COTTON AND WOOL YEARBOOK November 27, 2001 November 2001, ERS-CWS-2001 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 will be available from the USDA order desk in about 3-4 weeks. Call toll-free, 1-800-999-6779 and ask for stock # ERS- CWS-2001, $21. ERS-ERS-NASS accepts MasterCard and Visa. --------------------------------------------------------------------------- Contents Summary Cotton Situation and Outlook World Cotton Situation and Outlook Situation and Outlook for Other Fibers Special Articles: Regional Shifts in China's Cotton Production and Use The Agreement on Textiles and Clothing: Impact on U.S. Cotton Report Coordinator Leslie Meyer (202) 694-5307 Fax (202) 694-5823 E-mail: LMEYER@ers.usda.gov Other Contributors Stephen MacDonald (202) 694-5305 Robert Skinner (202) 694-5313 Mae Dean Johnson (202) 694-5299 (Statistics) Editor Lou King Layout and Design Wynnice Pointer-Napper Summary U.S. cotton production in 2001 is forecast at a record 20.2 million bales, 3 million bales or 17 percent above last season's cotton crop. This season's output is based on larger area and a higher national yield. U.S. cotton planted area jumped almost 700,000 acres from 2000 to 16.2 million this season, the second largest area planted to cotton in four decades. While drought conditions this season raised abandonment and limited production in parts of the Southwest, conditions in other parts of the Cotton Belt were above average. Nationally, abandonment reached 13 percent this year, compared with 16 percent in 2000. Harvested cotton area is estimated at 14.1 million acres, 1 million above last season and the highest since 1995. The national yield is projected at 685 pounds per harvested acre, well above last season and the highest in 5 years. U.S. cotton exports in 2001/02 (August/July marketing year) are projected to jump dramatically, rising nearly 40 percent from last season to 9.4 million bales. Large U.S. supplies of exportable cotton coupled with the strong sales and shipments to date are expected to push exports to match the recent high achieved in 1994/95. As of early November, U.S. cotton export commitments totaled 7.7 million bales, or 82 percent of the shipment target. With U.S. exports expanding and prospects for world trade rising modestly in 2001/02, the U.S. share of global trade is expected to reach 33 percent, well above last season's 26 percent and near the 1994/95 level. In contrast, U.S. cotton mill consumption is forecast to reach only 8.1 million bales in 2001/02, 9 percent or nearly 800,000 bales below last season. The continued growth of U.S. cotton textile and apparel imports, along with the slowdown in the U.S. economy, has kept tremendous pressure on the U.S. spinning industry and has forced major restructuring in the U.S. textile and apparel industry over the last several years. U.S. cotton stocks at the beginning of 2001/02 were estimated at 6 million bales. With the record crop forecast this season, total U.S. cotton supply in 2001/02 is projected at 26.2 million bales, or 24 percent above a year ago. Meanwhile, U.S. cotton demand is forecast to expand to 17.5 million bales, nearly 2 million above 2000/01. However, based on these supply and demand estimates, U.S. ending stocks for 2001/02 are estimated to climb significantly to 8.7 million bales, the highest since 1985/86. As a result, the U.S. stocks-to-use ratio is expected to climb to 50 percent in 2001/02. World cotton consumption in 2001/02 is forecast at 91.6 million bales, about 200,000 bales below a year earlier as world economic growth drops to its slowest in nearly a decade. However, foreign mill use is expected to rise to its third consecutive record-high while the United States is forecast to undergo the largest decline in mill use of any country. The largest gains are expected in Turkey, Uzbekistan, Pakistan, and Thailand. In contrast with stagnant consumption, world cotton production in 2001/02 is forecast at a record 96.9 million bales, 8.5 million bales above last season. Production is estimated higher across much of the Northern Hemisphere, with by far the largest increases expected in China (3.2 million bales higher) and the United States (3 million). Increased plantings and improved weather are also expected to lead to substantially larger crops in India, Central Asia, and West Africa's Franc Zone. Foreign production is forecast to be its second highest ever, 76.7 million bales, with only a few countries, particularly Brazil and Australia, expected to harvest lower crops than the year before. World cotton exports are expected to increase 7 percent in 2001/02, to 28.1 million bales, their highest since 1994/95. However, foreign exports are expected to fall 5 percent to 18.7 million bales, their lowest since 1983/84. Australia's exports are expected to be almost 800,000 bales lower than the year before in 2001/02, and smaller declines are expected from Central Asia and South America. With rising world production and stagnant consumption, stocks are projected to rise this season to 44.4 million bales, their highest in 3 years. The 5.5-million-bale stock increase is expected to be split evenly between the United States and the rest of the world, with India accounting for the largest increase outside the United States. Little change is foreseen in China's ending stocks in 2001/02, in sharp contrast to the last 2 years when stocks there fell by 6.2 million bales in 1999/00 and by 3.4 million in 2000/01. U.S. Cotton Situation and Outlook U.S. Cotton Review, 2000/01 The 2000/01 (August/July) marketing year began with U.S. cotton stocks estimated at 3.9 million bales, similar to the previous three seasons. The 2000 cotton crop was the fifth planted under the Federal Agriculture Improvement and Reform (FAIR) Act, which provides total planting flexibility to producers. In 2000, this flexibility facilitated the planting of additional acreage to cotton as relative net returns at planting time, the marketing loan program, and the insurance program seemed to favor cotton over competing crops. The U.S. cotton area in 2000 increased to 15.5 million acres, 4 percent above the 14.9 million planted in 1999. However, the national abandonment rate jumped from near 10 percent in 1999 to 16 percent in 2000, the result of drought conditions in Texas where 75 percent of the lost acreage occurred. Consequently, total cotton harvested area was only 13.1 million acres, slightly below that of the previous year. With fewer harvested acres, the U.S. yield averaged 632 pounds per harvested acre, the highest in 3 years. With the higher yield more than offsetting the lower harvested area, U.S. 2000 cotton production totaled nearly 17.2 million bales, slightly higher than the crop produced in 1999. In 2000/01, total U.S. cotton demand slipped to 15.6 million bales, a 1.3-million-bale reduction. While U.S. raw cotton exports remained stable, U.S. mill use continued a third year of decline. In 2000/01 U.S. cotton exports reached 6.8 million bales, slightly above the previous 5-year average. Despite an increase in foreign cotton production, higher foreign mill use helped sustain U.S. exports. With world trade contracting to 26.4 million bales in 2000/01, the U.S. share of global trade increased one percentage point to nearly 26 percent. U.S. cotton mill consumption in 2000/01 dropped to 8.9 million bales from 10.2 million the year before, a 13- percent reduction. In addition, the 2000/01 mill use was the first in a decade to fall below 9 million bales. The decline has resulted from competition from cotton textile and apparel imports that has provided U.S. consumers with relatively inexpensive products. The situation was exacerbated by the strength of the U.S. dollar, providing a home for foreign goods in the United States while hindering U.S. cotton textile exports to the world. During the 2000/01 marketing year, U.S. textile and apparel manufacturers were forced to reduce their output dramatically, resulting in a record number of industry employees affected by plant layoffs and closings. With U.S. cotton supply at 21.1 million bales in 2000/01 and total demand at only 15.6 million, stocks at the end of last season jumped significantly. Ending stocks were placed at 6 million bales, a gain of more than 2 million bales during the season. As a result, a stocks-to-use ratio of more than 38 percent was the highest since 1988/89. Despite a reduction in demand and an increase in stocks, the average upland price received by producers improved from 45 cents per pound in 1999/2000 to 50 cents. Although farm prices had improved, upland producer returns in 2000/01 were lower as the cotton program provided a loan deficiency payment averaging only 4 cents per pound, compared with 20 cents in 1999/2000. Similarly, the extra-long staple (ELS) price rebounded from 1999/2000's 85 cents per pound to average $1.00 in 2000/01. Supply Outlook for 2001/02 As planting time for the 2001 crop approached, U.S. cotton prices had fallen but alternative crops--like corn and soybeans--moved lower as well and were not attractive enough to pull area away from cotton. Also, the cotton marketing loan program and the insurance program for cotton provided incentives for U.S. producers to plant additional acreage to cotton in 2001. Favorable springtime weather allowed producers to plant 16.2 million acres of cotton this season, 4 percent higher than in 2000 and the second largest in nearly four decades (table A). Upland cotton acreage this season is estimated at 16 million acres, compared with 15.3 million in 2000. Similarly, ELS cotton area climbed to 235,000 acres in 2001, 65,000 acres above a year earlier. The rise in ELS acreage came in California, where some area moved out of upland production as prospects for ELS looked more favorable in 2001. With the increase in cotton planted area this season, harvested area rose about 1 million acres from 2000 to 14.1 million. Although the abandonment rate based on November estimates of 13 percent--2 million acres--was below last season's 16 percent, the loss remains above average. Despite this, U.S. cotton harvested area is the second highest since 1963/64. After a favorable start to the growing season and cotton crop conditions more favorable than last year, this summer's drought across the Southwest region lowered overall conditions during the growing season and limited the potential of the cotton crop there. Based on November 1st conditions, the 2001 U.S. cotton crop was estimated at nearly 20.2 million bales, 3 million bales or 17 percent above last season and the largest crop on record. Compared with 2000, this season's larger crop was aided not only by increased harvested area but also a higher yield. The U.S. average cotton yield is currently projected at 685 pounds per harvested acre, 53 pounds above last season and the highest in 5 years (fig. 1). Upland production is projected at a record 19.6 million bales in 2001, with an average yield forecast at 675 pounds per harvested acre. On a regional basis, forecast production is larger in three of the four Cotton Belt regions this season. In the Delta, upland output is projected at 6.8 million bales, the highest since 1994/95's record of 6.9 million. In the Southeast, production is forecast at 5.3 million bales, the first 5-million-plus season there since 1937/38. Drought conditions in the Southwest have limited upland output expectations to near the 5-year average of 4.6 million bales. In the West, the