FFIEC’s 2017 CAT update let banks answer ‘yes with compensating controls,’ a claim of equivalent protection
On May 31, 2017, the FFIEC updated its Cybersecurity Assessment Tool with a new answer: yes with compensating controls. In our reading, that answer asserted equivalent protection, and so needed an owner, a record and someone independent to test it.
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
On May 31, 2017, the FFIEC updated its Cybersecurity Assessment Tool with a new answer: yes with compensating controls. In our reading, that answer asserted equivalent protection, and so needed an owner, a record and someone independent to test it.
- The week The first was a revised mapping.
- The claim: what the new answer asserted The first part asks what an institution said when it answered Y(C).
- The judge: who decided equivalence The second part asks who in the institution made the call.
- The record: what supported the answer The third part asks what an institution kept.
- The test: how answers were checked over time The fourth part asks who checked the claims and when.
- How the update reached different institutions The FFIEC’s members are the Federal Reserve, the FDIC, the Consumer Financial Protection Bureau, the OCC, the NCUA and the State Liaison Committee, and the release addressed financial institutions generally. [1] The descriptions below are our reading of how the four parts fit different institutions.
- Objections, in order It could.
- What security, risk and audit leaders could check The four parts translate into checks any institution can run on its own assessments, whatever framework it uses today.
- With hindsight The tool’s later history ended in a sunset.
The week
On May 31, 2017, the Federal Financial Institutions Examination Council released an update to its Cybersecurity Assessment Tool, on behalf of its members. [1] The press release named two changes. [1]
The first was a revised mapping. Appendix A of the tool maps its baseline statements to the FFIEC IT Examination Handbook. The update revised that mapping to reflect the updated Information Security and Management booklets of the handbook. [1]
The second was new answer options. The release said the updated tool would provide additional response options, allowing financial institution management to include supplementary or complementary behaviors, practices and processes that represent the institution’s current practices. [1] The updated user’s guide put the change in concrete form. Each declarative statement could be answered Y, Y(C) or N, meaning yes, yes with compensating controls, or no. [2]
The release restated the tool’s purpose. The FFIEC members developed it to help management determine the institution’s risk profile, inherent risks and cybersecurity preparedness. [1] It provided a repeatable and measurable process to measure preparedness over time. [1] Its use was voluntary. Management could choose the tool, another framework, or another risk assessment process. [1] The release also said that management of financial institutions, and of third-party service providers, are primarily responsible for assessing and mitigating their entities’ cybersecurity risk. [1]
The FFIEC’s web page for the tool marked six documents as updated in May 2017. [3] They were the full tool, the user’s guide, the inherent risk profile, the cybersecurity maturity section, Appendix A and a print-all version. [3] The overview for chief executive officers and boards, Appendix B mapping to the NIST Cybersecurity Framework and Appendix C, the glossary, were not marked as updated. [3] The page set out five steps: read the overview, read the user’s guide, complete the inherent risk profile, complete the cybersecurity maturity section, and interpret and analyze the results. [3] The page showed it was last modified on June 2, 2017. [3]
The May 2017 tool kept its two-part structure. [2] Part one, the inherent risk profile, rated activities, services and products in five categories: technologies and connection types, delivery channels, online and mobile products and technology services, organizational characteristics, and external threats. [2] Each item was rated from least to most inherent risk. Where an item fell between two levels, management was told to select the higher. [2]
Part two, cybersecurity maturity, covered five domains. [2] They were cyber risk management and oversight, threat intelligence and collaboration, cybersecurity controls, external dependency management, and cyber incident management and resilience. [2] Each domain held assessment factors, components and declarative statements across five maturity levels: Baseline, Evolving, Intermediate, Advanced and Innovative. [2]
The user’s guide set the scoring rule. All declarative statements in a maturity level, and in previous levels, must be attained and sustained to achieve that domain’s level. [2] Attained and sustained required an answer of yes or yes with compensating controls to each statement. [2] The tool was not designed to produce an overall maturity level for the institution. [2]
A footnote defined the new term. A compensating control is a management, operational or technical control employed in lieu of a recommended security control that provides equivalent or comparable protection for an information system. [2]
A self-assessment is a set of claims an institution makes about itself. The May 2017 update added what the FFIEC called additional response options. [1] The user’s guide listed three answers for each declarative statement: yes, yes with compensating controls, and no. [2] The tool defined a compensating control as one employed in place of a recommended control that provides equivalent or comparable protection. [2] So the new answer was a claim of equivalence.
