9F Dividend Quality Scores : Factor 1 – Dividend Safety

I've been building a new quality scoring system that rates dividend stocks on a 10-point scale across nine weighted factors. Each factor blends qualitative expert assessments from independent sources with quantitative financial metrics, such as payout ratios, growth rates, and leverage. The goal is a single, defensible number that captures a dividend stock's overall quality.

Here are the nine factors that contribute to the quality score of a stock:

1. Dividend Safety
2. Dividend Track Record
3. Financial Strength
4. Earnings & Revenue Growth
5. Profitability & Earnings Quality
6. Risk-Adjusted Returns
7. Credit Rating & Capital Structure
8. Moat & Competitive Advantage
9. Governance & ESG

I'll reveal the specific factor weights in the final article of this series when I assemble the composite score. For now, I want to present each factor individually, one per article, starting with the factor I consider most fundamental: Dividend Safety.

The composite score (out of 10) maps to quality labels with an investment-grade threshold of 6:

My database covers 727 dividend stocks across the eleven GICS sectors. Most of these are dividend growth stocks, as defined by the late David Fish, who created the popular CCC list of Dividend Champions, Contenders, and Challengers. The CCC list includes stocks with dividend-increase streaks of at least five years. My database also includes a few dividend-paying stocks, mainly because I own them in my DivGro Portfolio and want to track them as they become dividend growth stocks.

This article covers Factor 1: Dividend Safety and explains why dividend safety matters, which metrics I use, and how I score them.

The Principles Behind the Scoring

The following design principles shape how every factor in this system operates:

• Qualitative anchors, quantitative depth. Expert assessments from independent sources serve as the anchor for each factor. I use their ratings, grades, and scores as the primary signal, then layer in quantitative financial metrics to add granularity and catch things the qualitative assessments might miss.
• Multiple opinions over single sources. Where possible, I combine three or more independent assessments rather than trusting a single source. If three sources agree that a dividend is safe, I have higher confidence than if only one source says so.
• Sector-relative scoring. Some metrics, such as payout ratios, vary widely across sectors. A Utility paying out 70% of earnings is perfectly normal, whereas a Technology company at 70% would be alarming. To accommodate such variation, I use a blend of absolute and sector-relative scores: 70% of the score is based on a piecewise-linear mapping of the metric's actual value. In comparison, 30% is based on the metric's sector percentile.
• Graceful degradation. When a metric is missing, the remaining metrics automatically absorb its weight, with a small coverage penalty applied. This means every stock gets scored, even those with incomplete data, but stocks with fewer data points receive slightly lower scores to reflect my reduced confidence.
• Penalty-only modifiers. One of my safety sources has a weak correlation with the other safety metrics. Rather than dropping it entirely, I use it as a penalty-only modifier: a bad grade can hurt, but a good grade cannot help. If a metric isn't reliable enough to reward, it can still serve as a red flag.

Why Dividend Safety Matters

Dividend cuts are devastating for income-focused portfolios. A cut doesn't just reduce your income stream. It typically coincides with a steep price decline as the market reprices the stock. The double hit of lost income and capital loss is exactly what dividend growth investors are trying to avoid.

Safety is also a forward-looking signal. A company maintaining a safe dividend today is implicitly telling you that management has confidence in future earnings and cash flow. The dividend is a commitment, and the degree to which a company can comfortably cover that commitment tells you a lot about its financial health.

For income-focused investors, especially those in or approaching retirement, avoiding cuts matters more than chasing yield. A 2% yield that grows reliably for 20 years will outperform a 6% yield that gets cut in half after three years.

The Metrics I Use

Factor 1 consists of two parts: a qualitative sub-score (0-5 points) assigned through expert safety assessments and a quantitative sub-score (0-5 points) derived from payout ratio analysis. These combine to form a total Factor 1 score ranging from 0 to 10 points.

Qualitative Metrics (max 5 points)

I use three primary sources plus one penalty-only modifier.

Simply Safe Dividends [SS] Dividend Safety Scores

Simply Safe Dividends provides the most comprehensive dividend safety assessment I've found. Their scores range from 0 to 100, blending payout ratios (trailing and forward), industry-specific metrics (AFFO for REITs, DCF for MLPs), off-balance sheet liabilities, and forward analyst estimates. Analysts review ratings after quarterly earnings or when a change in a company's fundamentals alters its dividend risk profile.

I give SS the highest weight at 0.45 and map scores using piecewise linear interpolation with an inflection point at 50 (my safety threshold):

The ramp from 80 to 99 is more gradual, reflecting that the difference between 80 and 99 is meaningful but less dramatic than the difference between 40 and 50.

Sure Dividend [SD] Dividend Risk Score

Sure Dividend assigns A-F letter grades based on a fully transparent, published formula. The score is based on three components: a conservative growth estimate, dividend increase streak, and payout ratio. They use EPS for stocks, FFO for REITS, and DCF for MLPs.

I give SD a weight of 0.15 and map letter grades using piecewise linear interpolation:

The rapid decline between C and D is intentional. A C-rated stock still gets half-credit, but a D drops to near-zero, reflecting Sure Dividend's interpretation that a D signals genuine dividend risk.