upland crop is estimated at 2.8 million bales this season. Despite excellent growing conditions in 2001, higher yields could not offset the acreage shift from upland to ELS and as a result, upland production declined from a year ago. ELS production is forecast at 618,000 bales in 2001, nearly 60 percent above the 2000 crop. Larger area and a record yield provided the jump in output this season as ELS cotton provided a better alternative to some producers in the West region. The national ELS yield is projected at 1,233 pounds per harvested acre. California continues to dominate ELS production, accounting for about 90 percent of the area and output. U.S. cotton stocks on August 1, 2001 were estimated at 6 million bales, well above the 3.9 million seen during the previous three seasons and the largest beginning stocks since 1989/90. Like 2000/01, only a small volume of raw cotton imports is expected to enter the United States this season, as the record production and decline in mill use limits the need for foreign supplies. As a result, total U.S. cotton supply in 2001/02 is projected at 26.2 million bales, 5.1 million above last season and the largest since 1966/67. Demand Outlook for 2001/02 The U.S. cotton demand outlook for 2001/02 is projected to improve modestly from a year earlier. This season's total use is forecast at 17.5 million bales, nearly 12 percent above 2000/01, with a jump in exports offsetting a decline in mill use. U.S. cotton exports are forecast to rise nearly 40 percent from last season to 9.4 million bales. Large U.S. supplies of exportable cotton coupled with the strong sales to date are expected to push shipments to match the recent high achieved in 1994/95. And, despite stagnant global demand, the U.S. program has kept U.S. cotton competitive around the world. While world cotton prices decreased significantly in 2001--falling about 30 cents per pound between December 2000 and October 2001--U.S. cotton prices declined a similar amount (table B). This movement allowed U.S. sales to continue to build this fall. Upland cotton shipments are projected to surpass 8.9 million bales, while ELS exports are expected to reach 460,000 bales, both near their respective highs during the mid-1990s. With U.S. exports expanding dramatically and prospects for world trade rising modestly in 2001/02, the U.S. share of global trade is expected to reach 33 percent, well above last season's 26 percent and near the 1994/95 level (fig. 2). Based on U.S. Export Sales data through early November, U.S. cotton export commitments (shipments plus outstanding sales) totaled 7.7 million 480-pound bales, significantly above the 4.2 million reported a year earlier. Upland cotton shipments to date have reached 2.5 million bales, compared with only 1.2 million in 2000/01. Likewise, upland outstanding sales are reported at 4.9 million bales, or more than 2 million above the comparable period last season. However, ELS shipments and outstanding sales are a combined 30 percent below year-earlier levels. As of early November, ELS commitments were approximately 256,000 bales, or about 56 percent of the current export forecast. U.S. cotton mill consumption, on the other hand, is forecast to reach only 8.1 million bales in 2001/02, 9 percent or nearly 800,000 bales below last season. The current projection is more than 3 million bales below the recent high seen in 1997/98 as growth of U.S. cotton textile and apparel imports has continued to pressure the U.S. spinning industry along with the slowdown in the U.S. economy. As a result, major restructuring in the U.S. textile and apparel industry has occurred over the last several years, forcing many in the industry to limit output or close operations. The level and origin of these imports will help determine the amount of raw cotton consumed by U.S. mills. Upland mill use is forecast at 8 million bales in 2001/02, while ELS consumption is projected at 110,000 bales, both below a year ago. With mill use declining, its share of total demand is also reduced and stresses the importance of U.S. exports (fig. 3). Based on the first 3 months of data from the Department of Commerce, the seasonally adjusted annual rate of cotton consumption averaged only 7.8 million bales. Actual cotton mill use for August through October 2001 totaled 2.10 million bales, compared with 2.57 million a year earlier. Like cotton, manmade fiber mill use has fallen, but at a faster rate during the first three months of 2001/02. Consequently, cotton's share of fiber consumption on the cotton system has risen from a year earlier to 81 percent, compared with a 2000/01 average of 78.5 percent. Based on these projections of U.S. cotton supply and demand, ending stocks for the 2001/02 marketing year are estimated to climb to 8.7 million bales, 45 percent above the beginning level and the highest stocks since 1985/86 when 9.3 million bales were held in the United States (fig. 4). And, despite the healthy demand increase expected this season, the record U.S. crop has more than offset the rise in demand and has pushed the U.S. stocks-to-use ratio from 38 percent in 2000/01 to 50 percent. Upland stocks are estimated at 8.5 million bales, while ELS stocks are projected to rise to 189,000 bales. U.S. Textile Trade and Domestic Consumption The overall volume of U.S. textile trade in calendar 2001 is expected to decline slightly (based on preliminary data) for the first time in over a decade. Although the liberalization of world textile and apparel trade continues, the sluggish U.S. economy in 2001 has kept both imports and exports from expanding significantly this year. However, the U.S. textile trade deficit will increase again in 2001, as imports remain stable and exports decline. In calendar 2000, the textile trade deficit for all fibers reached a record of 8.3 billion (raw-fiber-equivalent) pounds, 12 percent above 1999. U.S. textile and apparel imports during the first 9 months of 2001 were nearly identical to the comparable period in 2000, at 10.3 billion pounds. However, textile exports through September 2001 have slipped below last year's 4 billion pounds to 3.8 billion as the strength of the U.S. dollar makes exports more expensive to foreign customers. As a result, the textile trade deficit for all fibers during the first 9 months of 2001 approached 6.6 billion pounds, about 4 percent above the same period in 2000. For cotton, U.S. textile trade has followed a similar pattern so far in 2001. Cotton textile imports totaled 5.7 billion pounds through September 2001, equivalent to a year earlier. Meanwhile, cotton textile exports have slipped 9 percent from 2000 to approximately 1.7 billion pounds during the January through September period. For calendar 2001, cotton textile imports are expected to rise for the 13th consecutive year, while exports may fall for the first time in 17 years. Imports for the year will likely approach the equivalent of 16 million bales (7.6 billion pounds) of raw cotton, about 1 percent higher than in 2000. On the other hand, U.S. cotton textile exports are likely to slip below the 5-million-bale level (2.2 billion pounds), a decline of 8 percent from a year ago. With cotton textile exports decreasing and imports steady, the U.S. cotton textile trade deficit is expected to rise again in 2001 to the equivalent of 11.3 million bales of raw cotton (fig. 5). Meanwhile, imported textile and apparel products continue to displace U.S. fiber mill use. In 2001, total U.S. fiber mill use is expected to fall to approximately 15 billion pounds, the fourth consecutive annual decline and perhaps the lowest since 1992. While total U.S. manmade fiber mill use is expected to fall for the second year in a row in 2001, cotton mill use will decline for the fourth consecutive year to about 4 billion pounds, the lowest in a dozen years. With a reduction in cotton-rich apparel output in the United States, cotton's share of U.S. fiber mill use continues to erode and will likely decline to about 26.5 percent in 2001 while manmade fiber share climbs above 70 percent. In addition to U.S. mill use, 2001 total domestic consumption (mill use plus net textile trade) of cotton is expected to fall for the first time in 5 years as the slowdown in the U.S. economy persists. Based on data for the first 9 months of 2001, domestic consumption totaled 7.1 billion pounds, 6 percent below the comparable period in 2000. By year's end, cotton domestic consumption is likely to approach 9.4 billion pounds, the lowest since 1998 (fig. 6). Likewise, the U.S. per capita cotton consumption could fall about 2 percentage points from 2000's 35.8 pounds. In addition, only about 40-45 percent of this total is being produced in U.S. mills, compared with 65 percent just 5 years ago. World Cotton Situation and Outlook Slowing World Economy, Favorable Weather, and Policies Hit World Price in 2001 A slowing world economy cut into cotton consumption during 2000/01, and also reduced prospects for consumption in 2001/02 just as cotton output around the world jumped in response to last year's relative prices and this year's improved weather (table C). China's long-standing efforts to reform its cotton sector, and its imminent accession to the WTO, resulted in subsidized auctions from government stockpiles during 2000/01. As the government auctions released more than 6 million bales of cotton to China's mills and trading companies, the need for imports during 2000/01 diminished significantly. With the world's consumption falling, production rising, and yet another postponement of significant Chinese imports, the A-Index fell from a 64-cents/lb average during January 2001 to 37 cents/lb in October. Agricultural commodity prices were low in general as the 2001 U.S. harvest drew to a close, but cotton turned particularly weak compared with other major U.S. field crops. Compared with their average during the first half of the 1990s, U.S. Gulf port prices for wheat, corn, and soybeans during August-October of 2001 were down 13-19 percent, about the same as the 2000/01 level. The A-Index, however, during the first 3 months of marketing year 2001/02, averaged 45 percent below its 1990-94 average. During marketing year 2000/01, the A-Index was only 23 percent below average. While the shift in relative prices against cotton suggests that in 2002/03 resources might shift toward producing other crops, it remains to be seen if an improving macroeconomic scenario will allow coinciding consumption gains to complete cotton's recovery. Clothing's income elasticity is greater than food's, indirectly making cotton more sensitive to economic volatility. World Consumption Stagnates During the 1990s, world cotton consumption grew 0.8 percent annually on average, down from 3.3 percent during the 1980s. In 2000/01, consumption shrank slightly, by one-tenth of a percent, and is forecast to decline two-tenths of a percent in 2001/02. Cotton consumption is affected by changes in consumer taste and government policy as well as in world economic growth, but the recent economic slowdown has affected cotton significantly. According to the International Monetary Fund (IMF), world economic growth averaged 3.4 percent annually during the 1980s and 3.0 percent during the 1990s. However, cotton consumption rose 3 times as quickly during the 1980s than during the 1990s. World consumption during the 1980s was boosted by changes in taste and the shift of U.S. agricultural programs away from market