Our argument is that this claim changed the work behind the tool. A plain yes says a control exists. A yes with compensating controls says a different control exists and that it protects as well. The second claim needs more: someone who decides the substitute is equivalent, a record of why, and someone independent who tests it. Without those, the new answer let an institution count gaps as attained. With them, it gave an honest way to record controls that fit the institution better than the standard ones.
We take the update in four parts:
- The claim: what a yes with compensating controls asserted.
- The judge: who in the institution decided the substitute was equivalent.
- The record: what evidence supported the answer.
- The test: how internal audit, the board and examiners checked it over time.
The first part is about meaning. The second is about people and authority. The third is about documentation. The fourth is about assurance. Together they cover the answer from the moment it is given to the moment someone else relies on it. After the four parts, we look at how the update reached different institutions, answer objections in order, and list what security, risk and audit leaders could check.
A note on sources and our reading. We describe the update from the FFIEC’s release and the May 2017 tool as archived. [1] [2] [3] We describe the tool’s later sunset from the OCC and FDIC notices of 2024. [4] [5] The tool was voluntary, and examiners’ use of it is not something the release described. [1] How an institution should govern its answers is our reading.
The claim: what the new answer asserted
The first part asks what an institution said when it answered Y(C). The tool’s footnote gives the answer. It said the institution had a control, in place of the recommended one, that provides equivalent or comparable protection. [2]
Equivalent or comparable
The definition sets a standard. Equivalent or comparable protection is a comparison between two controls against the same risk. [2] To make the comparison, someone has to know what the recommended control protects against, how well, and whether the substitute does the same. In our reading, a Y(C) answer without that comparison is a yes in name only.
The footnote also says what kind of control may compensate: management, operational or technical. [2] A management control is a policy or oversight step. An operational control is a procedure carried out by people. A technical control is implemented in systems. The breadth matters. It lets an institution offer, say, a manual review in place of an automated one. It also makes the equivalence claim harder to test, because the substitute may work in a different way.
Supplementary, complementary and compensating
The release and the guide used different words. The release spoke of supplementary or complementary behaviors, practices and processes. [1] The guide’s answer key and footnote spoke of compensating controls, defined by equivalent or comparable protection. [2] In our reading, the difference matters. A supplementary practice adds to a control. A compensating control replaces one. An institution that answered Y(C) for a practice that only added to a missing control would be claiming more than it had.
A careful institution would sort its alternative practices into the two kinds before answering. Where a practice replaced the recommended control and met the equivalence test, Y(C) fit. Where a practice helped but did not replace it, the honest answer was no, with the practice noted as a step toward the control.
Not applicable is a different answer
The tool marked some statements as possibly not applicable, for example where an institution had no wireless networks or no high-risk systems. [2] The guide said statements not applicable to all institutions are clearly designated and would not affect the maturity determination. [2] That is a separate answer from Y(C). Not applicable says the risk does not exist. Y(C) says the risk exists and a different control addresses it. Confusing the two would misstate the institution’s exposure in either direction.
Why the option was added
The release explained the purpose in its own terms. The new options let management include supplementary or complementary behaviors, practices and processes that represent the institution’s current practices. [1] In our reading, that responded to a real problem. Declarative statements describe one way to meet an objective. Without an alternative answer, an institution that met the objective another way would have to answer no. The no would understate its preparedness.
What the answer does to scoring
The scoring rule raised the stakes. A domain’s maturity level requires every statement at that level, and at all levels below, to be attained and sustained. [2] Y(C) counts as attained. [2] So one compensating control can decide whether a domain reaches a level. In our reading, that makes each Y(C) answer a possible hinge for the whole domain’s result, and worth more scrutiny than a plain yes.
