7 Questions Before You Trust a Dividend

A dividend looks like a fact. It sits on the brokerage screen next to the price, quoted to two decimal places, paid on a date you can put in a calendar. It feels more solid than an earnings estimate or a price target.

It is not a fact. It is a promise, and a company can break it at any board meeting. In 2009, about one dividend payer in five cut its payment, including household names with decades of increases behind them.

Investors who got hurt were rarely the ones who failed to predict a crisis. They were the ones who bought a yield without asking what stood behind it. Here is what stands behind it, in seven questions.

Question 1: Is the yield high because the dividend is good, or because the price fell?

Yield is a fraction. The dividend sits on top, the price sits underneath. So a yield rises for two opposite reasons: the company raised its payment, or the market marked the shares down.

Only one of those is good news, and the second is far more common at the top of a yield screen. A stock that falls 60% while holding its dividend steady shows a yield nearly three times higher than a year ago. Nothing improved. The market is pricing in a cut, and the screen reports that fear back to you as an opportunity.

In practice: CenturyLink, now Lumen, paid $2.16 a share every year from 2016 to 2018 without raising it once. As the shares fell, the yield climbed from about 11% to 14%, and it sat near the top of every high-yield screen. In early 2019 the company cut the payment by 54%, to $1.00. It stopped paying altogether after 2022.

What good looks like: a yield that rose because the payment rose. Put dividend per share and share price on the same ten-year chart and see which one moved.

Question 2: Can it pay for the dividend out of what it earns?

The payout ratio is the dividend divided by earnings. At 30%, a company keeps 70 cents in every dollar for reinvestment, debt repayment, or a bad year. At 95% there is nothing spare, and the first earnings stumble comes out of your income.

Under 60% is the usual comfort zone, but the number means different things in different places, which is where simple screens go wrong. Utilities are built to distribute most of what they earn against regulated, predictable revenue. Cyclicals look safest at the top of a cycle, when peak earnings flatter the ratio right before earnings fall. And REITs are the biggest trap of all: they must distribute most of their taxable income, and accounting depreciation on buildings that are not really wearing out pushes reported earnings down, so the ratio reads as alarming when the business is fine.

In practice: Realty Income has raised its dividend every year for 31 years. Measured against earnings per share, its payout ratio was over 300% in 2024, a number that would look like a company paying three times what it makes. Measured against funds from operations, the figure REITs are actually judged on, it is about 82%. Same company, same dividend, completely different verdict, and the only thing that changed was the denominator.

What good looks like: a payout ratio that is normal for that kind of business, measured against the right earnings figure.

Question 3: Does the cash back up the earnings?

Earnings are an accounting result. Dividends are paid in cash. When the two diverge for long, the payment is being funded by something other than the business: the balance sheet, asset sales, or borrowing.

Compare free cash flow with the total dividend paid. A company generating $3bn and paying out $1bn has room. One generating $900m and paying $1bn is covering the gap somewhere, and every source of that gap is finite.

In practice: Exxon Mobil lost $22bn in 2020 and its free cash flow turned negative, yet it kept paying about $3.48 a share. The money came from the balance sheet: total debt went from $47bn to $68bn in a single year, an extra $21bn. That was a deliberate choice by a company with the capacity to make it, and it worked out, as oil recovered and Exxon has now raised its dividend for 42 straight years. Smaller producers made the same choice in the same year and cut anyway.

What good looks like: dividends comfortably covered by free cash flow across a full cycle, not only in the good years.

Question 4: Does the balance sheet leave room?

Dividends are the last thing paid and the first thing cut. Suppliers, staff and lenders are paid first, and none of those payments are optional. Yours is.

That ordering makes the balance sheet a dividend question. Interest cover (operating profit divided by interest expense) and net debt against EBITDA tell you how much room there is between a bad year and a hard choice. A company covering interest twelve times over can absorb a downturn without going near the dividend. One covering it twice is one bad year away from choosing between its lenders and its shareholders, and that is not a real choice.

In practice: Boeing paid $8.22 a share in 2019, having raised the dividend every year for the previous decade. It paid one quarter in 2020 and then suspended the dividend entirely. The company had spent years buying back stock, the 737 MAX grounding and the travel collapse hit a balance sheet already carrying heavy debt, and shareholders were the first thing cut. Six years on, they have not been paid again.

What good looks like: interest comfortably covered, debt maturities spread out, and no refinancing wall in the next couple of years.

Question 5: How long has it raised the payment, and what was it raising through?

A long record of increases is the most-quoted dividend metric, and the most useful thing about it is behavioural. Management teams know that breaking a thirty-year streak becomes the headline, so they defend it, often at real cost. That determination is what you are buying.