price support. During the early 1990s, world consumption was depressed by the collapse of Russia's textile industry, an industry that relied on cotton to a far greater extent than the rest of the world. By the mid-1990s, world cotton consumption had largely worked through these shifts, and embarked on a period of faster, albeit more volatile, average growth driven by an expanding world economy. The economic slowdown in 2001 followed a 7-year period of relatively strong economic growth. Excluding 1998, when the Asian financial crisis brought world Gross Domestic Product (GDP) growth down to 2.8 percent, GDP growth averaged 4 percent during 1994-2000. During the decade before 1994, the world economy expanded only 3.3 percent annually on average. However, for both 2001 and 2002, the IMF forecasts world GDP growth at 2.4 percent per year, and most other forecasts suggest a prospective slowdown at least as great. The geographic distribution of the current slowdown has also been unfavorable for sustaining cotton consumption. During the 1990s, the economic strength of the United States relative to the rest of the world was its highest in perhaps a generation. The analogy of the U.S. economy to a 'locomotive,' pulling the rest of the world along, is well known with respect to overall economic growth, but it is particularly apt for cotton consumption. About 80 percent of the increase in world cotton consumption between 1995 and 1999 is attributable to increased purchases by U.S. consumers. In part, this represented a long-term trend where a long-standing consumer and technical promotion program unique to the United States has resulted in a growing preference for cotton. But, in part it was also fueled by an unusually strong U.S. economy, which began slowing in 2001. Mill Use Down Most in U.S. in 2000/01, Up Most in China In 2000/01, U.S. end-use of cotton declined for the first time since 1995/96 as U.S. GDP growth dropped from 4.1 percent in 2000 to 1.1 percent in 2001. U.S. end-use of cotton fell 3 percent to 19.9 million bales, a 600,000-bale decline. U.S. mills bore the brunt of this contraction, and U.S. mill use fell more than any other country in the world, 1.4 million bales. U.S. net imports of cotton products continued to rise in 2000/01, but only by 600,000 bales (fiber-equivalent), down substantially from the 1.2-1.5 million-bale gains of the preceding 3 years. The European Union (EU) and Japan experienced similar shifts in GDP growth, with Japan actually contracting by 1.2 percent in 2001. These slowdowns reduced the opportunities for textile exporters in developing countries, and helped cut mill use in Japan and in some EU members as well. After the United States, the country with the next largest year-to-year decline in cotton consumption was Turkey, with a 600,000-bale drop to 5 million bales. In addition to difficulties exporting to a slowing EU market, Turkey experienced a financial crisis during 2001, and its GDP is estimated to have declined more than 8 percent. Mexico has a trading relationship with the United States comparable to that between Turkey and the EU, and Mexico's was the next largest decline in cotton use. Mexico's cotton use fell 300,000 bales from the year before in 2000/01, its first decline since 1992/93. Between 1992/93 and 1999/2000, Mexico's integration with the rest of North America drove mill use there 226 percent higher, a 1.7- million-bale increase. Textile Exporters' Cotton Consumption Jumps in 2000/01 Not every country suffered a decline in mill use in 2000/01, with China's cotton consumption estimated 1.3 million bales higher than the year before, at 23.5 million bales. This occurred even though China's textile and apparel exports to the United States rose only slightly, from 865,000 bales (fiber-equivalent) in 1999/2000 to 906,000 bales in 2000/01. Only about 10 percent of China's total textile and apparel exports go directly to the United States. While a portion of what Hong Kong ships to the United States is derived from cotton spun in China, Hong Kong's exports to the United States were essentially unchanged from the year before at about 750,000 bales (fiber-equivalent). China's increase in cotton consumption reflected improved sales to Japan and other destinations, increased purchases by consumers in China, and increased inventories of yarn and fabric. Pakistan achieved the next largest increase in cotton consumption in 2000/01, up 450,000 bales to 8.1 million bales. Investment in Pakistan's textile industry was revitalized when cotton yields and output rebounded there in 1999/2000. After strong growth in cotton consumption between 1985 and 1995, Pakistan's cotton textile industry stagnated as what had been a rapidly soaring level of cotton output was hampered by disease and pest problems. Pakistan's switch to a policy of relatively free exports since 1995 probably also necessitated a period of adjustment for the industry, which had relied on export restrictions to indirectly subsidize its cotton purchases. By 2000/01, Pakistan had increased its consumption of cotton by 1 million bales over 2 years, and continued increases in its imports of spinning equipment suggested further gains might be ahead. Indonesia's estimated increase in cotton consumption in 2000/01, at 400,000 bales, was comparable in size to Pakistan's, but with only 1 year of rebounding consumption there the trend is less clear. Export data from the EU, Japan, Taiwan, South Korea, Switzerland, and the United States indicate that Indonesia's imports of spinning machinery actually fell about 25 percent in 1999, but rose about 85 percent in 2000. Preliminary data for 2001 suggests they continued rising after 2000. Therefore, even in the absence of reliable government statistics from Indonesia, there is evidence supporting the estimate that mill use increased from 2 million bales to 2.4 million in 2000/01. This was Indonesia's largest annual increase in percentage terms since 1988/89, and marked the first year that Indonesia's consumption has broken out of the 1.8-2.1 million-bale range in which it has languished since 1992/93. Brazil's consumption also broke through a ceiling of sorts recently, rising from 3.9 million bales in 1998/99 to 4.1 million in 1999/2000, and then rising another 250,000 bales in 2000/01 to reach 4.4 million bales. Somewhat like Indonesia, Brazil's consumption fluctuated between 3.4 million and 4 million bales during 1992-98, and was in part boosted by a significant currency depreciation that revitalized the competitiveness of its textile industry. However, Brazil avoided the political volatility that has frequently affected Indonesia since the onset of the Asian financial crisis, enabling it to capitalize on its improved competitiveness even faster than Indonesia, even though its currency began depreciating about 1 year later. Finally, Uzbekistan matched Brazil's 250,000-bale increase in cotton consumption in 2000/01, rising to 1.1 million bales. Uzbekistan's government has made a priority of building a textile industry. Pervasive government control of cotton marketing, exports, and access to foreign exchange have enabled Uzbekistan to boost export-oriented textile output despite what is widely considered to be a highly overvalued official exchange rate. Uzbekistan has been liberalizing its regulations requiring the surrender of foreign exchange at the overvalued official and commercial bank rates, and has stated its intention to achieve currency convertibility in the relatively near future. For example, according to the IMF, as of January 1, 1999 the percent of export proceeds that must be surrendered at the commercial bank rate was reduced from 50 percent to 30 percent, and the government pursued a policy of more aggressive official devaluation during 2001. This suggests the profitability of exporting textiles may have been improving recently, and investors may be counting on future improvements as liberalization proceeds. With investment from a wide variety of sources including Korea, Japan, Russia, and the United States, Uzbekistan's consumption was finally able to surpass the 750,000 to 950,000 bale range for the first time in more than a decade. World Consumption Remains Weak in 2001/02 At 91.6 million bales, 2001/02 world consumption is forecast 155,000 bales lower than the year before. A forecast of essentially unchanged consumption might be considered low given the combined effects of lower prices and stable expected macroeconomic outlook. While the IMF's 2002 forecast (released November 15) of world GDP growth of 2.4 percent is consistent with USDA's 2001/02 world consumption forecast, the significant decline in cotton prices in the last year would suggest cotton consumption could grow even with relatively slow world economic growth. Not only are cotton prices low with respect to recent history, they are also low with respect to polyester prices. The average of the international polyester quotes reported in Cotton Outlook, weighted by chemical fiber output in each of the countries quoted, actually rose 4 percent between January and October 2001, while the A-Index fell 42 percent. The ratio between these prices had fallen to a level last seen briefly in mid-1999/2000 and last seen for any extended period of time during 1991/92-1992/93 (fig. 7). However, several factors suggest that world cotton consumption could decline slightly rather than increase in 2001/02. One is that although most macroeconomic forecasters are expecting steady-to-improving world GDP growth in calendar 2002 compared with 2001, the first half of 2002 is expected to be less robust than the second half. While one can generally capture the impact of GDP growth on a given marketing year by pairing it with a calendar year estimate for GDP corresponding to the latter part of the marketing year (e.g. pairing marketing year 2001/ 2001/02 and calendar year 2002), the match is not always perfect. In 2001/02, most of the marketing year will be over before the foreseen improvement in calendar year economic activity raises consumption, so that marketing year 2001/02 will be a period when the global economy is weaker than during either calendar 2001 or 2002. Another factor is that increased concerns with terrorism since early in fall 2001 have increased the real or perceived costs of international trade. Since a substantial portion of clothing consumed around the world is traded across international borders at some stage, it is possible that cotton, clothing, and other traded goods might not share in any economic rebound as quickly as they have in the past. Finally, the world's largest cotton consumer, China, built inventories of yarn and fabric during 2000/01, suggesting that increased demand for textiles in China could be met without increasing cotton consumption. At 23.5 million bales, China's cotton consumption is forecast to remain at a record high, and more than 4 million bales above its recent 1998/99 low. China has accounted for a substantial portion of the increase in world cotton consumption during the last 2 years and an outlook for unchanged consumption in China reduces the likelihood world consumption can increase. U.S. Consumption Decline Again World's Largest While it is difficult early in the season to foresee exactly which countries will be in a position to supply more or less of the world's demand for cotton textile products in 2001/02, it seems likely that U.S. mill use of cotton will again decline more than