An illustration from the tool
One Domain 3 statement, mapped to the Information Security booklet, says remote access to critical systems by employees, contractors and third parties uses encrypted connections and multifactor authentication. [2] Consider a bank, hypothetically, with a legacy remote access path that cannot support multifactor authentication. It might restrict that path to a fixed set of addresses, route it through a monitored jump server, and review every session. Is that equivalent protection? The answer depends on the threat the statement addresses and on evidence about how the substitute performs. The tool did not answer that question. It made the institution answer it.
Compensating at the Baseline
The level of the statement matters. The tool defined Baseline maturity as minimum expectations required by law and regulations or recommended in supervisory guidance. [2] Baseline statements carry handbook citations for that reason. [2] A Y(C) answer at Baseline therefore claims that a substitute meets a minimum supervisory expectation. In our reading, that is the strongest kind of claim the tool allows, and the one an institution should be most ready to defend. At higher levels, the statements describe practices beyond the minimum, and a substitute carries less weight if it falls short.
Attained and sustained
The rule also said attained and sustained. [2] The guide added that management may determine a statement has been sufficiently sustained based on proven results. [2] In our reading, that applies with extra force to compensating controls. A manual review that works in its first month and lapses in its sixth is not sustained. The claim of equivalence has to hold over time, beyond the day of the assessment.
Where compensating controls cluster
The distribution of Y(C) answers across the five domains says something too. In our reading, a cluster in cybersecurity controls points to technology constraints, such as legacy systems. A cluster in external dependency management points to reliance on providers. A cluster in risk management and oversight points to governance substitutes, such as informal reporting in place of formal metrics. Each cluster suggests a different remedy, and a different owner for it.
The judge: who decided equivalence
The second part asks who in the institution made the call. The tool’s answer was management. Management determines which declarative statements best fit the institution’s current practices. [2]
Management’s role and its limits
Management is a broad word. In our reading, the people best placed to describe a control are those who run it, and they are not the best placed to judge it. An infrastructure team that built a workaround has reason to think it works. A security architect who understands the threat can test the comparison. A risk officer can weigh what residual risk remains. The tool left the allocation of those roles to the institution. [2]
A named approver
We would assign each Y(C) answer a named approver with security expertise and no ownership of the control in question. The approver signs that the substitute provides equivalent or comparable protection, and states why. [2] That separation is cheap. It also changes the conversation. A control owner who knows an independent approver will review the claim writes a better description of it.
Where the board fits
The tool said assessment results should be communicated to the chief executive officer and board. [2] Its baseline oversight statements included designated members of management being held accountable by the board for the information security program, and an annual written report to the board on its status. [2] In our reading, a board receiving a maturity result should see how many Y(C) answers supported it. A domain at Evolving with no compensating controls is a different fact from a domain at Evolving that depends on several.
Staffing for the judgment
The tool’s own staffing statements bear on this part. At Baseline, information security roles and responsibilities have been identified, and processes exist to identify additional expertise needed. [2] At Evolving, management with appropriate knowledge and experience leads cybersecurity efforts, and staff with cybersecurity responsibilities have the qualifications to perform their tasks. [2] Judging equivalence is one of those tasks. An institution that lacks the expertise in house can bring it in, but the decision still belongs to management. [1]
Conflicts in the answer
The people answering a self-assessment have an interest in its result. A security team whose budget depends on showing gaps may answer conservatively. A team whose performance is judged on maturity may answer generously. In our reading, Y(C) is where either bias has the most room, because the answer turns on judgment. An approver from outside the reporting line of the control owner, with no stake in the score, reduces that room.
The chief information security officer’s position
The chief information security officer, where one exists, is the natural approver for equivalence claims, in our reading. The exception is a control the security function runs itself, such as monitoring. For those, the approver should come from risk management or an outside reviewer. The tool’s oversight statements hold designated members of management accountable to the board for the information security program. [2] That accountability is easier to exercise when the board can see who approved each claim.
