The Stress Test: Concentration, Liquidity, and Sector Exposure

Private credit’s ability to cushion leveraged-loan stress is only as strong as the borrowers it can reach and the vehicles that supply it. This chapter tests that claim in two ways: first through software borrowers, where industry-level overlap coexists with limited movement back into LL; then through the fund-level boundaries that can make PC less resilient than its flexible loan contracts suggest. The goal is to identify transmission paths and monitoring questions, not to predict a systemic crisis from a single tightening.

The previous chapters established that private credit can expand through tailored underwriting and changing capital conditions. That expansion creates a larger possible alternative. It does not remove concentration, liquidity, valuation or leverage risk. Those risks matter because a PC shock can reduce the availability of new loans at exactly the point when PC-dependent borrowers are least able to find another route.

Software shows why market overlap is not borrower portability

Software is a useful test of the access model because it is important in both financing markets. The Federal Reserve reports that software firms accounted for about 21% of private-credit issuance and 16% of leveraged-loan issuance since 2020. The sector has also been part of recent credit-quality and valuation concerns. Yet the market-level presence of software does not answer whether an individual software borrower can switch.

At the end of 2025, about 46.3% of software borrowers in the Federal Reserve sample were PC-only, 30.6% had accessed both markets and 23.1% were LL-only. The PC-only group is the crucial fact. It means that nearly half of the observed software borrowers had no recorded LL borrowing in the 2020–25 history, even though LL investors finance software firms in the aggregate. The evidence does not reveal whether each PC-only firm was rejected by LL lenders, preferred PC, never sought an LL loan or faced a transaction that did not fit syndicated distribution. It does show that industry membership cannot be used as a substitute for borrower-level access.

The directional switching evidence makes the point sharper. Among software firms that had access to both markets, LL-to-PC switching tended to increase when LL conditions tightened. PC-to-LL switching was more muted, especially in recent years. This is consistent with the broader size pattern: software borrowers are heterogeneous and smaller on average than non-software borrowers, and a substantial share may rely more heavily on PC. Firms that can reach LL may use PC when the syndicated market becomes difficult, but the reverse route is not equally available to every PC borrower.

The sector case also guards against a misleading reading of “diversification.” If a lender’s portfolio contains technology or software loans, that does not mean the firms in that portfolio have diversified financing options. Portfolio diversification belongs to the lender; market access belongs to the borrower. A fund can spread its exposures across industries while an individual company remains dependent on one lender type. Conversely, a software company can have a strong operating business and still face a financing problem if the market it knows best reduces new commitments.

The same distinction applies to valuation concerns.