of bonds – driving reserves,
it’ll be the other way around. Demand for reserves will drive the size of OMO repos, which are a
short-term asset for the RBA. That has two important implications. First, we don’t need to accurately predict the exact level of underlying demand, at any point, for an
effective transition. If system reserves dropped below the level of underlying demand, there would be
upward pressure on money market rates, incentivising banks to come to OMOs for more liquidity. In that
way, demand for reserves automatically generates, in a timely way, the supply needed to keep the cash
rate close to target. Second, our pandemic-era bond portfolio can continue to run off uninterrupted. The level of our balance
sheet will become responsive to demand, but its asset composition will shift from long-term bonds to
short-term repo. In short, we don’t need to know the unknowable or make active choices to ensure a smooth transition. Other operations as a complement to OMOs While OMOs will be our key liquidity tool, we wouldn’t want them to reach a size that inhibits the
healthy functioning of the repo market. That may not become an issue. OMOs can safely scale up significantly and automatically from their size
today without impairing market functioning. Our repo book is much smaller than it was before the
pandemic. Meanwhile eligible collateral in the system has roughly doubled and been accompanied by a
strengthening and deepening of the private repo market (Graph 8). Indeed, as our bond holdings
mature that makes more bonds available for the market to use as collateral at OMOs. 12 The price
and other terms of OMO have also been set with the market’s health in mind. 13 So, it’s quite possible that system-wide reserves demand can be fully met by OMOs, particularly given
the lower demand estimate. Graph 8 But it’s also possible that at some point the size of the OMO inhibits market function by encumbering
a large amount of collateral, or involves operational risks. For those reasons, other central banks with
a similar framework to us are preparing to use a wider range of operations, if required, including the
European Central Bank (ECB), Bank of England (BoE) and Reserve Bank of New Zealand (RBNZ). 14 In that situation, we have other liquidity tools in place to complement OMOs. Those include foreign
exchange (FX) swaps, cross-currency basis swaps, and purchases of short-term Australian Government
Securities (AGS). These would allow us to inject reserves through other deep, liquid markets where the
associated risks to the RBA remain low. These types of operations are not new. Most were a routine part of our liquidity operations before the
pandemic (Graph 9). 15 Indeed, FX swaps were larger than our OMO book at
times in the early 2000s, when government deposits were draining reserves and the repo market wasn’t
large enough for us to fully offset that through repos alone. Graph 9 It’s important to be clear what the objective of these other operations would be. They would aim to
supply a relatively stable quantity of reserves to the system, so that OMOs do not inhibit
the functioning of private repo markets and to support operational resilience (Graph 10). In contrast to OMOs, they would not aim to influence an interest rate in FX swap,
cross-currency swap or AGS markets. Graph 10 To meet their objective, these operations would follow three key principles: Keep OMOs as the key source of liquidity. Other operations may provide substantial
reserves, but they should be no larger than needed. OMOs would continue to play their key role as the
marginal source of supply to meet the demands of banks and control short-term rates. Transact as a price-taker. We would transact at prevailing market rates in these
other operations. Though we’d avoid transacting at prices that might be inconsistent with our
monetary policy implementation objectives. Take a gradual and steady approach. If OMOs require any support – and they
might not – we would probably start with FX or cross-currency swaps. Our long-term bond
portfolio is still large and will still be rolling down when we reach ample. The size and composition
of these operations would be adjusted as conditions evolve. 16 So these other operations would help us to accommodate a wide range of scenarios for reserves demand and
market function. Should such operations begin, we would provide an appropriate level of transparency
around their nature and goals. Why open repo is no longer needed The final element of our operational toolkit is standing facilities. These play an important role in
supplying reserves for short terms between weekly OMOs. They are available bilaterally, on-demand. But one of our standing facilities – open repo – is no longer needed because banks can now
readily obtain as many reserves as they need from full-allotment OMOs. 17 Open repo was introduced in 2013 to overcome the constraints of the scarce reserves system. Its aim was to
provide an extra source of liquidity so that banks had the minimum reserves they needed to ensure smooth
after-hours payments, at zero-cost. At the time, we estimated the banks would need around $ 20–30 billion to achieve this, well above reserves supply of around
$1 billion. 18 Use of open repo has declined significantly in recent years, to below $5 billion (Graph 11).
