Aller au contenu
Corporate & Investment Banking Mon parcours English

Accueil / Corporate Banking / Fonctions support / Module 15

15Operations, Finance, Data, Technology & Reporting

The Machinery Behind Corporate Lending · Loan Operations, IFRS 9 Accounting and Provisioning, Client and Contract Data, Lending Platforms and Credit Reporting

La machinerie qui enregistre, administre, comptabilise et déclare chaque crédit.

Lecture
50 min
Sections
49
Exercices
19
Questions de rappel
30
Questions de QCM
15

Le texte de ce module est encore en anglais : la version française arrive. Les outils interactifs, eux, sont déjà en français.

Sa place dans le Corporate Banking
  1. Origination
  2. Structuration
  3. Analyse crédit et notation
  4. Approbation
  5. Exécution et enregistrement
  6. Gestion courante
  7. Suivi
  8. Restructuration et sortie
  • KYC
  • Conformité
  • Risque
  • Juridique
  • Opérations
  • Finance
  • Données
  • Technologie
  • Reporting
  • Trésorerie / ALM

Ponts vers Capital Markets & Securities Services

This module teaches the machinery of corporate banking from zero, but goes deep enough for a professional

It assumes no prior knowledge of loan operations, banking accounting or data management and builds every idea from its foundations. Yet it goes deep enough to be useful to someone who runs a loan operations team, computes provisions, owns client or contract data, implements a lending platform, prepares credit reporting for supervisors, or leads the transformation of a corporate bank's operating model.

Wherever a concept is technical, we stop and explain it in plain language before using it again. Wherever a term is jargon, we define it once and then rely on it.

Two threads run through everything: the business and the technology

The business thread covers the five enablers that turn corporate banking decisions into reality. Operations books and services every loan, drawing, rate fixing, fee and payment. Finance decides how loans are measured, how income is recognised and how expected credit losses are provisioned under IFRS 9, and measures the business's profitability. Data holds the client, contract, collateral and exposure information on which every other function relies. Technology provides the platforms. Reporting turns all of it into management information, regulatory returns and client statements.

The technology thread runs through all five: loan servicing platforms, origination and workflow tools, accounting engines and provisioning models, data platforms and lineage, integration and APIs, and the reporting chains that deliver AnaCredit, FINREP and COREP returns.

Where this module sits in the Corporate Banking track

This is the second and last module on the shared enablers and closes the Corporate Banking track. It follows the Treasury and ALM module and the Risk, Compliance and Legal module. It does not repeat the CIB series' modules on trade capture, settlement, reconciliation, reference data and regulatory reporting, which cover those disciplines for markets businesses; it focuses on what is specific to corporate lending and transaction banking.

How to read this module

Read it as a narrative, not a reference card

Each part assumes you have read the ones before it. If you are short on time, read the section titles and the "In a nutshell" boxes first.

Each coloured box plays a fixed role

The boxes play the same roles as in the earlier modules: Key insight, Going deeper, IT & data lens, Worked example, Watch out, In a nutshell, Common misconception, Practice and Explain it back.

Two running cases: a performing loan and a deteriorating one

We follow the two credits of the previous module through the machinery. Helion's new 150 million euro term loan is booked, drawn, serviced, accounted for at amortised cost with its fee spread over its life, and reported to the central bank's credit register. Vela Packaging's term loan deteriorates: it moves from stage 1 to stage 2 and later stage 3 under IFRS 9, its provisions rise, its forbearance must be reported, and its data must be right at every step. Every number is invented to teach a mechanic.

The module is built to be worked, not only read

Practice boxes contain the calculations and designs these functions actually work on: interest accruals and rollovers, effective interest rates, expected credit loss by stage, significant increase in credit risk, reconciliation breaks, AnaCredit data attributes, data quality rules, and the design of a target operating model.

A suggested path through the document

Four sessions work well: Part 1, the five enablers and their frameworks; Part 2, the end-to-end flow from booking to reporting, in two sittings; Parts 3 and 4, the actors and the systems; Part 5, two episodes and the transformation agenda.

What you will master

You will understand loan operations

You will know how a corporate loan is booked, drawn, rolled over, accrued, paid and reconciled, and how operations controls the risks of manual processing.

You will understand the accounting of corporate loans

You will know how IFRS 9 classifies and measures loans, how fees are recognised through the effective interest rate, how expected credit losses are computed in three stages, and how forbearance and non-performing exposures are defined.

You will understand the data a corporate bank runs on

You will know the core data domains — client, hierarchy, contract, collateral, exposure — what golden sources and lineage are, and what supervisors expect under the BCBS 239 principles.

You will know the technology landscape

You will know the main platforms of a corporate bank — origination, loan servicing, payments, accounting, data and reporting — and how they integrate.

You will understand credit reporting

You will know how internal management information, AnaCredit, FINREP, COREP and other returns are built from the same data, and why consistency matters.

You will recognise the failures and the transformation agenda

You will know what the Covid provisioning test and the Citigroup consent orders taught banks, and how operating models, data and technology are being rebuilt.

1Context and the Big Picture

Before following Helion's loan from booking to the central bank's credit register, you need to see the five enablers as one chain, and understand the frameworks that govern each: loan operations, IFRS 9 accounting and provisioning, data principles, the technology landscape and the reporting obligations of a corporate bank.

Five enablers, one chain

From decision to report

A corporate banking decision travels through a chain of enablers. Operations books the facility and processes every event. Finance measures it in the accounts and provisions for its credit risk. Data holds the facts about the client, the contract and the exposure. Technology runs the platforms that do all this. Reporting delivers the results to management, supervisors and clients. A weakness at any link propagates downstream: a wrong booking becomes a wrong accrual, a wrong provision and a wrong regulatory return.

Figure 1 The enabler chain. Decisions from the front office are booked and serviced by operations, measured by finance, held as data, run on technology and delivered as reports.

