When liquidity stops being liquid
28 April 2026
In the third instalment of this ongoing series, Cyril Louchtchay de Fleurian, head of securities finance and balance sheet strategy at Capteo: Strategy & Management Consulting, on gatekeepers, collateral circulation, and excess reserves at the core of the new liquidity regime
Image: stock.adobe.com/ivandanru
Liquidity is only as strong as the chain behind it
First constraint: infrastructure. In stress, average volumes are a poor proxy for liquidity. What matters is what you can actually get done 鈥� here and now: size, speed, price.
It all comes down to execution capacity: how much risk can be intermediated, how fast, and at what cost, through players and infrastructures that are themselves balance sheet, funding, and margin constrained. You do not see that in the dashboards and KPIs. Liquidity runs through an intermediation chain and that chain tightens under stress.
Investment banks make markets and transform collateral. Dealers provide liquidity by putting balance sheet at risk; their capacity is a direct function of secured funding and balance sheet constraints. The European Central Bank (ECB) has been explicit on the market/funding link and the role of dealer balance sheets, with repo sitting at the core as the funding layer for inventories. In practice, it all comes down to how much balance sheet a small set of players can actually deploy at a given point in time.
Hedge funds run on prime financing, with collateral constantly in motion. The package is standard: position financing (repo, margin lending), clearing, settlement, collateral management. In that setup, margining and haircuts turn procyclical very quickly: they rise when volatility picks up and liquidity deteriorates, driving higher collateral needs and forced selling, exactly when the system starts to seize.
Central clearing adds another transmission channel through margin calls and clearing member liquidity needs. Market infrastructures are not neutral pipes; they shape how liquidity actually behaves. At the start of the day on Target2-Securities (T2S), banks need to pre-position securities and/or cash on their securities accounts and dedicated cash accounts (DCAs) with international central securities depositories (ICSDs)/CSDs to build start-of-day liquidity, ensuring immediate intraday settlement.
Liquidity is driven by three constraints 鈥� all at once. First, the intermediation chain, which sets the real conditions of access and execution. Second, collateral circulation, whose velocity drives system elasticity. Third, the excess reserves regime, which shifts the equilibrium. Together, they define the new liquidity regime.
You can already see it in the market. Holding collateral is no longer enough to guarantee liquidity: capacity depends on depth, intermediation, and endogenous margin and haircut constraints. Stress does not build linearly anymore; it comes in bursts, driven by margin calls and balance sheet pressure. Liquidity does not exist in isolation either: it sits in a coupled system 鈥� market, funding, collateral 鈥� where dealers, prime brokers, central counterparties (CCPs), and infrastructures jointly drive feedback loops and contagion.
When the chain tightens, it all comes down to collateral velocity. At that point, is it still a value chain 鈥� or already a kill chain? More on that later.
Collateral only matters at the speed it moves
Second constraint: collateral circulation. The same asset can be reused multiple times along the intermediation chain. That reuse defines collateral velocity. The key proxy is the re-use rate: how smoothly collateral moves across players, jurisdictions, and infrastructures, and how fast it can be turned into executable liquidity. That speed is critical: it determines whether the system absorbs shocks or amplifies them.
Collateral velocity has fallen sharply since 2008.
Three forces have slowed it down. Post-crisis regulation has locked up more collateral (margin segregation, buffers, balance sheet constraints). Infrastructure fragmentation (CCPs, triparty, CSDs/ICSDs, jurisdictions) has reinforced silos and reduced mobility. And concentration of collateral on certain balance sheets, especially public sector, has taken a share of the most reusable assets out of circulation: 20鈥�25 per cent of European government bonds still sit with the Eurosystem as of end-2025.
Bottom line: there is collateral 鈥� just not where it is needed. The system holds plenty of collateral, but less of it is mobilisable, circulating, and usable when it matters. The repo market makes that visible: by end-2024, around 80 per cent of European sovereigns were trading through repo below the deposit facility rate (source: MMSR, ECB 2025). The rest follows: higher and more volatile funding costs, tighter substitution, and greater reliance on a handful of monetisation channels and dominant players. Fail rates become non-linear, more a sign of system stress than operational issues.
