Summary
The official record supports serious institutional criticism, not the broadest allegation. Reviews identified widespread inappropriate treatment and damaging conduct. The FCA also said it found no evidence of a general practice of artificially distressing and transferring otherwise viable SMEs to profit from restructuring or insolvency.
Distressed lending requires a declared objective. A recovery unit can protect the bank and try to restore a borrower, but those objectives can conflict. Transfer, pricing, covenant, valuation and exit decisions need evidence showing which objective controlled each action.
Profit-centre incentives amplified conduct risk. Income targets and complex fees can make a vulnerable borrower valuable to the unit managing its distress. Governance must prevent revenue from becoming the unstated measure of turnaround success.
Valuation and asset-acquisition conflicts require independence. When a bank unit can influence a valuation and a related vehicle can acquire property, the borrower needs transparent instructions, independent evidence and a documented conflicts decision.
The regulatory perimeter limited enforcement. Much commercial lending was outside conduct regulation during the relevant period. Limited power to discipline is not a finding that treatment was acceptable, and a conduct failure is not automatically a regulatory offence.
Durable repair needs outcome proof. A bank should be able to replay transfer rationale, options, communication, fees, valuations, conflict decisions, complaints, refunds, direct and consequential loss, exit and the survival or orderly failure of the business.
The final FCA account requires two conclusions to remain together
The FCA's announcement of its final GRG report states that the unit clearly fell short of the standards customers expected and that the impact damaged affected customers and trust in banking. It also says the FCA found no evidence that RBS artificially distressed and transferred otherwise viable businesses to GRG to profit from their restructuring or insolvency. Both conclusions are part of the official record.
Removing either conclusion distorts accountability. Repeating only the most serious allegation overstates what the investigation established. Repeating only the rejected general-practice allegation understates the inappropriate treatment that reviews did find. A reliable analysis distinguishes a system's harmful operation from a specific theory of deliberate targeting.
The same discipline applies to individual responsibility. The FCA considered whether action could be taken against senior management or RBS within its powers. It concluded that disciplinary powers did not apply to much of the conduct and that a fitness-and-propriety action would not have reasonable prospects of success. That is a legal and evidentiary outcome, not a certificate that every decision was good or every customer was treated fairly.
Institutional repair can proceed without converting criticism into personal guilt. The bank can redesign incentives, valuation controls, complaints and governance because the system produced unacceptable outcomes. Individual action, if considered, requires a separate record of role, knowledge, authority, conduct and applicable law. That separation protects fairness while preventing the limits of enforcement from becoming an excuse for inaction.
The core control lesson is therefore balanced but not weak: a distressed-business unit can cause widespread harm through incentives, discretion and poor controls even if investigators do not establish a general strategy to manufacture distress.
GRG's operating model combined objectives that needed explicit priority
The FCA's full report on further investigative steps describes GRG's evolution, governance, objectives, senior-management questions and the limits of possible action. It is the regulator's final extended account of its investigation, distinct from the earlier skilled-person review and from individual customer litigation.
A recovery unit normally has several objectives. It protects the lender from loss, preserves capital, enforces contracts, identifies viable turnaround options and exits exposures that cannot be restored. Those goals can align, but they can also conflict. A rapid asset sale may reduce the bank's exposure while destroying a business that could recover. Extending credit may preserve the enterprise while increasing loss if assumptions fail. Charging for risk may be legitimate while making recovery less likely.
Governance must specify how such conflicts are decided. A transfer paper should identify the customer's condition, covenant position, liquidity, business plan, collateral, alternatives and the reason specialist management is needed. It should state whether the primary objective is turnaround, consensual restructuring, exposure reduction or enforcement. Changing that objective later should require a new decision and communication.
The unit's scorecard should not allow bank income to stand in for customer recovery. Metrics need to include businesses returned to mainstream banking, sustainable restructurings, orderly exits, time, realised loss, customer communication, complaints and exceptions. Revenue may be monitored, but it should not dominate staff evaluation where the unit controls vulnerable customers.
Board oversight should compare stated purpose with outcomes. If only a small share of customers return to ordinary banking, the board should ask whether the unit is really a turnaround function. If income rises as customer viability falls, it should examine fees, margins and conflicts. A name cannot govern a function; evidence must.
