COTTON AND WOOL YEARBOOK December 04, 1998 November 1998, CWS-1998 Approved by the World Agricultural Outlook Board --------------------------------------------------------------------------- COTTON AND WOOL YEARBOOK is published annually by the Economic Research Service, U.S. Department of Agriculture, Washington, DC 20036-5831. This release contains only the text of the COTTON AND WOOL YEARBOOK -- tables and graphics are not included. Printed copies of this Yearbook are available from the ERS-ERS-NASS order desk. Call, toll-free, 1-800-999-6779 and ask for stock # CWS-1998, $21. ERS-ERS-NASS accepts MasterCard and Visa. --------------------------------------------------------------------------- Contents Summary U.S. Cotton Situation and Outlook World Cotton Situation and Outlook U.S. Wool Situation and Outlook Foreign Wool Situation and Outlook Manmade Fibers Special Articles Factors Affecting the U.S. Farm Price of Upland Cotton China's Cotton Sector Under Stress List of Tables Situation Coordinator Leslie A. Meyer (202) 694-5307 Principal Contributors Leslie A. Meyer (202) 694-5307 Steve MacDonald (202) 694-5305 Robert Skinner (202) 694-5313 Mae Dean Johnson (statistics) 202) 694-5299 John V. Lawler Layout & Text Design Summary Based on November 1 crop conditions, U.S. cotton production in 1998 is forecast at 13.23 million bales, 30 percent below last season's 18.8 million and the lowest since 1989. Although planted area fell nearly 1 million acres to 12.9 million, higher abandonment this season is expected to lead to the smallest harvested area in a decade. Harvested area is estimated at 10.4 million acres and the national average yield is forecast to decline 68 pounds from last season to 612 pounds per harvested acre. U.S. cotton exports in 1998/99 (August/July marketing year) are projected to fall 3 million bales (40 percent) from last season to 4.5 million, due to the small U.S. crop and larger foreign exportable supplies. U.S. cotton has been less price competitive overseas this year as evidenced by the slow pace of early-season export sales. During the first 3 months of 1998/99, U.S. export commitments have reached only 3.3 million bales. And with a decrease in world cotton trade this season, the U.S. share of the global market is projected to fall to 18 percent, compared with last season's 28.5 percent. U.S. mill use of cotton in 1998/99 is also projected to decline 750,000 bales (7 percent) from last season to 10.6 million. A slowdown in U.S. economic growth and the surge in U.S. cotton textile imports this year are expected to limit mill consumption. In 1998, cotton textile exports are expected to advance for the 14th consecutive year, reaching the equivalent of 4.1 million bales. On the other hand, cotton textile imports will rise for the 10th year in a row, and could reach the equivalent of 12.5 million bales, further widening the cotton textile trade deficit. Meanwhile, total domestic cotton consumption (mill use plus net textile trade) could approach a record 9.4 billion pounds, with per capita consumption surpassing last year's 32.5 pounds. The U.S. cotton supply in 1998/99 is expected to include a significant volume of imports, unlike the previous season, as abundant foreign supplies and the declining U.S. crop have kept U.S. prices well above world prices. And with funding to end shortly for the "Step 2" provision of the cotton competitiveness program, the gap between U.S. and world prices will likely trigger the "Step 3" import quotas 10 weeks after "Step 2" funds are exhausted. However, even with imports, the total U.S. supply is estimated at only 17.4 million bales, more than 5 million below a year earlier. Meanwhile, total use is forecast at 15.1 million bales, 20 percent below 1997/98. Based on these supply and demand estimates, U.S. ending stocks for 1998/99 are projected at 2.3 million bales. With U.S. stocks expected to fall substantially this season, the stocks-to-use ratio is estimated at 15 percent, 3 percentage points below the 5-year average. World cotton production in 1998/99 is forecast to decline to 83.7 million bales, 7.5 million below last season's outturn. Foreign production, projected at 70.4 million bales, accounts for one-quarter of the decrease. Lower yields in China account for most of the decline in foreign production. In addition to China, lower production is anticipated in Uzbekistan and Egypt. Output in India and Pakistan is expected to rise. World consumption is projected to decline to 86.4 million bales in 1998/99, its second consecutive decline. In addition to lower use in the United States, consumption in China and India (the world's two largest consumers) is forecast to decline 1 million and 100,000 bales, respectively. Russia, Brazil, and Turkey are also expected to consume less cotton due in part to the repercussions of Asia's financial problems. World cotton exports in 1998/99 are expected to decline 5 percent to 25 million bales. However, foreign exports are forecast up from last season to 20.5 million bales, as China becomes a net exporter for the first time since 1992/93. Australia is also projected to export more cotton, while Uzbekistan is expected to export substantially less. With the large decline in world cotton production, world stocks are projected to decline 2.8 million bales, to 37.3 million. Both U.S. and foreign stocks are expected to fall in 1998/99, but lower foreign stocks are entirely attributable to China. Excluding China, foreign stocks are expected to increase 900,000 bales. China is forecast to hold 40 percent of the world's cotton stocks at season's end, down from 42 percent in 1997/98. U.S. raw wool production in 1998 is estimated near 27 million pounds, clean, 6 percent below last year and about half the production of the early 1980's. Wool imports, however, are projected to rise to 80 million pounds due to the smaller production and despite lower mill use. Mill demand is forecast at 120 million pounds this season, 24 million below 1997. As a result, 1998 carryover supplies are forecast to increase to 46 million pounds, similar to 1996. U.S. Cotton Situation and Outlook U.S. Cotton Review, 1997/98 The 1997/98 (August/July) season began with U.S. cotton stocks at 3.97 million bales, up 1.4 million from the previous season and the highest in 4 years. The 1997 crop was the second produced under the Federal Agriculture Improvement and Reform (FAIR) Act, which allows producers to respond to the marketplace with total planting flexibility. At planting time, cotton producers did respond to market signals as cotton prices were below those of the previous 2 years. In addition, the attractiveness of net returns for competing crops pulled additional acreage away from cotton in 1997. Cotton planted area was lower in each region than in 1996, except the Southeast, where it was marginally higher. In 1997, only 13.8 million acres were planted to cotton in the United States, nearly 6 percent below the 14.6 million planted in 1996. However, relatively low abandonment in 1997 provided a larger harvested area than in 1996. Harvested cotton area totaled 13.3 million acres and the national yield averaged 680 pounds per harvested acre. As a result, 1997 cotton production totaled 18.8 million bales, only 150,000 bales below 1996 and the fourth largest crop on record. In 1997/98, U.S. cotton demand rose from the previous year to match production after failing to do so in 1996/97. Although total demand for U.S. cotton was nearly 18.9 million bales, the relatively high production kept stocks virtually unchanged. U.S. mill consumption last season approached 11.4 million bales, 2 percent above 1996/97 and the highest in over 50 years. The higher demand for U.S. cotton products was due to the expansion of the U.S. economy and increased trade under the North American Free Trade Agreement (NAFTA). Similar to mill consumption, U.S. raw cotton exports increased in 1997/98 to an estimated 7.5 million bales, 9 percent above the previous season. Despite a reduction in total foreign imports and consumption, the United States had exportable supplies of more competitively priced cotton; U.S. exports also benefited from the "Step 2" competitiveness program and from GSM-102 export credits. Due to the decline in world trade and larger U.S. cotton exports, the U.S. share of world trade rose 3 percentage points to 28.5 percent. The 1997/98 export volume remains an estimate rather than a "final" number as of November because of the large discrepancy between Census and several USDA sources that report cotton export data. USDA is working with Census to resolve the reporting differences and to help prevent these discrepancies in the future. Therefore, the export estimate for 1997/98 may change. With a U.S. cotton supply of 22.8 million bales in 1997/98 and total demand at 18.9 million, stocks at the close of the season on July 31 remained near the beginning level. Ending stocks were placed at 3.9 million bales, a stocks-to-use ratio of 21 percent. Because of the relatively large crop and adequate carryover stocks, producer prices declined. In 1997/98, the average upland farm price was 65.2 cents per pound, compared with 69.3 cents a year earlier, while the extra-long staple (ELS) price was $1.014 per pound, compared with $1.07 in 1996/97. Supply Outlook for 1998/99 U.S. cotton supplies at the start of this season were 3.9 million bales, similar to 1997/98. Meanwhile, at planting time, cotton prices were below those of the previous year and competing crop prices enticed some cotton producers to increase grain acreage at the expense of cotton in 1998. A cool, wet spring across much of the southern United States may have forced some acreage intended for grain back into cotton but may also have discouraged cotton acreage in some regions. Nonetheless, U.S. cotton area fell nearly 1 million acres or 7 percent from 1997 to 12.9 million acres. Upland acreage this season is approximately 12.6 million acres, compared with 13.6 million in 1997. In contrast, ELS area increased to 313,500 acres from 250,000 in 1997. However, much of the increase in ELS area came from a non-traditional growing area where most of the acreage was abandoned. The reduction in total cotton area in 1998 and the weather-related problems experienced this season pushed the national average abandonment rate to its highest in 65 years. In 1933, the cotton abandonment rate was 27 percent. Based on November estimates, the rate for this season is forecast at 19.4 percent, or nearly 2.5 million acres. With the higher abandonment, 1998 harvested area is estimated at only 10.4 million acres (table A). Cotton crop conditions have been unusually poor this year, as the cool, wet start turned into a long, hot drought and played havoc with crop development in many areas of the cotton belt. Based on November 1 conditions, the 1998 U.S. cotton crop is estimated at 13.23 million bales, 30 percent below last season's final crop estimate and the lowest since 1989 (figure 1). Compared with 1997, this season's production estimate is based on a much smaller harvested area and yields. The national average yield is projected at 612 pounds per harvested acre, 68 pounds below 1997 and the lowest since 1995. Upland production is forecast at 12.8 million bales in 1998, with the average yield estimated at 606 pounds per harvested acre. Expected production is lower this year in all regions with the largest decline indicated in the Southwest, where a 2-million-acre decrease in harvested area lowered the upland crop there 2 million bales. Likewise, ELS production is forecast at 446,000 bales, 100,000 below 1997. The decrease in the ELS crop stems from a lower national yield as harvested area is virtually unchanged. California continues to dominate ELS production, accounting for nearly 75 percent of the area. The ELS yield is estimated at 868 pounds per harvested acre, well below last season's record of 1,056 pounds. By mid-November, 80 percent of the total U.S. cotton crop had been harvested, slightly ahead of the 5-year average. Unlike the previous season, the U.S. cotton supply in 1998/99 is expected to include a significant volume of raw cotton imports, which are forecast at 300,000 bales. The small U.S. crop, abundant foreign supplies, and the prospect of exports by China have kept world market prices significantly below those for U.S. cotton and have provided the