Loan operations

What operations does for a corporate loan

Activity Description
Set-up and booking Creating the facility, its tranches, limits, pricing and schedules in the loan system after CPs are satisfied
Drawdowns and rollovers Processing drawing requests, rate fixings, rollovers of loans at the end of interest periods
Accruals and billing Calculating interest and fees daily, issuing notices, collecting payments
Payments and reconciliation Settling cash with borrowers and lenders, reconciling accounts and nostros
Amendments and events Processing margin changes, waivers, extensions, transfers, prepayments
Collateral and documents Recording security, tracking perfection and renewals, holding original documents
Agency For syndicated loans, the agency activity described in the agency module

The operational risks of lending

Corporate loan operations remain more manual than payments or securities processing: facilities are bespoke, events are triggered by notices and emails, and many calculations depend on terms read from long agreements. The main risks are wrong set-up of terms, missed or wrong rate fixings, wrong interest calculations, payments to wrong accounts, unreconciled breaks, and lost track of collateral perfection or document custody. The Revlon case in the agency module showed how expensive a single processing error can be.

Finance: IFRS 9 and corporate loans

Classification and measurement

Under IFRS 9, a loan's accounting depends on two tests. The business model test asks whether the bank holds the loan to collect its contractual cash flows, holds it to collect and sell, or holds it for sale or trading. The SPPI test asks whether its contractual cash flows are solely payments of principal and interest — consistent with a basic lending arrangement.

Situation Classification and measurement
Held to collect, cash flows pass SPPI Amortised cost, with expected credit loss provisions
Held to collect and sell, SPPI Fair value through other comprehensive income
Held for sale — e.g., the part of an underwritten loan the bank intends to syndicate Fair value through profit or loss
Cash flows fail SPPI — e.g., returns linked to the borrower's equity value Fair value through profit or loss

Most corporate loans are held to collect and measured at amortised cost. The portion of an underwritten leveraged loan intended for syndication is usually held for sale and measured at fair value, which is why hung deals in 2022 produced immediate losses, as the distribution module explained.

Fees and the effective interest rate

Fees that are an integral part of a loan's yield — such as arrangement or upfront fees on a loan the bank keeps — are not recognised immediately but spread over the loan's life through the effective interest rate. Fees for services — such as agency fees — are recognised as the service is provided, and syndication fees are recognised when the syndication is completed, for the part the bank does not keep.

Expected credit losses in three stages

IFRS 9 requires provisions for expected credit losses, not only for losses already incurred. Each loan is placed in a stage:

Stage Condition Provision
Stage 1 No significant increase in credit risk since origination 12-month expected credit loss
Stage 2 Significant increase in credit risk since origination, not credit-impaired Lifetime expected credit loss
Stage 3 Credit-impaired, generally defaulted Lifetime expected credit loss; interest on net carrying amount

The move from stage 1 to stage 2 — a significant increase in credit risk — is judged by comparing the current lifetime probability of default with the one at origination, with backstops such as 30 days past due, and qualitative triggers such as watchlist status or forbearance. Expected losses are probability-weighted across several economic scenarios.

Forbearance and non-performing exposures

Regulatory definitions complement IFRS 9. A forborne exposure is one whose terms have been modified because the borrower is in financial difficulty. A non-performing exposure is one that is 90 days past due or unlikely to be paid in full without realisation of collateral. These definitions drive regulatory reporting and supervisory expectations on provisioning, including the EU's prudential backstop that requires minimum coverage of non-performing loans over time.

Data principles

The data a corporate bank runs on

Domain Contents Used by
Client and hierarchy Legal entities, identifiers, ownership, groups of connected clients Credit, KYC, limits, reporting
Contract and facility Facilities, tranches, terms, pricing, covenants, schedules Operations, finance, risk, reporting
Collateral and guarantees Security, valuations, perfection, links to facilities Credit, finance, reporting
Exposure and position Drawn and undrawn amounts, accruals, past-due status Risk, finance, reporting
Risk parameters Ratings, PD, LGD, stages Risk, finance, capital, reporting
Market and reference data Rates, FX, calendars, industry codes All

BCBS 239

In 2013, after the financial crisis showed that many banks could not aggregate their risk exposures quickly and accurately, the Basel Committee published fourteen principles for effective risk data aggregation and risk reporting, known as BCBS 239. They require governance, data architecture and IT infrastructure that allow accurate, complete, timely and adaptable risk data, and reports that are accurate, comprehensive, clear and timely. Global systemically important banks were expected to comply from 2016; supervisors have extended the expectations to other significant banks. The ECB published a guide on risk data aggregation and risk reporting in 2024, noting that most banks still fell short.

The technology landscape

The main platforms

Platform Purpose
Origination and credit workflow Applications, approvals, conditions
Loan servicing Facilities, drawings, accruals, billing, agency
Trade finance and cash management The platforms of the transaction banking modules
Payments Execution and screening of payments
Accounting and subledgers Postings, general ledger, effective interest, provisioning
Risk engines Ratings, capital, limits, expected credit loss models
Data platforms Integration, golden sources, lineage, analytics
Reporting Management information, regulatory and client reporting

Reporting obligations

Internal and external

The corporate bank reports internally — profitability, portfolio quality, pipeline, limits — and externally. External reporting includes COREP templates on capital, credit risk and large exposures; FINREP templates on financial information, including forborne and non-performing exposures; AnaCredit, the euro area's granular credit register; liquidity and interest rate risk returns; and Pillar 3 public disclosures, now including ESG information. The CIB regulatory reporting module covers the overall framework; this module focuses on what corporate lending contributes.

AnaCredit — analytical credit datasets — requires euro-area banks to report to their national central banks, loan by loan, credit exposures to legal entities where the debtor's total commitment with the bank is at least 25,000 euros, with close to a hundred attributes on the instrument, the counterparties, the protection and the accounting. It has applied since 2018. Its granularity means that every data error in the loan system appears in the central bank's register, and central banks run quality checks that come back to the bank as queries.

Work these through before reading on. Full solutions are in the Answer Key.

1.1 Which enabler?

Assign each task to operations, finance, data, technology or reporting: processing Helion's rollover; deciding whether Vela moves to stage 2; maintaining the Helion group hierarchy; upgrading the loan servicing platform; submitting AnaCredit records.

Rollover: operations. Stage decision: finance with risk. Group hierarchy: data, owned by client data management. Platform upgrade: technology. AnaCredit submission: reporting.

1.2 Classification

How would IFRS 9 classify: (a) Helion's term loan, held to maturity; (b) the 200 million euro portion of an underwritten LBO loan the bank intends to sell within weeks; (c) a loan whose interest includes a share of the borrower's profits?

