The weapon that never fires
23 June 2026
In the fifth instalment of this ongoing series, Cyril Louchtchay de Fleurian, head of securities finance and balance sheet strategy at Capteo: Strategy & Management Consulting, looks at repo, collateral, and the chokepoints nobody put on the risk committee agenda
Image: stock.adobe.com/Nart
Liquidity is execution. Execution is resilience. And resilience, it turns out, is now a chokepoint — owned, gated, and deliberately opaque.
Repo, collateral, clearing, market infrastructures. Access points. Filters. Pressure levers. In the right hands — or the wrong ones — they function as sanction tools, scarcity engines, hard stops. The risk has not just moved. It has mutated. No longer technical. No longer prudential. Strategic now. Exogenous. Sometimes hostile. These instruments are no longer neutral. They were never truly neutral.
I have argued in this series that the liquidity regime has changed. That argument was too modest. The gap is not a regime shift. It is existential — and it is already open.
The traditional view of repo — technical tool, apolitical, interchangeable — rests on one silent assumption: that any funding channel can always be substituted for another. That assumption is broken. Repo is the infrastructure layer on which collateral transformation, funding price formation, and liquidity transmission all depend. Banks use it to run their balance sheets and buffers. Dealers use it to provide intermediation. Hedge funds use it to carry risk. Central banks use it to inject and sterilise. Market infrastructures use it to enforce discipline.
Disrupt the repo market, and you break the flow that keeps the whole system alive. Once access becomes conditional, whoever controls the channels — the eligibility rules, the chokepoints — holds the control. Potentially a systemic weapon.
You think you own it. Read the small print
An asset is liquid only where it sits — and within the infrastructure where it is held. I think of liquidity in three dimensions: the nature of the collateral, the nature of the underlying legal contract, and the nature of the processing and settlement infrastructure. In stress, executable liquidity depends less on accounting classification — high-quality liquid asset (HQLA) — or on contract type, than on the operational mechanics: which custodian, which triparty agent or central counterparty (CCP), under which law, within which cut-off window, with what mobilisation lag. The Liquidity Coverage Ratio (LCR) does not see this geography. Yet this is precisely the geography that creates points of failure.
Location matters because mobility is neither free nor instant. Europe has long suffered from fragmentation in clearing and settlement. Cross-border settlement costs run on average 65 per cent higher than in the US — a structural symptom of a post-trade market that remains fragmented, expensive, and operationally inefficient. Even after years of harmonisation, the authorities themselves acknowledge that work remains. Standardisation does not eliminate friction. AFME — October 2025, ‘Analysis of CSD fees in major European markets’.
TARGET2-Securities (T2S) lowered barriers. It did not solve trapped collateral. The platform now connects 24 central securities depositories (CSDs) and processes large settlement volumes — a genuine improvement in technical integration. But two structural facts shape executable liquidity. First: cross-CSD activity remains a small fraction of total flows. Operations are still largely organised in CSD silos, even where the technology is common. In 2024, cross-CSD transactions accounted for 3.8 per cent of T2S volumes. Second: even on centralised infrastructure, settlement efficiency is not 100 per cent. In 2024, end-of-day efficiency stood at 94.4 per cent by volume. Roughly 5.5 per cent of volumes did not settle by EOD — spilling into fails, queues, and delays, with direct consequences for collateral mobilisation and intraday management. All this is according to the ECB Target Services Annual Report 2024.
Collateral that is unsettled, misaligned, or stranded in the wrong location is not mobilisable — regardless of its credit quality. The challenge is the ability to locate the security and mobilise it in time through the execution chain. What most risk managers still miss: owning an asset and controlling it are two different things. The gap between the two is exactly where executable liquidity disappears — quietly, without a warning signal.
That gap shows up in very specific places. It sits between the local CSD and the international central securities depository. Between a house account and a triparty account. Between segregation and encumbrance. It widens with each instrument choice — bilateral repo, synthetic repo, cleared triparty, securities lending, buy and sell-back, central bank facility. It extends with the governing documentation — Global Master Repurchase Agreement, Credit Support Annex, eligibility schedules, substitution conditions — which determine what you can do, how fast, and with whom. It closes in around cut-off windows, settlement cycles, realignment frictions. And it becomes a trap in post-trade fails and queues.
The real questions are simple: who and what are you dependent on? Is that dependency chosen or implicit? Running a liquidity buffer without managing how your dependencies actually work is like switching on a generator without checking whether it has fuel. The collateral is yours. The question is — from when?