opportunity for imports. Also, the imminent exhaustion of "Step 2" funds, which are a part of the cotton competitiveness program, and subsequent opening of "Step 3" quotas shortly after the loss of "Step 2" provide additional opportunity for U.S. raw cotton imports to boost supplies. However, the timing of these imports is difficult to predict and additional imports may occur in the early months of the 1999/2000 marketing year to bridge the gap to "new crop" supplies. Despite imports, the total U.S. cotton supply in 1998/99 is projected at only 17.4 million bales, more than 5 million below last season. Demand Outlook for 1998/99 The U.S. cotton demand outlook points to a smaller total offtake in 1998/99. This season, total demand is projected at 15.1 million bales, 20 percent below 1997/98, with both exports and mill use projected to decline. U.S. exports are expected to fall because of the small U.S. crop and the abundant supplies overseas that have kept U.S. cotton prices less competitive than a year ago. Despite an approximate 20-cent decline in world prices from a year ago, U.S. prices have not decreased as dramatically. Early last season, A Index prices were about 3 cents per pound below the comparable U.S. price quote, while early this season, the difference has risen close to 15 cents. Meanwhile, U.S. spot and futures prices have fallen only about 4 cents compared with early 1997/98 (table B). Currently, 1998/99 U.S. exports are projected to fall 3 million bales from last season to 4.5 million, the lowest since 1985/86. Upland shipments are expected to reach only 4.2 million bales, while ELS exports are forecast lower at 340,000 bales due to increased competition from Egypt and possibly China. With the 40-percent decline in U.S. exports and prospects that global cotton trade will drop below 1997/98, the U.S. share of world trade will fall substantially from a year ago. As of November, the U.S. share of world trade is estimated at 18 percent for 1998/99, compared with 28.5 percent last season (figure 2). Based on U.S. Export Sales data through early November, 1 million 480-pound bales of upland cotton have been shipped, compared with 1.2 million in 1997/98. In addition, sales on the books are running well below a year earlier. Upland commitments (shipments plus outstanding sales) for 1998/99 have reached only 3.1 million bales thus far, compared with 4.9 million a year ago. ELS commitments are also below last season with only 233,000 bales committed, compared with 375,000 by early November 1997. U.S. cotton mill consumption is estimated to reach only 10.6 million bales in 1998/99, nearly 7 percent or 750,000 bales below last season (figure 3). A slowdown in the exceptional growth seen in the U.S. economy over the last 2 years and the surge in cotton textile imports this year will keep mill consumption from reaching the 50-year high seen in 1997/98. Upland mill use is projected at 10.5 million bales in 1998/99, while ELS consumption is estimated at 100,000 bales. Based on the first 3 months of data from the Department of Commerce, the seasonally adjusted annual rate of cotton consumption averaged nearly 10.7 million bales. Actual cotton mill use for August through October 1998 reached 2.8 million bales, compared with 3 million a year earlier. Although cotton mill use has fallen from a year ago, manmade fiber use on the cotton system has declined faster during the first 3 months of 1998/99. As a result, cotton's share has averaged 79.3 percent, compared with 78.5 percent for the entire 1997/98 season. Based on these projections of U.S. cotton supply and demand, ending stocks for the 1998/99 season are estimated at a relatively low 2.3 million bales, similar to the 1990 season. With stocks expected to decrease about 1.6 million bales from the beginning level, the stocks-to-use ratio is projected at 15.2 percent, compared with 20.6 percent last season (figure 4). U.S. Textile Trade and Domestic Consumption The overall volume of textile trade in calendar 1998 has risen significantly from 1997, and the trade deficit for textiles will widen in 1998 as the increase in imports outpaces the rise in exports. In calendar 1997, the textile trade deficit reached a record 5.3 billion (raw-fiber equivalent) pounds, 17 percent above 1996. Meanwhile, textile exports during the first 9 months of 1998 expanded to 3.4 billion pounds, compared with 3.2 billion in 1997. However, textile imports through September 1998 increased to nearly 8.3 billion pounds, compared with 7.2 billion in 1997. As a result, the textile deficit for all fibers through the first 9 months of 1998 has approached 4.9 billion pounds, 22 percent above the same period in 1997. Similarly, cotton textile trade for January through September 1998 has increased. Cotton textile exports are running 9 percent ahead of last year, reaching 1.5 billion pounds by September. Meanwhile, cotton textile imports have jumped dramatically, rising 20 percent to 4.5 billion pounds during the first 9 months of 1998. For calendar 1998, cotton textile exports will rise for the 14th consecutive year while imports expand for the 10th year. Exports for the 12 months will likely approach the equivalent of 4.1 million bales (2 billion pounds), 11 percent above 1997. On the other hand, cotton textile imports could reach the equivalent of 12.5 million bales (6 billion pounds), 19 percent higher than 1997. With imports rising faster than exports, the cotton textile trade deficit is expected to expand once again in 1998 to the equivalent of 8.3 million bales of raw cotton (figure 5). On the other hand, total U.S. fiber mill use is expected to decrease from 1997. Mill use of all fibers in 1998 should range between 16 billion pounds and last year's 16.8 billion. Although U.S. cotton mill use is expected to decline from the level reached in calendar 1997, cotton's share of total fiber mill use is likely to remain near 32 percent. Meanwhile, manmade fiber mill use is expected to account for about 66 percent of the total fiber used by U.S. mills in 1998, similar to the past 7 years. With cotton mill use lower and the continued strength of textile imports, calendar 1998 marks the first year in which raw-fiber equivalents of cotton textile imports will exceed the quantity consumed by domestic mills. While the imported quantity has been moving closer to that used by U.S. mills for a number of years, there is little doubt that trade agreements such as NAFTA and the Caribbean Basin Initiative (CBI) have accelerated the situation. In addition, the 1998 jump may also be attributable to the Asian crisis, which provided relatively cheap textile products to the U.S. market. Total domestic consumption (mill use plus net textile trade) of cotton is projected to increase for the second consecutive year. Based on data for the first 9 months, domestic consumption of cotton totaled 7.1 billion pounds, nearly 9 percent above a year ago. By year's end, domestic consumption is expected to approach 9.4 billion pounds, a new record. In addition, the U.S. per capita consumption of cotton will likely exceed last year's 32.5 pounds. In 1998, per capita cotton consumption could approach 35 pounds, with about 14 pounds per capita being produced in U.S. mills (figure 6). World Cotton Situation and Outlook World Consumption Slows Slightly in 1997/98 While world financial markets were preoccupied by the Asian crisis and its effects on the rest of the world in 1997/98, global cotton consumption suffered only a small decline as shrinking mill use in importing East and Southeast Asian countries was nearly offset by increases in cotton producing countries. At 88.2 million bales in 1997/98, world consumption was virtually unchanged from the year before. Foreign consumption fell 1 percent to 77 million bales, but China accounted for all of this change. World cotton production grew 2 percent compared with a year earlier in 1997/98, to 91 million bales. At 72.4 million bales, foreign cotton production was 3 percent higher than the year before, led by a 1.8-million-bale increase in China's crop. Syria, Uzbekistan, and Brazil each harvested about one-half million bales more than the year before, while Australia, Burkina Faso, Turkmenistan, and Turkey each harvested at least 200,000 bales more. India's crop fell 1.8 million bales, while Pakistan's was about 300,000 bales lower and Iran's about 200,000 bales lower. World trade fell in 1997/98, as China's imports contracted by nearly 2 million bales. Southeast Asia's imports fell about 600,000 bales, and a larger domestic crop cut Brazil's imports by nearly 600,000 bales. Buoyed in part by offsetting increases in imports by Mexico, India, and Russia, world trade dipped only 2 percent from the year before, to 26.3 million bales. The United States increased its exports the most in 1997/98, followed by Africa's Franc Zone, Syria, Australia, and Pakistan. India's exports plunged 1.1 million bales from a year earlier, to 150,000 bales, and Argentina, Greece, and Mexico each exported less. World ending stocks rose 3.4 million bales from the year before in 1997/98, to 40 million bales. Most of the gains came in China, where stocks rose 2.1 million bales, but ending stocks also rose by 200,00 to 300,000 bales each in Egypt, Mexico, Turkmenistan, and Brazil. Stock declines of similar magnitudes were recorded in India and Pakistan. Excluding China, world ending stocks as a share of world consumption (but, with China's net imports included in consumption) rose to 33.6 percent, the highest since 1992/93's 37.5 percent. Production, Consumption, and Trade Fall in 1998/99 World cotton consumption is expected to reach 86.4 million bales in 1998/99, while world production is forecast to reach 83.7 million (table C). Consumption and production are each forecast below the year before: consumption 1.8 million bales lower, and production 7.5 million lower (figures 7 and 8). Consumption is expected to decline largely due to the lingering effects of Asia's financial problems, and the production outlook is largely due to a smaller U.S. crop. Foreign production is also foreseen down, mainly reflecting smaller crops in China and Uzbekistan. World cotton trade is expected to reach 25 million bales in 1998/99, while foreign exports are forecast to reach 20.5 million. World trade is forecast below the year before, as U.S. exports shrink and foreign exports rise: world trade is 1.3 million bales lower, and foreign exports are 1.7 million higher. Virtually all of this year's increase in foreign exports can be attributed to China, whose exports are expected to rise 1.8 million bales, from virtually nothing in 1997/98. In addition, year-to-year export gains of between 100,000 and 200,000 bales are foreseen for Australia, Tajikistan, Turkmenistan, Egypt, and Mali. Lower exports are forecast for Pakistan, Mexico, and Syria, while a 600,000-bale decline is forecast for Uzbekistan following a weather-reduced crop. At 3.8 million bales, Uzbekistan's exports are forecast marginally below those of Africa's Franc Zone for the first time ever, and Uzbekistan is forecast to be the world's third largest exporter, following the United States and the Franc Zone. The world's largest increase in imports in 1998/99 compared with the year before is expected in the United States. U.S. imports are expected to increase about 300,000 bales, substantially more than the next largest expected gain, 125,000 bales in Bangladesh. Brazil, Russia, India, and Turkey are each expected to reduce their imports between 200,000 and 400,000 bales. The world's largest importer is expected to be Indonesia, at 1.9 million bales, followed by Italy, Korea, and Brazil. The United States is expected to rank 21st among world importers, accounting for about 1 percent of world imports. World cotton ending stocks are expected to fall to 37.3 million bales in 1998/99, down 2.8 million from a year earlier. Foreign ending stocks are expected to be 1.2 million bales below the year before, slightly less than the decline foreseen for U.S. stocks. However, excluding China, foreign cotton ending stocks are expected to increase about 900,000 bales, to 20.2 million. Small stock increases are foreseen for a number of countries, led by Turkmenistan's 200,000-bale gain. Only Greece, Argentina, and Mali are expected to have stocks increase by as much as 100,000 bales. Lower stocks are expected in Egypt as