The courage not to act: Moving towards a demand-driven operational mode

Bank of Finland stakeholder event on 14 October 2026

Translation of the draft script.

Presentation (pdf)

The courage not to act: Moving towards a demand-driven operational mode

Dear Friends and Colleagues,

Today, the Bank of Finland is publishing three articles in both Finnish and English on a core central bank task: monetary policy implementation. The topic is timely as, following the March 2024 decision of the European Central Bank’s Governing Council, the Eurosystem is moving towards a demand-driven framework for steering short-term money market rates.[1]

The three articles approach the issue from different angles, but they all address the same fundamental question: How will the Eurosystem’s implementation of monetary policy perform in an environment where the financial system is changing and the level of central bank reserves is declining?[2]

To complement these articles, I would like to focus today on a more immediate question: What will the transition to this new way of implementing monetary policy actually look like?

Let me stress at the outset that this is not a discussion about the monetary policy stance. My focus is on the plumbing: how the desired monetary policy stance is implemented in short-term money markets.

And my main message is simple.

A genuinely demand-driven operational framework requires us to allow market prices and market activity to reveal banks’ demand for reserves.

As central bank reserves decline, this is likely to involve some increase in money market rates relative to the deposit facility rate (DFR) and some return of volatility. Neither should automatically be interpreted as a sign that monetary policy implementation is malfunctioning. They may instead be part of the mechanism through which the framework starts to operate as intended.

This creates an important tension for central banks. We must remain ready to intervene when market functioning or policy transmission becomes impaired. But if we smooth every market movement too quickly, we may prevent the very adjustment on which a demand-driven framework relies.

Our challenge is to tell the difference.

The euro area banking system is moving to a new liquidity regime

Former Federal Reserve Chair Ben Bernanke called his memoir The Courage to Act. It describes a crisis in which the Federal Reserve (and of course othercentral banks) had to act forcefully and quickly to prevent the financial system and the economy from collapsing.

Fortunately, the Eurosystem’s transition towards a demand-driven operational mode is taking place under very different circumstances.

But it may require a different kind of courage: the courage not to act prematurely.

That does not mean being passive. We must monitor markets closely, explain our intentions and remain ready to act if interest rate steering or monetary policy transmission becomes impaired.

But not every movement in market rates is a problem for the central bank to solve. Some movements are precisely what we should expect when markets regain a greater role in allocating liquidity.

To see why, let us briefly recall how we arrived here.

When the euro was introduced, the Eurosystem operated a scarce-reserves corridor system and maintained a lean balance sheet. The global financial crisis changed that. We acted with courage and introduced fixed-rate full allotment in our refinancing operations. Later, asset purchases expanded both our balance sheet and the supply of reserves. The result has been a long period of a de facto supply-driven floor system, with abundant excess liquidity keeping short-term money market rates close to, and actually below, the DFR.

Ceteris paribus, that era is gradually coming to an end.

As our monetary policy portfolios run off, excess liquidity is declining. Under the operational mode communicated by the Governing Council in March 2024, the DFR will continue to signal the monetary policy stance, money markets will play a greater role in redistributing liquidity, and banks will be able to meet their reserve needs elastically through our standard refinancing operations.

So we are not returning to the original wide-corridor system. But neither are we staying in the present floor system.

We are moving somewhere in between, to a regime in which market activity will matter more than it has for many years.

Reserves will decline until banks turn to standard refinancing operations on a larger scale

So what should we expect as the transition proceeds?[3]

Today, excess liquidity remains abundant, at around 2.1 trillion euros.

Most banks still hold considerably more reserves than they need. If a bank needs additional liquidity, it can generally obtain it in the money market at a rate close to the DFR. There is therefore little economic incentive to borrow from the Eurosystem at the higher main refinancing operations (MRO) rate.

In that sense, the framework still behaves very much like a supply-driven floor system.

This will change as our monetary policy portfolios continue to run off.

Banks can redistribute reserves among themselves, but they cannot change the aggregate amount of reserves in the banking system. Trading merely moves reserves from one bank to another. At the aggregate level, reserves are created or extinguished through the central bank balance sheet.

As that balance sheet shrinks, reserves become less abundant and banks have stronger incentives to redistribute them actively.