Pick your infrastructure. Pick your master
No execution venue choice is neutral. It determines jurisdiction, eligibility rules, margin models, default rights, and exposure to sovereign-level decisions — sanctions, ring-fencing, local collateral requirements. Infrastructure selection is a regime choice. Most people in the room do not want to say that out loud. Does it take getting blocked before you ask the question?
If repo access is silo-dependent, then repo is far more than a funding tool. It is a chokepoint. And a single chokepoint creates filtering capacity and constraint — over who can execute, at what speed, with what collateral, and under which law. That is the operational definition of a control lever.
Your ability to mobilise collateral, raise cash, and hold positions depends on constraints you do not fully control. Power here is not exercised through prohibition. It is exercised through parameter adjustment. Under stress, those parameters determine who stays in the market and who is forced out. Understanding a silo is not about mapping its features. It is about identifying where it constrains, when it breaks, and how far it exposes you to an external decision — regulatory, political, or purely mechanical.
All of the above can be tested. Not in a committee. Not in a stress test. On an ordinary Tuesday morning.
08:47. Your desk has a margin call due at 09:00. Right in the window where two-thirds of repo-margin-CCP activity concentrates. The collateral is there — on paper. Wrong custodian. Wrong governing law. Wrong cut-off window. In 13 minutes, you will find out exactly who and what you depend on.
08:48. Cut-offs, settlement windows: that is your first problem. ‘Usable now’ does not mean the same thing across chains and silos. Next up: cross-product netting rules — and they vary. Without netting, your 5–10 basis point spread is gone. Not every CCP allows netting of repo exposures against derivatives on the same underlying. Euronext Clearing and Eurex Clearing offer this on certain segments. LCH less so. That cross-margining gap is a real, recurring capital cost — independent of transaction pricing, and entirely a function of which silo you are in. Silo dependency also forces you to carry extra liquidity buffer to absorb operational friction: missed cut-offs, refused substitutions, unanticipated margin calls. That over-buffering has a real annual cost. It is rarely calculated. Never reported as such. Its value is destroyed silently, every quarter. Time-to-cash and opportunity cost are the right questions to ask before you are in this situation. Which intraday windows turn theoretical liquidity into unusable liquidity? What does your current silo dependency actually cost in over-buffering and netting leakage? This test is the most directly P&L-linked — and the one least often run.
08:49. Eligibility hits you next. Eligibility schedules are not identical across silos. Under time pressure, they are barely interoperable. An asset that is mobilisable in one silo can be downgraded in another or hit with a steeper haircut. What is repo-eligible at Euroclear is not necessarily eligible at Clearstream or in a BNY triparty basket. The choice of triparty agent or custodian — European or non-eurozone — changes the risk profile entirely: real execution capacity on one side; platform reliance, legal exposure, compliance constraints, and potential extraterritoriality on the other. In stress, assets can flip to non-mobilisable or prohibitive-haircut status without any formal prohibition. A schedule adjustment is enough. This is liquidity risk that is structurally hard to quantify. The International Swaps and Derivatives Association and the Basel Committee on Banking Supervision have been working on it since 2020. No resolution. The right question: which assets are mobilisable in each silo, and which flip to non-mobilisable in stress? This mapping needs to be done asset by asset, silo by silo — not in theory, but on your actual portfolio.
08:50. You go straight to margins and haircuts — as you should. Every CCP runs its own models and parameters. In stress, requirements can shift fast — the margin cliff. CCPs publish their methodologies. Not their actual internal parameters. Not their proprietary stress scenarios. A clearing member never knows precisely what a CCP will call in a specific stress event. The structural problem: infrastructure wrong-way risk. The CCP raises margin in stress — precisely when your HQLA is under pressure. A triparty agent restricts substitutions during peak activity — precisely when you need to mobilise. The correlation between silo constraint and your own liquidity need is rarely analysed. In BAU, a wrong-way dependency test is the right starting point: is your primary silo most constrained exactly when you need it most? It is the most uncomfortable test. It is the one that never gets done.
08:51. That is when you see it: access — direct or sponsored — netting capacity, and entry terms effectively define a club. You are either in, or you are out — or relying on someone to get you in. You are still in — but tracing back through the custody, clearing, and triparty chains, you see that 70–80 per cent of mobilisable collateral flows through one or two actors: BNY, Euroclear, and a handful of others. Apparent diversification masks real concentration. BNY and State Street hold a material share of global collateral. If either suffers an operational incident — as nearly happened in the US repo market in September 2019 — the entire mobilisation chain seizes, regardless of collateral quality or CCP selection. The right preparatory action: concentration testing and silo-mapping. For each key repo activity, what is the actual infrastructure-jurisdiction pairing in use? How many silos do you genuinely have? This test reveals the real single point of failure — not the theoretical one on the risk management slide.