in any other country. U.S. end-use is unlikely to resume significant growth during 2001/02, and net imports are likely to grow for the sixth consecutive year, albeit at a reduced pace. U.S. mill use is forecast at 8.1 million bales, about 800,000 bales lower than the year before. Brazil's mill use is expected to drop 150,000 bales as electricity rationing due to reduced hydroelectric output offsets the impact of continued currency depreciation. Mexico's consumption is expected to drop 100,000 bales due to continued economic sluggishness in the United States, its largest market. Similarly, India's consumption is expected to drop 100,000 bales due to the impact of a slowing world economy and reduced textile exports. On the plus side, Turkey's consumption is expected to rebound--climbing 400,000 bales to 5.5 million--as it emerges from its recent period of financial instability with a more competitive exchange rate. Another 200,000-bale increase is foreseen for Uzbekistan's consumption as investment continues to grow in its textile industry, and a 150,000- bale increase is similarly foreseen for Pakistan, despite the recent disruption to the region's shipping. Smaller increases are foreseen for Thailand and Bangladesh as well. World Crop Grows in 2000/01 and 2001/02 In a prelude to 2001/02's surge, world production rose 1 million bales in 2000/01, to 88.4 million bales. Dominating the production changes of the year, China's output rose 2.7 million bales as cotton area rose for the first time since 1994/95. Brazil's crop rose 1 million bales as the new Mato Grosso frontier continued the phenomenal growth, leaving Brazil's 2000/01 output at 4.1 million bales, compared with only 1.3 million in 1996/97. Partly offsetting these increases were losses in India (1.3 million bales) Uzbekistan (780,000), Mali (420,000), Pakistan (400,000), and Mexico (306,000), where various combinations of low prices and poor weather resulted in lower yields and/or area. While the 2000/01world crop surpassed output in the previous 2 years, it was not particularly large. Output in 2000/01 was 1 million bales below the average of the preceding 5 years and was smaller than global output peaks reached as far back as 1984/85. On the other hand, world consumption in 2000/01 was at a near-record high, 4.1 million bales above the average of the preceding 5 years, and exceeding world production by 3.4 million bales. With world consumption exceeding production for the second consecutive year, cotton prices improved. The A-Index's average for the first half of the marketing year rose 32 percent from the year before. During the same time, U.S. Gulf Port prices for wheat rose 17 percent, but soybean prices rose only 2 percent and corn prices were unchanged. Perhaps more importantly for world cotton production, a microcosm of these developments also occurred within China during 2000/01. Almost half of the global surge in consumption during 1999/2000 occurred in China, and during 2000/01 China's consumption continued rising while the rest of the world's fell. China's government suspended guaranteed cotton procurement beginning with the 1999/2000 crop, and although area devoted to cotton in 2000/01 partly rebounded from the 37-year low reached during the first year of this reform, production still fell short of consumption by 3.5 million bales. With the government reluctant to allow significant imports while it struggled with huge stockpiles of cotton, and with government policy discouraging farmers from producing grain, prospective returns from cotton appeared more favorable at planting than during the year before in 2001/02. This was particularly true in eastern China, where the adoption of Bt-cotton resulted in substantially lower production costs. China's cotton area soared 800,000 hectares in 2001/02, its largest annual increase since 1991/92. While forecasting China's yields and output has proven to be a humbling exercise in many years, weather analysis and credible press reports suggest that yields in most provinces should be comparable to the those of the year before. Given that most of the increased plantings probably occurred outside of Xinjiang, and that yields in those provinces are historically lower than Xinjiang's, China's national average yield is forecast slightly lower than the year before, despite the assumption that yields look as good as last year on a province-by-province basis. With a 20-percent increase in planted area, and a slightly lower average yield, China's 2001/02 cotton crop is forecast 3.2 million bales higher than during the year before, and exactly equal to consumption at 23.5 million bales. Weather Improves for 2001/02 Crop As in China, India's cotton producers also received higher prices than the year before as they harvested their 2000/01 crop, and this, combined with large supplies of Indian rice, led to a 618,000-hectare increase in cotton area. With the return this year of a favorable monsoon to Gujarat (India's largest cotton-producing state), India's yields are expected to be their highest since 1996/97, and India's output is expected to increase 1.3 million bales to 12.2 million. West of India, a prolonged drought has entered its third year, but, ironically, Pakistan's output is expected to increase as planted area shifts from rice to cotton, a less water- intensive crop. An increase of only 100,000 bales from the year before is foreseen in Pakistan's cotton crop, to 8.3 million bales. In Central Asia, yields appear to have increased in 2001/02 despite the continued drought. Uzbekistan's westernmost districts suffered from poor irrigation supplies last year, so area shifted closer to irrigation sources in 2001/02, and output is forecast 300,000 bales higher at 4.7 million bales. Larger crops have also been realized in Kazakhstan, Tajikistan, and Turkmenistan. Cotton production in Central Asia has stabilized since 1996, following a 50-percent reduction over the preceding 8 years. Area has actually trended upwards in the region despite the steady decline in world prices as state monopolies determine producer payments independent of world events. West Africa's Franc Zone has also seen increasing area since the mid-1990s, but largely because of a 1994 exchange rate correction. In 2001/02, area is estimated up 19 percent from the year before and record output is expected. Last year saw one of the sharpest-ever declines in the region's output, in part due to poor weather, and in part due to a strike by producers in Mali, the largest Franc Zone producer. World prices rose slightly last year, and producers and marketing boards in the region pursued increased output. Ironically, Mali's producers refused to plant cotton in many cases in 2000/01 due to unrenumerative prices, but returned in force to the crop in 2001/02. Mali's area is estimated up 89 percent from the year before, and with favorable rains West Africa's largest cotton producer is expected to harvest a crop of 620,000 bales, or 129 percent higher than during the year before. Overall, Africa's Franc Zone is expected to produce 1.2 million bales more cotton in 2001/02 than during the year before, and the region's exports are expected to overtake Uzbekistan's for the first time ever. Lower production is expected from a few producers in 2001/02, largely in Southern Hemisphere countries responding to the recent weakness in world prices. Brazil's crop is expected to be 800,000 bales lower as Mato Grosso's area expansion is devoted entirely to soybeans, and some area previously planted to cotton is also diverted to soybeans. Lower plantings are also expected in Australia, where the crop is expected to fall 500,000 bales, Argentina (down 185,000 bales), and Paraguay (down 50,000 bales). Iran reportedly also planted less cotton and had output further reduced by drought, resulting in a 235,000-bale decline from the year before, and output is also estimated lower in Greece (down 135,000 bales), Sudan (90,000), and Syria (75,000). World Ending Stocks Fall in 2000/01 as China Sells Government Stocks Led by a 3.4-million-bale decline in China, world ending stocks declined for the second consecutive year in 2000/01, to 38.9 million bales, equaling 42 percent of world consumption. Ending stocks also declined by 1 million bales in India, and by 300,000 bales in Australia. On average, non-U.S. ending stocks fell compared with the year before, although increases of between 100,000 and 200,000 bales were reported for Indonesia, Turkey, and Brazil. U.S. ending stocks rose 2.1 million bales as U.S. mill use plunged and exports only began to pick up late in the season. With China's stocks falling and those in the U.S. rising, China ended the year with its lowest share of world ending stocks since 1994/95, and the United States ended with its largest share since 1988. Between 1975/76 and 1994/95, world ending stocks as a share of consumption generally ranged between 37 and 42 percent (the highest and lowest 5-year averages during this period). During 1995-99 this rose to an average of 47 percent, almost entirely due to stock-building in China. China's share of world ending stocks had typically averaged about 25 percent from 1980 to 1994, but jumped to 42 percent on average during 1995-99. Through 1995, it seemed reasonable to assume that outside of China and the United States, stocks as a share of use were on a downward trend. This was consistent with the increased economic efficiency that should be possible with improved transportation, communication, and legal protection. It also reflected changes in government policies in countries like Brazil, which no longer carried large government-owned stocks. The U.S. stocks-to-use ratio was also on a generally downward trend, particularly compared to the large peaks that occurred during the 1980s. China's stockholding behavior has been more obscure, in part because there has always been substantial doubt about the accuracy of relevant statistics. China's stock levels have also been volatile, with year-to-year changes of 5 to 10 million bales punctuating its policy shifts with alarming frequency. Generally speaking, since China embarked on its opening to the world economy in 1978, its stock-to-use ratio has trended up, with 1995-98 marking its highest and most prolonged peak. Since 1998/99, China has substantially reduced its stocks, although they still remain high compared with much of the rest of the world and with much of China's own pre-1995 history. China reduced stocks after 1998/99 by reversing the policies that caused a buildup during the preceding years. Stocks rose after 1994/95 as China raised its farm prices while maintaining an open trade regime. China's government- mandated farm prices proved difficult to reduce as world prices fell, and restricting imports seemed inconsistent with ensuring the profitability of its huge textile industry. Also, government policy locked older cotton in stocks in order to prevent bookkeeping losses as the market value of procured cotton tumbled below the cost of purchasing, processing, and storage. Stocks reached a staggering 106 percent of use in 1998/99, and China accounted for 47 percent of the entire world's cotton ending stocks. Then, starting in 1998/99, the government began applying quantitative restrictions to cotton imports and subsidizing exports. In 1999/2000 the government effectively cut farm prices by refusing to guarantee procurement, and in 2000/01 a program to allow the central government