Eventually, we approach a point at which the marginal value of reserves rises, money market rates come under upward pressure, and some banks find it attractive to obtain additional reserves directly from the Eurosystem.

When exactly will that happen, and at what aggregate level of reserves?

That is not the million-euro question from Who Wants to Be a Millionaire? For central banks today, it is more like a trillion-euro question.

Banks’ demand for reserves in the euro area has never really been tested under the regulatory, institutional and financial environment that has emerged since the global financial crisis. Judging by international experience and bank surveys, their demand is likely to be higher than it was before that crisis, reflecting precautionary and payment-related needs as well as liquidity regulation. But there is no fixed quantity called “the banking system’s neutral demand for reserves” waiting to be discovered. Banks differ in their business models, payment flows, funding access, collateral positions and liquidity-risk frameworks. Their demand for reserves also depends on the opportunity cost of holding them and on the price and other terms of obtaining reserves through central bank operations.

Of course, we estimate reserve demand and should continue improving those estimates.

But the uncertainty is too large for us to determine the appropriate size and composition of the Eurosystem balance sheet reliably ex ante.

And this is exactly why the new operational mode is demand-driven.[4]

Our standard refinancing operations are conducted under fixed-rate full allotment. Banks can obtain as much liquidity as they require at the MRO rate, provided they have sufficient eligible collateral. The MRO rate is currently 15 basis points above the DFR.

The framework therefore kills two birds with one stone: it allows banks’ actual behaviour to reveal their demand for reserves, and it allows the Eurosystem to accommodate that demand without having to know it in advance.[5]

But this mechanism works only if two conditions are met.

The first is operational readiness. To borrow from Hamlet, “the readiness is all”. After many years of abundant liquidity, some banks have little recent experience of regular Eurosystem borrowing. So those of you working in bank treasury functions need to remember your tender platform passwords, navigate our somewhat demanding − but secure − login procedures, place a bid and have the right collateral in the right place at the right time.

This can be particularly demanding when mobilising non-marketable assets.

For this reason, the Eurosystem has encouraged banks to participate regularly in standard refinancing operations simply to maintain operational readiness.[6]

I am pleased that many banks have responded. Participation has increased without a material rise in aggregate borrowing, suggesting that much of the participation still reflects preparedness rather than genuine reserve demand.

The second condition is at least as important. It is that banks must regard Eurosystem refinancing operations as a normal part of day-to-day liquidity management and feel entirely comfortable using them whenever they are economically attractive. This has been communicated clearly both by the monetary policy side of the Eurosystem and by European banking supervision.

In the new framework, regular central bank borrowing should not in itself be interpreted as a sign of  bank stress. It is an intended part of normal liquidity management.[7]

Individual banks will naturally make different choices. Some will rely more on Eurosystem credit; others more on private markets.

But at the aggregate level, the banking system will ultimately need to obtain reserves from the Eurosystem. Regular use of our refinancing operations should therefore become part of the normal state of affairs.

That is a genuine regime change in liquidity management.

And banks will make that change when relative prices give them a reason to do so.

Different money market rates tell different stories

This brings me to short-term rate steering.

Today, unsecured overnight rates such as the euro short-term rate (€STR) trade below the DFR.

At the same time, many secured money market rates − repo rates − have risen closer to the DFR, and at times slightly above it.

At first sight, this may appear counterintuitive.

How can it be cheaper to borrow without collateral than with collateral?

The answer is that the two markets price different things and involve different participants.

The €STR reflects unsecured overnight borrowing by banks, much of it from money market funds and other non-bank financial institutions that cannot deposit directly with the Eurosystem.

When reserves are abundant, banks do not need to compete aggressively for these deposits. They may also pass some of the balance-sheet and regulatory costs of accepting them back to their counterparties.

This helps explain why the €STR can trade below the DFR.

Repo transactions are different. They exchange cash against securities, so repo rates reflect not only the supply and demand for cash but also the value and availability of collateral.

Their participants also differ. Hedge funds, for example, are important borrowers in repo markets, for purposes that include financing government bond positions.

The recent increase in repo rates therefore reflects several factors: declining reserves, greater availability of collateral, and the particular composition and incentives of borrowers and lenders in secured markets.