production falls and exports rise in the wake of new price policies. The ratio of world (excluding China) ending stocks to consumption (with China's net imports included in consumption) in 1998/99 is forecast at 34.2 percent. Compared with a year earlier, this estimate is about 0.5 percentage points higher, and is slightly above the 32.9-percent average of the preceding decade. Global comparisons of stocks to consumption often are made by first excluding China because China's patterns of consumption and stockbuilding are only weakly integrated with those of the rest of the world. When China builds stocks, the rest of the world cannot readily access these supplies, and when mills in China consume cotton, they often cannot readily access supplies in the rest of the world. China does import and export cotton, affecting supplies for the rest of the world, so net trade information for China is substituted into the rest-of-world consumption/stocks figures to arrive at a global number. (See the special article, China's Cotton Sector Under Stress, in this issue). Smaller Foreign ELS Production, Larger Exports Projected According to the International Cotton Advisory Committee (ICAC), foreign ELS production is expected just below 2 million bales in 1998/99, down 250,000 bales from the year before (table D). Production in Egypt is estimated to be substantially lower than a year ago, due to the introduction of various reforms over the last year. Reduced pesticide subsidies to farmers, liberalization of land rental rates, and reduced purchase prices resulted in lower area and yields compared with 1997/98. Lower production is also foreseen in Central Asia in 1998/99, but larger crops in India and China are expected to offset some of the losses in Egypt and Central Asia. In addition to reducing production, Egypt's new price policies are expected to increase exports in 1998/99. Total foreign exports are expected to rise almost 100,000 bales, to 757,000 bales. After Egypt, the next largest increase is foreseen for China, which had no ELS exports during 1995/96-1997/98. Lower exports are forecast for Central Asia, Sudan, and a few others. Egypt's decline in ending stocks is expected to account for a substantial portion of the entire 500,000-bale decline expected for all foreign ELS ending stocks in 1998/99. With foreign ELS ending stocks forecast by the ICAC at 770,000 bales, and foreign consumption forecast at 1.9 million, a foreign ending stocks/consumption ratio of 41 percent is foreseen. Including USDA's U.S. forecasts to reach a global total, gives an expected global ratio of 42 percent, compared with 56 percent in 1997/98. U.S. Wool Situation and Outlook U.S. Sheep and Wool Industry Continues To Contract The U.S. sheep and wool industry continued to contract for the eighth consecutive year. Sheep stocks on July 1, 1998, were 5.6 million head, 5 percent below July 1997 and 42 percent below 1990. Sheep shorn during 1998 are expected to decline to 6.6 million head, compared with 7.0 million during 1997. Wool production is estimated at 28.6 million pounds, clean (53.9 million, greasy) for the 1997 marketing year (table E). Lower sheep numbers in 1998 imply production could decline to near 27 million pounds, clean, nearly 6 percent below last year. Expected production this year would be the lowest on record and significantly below the early 1990's (figure 9). During 1997, U.S. raw wool imports totaled 76.4 million pounds, clean, 1 percent above a year earlier (table F). With lower production forecast this year, imports are expected to rise slightly to 80 million pounds. Shipments through the end of September have reached nearly 54 million pounds, compared with 52 million last year. Imports of fine wool accounted for 65 percent of total 1998 shipments. Imports of unimproved and other grades not-finer-than 46's were 19 million pounds through September. Shipments from New Zealand represented 80 percent of these coarser grades while Australia accounted for 84 percent of the finer wools. Total U.S. supplies are forecast at 168 million pounds, clean. Raw wool mill consumption is forecast at 120 million pounds, 24 million below 1997. Mill use during the first 6 months of 1998 totaled 66 million pounds. Apparel wool consumption, at 58.8 million pounds, was 12 percent below 1997, while carpet mill use was nearly 500,000 pounds above 1997 at 7.3 million (table G). The woolen system used nearly 29 million pounds and the worsted system used 30 million during January through June of this year. Wool top production in the first 6 months totaled 28 million pounds compared with 29 million in 1997. Raw wool exports during the first 9 months of 1998 were 1.6 million pounds, clean, compared with 4.2 million in the corresponding period of 1997. Shorn wool shipments were 1.3 million pounds and not-shorn (pulled) exports were 244,960 pounds. Carbonized wool exports totaled only 77,254 pounds through September, compared with 1.7 million in 1997. The majority of shorn wool exports went to Germany, Mexico, France, United Kingdom, and Canada. Exports of not-shorn wool went primarily to Canada. Exports of wool top during January through September 1998 were 4.7 million pounds, 16 percent below a year earlier. The average price received was $3.28 per pound, compared with $2.83 last year. The value of wool top shipments totaled $15.5 million with the majority of shipments going to Mexico, China, and Italy. Top imports were 1.9 million pounds, 16 percent above January-September 1997. Australia and Germany were the largest sources of top imports, accounting for 44 and 28 percent, respectively, of total shipments. U.S. prices for clean, mill-delivered territory raw wool declined sharply during the first 8 months of 1998. Finer grades 64's and 62's averaged $2.36 and $2.06 per pound, respectively, last January. By October, prices of 64's averaged $1.15 per pound and the 62's averaged $1.05. The 60's and 58's have followed a similar trend, decreasing throughout the year. However, while the finer grades have declined about 50 percent since January, the coarser grades have declined 30 to 40 percent. The 60's and 58's fell to $1.00 per pound and $0.98, respectively, during October. By October, the 56's had declined to $0.90 per pound and the 54's had fallen to $0.80. Domestic prices of Australian raw wool of all grades have generally moved lower each month since July 1997. The 80's dropped to $1.90 per pound and the 70's declined to $1.51 in October. The 64's established a season low in October of $1.40 per pound. Grade 58's prices have declined from $1.85 per pound in January to $1.32 in October. Foreign Wool Situation and Outlook Weak Wool Demand, Lower Production During the 1997 Season Despite slightly larger world wool production during the 1996 season, the industry continued the long-term downward trend in 1997. World sheep numbers declined by 9 million head in 1997 to 1 billion. With lower sheep numbers, world wool production fell to 3.2 billion pounds, clean, 3 percent below the 1996 season (table H). Lower production occurred in all the major producing countries. Australian production dropped 45 million pounds to 996,000 pounds, clean. Wool production in the New Independent States (NIS) continued its long-term decline during the 1990's. NIS production, at 205 million pounds during the 1997 season, is 60 percent below the 1990 output. Production declines of nearly 10 percent occurred in both Uruguay and Argentina in 1997. World exports of raw wool, at 1.3 billion pounds, decreased nearly 10 percent from 1996 shipments. Australian wool exports declined nearly 100 million pounds to 845 million. World wool consumption declined for the sixth consecutive year in 1997. Raw wool consumption, at 3.1 billion pounds, clean, was 3 percent below 1996. Although wool mill use declined in 1997, world carryover supplies dropped for the seventh consecutive year. World carryover at the end of the 1997 season is estimated at 558 million pounds, clean. However, a substantial percentage of stocks is owned by Wool International. On October 15, Wool International stopped selling its stockpile. The Australian government, under pressure from the wool industry, is in the process of transferring ownership of the wool (1.05 million bales) and assets of Wool International to growers who have paid wool levies between 1993 and 1996. The transfer is expected to be completed by July 1, 1999. As a result, a significant amount of wool stocks will not be available until next summer. Weak mill consumption during 1997 was in part due to the slump in Asian demand. Most Asian countries imported less raw wool than a year earlier. The largest importer, China, purchased 343 million pounds of foreign raw wool in 1997, 11 percent below a year earlier. Raw wool import demand was also lower in Japan, Taiwan, and South Korea. However, larger shipments to EU countries helped offset lower imports by Asia. Italy's wool imports increased 17 percent to 313 million pounds. Larger wool imports also occurred in United Kingdom, Italy, Germany, and Belgium. Wool prices moved lower during the 1997 season. The Australian market indicator (a weighted average index of 15 categories) averaged A 717 cents per kilogram during July 1996, then fell to A 585 cents per kilogram by June 1997. The New Zealand market indicator also declined from July 1996 through February 1997. However, prices firmed in March and by June, rose to NZ 480 cents per kilogram, about the same level as the beginning of the 1997 season. The South African market indicator followed a similar pattern as Australian prices, declining throughout the season. The market indicator declined from near SA 2,600 cents per kilogram during July and August to SA 1,738 cents at the end of the 1997 season (June). Manmade Fibers The U.S. manmade fiber industry during the first three quarters of 1998 declined slightly from a year earlier. Total shipments were 7.79 billion pounds, down 1.6 percent from a year ago. Noncellulosic shipments, at 7.52 billion, were down 0.5 percent. Total manmade fiber production, 7.86 billion pounds, was 1.3 percent below a year earlier. Noncellulosic fiber production, 7.58 billion, was 0.2 percent less than in 1997 (appendix table 42). Manmade fiber stocks in fiber producers' plants at the end of the third quarter were 0.83 million pounds, 14.5 percent above a year earlier. Domestic shipments of noncellulosic fibers during January-September 1998 were 6.99 billion pounds, 0.1 percent more than in 1997. An increase in filament fiber shipments more than offset a decrease in staple shipments. Domestic shipments of noncellulosic filament fibers were 4.14 billion pounds, 1.8 percent above a year earlier. Filament shipments of nylon were 1.46 billion pounds; polyester, 1.08 billion; and olefin, 1.60 billion. In contrast, noncellulosic staple domestic shipments, at 2.85 billion pounds, were 2.1 percent less than a year earlier. Staple shipments of polyester were 1.62 billion pounds; nylon, 0.58 billion; olefin, 0.43 billion; and acrylic, 0.22 billion pounds. According to the capacity survey by Fiber Economics, Inc. in May 1998 (the latest data available), total in-place capacity to produce manmade fibers in 1998 is 12.6 billion pounds, up 6.1 percent from 1997 (appendix table 42). Most of this increase resulted from the 1997-1998 capacity increase of polyester fibers, 8.5 percent, and olefin fibers, 9 percent. Two fibers had slower growth, nylon at 1.5 percent and acrylic at 0.2 percent. The planned capacity for total manmade fibers by the year 2000 is expected to be 13.5 billion pounds, 3.6 percent (an average annual increase) above the 1998 level. Most of the planned capacity growth comes from polyester, 6.3 percent, and olefin, 3.2 percent. Nylon fiber capacity is expected to grow 0.6 percent by 2000, with no growth expected for acrylic fibers. The carpet market continues to consume more fibers in facing and backing uses than any other fiber market (appendix table 43). During January-June 1998 (the latest data available) this market took 1.98 billion pounds, 4.2 percent more than a year earlier. Noncellulosic carpet use accounted for more than 42 percent of total noncellulosic fiber domestic shipments. Nylon dominates the carpet market, constituting 52 percent of the total use of noncellulosic carpet fibers. Nylon staple carpet fibers were 94 percent of nylon staple domestic