COTTON AND WOOL YEARBOOK December 03, 1999 November 1999, ERS-CWS-1999 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 ERS-NASS order desk in about 3-4 weeks. Call toll-free, 1-800-999-6779 and ask for stock # ERS- CWS-1999, $21. ERS-ERS-NASS accepts MasterCard and Visa. --------------------------------------------------------------------------- Contents Summary U.S. Cotton Situation and Outlook. World Cotton Situation and Outlook U.S. Wool Situation and Outlook World Wool Situation and Outlook Manmade Fibers Special Articles: Analyzing U.S. Cotton Acreage Response Under the 1996 Farm Act An Economic Analysis of U.S. Total Fiber Demand and Cotton Mill Demand List of Tables Situation Coordinator Leslie A. Meyer (202) 694-5307 Principal Contributors Leslie A. Meyer (202) 694-5307 Steve MacDonald (202) 694-5305 Robert Skinner (202) 694-5313 Mae Dean Johnson (202) 694-5299 (Statistics) William Lin (202) 694-5303 Editor Diane Decker Layout and Text Design Wynnice Pointer-Napper Summary U.S. cotton production in 1999 is forecast at 16.53 million bales, 19 percent above last season's 13.92 million. This season's output is based on a much larger area but a lower national yield. U.S. cotton acreage jumped 1.2 million acres from 1998 to 14.6 million, the third largest area in the past 35 years. After approaching 20 percent in 1998, abandonment returned to a more "normal" level this season, and harvested cotton area is estimated at 13.4 million acres, nearly 3 million above 1998. The national average yield is forecast to decline 33 pounds from last season to 592 pounds per harvested acre, the lowest since 1995. U.S. cotton exports in 1999/2000 (August/July marketing year) are expected to rebound from last season's 13-year low. Shipments are projected to reach 5.7 million bales this season as an improving world economy in the aftermath of the Asian crisis and more competitive prices resulting from the reinstatement of the "Step 2" program are coupled with the larger exportable supply. In early November, U.S. export commitments had already reached 4.2 million bales, nearly 1 million above a year earlier. Even with world cotton trade rising this season, U.S. exports are climbing faster. The U.S. share of the global market is projected to climb to 22 percent in 1999/2000, compared with 18.5 percent last season. In contrast, U.S. cotton mill use is expected to fall slightly (2 percent) this season to 10.2 million bales. The continued rise in cotton textile imports, a projected slowdown in the robust U.S. economic growth of the last several years, and a decline in denim demand are expected to moderate mill consumption. In 1999, cotton textile exports are expected to increase for the 15th consecutive year, reaching the equivalent of 4.3 million bales. On the other hand, cotton textile imports will rise for the 11th year in a row, and could approach the equivalent of 13.5 million bales, further widening the textile trade deficit. In addition, total domestic consumption (mill use plus net textile trade) could reach a record 9.5 billion pounds, with per capita consumption surpassing last year's 34.4 pounds. Similar to 1998/99, U.S. cotton beginning stocks were 3.9 million bales. However, the larger production this season has mitigated the need for foreign supplies and imports are expected to be small in 1999/2000. As a result, total U.S. cotton supply is projected at 20.5 million bales, more than 2 million above 1998/99. Meanwhile, demand is forecast at 15.9 million bales, 8 percent above last season. Based on these supply and demand estimates, U.S. ending stocks for 1999/2000 are estimated to rise to 4.6 million bales, 17 percent above last season and similar to the 1992 season. With both stocks and use projected to rise in 1999/2000, the stocks-to-use ratio is expected to increase only 2 percentage points to 29 percent. World cotton production in 1999/2000 is forecast to rise to 87.3 million bales, up 2.8 million, primarily due to the larger U.S. crop. Foreign production, projected at 70.8 million bales, is only marginally higher. Production increases in 1999/2000 for Pakistan and Uzbekistan stem from higher area and yield, but are offset by lower production in China and Mexico, where planted area declined. Likewise, world cotton consumption is projected to rebound after two consecutive declines to 87.8 million bales in 1999/2000, a gain of 2.8 million bales. Although lower use is expected in the United States, consumption in China and India--the world's two largest consumers--is forecast to rise a combined 1.2 million bales and account for over 40 percent of the rise in consumption. Brazil, Turkey, Mexico, and Pakistan are also projected to consume more cotton this season as world economic growth rebounds. World cotton exports in 1999/2000 are forecast to increase 9 percent to 25.7 million bales. Foreign shipments are expected to reach 20 million bales, above the last two seasons and equal to 1996/97. Rising exports from Uzbekistan--the leading foreign exporter--coupled with China's return as a major shipper in 1999/2000 account for nearly all of the projected 800,000-bale increase in foreign exports this season. In addition, China's 1.2-million-bale export estimate is the highest in over a decade. With the increase in world cotton consumption, stocks are projected to decline 600,000 bales to 41.3 million. Despite the expected increase in U.S. stocks, foreign stocks in 1999/2000 are forecast to fall nearly 1.3 million bales to 36.7 million, the lowest in 3 years. However, the foreign decline is attributable to China, where stocks are forecast to fall nearly 2.5 million bales this season. Excluding China, foreign stocks are expected to increase 1.2 million bales to 21.8 million, equaling those of 1991/92. At season's end, China is forecast to hold 36 percent of the world's cotton stocks, down from 41 percent in 1998/99. U.S. raw wool production in 1999 is estimated at 25 million pounds, clean, significantly below the early 1990s and the lowest on record. Despite lower production, U.S. wool imports are also projected to fall in 1999 as consumption continues its recent descent. Imports of raw wool are projected at only 40 million pounds, significantly below 1998. Likewise, U.S. raw