Cyber risk can turn a technology failure into a funding problem. Banks depend on connected payment systems, cloud services and outside technology providers, so an attack that disrupts one institution may affect others. If the disruption shakes creditor confidence, withdrawals or failed financing renewals can add liquidity pressure. The financial stability question is therefore broader than whether each bank can defend its own systems: it also concerns how those defenses interact across the network.
A 2026 paper by Anand, Duley and Gai, published in the Review of Finance, examines how banks choose between preventing successful attacks and maintaining operations when prevention fails. Its central point is that private incentives can pull in different directions. Banks may spend too little on protection because other institutions benefit from their investments. Yet the prospect of a run can make a bank spend more to reduce its own exposure to an attack. In some conditions, that private response can exceed the level of cybersecurity that would be socially optimal.
Shared infrastructure spreads both protection and exposure
Cybersecurity has features of a weakest-link public good. A bank that improves its defenses may reduce the chance that attackers use it as an entry point into shared systems, which can benefit other institutions as well. The reverse also holds: a weakness at one bank or technology provider may expose organizations connected to the same infrastructure.
That spillover creates a free-riding incentive. Security spending is costly, while some of its benefits accrue outside the institution paying for it. A bank weighing its own costs may therefore invest less than would be desirable for the financial system as a whole. When many institutions make that calculation independently, the resulting network may be more vulnerable than any one bank’s own risk assessment suggests.
Shared technology also complicates the meaning of diversification. Banks, custodians and other financial service providers can have different balance sheets while relying on the same digital service. A disruption at a common provider could affect multiple institutions at once. Exposure to that provider is a source of concentration across firms, even when the institutions themselves appear distinct.
Run risk can strengthen prevention incentives
A cyberattack can threaten a bank through more than direct operational damage. If systems fail or customers cannot access services, creditors may question whether the institution can meet its obligations. Depositors could withdraw funds, and other creditors might decline to renew financing. Anticipating those consequences can make a bank more willing to invest in cybersecurity, because preventing an attack may protect its funding as well as its operations.
The model’s result depends on how a bank fails. When failure is driven by insolvency, cybersecurity and creditor decisions can reinforce one another: stronger protection can support a higher debt burden, and that burden can increase the bank’s incentive to protect itself. When failure is driven by illiquidity, the relationship differs. More cybersecurity can reduce the debt burden, removing the bank’s private risk-return trade-off in the model.
Across the broader system, the run channel and free-riding work in opposite directions. Free-riding tends to weaken private investment because a bank cannot capture all the benefits of its spending. Rollover risk can strengthen an individual bank’s prevention efforts because it faces a direct cost if creditors lose confidence. If that funding risk is high enough, a bank may invest more in cybersecurity than the social optimum would call for. The result is not a simple prediction that banks always spend too little; it is a set of incentives that depends on the source of failure and the way institutions are funded.
Prevention and recovery compete for resources
The paper separates cybersecurity from operational resilience. Cybersecurity reduces the likelihood of a successful intrusion. Resilience measures—such as backup systems and recovery procedures—help a bank keep operating or restore services after an intrusion succeeds.
Both matter, but they use resources that could otherwise be directed to the other objective. The model suggests that more intense cyber threats can increase the value of operational resilience, while higher rollover risk can sharpen incentives for preventative security. That distinction matters when assessing preparedness: evidence of spending on access controls or other defenses does not by itself show that a bank can maintain critical operations during a disruption.
A useful assessment therefore asks two related questions. How difficult is it for an attacker to compromise the institution or its providers? And if defenses fail, how quickly can the bank restore essential services? Recovery plans, backup systems and business continuity arrangements speak to the second question. The capacity to recover can limit the financial consequences of a successful attack, even though it cannot prevent the intrusion itself.
System-wide dependencies shape the policy problem
Because one institution’s security choices can affect others, the authors examine ways to align private decisions with system-wide stability. They discuss minimum cybersecurity standards, negligence rules, cyber hygiene requirements and measures that affect banks’ funding structures. The model also points to the value of mapping connections between banks and technology providers. That information could help identify dependencies that an individual institution may not see in full and support cybersecurity stress tests.
The policy challenge is to address the relevant margin. A measure aimed at improving prevention has a different purpose from one designed to strengthen recovery after an attack. The paper’s framework suggests that corrective tools need to account for both choices, as well as the way bank failure arises. A single institution-level view may miss the spillovers created by shared infrastructure; a system-wide view can reveal where one weak point creates exposure for others.
What the research means for financial risk analysis
Capital, profitability, credit quality and liquidity remain central to evaluating a financial institution, but they do not capture every operational vulnerability. A bank can appear financially sound and still face material risk if a key technology service fails or if recovery arrangements prove inadequate. Cybersecurity governance, provider dependencies and continuity planning belong in the broader assessment of operational and financial risk.
Scenario analysis can also account for the path from disruption to funding pressure. A cyber event may interrupt services, undermine confidence and prompt withdrawals or financing nonrenewals. Those responses could amplify an operational shock into a liquidity problem. The relevant exercise is to trace how the disruption might move through the bank and its connections, rather than treating cyber risk as a separate technical issue.
Figure 1 in the paper sets out conditions under which a bank may fail after a successful cyberattack. It illustrates the paper’s focus on the link between an operational event and financial outcomes.

The results are hypothetical and are not an indicator of future results or returns attained by any investor. Indexes are unmanaged, do not reflect management or trading fees, and cannot be invested in directly.
The paper’s practical takeaway is that financial resilience depends on more than preventing attacks. Banks operate through connected systems, and their choices affect the institutions around them. A strong assessment should consider both the chance of compromise and the ability to sustain or restore operations if a breach occurs. As technology dependencies deepen, those shared links will remain an important part of how cyber disruptions can become financial shocks.