(a) Amortised cost. (b) Fair value through profit or loss, held for sale. (c) Fair value through profit or loss, because profit-linked returns fail the SPPI test.

1.3 Effective interest

A bank keeps a 200 million euro four-year loan with an upfront fee of 0.40 per cent. How much fee income is recognised at signing and, approximately, in each year?

Fee: 800,000 euros. Nothing is recognised as income at signing; about 200,000 euros a year is recognised through the effective interest rate over four years.

Explain to a friend why a bank cannot treat the fee it charges for a five-year loan as profit on the day the loan is signed.

      1. Name the five enablers and what each does.

      2. List the main activities of loan operations and their risks.

      3. What two tests determine a loan's IFRS 9 classification?

      4. Describe the three stages of expected credit loss.

      5. What are forborne and non-performing exposures?

      6. What are the BCBS 239 principles, and what did the ECB find in 2024?

      1

      1. Which two tests set a loan's IFRS 9 classification?

      2. What does finance do among the five enablers?

      3. Which is a main loan operations risk?

      2The End-to-End Flow

      This part follows Helion's term loan from booking to the central bank's credit register, through drawdown, accruals, payments, the monthly close, data checks and reporting. It then follows Vela's loan through stage transfers, forbearance and default, showing how operations, finance, data and reporting must move together.

      From booking to reporting in eight steps

      Step Lead enabler What happens Key control
      1. Set-up and booking Operations, finance Facility created with terms, schedules and accounting classification Four-eyes check against the agreement
      2. Drawdown and rate fixing Operations Drawing processed, rate fixed, cash paid Request validation, rate source
      3. Daily accrual and accounting Operations, finance Interest and fees accrued; effective interest applied Accrual reconciliation
      4. Billing, payment and reconciliation Operations Interest billed and collected at period end Cash-to-system reconciliation
      5. Monthly close and provisioning Finance, risk Staging, ECL calculation, overlays, P&L, FTP Model governance, review
      6. Data quality and aggregation Data Data checked, aggregated, reconciled to the ledger Quality rules, lineage
      7. Reporting Reporting Management information, COREP, FINREP, AnaCredit Reconciliation, validation rules
      8. Events and lifecycle changes All Amendments, forbearance, stage transfers, default, write-off Consistent updates across systems
      Figure 2 Helion's loan through the enabler chain. Operations events drive accounting; the monthly close adds provisions; data checks feed every report.

      Step 1 — Set-up and booking

      Translating the agreement into the system

      Once the conditions precedent are satisfied, loan operations creates the facility in the loan servicing platform: borrower and guarantor identifiers, facility type, currency, amount, availability period, maturity, repayment schedule, interest basis — six-month Euribor, actual/360, the margin grid — fees, the agent and lender shares if syndicated, and the links to collateral and limits. Finance's classification — amortised cost, business model "hold to collect", SPPI passed — is recorded, so that the accounting engine treats the loan correctly. A second person checks the set-up against the agreement and the credit approval.

      Step 2 — Drawdown and rate fixing

      Cash out

      Helion submits a utilisation request for the full 150 million euros, for an interest period of six months. Operations validates the request — notice period, amount, conditions — fixes the rate from the Euribor administrator's publication two business days before the period starts, books the drawing and pays the funds to Helion's account. The loan system creates the outstanding, the interest period and the accrual schedule.

      Step 3 — Daily accrual and accounting

      Interest every day

      Each day, the loan system accrues interest on the outstanding amount at the fixed rate: six-month Euribor of 2.30 per cent plus the 1.20 per cent margin, 3.50 per cent, actual/360. The accounting engine posts the accrual to the general ledger, together with the daily amortisation of the upfront fee through the effective interest rate. Undrawn commitments accrue commitment fees in the same way.

      Step 4 — Billing, payment and reconciliation

      End of the interest period

      At the end of the six-month period, the loan system calculates the interest due and issues a notice to Helion. Helion pays; operations matches the payment to the notice and closes the period; a new rate is fixed for the next period — a rollover.

      Step 5 — The monthly close and provisioning

      Staging and expected credit loss

      At each month end, finance and risk compute provisions. For each loan, the provisioning engine takes the current rating, the lifetime probability of default at origination and today, the loss given default, the exposure profile and the macroeconomic scenarios, determines the stage, and calculates the expected credit loss.

      Overlays and judgement

      Models are calibrated on history and may miss new risks — a pandemic, a sector shock, an energy crisis. Banks therefore apply post-model adjustments, or overlays, approved by governance committees and documented, to reflect risks the models do not capture. Supervisors scrutinise overlays closely, as Part 5 explains.

      Profitability and FTP

      The close also produces the corporate bank's management results: client margins after FTP, fees, provisions, operating costs, capital and RAROC, as the Treasury and ALM module described. These must reconcile to the general ledger, so that management and financial accounts tell the same story.

      Step 6 — Data quality and aggregation

      Checking before reporting

      Before reports are produced, data quality rules check the loan data: every exposure has a valid counterparty identifier and hierarchy; every facility has a classification, stage and rating; outstandings reconcile to the ledger; collateral links are complete; dates and amounts are plausible. Failures are routed to the owners — operations, risk, client data — for correction. Lineage tools record where each reported figure came from, so that errors can be traced to their source.

      Step 7 — Reporting

      One loan, many reports

      Helion's loan appears in many reports, each with its own definitions:

      Report What Helion's loan contributes
      Internal management information Volumes, margins, RAROC, portfolio quality
      COREP credit risk and capital Exposure, risk weight, risk-weighted assets
      COREP large exposures Contribution to the Helion group's exposure
      FINREP Carrying amount, stage, provisions, forbearance status
      AnaCredit Loan-level record with instrument, counterparty, protection and accounting attributes
      Pillar 3 and ESG disclosures Sector, stage, and climate-related information

      AnaCredit in practice

      Step 8 — Events and lifecycle changes

      Keeping everything in step

      When something changes, every enabler must move together. When Vela's lenders agree to waive a covenant and extend maturities in exchange for new equity, operations updates the facility; risk assesses whether the change is a forbearance measure — yes, because Vela is in financial difficulty — and whether Vela is now non-performing or unlikely to pay; finance assesses whether the modification is substantial enough to derecognise the old loan and recognise a new one, and computes any modification gain or loss; data updates flags; reporting reflects forborne and non-performing status in FINREP and AnaCredit.

