IT cost transparency in financial services: why resilience rules change the model

In brief

Banks, payment firms and insurers must map technology and suppliers to critical services. Why that changes IT cost transparency, and how a TBM model helps.

Steve Elliott

Steve Elliott · · 6 min read


On this page
  1. What do regulators expect firms to map?
  2. Why is the cost question harder in a bank?
  3. Where do the cost model and the resilience map overlap?
  4. How does one model support both cost and resilience decisions?
  5. What should a TBM office build first?
  6. Sources

In financial services, IT cost transparency answers to regulators as well as to the CFO. UK and EU rules, and US banking guidance, expect firms to know which people, technology and suppliers support each critical service, and the EU's register of ICT contracts asks for each contract's annual cost.

That makes the map behind a TBM cost model and the map behind a resilience self-assessment largely the same map. In our view, firms that build it once, through the application inventory, spend less answering both questions and make better trade-offs between cost and resilience.

What do regulators expect firms to map?

Four sets of rules and guidance ask for much the same thing: a link from the services that matter to the people, technology and third parties behind them.

Who it coversWhat must be mapped or trackedLink to costKey dates
UK: PRA SS1/21 and FCA PS21/3Banks, building societies, PRA-designated investment firms, insurers; the FCA's scope also includes payment and e-money institutions, among othersPeople, processes, technology, facilities and information behind each important business service, including third partiesThe PRA says impact tolerances should guide investment prioritiesIn force March 31, 2022; firms had to be able to stay within impact tolerances by March 31, 2025
EU: DORAMore than 20 types of financial entityA register of all contractual arrangements with ICT third-party service providersThe register records the annual expense of each contract and each providerApplies from January 17, 2025; registers reported each year
US: interagency third-party guidanceBanking organizations supervised by the Fed, FDIC and OCCA complete inventory of third-party relationships, with critical activities identifiedPlanning should weigh direct and indirect costs; the termination stage covers termination costs and feesIssued June 6, 2023
Basel Committee principlesBanks (international principles, not rules)Interconnections and interdependencies behind critical operations, including third parties and intragroup arrangementsNone statedPublished March 31, 2021

A few details matter for IT finance.

  • UK. The PRA's supervisory statement requires firms to identify and document the resources needed to deliver each important business service and calls this process mapping. An impact tolerance is the maximum tolerable disruption to a service, measured at least in time. The PRA describes impact tolerances as "a standard which boards and senior management should use for prioritising investment."
  • EU. The European Supervisory Authorities (ESAs) noted that DORA has no transitional period. Implementing Regulation (EU) 2024/2956 sets the register templates. They link each contract to a function the firm defines, record whether that function is critical or important, and include a field for "Annual expense or estimated cost of the contractual arrangement for the past year." Other fields cover total annual expense per provider and, for services supporting critical or important functions, whether the provider can be substituted and whether an exit plan exists. National supervisors were due to pass the first registers to the ESAs by April 30, 2025, and from 2026 must do so by March 31 each year.
  • US. The guidance says planning should include evaluating "estimated direct contractual costs and indirect costs." It also says using third parties doesn't diminish a bank's own responsibilities. The Fed's cover letter says the guidance imposes no new requirements, and practices are meant to scale with a bank's size, complexity and risk profile.
  • Basel. Principle 4 asks banks to map the people, technology, processes, information and facilities behind critical operations. Principle 5 covers dependencies on third parties and intragroup entities.

This is a summary of public texts, not legal or compliance advice. Confirm how each regime applies to your firm with your legal and compliance advisers.

Why is the cost question harder in a bank?

Three features of large financial firms make the cost model harder to build and more useful once built.

The money is large. JPMorgan Chase's chief operating officer put its 2026 technology budget at about $19.8 billion. CEO Brian Moynihan's letter in Bank of America's 2025 annual report says it spends nearly $13 billion a year on technology, of which more than $4 billion went to new initiatives in 2025. At that scale, even a small misallocation is a large number.

Legacy cores sit under everything. We think this is the hardest part. Core banking, policy administration and card processing systems can run for decades and support many business lines at once. In our view, their cost often sits in mainframe and data center pools and is spread by a flat key rather than by consumption. The same systems usually appear in the resilience map under many important business services. A model that can't trace them to services can't say what retiring one would save or what it would put at risk.

Shared services cross legal entities. Large groups run technology through central functions and intragroup providers that serve retail, commercial, markets and insurance arms. Basel's Principle 5 names intragroup dependencies, and DORA registers are kept at entity, sub-consolidated and consolidated levels. So shared services cost has to be allocated by business line and by legal entity, and the two views have to reconcile. We think this is where many financial services cost models are weakest.

Where do the cost model and the resilience map overlap?

