Keeping Your Head Above Water

When a client experiences financial distress, especially SMEs and corporates, banks would normally consider avenues of rehabilitation prior to undertaking definitive recovery and liquidation activities. This can involve granting certain concessions, with the hope that these will aid rehabilitation and reduce potential credit losses. The difference between business-as-usual (BAU) and distressed restructures is the circumstances surrounding the restructure. If the borrower or loan in question has any indication of recent, ongoing or impending financial distress, even if repayments are up to date, then the restructure should be seen as distressed. Financial distress can nonetheless be ambiguous, and imperfect information on the bank's side can result in misclassification. The accuracy and objectivity of classification largely depend on the capabilities of the lender’s credit risk management framework.

D11 casts a wider net on the meaning of financial distress. The examples provided in the directive include irrefutable indicators of distress such as arrears, excess, covenant breaches, bankruptcy, liquidation, stock delisting and business rescue. But it also introduces factors that are more quantitative and predictive such as limit changes on revolving facilities, deterioration in risk rating (or credit score) in the past 12 months, and the high-risk watch list. This expansion makes financial distress somewhat similar to the IFRS 9 concept of significant increase in credit risk (SICR).  

Going forward, the identification of distressed exposures would therefore incorporate a combination of qualitative and quantitative factors. Banks should consider that using predictive distress factors can yield false positives as deterioration in credit quality does not unequivocally mean the borrower is in distress. Just like with the IFRS 9 SICR assessment, Directive 11 does not define a threshold of significance for the 12-month deterioration in credit score. This would result in varying interpretations across banks.  

The effectiveness of predictive factors in identifying distress would depend on the performance of underlying models. The capacity to detect an increase in risk is limited by the rank ordering ability of underlying scorecard(s).  

The new 1% relative materiality threshold for non-retail exposures is a precaution against inconsequential facilities contaminating an entire counterparty. This materiality threshold is similar to the 1% relative default materiality thresholds in place in the UK and EU. The exception here is that the D11/2025 measure considers both draw and undrawn exposures. 

For corporate entities, indicators of financial distress may be a lot more subtle, as precursory signs of financial difficulties can begin to appear in a company’s financials or sector long before distress crystallises. The servicing of debt and compliance with contractual covenants should not be regarded as contradictions to current or potential financial difficulties. A breach or near-breach to a covenant should be duly considered by the applicable credit committee. Even seemingly harmless concessions following a breach of covenant, e.g. failed semi-annual financial reporting, can be a sign of trouble brewing.

Reduced financial obligation and default classification 

Beyond financial distress, D11 provides further clarity on the concept of “reduced financial obligation”. There were different interpretations on the default classification of distressed exposures. According to D11, a distressed restructure should be classified as defaulted only if the concessions granted reduce the economic value of the instrument. Hence, if the modification has a neutral or positive effect on the carrying value of the loan(s) then there is no default. Essentially, banks should perform an impairment assessment on the borrower or facility. Some undisputable indicators of default for D11 restructures are: Reduced contractual rate; Partial write-off; Raising of a specific impairment; Reduced instalment without a commensurate term extension. 

The requirement for a reduced financial obligation means that undertaking a no-loss restructure can keep a borrower or facility from entering default e.g. a term extension without interest reduction, or a payment holiday with a term extension. The Risk function should establish stringent safeguards against the use and abuse of D11 restructures to mask credit risk deterioration. Ultimately, the primary purpose of distressed restructuring is to rehabilitate. Failure to rehabilitate after one or two attempts should probably mean the client is unlikely salvageable. 

In principle a client should only be rehabilitated once it has fully recovered from the circumstances which hampered their ability to service the debt obligation. At the very least, this recovery is presumed when the borrower can demonstrate a return to consistent repayment behaviour in line with the revised contract terms. D11 increases the minimum probation period from 6 to 12 months. Retail loans repayable monthly should observe at least 12 monthly instalments in a 12-month period to be eligible for curing. For corporate borrowers, rehabilitation is at the discretion of the relevant credit risk committees. A longer probation period is more stringent and should thereby result in more sticky defaults as there would be a lower incidence of premature rehabilitations. 

