The FFIEC’s 2018 cyber insurance statement made a policy application one more record of a bank’s controls
The April 2018 joint statement required no insurance, but said a purchase should start from an analysis of residual risk and asked institutions to confirm they could meet the policy’s control requirements. That made the application one more description of controls.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
The April 2018 joint statement required no insurance, but said a purchase should start from an analysis of residual risk and asked institutions to confirm they could meet the policy’s control requirements. That made the application one more description of controls.
- The week On Tuesday, April 10, 2018, the members of the Federal Financial Institutions Examination Council issued a joint statement titled “Cyber Insurance and Its Potential Role in Risk Management Programs.”
- Residual risk: insurance priced against what the controls leave The statement’s first substantive instruction was about sequence.
- The decision: who takes part, and who answers for it The second part of the statement concerned people.
- The contract: coverage, conditions and the insurer’s ability to pay The third part of the statement was the longest.
- The renewal: an annual decision, not a one-time purchase The fourth part placed cyber insurance inside the annual insurance review and budgeting process.
- Why the application tied the four parts together The four parts could be read as a general purchasing checklist.
- What examiners could ask about cyber insurance The statement said it contained no new regulatory expectations, and the agencies did not require cyber insurance.
- How the statement fell on different institutions The same three pages read differently depending on the institution’s size, charter and operating model.
- Objections: cost, standardization, moral hazard and who sets the bar Four objections deserve a direct answer.
- What a bank or credit union leader could check The statement’s four parts translate into checks that need no new program.
- With hindsight The statement’s central caution, that insurance was no substitute for controls, became sharper as ransomware demands rose, by Treasury’s account.
The week
On Tuesday, April 10, 2018, the members of the Federal Financial Institutions Examination Council issued a joint statement titled “Cyber Insurance and Its Potential Role in Risk Management Programs.” The FFIEC brings together the Federal Reserve Board, the Consumer Financial Protection Bureau, the FDIC, the National Credit Union Administration, the OCC and the State Liaison Committee. The statement was three pages long.
The FFIEC’s press release was plain about what the statement did not do. “The FFIEC members do not require financial institutions to maintain cyber insurance.” It said the statement described matters institutions should consider if they were deciding whether to use cyber insurance as part of their risk management programs. It added that cyber insurance “does not diminish the importance of a sound control environment.”
The statement itself said it did “not contain any new regulatory expectations.” It pointed institutions to five booklets of the FFIEC Information Technology Examination Handbook for those expectations: Audit, Business Continuity Planning, Development and Acquisition, Information Security, and Management. It listed two Department of Homeland Security resources on cybersecurity insurance as further reading.
Its background section explained the gap the statement addressed. Traditional policies for general liability or basic business interruption “may not fully cover cyber risk exposures without special endorsement or by exclusion not cover them at all.” Coverage might also be limited and “not cover incidents caused by or tracked to outside vendors.” The statement said cyber coverage could be sold on its own or endorsed onto existing general liability, business interruption, errors and omissions, or directors’ and officers’ policies.
It distinguished two kinds of coverage. First-party coverage insured against the institution’s own direct expenses, and might address customer notification, event management, business interruption and cyber extortion. Third-party coverage protected against claims by the institution’s customers, partners or vendors arising from a cyber incident at the institution. The statement listed the risks an institution faced from cyber incidents as financial, operational, legal, compliance, strategic and reputation risks.
The core of the statement was a list of considerations in three groups. The first was involving multiple stakeholders: legal, enterprise risk management, operational risk management, finance, information technology and information security. The second was due diligence on coverage. It named policy terms, exclusions, triggers, sub-limits, the financial strength and claims-paying history of insurers, and the risk management and control requirements in the policy. The third was evaluating cyber insurance in the annual insurance review and budgeting process, including engaging the board.
Before those lists, the statement set a starting point. If management was considering cyber insurance, “the assessment of cyber insurance benefits should include an analysis of the institution’s existing cybersecurity and IT risk management programs to evaluate the potential financial impact of residual risk.” It also said: “An effective system of controls remains the primary defense against cyber threats.”
