What is private credit?

Private credit is company lending arranged directly with private funds, insurers, pension funds or other institutional investors instead of through a standard bank loan or a public bond issue. These loans are generally not bought and sold every day on a public market. The borrower and lender negotiate the interest rate, collateral and repayment terms directly.

That does not mean private credit has no connection to banks. Banks can provide credit lines to funds, help finance their portfolios or lend to the same borrower through another channel. The difference is that the initial ownership and day-to-day valuation of the loan usually sit with private investors rather than in a publicly traded bank-loan or bond market.

Quick summary:

  • Borrower: Usually a middle-market company or a business with limited access to public bond markets takes the loan.
  • Lender: A private fund, publicly traded business-development company, insurer or other institutional investor holds the loan in a portfolio.
  • Investor: When fund investors want their money back, the loan may not be easy to sell, creating a redemption limit, waiting period or valuation gap.

The short answer

The narrowest answer is this: economic risk begins with the borrower and becomes visible in the fund; redemption pressure, insurance links and bank connections can amplify it only under specific conditions.

The Federal Reserve estimates U.S. private credit at approximately $1.4 trillion as of the second half of 2025. That size is not, by itself, evidence of a crisis. The more useful question is this: when a borrower struggles to pay cash interest, where does the first loss appear, and through which channel could the pressure spread?

As of August 1, 2026, the evidence points to concentrated stress that deserves monitoring for liquidity and valuation risk, not to a generalized private-credit collapse.

How does private credit work?

Private credit generally refers to company loans that are not held on a bank's balance sheet or traded as public bonds. Private-credit funds, publicly traded business-development companies, insurers, pension funds and other institutional investors can finance these loans directly or through related vehicles.

The relationship between borrower and lender is often more direct. The parties negotiate the loan agreement, interest rate and collateral terms. Loans are frequently secured and floating rate. Those features can provide some protections compared with public bonds.

But a more direct relationship does not make risk disappear. A loan price is not formed on a public exchange every day, so problems may become visible more slowly. The first signs may appear in negotiations, changes in interest income or a lower portfolio valuation before anyone uses the word default.

That is why “private credit sits outside banking” is an incomplete description. The risk may have moved outside a bank's balance sheet, but it still exists across the borrower, vehicle, insurer and bank connections around it.

The chain through which risk moves

The chain to monitor is:

borrower cash flow weakens -> PIK or non-accrual appears -> fund marks the loan down -> net asset value (NAV) and income come under pressure -> investors request redemptions -> the vehicle needs a gate, asset sale or new capital -> insurer and bank connections matter

The terms in this chain do not mean the same thing.

Non-accrual means that a lender no longer treats all contractual interest as collectible and stops recognizing some interest income. It is a strong credit warning, but it is not by itself a default or bankruptcy finding.

PIK interest is interest that is not paid in cash and is instead added to the loan principal. It may be planned in the original agreement. It may also be used when a borrower is finding cash payments difficult. The PIK share should therefore be read alongside cash flow, restructuring and valuation information.

Fair value is the lender's current estimate of what an asset is worth. A mark-down is not necessarily a realized cash loss. It does, however, affect the vehicle's net asset value and the return shown to investors.

A redemption limit or gate restricts how much an investor can withdraw from a vehicle during a particular period. It can spread liquidity pressure over time. It does not make an illiquid loan liquid.

What does market size tell us?

The Fed puts U.S. private credit at approximately $1.4 trillion. It says that amount is about 10% of U.S. nonfinancial corporate debt and roughly one-third of below-investment-grade corporate debt excluding bank loans.

The Financial Stability Board estimates global private credit at approximately $1.5 trillion to $2 trillion. These figures should not be added or treated as one comparable series. The Fed is measuring a U.S. market, while the FSB is using a broader global scope.

The Fed's more important warning concerns the structure of the vehicles. Perpetual-life business-development companies and interval funds together hold approximately $241 billion of net assets. Unlike traditional closed-end vehicles, these funds can receive investor requests for cash more frequently while holding assets that are not easy to sell quickly.