Seeking Alpha [SA] Dividend Safety Grade

Seeking Alpha's quant system assigns A+ through F grades based on 13 underlying financial metrics, including leverage, cash flow coverage, institutional ownership, sustainable growth rate, and trends in DPS revisions. All comparisons are relative to sector peers. The system is fully quantitative with no human analysts in the loop. 

I give SA a weight of 0.40 and map letter grades using piecewise linear interpolation:

The A and B tiers are compressed (with 0.25 spacing), reinforcing SA's finding that 98% of dividend cuts were averted by owning stocks with Dividend Safety Grades from A+ to B−. Cliffs separate B− from C+ and C− from D+, representing a rapid deterioration of safety. 

Finbox [FB] Cash Flow Grade

Finbox assigns A-F cash flow health grades by ranking a company's cash flow metrics against sector peers in the same economic risk region. It's one component of their broader financial health analysis.

In my system, this grade is a penalty-only modifier. Its low correlation with the other safety metrics (r = 0.188) makes the metric unsuitable for boosting scores. However, in 92% of cases where FB disagreed with the other sources, it was the low outlier, making it useful as a red flag:

How the Qualitative Sources Combine

I compute a weighted average of the scores from available sources and apply a 10% penalty for missing data. The FB penalty, if triggered, is subtracted from the result and floored at zero.

The maximum qualitative sub-score is 5.0 points.

Coverage

Here's the coverage of each source:

• SS Dividend Safety Scores: 499 of 727 tickers (68.6%)
• SD Dividend Risk Score: 688 of 727 tickers (94.6%)
• SA Dividend Safety Grade: 649 of 727 tickers (89.3%)

The combined coverage is 717 tickers (98.6%), so only 10 tickers (1.4%) lack any expert safety signal. For these stocks, Factor 1 relies entirely on quantitative metrics (presented in the next section). All three sources cover 452 stocks (62.2%). Those are the stocks that can earn the maximum qualitative sub-score of 5 points. 

Quantitative Metrics (max 5 points)

I score two metrics and then scale their weighted average to 5 points. For all quantitative metrics, I use the 70/30 blend of absolute and sector-relative scores mentioned earlier.

Payout Ratio (weight 0.45)

The earnings payout ratio measures the percentage of earnings paid out as dividends. For REITs, I use the AFFO or FFO payout ratio rather than the earnings-based payout ratio. For BDCs and MLPs, I use AFFO if available, or the earnings payout ratio if it's below 1.0; otherwise, I skip it altogether. There are only 2 BDCs and 8 MLPs in my database.

For 70% of the score, I use the following piecewise linear mapping of the payout ratio:  

The remaining 30% of the score is based on the stock's percentile rank within its GICS sector. Higher percentiles mean lower payout ratios and, therefore, higher scores.

FCF Payout Ratio (weight 0.55)

The free cash flow payout ratio measures the percentage of free cash flow paid out as dividends. FCF payout has slightly greater weight than earnings payout because it's arguably the more forward-looking metric: free cash flow reflects actual cash generation after capital expenditures.

For 70% of the score, I use the following piecewise linear mapping of the FCF payout ratio, with breakpoints tuned for FCF's naturally wider distribution: 

For negative FCF payout ratios (the company is burning cash), a stock receives an absolute score of 0.0, though the sector percentile component still contributes to the blended score.

How the Quantitative Sources Combine

I compute the weighted average of the Payout Ratio and FCF Payout Ratio, and apply the same 10% penalty for missing data. The result is then scaled to 5 points to obtain the quantitative score.

Combining the Sub-Scores

Calculating the final Factor 1 score is straightforward:

• If both sub-scores are available: qual₁ (0-5) + quant₁ (0-5) = Factor 1 score ranging from 0 to 10.
• If only one sub-score is available, it is multiplied by 1.9 rather than doubled, which enforces the 10% penalty for missing data.  

This means a stock with only qualitative data can still score up to 9.5 (5.0 x 1.9), but never a perfect 10. Full marks require both halves.

Finally, I calculate a factor confidence score: conf₁ = number of metrics ÷ 5.

Results: How the Scores Look

Let's now look at how the Factor 1 scores look for the 727 stocks in my database.

Distribution

Here is a chart showing the distribution of Factor 1 scores:

The distribution has a satisfying spread. It's slightly skewed to the high side with the median above the mean. With 13.1% of stocks scoring 9-10, the system identifies genuine safety leaders without being overly generous. At the other end, 14.5% of stocks score below 4.0, flagging real dividend risk.

