Living-will rules keep moving; the systems map stays put
The October 2019 resolution-plan rule eased living wills for large banks and let off-cycle plans skip materially unchanged system descriptions. Every later swing kept one demand: show what systems you run and who owns them.
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 October 2019 resolution-plan rule eased living wills for large banks and let off-cycle plans skip materially unchanged system descriptions. Every later swing kept one demand: show what systems you run and who owns them.
- The week On Monday, October 28, 2019, the Federal Reserve Board and the Federal Deposit Insurance Corporation announced that they had finalized changes to their resolution plan requirements for large banking organizations.
- What the 2019 rule changed about technology content The 2019 rule kept the full-plan content requirements on management information systems largely as they were.
- The map: which systems exist, who owns them, what depends on them The first question is the most basic and the hardest to keep current.
- Access: keeping the use of what the bank does not own The second question is about continuity.
- Data on demand: producing accurate information fast The third question is speed.
- Who is covered, and by which rule Resolution planning in the United States runs through two rules that apply to different entities.
- The case against treating the inventory as permanent The strongest objection is that the inventory is a compliance cost with little benefit outside a failure.
- What to build, whatever the next rule says A bank that wants to be ready for the next swing in resolution rules, in either direction, can focus on capabilities rather than documents.
- With hindsight The 2019 rule’s narrower scope soon applied to a firm that failed.
The week
On Monday, October 28, 2019, the Federal Reserve Board and the Federal Deposit Insurance Corporation announced that they had finalized changes to their resolution plan requirements for large banking organizations. Resolution plans, also called living wills, describe how a firm could be resolved in an orderly way under the bankruptcy code if it suffered material financial distress or failed. Section 165(d) of the Dodd-Frank Act requires them, and the agencies’ first joint rule dated from November 2011.
The agencies said the requirements had taken effect in 2012. Since then, they said, the largest firms had improved their resolution strategies and governance, refined their estimates of liquidity and capital needs in resolution, and simplified their legal structures. The final rule was substantially the same as a proposal issued earlier in 2019. It relied on the separate framework of risk-based categories the banking agencies had developed for applying prudential standards, and it implemented changes Congress had made in the Economic Growth, Regulatory Relief, and Consumer Protection Act.
The rule sorted covered firms into groups. The most systemically important firms would keep filing on a two-year cycle, as they already did in practice. Other covered firms would move to a three-year cycle. Both groups would alternate between full resolution plans and shorter targeted plans. Foreign banking organizations with relatively limited U.S. operations would file reduced plans. All plans would be due on a single date, July 1.
Targeted plans were the main change in content. According to the agencies’ announcement, a targeted plan would cover core elements related to capital, liquidity and recapitalization, along with material changes at the firm and any areas the agencies identified. It would not include certain areas that were materially unchanged since the last plan. The announcement named two examples: descriptions of management information systems and of corporate governance systems.
The rule also narrowed who had to file at all. Firms with less than $250 billion in total consolidated assets that did not meet certain risk criteria would no longer be subject to the requirement. In the rule’s preamble, the Board noted that some commenters had argued the requirement should apply to every firm with $100 billion or more in assets. One argued that the FDIC’s own resolution authority does not reach beyond a firm’s insured bank. The Board finalized this part as proposed.
One change from the proposal tightened the rule. Only smaller and less complex firms could ask to change what their full plans contained, and both agencies would have to approve any such request before it took effect. The rule also lengthened the default filing cycle and allowed more focused submissions, which the agencies presented as improvements identified since 2011.
The announcement came with charts setting out the new requirements and listing the firms expected to be in each category.
The agencies stressed one limit. The changes did not affect the FDIC’s separate rule requiring resolution plans from large insured depository institutions, the banks themselves. The final rule was published in the Federal Register on November 1, 2019, and took effect on December 31, 2019.