The skilled-person review found material problems across a broad sample
The FCA's 2017 update on the independent review explained that Promontory, assisted by Mazars, reviewed RBS's treatment of SME customers transferred to GRG between 2008 and 2013. It reported that the most serious allegations were not upheld but identified other serious concerns and led to voluntary fee refunds and a complaints process.
Sampling is central to interpretation. A skilled-person review can identify patterns and evaluate whether practice matched policy. It does not adjudicate every customer case. Findings about a significant proportion of reviewed cases should not be converted into a claim that every transfer or fee was improper. Conversely, individual complaint outcomes cannot be used to dismiss system-level patterns.
The review design should be replayable. The population, sampling method, thematic tests, file completeness, customer input and judgment criteria need to be clear. Distressed-business files are especially difficult because viability is uncertain and records may span relationship management, credit, valuation, legal and recovery systems. Missing data should be treated as a control weakness, not filled with confident hindsight.
For the bank, the review points to a permanent assurance model. Each year, an independent function should sample transfers, fee changes, valuations, forbearance, security enforcement, asset sales and exits. It should compare policy with actual communication and customer outcome. High-risk themes should be tested across the full population using data, then reviewed in files.
The board should see both prevalence and severity. A rare conflict that destroys a viable business may deserve urgent action. A common communication failure may require system redesign even if direct loss is small. Assurance becomes useful when it leads to named remediation, deadlines and retesting rather than a one-time report that sits outside ordinary governance.
The detailed review preserves important boundaries around the strongest allegations
The published interim summary of the independent review describes GRG's practices, the review methodology and findings across transfer, turnaround, pricing, communication, West Register and governance. It records widespread inappropriate treatment in a significant proportion of cases while distinguishing those findings from allegations the review did not substantiate as general or systematic practice.
That distinction should shape control design. If the problem were only rogue acts, targeted discipline might be enough. If the problem were an explicit strategy to destroy viable businesses, the response would focus on senior intent and policy. The record instead supports a more institutional explanation: broad discretion, income pressure, weak objectives, poor communication, inadequate oversight and conflicts combined to create harmful treatment.
Institutional failures can be harder to repair because no single prohibited instruction needs to exist. Staff may follow local incentives, use available pricing tools and pursue exposure reduction, while no control tests the combined effect on the customer. The absence of a written strategy to cause harm does not make the resulting system acceptable.
A bank should therefore test cumulative burden. A customer may face a higher margin, review fee, valuation cost, professional adviser cost, reduced facility and additional security in the same period. Each item may have an individual rationale, yet together they can make turnaround impossible. Decision papers should show the combined cash impact and the alternative path considered.
The bank should also preserve counterfactual uncertainty. It may be impossible to know whether a distressed business would have survived under different treatment. Complaint and redress decisions should distinguish documented direct loss, lost-opportunity claims and speculative future value. Honest uncertainty protects both customer and bank while still allowing compensation where evidence supports it.
Parliamentary publication made transparency itself an accountability issue
The Treasury Committee's publication of the unredacted section 166 report followed dispute over whether and how the report could be made public. The Committee argued that public interest in the findings outweighed the normal confidentiality surrounding skilled-person reports. Its statements are parliamentary positions; the underlying report and FCA investigation retain their own status.
Transparency has competing purposes. It enables customers to understand systemic findings, allows Parliament to scrutinise the regulator and bank, and prevents selective quotation. But reports may contain confidential information, personal data and disputed claims. Publication governance needs a lawful route, representations where required, redaction principles and a stable authentic version.
The episode shows why firms should not depend on confidentiality to manage reputation. A control record should be written on the assumption that an independent reviewer may later examine it. That does not mean stripping candid analysis from minutes. It means avoiding language and incentives that cannot be defended as serving a legitimate purpose.
Regulators also need a publication policy for exceptional reviews. It should distinguish a skilled person's observations, the regulator's adopted conclusions and later enforcement. Summaries must preserve material adverse and exculpatory points. If a fuller report cannot be published, the reasons and validation process should be explained.