In anticipation of D11/2025, many banks in South Africa had already adopted a 12-month probation period. This extension of probation period should therefore have a notable effect on only some banks. D11 allows for exceptional cases for the rehabilitation of corporate exposures in less than 12 months, but not less than 6 months. This is at the discretion of the credit risk committee, with the support of tangible evidence. However, it would indeed be extraordinary for a corporate entity to fully recover from a notable credit risk inducing event less than a year after the debt was restructured. 

D11-compliant models are scheduled for phase 4 of the SARB’s credit risk roadmap from 2029 i.e. Directive 12 of 2025. Banks will have at least two full years from 1 January 2027 to 31 December 2028 to analyse their reported D11 population prior to refreshing their models. 

• Distress and the definition of default  

D11 has expanded financial distress beyond the obvious factors. Where suitable, banks will need to identify and test additional indicators of financial distress.  The definition of default (DoD) will need to be updated to incorporate the reduced financial obligation requirements under D11. The updated DoD will need to be applied retrospectively in the data over their entire modelling period.  

The performance of a distressed restructure does not necessarily result in the default of a client or account. However, due to an increase in risk distressed exposures should at the very least be classified as IFRS 9 stage 2 exposures while they remain in probation. 

• Update of the probation period  

Banks that are not yet on a 12-month probation period will need to update their credit risk management policies and retrospectively apply it in their data and models. Banks would also need to ensure that rehabilitation requirements, whether quantitative or judgmental, are correctly implemented. This includes the requirement for at least 12 instalments in 12 months for monthly repayment contracts, and the effect of default and subsequent modifications while in probation. 

• Refresh of models

All capital and impairment models will need to be redeveloped to align with the amended DoD and curing rules.  

Banks will require more granular and better-connected data on loan modifications, borrower financial health, restructuring triggers and repayment performance. This includes consistent Days-Past-Due (DPD) and Unlikeliness-to Pay (UTP) indicators, original and amended cash flows, behavioural and risk rating histories, and cure and re-default events. On commercial and corporate portfolios, banks must also aggregate on- and off-balance-sheet exposures at counterparty level to apply the 1% materiality threshold.

In addition to needing to cater for new data points, the historical data will also require re-labelling. This means that all portfolios will need to be reviewed to determine which assets require retrospective reclassification as distressed given the new D11 distress indicators. Therefore, a revised default dataset needs to be created, catering for both forward-looking as well as backward-looking (historical) assets. These new default datasets across the bank’s portfolios will need to be tested, validated and the bank’s Data team will need to be able to defend that the logic has been consistently applied to the re-labelling. The reconstructed or relabelled datasets will support the subsequent redevelopment and validation of Probability of Default (PD), Loss Given Default (LGD) and Exposure at Default (EAD) models.

For some banks, implementation may require system redesign. Credit risk and retail banking systems will need to capture the expanded early-warning indicators, automate relevant DPD and UTP triggers, and apply consistent restructure, default, cure and re-default logic across portfolios. Core banking, credit workflow, impairment and regulatory reporting platforms must also maintain separate but connected indicators for arrears, distressed restructuring, probation, regulatory default and BAU classification. Legacy system limitations and fragmented DPD histories are likely to increase the scale of integration and remediation required. 

A key element of D11 is the consistent restructure and default definitions across all portfolios. This requires system-wide consistency, specifically, requiring automation, integration and tracking of these indicators, including UTP and DPD, across the various systems. This may pose a key challenge for legacy systems.

Given the wide-ranging effect of the directive on credit risk processes as a whole, it is imperative that banks update the relevant credit policies and frameworks to ensure alignment with the new requirements. At the top of the list of policies which need updating we have the policies on (i) restructuring and forbearance, (ii) arrears and default management, as well as those on (iii) IFRS 9 impairments and regulatory capital. 

Banks will need to combine automated triggers with documented expert judgement when assessing financial distress. Restructuring processes must incorporate clearer and more objective decision criteria, evidence of repayment capacity and controlled approval authorities. Probation and rehabilitation will require active monitoring, while repeat restructures and re-defaults must be identified and escalated consistently. For Internal Ratings-Based (IRB) banks, revised default datasets and models will also need to pass the scrutiny of model technical committees, independent validation and other governance processes before the relevant model changes are submitted to the PA for approval.