Why the Economics of Operating a Bank Are Being Rewritten
Cost efficiency can improve while structural cost position deteriorates. This article examines how technology, shared infrastructure, operating-model redesign, scale, resilience and ownership architecture are changing the economic frontier against which banks must compete.
An improving cost-to-income ratio does not necessarily mean the bank is moving closer to the future industry cost frontier.
For decades, cost efficiency in banking has been governed through a familiar financial relationship. Management seeks to grow revenue faster than operating expenses, reduce avoidable overhead, automate labour-intensive processes, rationalise distribution, capture scale economies and ultimately improve the cost-to-income ratio. The underlying logic remains sound because an institution that repeatedly consumes more resources to produce the same economic output will eventually weaken its ability to compete, invest and generate acceptable returns.
The difficulty begins when improvement in reported efficiency is mistaken for improvement in structural cost position. A bank can lower its cost-to-income ratio while the economic frontier around it moves faster. It can remove branches while accumulating substantial technology, cyber, data and compliance costs. It can automate thousands of tasks without changing the process architecture that created those tasks, reduce headcount while competitors redesign the unit economics of customer acquisition and servicing, or celebrate productivity gains while the minimum economic cost of operating a competitive bank is being rewritten elsewhere.
Recent European experience illustrates the distinction. Euro area banks improved their aggregate cost-to-income ratio from approximately 66 percent in 2020 to 54 percent in 2025, but ECB analysis concluded that the improvement since 2021 was entirely attributable to increasing revenues rather than a reduction in the underlying cost base. The ratio improved while the structural question around cost remained substantially unresolved.
This is not a criticism of the cost-to-income ratio. It is an argument for understanding what the ratio cannot tell management. Cost-to-income describes the relationship between expenditure and revenue produced by the current economic architecture. It does not determine whether that architecture will remain competitive when the industry cost frontier changes.
The future banking cost question is therefore not simply how much expense an institution can remove. It is whether the bank understands which costs will remain economically necessary, which costs should disappear, which new costs will become unavoidable, where shared infrastructure will reduce the need for proprietary investment, where scale will become more important, where scale will become less important and what operating architecture will ultimately be required to produce acceptable returns.
Bancly describes this condition as Structural Cost Position: the degree to which the bank’s operating architecture can produce the required customer, risk, control and financial outcomes at a cost level that remains competitively viable as the economics of the industry change. A strong Structural Cost Position does not necessarily mean having the lowest current operating expense. It means possessing an architecture capable of moving towards the future cost frontier without destroying service quality, resilience, strategic optionality or the ability to generate revenue. That is where cost management becomes Future Banking Economics.
Cost Efficiency And Structural Cost Position Are Not The Same Thing
The distinction can be understood through two banks reporting identical cost-to-income ratios. Suppose both institutions operate at 45 percent. The first bank has reached that level after several years of process redesign, technology modernisation and simplification. Customer activity is increasingly digital, operational exceptions are declining, product complexity is controlled, legacy systems are being retired and the institution can absorb additional customer volume without expanding costs proportionately.
The second bank also reports 45 percent, but the economics underneath the ratio are different. Revenue has benefited from favourable interest rates, the organisation continues carrying duplicated processes and systems, branches remain economically underutilised, manual exceptions remain embedded in operations and technology expenditure is rising merely to maintain the existing architecture. The reported efficiency is the same, but the structural cost position is not.
This is why cost management cannot be separated from the operating model that produces the cost. A bank with a structurally scalable architecture can tolerate periods of elevated expenditure because those costs may be building a lower future unit-cost position. Another institution can report comparatively low expenditure while continuing to preserve complexity that will eventually become expensive to operate, control and change. The management challenge is therefore to distinguish cost level from cost architecture.
The first tells leadership how much the institution is spending today. The second explains why that expenditure exists, whether it changes with scale, whether it creates future economic capacity and how difficult it will be to remove when the competitive environment changes. This distinction becomes particularly important when revenue conditions are supportive. A strong margin environment can improve the cost-to-income ratio without changing operating economics materially, while a weaker revenue environment can make an otherwise improving operating model appear temporarily less efficient.
Recent EBA data show why the ratio requires careful interpretation. EU and EEA banking costs rose in absolute terms by around 4 percent over the year covered by its December 2025 assessment, while cost-to-income increased from 51.8 percent to 52.4 percent as income dynamics changed. A bank therefore needs to know not merely whether its ratio is improving, but what is causing the improvement.
