RESEARCH DOSSIER
Offshore Financial Contagion
Abstract. This research dossier examines financial crisis transmission through jurisdictionally distributed financial claims and balance sheets. It analyzes direct counterparty losses, common lenders, currency and maturity mismatches, collateral and fire sales, nonbank intermediation, margin, information and statistical exposure attribution. Comparative evidence from the global financial crisis, March 2020, the 2022 gilt market disturbance and cross-border Treasury holdings, together with illustrative accounting exercises, distinguishes common shocks from endogenous amplification and liquidity stress from realized loss. The synthesis considers empirical identification, cross-border supervision, and real-economy incidence. Its evidence base is selective; no new empirical estimate, independent replication, or named theory is asserted.
Keywords: offshore finance; financial contagion; international banking; systemic risk; common lenders; dollar funding; financial networks; market liquidity; nonbank intermediation; financial statistics
Introduction
Offshore financial contagion is the transmission and amplification of financial distress through claims, funding relationships, market valuations, and payment obligations distributed across jurisdictions. Its central object is the economic effect of these connections conditional on a disturbance. A financial contract may be administered in one economy, denominated in the currency of a second, guaranteed by an entity in a third, and ultimately finance activity elsewhere. The location at which it is recorded is therefore insufficient to explain where the consequences of a shock materialize.
This study asks how distress travels, how financial institutions respond, how transmission can be distinguished from simultaneous exposure to common shocks, and how public policy can reduce destructive externalities while retaining the benefits of international intermediation. General offshore financial stability encompasses the prevention of initial failure and the resilience of institutional arrangements; this dossier concentrates on the subsequent interaction of balance sheets and financial markets.
The research method consists of a selective review of original scholarship and official source records, comparative interpretation of historical crises, and transparent hypothetical accounting exercises. Some primary records were accessible as indexed extracts or abstracts rather than complete texts. The study does not claim a systematic literature review, independently replicated estimates, or a newly constructed global data set. All numerical examples introduced as illustrations are hypothetical.
The Offshore Dimension
The international organization of finance creates several relevant boundaries: the residence of an immediate counterparty, the headquarters of its corporate group, the currency of a contract, the location of collateral, and the location of supervision and settlement. Offshore arrangements involve financial activities outside a relevant home jurisdiction or currency area. They can support legitimate financing, investment, international treasury management, and diversified access to markets.
Pogliani and Wooldridge (2022) develop a statistical identification of cross-border centers based on the scale and intermediation characteristics of external balance sheets. The historically distinct offshore assessment program was integrated into the Financial Sector Assessment Program following the IMF Executive Board's May 30, 2008 decision, recorded in its July 2008 public notice. The separate jurisdiction list was no longer needed under the integrated approach. A statistical classification of intermediation and a supervisory assessment of risks consequently answer different questions.
A fully equity-funded holding company may pass a valuation loss to shareholders without amplifying it. A highly leveraged institution can transform a minor disturbance into urgent demands for cash. A foreign affiliate may stabilize a local market through support from its parent, or transmit the parent's financing constraints to that market. The economically relevant feature is the arrangement's exposures, maturity structure, access to funding, and capacity to absorb losses.
A country's offshore activity is thus not a sufficient statistic for its contribution to global contagion. Geographic labels may identify the population to study; they cannot establish the presence, direction, magnitude, or causal relevance of a transmission mechanism. Analysis starts from contractual and institutional facts rather than a presumption of misconduct.
Contagion and Interdependence
Financial stress can arise simultaneously because several economies are directly exposed to the same fundamental shock. A global interest-rate increase may reduce bond valuations even in markets whose participants do not transact with one another. Propagation occurs when changes in one participant's condition affect the decisions or constraints of another. Excess transmission refers more narrowly to a supported increase in transmission relative to a specified baseline.
The Forbes–Rigobon correlation critique (2002) demonstrates that raw correlation comparisons can confuse volatility changes with changes in underlying interdependence. Their adjusted tests, under stated assumptions, substantially change the inference for the crisis episodes examined. This finding is consistent with the existence of identifiable contractual and funding channels: correlation tests and balance-sheet propagation answer different questions.
For this dossier contagion encompasses path-dependent financial propagation, with excess transmission reserved for a documented structural or conditional increase. An initiating shock, a prior exposure, a subsequent response, and competing explanations should all be specified. A fall in lending can reflect worsened borrower prospects, weakened lenders, or both. Mere coincidence or temporal succession is insufficient.
A credible counterfactual compares otherwise similar borrowers or markets exposed to the same disturbance but with differing pre-existing dependence on an affected financial connection. The strength of the inference depends on the comparability of these groups, the quality of exposure measures, and the treatment of spillovers and simultaneous policy interventions.
Financial Geography and Exposure Attribution
Residence-based positions allocate financial claims to institutional units where they reside. Consolidation assigns entities to corporate groups under a particular perimeter. Risk-transfer adjustments can reallocate an obligation to a guarantor or other ultimate obligor. The BIS residence and nationality analysis (McGuire et al., 2024) explains the distinct uses of these perspectives.
Consider an operating firm in A, a financing subsidiary in B, and a fund in C. The subsidiary issues a bond of 100 units to the fund and lends the proceeds to the parent. B is the immediate issuer's residence. A is the ultimate corporate group's location. The investor's legal repayment rights depend on the security's contractual terms, including recourse and guarantees. Counting the 100-unit bond both as exposure to B and as an additional 100-unit final exposure to A would double-count a reallocation.
The portfolio reallocation study by Coppola et al. (2021) maps securities issued through offshore entities to their ultimate parents. Its published abstract reports that official US positions in Chinese firms were understated by nearly $600 billion in the authors' reconstruction. That estimate relates to a specific position measurement exercise, not a measure of crisis loss or current outstanding exposure.
A group-nationality view provides insight into control and consolidated exposures; it cannot replace all entity-level information. A parent may not guarantee a subsidiary's debt, local authorities may restrict resource transfers, and actual sales and assets may be distributed across many economies. Aldasoro et al. (2026), in their offshore activity primer extend the statistical rationale for looking through foreign affiliates in banking and securities markets.
The analytical requirement is to preserve multiple linked records of the same claim, assigning each its appropriate purpose. Residence identifies immediate financial and supervisory locations; corporate group information helps identify common decisions; guarantee information clarifies where contractual credit risk may fall; dated obligations identify immediate liquidity needs. Reconciliation does not mean treating the records as independent stocks.
Direct Counterparty Losses and Capital Constraints
A default reduces the value of the creditor's claim after applicable recovery and credit protection. The creditor's resulting equity loss can impair its capacity to maintain other positions. The first-round loss and subsequent adjustment must be recorded separately.