The Fed reported that redemption requests in some vehicles rose above the normal 5% quarterly limit during the first quarter of 2026, while describing the requests as manageable at that point. That wording matters. “Manageable” is an assessment of that period. It does not prove that a future stress wave will remain contained.

The first risk appears in the borrower

Many private-credit borrowers are middle-market companies with higher leverage or limited access to public credit markets. An NBER study of a historical middle-market loan sample found that nonbank borrowers were less profitable and more levered than bank borrowers. After controlling for interest rates and other characteristics, the probability of borrowing from a nonbank rose materially when a company's earnings before interest, taxes, depreciation and amortization fell below zero.

This does not mean that every private-credit borrower is weak. It does mean that the selection of borrowers entering private credit can differ from bank lending or the market for large-company bonds. Another study in the Review of Financial Studies finds that nonbank credit supply can be more cyclical than bank credit over credit cycles.

In real cases, the first signal is often not a default headline. It may be the loss of cash interest, a higher PIK share, a lower loan valuation or a restructuring.

Medallia: restructuring through debt reduction

Software company Medallia announced a restructuring agreement with its lenders in June 2026. The company said the agreement materially reduced outstanding debt, provided $150 million of new capital and transferred ownership to an investor group led by Blackstone.

It would be wrong to summarize this as “Medallia went bankrupt.” The company described a recapitalization and restructuring involving debt reduction and new capital, not a bankruptcy filing.

The mechanism is narrower and more useful. When a borrower can no longer comfortably carry its existing debt, lenders may put in new capital, reduce debt or change ownership to preserve value. The economic risk begins with the borrower, but the solution and the valuation move into the lenders' portfolios.

Affordable Care: one borrower, different credit layers

FS KKR Capital's March 2026 portfolio schedule lists different positions in Affordable Care separately. One position with $12.7 million of amortized cost had a fair value of $9.2 million. Another with $58.1 million of amortized cost had a fair value of $42.3 million. A preferred-stock position showed $77.8 million of cost and $1.7 million of fair value.

It would be wrong to combine these figures into one “recovery rate.” Each position can have a different seniority, collateral package, contract and payment priority. The table does, however, show how exposure to one borrower can be split into different layers inside a lender's portfolio.

The risk then becomes visible in the fund

A private-credit vehicle takes the borrower's cash flow and loan value and reflects them in its net asset value (NAV). This creates two separate risks:

  • The borrower may become less able to pay interest and principal.
  • Fund investors may question both the value of the assets and the time needed to turn them into cash.

Blackstone Private Credit Fund, known as BCRED, reported more than $80 billion of portfolio assets, 97% senior-secured debt and an average loan-to-value (LTV) of 41%, nearly 700 issuers and more than 50 industries at March 31, 2026. This structure explains why private credit can appear resilient: collateral, a relatively low average LTV and diversification.

The same filing said BCRED's non-accrual rate rose from 0.6% to 2.4% at cost and stood at 1.4% at fair value. The fund identified Medallia and Affordable Care as the main contributors. The average mark for the bottom 5% of the private-debt portfolio was 69.6.

These figures do not support the conclusion that BCRED collapsed. They do show that a highly secured and diversified portfolio can still contain borrowers that create stress. Risk is not evenly distributed across the market. It can be concentrated in a small number of names.

How does liquidity pressure emerge?

In a June 2026 repurchase letter filed with the SEC, BCRED said second-quarter repurchase requests were approximately 10% of shares outstanding, while the designed limit was 5%. In other words, requests were about twice the normal quarterly limit. The fund also reported roughly 2% of NAV in inflows and roughly 3% of NAV in net outflows.

This information can be misread in two ways. The first is to call it a run or a fund failure immediately. We would also need to know how much was met, in what order, and which capital sources the vehicle used.

The second is to treat the 5% limit as a guarantee that investors can always exit at NAV. The limit is a tool for managing asset sales and cash. If requests exceed it, an investor may receive less than requested or carry the request into a later period.

Golub: the difference between valuation, NAV and cash income

Golub Capital BDC is a publicly traded business-development company that lends to middle-market businesses. At March 31, 2026, it reported a portfolio fair value of $8.317 billion, 420 portfolio companies, 92% first-lien exposure and 99% floating-rate exposure.