Training the people who answer
The tool’s training statements set a Baseline of annual information security training, including incident response and current threats. [2] Answering an assessment well is a different skill. In our reading, the staff who complete the maturity section need guidance on what each answer means, especially Y(C) and not applicable. A short internal guide, with examples of answers the institution has accepted and rejected before, makes the next assessment more consistent and less dependent on who happens to fill it in.
Third-party providers
The release said management of third-party service providers is primarily responsible for assessing and mitigating their entities’ cybersecurity risk. [1] For an institution that outsources core processing, many controls belong to the provider. In our reading, a Y(C) answer that relies on a provider’s control is a claim about someone else’s system. The institution needs the provider’s evidence before it can make that claim, and its contract should give it access to that evidence.
The record: what supported the answer
The third part asks what an institution kept. The tool did not prescribe documentation for Y(C) answers. [2] It did describe itself as a repeatable and measurable process, and recommended periodic reassessment. [1] [2] Repetition and measurement require records.
The minimum record
In our reading, each Y(C) answer needs a short record with six elements. The statement it answers. The recommended control it replaces and why that control is not used. The substitute control, described specifically. The reasoning for equivalence. The evidence that the substitute operates. And the approver and date. A one-page form would hold all six.
Mapping to the handbook
The update’s first change was a revised Appendix A mapping baseline statements to the IT Examination Handbook. [1] Each baseline statement in the tool carries a citation to a handbook booklet and page. [2] For the remote access statement above, the citation is to the Information Security booklet. [2] In our reading, that mapping gave the record a natural anchor. A Y(C) answer to a baseline statement could cite the handbook expectation it addresses, which is the expectation an examiner would also start from.
Evidence that the substitute operates
A description of a control is not evidence that it works. Evidence is a log of reviews performed, a report from a monitoring tool, a test result, or a sample of transactions checked. In our reading, the record should name the evidence and where it is kept, so a reviewer can pull it. The tool’s audit statements point the same way. At Baseline, issues and corrective actions from internal audits and independent testing are formally tracked. [2]
Inventory and change
A compensating control can depend on a specific system or configuration. The tool’s Baseline statements include maintaining an inventory of organizational assets and a change management process for systems and security tools. [2] In our reading, each Y(C) record should list the assets it depends on. When those assets change, the change process should flag the record for review. Otherwise a system upgrade can remove a compensating control without anyone updating the answer that relied on it.
The inherent risk side of the record
The record also connects to part one of the tool. The inherent risk profile includes end-of-life systems as an item under technologies and connection types. [2] It rates an institution higher where systems are past or near end of life and support critical operations. [2] Where management is unsure between two levels, the guide says to select the higher. [2]
In our reading, end-of-life systems are one reason an institution may need compensating controls, because such systems may not support some recommended controls at all. An institution that answered Y(C) for such systems on the maturity side should see them reflected on the inherent risk side. The two parts of the tool were designed to be read together: management compares its maturity in each domain with its inherent risk. [2] A record that links each Y(C) answer to the risk profile items it touches makes that comparison honest.
Keeping the record proportionate
Not every Y(C) needs the same depth. A compensating control on a Baseline statement, which the tool maps to a handbook expectation, deserves a full record. [2] One on an Innovative statement, a level the tool describes in terms of driving innovation for the institution and the industry, needs less. [2] In our reading, depth should follow two factors: how close the statement is to a supervisory expectation, and whether the answer decides the domain’s maturity level. A record sized to those two factors keeps the paperwork in proportion.
The record as a workforce asset
Staff turn over. A compensating control designed by one engineer may be understood by no one else. The record preserves the reasoning. In our reading, it is also a training document. A new security analyst who reads the Y(C) records learns where the institution departs from standard practice and why. That knowledge is otherwise held in a few heads. When those people leave, the institution may keep the control and lose the reason for it, which is how a sound substitute decays into an unexamined habit.
