Creating the organisational capability, specialist ownership and cross-functional discipline that allow Expected Credit Loss to operate as a sustained institutional competence rather than a periodic reporting exercise
An Expected Credit Loss framework can be strong in methodology, data, models, controls and technology, yet still remain fragile if the organisation around it is weak. Many institutions eventually discover that their true constraint is no longer the formula, the staging threshold or the scenario file. It is capability. Who really understands the framework end to end Who owns the policy Who challenges the models Who interprets portfolio deterioration Who manages overlays with discipline Who bridges risk, finance, credit, collections, data and technology Who preserves continuity when individuals move roles Who ensures that ECL evolves as the portfolio and environment evolve
Building an ECL Centre of Excellence explains how institutions can create a durable organisational capability around expected credit loss by bringing together policy stewardship, methodology coordination, movement analysis, overlay governance, issue remediation, cross-functional collaboration and knowledge continuity. A strong Centre of Excellence turns ECL from a fragmented periodic exercise into an institutional competence that can be sustained, improved and explained over time.
An Expected Credit Loss framework can be strong in methodology, data, models, controls and technology, yet still remain fragile if the organisation around it is weak. Many institutions eventually discover that their true constraint is no longer the formula, the staging threshold or the scenario file. It is capability. Who really understands the framework end to end Who owns the policy Who challenges the models Who interprets portfolio deterioration Who manages overlays with discipline Who bridges risk, finance, credit, collections, data and technology Who preserves continuity when individuals move roles Who ensures that ECL evolves as the portfolio and environment evolve
This is why building an ECL Centre of Excellence deserves a pillar article of its own.
An ECL Centre of Excellence is not necessarily a large standalone department. Nor is it simply a project office, a modelling team or a finance committee. It is a structured organisational capability that brings together ownership, knowledge, governance, technical depth and operating discipline around the Expected Credit Loss framework. It ensures that ECL is not carried in fragments across the institution, with one team knowing the data, another the model, another the accounting, another the scenarios and no one truly seeing the whole. Instead, it creates a centre of coordinated competence from which the framework can be maintained, challenged, improved and explained over time.
This matters because ECL is a cross-functional framework by nature. It sits between finance and risk. It draws on credit policy, data architecture, modelling, workout practice, macroeconomic judgment, technology operations, controls and external reporting. If these capabilities remain disconnected, the framework may still function, but often at high coordination cost and with weak continuity. A small number of individuals become indispensable. Knowledge remains tacit. Governance becomes person-dependent. Improvements happen reactively. The institution can run ECL, but it cannot yet say that ECL capability is institutionalised.
A mature institution takes the next step. It builds a Centre of Excellence, whether formally named or not, that gives ECL a durable home.
This article explores what that means in practice: why institutions need such a capability, what functions it should perform, how roles and ownership should be structured, how cross-functional participation should work, what skills matter most, how the Centre should engage with governance forums, how it should support continuous improvement, and what common organisational weaknesses prevent ECL from becoming a true institutional strength.
Expected Credit Loss frameworks often begin as projects. A task force is formed. Policies are drafted. Consultants or internal specialists help design the model. The framework goes live. But after implementation, institutions sometimes fail to define where ECL truly lives.
This creates a familiar problem. Finance may own the booked number. Risk may own the model logic. Data teams may own inputs. Technology may own platforms. Credit teams may own borrower insight. Collections may own recovery evidence. Validation may own challenge. Yet no single function owns the framework as a living whole.
This is where an ECL Centre of Excellence becomes essential. It gives the framework an organisational home. Not in the sense that all work must be centralised there, but in the sense that someone is responsible for coherence across the moving parts.
Without that home, ECL often remains a shared responsibility that is not fully anybody’s responsibility.
The phrase “Centre of Excellence” can sometimes sound abstract or fashionable. It should not be.
In the ECL context, it means something concrete: a clearly defined organisational capability that concentrates expertise, coordinates responsibilities, standardises methods and supports governance across the full impairment lifecycle.
This may take different forms depending on institution size and complexity. In a large bank, it may be a formal dedicated unit with methodology, reporting and governance specialists. In a mid-sized institution, it may be a smaller cross-functional hub anchored in finance or risk with clearly defined roles. In a corporate context, it may be a lean impairment excellence team supported by finance, treasury and business controllers.
The exact shape can vary. The essential feature is the same: ECL capability should be intentionally structured, not left to emerge informally from whichever teams happen to touch the process.
One of the main reasons a Centre of Excellence becomes necessary is that fragmented ownership creates cumulative weakness.
When the data team understands the feeds but not the staging logic,when finance understands the journal but not the scenario transmission,when the risk model team understands PD but not disclosure narrative,when collections understands recoveries but not LGD architecture,and when macroeconomic judgment sits in a separate forum with little connection to model design,the institution begins to rely on coordination heroics rather than designed capability.
This may still work for a while. But over time it creates:
slow issue resolution,confusion over decision rights,knowledge silos,repeated misunderstandings,weak change management,and high dependency on a few bridge individuals.
