In Japan, the March Fiscal Year Is a Minority Position
A launch date gets chosen early, and it usually gets chosen on the calendar of whoever is doing the choosing. A quarter that suits the home market. A window after a global campaign. A month when the team has capacity.
Somewhere on the Japanese side there is a partner, a distributor, a retailer or a client whose own year ends on a date nobody in the planning meeting has looked up. The common shorthand is that it ends in March.
That shorthand is checkable, and Japan’s National Tax Agency publishes the numbers.
More than four in five close in another month
The agency’s statistical yearbook records how many corporations filed returns, broken down by the month their accounting period ended. In the latest published edition, among the corporations that settle accounts once a year:
| Ending month of accounting period | Number of corporations that filed returns | Share |
|---|---|---|
| March | 535,064 | 17.8% |
| September | 329,243 | 11.0% |
| December | 317,140 | 10.6% |
| June | 291,067 | 9.7% |
| August | 261,798 | 8.7% |
| May | 247,013 | 8.2% |
| July | 229,737 | 7.7% |
| April | 207,892 | 6.9% |
| February | 198,161 | 6.6% |
| October | 155,662 | 5.2% |
| November | 119,670 | 4.0% |
| January | 110,368 | 3.7% |
| Total | 3,002,815 | 100% |
The share column is ours, calculated from the published counts; the yearbook prints the counts only.
No month exceeds March, and March reaches 17.8%. More than four in five of these corporations close their books in another month. September and December each clear ten percent. Every month of the year has at least a hundred thousand corporations ending an accounting period in it.
One row in that table is worth a second look by anyone planning from a December year end. December is third, at 317,140 corporations. A brand whose own books close in December is not looking at a country that closes in March instead — it is looking at a country where its own closing month is third on the list, and where no month reaches a fifth of the total.
So the axis is not December against March. The shorthand is not wrong about which month is biggest. It is wrong about how much that settles.
What this table is counting
Before the number gets used for anything, it is worth being exact about what was counted. The yearbook states its own scope, in English, on the same page:
This table shows the statistics of domestic ordinary corporations (companies, etc., joint enterprises cooperatives and medical corporations) and totalized corporations of which the accounting period ended between April 1, 2024, and March 31, 2025, and for which cases declared until July 31, 2025, or processed between July 1, 2024 and June 30, 2025, based on “Administrative documents for corporation tax (returns and resolutions, etc.)”
Three things follow from that sentence.
It counts filings rather than companies in general — corporations that filed a return in the window described. It counts domestic ordinary corporations, a defined tax category that includes companies, cooperatives and medical corporations. And it covers one year, the accounting periods that ended between April 2024 and March 2025.
The figures above are also only the corporations settling accounts once a year. The same table lists a smaller group settling twice a year, and a residual category, which together bring the grand total to 3,025,599.
What the table records is the month a set of books closed. It does not record when a budget was set, when an approval was granted, or why any month is more common than another. Those are different questions, and this table is not evidence about any of them.
The picture changes with size
The same table breaks the counts down by capital range, and this is where the shorthand starts to earn its keep — for some counterparties.
| Capital range (as published) | Ending in March | All months | Share ending in March |
|---|---|---|---|
| Under 100 million yen (re-grouped) | 515,689 | 2,967,002 | 17.4% |
| 100 million yen or more (re-grouped) | 19,375 | 35,813 | 54.1% |
| Over 10 billion yen | 1,061 | 1,454 | 73.0% |
The share column here is worked out the same way, from the counts as printed. The first two rows divide the table exactly — they add to 3,002,815. The third is a band inside the second, not a fourth group.
Among corporations capitalized over ten billion yen, close to three in four end their year in March. Among those capitalized under a hundred million yen, it is fewer than one in five — almost exactly the all-company figure, which is what the arithmetic requires, because that band is 98.8% of the table.
That last point is worth stating plainly. Of the 3,002,815 corporations counted, 2,967,002 are in the under-100-million band — 98.8% of them. The all-company figure of 17.8% is, in effect, a report on small companies with a small number of large ones mixed in.
So the March shorthand describes a real and strong pattern. It describes it at the top of the size distribution, where a small number of corporations sit.
What the size figures do not settle
Here is where this data stops being useful, and the stopping point matters more than anything above it.
The table has nothing to say about the company on the other side of a particular contract.
There is a second limit inside the size figures themselves. The bands are the ones the yearbook publishes, and “under 100 million yen” is a single bucket holding almost three million corporations — a two-person company and a substantial private firm sit in it together. Knowing a counterparty is under that line narrows very little.
The 98.8% is a share of all domestic ordinary corporations that filed. A brand entering Japan does not draw its partners at random from that population. It works with distributors, retailers, agencies, media owners, production companies, logistics providers — and the capital distribution of the firms a foreign brand actually ends up contracting with is not something this table measures. It is not in here at all.
Which means the numbers above cannot be run in either direction:
- They do not support “the counterpart is probably not a March company.” Four in five corporations are not, but a specific counterparty is not a draw from that population.
- They do not support “the counterpart is probably a March company.” Even the figure for the over-ten-billion band is a share rather than a certainty, and nothing here establishes which band a given counterparty sits in.
What the table establishes is that one number cannot stand in for the answer. The spread between 17.4% and 73.0% is the finding. It is large enough that a guess in either direction is worth less than a question.
The question is cheaper than the assumption
Which turns this from a statistics problem into a scheduling one.
A Japanese company’s fiscal year end is not confidential. It appears in the published financial statements of listed companies, on company websites under investor or corporate information, and it is an ordinary thing to be told when asked. For a specific counterparty this is a question with an answer, not a probability that has to be carried.
The cost of asking is one line in an email. The cost of not asking is that a date already fixed in a global plan turns out to sit on top of somebody’s closing period — and by then the date is in a deck, and moving it is a conversation about credibility rather than about calendars.
The economics match what happens to the decision not to change something once production is commissioned — cheap to raise while the plan is still soft, expensive once it has hardened. The difference is that here the answer is a published fact rather than a judgment, so nobody has to be persuaded of it.
So the practical move, before a launch window is agreed rather than after, is to establish for each Japanese counterparty in the plan the month their accounting period ends, and to hold the plan next to those dates. Not to predict them from a national average — the table above shows how far that average can be from any given company.
Fixing that sequence — dates checked while they are still proposals — is one of the things a Japan readiness step is for.
Source: 国税庁統計年報 令和6年度分 / National Tax Agency Statistical Yearbook, FY2024, table 4-2 (1), “Number of ordinary corporations and totalized corporations by settlement term,” page 186. https://www.nta.go.jp/publication/statistics/kokuzeicho/r06/R06.pdf — the table’s title, column headings and scope statement are quoted from the English printed alongside the Japanese in the source; one column heading is printed there as “100 million yen or more than” and is given here as “or more.” Share columns are our own calculations from the published counts.