shipments, while nylon filament carpet fibers were 69 percent. Preliminary data for third-quarter 1998 indicate that about 1.57 billion pounds of nylon were used in carpets during January-September 1998, 6.4 percent more than a year earlier. The use of olefin fibers in facing and backing during the first half of 1998 was 822 million pounds, 2.7 percent above a year ago. Olefin fibers constitute 41 percent of the noncellulosic fibers used in carpets. Carpeting is the most important use of olefin fibers at 61 percent. Woven textile production remained the second largest outlet for noncellulosic fibers, taking 22 percent of the January-June 1998 domestic shipments. The woven market used 1.02 billion pounds, 1.3 percent less than a year ago. Two fibers made up 84 percent of this market, polyester at 68 percent and olefin at 16 percent. The knit market took 583 million pounds, 13 percent below a year earlier. Shipments of manmade fibers to the knit market were 12 percent of total domestic shipments. Three fibers dominated the knit market; polyester at 355 million pounds constituted 61 percent; nylon at 112 million equaled 21 percent; and acrylic at 96 million totaled 16 percent. Almost all chemical prices were affected by the Asian financial crisis, which deepened during the year. Its negative effect was not only on chemical exports but also on exports of the various noncellulosic fibers. The price of benzene (a precursor to many chemicals) began the year at about $1.10 - $1.15 per gallon but expanded capacity and higher inventories caused the price to decline to 75-80 cents during February-May. During the third quarter, the price averaged 70-75 cents per gallon because of slower demand (figure 11). The price of cyclohexane, a basic chemical used in nylon production, somewhat follows the price of benzene. During January-April the price was $1.20 per gallon. Later in the year it declined to $1.04, reflecting the lower price of benzene. The price of paraxylene, a precursor to polyester fibers, was 19 cents per pound during the first quarter, then declined to 15.76 cents in July-October (table I). The price of polymer grade propylene, a precursor for acrylonitrile (a raw material for acrylic fibers) and olefin fibers, fell during the year from a high of 19.5 cents per pound in January to 13.5 cents in September-October. The price of acrylonitrile listed at 53 cents all year, but it was reported to be discounted as much as 50 percent. The price of ethylene glycol (a raw material used to make polyester fibers) varied during the year from a high of 32-38 cents per pound to a low of 20-24 cents in October. The price of caprolactam (a raw material used to make nylon fibers) remained between 90 and 93 cents per pound, although heavy discounting has been reported. Special Article Factors Affecting the U.S. Farm Price of Upland Cotton by Leslie A. Meyer 1/ Abstract: An annual model that explains the U.S. upland cotton farm price includes various market components, as well as government loan and storage programs in selected years. With the declining role of government agricultural programs, however, market supply and demand conditions, both in the United States and around the world, are now more influential in determining U.S. cotton prices. An overall measure of supply and demand is provided by the stocks-to-use ratio, which is a useful tool in explaining inter-year price movements. The regression model presented explains 92 percent of the variation in upland cotton prices during the 1978 to 1996 marketing years. In addition, the effect of shifting one variable at a time and its impact on upland farm prices is highlighted for each explanatory variable. Keywords: Upland cotton, farm price, price determination, ending stocks, stocks-to-use ratio. Cotton is one of the world's most important textile fibers, accounting for about 45 percent of fiber production. In 1997/98, U.S. farmers produced over 20 percent of the world's cotton, while U.S. mills used 13 percent of the world total or 20 pounds for each person in the country. And, because cotton is a major raw material for the textile and apparel industries, cotton production, marketing, and manufacturing affect the lives of many people from producers to consumers. Cotton is a vast and dynamic industry that accounts for more than $25 billion in products and services annually in the United States. 1/ Agricultural Economist, Economic Research Service, United States Department of Agriculture. The author would like to acknowledge the Economic Research Service's Paul Westcott and Linwood Hoffman for their recommendations during this project. Many interrelated factors help determine the price of U.S. cotton, such as the fundamental elements of supply and demand and the effects of agricultural policy. These factors have various implications for the cotton industry, as well as many other industries, that extend well beyond the farm gate. Background Recent agricultural legislation has altered the nature of U.S. Government commodity programs, advancing the efforts toward increased market orientation. In particular, the 1996 Farm Act decoupled the income support programs that were in place for many years, and shifted the price volatility risk from the government to producers (see Young and Westcott). As a result, market information has become increasingly essential as producers and other market participants seek to make informed pricing decisions under a more market-oriented agricultural environment. The U.S. Department of Agriculture (USDA) analyzes and publishes monthly supply and demand data and information pertaining to the major agricultural commodities. In addition, USDA publishes forecasts of season-average farm prices for these commodities. Cotton is the exception, however, as USDA is prohibited by law from publishing cotton price forecasts (see Townsend, 1989). Nonetheless, USDA calculates unpublished cotton price estimates for internal Departmental use each month along with the other reported U.S. and foreign cotton supply and demand estimates. Additionally, USDA analysts examine reasons explaining historical movements in commodity prices, including cotton prices. This paper analyzes some of the factors that have historically influenced U.S. farm level cotton prices. An annual upland cotton price model is developed and designed to be used by USDA with other price estimation techniques in the analysis of market forces affecting the supply and demand for cotton. For an annually produced crop like cotton, ending stocks for a particular year summarize these supply and demand forces and can be a useful indicator of price movements. Annual cotton prices tend to be negatively correlated with ending stocks; high stocks of cotton tend to depress prices while low stocks tend to support prices, all other things being equal. Agricultural policies and programs, such as the nonrecourse loan program, have also influenced prices, although the results of these programs have differed under the various farm policy environments. The basic loan program allows producers to obtain loans for a commodity, in exchange for placing that commodity as collateral in approved government storage. At the producer's option, the loan, including any accrued interest and/or storage charges, can be repaid at any time during the loan period, or the commodity can be forfeited to the government at the end of the loan period if market prices are not high enough to make economic sense for the producer to repay the loan. For upland cotton, significant forfeitures occurred in the early 1980's. However, with the passage of the 1985 farm legislation, a new program, the marketing loan, provided a repayment option below the loan rate for upland cotton when the U.S. price was not competitive on the world market. The 1985 Act allowed upland cotton producers to repay loans at the lower of the loan rate or the adjusted world price (AWP). The marketing loan remains in effect under the 1996 Act and has eliminated the large forfeitures seen in the past. Previous research addressing factors affecting commodity prices has included the stocks-to-use ratio. This ratio is defined as the stock level of a given commodity at the end of a particular period divided by the total use of that commodity for that same period. Thus, the stocks-to-use ratio provides a good summary of the year's supply and demand situation and is comparable over time. Similar to other research, the stocks-to-use approach was employed as the basis for this price determination analysis. There is no single price for cotton, as numerous daily, monthly, and annual price series reflect different markets, qualities, and/or delivery times. Nevertheless, each series is linked by the fundamental elements of supply and demand. This paper presents efforts by the Economic Research Service to reevaluate factors affecting farm price movements for major commodities under the more market-oriented agricultural environment. While the annual estimate of the average U.S. upland cotton farm price, in and of itself, may be less useful to those who must follow monthly or daily prices, the factors that influence the annual price provide a framework that allows a better understanding of intra-year price movements and perhaps a more informed decision about cotton price movements in general. Model Specification The model specification presented here follows a general equilibrium model with a basic framework relating prices to ending stocks. In its simplest form, which excludes government price support programs, stocks are a function of price; and, in equilibrium, prices are inversely related to stocks (Westcott). This basic model is used, to which adjustments are introduced that shift the price relationship. The following functional form was used in estimating annual upland cotton farm prices: Ln (PRICE) = a + b Ln (S/U) + c CHFSTKS + d INDEX + e DSU + f Ln (LDP)*D1 + g Ln (1+CCC/USE) where the dependent variable is the natural log of the marketing year average price received by upland cotton producers. This price is a weighted annual price based on marketings throughout the year and is reported in cents per pound. The price model's independent variables hypothesized to affect prices account for both U.S. and foreign market supply and demand conditions and U.S. agricultural policy programs, which have altered the supply and demand of cotton in the past. The basic model relates stocks-to-use ratios to prices. The S/U variable is the upland stocks-to-use ratio for a given year, and is reported as a percentage. This variable indicates the tightness of U.S. supplies relative to demand. Figure A-1 shows a historical plot of U.S. farm prices for upland cotton and stocks-to-use ratios for the 1978 through 1996 marketing years. As the stocks-to-use ratio changes, the effect on prices is expected to be in the opposite direction. The variable is included in the model in logarithmic terms. Other supply and demand conditions affecting price are introduced with the next three variables. The CHFSTKS variable is the percentage change for a given marketing year, from the previous year, of total foreign ending stocks, excluding China's. This variable is representative of the foreign cotton supply and demand situation outside of China. As this variable changes, the effect on prices is expected to be in the opposite direction. For example, if the percentage change in foreign stocks (less China) declines from the previous season, this indicates that global cotton demand exceeded global supplies, so U.S. farm prices are expected to increase, all other things being equal. Stocks in China are excluded from this variable partly because there has been much debate in the cotton industry over the size of China's stocks. Market analysis of the foreign situation and outlook often omit China, as policy actions there pertaining to cotton do not always reflect responses to market supply and demand conditions, particularly related to stocks. However, because China is the largest producer and consumer of cotton, changes in China are often reflected in the supply and demand balance sheets of other countries, including the United States. Therefore, omission of China stocks is not expected to reduce the overall effectiveness of the specified price model. The INDEX variable weights the percentage of cotton crop acreage forward contracted by the end of September