The member agencies distributed the statement through their own channels. The FDIC issued it as Financial Institution Letter FIL-16-2018, dated April 10, 2018, applicable to all FDIC-supervised institutions. On Wednesday, April 11, the OCC issued Bulletin 2018-8. It addressed the bulletin to the chief executive officers of national banks, federal branches and agencies and federal savings associations, and to technology service providers. The OCC said the statement applied to all institutions it supervised. It referred banks to the “Corporate and Risk Governance” booklet of the Comptroller’s Handbook for its general expectations on insurance programs.
A statement that announces no new expectations is easy to file and forget. We read the April 2018 cyber insurance statement as more useful than that. Its considerations, taken together, put the purchase of a cyber policy inside the institution’s control program rather than beside it. That had a consequence the statement did not spell out. A policy application, with its questions and attestations about controls, became another written account of how the institution protected itself. Examiners would read one account and the insurer another.
The statement was addressed to financial institutions supervised by the FFIEC’s members, such as banks, savings associations and credit unions. It did not reach broker-dealers or investment advisers, whose supervisors were outside the council. Those firms faced the same buying decision without this statement as a reference. We focus here on the banks and credit unions it addressed.
We take the statement in four parts. They follow the life of a cyber policy from the question of whether to buy one through each renewal, without overlap:
- Residual risk: what the institution’s controls leave uncovered, which the statement made the starting point for any purchase.
- The decision: who inside the institution should take part in deciding whether, and how much, to buy.
- The contract: what the policy actually covers, on what conditions, and from whom.
- The renewal: how the decision is revisited each year as exposures, products and controls change.
Our argument is that the third part, the contract, tied the other three together through one document: the application and the control requirements the policy imposed. An institution that described its controls one way to its examiner and another way to its insurer would create a problem in both places. After the four parts we look at why the application mattered, what examiners could ask, how the statement fell on different institutions, the main objections, and what a leader could check.
Residual risk: insurance priced against what the controls leave
The statement’s first substantive instruction was about sequence. An institution considering cyber insurance should start with an analysis of its existing cybersecurity and IT risk management programs. The purpose was to evaluate “the potential financial impact of residual risk.” Insurance would then be weighed against that residual risk.
This sequence matters because it reverses a familiar buying pattern. A buyer can start with a broker’s quote, a peer’s coverage limit or a board member’s question and work backward. The statement asked institutions to work forward from their own control assessment. The amount of risk left after controls, not the size of the institution or the policy on offer, was to drive the decision.
What a residual risk analysis would contain
For a bank or credit union, the inputs already existed in examination materials. The information security program, the business continuity plan, the vendor management records and the incident response plan each described controls and their gaps. A residual risk analysis would take the scenarios those documents addressed and ask what loss would remain if each control worked as documented.
The statement’s own categories gave the analysis a structure. First-party losses included notification costs, event management, business interruption and extortion. Third-party losses included claims from customers, partners or vendors. A credit union could estimate, for each category, what its controls would limit and what they would not. The result would be a range of plausible losses, not a single number.
Scenarios a bank or credit union could test
The statement did not list scenarios, so the choice is the institution’s. Three illustrations show how the analysis could work. They are hypothetical and chosen only to show the method.
In the first, ransomware encrypts the institution’s own servers and file shares. Controls that limit the loss include segmented networks, offline backups and a tested restoration plan. The residual risk is the cost of restoration, the business interruption during it, forensic work and any notification if data was taken. Each item maps to a first-party coverage category the statement named.
In the second, a service provider that hosts online banking suffers an outage or intrusion. The institution’s own controls do little to prevent it. What limits the loss is the provider’s resilience, the contract’s terms and the institution’s ability to communicate with members or customers. The residual risk here is large relative to the institution’s control over it, and the statement’s vendor warning applies directly.
In the third, attackers take over customer accounts and move funds. Here the loss may fall under other policies, such as a financial institution bond, depending on their wording. The residual risk analysis would need to show which policy responds to which part of the loss. Otherwise, the institution could buy overlapping coverage or leave a gap between two policies.
Why vendors belonged in the analysis
The statement flagged a specific gap: traditional coverage might not cover incidents caused by or traced to outside vendors. For institutions that outsourced core processing, online banking or card services, a large share of the relevant controls sat with the vendor. A residual risk analysis that counted only the institution’s own systems would miss the exposure the statement warned about.