08:53. You feel it building. Surprises keep coming. What varies by silo — who can liquidate what, when, and how, in terms of default management and auction procedures — directly conditions executable liquidity in dislocation. A bank with positions across two CCPs managing a collateral substitution between them faces settlement fail risk that sits not inside either CCP, but in the gap between their two settlement cycles. That inter-silo risk is invisible in individual reports. Nobody consolidates it. Hence the surprise. The question: how quickly can you replace your primary silo with an operational alternative? Not in theory — with actual teams, live technical connections, existing legal agreements, and real access limits. For most banks, the honest answer is several weeks. Your Contingency Funding Plan (CFP) must include a genuine substitution test. Does the switchover plan actually work? For most firms, the honest answer is that it is a fiction under acute stress. This scenario needs to be simulated concretely — not documented theoretically.
08:56. You have got the operational pressure under control. For now. You zoom out to the strategic and political dependency across your book. Exposure to public-sector measures — sanctions, restrictions, collateral localisation requirements, access limitations — varies materially by infrastructure location. A CCP or custodian can be used by its national regulator to impose constraints on foreign members without ever issuing a formal sanction. Adjust eligibility rules, tighten stress parameters, require local collateral — that is sufficient. Post-Brexit, LCH showed this in real time. The question for your institution: what share of your collateral sits under a non-European jurisdiction in a crisis? The relevant law is not standard contract law — it is emergency law. How much of your collateral is held in structures where a unilateral decision by a foreign regulator — the Fed, Office of Foreign Assets Control — could block access or impose ring-fencing? That number is rarely calculated. Almost never reported to the risk committee. This is the jurisdiction stress test: how much of your collateral sits under non-European jurisdiction, and how exposed each infrastructure is to regulatory capture. In a dollar stress scenario, converting euro collateral into US dollar liquidity depends on Fed/European Central Bank (ECB) swap lines — discretionary, political, potentially conditional lines that quietly structure your dependency.
08:59. Then you hit the last step: law and enforceability. Close-out, netting, and set-off — their speed and robustness depend entirely on the legal framework governing the infrastructure. For CCPs, the law differs across LCH SA (Paris), LCH Ltd (English law, post-Brexit), Eurex Clearing (German law), and Euronext Clearing (Dutch parent, equity markets across Paris, Amsterdam, and Dublin, CCP incorporated in Italy). Choosing between them is not just a question of pool depth, netting/compression conditions, or settlement optionality. It is a choice of jurisdiction, margin rules, access conditions, collateral treatment — and a choice of resilience and dependencies. In stress or default, your ability to recover positions depends entirely on the portability rules of the relevant CCP. A trader who has not checked this before entering the silo discovers at the worst possible moment that recovery is not available. This is not theoretical. It is exactly what happened to certain portfolios during Archegos. Repo becomes non-neutral the moment the chosen infrastructure filters access through membership, recalibrates requirements through margins, or requalifies assets through eligibility — precisely the dimensions that matter under stress. The BAU challenge is the portability test: if your clearing member defaults tomorrow morning, how long does position recovery take? For most banks, the honest answer is several days — during which you are out of the market. Most firms have not properly tested this scenario.
One question runs through all seven dimensions — all the way from operational to political. Is the dependency you are carrying chosen or implicit? A chosen dependency can be managed. An implicit one surfaces at the worst possible moment. It is 09:00. The margin call is met. This time.
No decree. No embargo. No recourse
What just happened on your desk is happening simultaneously across hundreds of firms. Aggregated, it is a systemic crisis. The most striking feature: the pressure is being applied without a single decision being taken.
Repo can be more powerful than a conventional sanction because it strikes at the most vulnerable point of any institution — immediate access to liquidity and leverage — and it does so with superior operational efficiency. A conventional sanction is slow, visible, legal, and therefore contestable. It is typically binary: permitted or prohibited. Repo works quickly, opaquely, through technical parameters — eligibility, haircuts, CCP and triparty access, margins — that are difficult to challenge and, critically, graduated. You can degrade without prohibiting. Ration without announcing. Isolate without naming.
That changes everything for everyone in the market. For banks, collateral becomes a political asset in the strict sense. Its funding value depends on acceptance criteria that go well beyond credit quality. Prudential metrics — LCR, Net Stable Funding Ratio — cease to be pure compliance ratios and become management instruments exposed to power dynamics: what is fundable today may not be tomorrow.
