GSX Monthly Cotton Report

August should have been a win for the bulls, the chart above does most of the talking,iIn the historical sample 2010 - 2026, it has the widest revision range and the largest average move. The 13.61m-bale estimate we have now is therefore a starting point, not a finished crop number, where is the next opportunity for the bulls, it is September and from there just simply looking at what has happened in the past, the coming reports will have less influence as we start approaching the harvesting period. At that point in time, all things being equal if the SnD is the same will the bulls still want to be long? Will the bears still want to be short? Will the dungeon masters armed with their shiny new latest GPU named after a distant galaxy still have their piggie algorithm working in a cross sectional carry, momentum strategy,….., black box algorithm still be interested in Cotton?

It would be more stressful to be a bull than a bear. How does that translate into market psychology and willingness to take prices higher? Well if you are long and your case depends on a weather forecast showing poor weather in Texas and getting worse by the next WASDE that would be stressful wouldn’t it?

The August WASDE was the first report this season to include survey-based yield information. The key question is not simply whether the headline numbers moved higher or lower, but what the revisions could imply for the range of possible outcomes.

The weather picture is uneven rather than uniformly bad. The largest cotton states sit across a wide range of rainfall and condition readings, with Texas standing out because of its acreage and weak current condition.

At the world level, India and China account for the largest changes in domestic use and stocks. Brazil remains the key export-side swing factor, while the U.S. revision is comparatively small.

The global balance is tighter because use rose faster than production. Stocks are now 69.7m bales and the stocks-to-use ratio is 56.7%. That is constructive, but Brazil remains the main supply risk. Its export base has expanded steadily over the past two decades, and the current data may still capture only part of that growth.

The August production estimate is 13.61m bales, but the 90% historical range runs from 11.62m to 15.60m. Ending stocks carry an even wider error band. The point is not to disregard USDA, but to treat the estimate as one input and do the independent work around it.

Earlier August estimates tend to lean lower by the final estimate. In the last 14 years shown, 10 were revised down, with a mean final revision of negative 0.59m bales. The current estimate should not yet be treated as settled.

At the world level, the revision is constructive on use, but it is not a shortage call by itself. Stocks to use is 57%, below its long-run average, while yield remains above trend and production is still close to mill use.

Texas carries the largest production base, while Oklahoma and several western states show the sharpest year on year declines in production and yield.

The state-by-state area, yield and production mix has changed. That broadens the stability of the production estimate. For the bull case to rely on Texas, weather and especially subsoil moisture would need to deteriorate further.

The weather risk is concentrated in the soil profile, not just the weekly condition number. Texas, Oklahoma and New Mexico are among the driest large-acreage states, and the 2026 topsoil and subsoil paths are running above 2025 through the current window.

Rainfall has not been a uniform answer either. The observed map is dry across parts of the central and western belt, while the forecast balance through late August is more negative in Texas and the Plains and more favourable in parts of the East.

At county level, the same split is easier to see: much of Texas and the Plains remains in a negative water balance, while the eastern cotton states have more neutral or positive readings. This is a diagnostic map, not a yield forecast.

That leaves a regional question for September. Texas and Oklahoma still carry high short or very short readings in both topsoil and subsoil, while the comparison with the five-year average shows that the pressure is not evenly spread across the belt.

USDA’s 2026/27 U.S. production estimate is 13.61m 480-lb bales, down from 13.70m in July. Harvested area increased by 0.65m acres, but the 74 lb/acre yield reduction more than offset it. Ending stocks fell from 4.10m to 4.00m bales because exports stayed at 12.30m and mill use at 1.60m.

The extra acres were not enough to absorb the yield cut, and demand was not revised higher to cushion the smaller crop. The change therefore came through yield rather than demand. That is consistent with upland cotton rated 40% good to excellent for the week ending 9 August, seven points below the five-year average. Season-to-date rain across the cotton states was 80% of normal, with 43% of topsoil and 43% of subsoil rated short or very short.

The global revision is more constructive than looking at the U.S. number alone: world production rose 0.37m bales to 117.63m, but domestic use rose 0.97m to 122.92m. Ending stocks fell 1.53m to 69.69m, taking the world stocks-to-mill-use ratio to 56.7%, below 61.9% in 2025/26.

China’s MARA/CASDE places the local crop at 30.5m bales, but its latest figure was published in April and its stocks-to-use ratio is 103.7%. India’s CAI/CAB balance is much tighter at 12.8%, although it is an industry figure rather than a government estimate. Brazil is around 18m bales in the latest CONAB report and is the export supplier that can move the most volume. Australia is smaller and less relevant, but its production swings are more than seven times the world’s typical year-on-year move.

Our view: August often produces some of the year’s largest revisions, so the 13.61m-bale estimate should be treated as a starting point rather than a final crop number. We are skewed to the bearish side, but the risk/reward is skewed lower unless weekly AYP data, exports and on-call reporting confirm additional supply stress, particularly in Texas. By September, the crop will be beyond much of its most weather-sensitive window, so the focus should shift from weather headlines to realised yield, shipment conversion and origin competition.

The practical implication is that the first estimate should be treated as a range. September is the next scheduled test, but the size of the August revision history means weekly weather and crop reports remain important in the meantime.

The U.S. production cut was close to the historical average in percentage terms, but the ending-stocks cut was more unusual. The global ending-stocks revision was also on the tighter side.

Price has not responded in a straight line to these revisions. The 63-day returns after August releases range from strong gains to sharp losses. Subsequent direction tends to depend on the size of the revision relative to the U.S. and global balance sheets.

What does the market need to go higher? It needs a larger wave of speculative buying. Commodity markets are often driven by positioning and trend-following flows, especially near the extremes. Natural hedgers matter, but they do not explain every move.

In the chart below, managed-money net positioning contributed to the move from 63 to 88, but positioning remains below the upper end of the 20-year range. A move toward 90 or 100 would require the four large trader groups to keep increasing their positioning in percentile terms.

Their conviction would likely need three inputs: worsening weather and AYP or production signals in the Texas High Plains; further price strength that attracts trend-following flows; and a constructive U.S. macro momentum thump. Macro conditional, on lower oil translating into core PCE, together with a more dovish Federal Reserve, which could improve the environment for macro-sensitive commodities such as cotton.

A statistical model of the baseline USDA balance against flat price would yield a price of ~ 73 c/lb.

A bullish SnD with 12.5 million bales of production would yield a price of ~85 c/lb.

A bearish SnD would yield a price of 63 c/lb.

If the SnD does continue to tighten at the U.S. level and especially the global level which would then feed back into the U.S. balance sheet, then we would expect the SnD to closely resemble the situation we were in last year.

The views, opinions and market commentary expressed in this article are solely those of the author and are provided for general information purposes only.

Whether you're looking for weekly market updates or comprehensive commodity research and technical analysis, there's a GSX Research package to suit your needs.

Basic – Essential weekly market intelligence.

Premium – Full research reports, technical analysis, proprietary data and cross-commodity insights.

Exclusive – Bespoke research, monthly strategy calls and direct access to GSX analysts.

👉 View Plans & Subscribe Now

GSX Commodities Research podcast: conversations on commodity markets, trade and global supply chains.