For customers, authenticity matters. Leaked fragments and competing summaries create confusion about what was actually found. A canonical document with version, scope, boundaries and links reduces misinformation. Transparency is therefore not just disclosure after a scandal. It is a control over how evidence enters the public accountability process.
SME finance policy exposed the gap between commercial freedom and customer protection
The Treasury Committee's SME Finance report placed GRG within a wider examination of lending, disputes, the regulatory perimeter and access to redress. Parliamentary conclusions can recommend policy and describe evidence but do not replace courts, regulators or individual complaint processes.
Commercial lending involves negotiation between parties with different sophistication and bargaining power. Large companies can obtain advice and diversify banks; smaller firms may depend on one lender for working capital, payment services and property security. When a business enters distress, switching becomes harder and the bank's leverage increases.
The regulatory perimeter should therefore be explicit. A customer needs to know which activities are regulated, which complaints can go to the Financial Ombudsman, which contractual rights remain and what voluntary review exists. Bank staff should not imply that the same protections cover every product and borrower if they do not.
Perimeter gaps create governance risk even where conduct is legal. A bank's board can set a customer standard that applies to all SME activity, including unregulated lending. That standard can require clear communication, reasoned decisions, conflict controls and complaint review. Internal standards should not be presented as statutory rights, but they should be auditable and enforceable within the firm.
Policy makers should measure outcomes before expanding law. They need data on business size, complaint type, loss, court accessibility, lender concentration and the effect of new duties on credit availability. Protection that exists only on paper or makes distressed credit unavailable may not help SMEs. The GRG record makes the gap visible; durable reform requires a design that works during actual financial stress.
Earlier parliamentary work showed that the concerns predated the final review
The House of Commons Conduct and competition in SME lending report discussed GRG, the Tomlinson allegations, bank responses and wider SME lending practices. It demonstrates that questions about turnaround purpose, profit and customer treatment were present years before the final FCA account.
Early signals need a governed intake process. A regulator or bank may receive complaints, parliamentary evidence, whistleblowing, litigation and media reports of different quality. Dismissing them because some allegations are overstated loses useful evidence. Accepting them wholesale prejudges the case. The control response is triage: define the allegation, identify available data, look for patterns and preserve contrary evidence.
Boards should see signal convergence. A single complaint may be idiosyncratic. Repeated claims about transfer, unexplained fees, low valuations or related acquisitions across regions should trigger thematic review. The escalation threshold should consider severity and common mechanism, not only the number of upheld complaints.
Management responses also need verification. A statement that the unit turns around businesses should be supported by a definition, population and outcome data. A statement that it is not a profit centre should reconcile internal budgets, revenue allocation and performance measures. Assertions become controls only when independently testable.
The time between early warning and final review matters. During delay, more customers can be affected and records can disappear. A preliminary protective response may be appropriate before every allegation is resolved: pause a fee, require independent valuation, strengthen approval or widen complaint access. Such action need not imply guilt. It acknowledges uncertainty while reducing the chance of repeated harm.
Oral evidence revealed how definitions can obscure accountability
The Treasury Committee's GRG inquiry publication record brings together correspondence, oral evidence and written submissions. It is a primary accountability record of what witnesses and institutions told Parliament, but testimony remains testimony and should be attributed.
Terms such as “turnaround,” “viable,” “distressed,” “material financial distress” and “profit centre” can change the apparent conclusion. A business may leave GRG through repayment, sale, insolvency or return to mainstream banking; counting all exits as turnaround would mislead. A bank can allocate income to a unit without calling it a profit centre. A borrower can be distressed yet still capable of recovery.
Governance needs a controlled dictionary. Each outcome metric should have a definition, denominator and date. Viability assessments should identify the forecast horizon, financing assumptions and decision-maker. Transfer criteria should distinguish covenant breach, expected default, information failure and strategic risk. Exceptions should be visible rather than absorbed into professional judgment.
Parliamentary correspondence also demonstrates the importance of correction. If a witness later discovers that evidence was incomplete or inaccurate, the correction should be prompt, specific and linked to the original. Firms should apply the same rule to board information and customer letters.
Definitions are not a semantic side issue. They determine which customers enter a high-control unit, how success is reported and whether complaints are recognised. A bank that cannot reproduce its definitions across systems will struggle to prove fair treatment even where staff acted in good faith.