The test: how answers were checked over time
The fourth part asks who checked the claims and when. The tool named three moments: periodic reassessment, reassessment when planned changes affect the risk profile, and optional independent validation. [2]
Independent validation
The guide said management may also seek independent validation, such as by the internal audit function, of the assessment process and findings. [2] The tool’s audit statements went further at higher levels. At Evolving, the independent audit function validates that the risk management function is commensurate with the institution’s risk and complexity. [2] In our reading, Y(C) answers are where independent validation adds the most value. A plain yes can be confirmed by observing the control. An equivalence claim needs judgment, and audit brings judgment from outside the team that made the claim.
What an auditor would test
An auditor testing a Y(C) answer would test three things, in our reading. First, that the substitute control exists and operates as described. Second, that it addresses the same risk as the recommended control. Third, that the evidence supports equivalent or comparable protection, as the tool’s definition requires. [2] The first is a standard test of operating effectiveness. The second and third are design questions, and they need the reasoning in the record.
Sampling for the test
An independent reviewer cannot always test every Y(C) answer each year. In our reading, a sensible sample takes all Baseline answers, all answers that decide a domain’s level, and a rotating share of the rest. Over a few cycles, every answer is tested at least once. The reviewer’s findings go into the same tracking the tool’s Baseline audit statement describes, where issues and corrective actions from audits and independent testing are formally tracked. [2]
The test has two possible results beyond pass. The substitute may operate but fall short of equivalence, in which case the answer becomes no and the gap enters the plan. Or the substitute may not operate at all, which is a control failure and a finding in its own right. Both results are useful. They keep the reported maturity tied to what the institution actually does.
Reassessment
The tool said management should consider reevaluating its inherent risk profile and maturity periodically and when planned changes, such as new products, services or connections, could affect the profile. [2] In our reading, each reassessment should revisit every Y(C) answer. Threats change, and a substitute that was comparable two years ago may fall behind. The record’s date tells the reassessment team which answers are oldest.
What to watch between assessments
The tool asked management to consider reassessment when planned changes could affect the inherent risk profile, such as new products, services or connections. [2] In our reading, the same events can undo a compensating control. A new connection can bypass a network restriction that compensated for weak authentication. A new product can route data around a manual review. The change process that triggers a risk profile review should also list the Y(C) records the change might affect.
Target levels and gap analysis
The tool described a process for improvement: determine target maturity levels, conduct a gap analysis, prioritize and plan actions, implement changes, reevaluate over time, and communicate results. [2] In our reading, a Y(C) answer can be a permanent design choice or a step toward the recommended control. The record should say which. A temporary compensating control belongs in the gap analysis, with a date for replacing it. A permanent one belongs in the institution’s standards, so it is maintained.
Reporting to the board
A board report on the assessment could show, for each domain, the maturity level reached and the number of Y(C) answers supporting it. [2] It could flag which of those are temporary and when they are due for replacement. It could show which have been independently tested. In our reading, that format gives directors a way to ask about the quality of the result without reading every statement. It also makes a rising count of compensating controls visible as a trend. A board that sees the count rise for three cycles can ask whether substitutes are standing in for investment that has been deferred. That question is easier to ask when the number is on the page, and harder to avoid. A falling count, with temporary controls replaced on schedule, shows the opposite: a plan being carried out.
Examiners
The release described the tool as voluntary, and said institutions could use it or another framework. [1] It did not describe how examiners would treat Y(C) answers. In our reading, an institution that chose to share its assessment, or whose examiners asked how it measured preparedness, would want each Y(C) answer to survive the same questions an auditor would ask. The record is what makes that possible.
How the update reached different institutions
The FFIEC’s members are the Federal Reserve, the FDIC, the Consumer Financial Protection Bureau, the OCC, the NCUA and the State Liaison Committee, and the release addressed financial institutions generally. [1] The descriptions below are our reading of how the four parts fit different institutions.
Community banks
A community bank with a small IT staff and outsourced core processing could depend on providers for many controls. Its Y(C) answers could rest on provider evidence. The judge and the record parts were hardest for it, because it had few people to separate design from approval. An outside firm could fill the approver role, under the bank’s management.
