What the rule actually selects for is direction, not size or safety. A company that raises its dividend from $2.00 a share to $2.01 has increased it, and its streak continues. A company that holds a generous dividend flat for a single year is removed, even if the payment is larger, better covered by earnings and more likely to persist. The screen is entirely about the shape of a payment history, which is a real signal about management's priorities and a weak signal about a company's financial condition. It says nothing about payout ratio, debt, cash flow coverage or the price paid for the shares.
It is not a yield screen, and treating it as one is the commonest misreading. Nothing in the membership rule mentions the dividend yield. Companies whose share prices have risen faster than their dividends stay in the index with modest yields, and high-yielding companies with a recent cut are excluded regardless of what they pay now. An investor looking for current income and an investor looking for dividend growth want different things, and this index is built for the second.
Equal weighting and the sector cap change the portfolio more than the dividend rule does. Weighting every holding equally means the index rebalances toward whatever has fallen and away from whatever has risen, four times a year, which produces a different return pattern from a capitalization-weighted index regardless of dividends. The 30 percent sector cap is a second structural constraint. Together they make comparisons with the S&P 500 partly a comparison of weighting schemes rather than of dividend policies.
A 25-year requirement is a filter on age as much as on behavior. No company can qualify without a quarter-century of history as a dividend payer, which structurally excludes younger companies, including several of the largest firms in the market. That is a deliberate feature of the rule and it gives the index a persistent tilt toward mature, established sectors. It also means the index is defined by a record, which is the setting where survivorship bias lives: the members are, by construction, the companies that did not fail or cut, and a list assembled that way always looks steadier than the population it came from.
The escape hatch in the rule is worth knowing. Because the index must hold at least 40 companies and no sector may exceed 30 percent of its weight, the methodology relaxes the dividend-growth requirement when those conditions cannot otherwise be met, admitting companies with shorter records. So in a period of widespread dividend cuts, the index that is defined by 25-year streaks will contain members without them.
A long history shown for the index is not a track record in the usual sense. Any performance chart that predates an index's launch is a backtest of the rule applied to past data, and no investor earned it. That is worth keeping in mind separately from the survivorship point, because the two reinforce each other: a rule defined by past durability, tested on the past, will look good on the past.