Insurance by BogdanovaThe Encyclopedia of Insurance and Risk Science

RESEARCH DOSSIER

Cross-Sector Insurer Bank Asset Manager Risk Transmission Networks

Abstract. Financial risk travels between insurance undertakings, banks, and asset management structures through legal obligations, contingent financing, securities holdings, derivatives, collateral, common investment strategies, and organizational dependencies. The institutional form of these links determines who bears a loss, who must deliver liquidity, and when financial stress becomes a threat to essential services. This monograph investigates direct insurer-bank exposures, asset manager common holdings, and financial contagion as interdependent mechanisms. It compares international surveillance with European Union, United Kingdom, and United States evidence, using public supervisory documents, legal texts, selected scholarly literature, and reproducible analytical examples. The evidence distinguishes contingent and realized exposures, common-cause losses and endogenous amplification, regulated entities and managed portfolios, and insolvency from liquidity interruption. It establishes the importance of identifying the ultimate legal risk bearer, examining cash-flow obligations at operational horizons, and evaluating the compatibility of simultaneous management actions. System-wide stress testing is assessed as a conditional method for identifying vulnerabilities rather than estimating event frequency. The conclusions are an original comparative synthesis of established mechanisms and public evidence. No proprietary theory, new empirical network estimate, or independently validated forecasting model is claimed.

Keywords: Insurance; banking; asset management; financial contagion; systemic risk; liquidity risk; common asset holdings; nonbank financial intermediation; derivatives; collateral; repo; asset-liability management; fund redemptions; private credit; network analysis; fire sales; macroprudential supervision; stress testing.

Introduction

An insurance promise acquires economic value through the financial capacity of the institution that undertakes it. Premium receipts, investments, reinsurance arrangements, and access to liquidity connect that promise to institutions outside the insurance contract. Banks supply payment, collateral, credit, and custody infrastructure while also issuing securities bought by insurers. Asset managers allocate insurer assets, direct collective investment vehicles, and execute market transactions whose consequences depend on the balance sheets of clients, intermediaries, and ultimate borrowers. A loss initially localized in one of these relationships can eventually change the availability or cost of risk protection and finance.

The research object is a network of financially consequential relationships between insurers, deposit-taking and investment banks, and asset management structures. The network includes the vehicles and infrastructures through which their activities meet: investment funds, special-purpose issuers, central counterparties, securities markets, payment arrangements, and intra-group contracts. Households, public authorities, and corporate borrowers belong to the boundary of the investigation when their behavior or economic losses initiate or transmit the cross-sector shocks.

The three subject-specific questions concern the observable structure of insurer-bank exposures, the ownership and incentive mechanisms shaping common holdings under asset management, and the methods by which financial contagion can be identified and evaluated across place and time. These questions require complementary forms of evidence. A contract specifies who owes a payment; a holdings report describes ownership at an observation date; a supervisory stress exercise evaluates what might happen under assumptions about shocks and responses; a historical episode provides temporal evidence of actual market disorder. The evidentiary value of each depends on the claim being examined.

The comparative research cutoff is November 19, 2026. The selected documentary evidence emphasizes 2020, 2022, 2024, 2025, and interim 2026 findings. It does not present an exhaustive census of every national market or the latest value of every exposure. Public records support institutional reconstruction and comparisons of mechanisms; confidential counterparty-level datasets would be necessary for a globally calibrated network. Mathematical examples are constructed for explanation and can be independently reproduced from the assumptions in this text.

Conceptual and Historical Foundations

Financial institutions and the functions of intermediation

A commercial bank generally operates with assets such as loans and securities and liabilities including deposits and wholesale funding. Its liquidity transformation links the timing of borrowers' repayments to depositors' and creditors' access to funds. An insurer takes on contingent obligations in exchange for premiums or other contractual contributions. Its balance sheet depends on claims, guarantees, mortality, longevity, policyholder options, discount rates, and capital. An asset management firm directs investments under contracts and may receive management or performance fees. Its own corporate balance sheet must be distinguished from the assets and liabilities of the funds or accounts it manages.

The economic function of an institution provides an initial description of its risk channels. The legal contract, however, determines the allocation and enforceability of the claims. A segregated mandate may leave ownership of a bond with an insurer; an investment fund may own an identical bond while issuing redeemable units to investors; a bank may provide custody to both. Combining these positions under the name of a management company misrepresents financial exposure. A manager's corporate creditworthiness, the value of a managed portfolio, and the operational continuity of its investment service are separate objects.

Insurer liabilities can have very different liquidity characteristics. A property insurer's catastrophe obligations may generate concentrated payments after a physical event, while a life insurer may have liabilities payable over decades with options that permit earlier surrender. A unit-linked policy transfers much of the investment performance risk according to the policy terms, but guarantees, operating obligations, and liquidity arrangements can leave other risks with the insurer. Differences in the behavior of these liabilities are indispensable when analyzing financial transmission: a twenty-year contractual maturity does not imply a twenty-year interval before every possible cash outflow.

Direct, indirect, and structural connections

Direct exposure arises from an enforceable payment or performance relationship between two legal entities. Deposits, bank-issued bonds, loans, derivatives, repo, and guarantees are examples. Indirect exposure is transmitted through variables such as securities prices and market liquidity. Structural connections arise through ownership, joint arrangements, outsourcing, and the provision of services essential for an institution to operate.

The FSB's 2025 mapping of bank-NBFI linkages identifies these overlapping categories and describes how they transmit market, credit, and funding stress (FSB, 2025). The categories sometimes describe different attributes of the same transaction. An affiliated manager may manage securities issued by the parent bank, use a bank subsidiary for custody, and obtain a credit line for a fund. A group tree describes ownership, while securities holdings, collateral, and credit contracts describe the financial links. Each needs its own records.