For sovereigns, securing the repo financing of government collateral is existential. Building domestic repo markets and local infrastructure chains is a strategic and geopolitical choice. Dependence on foreign infrastructure is a material vulnerability — funding cost and access can be modulated externally without any explicit political decision. For investors, repo risk becomes country risk. Liquidity that looks deep can disappear instantly if haircuts widen or eligibility tightens. Collateral pricing is increasingly political — sensitive to regulatory expectations, ESG positioning, and sanctions anticipation.
Repo is a uniquely potent pressure mechanism because it operates simultaneously across four channels: confidence — counterparty and collateral acceptance; liquidity — monetisation capacity; leverage — roll-over and margins; and systemic stability — forced selling, contagion. No other financial infrastructure combines these four channels with the same immediacy and granularity.
The new weapons don’t fire — they filter
Local jurisdictions create dependencies — deeper than day-to-day flows suggest. Repo is a foundational layer of the sovereignty silos that market infrastructures represent. And repo can be more effective than a formal sanction — and infinitely less visible. Repo and securities financing transactions (SFTs) must be read for what they are: power infrastructure. In a liquidity regime built on collateral, constrained balance sheets, and margin dynamics, whoever shapes the repo and SFT channels shapes collateral mobility, liquidity circulation, and ultimately the funding conditions of sovereigns and banks. Without repo, even the most apparently liquid asset can become dysfunctional.
Repo acts directly on sovereign funding costs. Without robust repo capacity, a bond is less easily financed, its investor base narrower, its liquidity premium higher, and its funding cost heavier. The reverse is equally true: the more repo-financeable a bond, the more portable it becomes — broader investor base, compressed spreads. Look at the African Eurobond market — one-third of the continent’s financing. No repo market means issuance rates 150–300bps higher at equivalent sovereign rating. That is a macroeconomic lever, not a technical detail.
Repo creates a collateral hierarchy — an implicit financial ranking of what is eligible, under what conditions, and at what haircut. Control what gets accepted and you control what counts as prime collateral. You define the boundaries of global liquidity. Eighty per cent of assets financed through European triparty repo are rated at least single-A.
Repo can function as an invisible financial sanction — deliberate or systemic. A country, sector, or actor can be cut off from market funding without a formal embargo. Make its collateral unfinanceable. Repo does not block — it degrades. And degradation is sufficient to sever liquidity access. The 2022 sanctions made this explicit: by reclassifying assets as ineligible and restricting access to funding channels, the authorities triggered a near-instantaneous collapse in liquidity and severe dislocation in funding conditions. Exclusion did not come through direct prohibition. It came through the transformation of eligibility rules. The sanction moved through the pipes.
But this mechanism requires no political intent. The UK LDI crisis in autumn 2022 demonstrated the endogenous version: a brutal surge in margin calls and a contraction of repo funding triggered a forced-selling spiral that turned a market adjustment into a systemic crisis. Nobody sanctioned the pension funds. The outcome was identical — loss of leverage access, forced liquidation, market dislocation. September 2019 in the US repo market showed the same dynamic: a cash shortage concentrated among a handful of actors, combined with operational constraints, was enough to blow out rates. No political decision. A chokepoint that de facto rationed liquidity access.
The mechanism is identical across all three cases. Funding access does not disappear by decree. It deteriorates through parameters — eligibility, haircuts, margins, balance sheet capacity, operational constraints. That deterioration produces the same effects as an explicit sanction. This is what makes repo singular: a system capable of excluding without prohibiting, constraining without declaring, and sanctioning without accountability.
Repo can stabilise or amplify a systemic crisis. It is either a shock absorber — through facilities, interventions, liquidity support — or a crisis multiplier, through haircut spirals, margin calls, and forced deleveraging. When funding parameters tighten, roll-over capacity seizes. The crisis propagates not through insolvency but through loss of transformation capacity. Procyclicality makes repo a systemic accelerator.
Finally, repo is a lever of financial sovereignty. The question turns geopolitical the moment clearing, triparty, custody, or repo access are concentrated in specific jurisdictions and infrastructures. Depending on another bloc for your own critical chokepoints — clearing, collateral, repo — means depending on its rules, its stability priorities, and potentially its political decisions. This is why the debates on strategic autonomy and market architecture cannot stay in the technical lane — because repo is a weapon of control and controlling the system’s chokepoints is often worth more than controlling the underlying assets.
The ECB has, in its own way, acknowledged this — launching in December 2025 its first geopolitical stress tests on European banks, results due July 2026. A commendable initiative, an ambitious timeline. One can only hope that someone in the scenario design process thought to model the financial equivalent of a blockade of the Strait of Hormuz: a repo-collateral infrastructure chokepoint that declares nothing, prohibits nothing, and cuts liquidity access with the quiet efficiency of a maître d’hôtel who informs you, with a warm smile, that your table is no longer available.
