Private Credit and Bank Risk

By Jonas Osman Abdelghafour.

The growth of private credit is often described as risk migrating out of the banking system. From a bank risk manager's perspective this is only partly true. Banks remain connected to the asset class through several channels simultaneously, and because those channels sit in different parts of the balance sheet and different risk reports, the aggregate exposure is frequently larger and more correlated than any single report suggests.

The channels

Direct lending to funds and managers. Subscription lines secured on limited-partner commitments, NAV facilities secured on the fund's underlying portfolio, and general-purpose lending to management companies. Subscription lines are typically strong credit; NAV facilities depend on the valuation of illiquid assets and are structurally more exposed to the cycle they are meant to survive.

Lending to non-bank financial intermediaries more broadly. Warehouse lines, repo, and facilities to vehicles that in turn hold private credit assets. This is often reported as financial-institution exposure rather than corporate credit, which obscures the underlying sector concentration.

Shared borrowers. The same mid-market corporate may have a bank revolver alongside a private credit term loan. The bank's exposure is smaller but its position in the capital structure and its influence in a restructuring may be weaker than the size implies.

Origination and distribution relationships. Banks that originate to distribute retain pipeline risk, and in stress may find distribution channels closed at exactly the point when the pipeline is largest.

Indirect exposure through clients. Insurers, pension schemes and wealth clients holding private credit create counterparty, collateral and reputational linkages that credit reporting rarely aggregates.

Why the risk is hard to measure

Valuation is the core issue. Private credit assets are marked infrequently and with substantial judgement. Reported volatility is therefore low, and correlation with public credit appears modest — not because the underlying risk is lower, but because the measurement is smoother. Any risk model that ingests reported NAV series as if they were market prices will understate both volatility and correlation, and will do so most severely in the scenarios that matter.

Related complications include payment-in-kind interest masking borrower stress, covenant structures that delay recognition of deterioration, leverage applied at multiple layers of the same structure, and limited transparency into subordination and cross-holdings. Each of these delays the arrival of information rather than removing risk.

What risk functions should do

• Aggregate across channels. Build a single view of exposure to the asset class combining direct fund lending, NBFI facilities, shared borrowers, pipeline and client-driven exposure. The first run of this analysis is usually the finding.
• Look through to sectors. Private credit portfolios are concentrated in a handful of sectors — software, business services, healthcare — that are more cyclically correlated than a diversified label suggests. Add the look-through concentration to the bank's own corporate book.
• Treat NAV facilities as portfolio exposures. Test advance rates against stressed asset valuations, not reported ones, and model the correlation between a fund's inability to repay and the deterioration of its collateral.
• Stress liquidity, not just credit. In stress, subscription lines are drawn, distributions slow, and refinancing demand arrives simultaneously across the sector. Model the drawdown correlation.
• Interrogate valuation. Ask managers for mark methodology, frequency, independent verification and realisation experience against carrying value. Discount reported volatility accordingly in internal models.
• Set explicit appetite. Concentration limits by manager, by fund vintage, by facility type and by look-through sector — expressed together rather than scattered across separate policies.

The systemic question

Supervisors have flagged the opacity of bank–non-bank interconnections as a financial stability concern, and the honest position is that the data to resolve it does not yet exist at system level. Individual banks cannot fix that, but they can make sure they are not the institution that discovers its own aggregate exposure during a stress. The work is unglamorous: aggregate the channels, look through the labels, and stop treating smooth reported valuations as evidence of low risk.

Primary sources: ECB Financial Stability Review, May 2026; EBA Risk Assessment Report, June 2026; IMF and FSB analysis of non-bank financial intermediation.