Financial contagion is the propagation of distress through such interactions. In empirical research, a distinction is needed between dependence created by a shared exogenous shock and dependence created by the response of another participant. The same sudden increase in yields may lower the value of two portfolios independently. If sales by one institution subsequently push prices further down and induce a loss at the other, the increment attributable to those sales is a candidate for contagion. A credible estimate must specify the counterfactual market path without that response.

Systemic risk concerns the disruption of financial functions whose impairment matters for the broader economy. A contractual loss is relevant to systemic risk when its scale, distribution, liquidity consequences, or interference with a critical function threaten wider provision of credit, insurance, payments, investment liquidity, or risk transfer. A network measures interdependence; materiality to those functions supplies the economic interpretation.

Development of the scientific problem

The study of contractual clearing is a foundational component of financial network research. In Systemic Risk in Financial Systems, Eisenberg and Noe (2001) derive simultaneous clearing of obligations within a system with limited liability and priority of debt. Their publisher abstract establishes the existence of a clearing-payment vector and uniqueness under stated regularity conditions. It supplies a conceptual baseline for bilateral default propagation; the results are not automatically transferable to insurance resolution, where claim priority, collateral, netting and timing require institutional detail.

A different line considers price-mediated transmission. Greenwood, Landier, and Thesmar's Vulnerable Banks (2015) describes the consequences of asset sales undertaken to restore leverage and the resulting exposure of other holders to price pressure. The institutional repository abstract documents applications to European sovereign-risk exposures and US financial institutions. It illustrates why standard bilateral credit accounting is incomplete for analyzing a market in which participants must transact under common constraints.

The transition from a banking-centered perspective toward system-wide analysis has been driven by crises in which market-based finance mattered to the supply of liquidity and credit. The FSB's Holistic Review (2020) documented the combined stress of March 2020. Its relevance to insurance is the interaction of cash demands and the capacity of market intermediaries, even where the original disturbance occurred outside insurance underwriting. The Bank of England's subsequent investigations into gilt market dysfunction and system-wide scenarios extend this inquiry to the interaction of institutions with different business models and trading strategies.

Historical development therefore changes the research unit from the isolated institution to a collection of contractual and market relationships operating under a specified shock. The classical inquiry into the soundness of each risk-bearing firm remains essential. Network analysis adds the question of whether the decisions made by those firms can be jointly executed.

Primary Institutional Documents and Comparative Research Design

Comparative evidence and institutional perimeter

The documentary method records the document's issuer, publication date, reference period, population, measure, and level of certainty. It treats legislative enactment, supervisory recommendation, statistical report, exploratory scenario, and research paper as different forms of authority. International policy documents are useful for identifying common supervisory objectives; domestic law and entity contracts determine the actual rights and duties in a particular transaction.

The Insurance Core Principles and ComFrame (IAIS, December 2024 update) provide the international reference for insurance supervision and an additional group-wide framework for internationally active insurance groups. They establish the institutional importance of policyholder protection and financial stability. The Basel Committee's bank-NBFI horizon-scanning report (2025) addresses links and potential risks from the perspective of bank supervision. Its published overview supports classifications and scenario inquiry; a description of risks in a horizon-scanning report is not itself an enforceable capital rule in every jurisdiction.

For the EU, the original Solvency II Directive (European Parliament and Council, 2009), Article 132, establishes the prudent-person principle and the need to identify, measure and manage investment risks and maintain appropriate portfolio liquidity and availability. Citation to the original enacted instrument establishes the origin and structure of that obligation. A current firm-specific compliance opinion would require review of subsequent amendments, transposition and relevant delegated instruments; this study does not issue such opinions.

For market evidence, the Financial Stability Board assembles jurisdictional financial statistics. The IAIS conducts insurance-specific monitoring and group-level systemic risk analysis. Their reporting populations, classifications, observation dates, and measurement objectives differ. The 2025 Global Insurance Market Report (IAIS, sections 4.2 and 4.2.1) reports a decline of 1.2% in the aggregate systemic risk score of its insurer pool between year-end 2023 and year-end 2024 on a controlled sample. The metric is a supervisory composite score, not a realized loss or insolvency probability. Its cross-sector comparison gives the bank pool a higher aggregate score under the specified common-indicator methodology.

The interim 2026 GIMAR findings (IAIS, July 2026) indicate a slight increase in aggregate systemic risk scores of global insurance groups at year-end 2025 and point toward increased roles for liquidation and interconnectedness. The 2025 edition documents a scheduled methodology revision for the 2026-2028 monitoring cycle. A numerical trend across the methodological boundary requires checking the indicator weights and normalization before treating the scores as one stable series. A direction reported within an exercise can be quoted with its reference period; independently calculating a long-run trend requires further work.

National and regional observation

European Union evidence includes banking reporting, security holdings, and insurer regulatory submissions. In its intersectoral stress analysis (ECB, 2025), the central bank combines holdings and supervisory data in the ISA framework. The published data-source description includes AnaCredit, Securities Holdings Statistics, COREP, FINREP, Solvency II, and commercial fund information. This demonstrates the possibility of integrating balance-sheet and market information without implying that all underlying entity records are publicly available.

The United Kingdom provides an important complementary observation through institutions' reported stress responses in the system-wide exploratory scenario (Bank of England, 2024). Its particular focus on core UK financial markets permits comparison of funding and trading behavior among different sectors under a common scenario. The experiment's scope, participants and scenario dates limit direct extrapolation to another currency area.

United States public information offers balance-sheet aggregates and fund filings. The SEC's Form N-PORT Data Sets cover public information extracted from eligible registered investment-fund filings, with explicit coverage exceptions, including money market funds. A researcher must identify what was publicly available on each observation date. The filing system does not supply a comprehensive database of every private fund or every insurer's beneficial ownership. A public dataset is also subject to registrant errors and extraction differences, as the SEC explains.