      Default and write-off

      If Vela defaults, the loan moves to stage 3: interest is recognised on the net carrying amount, the provision reflects expected recoveries from the restructuring, and the exposure is reported as non-performing. When recovery prospects are exhausted, the uncollectible part is written off — removed from the balance sheet against the provision — while recovery efforts may continue.

      Work these through before reading on. Full solutions are in the Answer Key.

      2.1 Interest accrual

      An 80 million euro loan bears three-month Euribor of 2.10 per cent plus 1.50 per cent, actual/360. Compute interest for a 91-day period.

      80,000,000 × 3.60% × 91 / 360 = 728,000 euros.

      2.2 ECL by stage

      A loan of 200 million euros has an LGD of 40 per cent, a 12-month PD of 0.8 per cent and a lifetime PD of 6 per cent. Compute the stage 1 and stage 2 provisions, using the simplified method of the worked example.

      Stage 1: 200 × 0.8% × 40% = 0.64 million euros. Stage 2: 200 × 6% × 40% = 4.8 million euros.

      2.3 Significant increase

      A loan's lifetime PD was 2 per cent at origination and is now 5 per cent; the bank's threshold is a doubling. It is 10 days past due and not on the watchlist. Which stage is it in, and why?

      Stage 2: the lifetime PD has risen 2.5 times, above the doubling threshold, which is a significant increase in credit risk even though the 30-days-past-due backstop is not reached.

      2.4 Reconciliation

      The loan system shows 2,000,000 euros of interest due from a borrower; 1,986,111.11 euros arrive. The loan is 100 million euros at 5.00 per cent, actual/360. What is the likely cause of the difference?

      Daily interest: 100,000,000 × 5.00% / 360 = 13,888.89 euros. The difference of 13,888.89 euros is one day's interest: likely a day-count or business-day convention difference between the borrower's and the bank's calculations.

      2.5 Forbearance

      For each case, say whether it is forbearance: (a) a healthy borrower negotiates a lower margin because market conditions have improved; (b) a borrower struggling with cash flow obtains a one-year payment holiday; (c) a borrower in breach of covenants obtains a waiver in exchange for a higher margin.

      (a) Not forbearance: the borrower is not in financial difficulty. (b) Forbearance: a concession to a borrower in difficulty. (c) Forbearance: a waiver granted to a borrower in difficulty is a concession, even with a higher margin.

      Explain to a relative why a bank must set aside money for losses on a loan that is still being paid on time.

          1. Name the eight steps from booking to reporting.

          2. Why are set-up errors so dangerous, and how are they prevented?

          3. What happens at a rollover?

          4. How is a loan's IFRS 9 stage determined, and what changes in the provision?

          5. What are post-model adjustments, and why are they scrutinised?

          6. How must the enablers act together when a loan is forborne or defaults?

          2

          1. Why are set-up errors so dangerous?

          2. What happens at a rollover?

          3. Which step follows monthly close and provisioning?

          3Roles, Actors, and Interactions

          The enablers are large organisations: operations centres, finance and accounting teams, data offices, technology departments and reporting functions, often spread across locations and supported by vendors and outsourcers. This part maps the cast, explains who owns what, and shows where the hand-offs between them create risk.

          The cast of the enablers

          Function Teams Role for corporate banking
          Operations Loan operations, agency operations, documentation and CP teams, collateral operations, client service Book, service and control every facility and event
          Finance Financial control, accounting policy, provisioning, management accounting, regulatory reporting Measure, provision, report and analyse
          Data Chief data office, data owners and stewards, data quality teams Define, own and improve data
          Technology Platform owners, architecture, engineering, infrastructure, security Build and run the systems
          Reporting Management information, regulatory reporting, client reporting Deliver reports from shared data
          External parties Vendors, outsourcers, external auditors, supervisors, central banks Provide, audit, supervise, collect
          Figure 3 The cast of the enablers around the corporate bank. Operations, finance, data, technology and reporting serve the business and answer to auditors and supervisors.

          Operations

          Organised by product and by location

          Corporate banking operations are typically organised by product — loan operations, agency, trade, cash management — and run from a mix of locations: teams close to clients for complex events, and shared service or offshore centres for high-volume processing. Loan operations works closely with the relationship manager and credit on conditions precedent, with the agent on syndicated loans, and with finance on accounting events.

          The business control office

          Many corporate banks have a business control or middle office function that sits between the front office and operations: it checks that deals are booked as approved, monitors operational risk indicators, follows up breaks and incidents, and coordinates remediation. It is the corporate bank's first-line control on its own processes.

          Finance

          Accounting, provisioning and management information

          Financial control owns the general ledger and the accuracy of the accounts for the corporate bank's products. Accounting policy interprets IFRS 9 for new products and structures — SPPI tests for unusual loan features, derecognition of restructured loans. The provisioning team, often shared between finance and risk, runs the expected credit loss process, the overlays and the disclosures. Management accounting produces profitability by client, product and team, with FTP and cost allocations. Regulatory reporting produces FINREP, COREP and other returns, usually with risk.

          Finance and risk together

          Provisioning is a joint responsibility: risk provides models, ratings and parameters; finance owns the accounting outcome and the disclosures. Disagreements — on staging criteria, overlays, the weight of scenarios — are resolved in provisioning committees chaired by the CFO or CRO.

          Data

          Ownership and stewardship

          Good data management rests on clear ownership. The business that creates the data — the corporate bank for client and facility terms, risk for ratings and parameters, finance for accounting classifications — is its data owner, accountable for its quality. Data stewards manage day-to-day definitions and quality. The chief data office sets policies, standards, the data catalogue and the quality framework, and coordinates compliance with BCBS 239.

          Technology

          Run and change

          Technology teams run the bank — keeping platforms available, secure and patched — and change the bank — implementing new products, regulations and modernisation programmes. For corporate banking, platform owners are responsible for the loan servicing system, origination tools, trade and cash management platforms and their integrations. Architecture teams decide how systems connect and which are strategic.