of a given year multiplied by the September average of the December futures contract of the same year. The index is inversely related to early season supply expectations. For example, if supply expectations are low, December futures tend to be higher and therefore more acreage tends to be forward contracted. And, because the farm price is based on marketings (including forward contract deliveries), the larger the percentage of acreage forward contracted, the greater role the index plays in determining the marketing year average price. Thus, the index is positively related to price. The DSU variable represents the effects of starting a new marketing year with very tight beginning stocks. DSU is a dummy variable equal to one in years when the stocks-to-use ratio of the previous year is less than or equal to 22.5 percent and zero in all other years. During the sample period, the stocks-to-use ratio was less than or equal to 22.5 percent in 1979-1980, 1983, 1989-1991, and 1993-1996. Therefore, DSU is equal to one in the marketing year following each of these years. When the stocks-to-use ratio of a given year is less than or equal to 22.5 percent, the subsequent marketing year's prices may be high and total use may be limited, particularly in the early months before the new crop is harvested and becomes available. This variable is an intercept shifter and is expected to have a positive sign. The next two variables are program variables related to current or past agricultural policy. The LDP variable is the loan deficiency payment rate for upland cotton, equal to the loan rate less the effective loan repayment rate. This program was first implemented for the 1986 marketing year as established in the 1985 Act. Since implementation, loan deficiency payments have been issued in only 4 years (1986 and 1991-1993) during the sample period and ranged from 6.35 to 11 cents per pound. Natural logs of this variable are used in the model. The transformed loan deficiency payment rate is multiplied by D1, a dummy variable equal to one during years when the loan repayment rate was below the loan rate and zeroes in all other years. Since loan deficiency payments are not included in the reported average price received by producers, the addition of this payment to the reported price would reflect a "more accurate" effective price received. Therefore, when loan deficiency payments are made, market prices tend to be lower, so there is a negative correlation to the price received by producers. This variable is an intercept shifter for the years when loan deficiency payments are made. The final variable is CCC/USE, which is the Commodity Credit Corporation (CCC) stocks divided by total use for a given year and expressed as a percent. Natural logs of one plus CCC/USE are used in the model, keeping the transformed variable from falling below zero. This program variable becomes relevant when CCC inventories are large, as during the 1982 through 1985 crop years when the loan program for upland cotton was influential in forfeitures to the CCC. Between 1982 and 1985, CCC inventories of upland cotton, ranging from 124,000 to 775,000 bales, were substantially larger than at any other time during the data period analyzed. The transformed variable is positively related to price and is an intercept shifter for years when CCC inventories of upland cotton are held. Results The upland farm price model was estimated using ordinary least squares regression, using annual data for marketing years 1978 through 1996. The estimated regression equation is: Ln (PRICE) = 4.32596 - 0.09740 Ln (S/U) - 0.00276 CHFSTKS + 0.00559 INDEX (-4.71) (-3.90) (5.17) + 0.06293 DSU - 0.04377 Ln (LDP)*D1 + 0.07212 Ln (1+CCC/USE) (3.24) (-4.27) (3.97) Adjusted R-squared = 0.9247 F-statistic = 37.84 Standard error of the estimate = 0.03102 Durbin-Watson statistic = 1.794 Degrees of freedom = 12 Over 92 percent of the variation in (the log of) annual upland cotton prices is explained by the equation. The numbers in parentheses below each coefficient are the t-statistics. Each coefficient has the expected sign and each is significant at the 1-percent level. Model Performance Figure A-2 shows historical upland cotton farm prices and the associated predicted values derived from the estimated equation. As illustrated, the price model tracks actual cotton prices well. Most differences between the actual upland farm price and the model estimate are less than 2 cents per pound. The largest difference (3.2 cents) occurred in 1996, a recent year where revisions to foreign supply, demand, and stock data are still possible. In addition, mean absolute errors and mean absolute percentage errors were calculated for the full estimation period, and for selected subsamples of recent marketing years. Throughout the entire estimation period, the mean absolute error was 1.3 cents per pound, while the mean absolute percentage error was 2.1 percent. During one subsample period covering the 1991-1996 marketing years, the errors are slightly higher due to the difference in 1996. For the 1991-1996 period, the mean absolute error was 1.5 cents per pound, while the mean absolute percentage error was 2.3 percent. However, if 1996 is excluded from the subsample, the mean absolute error declines to 1.1 cents per pound and the mean absolute percentage error decreases to 1.8 percent. These statistical measures indicate good performance of the upland cotton pricing model. Model Features Upland stocks-to-use ratios have been below 30 percent since 1988, although they ranged from 13 to 113 percent during the estimation period. While stocks-to-use ratios for upland cotton are expected to remain below the 30-percent level under the current policy environment, the various features of the regression equation results are illustrated using stocks-to-use ratios ranging from 5 to 100 percent. A base model relationship was first determined by varying stocks-to-use while holding the CHFSTKS and INDEX variables at their sample means, DSU equal to one, and zeroes for the other variables (see table A-1). The base model results (solid-line curve) and the different features of the regression model are illustrated in figures A-3 through A-7. For each graph, upland cotton prices are plotted against ending stocks-to-use ratios, adjusting the variables from logarithms to levels. The base model curve is identical in each figure. Therefore, each graph illustrates the effect of shifting one explanatory variable at a time from the base model values, highlighting that variable's influence on prices. The base model equation and the effects of the previous year's stocks-to-use ratio on cotton prices are shown in figure A-3. The higher (base model) price curve incorporates the effect of upland stocks-to-use ratios in the previous year of less than or equal to 22.5 percent (DSU=1). The lower dotted-line price curve represents the price effect of upland stocks-to-use ratios in the previous year of greater than 22.5 percent (DSU=0), holding all other independent variables at their base model values. Price impacts shown in figure A-3 range from -3.4 to -4.5 cents per pound when the previous year's stocks-to-use ratio is greater than 22.5 percent when compared with the base model. Figure A-4 illustrates the effects on cotton prices for different percentage changes in foreign stocks less stocks in China. The base model is again represented by the solid-line curve, while the two dotted-line curves represent one standard deviation above and below the sample mean of the variable, CHFSTKS. This deviation corresponds to a 13.6-percentage-point increase in foreign stocks (less China) and an 11.9-percentage-point decrease. Compared with the base model, price impacts illustrated here are -2 to -2.6 cents per pound with the increase in competitor stocks and 2 to 2.7 cents per pound with the decrease in competitor stocks. Figure A-5 indicates the sensitivity of the upland cotton price function for different INDEX values related to the amount of the crop that was forward contracted and the level of the December futures contract. The base model along with the two dotted-line curves, representing one standard deviation above and below the sample mean, are pictured. One standard deviation above the mean corresponds to an index of 26.1, while one standard deviation below corresponds to an index of 5.0. Price impacts shown here range from 3.4 to 4.6 cents per pound with the higher index and -3.2 to -4.2 cents per pound with the lower index when compared with the base model. As indicated from the results in figures A-4 and A-5, one standard deviation around the INDEX variable causes a greater impact on prices than one standard deviation around the CHFSTKS variable. However, the INDEX variable would be finalized early in the marketing year (early October), while the competitor stocks variable would be expected to change throughout the season as better information and data became available. Therefore, the effect on prices for the CHFSTKS variable may shift several times throughout the season. The effect of loan deficiency payments on upland cotton prices is shown in figure A-6. The upper curve represents the base model when no loan deficiency payments are made. The lower dotted-line curve, on the other hand, illustrates the price effect of loan deficiency payments averaging about 9 cents per pound, corresponding to the mean logarithmic value of these payments for the 1986 and 1991-1993 marketing years. As discussed earlier, these payments are not included in the reported farm price for upland cotton. Therefore, the payment value should be added to the reported price to get a "more representative" indicator of the effective price received by producers in years when loan deficiency payments are made. Price impacts shown in figure A-6 range from -5.1 to -6.8 cents per pound, compared with the base model, corresponding to average loan deficiency payments of about 9 cents. Figure A-7 illustrates the effect of CCC stocks on cotton prices. The solid-line curve represents the base model when there are no inventories of CCC stocks. The upper curve indicates the price supporting effect of having stocks unavailable to the marketplace and held as CCC inventory. The dotted-line curve represents a CCC stocks/use ratio of about 3 percent, corresponding to the mean logarithmic value of the 1982-1985 marketing years, the period when CCC inventories of cotton were much higher than at any other time during the estimation period. Price impacts illustrated on the graph range from 5.7 to 7.8 cents per pound when average CCC stocks/use ratios of about 3 percent are present when compared with no CCC stocks. Out of Sample Estimate The largest error in the model, as discussed earlier, occurred in the last year of the estimation period, 1996. So, there was a concern about the model's performance in future years. Did the 1996 farm legislation provide additional factors, not accounted for in this model, that were more influential in determining farm prices or was 1996 just an outlier? While USDA, by law, cannot forecast cotton prices, the first out-of-sample estimate for this model was the 1997 marketing year that ended in July 1998. Model variables, therefore, are not expected to change significantly and the average upland price received by farmers has been reported by USDA for the 1997 season. The price model presented here was used to estimate a farm price for upland cotton using the latest available data (November 1998) for the 1997 marketing year. The model estimated the 1997 marketing year average price for upland cotton to be 63.2 cents per pound, while the actual farm price, reported by USDA, was 65.2 cents per pound. Although the regression equation underestimated the price by 2 cents per pound, the first out-of-sample estimate is within one standard error of the model estimate. And with limited data, it is difficult to determine whether the 1996 farm legislation has introduced "new" factors that influence farm prices that are not already captured by the model estimated here. Perhaps additional research into the model's underestimation of upland farm prices in 1996 and 1997 may be useful for future work. Conclusions The upland cotton price determination model presented in this article uses a stocks-to-use ratio framework. In addition, the model addresses issues regarding the historical influence of government