The Federal Reserve's May 2026 Funding Risks assessment provides selected US insurance indicators and describes developments in private credit. These records combine data from public filings, commercial compilations and official statistics. The distinction between a published chart, its restricted source population, and the researcher's own analysis matters: the availability of chart values does not create access to the underlying contracts or permit causal estimates about particular firms.

Data model and reconciliation

A usable exposure register starts with a legal entity identifier, role in the transaction, instrument, amount, currency, valuation approach, observation date, maturity, collateral status and applicable netting set. For a fund investment it also needs the fund and share class, redemption rules, investor ownership fraction and available portfolio holdings. For an intragroup commitment it needs the legal obligor, beneficiary, contractual trigger, and any transfer restrictions. The data should preserve the source jurisdiction and consolidation perimeter.

Observation Preferred unit Material qualification
Bank bond held by insurer Security and owner Par, market value, priority and maturity differ
Repo Counterparty and agreement Haircut, collateral and rollover date
Fund unit Fund, class and beneficial owner NAV, redemption terms and look-through
Derivative Contractual netting set Current exposure, margin and potential future exposure
Group support Legal entity and commitment Transferability and activation conditions
Market sale Security, time and venue Flow, direction, execution price and depth

Reconciliation first tests whether assets and liabilities are expressed on comparable terms. A published value for an insurer's bank securities cannot be equated with the bank's funding exposure to insurers without reconciling valuation, coverage and maturity. A banking group may include an asset manager; group consolidation removes internal entries but can obscure legal-entity liquidity restrictions. Accordingly, both a group view and a legal-entity view are needed. Gross positions may be reported alongside legally enforceable net positions, because collateral movements and timely payment obligations can depend on gross flows.

The FSB (2025) cross-border data discussion explains why residence-based reporting, bank consolidated statistics, and voluntary commercial records may fail to identify specific foreign counterparties. Unknown entries should remain unknown. Treating an unavailable exposure as zero systematically removes channels from the model and can understate stress. Bounds, alternative plausible network reconstructions, and transparent coverage ratios are preferable where missingness matters.

Research validity and source discipline

This monograph uses institutional sources to support limited claims about reported evidence. Journal publisher and institutional repository abstracts support the stated conceptual landmarks only. The cited official summaries are not represented as examinations of their full underlying reports. Selected sections of the 2025 IAIS and FSB reports were inspected directly. No original security-level dataset was assembled, no confidential exposure matrix was recovered, and no regression or empirical stress estimation was performed. Those are limits of the empirical scope rather than invented results.

The international comparison tests three interpretations: differences in legal liability and liquidity timing determine the direction of direct contagion; common holdings become particularly consequential when simultaneous transactions exceed market absorption; and subgroup vulnerabilities may remain material despite favorable sector-wide aggregates. These are established research questions operationalized through source comparison and arithmetic, not assertions of independently discovered causal laws.

Insurer Bank Exposures

Funding and credit relationships

Insurers can extend funding to banks through deposits, certificates, bonds, and money market instruments. The bank supplies a claim against its balance sheet; the insurer accepts risks associated with payment, recovery, maturity and access to funds. Deposits held for operational purposes differ from securities held as investments. A short-maturity placement may be inexpensive to replace in ordinary conditions but becomes important when many holders simultaneously withdraw or refuse renewal.

Banks can provide credit to insurers or affiliated vehicles through revolving facilities, secured loans, repo and other contractual arrangements. A drawn bank credit facility appears as a loan asset and a borrower liability. An undrawn commitment is a contingent source of cash to the insurer and a possible future liquidity demand on the bank. An available line on a reporting date supplies incomplete evidence of future capacity because contractual eligibility, collateral, covenant status, currency and operational processes govern actual drawings.

The FSB (2025) documents two-way domestic funding channels and the dominance of loans and debt securities in bank exposures to nonbanks, while nonbank claims on banks are often deposits and debt instruments. The observation is sectoral; the aggregate category includes institutions other than insurers. It supports the direction of possible contractual transmission and the need to disaggregate by instrument rather than a claim that all insurers are net bank funders.

Credit losses travel through the enforceable financial claim. If a bank defaults on a senior unsecured bond of face value 20 million monetary units and recovery is 60%, the holder's loss relative to face value is 8 million. If its initial carrying value was 18 million, a recovery of 12 million would instead imply an incremental carrying-value loss of 6 million before other adjustments. Distinguishing these loss baselines avoids counting the earlier discount twice. A separate temporary inability to sell an asset at its expected economic value can cause a liquidity problem without the assumed terminal default.

The transmission in the opposite direction is equally important. A bank's financing of an insurer's affiliated credit vehicle may expose the bank to vehicle-level collateral and repayment performance. Credit protection furnished by an insurer can transfer some borrower-default risk to the insurer. The recipient's risk remains dependent on the terms and enforceability of the contract and the protection provider's ability to pay under the same macroeconomic conditions. A risk transfer redistributes the obligation; it does not make the underlying corporate loss vanish.

Repos and collateral dynamics

Repo is secured financing in economic substance. Let a borrower pledge collateral worth 100 million, receive 95 million in cash and agree to a 5% initial haircut. Suppose the collateral falls to 90 million and the haircut becomes 10%. At the new haircut the collateral supports only 81 million of funding. To maintain the original 95 million borrowing, the borrower must either reduce it by 14 million or provide additional eligible collateral worth 15.556 million at the stressed haircut. The example assumes immediate repricing, no interest or other collateral, and a lender requiring full restoration of the agreed terms.

The calculation is an accounting relationship, not an estimated haircut or prediction about a specific market. It isolates the simultaneous effects of repricing and lending terms. When cash is scarce, an insurer might prefer to provide eligible securities; where the lender declines those securities, selling assets or drawing a credit line becomes necessary. Each option transfers demands to another portion of the network.