Suppose a bank has assets of 100, liabilities of 92, and equity of 8. A loan of 20 suffers a loss of 30 percent, reducing assets to 94 and equity to 2. Its simple equity-to-assets ratio falls from 8 percent to approximately 2.13 percent. Without new equity, a frictionless reduction of assets accompanied by equal debt repayment would have to reduce assets to 25 to restore the original ratio. The illustrative contraction of 69 from post-loss assets is not an additional loss of 69. It represents one extreme response under explicitly restrictive assumptions. Actual banks may issue equity, change asset composition, retain earnings, or receive support; regulatory ratios also depend on risk-weighted exposures and eligible capital.
A creditor's exposure to a remote operating firm may be recorded as a claim on an offshore issuer. If the immediate debtor is a special-purpose vehicle with limited recourse, the appropriate loss depends on the underlying assets and actual contractual protection. A group's consolidated loss recognition and a legal entity's insolvency resources are different questions.
Acemoglu, Ozdaglar, and Tahbaz-Salehi (2015) demonstrate in their network model that dense liability connections can support diversification when shocks are sufficiently small while allowing sufficiently large shocks to propagate widely. This conditional result cautions against interpreting network density in isolation. Capital, liquidity, concentration, and the initiating shock are jointly relevant.
Global Banks and the Common Lender Channel
Two borrowers may be connected economically through a common lender despite having no direct relationship. An institution suffering losses in A can cut credit to borrowers in B because its own capital, funding, or risk constraints have tightened. The reduction in lending can occur in an otherwise sound market.
Consider a bank with loans of 60 to customers in A and 40 in B, funded with 90 of liabilities and 10 of equity. A realized four-unit loss in A reduces equity by 40 percent. If its management subsequently decides to reduce assets by 20, the institution may concentrate the adjustment where maturities are shortest or sale costs lowest. The loss and the location of the contraction need not coincide.
The global-bank transmission study by Cetorelli and Goldberg (2010) investigates emerging-market loan supply during the global financial crisis. It identifies effects through direct cross-border lending, local lending by foreign affiliates, and domestic lenders dependent on external interbank funds. Policies aimed exclusively at direct international loans therefore omit other possible transmission paths.
The common-lender comparison by Koch and Remolona (2018) traces changes in international creditor composition across three crises affecting emerging Asia. The set of relevant creditor institutions is historically contingent. Exposure maps built from one episode cannot be carried unchanged into another.
A host-country affiliate funded primarily by local deposits may face different responses to parent stress from an affiliate dependent on short-term interoffice funding. Parent institutions may also allocate scarce liquidity across subsidiaries according to strategic priorities, contractual restrictions, and local market opportunities. Empirical work should distinguish group risk from affiliate-specific dependence rather than use headquarters nationality as a complete explanation.
Currency and Maturity Transformation
An institution can have a balanced foreign-currency valuation exposure and still face a shortage of cash in that currency. Current values and dated payment obligations measure different things. A long-term asset financed with short-term funding requires repeated renewal even when interest-rate and currency valuation risks are hedged.
Consider a fund holding a five-year dollar bond and renewing a three-month currency hedge. Its protection against exchange-rate movements depends on the contracts and may entail collateral transfers and repeated access to the swap market. When dealers cannot provide the required funding at former prices, the fund may need alternative financing, unencumbered cash, or asset sales.
The BIS dollar-debt study by Borio et al. (2022) reports over $80 trillion in outstanding worldwide dollar payment obligations from FX swaps, forwards, and currency swaps around mid-2022. This gross amount is neither expected credit loss nor net dollar debt payable at a single time. The measurement issue is the existence of significant forward dollar payment commitments outside conventional debt statistics.
A widening cross-currency basis may reflect changes in hedging demand, balance-sheet intermediation costs, credit considerations, or funding scarcity. It is not a sufficient diagnostic of insolvency. Its effect on a given institution depends on how much financing it must renew, the direction of its obligations, and available collateral.
A credible liquidity assessment tracks inflows, outflows, and accessible funding by currency, maturity, and legal entity. An aggregate group surplus cannot demonstrate that a separate affiliate can make a payment if transfer restrictions, settlement times, or legal constraints prevent timely access. Conversely, contractual and operationally credible parent support can mitigate liquidity transmission.
Market Liquidity, Collateral, and Fire Sales
Funding liquidity concerns an institution's ability to obtain cash; market liquidity concerns the ability to execute transactions without severe price impact. The connection between these forms of liquidity is the subject of Market Liquidity and Funding Liquidity by Brunnermeier and Pedersen (2009). Collateral values influence funding capacity, and funding constraints influence demand for assets.
An intermediary holds 100 units of collateral financed with 90 of secured borrowing and 10 of equity. A five-percent price decline reduces its asset value to 95 and equity to 5. If the required haircut simultaneously increases from 10 to 20 percent, financing capacity falls to 76. It must repay or replace 14 units of borrowing. The 14-unit funding deficit is not an additional valuation loss of 14.
If the institution must sell, market prices may decline further. Other holders then experience valuation changes, and other borrowers may encounter new margin or collateral requirements. A shock can travel from the original holder to a shared market and from that market to multiple balance sheets without a direct credit relationship among them.
Offshore fund domiciles and booking locations can give the appearance of geographic diversification while the portfolios remain concentrated in the same liquid securities. Strategies as well as holdings matter. Different securities may be sold together if institutions face similar leverage or volatility constraints. Conversely, investors with longer horizons and unencumbered funds may absorb sales, interrupting amplification.
A stress test must therefore state its assumptions about market depth, risk limits, collateral eligibility, and the ability of other investors to buy. Fixed-price first-round tests and price-feedback exercises serve different purposes. An intervention designed to preserve market functioning should also be distinguished from a policy that protects investors from fundamental losses.
Nonbank Intermediation and Redemption
International banks interact with investment funds, pension funds, finance companies, insurers, and other nonbank participants through loans, securities, derivatives, deposits, and other funding arrangements. Their functions and obligations differ materially. The economic analysis should identify which liability can generate a demand for cash.
The BIS bank–nonbank study (Aldasoro et al., 2020) documents the international relationships between banks and nonbanks, including financial centers' role in dollar funding. A fund can affect a bank as a creditor or derivative counterparty, not only as a borrower. The direction of a financing relationship is as important as the presence of the relationship.
Suppose an open-ended fund has assets of 100, including 10 of available cash, and receives redemptions of 25. If these requests are fulfilled and no new financing is available, the fund must raise at least 15 through asset sales. Where redemption terms leave liquidation costs to remaining investors, early withdrawal can be privately advantageous. Such incentives depend on pricing, notice, and portfolio terms. The arithmetic does not demonstrate that any actual fund is experiencing a run.
The IMF nonbank financial stability assessment (2023) emphasizes leverage, liquidity mismatches, and interconnectedness. Money-market funds present a specific link between redemption and the availability of short-term financing for banks and firms. The FSB money-market fund review (2024) emphasizes variation in fund and market structures relevant to resilience. Fund domiciles alone provide little information about cash buffers, investor behavior, or liquidity of underlying holdings.
