COTTON: The 2026/27 U.S. cotton balance sheet shows reduced beginning and ending stocks, due to a 200,000-bale decrease from the previous year. Production, consumption, and trade forecasts are unchanged this month, and the projected season-average price remains at 73 cents per pound. Exports for 2025/26 are now projected at 12.20 million bales, an increase of 200,000 from last month, while mill use is reduced 50,000 bales to 1.55 million.
As a result, ending stocks are now forecast at 4.20 million bales, for a stocks-to-use ratio of 31 percent. The 2025/26 season-average farm price remains estimated at 63 cents per pound. World cotton supply for 2026/27 is slightly lower due to reduced beginning stocks.
Production remains at 116.0 million bales, with trade marginally down. Consumption edges up to 121.8 million bales, driven by increased demand from China despite declines for Bangladesh, Pakistan, and South Korea. Ending stocks for 2026/27 are reduced slightly this month to 71.1 million bales, mainly because of lower beginning stocks.
For 2025/26, higher world exports reduce ending stocks, with global production and use largely unchanged. Exports are raised by over 1 percent, led by Brazil, the United States, Kazakhstan, and Turkey. Global production is increased by 15,000 bales due to Egypt, offsetting Argentina's decline.
Consumption is lowered 25,000 bales as decreases in several countries outweigh gains in China and Vietnam. Ending stocks are reduced by more than 600,000 bales, lowering the stocks-to-use ratio to 64 percent.
Select a region and attribute, then choose which marketing years to overlay. The table shows the estimate published each month; the chart shows the revision trend over time. Large swings indicate high uncertainty or late-breaking supply/demand shifts.
How to read it: A rising line through the season means USDA progressively raised estimates. A dip mid-season often reflects adverse weather or demand revisions. Compare years to spot patterns — e.g. whether U.S. production estimates tend to be cut after planting season.
This page measures forecast error for each attribute (Production, Exports, Ending Stocks, etc.) by comparing the May estimate (first projection of the season) against the final August estimate of the following year — the closest thing to a "true" outcome.
RMSE (Root Mean Square Error) — the average size of the error in million bales. A lower RMSE means the early estimate was more accurate.
90% CI — 90% of final outcomes fell within this range of the early estimate. For example, a 90% CI of ±2.5M on U.S. Production means the May forecast was within 2.5M bales of the final number 9 times out of 10.
Key insight: U.S. production is highly uncertain early in the season (planting/weather risk), while World Ending Stocks tends to be revised significantly as consumption data from Asia arrives late.
| Category | Region | RMSE | ±90% CI | Bias | Accuracy |
|---|