wool mill demand is expected to reach about 85 million pounds in 1999, 25 percent below a year ago. As a result, 1999 carryover supplies are forecast higher at 48 million pounds, the largest in 4 years. U.S. Cotton Situation and Outlook U.S. Cotton Review, 1998/99 The 1998/99 (August/July) season began with U.S. cotton stocks reported at 3.89 million bales, a slight 2 percent below the 4-year high a year earlier. The 1998 crop was the third planted under the Federal Agriculture Improvement and Reform (FAIR) Act, which allows producers total planting flexibility in response to the marketplace. This flexibility enticed some cotton producers to increase their grain acreage at the expense of cotton area. At planting time, producers were faced with cotton prices that were below those of the previous year and relatively higher net returns for competing crops. As a result, cotton planted area in 1998 was lower in each region except the Southwest, where it was marginally higher. Only 13.4 million acres were planted to cotton in the United States in 1998, nearly 4 percent below the 13.9 million planted in 1997. In addition, a relatively large abandonment near 20 percent in 1998 reduced harvested area to only 10.7 million acres, the lowest since 1989. Despite the increased abandonment, the national yield averaged only 625 pounds per harvested acre, well below the previous two seasons. As a result, 1998 cotton production totaled 13.9 million bales, nearly 5 million below 1997 and the smallest crop in 9 years. In 1998/99, U.S. cotton demand declined a dramatic 4 million bales from the previous year and nearly matched the production drop. Total demand hit a decade low of 14.7 million bales. U.S. mill consumption last season decreased to 10.4 million bales, 8 percent or nearly 1 million bales below a year earlier, when it was the largest in over 50 years. Despite this decline, U.S. mills imported nearly 450,000 bales of raw cotton in 1998/99--the most since the 1928 season--to replace desirable qualities not available with the U.S. crop shortfall. Reduced mill use also resulted partly from the dollar's strength that provided relatively inexpensive cotton textile and apparel imports to compete with U.S.-made products. U.S. raw cotton exports fell 3 million bales in 1998/99 to 4.3 million. A reduction in total foreign imports and consumption--related to the Asian crisis--and the short U.S. crop led to the lowest exports since the 1985 season. Also weighing on shipments was the loss of "Step 2" funding that reached the statutory limit on December 15, 1998. U.S. cotton quickly became uncompetitive in an abundant global market. Even as world trade slipped to its lowest level in nearly three decades, U.S. exports fell faster and the U.S. share of global trade reached only 18.5 percent, well below 1997/98's 28.1 percent. With a U.S. cotton supply of 18.2 million bales in 1998/99 and total demand at 14.7 million, stocks at the close of the season on July 31 rose slightly from the beginning level. Ending stocks were placed at 3.9 million bales, for a stocks-to-use ratio of 26.7 percent. Due to the decrease in demand and adequate carryover stocks, 1998/99 producer prices declined once again. The average upland farm price was 60.2 cents per pound, compared with 65.2 cents in 1997/98, while the extra-long staple (ELS) price was $0.924 per pound, compared with $1.014 a year earlier. Supply Outlook for 1999/2000 As planting time approached for the 1999 crop, U.S. cotton prices were well below those of the 1998/99 season. However, depressed prices of alternative crops, concerns about aflatoxin in corn, and the safety net supplied by the cotton marketing loan program provided producers enough incentive to plant additional acreage to cotton. In 1999/2000, U.S. cotton acreage jumped 1.2 million acres--9 percent--from 1998 to 14.6 million, similar to the 1996 season and the third largest in the last 35 years. Upland acreage this season is estimated at nearly 14.3 million acres, compared with 13.1 million in 1998. In contrast, ELS area fell 10,000 acres from last season to 318,000 in 1999. The increase in total acreage this season and a return to a "near normal" national abandonment kept harvested area above last season's planted area. Based on November estimates, abandonment is estimated at 8 percent--1.2 million acres--well below 1998's 20 percent. Harvested area this season is estimated at 13.4 million acres, nearly 3 million above 1998 (table A). Although U.S. cotton crop conditions were generally more favorable than a year ago, a persistent drought across much of the cotton belt and wind and flood damage related to hurricane activity in North Carolina trimmed cotton output this year. Based on November 1 conditions, the 1999 U.S. cotton crop is estimated at 16.5 million bales, below USDA's August estimate of 18.3 million but nearly 19 percent above last season's final production estimate. Compared with 1998, this season's cotton production is based on a much larger area but lower national yields. The U.S. average cotton yield is currently projected at 592 pounds per harvested acre, 33 pounds below 1998 and the lowest since 1995 (figure 1). Upland production is forecast at 15.8 million bales in 1999, with an average yield projected at 581 pounds per harvested acre. On a regional basis, expected production is higher in three of the four regions. Twenty-percent gains in the Delta and West regions combine for a 1.2-million-bale increase in production this season, a similar quantity to the one-third higher output seen in the Southwest. Output in the Southeast region is expected to decline about 3 percent. In addition, ELS production is forecast at 684,500 bales, up more than 200,000 bales from 1998 and the second largest ELS crop on record. The increase in ELS output stems from the domination by California, which now accounts for over 80 percent of the area and production. The ELS yield is estimated at a record 1,063 pounds per harvested acre. By mid-November, 77 percent of the total U.S. cotton crop had been harvested, slightly ahead of the 5-year average. Similar to the previous season, U.S. cotton supplies on August 1, 1999, were 3.9 million bales. But unlike 1998/99, only a small volume