Much of the current use is by banks that do not participate in after-hours settlements, who were also
given access. Graph 11 In the new liquidity framework, banks will still be expected to hold minimum ES balances to support
after-hours payments. But open repos are no longer needed to meet those. The RBA has decided that the facility will cease from early next year. We’ll communicate more with
market participants about this change in coming months. And we would retain the ability to reintroduce
the facility if future developments created a clear need for it again. Preparing for a more active liquidity environment How can market participants be well prepared as we get closer to ample? Money market rates are likely to become more variable, which is a normal feature of an active market. They
should become more responsive to fluctuations in the level and distribution of reserves, particularly
between weekly OMOs. That sort of variation provides useful signals and has been the recent experience in
the United Kingdom, New Zealand and Canada when transitioning to similar systems. 19 As we get closer to ample, and these price signals again begin to play more of a role, market participants
should be prepared to again manage their liquidity positions more actively. While we will supply the
reserves needed at a system level, that doesn’t displace the need for individual firms to prudently
manage their liquidity. Being prepared means: monitoring and forecasting their liquidity needs being ready to actively source liquidity through money markets and provide liquidity to others across
markets and having the operational arrangements in place to do that (such as appropriately calibrated
counterparty limits). Counterparties should also be ready to use the full range of RBA facilities, drawing on those when they
need and recycling liquidity to others through markets, as well as regularly testing these facilities. I
want to emphasise that our overnight and intraday standing facilities remain very much open for business,
and we expect their use to rise as reserves decline. Last year we made a joint statement with the
Australian Prudential Regulation Authority to underscore that use of the overnight facility is seen as
routine liquidity management, and we have since seen an encouraging lift in its use (Graph 12). 20 Graph 12 Conclusion Let me conclude. The road to ample is a transition from a system in which the RBA determines the quantity of reserves to
one in which the banking system does. We are not there yet, and when we will arrive is uncertain. Market behaviour will signal
when we are getting close, and today suggests that reserves still reflect our pandemic-era assets more
than the underlying demand of the banking system. Most importantly, reserve demand can and will evolve
over time, in both directions. Our framework is designed so that a smooth transition does not depend on knowing when we will
reach ample reserves. Ample is not about reaching a particular quantity of reserves.
It’s about a system that can flexibly supply whatever quantity the banking system demands, while
keeping the cash rate close to the Board’s target. Our full-allotment OMOs provide that flexibility.
And other liquidity tools will be available to provide a quantity of reserves, if OMOs were to reach a
size that inhibited repo market functioning. Under that framework, open repo is no longer needed. It was introduced to overcome the
constraints of a scarce-reserves system. But our other standing facilities remain very much open for
business. Finally, market participants have an important role to play. As reserves become more
demand-driven, active liquidity management will become more important for financial institutions.
Participants should be prepared to deal with the full range of RBA operations, monitor their needs, and
obtain and provide liquidity through private money markets. A flexible framework, supported by well-prepared market participants, will help ensure that transition is
a smooth one. Thank you for your time, and I look forward to your questions. Endnotes * I thank Ahmet Aziz, Alexandra Baker, Joel Bank, Matthew Carter,
Anna-Sophia Cetin, Nakshi Chowdhury, Sean Dowling, Jacob Harris, Aidan Penman, Ewa Pluciennik,
Dmitry Titkov, Michael Thornley, Nicholas Twaddle and Benjamin Ung for the work underlying this
speech. I would also like to thank numerous RBA staff for their helpful comments and suggestions.
The views here are my own. 1 See Kent C (2024), ‘ The Future System for Monetary Policy
Implementation ’, Bloomberg Australia Briefing, Sydney, 2 April. This is
essentially the ‘Friedman rule’ at work – making the supply of money relatively
unconstrained is socially optimal. 2 The opportunity cost of reserves is the spread
between the OMO rate and ES rate. Similar systems have been adopted by the BoE, ECB and RBNZ (as
well as Riksbank albeit from a structural surplus). See Hauser A (2019), ‘Waiting For the
Exit: QT and the Bank of England’s Long-term Balance Sheet’, Speech at the European
Bank for Reconstruction and Development, London, 17 July; BoE (201), ‘The Bank of
England’s Future Balance Sheet and Framework for Controlling Interest Rates’,
Discussion Paper; Schnabel I (2023), ‘Back to Normal? Balance Sheet Size and Interest Rate
Control’, Speech at an Event Organised by Columbia University and SGH Macro Advisors, New
York, 27 March; Schnabel I (2024), ‘The Eurosystem’s Operational Framework’,
Speech at the Money Market Contact Group, Frankfurt am Main, 14 March; and Silk K (2024),
‘Liquidity Management: Principles for Liquidity Provision and the End of an Abundant
