When to buy: the calendar
The one seasonal pattern in textile prices that survives an honest test, and the dates that move shipping: Chinese New Year, Ramadan, the two Eids and the container peak.
There is one seasonal pattern in textile prices that survives an honest statistical test, and a great many that do not. This page separates them, and then sets out the recurring dates that move shipping rather than prices: the Chinese New Year shutdown, Ramadan and the two Eids, and the container peak before the western holidays. If you are planning a first order, the useful question is rarely whether cotton is cheap this month. It is whether the factory is open, whether the vessel has space and whether your goods will sit in a port over a holiday.
When to buy
Two completely different kinds of statement get muddled together whenever anyone talks about the best time to buy. We keep them apart and label both, because one is arithmetic about the past and the other is a diary.
Computed A month-of-year pattern worked out from data we hold. It says what happened, over a stated number of years. It is not a forecast.
Calendar A date or a window that is documented and recurring. It says what is scheduled to happen, and carries a source.
Computed Which months are usually dear, and which cheap
Computed from US producer price indexes over 9 complete calendar years, 2017 to 2025. Each month is expressed as a percentage of its own calendar year average, and those figures are then averaged across the complete years in the window. Expressing a month against its own year removes the trend, so what is left is the shape of the year rather than inflation. 103 means that month typically sat three per cent above the year it belonged to.
Raw cotton: a clear yearly pattern
| Month | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Typical level | 102.6 | 104.1 | 103.5 | 101.7 | 100.3 | 100.9 | 95.8 | 96.8 | 98.3 | 97.6 | 98.9 | 99.6 |
| Years above its own average | 6 of 9 | 7 of 9 | 6 of 9 | 5 of 9 | 3 of 9 | 4 of 9 | 2 of 9 | 3 of 9 | 5 of 9 | 3 of 9 | 2 of 9 | 3 of 9 |
Read 100 as the average of the year the month belongs to. The dearest month was February at 104.1, above its own year in 7 of 9 years. The cheapest was July at 95.8, below its own year in 7 of 9. Top to bottom that is a spread of 8.3 points, and the consistency is what makes it worth reading rather than the size: an average can be dragged up by one extreme year, and this one is not.
The shape matches the harvest and nothing else needs to be invented to explain it. Lint is at its most plentiful once the northern hemisphere crop is picked and ginned, and at its scarcest in the late winter months before the next one arrives. That is the calendar below, showing up in the arithmetic above.
Weak patterns, shown for honesty
- Artificial and synthetic fibers and filaments manufacturing. Spread of 3.9 points over 9 years, dearest in October (102, above in 7 of 9 years), cheapest in January (98.1, below in 5 of 9). One end of that is reasonably consistent and the other is not, so it does not meet our bar for calling it a pattern.
Where we looked and found no pattern
These are shown because a null result is a result. If we only published the series with a pattern, a reader would reasonably conclude that everything in this trade is seasonal, and it is not.
- Noncellulosic fiber and other natural fiber spun yarns. Top to bottom spread of only 2.6 index points across 9 years. That is inside the ordinary month-to-month noise of the series, so there is no month-of-year effect here worth timing an order around.
- Textured and twisted yarns. Top to bottom spread of only 2.3 index points across 9 years. That is inside the ordinary month-to-month noise of the series, so there is no month-of-year effect here worth timing an order around.
- Fiber, yarn, and thread mills. Top to bottom spread of only 2.8 index points across 9 years. That is inside the ordinary month-to-month noise of the series, so there is no month-of-year effect here worth timing an order around.
Not enough history to try: Cotton spun yarns, Rayon, acetate, and lyocell spun yarns. Our rule is a minimum of 5 complete calendar years, and these series do not go back that far. They are not shown at all rather than shown with a warning.
A seasonal average is a description of the past. It is not a forecast, it carries no confidence interval, and any single year can and does ignore it. It is computed from US producer price indexes, so it describes US producer prices and nothing else.
Calendar Dates that matter every year
None of these is a prediction about price. They are things that are scheduled, or that recur for a structural reason, and each one carries the source that documents it.
-
August to January
The northern hemisphere cotton harvest
The United States marketing year begins on 1 August. Picking runs through autumn, and weekly crop progress reports track it field by field until the crop is off. Pakistan and India pick earlier and later than the US respectively, so the three overlap rather than queue. Physical lint is at its most plentiful once each crop is ginned.
Why it matters to a buyer. Supply arrives in a lump rather than evenly, so the physical market is loosest just after the harvest and tightest in the months before the next one.
Source: USDA NASS Crop Progress (opens in a new tab), USDA Economic Research Service, cotton and wool (opens in a new tab)
-
1 September onwards, fortnightly
Pakistan: cotton reaching the ginneries
The Pakistan Cotton Ginners Association publishes seed cotton arrivals at ginneries twice a month from the start of September. It is the earliest hard read on the Pakistani crop, well before any official estimate.
Why it matters to a buyer. If you buy Pakistani lint or Pakistani yarn, this is the number that moves your supplier’s cost before it moves your quotation.