          Vendors and outsourcing

          Corporate banks rely heavily on vendors: loan servicing platforms such as Finastra's Loan IQ, trade and payment platforms, data and reporting tools, cloud providers. Under the EU's Digital Operational Resilience Act, banks must manage the risks of these ICT third parties, maintain registers of arrangements, include required clauses in contracts, and plan for exit; critical providers are subject to direct oversight by EU authorities.

          Reporting

          Three audiences, one source

          Reporting serves management, supervisors and clients. The strongest reporting functions build all three from the same data, reconciled to the ledger, rather than separate extractions that diverge. Regulatory reporting for corporate lending — credit risk, large exposures, forborne and non-performing exposures, AnaCredit — is especially demanding because it is granular, frequent and checked by supervisors against other returns.

          External parties

          Auditors, supervisors and central banks

          External auditors audit the financial statements, with particular attention to expected credit losses, which are a key audit matter in most banks' reports. Supervisors review provisioning, data aggregation and reporting quality, operational resilience and outsourcing. National central banks collect AnaCredit data, run quality checks and send queries back to banks.

          Hand-offs and accountability

          Hand-off What can go wrong Good practice
          Front office to operations Terms booked differently from the agreement Structured term sheets, four-eyes, business control checks
          Operations to finance Events not reflected in accounting Automated postings, daily reconciliations
          Risk to finance Ratings or parameters not updated before close Calendars, cut-offs, governance
          Business to data Data created without ownership Data owners accountable for quality
          Data to reporting Reports built from inconsistent extracts Single source, lineage, reconciliation
          Bank to vendors Dependencies unmanaged DORA registers, contracts, exit plans

          Work these through before reading on. Full solutions are in the Answer Key.

          3.1 Data ownership

          Assign an owner for each data element: Helion's legal name and LEI; the margin grid of its term loan; its internal rating; its IFRS 9 stage; the value of Vela's pledged shares.

          Legal name and LEI: client data management, on behalf of the corporate bank. Margin grid: the corporate bank, booked by loan operations. Internal rating: credit risk. IFRS 9 stage: finance with risk. Value of pledged shares: credit risk, through collateral management.

          3.2 Hand-off failure

          A margin step-down on a loan, triggered by an improved leverage ratio, was agreed by the agent but not applied in the bank's loan system for two quarters. Which hand-off failed, what are the consequences, and how should it be prevented?

          The hand-off between the agent's notice and the bank's loan operations — or between the front office and operations — failed. The bank overcharged interest for two quarters, owing a refund, with wrong income and wrong data in reports. Prevention: automated ingestion of agent notices, covenant-linked margin grids monitored against compliance certificates, and business control checks.

          3.3 DORA

          List four obligations a bank has under DORA regarding the vendor of its loan servicing platform.

          Record the arrangement in the register of ICT third-party arrangements; include the required contractual clauses on service levels, security, audit and termination; assess concentration and the criticality of the function supported; maintain and test an exit strategy and business continuity plans.

          Explain to a friend why the people who type a client's details into a system should be responsible for the accuracy of a report they never see.

              1. Name the teams within operations, finance, data, technology and reporting that serve corporate banking.

              2. What does a business control office do?

              3. How do finance and risk share responsibility for provisioning?

              4. Who owns data, and what does the chief data office do?

              5. What does DORA require regarding ICT third parties?

              6. Where do the enablers most often fail, and why?

              3

              1. What does a business control office do?

              2. How do finance and risk share provisioning?

              3. Which team is part of operations?

              4Systems, Data, and IT

              This part opens the corporate bank's technology architecture as a whole: the platforms each earlier module described, how they connect, where data lives, and how the accounting, provisioning and reporting chains are built on top. It ends with the target architecture most banks are moving toward.

              The corporate bank's architecture in eight layers

              Layer Contents Examples from this track
              Channels Client portals, host-to-host, APIs, relationship manager tools Trade and cash portals, lender portals
              Origination and workflow CRM, credit workflow, pricing tools, KYC onboarding Pricing tool, credit workflow
              Product platforms Loan servicing, agency, trade finance, cash management, payments Loan IQ, trade platform, payment hub
              Risk engines Ratings, capital, limits, collateral, ECL models, early warning Capital engine, provisioning engine
              Finance engines Subledgers, effective interest, general ledger, consolidation Accounting engine
              Data platform Golden sources, integration, data warehouse or lake, lineage, catalogue Client master, facility data hub
              Reporting Management information, regulatory and client reporting AnaCredit, FINREP, COREP chains
              Infrastructure and security Data centres, cloud, identity, resilience DORA-compliant operations
              Figure 4 The corporate bank's layered architecture. Product platforms sit between channels and the risk, finance and data layers; reporting is built on the data platform.

              Product platforms

              Loan servicing at the core

              The loan servicing platform is the system of record for corporate loans: facilities, tranches, drawings, schedules, rates, accruals, fees, payments and, for agent banks, lender shares. Large banks often use a vendor platform — Finastra's Loan IQ is the most widespread for syndicated and corporate lending — sometimes alongside older in-house systems for specific portfolios. Trade finance, cash management and payment platforms, described in the transaction banking modules, sit beside it.

              Integration between platforms

              A single client relationship touches several platforms: a revolving facility in the loan system, letters of credit in the trade platform, accounts and pools in cash management, derivatives in markets systems. Integration — through messaging, files and APIs — must keep identifiers, limits and exposures consistent. A client that appears under different identifiers in each platform cannot be aggregated reliably.

              Risk and finance engines

              Expected credit loss engine

              The ECL engine takes exposures and their cash-flow profiles from the product platforms, ratings and lifetime PD curves, LGD models and collateral values from risk, and macroeconomic scenarios with their weights; determines stages; calculates expected losses; applies approved overlays; and feeds the accounting engine and disclosures. It must keep a full audit trail: which model version, which inputs, which overlay, for each month.

              Accounting engine

              The accounting engine — often a subledger fed by product platforms — turns events into accounting entries: drawings, accruals, effective interest amortisation, fees, provisions, write-offs, modifications. It applies the accounting rules for each product and classification and posts summarised entries to the general ledger. Because regulatory returns such as FINREP are built on accounting data, the subledger is also a key source for reporting.

              The data platform

              Golden sources

              For each data domain, the bank designates a golden source: the authoritative system or dataset from which all others take the data. For corporate banking, typical golden sources are a client master with legal entities, identifiers and hierarchies; the loan platform for facility terms and balances; the rating system for ratings; the collateral system for security; the ledger for accounting balances. Other systems should consume from the golden source rather than maintain their own copies.