The Bank Cost Base Is Being Recomposed, Not Simply Reduced
Banking has spent decades moving expenditure away from physical distribution and manual processing towards technology. The simplistic interpretation is that digitalisation should steadily make banks cheaper because software replaces branches, paper and labour. The evidence suggests a more complicated transition.
Some traditional costs are undoubtedly declining. Branch networks have been reduced in many markets, routine activities have moved to digital channels and automation has lowered labour requirements across numerous processes. At the same time, a different category of fixed and semi-fixed expenditure has expanded around technology platforms, cyber security, data management, cloud infrastructure, fraud prevention, regulatory controls, third-party governance and continuous software development. The result is not necessarily a smaller cost base. It can be a different cost base.
The European Banking Authority’s June 2026 assessment illustrates the direction of travel. ICT-related expenditure represented 32.1 percent of other administrative expenses at EU and EEA banks in 2025, increasing from 31.2 percent a year earlier. Banks expect staff and administrative expenses to decline over the coming years, partly because of automation, yet the EBA specifically questioned how easily administrative costs can fall when technology has become such a large component of the operating base.
This matters strategically because the economic behaviour of technology cost differs from the economic behaviour of many traditional costs. A branch network is geographically distributed and directly visible. Technology expenditure can be more concentrated, less visible to customers and increasingly foundational to every business line. Once the institution depends on a platform, data environment or third-party provider, the associated expenditure may become difficult to remove without changing the operating architecture itself.
Cyber and operational resilience add another dimension. The EBA’s 2025 risk assessment reported rising dependencies on third-party providers, particularly cloud and payment providers, while regulation such as the Digital Operational Resilience Act has strengthened expectations around incident management and resilience. The bank can therefore automate operational activity while simultaneously increasing the cost of governing the infrastructure on which automation depends.
The future cost base can therefore be interpreted through several economic movements. Some costs are eliminated because activities no longer need to exist, some are compressed because the same activity can be delivered with fewer resources, some are mutualised because common infrastructure can replace institution-specific investment, some are transferred from labour or premises into technology and third-party expenditure, and some are created because digital banking requires new capabilities in cyber security, data, resilience, model governance and fraud management. The economic task is not to assume that digitalisation produces lower expense automatically, but to understand how each category changes the relationship between cost, scale and economic output.
The Digital Bank Paradox
Digital-only banking provides an important warning against simplistic assumptions about cost. A bank without branches would appear, intuitively, to possess a structural cost advantage over an institution maintaining a large physical network. In some areas it clearly does because the absence of a traditional distribution estate can reduce premises expenditure, simplify servicing and permit national or cross-border customer acquisition without comparable physical infrastructure.
Digital banking also reveals the importance of scale. ECB research published in 2025 found that euro area digital banks were, on average, less profitable than traditional institutions. Their business models benefited from the absence of large branch networks but faced comparatively high fixed technology and marketing expenditure, while their smaller scale prevented those costs from being spread across a sufficiently large revenue base. Administrative expenditure excluding staff was approximately twice as high relative to assets as at traditional banks, with IT infrastructure and marketing accounting for much of the difference.
This is a critical economic insight because digital does not eliminate fixed cost. It changes the fixed-cost structure. A branch-led bank needs sufficient economics to support physical distribution, while a digital institution needs sufficient economics to support technology infrastructure, continuous development, customer acquisition, cyber security, regulatory compliance and brand-building. The competitive advantage appears only when the operating architecture reaches sufficient scale.
This creates what might be called the Digital Scale Paradox. Technology can reduce the marginal cost of serving the next customer while simultaneously increasing the importance of achieving enough scale to absorb the fixed cost of the platform. The result is a changing relationship between scale and efficiency. In some parts of banking, technology lowers the minimum economically viable size because institutions can reach customers without large physical networks. In other parts, the sophistication required in technology, cyber, data and regulation can increase the fixed-cost burden and make scale more important.
The question is therefore not whether digital banks are cheaper than traditional banks. It is which activities become cheaper at scale, which new fixed costs replace traditional ones and how large the institution needs to become before those economics become attractive. That question applies equally to incumbent banks, which cannot assume that reducing branches automatically creates a digital cost advantage. If physical costs fall while duplicated systems, legacy architecture and digital investment remain, the bank can temporarily carry both cost structures at once, making the transition period itself expensive.