to absorb the cost of marketing losses for stockpiled cotton went into high gear, opening the floodgates for enormous government stocks to flow into the market. By 2000/01, China had cut its ending stocks by nearly 10 million bales, mostly from government stocks. World Ending Stocks Higher in 2001/02 World ending stocks are forecast to rise substantially in 2001/02, up 5.4 million bales to 44.4 million. As a share of consumption, world ending stocks are not expected to be above the levels seen just a few years ago--48 percent, compared with more than 50 percent in 1997/98 and 1998/99. But, with the shift in stockholding out of China, more of those stocks are available to the market, and their impact is greater. If China is excluded, world stocks as a share of consumption in 2001/02 are forecast to rise from 29 percent in the year before to 34 percent. During the 5 years before 2000/01, the average share was 26 percent, and it last exceeded 30 percent 1985/86. Global stock-building in 2001/02 is primarily concentrated in the United States, although larger rather than smaller stocks are the expected norm throughout the world. U.S. ending stocks are expected to increase 2.7 million bales, India's stocks are expected to rise by 500,000 bales, and increases of between 150,000 to 300,000 bales are foreseen for Turkey, Mali, Uzbekistan, Pakistan, Egypt, and China (fig. 8). Virtually no significant decline in ending stocks is foreseen for any country in 2001/02, with Iran proving the most significant exception with a 50,000-bale decline. With prices recently at historical lows, it makes sense that importers might purchase at a rate at least slightly above their annual use, and that exporters might delay sales to the extent possible. Ending stocks outside of China and the United States are forecast to equal 26 percent of world consumption, compared with 23.3 percent in 2000/01. This would be similar to stockholding behavior during 1991/92, when stocks in these countries rose from 22.8 percent of world consumption to 25.5 percent. Similarly, in recent years the trend in these countries has been towards higher rather than lower stocks. It is difficult to tell if this trend represents a shift in the location of stocks or buildup in stocks across a set of countries that will later require economic adjustment to bring these stocks down. One factor supporting the forecast increase in foreign stocks is the pace of U.S. export sales during the last half of calendar 2001. Current circumstances, including unusually low prices resulting in large marketing loan payments may change the seasonality of U.S. export sales; however, past experience with export sales through mid-November suggests that current activity is consistent with near-record U.S. cotton exports during 2001/02. With current forecasts for global consumption, 9.4 million bales of U.S. exports will lead to higher stocks in the rest of the world. Foreign ELS Output, Consumption, Exports, and Stocks Rising According to the International Cotton Advisory Committee (ICAC), foreign ELS production is expected to see its largest increase since at least the mid-1980s in 2001/02. At 2.5 million bales, foreign production is forecast 27 percent above the year before, largely because of a rebound in Egypt. Foreign ELS output fell steadily for a decade until 1995/96. Since then, while volatile, foreign production has been relatively steady, although 2001/02 is forecast above the 2.1-million-bale average of the preceding 5 years (table D). While Egypt's output is rebounding from 2000/01 to 1.3 million bales, the 2001/02 production is still forecast below the 1.6-million-bale level achieved during 1996/97 and 1997/98. China, on the other hand, is expected to see a year-to-year increase in output of 70 percent--to 390,000 bales--that will leave its 2001/02 crop 6 times the size of its 1996/97 and 1997/98 output. China's price reforms have restored the premium paid for ELS cotton in China, permitting a recovery in output. ELS consumption in 2001/02 by foreign producers is also expected to rise, but substantially slower than production. At 1.64 million bales, 2001/02 foreign producers' consumption is forecast at its highest level since 1997/98. As is the case with production, Egypt is both the largest foreign consumer and is expected to account for most of the total increase. Egypt is expected to consume 675,000 bales, compared with 560,000 bales by India, the next largest consumer. Foreign ELS producers' exports are expected to rise sharply in 2001/02, up 46 percent to 928,000 bales. This would be their highest export volume since 1993/94. Although exports are forecast to rise for all major ELS producers in 2001/02, Egypt, the world's largest exporter, accounts for much of the expected increase. Foreign ELS producers' ending stocks are also expected to rise sharply in 2001/02, up 21 percent to 1.3 million bales. Egypt's share of ending stocks is expected to fall as China's stocks more than double. However, Egypt is still expected hold more than half of the foreign ELS ending stocks in 2001/02. Situation and Outlook for Other Fibers U.S. Wool Production and Mill Use Continue Decline U.S. sheep numbers and wool production declined for the eleventh consecutive year in 2000. The U.S. farm price for shorn wool averaged $0.33 per pound, greasy, in 2000, 13 percent below a year earlier. Sheep numbers on January 1, 2001 were 4.9 million head, 6 percent below January 2000 and 51 percent below 1990. Wool production is estimated at 24.5 million pounds, clean (46.4 million, greasy) for the 2000 marketing year (table E). Lower sheep numbers in 2001 imply production this year could decline to nearly 23 million pounds, clean. Expected production this year would be the lowest on record and significantly below the early 1990s (fig. 9). During 2000, U.S. raw wool imports totaled 45.0 million pounds, clean, 4 percent above a year earlier (table F). During 2000, imports of fine wool represented 53 percent of total shipments. Imports of unimproved and other grades not- finer-than 46's totaled 21 million pounds in 2000. Shipments from New Zealand represented 71 percent of these coarser grades, while Australia accounted for 85 percent of the finer wools last year. With beginning stocks of 42.9 million pounds, total 2000 supplies were 132.4 million pounds, clean, 7 percent above a year earlier. Raw wool mill consumption totaled 77.2 million pounds, slightly below 1999. Apparel wool consumption, at 62.0 million pounds, was 2 percent below a year earlier, while carpet mill use was 13 million pounds above 1999 at 15.2 million pounds (table G). In 2000, the woolen system used 31.9 million pounds and the worsted system used 30.1 million. Raw wool exports during 2000 were 6.6 million pounds, clean, compared with 3.7 million a year earlier. Shorn wool shipments were 4.5 million pounds and not-shorn (pulled) exports were 965,287 pounds. Carbonized wool exports totaled 1.1 million pounds, compared with 173,700 pounds in 1999. The majority of shorn wool exports went to Germany, Mexico, Spain, and Italy. The majority of not-shorn wool went to Canada. U.S. prices for clean, mill-delivered territory raw wool declined during 2000, compared with the previous year. Finer grades 64's and 62's averaged $1.08 and $0.92 per pound, respectively, while the 60's and 58's averaged $0.75 and $0.65 per pound, respectively. The finer grades are down about $0.05 per pound from 1999, while the coarse grades are down $0.10. Current domestic prices of Australian raw wool of all but the finest grade averaged below market prices of a year earlier. Prices for Australia's finest wool, the 80's, averaged $2.80 per pound in 2000, 8 percent above 1999. For all other grades, prices declined about $0.30 per pound, compared with a year earlier. The 70's averaged $1.69 per pound and the 64's averaged $1.50 per pound. Noncellulosic Fiber Industry Weak The U.S. noncellulosic fiber industry during the first three quarters of 2001 declined significantly from a year earlier. Total shipments were 6.5 billion pounds, 11 percent below a year ago. Total noncellulosic production, at 6.4 billion pounds, was 13 percent less than a year earlier. Noncellulosic stocks at fiber producers' plants were 0.65 billion pounds at the end of the third quarter, down 8 percent from last year. Domestic shipments of noncellulosic fibers during January- September 2001 were 6.17 billion pounds, 10 percent below a year earlier. Filament fiber domestic shipments were 3.81 billion pounds, 9 percent less than a year ago. Filament fiber domestic shipments of olefin totaled 1.70 billion pounds, 1.19 billion of nylon, and 0.92 billion of polyester. Noncellulosic staple domestic shipments were 2.37 billion pounds, 12 percent below last year. Staple shipments were 1.35 billion pounds of polyester, 0.54 of olefin, and 0.47 billion of nylon. The carpet market continued to consume more fibers in facing and backing uses than any other fiber market (appendix table 43). During January-June 2001 (the latest data available), the carpet market used 1.81 billion pounds, 12 percent below last year. Noncellulosic carpet use accounted for more than 42 percent of total noncellulosic fiber domestic shipments. Nylon dominated the carpet market, constituting more than 49 percent of the total use of noncellulosic carpet fibers. Nylon staple carpet fibers accounted for more than 94 percent of nylon staple domestic shipments, while nylon filament carpet fibers totaled almost 76 percent of nylon filament domestic shipments. Preliminary data for third- quarter 2001 indicated that about 1.36 billion pounds of nylon were used in carpets during January-September 2001, 10 percent below a year earlier. The use of olefin in facing and backing during the first half of 2001 was 0.79 billion pounds, 14 percent less than last year. Olefin fibers constituted 44 percent of the noncellulosic fibers used in carpets. Carpeting was the most important use of olefins, at more than 52 percent. Woven textile production remained the second largest outlet for noncellulosic fibers, taking 20 percent of the January- June 2001 domestic shipments. The woven market used 0.86 billion pounds, almost 8 percent below a year ago. Two fibers made up more than 92 percent of this market: polyester accounted for 68 percent while olefin contributed 24 percent. Similarly, the knit market used almost 0.42 billion pounds, 20 percent below last year. Shipments of noncellulosic fibers to the knit market were almost 10 percent of total domestic shipments. Three fibers dominated the knit market: polyester at 0.29 billion pounds, constituted 69 percent, nylon at 0.08 billion accounted for 19 percent, and acrylic at 0.05 billion, totaled 12 percent. According to the capacity survey by Fiber Economics, Inc. in May 2001, (the latest data available), total in-place capacity to produce noncellulosic fibers in 2001 was 11.5 billion pounds, down 3.5 percent from 2000 (appendix table 42). Of the major noncellulosic fibers, only olefin showed an increase in 2001. Olefin capacity is also projected to rise in 2002, surpassing for the first time the estimate for polyester. Meanwhile, total capacity in 2002 is expected to be 11.4 billion pounds, 1 percent below 2001. Special Article Regional Shifts in China's Cotton Production and Use Hsin-Hui Hsu and Fred Gale -----1/ -----1/ 1/ Agricultural economists at USDA's Economic Research Service. The authors wish to acknowledge