The Safest Dividend Stocks by Sector

Here are the highest Factor 1 scores in each GICS sector, with a full breakdown of how they got there: 

A few things jump out:

• WST and HEI both got a perfect 10.0. These companies have SS=99, SD=A, SA=A+, and near-zero payout ratios (4.7% for HEI, 12% for WST), placing them at the top of their sector's percentile ranks. Both businesses retain nearly all of their earnings.
• The qualitative sub-score is the primary differentiator among sector leaders. Eight of eleven sector leaders achieve a Quant score of 5.0, indicating the quantitative ceiling is easy to hit for low-payout companies. The ranking within a sector is almost entirely determined by whether the safety rating agencies (SS, SD, SA) cover and rate the stock highly.
• TPB tops Consumer Staples, not KO, PG, or CL. TPB wins on ultra-low payout (9.5% EPS, 12.8% FCF) and A-grades from both SD and SA. The well-known Consumer Staples stocks have much higher payout ratios (60-70%+), which compresses their quantitative sub-score even with excellent qualitative coverage.
• GOOG leads Communication Services with partial coverage. GOOG only initiated dividends in 2024, so SD and SA have no coverage yet (confidence = 0.667). Its 7.8% EPS payout ratio is the lowest in the sector by a wide margin. If SD and SA fill in with A-grades, GOOG could approach 9.5+.
• VST illustrates the Utilities tension. VST wins Utilities but with a notably lower F1 (7.98) than leaders in other sectors. The FCF payout percentile of 0.185 is the key drag: Vistra generates significant FCF, but its 30.7% FCF payout ranks poorly within a sector where many utilities have much lower FCF payouts. Traditional regulated utilities like SO or D likely score worse on absolute payout ratios despite their reputation for safety.

Worked Examples Across the Quality Spectrum

Here I show ten stocks ranging from the highest composite quality scores down to the mid-range, with their Factor 1 breakdown:

Note the following:

• The qual/quant divergence tells the story. MA, MSFT, and NVDA score well on both halves. But from CTAS downward, the two sub-scores decouple sharply.
• NVDA has the second-highest F1 in the entire table at 9.66, higher than MSFT. This is entirely mechanical: a 0.9% EPS payout and 1.0% FCF payout put it at the 100^th and 97.5^th percentiles within IT. NVDA barely pays a dividend, so the safety factor treats it as nearly bulletproof. The high F1 is not a judgment that NVDA is a dependable dividend grower. It reflects that when you pay almost nothing out, you won't be forced to cut it.
• LLY and JNJ show the Health Care payout squeeze. Both are elite companies (SS=96 for LLY, SS=99 for JNJ) with near-perfect qualitative scores. But LLY's FCF payout of 75.2% sits at only the 8.6^th percentile within Health Care, and JNJ's 63.8% FCF payout is at the 20^th percentile. Pharmaceutical R&D spending compresses reported FCF, dragging the quantitative half down even for the sector's best names.
• KO is the most striking case. Arguably the canonical dividend growth stock (62-year streak, AA− credit, SD=A), yet F1 = 4.83 falls below the 6.0 investment-grade threshold. The entire problem is the FCF payout at 165.9%: Coca-Cola distributes more in dividends than it generates in free cash flow, relying on debt capacity and the certainty of its cash flows to sustain payments. The system scores this as near-zero quant (0.76) because it evaluates coverage ratios at face value. Whether that's conservative wisdom or mechanical blindness depends on your view of KO's moat, which is why Factor 8 (where KO scores 9.30 on a Wide moat) exists.
• DPZ is a rare case where quant outscores qual. Its SS=55, SD=B, SA=D+ reflect genuine concern about the heavily leveraged balance sheet. But its 39.9% EPS and 35.4% FCF payouts are mid-pack for Consumer Discretionary, not alarming. The market and rating agencies are nervous about the debt structure, whereas the payout ratios themselves are moderate.

REITs, Utilities, and the Sector Adjustment

These sectors deserve special attention because their payout characteristics differ structurally from the market:

Here are some observations about these stocks:

• EQIX demonstrates the AFFO substitution working as intended. Its 52.9% AFFO payout is well-covered, and at the 91.7^th percentile among REITs, it scores strong quant (3.82). The qualitative half is also solid at 4.35.
• O and MAA both score below 5.0 despite excellent SS scores (80 and 90, respectively). The culprit is the SD=D grade, which maps to 0.5 on a 0-10 scale and significantly drags down the qualitative weighted average. Sure Dividend's formula penalizes REITs with high payout ratios and modest growth, even when those payouts are structurally normal for the REIT model.
• ED shows the 70/30 blend at work. Its 59.2% earnings payout scores moderately on the absolute piecewise function. But the 57.4^th sector percentile places ED just above the midpoint for Utilities, giving it a modest boost from the sector-relative component. Without the blend, ED's payout would look worse against the broader market; within Utilities, it's unremarkable.

Stocks That Score Poorly

Some stocks look bad under the Factor 1 lens:

One additional note:

• DKL is an interesting case. As an MLP with pass-through distributions, its 134.9% earnings payout is structural and gets skipped. But no AFFO or FFO data are available, so the only quantitative input is FCF payout. Combined with SS=24, SD=F, and a CF penalty for a D grade, the qual score drops to 0.00, and the overall F1 is 0.72 out of 10.

What's Next

Dividend Safety is one of nine factors that will combine into the final composite quality score. Next in the series, I'll cover Factor 2: Dividend Track Record, which evaluates how long and how consistently a company has grown its dividend.

The final article will reveal the specific weights assigned to each factor and the distribution of the overall ranking of the 727 stocks. Until then, these factor-by-factor articles are building the foundation so that when the full system comes together, every component will already be familiar.