A collaboration between SAS, Man Group, Pension Insurance Corporation, and Stanford University has developed a pioneering machine learning-based model that significantly enhances the forecasting of corporate credit rating movements, providing investors with crucial early warning signals amid rising market uncertainties.
A collaborative effort between SAS, Man Group, Pension Insurance Corporation, and Stanford University has resulted in a pioneering machine learning-based corporate credit risk model that significantly improves the forecasting of credit rating changes. The model offers early warning signals for credit rating upgrades and downgrades, aiming to provide investors and portfolio managers with a valuable lead time to adjust before such changes are fully priced by markets or updated by rating agencies.
By harnessing more than 20 years of historical data, the model integrates KRIS default probability metrics alongside bond spreads, yields, equity performance indicators, and macroeconomic variables. This breadth of data allows the model to effectively predict the likelihood of a company’s credit rating moving up, down, or remaining stable. Backtesting results show that this approach outperforms traditional risk assessment tools, ranking firms more accurately by their future rating changes.
The development team emphasises the model’s proactive role in risk management, especially for investors regulated to maintain portfolios of investment-grade securities. Early identification of rating transitions is crucial because downgrades below investment grade can force institutional investors to divest affected bonds, potentially triggering price drops. The model’s capability to flag such risks before market prices reflect them is expected to provide greater flexibility and strategic advantage in portfolio management.
“The breakthrough approach of our new model shows that investors can do a lot better than the current best practices,” said Stas Melnikov, Head of Quantitative Research and Risk Data Solutions at SAS. “The model’s early-warning signals give them critical time to act before the market fully prices in the event, helping better manage risk, reduce losses, and seize opportunities.”
The timing of this model’s introduction is noteworthy, coinciding with a period of sustained higher borrowing costs and pressure on corporate credit profiles. While the Bank of England’s interest rates have been steady at 4%, refinancing remains costly for companies, intensifying default risks. The model could serve as a key tool in forecasting such stress before it translates into market-wide asset valuation adjustments.
The dataset underpinning the model is extensive, comprising over 500,000 credit event records from 2001 to 2024. Notably, KRIS one-year default probability emerged as a highly influential input, ranking just behind option-adjusted spreads and yield to maturity in predictive importance. This aligns with broader academic and industry research underscoring the value of advanced machine learning techniques, such as artificial neural networks, gradient boosting, and ensemble models, in surpassing traditional logistic regression and support vector machines for credit risk forecasting.
Industry leaders acknowledge the model’s potential impact. Steven Desmyter, President of Man Group, remarked that the new approach significantly outperformed existing methods, providing clearer prediction of which companies face upgrade or downgrade risks. The model is expected to be integrated with SAS’s wider risk management platforms, augmenting capabilities in asset-liability management, credit risk oversight, and the provisioning of expected credit losses.
Independent research reports further corroborate the effectiveness of machine learning in credit rating transitions. Studies published in journals such as MDPI and Risk.net highlight that ensemble methods like XGBoost and random forests consistently enhance accuracy and reliability, improving financial institutions’ ability to assess credit risk under varying market conditions.
Ultimately, this machine learning model represents a crucial advancement in credit risk prediction, with the potential to deliver material benefits for institutional investors operating under regulatory constraints and seeking to optimise capital allocation amidst uncertain credit markets. The model’s forward-looking insights may prove essential in navigating an environment marked by volatile borrowing costs and credit stresses, equipping market participants to better anticipate and respond to shifts in corporate creditworthiness.
- https://itbrief.co.nz/story/machine-learning-model-gives-early-warning-on-credit-rating-shifts – Please view link – unable to able to access data
- https://www.sas.com/en_us/news/press-releases/2025/november/model-predicts-credit-rating-changes.html – SAS, in collaboration with Man Group, Pension Insurance Corporation, and Stanford University, has developed a machine learning model that accurately forecasts corporate credit rating changes. By analysing over two decades of data, including KRIS default probability metrics, bond spreads, yields, equity performance indicators, and macroeconomic variables, the model provides early warning signals for credit rating upgrades and downgrades. This proactive approach enables investors and portfolio managers to act before such changes are fully priced by markets and reflected by ratings agencies, thereby enhancing risk management and capital allocation decisions.
- https://www.prnewswire.com/news-releases/new-risk-model-sets-benchmark-for-accuracy-in-predicting-credit-rating-changes-302606447.html – A new machine learning-based corporate credit risk model, developed by SAS, Man Group, Pension Insurance Corporation, and Stanford University, offers early warning indicators for credit rating upgrades and downgrades. By leveraging over twenty years of historical data, including KRIS default probability metrics, bond spreads, yields, equity performance indicators, and macroeconomic variables, the model enables investors and portfolio managers to act before such changes are fully priced by markets and reflected by ratings agencies, thereby improving risk management and capital allocation decisions.