If that scenario is in the exercise, it is reassuring. If it is not, European banks will be perfectly prepared for the next crisis — provided it resembles the previous one.
Repo, collateral, clearing, market infrastructures. Access points. Filters. Pressure levers. In the right hands — or the wrong ones — they function as sanction tools, scarcity engines, hard stops. The risk has not just moved. It has mutated. No longer technical. No longer prudential. Strategic now. Exogenous. Sometimes hostile. These instruments are no longer neutral. They were never truly neutral.
I have argued in this series that the liquidity regime has changed. That argument was too modest. The gap is not a regime shift. It is existential — and it is already open.
The traditional view of repo — technical tool, apolitical, interchangeable — rests on one silent assumption: that any funding channel can always be substituted for another. That assumption is broken. Repo is the infrastructure layer on which collateral transformation, funding price formation, and liquidity transmission all depend. Banks use it to run their balance sheets and buffers. Dealers use it to provide intermediation. Hedge funds use it to carry risk. Central banks use it to inject and sterilise. Market infrastructures use it to enforce discipline.
Disrupt the repo market, and you break the flow that keeps the whole system alive. Once access becomes conditional, whoever controls the channels — the eligibility rules, the chokepoints — holds the control. Potentially a systemic weapon.
You think you own it. Read the small print
An asset is liquid only where it sits — and within the infrastructure where it is held. I think of liquidity in three dimensions: the nature of the collateral, the nature of the underlying legal contract, and the nature of the processing and settlement infrastructure. In stress, executable liquidity depends less on accounting classification — high-quality liquid asset (HQLA) — or on contract type, than on the operational mechanics: which custodian, which triparty agent or central counterparty (CCP), under which law, within which cut-off window, with what mobilisation lag. The Liquidity Coverage Ratio (LCR) does not see this geography. Yet this is precisely the geography that creates points of failure.
Location matters because mobility is neither free nor instant. Europe has long suffered from fragmentation in clearing and settlement. Cross-border settlement costs run on average 65 per cent higher than in the US — a structural symptom of a post-trade market that remains fragmented, expensive, and operationally inefficient. Even after years of harmonisation, the authorities themselves acknowledge that work remains. Standardisation does not eliminate friction. AFME — October 2025, ‘Analysis of CSD fees in major European markets’.
TARGET2-Securities (T2S) lowered barriers. It did not solve trapped collateral. The platform now connects 24 central securities depositories (CSDs) and processes large settlement volumes — a genuine improvement in technical integration. But two structural facts shape executable liquidity. First: cross-CSD activity remains a small fraction of total flows. Operations are still largely organised in CSD silos, even where the technology is common. In 2024, cross-CSD transactions accounted for 3.8 per cent of T2S volumes. Second: even on centralised infrastructure, settlement efficiency is not 100 per cent. In 2024, end-of-day efficiency stood at 94.4 per cent by volume. Roughly 5.5 per cent of volumes did not settle by EOD — spilling into fails, queues, and delays, with direct consequences for collateral mobilisation and intraday management. All this is according to the ECB Target Services Annual Report 2024.
Collateral that is unsettled, misaligned, or stranded in the wrong location is not mobilisable — regardless of its credit quality. The challenge is the ability to locate the security and mobilise it in time through the execution chain. What most risk managers still miss: owning an asset and controlling it are two different things. The gap between the two is exactly where executable liquidity disappears — quietly, without a warning signal.
That gap shows up in very specific places. It sits between the local CSD and the international central securities depository. Between a house account and a triparty account. Between segregation and encumbrance. It widens with each instrument choice — bilateral repo, synthetic repo, cleared triparty, securities lending, buy and sell-back, central bank facility. It extends with the governing documentation — Global Master Repurchase Agreement, Credit Support Annex, eligibility schedules, substitution conditions — which determine what you can do, how fast, and with whom. It closes in around cut-off windows, settlement cycles, realignment frictions. And it becomes a trap in post-trade fails and queues.
The real questions are simple: who and what are you dependent on? Is that dependency chosen or implicit? Running a liquidity buffer without managing how your dependencies actually work is like switching on a generator without checking whether it has fuel. The collateral is yours. The question is — from when?
Pick your infrastructure. Pick your master
No execution venue choice is neutral. It determines jurisdiction, eligibility rules, margin models, default rights, and exposure to sovereign-level decisions — sanctions, ring-fencing, local collateral requirements. Infrastructure selection is a regime choice. Most people in the room do not want to say that out loud. Does it take getting blocked before you ask the question?