Insurance and pension obligations must be analyzed under their distinct contractual schedules. A withdrawal right in an investment fund cannot be assumed to exist for a different institution. Common analysis concerns collateral, liquidity, and asset-market spillovers; sector-specific measurement remains essential.
Derivatives, Margin, and Settlement
Derivatives redistribute risk and create obligations whose timing can differ from underlying asset returns. A central counterparty may reduce some counterparty risks through netting and collateral, while margin obligations generate demands for eligible liquidity during volatile periods. Credit protection and liquidity preparedness must be assessed jointly.
Two transactions yielding mark-to-market values of plus 10 and minus 7 units for one participant produce a net replacement-value exposure of 3 if enforceable netting applies and before collateral. Positions with different counterparties may be economically offset but require independent payments. A payment received tomorrow cannot automatically finance one due today. Gross settlement amounts, net replacement values, and notional amounts address different risks.
The Review of Margining Practices (BCBS–CPMI–IOSCO, 2022) examines transparency, predictability, volatility, and preparedness. Offshore contagion analysis requires these considerations because a financial group may obtain dealing or clearing services in one location while holding eligible collateral elsewhere. Operational transfer times and currency constraints are relevant.
A stress scenario should represent simultaneous movements in rates, prices, volatility and exchange rates; calls at different time horizons; and financing access by legal entity. Collateral segregation and contractual restrictions influence whether resources received by one institution are immediately available to another. Margin is a mechanism of risk management and a potential liquidity transmission channel. Weakening it indiscriminately would trade liquidity risks against counterparty risks.
Comparative Crisis Evidence
The Asian financial crisis, the global financial crisis, and the European sovereign debt crisis involved distinct interactions between internationally active lenders and regional borrowers. The three-crisis study (Koch and Remolona, 2018) illustrates how the origin and composition of lenders changed over time. A similar reduction in external credit can arise from creditor losses in different markets, adjustments in local borrower demand, and policy responses.
During the global financial crisis, the Cetorelli–Goldberg staff report (2010) identifies multiple bank credit channels that transmitted advanced-economy financial distress to emerging economies. The study is relevant to offshore lending arrangements, but its conclusions cannot be transformed into a quantitative estimate of the global crisis attributable to offshore entities.
March 2020 began with a different shock: a pandemic and broad reassessment of risks. The FSB March turmoil review (2020) identifies severe liquidity imbalances, dealer capacity constraints, Treasury sales, margin demands, and redemption incentives. The simultaneous fundamental shock and subsequent amplification coexist. The episode therefore illustrates why the distinction between an exogenous common shock and endogenous response matters.
Comparative analysis should record the initiating event, the financial institutions concerned, assets and currencies, relevant funding contracts, the timing of credit and price responses, and policy interventions. It should include episodes of successful absorption as well as widely documented crises to reduce selection bias. The historical evidence supports the existence of several pathways, rather than an invariant ranking of offshore jurisdictions or a single universal propagation mechanism.
Information and Statistical Opacity
Complex structures can make it costly to identify the obligations that connect intermediaries. Market participants update assessments of borrowers' risks when they obtain new information. Simultaneous repricing may reflect previously unknown common exposures rather than a contractual propagation of an existing loss. It can subsequently generate financing stress if creditors reduce funding access.
Financial confidentiality and supervisory information are distinct issues. A system can preserve legitimate confidentiality while providing relevant exposure information to competent authorities. A public ownership register alone may not contain the currencies, maturities, guarantees, collateral or legal obligations needed for a contagion assessment.
The BIS statistical perspective (2024) explains why grouping exposures along the wrong perimeter can conceal economic dependencies even where individual reported balances are correct. Reducing classification errors and improving exposure information address related but distinct problems.
Transparency can initially cause prices to respond to recognized risks. A decline after a disclosure does not establish that disclosure caused the underlying risk. A policy study should distinguish information revelation, changes in lending or investment behavior, realized losses, and longer-run improvements in risk allocation.
Measurement and Empirical Identification
A statistical investigation begins with the unit, reporting perimeter, date, currency, and valuation basis of each financial quantity. External debt, bank claims, securities holdings, transaction flows, derivatives notional amounts, and cash obligations cannot be aggregated as if they represented the same measure.
A position falling from 100 to 85 may reflect a sale, valuation movements, exchange-rate changes, a reclassification, or a combination. Whenever possible, research uses transaction data or appropriately adjusted changes. Consolidated statistics remove relevant intragroup positions; entity-level records remain necessary for liquidity analysis.
| Dimension | Quantity | Limitation |
|---|---|---|
| Residence and nationality | Immediate and reallocated positions | Alternative attribution is not a second stock |
| Counterparty | Claim, guarantees and collateral | Legal rights differ across contracts |
| Currency and maturity | Dated payments and funding capacity | Net exposure can hide rollover risk |
| Portfolio | Holdings and trading strategies | Common holdings do not prove liquidation |
| Derivatives | Netting sets and gross settlement | Notional does not measure expected loss |
An illustrative panel strategy regresses changes in credit or funding spreads on pre-shock exposures interacted with a crisis-period indicator, together with appropriate unit and time effects. Identification requires assumptions about pre-trends, selection into funding relationships, and confounding disturbances. Exposure predetermined before a crisis may remain endogenous; network spillovers may contaminate control groups. The Forbes–Rigobon correction (2002) cautions specifically against interpreting changes in conditional correlation as autonomous shifts in transmission.
Stress tests ask a different question: what outcomes follow from specified shocks under explicit behavioral assumptions? They may incorporate first-round losses, capital and liquidity constraints, price impacts, and subsequent adjustments. Missing bilateral exposures are not necessarily zero; different networks compatible with aggregate totals can generate different simulated outcomes. Sensitivity analysis should document dependence on reconstruction and price-impact assumptions.
A reproducible empirical study preserves exact data vintages, variable construction, sample inclusion, currency conversions, analytical code, diagnostics and robustness checks. A public alternative can illustrate the method where confidential exposure data are unavailable. The present work offers an analytical research design and illustrative arithmetic, not a new empirical estimate.
Supervision and Policy Spillovers
Policy should address the mechanisms through which financial stress imposes costs on other market participants. Capital buffers reduce the need to respond to losses by contracting assets. Reliable liquidity resources reduce exposure to refinancing disruption. Appropriate margin preparedness and information about counterparties can reduce disorderly adjustments.
Consolidated supervision and entity-level supervision have complementary tasks. A local authority may need to preserve resources for a domestic institution even when the wider group could deploy those resources elsewhere. Constraints on transfers can protect one location while worsening stress in another. Cooperation should consider the distribution of these effects and the possibility of coordinated responses where compatible with applicable legal responsibilities.
Emergency liquidity provision addresses cash shortages, subject to eligibility, collateral, maturity, and credit-risk arrangements. Foreign-currency shortages may require cooperation with the issuing monetary authority. Such support cannot itself restore the fundamental economic value of impaired assets. Public policy must distinguish temporary financing assistance from the allocation of credit losses.