of raw cotton imports are expected this season as the larger production appears to have mitigated the need for foreign supplies. In addition, U.S. and world prices are more closely aligned than a year ago. As a result, total U.S. cotton supply in 1999/2000 is projected at 20.5 million bales, more than 2 million above 1998/99. Demand Outlook for 1999/2000 The U.S. cotton demand outlook points to an improved offtake in 1999/2000. This season, total cotton use is projected at 15.9 million bales, 8 percent above 1998/99, with exports and domestic mill use moving in opposite directions. U.S. exports are projected to rebound from last season's disappointing 13-year low, but remain well below shipments of 2 years ago. Two key factors keeping U.S. exports from returning to the robust 1997/98 level include China's return as a raw cotton exporter and the largest foreign beginning stocks since 1986/87. Supporting U.S. shipments, however, is an improving world economy in the aftermath of the Asian crisis. U.S. prices are also more competitive this season. Despite an approximate 20-cent per pound decline in world prices from a year ago, U.S. prices have decreased similarly (table B). A-Index prices fell to about 47 cents per pound in October, while the comparable U.S. price quote was near 55 cents. Demand for U.S. cotton overseas is also expected to improve due to legislative funding for the "Step 2" program. See the next section for more details on this and other legislative actions affecting cotton this season. Currently, 1999/2000 U.S. exports are projected to rise over 1 million bales from last season to 5.7 million, but still nearly 2 million bales below 1997/98. Upland shipments are expected to reach 5.3 million bales, while ELS exports are forecast at 400,000 bales. With the 30-percent rebound in U.S. exports and prospects that global cotton trade will improve to 25.7 million bales, the U.S. share of world trade will rise modestly. As of November, the U.S. share of world trade is estimated at 22 percent for 1999/2000, compared with 18.5 percent last season (figure 2). Based on U.S. Export Sales data through early November, about 650,000 480-pound bales of upland cotton have been shipped, compared with 1 million in 1998/99. While the pending "Step 2" funding may have delayed some shipments, sales "on the books" are running well ahead of a year earlier. Upland commitments (shipments plus outstanding sales) for 1999/2000 have reached 3.9 million bales thus far, compared with 3.1 million a year ago. ELS commitments are also above last season with 255,000 bales committed, compared with 233,000 by early November 1998. U.S. cotton mill consumption is estimated to reach only 10.2 million bales in 1999/2000, 2 percent or 200,000 bales below last season (figure 3). The continued rise in cotton textile imports, a slowdown in the "exceptional" growth seen in the U.S. economy over the last 3 years, and a decline in denim demand will help moderate mill consumption this season. Upland mill use is projected at 10 million bales in 1999/2000, while ELS consumption is estimated at 160,000 bales. Based on the first 3 months of data from the Department of Commerce, the seasonally adjusted annual rate of cotton consumption averaged nearly 10.1 million bales. Actual cotton mill use for August through October 1999 reached 2.6 million bales, compared with 2.8 million a year earlier. Cotton mill use, like manmade fiber use, has fallen from a year ago, but cotton has declined faster during the first 3 months of 1999/2000. As a result, cotton's share of fiber consumed on the cotton system has averaged only 78.7 percent, compared with 79.3 percent for the entire 1998/99 season. Based on these projections of U.S. cotton supply and demand, ending stocks for the 1999/2000 season are estimated to rise to 4.6 million bales, 17 percent above last season and similar to the 1992 season. Although stocks are expected to increase about 700,000 bales from the beginning level, the ratio of stocks to use is projected to rise only 2 percentage points from last season to 29 percent (figure 4). Due to much larger production, ending stocks of ELS cotton are projected to more than double during 1999/2000. The ELS price support loan rate (79.65 cents per pound) is currently near the world price level. As a result, a significant quantity of ELS cotton may be forfeited to the Commodity Credit Corporation (CCC) under the loan program. Legislative Changes Affecting Cotton Discussions about refunding the "Step 2" program began well before the funds were exhausted in December 1998 when the FAIR Act's statutory spending limit of $701 million was reached. At that time, U.S. cotton became less competitive on the world market and soon the "Step 3" import quotas began to trigger. Despite a significant drop in cotton demand last season, this spring saw U.S. cotton plantings increase--due to less attractive alternatives--leading to prospects of excess supplies. And while U.S. agriculture has not experienced the prosperity of the general economy, conditions were compounded by weather-related problems once again in 1999. As a result, additional funding for agricultural aid was developed in Congress--and later signed by the President--for fiscal year 2000 which began October 1, 1999. In general, the legislation increased the payment limitation from $75,000 to $150,000 per person and also doubled the production flexibility contract payments for the various eligible commodities. For cotton, the major legislative change was the reinstatement of the "Step 2" program through July 31, 2003. The legislation left the calculation of the "Step 2" payment essentially unchanged and funding for the program was not limited to a spending cap as had been previously established. In addition, changes to the "Step 3" program were made that altered the trigger mechanism and also set a maximum import level for a given marketing year. Under the new legislation, a special upland cotton ("Step 3") import quota will be triggered after any consecutive 4-week period that the average price quote for the lowest U.S. growth--adjusted for any "Step 2" payment in effect--exceeds the northern Europe price by more than 1.25 cents per pound. However, the U.S. growth