Private credit: when borrowers find a cheaper exit

On 17 September 2026, Mercer Advisors priced a $1.65 billion, seven-year syndicated loan intended to refinance roughly $1.6 billion of private debt, according to Bloomberg. The reported spread was 2.75 percentage points above the benchmark, compared with 4.50 points on the existing financing. The report, published on 18 September, relies in part on an unnamed source. S01

The borrower’s incentive is straightforward: obtain cheaper financing. For the outgoing lenders, recovering principal would be a normal and often welcome outcome. Their interest-earning investment would, however, need a replacement. Repayment can improve a fund’s liquidity while reducing its future income. Reinvestment risk is expressly disclosed in fund prospectuses, separately from default risk. S05

A second question follows. Companies able to secure better terms elsewhere may not be a representative sample of the portfolio. If they leave more readily than other borrowers, the composition of the remaining loans changes. That possible selection effect deserves scrutiny. The available documents do not establish it across private credit.

Add the benchmark to the spread

Private credit here means loans negotiated directly with funds or other nonbank financing companies. Many of these loans have a floating benchmark plus a contractual spread, a structure described in the Federal Reserve’s research on the sector. S03

The Mercer position reviewed in the financial statements uses SOFR as its reference rate. SOFR measures overnight dollar borrowing costs against US Treasury collateral; the way it feeds into a particular loan’s coupon depends on the contract. The spread comes on top. It is neither the borrower’s total interest rate nor the lender’s net profit. S02 S09

The difference between 4.50 and 2.75 percentage points is 175 basis points. Applied to an unchanged $1.65 billion balance, that produces $28.875 million of annual interest. Applied to $1.6 billion, it produces $28 million. These calculations isolate the spread change, assuming identical benchmark rates and excluding contractual floors, principal amortisation and other costs.

They are not a verified estimate of Mercer’s net savings. The old balance and new face amount differ. The reported issue price is 99.75 per 100, a $4.125 million discount on $1.65 billion before other fees. The package also includes a separate $250 million delayed-draw facility, which should not be treated as cash already advanced. S01

A complete comparison would require drawdown and repayment schedules, arrangement fees, any exit premium and both contracts’ rate provisions. Pricing a loan does not establish that every existing lender has already received payment. No settlement document detailing those payments was identified in the sources reviewed as of 20 September.

A disclosed Mercer position

The quarterly filing of KKR FS Income Trust Select, a US business development company, provides an independent documentary check on the old financing. Its 30 June 2026 investment schedule lists a Mercer Advisors position with $17.633 million of principal, $17.563 million of amortised cost, reflecting the gradual recognition of fees and discounts, and $17.525 million of fair value. The listed spread is SOFR + 4.5%, with an October 2030 maturity. The original table is denominated in thousands of dollars. The amount refers to the first Mercer row; two further rows are unfunded commitments, marked with note (i). S02

That confirms an identified exposure and the old spread. It does not provide Mercer’s complete creditor list, a credit rating or the terms of the September refinancing. Fair value is a portfolio estimate at a particular date; a sale still requires a buyer and an agreed price.

Legal entities matter. KKR FS Income Trust Select is not FS KKR Capital Corp., the separate BDC discussed later. Nor is a fund’s interest income the same thing as the fees earned by its asset manager. Working only with the KKR or Apollo brand name can obscure whose capital actually bears the exposure.

Follow the new loan’s holders

A syndicated loan is arranged for distribution among multiple lenders or investors. Buyers can include funds and collateralised loan obligations, or CLOs: vehicles that purchase loan portfolios and finance them by issuing securities with different payment priorities. BIS research describes that structure and its use of leveraged loans. S08

Describing a refinancing as a return to banks can therefore be misleading. A bank may arrange and distribute the financing while retaining a limited position. Naming the arranger does not identify the ultimate holders. An asset manager active in private credit could even encounter the same borrower in another strategy, where its mandates permit.

Mercer would not be repaying the refinanced debt with an equivalent operating cash surplus. New financing would replace the old. The outgoing creditor receives money, while the company remains indebted. Understanding the overall risk requires following the new debt, its cost and its owners, rather than stopping at the old fund’s exit.

The borrower can shorten the attractive part

Consider a lender accepting corporate risk in exchange for a 4.5-point spread over several years. If competing financing becomes available at a 3-point spread, the borrower has reason to renegotiate or leave. The lender can lower its price to retain the exposure or take its capital back and look for another investment.

That exit option is contractual. Call protection can require a premium or compensation for interest expected over a specified period. Some provisions apply only to refinancings undertaken to reduce borrowing costs. Proskauer’s 2023 analysis, focused mainly on Europe, describes these structures and their exceptions. Mercer’s own protection terms remain unavailable in the documents reviewed. S04

The asymmetry is worth understanding. When financing conditions improve, the borrower may be able to shorten the period for which it pays a high spread. If its own condition deteriorates, finding a replacement lender becomes harder. The creditor does not acquire a matching right to demand repayment at par whenever a more attractive investment becomes available.