Shared Infrastructure Changes What The Bank Needs To Own
Another structural force is moving in the opposite direction. While some technology requirements are becoming more complex and expensive, other forms of infrastructure are becoming increasingly shared. Fast-payment systems, digital identity, data-sharing frameworks and other forms of digital public infrastructure create common foundations on which multiple institutions can operate. The World Bank describes digital public infrastructure as shared foundations including digital identity, payments and secure data exchange, and its global programme now supports work across more than 80 countries.
For banks, the economic significance lies in mutualisation. When every institution must independently construct a capability, advantage can come from owning better infrastructure than competitors. When the capability becomes common national or industry infrastructure, part of that investment ceases to be differentiating. If payments become faster and interoperable for everybody, competitive advantage shifts towards what the bank builds on top of those rails. If digital identity reduces onboarding friction across the system, advantage moves from possessing a proprietary verification process towards decision quality, customer experience and the economics of the relationship being opened. If data can be transferred through standardised open-finance infrastructure, the value of merely possessing customer information declines relative to the ability to interpret it better.
The cost implication is important because banks can eventually avoid replicating parts of the infrastructure that the system provides collectively. This does not imply that shared infrastructure is free. Banks still incur integration, compliance, resilience and operating costs, while new infrastructure can require substantial initial investment. The ECB’s work on the proposed digital euro, for example, explicitly identifies opportunities for cost mutualisation and infrastructure synergies while also recognising significant implementation costs for financial institutions.
The strategic principle is therefore not that shared infrastructure always lowers banking cost. It is that shared infrastructure changes the boundary of what banks need to own themselves. Banks that continue treating increasingly common infrastructure as proprietary sources of differentiation may overinvest in capabilities customers and regulators increasingly experience as utilities. Banks that externalise too much can become dangerously dependent on shared providers and lose the capabilities required to control operational resilience.
The correct architecture requires deliberate ownership choices. Some capabilities should remain proprietary because they create meaningful customer, risk or economic advantage. Some should be shared because ownership creates little differentiation. Some should be outsourced because specialist providers possess superior economics. Some should remain internally controlled even when external infrastructure is used because failure would create unacceptable financial or regulatory consequences. The future cost curve will therefore be shaped partly by ownership architecture.
Automation Changes The Economics Only When The Operating Model Changes
Automation is already central to bank cost programmes, and the next phase will undoubtedly involve far more intelligent forms of automation. The strategic error would be to interpret the opportunity principally through the number of tasks or roles that technology can replace. The larger economic question concerns the architecture of work.
Consider a lending process containing twenty operational steps, six manual approvals, repeated data entry, multiple reconciliations and several control points created because systems do not communicate effectively. Automating individual tasks can reduce effort, but the process still contains twenty steps. The bank may therefore improve productivity without fundamentally changing the cost architecture.
Structural cost improvement occurs when management asks whether the process should still exist in its current form, which decisions need human intervention, which information can be verified once rather than repeatedly, which controls can be embedded within the workflow, which exceptions genuinely require escalation and whether customer and regulatory outcomes can be delivered through a substantially simpler architecture. The distinction is between automating work and redesigning the economics of work.
The same principle applies across customer service, compliance, operations, underwriting, collections, finance and management activity. A bank with hundreds of automated processes can remain structurally expensive if the underlying institution continues generating unnecessary activity, exceptions and complexity. This is why the next cost curve cannot be reduced to technology adoption. Technology is one of the forces moving the frontier, but the bank captures the economic advantage only when operating models, decision rights, organisational structures and control systems change with it.
Recent European bank planning reflects the expectation that automation can reduce personnel costs over time, but regulators have also noted that technology expenditure itself remains substantial. The balance between these forces will depend on whether institutions merely substitute one form of cost for another or actually reduce the number of economic resources required to produce the same banking outcome. The bank should therefore ask not only how many employees it has per branch, customer or asset base, but how much customer value, revenue, risk capacity and decision throughput the operating architecture can produce for each unit of economic cost.
Scale Is Being Rewritten, Not Eliminated
Digitalisation is sometimes described as a force that democratises banking by reducing the advantages associated with large physical networks. There is truth in that proposition, but the economics are more nuanced because scale still matters wherever banking contains substantial fixed costs. Regulation, cyber security, technology platforms, data, financial crime controls, model governance, product infrastructure and operational resilience all require investment before the next customer is served.
The relevant question is what kind of scale matters. Traditional banking scale was heavily connected to physical reach, customer numbers, branches, deposits and balance-sheet size. The next model increasingly includes technology scale, data scale, transaction scale, control scale and ecosystem scale. These forms of scale do not behave identically.