The hearing record exposed both poor conduct and limits on inference
In oral evidence on 30 January 2018, the skilled-person reviewer and RBS leaders were questioned about widespread inappropriate treatment, transfer decisions, material financial distress, internal income language, governance and the difference between a general practice and individual concerning cases. The hearing includes strong political characterisations, witness concessions and disputed interpretations.
The evidence is valuable because it tests institutional explanations in public. It is not a judicial finding. A committee member's question may contain an allegation. A witness's answer may accept, reject or qualify it. Analysis should not quote the premise as though the witness admitted it.
The hearing reinforces the need for case-level evidence. Reviewers examined customer files and, where possible, customer input. They encountered fragmented systems and had to triangulate information. That is itself a governance failure: the bank should be able to assemble the complete decision history for a distressed customer without a forensic reconstruction.
A canonical customer record should connect relationship notes, credit papers, valuations, facility changes, fees, legal steps, security, complaints and communications. It should preserve the state available at the time, not only later summaries. Access should be controlled, but relevant functions should not maintain incompatible versions.
Public questioning also shows why culture cannot be measured only through values statements. Internal phrases focused on hitting budget or generating income can reveal the practical message received by staff. Boards should sample working documents, training, incentives and communications, then compare them with the formal purpose. Culture is the pattern of choices the system rewards.
Fee refunds addressed one mechanism of harm but not every loss claim
RBS's background information on the GRG complaints process describes automatic refunds of complex fees, interest, the scope of eligible customers, complaint volumes and the later closure of the special process. It is a bank account of its voluntary programme, not an independent adjudication of every claim.
Automatic refund can be effective where a fee category was not properly communicated and individual causation would make review slow. It reduces burden and provides consistent treatment. But refunding a fee does not necessarily address margin increases, lost opportunity, business failure, professional costs or distress. Each category needs its own legal and evidentiary basis.
The programme should reconcile population to payment. Who was eligible? How were dissolved companies, former directors and unreachable customers handled? Were refunds offset against due debt? How was interest calculated? What remained unpaid? Aggregate offers should be distinguished from accepted and delivered payments.
The bank should also analyse why the fee existed. If a complex instrument compensated for risk, did removing it cause the same economics to appear through another charge? If communication was the failure, were customer letters and staff training corrected? If incentives drove use, did scorecards change? Redress without root-cause repair can move the problem.
Complaint access after closure needs clarity. Customers may still use ordinary channels, but they may lose the special process or independent appeal. Communications should state that distinction and any time limits. A closing date should not erase unresolved exceptions or the bank's obligation to retain records.
Independent appeal can improve legitimacy only if its remit is clear
RBS's description of the Independent Third Party appeal process explains assurance, appeal review, direct-loss decisions and reports. The arrangement created an additional layer outside the bank's initial complaint team, but it remained a defined voluntary mechanism with eligibility and scope limits.
Independence depends on appointment, terms, information access, remuneration and freedom to reach decisions. A retired judge or external reviewer adds credibility, but title alone is not enough. The process should publish its methodology, service standards, overturn rates, themes and recommendations without exposing customer information.
The appeal body needs the complete file and the customer's evidence. It should identify what standard it applies: contractual obligation, bank policy, fair-and-reasonable treatment or a bespoke scheme rule. A customer should know whether consequential loss can be considered and how causation will be assessed.
Appeal outcomes should feed systemic remediation. If many decisions are overturned for the same reason, the bank should re-review similar cases rather than wait for each customer to appeal. The independent party should be able to raise thematic concerns with the board and regulator.
The design should also avoid false finality. A voluntary appeal may not remove court rights unless the customer agrees to a release. It may sit alongside ordinary complaints or the Ombudsman for eligible cases. Clear route maps prevent customers from missing a remedy because they assumed one process suspended another.
Legitimacy comes from demonstrable fairness, not merely an extra stage. The bank should prove that eligible customers knew about the process, could participate without prohibitive cost, received reasoned outcomes and obtained payment when awarded.