is not adjusted for the "Step 2" payment rate if the U.S. upland stocks-to-use ratio--excluding projected raw cotton imports but including the quantity already imported during the marketing year--is below 16 percent. Additionally, the new import limitation for a given marketing year is equal to 5-weeks' consumption of upland cotton by domestic mills at the seasonally adjusted average rate of the 3 months immediately preceding the first special import quota that is established in the marketing year. U.S. Textile Trade and Domestic Consumption The overall volume of U.S. textile trade has increased once again during calendar 1999. Because of the Asian crisis and the continued trade liberalization of textile and apparel products, the U.S. trade deficit will widen in 1999 as the rise in imports outpaces the gain in exports. In calendar 1998, the textile trade deficit reached a record of nearly 6.5 billion (raw-fiber equivalent) pounds, 21 percent above 1997. Textile exports during the first 9 months of 1999 expanded to 3.45 billion pounds, compared with 3.38 billion in 1998. However, textile imports through September 1999 increased to 9 billion pounds, compared with 8.3 billion in 1998. As a result, the textile deficit for all fibers through the first 9 months of 1999 has approached 5.6 billion pounds, 14 percent above the same period in 1998. Similarly, cotton textile trade for January through September 1999 has increased. Cotton textile exports are running 7 percent ahead of last year, reaching 1.6 billion pounds by September. Meanwhile, cotton textile imports have jumped dramatically, rising 11 percent to 5 billion pounds during the first 9 months of 1999. For calendar 1999, cotton textile exports will rise for the 15th consecutive year while imports expand for the 11th year. Exports for the 12 months will likely approach the equivalent of 4.3 million bales (2.1 billion pounds), 5 percent above 1998. On the other hand, cotton textile imports could exceed the equivalent of 13.5 million bales (6.5 billion pounds), 8 percent higher than in 1998. With imports rising faster than exports, the cotton textile trade deficit is expected to expand again in 1999 to the equivalent of 9.2 million bales of raw cotton (figure 5). Meanwhile, total U.S. fiber mill use is expected to decline slightly from 1998. Mill use of all fibers in 1999 should range between 16.5 billion pounds and last year's 16.7 billion. With U.S. cotton mill use expected to decrease from the level reached in calendar 1998 (see special article in this issue), cotton's share of total fiber mill use is likely to slip to near 30 percent. On the other hand, manmade fiber mill use is expected to account for about 68 percent of the total in 1999, slightly above 1998. With cotton mill use lower and the continued strength of textile imports, calendar 1999 marks the second consecutive year in which the raw-fiber equivalent of cotton textile imports exceeds the quantity consumed by domestic mills. While the imported quantity had moved closer to that used by U.S. mills for a number of years, the current trend is expected to continue as textile and apparel trade liberalization accelerates in the future. Total domestic consumption (mill use plus net textile trade) of cotton is projected to expand slightly for the third consecutive year. Based on data for the first 9 months of 1999, domestic consumption totaled 7.2 billion pounds, nearly 1.3 percent above a year ago. By year's end, domestic consumption is expected to approach 9.5 billion pounds, a record high. In addition, the U.S. per capita consumption of cotton could reach 35 pounds in 1999, up from last year's 34.4 pounds, with about 19 pounds being produced in U.S. mills (figure 6). World Cotton Situation and Outlook World Cotton Consumption, Trade, and Prices Fall in 1998/99 In 1998/99 world cotton markets reeled under a nearly unprecedented decline in consumption, one of the largest declines in world trade in the last 40 years, and the largest ending stocks since the mid-1980s. Although production also fell from the year before, the 1998/99 average A-Index, in inflation- adjusted terms, fell 19 percent. U.S. exports suffered as well, dropping 42 percent to 4.3 million bales. This marked only the second time in more than 20 years that U.S. exports dropped below 5.0 million bales or below a 25-percent share of U.S. beginning supplies (beginning stocks plus production). World production fell 7 million bales from the year before in 1998/99, to 84.5 million bales. Declines were led by a 5- million-bale drop in U.S. production. Foreign production fell only 2.1 million bales, to about its 1996/97 level of 70.6 million bales (table C). In the largest foreign declines, Pakistan's crop fell 875,000 bales and Uzbekistan's fell 628,000, as unfavorable weather at the beginning and the end of the season, respectively, led to yet another year of disappointing yields in each country. Weather problems also helped trim Argentina's crop by about 500,000 bales, and Egypt's production was also about 500,000 bales lower due to price reforms and an unusually hot late season there. China's production fell only 400,000 bales from the year before in 1998/99 as larger area and higher yields in Xinjiang partly offset reduced crops in most other provinces. Sub-Saharan Africa harvested 579,000 fewer bales than during the year before, due to a 270,000-bale decline in Franc Zone production and declines in Sudan, Tanzania, and Zambia. Among the handful of countries with notably larger crops were India (up 463,000 bales), Brazil (up 355,000), Turkey (199,000), and Turkmenistan (100,000). World consumption fell 3.5 million bales in 1998/99 to 85 million (figure 7). This was a 3.9-percent decline, the largest since 1974's 5.4-percent decrease. Outside the United States, foreign consumption fell 3.3 percent, the largest since 1961. By country, the largest drop came in China, 1 million bales lower than the year before, a decline slightly larger than the 950,000- bale drop in U.S. mill consumption. Other large declines were 545,000 bales in Turkey, 300,000 bales in Russia, and about 200,000 bales each in the EU, Pakistan, and India. Countries where consumption rose from a year earlier included Indonesia (up 300,000 bales), Mexico (200,000), Korea (122,000), Hong Kong (51,000), and Bangladesh (50,000). World trade fell 3.2 million bales to 23.5 million in 1998/99 as world consumption shrank. U.S. exports also shrank 3.2 million bales as the large drop in U.S. yields, exhaustion of "Step 2" funding, reduced sales prospects in importing countries, and changes in Chinese trade policy drove U.S. exports to their lowest since 1985 (figure 8). Uzbekistan, Pakistan, Argentina, and Chad also exported less cotton in 1998/99, with declines ranging from 770,000 bales to 175,000 bales (appendix table 17). China's exports rose 647,000 bales and Turkey, Turkmenistan, and Syria each exported 250,000 bales more than the year before. Exports by Australia and Egypt were also higher. Import declines were led by China (down 1.5 million bales), Turkey (530,000) , Brazil (519,000), Russia (375,000), and the EU (366,000) (appendix table 18). Surprisingly, the largest import gains from a year earlier in 1998/99 were all by traditional exporters: Pakistan 780,000 bales higher, the U.S. 430,000 higher, and India, 305,000 higher. Larger imports were also reported by several traditional importers, including Indonesia (up 277,000 bales), Taiwan (166,000), and South Korea (150,000). The changes in imports highlight the difficulties facing the United States on world markets in 1998/99. At the same time that an extraordinary drop in U.S. yields reduced the availability of U.S. cotton, China reversed its trade policy of the preceding half-decade, and reverberations of the financial shocks that began in Southeast Asia during 1997 continued to affect importing markets. While Asia's financial crisis did far less economic damage than many feared--highlighted by rising imports during 1998/99 by Indonesia, Taiwan, and South Korea--late 1998 saw Russia and Brazil suffer severe currency devaluations and economic contractions. As a result, textile industries and cotton imports contracted, but indirect effects showed up in lower imports by Turkey and the EU. Turkey's textile industry had benefited substantially from "suitcase trade" with Russia for several years as small traders shipped clothing--and many other goods--in response to an overvalued Russian exchange rate. With the end of this opportunity, Turkey's textile production shrank, and traders sought new markets in the EU. EU textile production, already depressed by increased competition from Asia following substantial devaluations in Southeast, East, and South Asia during the early phase of the Asian crisis, was forced to contract still further. China's change in trade policy also had serious consequences for U.S. exports, as well as world prices. On the one hand, the 1.5- million-bale decline in China's imports from all sources represented a direct loss in U.S. exports to what was the United States' second largest market during 1997/98 and its largest in 1996/97. During 1998/99, U.S. exports to China fell 666,000 bales from the year before. Moreover, indirect competition in other markets increased as other exporters also responded to reduced Chinese purchases, and as China returned to the world market as a competitor. While the increase in China's exports was not extraordinarily large by historical standards, and its decrease in imports was not its largest ever, on a net basis, the impact of these two changes was extraordinary. China's year-to-year decline in imports and increase in exports in 1998/99 summed to a 2.2- million-bale shift in net trade. This was the largest decline in China's net imports in a single year, and was equivalent to 9 percent of world trade in 1998/99. Thus, the United States and other exporters faced not only the difficulty of competing during a year when world trade shrank to its lowest since 1970, but also faced the difficulty of accommodating the efforts of the world's largest cotton producer and consumer to correct its supply imbalances through trade policy. While the level of world trade is expected to rebound in 1999/2000, China's supply imbalance could be expected to continue to haunt world markets. World Production and Consumption Higher in 1999/2000 With a rebounding U.S. crop and foreign production about unchanged from the year before, world cotton production is estimated at 87.3 million bales in 1999/2000, up 2.8 million bales, or 3.3 percent, from the year before. Beneficial weather across a substantial part of Asia contributed to a 4-percent increase in foreign production outside of China, despite a 19-percent decline in the inflation-adjusted A-Index during the previous marketing year. Central Asia is expected to harvest 880,000 bales of additional cotton in 1999/2000 compared with the previous year. This is by far the biggest upward spike in Central Asia's production since the economic upheaval associated with the breakup of the Soviet Union drove the region's production into a downward spiral. Pakistan's crop is expected to jump 1.5 million bales to its third largest ever, and excellent yields in the Punjab are expected to sustain India's crop close to its 1998/99 level, for its fourth largest crop ever. In Africa's Franc Zone, untimely rains have trimmed forecasts recently, but the 1999/2000 crop is expected to be 110,000 bales above the year before, and the second biggest ever. In the Southern Hemisphere, Australia's crop is expected to fall only 100,000 bales in 1999/2000--for the second largest crop ever--and no change is expected from the year before in Brazil's and Argentina's output. China's production is forecast 1.7 million bales lower than the year before in 1999/2000, but early season procurement prices there have fallen around 30 percent from the preceding year's official levels as China begins its first year without guaranteed prices or procurement. In anticipation of this change, cotton area in China fell nearly 13 percent, but with area largely unchanged in high-yielding Xinjiang province, China's production is expected to be only 8 percent lower than the year before, at 19 million bales. Smaller crops area also expected in Mexico (down 400,000 bales), Syria (139,000), Israel (110,000), Zimbabwe (68,000), and Nigeria (50,000). World consumption is forecast to be slightly larger than