That is a constructive starting point. The same Form 10-Q reported 19 non-accrual investments, equal to 2.3% at amortized cost and 1.4% at fair value. The amortized-cost base was $198.441 million and fair value was $118.510 million.

The simple ratio of fair value to the amortized-cost base was approximately 59.7%. This is an accounting comparison between two disclosed bases. It is not a recovery rate or an estimate of loss. Seniority, collateral, restructuring and valuation differences prevent us from turning the ratio into a direct recovery assumption.

Golub's NAV per share declined from $14.84 to $14.35, a fall of approximately 3.30%. The company reported a first-quarter net loss of $0.18 per share and a net unrealized loss of $0.51 per share.

PIK income shows another important distinction. Golub reported $38 million of PIK income, equal to 12.5% of total investment income. PIK is not cash interest collected. Investors should therefore ask how much of the reported income can actually support distributions and how much has been added to loan balances.

The market is not homogeneous: Goldman Sachs as a counterexample

Goldman Sachs Private Credit reported a 0.2% non-accrual rate at amortized cost and only one company on non-accrual at March 31, 2026. That is materially different from the higher rates reported by some other private-credit vehicles in the same period.

The counterexample shows two things. First, private credit does not behave like one portfolio. Borrower selection, loan seniority, sector concentration, restructuring approach and valuation practice can produce different results. Second, it is not sound to infer the condition of the whole market from one or two stressed vehicles.

Goldman's distinction between “good PIK” and “bad PIK” is the company's own classification, not a market-wide standard. It is still useful as a reminder that PIK needs context rather than an automatic label.

When do insurer and bank connections matter?

Insurers can hold long-duration and less liquid assets to match long-term liabilities. NAIC says one reason life insurers participate in private credit is that the asset duration can match their liabilities. This structure can reduce the likelihood that an insurer will be forced to sell quickly.

But less liquid assets are also harder to price. NAIC points to limited secondary-market liquidity and less frequent valuations as transparency and pricing concerns in private credit. Its 2026 reporting changes seek more detail on fair value, Level 2 and Level 3 classifications, PIK and private credit ratings.

The Financial Stability Board says private credit's connections with banks, insurers and private-equity firms are deepening. That statement does not mean insurers or banks will necessarily take losses. For risk to grow, the same assets would need to appear across multiple institutions, shared credit lines would need to come under pressure, collateral calls would need to rise or investor redemptions would need to force asset sales.

The distinction is important: insurer exposure is a transmission channel to monitor, not evidence of proven contagion. The IMF likewise emphasizes that private-credit liquidity mismatch and retail access are rising, but that systemic risk depends on how leverage, interconnections and forced deleveraging combine rather than on market size alone.

NAIC's $276.8 billion figure for insurers' CLO holdings should be read in the same way. CLOs are structured-credit instruments related to private credit, but they are not the same as direct private-credit funds. The figure should not be added to the Fed's $1.4 trillion estimate.

The strongest counterargument

The strongest objection to the risk story is that private-credit loans are often first-lien, secured, floating rate and held by investors with longer horizons. Lenders may also be able to speak with borrowers earlier and amend terms more quickly than investors in public bonds.

This objection is real. BCRED's 97% senior-secured share and 41% average LTV show why private credit may be more resilient than parts of the public leveraged-loan market. The Goldman Sachs example also shows that stress has not spread evenly across managers.

But these features do not eliminate risk. Collateral can lose value, borrowers can stop paying, PIK income can replace cash income and fund investors may not be able to withdraw on demand. The better conclusion is that private credit changes the form of risk. It does not remove it.

Five questions for locating the risk

When reading a new private-credit disclosure, ask:

1. Is the borrower still paying cash interest?

If PIK rises while cash interest falls, the income statement may show interest income while cash flow weakens. Check whether the PIK increase was planned in the agreement or introduced through a restructuring.

2. Are non-accruals and fair-value marks deteriorating together?

Non-accrual provides information about credit quality and income recognition. Fair value provides information about valuation and expected collection. A simultaneous deterioration is a stronger signal. A move in only one measure needs a different explanation.