The October 2019 rule was part of the tailoring package: relief for large regional banks, longer filing cycles and shorter plans. Read from a technology seat, its most telling sentence was a small one. Targeted plans would not need to describe management information systems if those systems had not materially changed. The systems that a resolution depends on became something a firm would describe in full only every other filing.
Living-will rules have kept moving since then. The FDIC tightened its separate rule for large banks in 2024, the agencies issued new guidance the same year, and the FDIC proposed to loosen its rule again in 2026. Through all of it, one demand never left. Each version asked banks to show what systems they run and who owns or licenses them. The push for fast data stayed too, though the 2026 proposal would move it from the filing to a proposed adjustment in deposit insurance pricing. The analysis around that inventory came and went. The inventory itself stayed.
That inventory answers three questions, and a resolution needs all three answered. The first is the map: which systems exist, who owns them, and which legal entities and business lines depend on them. The second is access: whether the bank, or the FDIC acting as receiver, can keep using systems and services the bank does not own. The third is data on demand: whether the bank can produce accurate information quickly enough for regulators and bidders to act. After a short account of what the 2019 rule changed, the sections below take those three questions in turn. Three further sections follow: who is covered by which rule; the strongest case against treating the inventory as permanent; and what a bank should build. Nothing here is legal advice. Firms should confirm their own filing obligations with their supervisors and counsel.
What the 2019 rule changed about technology content
The 2019 rule kept the full-plan content requirements on management information systems largely as they were. A full plan had to include a detailed inventory and description of key systems and applications, including those for risk management, accounting, and financial and regulatory reporting. For each, it had to identify the legal owner or licensor, the system’s use or function, related service level agreements, software and system licenses, and associated intellectual property.
The full plan also had to map those systems to the material entities, critical operations and core business lines that relied on them. It had to describe the key internal reports senior management used and explain how supervisors could access the systems. It also had to analyze whether the systems could collect, maintain and report the data underlying the plan in a timely way. A separate section on interconnections asked firms to map shared personnel, facilities and systems, including technology platforms and service level agreements, to their material entities.
The rule made one wording change in this area. It removed the word deficiencies from the requirement to describe gaps in system capabilities, because the rule now used deficiency as a defined term for serious plan weaknesses. The agencies said they did not intend to change what firms had to report about their systems.
What a targeted plan still asks
A targeted plan was shorter, not empty. It had to cover the core elements of capital, liquidity and recapitalization. It had to describe material changes since the last plan. And it had to address any areas of interest the agencies identified in advance. That last element gave supervisors a way back into technology content. If they had concerns about a firm’s systems, they could name management information systems as an area of interest, and the targeted plan would have to cover them.
For a technology organization, this meant the targeted cycle was lighter only if supervisors were satisfied. The signal to watch was the agencies’ feedback on the previous plan and their notice of areas of interest. A bank that read those documents only in its resolution planning office could be surprised by a systems request with little time to answer it.
Cadence, not content
The change that mattered was cadence. Under the new cycles, firms alternated full and targeted plans. A targeted plan could leave out system descriptions that were materially unchanged. A firm could therefore go a full filing cycle, or longer, between complete descriptions of its systems, with targeted plans in between relying on the earlier description.
That is a defensible regulatory choice. Rewriting a full system inventory every filing is expensive, and much of it does not change. The risk lies in the word materially. Inside a large bank, systems can change often. A core platform is upgraded, a reporting tool moves to a cloud provider, a subsidiary starts using a different payments processor. None of those changes may look material on its own. Together, over several years, they can leave the filed description a long way from reality.
Who stopped filing altogether
The rule’s other change removed some firms entirely. U.S. firms with assets between $100 billion and $250 billion that did not meet the risk criteria for a higher category no longer had to file holding company resolution plans. Their insured banks, if large enough, still filed under the FDIC’s separate rule. For a technology organization, that split meant the holding company level, where shared services and intercompany arrangements may live, no longer needed mapping for resolution purposes.
