Executive summary

• Recent rapid expansion in private credit markets has driven parallels with the sub-prime MBS crisis in 2008-09.
• Increased software exposure and growing retail investor participation have caused these fears to deepen.
• To assess these claims, we look at Private Credit (PC) collateralised loan obligations (CLOs), which offer a window on the private credit market.
• We use the LSEG Yield Book CLO credit model and find that although PC CLO loans have (marginally) higher projected default rates vs syndicated CLO loans, they also have higher projected loss recovery rates, due to stronger convenants.
• The other striking feature is how much higher cumulative defaults were, and how much lower recovery rates were, on sub-prime MBS during the GFC, compared with PC loans in 2025-26. Recovery rates were only 25% on 2008 sub-prime deals, versus 65% on PC CLOs.
• Finally, although bank exposure is growing quickly, systemic risk appears far less pronounced than between sub-prime and the financial system in 2008. As of Q4, 2024, outstanding bank loans to PC vehicles amounted to $56bn, only about 5% of the private credit industry size in the US.
• We note that PC largely withstood the Covid and Ukraine shocks in 2020-22 and that both lenders and borrowers are well aware of the risks involved in these loans, whence the covenant-protection is generally greater.

Parallels are frequently drawn with sub-prime MBS during the GFC
Following the high-profile credit default events of First Brands and Tricolor in late 2025, 2026 has seen more pressure on private credit markets as redemptions of private credit funds increased  and government agencies increased scrutiny. Software exposure and growing retail investor participation have also caused AI disruption fears to intensify, and increased regulator scrutiny. As a result, parallels have been drawn with the experience of subprime mortgages in the global financial crisis (GFC), based on the underlying structure of the private credit market.

These parallels include (1) a surge in loans to lower quality borrowers, generally single B companies for private credit, or individuals with low credit scores and high debt/income ratios for sub-prime, (2) concern about a deterioration in underwriting standards, and rising defaults, (3) migration of risk to non-bank entities, outside banking system regulation, like the SPVs and securitized vehicles that were used during the GFC, (4) systemic risks via bank credit lines to private credit vehicles, and (5) the threat to financial stability from a lack of data and transparency, making leverage in the financial system and asset quality hard to detect. How realistic are these parallels, and claims, and will there be widespread credit catastrophe or another financial crisis?

Private credit CLOs are reasonably transparent and offer a means to assess PC market
To assess these claims in this paper, we look at Private Credit (PC) collateralized loan obligations (CLOs), which offer a window on the health of the private credit market. Even though Private Credit (PC) CLOs (currently at ~$150 billion size) comprise only approximately 10% of the private credit universe, they have greater data transparency for the analysis of private credit lending. PC CLO loans are typically middle market loans by direct lenders to smaller companies. They tend to be loans to smaller companies, who could not gain access to public debt markets or the banking system, post-GFC. Indeed, both sides to the transaction are fully aware of the credit risks. 

Compared to Broadly Syndicated Loan (BSL) CLO loans, which are syndicated/sold to large institutional investors, PC CLO loans are usually of lower quality with a higher share of CCC-rated loans and a higher Weighted Average Rating Factor (WARF). More notable is the increasing presence of PIK (pay-in-kind) toggles in PC facilities in recent years, with which struggling borrowers may skip interest payments and instead add to the principal balance - essentially a hidden form of additional leverage.

Current default rates of PC CLOs remain low, even if spreads have widened…
But despite that, the current default rate of PC CLO loans has remained only marginally higher than for BSL CLO loans (0.52% vs. 0.43%), reflecting the resilience of the sector and solid underwriting, even if liability spreads of PC CLOs have been widening versus BSL CLOs in recent months as investors became more cautious about private credit. 

For an objective, model-based assessment, we use the LSEG Yield Book CLO credit model to predict default and loss for loans in all CLO deals issued between 2020 and 2025. We find that although private credit CLO loans have higher projected default rates vs BSL CLO loans, they also have higher projected loss recovery rates.