special assistance of data collection and discussion from Haowu Ding, Min Du, and Zuoyu Guo. ----- Abstract: This article examines the geographic distribution of cotton production in China. While overall cotton production was fairly stagnant in recent years, there have been important geographic shifts that are hidden in the total production numbers. This timely update on the geography of cotton production will help analysts make a better assessment of China's cotton production capacity. Keywords: China, cotton, production, consumption, WTO. China has maintained its position as the largest producer and consumer of cotton in the world, but growth in cotton production did not keep up with textile output growth during the 1990s. From 1990 to 2000, annual cotton output fluctuated around 4.4 million tons while estimated use of cotton in yarn production (including synthetic fibers) grew 25 percent (fig. A-1). With textiles expected to be one of the chief beneficiaries of China's accession to the World Trade Organization (WTO), further growth is expected. Since 1998, China has avoided large cotton imports by drawing down domestic cotton stocks and increasing the use of synthetic fibers, but continued growth in textile production expected after WTO accession suggests that imports will have to grow unless domestic production can be expanded. Regional Characteristics of China's Cotton Cotton is grown in most of China's provinces, municipalities, and autonomous regions, from the border areas of southern Yunnan to northeastern Liaoning and northwestern Xinjiang (fig. A-2). The major producing areas can be divided into three regions: the Yellow River valley, the Yangtze River valley, and the Northwest (fig. A-3 and table A-1). The Yellow River region encompasses the northern China plain, extending south from the Great Wall in northern Hebei province to the Huai River that flows through central Jiangsu and Anhui. The Yellow River region includes the northern provinces of Shandong, Hebei, Henan, Shanxi, and Shaanxi, and the municipalities of Beijing and Tianjin. In 2000, Henan was the leading cotton-producing province, with 0.7 million tons. The weather is often dry in the spring and irrigation is needed for cotton production (see details in table A-1). The major portion of the rainfall comes in summer, while the weather is usually dry in the fall during harvests. Because of its northern location, this region has about 180 days in the growing season and has to adopt early- maturing cotton varieties, which are usually double-cropped with winter wheat. American upland species, especially the early big boll types, were introduced and found to be well- adapted to this region in the 1930s (Shen). Most cotton is shipped by truck or rail to nearby textile mills. The Yangtze (Changjiang) River originates in Sichuan and flows east through Hubei, Anhui, and Jiangsu provinces. The Yangtze valley cotton area is bordered on the north by the Qinling Mountains in southern Shaanxi and the Huai River. The region includes Jiangsu, Anhui, Hubei, Hunan, Jiangxi, and Zhejiang provinces. Shanghai, a province-level municipality, is geographically within this region, but it has little cotton production, so we exclude Shanghai from the Yangtze cotton region. As we will see below, Shanghai was historically an important yarn production center, but its production fell steeply in the 1990s. In contrast with the Yellow River and Xinjiang regions, rainfall is relatively abundant in the Yangtze region. Annual rainfall averages more than 1,000 millimeters (or 3.38 feet), over 85 percent of which come during the cotton- growing season. Excessive rainfall in late summer and early fall often hurts cotton quality by fostering cotton pests and diseases. With long growing season, cotton is usually double-cropped with a winter crop (wheat or rapeseed) in the region. Transplanting of seedlings is a common practice in double cropping, which saves about 2 weeks of seed germination and growing time. Jiangsu and Hubei are the two leading cotton-producing provinces in the region, each yielding 0.3 million tons in 2000. The abundant water in the Yangtze region also provides transportation for cotton, both on the river itself and on a network of canals cut through rivers and streams and connected to the Huai and Yangtze rivers. Cotton gins and mills are situated along the rivers largely for the convenience of transportation. The Northwest region includes primarily the Xinjiang Uighur autonomous region plus northwestern Gansu province. Xinjiang covers one-sixth of the entire area of China and borders Tibet, Mongolia, and Central Asia. The Northwest climate is arid, with annual precipitation below 200 millimeters and wide daily swings in temperature, but dryness has kept pest and disease problems to a minimum. Xinjiang mainly grows upland cotton, with high quality color and fiber length due to its favorable climate conditions compared with eastern regions. Xinjiang has the only long-staple cotton production base in China. Xinjiang's remote location makes transportation vital. More than 70 percent of Xinjiang's cotton crop is shipped to eastern provinces or to foreign destinations. Xinjiang's capital, Urumqi, is linked to the Chinese rail network by a dual-track system to Lanzhou in Gansu province, a hub that connects with China's main east-west rail network. To the west and south of Urumqi, transportation is mainly by highways and segmented single-track railroads. Farmers use mules and light trucks for shipping cotton from their farms to local gins. Varying Yellow River Production Except for a brief period in the late 1970s and early 1980s, China's cotton production has not shown an upward trend, but there were important regional changes, including wide year- to-year swings in production in the Yellow River region and gradual but steady growth in Northwest production (fig. A- 4). Expansion of cotton production in the late 1970s and early 1980s took place primarily in the Yellow River region. Prior to that time, the Yangtze region was the primary production base and production in the Northwest was insignificant. Encouraged by government support policies, the Yellow River region's share of production rose dramatically from 30 percent in 1978 to over 60 percent in 1984, when China's cotton production peaked at over 6 million tons. However, production in the Yellow River region fell again in subsequent years before peaking again in 1991 at over 3 million tons. Production in the Yellow River region then plunged again to 1.4 million tons two years later, in 1993. This was largely due to a severe bollworm infestation, as well as increased labor costs in the region and changes in relative crop returns. In the late 1990s, transgenic bollworm-resistant cotton varieties--also referred as Bacillus thuringiensis (Bt) cotton--became available to farmers in Hebei and Shandong provinces, reviving production in the region (Hsu and Gale). Area planted to Bt cotton (mostly in these two provinces) increased from 100,000 hectares in 1998 to more than 1 million hectares in 2000 (Du). Bt cotton's share of total cotton area increased from 2.2 percent in 1998 to 28 percent in 2000. In additional to controlling bollworms, the new varieties reduce labor requirements and health hazards by cutting the number of sprayings of dangerous agricultural chemicals from 10 times to twice a year. While the cost of Bt cotton seeds is 5-6 times that of non-Bt seeds, the reduced costs and higher yields far exceed the difference in seed cost. Farmers in Shandong and Hebei eagerly adopted Bt cotton, even though the government in 1999 stopped guaranteeing that the cotton would be sold at a 'protected' (support) price. Bt cotton has led to a modest recovery in the Yellow River region's cotton production share. The share bottomed out at 32 percent in 1997 and rose to 38 percent by 2000, and was estimated even higher in 2001, but that was still well below the 60-percent share produced in the Yellow River region during the late 1980s. Cotton production in the Yellow River region is sensitive to prices of competing crops. Data from China's agricultural census indicate that the average Yellow River region cotton farm devoted only 12.8 percent of its acreage to cotton in 1996 (table A-2). About 43 percent of the average cotton farm's area was devoted to wheat, often winter wheat double- cropped with cotton. (On average, in this region about 70 percent of a farm's area was planted in two crops per year.) In the Yellow River region, corn is the chief crop competing for cotton acreage. Corn accounted for 22 percent of the average Yellow River cotton grower's sown area in 1996. Soybeans, peanuts, and vegetables each accounted for about 5 percent or less of the average cotton farm's acreage in this region. Farmers' cotton planting decisions are influenced by profitability of competing crops. During the 1980s and early 1990s, cotton production in Zouping county in Shandong province was usually well below government-set cotton production quotas because cotton procurement prices were not high enough to make growing cotton as profitable as growing other crops (Sicular). In years when cotton procurement prices were raised, farmers raised production. In 1990s, the cost of high labor and pesticide requirements also reduced the relative profitability of cotton production. In addition to its pest resistance, Bt cotton's lower labor and pesticide requirements have played a major role in encouraging cotton production in the Yellow River region since 1999. Steady Rise in Northwest Northwestern production has steadily increased and now accounts for about a third of national production. Xinjiang accounted for only 2.5 percent of production in 1978, but is now the largest producer, with 1.46 million tons in 2000. Gansu's production was less than 1,000 tons annually prior to 1991, but rose to 5,700 tons by 2000. High yields, low production costs, improved transportation, relatively few pest problems, and a strong government-led push to develop Western provinces have combined to encourage production. Compared with farmers in other regions, northwestern farmers have fewer alternatives to cotton, which accounted for 42 percent of the average northwestern cotton farm's acreage in 1996 (table A-2). Wheat and corn account for most of the remaining area. Production in the Yangtze River region has been more stable than in other regions. Production has varied between 1.2 million and 1.9 million tons since 1978. This region's share of national production has been cut in half from about 60 percent in the late 1970s to 27 percent in 2000. Based on China's cost of production surveys in 1998, the Yangtze River region has the least seed use per hectare (33.2 kg). But, average labor input in Yangtze River area, per mu----- 2/ ----- 2/ 15 mu = 1 hectare = 2.471 acres. ----- basis, was the highest among the three regions with 442 yuan in 1999. Rice is the primary crop in the Yangtze region. Cotton's share of acreage on Yangtze River region farms was 13 percent, the same as that of Yellow River farms. However, Yangtze farms had a much higher share of land devoted to rice (29 percent in 1996), which generally does not compete with cotton for land since it is a staple food crop and is grown in irrigated paddies. Farms in this region devoted 23 percent of land to wheat, 10 percent to rapeseed, 8 percent to vegetables, and 6 percent to corn. This region has the highest multiple cropping index, reflecting the intensive use of land in this region. Textile Industry Concentrates in Yangtze River Region Yarn-spinning and cloth-weaving industries are primarily located in cotton-producing provinces. In 1999, 45 percent of yarn production was in the Yangtze region, 35.7 percent was in the Yellow River region, and 5.4 percent was in the Northwest. Only 11 percent of yarn was spun outside the main cotton-producing provinces. The leading provinces in yarn production are Jiangsu (over 1 million tons in 1999), a coastal province in the Yangtze region, and its northern neighbor, Shandong (790,000 tons), in the Yellow River region. Two neighboring central cotton-growing provinces, Henan (Yellow River) and Hubei (Yangtze River), also produced over 500,000 tons of yarn each. Comparison of regional shares of yarn and cotton production for 1999 indicates that most Northwestern cotton is shipped eastward for spinning, while the Yangtze region spins a significant quantity of cotton produced in other regions (fig. A-5).