A repo ladder should record the contractual maturity for each position, expected collateral movements and whether replacement funding can be executed with the same counterparty, another lender, or a central bank facility for which the institution is eligible. Overstating rollover capacity in risk models transforms a stressed financing relationship into an assumed independent source of liquidity. Forecasts should separately examine unchanged haircuts, higher haircuts, reduced collateral eligibility and complete failure to roll selected positions.

Derivative hedges and liquidity transformation

Life insurers use derivatives to manage interest-rate, currency and other market risks. A hedge may improve the sensitivity of economic surplus to a reference rate while introducing mark-to-market settlements. Let an insurer's assets be 500 million with modified duration 7, and its liabilities 470 million with modified duration 9. Under an instantaneous parallel increase in yields of 100 basis points, the first-order changes are approximately minus 35 million in assets and minus 42.3 million in liabilities. Economic surplus increases by 7.3 million, from 30 to 37.3 million, before convexity, behavioral effects or hedging.

Suppose a receiver-fixed derivative hedge has a negative mark-to-market change of 7 million under this shock. With immediate cash variation margin, it creates an outflow of 7 million, reducing the surplus improvement to 0.3 million if its change was excluded from the initial asset calculation. The transfer also changes liquid resources, even though the economic hedge nearly neutralizes the mismatch. Variation margin pays the mark-to-market obligation; it should not be counted as a second independent loss on top of the derivative valuation.

The example illustrates why the balance-sheet and liquidity horizons must be simulated together. The decrease in the present value of long-dated liabilities does not place cash into the treasury account. Funding stress can arise while the liability valuation improves. The FSB's margin and collateral recommendations (2024) address preparedness for these calls, including governance, scenarios and collateral arrangements. Implementation requires distinguishing legally eligible collateral from assets that may take time to deliver or monetize.

Initial margin covers potential future exposure on stated terms, while variation margin settles current changes in value. Central clearing reallocates bilateral obligations through a central counterparty and its members but retains dependence on settlement, collateral and liquidity. The network must record exposure to the clearing member and the clearing infrastructure without counting the underlying derivative several times. Treatment of excess collateral and rights of reuse also affects recovery and liquidity in a default.

Group relationships and private credit

Ownership can connect a bank or insurer to an asset management business independently of contractual positions. Capital investment in the management firm, contingent support for a product, common branding, shared treasury and distribution relationships all have distinct implications. A guarantee gives its beneficiary a contractual claim; reputation may create an incentive for discretionary support. Analysts should not mechanically treat all reputational commitments as certain legal liabilities.

The Federal Reserve staff's life-insurer credit intermediation study (2025) documents a stylized factual example of a US insurer group involving an affiliated manager, loan originators, securitization vehicles and offshore reinsurance. It establishes that several legal structures may mediate the same ultimate corporate exposure. The source's estimates and example concern that structure; they cannot serve as a representative leverage estimate for all US life insurers.

Private lending may constitute diversified long-term financing suited to long insurance liabilities. Illiquidity, uncertain valuation, concentration and complex arrangements can become material when investors need cash or when deteriorating borrower quality affects several positions. The IAIS's private credit analysis (2025, private credit section) identifies these issues and the importance of considering exposures through asset-intensive reinsurance. Reinsurance transfers contractually specified obligations; the cedent's counterparty risk and recapture conditions can reconnect the exposure to its balance sheet in adverse circumstances.

Stressed group analysis should investigate the flow of claims from borrower to vehicle, from vehicle to funding provider, and from the investing insurance entity to policyholders and owners. The same loan can support multiple accounting positions, for example when a vehicle holds it and issues a tranche to an insurer. The downstream tranche and upstream loan cannot both be treated as separate ultimate corporate credit in an aggregate exposure count. They remain distinct legal liabilities for the purpose of default propagation.

Liquidity horizon and national evidence

The Federal Reserve (2026) nontraditional insurer liabilities data reports 531 billion dollars in constant 2025 dollars at the fourth quarter of 2025 and a real annual increase of 15% for the defined US life-insurer funding categories. These include funding-agreement-backed securities, Federal Home Loan Bank advances, and funds from securities lending and repo. The report also notes that the combined liabilities remain relatively small beside insurer assets. The specific size and growth refer to the source's series, not to all insurer liabilities.

In the same assessment, estimates of illiquid assets as a proportion of assets in 2024 are about 37% for life insurers and 14% for property and casualty insurers. Different liability structures and portfolio choices help motivate product-specific analysis. That difference does not identify the liquidity gap, since the horizon, contractual cash demands, hedges and readily mobilizable assets must be measured separately. An insurer with a higher illiquid share and low short-horizon obligations could be more resilient than one with a lower share but substantial surrender or financing pressure.

A cash-flow schedule should ask how much can be paid today, within five business days, and through the remaining stress horizon. A surplus forecast and a cash forecast must be reconciled but retained separately. The first attributes economic gains and losses; the second identifies receipts, payments, settlement lags and funding availability. The meaningful insurer-bank relationship is therefore a combination of valuation exposures and timed commitments.

Asset Manager Common Holdings

Managed assets and legal ownership

Asset management connects many investors to securities markets. The relationship may be a discretionary mandate, advisory service, pooled fund or delegated management arrangement. An investor can retain legal or beneficial ownership of an individual security in a managed account, while a fund investor holds an interest in a vehicle owning the portfolio. The economics of fees, securities selection and execution may be similar across these structures, but the legal capacity to redeem, change the mandate, transfer the securities or meet creditors differs.

Assets under management are consequently not equivalent to a manager's own assets or capital. The same manager can administer portfolios with different client objectives and legal constraints. Calculating the systemic significance of a manager purely from assets under management can mistake a service provider's operational reach for financial loss-absorbing obligations. Conversely, an operational outage at a broadly used manager can interrupt trading or collateral processing across separate owners, even if those clients' investment assets remain legally segregated.