But a streak only tells you what a company survived, and a decade of good conditions is not a test. A record built through a recession, a rate shock, or a collapse in the company's own end market means something. A record built through a benign stretch for that particular industry means considerably less, and the two look identical on a screen.

In practice: in 2007 plenty of large banks carried long, unbroken records of dividend increases. Bank of America had raised its payment for decades, then cut it to a penny a share in 2009. The streak was real, but it had been earned across a credit cycle that never got tested, so it measured the era rather than the durability of the business.

What good looks like: a long record that includes at least one genuine crisis for that company's own industry.

Question 6: Has it cut before, and what happened afterwards?

A past cut matters for two reasons. It tells you what a board does under pressure, and it tells you what recovery looks like, which is slower than most income investors expect.

A cut is not a bad quarter. The income drops immediately, the share price usually drops with it, and getting back to the old payment routinely takes the better part of a decade. Some companies never do.

In practice: General Electric cut its dividend in 2009, cut it again in 2017, then cut it to a token penny a share in 2018. The payment fell by about 95% in two years. GE has since recovered as a business and is raising again, but the dividend today is still roughly a third of what it paid in 2017. An income investor who held through it waited eight years to get a third of their income back.

What good looks like: no cut in the recent past, and if there was one, a clear account of what changed in the business since.

Question 7: Is the dividend growing, or merely surviving?

A dividend that never rises is a dividend that shrinks. At 3% inflation, a payment held flat for a decade loses about a quarter of its purchasing power, and a company that cannot raise its payment is usually telling you something about its growth as well.

Raising is also the cleanest confidence signal a board can send, because it commits the company to funding the higher number every year from here. Boards do not do that casually, which is why a rising payment carries more information than a high one.

In practice: Verizon and Snap-on have both raised their dividends for about two decades, so both clear question 5 and both appear on the same "dividend grower" screens. Over the past ten years Verizon's payment has grown about 2% a year, roughly in line with inflation and no better. Snap-on's has grown about 15% a year, so its payment has quadrupled. Same label, entirely different outcome for the income.

What good looks like: dividend per share growing over five and ten years, and growing no faster than earnings, because a payment outrunning profits is just the payout ratio climbing towards a ceiling.

Putting the seven together

No company clears all seven perfectly, and they do not carry equal weight. The safety questions matter more than the yield question that most screens lead with, because the downside is asymmetric. A yield one point lower costs you a little every year. A cut costs you the income and the capital in the same week, and can take a decade to undo.

Read in order, the seven answer a single question: if the next two years are bad, does this dividend survive?

Answering all seven without doing it by hand

Assembling this for one company takes an afternoon of filings. Doing it across a thousand payers is what a tool is for, and it is what our dividend scoring does.

The dividend strategy scores every dividend payer in the universe on four pillars, Yield, Growth, Safety and Consistency, and ranks the market on the blend. The pillars map onto the seven questions directly:

Question	Where it is answered
1. Is the yield real?	Yield, which scores a payment highest in the 2% to 6% range and marks down unusually high yields rather than rewarding them
2. Covered by earnings?	Safety, using payout ratios with sector adjustments, and funds from operations rather than EPS for REITs
3. Covered by cash?	Safety, using free-cash-flow coverage, with earnings coverage as a fallback where cash-flow data is missing
4. Balance-sheet room?	Safety, reading current ratio, debt to equity, and interest coverage
5. Length of the record	Consistency, counting consecutive increases with no cap, so a thirty-year record scores on its full merit
6. Past cuts	Consistency, where a recent cut disqualifies the pillar outright
7. Growth	Growth, the heaviest pillar by default, reading multi-year dividend growth

Two things about it are worth knowing. The weights are a disclosed starting point, not a verdict: if you think safety should outrank growth, or that a 2% yield is not worth owning, you move those levers and the ranking rebuilds around your definition. And the filters are separate from the weights, so you decide what qualifies for the list before you decide what ranks highly on it.

If you would rather start from a finished list than build one, two of the preset screens are dividend-specific. High Yield, Cash-Covered takes questions 1 to 3 and turns them into a single filter: a yield above 2%, a payout ratio between 20% and 75%, free cash flow yield above the dividend yield, and a quality floor underneath. Dividend Growth comes at it from question 7, looking for a payout under 50% alongside a high return on equity, on the logic that a company retaining most of its earnings has room to keep raising.

For a single company rather than a ranked list, the dividend section of any stock's Quality page carries the same inputs: yield, payout ratio, coverage, growth rate, and the payment history behind the streak. The Apple page opens without an account, if you want to see the shape of it first.