The FSB market turmoil review (2020) and joint margin review (2022) motivate attention to system-wide liquidity rather than measuring resilience solely at the level of each institution. The money-market fund peer review (2024) further illustrates the need to account for local market institutions. A rule that reduces recorded risk in one segment may shift the same exposures into another. Monitoring policy spillovers requires consistent data across institutions and jurisdictions.
Assessment should consider realized losses, continuity of financing to viable borrowers, asset market functioning, liquidity needs, public balance-sheet exposure, and migration of activity. A decline in recorded offshore obligations is not sufficient evidence of reduced vulnerability if the same financing has moved to instruments or entities outside the reporting perimeter.
Real-Economy Incidence
Financial contagion can change firms' access to working capital, investment financing, and refinancing. Those changes may affect production and employment depending on substitute sources of funding. An increase in market yields can have different consequences for a firm with long-dated fixed-rate liabilities and for a firm facing an immediate maturity.
The economic burden can fall on shareholders, creditors, borrowers, workers, service providers, or public authorities. A fall in assets administered by an offshore center can reduce fee income locally while imposing portfolio losses on foreign beneficial owners and financing costs on operating firms elsewhere. These effects should be measured separately. Quantities recorded in the financial center do not by themselves locate final welfare incidence.
Large international positions relative to local economic activity create potential sensitivity to disturbances originating abroad. International integration can also enable access to diversified sources of finance and risk sharing. The policy objective is to retain these benefits while reducing external costs from disorderly adjustments. A welfare assessment requires a counterfactual arrangement and the associated benefits, losses, and costs of mitigation.
Research should distinguish temporary liquidity disruption from persistent changes in investment and employment. A measured change in loan supply cannot be directly converted into an employment effect without additional evidence. Distributional conclusions require explicit identification of the affected economic population, time horizon and alternative financing conditions.
Clearing Networks and the Conditional Value of Connectivity
Financial interdependence is often studied using a network of payment obligations. Each edge identifies a claim of a creditor on a debtor; the same claim enters the liability side of one entity and the asset side of another. The relevant network is directed and weighted. It cannot generally be reconstructed from the number of counterparties or the ranking of financial centers by recorded activity. A very large interbank exposure and a trivial deposit relationship are different edges even though both appear in a simple adjacency matrix.
The theoretical literature makes the welfare significance of connectivity conditional. In Financial Contagion, Allen and Gale (2000) study interregional liquidity insurance and show how incomplete structures of claims can create fragility. Their model concerns a particular economic setting in which financial claims support the allocation of liquidity across regions. Completeness improves risk sharing in that setting. The result does not imply that every expansion of financial interconnection is stabilizing regardless of the instruments, shock size or leverage involved.
Consider a stylized system of four intermediaries, each with equity of eight and a total of 12 in claims on counterparties. In one arrangement each holds three separate claims of four; in another each holds a single concentrated claim of 12. If the affected debtor supplies a recovery of 50 percent, an intermediary holding a four-unit claim loses two, whereas an intermediary holding a 12-unit claim loses six. Both losses are below the assumed equity buffer, but their implications for the margin of safety differ. A second loss or a binding liquidity requirement can push the concentrated lender into adjustment more quickly. The comparison changes if the dispersed creditors hold similar securities or if all debtors face a shared shock. It illustrates a distribution of exposure, not a quantified model of actual offshore centers.
The challenge increases when losses alter the payments that counterparties can make. Eisenberg and Noe's clearing analysis (2001) formalizes jointly determined payments in an interconnected set of firms subject to limited liability and the priority structure represented by the model. A useful simplified statement takes total nominal liabilities of each entity as fixed, specifies its exterior assets, and makes its actual payment the lesser of nominal debt and available resources from exterior assets plus payments received from other entities. Settlement is simultaneously determined rather than calculated by independently imposing full payment for all creditors. The existence and uniqueness results are conditional on the authors' assumptions, and they do not directly solve real-world insolvency procedures or intraday settlement.
A two-entity example demonstrates how recovery assumptions influence apparent propagation. Suppose entity A owes B 10, owes outside creditors another 10, and holds exterior assets worth 12 with no other receipts. Under a hypothetical proportional distribution across creditors of equal rank, B receives six rather than ten. B has exterior assets worth nine and nominal obligations to outsiders of sixteen; it can pay fifteen after receiving six from A, leaving a one-unit shortfall. A's payment shortfall of eight has produced a further shortfall of one at B. If B instead has three more units of exterior assets, it can meet all its obligations. Adding the total contractual shortfalls across the two balance sheets would measure unpaid gross obligations, not necessarily the final loss borne by outside investors. Under the stated example, A's outside creditors lose four and B's outside creditors lose one. This distinction matters in networks that include multiple intragroup claims.
Actual offshore networks also contain contingent guarantees, repo financing, securities loans, derivative netting sets, and counterparties with different priorities. Different layers can have different directions: an entity can provide cash to a bank while receiving credit protection from the same group. A model that treats these opposite relationships as a single net credit edge may conceal the need to fulfill each cash payment at its own time. The model's analytical perimeter should therefore follow the particular shock rather than combine all exposures into a universal graph.
The original synthesis here uses established network results as a discipline for inquiry. The size of the exposed balance sheet relative to available resources, the conditional correlation of counterparties' vulnerabilities, and the ability to replace funding matter alongside graph topology. It is rational to examine a central intermediary closely, but centrality alone is not a welfare ranking. An institution with many small and readily substitutable connections may impose less marginal systemic risk than an institution with a few very large, difficult-to-replace contracts.
A Comparative Economics of Common-Lender Retrenchment
A common creditor channel depends on three economically separate variables: the lender's initial exposure to the shock, the dependence of other borrowers on that lender, and the availability of alternative suppliers of financing. A creditor's global market share may be an incomplete indicator of all three. Its distress can be large relative to equity while its borrowers elsewhere depend on credit that can be replaced; in that case the effects on aggregate financing may be limited even though the original bank is severely constrained.
The IMF working paper by Van Rijckeghem and Weder (2000), Spillovers Through Banking Centers, studies bank-flow data covering 30 emerging markets and 11 banking centers. The research approach is useful because it looks at creditor-specific variation rather than only returns on assets in borrower economies. The sample and method are historical; the study does not provide a timeless estimate for offshore financial centers as a class. A regression on flows needs to distinguish reduced credit supply from falling borrowing demand, and creditor composition can change between reporting periods.
Koch and Remolona (2018) document shifts in creditor composition across three crises in their comparison of emerging Asia. Japanese banks were especially important creditors around the Asian crisis, whereas euro-area banks retrenched during the subsequent global and European crises. Other lenders' behavior sometimes offset part of the withdrawal. These episodes illustrate why a creditor-specific shock and a country-wide drop in bank claims should not be treated as synonyms.
