Sort a stock screener by dividend yield, descending. The top of that list is almost never a list of the best businesses. It's a list of companies whose share price has fallen furthest, fastest. That's not a quirk of the screener. It's arithmetic, and it's the single most misread number in income investing.
Yield is a ratio. Annual dividend divided by share price. There are two ways for it to go up, and only one of them is good news.
The denominator does most of the work
Say a company pays $2.00 a share per year and trades at $50. That's a 4% yield. Nothing changes about the dividend, but the stock drops to $25 because a lawsuit landed or a segment stopped growing. The yield is now 8%.
Yield = annual dividend / share price
$2.00 / $50.00 = 4.0%
$2.00 / $25.00 = 8.0% <- same dividend, worse companyThe screener now ranks this stock twice as attractive as it did last quarter, on the strength of the thing that should worry you. This is the pattern behind what income investors call a yield trap: cash generation weakens, the payout ratio climbs, confidence erodes, the price falls, and the falling price manufactures a headline yield that pulls in buyers right before the dividend gets cut.
It's been a live pattern through 2026 across mortgage REITs, office REITs, business development companies, and parts of telecom and pharma. The shape is consistent. The payout stops being covered by cash, the market prices that in, and the yield looks its best in the weeks before the announcement that ends it.
The ex-dividend date is not free money
A recurring idea for new dividend investors is dividend capture: buy the day before the ex-dividend date, collect the payout, sell after. It's a clean-sounding trade that mostly doesn't work.
On the ex-dividend date, the stock generally opens lower by roughly the dividend amount, because anyone buying that morning is buying a share that no longer carries the pending payment. A $0.30 dividend takes about $0.30 out of the price. You haven't gained anything, you've moved value from one column to another and picked up a taxable event on the way.
The mechanics are worth knowing anyway, just for settling trades correctly. You have to own the shares before the ex-dividend date to be on the record. Under T+1 settlement, that means buying at least one business day ahead of it, not on it.
For someone holding a position for years, none of this matters. The ex-date wobble is noise. It only becomes a real cost if you're trading around it.
Coverage is the number that decides everything
The question a yield can't answer is whether the company can keep paying. Two metrics get you most of the way there.
The payout ratio is dividends divided by earnings. Somewhere in the 40-60% range is generally considered comfortable for a mature company: enough margin that a bad quarter doesn't force a decision. A 5% yield with a 50% payout ratio and a 5% yield with a 95% payout ratio are entirely different securities. The second one has no slack.
Free cash flow coverage is the harder test, and the more honest one. Earnings can be managed. Cash is cash. If dividends paid have been running ahead of free cash flow for several quarters, the payout is being funded by the balance sheet, by debt, or by asset sales. That works until it doesn't. Layer high leverage and near-term debt maturities on top and the cut becomes a scheduling question rather than a possibility.
The tax layer nobody screens for
Dividends land in one of two buckets. Qualified dividends are taxed at long-term capital gains rates, currently 0%, 15%, or 20% depending on income, with an additional 3.8% net investment income tax above certain thresholds. Ordinary (non-qualified) dividends are taxed as ordinary income, which for most people is a materially worse rate.
The qualification test has a holding period attached: you generally need to have held the stock more than 60 days during the 121-day window that starts 60 days before the ex-dividend date. Count the day you sold, not the day you bought. This is the other reason dividend capture struggles. The strategy that gets you in and out around the ex-date is the one that guarantees the payout is taxed at the worse rate.
None of this makes dividends bad. Cash returned to shareholders is a real thing, and a long record of raising it is a genuine signal about how a business is run. But the yield itself is a ratio between two numbers, and the one you should be looking at is the one underneath the line. Before the yield, pull up the payout ratio, the free cash flow trend, and the dividend history. If the number looks too good, it's usually the price telling you something the payout hasn't caught up to yet.