In our view, the TBM Taxonomy already has most of the objects a resilience map needs. Version 5.0 renamed the Business layer the Consumer layer and added value streams, products, portfolios and digital services. It also says "Applications are no longer standalone solutions"; they now support products and services. That puts the application between the technology and the service, which is where resilience mapping puts it too.

TBM model objectResilience question it helps answer
Cost pools and vendor contractsWhich ICT third parties do we depend on, and what does each cost a year?
IT towersWhich technology resources (compute, network, data, security) sit behind a service?
Application inventory, linked to solutionsWhich applications deliver each important business service or critical function?
Consumer layer (products, digital services, business units)Which customers and business lines are affected if a service fails?

The join is the application inventory, kept current by enterprise architecture. Each application record should point down to the towers and contracts it consumes and up to the services it supports. In our view, a CMDB alone rarely gives you a reliable upward link, because in many firms the business-service layer is thin compared with the infrastructure discovered beneath it.

The objects don't line up exactly. A DORA "function" is defined by the firm, and an important business service is defined from the customer's or market's point of view. Neither has to match a TBM product or service one for one. Keep a crosswalk rather than forcing one list to fit the other.

How does one model support both cost and resilience decisions?

The payoff is that cost decisions and resilience consequences show up in the same view. Take an illustrative payments firm whose card payments service runs at $48 million a year:

ComponentAnnual cost (illustrative)Resilience note
Card authorization application on the legacy core$21 millionIn-house; supports four other services
Payments gateway (cloud provider)$9 millionThird party; supports a critical function
Fraud scoring (SaaS provider)$6 millionThird party; supports a critical function
Shared network, security and data, allocated$12 millionShared with every service
Total$48 million

Suppose procurement proposes moving fraud scoring to the gateway provider to save $2.5 million a year. A cost-only model approves it. A linked model also shows that two critical functions would then depend on one provider, and it prompts the questions the register already asks: is the provider substitutable, and is there an exit plan? The saving may still be right. The point is that the trade-off is visible before the contract is signed.

The same logic works in reverse. The PRA expects impact tolerances to steer investment, so remediation spend from scenario testing belongs in the budget alongside run and change spend, tagged to the service it protects. A legacy core retirement case can show both the run cost it removes and the services whose resilience it changes.

What should a TBM office build first?

  1. Adopt the firm's service list. Use the important business services and critical functions the risk team has already defined, and crosswalk them to Consumer layer objects. Don't create a third list.
  2. Make the application inventory the join. Give each application an owner, the services it supports, the towers it consumes and the contracts it relies on.
  3. Reconcile vendor cost to the register. The annual expense per contract in the DORA register, or any cost data your US third-party inventory holds, should match what the cost model holds for the same supplier. Differences are worth explaining.
  4. Allocate shared cost by legal entity as well as business line. Use consumption drivers for the large shared pools, especially the legacy core, and test that both views add up to the same total. See our cost allocation entry for methods.
  5. Put resilience spend in the plan. Tag remediation and testing work to the services they protect, so the budget shows what it costs to stay within impact tolerances.

If the program stalls, the usual causes apply; we cover them in why IT cost transparency projects stall.

We think financial services firms have a reason other industries lack to connect cost to services: supervisors already require most of the map. An operating model that holds cost, applications, suppliers and services in one place lets the same facts serve the CFO, the chief risk officer and the regulator.

Sources

Frequently asked questions

Does DORA require firms to report IT costs?

Not IT costs in general. DORA requires a register of contractual arrangements with ICT third-party service providers. Its templates, set by Implementing Regulation (EU) 2024/2956, include the annual expense or estimated cost of each contract and the total annual expense per provider. Firms keep the register and national supervisors pass it to the European Supervisory Authorities each year.

What is an important business service under UK rules?

It is a service whose disruption could cause intolerable harm to consumers or markets, or threaten a firm's safety and soundness, financial stability or, for insurers, policyholder protection. Firms set an impact tolerance for each one, map the people, processes, technology, facilities and information behind it, and test whether they can stay within tolerance in severe but plausible scenarios.

Do US banks have to keep an inventory of third parties?

In practice, yes, as a supervisory expectation rather than a new rule. The June 2023 interagency guidance from the Fed, FDIC and OCC says a complete inventory of third-party relationships helps a bank see whether risks have changed, and the Fed's cover letter says the guidance imposes no new requirements. Practices scale with a bank's size, complexity and risk profile.

How does a TBM model help with operational resilience?

A TBM model links cost pools and vendor contracts to towers, applications and the services consumers use. Resilience mapping needs the same links from services to technology and third parties. In our view, building them once through an application inventory kept by enterprise architecture serves both, and puts cost and resilience consequences in the same view.

Free playbooks

In-depth guides for product and operations leaders.

Browse the playbooks
Steve Elliott

Steve Elliott

Published November 22, 2026

More on TBM & IT Financial Management

Cost pools, towers, showback, unit cost.