In the past, the easing of collateral scarcity may have dominated. Going forward, declining reserve abundance may become increasingly important.

So the divergence between secured and unsecured overnight rates is not in itself a market anomaly.

Different rates tell different stories. There is no reason to expect them always to coincide.

The relevant questions are whether we understand the economic forces behind rate movements and whether the markets in which they are formed continue to function well.

For the moment, the answer to both is yes.

Money market rates must adjust before demand for refinancing operations rises

As reserves decline further, banks will make greater efforts to redistribute them among themselves in money markets. This should put upward pressure on at least some money market rates. With the MRO rate currently 15 basis points above the DFR, central bank borrowing will eventually become economically attractive relative to market alternatives.

The Eurosystem’s broad collateral framework matters here.

A bank borrowing against non-high-quality liquid assets (non-HQLA) that are eligible for use as collateral can transform those assets into central bank reserves, thereby improving its liquidity coverage ratio (LCR) position.[8] MRO borrowing may therefore become attractive before the MRO rate is reached by repo rates backed by HQLA collateral.

Different banks will reach this point of MRO borrowing at different times.

Their funding costs, collateral positions, operational costs, internal liquidity rules and regulatory constraints all differ.

We should therefore expect a gradual transition, instead of a rainy Tuesday morning on which everyone suddenly remembers that the MRO exists.

Nor do we know exactly how much rates need to adjust before meaningful demand for central bank credit emerges.

And that uncertainty is not a flaw in the framework.

The central bank does not need to target the precise configuration of money market rates at which banks begin to use its refinancing operations. Individual rates reflect differences in liquidity and collateral demand, counterparties, balance-sheet costs and regulatory incentives. The resulting spreads are the market price of those differences. As long as movements remain gradual, understandable and consistent with functioning markets, there is no reason for the central bank to eliminate those spreads or steer every rate to a particular level relative to its policy rates.

The DFR anchors the monetary policy stance. This does not imply that every overnight rate must trade exactly at the DFR. Indeed, some adjustment is necessary for the framework to become genuinely demand-driven.

Banks will increase their use of refinancing operations when market conditions make central bank borrowing sufficiently attractive. Their behaviour will reveal the demand for reserves, and fixed-rate full allotment will allow the supply of reserves to respond.[9]

If we prevent market rates from adjusting before this mechanism can operate, we risk short-circuiting the very framework we have chosen.

This is why the transition is also likely to involve somewhat greater volatility than we have been used to.

The floor system has compressed money market volatility to unusually low levels. As reserves become less abundant and private markets regain a larger role in redistributing liquidity, some volatility should return. After years of being close to The Sound of Silence in money markets, a little more noise need not be a cause for concern. It may simply mean that markets are doing more of the work. The relevant distinction is therefore not between volatility and no volatility, but between volatility that reflects orderly price discovery and volatility that signals impaired market functioning or impaired monetary policy transmission.

Quarter ends provide a useful example.

Quarter-end repo volatility may increase as dealer banks reduce balance-sheet-intensive intermediation at regulatory reporting dates.[10] When such movements are temporary, anticipated and attributable to balance-sheet or regulatory constraints rather than aggregate reserve scarcity, they need not trigger a central bank response. By contrast, recurrent sharp spikes, persistent fragmentation or unstable transmission from policy rates to broader funding conditions would be more concerning.

The central question is therefore diagnostic: Are we observing orderly price discovery, or impaired monetary policy transmission?[11]

Which money market variables should we monitor?

To answer that question, we need to look at more than one rate, and at more than rates alone.

I would highlight three dimensions.

First, developments across secured and unsecured overnight markets.

The €STR and repo rates need not move together, but their movements should remain explainable and orderly.

Second, rates should be assessed together with volumes, dispersion and participation in central bank operations.

Effective liquidity redistribution normally requires meaningful trading volumes. Rising dispersion accompanied by declining volumes could point to fragmentation or impaired market functioning.

Participation in Eurosystem refinancing operations also indicates to us, initially, whether banks are operationally ready, and over time it should increasingly reveal genuine reserve demand.

Third, we need to monitor transmission beyond the overnight segment.