Lower velocity shows up quickly in the market: liquidity pools fragment, choke points emerge (liquidity bottlenecks, effectively toll gates), and speed asymmetries widen. Some names, assets, or currencies remain liquid; others quickly lock up. As re-use falls, the same collateral supports less intermediation: repo capacity shrinks, and pressure builds on repo rates, spreads, and haircuts.
The system can look liquid on aggregate but be tight locally. Banks can show strong high-quality liquid asset (HQLA) buffers and still be vulnerable if assets are encumbered, poorly located, not transferable in time, or not eligible in the right channel. Stress is less about collateral scarcity and more about breaks in circulation: haircuts, margin calls, roll-over stress, and congestion create fast, non-linear dislocations, often invisible in prudential metrics until execution itself fails.
Collateral velocity
Source: IMF 鈥撀燙ollateral re-use and balance sheet space聽(Singh, 2023); Bank of England/market data
Velocity dropped from 3.0 to 1.0 between 2017 and 2021, effectively wiping out around 鈧�1 trillion of collateral per point, roughly 20 per cent of current European repo outstanding (鈧�12 trillion in 2025). That is significant! We simply recycle less collateral than before. Less recycling means less or worse funding. The same pool of collateral now generates less liquidity, and the system becomes more dependent on repo intermediation, infrastructures, and central banks. That is the trap: you can have plenty of reserves and collateral and still be fragile, because it does not move fast enough.
For banks, this is day-to-day reality: collateral holdings, eligibility, and circulation all have to work together. It is about mobilising, allocating, and routing collateral through the right channels. Liquidity is a function of re-use: mobilise, transform, reallocate, and execute within a tight operational window.
Excess reserves have shifted the liquidity equilibrium
Third constraint: excess reserves. Liquidity is now more conditional, more discontinuous, and more sensitive to collateral frictions because the production regime has changed. In less than 20 years, the system has moved from near-zero excess reserves to structurally abundant central bank liquidity.
Source: ECB and Fed 2025
What that means on the ground.
Repo becomes the engine of a circulation-driven liquidity: funding positions, financing inventories, and anchoring non-bank financial institutions (NBFI) intermediation. The European repo market has doubled between 2015 and 2025 (from 鈧�5.5 trillion to 鈧�12 trillion 鈥� source: ICMA/ERCC).
This shift changes the landscape. Safe assets become the third scarce resource for banks, alongside balance sheet and liquidity coverage ratio (LCR). Collateral management moves from support to core function. Infrastructures act as both toll gates and congestion points. Liquidity fragments 鈥� abundant globally, tight locally. It becomes procyclical: haircuts, margins, and roll-over dynamics amplify shocks. Central banks now steer liquidity as much through repo as through reserves.
In an excess reserves regime, liquidity is largely a stock of central bank money. In that environment, collateral velocity dropped from 3.0 to 1.0 (2017鈥�2021), in line with peak quantitative easing (QE). The 2021-22 period marks the trough, with velocity around 1.0鈥�1.5. As excess reserves build up, the need to mobilise collateral falls, re-use chains shorten, and velocity compresses.
From 2022, with the shift to quantitative tightening (QT), central bank balance sheets shrink, policy normalises, and excess liquidity is gradually withdrawn. That is when velocity stops falling, rebounds, and roughly doubles to 2.0鈥�2.5 in 2023. As reserves decline, banks rely more on market funding and repo becomes the adjustment mechanism. Net-net: pressure on collateral circulation is back.
In a repo-driven system, liquidity depends on the ability to mobilise collateral. It becomes a function of circulation rather than money supply. In the current normalisation phase, higher velocity is the base case 鈥� and with it, higher system pressure.
This is not just a monetary cycle 鈥� it is a regime shift
The build-up and unwind of excess reserves have not just moved liquidity around; they have changed its nature. In this regime, equilibria are more fragile. The same collateral pool can generate very different liquidity depending on velocity, location, and intermediary constraints. Liquidity is no longer a system property; it is a function of access conditions and gatekeepers.
Stress no longer builds gradually; it is non-linear and requires constant BAU readiness. Liquidity does not disappear: it turns intermittent.
More fundamentally, liquidity is now about timing as much as volume. Mobilising an asset is no longer just about quality or eligibility; it is about whether it can be activated within a tight window, exactly when needed. Misalignments become critical: between holding and access, between need and mobilisation, between asset deterioration and monetisation. That is where the sharpest dislocations now form. Liquidity no longer breaks on volume alone; it breaks on timing mismatches.
