Method
Data and method in brief
The Private Credit Radar tracks credit stress in the loan books of US business development companies (BDCs), the largest group of private-credit lenders that publish loan-level marks every quarter. It reads the schedules of investments filed with the SEC, follows each loan from quarter to quarter, and reports six indicators through which stress becomes visible. Version 1.0 covers Q1 2023 to Q2 2026. This page says where the data comes from, what was done to it, and what it cannot show.
The data
Every number in the Radar comes from public SEC filings. The source is the schedule of investments in each 10-Q and 10-K that a BDC files on EDGAR, read from the XBRL instance and the HTML document, for the 14 quarter-ends from 31 March 2023 to 30 June 2026 (filings pulled 19 September 2026). Reference data on fund structure comes from the filings and from a listing history built from N-54A, prospectus and 8-K filings.
| Q1 2023 | Q2 2026 | |
|---|---|---|
| BDCs with a schedule of investments | 129 | 174 |
| BDCs in scope (funded debt book, not a venture lender) | 105 | 162 |
| Schedule rows read (all instruments) | 28,245 | 57,762 |
| Funded debt positions in scope | 15,523 | 36,687 |
| Distinct borrowers | 4,247 | 7,139 |
| Fund managers in scope | 61 | 88 |
| Positions followed from the previous quarter | 86.6% |
A position is one debt facility held by one fund. Indicators are published for all funds, for a constant panel of the 87 funds present at every quarter-end, and for listed and non-traded funds separately. Equity, CLO tranches, joint-venture certificates and cash lines are read but excluded from the indicators.
Listed and non-traded funds
A BDC is listed when its common stock trades on the New York Stock Exchange or Nasdaq at the quarter-end; its shareholders can sell at the market price. Every other BDC in scope is non-traded: its shares do not trade on an exchange. Most non-traded BDCs are perpetual-life funds that issue shares continuously at net asset value and offer to buy back a limited amount each quarter, sold either through a registered public offering to wealth-management clients or privately to institutions. The rest are finite-life private BDCs that call committed capital from institutional investors and return it as loans are repaid, and a few that intend to list. Listing status is set quarter by quarter from each fund's listing history (Form N-54A, prospectuses, 8-K filings), so a fund that lists moves to the listed group from its first quarter-end as a traded stock.
| Listed | Non-traded | |
|---|---|---|
| BDCs in scope, Q2 2026 | 44 | 118 |
| Funded debt positions | 10,417 | 26,270 |
| Principal, USD billion | 134 | 365 |
Most funds that entered the sample after Q1 2023 are non-traded and hold young loans marked near par, which is the main reason the non-traded group shows less stress. Where a listed and a non-traded fund hold the same borrower, they mark it alike; the difference between the groups lies in the age and composition of their books, not in how they value the same loan.
What we did with it
- Extraction. Each schedule is parsed row by row: borrower, instrument, principal, cost, fair value, spread and reference rate, PIK rate, maturity, footnotes. Where the XBRL tags are incomplete the HTML table fills the gap, and every field records which source it came from.
- Cleaning. Scale errors in the filings are corrected by rule: figures tagged in thousands instead of units, spreads tagged at a hundredth, and principal written into the share-count column. Each repair must be corroborated by the other amounts on the row before it is applied, and a fund whose positions fall out of scope against its own previous quarter is flagged before the indicators are computed. Loans in euro, sterling, Canadian or Australian dollars are converted at the period-end rate so that fair value over principal is a mark and not an exchange rate. The mark is fair value divided by principal, times 100.
- Scope. A position is in scope when it is funded debt with principal, cost and fair value present. Equity, structured products and cash-like rows stay out; pure venture lenders stay out at fund level.
- Identifiers and links. Each position is followed across quarters, and each borrower carries one identifier across all funds, so that the same company held by several lenders is seen once and the dispersion of their marks can be measured.
- Checks against the filings. Research assistants read a random sample of positions and every named figure back from the filing each release; a field whose error rate exceeds 2% blocks the release. Corrections found this way (non-accrual flags, borrower identities, mis-scaled values) are recorded with the filing page that supports them and applied before the indicators are computed.
- Indicators. Six series are computed each quarter on the position panel, for all funds, for the constant panel and by listing status.
The indicators
All funds, 30 June 2026 and a year earlier. Each indicator has its own page with the definition and the full series.
| Indicator | What it captures | Q2 2026 | Q2 2025 | |
|---|---|---|---|---|
| 1 | Marked below 90 and below 80 | The share of loan positions that funds mark below 90% of face value, and below 80%. | 10.3% / 5.9% | 7.7% / 4.4% |
| 2 | Markdown breadth | The share of positions marked down by five points or more during the quarter. | 4.0% | 3.1% |
| 3 | Cross-manager mark gap | The share of borrowers held by two or more managers whose marks differ by five points or more. | 12.3% | 9.3% |
| 4 | PIK switch rate | The share of loans that moved from cash interest to payment-in-kind (PIK) interest during the quarter. | 1.44% | 1.15% |
| 5 | Distress amendments (Tier 2) and other amendments (Tier 3) | The share of loans that were amended under stress (Tier 2) or amended without a stress marker (Tier 3) during the quarter. | 0.53% / 2.74% | 0.40% / 2.11% |
| 6 | Non-accrual rate | The share of debt, at cost, on which funds have stopped accruing interest. | 1.84% | 1.29% |
Indicators 2, 4 and 5 are measured on positions followed from one quarter to the next. Thresholds (90, 80, 5 points, 50 basis points, 180 days, 12 months) are parameters; the series are published with them fixed and any change is reported with its effect.
What the data does not see
- The Radar measures US BDC loan books. European borrowers enter only where a US BDC holds them: 690 borrowers at 30 June 2026. There is no European lender data in version 1.0.
- Waivers, covenant resets, equity cures and amendment fees are not printed in a schedule of investments; the amendment indicators rest on PIK switches, spread increases and maturity extensions.
- A fund that reports non-accruals only in running text shows none.
- A loan held by several funds counts once per fund. Realised losses and restructurings with principal written down are outside the six indicators.
- Positions at funds that entered after Q1 2023 are younger and marked higher; the all-fund series falls when new funds enter without any loan improving, which is why the constant panel is shown alongside.
- Between 13% and 22% of positions in a quarter cannot be followed from the previous quarter, mostly new positions and funds with off-calendar fiscal quarters.
Versions and citation
The Radar is published quarterly, about six weeks after the last BDC filing of the quarter. Each release names its data version. Definitions do not change between releases without a note stating the change and its effect on the history; corrections are listed on the corrections page with their date. No borrower is named.
Cite as: Steffen, S. (2026). Private Credit Radar, version 1.0. privatecreditradar.eu.