The map: which systems exist, who owns them, what depends on them
The first question is the most basic and the hardest to keep current. A resolution authority taking over a failing bank needs to know which systems run deposits, loans, payments, general ledger and regulatory reporting. It needs to know which legal entity owns or licenses each one, and which business lines would stop if it stopped. Without that map, nobody can plan the sale of a business line or keep a critical operation running through a weekend.
The living-will rules have always framed the map in legal-entity terms. A technology inventory organized by application, team or platform does not answer that framing. A bank’s configuration management database may list thousands of applications with their technical owners without recording which legal entity holds the license, which entity pays for the hosting, or which entities’ customers the system serves.
Building a resolution-grade inventory
A resolution-grade inventory adds legal and commercial facts to technical ones. For each key system it records the owning or licensing entity and the contract under which it is provided. It also records the business lines and legal entities that use it, the people needed to run it, and where it physically runs. The FDIC’s 2024 rule for large banks spelled out almost exactly this list, adding key personnel and physical location to the earlier requirements.
The work spans three functions that may not share a data model. Technology knows the systems. Procurement and legal know the contracts. Finance knows which entity pays and which entity’s books carry the asset. A bank that keeps these in separate tools rebuilds the join every time a plan is due. A bank that links them once, and keeps the links current through ordinary change processes, produces the map as a by-product.
Keeping the map current
The inventory decays through ordinary change. The practical defense is to attach the resolution attributes to events that already happen. When a new application is approved, the approval captures its owning entity and the entities it serves. When a contract is signed or renewed, procurement records which systems it covers. When a system moves hosting, the change record updates its location. A quarterly reconciliation then catches what slipped through.
One way to test the map is by sampling. Pick a critical operation and trace every system it depends on, the contracts behind them and the people who run them. A bank that can do that within a day has a working map. A bank that needs a project team for a month has a filing.
Cloud accounts and legal entities
Cloud hosting adds a twist to the map. When a system runs in a cloud provider’s data centers, its physical location is a set of provider regions, and the relationship that matters is the account or subscription under which it runs. Which legal entity holds that account, and pays for it, determines who controls the system if the group breaks apart.
Where cloud accounts are organized by team or application, a resolution-ready structure also records the owning legal entity for each account, and keeps workloads that serve different material entities separable. That does not require a separate cloud estate for every entity. It does require tags or account structures that let the bank say, quickly and accurately, which entity’s systems sit where and which contract covers them.
Access: keeping the use of what the bank does not own
The second question is about continuity. In a resolution, the failing bank’s systems need to keep running long enough for the FDIC to operate the bank, sell it, or transfer its deposits. Some of those systems may not be owned by the bank. They may be licensed from vendors, provided by affiliates, or run by service providers under contracts that may allow termination when a customer fails.
The FDIC’s 2024 rule made this explicit. For any key system or application the bank did not own or license directly, a full submission had to describe the obstacles to keeping access in resolution, approaches for keeping it, and the projected cost. The agencies’ 2024 guidance for large holding companies asked for arrangements to support the continuity of shared and outsourced services, including plans to retain key personnel needed to carry out the resolution strategy.
Affiliates and shared services
Some banking groups run technology through a service company or the parent, which then provides it to the bank. That structure is efficient in normal times. In resolution it creates a dependency that crosses legal entities, and possibly crosses the line between the failed bank and a parent going through bankruptcy. If the service entity stops being paid, or is sold separately, the bank can lose its systems.
The answers are contractual and structural. Intercompany service agreements can state that services continue for a set period in resolution, at stated prices. Shared service entities can be set up so they can operate on their own, with enough funding and staff to keep serving whichever part of the group survives. The 2019 rule’s interconnections section asked firms to map exactly these relationships. Firms that dropped out of the holding company requirement in 2019 lost the regulatory prompt to keep that map current, though not the dependency itself.