Two counterfactuals make the interpretation clearer. In a stylized borrower economy, the outstanding external bank credit is 100, with 40 lent by the stressed group. Suppose it refuses to renew 20 of those 40 when contracts mature. If two replacement groups each provide 10 at comparable maturity and price, the total stock can be restored; the identity of lenders changes. If replacement lenders instead supply only eight, twelve units of credit remain unavailable. The original withdrawal is the same in both scenarios. The actual financing outcome differs because of market substitution.
Aggregate credit can also remain stable while its composition deteriorates. If an affected long-term loan is replaced by a short-term claim, the position at one observation date may be unchanged even as rollover risk increases. If new finance is available only against additional collateral, a borrower may conserve liquidity by canceling investment. A study based on nominal aggregate credit therefore needs complementary maturity, pricing and borrower-level measures.
For host economies, the location of the branch through which a loan was booked is more than a bookkeeping detail if the branch operates under a rule restricting transfers of assets or liquidity. At the same time, the bank's parent nationality tells researchers something about governance, risk limits, and financial conditions elsewhere in its network. These are complementary facts. An exposure study should ideally identify lending decisions by group and the actual liability and payment relationships by entity.
The most defensible policy implication is conditional on evidence of concentrated dependence. Measures improving access to substitutable financing have different economic effects from preventing all international credit contraction. A lender that withdraws credit because a borrower has become insolvent is performing a different economic function from a lender that withdraws otherwise viable credit solely because of a foreign funding shock. Policy needs to preserve this distinction to avoid sustaining fundamentally impaired positions under the label of crisis mitigation.
Dollar Funding Crises Across Two Historical Episodes
The global financial crisis demonstrated the importance of the currency composition of international-bank funding, not simply the national balance sheet of the country in which the borrowing institution was located. McGuire and von Peter (2009), in The US Dollar Shortage in Global Banking, reconstruct consolidated balance sheets and examine the dollar funding requirements generated as European and Japanese banking groups accumulated dollar assets before the crisis. The key analytical issue is the financing of those assets along the group's international network.
A dollar-denominated asset can be financed by issuing a dollar liability, drawing dollar funding from another office, or obtaining dollars through a foreign-exchange transaction combined with borrowing in another currency. These methods differ in maturity, counterparties, collateral and availability in stressed markets. An institution that has hedged the value of its exchange-rate exposure may have substantial replacement and settlement obligations. Consequently a bank's ability to fund its dollar positions depends on both the stock of its assets and the market institutions that support the recurring cash flows.
The 2020 pandemic crisis placed even stronger emphasis on the interaction between nonbank finance and internationally active banks. In their April 2020 FX swap funding study, Avdjiev, Eren, and McGuire evaluate market indicators of dollar financing costs and discuss pressures from hedging demand as well as constrained dealer supply. They report a narrowing of funding-cost indicators after central banks employed swap facilities, while noting the difficulty of distributing liquidity outside the banking system. This chronology and analysis support the relevance of international funding access; they do not alone identify the entire causal effect of the policy.
A second BIS bulletin by Eren, Schrimpf, and Sushko (2020a), the money-market fund turmoil, documents the link from withdrawals at prime funds to the availability and maturity of unsecured dollar finance received by banks. This pathway has a different initiating balance sheet from the credit losses that weakened some banks in the earlier crisis. The companion international funding analysis (2020b) finds that pressures varied between foreign banks, with differential relief for banks headquartered in jurisdictions with standing Federal Reserve swap arrangements. The findings motivate evaluating funding access by institution and market segment, not by a single global average spread.
A swap facility between monetary authorities can change which intermediaries are able to obtain the relevant currency and at what terms. The receiving central bank can allocate financing within its legal and operational framework. Pass-through from that central bank to banks and onward to nonbanks depends on institutional arrangements and intermediaries' willingness and ability to expand their balance sheets. A decline in an aggregate exchange-swap funding spread is thus an imperfect measure of the improvement in cash availability for particular firms or investment funds.
The two crises suggest complementary research designs. The 2007–09 episode is suited to tracing accumulated currency-specific balance-sheet vulnerabilities before bank funding markets became disrupted. The March 2020 episode is suited to studying a sudden broad demand for liquidity and its propagation through the interaction of funds, banks and securities markets. In both, it is necessary to separate the preexisting financial structure from the trigger and the official policy response. The shared role of dollar funding does not make the crises identical.
This economics of currency access also matters for small and emerging economies that have no direct claim on a major currency issuer. Local central-bank reserves, access to official borrowing, credit lines with private institutions, and the ability of commercial entities to maintain foreign-currency cash all vary. A useful analysis asks not only how many dollars a country or firm owes but also how its obligations are distributed over time, how committed its prospective funding is, and where its available dollars can legally and operationally be deployed.
The British Gilt Crisis and Cross-Border Nonbank Intermediation
The United Kingdom's September–October 2022 gilt market disturbance provides an unusually clear illustration of the difference between economic hedging and liquidity capacity. Liability-driven investment (LDI) strategies are intended to manage the interest-rate characteristics of defined-benefit pension liabilities. Investors can obtain duration exposure using gilts and derivatives. Leverage allows other resources to be invested elsewhere, while generating sensitivity to collateral demands and market access.
The December 2022 Bank of England Financial Stability Report describes the interaction of sharp gilt repricing, liquidity calls, and asset liquidation. The report estimates combined margin and collateral calls on LDI funds and pension schemes in excess of £70 billion for September 23 through October 14. It separately reports gilt sales of about £23 billion by LDI funds and £14 billion by defined-benefit schemes over that interval. These quantities have distinct meanings. The margin total is not an aggregate economic loss; gilt sales meet only part of the cash requirements, because investors can use existing cash or sell other assets. Netting the two official estimates against one another as if they described all transactions would misstate the mechanism.
A defined-benefit pension scheme may benefit economically from a rise in yields because the present value of its liabilities falls, while its separate LDI fund faces an urgent collateral demand. A stylized fund has a duration hedge whose mark-to-market loss is 15; the scheme's corresponding liabilities fall in value by 18. If the hedge offsets the targeted economic exposure, the change in net funding position can be favorable while the fund still needs 15 in cash under the hypothetical variation-margin terms. The scheme's improvement in long-run funding status does not itself furnish the required funds at the necessary hour.
Pooled investment vehicles add further dimensions. Large numbers of investors may need to authorize or transfer funds, and contractual arrangements may limit the speed of the transfer. The policy report describes the role of operational delays and concentrated positions. Risk management therefore requires verification that resources are both economically available and usable within a relevant settlement window. An assumed committed contribution that can arrive only next week will not satisfy today's collateral requirement.
The transaction-level gilt market study by Pinter (2023) relates greater exposure to repo and swaps before the disturbance to more extensive gilt selling within the LDI-pension-insurance sector. Its abstract also documents deterioration in trading conditions beyond that sector. The evidence supports a path from leveraged positions and financing obligations to liquidity pressure in shared markets. It does not establish that every transaction of an offshore-domiciled vehicle produced systemic harm.