First constraint: infrastructure. In stress, average volumes are a poor proxy for liquidity. What matters is what you can actually get done 鈥� here and now: size, speed, price.
It all comes down to execution capacity: how much risk can be intermediated, how fast, and at what cost, through players and infrastructures that are themselves balance sheet, funding, and margin constrained. You do not see that in the dashboards and KPIs. Liquidity runs through an intermediation chain and that chain tightens under stress.
Investment banks make markets and transform collateral. Dealers provide liquidity by putting balance sheet at risk; their capacity is a direct function of secured funding and balance sheet constraints. The European Central Bank (ECB) has been explicit on the market/funding link and the role of dealer balance sheets, with repo sitting at the core as the funding layer for inventories. In practice, it all comes down to how much balance sheet a small set of players can actually deploy at a given point in time.
Hedge funds run on prime financing, with collateral constantly in motion. The package is standard: position financing (repo, margin lending), clearing, settlement, collateral management. In that setup, margining and haircuts turn procyclical very quickly: they rise when volatility picks up and liquidity deteriorates, driving higher collateral needs and forced selling, exactly when the system starts to seize.
Central clearing adds another transmission channel through margin calls and clearing member liquidity needs. Market infrastructures are not neutral pipes; they shape how liquidity actually behaves. At the start of the day on Target2-Securities (T2S), banks need to pre-position securities and/or cash on their securities accounts and dedicated cash accounts (DCAs) with international central securities depositories (ICSDs)/CSDs to build start-of-day liquidity, ensuring immediate intraday settlement.
Liquidity is driven by three constraints 鈥� all at once. First, the intermediation chain, which sets the real conditions of access and execution. Second, collateral circulation, whose velocity drives system elasticity. Third, the excess reserves regime, which shifts the equilibrium. Together, they define the new liquidity regime.
You can already see it in the market. Holding collateral is no longer enough to guarantee liquidity: capacity depends on depth, intermediation, and endogenous margin and haircut constraints. Stress does not build linearly anymore; it comes in bursts, driven by margin calls and balance sheet pressure. Liquidity does not exist in isolation either: it sits in a coupled system 鈥� market, funding, collateral 鈥� where dealers, prime brokers, central counterparties (CCPs), and infrastructures jointly drive feedback loops and contagion.
When the chain tightens, it all comes down to collateral velocity. At that point, is it still a value chain 鈥� or already a kill chain? More on that later.
Collateral only matters at the speed it moves
Second constraint: collateral circulation. The same asset can be reused multiple times along the intermediation chain. That reuse defines collateral velocity. The key proxy is the re-use rate: how smoothly collateral moves across players, jurisdictions, and infrastructures, and how fast it can be turned into executable liquidity. That speed is critical: it determines whether the system absorbs shocks or amplifies them.
Collateral velocity has fallen sharply since 2008.
Three forces have slowed it down. Post-crisis regulation has locked up more collateral (margin segregation, buffers, balance sheet constraints). Infrastructure fragmentation (CCPs, triparty, CSDs/ICSDs, jurisdictions) has reinforced silos and reduced mobility. And concentration of collateral on certain balance sheets, especially public sector, has taken a share of the most reusable assets out of circulation: 20鈥�25 per cent of European government bonds still sit with the Eurosystem as of end-2025.
Bottom line: there is collateral 鈥� just not where it is needed. The system holds plenty of collateral, but less of it is mobilisable, circulating, and usable when it matters. The repo market makes that visible: by end-2024, around 80 per cent of European sovereigns were trading through repo below the deposit facility rate (source: MMSR, ECB 2025). The rest follows: higher and more volatile funding costs, tighter substitution, and greater reliance on a handful of monetisation channels and dominant players. Fail rates become non-linear, more a sign of system stress than operational issues.
Lower velocity shows up quickly in the market: liquidity pools fragment, choke points emerge (liquidity bottlenecks, effectively toll gates), and speed asymmetries widen. Some names, assets, or currencies remain liquid; others quickly lock up. As re-use falls, the same collateral supports less intermediation: repo capacity shrinks, and pressure builds on repo rates, spreads, and haircuts.