Third-party contracts
Vendor contracts raise a different issue. Some contain clauses allowing the vendor to terminate if the customer becomes insolvent or enters resolution. A commenter on the 2024 guidance argued that such ipso facto clauses would be stayed by the Federal Deposit Insurance Act in an FDIC receivership. The same commenter said many firms add resolution-resilient terms to vendor contracts as they come up for renewal. Whether a clause is enforceable is a legal question for counsel. The technology question is simpler: does the bank know which contracts cover its key systems, and what each one says about resolution?
A practical program reviews contracts for key systems at renewal, adds terms that commit the vendor to continue service during resolution for a stated period on existing terms, and records the result in the inventory. Contracts that cannot be amended are flagged as residual risk, with a fallback plan or an estimate of the cost of keeping access.
Payments and market infrastructure
Some of the most important things a bank does not own are its connections to payment systems and other market infrastructure. Access to wire transfer, automated clearing house and card networks may run through memberships, sponsor banks, correspondents or processors. If those relationships pause when a bank fails, customers cannot move money, which is the fastest way to destroy the value of a deposit franchise.
The inventory should treat each of these as a key dependency. It should record the legal entity that holds the membership or contract, the systems that connect to it, and what the provider’s rules say about a member in resolution. Where those rules are set by the infrastructure operator rather than negotiated, knowing them in advance matters all the more.
Software as a service
Software-as-a-service applications raise access questions in a sharper form. The bank does not run the software at all, and its data sits in the vendor’s systems. In resolution, the receiver needs both continued use of the application and the ability to get the data out. A subscription that lapses because an invoice goes unpaid during a chaotic weekend can cut off a business line as surely as a server failure.
Contract terms for these services should cover continuity in resolution, data export in a usable format within a stated time, and the vendor’s cooperation with a receiver or acquirer. The inventory should record, for each such service, which business lines rely on it and how the data would be retrieved. The 2026 proposal’s emphasis on core processors for deposit and loan data points to the services where this matters most.
People
Systems need people. The 2024 rule required the inventory to name the key personnel needed to support and operate each key system. In a failing bank, those people may leave first, because their skills are portable and their employer’s future is uncertain. Retention plans for critical technology staff are part of operational continuity. They may involve stay bonuses or other incentives approved by the board in advance, so that a resolution does not begin with the core platform team’s resignations.
Data on demand: producing accurate information fast
The third question is speed. When a bank fails, the FDIC needs to know quickly which deposits are insured, what the loans are worth, and what a buyer would be acquiring. Bidders need data to price a bid. Every hour spent assembling data is an hour in which the franchise loses value.
The 2019 rule asked for an analysis of whether the firm’s systems could collect, maintain and report the data underlying the plan in a timely way. The agencies’ 2024 guidance for large holding companies was more concrete. It said firms should have the capabilities to readily produce data on a legal-entity basis, with controls to ensure its integrity and reliability. It also asked them to analyze the specific types of financial and risk data the resolution strategy required and how often they would need to produce it.
The virtual data room
The FDIC’s 2024 rule turned speed into a specific capability. Large banks had to be able to set up a virtual data room containing enough information for a bidder to make an initial bid on the bank or its franchise components. The rule did not require the data room to be maintained at all times. It required the capability to populate one quickly, and said the FDIC expected the time frame to be brief and measured in days. Banks had to describe how long it would take and what obstacles stood in the way.
A data room is a test of data engineering. It needs loan-level data, deposit data, contracts, financial statements and operational information, extracted from systems of record, checked and organized for outsiders. A bank that keeps its key data well defined, with known owners and lineage, can assemble a data room from repeatable extracts. A bank that relies on analysts building one-off spreadsheets will struggle to meet a time frame measured in days.
Data quality is the constraint
Speed depends on quality. Data that has to be reconciled, corrected and explained before anyone outside the bank can use it will never be produced in days. Three problems slow it down. Definitions differ between systems, loan or deposit records do not reconcile to the general ledger, and records are not tagged with the legal entity that owns them.