The Bank announced a temporary targeted purchase operation on September 28 and stopped it on October 14. The intervention was designed to restore market functioning and give funds time to adjust. This specific objective matters when evaluating success: the counterfactual is a more disorderly market with tighter financial conditions, not the restoration of every pre-shock asset price. Fundamental repricing of government debt and additional price effects from forced selling need to be differentiated.
The cross-border relevance of the episode arises where fund vehicles, investors, managers, financing intermediaries and supervisors operate in different jurisdictions. The Bank's report explicitly calls for coordination with overseas regulators. The available source evidence supports the importance of such coordination but does not independently quantify the foreign domicile share or establish that foreign incorporation caused the crisis. A rigorous jurisdiction-level extension would require verified fund-level domicile, investment, collateral and transaction records.
Treasury Basis Trades and the Geography of Market Exposure
The US Treasury market offers a second case in which the underlying asset's credit quality and the resilience of the financing strategy are different objects of analysis. A market participant may purchase a Treasury security and take an offsetting futures position to capture a relative price difference. The position requires financing and margin management. A low expected spread can coexist with large gross positions and dependence on repeated access to secured borrowing.
In a hypothetical balance-sheet exercise, a fund buys securities worth 100 and finances 95 through repo, leaving five funded by equity. Separately it takes an offsetting futures position. If the financing expense rises, collateral values change, or the futures position generates a cash margin requirement, the fund must evaluate the liquidity of the combined position even when duration risk is approximately hedged. A cash demand of three is small relative to the gross securities position but represents 60 percent of the assumed initial equity. This percentage does not measure the position's credit loss; it illustrates the significance of immediate cash relative to a buffer. Actual financing and margin arrangements differ across institutions.
Barth et al. (2025), in The Cross-Border Trail of the Treasury Basis Trade, use regulatory and external-account information to examine US Treasury holdings of Cayman-domiciled hedge funds. Their comparison indicates an undercount in Treasury International Capital data of around $1.4 trillion at end-2024 relative to estimates of hedge-fund holdings. They explicitly caution that the entire gap is not attributable to the basis trade. Securities used as repo collateral and the institutional structure of sponsored clearing help explain the reporting difficulty.
Three distinct interpretations of the reported gap need to be kept separate. The first is a statistical discrepancy between measures intended to capture economic ownership and actual positions identified through regulatory data. The second is a potential measure of gross market exposure of a particular investor population. The third is the amount of prospective contagion from that population in a hypothetical shock. The first two can be studied using data reconciliation; the third additionally requires financing terms, liquidity buffers, price sensitivity and likely trading behavior. A $1.4 trillion discrepancy is therefore not a systemic-loss estimate.
The report also shows why misclassification can propagate through national financial accounts. If residual household holdings are calculated after accounting for other sectors, an understated rest-of-world position can overstate assets attributed to domestic households. This is an accounting transmission of measurement error, distinct from a financial contagion channel. Its policy significance is that the resulting sectoral misidentification can distort assessments of who supplies financing to a market and who might liquidate holdings under stress.
An accurate reconstruction must observe how sponsored repo separates the original trade from its cleared legal counterparties. The contractual intermediary, the actual cash lender, the economic owner of collateral, and the reporting institution can be different entities. Central clearing may improve risk management while also changing the visibility of the original cross-border relation in available statistics. The appropriate remedy is to reconcile these perspectives, not to infer that novation itself is inherently dangerous.
This case is more directly connected to offshore finance than the general gilt-market example because the research identifies a specific fund-domicile population. Even here, domicile does not reveal the nationality of ultimate investors, the location of the management team, the liquidity provision mechanism, or the location of final losses. The evidence justifies scrutinizing the statistical gap and the financing chain rather than attributing risk to a jurisdiction as an undifferentiated whole.
Nonbank Scale, Concentration, and Current Policy Evidence
The scope of nonbank intermediation matters when evaluating contagion because payments, securities, credit and derivatives connect banks to institutions with different liability and investment arrangements. The FSB's 2025 global monitoring report (2025b) records growth of 9.4 percent in the broad NBFI sector during 2024 and a share of 51.0 percent of total financial assets in its monitoring population. The report covers 29 jurisdictions accounting for more than 90 percent of global GDP. Its broad sector is not synonymous with high-risk activity and its aggregate asset total is not a measure of global offshore exposure.
The BIS bank–nonbank cross-border analysis by Aldasoro et al. (2020) reports growth in bank claims on nonbanks from $4.6 trillion at the start of 2015 to $7.5 trillion at the start of 2020. These are cross-border bank claims in the reporting statistics, not total nonbank assets. Combining the claim amount with the broad monitoring total would risk double-counting financial assets and mixing populations and dates. The two series illuminate different perspectives: the general scale of nonbank balance sheets and a particular channel linking them to banks.
Exposure to nonbanks arises in several directions. Banks finance funds through loans or secured transactions, while funds purchase bank securities or place deposits. Banks may also act as derivative counterparties, clearing providers, or custodians. Custody of a security alone need not imply a bank's exposure to its market value. Conversely, a guarantee or credit commitment may generate a contingent exposure despite the institution's initially small balance-sheet position. Analysis should use economic function and contract type rather than equate all service relationships with credit exposure.
Modern supervisory policy increasingly treats system-wide interactions as important. The FSB's 2025 leverage recommendations (2025a) address risk monitoring, core-market resilience, the conduct of leverage providers, regulatory differences and cooperation among jurisdictions. These are policy recommendations from an international coordination body, not direct enactments in every national jurisdiction. Their economic motivation is the potential for leveraged financial activities to amplify market stress or impose exposures on institutions providing credit.
Suppose a regulator imposes a constraint on one legal form of secured financing. Some transactions might become less leveraged, but others could move to instruments with economically similar exposures or to institutions outside the original regulatory perimeter. A lower reported aggregate in the constrained segment would then be ambiguous. The regulator needs evidence about total financing, substitute activities, counterparties, and liquidity risk. The costs of monitoring, disclosure and regulation should be weighed against the externalities that the interventions seek to limit.
Concentration can be portfolio-specific rather than institutional. A large number of funds with different shareholders and domiciles can hold the same securities, finance them through the same small group of dealers and be exposed to a common change in haircuts. A stress test that assumes an independent sale decision for each fund can underestimate correlated behavior. At the same time, restrictions that discourage a broad set of investors from holding a security can remove potential buyers in stressed markets. The research problem is to distinguish correlated constrained sellers from investors with capacity and willingness to absorb risk.
The FSB's 2024 money-market fund peer review should also be interpreted as a report on adoption and planned policy measures, not as an evaluation proving that the policies were effective. That limitation matters methodologically. Policy implementation is an observable institutional event; effectiveness requires evidence about behavior and counterfactual crisis outcomes. The dossier's conclusions about desirable evaluation criteria are analytical judgments rather than assertions of empirically established policy effects.