- https://blogs.sas.com/content/sascom/2025/10/24/what-happens-when-machine-learning-meets-credit-ratings/ – SAS has developed a machine learning model that accurately forecasts corporate credit rating transitions by integrating over two decades of data, including KRIS default probability metrics, bond spreads, yields, equity performance indicators, and macroeconomic variables. This model provides early warning signals for credit rating upgrades and downgrades, allowing investors and portfolio managers to act before such changes are fully priced by markets and reflected by ratings agencies, thereby enhancing risk management and capital allocation decisions.
- https://www.mdpi.com/2227-9091/13/6/116 – This study evaluates the predictive performance of various machine learning models—Artificial Neural Networks (ANN), Gradient Boosting (GB), Logistic Regression (LR), Random Forest (RF), and Support Vector Machines (SVM)—in credit rating classification across different countries. The findings reveal that ANN and ensemble methods like GB and RF consistently outperform traditional approaches such as LR and SVM, demonstrating superior predictive accuracy in financial forecasting tasks. These results underscore the effectiveness of advanced machine learning techniques in enhancing credit risk assessment and management.
- https://www.mdpi.com/1999-4893/18/10/608 – This research investigates the effectiveness of various machine learning algorithms—Logistic Regression (LR), Random Forest (RF), XGBoost, and Support Vector Machines (SVM)—in predicting corporate credit rating transitions and defaults. The study finds that ensemble models, particularly XGBoost, consistently outperform traditional approaches across different rating segments, highlighting the potential of machine learning in enhancing credit risk assessment and management. These findings suggest that integrating advanced machine learning techniques can significantly improve the accuracy and reliability of credit rating predictions.
- https://www.risk.net/journal-of-credit-risk/7954116/an-effective-credit-rating-method-for-corporate-entities-using-machine-learning – This study introduces a new approach to designing credit risk rating models for corporate entities, aiming to enable banks to make accurate and cost-effective rating decisions by maximising risk-adjusted returns under model uncertainty. The proposed meta-algorithm effectively utilises ordinal information embedded in expert-assigned credit ratings to accurately rank customers. Empirical results on North American medium and large-sized companies demonstrate the strong performance of the proposed approach, suggesting its potential for enhancing credit risk assessment and management in the banking sector.
Noah Fact Check Pro
The draft above was created using the information available at the time the story first
emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed
below. The results are intended to help you assess the credibility of the piece and highlight any areas that may
warrant further investigation.
Freshness check
Score:
9
Notes:
The narrative was published on 14th November 2025, making it current. The earliest known publication date of substantially similar content is 14th November 2025. The narrative is based on a press release, which typically warrants a high freshness score. No discrepancies in figures, dates, or quotes were found. No similar content appeared more than 7 days earlier. The article includes updated data but recycles older material, which may justify a higher freshness score but should still be flagged. ([itbrief.co.nz](https://itbrief.co.nz/story/machine-learning-model-gives-early-warning-on-credit-rating-shifts?utm_source=openai))
Quotes check
Score:
10
Notes:
The direct quotes from Stas Melnikov and Steven Desmyter appear to be original, with no identical matches found in earlier material. No variations in quote wording were noted. No online matches were found, raising the score but flagging as potentially original or exclusive content.
Source reliability
Score:
8
Notes:
The narrative originates from IT Brief New Zealand, a reputable organisation. The report mentions collaboration with SAS, Man Group, Pension Insurance Corporation, and Stanford University, all of which are verifiable entities. No unverifiable entities were mentioned.
Plausability check
Score:
9
Notes:
The claims about the machine learning model’s capabilities are plausible and align with existing research in credit risk prediction. The narrative lacks supporting detail from other reputable outlets, which is flagged. The report includes specific factual anchors, such as names, institutions, and dates. The language and tone are consistent with the region and topic. The structure is focused and relevant, with no excessive or off-topic detail. The tone is formal and appropriate for corporate communication.
Overall assessment
Verdict (FAIL, OPEN, PASS): PASS
Confidence (LOW, MEDIUM, HIGH): HIGH
Summary:
The narrative is current, with no significant issues found in freshness, quotes, source reliability, or plausibility. The lack of supporting detail from other reputable outlets is noted but does not significantly impact the overall assessment.