Era’, Speech at CBA Global Markets Conference Sydney, 22 October. 3 Their design aims to provide sufficient control of
short-term interest rates, while still encouraging activity in private markets and not exposing
the RBA balance sheet to more risk than necessary. See Kent C (2025), ‘ The RBA’s Monetary Policy Implementation
System – Some Important Updates ’, Speech at KangaNews Debt Capital Markets
Summit, Sydney, 2 April. 4 The ES rate is set 10 basis points below the
cash rate target and the OMO rate 10 basis points above the cash rate target. 5 The cash rate has also risen back to target, but
this is so far off quite narrow activity. Rates on some repos have moved closer to the OMO rate,
including for broader (GC2) collateral and terms of a few weeks. 6 To some extent, the decline is likely to also
reflect a shift in survey wording. Previously, banks were asked about the reserves they
‘expect’ to hold as system reserves reach ample levels. The most recent survey asks
banks their ‘minimum’ and ‘target’ holdings (which inform the lower and upper
estimates of demand presented here, respectively). Even so, the survey results are consistent
with a decline in demand, while market prices, OMO usage, and liaison discussions with banks also
point to the same conclusion. 7 For a discussion of the drivers of reserves demand,
see Bristow L (2024), ‘ Modelling Reserve
Demand with Deposits and the Cost of Collateral ’, RBA Research Discussion Paper No
2024-08. 8 Many central banks have been looking to make sure
their frameworks are robust to a world of higher, but uncertain and variable reserves demand. It
is worth noting that the US Federal Reserve is currently reviewing its own ‘ample’
reserves framework. While similar in name, the Fed operates quite differently to the system in
Australia and several other central banks – with a system where reserves are determined by
the actions of the central bank. See Warsh K (2026), ‘Semiannual Monetary Policy Report to
the Congress’, Testimony to US Committee on Financial Services, 14 July; Anderson AG, A
Barbarino, AM Diercks and S Miran (2026), ‘A User’s Guide to Reducing the Federal
Reserve’s Balance Sheet’, FEDS Paper, March; and for a comparison with full-allotment
systems: Williams JC (2025), ‘Theory and Practice of Monetary Policy Implementation’,
Remarks at the ECB Conference on Money Markets, Frankfurt, 7 November; and Beckworth D
(2025), ‘A New Plumbing System for Central Banks?’, 9 April. 9 There is evidence that extended periods of excess
reserves can lead to market atrophy, facility stigma, rise in banks’ liquidity risk
exposures, resulting in reliance on high levels of reserves. See Acharya VV, RS Chauhan, R Raja
and S Steffen (2024), ‘Liquidity Dependence and the Waxing and Waning of Central Bank
Balance Sheets’, NBER Working Paper No 31050; and Borio C (2023), ‘Getting Up From the
Floor’, BIS Working Paper No 1100. 10 Or demand might prove quite rigid – in which
case there might be little signal that we’re approaching it and conditions could change
quickly. 11 When an AGS bond held by the RBA matures there is
no direct effect on reserves (RBA bond holdings and Australian Office of Financial Management
(AOFM) deposits with the RBA both decline). But there is an indirect drain of reserves when a
replacement bond is issued in the market by the AOFM. When a semi-government bond matures, there
is a direct withdrawal of liquidity as the borrowing authority pays from a commercial bank
account to the RBA. 12 When an AGS bond held by the RBA matures, the AOFM
will raise the funds to repay the RBA by issuing new AGS bonds to private market investors. This
increases the supply of AGS bonds available to the private market. 13 See Kent, n 3. 14 For example, see Schnabel I (2025), ‘Towards
a New Eurosystem Balance Sheet’, Speech at the ECB Conference on Money Markets, Frankfurt am
Main, 6 November; Saporta V (2024), ‘Let’s Get Ready to Repo!’, Speech at
Association for Financial markets in Europe, London, 22 July; and Silk, n 2. 15 Other than cross-currency basis swaps (CCBS),
which we have used in more recent years to obtain foreign currency reserves. 16 While these operations present a low risk to the
RBA, we would consider different exposures when deciding the mix of operations. For example,
short-term AGS involve a small amount of interest rate risk. But shorter-term operations such as
repo or FX swap involve more turnover and so some more operational risk. 17 Open repos are contracted on an ‘open
basis’ – that is, without an agreed maturity date. The interest rate on these open
repos is set at the ES rate, meaning there is no opportunity cost to banks from
obtaining reserves under open repo. The facility is drawn in advance, unlike other standing
facilities. 18 See Debelle G (2013), ‘ The Impact of Payments System and Prudential Reforms
on the RBA’s Provision of Liquidity ’, Address to the AFMA and RBA Briefing,
Sydney, 16 August. 19 See, for example, Dolan S and M Roberts-Sklar
(2025), ‘Learning to Navigate Bumps in the Road’, BoE Insights, 8 December;
Fontaine JS, N Maru and S Tchamova (2026), ‘A Buoy on Funding Tides: How Client Repo Demand
and Dealer Constraints Lifted CORRA’, Bank of Canada Staff Analytical Paper No 2026-15; and
Thedéen E (2025), ‘The Banks Need to Have More Active Liquidity Management’, Speech at
Sveriges Riksbank, Stockholm, 11 September. 20 The intraday facility has also continued to be
used throughout the period of excess reserves, and is used on a daily basis by counterparties. Underlying data You can download the underlying data file
for all data that are available for public release. Some graphs in this article were generated using Mathematica.