Home › Research
Dividend timing: what 1,882 US payers actually do
Original analysis of our own data — 178,528 dividend payments across 1,882 US-listed companies, reaching back to January 1962. Rebuilt with every update of the site.
Individual stock pages answer questions about one company. This page answers the ones that only make sense in aggregate: how often US companies pay, when in the year dividend activity clusters, which day of the week ex-dates land on, and how much notice a board actually gives before its shares go ex. Every figure below is measured, not sourced.
How often US companies pay
Payment frequency is inferred from each company's recent record rather than a static label, so a company that has changed cadence is counted as what it does now.
| Frequency | Companies | Share |
|---|---|---|
| Monthly | 155 | 8.5% |
| Quarterly | 1,594 | 87.0% |
| Semi-annually | 59 | 3.2% |
| Annually | 24 | 1.3% |
Quarterly payment dominates at 87% of covered companies. That is a US convention rather than a rule — semi-annual payment is standard across much of Europe, and annual payment is common in Japan — and it is why the whole market's dividend rhythm is organised around a four-times-a-year cycle. Monthly payers are a small minority, concentrated in REITs and business development companies, which collect income monthly and pass it through.
Dividend activity is seasonal
Because the quarterly cycle dominates, and most companies align it to their fiscal quarters, ex-dividend dates are far from evenly spread. Across 24,893 ex-dates in the past 3 years:
| Month | Ex-dates | Share of year |
|---|---|---|
| January | 1,180 | 4.7% |
| February | 2,188 | 8.8% |
| March | 2,719 | 10.9% |
| April | 1,290 | 5.2% |
| May | 2,449 | 9.8% |
| June | 2,559 | 10.3% |
| July | 1,382 | 5.6% |
| August | 2,482 | 10.0% |
| September | 2,191 | 8.8% |
| October | 1,339 | 5.4% |
| November | 2,480 | 10.0% |
| December | 2,634 | 10.6% |
The spread is close to 2.3× between the busiest month (March, 10.9%) and the quietest (January, 4.7%). An evenly spread calendar would put about 8.3% in each month.
The pattern is the second and third months of each quarter carrying the load, with the first month — January, April, July, October — comparatively empty. For anyone living on dividend income from a portfolio of quarterly payers, that means cash arrives in waves rather than smoothly, and the troughs are predictable.
Ex-dates cluster on particular weekdays
Less obvious, and rarely reported: ex-dividend dates are not spread evenly across the trading week.
| Weekday | Ex-dates | Share |
|---|---|---|
| Monday | 5,297 | 21.3% |
| Tuesday | 4,271 | 17.2% |
| Wednesday | 3,359 | 13.5% |
| Thursday | 4,446 | 17.9% |
| Friday | 7,520 | 30.2% |
Friday is the most common ex-dividend day, at 30.2% of all ex-dates, against 13.5% for Wednesday — an even split would be 20% each. The driver is that companies anchor their schedules to month-ends and quarter-ends, and those anchors, filtered through the trading calendar, land disproportionately on certain weekdays. It is a useful sanity check on any predicted date: an estimate that lands mid-week is going slightly against the grain of how these schedules are actually set.
How much notice companies give
A dividend only becomes official on the declaration date. Measuring the gap from declaration to ex-date across the 1,023 companies for which we hold enough announcements to compute a reliable median each:
| Median notice before the ex-date | 14 days |
|---|---|
| Middle half of companies | 12 to 23 days |
| Median wait from ex-date to payment | 16 days |
Roughly three weeks of notice is the norm. That is the window this site exists to widen: because dividend schedules repeat, the next announcement can often be estimated from a company's own history, months before the board actually declares anything — see how we predict it.
What we looked for and didn't find
Not every result is a positive one, and the negative findings are worth publishing too.
Deriving the announcement from our former ex-date prediction was the worst route we tried. It seemed like the obvious approach — we used to predict the ex-date directly, and the declaration comes a fairly steady interval before it — but the gap between the two is less stable than either date is on its own, so the method compounded our own error instead of reusing it. A single seasonal-and-cadence rule did better, calling the declaration to within three days on 77% of the quarters it spoke about, though it only spoke about 45% of them. What eventually worked was neither: requiring several independently-computed estimates to agree before publishing anything. That is the method behind our announcement predictions, and its accuracy is on the accuracy page.
Anticipating a dividend raise doesn't pay. Companies announcing an increase see their shares move about +0.8% on the news. But when we restricted this to raises that were predictable in advance from the payment pattern, the move was +0.15% and not statistically significant. The market reaction is concentrated in the surprises — which are, by definition, the ones we cannot anticipate.
Dividend cuts are already priced. Companies that cut had typically already underperformed by around 3% over the previous six months. The information is in the price before the announcement, so a cut cannot usefully be predicted from the payment record.
We publish these as findings rather than strategies. They describe historical data, we have not established that any of them survive real-world costs, and none of them is a recommendation.
Method and caveats
- Coverage is US-listed dividend payers with enough history to model — currently 1,882 companies. It is not the whole market, and it excludes companies that have stopped paying.
- Seasonality and weekday figures use the past 3 years only. The store reaches back much further, but how companies scheduled dividends decades ago says little about current practice.
- Announcement-timing figures cover only companies with enough declarations on file to compute a median each. That ledger accrues daily and is still growing, so the sample will widen over time.
- Payment amounts in our source data are as reported and are not restated for later share splits, so this page deliberately reports on timing rather than amounts.
Everything here is derived from public dividend records. See how the predictions are made, the full accuracy breakdown, or browse the announcement and ex-dividend calendar.
Research published for information only. Nothing on this page is financial advice, and historical measurements are not predictions of future behaviour.