If repo access is silo-dependent, then repo is far more than a funding tool. It is a chokepoint. And a single chokepoint creates filtering capacity and constraint — over who can execute, at what speed, with what collateral, and under which law. That is the operational definition of a control lever.
Your ability to mobilise collateral, raise cash, and hold positions depends on constraints you do not fully control. Power here is not exercised through prohibition. It is exercised through parameter adjustment. Under stress, those parameters determine who stays in the market and who is forced out. Understanding a silo is not about mapping its features. It is about identifying where it constrains, when it breaks, and how far it exposes you to an external decision — regulatory, political, or purely mechanical.
All of the above can be tested. Not in a committee. Not in a stress test. On an ordinary Tuesday morning.
08:47. Your desk has a margin call due at 09:00. Right in the window where two-thirds of repo-margin-CCP activity concentrates. The collateral is there — on paper. Wrong custodian. Wrong governing law. Wrong cut-off window. In 13 minutes, you will find out exactly who and what you depend on.
08:48. Cut-offs, settlement windows: that is your first problem. ‘Usable now’ does not mean the same thing across chains and silos. Next up: cross-product netting rules — and they vary. Without netting, your 5–10 basis point spread is gone. Not every CCP allows netting of repo exposures against derivatives on the same underlying. Euronext Clearing and Eurex Clearing offer this on certain segments. LCH less so. That cross-margining gap is a real, recurring capital cost — independent of transaction pricing, and entirely a function of which silo you are in. Silo dependency also forces you to carry extra liquidity buffer to absorb operational friction: missed cut-offs, refused substitutions, unanticipated margin calls. That over-buffering has a real annual cost. It is rarely calculated. Never reported as such. Its value is destroyed silently, every quarter. Time-to-cash and opportunity cost are the right questions to ask before you are in this situation. Which intraday windows turn theoretical liquidity into unusable liquidity? What does your current silo dependency actually cost in over-buffering and netting leakage? This test is the most directly P&L-linked — and the one least often run.
08:49. Eligibility hits you next. Eligibility schedules are not identical across silos. Under time pressure, they are barely interoperable. An asset that is mobilisable in one silo can be downgraded in another or hit with a steeper haircut. What is repo-eligible at Euroclear is not necessarily eligible at Clearstream or in a BNY triparty basket. The choice of triparty agent or custodian — European or non-eurozone — changes the risk profile entirely: real execution capacity on one side; platform reliance, legal exposure, compliance constraints, and potential extraterritoriality on the other. In stress, assets can flip to non-mobilisable or prohibitive-haircut status without any formal prohibition. A schedule adjustment is enough. This is liquidity risk that is structurally hard to quantify. The International Swaps and Derivatives Association and the Basel Committee on Banking Supervision have been working on it since 2020. No resolution. The right question: which assets are mobilisable in each silo, and which flip to non-mobilisable in stress? This mapping needs to be done asset by asset, silo by silo — not in theory, but on your actual portfolio.
08:50. You go straight to margins and haircuts — as you should. Every CCP runs its own models and parameters. In stress, requirements can shift fast — the margin cliff. CCPs publish their methodologies. Not their actual internal parameters. Not their proprietary stress scenarios. A clearing member never knows precisely what a CCP will call in a specific stress event. The structural problem: infrastructure wrong-way risk. The CCP raises margin in stress — precisely when your HQLA is under pressure. A triparty agent restricts substitutions during peak activity — precisely when you need to mobilise. The correlation between silo constraint and your own liquidity need is rarely analysed. In BAU, a wrong-way dependency test is the right starting point: is your primary silo most constrained exactly when you need it most? It is the most uncomfortable test. It is the one that never gets done.
08:51. That is when you see it: access — direct or sponsored — netting capacity, and entry terms effectively define a club. You are either in, or you are out — or relying on someone to get you in. You are still in — but tracing back through the custody, clearing, and triparty chains, you see that 70–80 per cent of mobilisable collateral flows through one or two actors: BNY, Euroclear, and a handful of others. Apparent diversification masks real concentration. BNY and State Street hold a material share of global collateral. If either suffers an operational incident — as nearly happened in the US repo market in September 2019 — the entire mobilisation chain seizes, regardless of collateral quality or CCP selection. The right preparatory action: concentration testing and silo-mapping. For each key repo activity, what is the actual infrastructure-jurisdiction pairing in use? How many silos do you genuinely have? This test reveals the real single point of failure — not the theoretical one on the risk management slide.