Credit unions
Credit unions faced the same structure. The FFIEC’s 2024 sunset statement, as the OCC published it, noted that the NCUA would continue to support and encourage credit unions to use its Automated Cybersecurity Examination Tool, derived from the CAT. [4] In our reading, the same governance questions apply to any alternative-control answer in that tool, and a credit union should confirm how it treats them.
Large banks
A large bank had more staff and more frameworks. It might use the tool alongside others, as the release allowed. [1] Its challenge was consistency. Different teams might answer similar statements with different compensating controls. A central register of Y(C) answers would show where the bank departed from standard practice across units.
Small credit unions with volunteer boards
For a small credit union, the questions in this article apply to whichever tool it uses. The tool asked that assessment results be communicated to the chief executive officer and board. [2] A small credit union with a volunteer board needs the board report in the fourth part to be short and clear. A count of compensating controls by domain, with dates for the temporary ones, fits on a page.
Banks with fintech partners
A bank whose products ran on a partner’s platform faced the provider question in sharp form. Delivery channels and online products were inherent risk categories in part one. [2] The controls over those channels might sit with the partner. A Y(C) answer there was a claim about the partner’s system, supported only by what the partner would share.
State-chartered institutions
The State Liaison Committee is an FFIEC member. [1] State-chartered banks and credit unions could use the tool alongside their federal supervisors’ expectations and any state requirements. In our reading, a state institution facing a state cybersecurity rule would want its Y(C) records to serve both purposes. A rule that asks for documented risk decisions is answered in part by the same record.
Holding companies with several charters
A holding company with several banks, or a bank and a trust company, may run one security program across them. In our reading, it should still record Y(C) answers by entity where the controls differ. A compensating control that works in one charter’s environment may not exist in another’s. A single consolidated answer could hide that difference from each entity’s board. Each board is responsible for its own institution, and the record should let it see its own position.
Core and service providers
Providers were outside the tool’s direct audience, but the release named their management as responsible for their own risk. [1] In our reading, providers that anticipated client questions could supply standard evidence packages for common compensating controls. That would lower each client’s cost of the record and the test.
Objections, in order
1. The new option just made scores higher
It could. A Y(C) answer counts as attained. [2] An institution that used the option loosely would report higher maturity than it had. That is the risk our argument addresses. With a named approver, a record and independent testing, the option measures real controls that a yes-or-no answer would miss.
2. Equivalence is too subjective to govern
The definition does require judgment. [2] Judgment can still be governed: by who makes it, how it is recorded and who reviews it. The tool defined the concept and left that governance to institutions. [2]
3. The tool was voluntary, so its answers did not matter
The tool was voluntary. [1] Its results were still meant to reach the chief executive and board. [2] A board that relies on a maturity result relies on each answer behind it. Voluntary use did not make the answers unimportant to the people who read them.
4. Internal audit lacks the expertise to test equivalence
Some audit teams may. The guide offered internal audit as one example of independent validation, not the only one. [2] An institution can bring in outside specialists. The independence of the test matters more than who supplies it. A reviewer from outside the team that made the claim, with enough security knowledge to understand the threat, meets the need.
5. A record for each answer is too much paperwork
A short record with six elements is less work than reconstructing the reasoning when someone asks. In our reading, the institutions with the fewest Y(C) answers would bear the least burden. Those with many would face the most, and they are the ones whose results depend most on the claims.
6. The tool was later sunset, so the update is history
The FFIEC set the tool’s sunset for August 31, 2025. [4] [5] The governance of compensating controls did not go away with it. Any framework that allows an alternative to a recommended control raises the same four questions.
7. Compensating controls hide technical debt
They can, if they become permanent by default. The tool’s improvement process asks management to set target levels, analyze gaps and plan actions. [2] A temporary compensating control with a replacement date belongs in that plan. In our reading, the record should say whether each Y(C) is temporary or permanent, so debt is visible and scheduled.
8. Answering no is always more honest
A no is honest when the objective is not met. In our reading, it is misleading when the objective is met another way, because it understates preparedness and can misdirect investment. The honest answer depends on the facts. The governance in our four parts is how an institution shows which answer the facts support.
