Systemic risk is the possibility that disruption to part or all of the financial system will seriously impair the provision of financial services and harm the wider economy. Its defining feature is not simply a large financial loss but a breakdown in functions such as payments, credit provision, risk transfer, and the financing of economic activity. A disturbance can become systemic through interconnected exposures, common asset holdings, fragile funding arrangements, or feedback between financial conditions and production.
The concept became central to financial policy after the global financial crisis of 2007–2009, although the underlying mechanisms are much older. Individually prudent decisions—such as withdrawing deposits, selling assets, or reducing lending—can collectively destabilize a system. Consequently, the safety of each institution considered separately does not guarantee the safety of the financial system as a whole.
Systemic-risk analysis examines both the accumulation of vulnerabilities over time and the distribution of exposures across institutions and markets. It connects the economics of financial crises with network analysis, institutional design, regulation, and crisis management. Although most commonly applied to finance, related ideas also appear in the study of infrastructure, supply networks, and other systems whose components depend on one another.
Definition, boundaries, and economic foundations
A widely used policy formulation, developed jointly by the International Monetary Fund, the Bank for International Settlements, and the Financial Stability Board in 2009, describes systemic risk in terms of disruption to financial services that can have serious negative consequences for the real economy. This approach makes economic function the central criterion. Failure of a small institution may be systemic if it interrupts an indispensable service, while a large investment loss need not be systemic if investors absorb it without disrupting credit or payments. Systemic importance therefore depends on size, interconnectedness, complexity, substitutability, and institutional context rather than size alone.
Systemic risk differs from credit risk, which concerns a borrower’s failure to meet obligations, and from liquidity risk, which concerns difficulty meeting payments or trading assets without substantial loss. Either can become a source of systemic disruption, but neither is inherently systemic. The term also differs from systematic risk in investment theory. Systematic risk refers to exposure to broad market movements that diversification cannot eliminate. Systemic risk concerns impairment of the system itself. A market-wide decline in share prices can involve considerable systematic risk without causing a banking or payments crisis.
An important analytical distinction separates the time dimension from the cross-sectional dimension. The time dimension concerns the buildup and subsequent reversal of leverage, credit, maturity mismatch, and asset-price optimism. The cross-sectional dimension concerns the concentration and distribution of exposures at a particular time: which institutions depend on which counterparties, which assets are widely held, and which services lack substitutes. These dimensions interact. A credit boom may make institutions simultaneously more indebted and more exposed to the same property market, increasing both cyclical fragility and the potential for correlated losses.
The economic foundation is often an externality: an institution does not bear the full social cost of its actions. A bank deciding whether to sell securities considers its own funding needs but not necessarily the price losses imposed on other holders. A lender withdrawing short-term credit may protect itself while worsening a borrower’s distress and losses elsewhere. Such effects can generate a market failure even when contracts are individually voluntary and institutions satisfy existing rules. Financial stability consequently has important public-good characteristics, although particular services such as deposits and payment processing remain privately supplied. Systemic-risk policy addresses the gap between private incentives and the resilience of the system.
Historical development and major episodes
Financial panics long predate the terminology of systemic risk. During the nineteenth century, banks commonly financed illiquid loans with deposits redeemable on demand. Doubts about one institution could prompt withdrawals from others, especially when depositors lacked reliable information about bank assets. Walter Bagehot’s Lombard Street, published in 1873, examined the responsibilities of the Bank of England during financial panic and helped establish the intellectual basis of emergency central-bank lending. The United States panic of 1907, involving trust companies and strains in money markets, contributed to the creation of the Federal Reserve System in 1913.
The Great Depression demonstrated how financial disruption could reinforce a collapse in economic activity. Banking panics in the United States between 1930 and 1933 destroyed or immobilized deposits, disrupted lending relationships, and intensified monetary contraction. Irving Fisher’s 1933 account of debt deflation emphasized a feedback mechanism in which liquidation depresses prices, increasing the real burden of outstanding debt and encouraging further liquidation. Later research distinguished several transmission channels, including changes in the money supply, deterioration in borrower balance sheets, and the destruction of information embedded in banking relationships. These explanations are complementary rather than mutually exclusive.