Overnight indexed swaps, Euribor rates and broader bank-funding indicators help us assess whether changes in short-term money markets are transmitted predictably along the yield curve and ultimately to the real economy.

These are not operating targets for the Eurosystem, but are important indicators of policy transmission.

And Euribor rates are naturally of particular interest to us at the Bank of Finland, given their central role in pricing Finnish mortgages.

An orderly transition therefore does not mean unchanged rates. It means understandable rate movements, functioning markets, effective liquidity redistribution and predictable monetary policy transmission.

What, then, should the Eurosystem do if volatility increases sufficiently to raise concerns?

Broadly speaking, there are three possible responses.

The first is to add liquidity through fine-tuning operations. These can be appropriate in the face of temporary and severe disturbances. But repeated discretionary liquidity injections would weaken incentives to redistribute liquidity or use standard refinancing operations. At the extreme, this could prevent the framework from becoming genuinely demand-driven.

Launching structural liquidity operations too early could create a similar problem: it could reduce the role of standard refinancing operations in revealing banks’ underlying demand for reserves.[12]

A second possibility would be to narrow the interest-rate corridor.

That would limit the potential range of short-term rates, but it would also reduce the economic incentives for banks to redistribute reserves in private markets.

The third option is to tolerate some volatility and allow the adjustment mechanism to work. Higher relative market funding costs strengthen banks’ incentives to trade liquidity among themselves and, when appropriate, to borrow from the Eurosystem.

That is precisely how reserve demand is supposed to reveal itself.

Sometimes the appropriate response is simply to – as the Beatles put it − let it be, at least for a while as the market adjusts.

In my view, adopting a demand-driven framework implies greater tolerance for endogenous movements in market rates than under a supply-driven floor system. That is not an incidental feature of the framework but in fact the cornerstone of its very logic.

The more quickly the central bank smooths every market movement, the less room there is for the adjustment on which a demand-driven framework relies.

This calls for patience when movements are temporary and understandable, and consistent with functioning markets.

It calls for action when movements become disorderly, markets fragment, or the transmission of the intended monetary policy stance becomes durably impaired.

That is the line we need to draw.

Conclusion

Let me conclude.

The Eurosystem balance sheet is shrinking. We do not know exactly when this transition will be complete, because we don’t know the level of central bank liquidity that banks will ultimately choose to hold.

Nor do we need to know that in advance.

The operating mode is designed precisely to deal with this uncertainty by allowing reserve demand to reveal itself through banks’ behaviour and by accommodating that demand through our refinancing operations.

For this mechanism to work, however, market conditions must be allowed to adjust.

As reserves become less abundant, some money market rates are likely to rise relative to the DFR, and volatility is likely to increase from the exceptionally low levels of recent years. This should not automatically be viewed as evidence that the framework is failing; it may be entirely consistent with the transition proceeding as intended.

The relevant threshold for central bank action is therefore not whether rates move. It is whether those movements become disorderly and threaten market functioning or the effective transmission of monetary policy.

The Eurosystem must remain alert and ready to respond if and when that happens.

But trying to suppress every increase in rates or every episode of volatility would carry its own risk: the operational framework might never complete its transition to a genuinely demand-driven state.

Ben Bernanke wrote about the courage to act in a crisis.

The transition ahead of us may require a different kind of courage:

the courage to monitor carefully, diagnose correctly, tolerate some uncertainty − and allow money markets and our refinancing operations to do their job.

Footnotes

[1] See the ECB Governing Council’s statement of 13 March 2024, a speech by ECB Executive Board member Isabel Schnabel, and a 14 March 2024 blog post (in Finnish) by Member of the Board of the Bank of Finland Tuomas Välimäki.

[2] The first article, by Herrala, Paavola and Tötterman, examines how the Eurosystem can steer interest rates in its new demand-driven operational framework. The second article, by Ilmanen, Palmén and Rannikko, discusses the choices and tradeoffs involved in designing the structural liquidity provision instruments that will eventually support the framework. The third article, by Järvinen, Koukkunen and Liikkanen, outlines how emerging technological developments, such as tokenization and distributed ledger technologies, may affect payment systems and monetary policy implementation.