PC CLOs do have higher default rates, but stronger covenants help recovery rates
Why is this? The results may reflect the fact PC loans typically have more, and stronger, covenants. This is largely due to the history of PC loans as bank loan replacements. PC lenders also tend to have closer direct relationships with borrowers, and maintenance covenants and active monitoring of the loans, leading to higher covenant breach-related defaults, but also earlier detection and re-structurings. In addition, covenant terms give lenders greater negotiating power (i.e., in Liability Management Exercises). 

…and default rates were substantially higher on sub-prime MBS during GFC
This has enabled lenders to secure higher recovery values vs BSL loans, which are more likely to be covenant-lite. The higher loss recovery is shown in Table 1. The other striking feature of the table is how much higher cumulative defaults were on sub-prime MBS during the GFC, and how much lower recovery rates were on sub-prime MBS during this period, compared with PC loans in 2025-26.

Main risks of private credit CLO loans are asset quality and lack of price discovery
We concede that a key risk or uncertainty of private credit CLO loans is asset valuation and limited liquidity, since many of these loans lack secondary trading price transparency, so uncertainty arises from the mark-to-model valuations and limited price discovery. However, we would note credit fundamentals of PC CLO loans have been stable in recent months, and there are few signs of major dislocation. In addition, LSEG Yield Book’s recent research on “quantifying the software risk in CLOs” found that while PC CLOs do hold significant software exposure at 20.5%, only 16.3% of those software loans are highly vulnerable to AI disruption, based on our analysis of loan issuers’ business model and data moat. 

Furthermore, changes in the structure of CLOs since the GFC may have reduced the risk in PC CLOs. For example, over-collaterisation (OC) cushions for PC deals remain healthy at 6%, and the share of PC deals failing the OC test is currently at negligible levels.

Regulator concerns about private investors and systemic risks increased…
While private credit concerns are real and defaults may pick up in the future, systemic risks are likely to remain low.  Although banks have exposure to private credit, the overall share of exposure is low, as the Federal Reserve points out. As of Q4, 2024, outstanding bank loans to private credit vehicles amounted to $56bn, only about 5% of the private credit industry size in the US. Interconnectedness to the banking system was much greater during the sub-prime crash, through contagion and leveraged SPVs, so liquidity dried up very quickly once it became clear correlation risk in sub-prime MBS structures was much higher than realised, and asset quality much lower. In contrast, there are liquidity lock-ups on many PC funds, and early redemption penalties, so the type of sudden liquidity freeze seen in the GFC is less likely.

…but lower bank exposure, less leverage & maturity transformation make risks lower than GFC
In addition, private credit vehicles are less engaged in maturity transformation as they use long-term capital for long-term lending. A sizeable proportion of private credit end-investors are also long-only institutional investors, who are likely to retain PC structures to maturity. Based on US Fed Financial Stability reports, pension funds, insurance companies and sovereign wealth funds are the dominant suppliers of capital to private credit, with a market share of 70-80%, with retail investors and family offices growing, but remaining smaller investors. Finally, we note that PC largely withstood the Covid and Ukraine shocks in 2020-22 and that both lenders and borrowers are well aware of the risks involved in these loans, whence the covenant-protection is generally greater. None of which means that there will not be re-financing issues periodically, and defaults. Indeed, the most recent PitchBook data indicates new loan issuance by private credit lenders fell to $44.76 billion ‌in the three months ended May 2026, down about 40% from $74.56 billion in the first quarter.

About LSEG Yield Book 
Yield Book is a trusted and authoritative source for fixed income analytics that enables market makers and institutional investors to perform complex analysis of their portfolios, benchmarks, trading decisions, historical performance, and risk. Yield Book products offer analytical insight into an extensive range of financial products in the fixed income space including governments, agencies, corporates, high yield, emerging markets, mortgages, ABS, CMBS, CMOs, CLOs, and derivatives. The platform utilizes dedicated centralized servers that help ensure reliable, prompt data delivery. Yield Book forms part of London Stock Exchange Group (LSEG)'s Data and Analytics division. 

Read the original article here.