A Centre of Excellence exists partly to solve this fragmentation by creating a place where the pieces are connected with intention.
A strong Centre of Excellence does not “own everything” in a narrow sense. Rather, it orchestrates capability across multiple disciplines.
Key participating areas often include:
finance,credit risk,modelling,data and technology,collections or workout,treasury or economics,reporting and controls,and internal validation or independent review.
The Centre of Excellence should therefore be designed for collaboration. It is not a silo that replaces these functions. It is a focal point that ensures they work together with consistent language, clear interfaces and shared priorities.
This is one of the deepest reasons the Centre matters. ECL cannot be sustained well as a purely local capability in one function, because the framework itself is inherently cross-functional.
A mature ECL Centre of Excellence often takes responsibility for several core activities.
These may include:
maintaining the impairment policy framework,coordinating model and methodology consistency,owning movement analysis and interpretation,supporting overlay and emerging-risk discipline,coordinating scenario and judgment integration,ensuring portfolio coverage remains current,driving issue remediation and continuous improvement,supporting reporting and disclosure consistency,facilitating validation responses and model change governance,and preserving institutional memory of decisions, assumptions and evolution.
The key point is that the Centre is not just a calculation team. It is the keeper of coherence.
Policies can become stale quickly if no one actively stewards them. This is especially true in ECL, where changes in portfolio mix, systems, macroeconomic environment and governance expectations can gradually make existing policy wording incomplete or misaligned with practice.
A strong Centre of Excellence acts as policy steward. It ensures that:
definitions remain current,methodology language matches the actual framework,portfolio coverage remains complete,governance roles are current,and policy changes are integrated with model and reporting change.
This stewardship matters because many implementation weaknesses begin as drift between policy and practice. The Centre helps prevent that drift from becoming normalised.
Even if individual models are owned by specialist teams, methodology coordination should usually sit close to the Centre of Excellence.
This means the Centre should understand and help align:
default and cure definitions,SICR philosophy,stage interpretation,portfolio segmentation,PD-LGD-EAD relationships,provision matrix design,scenario application,overlay principles,and disclosure logic.
This coordination role is important because ECL frameworks often evolve in parts. One team updates PD. Another refines stage logic. Another improves reporting. Without a coordinating function, conceptual consistency can weaken. The Centre helps preserve the overall architecture while still allowing specialised teams to deepen particular components.
One of the strongest signs of an immature ECL function is when the process produces numbers, but no one clearly owns the interpretation of those numbers.
A good Centre of Excellence should be able to answer questions such as:
Why did the reserve move this periodWhich segments drove the changeHow much of the movement was macro-drivenWhere are overlays concentratedWhich stage migrations matter mostWhat has changed since last quarter in portfolio riskWhich model limitations remain most material
This role matters because ECL is not only a mechanical reporting process. It is a credit insight framework. The Centre should therefore sit close to the interpretive layer of the process, not just the operational one.
A strong Centre of Excellence is rarely made up of one skill type alone.
Useful capability often includes a mix of:
credit risk understanding,financial reporting understanding,quantitative modelling literacy,data and technology literacy,portfolio analysis ability,control and governance discipline,and communication skill.
This mix matters because ECL sits at the intersection of all these domains. A team made up only of modellers may underweight reporting implications. A team made up only of finance professionals may underweight credit mechanics. A team made up only of control specialists may struggle with economic interpretation.
The Centre becomes powerful when it combines these viewpoints rather than choosing among them.
One of the most practical reasons to build an ECL Centre of Excellence is to reduce reliance on a few individuals who “know how it all fits together.”
In many institutions, one person knows the staging logic, another knows the overlay history, another understands the accounting bridge and another can explain the scenario mechanics. As long as those people are available, the process seems stable. But the institution itself remains fragile.
A mature Centre of Excellence turns that fragile network into institutional capability. It does so through:
clear role definition,shared documentation,repeatable governance routines,cross-training,and visible ownership of key decisions and frameworks.
This is one of the strongest signs that ECL has moved from project mode to institutional maturity.
ECL governance often involves committees, review forums and senior approvals. These forums can function well only if the information coming into them is coherent, timely and interpretable.
A strong Centre of Excellence supports this by:
preparing clear packs,highlighting material issues,distinguishing model movement from portfolio movement,summarising overlay rationale,flagging unresolved exceptions,and translating technical detail into decision-ready form.
This matters because governance committees are most effective when they challenge substance, not when they spend their time untangling fragmented explanations from multiple teams. The Centre helps create that discipline.
Even where independent validation sits outside the Centre, the response to validation findings should usually connect back into it.
Why Because findings often require coordinated action:
a policy clarification,a segmentation review,a data improvement,an overlay release,a model redevelopment,or a reporting change.
If validation remains a separate technical conversation, the framework may not improve holistically. A mature Centre of Excellence acts as the coordinating point through which validation becomes remediation rather than just commentary.
The Centre does not need to own data engineering or platform support directly. But it should work closely with those teams through structured partnership.