The fixes are ordinary data management. Common definitions for the data elements a resolution needs, regular reconciliation of subledgers to the general ledger by legal entity, and lineage showing which system each figure came from. These practices also serve regulatory reporting and risk management. The guidance’s phrase about producing data on a legal-entity basis with controls for integrity and reliability describes the same work.
A useful starting point is the bidder’s side. A bank or investor considering a bid on a failed bank’s franchise wants to know what deposits and loans it would acquire and how they perform. It also wants to know what contracts come with them and what systems would have to be converted. The FDIC’s 2024 rule asked large banks to describe the time and obstacles involved in providing that information. Working backward from those questions shows which extracts a bank needs to have ready.
Deposit data and insurance determinations
Deposit data is the most time-sensitive of all. To pay insured depositors promptly, the FDIC needs accurate account and ownership data. A separate FDIC rule, part 370, adopted in 2016, applies to banks with two million or more deposit accounts. It requires them to configure their systems so they can calculate deposit insurance coverage themselves.
In November 2026 the FDIC published a proposal to remove part 370. Under the proposal, those banks would no longer need systems that calculate coverage. They would instead have to maintain depositor and customer data in a standard format and provide it to the FDIC, which said its own systems had improved enough to do the calculation. The proposal moves the burden from computing an answer to producing clean data. That is the same shift visible across the resolution rules.
Who is covered, and by which rule
Resolution planning in the United States runs through two rules that apply to different entities. The holding company rule, under section 165(d), is joint between the Federal Reserve and the FDIC. After 2019, it applied to the largest firms on a two-year cycle, to other large firms in higher risk categories on a three-year cycle, and to some foreign banks with limited U.S. operations through reduced plans. It no longer applied to U.S. firms with assets between $100 billion and $250 billion that met none of the higher risk criteria.
The FDIC’s rule for insured depository institutions applies to the banks themselves. Under the 2024 version, banks with $100 billion or more in total assets file full resolution submissions, and banks with at least $50 billion but less than $100 billion file informational filings. That rule was unaffected by the 2019 changes. A regional bank that left the holding company requirement in 2019 could still have a full IDI obligation, with systems content of its own.
Thresholds create transitions, and transitions are where capability gaps open. A bank growing past $50 billion, or past $100 billion, picks up new filing obligations on a schedule set by the rule. The Federal Reserve’s SVB review noted how long transition periods delayed several requirements for that firm. A bank approaching a threshold can start building the inventory and data capabilities before the obligation arrives, so the first filing describes something that already works.
Foreign banking organizations
Foreign banks with limited U.S. operations file reduced plans under the holding company rule. Their U.S. branches and subsidiaries may depend on systems run by the parent abroad, under intercompany arrangements governed by foreign law. The map for these firms crosses borders. A U.S. operation that cannot say which of its critical systems are run from outside the country, and under what terms, has a dependency that no U.S. receiver could easily manage.
Firms outside both rules
Banks below the FDIC rule’s $50 billion threshold do not file resolution submissions. When such a bank fails, the FDIC still has to resolve it, and still needs the same three answers. The difference is that the institution has never been asked to prepare them. For smaller institutions, the practical equivalents are good core processor contracts, clean deposit data and an up-to-date list of critical vendors. Those are worth having for ordinary operational resilience whether or not any resolution rule applies.
A smaller institution can apply the same three questions at a fraction of the effort. A one-page list of critical systems with their providers and contracts answers the map question. A check of the core processor and payments contracts for what happens in a failure answers the access question. A tested export of deposit and loan data answers the data question. None of it requires a filing, and all of it shortens the work a receiver would face.