Postwar deposit protection and financial regulation reduced some traditional panic mechanisms without eliminating systemic vulnerability. The 1997–1998 Asian financial crisis highlighted the danger of short-term foreign-currency borrowing used to finance longer-term domestic assets. Currency depreciation increased debt burdens measured in domestic currency, while foreign lenders’ withdrawal made refinancing difficult. In 1998, the near-collapse of Long-Term Capital Management illustrated how leverage, crowded trading positions, and counterparty exposures could create concern beyond the institution’s own investors. Its private-sector recapitalization was coordinated by the Federal Reserve Bank of New York.
The 2007–2009 crisis extended systemic-risk analysis beyond insured commercial banks. Falling United States housing prices undermined mortgage-related securities, while uncertainty over exposures disrupted interbank lending and wholesale funding. Lehman Brothers filed for bankruptcy on 15 September 2008; the Reserve Primary Fund’s subsequent losses helped trigger withdrawals from money market funds. The European sovereign debt crisis then demonstrated feedback between banks and public finances. In March 2020, pandemic-related demand for cash disrupted even major government-bond markets. In 2023, the failure of Silicon Valley Bank highlighted the interaction of interest-rate exposure, concentrated uninsured deposits, and rapid withdrawals. Each episode exposed different vulnerabilities rather than a single universal crisis mechanism.
Balance-sheet amplification, runs, and asset sales
Financial fragility frequently begins with the structure of the balance sheet. An institution’s assets are financed by liabilities and equity, with equity absorbing losses before creditors do. Financial leverage magnifies the consequences of changes in asset values. If assets worth 100 are financed by debt of 95 and equity of 5, a loss of 3 reduces equity to 2, assuming liabilities remain unchanged. An asset decline of only 3 percent has therefore eliminated 60 percent of the initial equity. The institution may remain solvent, but its capacity to absorb further losses is sharply reduced.
Attempts to restore leverage can amplify the original shock. If the institution in this example seeks to return to an assets-to-equity ratio of 20 without raising new equity, it must reduce assets from 97 to 40 and use the proceeds to repay debt. That entails selling 57 of assets under the simplifying assumption that sales occur at unchanged prices. When many institutions make similar adjustments, prices may instead fall. A fire sale then reduces the marked value of assets held by institutions that have not sold anything, potentially forcing them to deleverage as well. Raising new capital can avoid some asset liquidation, but may be costly or unavailable during a panic.
Funding structure creates another amplification mechanism. Banks and other intermediaries often undertake maturity transformation, financing long-lived or illiquid assets with short-term claims. This supports useful lending but exposes the intermediary to withdrawals before its assets mature. A bank run can arise because creditors expect others to withdraw first. In the model developed by Douglas Diamond and Philip Dybvig in 1983, demandable deposit contracts can support both an efficient risk-sharing outcome and a destructive run equilibrium. Actual runs may reflect both coordination problems and genuine information about losses; the two explanations are not mutually exclusive.
Secured financing can also become unstable. In a repurchase agreement, lenders protect themselves through collateral and a haircut, the difference between collateral value and the amount lent. Rising haircuts require borrowers to provide additional collateral or repay funding. Similarly, a margin call on derivatives can generate immediate cash needs even when a position is intended to hedge a longer-term exposure. Funding liquidity and market liquidity thus interact: deteriorating trading conditions produce funding demands, while funding demands force sales that worsen trading conditions. Accounting losses, collateral rules, and contractual triggers can transmit this process rapidly across institutions without any initial default.
Financial networks and critical infrastructure
Direct contractual connections create channels through which losses and payment shortfalls can spread. If one bank defaults on an interbank loan, its creditor suffers a loss that may weaken its own ability to pay. Derivatives, guarantees, committed credit lines, and securities-financing transactions add further connections. Gross exposures matter for immediate payment demands, while legally enforceable netting and collateral influence ultimate losses. A network map based only on final net claims can therefore overlook intraday liquidity needs or the practical difficulties of obtaining collateral during a counterparty’s insolvency.
Network structure has no simple, universally stabilizing form. More connections can diversify small shocks across many creditors, reducing the consequences for any single institution. Yet those same connections can spread sufficiently large shocks widely. A concentrated network may allow core institutions to absorb disturbances efficiently, but it can also make their failure particularly disruptive. Studies of financial contagion therefore distinguish the size of the initial shock, the loss-absorbing capacity of institutions, and the topology of obligations. The number of connections alone is not a reliable measure of either safety or danger.
