Skip to content
Dividend Coverage
Model v1 · published 2026-11-15

How the score works

Five factors, fixed thresholds, 100 points. Every rule is published here, so anyone can recompute any score from the company's own release.

Five factors

  • Coverage
  • NAV trend
  • Non-accruals
  • Leverage
  • Earnings quality

Coverage

35 pts

Measures: Core NII per share ÷ regular dividend per share.

The single best predictor of a cut. We strip disclosed one-time items (dividend income from restructurings, fee waivers, prepayment spikes) and exclude supplemental and special dividends from the denominator.

115%+35
105-115%30
100-105%24
95-100%15
90-95%8
Under 90%0

NAV trend

25 pts

Measures: Change in NAV per share since Dec 31, 2025.

Falling NAV means credit losses or over-distribution. Either way, the earnings base is shrinking.

+1% or better25
0 to +1%20
0 to -2%15
-2 to -4%10
-4 to -7%5
Worse than -7%0

Non-accruals

20 pts

Measures: Loans on non-accrual as % of the portfolio at fair value.

Loans that stopped paying interest are tomorrow’s NII shortfall.

Under 1%20
1-2%15
2-3%10
3-5%5
5%+0
Not found10

Leverage

10 pts

Measures: Debt ÷ equity (net of cash where reported).

High leverage leaves less room to absorb losses before covenants or the 150% asset-coverage limit bind.

1.00x or less10
Up to 1.15x8
Up to 1.25x5
Above 1.25x2
Not found5

Earnings quality

10 pts

Measures: Analyst judgment, rules-based.

Catches what ratios miss: spillover cushion, one-time income and announced cuts.

Spillover $0.50+/sh, clean10
Neutral or unknown5
Minor one-time income3
Material one-time income or cut announced0

Ratings

Safe80-100

Covered with room to spare. No sign of stress.

Watch60-79

Covered today, but one factor is slipping.

At risk40-59

Thin or negative cushion. A bad quarter could force a cut.

Danger0-39

Not covered, or a cut is already announced.

Data rules

  • Primary sources only. Company earnings releases, 10-Q/10-K filings and earnings calls. Each ticker page cites its source.
  • Missing data scores neutral. If a figure isn't disclosed, that factor gets its middle score and is flagged "est." on the page.
  • Regular dividend only. Supplementals and specials are excluded from coverage, so a generous extra payout can't hide a thin base.
  • No relabeled cuts. If the base dividend was cut in the last four quarters and a supplemental now keeps the total at 90% or more of the old base, the supplemental is scored as regular. Otherwise a cut could be hidden by renaming part of the dividend.
  • Core NII excludes temporary fee waivers. A waiver that ends on a set date, or is granted quarter by quarter, is stripped from core NII along with other one-time items. Permanent waivers written into the advisory agreement stay in. Waivers under 1c/sh are ignored.
  • Leverage includes all debt. SBA debentures and baby bonds count. Where a company reports a lower regulatory figure, the ticker page shows both.
  • Every score is timestamped. Changes are logged on the track record, including misses: a cut we didn't flag stays on the record.
  • The build checks itself. Scores are computed from inputs by code that mirrors these rules; if a published number ever disagrees with the rules, the site refuses to publish.

Rule changes

  • Nov 19, 2026. Added the relabeled-cut rule and the fee-waiver rule above. Scores published the same day were computed under these rules before going live; the nine Nov 15 scores were unaffected except FSK, whose core NII fell from $0.38 to $0.34 with no change to its score.

Backtest

We scored 46 listed BDCs every quarter from Sep 2022 to Sep 2025 (467 company-quarters, including BDCs that later merged) and checked whether the regular dividend was cut in the following 12 months. 31 cuts occurred. The base rate was 19%.

Rating at the timeCompany-quartersCut within 12 months
Safe648%
Watch2037%
At risk13127%
Danger6949%

At risk and Danger ratings came before 78% of the cuts. A plain payout-ratio test (NII under the dividend) caught 58%. Safe and Watch did about equally well, so the model separates "cut likely" from "cut unlikely" better than it ranks the safest names.

What the backtest can and can't show. It tests coverage and NAV trend only. Non-accruals, leverage and earnings quality aren't in the SEC's machine-readable history, so they scored neutral, as the live model does when data is missing. BDC filings only became comparable in that data around 2021, so the window is short and dominated by the 2024-26 rate cuts. Cuts and regular dividends are inferred from total distributions, which include specials. Quarters overlap, so the 467 observations are not independent. The code is in scripts/backtest.py and reruns from public data.

Limits

v1 thresholds are a first-draft judgment, not a statistically fitted model; the backtest above checks them but did not tune them. A score describes the dividend's safety, not whether the stock is a good buy: price, valuation and your own situation are outside it.

Coverage roadmap

One sector at a time, done right
  1. BDCsNow

    All ~47 listed BDCs. Core NII coverage, NAV, non-accruals.

  2. REITs+3-6 months

    AFFO coverage instead of EPS. Rate-sensitivity stress test.

  3. MLPs+6-12 months

    Distributable cash flow coverage and leverage.

  4. Preferreds & CEFs+6-12 months

    Issuer coverage, call risk, distribution sourcing (return of capital).