Technology platforms can make the incremental cost of serving additional customers very small once infrastructure has been built, while cyber and regulatory requirements can impose large fixed costs that are easier for larger institutions to absorb. Shared infrastructure can simultaneously reduce the need for proprietary investment, allowing smaller institutions to access capabilities that previously required greater scale.
The industry can therefore experience two forces at once: some barriers to scale decline while other minimum efficient-scale requirements increase. This helps explain why digital competitors can enter banking without traditional distribution infrastructure while still struggling to generate acceptable returns until they reach sufficient customer and revenue scale. It also explains why traditional institutions with very large balance sheets can remain inefficient if complexity prevents scale from translating into lower unit cost.
Scale is valuable only when the operating architecture converts it into economics. A large bank running thousands of duplicated products, systems and processes can possess enormous accounting scale without achieving comparable economic scale. The relevant measure is therefore scalable output rather than institutional size.
The Quality Of Cost Matters As Much As The Quantity
The next cost curve also requires management to distinguish different categories of expenditure. Reducing every cost by the same percentage may improve the next reporting period while damaging the institution’s future economics. A more useful architecture separates cost according to what the expenditure actually does for the bank.
Run Cost is the expenditure genuinely required to serve customers, operate infrastructure, process transactions and maintain the current institution. Build Cost is investment intended to create future revenue capacity, lower structural unit cost, stronger risk capability or strategic advantage. Control Cost is expenditure required to keep the institution safe, compliant, resilient and trustworthy. Legacy Cost is expenditure created primarily by complexity inherited from the past, including duplicated systems, unnecessary process steps, obsolete infrastructure, organisational fragmentation and work required solely because the architecture has not been simplified.
These categories behave very differently. Cutting Run Cost without redesign can weaken service, cutting Build Cost can improve current efficiency while reducing future competitiveness, and cutting Control Cost indiscriminately can create unacceptable risk. Reducing Legacy Cost should generally create economic value, but identifying it accurately can be difficult because legacy expenditure is often embedded inside ordinary operations.
The central management problem is therefore not simply deciding how much the bank should spend. It is determining what the expenditure is purchasing. A bank whose expenses increase because it is temporarily replacing legacy architecture can be improving its Structural Cost Position even while the reported cost-to-income ratio deteriorates. A bank whose expenses remain flat because it has deferred necessary investment can appear efficient while its future cost position weakens.
Cost quality therefore becomes a strategic variable, particularly for boards because short-term financial discipline and long-term operating economics can appear to conflict even when the institution is making the correct choice. The answer is not to relax cost discipline. It is to make cost discipline more economically precise.
Structural Cost Position Requires A Different Scorecard
Traditional operating metrics remain important. Cost-to-income, operating expense growth, staff cost, revenue per employee, branch productivity and cost per transaction all provide valuable information and should be retained. They should also be supplemented because none of them individually establishes whether the operating architecture is becoming more competitive.
A Structural Cost Position scorecard would examine at least six dimensions. Current Efficiency measures the cost required to support today’s revenue and balance sheet. Unit Cost Scalability examines whether customer and transaction growth can occur without proportional cost growth. Legacy Intensity identifies how much expenditure exists because of duplicated systems, processes, products and organisational complexity. Technology Economics examines whether technology investment is producing lower unit cost, greater capacity or simply maintaining the current architecture. Control Economics measures whether cyber, compliance, fraud and resilience capabilities are becoming more effective as they become more expensive. Ownership Architecture identifies which capabilities the bank should own, share, procure or outsource as the infrastructure surrounding banking evolves.
The scorecard changes the management question from whether costs are declining to whether the institution is moving towards a stronger economic architecture. A bank can therefore be considered structurally improving even when absolute expenditure rises temporarily, provided that the investment reduces future unit cost, removes legacy complexity or expands economically productive capacity. Conversely, a declining expense base can represent structural deterioration if the reduction results from underinvestment, deferred change or indiscriminate cuts that preserve an increasingly uncompetitive architecture.
The Cost-To-Income Ratio Needs An Economic Overlay
This brings the analysis back to the metric bank leadership already knows best. There is nothing wrong with the cost-to-income ratio. The danger lies in asking it to answer a question it was never designed to answer. A cost-to-income ratio of 40 percent tells management that the institution currently spends forty units of operating expense for every hundred units of operating income under the prevailing accounting definitions.