Indirect connections can be as important as direct claims. Institutions that own similar securities are linked through market prices even if they have no contracts with one another. Institutions can also depend on the same custodian, settlement bank, data provider, or source of short-term financing. Uncertainty about these connections creates an informational channel: creditors may withdraw from otherwise sound institutions because they cannot distinguish their exposures from those of distressed firms. Such responses need not be irrational when balance sheets are opaque and the cost of delayed withdrawal is high. The distinction between common exposure and contagion is therefore central to interpreting simultaneous losses.
Critical infrastructure introduces a functional dimension. A payment system must settle obligations reliably; disruption can prevent otherwise solvent participants from making payments. A central counterparty substitutes itself between buyers and sellers, nets exposures, and manages default risk through margin and prefunded resources. Central clearing can reduce bilateral complexity, but also concentrates risk-management responsibilities and liquidity demands. Its resilience depends on margin practices, default-management arrangements, operational continuity, and access to settlement assets. Cyber incidents or the failure of a shared technology provider can likewise interrupt essential services without first producing conventional credit losses. These operational channels show why systemic risk cannot be measured solely by adding up expected investment losses.
Non-bank finance and market-based intermediation
Non-bank financial intermediation, often discussed under the older label “shadow banking,” encompasses activities performed by investment funds, finance companies, securitization vehicles, and other entities outside conventional deposit-taking banks. The sector is heterogeneous: a closed-ended investment fund financed by long-term equity differs fundamentally from a leveraged vehicle dependent on overnight borrowing. Systemic-risk analysis therefore focuses on economic functions—leverage, liquidity transformation, maturity transformation, and risk transfer—rather than treating all non-bank institutions as equally fragile. Connections to banks through credit facilities, derivatives, and financing arrangements remain important.
The expansion of securitization before 2007 illustrates how risk can be redistributed while funding fragility accumulates. Mortgage loans were pooled into securities, and exposures were further repackaged into structured products such as collateralized debt obligations. Some vehicles financed longer-term assets by repeatedly issuing short-term commercial paper. Investors’ refusal to roll over that paper created funding gaps, while bank liquidity commitments brought some exposures back onto bank balance sheets. Credit ratings and contractual structures did not eliminate the underlying dependence on mortgage performance or the possibility that market liquidity would disappear across many products simultaneously.
Open-ended funds face a different problem when investors can redeem frequently but portfolio assets trade infrequently or at substantial cost. If redeeming investors receive a price that does not fully reflect the trading costs their withdrawals impose, remaining investors bear those costs. This can create a first-mover advantage and encourage further redemptions. It is not a feature of every fund: asset liquidity, redemption terms, valuation methods, and mechanisms for allocating transaction costs all matter. A mutual fund holding highly liquid short-term instruments therefore has a different risk profile from one offering daily redemption against illiquid credit assets.
Leveraged investment strategies can also generate urgent liquidity demands. During the United Kingdom’s liability-driven investment episode in September 2022, sharp increases in gilt yields led to collateral calls on leveraged strategies used by pension schemes. Gilt sales to raise cash contributed to further market pressure, and the Bank of England conducted temporary purchases to restore orderly conditions. The episode illustrated that an institution’s long-term funding position and its immediate cash requirements can move differently. Stablecoins create another interface between redemption promises, reserve assets, and market confidence. Their systemic relevance depends on scale, reserve composition, redemption arrangements, and connections to payments and conventional finance, rather than on the use of distributed-ledger technology alone.
Transmission to the real economy and across borders
Financial disruption becomes economically systemic when it impairs spending, production, employment, or essential transactions. One major channel is the supply of credit. Losses can reduce bank capital, leading lenders to tighten standards, reduce credit lines, or avoid renewing loans. Borrowers dependent on relationship lending may be unable to replace this financing quickly because a new lender lacks the incumbent’s information. Small and medium-sized enterprises can be particularly exposed to this problem. Even firms with viable long-term projects may cut investment or employment when external finance becomes unavailable.
