The Dividend Aristocrats That Are Quietly Shrinking
3M has paid a dividend for 56 consecutive years. Its dividend is 9.9% a year smaller than it was five years ago.
Both statements are true, and only the first one appears on any list of dividend aristocrats. Across the 364 companies on DividendAtlas with an unbroken payment record of 25 years or more, 21 are paying less today than they were five years ago. About one in seventeen. Drop the bar to a 10-year record and the count is 39.
Why a screen cannot see this
The number that builds every aristocrat list counts the wrong thing.
An unbroken record counts years in which a dividend was paid. It does not count years in which the dividend rose, and it says nothing at all about the amount. A company can cut its payment by 90% and extend its streak in the same year, because it still paid something.
That is not a flaw anyone hid. It is what the definition says, and it works perfectly well as a measure of what it actually measures: survival. The problem is that it gets read as a measure of something else, namely a dividend that keeps growing, and those two things come apart more often than the lists suggest.
Our own European dividend aristocrats list uses consecutive years of growth for exactly this reason, which is a stricter and much shorter list than the payment-record version.
Two long records, both going backwards
Leggett & Platt and International Flavors & Fragrances are both scored on the
Cyclical profile, so our score judges them on the same terms.
| Leggett & Platt | Int'l Flavors & Fragrances | |
|---|---|---|
| Years paid without interruption | 41 | 52 |
| Dividend growth, 5y | -34.0% a year | -12.1% a year |
| Dividend per share, FY2023 | 1.82 USD | 3.24 USD |
| Dividend per share, current year | 0.20 USD | 1.60 USD |
| Free cash flow per share, 3y average | 2.15 USD | 2.34 USD |
| Dividend Health Score | 52, borderline | 43, borderline |
All figures as of 29 August 2026.
Leggett & Platt Incorporated (LEG.US)Data as of 2026-08-29- Price
- $9.20
- Dividend yield
- 2.17%
- Annual dividend
- $0.20
Key statistics
Leggett & Platt has paid without interruption for 41 years. The payment went from
1.82 USD in FY2023 to 0.61 USD, then to 0.20 USD, where it has stayed. That is a
fall of about 89% in two years, inside a record that never broke. Its score of 52
sits in the borderline bucket, with high confidence.
International Flavors & Fragrances Inc (IFF.US)Data as of 2026-08-29- Price
- $86.85
- Dividend yield
- 1.84%
- Annual dividend
- $1.60
Key statistics
International Flavors & Fragrances has the longer record at 52 years and the
gentler decline. Its dividend halved once, from 3.24 USD in FY2023 to 1.60 USD,
and has held there since. The five-year growth rate of -12.1% a year is the
arithmetic of that single step spread across the window, not a slow bleed. It
scores 43, also borderline.
The part most coverage gets wrong
A cut is bad news for the shareholder who owned it. It does not follow that the dividend is now more dangerous. Often it is considerably less so.
Look at what Leggett & Platt pays now against what it generates: 0.20 USD a share, against 2.15 USD of free cash flow per share averaged over its last three completed financial years. Whatever pressure produced that cut, the payment left standing is covered roughly ten times over. The same holds at International Flavors & Fragrances, at 1.60 USD against 2.34 USD.
This is why both still score in the middle of the range rather than at the bottom. Our score penalises a cut through the track record, and it reads current coverage too. A company that rebased two years ago and now covers its payment comfortably is in a genuinely better position than one still paying more than it earns while its streak stays intact. Ranking those two by streak length puts them in the wrong order.
The uncomfortable implication is that the aristocrat screen is at its least reliable exactly where it feels most authoritative. A 50-year record tells you the company has never missed. It does not tell you the payment is the same size, that it is growing, or that it is covered.
The most famous case
3M Company (MMM.US)Data as of 2026-08-29- Price
- $174.12
- Dividend yield
- 1.79%
- Annual dividend
- $3.12
Key statistics
3M is scored on the Standard profile rather than the
Cyclical one above, so it is not a like-for-like comparison with those two. It
is here because it is the record everyone knows: 56 consecutive years, one of the
longest anywhere.
The dividend went from 5.02 USD in FY2023 to 3.36 USD, then 2.92 USD, and stands at 3.12 USD for the current year. So the payment is down about 38% from its peak, and it rose again this year. A dividend that falls and then resumes rising is a different animal from one still falling, and the five-year growth rate of -9.9% a year cannot distinguish them. That is a real limit of the metric, not a criticism of the company.
3M also shows a payout ratio of 980%, which needs reading rather than reacting to. That figure divides the dividend by an average of GAAP earnings across three completed financial years, and those earnings were pushed close to zero by accounting charges. Free cash flow per share averaged 4.30 USD over the same window against a 3.12 USD dividend. The cash covers it; the earnings line does not. Our guide to a payout ratio above 100% works through how to tell those two situations apart.