[3] The Eurosystem is continuously monitoring the transition. See, for example, the following ECB blog posts: The first year of the Eurosystem’s new operational framework, posted on 25 April 2025, and How banks are adjusting to declining reserves, posted on 2 April 2026.

[4] See Schnabel, I. (2024), “The Eurosystem’s operational framework”, speech at the Money Market Contact Group meeting, Frankfurt am Main, 14 March 2024.

[5] See Välimäki, T. (2025), “A new dawn, a new day, a new life for monetary policy implementation”, opening keynote at the Bank of Finland–SUERF Conference on Monetary Policy Implementation: Old Wisdoms and New Trends, Helsinki, 11 June 2025.

[6] See, for example, the blog post (in Finnish) by Sivi Palmén, posted on 12 May 2026.

[7] See the blog post of 18 March 2025 by Claudia Buch, Chair of the Supervisory Board of the ECB, and Isabel Schnabel, member of the Executive Board of the ECB: Managing liquidity in a changing environment.

[8] See Schnabel, I. (2025), “No longer convenient? Safe asset abundance and r*”, speech at the Bank of England BEAR conference, London, 25 February 2025.

[9] See Schnabel, I. (2025), “Towards a new Eurosystem balance sheet”, speech at the ECB Conference on Money Markets, Frankfurt am Main, 6 Novermber 2025.

[10] See European Central Bank (2025), Euro money market study 2024, in particular Section 2.1.4 “Maturities and calendar effects”.

[11] See, for example, Holm-Hadulla and Pool (2025) and a related ECB blog post: Low money market volatility benefits monetary policy transmission, posted on 18 April 2025.

[12] See Schnabel, I. (2025), “Towards a new Eurosystem balance sheet”, speech at the ECB Conference on Money Markets, Frankfurt am Main, 6 Novermber 2025.

References

Bernanke, B. S. (2015), The Courage to Act: A Memoir of a Crisis and Its Aftermath, W. W. Norton & Company, New York.

Buch, C. and Schnabel, I. (2025), “Managing liquidity in a changing environment”, The ECB Blog, European Central Bank, 18 March 2025.

European Central Bank (2024), “Changes to the operational framework for implementing monetary policy”, statement by the Governing Council, 13 March 2024.

European Central Bank (2025), Euro Money Market Study 2024, April 2025.

Hartung, B., Linzert, T., Rahmouni-Rousseau, I., Schneider, Y. and Skrzypińska, M. (2025), “The first year of the Eurosystem’s new operational framework”, The ECB Blog, European Central Bank, 25 April 2025.

Holm-Hadulla, F. and Pool, S. (2025a), “Interest rate control and the transmission of monetary policy”, ECB Working Paper Series, No 3048, European Central Bank, April 2025.

Holm-Hadulla, F. and Pool, S. (2025b), “Low money market volatility benefits monetary policy transmission”, The ECB Blog, European Central Bank, 18 April 2025.

Iskaki, V., Linzert, T., Schneider, Y., Skrzypińska, M. and Vergote, O. (2026), “How banks are adjusting to declining reserves”, The ECB Blog, European Central Bank, 2 April 2026.

Palmén, S. (2026), “Pankkeja kehotetaan osallistumaan keskuspankkien luotto-operaatioihin keskuspankkitalletusten vähentyessä”, Bank of Finland Bulletin blog, Bank of Finland, 12 May 2026.

Schnabel, I. (2024), “The Eurosystem’s operational framework”, speech at the Money Market Contact Group meeting, Frankfurt am Main, 14 March 2024.

Schnabel, I. (2025a), “No longer convenient? Safe asset abundance and r*”, keynote speech at the Bank of England’s 2025 BEAR Conference, London, 25 February 2025.

Schnabel, I. (2025b), “Towards a new Eurosystem balance sheet”, speech at the ECB Conference on Money Markets 2025, Frankfurt am Main, 6 November 2025.

Välimäki, T. (2024), “EKP uudisti rahapolitiikan toimeenpanon menettelyjä”, Euro & talous blog, Bank of Finland, 14 March 2024.

Välimäki, T. (2025), “A new dawn, a new day, a new life for monetary policy implementation”, opening keynote at the Bank of Finland–SUERF Conference on Monetary Policy Implementation: Old Wisdoms and New Trends, Helsinki, 11 June 2025.