commodity loan and storage programs on cotton prices. These programs were shown to have affected upland cotton price determination during the early 1980's, prior to the passage of the 1985 farm legislation. With the implementation of the 1985 Act, however, storage programs have not influenced upland cotton prices significantly, but the cotton loan program remains an important component. As U.S. prices are more closely tied to world market conditions, foreign market supply and demand expectations, as well as U.S. conditions, have played a larger role recently in affecting the price received by U.S. upland cotton producers. The stocks-to-use ratio and other variables identified in the model have shown the importance of market supply and demand factors on upland cotton price determination. The statistical model's evaluation measures and the graph illustrating the actual prices and model estimates indicate the effectiveness of the regression model in the determination of upland cotton prices. This is particularly relevant given the wide range in upland cotton prices over the sample period (1978-1996), as well as the changes in agricultural policy that have had varying impacts on prices. Carryover stocks of upland cotton as a percent of total use are typically smaller in the 1990's than in any previous period. Therefore, price determination is in the steeper portion of the price function and implies more price responsiveness to shocks. However, with the continued full planting flexibility currently in place, market signals and producers' responses to these signals may help mitigate the large annual variability in upland cotton prices seen in the past. Finally, the relatively simple structure and the limited data needs of the regression model presented in this article allow for sensitivity analysis under various market supply and demand conditions that may develop during a given year or from one year to the next. While USDA does not publish cotton price forecasts, this model, along with other models, is used to analyze historical cotton price movements and is used in USDA's short-term market analysis as well as long-term baseline projections. References Collins, Keith. "Cotton Price Analysis," Cotton and Wool Situation and Outlook, CWS-28, August 1981, pp. 18-21. Glade, E.H., Jr., L.A. Meyer, and S. MacDonald. Cotton Background for 1995 Farm Legislation, Agricultural Economic Report 706, April 1995. MacDonald, Stephen. "Forecasting World Cotton Prices," The Ninth Federal Forecasters Conference - 1997: Papers and Proceedings. D.E. Gerald (ed.). National Center for Education Statistics. Washington, DC. 1998. Townsend, Terry. "Accounting for Inflation When Estimating Cotton Prices," 1984 Proceedings of the Cotton Beltwide Conferences. National Cotton Council. 1984. Townsend, Terry P. "This Was a High-Powered Instrument That Sent Its Projectile to the Vitals of the Industry--Why USDA Does Not Forecast Cotton Prices," 1989 Proceedings of the Cotton Beltwide Conferences. National Cotton Council. 1989. U.S. Department of Agriculture. Economic Research Service. Cotton and Wool Situation and Outlook Yearbook, CWS-1997, November 1997. U.S. Department of Agriculture. Economic Research Service. The Cotton Industry in the United States, Agricultural Economic Report 739, July 1996. U.S. Department of Agriculture. Foreign Agricultural Service. Cotton: World Markets and Trade, Various issues. U.S. Department of Agriculture. World Agricultural Outlook Board. World Agricultural Supply and Demand Estimates, Various issues. Westcott, Paul C. "Market Factors and Government Programs Affecting U.S. Corn Prices," Feed Situation and Outlook Yearbook, FDS-1998, USDA-ERS, April 1998, pp. 46-51. Young, C. Edwin and Paul C. Westcott. The 1996 U.S. Farm Act Increases Market Orientation, Agricultural Information Bulletin 726, August 1996. Special Article China's Cotton Sector Under Stress by Hunter Colby and Stephen MacDonald 1/ Abstract: China maintains a government monopoly on cotton purchases, but there are indications this policy could be relaxed in the future. This monopoly is a vestige of the centralized planning process formerly applied to the entire economy, and in recent years its enforcement has been tightened and relaxed as cotton surpluses have ebbed and flowed. Similarly, China's trade policy has alternated between import and export restrictions to accommodate perceived scarcity and surfeit. Currently, China's domestic supplies of cotton and other agricultural products are large. How China's policymakers address these circumstances will also be affected by the financial health of the country's numerous state-owned enterprises, particularly those producing textiles, and the costs associated with purchasing cotton through the government's Cotton and Jute Company. Keywords: Cotton, China, policy, procurement, textiles, state-owned enterprises, Cotton and Jute Company Cotton is still a "planned" commodity in China, but progress towards more market-driven decisionmaking has been accelerating. China's cotton stocks are at nearly their highest level ever, government procurement is clogged with cotton purchased above world prices, and the state-owned segment of the textile industry is reeling from growing inventories and debts. To alleviate this situation, China's government has resorted to price bands, cotton fairs, import restraints, and procurement price cuts. Now, China may be on the brink of taking some of the biggest steps in 20 years toward liberalizing its cotton production and purchasing. This paper lays out the institutional and policy framework that has directed cotton supply and use in China in recent years, and indicates possible directions for its continuing evolution. 1/ Agricultural Economists, Economic Research Service, United States Department of Agriculture. The authors would like to acknowledge the support provided by Foreign Agricultural Service's Office of International Cooperation & Development for making possible a 1996 U.S-China Scientific Exchange, and the Emerging Markets Office for its support of the ERS Technical Assistance project for cotton in 1997. China's economy is guided by both independent market events and by government decisions regarding--for example--purchasing, lending, and prices. About 20 years ago, led by the agricultural sector, the role of market-driven decisions began steadily growing, and cotton is now one of the few remaining goods with such obvious vestiges of central planning as the prohibition of private trading. Tobacco and cotton are the only agricultural products for which the state maintains a monopoly on purchases from farmers. For cotton, this monopoly has not been perfectly enforced, and has been relaxed during some years of relative surplus. Planning and Planting China's government develops economic plans annually and reviews them quarterly. The role of these plans has declined tremendously in the last 20 years, but for cotton, in particular, the role is far from trivial. Ultimately, China's economic decisionmaking is guided by the Communist Party, which controls the government. Government decisionmaking is implemented through the State Council. To formulate a plan for cotton production, the State Development Planning Commission receives information on expected domestic cotton demand from the All-China Federation of Supply and Marketing Cooperatives (SMC) and the State Textile Industry Bureau. Information from the Ministry of Agriculture helps determine likely production levels in the coming year. A plan covering how stock adjustments and trade can be used to balance consumption and production is then developed. Appropriate prices for cotton purchases and sales are also determined to balance the needs of cotton production, competing crop production, farmer welfare, and the financial health of the textile industry. The cotton plan is also developed in the context of the government's plans for the rest of the economy. Price is one of the most important of the government's tools for fulfilling its goals. In a mixed economy like China's, farmers can make substantial adjustments in production depending on prices. However, history suggests that government prices have not always fulfilled government objectives: price announcements have sometimes come too late to influence farmer decisions during the current year, and have sometimes come too early, leading farmers to withhold their crop until the following year. Problems with the timing and level of procurement prices probably reflect the need to find consensus among the different interests represented in the planning process as well as the inherent difficulties of accurately forecasting future supply and demand trends. While prices are one signal to farmers regarding how much cotton to produce, local governments may also at times pressure farmers to achieve specific planting and production targets. Under the plan, achieving a planned production level becomes government policy, and local governments, particularly in areas designated as cotton production bases, take responsibility for ensuring an appropriate level of planting. While government control of day-to-day life in China has declined in recent years, government influence remains large in many aspects of farmers' lives, and may affect their agricultural production decisions. In China's largest producing province, Xinjiang, the Production and Construction Corps (PCC) accounts for about one-third of production. The PCC is a military entity charged with developing and defending China's most remote province. As state farms, PCC producers are more tightly integrated into the economic planning system, and more quickly and completely responsive to central government policy. Purchasing, Procurement, and Distribution As cotton is harvested in the fall, farmers bring both ginned and unginned cotton to SMC and the China Cotton and Jute Company (CJC, a subsidiary agency of the SMC) purchasing stations. The cotton is graded and the farmer receives payment from CJC based on fixed procurement prices for that quality of cotton. Official policy during much of the 1990s is that farmers are to be paid immediately rather than over the following months, and CJC therefore borrows from the China Agricultural Development Bank to ensure sufficient capital to make these initial purchases. The China Agricultural Development Bank is what is known as a policy bank, lending to government agencies in accordance with government policy and assuring they can fulfill their policy obligations. The CJC is expected to receive payment at the time it transfers cotton to textile firms and then repay its bank loans. How closely practice actually follows the above outline depends on the relative supply and demand for cotton at prices fixed by the government for purchases from farmers and sale to mills. During the mid-1990s, insect infestations, rising grain prices, and a legacy of several years of low procurement prices and late payments to farmers, led to crop shortfalls. With the economy booming, farmers found alternative marketing channels that paid prices well above procurement prices, and mills found it difficult to buy as much cotton as they would have liked from the CJC at official prices. CJC was able to earn profits by procuring below the black market price and quickly selling to textile mills. Government efforts to alleviate the shortage through releases from stocks in the State Reserve proved ineffectual. On the other hand, when supplies are relatively large, as they are currently, textile mills are reluctant to buy from CJC at fixed official prices, meaning that CJC cannot repay its loans, hindering its further procurement efforts. In effect, the price to farmers may then be reduced by payment delays or surreptitious discounts and quality downgrading. At such times, the government has occasionally announced procurement ceilings, and encouraged farmers to sell cotton elsewhere. During the mid-1980s, record cotton production led to such ceilings, and the government encouraged the development of textile mills organized as township and village enterprises (TVE) to purchase and process this "excess" cotton. This simultaneously helped raise incomes and use excess labor in rural areas, and provided an outlet for cotton that could not be effectively used by the mills that were state-owned enterprises (SOEs). The system of government monopoly on