Progress reporting must distinguish cases, decisions, offers and cash paid
RBS's monthly GRG complaints progress page was designed to report complaint and consequential-loss progress. Such reporting is essential in a large remediation, but metric definitions determine whether it informs or reassures.
A complaint “concluded” may be upheld, partially upheld, rejected or withdrawn. An “offer” may be disputed or unpaid. A direct-loss payment may not resolve a consequential-loss claim. Appeals may reopen outcomes. The dashboard should show movement through each state and the age of cases.
The population denominator also matters. Complaint count compared with eligible customers can show participation, but low participation may reflect satisfaction, lack of awareness, business dissolution, cost or distrust. The bank should test contact success and reasons for non-response rather than assume silence means no harm.
Quality metrics should accompany throughput. File completeness, rework, appeal overturn, calculation errors, missed service standards and customer feedback reveal whether speed is being bought at the expense of fairness. Independent assurance should sample both upheld and rejected cases.
Completion requires an exceptions inventory. Unreachable customers, dissolved companies, disputed authority, missing documents and complex causation can remain after ordinary cases close. The board should approve how each class is resolved and publish aggregate residual value where lawful.
Transparent progress reporting also protects staff. It makes workload and dependencies visible, reducing pressure to classify unresolved cases as complete. A remediation should end because entitlements are paid and routes exhausted, not because a target date arrived.
Public debate after the FCA report preserved the question of reform
The House of Lords Library's briefing on the FCA report assembled material for parliamentary debate on the final investigation and measures to prevent future mistreatment. A library briefing synthesises official and public material; it is not a new finding against the bank or a determination of a customer's case.
Its value is institutional memory. Conduct events often lose urgency once enforcement closes and leadership changes. A consolidated public record helps future policy makers see which problems remained: perimeter limits, SME bargaining power, complaint access, senior accountability and confidence in business lending.
Banks should maintain the same memory internally. Remediation decisions, discontinued controls and residual risks need a durable archive. New leaders should understand why a fee was prohibited, why valuation independence was strengthened or why a transfer committee exists. Otherwise, commercial pressure can gradually recreate the old model under new names.
Policy review should use outcome data. Has Ombudsman access increased? Are SMEs receiving reasoned decisions? Do recovery units return viable firms to mainstream banking? Are valuation and asset-acquisition conflicts declining? Have complaints shifted into products outside the revised perimeter? Reform should be tested rather than announced once.
Institutional memory also protects proportionality. Not every distressed-lending decision should be second-guessed because GRG existed. Lenders need to manage risk and enforce agreements. The lesson is to make those powers explainable, independently challenged and accessible to review where customer vulnerability and conflicts are high.
The 2018 enforcement update shows that jurisdiction and merits are different
The FCA's July 2018 statement on further investigative steps explained why misconduct discipline was difficult and why a fitness-and-propriety case did not have reasonable prospects. It also noted that later accountability frameworks, including the Senior Managers Regime, changed the position for bank activities.
Jurisdiction asks whether the regulator has power over the activity, firm, person and period. Merits ask what happened and whether it breached the applicable standard. A strong factual criticism can coexist with limited enforcement power. Conversely, jurisdiction does not guarantee that evidence supports action.
Boards should not outsource ethics to the perimeter. An activity can be unregulated but still create legal, credit, reputation and customer risks. Enterprise conduct standards should cover the bank as a whole, with controls proportionate to customer vulnerability and discretion.
Regulators should make perimeter limits visible early. Customers and Parliament may otherwise interpret a closed case as exoneration. A decision statement should separate findings, statutory constraints, evidence and policy implications. The GRG statements largely attempted that distinction, though public debate shows how difficult it is to communicate.
Future accountability frameworks should also avoid retroactivity. A responsibility regime introduced later can improve role clarity but cannot simply be applied to earlier conduct. The lesson is prospective: assign senior managers clear areas, require reasonable steps and retain evidence. It is not a shortcut to liability for the past.
Publication controversy revealed the governance of regulatory evidence
The FCA's December 2017 statement on publication of the GRG report addressed speculation about why the full skilled-person report had not been published and the constraints affecting disclosure. The subsequent parliamentary publication demonstrates that evidence governance can become an accountability issue separate from the underlying conduct.