Borrower balance sheets provide a second channel. Falling property or securities prices reduce available collateral and may raise the cost of borrowing. Lower revenue then weakens borrowers further, increasing expected losses for lenders and reinforcing credit contraction. This interaction is often described as a financial accelerator. Household responses can operate through both debt burdens and wealth: declining asset values, tighter credit, and greater uncertainty can reduce consumption. Financial distress can also interrupt working capital and trade finance, affecting otherwise profitable firms that need short-term funding to purchase inputs or bridge the interval between delivery and payment.
Government finances and banking systems can reinforce each other. Banks often hold substantial quantities of domestic public debt, so doubts about sovereign repayment can reduce the value of their assets and collateral. Conversely, public guarantees or bank recapitalizations can increase government liabilities. A weakening economy lowers tax revenue and raises some expenditures, further tightening the connection. This bank–sovereign feedback was prominent during the euro-area crisis. Its strength varies with bank portfolios, fiscal capacity, monetary arrangements, and the credibility of institutions responsible for crisis management; it is not an automatic consequence of banks owning government bonds.
Cross-border finance adds currency, jurisdictional, and coordination problems. An institution can be solvent in aggregate yet unable to obtain the foreign currency required to meet immediate obligations. During global stress, demand for dollar funding may rise even outside the United States, reflecting the currency denomination of contracts rather than domestic payment needs alone. Parent banks may reduce lending by foreign affiliates, while investors sell assets across countries to meet withdrawals elsewhere. Flexible exchange rates can absorb some shocks but also increase domestic-currency debt burdens where borrowers owe foreign currency. International spillovers consequently depend on the composition and maturity of external claims, not merely their total value.
Measurement, early warning, and stress testing
Systemic risk is not directly observable as a single statistic. It concerns the distribution of possible future system outcomes, including events that occur infrequently and change institutions’ behavior. Measurement therefore combines indicators of vulnerability with estimates of transmission and economic consequences. Credit growth, debt-service burdens, property prices, leverage, funding concentration, and maturity mismatch are commonly monitored. The credit-to-GDP gap, which compares the credit-to-GDP ratio with an estimated long-run trend, is one reference indicator for countercyclical capital decisions. Its interpretation is sensitive to trend estimation, structural change, and the availability of historical data.
Market-based measures use prices to infer distress dependence. CoVaR, developed by Tobias Adrian and Markus Brunnermeier, measures a conditional tail outcome for a financial system or portfolio when another institution is in a specified condition. Changes between distress and benchmark conditions are used to assess association with system vulnerability. Marginal expected shortfall measures an institution’s expected equity loss when the broader market experiences severe losses. SRISK estimates the capital shortfall an institution could face in a substantial market decline, combining information on size, leverage, and equity sensitivity. These measures answer different questions and are not interchangeable rankings of systemic importance.
A bank stress test examines how specified adverse scenarios affect losses, earnings, capital, and sometimes liquidity. A solvency exercise might combine a recession, falling property prices, rising unemployment, and borrower defaults over several years. A liquidity exercise instead examines withdrawals, collateral demands, and the capacity to obtain cash over shorter horizons. System-wide exercises attempt to include common exposures and institutions’ responses, such as asset sales or reductions in lending. Crucially, a stress scenario is generally a conditional hypothetical, not a forecast with an assigned probability. Its usefulness depends on coverage of relevant vulnerabilities and credible behavioral assumptions.
Measurement faces several identification problems. Institutions can fall together because they share an exposure rather than because one transmitted losses to another. Market prices can signal emerging fragility, but may appear unusually calm during a credit boom. Models estimated from normal periods may understate correlations, illiquidity, or nonlinear losses during stress. Data may omit off-balance-sheet commitments, foreign-currency mismatches, private financing arrangements, or dependencies on common service providers. Supervisory balance-sheet data, transaction records, market indicators, and qualitative institutional knowledge are therefore complementary. No single metric reliably identifies every systemic institution, predicts crisis timing, or captures all forms of disruption.
Prevention, regulation, and institutional design
Macroprudential policy seeks to protect the financial system as a whole, complementing supervision focused on individual institutions. A central objective is to build resilience before stress occurs rather than require simultaneous adjustment after losses have already materialized. The Basel III framework, developed after the global financial crisis, strengthened bank capital and liquidity standards and introduced additional safeguards against leverage and system-wide vulnerability. Capital requirements absorb losses; liquidity requirements address the ability to meet cash demands. Neither substitutes completely for the other, because a solvent institution can become illiquid and a liquid institution can still have losses exceeding its equity.