9. Providers’ assurance reports settle provider-based answers
A provider’s independent report can be strong evidence. It may not cover the specific control the institution relies on, or the period the assessment covers. In our reading, an institution should map each provider-based Y(C) answer to the section of the provider’s report that supports it, and note any gap. The release placed responsibility for a provider’s risk on the provider’s management, and the institution still decides what to claim about it. [1]
What security, risk and audit leaders could check
The four parts translate into checks any institution can run on its own assessments, whatever framework it uses today.
- Can the institution list every answer that relies on a compensating or alternative control?
- Does each such answer have a named approver who does not own the control?
- Does each record state the recommended control, the substitute, the reasoning for equivalence and where the evidence is kept?
- Has an independent function tested a sample of those answers in the past year?
- Does a change to an underlying system trigger a review of the answers that depend on it?
The checks also travel. An institution moving to an outcome-based framework will choose its own controls for each outcome. Every one of those choices raises the equivalence question in a new form: does the chosen control achieve the outcome? The approver, the record and the independent test carry over unchanged.
The first check is where to start. An institution that cannot list its compensating controls cannot govern them, and cannot tell its board how much of its reported maturity rests on substitutes. [2]
With hindsight
The tool’s later history ended in a sunset. On August 29, 2024, the OCC issued Bulletin 2024-25, carrying an FFIEC statement that the FFIEC would sunset the Cybersecurity Assessment Tool on August 31, 2025. [4] The FDIC issued the same message in FIL-61-2024 on September 5, 2024. [5] Both said the FFIEC would remove the tool from its website on August 31, 2025. [4] [5]
The statement gave the history and the reason. The tool was released in June 2015 as a voluntary assessment tool to help institutions identify their risks and determine their cybersecurity preparedness. [4] The FFIEC said the fundamental security controls in its maturity levels are sound. [4] It said new and updated government and industry resources were available to help institutions manage cyber risk. [4]
The FFIEC decided not to update the tool to reflect new government resources, including the NIST Cybersecurity Framework 2.0 and CISA’s Cybersecurity Performance Goals. [4] It said supervised institutions could refer directly to those resources. [4] CISA had released cross-sector performance goals in 2023 and was preparing goals for the financial sector, according to the statement. [4] The FDIC added that institutions may consider industry-developed resources to assist in self-assessment. [5]
NIST had published version 2.0 of its framework on February 26, 2024. [6] It describes itself as a taxonomy of high-level cybersecurity outcomes that any organization can use to understand, assess, prioritize and communicate its efforts. [6] It does not prescribe how outcomes should be achieved, and links to resources on practices and controls that could achieve them. [6]
That design bears on our argument. An outcome-based framework leaves the choice of control to the organization. [6] Every control an institution chooses is, in effect, its own way of meeting the outcome. In our reading, the question the 2017 update raised for one answer option, whether the chosen control provides the protection the outcome needs, now applies to the whole assessment.
For credit unions, the statement noted that the NCUA would continue to support and encourage use of its Automated Cybersecurity Examination Tool, derived from the CAT. [4] The OCC’s bulletin also said the FFIEC planned a webinar for bankers on new and updated government and industry resources, scheduled for October 17, 2024. [4]
What held up from May 2017? The need to govern alternative controls. The tool’s definition of a compensating control asked for equivalent or comparable protection. [2] An outcome-based framework asks the same of every control. [6] What did not hold up was the tool itself. The FFIEC chose to retire it rather than update it. [4] Institutions that built governance around their Y(C) answers could carry that governance to whatever framework they used next. Those that treated Y(C) as a scoring option had less to carry. The FDIC’s note that institutions may consider industry-developed resources for self-assessment left the choice of tool to each institution. [5] Whatever tool an institution picks, the governance of its alternative-control answers is its own to design.
The 2017 update was small in size: a revised appendix and a new answer. [1] Its effect on governance was larger than its size, in our reading, because it moved a judgment from the FFIEC’s text into each institution. Judgments of that kind need owners, records and review, whatever framework frames them.
Spotted an error? Tell us and we will check it and correct the article where needed.