procurement from farmers and a monopoly on allocation to the textile mills was developed to implement planned levels of textile production. CJC owns nearly all the cotton ginning capacity in China, and until the mid-1980s, virtually all of the textile firms were SOEs. In the course of the economy's liberalization, joint-venture enterprise (JVE) mills with foreign investors have been established, and later the TVE mills grew as well. The JVE textile mills are much less tightly integrated into the planning process than the SOEs, and the TVEs are largely outside it. The share of cotton production purchased by the government has fluctuated significantly in recent years, indicating the official monopoly is not complete. Some observers believe part of this gap between reported cotton production and procurement reflects inflated production data. However, it is universally acknowledged that cotton is sold outside of the official, fixed-price procurement channels. There are numerous cotton gins not under control of government agencies officially tasked with ginning (CJC, the Ministry of Agriculture, and the Xinjiang PCC), and the TVEs are advantageously positioned to make direct purchases from local farmers, bypassing procurement. Textile Industry Consumption Garments and other consumer textile goods are essentially freely traded in China, and China has been a major exporter of yarn, fabric, and garments for a number of years. The state-owned mills, which are directly influenced by government planning, play a major role in textile production--particularly at the initial, capital-intensive stage of spinning--but textile production and sales in China are under less government control than raw cotton. Garment production is particularly far removed from the planning process since domestic consumers' purchasing decisions are freely made, and export demand is beyond the control of China's government. TVEs account for 60-80 percent of China's garment production, and, since TVEs are excluded from government planning ( although in many cases they are owned and managed by local governments), the production as well as the consumption side is largely outside of direct state control for garments. However, in 1997, SOEs accounted for 32 percent of the output of China's textile industry, while accounting for 15 percent of all the firms. As their above-average size indicates, SOEs are more involved in the capital-intensive spinning portion of the industry. JVEs are particularly prominent in producing for export, and are also more oriented towards garment production than the state-owned sector. With their foreign participation and export orientation, the JVEs are much less integrated into the plan than the SOEs. Participating in the plan sometimes means that production continues, despite lagging sales and rising inventories, simply to fulfill a planned level of output. SOEs often have access to government-backed loans that have enabled them to remain in business despite rising financial losses. On average, the state-owned textile sector lost a reported 5 yuan for every 100 yuan of sales in 1997. Economy-wide, China's SOEs reported losses rose from 3 billion yuan in 1985 to 83 billion yuan in 1997. These losses would probably be substantially higher if U.S. accounting practices were used to develop the estimates. In principle, SOEs are allocated cotton from CJC at a fixed price. The SMC and the government textile administration determine the appropriate allocation of cotton fiber from gins to spinning mills or the local government Raw Material Company supplying the mills, and the mill pays the supplying CJC. Joint-venture mills also participate in this allocation system, although during periods of import liberalization many JVEs switch solely to imported cotton. In addition to imported cotton, manmade fibers are an important alternative for China's textile industry. Manmade fiber production in China has grown rapidly, and imports are large as well. The government has encouraged increased production of manmade fibers as part of a strategy to maintain a high level of self-sufficiency in food production: Increased availability of alternative fibers helps reduce the need for cotton--which competes with grains for limited agricultural resources. The share of manmade fiber use in total fiber use has reportedly reached about 40 percent in the most industrialized regions of the coastal provinces, and may total 30-35 percent as a national average. Trade Policy Development Imports of raw cotton have accounted for as much as 19 percent of China's domestic mill consumption in recent years, while in the 1980s as much as 19 percent of production was exported. China became the world's largest cotton importer during the mid-1990s as import quotas increased sharply, and the right to purchase for import was extended to new entities in addition to the China National Textiles Import & Export Corporation (Chinatex). Later, as crop production rebounded in China, cotton import policy tightened. First, SOE mills were prevented from importing, while JVE mills were allowed to continue, but only for processing and reexport. Cotton intended for reexport as processed goods was exempt from import tariffs and from the value-added tax (VAT). However, JVEs often served as a conduit to their SOE partners, and imported cotton intended for reexport as processed goods filtered into domestic consumption. In 1998 JVE imports fell sharply in response to tighter government restrictions, and subsequently China initiated its first major export sales of cotton since the early 1990s. Chinatex serves as the primary agent for cotton imports for SOE mills and for most of 1998 was China's sole export agent for raw cotton. It is a specialized foreign trade corporation attached to China's Ministry of Foreign Trade and Economic Cooperation (MOFERT). With its periodic monopoly role in exports and imports, Chinatex is the key actor in implementing the government's cotton trade policy. Trade policy for cotton in China is generally built around trade volume targets. These targets or quotas are developed each year by the State Council, in consultation with the CJC and others, based in part on expected cotton supply and demand. The State Council also determines the distribution of importing and exporting rights, permitting entities in addition to Chinatex to import or export during periods of relative scarcity and surplus, respectively. Textile trade policy has also been used to indirectly affect the balance of fiber supply and demand in China. China's provinces often pursue self-sufficiency vis-a-vis other provinces, and cotton-producing provinces seek to consume domestic cotton before importing from other provinces. As cotton stockpiles rose in China, Xinjiang province found its stockpiles mounting particularly rapidly since it primarily grows cotton for other provinces. In 1997 China announced a basket of policies to reduce Xinjiang's stocks, boost textile exports, and ease financial pressure on the textile industry. The VAT rebate for textile products made with cotton produced in Xinjiang Province was increased to 17 percent, compared with 9 percent for products made with cotton from other provinces. In early 1998, the 9 percent was increased to 11 percent. Trade policy decisions for cotton are made not only with the performance of the cotton sector in mind, but also in the broader contexts of agricultural policy, macroeconomic management, and international economic diplomacy. Cotton exports and imports by China are positively correlated with China's trade in soybeans, rice, and corn. These crops compete for farm area in China, and their production and trade fluctuates with policymakers' concern with food security and rural incomes. Correlations between China's net trade in cotton and trade in soybeans, rice, and corn during 1980-96, were, respectively, 73, 47, and 41 percent. During the mid-1990s, imports of both cotton and grain soared in part due to government efforts to rein in double-digit inflation. Another macroeconomic concern during past years was the level of foreign exchange reserves, a less important issue during the 1990s due to China's investment inflows and large trade surpluses. Finally, as a country with large trade surpluses, China finds its trade policies inevitably come under international scrutiny. For a number of years, it seemed China calibrated its trade policy changes with an eye towards achieving its goal of membership in the World Trade Organization. More recently, the Asian crisis has raised more immediate concerns as China maintains its currency peg against the U.S. dollar--despite Southeast Asia's devaluations--and China's growing trade surpluses have raised concerns about the outlook for various industries in the United States and Europe. Rigidities and Lagged Responses Lead to Problems While most of the problems facing China's cotton sector are self-inflicted, the financial difficulties that started in Southeast Asia last year have exacerbated China's difficulties. This year China's textile exports are lagging in part due to strong competition from textile exporters in South and Southeast Asia. Since China's textile exports account for 25-40 percent of total cotton consumption, these trade difficulties, coupled with the general slowdown in China's economy, have increased enterprise debt levels and aggrevated the longstanding weaknesses in China's SOE sector. The losses of state textile mills stem not only from growing export competition, but because SOE mills must purchase cotton inputs from the CJC largely at state-set prices that are generally much higher than if they were true market-clearing prices. The losses of textile mills lead to delinquent payments to CJC, which has also accumulated an extremely large debt overhang. The need for reform of China's SOEs has a corollary in the need for reform in China's banking sector, which is largely state-owned. The Asian financial crisis has highlighted the role of healthy financial intermediaries in a developing economy, and with an enormously large volume of non-performing loans, many of China's banks have been far from healthy since well before the recent developments in Southeast Asia. Centralized decisionmaking for China's cotton sector can perhaps best be characterized as "lagged decisionmaking." Procurement prices were raised substantially in 1995 in response to the shortages earlier in the 1990s. Increasing procurement prices helped raise production in China, and helped reduce world prices, but the downward price response to a recovery in world and domestic supplies was then delayed. As of late 1998, it is not clear that this price response has occurred; while China's explicit reduction of its cotton procurement price in April 1998 was unprecedented, it was not large. Similarly, delays in changing trade policy to open the sector to imports early in the 1990s exacerbated the difficulties of the fixed-price system, and delays in restraining imports added to the stock build-up through 1997/98 that has kept black market cotton prices in China well below official prices. The Search for Solutions Since 1996, China has pursued a number of options to add limited price and purchasing flexibility, including acceptable bands around official prices, cotton trade fairs to encourage interaction between mills and government agencies holding cotton, increased credit provisions for transactions between CJC and the textile industry, debt write-downs, and finally reduced procurement prices in 1998. While the national procurement price cut was a modest 7 percent, the procurement price in Xinjiang has, at least in principle, been completely freed. Similarly, the sale price to mills has been completely freed, although government edicts also prohibit CJC from selling at a loss, which presumably places a floor on the prices CJC can offer. Reform of the consumption sector also accelerated in early 1998 with a renaissance of China's perennial spindle-reduction program. A large number of obsolete spindles from loss-making SOE mills were reportedly destroyed during 1998 in response to bounties and debt relief. However, a number of retired spindles were reportedly transferred to TVE mills rather than destroyed, and it remains to be seen if the early ardor for spindle destruction can be maintained as unemployment mounts. However, textile mills account for a significant