08:53. You feel it building. Surprises keep coming. What varies by silo — who can liquidate what, when, and how, in terms of default management and auction procedures — directly conditions executable liquidity in dislocation. A bank with positions across two CCPs managing a collateral substitution between them faces settlement fail risk that sits not inside either CCP, but in the gap between their two settlement cycles. That inter-silo risk is invisible in individual reports. Nobody consolidates it. Hence the surprise. The question: how quickly can you replace your primary silo with an operational alternative? Not in theory — with actual teams, live technical connections, existing legal agreements, and real access limits. For most banks, the honest answer is several weeks. Your Contingency Funding Plan (CFP) must include a genuine substitution test. Does the switchover plan actually work? For most firms, the honest answer is that it is a fiction under acute stress. This scenario needs to be simulated concretely — not documented theoretically.
08:56. You have got the operational pressure under control. For now. You zoom out to the strategic and political dependency across your book. Exposure to public-sector measures — sanctions, restrictions, collateral localisation requirements, access limitations — varies materially by infrastructure location. A CCP or custodian can be used by its national regulator to impose constraints on foreign members without ever issuing a formal sanction. Adjust eligibility rules, tighten stress parameters, require local collateral — that is sufficient. Post-Brexit, LCH showed this in real time. The question for your institution: what share of your collateral sits under a non-European jurisdiction in a crisis? The relevant law is not standard contract law — it is emergency law. How much of your collateral is held in structures where a unilateral decision by a foreign regulator — the Fed, Office of Foreign Assets Control — could block access or impose ring-fencing? That number is rarely calculated. Almost never reported to the risk committee. This is the jurisdiction stress test: how much of your collateral sits under non-European jurisdiction, and how exposed each infrastructure is to regulatory capture. In a dollar stress scenario, converting euro collateral into US dollar liquidity depends on Fed/European Central Bank (ECB) swap lines — discretionary, political, potentially conditional lines that quietly structure your dependency.
08:59. Then you hit the last step: law and enforceability. Close-out, netting, and set-off — their speed and robustness depend entirely on the legal framework governing the infrastructure. For CCPs, the law differs across LCH SA (Paris), LCH Ltd (English law, post-Brexit), Eurex Clearing (German law), and Euronext Clearing (Dutch parent, equity markets across Paris, Amsterdam, and Dublin, CCP incorporated in Italy). Choosing between them is not just a question of pool depth, netting/compression conditions, or settlement optionality. It is a choice of jurisdiction, margin rules, access conditions, collateral treatment — and a choice of resilience and dependencies. In stress or default, your ability to recover positions depends entirely on the portability rules of the relevant CCP. A trader who has not checked this before entering the silo discovers at the worst possible moment that recovery is not available. This is not theoretical. It is exactly what happened to certain portfolios during Archegos. Repo becomes non-neutral the moment the chosen infrastructure filters access through membership, recalibrates requirements through margins, or requalifies assets through eligibility — precisely the dimensions that matter under stress. The BAU challenge is the portability test: if your clearing member defaults tomorrow morning, how long does position recovery take? For most banks, the honest answer is several days — during which you are out of the market. Most firms have not properly tested this scenario.
One question runs through all seven dimensions — all the way from operational to political. Is the dependency you are carrying chosen or implicit? A chosen dependency can be managed. An implicit one surfaces at the worst possible moment. It is 09:00. The margin call is met. This time.
No decree. No embargo. No recourse
What just happened on your desk is happening simultaneously across hundreds of firms. Aggregated, it is a systemic crisis. The most striking feature: the pressure is being applied without a single decision being taken.
Repo can be more powerful than a conventional sanction because it strikes at the most vulnerable point of any institution — immediate access to liquidity and leverage — and it does so with superior operational efficiency. A conventional sanction is slow, visible, legal, and therefore contestable. It is typically binary: permitted or prohibited. Repo works quickly, opaquely, through technical parameters — eligibility, haircuts, CCP and triparty access, margins — that are difficult to challenge and, critically, graduated. You can degrade without prohibiting. Ration without announcing. Isolate without naming.
That changes everything for everyone in the market. For banks, collateral becomes a political asset in the strict sense. Its funding value depends on acceptance criteria that go well beyond credit quality. Prudential metrics — LCR, Net Stable Funding Ratio — cease to be pure compliance ratios and become management instruments exposed to power dynamics: what is fundable today may not be tomorrow.
For sovereigns, securing the repo financing of government collateral is existential. Building domestic repo markets and local infrastructure chains is a strategic and geopolitical choice. Dependence on foreign infrastructure is a material vulnerability — funding cost and access can be modulated externally without any explicit political decision. For investors, repo risk becomes country risk. Liquidity that looks deep can disappear instantly if haircuts widen or eligibility tightens. Collateral pricing is increasingly political — sensitive to regulatory expectations, ESG positioning, and sanctions anticipation.