3. Is the problem concentrated in a few borrowers?

Look at the share of the portfolio represented by the lowest-marked positions, sector concentrations and the same borrower across different funds. A loss in a few names is not the same risk as deterioration across a portfolio.

4. What happens when fund investors want out?

Monitor the size of requests, the normal limit, how much is met, what is carried forward and whether the vehicle sells assets. “Redemption requests arrived” is not the same information as “investors received all their money.”

5. Whose balance sheet holds the same asset?

Separate the borrower, fund manager, insurer, bank credit line and final investor channel wherever possible. To understand who bears the risk, it is not enough to look at the borrower's name or the fund's brand.

What would change the conclusion?

The conclusion would become more severe if several signals appeared together: non-accruals and PIK rising across unrelated managers, persistent marks across different portfolios, redemption requests above limits, forced sales below carrying values, pressure on insurer capital or bank collateral calls.

The conclusion would become less severe if cash-interest payments remained stable, marks proved to be temporary spread movements, restructurings produced high recoveries, redemption limits were rarely tested and bank and insurer links remained small and well capitalized.

The FSB says private credit has not yet been tested through a severe downturn. Current data therefore tell us more about which indicators may provide an early warning than about the eventual size of a future crisis.

What to monitor next

  • Whether private-credit managers report non-accrual, PIK and fair-value data using comparable definitions.
  • Whether borrower cash-interest payments are diverging from total reported interest income.
  • How frequently restructurings involve debt reduction, new capital or ownership changes.
  • Whether the gap between redemption requests and vehicle limits is widening.
  • Whether fair-value marks remain concentrated in a few borrowers or spread across sectors.
  • How life insurers disclose concentration, valuation and capital treatment for private credit and related structured-credit assets.
  • Whether collateral and liquidity terms are changing on bank credit lines to private-credit managers and portfolios.

Conclusion

There is no one-word answer to where risk accumulates in private credit.

The first economic risk appears when a borrower cannot produce enough cash. The fund sees it through non-accruals, PIK, fair-value marks and NAV pressure. When investors request redemptions, a credit problem can become a liquidity problem. Insurers and banks connect to the structure through investments, funding, collateral and credit lines.

Medallia and Affordable Care show how risk can move from borrower to lender portfolio. BCRED and Golub show how it becomes visible at the vehicle level. Goldman Sachs Private Credit reminds us that stress has not spread to every fund.

The most accurate sentence today is not “private credit is collapsing.” The risk did not disappear when it moved outside banks. It was redistributed across borrowers, funds and liquidity connections. Whether it becomes a market-wide crisis will depend on how non-accruals, PIK, valuations, redemptions and institutional links move together.

Methodology and limitations

This Guide uses the Federal Reserve's May 2026 Financial Stability Report, the Financial Stability Board's report on private-credit vulnerabilities, IMF and NAIC materials, NBER and Review of Financial Studies research, filings from BCRED, Golub Capital BDC, Goldman Sachs Private Credit and FS KKR Capital, and Medallia's company announcement. The cutoff is August 1, 2026 at 18:39 Istanbul time.

The Fed's approximately $1.4 trillion figure is used for the U.S. market. The FSB's $1.5 trillion to $2 trillion range is global. I did not add them because their scopes are different.

I calculated the 2.0x ratio of BCRED's roughly 10% redemption request to its 5% designed limit, the Golub and FS KKR NAV changes, and the ratio of fair value to amortized-cost base for Golub's non-accrual investments using the figures in the cited sources. The 59.7% ratio is not a recovery or collection rate. NAV changes are not realized credit-loss rates.

Non-accrual, PIK, fair value, redemption request, redemption limit and restructuring are not synonyms. Company and fund disclosures can present portfolio quality favorably, so the evidence is read alongside official stability assessments, academic research and a lower-stress counterexample.

This Guide does not run a severe-recession stress test. A comparable manager-level panel combining non-accrual, PIK, fair-value, redemption and post-restructuring recovery data is not available. The Guide therefore does not forecast a private-credit collapse. It explains where risk may become visible and which indicators could show that a concentrated problem is spreading.

Sources and further reading

Not investment advice; for research and educational purposes.