In our reading, this is where the statement connected to existing third-party risk expectations. An institution already had to understand what its service providers did with its data and systems. The insurance question added a financial layer to that understanding. If a core processor’s outage stopped the institution’s operations, would any policy respond? The answer depended on contract wording the institution had to read, not on assumptions.
From residual risk to a coverage limit
The analysis also gives the limit question a basis. Without it, a limit is chosen by price or by comparison with a peer. With it, the institution can ask how much of each scenario’s residual loss it can absorb from earnings and how much it wants to transfer. The answer may differ by scenario, which is why sub-limits matter.
An institution may conclude that some scenarios are cheaper to retain than to insure. The statement allowed for that. It asked institutions to consider benefits relative to cost, and it said insurance was not required. A documented decision to retain a risk is as defensible as a decision to transfer it, if the analysis supports it.
Where residual risk analysis meets capital and budgeting
The residual risk figure also had a home outside insurance. Operational risk losses affect earnings and, for larger institutions, capital planning. An analysis built for the insurance decision could inform those processes too. The statement did not say so, and we draw the link only as our reading. A finance team that already modeled operational losses would recognize the inputs.
The decision: who takes part, and who answers for it
The second part of the statement concerned people. It listed the departments that should be involved in the cyber insurance decision: legal, enterprise risk management, operational risk management, finance, information technology and information security management. It asked that the decision-making process, including the assessment of options, be communicated to the appropriate level of management.
Why the list was broad
Each department held a different piece of the question. Legal could read policy wording and judge how exclusions might apply. Risk management could place cyber risk among the institution’s other risks. Finance could weigh premium against budget and expected loss. IT and information security could say whether the controls the insurer would ask about actually existed and worked.
The statement also asked those stakeholders to assess the sufficiency of existing controls to address cyber risk exposures “and attestation requirements for the insurance policy.” That phrase brought the security team into the purchase directly. If the policy required the institution to attest to controls, someone had to confirm the attestation was true.
Records of the decision
The statement asked that the decision-making process, including the assessment of options, be communicated to the appropriate level of management. That implies a record. A short memorandum would serve: the residual risk analysis, the options considered, the reasons for the choice and the names of those who reviewed the application. The memorandum would also note any control condition the institution could not yet meet, and the plan to meet it.
Such a record has uses beyond the purchase. If an examiner asks how the institution decided whether to insure cyber risk, the memorandum answers. If a claim is later disputed, it shows what the institution understood at signing. If staff change, it tells successors why the program looks as it does.
The risk of a decision made in one department
Insurance purchasing at a financial institution can sit with finance or a corporate risk function that buys property, liability and fidelity coverage. The statement’s list implied that cyber coverage should not be bought the same way. A finance team acting alone could buy a policy whose conditions the security team could not meet. A security team acting alone could recommend coverage without knowing how it fit the institution’s other policies.
For a community bank or small credit union, the departments on the list might be a handful of people, some wearing several hats. The principle still applied. The person who could say whether multifactor authentication covered remote access, for example, needed to see the application before it was signed. That is a control question answered by the security function, whatever its size.
The board’s role
The statement returned to governance in its third group of considerations, which asked institutions to engage the board in assessing the factors in insurance program reviews. The board did not need to choose a carrier. It did need to understand what the institution was transferring, what it was retaining, and why. The OCC bulletin pointed national banks to the section on an appropriate insurance program in the “Corporate and Risk Governance” booklet of the Comptroller’s Handbook. In our reading, that reference placed insurance within the board’s oversight of risk governance.
A board could ask management for three things to meet that role. The first is the residual cyber loss the analysis estimated, and the scenarios behind it. The second is how much of that loss the policy transfers after sub-limits and exclusions. The third is the list of control conditions in the policy the institution would find difficult to meet. The answers fit on one page and give the board a basis for its review.
The contract: coverage, conditions and the insurer’s ability to pay
The third part of the statement was the longest. Its due diligence list asked institutions to understand what they were buying in specific terms. We group the items into three questions: what the policy covers, what the policy requires, and whether the insurer can pay.