production in 1999/2000, and, at 87.8 million bales, has matched the 2.8-million bale and 3.3 percent year-to-year gain achieved by world production. China's expected increase in use is the largest, at 700,000 bales, reflecting the improved competitiveness of cotton versus manmade fibers following price reform. India, Brazil, and Turkey are expected to increase their cotton consumption by 500,000, 350,000, and 345,000 bales, respectively, from the year before in a rebound from the effects of the Asian financial crisis. Smaller gains are expected for similar reasons in Indonesia (150,000 bales) and Russia (100,000), and Mexico and Pakistan are each expected to increase consumption by 300,000 bales. World cotton trade is expected to increase in 1999/2000 in line with the expected rebound in consumption. At 25.7 million bales, world exports are forecast 2.2 million bales or 9.3 percent higher than the year before. The United States accounts for much of the expected increase, with exports forecast 1.4 million bales higher than the year before, at 5.7 million bales. The second largest export gain is expected in China, where exports are projected to rise 519,000 bales to 1.2 million, the highest in a decade. Rebounding crops in Pakistan, Uzbekistan, and Turkmenistan are expected to boost exports from these countries by 440,000, 300,000, and 100,000 bales, respectively. Lower exports are expected from Turkey (225,000 bales), Argentina (150,000), and Syria and Australia (100,000 each). Expected import gains from the year before in 1999/2000 are led by those in Mexico (up 650,000 bales), Brazil (285,000), and Indonesia (200,000). Imports by Turkey, Russia, Korea, and Thailand are expected to increase between 100,000 and 200,000 bales. The largest declines in imports are all foreseen in exporting countries in 1999/2000: Pakistan (down 600,000 bales), the United States, (368,000), and China (109,000). World ending stocks are expected to fall slightly in 1999/2000, to 41.3 million bales. However, the 619,000-bale global decline is comprised of an expected decline of 2.5 million bales in China's ending stocks, and a 1.8-million-bale increase in ending stocks for the rest of the world. Outside of the United States (where a 661,000-bale increase is expected), the largest increases are foreseen in Pakistan (up 325,000 bales), Uzbekistan (255,000), Turkmenistan (200,000), and Brazil (150,000). India's stocks are expected to fall 250,000 bales, and Egypt's by 123,000 bales. As a share of global consumption, world ending stocks are expected to be their highest since the mid-1980s. Excluding China, the stocks-to-consumption ratio is expected to increase from 37 percent in 1998/99 to 38.5 percent in 1999/2000, comparable to levels seen in the early 1990s, but well below those of the mid-1980s. However, as of October 1999, the monthly average A-Index in nominal terms was well below that of 1992 and 1993. In inflation-adjusted real terms, the A-Index was even slightly below the lowest points of 1986, and perhaps the lowest ever. Adjusting the August 1986 A-Index of 37.2 cents to account for inflation (as measured by the U.S. Gross Domestic Product deflator), suggests a price in 1999 dollars of 52.7 cents, compared with the 47.5 cents that the A-Index averaged during October 1999. Adjusted for inflation, the A-Index has fallen about 60 percent from its most recent peak in May 1995. In 1986, prices were low in part due to a shift in the United States to a more market-oriented farm policy, a series of staggering increases in China's cotton production during the first half of the 1980s, and a legacy of poor competitive performance for cotton vis-a-vis manmade fibers. Within a few years of these extreme difficulties, cotton consumption was soaring as the rest of the world followed a U.S.-led global economic recovery, and lower cotton prices and changing consumer tastes drove cotton's share of textile fiber demand back upward. In 1999, world cotton markets face the uncertainty of China's shift to a more market-oriented farm policy under the cloud of beginning stocks in excess of 17 million bales. Markets have not forgotten how China's exports soared in the years following the last time its stocks surpassed 17 million bales. Also, while world economic growth in calendar 1999 has not fared as poorly as many had earlier feared, the average growth rate of 2.6 percent over 1997-99 is well below the 3.8-percent average of 1984-86. Even more distressing for cotton, in 1986, the lowest annual increase in world consumption during the preceding 4 years had been 2.7 percent, while in 1999, the highest rate in the preceding 4 years was 2.5 percent. Thus, past experience could lead current market participants to expect less favorable future prospects than they would have foreseen in the mid-1980s. Finally, the legacy of past monetary policy and price expectations in general is very different in 1999 from 1986, with prospective price increases for any commodity now much more difficult to pass through to consumers. Wheat, corn, and soybean prices are all in similar circumstances. Current prices for these commodities are once again approaching, or even undershooting, nominal lows last seen in the last half of 1986, and it is unlikely that cotton could stray from general price conditions for very long in such an environment. What remains to be seen is the response in the following year by producers and consumers of all these commodities after this period of remarkably low prices. Stable Foreign ELS Production, Lower Use Projected According to the International Cotton Advisory Committee (ICAC), 1999/2000 foreign ELS production is projected slightly above last season at 1.9 million bales (table D). While production in Egypt, Sudan, and a few other countries is expected to rise this season, offsetting declines are seen for India and China. With production about unchanged and beginning stocks nearly cut in half from a year ago, ELS consumption in foreign-producing countries in 1999/2000 is estimated to fall 17 percent to about 1.5 million bales. Declines in Egypt account for most of the decrease, although India's consumption is projected to rise slightly this season. Like consumption, ELS exports in foreign-producing