-----3/ ----- 3/ China's cotton imports were negligible in 1999. ----- ----- The Yangtze River region's share of yarn production (47.2 percent) far exceeds its share of cotton production (27.3 percent), since that a large share of the yarn produced there is made from cotton shipped in from other regions. (Note: Shanghai, an important yarn-producing area, is excluded from the Yangtze River region since it is not a major cotton-producing area and Shanghai's trend differed from that in neighboring provinces). Conversely, the Northwest's share of cotton production (33 percent) far exceeds its share of yarn production (5.4 percent), indicating that most of the Northwest's cotton is shipped outside the region for spinning. The Yellow River shares of yarn and cotton production were similar, suggesting that most cotton produced there is also spun within the region. The share of yarn production in other regions (13.9 percent) exceeded the share of cotton produced outside the 3 major regions (2 percent), since that some cotton is shipped out of the major growing regions to be processed elsewhere, primarily in Guangdong, Fujian, and Shanghai. From 1990 to 1999, three-fourths of the increase in yarn production occurred in the Yangtze River region, primarily in Jiangsu (340,000 tons), Hubei (180,000 tons), and Zhejiang (120,000 tons). The Yangtze River region's share of yarn production rose from 38 percent to 45 percent between 1990 and 1999. While the Northwest continued to ship most of its cotton to other regions for spinning, its yarn production more than doubled, from 130,000 tons to 310,000 tons. The Northwest's share of yarn production also doubled from 2.8 percent to 5.4 percent. Yarn production in the Yellow River region also increased, but at less than the national rate, and its share of national production fell slightly. Yarn production outside the three main cotton regions fell. This was primarily due to a decrease of 220,000 tons in Shanghai, as production apparently shifted to lower-cost inland locations in neighboring provinces. In other provinces outside the three major cotton regions, yarn production was roughly constant between 1990 and 1999. Conclusion Many observers expect China's textile and garment exports to continue rising after WTO accession, which will likely boost cotton demand to feed the growing textile and apparel industries. A geographic view of China's cotton production and use points out several issues. One issue is the importance of transportation. Much of China's growth in cotton production has been in China's remote northwest corner, which now accounts for a third of all cotton production. Textile manufacturing has boomed primarily in eastern and central provinces. China's improved transportation network has facilitated this geographic separation of production and use. Rail shipments of cotton rose from 1.4 million tons in 1990 to 2.27 million tons in 1999. The average rail shipping distance for cotton was over 3,000 km in 1999, longer than for any other major commodity and up from 1,755 km in 1989. Can the transportation network handle even more production? Are subsidies necessary to make northwestern cotton cost-competitive on China's east coast, either at mills or ports? The spinning industry in Xinjiang has grown rapidly, and there is currently a focus on economic development in western provinces. Processing of cotton may increase, but yarn or other products must eventually be shipped eastward, since the final market in the northwest is limited in size by the region's relatively modest population and low average income. We also note the year-to-year variability in cotton production in the Yellow River region, which accounts for the largest share of production. In this region, cotton competes with other dryland crops, so cotton plantings are probably more sensitive to prices of grain, oilseeds, fruits, and vegetables that compete for cotton land. In the Northwest there are few other crops that can be grown profitably and in the Yangtze River valley the main crop is rice, which probably does not directly compete with cotton for land on a year-to-year basis. Thus, a rise in cotton prices would be required to bring forth more production in the Yellow River region, but that seems unlikely since domestic prices are currently above world levels. Bt cotton has played a role in reviving production in the Yellow River region. Some observers estimate that nearly all cotton in Shandong and Hebei was in Bt varieties by 2001. The low labor requirements could encourage widespread adoption in other provinces as well, helping to boost production in the region. Water resources may be a limiting factor on production growth in the Yellow River and Northwest regions. Falling water tables and unreliable surface water in the Yellow River region are becoming a serious problem. Lack of irrigation water may prevent cotton acreage from growing in this region. Cotton in the arid Northwest is heavily reliant on irrigation. Growth in Northwest cotton production will depend on whether water resources are adequate to support even more production and maintenance of irrigation infrastructure. References Du, Min, 'Bt Transgenic Cotton in China: Present and Prospect,' Prospects of the World Cotton Market and the Chinese Cotton Industry in the New Millennium, Guilin, China, June 2001. Hsu, Hsin-Hui and Fred Gale, 'China's Cotton Reserves To Meet Shortfall, Avert Price Rise,' U.S. Dept. Agriculture, Economic Research Service. China Agriculture and Trade Report, 2001. Shen, T.H., Agricultural Resources of China, Cornell University Press, Ithaca, New York, 1951. Sicular, Terry, 'Establishing Markets: The Process of Commercialization in Agriculture,' in Andrew Walder, ed. Zouping in Transition. Cambridge: Harvard University Press, 1998. Special Article Stephen MacDonald, Agapi Somwaru, Leslie Meyer -----1/ and ----- 1/ Agricultural economists at USDA's Economic Research Service. ----- Xinshen Diao -----2/ ----- 2/ Research Fellow, International Food Policy Research Institute. ----- Abstract: During the past 40 years, world textile trade has been in large part governed by the Multifiber Arrangement (MFA) and its predecessor agreements. Starting in 2005, in accordance with World Trade Organization (WTO) obligations, these restrictions must end, and the import restrictions developing countries have imposed largely must end as well. A dynamic computable general equilibrium (CGE) model finds that trade reform improves welfare in every region of the world, and causes world textile, apparel, and cotton production to rise. In the long run, world textile and apparel prices fall, while the world price of cotton rises. U.S. production declines for cotton as well as for textiles and apparel, although U.S. cotton exports rise. Keywords: Cotton, textiles, apparel, trade, policy, quotas, Multifiber Arrangement, WTO. The Uruguay Round's Agreement on Textiles and Clothing (ATC) mandates the end of the quotas established under the MFA and also the reciprocal termination of the restrictions imposed by developing countries on their imports of textiles and clothing. By 2005, restrictions that do not meet General Agreement on Tariffs and Textiles (GATT) standards are supposed to be phased out, and the strengthened dispute- settlement mechanism the Uruguay Round introduced to the World Trade Organization increases the likelihood that the agreed liberalization will in fact occur. Methodology The study is conducted in an intertemporal general equilibrium framework. The detailed description of the model can be found in Diao and Somwaru (2000, 2001). The data are from the GTAP database version 5, pre-release 3 (GTAP, 2001), including data about trade flows, production, and consumption in each country/region in 1997. The original data set includes 66 countries/regions and 57 aggregate sectors. For this study, we aggregate the data into 13 countries/regions (listed below) and 7 sectors, including cotton, other crops, livestock, processed food, textiles, apparel, and an aggregated manufacturing and services sector. The MFA phaseout and other changes in trade policy can be expected to affect textile and apparel (T&A) trade directly. Changes in a country's T&A exports can also affect the country's domestic economy as well as the world economy through input-output, supply-demand, and price linkages. The study tries to capture such linkages among economic activities, and hence to evaluate the general equilibrium impact of the MFA phaseout on the world economy. The intertemporal specification of the model captures the benefits not only due to resource re-allocation but also the dynamic gains due to growth in investment and capital flows. In the study, we first try to distinguish the exporting countries by whether or not trade is restrained under the MFA. Specifically, among the countries/regions whose exports are restrained by the MFA, we include: (1) China, (2) India, (3) the region of the other South and Southeast Asian countries, (4) the Middle East, (5) the region of former Soviet Union countries, and (6) the region of the Latin American countries (excluding Mexico and the Caribbean countries). We also include (7) the region of North African and East European countries and (8) the region of other African countries, representing the developing countries free from restraint in exporting to the European Union (EU); and (9) the region of Mexico and Caribbean countries, representing the countries free from restraint in the North American markets. The study also includes the following industrial countries as major importers and exporters in the world, including two restraining regions: (10) North America (U.S. and Canada), (11) the EU, and two non-restraining regions: (12) Australia and New Zealand, and (13) Japan, Taiwan, Hong Kong, and Korea. As Taiwan, Hong Kong, and Korea currently have unused quotas in their apparel exports, we treat them as developed countries in the model. This study abstracts from the question of whether importing or exporting countries capture the rents from MFA quotas, and assumes these rents are dissipated by rent-seeking behavior and inefficiency. That is, the MFA does not create either a price gap between domestic and border prices or quota rents for the restraining countries. Instead, the restraints cause difficulty for some developing countries to export their textile and apparel products to the restraining countries (North America and the EU in the model), and hence lower the efficiency of their exports. In a post MFA world, exports of textile and apparel products become relatively easy for the developing countries (included in countries/regions 1 through 6), and hence exports grow. We simulate the possible effect of MFA phaseout by improving the efficiency of textile and apparel exports from the countries/regions restrained by MFA (countries/regions 1 through 6). Technically, we exogenously increase the efficiency coefficient in the export function by 0.3 percent annually. Moreover, we assume other trade barriers (represented by tariff equivalent rates) on textile and apparel imports are reduced by 30-40 percent in all countries (including developing countries restrained by MFA), and the reduced tariff equivalent rates are close to each country's average tariff rate for other manufacturing imports. Integrating trade barrier reductions in the developing countries into the simulation means the simulation includes the reciprocal trade reform the ATC introduced with the MFA phaseout. Finally, based on an extension of Frankel and Romer's (1999) work we analyzed the relationship between T&A trade and national income of 91 countries over 37 years. Based on our econometric analysis, we exogenously increase the growth of T&A trade in the simulation. Freer Trade Brings More Trade World T&A trade increases in the model due to the MFA phaseout and other reforms under the ATC, increased efficiency, and increased growth in the T&A sectors. Compared with the base, world T&A trade increases by 5-16 percent annually in the simulated time period of 25 years. Consistent with the trend in the historical data and the higher protection generally in place for apparel compared with textiles, the model results show that world apparel trade will increase twice as fast as textile trade in the post-MFA world. It is no surprise that the increase in world trade is mainly due to more apparel exports from developing countries, as their exports become more efficient in the model. As liberalization improves trade opportunities, exports from countries formerly restricted by the MFA grow, but exports decline from regional trading partners that received preferential access from the U.S. and the EU before the MFA phaseout. By the end of the simulation period, the volume of China's apparel exports rises 37 percent due to liberalization, India's 36 percent, and the rest of Asia's 42 percent. Mexico and the Caribbean suffer an 8-percent decline as the value of their North American Free Trade Agreement (NAFTA) and Caribbean Basin Initiative preferences decline, and Eastern Europe and North Africa's apparel export volume falls 6 percent as competition in the EU increases. Welfare Gains: Some Gain More Than Others From the world's perspective, a more liberalized textile and apparel sector implies more efficient allocation of resources along with the 'dynamic,' long-term effects from increased savings and investment, resulting in higher global welfare. We use the well-accepted equivalent variation (often referred to as the willingness to pay) to measure the social welfare gains or losses in the post-MFA world. In a static analysis, the welfare effects are often measured by using the status-quo (pre-reform) prices as the base, and addresses the question: what income would be equivalent to the change brought about by, for example, liberalizing world T&A trade (Varian, 1984). We borrow this concept in our dynamic analysis: in other words, we evaluate the welfare effects within each time period and the entire time path by summing the discounted value of this measure over time. As expected, most countries whose textile and apparel exports are restrained by the MFA show post-MFA gains (table B-1). Global welfare increases by $103 billion in the short run (5 years) and by $204 billion in the long run (25 years). In percentage terms, the long-run global increase in welfare is not large, 0.9 percent. However, due to the difference in production and trade structure among countries, changes in global T&A trade policy can affect the rest of economy differentially in different countries. For this reason, the welfare gains can be different among countries that benefit directly from the MFA phaseout. For example, textile and apparel exports increase the most in the region of Other Asian countries. However, from a welfare point of view, China gains the most of any exporter both in absolute value and in percent of total consumers' expenditures. One reason is that the T&A sector contributed more to GDP in China than in the other countries. The benefits of trade liberalization are clearly not confined to exporters. The Former Soviet Union (FSU) has welfare gains comparable to China's in the long run, despite being a significant T&A importer and having a very low GDP share for T&A. After the United States, the FSU has the largest ratio of net T&A imports to consumption in the base period. With its large T&A deficit, and relatively low degree of protection (compared with developing Asian T&A exporters), the FSU is well placed to enjoy the benefits of increased efficiency in other countries as the MFA restraints are eliminated. Not surprisingly, the largest increase in apparel consumption of any region is in the FSU, a 12-percent increase in the value of consumer demand. The regions achieving the smallest welfare gains are the preferential trading partners of the U.S. and the EU. Mexico and the Caribbean and Eastern Europe and North Africa are the only developing regions where T&A output declines (table B-2) as trade reform under the ATC erases the benefit of their current preferential access to the largest import markets. Nonetheless, ATC reform raises welfare in these regions substantially more than it raises welfare in any developed region. In the long run, Mexico's welfare improves by 1 percent and Eastern Europe's by nearly 2 percent. A mix of larger output gains in non-T&A sectors, apparel consumption responses surpassing developed countries, and apparel consumption shares surpassing developed countries led to welfare increases several times as large as those in any developed region. Production Effects Globally, the value of textile and apparel production is higher after liberalization, each 2.5 percent higher than the base year by year 25 of the simulation. The value of cotton production is 3.5 percent higher. However, prices for textiles and apparel are lower in the simulation, and cotton prices higher, so that in quantity terms global textile and apparel production rises at least twice as much as cotton production in the long run. With fewer trade barriers, efficiency increases and the real price of apparel falls 4 percent in the long run. Textile prices fall by a smaller amount--2 percent in the long run-- since pre-reform barriers are lower for textiles than apparel. Cotton prices rise globally by about 2 percent. While world cotton production is relatively unchanged (about 2 percent higher), various regions have larger or smaller output, with the largest decline occurring in the United States. The volume of U.S. cotton production falls 1 percent by the end of the simulation, about half as much as the decline in U.S. textile production. The largest increases in cotton production occur in China and Other Asia. The only non-U.S. region where cotton output declines is the EU. The largest increase in cotton production is in China, up 9 percent in the long run, followed by a 6-percent increase in the Middle East. ATC Reduces U.S. Output With reduced import protection under the ATC, and with increased export efficiency in countries previously constrained by the MFA, the U.S. T&A trade deficit registers a substantial increase with reform. Consumption of apparel increases by about 7 percent as world prices fall and incomes rise, but the U.S. textile industry contracts by 2 percent as textile imports rise and the U.S. apparel industry shrinks. With the decline in output, the U.S. textile industry's demand for cotton decreases, helping reduce U.S. cotton production 1 percent. This decline contrasts with an increase foreseen in U.S. cotton production due to the Uruguay Round Agreements (URA) in at least one earlier study using partial equilibrium analysis (USDA, 1994). Two important differences between the assumptions of this study and the earlier USDA study are that, 1) this study includes China in the WTO for the purposes of T&A trade, and that, 2) other, non-T&A aspects of the URA are not incorporated in this study. An important set of policies not included here are changes in agriculture under the URA not implemented as of 1997. In addition, any changes in China's agricultural policy that might be associated with its WTO accession are not incorporated. This study was an attempt to isolate the impact of the MFA and other T&A trade barriers, not an attempt to do a complete URA or China accession analysis. U.S. welfare increases due to ATC reform in this study despite the declines in textile, apparel, and cotton production as these declines are more than offset by gains in other sectors of the economy. In agriculture, the output of other crops and livestock increases due to the indirect effects of the ATC reforms, and U.S. GDP increases as the rest of the economy's growth more than offsets the reduction of textile and apparel value-added. U.S. welfare is also improved by increased access to lower-priced apparel. Conclusions Liberalization of world T&A trade would enhance global welfare slightly, and prove particularly beneficial to developing countries, with China and the FSU achieving the largest gains. Some apparel exporting regions that now have preferential access to developed country markets would suffer a reduction in T&A output as a result of expanded access for countries like China and India, but would see increased welfare despite the setbacks in this sector. While Mexico and Eastern Europe would see a smaller increase in welfare than any developing region, their welfare is still estimated to improve more than in any developed region. World production of textiles, apparel, and cotton would rise, but prices of textiles and apparel would fall. The United States would benefit the least of any region from liberalization under the ATC, but would see welfare improve slightly despite lower output of textiles, apparel, and cotton. References Diao, Xinshen and Agapi Somwaru. 2001. 'A Dynamic Evaluation of the Effects of A Free Trade Area of the Americas--An Intertemporal, Global General Equilibrium Model,' Journal of Regional Integration 16(1): 21-47. Diao, Xinshen and Agapi Somwaru. 2000. 'An Inquiry on General Equilibrium Effects of MERCOSUR--An Intertemporal World Model,' Journal of Policy Modeling 22(5): 557-588. Frankel, Jeffrey A. and David Romer. 1999. 'Does Trade Cause Growth?,' The American Economic Review 89(3): 379-399. MacDonald, Stephen, Somwaru, Agapi, Diao, Xinshen, and Meyer, Leslie. 'The Agreement on Textiles and Clothing: Impact on U.S. Cotton Industries in an Intertemporal Global General Equalibrium Framework,' Selected Paper, American Agricultural Economics Association Summer Meetings, August 6, 2001. U.S. Department of Agriculture, Office of Economics and Economic Research Service. 1994. Effects of the Uruguay Round Agreement on U.S. Agricultural Commodities. Varian, Hal R. (1984) Microeconomic Analysis, Second Edition, New York and London: W.W. Norton & Company. END_OF_FILE