Source: Pakistan Cotton Ginners Association (opens in a new tab)
-
October to September
The Indian cotton season
India runs its cotton year from October. Spot rates are quoted in rupees per candy and published daily by the Cotton Association of India.
Why it matters to a buyer. An Indian yarn quotation is priced off an Indian lint market on an Indian calendar, not off the world index.
Source: Cotton Association of India, spot rates (opens in a new tab)
-
Monthly, usually between the 9th and the 12th
The monthly US supply report (WASDE)
USDA publishes World Agricultural Supply and Demand Estimates on a fixed monthly schedule. It is the single most watched revision to the world cotton balance.
Why it matters to a buyer. Prices can move on the day. If a decision can wait a week and a WASDE falls inside that week, waiting costs nothing and removes a known unknown.
Source: USDA WASDE, publication archive (opens in a new tab), International Cotton Advisory Committee (opens in a new tab)
-
Every Friday afternoon US time
Weekly reports on who holds cotton futures
The CFTC publishes the Commitments of Traders and the Cotton On-Call report weekly. On-call matters more than most buyers realise: it counts sales already made whose price is still to be fixed against a futures month.
Why it matters to a buyer. A large unfixed on-call position has to be fixed eventually, and the fixing itself is buying.
Source: CFTC Cotton On-Call report (opens in a new tab), ICE Cotton No. 2 futures (opens in a new tab)
-
Between 21 January and 20 February, varying
Chinese New Year and the rush before it
China’s State Council publishes the official holiday arrangement each year. Factories across China stop for roughly a fortnight, and much of the rest of Asia is disrupted with them. The weeks before are the busiest booking window of the year.
Why it matters to a buyer. Capacity and vessel space tighten ahead of it and both are slow to restart afterwards. Lead times quoted in January are not the lead times you will get.
Source: The State Council of the People’s Republic of China (opens in a new tab)
-
Moves about eleven days earlier each year
Ramadan and the two Eids
Working hours shorten across Pakistan, Bangladesh, Egypt, Turkey and the Gulf through Ramadan, and mills close for several days at each Eid. The dates are lunar and are announced locally rather than fixed in advance.
Why it matters to a buyer. A month of reduced hours plus two closures is a real dent in a delivery schedule, and it lands in a different Gregorian month each year, so last year’s plan does not transfer.
Source: Pakistan Ministry of Religious Affairs and Interfaith Harmony (opens in a new tab)
-
Roughly July to October
The container peak season
Autumn, winter and Christmas goods have to leave Asia in time to reach shelves, which concentrates bookings into the late summer and early autumn. Container rate indices are published weekly and are the place to watch it happen.
Why it matters to a buyer. Freight is a real part of a landed cost, and a peak season surcharge can move a per kilo figure more than the fibre did.
What we cannot show you. We do not publish a computed freight seasonal. Our own freight history is 14 weekly readings, which is not enough to compute anything, and the lane indices we can read lawfully do not cover South Asian origins.
Source: Shanghai Shipping Exchange (opens in a new tab), Drewry World Container Index (opens in a new tab)
-
Autumn for the following summer, spring for the following winter
European and US retail order placement
Retailers commit to a season roughly two to three quarters ahead. The European Commission publishes monthly order book and business expectation surveys for retail, which is the closest free read on whether those commitments are being made or held back.
Why it matters to a buyer. A garment order is placed long before the goods ship, so the demand signal that matters to a maker is the order book, not this month’s retail sales.
Source: European Commission business and consumer surveys (opens in a new tab)
What you will not find here
We will not tell you to buy in a particular month. The computed pattern above describes years that have already happened and carries no promise about this one. The calendar tells you what is scheduled, not what it will do to a price. Put together they tell you when to pay attention, which is a different and more useful thing than being told when to buy.
A seasonal average also cannot see the thing most likely to matter to you. One bad monsoon, one export ban, one currency move or one strike outweighs the whole nine year shape in the table above, and none of them keeps to a calendar.
What this means for the seasonality section
What this is Two separate answers to the question of when to buy: a month-of-year pattern computed from producer price history, and the recurring dates that are documented rather than computed.
How to read it In the month table, 100 stands for that year's own average price, so a month at 103 typically ran three per cent dearer than its year and one at 97 three per cent cheaper: brown marks the usually dear months, green the usually cheap ones. The row underneath is the consistency check, how many of the studied years that month really did sit above its own average, so 8 of 9 is a habit and 5 of 9 is a coin toss. The Calendar list further down is a diary of scheduled events with sources, not a computation at all.
What moves it The computed half updates when a new complete calendar year is available, so it changes once a year. The calendar half changes only when a schedule changes, except for the lunar dates, which move about eleven days earlier each year.
How to use it Deciding when to pay attention rather than when to buy. If a decision can wait a week and a WASDE falls inside that week, waiting removes a known unknown at no cost.
The trap Never read the two halves as one sentence. A computed average is a description of years that have already happened, with no confidence interval and no promise about this one, and any single year can ignore it completely. It is also computed from US producer prices, so it describes US producer prices and not your supplier.