countries are expected below a year ago. The ICAC estimates a decline of nearly 8 percent in 1999/2000, with Egypt's lower shipments more than offsetting modest increases projected for a few smaller producing countries. Despite reduced consumption and exports this season, stocks for these countries are projected to fall somewhat. For 2000/01, beginning stocks of foreign producers are projected under 600,000 bales, the lowest since 1995/96. With U.S. stocks of ELS cotton building significantly in 1999/2000, the United States will be in an advantageous position to supply any increases in demand that develop. U.S. Wool Situation and Outlook U.S. Wool Production and Mill Use Continue Decline U.S. sheep numbers and wool production declined for the ninth straight year in 1998. The U.S. farm price for shorn wool averaged 60 cents per pound (greasy) in 1998, 39 percent below a year earlier. Sheep numbers on July 1, 1999, were 5.4 million head, 4 percent below July 1998 and 44 percent below 1990. Shorn sheep during 1999 are expected to decline to 6.2 million head, compared with 6.4 million in 1998. Wool production is estimated at 26.1 million pounds, clean (49.2 million greasy) for the 1998 marketing year (table E). Lower sheep numbers in 1999 imply production this year could decline to near 25 million pounds, clean, 4 percent below 1998. Expected production this year would be the lowest on record and significantly below the early 1990s (figure 9). During 1998, U.S. raw wool imports totaled 70.5 million pounds, clean, 8 percent below a year earlier (table F). Despite lower production forecast this year, imports are expected to fall to 40 million pounds, clean, down 43 percent from 1998. Shipments through August 1999 reached only 29.5 million pounds, compared with 50.7 million in 1998. Imports of fine wool accounted for 52 percent of 1999 shipments through August. During 1998, imports of fine wool represented 65 percent of total shipments. Imports of unimproved and other grades not-finer-than 46's were 16 million pounds through August 1999 and totaled 25 million pounds in 1998. Shipments from New Zealand represented 80 percent of these coarser grades, while Australia accounted for 80 percent of the finer wools last year. With beginning stocks of 41 million pounds, total 1999 supplies are forecast at 136 million pounds, clean, 16 percent below last season. Raw wool mill consumption is forecast at 85 million pounds, nearly 30 million below 1998. Mill use during the first 9 months of 1999 totaled 63.7 million pounds. Apparel wool consumption, at 51.9 million pounds, was 36 percent below a year earlier, while carpet mill use was 124,000 pounds below 1998 at 11.8 million (table G). The woolen system used nearly 25 million pounds and the worsted system used 27 million during January- September 1999. Wool top production in the first 9 months of 1999 totaled 25 million pounds, compared with 39 million in 1998. Raw wool exports during the first 8 months of 1999 were 1.9 million pounds, clean, compared with 1.5 million in the corresponding period of 1998. Shorn wool shipments were 1.7 million pounds and not-shorn (pulled) exports were 162,053 pounds. Carbonized wool exports totaled 10,520 pounds through August, compared with 77,255 in 1998. The majority of shorn wool exports went to Germany, Mexico, and Italy. The majority of not- shorn wool went primarily to Canada. Exports of wool top during January-August 1999 were 4.3 million pounds, 9 percent below a year earlier. The average price received was $2.16 per pound, compared with $3.28 per pound in 1998. The value of wool top shipments totaled $9.3 million with the majority going to South Korea, Mexico, and Italy. Top imports were 1.0 million pounds, 47 percent below January-August 1998. Australia and Germany were the largest sources of top imports, accounting for 48 and 17 percent, respectively, of total shipments. U.S. prices for clean, mill-delivered territory raw wool have generally declined during the first 9 months of 1999. Finer grades 64's and 62's averaged $1.15 and $1.05 per pound, respectively, last January. By September 1999, prices of 64's averaged $1.05 per pound and the 62's averaged $0.92. The 60's and 58's have followed a similar trend, decreasing throughout the year. However, while the finer grades have declined about 10 percent since January, the coarser grades have declined 20 percent. The 60's and 58's fell to $0.79 per pound and $0.67, respectively, during September 1999, while the 56's had declined to $0.60 per pound and the 54's had fallen to $0.55. Current domestic prices of Australian raw wool of most grades are below prices established last January. However, prices for Australia's finest wool--the 80's--were $2.56 per pound in September 1999, above last January and a year earlier. For all other grades, current prices are below January 1999 levels. The 70's declined to $1.59 per pound in September. The 64's established a season low in September of $1.39 per pound. Grade 58's prices have declined from $1.45 per pound in January to $1.22 in September. World Wool Situation and Outlook Declining Wool Production, Weak Demand Highlight the 1998 Season World wool production continued the long-term downward trend in 1998 falling to 3.1 billion pounds, clean, the lowest in the 1990s. World sheep numbers declined by 4 million head in 1998 to 1 billion. With lower sheep numbers, production fell 2 percent below the 1997 season (table H). Lower production occurred in New Zealand, declining 26 million pounds, clean, in the New Independent States (NIS), declining 24 million pounds, and in Uruguay, declining 22 million pounds, from a year earlier. NIS production, at 163 million pounds during the 1998 season, was 64 percent below the 1992 output. However, production increased slightly in Australia, China, Argentina, and South Africa. World exports of raw wool, at 1.2 billion pounds, decreased nearly 13 percent from 1997 shipments. Australian wool exports declined 124 million pounds to 739 million. World wool consumption declined for the seventh consecutive year in 1998. Raw wool consumption, at 2.8 billion pounds, clean, was 7 percent below 1997. Although world production declined in 1998, world carryover supplies