What the policy covers
The statement asked institutions to review the scope of existing or proposed coverage to identify gaps. It asked them to understand policy terms, coverage, exclusions and costs for cyber events. It warned that policy terms and language “may not be standardized,” so coverage could differ among insurers and be tailored to each institution.
It singled out three features. The first was how coverage is triggered. The second was whether certain types of incidents, with cyber terrorism given as an example, are excluded. The third was the impact that sub-limits might have on total coverage and on the claims process. A policy with a large headline limit could cap specific categories, such as extortion or business interruption, at much lower amounts.
Each of these features can be mapped to the residual risk analysis from the first part. If the analysis showed that business interruption from a vendor outage was the main exposure, the institution needed to know whether the policy covered vendor-caused interruption, what waiting period applied and what sub-limit capped it. If the analysis showed customer notification as the main cost, the relevant terms were different. The statement’s sequence made this mapping possible.
What the policy requires
The statement asked institutions to “understand risk management and control requirements outlined in the policy and ensure the institution would be able to comply.” This is the item that, in our reading, gave the statement its practical weight. A cyber policy could impose conditions on the insured. If the institution did not meet them, coverage could be at risk when a claim arrived.
Those conditions and the application questions that preceded them were, in effect, a description of the institution’s controls written for the insurer. The same controls were described for examiners in the information security program and in responses during examinations. Two descriptions of one control environment, prepared by different people for different readers, can drift apart. The statement’s instruction to involve information security and to assess attestation requirements was a guard against that drift.
Matching coverage to the institution’s own exposures
The statement’s distinction between first-party and third-party coverage gave institutions a simple test. First-party coverage pays the institution’s own costs. Third-party coverage responds to claims by others. A bank whose main exposure was customer data would weigh notification costs and third-party claims. A credit union whose main exposure was dependence on one processor would weigh business interruption. The right mix followed from the residual risk analysis, not from a standard package.
The statement also noted that coverage could be added by endorsement to existing policies, such as directors’ and officers’ coverage or errors and omissions coverage. An institution could therefore already hold some cyber coverage without a stand-alone policy. The due diligence item on reviewing existing coverage to identify gaps addressed that possibility. An inventory of every policy that mentioned cyber events would be the first step.
Coverage that depends on the vendor
The statement warned that coverage may not extend to incidents caused by or traced to outside vendors. For a contract review, that meant asking how the policy defined the insured’s computer systems. It also meant asking whether systems operated by a service provider on the institution’s behalf were included, and what the policy said about losses that began at a provider. The answers determined whether the vendor scenario in the residual risk analysis was insured at all.
The vendor’s own insurance was a separate question. A service contract might require the provider to carry coverage, but that coverage would protect the provider first. The institution would need to know whether it could claim under the provider’s policy, and with what priority among the provider’s other customers. We raise this as a contract question for counsel, not as a statement of how such policies respond.
Whether the insurer can pay
The statement asked institutions to assess the financial strength, through ratings, and the claims-paying history of insurers. It added a specific concern: the insurer’s “ability to fulfill obligations under the policy if multiple institutions file claims.” That concern pointed to an exposure peculiar to cyber risk. A single incident at a shared service provider, or a widespread attack, could produce claims from many policyholders at once.
For financial institutions, the concern was concrete. Banks and credit unions can share core processors, card networks and other service providers. An incident at one of those providers could affect a large number of institutions at the same time, and their insurers could face correlated claims. The statement did not quantify this risk. It asked institutions to consider it when choosing an insurer.
The role of outside advisers
The statement said institutions might, as appropriate, engage outside advisers such as attorneys and brokers. It framed their role as helping assess benefits against cost. We read that as permission, not a requirement, and as a reminder that the institution remained the decision-maker. A broker’s recommendation did not replace the institution’s own residual risk analysis.
The renewal: an annual decision, not a one-time purchase
The fourth part placed cyber insurance inside the annual insurance review and budgeting process. The statement listed what that review should include: benefits relative to cost, the sufficiency of coverage as exposures and products evolve, confirmation that the policy includes the coverage the institution expects, and board engagement.
What changes between renewals
A year can change each input to the decision. The institution may add a digital service, change a core processor or move workloads to a cloud provider. Its controls may improve or lapse. The insurer may revise terms, add exclusions or change its application questions. The statement’s phrase about confirming that coverage includes what the institution expects addressed that last point directly. Terms present at purchase might not survive renewal.