              Lineage, catalogue and quality

              A data catalogue describes each data element: its definition, owner, golden source and uses. Lineage traces how each reported figure is derived from source data through transformations. Data quality rules check completeness, validity, consistency and timeliness, with dashboards showing scores by domain and owner. These are the practical tools of BCBS 239 compliance.

              A single facility identifier, assigned at origination and carried through the workflow, the loan platform, the risk and finance engines, the data platform and every report, is one of the most valuable — and most often missing — elements of a corporate bank's architecture. Without it, reconciling a loan's AnaCredit record with its FINREP carrying amount and its COREP exposure requires matching on names, amounts and dates. With it, lineage and reconciliation become automatic.

              Reporting chains

              From data to returns

              Regulatory reporting chains extract data from the data platform, apply regulatory definitions — exposure classes, counterparty sectors, forbearance flags — produce templates in the formats required, validate them against the regulator's rules, reconcile them with one another and with the accounts, and submit them. For AnaCredit, the chain handles loan-level records, monthly, with the central bank's feedback loop of queries and corrections.

              Target architecture

              Where banks are heading

              Most corporate banks are moving toward a similar target: fewer, modernised product platforms, often vendor-based and increasingly cloud-hosted; APIs connecting channels, platforms and engines; a data platform with golden sources, a catalogue and lineage, from which risk, finance and reporting all consume; event-driven flows that update downstream systems as events happen rather than in overnight batches; and automation — including AI — of manual operations such as reading notices, extracting terms and investigating breaks.

              Legacy and migration

              The obstacle is legacy: decades-old systems holding long-dated loans, with undocumented customisations and many downstream dependencies. Migrations are long, risky and expensive, as the export finance module's warning about preserving decades of loan history showed. Banks therefore modernise in stages, often building the data platform first so that reporting can be fed consistently while product platforms are replaced one by one.

              Work these through before reading on. Full solutions are in the Answer Key.

              4.1 Golden sources

              Name the golden source for: Helion's LEI and parent; the outstanding amount of its term loan; its internal rating; its IFRS 9 stage; the accounting carrying amount.

              LEI and parent: client master. Outstanding amount: loan servicing platform. Internal rating: rating system. IFRS 9 stage: ECL engine. Carrying amount: accounting subledger and general ledger.

              4.2 Data quality rules

              Write five data quality rules for corporate loan data used in AnaCredit.

              Every counterparty with a commitment of 25,000 euros or more has a valid LEI or national identifier; every instrument has a non-empty inception and maturity date with maturity after inception; outstanding nominal amounts reconcile to the loan platform within tolerance; each instrument has an IFRS 9 stage consistent with the ECL engine; each secured instrument links to at least one protection item with a value.

              4.3 Lineage

              Describe the lineage of the stage-2 provision on Vela's loan, from source data to the FINREP template.

              Vela's exposure and schedule come from the loan platform; rating and lifetime PD from the rating system and PD curves; LGD from the LGD model using collateral values from the collateral system; scenarios from economics; the ECL engine computes the stage and provision, with any overlay; the accounting subledger posts it; the general ledger holds it; the FINREP chain extracts it with the stage and counterparty sector and maps it to the template.

              4.4 Migration plan

              A bank wants to replace an old loan system holding 3,000 corporate facilities with a vendor platform. Propose the main phases and the controls for each.

              Data analysis and cleansing of the 3,000 facilities; mapping of terms and history, including effective interest data; build and configure the new platform; parallel runs comparing accruals and balances; reconciliation of every facility before cut-over; staged migration by portfolio; post-migration monitoring and decommissioning of the old system with archived history. Controls: reconciliations at each stage, sign-off by operations, finance and risk, rollback plans.

              Explain to a colleague why giving every loan one identifier from the day it is approved can save a bank thousands of hours of reconciliation.

                  1. Name the eight layers of the corporate bank's architecture.

                  2. What is the loan servicing platform's role?

                  3. What does the ECL engine take in and produce?

                  4. What are golden sources, a data catalogue and lineage?

                  5. Why must reports be reconciled with each other?

                  6. Describe the target architecture and the main obstacle to reaching it.

                  4

                  1. What is the loan servicing platform?

                  2. What does the ECL engine produce?

                  3. Which is NOT one of the eight architecture layers?

                  5Risks, Pain Points, and Transformation

                  The enablers fail quietly and expensively: wrong interest for years, provisions that miss a crisis, reports that supervisors cannot trust, data programmes that never finish. This part maps the risks, studies the Covid provisioning test and the Citigroup data consent orders, and sets out the transformation agenda that closes the Corporate Banking track.

                  A MECE map of enabler risk

                  Risk What goes wrong Typical control
                  Processing Wrong set-up, fixings, accruals, payments Four-eyes, automation, reconciliation
                  Accounting Wrong classification, income recognition, modification treatment Accounting policy review, automated rules
                  Provisioning and model Provisions miss risks or are mis-staged Model governance, overlays, back-testing
                  Data quality Incomplete, inconsistent or untimely data Ownership, quality rules, lineage
                  Reporting Wrong or inconsistent regulatory returns Validation, cross-report reconciliation
                  Technology and resilience Outages, failed migrations, cyber Resilience testing, DORA controls
                  Third-party Vendor failure or dependency Registers, contracts, exit plans
                  Change Programmes late, over budget, or failing to deliver Governance, staged delivery
                  Figure 5 Enabler risks along the chain: processing risk in operations, accounting and model risk in finance, quality risk in data, resilience risk in technology, consistency risk in reporting.

                  Two episodes that reshaped the enablers

                  The Covid provisioning test, 2020–2021 — when models met a crisis they had never seen

                  What happened

                  IFRS 9's expected credit loss model had been in force for only two years when the Covid-19 pandemic hit in March 2020. Mechanically applied, the models — driven by macroeconomic forecasts that had collapsed overnight — would have moved large portfolios into stage 2 and multiplied provisions, just as governments were supporting borrowers through payment moratoria and state-guaranteed loans. Standard-setters and supervisors, including the IASB, the ECB and the EBA, issued statements in March and April 2020 clarifying that IFRS 9 required judgement, that general payment moratoria did not automatically mean a significant increase in credit risk or forbearance, and that banks should give weight to long-run forecasts and to the effect of support measures. Banks raised provisions substantially but relied heavily on post-model adjustments. As support measures ended, supervisors pressed banks to identify borrowers in genuine difficulty, and later to maintain overlays for new risks such as energy prices and inflation.