The window you pick decides the answer
There is a second problem underneath the streak, and it affects the growth rate that was supposed to fix it.
A five-year growth rate is a comparison between two years, and moving either end changes the verdict. International Flavors & Fragrances above is a clean example: its dividend growth reads -12.1% a year over five years and +4.0% a year over twenty. Both figures are correct. The company really has compounded its dividend upward across two decades, and it really is paying half what it paid three years ago. A site quoting one number without saying which window it used has told you almost nothing.
That is not an isolated case. 18 companies on DividendAtlas show a negative five-year dividend growth rate alongside a positive twenty-year one. Screened on the long horizon they are growers. Screened on the short one they are shrinking.
The mirror image is more common and more dangerous. 31 companies show a positive five-year growth rate alongside a negative ten-year one. In those cases a cut has simply aged out of the shorter window, and what remains is the recovery measured from the low point. The most extreme reads +43.1% a year over five years and -16.6% over ten. Nothing about that company improved by 43% a year; the starting line moved.
So the growth rate is a better field than the streak, and it is not a verdict either. The pair of them, read with the actual payment history, is what tells you the shape. A dividend that fell and stayed down looks nothing like one that fell and recovered, and neither looks like one that never fell. All three can produce the same single growth number.
What the 39 look like as a group
They are not concentrated in one corner of the market.
The 39 companies with a 10-year-plus record and a shrinking dividend break down as 11 REITs, 10 cyclicals, 9 defensive businesses, 6 standard ones, 2 insurers and one the score does not classify. No sector explanation covers it, which is what you would expect if the cause is company-specific rather than structural.
Nine of the 39 are still rated safe or very_safe today. That is the rebasing
effect above, showing up as a number: roughly a quarter of the companies whose
dividend has shrunk are nonetheless well covered at the level they now pay.
Why we publish both numbers
This is the reason our instrument pages carry years paid and years increased as two separate fields rather than one figure called "streak".
They are genuinely different facts. Across the payers on DividendAtlas that publish both, they disagree for the large majority, and the median gap where they differ is 17 years. Collapsing them into a single headline number would mean picking one, and whichever we picked would be wrong for most of the corpus. A 41-year payment record beside a shrinking payment is an awkward pair of facts to put on one page, and putting both there is the only honest option.
The same reasoning applies to the growth rates. We publish five, ten and twenty year figures rather than a single "dividend growth" number, and we blank a horizon rather than filling it from a shorter window when the history does not reach back far enough. A company with eight years of completed history gets a five-year figure and no twenty-year one, instead of the same eight-year number printed three times under three different labels.
What to check instead
Three things, none of which takes long.
- Compare years paid against years increased. They are different fields and they disagree for most payers. A record quoted without saying which one it is should be treated as the flattering one.
- Read the five-year dividend growth rate. If it is negative, the payment has fallen, whatever the streak says. This single field would have caught all 39.
- Then check coverage anyway. A negative growth rate tells you a cut happened. It does not tell you whether what is left is safe, and those are different questions with different answers.
- Expect other sites to quote a different number. The window is one of four decisions behind why dividend growth figures disagree, and specials move the answer further than the window does.
- Look at the payments themselves before concluding anything. Every figure above is a summary of a series that takes about ten seconds to read directly. Leggett & Platt's 1.82, 0.61, 0.20, 0.20 says "cut hard, then held" more clearly than any growth rate can, and 3M's 5.02, 3.36, 2.92, 3.12 says "cut, then turned back up". No single number carries that shape.
Our five-step check on whether a dividend will hold runs the same sequence on a single company, and you can sort by dividend growth and Dividend Health Score together in the screener.
Frequently asked questions
- How many dividend aristocrats have a shrinking dividend?
- Of the 364 companies on DividendAtlas with an unbroken payment record of 25 years or more, 21 are paying less today than they were five years ago. That is about one in seventeen. Widening the net to 10-year records gives 39 companies.
- How can a company cut its dividend and keep its streak?
- Because the streak counts years in which a dividend was paid, not years in which it rose. A company can halve its payment and still extend an unbroken record, as long as it pays something.
- Does a cut always mean the company is in trouble?
- No, and this is the part most coverage gets wrong. A rebased dividend is often better covered than the one it replaced, because the payment fell and the cash flow did not. One company in this article now pays 0.20 USD a share against 2.15 USD of free cash flow per share.
- Why do so many of these companies still score reasonably?
- Because our score reads current coverage as well as the record. A cut is penalised through the track-record pillar, but a company that cut two years ago and now covers its payment comfortably is genuinely safer than one still paying more than it earns.
- How do I check whether a long record is still growing?
- Compare years paid against years increased, and check the five-year dividend growth rate. If growth is negative, the payment has fallen whatever the record says.
DividendAtlas provides data and research for informational purposes only. Nothing here is investment advice or a recommendation to buy or sell any security. Always do your own research.