A renewal process that only compared premium and limit would miss those changes. A process that re-ran the residual risk analysis and re-read the conditions would catch them. The statement did not prescribe a method, but its list implied the second approach.
Cyber insurance in the budget
The statement placed the evaluation in the budgeting process as well as the insurance review. That pairing matters. Premium is one cost. The controls a policy requires are another, and they appear in the technology budget rather than the insurance line. An institution that renewed a policy with a new condition, such as a control it did not yet have, would need budget to meet it. Reviewing both lines together avoids buying a policy whose conditions the institution cannot afford to satisfy.
The same pairing can show a trade-off. Spending on a control may reduce the premium, the deductible or the need for coverage in a scenario. The statement’s request to assess benefits relative to cost invites that comparison across both budgets.
Signals that should prompt a review between renewals
An annual cycle can miss changes that happen mid-year. The statement did not discuss mid-term reviews, but its logic supports them. A new core processor, a merger, a new digital channel or a significant audit finding each changes the residual risk the policy was bought against. Each could also make an earlier answer on the application inaccurate.
A simple rule would serve. When any of those events occurs, the person who owns the policy is told, and the application answers are checked against the change. If an answer is no longer true, the institution can ask its broker or insurer how to correct it before a claim tests it.
Keeping the application and the examination record aligned
Renewal was also the point at which the institution would answer the insurer’s questions again. That made it a natural checkpoint for consistency. An institution could compare its renewal answers against its current information security program, recent audit findings and examination results. Any difference would be a signal: either the application overstated the control, or the program understated it.
We think this check is a practical step the statement supports. It costs little. It uses documents the institution already has. And it reduces two risks at once: a claim denied for a misstatement, and an examination finding that the institution’s own descriptions of its controls were inconsistent.
Why the application tied the four parts together
The four parts could be read as a general purchasing checklist. What made them specific to cyber risk was the overlap between what the insurer asked and what the examiner asked. Both could ask about the same controls, such as access management, backups, patching, vendor oversight and incident response. Both relied on the institution’s own statements about them.
That overlap gave the application unusual weight. For property or general liability coverage, we would expect the insurer’s questions to concern buildings, operations and claims history. For cyber coverage, the questions concern the same control environment the FFIEC’s handbook booklets addressed. The application became a third description of those controls, alongside the institution’s internal documentation and what it told examiners.
The statement did not use this framing. It said the statement contained no new expectations, and we do not suggest otherwise. Our point is narrower. By asking institutions to start from their control programs and to assess attestation requirements, the statement made consistency between those descriptions part of a sound purchasing decision. An institution that followed its considerations would find the application in front of its security team before signature.
Who signs, and what they attest to
An application may be signed by an officer of the institution, and the signature may represent that the answers are true. If the answers concern controls, the signer is relying on the security team’s knowledge. The statement’s request to assess attestation requirements implies that the signer should know which answers rest on whose information.
A short attestation map would make that visible. For each control question on the application, the map would name the evidence: a policy, a configuration report, an audit result or a test record. It would also name the person who confirmed it. The map could be built from documents the institution already holds and would serve again at each renewal.
What happens when the descriptions diverge
Divergence between an application and the examination record can run in two directions. If the application overstates a control, the institution may face a dispute when it files a claim. The insurer could argue that the policy was issued on an inaccurate representation. How that argument would fare depends on the policy, the facts and the governing law, and is a question for counsel.
If the application understates a control, the risk is different. The institution may pay more than it needs to, or accept conditions it already meets. It may also have told examiners something more favorable than it told the insurer, which invites questions about which account is accurate. Either way, the remedy is the same: one owner for the facts about each control, used for every description of it.
What the statement left open
The statement did not say how much coverage an institution should buy, which coverage types mattered for which institutions, or how to price residual risk. It did not address how insurance interacted with incident reporting to supervisors or law enforcement. It named cyber extortion among the costs first-party coverage might address, but not what legal limits applied to paying a ransom. Those questions came later, from other authorities.
What examiners could ask about cyber insurance
The statement said it contained no new regulatory expectations, and the agencies did not require cyber insurance. That left a question for institutions: what, if anything, would examiners ask about it? We answer only from the documents, and our reading is not a description of examination practice.