                  Root cause

                  Models calibrated on historical relationships between the economy and defaults could not capture an unprecedented shock combined with unprecedented public support, so outcomes depended on judgement.

                  The lesson

                  Provisioning is a process, not just a model. Banks needed governance able to approve, document and later release overlays; data able to identify borrowers under moratoria and their underlying condition; and close cooperation between risk, finance and the corporate bank, whose relationship managers knew which clients were truly struggling. The episode also showed how much provisioning depends on the quality of operational data: banks that could not quickly flag every loan under a moratorium struggled to report and provision correctly.

                  Citigroup's consent orders, 2020–2024 — when data became a supervisory priority

                  What happened

                  In October 2020, US regulators — the Office of the Comptroller of the Currency and the Federal Reserve — issued consent orders against Citigroup for long-standing deficiencies in risk management, data governance and internal controls, and the OCC imposed a 400 million dollar civil money penalty. The orders followed years of supervisory findings and came shortly after the Revlon mistaken payment described in the agency module had illustrated weaknesses in the bank's processes. Citigroup launched a large transformation programme to simplify its systems and improve its data. In July 2024, regulators fined it a further 136 million dollars for insufficient progress, particularly on data management.

                  Root cause

                  A complex, fragmented technology and data landscape built up over decades of growth and acquisitions, with weak ownership and controls over data used for risk management and regulatory reporting.

                  The lesson

                  Supervisors now treat data governance and reporting quality as safety-and-soundness issues, with penalties comparable to those for misconduct. The ECB's 2024 guide on risk data aggregation and risk reporting sent a similar message to European banks. For corporate banking, the lesson is that the enablers — data, technology, reporting — are not back-office costs to minimise, but conditions for being allowed to do business.

                  Persistent pain points

                  Manual loan operations

                  Corporate loan operations still depend on reading notices and agreements, entering data by hand, and reconciling by email. Each manual step is a cost and a risk.

                  Regulatory reporting burden

                  Reporting requirements have multiplied — granular credit data, ESG disclosures, resolution data — often with overlapping but slightly different definitions. The European authorities' project for an integrated reporting framework, which aims to collect data once and derive many reports, is a long-term response.

                  Long, risky transformations

                  Platform and data programmes take years, cost large sums and often deliver less than promised. Many banks have several generations of incomplete programmes layered on top of each other.

                  Five transformation moves

                  Move What it means Benefit
                  Straight-through loan operations Structured notices, automated term extraction, event-driven processing Fewer errors, lower cost
                  Data as a product Owned, catalogued, quality-scored data domains with golden sources Trusted data for risk, finance and reporting
                  Integrated reporting Granular data from which many reports are derived Consistency, lower reporting cost
                  Modern platforms and cloud Fewer, modern product platforms; APIs; resilient infrastructure Agility, resilience
                  AI in operations and finance Reading documents, investigating breaks, drafting disclosures under control Productivity, with governance

                  AI in the back office

                  AI tools are entering the enablers: reading utilisation requests and agent notices to create bookings, extracting terms from facility agreements to check set-up, suggesting the causes of reconciliation breaks, explaining movements in provisions, and drafting parts of disclosures. Their value depends on the data they work with and on the controls around them: human review of outputs, model governance, and protection of confidential information.

                  Figure 6 Building the enabler platform in three phases: establish data ownership and golden sources, automate operations and reporting on top, then modernise platforms and add AI under control.

                  Work these through before reading on. Full solutions are in the Answer Key.

                  5.1 Overlay governance

                  In a sudden energy-price shock, the provisioning team proposes a 50 million euro overlay on energy-intensive corporate borrowers. List the elements of good governance for it.

                  A documented rationale for why models miss the risk; a defined scope of borrowers; a quantification method with sensitivity analysis; approval by the provisioning committee; disclosure where material; regular review with criteria for release; and auditor and supervisor access to the documentation.

                  5.2 Classify the risk

                  Name the risk: a loan system migration loses the effective interest rate history of 200 facilities; AnaCredit and FINREP report different stages for the same loans; the loan servicing vendor announces the end of support for the bank's version; an overlay is never released after the risk has passed.

                  Lost effective interest history: change and technology risk with accounting consequences. Different stages across reports: reporting and data quality risk. End of vendor support: third-party risk. Unreleased overlay: provisioning and model risk.

                  5.3 Data as a product

                  Define the "facility data product" for corporate banking: its owner, contents, consumers and quality measures.

                  Owner: the corporate bank, with loan operations as steward. Contents: facility identifier, borrower and group identifiers, terms, balances, schedules, rates, fees, collateral links, classification and stage. Consumers: risk, finance, treasury, reporting, agency, analytics. Quality measures: completeness, validity, reconciliation to the loan platform and ledger, timeliness, documented definitions.

                  5.4 Business case

                  A bank processes 120,000 loan events a year manually, at 20 minutes each. Automation would handle 70 per cent straight through. At 50 euros an hour, what is the annual saving? What other benefits and risks belong in the case?

                  Events automated: 84,000 × 20 minutes = 1,680,000 minutes = 28,000 hours; × 50 euros = 1.4 million euros a year. Other benefits: fewer errors and breaks, faster processing, better data for reporting. Risks: implementation cost and time, model and AI errors, dependence on vendors, need for controls and exception handling.

                  Explain to a relative why a large bank was fined hundreds of millions of dollars not for losing money or cheating clients, but for the state of its data.

                      1. Name the main enabler risks and a control for each.

                      2. Why did Covid challenge IFRS 9 provisioning, and how did banks respond?

                      3. What were the Citigroup consent orders about, and what do they signal?

                      4. Why are corporate loan operations still largely manual?

                      5. Name the five transformation moves.

                      6. What controls does AI in the back office need?

                      5

                      1. Why did Covid challenge IFRS 9 provisioning?

                      2. What do the Citigroup consent orders signal?

                      3. Which risk arises from relying on vendors and cloud providers?

                      Module Summary — The Six Ideas That Matter Most

                      If you retain only a handful of things, retain these. Each compresses a part of the module into a single, load-bearing idea.

