Fears that private credit markets are repeating the mistakes of sub-prime mortgage-backed securities (MBS) ahead of the global financial crisis (GFC) have intensified following the high-profile defaults of First Brands and Tricolor in late 2025, alongside rising redemptions and growing regulator scrutiny in 2026.
According to LSEG Data & Analytics, which recently delved into the topic, the comparisons rest on five parallels: a surge in lending to weaker borrowers, deteriorating underwriting standards, risk migrating to non-bank entities outside traditional regulation, systemic exposure via bank credit lines, and a lack of transparency that obscures leverage and asset quality across the system.
To test these claims, LSEG Data & Analytics examined private credit (PC) collateralised loan obligations (CLOs), a roughly $150bn segment that, despite representing only around 10% of the broader private credit universe, offers far greater data transparency than the market as a whole.
Using the LSEG Yield Book CLO credit model to assess loans issued between 2020 and 2025, the analysis found that while PC CLO loans carry higher projected default rates than broadly syndicated loan (BSL) CLOs, they also deliver stronger loss recovery rates, a result LSEG Data & Analytics attributes to tighter, more frequent covenants and closer lender-borrower relationships that allow earlier intervention.
The contrast with sub-prime MBS is stark. Cumulative defaults on 2007-2008 vintage sub-prime MBS deals reached 30%, with recovery rates of just 25%, compared with a 10.4% cumulative default rate and 64.8% recovery rate for PC loans. Current default rates also remain low across both CLO types, at 0.52% for PC versus 0.43% for BSL, even as liability spreads on PC CLOs have widened in recent months amid investor caution.
LSEG Data & Analytics does flag genuine risks, particularly around asset valuation and limited secondary market liquidity, which leave mark-to-model pricing exposed to uncertainty. Software exposure within PC CLOs also stands at 20.5%, though separate Yield Book research suggests only 16.3% of those loans are highly vulnerable to AI-driven disruption. Over-collateralisation cushions remain healthy at 6%, with few deals currently failing OC tests.
For more insights, read the full story here.
Copyright © 2026 FinTech Global