The statement directed institutions to the FFIEC IT Examination Handbook for expectations on IT risk management. Its five referenced booklets covered audit, business continuity, development and acquisition, information security, and management. The statement itself said insurance was not required. But each booklet covers controls an insurer would also ask about. An examiner reviewing those controls would see the same facts the application described.
The OCC’s bulletin added a pointer for national banks. It referred them to the “Corporate and Risk Governance” booklet of the Comptroller’s Handbook, in the section titled “Ensure an Appropriate Insurance Program.” That reference placed insurance decisions, cyber included, within the board’s and management’s governance of risk. An examiner reviewing governance could therefore ask how the institution decided its insurance program, without asking it to buy any particular coverage.
In our reading, the safest position for an institution was to be able to explain its decision either way. An institution that bought cyber coverage could show the residual risk analysis behind it. One that did not could show why it chose to retain the risk. The statement gave both a structure to follow.
How the statement fell on different institutions
The same three pages read differently depending on the institution’s size, charter and operating model. We look at four kinds of institution separately.
Community banks and small credit unions
For a small institution, the statement’s main value was permission and structure. It confirmed that cyber insurance was not required. It gave a short list of considerations that could be followed without a large staff. The risk for such institutions was a decision made by one person, without the security input the statement asked for. A small institution that outsourced much of its technology also faced the vendor gap the statement flagged.
Regional banks
A regional bank was likely to have every department on the statement’s list. Its challenge was coordination. The insurance buyer, the chief information security officer and the operational risk team might report to different executives. The statement’s request to communicate the decision process to the appropriate level of management gave a reason to bring them together on one document.
Large banking organizations
Large institutions faced a different question. Their potential cyber losses may exceed the limits available to them. For them, insurance could be a partial transfer at best, and the residual risk analysis would show how much remained with the institution regardless. The statement’s point about insurer capacity when multiple institutions file claims applied with special force. A large bank’s losses in a systemic event would compete with other policyholders’ claims.
Institutions outside the FFIEC’s reach
Broker-dealers, investment advisers and nonbank lenders faced the same purchasing decision without this statement addressed to them. Their own regulators set different expectations for cybersecurity programs. The statement’s structure was still usable as a reference, but firms outside the FFIEC agencies’ supervision would need to check with counsel how their own regulators viewed insurance in a risk program.
Objections: cost, standardization, moral hazard and who sets the bar
Four objections deserve a direct answer. Each has force, and none, in our view, undoes the argument that the application tied the decision together.
That the analysis costs more than the policy is worth
A small institution could object that a residual risk analysis is too elaborate for a modest policy. The statement did not prescribe the depth of the analysis. Its instruction was to include an analysis of existing programs, which a small institution could do in proportion to its size. The documents involved already existed. The incremental effort was in reading them against the policy, not in creating new ones.
That policies are too varied to compare
The statement itself said policy terms may not be standardized. A buyer could argue that this makes due diligence impractical. We read the warning the other way. Because terms varied, the institution could not rely on a policy’s label. It had to read the triggers, exclusions and sub-limits that mattered to its own residual risk. That is more work, but it is the work that determines whether a claim would be paid.
That insurance weakens the incentive to invest in controls
The third objection is moral hazard: an institution that buys coverage may spend less on prevention. The statement addressed this directly. It told institutions to avoid “overreliance on insurance coverage as a substitute for sound operational risk management practices.” It also said purchasing cyber insurance “does not remove the need for a sound control environment.” The application and its conditions could also work in the other direction, since an insurer asking about controls created a second outside reader of them.
That insurers will set the control bar instead of examiners
A fourth objection runs the other way. If insurers ask detailed control questions and attach conditions, they may effectively set security standards for their policyholders. That could look like a private regulator without accountability. We think the concern is real but the risk to the institution is manageable. The institution’s control standards still come from its own risk assessment and its supervisors’ expectations. An insurer’s conditions are a floor the institution agrees to by contract. Where an insurer asks for a control the institution does not need, the institution can decline the condition, negotiate it or choose another policy.
What a bank or credit union leader could check
The statement’s four parts translate into checks that need no new program. They are ways to confirm that a cyber policy, if the institution has one, matches the control environment it describes.