Broker-dealers and nonbank affiliates
Large banking groups may contain broker-dealers and other nonbank entities whose technology is shared with the bank. In a holding company plan, those entities appear as material entities with their own dependencies. A broker-dealer that relies on the bank’s market data or settlement systems needs those mapped too, because its customers’ accounts may need to be transferred under a different legal regime. A single inventory that covers every material entity avoids building separate maps for each rule.
The case against treating the inventory as permanent
The strongest objection is that the inventory is a compliance cost with little benefit outside a failure. A bank that never fails gets little return from it. A detailed map of systems by legal entity, with contracts, licenses and named personnel, is expensive to maintain. If the FDIC itself proposes to drop parts of its requirements, as it did in 2026, a bank might reasonably decide to do only what the current rule asks.
A second objection is that a real resolution may not follow the plan. A failure can unfold faster, and in a different shape, than any filed strategy assumed. A map built for a planned resolution may not match the resolution that actually happens.
A third objection is that legal-entity mapping fits poorly with how technology is run. Modern platforms are shared across business lines and entities by design. Forcing them into entity-by-entity descriptions creates a document that technologists do not use and regulators may not read.
A fourth objection is proportionality. A bank just above an applicability threshold carries much of the cost of a full inventory, with far less complexity than the largest firms. The FDIC’s 2026 proposal would raise and index its threshold. For a bank that falls out of scope, spending on resolution-grade mapping may look hard to justify.
Why the inventory still earns its keep
Each objection has force, and none of them removes the need for the three answers. The 2026 proposal shows this clearly. It would delete the analysis of obstacles to keeping access and the description of data production capabilities. It would also add a mapping of the bank’s information technology architecture and named core processors for deposit and loan data explicitly. The regulator was trimming analysis, not the map.
The same inventory serves purposes beyond resolution. Third-party risk management needs to know which contracts cover which critical systems. Operational resilience planning needs to know which business services depend on which platforms. Cyber incident response needs to know who runs each system and where it lives. A bank that builds the map once, for all of these, pays for it once. The resolution plan becomes one of several reports drawn from it.
The second objection is the more serious. Plans do not dictate how a real failure unfolds. Yet in any version of a failure, the authorities need data fast and need the systems to keep running. Those needs do not depend on which strategy is chosen. The map, access arrangements and data capability are the parts of a plan that remain useful when the strategy itself is set aside.
The proportionality objection is right about scale and wrong about substance. A smaller bank’s inventory is shorter, its contracts fewer and its data simpler. The work shrinks with the bank. What does not shrink is the need, in a failure, to know which systems matter and to get data out of them quickly. A bank that falls out of a rule’s scope can drop the filing and keep a lighter version of the capability.
The third objection, that legal-entity mapping fits poorly with shared platforms, is answered by keeping the entity view as an attribute rather than a structure. Platforms stay shared. The inventory records which entities each platform serves and under which agreements, so the entity view can be produced when needed without reorganizing how technology is run. Technologists keep working by platform, and the resolution view is a report drawn from the same records.
What to build, whatever the next rule says
A bank that wants to be ready for the next swing in resolution rules, in either direction, can focus on capabilities rather than documents.
- Link the application inventory to legal entities, contracts, licenses, hosting locations and key personnel, and keep the links current through existing change and procurement processes.
- Map each critical operation and core business line to the systems it depends on, and test the map by tracing one operation end to end each quarter.
- Review contracts for key systems at renewal for resolution continuity terms, and record residual risk where terms cannot be changed.
- Set intercompany service agreements so that shared technology services continue for a stated period, at stated prices, if part of the group fails.
- Build repeatable extracts for the data a resolution needs, and run a timed exercise that populates a data room from them.
- Agree retention arrangements for the staff who run core systems, approved in advance by the board.
Useful measures follow. Two are shares: key systems with a recorded owning entity and contract, and key vendor contracts with continuity terms. Two are times: how long it takes to trace a critical operation’s dependencies, and how long to populate a test data room. Each can be reported to a board risk committee and compared from year to year.
