A dividend is a promise the company can break any quarter, and many do. Investors who chase yield without checking the math behind it tend to find this out the hard way. The good news is that the signals that predict a dividend cut are mostly arithmetic, and you can compute them on a spreadsheet.

What actually makes a dividend safe

Three things drive whether a company can keep paying. First, payout ratio: how much of earnings the dividend eats. A 30 percent payout has plenty of slack. A 95 percent payout has none, and a 110 percent payout is funded out of the balance sheet or new debt.

Second, dividend history. Companies that have raised payouts for 25 years through recessions have shown that management cares about the dividend. A 35 year streak survived the dot com crash, the 2008 banking crisis, and the 2020 lockdowns. That is hard to fake.

Third, expected growth. If earnings per share are shrinking, today’s payout ratio gets worse on its own without anyone deciding anything. A 70 percent payout becomes a 90 percent payout in two years if profits drop fast enough.

If all three numbers look good, the dividend is probably safe. If two of three look bad, it is probably not.

How a risk score actually computes

A common scoring approach takes the three signals above, turns each into a number in roughly the same range, then averages them.

• Growth piece: take the lower of expected earnings growth and expected dividend growth, multiplied by 1,000. A company growing at 7 percent annually scores 70 on this leg.
• History piece: years of consecutive dividend growth, multiplied by 2. A 35 year streak yields 70.
• Payout piece: 100 minus the payout ratio in percent. A 30 percent payout yields 70.

Average those three. A score around 60 or higher is investment grade. Below 30 is risky. The exact thresholds vary by methodology, but the structure is similar across most published frameworks.

Why the constants? Growth rates are tiny decimals. History streaks are small integers. Payout ratios are percentages. The multipliers put all three legs on a comparable scale, so no single piece swamps the others. Frameworks that skip this step end up scoring almost entirely on whichever input has the widest natural range.

Why systematic ranking beats picking favorites

Peter Lynch once said the investor who turns over the most rocks wins. That sounds folksy, but it is a statement about sample size. If 10 percent of dividend payers are attractive at any given time, you find one in a watchlist of ten and ten in a universe of a hundred. Screening every payer on the same metrics surfaces candidates that gut feeling misses.

Systematic ranking also removes the recency bias that wrecks discretionary investors. A stock that just doubled feels safer than one that just halved. The numbers do not care. A company with a 95 percent payout ratio and three years of history is risky whether the chart is green or red.

This is the same logic credit rating agencies use on bonds. Nobody upgrades a junk bond because they like the CEO. They run the math.

Why REITs and MLPs need different math

A retail investor screening REITs (real estate investment trusts) by earnings per share will reject almost all of them. Reported EPS is distorted by depreciation on buildings that are appreciating in market value. The right metric is funds from operations, written as FFO, or adjusted FFO.

MLPs (master limited partnerships) are worse. Distributable cash flow per unit is the right number, not EPS. The accounting layer hides the economic one.

Any framework that scores income securities on the same metrics needs adjustments for these structures, or the rankings come out wrong. This is why throwing a generic stock screener at the question is dangerous. The screener computes the same payout ratio for an industrial firm and a REIT, which is meaningless for the REIT.

Total return is more than the dividend

A dividend yield is one of three sources of return. The other two are growth in earnings per share and the change in the price to earnings multiple. Add them up and you get an estimate of total return over the next five years.

A 4 percent yield with 6 percent earnings growth and a flat valuation is a 10 percent annualized return. A 7 percent yield with no growth and a slowly contracting multiple is closer to 4 or 5 percent. The headline yield is not what you take home.

This is why dividend cuts hurt twice. The cash flow shrinks, and the multiple compresses because the market reprices the stock as lower quality. A 30 percent price drop on a name yielding 6 percent wipes out five years of dividends in a day.

What to watch

For investors building dividend income for the long run, three rules cover most of the risk.

• Reject any payer with a payout ratio above 80 percent on the right metric for the structure (EPS for industrials, FFO for REITs, DCF for MLPs). The remaining 20 points of slack are the margin for a bad quarter.
• Reject any payer with a frozen or shrinking dividend. The point of holding dividend stocks is the growth. A flat dividend has the worst of both worlds: equity risk plus a yield that loses ground to inflation.
• Prefer companies that have raised through at least one recession. Two decades of growth covers 2008 and 2020. That is harder to fake than a short streak in a calm market.

None of this guarantees anything. Several large banks cut dividends in 2008 after raising them for decades. Major energy producers cut in 2020. But probability shifts when the math points the same way for years, and a portfolio built on those probabilities does better than one built on yield headlines.