The lender may have priced the option adequately, earned upfront fees or received an exit premium. Its economic outcome combines compensation for the risk actually borne with earnings on the returned capital. A contractual maturity date is not a guarantee of that many years of interest.

Reinvesting $100 million

Take a wholly fictional example, with no fund-level borrowing, defaults, taxes or management fees. A $100 million loan earns 8.5% a year. It is repaid at the start of the period under examination. The lender must decide how to use the proceeds over the next twelve months.

Keeping the same $100 million invested at 8.5% would produce $8.5 million of interest. Immediate reinvestment at 7% produces $7 million. Holding the money in cash at 4% for two months, then lending the $100 million at 7% for the remaining ten months, produces $6.5 million. The calculation uses simple monthly fractions without compounding. A fictional benchmark stays at 4%: the spread falls from 4.5 to 3 percentage points.

An equally hypothetical 1% prepayment premium adds $1 million: cash income reaches $7.5 million in the first year under the delayed-reinvestment scenario. The following year, with no further premium and the replacement loan still yielding 7%, income is $7 million. The premium can offset part of the shortfall, but it is not a permanent annual supplement.

These are not risk-adjusted comparisons. The replacement loan might be safer, or simply pay less for comparable risk. Cash also has value: it can prevent a forced sale or allow the manager to take a better opportunity. The $2 million interest shortfall in this example is not automatically $2 million of destroyed value.

A leveraged fund has another option: repay its own debt. The benefit depends on the cost of that funding, whether it can be repaid and any associated penalties. Comparing only asset coupons leaves half the balance sheet out of the analysis.

A repayment can flatter the quarter

Accounting introduces another complication. In its June 2026 quarterly report, MidCap Financial Investment Corporation (MFIC) states that early repayment brings prepayment premiums and unamortised origination fees or discounts into interest income. This is an example of MFIC’s accounting policy; no Mercer exposure is attributed to that entity here. S06

Consider a separate fictional transaction, ignoring fair-value changes. A fund advances 98 for a claim with a face amount of 100. Part of the difference is gradually recognised as interest income. When amortised cost has reached 99, the borrower repays 100. The remaining point of discount becomes income in that period instead of being spread over the loan’s expected remaining life.

That point is not a new premium paid on top of face value. It is already included in the repayment of 100. Principal recovery, ordinary interest, any additional premium and accelerated discount recognition must be distinguished. Treating them as interchangeable can lead to counting the same money twice.

The exit quarter can show strong interest income even though the loan will earn nothing thereafter. The accounting may be entirely proper. It simply means that the recurring income left after the transaction matters more than mechanically extrapolating the quarter’s result.

Early repayment may even improve the original loan’s annualised return when its discount or fees are earned over a shorter period. That still does not determine the fund’s return over the full year after reinvestment. An individual investment and a continuing portfolio operate on different timelines.

The remaining loans are not a random sample

Now consider the possible selection effect. Suppose a fictional portfolio holds $1 billion of loans, of which $100 million is on a watch list. This category is invented for the illustration. It is neither an observed default rate nor a regulatory classification. All loans are assumed to be carried at face value, with no new impairment.

If $250 million of loans outside the watch list is repaid and no new loans are made, the remaining loan book is $750 million, still including the same $100 million on watch. The share rises from 10% to 13.3%. No additional borrower has weakened. The denominator has changed.

The fund also holds $250 million of cash. While that cash is retained, total assets in this example remain $1 billion and the watch-list exposure is still 10% of loans plus cash. Claiming that fund risk has automatically increased by 33% would be wrong. Concentration within the remaining loans, liquidity and the risk borne by equity are different measures.

What happens next matters: retaining cash, distributing capital, repaying borrowings or making new loans. A larger share of difficult credits in the residual loan book can coexist with lower overall risk exposure. Conversely, rushing into riskier replacement loans to preserve a headline yield could genuinely weaken the portfolio. That is a scenario to test, not behaviour attributed here to the named managers.

Access to syndication is not a perfect ranking of borrower strength either. Financing size, transaction complexity and the need for flexibility all matter. Federal Reserve research notes that borrowers pay for private credit’s speed, execution certainty and customisation. A sound company may prefer to stay; a highly leveraged company may still find investors. S03

The accounts offer a counterpoint

Second-quarter disclosures provide a useful check, predating September’s refinancing. FS KKR Capital Corp. reported $590 million of purchases against $1.334 billion of sales and repayments. From March-end to June-end, non-accrual investments, where interest is no longer recognised in the normal way, fell from 4.2% to 3.8% of the portfolio at fair value. Debt principal fell from $7.290 billion to $6.491 billion. S07

Net exits coexist here with lower debt and a lower fair-value non-accrual ratio. Explaining the change requires separating sales, repayments, valuations and capital transactions. Marking down troubled assets can itself reduce their weight in a fair-value ratio.