Consider 100 million of an insurer's assets invested in a fund that holds 40% corporate bonds and 60% government securities. At the observation date, proportional economic look-through attributes 40 million to corporate bonds and 60 million to government securities. The insurer's balance sheet still holds a fund unit with its own valuation and redemption rights. An analysis that records both the 100 million fund investment and the 100 million underlying asset exposure as final economic holdings reports 200 million from a single investment. Use the original unit for contractual relations and a replacement look-through allocation for ultimate exposure analysis.

Fund investors may have different rights. Leverage, nonlinear derivatives, cash balances, unequal classes, gates and side pockets can invalidate simple proportional allocation. The underlying approach must therefore distinguish notional exposure, market value and contractual entitlements. If data on underlying holdings are missing for a material portion, state the uncovered percentage and test plausible ranges of concentration.

Common holdings and portfolio similarity

Let w(i,k) denote the nonnegative market-value share of asset k in institution i, with portfolio weights totaling one within a defined asset universe. A simple overlap statistic for two long-only portfolios is O(i,j) = sum over k of min(w(i,k), w(j,k)). This measure ranges from zero to one, is symmetric, and requires a common classification and observation date. It describes shared allocation, not the amount of losses or their causal origin.

Suppose a bank portfolio contains 60% government bonds, 20% corporate bonds and 20% cash. An insurer holds 30%, 60% and 10% respectively. An investment fund holds 20%, 80% and no cash. Their pairwise overlap values are 0.60 for bank-insurer, 0.40 for bank-fund, and 0.80 for insurer-fund. If cash is excluded and the remaining weights are renormalized, bank-insurer overlap becomes approximately 0.5833, bank-fund 0.45, and insurer-fund 0.8667. The asset universe changes the answer.

Broad categories may hide important differences between individual securities. An insurer holding long-dated government bonds and a bank holding short Treasury bills may look identical at sector classification level despite different duration, collateral usability and market depth. A holdings-based comparison needs security identifiers, issuer relationships and sensitivity to common credit and macroeconomic factors. For leveraged or short portfolios, signed exposure and gross risk measures should be recorded rather than forcing the long-only statistic onto unsuitable data.

Overlap is not a measure of illiquidity. Two institutions can share deep and highly liquid securities without producing substantial price pressure under ordinary sales. Their joint exposure can become costly when trading is large relative to available depth. Conversely, portfolios with few exact securities in common may share issuer risk or correlated strategies, and trades in one instrument may shift prices of closely related assets.

Incentives and trade execution

Portfolio mandates influence the size and direction of common positions. Benchmark tracking may generate overlapping securities by design. Risk budgets, asset classifications and prudential incentives may lead independently managed portfolios toward similar instrument characteristics. Contracts can also assign trading discretion to managers within limits that require action when ratings, liquidity or volatility change.

The existence of similar mandates is insufficient to infer synchronized trades. A passive holder might tolerate a price decline that triggers a leveraged investor's liquidation or a fund's redemption response. The distinction is between a shared stock of holdings and the dynamic elasticity of supply. In the same market, an insurer with stable cash flows may act as a buyer while the fund experiencing redemptions sells. Collectively these institutions may either amplify or absorb an initiating shock.

The ECB's cross-holding analysis (2025) reports significant insurer holdings of fund shares and emphasizes the feedback from investment-fund losses to insurer exposures. Its research model incorporates redemptions and forced sales as mechanisms under the specified exercise. The observation supports an institutional link between insurers and funds, not an assumption that every insurer redeems at the same threshold.

Fund redemption arrangements determine who bears transaction costs. If investors redeem at a value that fails to account for the market cost of liquidation, subsequent losses may concentrate on remaining investors. Under some contract and market conditions, this supplies an incentive to redeem early. A liquidity tool that allocates transaction costs to transacting investors may reduce that incentive, provided the valuation and cost estimates are appropriate. Restrictions on redemption can delay cash outflows but may transmit stress to investors expecting to use the proceeds for their own obligations.

The FSB's open-ended fund recommendations (2023) link redemption terms to the liquidity of fund holdings under normal and adverse conditions. The economic insight is that the terms on which investors can exit must be consistent with the portfolio's realizable liquidity. Comparing two fund structures therefore requires their specific terms and investor composition, rather than applying a single assumed run rate.

Private markets, valuation and leverage

Public bond marks permit frequent observation of price movements, even when transactions become sparse. Privately negotiated credit can have less frequent or more model-dependent valuation. A smoother valuation series is not evidence of lower economic risk. Price discovery may be delayed until a sale, credit event, independent valuation or financing renegotiation.

Leverage can be distributed across vehicles, bank finance, derivatives or investor arrangements. Consider a fund with assets of 120 million and borrowing of 20 million, so its net asset value is 100 million. A 10% loss on gross assets reduces them to 108 million. With debt unchanged the NAV is 88 million, a 12% loss. If borrowing is collateral-dependent, the fund may also need to sell assets or pledge additional collateral. This calculation excludes fees, interest and other flows; it illustrates the amplification of asset losses in a financed portfolio.

For an insurer holding 25% of that fund, and assuming a single ordinary equity share class with proportional losses, the fund stake falls from 25 to 22 million. The insurer's own creditors and policyholders are not automatically liable for the fund's entire borrowing. Any additional capital call, guarantee or recourse must appear separately as a contractual exposure. This distinction links look-through analysis to legal loss allocation.

A private credit structure can employ distinct legal entities to originate loans, manage portfolios, hold assets, borrow from banks and issue interests. The FEDS Notes intermediation example (2025) demonstrates why affiliated vehicles deserve examination alongside reported insurer portfolio holdings. Such arrangements may bring expertise or diversification while making measurement and incentives more complicated. Supervisory assessment should identify actual contract terms and economic exposure before attributing either benefit or vulnerability to a specific arrangement.