A Worked Cross-Border Stress Reconstruction
The following illustration integrates mechanisms discussed separately in earlier chapters. All numbers are hypothetical, denominated in a common unit, and chosen for arithmetic clarity. There is no empirical calibration, probability estimate, or representation of an actual financial center.
Consider an operating borrower headquartered in economy A that has issued debt through a financing affiliate in economy B. A fund resident in economy C holds securities worth 60 from that corporate group and 40 of highly liquid government securities. It owes repo financing of 70 to a banking group headquartered in economy D and has equity of 30, assuming no other assets or liabilities. For simplicity the secured lender accepts the two classes of assets at the same haircut and holds a perfected security interest over the entire 100-unit portfolio. Real arrangements may be materially different.
The initiating shock reduces the corporate securities' value by 10 percent, or six units. The government securities are initially unchanged. At unchanged obligations, the fund has assets of 94, liabilities of 70 and equity of 24. This initial loss is allocated to its equity. No debt default has yet occurred. The initial financing terms are consistent with a haircut of 30 percent and borrowing of 70 against collateral of 100.
If the secured lender increases the required haircut to 40 percent, the fund's maximum borrowing falls to 56.4, giving a 13.6-unit discrepancy relative to outstanding borrowing. The amount cannot simply be added to the six-unit equity loss as another loss. It represents a financing adjustment necessary under the stipulated new terms.
A fund with free cash outside its pledged portfolio could repay the 13.6 without selling collateral, but our simplified fund has no such resource. If it sells part of its portfolio at recorded values and applies all proceeds to repaying the lender, the collateral pool shrinks at the same time as the liability. Let the quantity sold, valued at the current recorded price, be x. The new condition is 70 minus x no greater than 0.6 times (94 minus x). Rearranging gives x no less than 34. This is larger than the initial 13.6 financing discrepancy because the fund loses collateral when it sells. After a frictionless sale and repayment of 34, assets are 60, liabilities 36 and equity 24. The haircut constraint is exactly satisfied.
This calculation is useful because an early approximation that identifies the shortfall with the securities sale would underestimate the requisite deleveraging under these conditions. It is also sensitive to assumptions. If the cash is raised by new equity not requiring collateral sales, the required injection and debt repayment can be 13.6 instead. If the lender supplies some unsecured funding or accepts new collateral from outside the pledged pool, the response changes. Asset sales at a discount reduce equity and can necessitate further adjustment. If the funding terms remain fixed, the initial price loss alone would still alter the collateral condition under the original 30-percent haircut.
Suppose several similarly exposed funds across jurisdictions all seek to sell government securities because they are more liquid than corporate debt. The first-round credit-related repricing is then transmitted to a common government-securities market through funding constraints, even though the government issuer was not the original source of loss. If aggregate sales reduce quoted prices, other leveraged institutions holding the government bonds may face additional collateral requirements. A dynamic stress test needs a price response dependent on market depth and the volume of sales. No numerical price effect can be inferred without specifying and estimating such a response.
The exercise also distinguishes parties involved in the contract. The issuing vehicle's legal residence determines part of the reporting trail. The operating parent is relevant to underlying cash flows and possibly guarantees. The fund's domicile identifies its residence for external statistics. The banking group's headquarters helps identify treasury policy and consolidated exposures. The government-securities issuer becomes exposed through market pricing without owing money to the distressed operating company. Ultimate investors in the fund bear the primary valuation loss subject to contractual arrangements. None of these roles is interchangeable.
A final judgment about systemic relevance would require knowing the scale of similar positions, funding concentration, substitutability of dealers, market liquidity and available resources. The value of an exercise at dossier level is to make accounting relationships explicit and to show where empirical evidence is needed. The values above teach the mechanism rather than establish a new financial model.
Evidence Adjudication and Scientific Research Priorities
The historical and contemporary evidence supports a layered treatment of offshore contagion. The strongest case-specific claims in the reviewed sources concern identifiable exposures, payments and changes in observed intermediation: international-bank claims can contract when a creditor becomes constrained; securities portfolios can be reallocated by ultimate issuer; margin and collateral obligations can generate urgent cash demands; and residence-based reporting can omit or misattribute positions. These claims are appropriately bounded by the institutional definitions, periods and data used in their respective sources.
Other inferences remain conditional. A larger stock of external claims may be associated with increased spillover potential, but the direction of the risk effect depends on diversification and shock properties. A specific offshore population may be highly exposed to a major securities market, but the economic implications of that exposure require information about financing and liquidation. A coordinated liquidity response may support market functioning, yet the full welfare effect also includes incentives, costs and alternative interventions.
The empirical question should be chosen before the statistical indicator. A lender-supply question calls for lender-level shocks and borrower dependence; a market-liquidity question calls for positions, financing terms and transactions; a currency-liquidity question requires dated currency-specific payments; a settlement question requires contractual obligations, netting rules and delivery timing. Seeking a common index first risks forcing different questions into an inappropriate common unit.
Several methodological improvements are particularly worthwhile. First, studies should record both immediate and ultimate counterparties, preserving the mapping rather than treating reclassification as new capital. Second, shocks and outcomes need comparable time horizons, because high-frequency funding pressure can disappear by the observation date of a quarterly stock. Third, the identification of common lenders should account for creditor switching and borrowers' substitution possibilities. Fourth, feedback from forced liquidation should be represented explicitly rather than inferred from contemporaneous correlations. Finally, the explanation of contagion should be separated from the normative determination of legal responsibility or misconduct.
Future reconstruction of the offshore financial system can develop a linked exposure dataset if regulators and data owners establish compatible access and confidentiality arrangements. A feasible research sequence would begin with individual instruments and reporting entities, reconcile corporate and legal identities, reconstruct cash requirements at several horizons, and test specific propagation hypotheses against observed episodes. It would then compare estimates under alternative aggregation and behavioral assumptions. The target is a set of testable mechanisms rather than an unobservable total of all offshore risk.
Source access in the present study ranges from complete HTML reports to authoritative abstracts and indexed primary records. The claims relying on restricted records are therefore stated at the level supported by the accessible material; no unobserved tables, causal coefficients or full-text model derivations are represented as reviewed. This limitation does not weaken the accounting identities or independently derived illustrative examples, but it limits the scope of empirical generalization. Source provenance and access levels are retained in the evidence package for subsequent scrutiny.
The author's substantive judgment is that exposure geography, duration and currency of financial commitments, and the mechanisms of balance-sheet adjustment should be analyzed together. This is a synthesis of the established literature and the dossier's comparative reasoning. It does not assert a new named theory or empirically verified novelty. Its usefulness is in organizing the evidence necessary to determine when international financial interdependence absorbs a shock and when the same interdependence transmits additional costs to the wider economy.
Conclusion
The economic significance of offshore financial contagion lies in the way contractual exposure, funding constraints, market valuations, and jurisdictionally distributed financial control interact. An offshore arrangement is relevant when it changes the paths through which an initiating disturbance is transmitted or the ability of institutions to absorb it.