share of the SOE sector, and the combined problems of state-owned industry and banking must be addressed to ensure the long-term health of the economy. There are few if any precedents of a developing country as large as China successfully negotiating a transition to a market economy, and the issues involved are beyond the scope of this paper. Raw cotton imports have been largely curtailed, and China's first large export sale since the early 1990s was consummated early in 1998, although shipments have been sluggish. Export rights have again reportedly been broadened to entities other than Chinatex, such as PCC state farms in Xinjiang. Exporting will be no panacea for China's cotton imbalance since a substantial portion of current stocks were presumably accumulated at prices well above likely revenues from exporting. Exporting cotton to other countries below the price officially paid to farmers could become the focus of economic diplomacy between China and other countries exporting cotton. Inevitably, other cotton exporters would want to be reassured that such exports are implemented in a manner consistent with international norms. There are also domestic considerations that make it difficult to export below prices officially paid to farmers, since doing so would require acknowledging an economic loss somewhere in the marketing system. China's accounting practices mean that indefinitely maintaining long-standing, uncollectible debts at face value may be preferable to retiring the debt at a loss, and concluding an export transaction would appear to require closing the books on the exported cotton. Debts may even be partly fictitious if price discounts were introduced along the way, further complicating the settlement process. Under these circumstances, it is not sufficient for the State Council to simply authorize the export of a given amount. Further enabling decisions are necessary to overcome the complications introduced by a chain of debts linking procurement, industry, and banking. These concerns do not necessarily wholly apply to Xinjiang during 1998/99. Effective procurement prices this year in Xinjiang are still unknown, and could even fluctuate for months. In addition to the price flexibility offered to the province as a whole, the PCC's organization as state farms enables it to delay payments to certain categories of farmers without imposing undue hardship, further increasing its price flexibility. Frequently in China, implicit regulations persist despite publicized policies such as price freedom, so procurement prices in Xinjiang are not necessarily able to decline without limit. But, it is conceivable that some portion of Xinjiang's 1998/99 crop could be purchased from farmers at prices that permit exporting with little or no subsidy. Finally, the issue most central to China's cotton sector is the marketing of farm production. Maintaining continuity potentially exposes the government to large subsidies to farmers. The official procurement price is still attractive to producers, and is too high for profitable sales to textile mills by CJC. Reports from China indicate that discussion is underway to further reduce, and perhaps even free the procurement price, and to possibly limit procurement volume. Limiting government procurement volume implies de jure recognition of the de facto alternatives already functioning alongside the government monopoly. It is unclear if CJC's role would then be to administer a vastly reduced procurement at above-market prices, or vigorously compete with other purchasers for a larger share of the crop. In the absence of any current legal price-discovery institutions like the Zhengzhou, Dalian, and Shanghai markets for some grains and oilseeds, alternatives to government-fixed prices will probably require a period of transition. However, if China successfully reembarks on the experiment last attempted in 1992/93 of legalizing alternatives to the 50-year old government monopoly, China's cotton sector will have taken one of its largest steps towards market-driven decisionmaking. More flexible and transparent changes in China's cotton sector could increase the stability of world markets as well as those in China, a welcome prospect for many cotton producers and consumers in the rest of the world. References State Statistical Bureau, China's Statistical Yearbook, 1997, Beijing, 1998. U.S. Department of Agriculture, China: International Agriculture and Trade Reports, various issues. _________________________, Economic Research Service, unpublished trip reports, 1996 and 1997. _________________________, Foreign Agricultural Service, Beijing Agricultural Office, China Cotton Reports, various issues. Table B-1--Government and Economic Participants in China's Cotton Sector SMC: Supply and Marketing Cooperatives A department of the Bureau of Internal Trade; manages and administers government procurement and distribution of agricultural products from farmers. Through 1997, the SMC interacted with 160 million farm households, about 85 percent of all those in China. CJC: Cotton and Jute Company A department of SMC; responsible for the administration of cotton purchasing, testing, processing, storing and marketing in China. It has 25 provincial companies, 110 prefectural companies, 830 companies at the county level, 15,000 cotton-purchasing stations, and 2,300 cotton gins, with a processing capacity of 6 million tons. PCC: Xinjiang Production and Construction Corps The PCC is a military entity tasked with both defense and economic development in Xinjiang-Uighur Autonomous Region. The Uighurs, a substantial portion of the province's population, are ethnically, culturally, and linguistically distinct from the rest of China. Governing a region so distant and different from China's eastern provinces poses unique difficulties, and the PCC's development has been one response. The PCC's cotton planning, financing, and marketing channels are distinct from those for the rest of Xinjiang. Chinatex: China National Textiles Import and Export Corporation Chinatex is a specialized foreign trade corporation directly attached to the Ministry of Foreign Trade and Economic Cooperation. Chinatex mainly specializes in all kinds of textile materials, garments, cotton and woolen products and other imports and exports approved by the state. Chinatex has 14 subsidiary companies, over 30 overseas enterprises, and over 20 domestic branches in the country's major provinces and cities. Cotton trade is handled by a subsidiary, Chinatex Cotton Import and Export Corporation. SOEs: State-owned enterprises SOEs are firms either nationalized by China's government after 1949 or owned by the government since inception. At one time virtually all of China's industry was state-owned. Profitability at such firms often lags due to overstaffing, a history of politically-guided decisionmaking, and extensive social obligations to the firms' employees and communities. JVEs: Joint-venture enterprises Beginning in 1978, China opened its economy to outside influences, permitting foreign investment. This has largely taken the form of joint ventures with SOEs. Local partners have often proven crucial, given the regulatory hurdles inevitable in China's rapidly evolving economic and legal system. Similarly, foreign partners provide management and technical expertise as well as access to overseas markets and sources of imported inputs for processing. TVEs: Township and village enterprises After rapid growth in the late 1980s and early 1990s, TVE industrial output in 1995 equaled nearly 30 percent of China's GDP. Food, clothing, building materials, and other products that can be produced with relatively low capital inputs are their major products. During the late 1980s, modernizing SOEs regularly transferred their older spindles to TVEs. Currently, competition from TVE textiles may be hurting the profitablity of SOEs. List of Tables Text Tables A. U.S. cotton supply and use, 1996/97-1998/99 B. World and U.S. cotton prices, August 1997 to present C. World cotton supply and use, 1996/97-1998/99 D. ELS cotton supply and use in foreign producing countries, 1996-99 E. Wool supply and disappearance, clean content, 1994-98 F. U.S. imports of raw wool for consumption, clean content, 1994-98 G. U.S. mill consumption of raw wool, clean basis, quarterly, 1994-98 H. World wool supply and disappearance, 1990/91-1998/99 I. Reported prices of raw materials for manmade fibers, 1998. Appendix Tables 1. U.S. cotton supply and use, 1960/61-1998/99 2. U.S. upland cotton supply and use, 1960/61-1998/99 3. U.S. ELS cotton supply and use, 1960/61-1998/99 4. Upland cotton: Planted acreage, by State, 1960/61-1998/99 5. Upland cotton: Harvested acreage, by State, 1960/61-1998/99 6. Upland cotton: Lint yield, by State, 1960/61-1998/99 7. Upland cotton: Production, by State, 1960/61-1998/99 8. ELS cotton: Planted and harvested acreage, by State, 1960/61-1998/99 9. ELS cotton: Production and yield, by State, 1960/61-1998/99 10. U.S. cotton supply and disappearance of all kinds, by months, 1994/95-1997/98 11. Upland cotton farm, spot, and mill prices, 1970/71-1997/98 12. Fiber prices: Landed Group B mill points, cotton prices, and manmade staple fiber prices at f.o.b. producing plants, actual and estimated raw-fiber-equivalent, 1960-98. 13. Index of prices of selected growths and qualities of U.S. cotton c.i.f. Northern Europe, monthly, 1989/90-1998/99 14. Index of prices of selected growths and qualities of U.S. cotton, c.i.f. Northern Europe, annual, 1975/76-1997/98 15. World cotton supply and use, 1965/66-1998/99 16. Foreign cotton supply and use, 1965/66-1998/99 17. Cotton exports, major foreign exporters, 1965/66-1998/99 18. Cotton imports, major importers, 1965/66-1998/99 19. Former Soviet Union cotton supply and use, 1965/66-1998/99 20. Brazil cotton supply and use, 1965/66-1998/99 21. Turkey cotton supply and use, 1965/66-1998/99 22. China cotton supply and use, 1965/66-1998/99 23. India cotton supply and use, 1965/66-1998/99 24. Pakistan cotton supply and use, 1965/66-1998/99 25. U.S. fiber consumption: Total and per capita, by type of fiber, 1990-98 26. Per capita domestic cotton consumption, 1975-97 27. Cotton and manmade staple fibers: Mill consumption on the cotton spinning system, 1960-97 28. U.S. wool supply and use, 1975-98 29. U.S. imports of raw wool for consumption, clean yield, 1970-97 30. U.S. raw wool imports by country of origin, clean yield, 1993-97 31. U.S. mill consumption of raw wool, scoured basis, annual, 1970-97 32. U.S. raw wool exports by country of destination, clean yield, 1995-97 33. U.S. trade in wool tops, 1994-97 34. Shorn wool prices: U.S. farm price, Australian offering prices, and graded territory shorn wool prices, 1978-97 35. U.S. consumption on the woolen system and worsted combing, annual, 1987-97 36. World wool supply and disappearance, 1989/90-1998/99 37. Sheep population, wool production, and exports, major foreign producing countries, 1991/92-1998/99 38. World wool trade by major importing and exporting countries, 1990/91-1997/98 39. Wool sales and government-owned stocks, major foreign exporters, 1988/89-1997/98 40. International wool prices, 1990/91-1998/99 41. U.S. mohair, clean, exports by country of destination, 1991-97 42. Manmade fiber production and capacity, 1997-2000 43. Domestic shipments of manmade fibers by major category, 1996-98 44. World textile fiber production, 1980-97 45. Raw-fiber-equivalent of textile manufactures, 1960-98 46. Raw-cotton-equivalent of U.S. imports of cotton-containing textile manufactures, 1994-98 47. Raw-cotton-equivalent of U.S. exports of cotton-containing textile manufactures, 1994-98 48. Raw-linen-equivalent of U.S. imports of linen-containing textile manufactures, 1994-98 49. Raw-linen-equivalent of U.S. exports of linen-containing textile manufactures, 1994-98 50. Raw-wool-equivalent of U.S. imports of wool-containing textile manufactures, 1994-98 51. Raw-wool-equivalent of U.S. exports of wool-containing textile manufactures, 1994-98 52. Raw-silk-equivalent of U.S. imports of silk-containing textile manufactures, 1994-98 53. Raw-silk-equivalent of U.S. exports of silk-containing textile manufactures, 1994-98 54. Raw-manmade-equivalent of U.S. imports of manmade-containing textile manufactures, 1994-98 55. Raw-manmade-equivalent of U.S. exports of manmade-containing textile manufactures, 1994-98 END_OF_FILE