Financial Contagion

Identifying transmission and constructing counterfactuals

A research model must specify the initiating shock and the channels activated after it. At the first round, independent balance-sheet revaluations capture a common exogenous shock. Later rounds introduce behavior such as defaults, drawdowns, redemptions, collateral movements and securities sales. Comparing the model with and without one specified response helps identify a conditional contribution. It does not automatically establish observed historical causation, since the model's counterfactual remains hypothetical.

Market returns across banks and insurers may have common drivers including sovereign rates, inflation, economic growth and corporate spreads. A simple correlation of equity prices cannot distinguish those drivers from direct counterparty loss. Event studies can compare return movements around unexpected disclosures, but simultaneous news and selection can create misleading associations. Security-level sales and ownership data, actual settlement records, and contractual timing can strengthen identification of mechanisms. Even with such data, price impact and trading decisions are jointly determined.

Cross-sectional tests should account for the fact that portfolio allocations reflect prior expectations, client preferences and risk limits. An association between shared holdings and subsequent losses can be driven by selection into the same risky assets. A researcher seeking causality would need a credible source of variation in forced sales, pre-shock exposure, and market depth, while checking the assumptions that connect that variation to the outcome. Liquidity policy announcements are especially difficult to use as exogenous events when authorities intervene in response to the same stress being measured.

Bilateral clearing and loss allocation

The clearing approach in Eisenberg and Noe's 2001 financial systems study formalizes obligations that depend on other entities' payment capacity. A simplified representation uses nominal liabilities L(i,j), total obligations P(i), and proportions a(i,j) = L(i,j)/P(i) for obligors with positive obligations. If x(i) is the value of available external assets and p(i) the total payment, a one-period proportional, limited-liability clearing relation is:

p(i) = min(P(i), x(i) + sum over j of a(j,i) times p(j)).

This expression is an explanatory statement of the standard clearing setup. It assumes homogeneous priority within each obligor's modeled liabilities and treats all payments as simultaneous. Applying it to real entities requires collateral, netting, different creditor ranks, maturities, stays and resolution arrangements to be represented or bounded.

For illustration, an insurer I owes a bank B 10 million and other creditors 20 million, and B owes external creditors 40 million. I has available external assets of 24 million. In a simultaneous proportional clearing with no other receipts, I pays 24 million overall and B receives 8 million. If B's other available assets equal 34 million, B has 42 million to pay its 40 million creditors, so its creditors receive full payment. If B instead has 30 million of other assets, capacity is 38 million; pass-through loss to its creditors is 2 million.

The example reveals the role of outgoing obligations and buffer capacity. The same 2 million shortfall on the insurer claim has different consequences for the bank depending on its other resources. The model's unpaid amount is not an actuarial estimate of expected default loss. It is the result of a specified scenario and priority assumptions.

A real insurer might be legally required to prioritize policyholders, while a bank's secured claim could have a different treatment from an unsecured bond. That could change B's recovery relative to the proportional calculation. The model should therefore start with legal priority and the identity of the obligor, rather than applying a uniform loss-given-default to every edge.

Fire sales and price-mediated propagation

A standard price-impact illustration makes the relationship between sales, market depth and overlapping exposures explicit. Suppose a fund must sell corporate securities with pre-sale market value of 20 million. At an assumed linear impact of 0.25% per 10 million sold, the additional end-of-round price decline is 0.5%. A bank holding 80 million of the same security experiences a 0.4 million mark-to-market loss; an insurer holding 200 million experiences a 1 million loss. Those two gains or losses belong to the portfolios on which the market mark is applied and do not imply that the fund paid either investor directly.

The assumed impact parameter is deliberately illustrative. Market depth can be nonlinear, state-dependent and influenced by intermediaries' capacity. At twice the assumed impact, the bank and insurer losses double to 0.8 and 2 million. With no incremental price impact, their modeled second-round loss from that sale is zero even though the initiating liquidity demand persists. The range shows why overlap must be interpreted alongside execution conditions.

A simple feedback arithmetic extends this scenario. Suppose bank and insurer losses together cause a further 5 million of selling per million of mark-to-market loss, and portfolios and impact remain fixed for the illustration. The first 20 million sale creates 1.4 million in combined loss and 7 million in additional sales. Each later increment is 35% of the previous sale. Under the unchanged linear assumptions, incremental sales after the first round approach 7/(1-0.35), or approximately 10.769 million. Total sales approach 30.769 million; the cumulative modeled price effect is approximately 0.7692%.

If the behavioral response rises enough that the feedback multiplier reaches or exceeds one, this geometric calculation no longer converges. Real markets have balance-sheet limits, prices have bounds, and management and authorities may react. A nonconvergent illustrative linear calculation signals that its local assumptions cannot be extrapolated to all rounds. It does not establish an actual threshold for an observed system.

The example follows the mechanism established in Vulnerable Banks (Greenwood et al., 2015), while applying different hypothetical portfolios expressly for cross-sector explanation. It is original analytical exposition using conventional arithmetic, not a calibrated new fire-sale model.

Liquidity demand and market capacity

For a given horizon, available cash is determined by initial usable cash, realizable inflows, collateral and funding that can be accessed during that horizon. Cash needs include claims and surrenders, financing maturities, margin and collateral calls, expenses and other commitments. A liquidity gap exists when obligations due exceed available resources at that time under specified assumptions. Different currencies and legal entities can maintain separate gaps even when the consolidated total appears sufficient.