Regulators receive confidential information to supervise effectively. Firms and customers need assurance that data will not be disclosed arbitrarily. At the same time, secrecy can prevent affected people from understanding a systemic review. The balance needs rules rather than improvised negotiation.
A publication decision should inventory legal restrictions, personal data, third-party fairness, market sensitivity and public interest. Independent counsel or review may validate that a summary is balanced. Redactions should be principled and versioned. If consent is sought, the regulator should explain whose consent is legally relevant and what happens if it is withheld.
The regulator also needs consistent records of its own governance. Committee papers, decisions and reasons should be retained. Freedom-of-information disclosure can reveal internal debate; that should be expected in a public authority. Candid challenge is not evidence of failure, but unexplained inconsistency can damage trust.
For banks, the parallel lesson is to design remediation reporting that can withstand later disclosure. Claims about volumes, payments and reform should reconcile to data. Legal review should protect rights without stripping the report of meaning. Trust is strengthened when adverse findings and limiting facts appear together.
Expanded Ombudsman access improved the route for later SME disputes
The Financial Ombudsman Service's eligibility guidance for small businesses explains which micro-enterprises, small businesses, charities, trusts and guarantors can use the service and notes the date limits for expanded small-business jurisdiction. The guidance is current route information, not a retrospective remedy for every GRG-era dispute.
Access to an informal adjudicator can reduce the cost imbalance between an SME and a bank. Court litigation may be too expensive after a business has failed. The Ombudsman can consider what is fair and reasonable within its jurisdiction and award limits. But eligibility, time and act dates remain decisive.
Banks should tell customers about applicable routes at the point of complaint, using the rules then in force. They should not assume a business is too sophisticated or too large without checking. Personal guarantors may have rights distinct from the company. Route decisions should be documented and quality-assured.
Expanded jurisdiction also generates learning data. Complaint themes can expose unclear transfer criteria, valuation disputes, forbearance communication and service failures across banks. Regulators and firms should use anonymised patterns while preserving adjudicator independence.
The route is not a substitute for internal control. An SME should not need external adjudication to receive a reasoned valuation or understand a fee. Ombudsman outcomes are a backstop and feedback mechanism. The strongest proof of reform is fewer repeatable failures, not merely a larger system for processing them.
Senior responsibility improves accountability only when authority and evidence align
The FCA's policy statement on the Duty of Responsibility explains how senior-manager responsibility interacts with a firm's contravention and the reasonable steps expected of the responsible manager. It is relevant prospective context, not a basis for retroactively deciding GRG-era individual cases.
Responsibility maps can solve one GRG lesson: collective structures should not make it impossible to identify who owned customer treatment in a recovery unit. The named manager needs actual authority, resources, information and escalation access. Assigning a title without those capabilities creates paper accountability.
Reasonable-steps evidence should be contemporaneous. It can include risk assessments, management information, challenge, resource decisions, remediation, audit findings and escalation. The record should also show obstacles and how they were addressed. A later narrative assembled for investigation is less reliable.
Boards need to test the map against real decisions. Who approved transfer policy? Who owned pricing? Who could stop a conflicted acquisition? Who monitored complaints? Who challenged outcome data? If answers move between committees, the map needs repair.
Senior accountability should not discourage delegation or candid debate. Managers can rely on competent teams if they define expectations, receive evidence and act on warning. The objective is not omniscience; it is a defensible control system with visible ownership.
State ownership and restructuring context did not remove customer-level duties
The government's review of RBS and the case for a bad bank addressed balance-sheet repair, asset reduction and the bank's role in supporting the economy. This macro-level policy context helps explain pressure to reduce risk after rescue. It does not determine whether any GRG customer was treated appropriately.
System stability and customer fairness can pull in different directions. A bank supported by taxpayers may need to shrink exposures rapidly, while government also wants lending to viable businesses. Those goals should be explicit in strategy and translated into controls, not left for front-line recovery staff to reconcile through income targets.
Ownership creates another interface. The state as shareholder should not direct individual credit decisions, but it can set high-level expectations and monitor strategy. Regulators retain independent roles. The bank's board remains responsible for conduct and risk. Clear boundaries prevent both political interference and accountability gaps.