MFIC’s activity table for the same quarter separates $47.2 million invested, $79.1 million sold and $128.3 million repaid, with reported net investment activity of negative $160.2 million. That net figure is not the gross volume of loans refinanced elsewhere. Rounding can prevent the table’s intermediate totals from reconciling to the last decimal place. S10

The aggregates combine several types of exit. Identifying syndicated refinancings and comparing the borrowers requires transaction-level detail that these tables do not supply.

Competition works even without a Fed rate cut

The reinvestment example deliberately holds the benchmark at 4%. Only the replacement loan’s spread changes. A floating-rate lender can therefore face lower contractual earnings without a central bank cutting its policy rate.

The spread charged to borrowers and the spread paid on the fund’s own financing need not fall together. Some fund liabilities are fixed-rate and others float; maturities and floors may differ from those on the assets. MFIC expressly discusses this asset-liability mismatch in its interest-rate risk disclosure. A sensitivity test that moves SOFR while leaving the portfolio unchanged cannot, on its own, measure the effect of repayments and lower-spread replacements. S06

For an income-seeking investor, the relevant number is earnings after financing costs, fees and losses, alongside how much is actually collected in cash. For the manager, the decision concerns the price at which to retain or replace a credit relationship. For the borrower, a lower interest bill may improve repayment capacity. The same refinancing can help a company and reduce its former lender’s revenue.

Track borrowers after they leave

Establishing adverse selection would require a cohort of borrowers present at a specified date, followed separately through retention, refinancing, sale and default. Departing companies should be compared using information observed before they leave: leverage, interest-paying capacity, available ratings, sector, size and contractual protections.

Replacement loans would then need to be examined. The exit of sound credits followed by equally sound lending is a different trajectory from a rotation towards weaker borrowers. Equally, a lower yield may reflect lower risk. Neither average portfolio yield nor the quarter’s default rate is enough to distinguish these outcomes.

Mercer’s case illustrates competition over loan pricing and the work awaiting a repaid lender. The quality of replacement loans and the use of cash will shape future income and risk. A sector-wide conclusion about the departure of stronger borrowers would require following those companies over time.

Repayment remains the basic promise of lending. The lender’s next test is how returned capital is used: what it earns, which risks it finances and how much borrowing it removes from the balance sheet. That is the level at which a borrower’s exit becomes a portfolio decision.

Further reading

Our analysis of private credit and software cash flows examines the income available to service debt. Private credit’s data gaps explain the limits of sector-wide comparisons.

Sources

Sources reviewed on 20 September 2026. Filings and releases describe issuer disclosures; the scenarios labelled fictional are l0g illustrative calculations.

1. S01 · Wealth Manager Mercer Takes Out Private Credit With New LoanBloomberg, via WealthManagement.com · 2026-09-18. Paragraphs 1 to 4: amount, tenor, spreads, issue price and delayed-draw facility.
2. S02 · Form 10-Q, quarter ended June 30, 2026KKR FS Income Trust Select / SEC EDGAR · 2026-06-30. Schedule at 30 June 2026, first Mercer row, page 9; notes (b), (c), (f), (h) and (i). Dollar amounts in thousands.
3. S03 · Private Credit: Characteristics and RisksFederal Reserve Board · 2024-02-23. Introduction; paragraphs on speed, execution certainty and customisation; Loan Type and Pricing and Illiquidity sections.
4. S04 · Private Credit Deep Dives: Call Protection (Europe)Proskauer Rose LLP · 2023-05-22. Introduction; three protection methods; section 2, Type of prepayment; sections 3 and 5, Net present value and Permitted refinancings.
5. S05 · Prospectus, April 29, 2024, as revised October 24, 2024BlackRock Credit Strategies Fund / SEC EDGAR · 2024-10-24. Printed pages 13 and 88: Prepayment Risk and Reinvestment Risk.
6. S06 · Form 10-Q, quarter ended June 30, 2026MidCap Financial Investment Corporation · 2026-06-30. Investment Income Recognition and Interest Rate Risk sections. Accounts at 30 June 2026.
7. S07 · Second Quarter 2026 ResultsFS KKR Capital Corp. / Future Standard · 2026-08-06. Financial and Operating Highlights; Portfolio Highlights; debt table, Debt outstanding (principal), $6,491m and $7,290m.
8. S08 · Structured finance then and now: a comparison of CDOs and CLOsBank for International Settlements · 2019-09. Introduction and Complexity and Spillover risks sections, explaining loan collateral and payment tranches.
9. S09 · Secured Overnight Financing Rate DataFederal Reserve Bank of New York · Definition only; no current fixing used. First methodological paragraph below the title, defining SOFR.
10. S10 · Financial Results for the Quarter Ended June 30, 2026MidCap Financial Investment Corporation · 2026-08-06. PORTFOLIO AND INVESTMENT ACTIVITY table: investments 47.2, sales 79.1, repayments 128.3, net activity −160.2 ($ millions).