                      Five enablers form one chain

                      Operations books and services, finance measures and provisions, data holds the facts, technology runs the platforms, reporting delivers the results. A weakness anywhere propagates downstream.

                      Loan operations turn agreements into events

                      Set-up, drawdowns, rate fixings, accruals, billing, rollovers, payments, reconciliations, amendments and collateral: each step must match the agreement, and set-up errors are silent and lasting.

                      IFRS 9 decides how loans appear in the accounts

                      Business model and SPPI tests set the classification; integral fees are spread through the effective interest rate; expected credit losses are provisioned in three stages, with judgement and overlays; forbearance and non-performing definitions drive regulatory views.

                      Data needs owners, golden sources and lineage

                      Client, hierarchy, contract, collateral, exposure and risk data must be owned by the functions that create them, held in golden sources, catalogued and traced — as BCBS 239 and supervisors expect.

                      One source, many reports

                      Management information, COREP, FINREP, AnaCredit and Pillar 3 must be built from the same reconciled data, with a single facility identifier making consistency automatic.

                      The plumbing is a licence to operate

                      The Covid provisioning test and the Citigroup consent orders show that judgement needs data and data needs ownership. Straight-through operations, data as a product, integrated reporting, modern platforms and controlled AI are the agenda.

                      Process Cheat Sheet — A Loan from Booking to Reporting

                      A single reference for the enabler chain. Everything below is developed in full in Part 2.

                      Step Lead Output Key control
                      Set-up and booking Operations, finance Facility in loan platform, classification Four-eyes against agreement and approval
                      Drawdown and fixing Operations Drawing, rate, payment Request validation, rate source
                      Accrual and accounting Operations, finance Daily accruals, effective interest Accrual reconciliation
                      Billing and payment Operations Notice, collection, rollover Cash reconciliation, break investigation
                      Monthly close Finance, risk Stages, ECL, overlays, profitability Model governance, provisioning committee
                      Data quality Data Validated, aggregated data Quality rules, lineage
                      Reporting Reporting MI, COREP, FINREP, AnaCredit, Pillar 3 Validation, cross-report reconciliation
                      Lifecycle events All Forbearance, NPE, modification, write-off Consistent updates across systems

                      Key numbers to remember

                      Item Rule or figure
                      IFRS 9 effective date 1 January 2018
                      Stage 1 provision 12-month expected credit loss
                      Stages 2 and 3 provision Lifetime expected credit loss
                      Non-performing 90 days past due or unlikely to pay
                      Common stage 2 backstop 30 days past due
                      AnaCredit reporting threshold Debtor commitment of at least 25,000 euros
                      AnaCredit start 2018
                      BCBS 239 principles Published 2013; G-SIB compliance expected from 2016
                      ECB guide on risk data aggregation 2024
                      DORA application January 2025
                      Citigroup penalties 400 million dollars (2020); 136 million dollars (2024)

                      The one-sentence version

                      Operations, finance, data, technology and reporting turn every corporate banking decision into a correctly serviced, accounted, provisioned and reported reality — and because supervisors, auditors and clients all depend on that reality, the bank's plumbing is not a cost to minimise but a condition of doing business.

                      Sources and further reading

                      Regulatory positions and market facts stated in this module reflect sources available in mid-2026. Treat every date and figure below as a fact with a shelf life and verify it before relying on it in client work.

                      How to read this section

                      Accounting and regulatory descriptions are simplified for learning; the standards, regulations and supervisors' guidance govern. Figures flagged as illustrative — Helion, Vela and all worked examples — are invented to teach a mechanic. For trade capture, settlement, reconciliation, reference data and regulatory reporting in markets businesses, see the corresponding CIB modules.

                      Primary instruments and standards referenced

                      Short name used in the module Instrument Relevance
                      IFRS 9 IFRS 9 Financial Instruments Classification, measurement, impairment
                      EBA forbearance and NPE definitions Implementing technical standards on supervisory reporting Forborne and non-performing exposures
                      Prudential backstop CRR provisions on minimum loss coverage for NPEs NPL provisioning
                      AnaCredit Regulation (EU) 2016/867 of the ECB Loan-level credit data
                      FINREP and COREP EU implementing technical standards on supervisory reporting Financial and prudential reporting
                      BCBS 239 Principles for effective risk data aggregation and risk reporting, 2013 Data governance
                      ECB RDARR guide Guide on effective risk data aggregation and risk reporting, 2024 Supervisory expectations
                      DORA Regulation (EU) 2022/2554 ICT risk and third parties
                      EBA moratoria guidelines EBA/GL/2020/02 Treatment of Covid moratoria

                      Dated facts and where they come from

                      IFRS 9 took effect on 1 January 2018. AnaCredit reporting began in 2018. The Basel Committee published BCBS 239 in January 2013. The ECB published its guide on risk data aggregation and risk reporting in May 2024. DORA has applied since 17 January 2025. Supervisory statements on IFRS 9 and Covid were issued in March and April 2020. The OCC and Federal Reserve consent orders against Citigroup were announced in October 2020, with a 400 million dollar OCC penalty; additional penalties of about 136 million dollars were announced in July 2024. Sources: official texts, supervisors' publications and announcements.

                      Where to go deeper

                      The IASB's IFRS 9 materials and the EBA's guidelines on credit risk management practices and expected credit losses explain provisioning in depth. The ECB's RDARR guide and the Basel Committee's progress reports on BCBS 239 describe supervisory expectations on data. The ECB's AnaCredit manuals describe each attribute. The CIB modules on reference data and regulatory reporting cover the wider reporting chain.

                      Closing the Corporate Banking track

                      This module closes the fifteen-module Corporate Banking track: coverage and the lifecycle; corporate and leveraged finance, restructuring, distribution, energy and real assets and agency; trade finance, receivables and supply chain finance, export finance, commodity finance and payments and cash management; and the enablers — treasury and ALM, risk, compliance and legal, and the operations, finance, data, technology and reporting machinery described here. Together with the thirty modules of the CIB series, they describe how a corporate and investment bank serves its clients and keeps itself safe.