It does not reveal whether the bank’s process architecture is becoming simpler, whether unit cost is falling, whether productive transformation investment is being confused with legacy expenditure, whether revenue has temporarily improved because of interest rates, whether competitors can serve the same customer at a materially lower economic cost, how much expense is necessary to support cyber and resilience requirements that did not exist at the same scale a decade earlier, or whether the current cost structure can support the institution the bank intends to become.
This is why the metric requires an economic overlay. Management should continue asking what the cost-to-income ratio is, while also understanding what caused it, whether the improvement is structural, what costs are migrating, where the future frontier may settle and which parts of the bank’s architecture prevent it from getting there.
The Questions At The Top Must Change
The changing cost curve requires a different conversation at CEO and board level because several traditional questions remain necessary but no longer go far enough. Leadership should continue asking whether the bank is meeting its cost-to-income target, but it should also understand how much of the improvement comes from operating change and how much comes from revenue conditions.
Management should continue governing headcount, but it should determine whether reductions represent genuine removal of work or simply redistribution of activity into technology, outsourcing and external providers. Branch rationalisation should continue where economics justify it, but leadership should understand whether branch savings are actually producing lower structural cost or are being absorbed by duplicated digital and physical architectures.
Technology expenditure should continue to be governed rigorously, but the relevant test should extend beyond project delivery. Management needs evidence that investment changes unit economics, decision capacity, service cost, control quality or revenue productivity. Automation should remain important, but leadership should distinguish between automating existing complexity and redesigning the operating model so that the complexity no longer needs to exist.
Outsourcing and cloud infrastructure can create attractive economics, but the bank should understand what dependency, concentration, resilience and control costs accompany those savings. Shared infrastructure should be embraced where it improves system economics, while management should continually reassess which capabilities still require proprietary ownership and which have become economically inefficient to replicate independently.
Most importantly, the bank should understand its distance from the emerging cost frontier. The central question is not whether operating expenses can fall by another 5 percent. It is whether the institution knows what level of operating economics will become competitively necessary over the next decade and whether its current architecture is capable of reaching that position without damaging the economics it needs to preserve.
The Next Banking Cost Curve
The banking cost base is not simply shrinking. It is being rewritten. Physical distribution is becoming less central in many markets, but technology infrastructure is becoming more economically important. Routine processing is being automated, while cyber security, data, fraud prevention and operational resilience require growing investment. Shared infrastructure is reducing the need for some forms of proprietary capability, while third-party dependencies create new governance obligations. Digital models can lower marginal servicing costs while making platform scale more important, and automation can reduce effort without improving Structural Cost Position unless the operating model itself changes.
These forces operate simultaneously. The future bank may therefore spend less on some activities and considerably more on others. It may employ fewer people in transaction processing while employing more specialists in risk, technology, data and control. It may operate fewer branches while maintaining a larger digital infrastructure. It may own fewer pieces of basic financial infrastructure while requiring greater capability to integrate, govern and orchestrate the systems it does not own.
The relevant strategic question is not whether the future bank will be cheaper in absolute terms. It is whether the bank will be able to produce more economically valuable output for each unit of cost. A future institution processing twice the activity, serving more customers, making decisions faster, controlling risk more effectively and generating materially more revenue could carry a larger absolute expense base while possessing a substantially stronger Structural Cost Position. Conversely, an institution can reduce expenditure while becoming structurally more expensive if complexity, legacy architecture and low productivity prevent cost from scaling with the business.
This is why cost strategy needs to move beyond reduction. The governing objective should be to build an operating architecture in which economically valuable activity scales faster than the cost required to produce it, while the institution retains enough resilience, control and strategic flexibility to remain credible as a regulated bank.
The metric underneath that architecture remains financial. Lower unit cost can improve operating leverage, better scalability can strengthen CIR, a more productive workforce can improve revenue generation, simpler architecture can release investment capacity, better control economics can reduce losses and operational risk, and more deliberate infrastructure ownership can reduce unnecessary operating commitments. Collectively, those changes reach ROE and enterprise value.
The future cost curve is therefore not an operational concern sitting beneath strategy. It is part of the economic architecture of the bank. The central question for leadership becomes whether the operating model that produced today’s cost structure will still be capable of producing competitive returns when technology, shared infrastructure, regulation, customer behaviour and new business models have moved the industry frontier. If the frontier is moving, the bank needs to know how far; if the institution is investing heavily, it needs to know what future economic position that investment is purchasing; and if costs are being reduced, leadership needs to know whether the institution is removing genuine economic waste or simply borrowing efficiency from the future. These are not questions about cost cutting. They are questions about the economics of operating the bank the institution is becoming.