Suppose an insurer has 12 million usable cash, 5 million reliable premium inflows, and 8 million of eligible assets that can be monetized immediately after haircut. It faces claim and surrender payments of 10 million, variation margin of 13 million and financing maturities of 9 million. Available resources are 25 million against obligations of 32 million; the gap is 7 million. An uncommitted facility of 10 million closes it only if at least 7 million can actually be drawn in time. Assuming the entire facility is certain to be provided converts counterparty behavior into a false guarantee.

This arithmetic is also relevant for risk transfer within a bank group or insurer conglomerate. Cash that is encumbered, held for segregated clients, trapped by local requirements, or denominated in a currency unavailable at the needed settlement date may not be mobilizable. A single total liquidity metric should be supplemented by legal-entity, currency, maturity and collateral schedules.

The relevant system-wide comparison asks whether the aggregate demands for bank funding, fund redemptions and market execution can be satisfied with the resources each counterparty can supply. If one institution's plan to redeem a fund relies on that fund selling to dealers already assumed to be providing repo to several other participants, stress testing must reconcile those assumptions. Individual plans may each be credible in isolation while the combined plans exceed available capacity.

Crisis evidence and comparative inference

The March 2020 shock supplies evidence of actual liquidity dysfunction. The FSB's March market turmoil review reports extraordinary demand for liquidity, elevated margin calls, early-redemption incentives in some funds and limited intermediary absorption. Treasury markets experienced pressure despite the assets' usual safe-haven role. Inference for insurance is conditional: an insurer holding assets traded in the same market can be affected by market-wide liquidity stress even without default on a policy or failure of another insurer.

In September 2022, severe repricing of UK gilts led to distress in leveraged liability-driven investment arrangements. The Bank of England's gilt intervention case study (2023) explains temporary purchases intended to give affected funds and pension investors time to reduce vulnerability. This was an LDI and pension-related episode. It provides a relevant collateral and market feedback comparison for insurers with interest-rate hedges, but it cannot be described as evidence that insurance companies initiated the episode.

The two events illustrate different origins of stress and the common importance of execution capacity. In 2020 the shock involved a broad reassessment and international liquidity demand; in 2022 a concentrated repricing and leveraged positioning in a particular sovereign market were crucial. The comparison supports analysis of cash needs and common liquidity suppliers without assuming identical sector involvement or equal policy responses. Historical episodes include interventions and changing strategies, making unconstrained estimates of what would have happened without those responses speculative.

System-wide and sector stress evidence

The 2024 Insurance Stress Test (EIOPA) reports an initial aggregate solvency ratio of 221.8% for its participating EEA insurers and a fall in capital of more than EUR 270 billion in the scenario. It also reports improved results with reactive management actions. These are scenario outcomes for a specified population. They are not realized capital losses, a worldwide statistic or an estimate of the probability of the scenario. The effect of an insurer's reactive action on other institutions is a further question beyond its isolated capital ratio.

The 2025 intersectoral stress exercise (ECB) estimates bank CET1 ratio depletion of 239 basis points in the first round and around 29 basis points of further average depletion from second-round effects. The ISA analysis includes investment-fund selling and insurer exposure through securities and fund units. The values belong to the specified adverse scenario and model perimeter, with assumptions about adjustments and hedging. They should not be added to an insurer's solvency ratio change, nor used as a fixed percentage of bank assets in an unrelated jurisdiction.

The 2024 exploratory scenario report (Bank of England) describes relatively conservative bank behavior under stress despite the banks not being acutely impaired in the scenario. Participants' additional financing needs could exceed bankers' willingness to lend, and some insurers could meet calls with eligible noncash collateral. The observations direct attention toward the feasibility of liquidity responses and collateral operations. They do not establish universal refusal to roll repo or availability of all noncash assets under every contract.

These distinct exercises should be interpreted together at the level of mechanisms: an insurance test safeguards understanding of insurer liabilities and solvency, an intersectoral model quantifies conditional amplification, and an exploratory scenario tests the compatibility of participants' intended behaviors. Their populations, units and horizons preclude mechanical pooling of their numerical results.

Sensitivity, validation and uncertainty

A credible network stress test must disclose the sources and dates of exposures, scenario paths, behavioral rules, valuation models, collateral treatment and assumed interventions. Calibration should be separated from validation. Model parameters fitted to one episode may be inappropriate for another with different policy measures and dealer inventories. A reconstructed network should be tested against plausible alternative structures, including clustered and diffuse counterparty relationships.

Sensitivity analysis should vary collateral haircuts, rollover probability or decision rules, market impact, surrender behavior, recoveries, and the accessibility of liquidity buffers. Stress models built from reporting snapshots should explore changes before and during the scenario. Behavioral heterogeneity is particularly important: a life insurer with long-dated predictable cash flows, a catastrophe-exposed property insurer, a daily-dealing fund, and a leveraged private vehicle have different response functions.

Validation begins with exposure reconciliation, conservation of payments under the chosen accounting perimeter, and checks that the model does not count the same market loss or underlying asset twice. It includes comparison of predicted directions with documented crisis behavior, tests against simple hand calculations, and independent replication where the necessary data exist. Scenario outputs should distinguish what was directly observed, what was estimated by others, what is assumed, and what is inferred.

Integrated Scientific Synthesis and Future Research

Institutional diversification and collective vulnerability

The research establishes that the roles of financial institutions in cross-sector transmission are contingent on the composition and timing of their obligations. Insurers can supply stable funding and absorb price shocks when policy payments are predictable and assets can be held over a long horizon. The same institutions can transmit pressure when guarantees, surrender options, financing relationships or collateral calls require accelerated liquidity. Banks supply essential payment and credit capacity but also face constraints on their willingness to provide additional financing during adverse conditions. Asset managers organize diversified exposure and price discovery while the vehicles they administer may generate redemption, leverage and transaction pressures.

Institutional diversification offers resilience when participants respond differently to a shock. The presence of many firms under different regulatory labels supplies limited diversification if they hold the same assets, rely on the same dealers and must sell at the same time. The relevant dimension of diversity is the distribution of financial capacities and behavior across conditions.