Four conclusions follow. Residence, nationality and contractual risk attribution are complementary perspectives. Connectivity can diversify risk or spread it depending on the structure of financial relationships and the magnitude of shocks. A change in liquidity requirements or balance-sheet size must be distinguished from a realized loss. Prudential policy requires information on currency, maturity, legal entity, and the timing and transferability of financial resources.
The author's synthesis applies established mechanisms to the specific problem of offshore transmission, combining financial geography with balance-sheet analysis, causal identification, and distributional reasoning. The illustrative arithmetic clarifies how valuation losses can lead to larger changes in funding or intermediation without implying equivalently large credit losses. The conclusions support targeted analysis of financial functions and contractual relationships rather than a blanket assessment of financial centers.
The evidence supports multiple propagation channels and the importance of measurement. It does not establish a single global offshore contagion coefficient, an independently estimated share of any crisis attributable to offshore arrangements, or unconditional rankings of jurisdictions. Further empirical research requires linked entity-level exposure records, explicit counterfactuals, and stress exercises that incorporate cross-border liquidity and endogenous asset prices. No new named theory or priority claim is made.
References
- Acemoglu, Daron, Asuman Ozdaglar, and Alireza Tahbaz-Salehi. 2015. Systemic Risk and Stability in Financial Networks. American Economic Review 105 (2): 564–608. doi:10.1257/aer.20130456. Source.
- Aldasoro, Iñaki, Bryan Hardy, Goetz von Peter, and Philip Wooldridge. 2026. International Finance through the Lens of BIS Statistics: Offshore Activity. BIS Quarterly Review, March 16. Source.
- Aldasoro, Iñaki, Wenqian Huang, and Esti Kemp. 2020. Cross-Border Links between Banks and Non-Bank Financial Institutions. BIS Quarterly Review, September, 61–74. Source.
- Allen, Franklin, and Douglas Gale. 2000. Financial Contagion. Journal of Political Economy 108 (1): 1–33. doi:10.1086/262109. Source.
- Avdjiev, Stefan, Egemen Eren, and Patrick McGuire. 2020. Dollar Funding Costs during the Covid-19 Crisis through the Lens of the FX Swap Market. BIS Bulletin 1, April 1. Source.
- Bank of England. 2022. Financial Stability Report, December 2022. Section 5, The Resilience of Liability-Driven Investment Funds. Source.
- Barth, Daniel J., Daniel O. Beltran, Matthew Hoops, R. Jay Kahn, Emily Liu, and Maria G. Perozek. 2025. The Cross-Border Trail of the Treasury Basis Trade. FEDS Notes, October 15. Washington, DC: Board of Governors of the Federal Reserve System. Source.
- Basel Committee on Banking Supervision, Committee on Payments and Market Infrastructures, and International Organization of Securities Commissions. 2022. Review of Margining Practices. September 29. Source.
- Borio, Claudio, Robert N. McCauley, and Patrick McGuire. 2022. Dollar Debt in FX Swaps and Forwards: Huge, Missing and Growing. BIS Quarterly Review, December 5. Source.
- Brunnermeier, Markus K., and Lasse Heje Pedersen. 2009. Market Liquidity and Funding Liquidity. Review of Financial Studies 22 (6): 2201–2238. doi:10.1093/rfs/hhn098. Source.
- Cetorelli, Nicola, and Linda S. Goldberg. 2010. Global Banks and International Shock Transmission: Evidence from the Crisis. Staff Report 446, May. Federal Reserve Bank of New York. Source.
- Coppola, Antonio, Matteo Maggiori, Brent Neiman, and Jesse Schreger. 2021. Redrawing the Map of Global Capital Flows: The Role of Cross-Border Financing and Tax Havens. Quarterly Journal of Economics 136 (3): 1499–1556. doi:10.1093/qje/qjab014. Source.
- Eisenberg, Larry, and Thomas H. Noe. 2001. Systemic Risk in Financial Systems. Management Science 47 (2): 236–249. doi:10.1287/mnsc.47.2.236.9835. Source.
- Eren, Egemen, Andreas Schrimpf, and Vladyslav Sushko. 2020a. US Dollar Funding Markets during the Covid-19 Crisis: The Money Market Fund Turmoil. BIS Bulletin 14, May 12. Source.
- Eren, Egemen, Andreas Schrimpf, and Vladyslav Sushko. 2020b. US Dollar Funding Markets during the Covid-19 Crisis: The International Dimension. BIS Bulletin 15, May 12. Source.
- Financial Stability Board. 2020. Holistic Review of the March Market Turmoil. November 17. Basel: FSB. Source.
- Financial Stability Board. 2024. Thematic Review on Money Market Fund Reforms: Peer Review Report. February. Basel: FSB. Source.
- Financial Stability Board. 2025a. Leverage in Nonbank Financial Intermediation: Final Report. July 9. Basel: FSB. Source.
- Financial Stability Board. 2025b. Global Monitoring Report on Nonbank Financial Intermediation 2025. December 16. Basel: FSB. Source.
- Forbes, Kristin J., and Roberto Rigobon. 2002. No Contagion, Only Interdependence: Measuring Stock Market Comovements. Journal of Finance 57 (5): 2223–2261. doi:10.1111/0022-1082.00494. Source.
- International Monetary Fund. 2008. IMF Executive Board Integrates the Offshore Financial Center Assessment Program with the FSAP. Public Information Notice 08/82, July 9. Source.
- International Monetary Fund. 2023. Global Financial Stability Report: Safeguarding Financial Stability amid High Inflation and Geopolitical Risks. April. Chapter 2: Nonbank Financial Intermediaries: Vulnerabilities amid Tighter Financial Conditions. Washington, DC: IMF. Source.
- Koch, Cathérine (now Catherine Casanova), and Eli M. Remolona. 2018. Common Lenders in Emerging Asia: Their Changing Roles in Three Crises. BIS Quarterly Review, March 11. Source.
- McGuire, Patrick, and Goetz von Peter. 2009. The US Dollar Shortage in Global Banking and the International Policy Response. BIS Working Paper 291, October 14. Source.
- McGuire, Patrick, Goetz von Peter, and Sonya Zhu. 2024. International Finance through the Lens of BIS Statistics: Residence vs Nationality. BIS Quarterly Review, March, 73–88. Source.
- Pinter, Gabor. 2023. An Anatomy of the 2022 Gilt Market Crisis. Bank of England Staff Working Paper 1019, March 31. Source.
- Pogliani, Pamela, and Philip Wooldridge. 2022. Cross-Border Financial Centres. BIS Working Paper 1035, July 25. Source.
- Van Rijckeghem, Caroline, and Beatrice Weder. 2000. Spillovers Through Banking Centers: A Panel Data Analysis. IMF Working Paper 00/88. doi:10.5089/9781451851151.001. Source.