The system can look liquid on aggregate but be tight locally. Banks can show strong high-quality liquid asset (HQLA) buffers and still be vulnerable if assets are encumbered, poorly located, not transferable in time, or not eligible in the right channel. Stress is less about collateral scarcity and more about breaks in circulation: haircuts, margin calls, roll-over stress, and congestion create fast, non-linear dislocations, often invisible in prudential metrics until execution itself fails.
Collateral velocity
Source: IMF 鈥撀燙ollateral re-use and balance sheet space聽(Singh, 2023); Bank of England/market data
Velocity dropped from 3.0 to 1.0 between 2017 and 2021, effectively wiping out around 鈧�1 trillion of collateral per point, roughly 20 per cent of current European repo outstanding (鈧�12 trillion in 2025). That is significant! We simply recycle less collateral than before. Less recycling means less or worse funding. The same pool of collateral now generates less liquidity, and the system becomes more dependent on repo intermediation, infrastructures, and central banks. That is the trap: you can have plenty of reserves and collateral and still be fragile, because it does not move fast enough.
For banks, this is day-to-day reality: collateral holdings, eligibility, and circulation all have to work together. It is about mobilising, allocating, and routing collateral through the right channels. Liquidity is a function of re-use: mobilise, transform, reallocate, and execute within a tight operational window.
Excess reserves have shifted the liquidity equilibrium
Third constraint: excess reserves. Liquidity is now more conditional, more discontinuous, and more sensitive to collateral frictions because the production regime has changed. In less than 20 years, the system has moved from near-zero excess reserves to structurally abundant central bank liquidity.
Source: ECB and Fed 2025
What that means on the ground.
Repo becomes the engine of a circulation-driven liquidity: funding positions, financing inventories, and anchoring non-bank financial institutions (NBFI) intermediation. The European repo market has doubled between 2015 and 2025 (from 鈧�5.5 trillion to 鈧�12 trillion 鈥� source: ICMA/ERCC).
This shift changes the landscape. Safe assets become the third scarce resource for banks, alongside balance sheet and liquidity coverage ratio (LCR). Collateral management moves from support to core function. Infrastructures act as both toll gates and congestion points. Liquidity fragments 鈥� abundant globally, tight locally. It becomes procyclical: haircuts, margins, and roll-over dynamics amplify shocks. Central banks now steer liquidity as much through repo as through reserves.
In an excess reserves regime, liquidity is largely a stock of central bank money. In that environment, collateral velocity dropped from 3.0 to 1.0 (2017鈥�2021), in line with peak quantitative easing (QE). The 2021-22 period marks the trough, with velocity around 1.0鈥�1.5. As excess reserves build up, the need to mobilise collateral falls, re-use chains shorten, and velocity compresses.
From 2022, with the shift to quantitative tightening (QT), central bank balance sheets shrink, policy normalises, and excess liquidity is gradually withdrawn. That is when velocity stops falling, rebounds, and roughly doubles to 2.0鈥�2.5 in 2023. As reserves decline, banks rely more on market funding and repo becomes the adjustment mechanism. Net-net: pressure on collateral circulation is back.
In a repo-driven system, liquidity depends on the ability to mobilise collateral. It becomes a function of circulation rather than money supply. In the current normalisation phase, higher velocity is the base case 鈥� and with it, higher system pressure.
This is not just a monetary cycle 鈥� it is a regime shift
The build-up and unwind of excess reserves have not just moved liquidity around; they have changed its nature. In this regime, equilibria are more fragile. The same collateral pool can generate very different liquidity depending on velocity, location, and intermediary constraints. Liquidity is no longer a system property; it is a function of access conditions and gatekeepers.
Stress no longer builds gradually; it is non-linear and requires constant BAU readiness. Liquidity does not disappear: it turns intermittent.
More fundamentally, liquidity is now about timing as much as volume. Mobilising an asset is no longer just about quality or eligibility; it is about whether it can be activated within a tight window, exactly when needed. Misalignments become critical: between holding and access, between need and mobilisation, between asset deterioration and monetisation. That is where the sharpest dislocations now form. Liquidity no longer breaks on volume alone; it breaks on timing mismatches.
NO FEE, NO RISK
100% ON RETURNS If you invest in only one securities finance news source this year, make sure it is your free subscription to Securities 麻豆影视传媒 Times
100% ON RETURNS If you invest in only one securities finance news source this year, make sure it is your free subscription to Securities 麻豆影视传媒 Times