The original analytical contribution of this study is the integration of legal risk allocation with the time-dependent execution of liquidity responses. A securities position yields a possible valuation loss, a financial claim yields contractual exposure, a liquidity schedule yields a timed funding requirement, and a market response yields a potential external effect. Treating these as four separate analytical views helps a researcher locate the channels and avoid counting them as four independent losses. This is an organizing interpretation of established evidence and methods rather than a new scientific theory.

Prudential objectives and trade-offs

Microprudential supervision emphasizes the resilience and responsible conduct of individual firms. Macroprudential assessment considers the interaction of those firms and the wider financial services they supply. The objectives are complementary but may generate different policy choices in a stress. A quick reduction of risky assets can strengthen one institution's ratio while increasing market pressure; a liquidity buffer can help the institution avoid abrupt sales but remains costly to hold; restrictions on redemption can reduce forced transactions while changing investors' access to cash.

Policy evaluation should follow the specific transmission channel. Concentrated counterparty exposures justify scrutiny of credit risk, collateral enforceability and recourse. Daily redemption with illiquid holdings justifies scrutiny of fund terms and costs. Margin-related pressure justifies liquidity preparation and the operational ability to mobilize collateral. Affiliated vehicle complexity justifies ownership and legal exposure tracing. A blunt reduction in interconnection could remove useful diversification or liquidity supply as well as vulnerability.

The international FSB and IAIS standards supply distinct yet complementary supervisory perspectives. The group-wide supervisory framework (IAIS, 2024) addresses the coordination of internationally active insurance groups, while the collateral preparedness recommendations (FSB, 2024) address market-wide cash demand mechanisms. Implementation must reflect domestic powers and institutional arrangements. Sector-tailored supervision and compatible cross-sector data definitions can preserve legal differences while improving shared assessment.

Monitoring and comparative stress design

A practical surveillance program would start with an instrument-level exposure inventory and its reporting perimeter. It would separately monitor issuer concentrations, counterparty credit and funding dependencies, redemption rights, collateral requirements, and the operation of shared services. Data aggregation would preserve the indicators' denominators, market-value basis, date, jurisdiction and entity scope. A high common holdings measure would trigger further analysis of concentration, market liquidity and possible sales, not an automatic declaration of systemic importance.

The scenario design should identify critical functions and compare a baseline shock, first-round valuation effects, and alternative response assumptions. Relevant scenario families include a sovereign-rate increase with collateral calls, a corporate spread widening with fund redemptions, bank counterparty default, loss of repo rollover, and operational interruption in a common service provider. Each scenario needs time and currency resolution; an overnight liquidity need and a multi-year reduction in insurance solvency cannot be represented by one common deadline.

Cross-border analysis must separate country of residence, country of supervision, ultimate obligor, currency, and consolidated parent. Crisis support or resolution can be constrained by different legal regimes. Shared identifiers and compatible data definitions improve the ability to reconstruct claims but cannot overcome legal restrictions on data sharing or liquidity transfers on their own.

Future comparative work should use independently observable public filings and supervisory datasets where access permits. An empirically meaningful project could compare securities holdings before and after a documented shock, match observed sales to investor categories, and estimate incremental price effects with stated identification assumptions. Another could compare simulated liquidity requirements with actual evidence from contractual settlements and financing renewals. Both require population definitions and rigorous treatment of missing identifiers, reporting delays and confidential observations.

Limits and research implications

The public evidence establishes the existence and importance of transmission channels and provides several documented case and scenario comparisons. It supplies no comprehensive current network of all insurers, banks, and asset managers. Official reports have different purposes and boundaries. The scholarly abstracts consulted establish the stated foundational concepts, not every quantitative result or formal condition in the associated papers. The stylized calculations intentionally omit taxes, complex valuation, changing risk weights, transaction costs unless specified, dynamic hedges and intervention.

The next empirical task is to connect security-level common holdings and contractual counterparty data with actual decisions over a defined stress horizon. Independent replication would need access to the inputs, a declared identification strategy, sensitivity analyses and a transparent comparison of observed outcomes versus counterfactual estimates. Regulatory reporting improvement matters because incorrect ownership or timing can generate structurally wrong inferences even when the subsequent model is mathematically precise.

Conclusion

Cross-sector risk transmission is intrinsic to the financial intermediation relationships that sustain insurance, banking, and investment management. Direct exposures distribute contractual gains and losses. Common holdings expose institutions to changes in asset prices. Financing and collateral arrangements determine how economic changes become cash demands. The resulting interactions can distribute a shock across firms and markets through more than one mechanism.

The evidence supports three specific conclusions. Insurer-bank exposures require bilateral and instrument-specific measurement that distinguishes secured from unsecured claims, current from contingent exposure, and contractual maturity from actual liquidity needs. Asset manager common holdings require look-through accounting and legal segregation of corporate and fiduciary assets, with trading behavior and market depth studied independently of overlap. Contagion analysis requires explicit separation of initial common shocks from incremental losses generated by participants' responses, and transparent acknowledgment of the limits of observed counterparty data.

International supervisory monitoring and stress exercises provide complementary evidence. Aggregate resilience can coexist with concentrated vulnerabilities, and reported scenario losses must retain their population, horizon, unit and assumptions. The most informative cross-sector assessments reconcile the plans of institutions against the financing and market resources on which they jointly depend.

The broader implication for insurance science is that claims-paying capacity includes the feasibility of the financial transactions needed to perform insured obligations when they fall due. The soundness of individual portfolios, the enforceability of contracts, and the capacity of shared markets to accommodate simultaneous liquidity needs are linked research objects. Their joint analysis strengthens the assessment of insurance's role in financial stability and clarifies the empirical questions that further research must resolve.

References