Repo is a uniquely potent pressure mechanism because it operates simultaneously across four channels: confidence — counterparty and collateral acceptance; liquidity — monetisation capacity; leverage — roll-over and margins; and systemic stability — forced selling, contagion. No other financial infrastructure combines these four channels with the same immediacy and granularity.
The new weapons don’t fire — they filter
Local jurisdictions create dependencies — deeper than day-to-day flows suggest. Repo is a foundational layer of the sovereignty silos that market infrastructures represent. And repo can be more effective than a formal sanction — and infinitely less visible. Repo and securities financing transactions (SFTs) must be read for what they are: power infrastructure. In a liquidity regime built on collateral, constrained balance sheets, and margin dynamics, whoever shapes the repo and SFT channels shapes collateral mobility, liquidity circulation, and ultimately the funding conditions of sovereigns and banks. Without repo, even the most apparently liquid asset can become dysfunctional.
Repo acts directly on sovereign funding costs. Without robust repo capacity, a bond is less easily financed, its investor base narrower, its liquidity premium higher, and its funding cost heavier. The reverse is equally true: the more repo-financeable a bond, the more portable it becomes — broader investor base, compressed spreads. Look at the African Eurobond market — one-third of the continent’s financing. No repo market means issuance rates 150–300bps higher at equivalent sovereign rating. That is a macroeconomic lever, not a technical detail.
Repo creates a collateral hierarchy — an implicit financial ranking of what is eligible, under what conditions, and at what haircut. Control what gets accepted and you control what counts as prime collateral. You define the boundaries of global liquidity. Eighty per cent of assets financed through European triparty repo are rated at least single-A.
Repo can function as an invisible financial sanction — deliberate or systemic. A country, sector, or actor can be cut off from market funding without a formal embargo. Make its collateral unfinanceable. Repo does not block — it degrades. And degradation is sufficient to sever liquidity access. The 2022 sanctions made this explicit: by reclassifying assets as ineligible and restricting access to funding channels, the authorities triggered a near-instantaneous collapse in liquidity and severe dislocation in funding conditions. Exclusion did not come through direct prohibition. It came through the transformation of eligibility rules. The sanction moved through the pipes.
But this mechanism requires no political intent. The UK LDI crisis in autumn 2022 demonstrated the endogenous version: a brutal surge in margin calls and a contraction of repo funding triggered a forced-selling spiral that turned a market adjustment into a systemic crisis. Nobody sanctioned the pension funds. The outcome was identical — loss of leverage access, forced liquidation, market dislocation. September 2019 in the US repo market showed the same dynamic: a cash shortage concentrated among a handful of actors, combined with operational constraints, was enough to blow out rates. No political decision. A chokepoint that de facto rationed liquidity access.
The mechanism is identical across all three cases. Funding access does not disappear by decree. It deteriorates through parameters — eligibility, haircuts, margins, balance sheet capacity, operational constraints. That deterioration produces the same effects as an explicit sanction. This is what makes repo singular: a system capable of excluding without prohibiting, constraining without declaring, and sanctioning without accountability.
Repo can stabilise or amplify a systemic crisis. It is either a shock absorber — through facilities, interventions, liquidity support — or a crisis multiplier, through haircut spirals, margin calls, and forced deleveraging. When funding parameters tighten, roll-over capacity seizes. The crisis propagates not through insolvency but through loss of transformation capacity. Procyclicality makes repo a systemic accelerator.
Finally, repo is a lever of financial sovereignty. The question turns geopolitical the moment clearing, triparty, custody, or repo access are concentrated in specific jurisdictions and infrastructures. Depending on another bloc for your own critical chokepoints — clearing, collateral, repo — means depending on its rules, its stability priorities, and potentially its political decisions. This is why the debates on strategic autonomy and market architecture cannot stay in the technical lane — because repo is a weapon of control and controlling the system’s chokepoints is often worth more than controlling the underlying assets.
The ECB has, in its own way, acknowledged this — launching in December 2025 its first geopolitical stress tests on European banks, results due July 2026. A commendable initiative, an ambitious timeline. One can only hope that someone in the scenario design process thought to model the financial equivalent of a blockade of the Strait of Hormuz: a repo-collateral infrastructure chokepoint that declares nothing, prohibits nothing, and cuts liquidity access with the quiet efficiency of a maître d’hôtel who informs you, with a warm smile, that your table is no longer available.
If that scenario is in the exercise, it is reassuring. If it is not, European banks will be perfectly prepared for the next crisis — provided it resembles the previous one.
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 Â鶹ӰÊÓ´«Ã½ 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 Â鶹ӰÊÓ´«Ã½ Times