The order matters. The map comes first, because access reviews and data extracts both depend on knowing which systems are key and who owns them. Access work comes second, starting with the systems that support deposit and payment operations. Data capability comes third, built on the map and tested against the operations that matter most in the first days of a resolution.
Who owns the capability
Resolution readiness falls between functions. The resolution planning office owns the filing, technology owns the systems, procurement and legal own the contracts, and finance owns the entity structure. Without a single owner for the capability, each function maintains its part for its own purposes and nobody maintains the joins. A named executive, reporting for example to the chief risk officer or chief operating officer, should own the inventory as a whole, with each function accountable for its fields.
The board’s role is oversight of the measures, not the detail. A short annual report can show coverage of the inventory, results of the trace and data room exercises, and the contracts still lacking continuity terms. That gives directors a view of readiness that does not depend on the filing calendar.
Evidence an examiner can use
Evidence that the capability works carries more weight than evidence that it is documented. Useful evidence includes the inventory itself with its change history, and records of quarterly traces and what they found. It also includes the results and timing of data room exercises, and a register of key contracts with their continuity terms. A bank that keeps these as ordinary operating records can answer a request in days. A bank that assembles them for each review is showing the same weakness the review is looking for.
With hindsight
The 2019 rule’s narrower scope soon applied to a firm that failed. The Federal Reserve’s April 2023 review of its supervision of Silicon Valley Bank recorded that SVB Financial Group crossed the $100 billion threshold in June 2021 and met the criteria for a Category IV firm. Under the 2019 revisions, it was not subject to a holding company resolution plan requirement. The bank itself became subject to the FDIC’s rule in 2021 and submitted its plan on December 1, 2022, with an as-of date of December 31, 2021. More than 98 percent of the group’s assets were in the bank. SVB was placed in resolution in March 2023.
The review said that without the changes made by the Economic Growth, Regulatory Relief, and Consumer Protection Act and the 2019 rules, the group would have faced tailored resolution planning requirements. It concluded that the changes from that act, the 2019 tailoring rule and related rulemakings combined to create a weaker regulatory framework for a firm like SVB Financial Group.
The FDIC tightened its own rule next. A final rule dated June 20, 2024, was published in July and took effect on October 1, 2024. It required full resolution submissions from banks with $100 billion or more in assets and informational filings from those with $50 billion to $100 billion. It expanded the system inventory to include key personnel and physical locations, and required analysis of access to systems the bank did not own. It also required the capability to populate a virtual data room within a time frame measured in days. In August 2024, the Federal Reserve and the FDIC issued final guidance for Category II and III firms’ holding company plans, covering management information systems and shared and outsourced services.
The direction then reversed. The FDIC published a proposal on June 30, 2026 to raise and index the threshold for its rule, reduce required content, and eliminate both its credibility assessment and its expectations for capabilities testing. The proposal would delete the analysis of access to non-owned systems and the description of data production capabilities. It would add a mapping of information technology architecture and name core processors for deposit and loan data. A companion proposal would adjust deposit insurance assessments for large banks that could populate a data room within a set time or give the FDIC direct access to the data it needs. Comments were due by August 31, 2026. When we checked the Federal Register on November 14, 2026, it had not been finalized.
Across these swings, the systems map is the constant. The 2019 rule let it go stale between full plans, the 2024 rule expanded it, and the 2026 proposal kept it while cutting the analysis around it. The capability to produce data quickly moved from a filing requirement toward a possible pricing incentive, which is a sign that regulators still value it. The FDIC’s November 2026 proposal on deposit insurance determinations followed the same pattern. It would drop the requirement that the largest banks by account count configure systems to calculate insurance coverage, while requiring them to maintain depositor and customer data in a standard format and provide it to the FDIC. Comments on that proposal are due by January 4, 2027.
Spotted an error? Tell us and we will check it and correct the article where needed.