- Residual risk: whether a written analysis links the institution’s main cyber loss scenarios to its controls, including scenarios that start at a vendor.
- The decision: whether information security reviewed the last application and any control conditions before signature, and who signed for each attestation.
- The contract: for each major loss scenario, which coverage applies, what triggers it, which exclusions could apply and what sub-limit caps it.
- The insurer: what the institution knows about the insurer’s financial strength and how it would handle claims from many policyholders after one shared-provider incident.
- The renewal: whether answers given to the insurer at renewal were compared with the current information security program, audit findings and examination results.
Each check produces a record. Together they show that the institution treated cyber insurance as one part of its control program, as the FFIEC members described it in April 2018.
With hindsight
The statement’s central caution, that insurance was no substitute for controls, became sharper as ransomware demands rose, by Treasury’s account. On October 1, 2020, Treasury’s Office of Foreign Assets Control issued an advisory on sanctions risks of facilitating ransomware payments. It updated and replaced that advisory on September 21, 2021. The updated advisory said companies that facilitate ransomware payments on behalf of victims, “including financial institutions, cyber insurance firms, and companies involved in digital forensics and incident response,” may risk violating OFAC regulations.
The advisory said license applications for such payments would be reviewed case by case “with a presumption of denial.” It said OFAC would treat a prompt and complete report of a ransomware attack to law enforcement or other relevant agencies as a significant mitigating factor. For a bank or credit union, that meant the extortion coverage the 2018 statement described as a first-party category now carried a legal question the statement had not addressed. In our reading, whether an insurer could reimburse a payment depended on who received it.
Incident reporting to supervisors also changed. On November 18, 2021, the Federal Reserve, the FDIC and the OCC approved a final rule on computer-security incident notification. A banking organization must notify its primary federal regulator of a significant computer-security incident as soon as possible and no later than 36 hours after determining that a qualifying incident occurred. A bank service provider must notify affected banking organization customers when an incident has materially affected, or is reasonably likely to affect, them for four or more hours. Compliance was required by May 1, 2022.
The NCUA followed for credit unions. Its final rule was published March 1, 2023 and took effect September 1, 2023. It requires a federally insured credit union to report a reportable cyber incident to the NCUA as soon as possible and no later than 72 hours after it reasonably believes the incident occurred. Both rules added a supervisory clock to the first hours of an incident. Those are the same hours in which an insured institution may also need to notify its insurer under the policy’s terms. In our reading, an institution’s incident plan now has to sequence both calls.
The question the statement raised about insurers’ ability to pay many claims at once moved to the federal level. On September 29, 2022, Treasury’s Federal Insurance Office published a request for comment on a potential federal insurance response to catastrophic cyber incidents. The notice cited a June 2022 GAO report recommending that FIO and CISA jointly assess whether risks to critical infrastructure from catastrophic cyber incidents warrant a federal insurance response. Both agreed to conduct the assessment. The notice described cyber insurance as a growing market, with approximately $4 billion in direct premiums written in 2020.
The same notice said insurers, through underwriting and pricing, “can encourage or even require policyholders to implement strong cybersecurity standards and controls.” That is the mechanism the 2018 statement pointed to when it asked institutions to understand control requirements in the policy. We think the insurer had become a second outside reader of a financial institution’s controls, alongside the examiner.
The statement itself has been revised once. The FDIC’s page for FIL-16-2018, as checked on November 15, 2026, says the attachment was revised on June 2, 2026 to remove references to reputation risk. The original 2018 text listed reputation among the risks institutions face from cyber incidents. The revised attachment linked from that page lists financial, operational, legal, compliance and strategic risks. The considerations on stakeholders, due diligence and annual review are unchanged, and the revised statement still says it contains no new regulatory expectations.
What held up was the sequence: controls first, then insurance for what remains. What did not hold up was the idea that insurance could be weighed apart from law enforcement, sanctions and supervisory reporting. Each of those has since attached its own conditions to the first hours of an incident. For a bank or credit union in 2026, the cyber policy, the incident notification rules and the sanctions advisory should all describe the same response plan.
Spotted an error? Tell us and we will check it and correct the article where needed.