Macro targets should be tested for customer effects. A rapid reduction in a portfolio may increase transfers, valuations, asset sales and complaints. Board papers should forecast that impact, allocate resources and define guardrails. If the strategy changes, customer-facing units need updated objectives rather than informal pressure.
The context can explain why exposure reduction mattered; it cannot excuse unclear fees, conflicts or poor communication. Accountability is strongest when it acknowledges the bank's genuine capital challenge and still asks whether the chosen customer-level actions were evidence-based, proportionate and reviewable.
A control model for distressed-business banking
A defensible recovery unit begins with purpose. It defines turnaround, exposure reduction, enforcement and exit, and states which objective governs each case. Transfer criteria use documented financial and behavioural evidence, with independent approval for exceptions. Customers receive a clear explanation of status, decision rights, information requirements and review routes.
Pricing controls show contractual basis, risk rationale, cumulative cash effect and approval. Complex or unusual fees receive enhanced communication and a conduct check. Valuations use independent instruction, appropriate comparables, documented assumptions and a route to challenge. A bank-related property acquisition is subject to a separate conflicts committee and evidence that the process did not exploit information or influence.
Forbearance decisions compare realistic alternatives and record why support is extended, changed or withdrawn. Personal guarantees, equity and participation instruments receive suitability and proportionality review. Staff scorecards balance loss management with sustainable customer outcomes, conduct and complaint quality. No local income document should contradict the formal purpose.
Complaints are identified consistently, including expressions of dissatisfaction embedded in relationship correspondence. Direct and consequential loss are separated, with causal standards and evidence. Appeals are independent within their remit. Population-level redress is used where a common mechanism makes case-by-case proof inefficient.
Boards receive outcome data with definitions: transfer reasons, returns to mainstream, insolvencies, sales, pricing changes, valuation disputes, related acquisitions, complaints, overturns, payments and aged exceptions. Independent assurance samples full case histories and tests cumulative burden. Supervisors examine interfaces and make perimeter limits clear.
The replay test is practical. Can an independent reviewer reconstruct the customer's condition, options, every material price and valuation, conflicts, communication, complaint and exit using records available at the time? Can the reviewer identify who decided and whether challenge changed the outcome? If not, recovery still depends too heavily on institutional power and retrospective explanation.
Outcome testing should extend beyond the bank's own closure date. A business returned to ordinary relationship management may fail soon afterwards because the restructuring left unsustainable debt or depleted working capital. A business sold or refinanced may survive while owners bear losses that were nevertheless contractually justified. The control record should therefore track a defined post-exit period and separate business survival, customer fairness and bank recovery rather than compressing them into one success rate.
Data quality must be tested at the same time. Transfer dates, fees, valuations, complaint status and exit reasons should reconcile across credit, finance and customer systems. Unexplained gaps should be reported as exceptions. If management changes a definition, historic metrics should be restated or clearly bridged so that apparent improvement is not produced by classification.
Finally, boards should compare recovery units across regions and products. A redesigned GRG may disappear by name while similar discretion migrates into special-assets, restructuring or intensive-care teams. Common minimum controls, cross-unit assurance and a central conflicts inventory prevent reform from being limited to one historic label. The objective is a bank-wide capacity to manage distress fairly, not a compliant successor unit surrounded by unexamined equivalents.
Conclusion
GRG's accountability lesson is not that banks should never enforce contracts or manage distressed exposures. Recovery decisions are necessary and often difficult. The failure was that wide discretion, competing objectives, income pressure, complex pricing, valuation and acquisition conflicts, fragmented records and weak complaint recognition combined in a unit dealing with customers at their least mobile and most vulnerable.
The official record must remain precise. It supports serious and widespread institutional criticism. It does not support the claim that investigators found a general practice of deliberately manufacturing distress in otherwise viable SMEs for profit. Limited regulatory power does not erase poor treatment, and poor treatment does not automatically establish a regulatory offence or individual guilt.
Durable reform is visible in decisions, not promises. A bank should prove why a customer entered specialist recovery, how each action supported the declared objective, how conflicts were controlled and how complaints and redress reached completion. Only then can a turnaround function protect the lender without making customer distress a source of ungoverned value.