GhostYield Methodology

GhostYield is a yield sleeve research dashboard for comparing income-producing funds that might sit around an existing portfolio. It helps you line up yield sources, NAV behavior, payout quality, and data quality—not just the headline number.

It is not a full portfolio builder and not a recommendation engine. It does not know your goals, tax picture, or the rest of your holdings.

• Not buy/sell advice or a timing signal
• Not personalized financial, legal, or tax advice
• Not a promise that high yield equals high total return
• Not live market data or a data feed product
• Not a substitute for sponsor documents, SEC filings, tax research, or your own due diligence

1. Yield source matters. The same percentage can mean bonds, options overwrite, leverage, return of capital, or something else entirely.
2. NAV behavior matters. For wrappers with a NAV, the path of net asset value is part of the honesty check on distributions.
3. Data quality matters. A row can be “interesting” and still have stale or incomplete fields—and vice versa.

Yield is not magic. It is usually compensation for some kind of risk.

Risk Score runs from 0–100. Higher means riskier under the economic sleeve / investment rules—not a prediction of the future.

Evidence quality (confidence, freshness, missing expected NAV) is not blended into Risk. It is shown separately as Evidence: Clear / Qualified / Insufficient. When Evidence is not Clear, Risk is still displayed with that qualifier.

The model considers economic / structural factors such as:

• Headline payout level, using current yield when keyed and approved structure-specific distribution-rate fallbacks for CEF and option-income rows
• Sleeve type and structural complexity
• NAV trend versus distributions
• Leverage (including structured BDC debt/equity and CEF effective leverage where present)
• Premium or discount to NAV where applicable
• Payout quality label from the snapshot
• Expense burden (including CEF expense ratio total when structured)
• Optional cefMetrics and bdcMetrics when present on a row

Bands: 0–24 Low · 25–49 Moderate · 50–69 Elevated · 70–84 High · 85–100 Extreme

Risk Score is not a forecast. Think of it as a structured warning system for sleeve and payout stress under the current GhostYield economic rules.

Fit Score runs from 0–100. Higher means a cleaner economic fit as a satellite yield sleeve under this model—not “you should buy this.”

Evidence quality is not blended into Fit. When Evidence is Insufficient, Fit is withheld from display (shown as —) even though an internal economic Fit may still be computed.

The model nudges fit up or down based on things like:

• Clarity / simplicity of the yield story (within the row text)
• Whether headline yield sits in a “reasonable” band versus extreme carry
• Stronger distribution quality labels
• Stable or positive NAV trend where NAV is available
• Discount/premium context for sleeves that price off NAV
• Expense ratio
• Sleeve role (e.g. cash-like ballast vs more complex sleeves)
• BDC dividend coverage and first-lien tilt when structured metrics exist
• CEF discount—only when payout quality is not weak and NAV trend is not badly deteriorating (per model rules)

Bands: 85–100 Strong Fit · 70–84 Good Fit · 50–69 Watchlist Fit · below 50 Weak Fit

High fit does not mean “buy.” It means the row has a cleaner economic profile under the current GhostYield rules and cited snapshot—not a verdict on your personal situation.

Data QA is the freshness/completeness badge for the manual row (fresh / caution / stale / missing / sample), measured against the static snapshot reference date—not today's date.

Evidence is a separate categorical gate derived from existing snapshot fields:

• Clear — fresh + high confidence, no critical expected-field gaps
• Qualified — stale, caution, or medium confidence (usable with caveat)
• Insufficient — missing expected NAV, low/illustrative confidence, or missing/illustrative freshness status

Data QA and Evidence are not investment-risk ratings. They answer snapshot-hygiene questions. Risk and Fit answer economic sleeve questions.

Fresh-in-snapshot does not mean a safe investment. Data gaps do not mean a bad fund.

The screener uses the best available sourced metric, in order:

1. currentYield when set
2. else distributionRate
3. else secYield

• Current yield is usually tied to market price where the source supports it.
• Distribution rate may follow fund or issuer definitions; read the row's source notes carefully.
• SEC yield can be more standardized for many ETFs/funds but may not exist for every CEF or listed BDC row.
• For listed BDCs, NAV-based distribution rates appear in the model with clear labeling in the detail panel when applicable.

Where a fund has a meaningful NAV, GhostYield treats NAV as part of the honesty check on high yield.

• Market price can move because sentiment and flows are noisy.
• NAV summarizes the underlying portfolio value on a per-share basis (per sponsor or filing definitions).
• Premium or discount shows whether the listed price is above or below last keyed NAV.
• A discount is not automatically a bargain; it can reflect real problems or tax complexity.
• A premium is not automatically “bad,” but it raises squeeze and mean-reversion risk in the model.

Some closed-end rows include structured cefMetrics, for example:

• Effective leverage
• Premium / discount (mirrored alongside generic fields)
• Distribution rate and frequency context
• Total expense ratio (structured)
• Coverage ratio, UNII per share
• Managed distribution policy, return-of-capital note (when sourced)

CEFs can combine leverage and managed distribution policies, so headline yield without context can mislead. The detail panel groups these fields under CEF-specific metrics.

Listed BDC rows may include structured bdcMetrics, for example:

• NAV per share and dividend amounts
• NII per share and dividend coverage
• Debt / equity
• Non-accruals
• First-lien exposure and portfolio yield at fair value (when sourced)
• Management structure notes

BDCs are operating lending companies, not generic ETFs. Dividend coverage and credit-quality metrics belong in the conversation alongside headline yield.

The candidate detail panel lists Score drivers: short explanations of selected economic contributors to Risk Score and Fit Score for that row under the current rules.

These drivers explain the model score. They are not buy/sell signals. Evidence gaps are handled by the Evidence gate, not by blending into Risk/Fit drivers.

GhostYield v0.1 ships with manually maintained rows in data/ghostyield/candidates.manual.json. Freshness labels are evaluated against a frozen snapshot reference date—not today's date or a live feed.

• Live price feeds and automated scraping are not part of this release.
• Automated source validation is not running; humans key what the citation supports.
• Unverified values stay null rather than guessed.
• Each row's sourceUrl and sourceLabel explain provenance.
• Some CEF figures cite CEF Connect–style summaries as interim secondary context; always confirm against the fund's own materials.

• Data can go stale the day after it is keyed.
• Sponsor sites change, paywalls happen, and not every field is always available.
• Tax character, ROC breakdown, and full fund coverage usually require documents beyond this page.
• Scores are only as good as the cited snapshot and the current rule set.
• The dashboard does not know your full portfolio, tax rate, liquidity needs, time horizon, or risk tolerance.