Ramadan and Eid, and why the shipping runs early
The Gulf retail year has two peaks and both of them move about eleven days earlier every Gregorian year. A seller working to a European calendar will be a month late without ever making a mistake they can point at.
Next: Ramadan begins 2027, about 4 months away
The Gulf retail year
The stocking runs ahead of Ramadan, not into it. Shelves for the Ramadan and Eid al-Fitr season are filled before Ramadan starts, so the shipment lands weeks earlier than the date in the table suggests. Work backwards from the retail date, not forwards from your production slot.
Eid al-Fitr is the larger of the two for apparel. New clothes at the end of Ramadan is the single biggest clothing occasion in the Gulf year. Eid al-Adha matters too and comes about ten weeks later, which for a first time exporter is close enough to be a second wave off the same production run rather than a separate season.
| Hijri year | Event | Projected date | Months away |
|---|---|---|---|
| AH 1448 | Ramadan begins | Monday, 8 February 2027 | 4 |
| AH 1448 | Eid al-Fitr | Tuesday, 9 March 2027 | 5 |
| AH 1448 | Eid al-Adha | Sunday, 16 May 2027 | 7 |
| AH 1449 | Ramadan begins | Friday, 28 January 2028 | 15 |
| AH 1449 | Eid al-Fitr | Saturday, 26 February 2028 | 16 |
| AH 1449 | Eid al-Adha | Friday, 5 May 2028 | 19 |
| AH 1450 | Ramadan begins | Tuesday, 16 January 2029 | 27 |
| AH 1450 | Eid al-Fitr | Wednesday, 14 February 2029 | 28 |
| AH 1450 | Eid al-Adha | Tuesday, 24 April 2029 | 30 |
| AH 1451 | Ramadan begins | Saturday, 5 January 2030 | 39 |
| AH 1451 | Eid al-Fitr | Monday, 4 February 2030 | 40 |
| AH 1451 | Eid al-Adha | Saturday, 13 April 2030 | 42 |
What this means for the Gulf calendar
What this is Projected Gregorian dates for Ramadan and the two Eids over the coming years, computed from the Umm al-Qura calendar and cross-checked two independent ways before publication.
How to read it Each row is one event in one Islamic year (AH is the Islamic year number): the event, its projected date on the ordinary calendar, and how many months away that is from today. The same events repeat down the table, one row per year, landing about eleven days earlier each time. The date to plan shipping against sits weeks BEFORE the printed date, because Gulf shelves are filled ahead of Ramadan, not during it.
What moves it Nothing but the lunar calendar: each event lands about eleven days earlier every Gregorian year, so a plan copied from last year is eleven days late by default.
How to use it Count backwards from the event. Gulf retail stocks ahead of Ramadan, not into it, so production and shipping have to land weeks before the date in the row.
The trap These are projections, not declarations. The actual start is set by moon sighting, can differ by a day between states, and is confirmed as a public holiday only weeks out; plan against the earlier date and treat the later one as slack.
These are projections and they can move by a day. They are computed from the Umm al-Qura calendar, which is the Saudi civil calendar and an arithmetic tabulation. Actual observance is declared by moon sighting one or two evenings ahead, it can differ by a day between GCC states, and governments confirm the public holiday only weeks out. Ramadan is tabulated at 29 days in most of the years above, so a single day of sighting difference moves Eid al-Fitr by a day. Plan the shipment against the earlier date and treat the later one as the slack.
How these dates were produced, so you can check them. Each date was worked out two independent ways, with two different published versions of the Umm al-Qura calendar (the calendar Saudi Arabia uses), and the two had to agree before the date went on the page. If they ever disagree, the page is not published rather than one being chosen. We stop at the Islamic year 1451 because beyond that the two versions start disagreeing about month lengths, and a date we cannot confirm twice does not go on the page.
There is no latest sailing date on this page and there will not be one until we can source a transit time. That calculation needs a published Asia to Gulf transit, and we have not found one from a source we would cite. Inventing a fortnight would undo the credibility the rest of this page buys.
Common questions
Is there a cheapest month to buy cotton?
Not reliably, and this page shows the working rather than asserting it. One seasonal pattern holds up when tested honestly and the rest are stories told after the fact. What does move predictably is the calendar of shutdowns and shipping peaks, which is why that half of the page is the useful half.
How far ahead of Chinese New Year do I need to order?
Far enough that your goods are on the water before the factories close, not far enough that they sit in a warehouse for a season. The date moves every year and the shutdown is longer than the holiday itself, because staff travel and lines restart slowly. The section on this page carries the timing.
Why does shipping to the Gulf run early before Ramadan?
Because the selling season is built around it and nobody wants stock arriving during it. Retail buying, customs staffing and delivery windows all shift, so the order that would normally be comfortable becomes late. The Gulf retail year section sets out what lands when.
What is the peak season, and does it affect me?
It is the run of months when everybody ships at once for the western holiday trade, and it affects you through freight availability rather than through the price of your goods. If your order is small, the risk is not cost, it is being rolled to the next vessel.