COTTON DIVISION

WEEKLY REPORT
by Alex Gansch -- Vice President / Senior Trader

September 9 – 13, 2002

The market started on a weak footing as some speculative players just liquidated modest long positions. During the first half of this week, participants kept their eyes primarily on the release of the important USDA supply/demand report, which was published on Thursday. Key for the market was to see whether the government would adjust the domestic- and foreign-demand categories, bearing the hope of reducing the burdensome US upland stocks. The USDA's September crop report, released the same day, was based on a crop survey as of September 1, causing private estimates to peg the U.S. cotton output for the 2002/2003 season at 18.222 million 480-pound bales, compared with the previous estimate of 18.44 million. Exports were seen at 10.842 million bales, domestic consumption at an estimated 8.085 million bales, leaving ending stocks at 6.978 million bales. The final figure for all cotton production arrived at 18.1 million 480-pound bales, down 2 percent from last month and 11 percent below last year's record high production.  Yields are expected to average 675 pounds per acre, the same as last month.  The lower production is due primarily to reduced harvested acreage in Arkansas, Louisiana, South Carolina, and Texas and lower yields along the Atlantic Coast.  Harvested acreage, at 12.9 million acres, was reduced based on administrative data.  The September harvested area reflects decreases of 40,000 acres in Arkansas, 70,000 acres in Louisiana, 9,000 acres in South Carolina, and 102,000 acres in Texas while general growing conditions improved in Arkansas, Missouri, and Texas, leading to higher yield expectations than those of last month. As for the Supply/Demand Report, this month's 2002/2003 U.S. forecast reflected slightly lower production and ending stocks. While production as mentioned above was slightly reduced, unfortunately, for many

traders, no changes were made to either domestic mill use or exports, leaving ending stocks down by only 300,000 bales, when compared to last month, at 6.7 million bales. As for the 2002/2003-world supply and demand estimate, it also featured lower production and ending stocks. World production was reduced by 1 percent, as reductions for Australia, the United States, Brazil, the African Franc Zone, and Central Asia were partially offset by an increase for India. With world consumption and trade nearly unchanged, ending stocks were reduced 1.5 percent to 39.2 million bales. While these figures were viewed as bearish by most analysts, one cannot ignore the current difference in ending stock calculation between the Census Bureau and the USDA. While the Census Bureau reported in its August report that as of July 27, stocks held in consuming establishments, public storage, compresses and elsewhere totaled 6.918 million 500-pound bales, which is equal to about 7.2 million 480-pound bales, the USDA's ending stock projection for 2001/2002 arrived at 7.6 million bales. The Census Bureau acknowledged that incongruity and said those figures are under scrutiny and that any potential revision will be unveiled in its next report due out September 27. But how the USDA - which usually adopts the Census numbers in September – resolves this problem is anybody's guess. Late in August of this year, a USDA representative said that the government will be considering to what extent it should reflect the Census data when the USDA releases its September estimate, possibly reducing ending stocks by yet another 400,000 bales, which would be most welcome. Meanwhile, US crop progress continues satisfactorily with 52 percent of all cotton bolls open versus 52 percent a year ago at the same time and the five-year average of 51 percent. As of September 8, 17 percent of

the crop were classified as “very poor to poor” compared with 18 percent the prior week, 29 percent as “fair” versus 29 percent as of September 1 and 54 percent as “good to excellent” versus 53 percent the week before. In other news, forecasts released on Tuesday of this week by the Australian Bureau of Agricultural and Resource Economics were that cotton production in Australia will fall 41 percent to 401,000 metric tons, or 1.841 million bales, this fiscal year ending June 30, 2003, from an actual 684,000 tons last fiscal year. This figure is well below the USDA current estimate of 2.2 million bales. Meanwhile, the spec/hedge report put speculators at a net long position of 32.7 percent of the open interest, compared with a net long stance of 33.8 percent the previous week. Brokers do not seem worried that the amount held by speculators could be a threat to prices, as there is no hurry to further liquidate long positions at this time. Interestingly, the weekly exports report issued by the USDA this past Thursday showed new sales of 364,400 bales, the largest since the new marketing year began, nearly six times the previous week and three and one-quarter times the 4-week average. Shipments arrived at 136,200 bales or 6 percent more than reported during the prior week. While the actual amount of new sales did not seem to surprise due to pre-existing rumors that a major Memphis merchant may have concluded sales of similar size during the third week in August, displaying a bullish market activity both in the futures and option pit, it was the destination of Mexico that raised eyebrows as the assumption had been that these sales were destined for China.

Despite a somewhat lethargic trading pattern in recent weeks and an overwhelming call for still lower prices by many participants, we believe that current values already present buying opportunities Having absorbed virtually all of the negative news a market can bear, there seems little room left to the downside. Gradually, yet substantially improving off-take causing a reduced ending

balance may well be ahead of us due to statistical adjustments as well as further sales of significant proportions to overseas markets.

As indicated in our last Weekly Report, the USDA decided to reduce the harvested Pima acreage in California for this month’s crop report by 20,000 acres from 229,000 to 209,000. This figure appears to be more in-line with many of the private estimates, however, as seen last season, the final figure can still end up closer to the top end of the range. Coupled with reduced yields in California from 1,300 to 1,286, this leaves the total production figure at 629,000 bales for the 2002/2003 season compared with 689,000 bales estimated by the USDA just a month ago. The only other changes involved Texas, where harvested acreage was reduced to 18,000 acres from 20,000 acres, yet yields were increased from 984 pounds per harvested acre to 1,093 pounds. At time of writing, this 60,000 bale swing in availability of US Pima cotton for the coming season did not affect spot prices nor interest to sign forward contracts by either growers or merchants. While during most other seasons such a decrease could well have caused a significant increase in prices and demand, the combined stocks of still nearly 1.0 million bales of US Pima alone, are not stirring ELS consumers’ minds. If nothing else, though, this adjustment is a reminder of how quickly significant portions of any presumed crop may just disappear, on paper. The weekly USDA export report came in with 5,000 b/c, lifting total sales to just above 172,000 bales for the current marketing year compared with 204,000 bales at the same time last year. Despite the yet to be released prices from Egypt for their respective, competitive varieties, we expect a noticeable increase in demand and fresh sales as we are getting closer to harvest and the season progresses.