This often means the Centre should be able to articulate:
which data issues are material,which fields are critical,which workflow improvements matter most,what reporting outputs are required,and what architectural changes will genuinely improve ECL capability.
Without this, technology teams may deliver features that are technically elegant but not well aligned with impairment priorities. The Centre gives the technology conversation business substance.
A good Centre of Excellence does not operate only period by period. It also maintains the strategic improvement agenda for the framework.
This may include:
recurring exceptions,persistent overlays,stale policy areas,model redevelopment needs,data-quality remediation,portfolio coverage gaps,new product inclusion,disclosure enhancement,and control improvements.
The Centre’s role is to ensure that known weaknesses do not simply remain known. They should become prioritised work items with ownership and follow-through.
This is one of the key differences between a mature Centre and a pure reporting team. The former improves the framework. The latter only runs it.
An ECL framework becomes more resilient when more people understand it properly.
A Centre of Excellence should therefore often support training across the institution, including for:
finance teams,credit teams,business heads,data and technology partners,audit and control functions,and sometimes board or committee participants.
This does not mean teaching everyone to build the model. It means ensuring that the people who rely on ECL outputs or participate in ECL decisions understand the framework well enough to engage meaningfully.
This is especially important because many ECL tensions arise from misunderstanding. Training reduces that friction and improves decision quality.
A common organisational mistake is to split the world into “run the process” and “change the framework,” with no durable bridge between them.
The team running monthly or quarterly ECL may become overwhelmed with close pressure. The team managing model change or transformation may lose touch with actual operating pain points. The result is weak continuity between current operation and future improvement.
A strong Centre of Excellence helps bridge these worlds. It remains close enough to period-end execution to understand the real process, but strategic enough to shape the next evolution of the framework.
This balance is one of the most important design features of the Centre.
If the Centre is measured only by whether the reserve closed on time, it may become operationally efficient but strategically shallow.
Better maturity indicators may include:
quality of movement explanation,reduction in manual interventions,number and ageing of recurring overlays,speed of issue remediation,improvement in reconciliation quality,model and policy update discipline,reduction in key-person dependency,quality of audit outcomes,and usefulness of reporting to management.
These indicators matter because the Centre exists not only to keep the process moving, but to make the framework stronger over time.
Several recurring organisational weaknesses arise when no true Centre exists.
One is fragmented ownership, where every team owns a piece and no one owns coherence.
Another is overdependence on a small number of bridge individuals.
A third is weak linkage between run-the-bank activity and improvement priorities.
A fourth is committee overload caused by poor synthesis, because no one is preparing a clean impairment story.
A fifth is drift between policy, model, reporting and controls, because no central capability is stewarding alignment.
A sixth is slow response to findings, as validation issues and audit themes move across functions without clear coordination.
These weaknesses often become visible only as the framework grows. A Centre of Excellence exists precisely to prevent that growth from turning into fragmentation.
Some institutions hesitate to create an ECL Centre of Excellence because they imagine it requires a large new department. That is not always true.
In many cases, a strong Centre may begin with:
a clear lead owner,a small core team,formal participation from partner functions,defined forums,shared issue tracking,and clear accountability for policy, methodology coherence and movement interpretation.
What matters most is not size. It is intentionality. A small but well-defined Centre is often better than a large loose network of partial owners.
Consider two institutions with similar ECL size and complexity.
The first runs ECL through a network of finance, risk, modelling and data teams. Each team performs its part competently, but no one owns the full story. Stage logic questions go to one group, overlays to another, disclosure language to finance, model limitations to a technical team and issue remediation to whoever has capacity. The framework works, but only through repeated coordination effort.
The second has established a small ECL Centre of Excellence. It does not replace specialist teams, but it owns the impairment policy, coordinates movement analysis, consolidates material issues, maintains the overlay inventory, supports governance packs and tracks remediation priorities. The same specialised work still happens, but it now has a centre of coherence.
Both institutions may produce a reserve. Only one is clearly building lasting institutional capability.
A strong Centre of Excellence often includes:
clear mandate and ownership,defined interfaces with finance, risk, modelling, data and technology,policy stewardship,methodology coordination,movement analysis ownership,overlay and issue governance,support for committees and approvals,training and documentation leadership,and a standing improvement agenda.
The strength of this structure lies in turning ECL from an activity into a competence.
Building an ECL Centre of Excellence is one of the clearest signs that an institution has moved beyond implementation into maturity. It means the institution no longer relies on the framework merely existing in documents, models and systems. It builds a durable organisational capability around it. It gives ECL a home. It gives the process a coherent voice. It gives senior management a place where the full impairment story can be understood, challenged and improved. It reduces fragmentation, lowers key-person dependency and creates continuity between current-period execution and future evolution.
A strong institution recognises that systems alone do not create excellence, and neither do committees in isolation. Excellence comes when specialist knowledge, governance discipline, interpretive insight and operating ownership are brought together intentionally. That is what a Centre of Excellence does.
In that sense, this pillar teaches a deeper organisational truth about ECL: the framework becomes truly sustainable not when one team can run it, but when the institution can carry it as a shared and well-coordinated capability over time.