The countercyclical capital buffer can require banks to accumulate additional capital when authorities judge cyclical systemic risks to be increasing. Releasing it during stress can create room for banks to absorb losses and maintain lending, although actual lending also depends on funding, profitability, and borrower demand. Additional capital requirements for global systemically important banks reflect the greater consequences of their distress. The assessment methodology considers size, interconnectedness, substitutability or financial-institution infrastructure, complexity, and cross-jurisdictional activity. These measures aim to reduce both failure probability and the social costs associated with especially consequential institutions.
Liquidity standards target different horizons. The Liquidity Coverage Ratio requires sufficient high-quality liquid assets to withstand a specified 30-day stress period. The Net Stable Funding Ratio promotes a more stable funding structure relative to assets and off-balance-sheet activities over a longer horizon. Borrower-based instruments, including loan-to-value and debt-service limits, address vulnerabilities in mortgage and other lending markets. Unlike a general increase in interest rates, such tools can target particular sources of fragility, although effects may migrate into less regulated lenders or alternative financing channels.
Policy design involves trade-offs and behavioral responses. Higher resilience can reduce expected crisis losses but alter the price and availability of financial services. Uniform risk models or collateral rules may inadvertently encourage similar portfolios and simultaneous trading. Regulation attached to legal categories can induce activity to move outside those categories, while nationally implemented measures can shift business across borders. Authorities therefore combine institution-based rules with activity-based oversight and system-wide surveillance. Emerging concerns, including climate change, cybersecurity, and dependence on common technology providers, require analysis of specific transmission mechanisms. Climate scenarios, for example, distinguish physical damage from transition-related changes in asset values; long horizons and uncertain adaptation make their results different from conventional short-term stress forecasts.
Crisis management, resolution, and policy limits
Crisis management attempts to preserve essential financial functions while allocating unavoidable losses. A central bank acting as lender of last resort can supply liquidity when private funding markets cease to function. The traditional principle associates emergency lending with solvent borrowers, adequate collateral, and terms that discourage routine reliance. Application is difficult because solvency is uncertain during a panic, collateral prices may be depressed by forced sales, and lending itself can change the outcome. Market-wide facilities may address a funding disruption without requiring the authority to support a particular institution’s business model.
Deposit insurance protects eligible depositors and reduces incentives for destabilizing withdrawals, but its effectiveness depends on credible funding, coverage rules, and timely access to money. It does not directly protect all wholesale creditors or eliminate losses on bank assets. Emergency guarantees, recapitalizations, and asset-support measures can prevent abrupt disruption, yet may transfer risk to the public sector and strengthen expectations of future rescue. This creates moral hazard: creditors and managers may take greater risks if they expect losses to be absorbed elsewhere. The challenge is not simply to choose between intervention and nonintervention, but to design interventions that preserve functions without unnecessarily protecting shareholders or management.
Bank resolution provides legal mechanisms for handling a failing institution outside ordinary corporate insolvency where necessary to maintain critical services. Authorities may transfer deposits and assets to another institution, establish a bridge bank, or impose losses on designated creditors through bail-in. The Financial Stability Board’s 2011 Key Attributes of Effective Resolution Regimes established an international framework for such arrangements. Recovery and resolution plans identify options before distress, while requirements for loss-absorbing resources seek to make restructuring financially feasible. Resolution still requires operational readiness: usable records, access to payment systems, adequate liquidity, and contracts that do not collapse immediately when the process begins.
Cross-border groups create additional difficulties because legal entities, assets, creditors, and supervisors are distributed among jurisdictions. A strategy that preserves a group internationally may conflict with a national authority’s incentive to retain local assets. International cooperation, information sharing, and recognition of resolution actions can reduce these conflicts, but do not remove differences in insolvency law or fiscal responsibility. A further unresolved issue is how much market-wide support should remain available when private risk bearing proves insufficient. Credible limits on rescue support market discipline, while an inflexible refusal to intervene can permit avoidable economic disruption. The appropriate boundary depends on the threatened financial functions, available loss-allocation mechanisms, and the capacity of public institutions to act without assuming unlimited liabilities.