increased for the first time in the past 8 years (figure 10). World carryover at the end of the 1998 season is estimated at 560 million pounds, clean, 3 percent above a year earlier. The Wool International (WI) stockpile accounted for nearly 61 percent of world carryover supplies in 1998, compared with 67 percent in 1997. The WI stockpile is expected to be eliminated by the 2002 season. Weak mill consumption during 1998 was in part due to the continuing economic problems in Asia. Most Asian countries imported less raw wool than a year earlier. The largest importer, China, purchased 296 million pounds, greasy, of foreign raw wool in 1998, 14 percent below a year earlier. Raw wool demand was also lower in Japan, Taiwan, and South Korea. Imports by Western Europe were also down in 1998. Italy, the largest European importer of raw wool, purchased 273 million pounds in 1998, 13 percent below a year earlier. Wool shipments to the United Kingdom and France were also sharply lower in 1998. World wool prices continued to decline during the 1998 season. The Australian market indicator (a weighted average index of 15 categories) declined in October 1998 to A483 cents per kilogram, then ranged between A488 to A549 cents during November 1998 through July 1999. For the season, the Australian market indicator averaged A527 cents per kilogram, 2.0 percent below the 1997 marketing year. The New Zealand market indicator established a low of NZ384 cents per kilogram in February 1999 and averaged NZ407 cents for the season, the lowest on record. The South African market indicator followed a similar pattern as prices established a seasonal low in October 1998 of SA1,620 cents per kilogram and averaged SA1,791 cents for the season, 20 percent below the 1998 season. Manmade Fibers The U.S. manmade fiber industry during the first three quarters of 1999 declined slightly from a year earlier. Total shipments were 7.72 billion pounds, down 2.2 percent from a year earlier, with noncellulosic shipments, at 7.49 billion pounds, down 1 percent. Total manmade fiber production, 7.61 billion, was 3.9 percent below a year earlier. Noncellulosic fiber production, 7.39 billion, was 3.2 percent below last year (appendix table 42). Manmade fiber stocks in fiber producers' plants at the end of the third quarter were 0.69 billion pounds, 17 percent less than last year. Domestic shipments of noncellulosic fibers during January-September 1999 were 6.98 billion pounds, 0.6 percent below a year earlier. Noncellulosic filament fiber domestic shipments were 4.15 billion pounds, 0.4 percent less than last year. Filament fiber shipments totaled 1.63 billion pounds of olefin, 1.45 billion of nylon, and 1.08 billion of polyester. Noncellulosic staple domestic shipments were 2.83 billion pounds, down 0.9 percent from last year. Staple shipments were 1.56 billion pounds of polyester, 0.58 billion of nylon, 0.51 billion of olefin, and 0.18 billion of acrylic. According to the capacity survey by Fiber Economics, Inc. in May 1999 (the latest data available), total in-place capacity to produce manmade fibers in 1999 is 12 billion pounds, up 1.2 percent from 1998, (appendix table 42). Almost all of this increase resulted from the 1998-1999 capacity increase of olefin fibers. Olefin staple capacity rose 6 percent while filament increased 5 percent. In addition, most of the planned capacity growth for 2000 comes from olefin fibers, which is expected to grow 2.3 percent. Capacity for total manmade fibers in 2000 is expected to be 12.9 billion pounds, 0.9 percent above 1999. The carpet market continues to consume more fibers in facing and backing uses than any other fiber market (appendix table 43). During January-June 1999 (the latest data available), the carpet market used 2.03 billion pounds, 1.8 percent more than a year earlier. Noncellulosic carpet use accounted for more than 43 percent of total noncellulosic fiber domestic shipments. Nylon dominates the carpet market, constituting more than 51 percent of the total use of noncellulosic carpet fibers. Nylon staple carpet fibers were 96 percent of nylon staple domestic shipments, while nylon filament carpet fibers were 70 percent of nylon filament domestic shipments. Preliminary data for third-quarter 1999 indicate that about 1.6 billion pounds of nylon were used in carpets during January-September, 1.1 percent more than a year earlier. The use of olefin fibers in facing and backing during the first half of 1999 was 0.87 billion pounds, 4.5 percent above a year ago. Olefin fibers constitute 43 percent of the noncellulosic fibers used in carpets. Carpeting is the most important use of olefin fibers at 51 percent. Woven textile production remained the second largest outlet for noncellulosic fibers, taking 21 percent of the January-June 1999 domestic shipments. The woven market used 0.98 billion pounds, 7 percent less than a year ago. Two fibers made up almost 86 percent of this market; polyester accounted for 68 percent while olefin contributed 17 percent. Similarly, the knit market used 0.54 billion pounds, 8 percent below a year earlier. Shipments of manmade fibers to the knit market were 11 percent of total domestic shipments. Three fibers dominated the knit market: polyester, at 0.35 billion pounds, constituted 65 percent; nylon, at 0.11 billion pounds, provided 21 percent; and acrylic, at 0.07 billion, totaled 13 percent. The price of benzene (a precursor to many chemicals) during the first 4 months of 1999 ranged from 70 to 75 cents per gallon. Higher oil prices in May pulled the price up to 90 cents, but lower energy prices during the summer caused the price to decline to 80-85 cents. Subsequently, rising oil prices in the fall brought the price up to $1.05 per gallon in October (figure 11). The price of cyclohexane, a basic chemical used in nylon production, somewhat follows the price of benzene (table I). During January-April 1999 the price remained in the $1.00-$1.04 per gallon range. In May, cyclohexane rose to $1.16 per gallon and then fell back to $1.06-$1.10. With rising oil prices, however, it peaked for the year at $1.20 per gallon in October. The price of paraxylene in the first quarter was 14.5 cents per pound. In the second and third quarters, it rose to 15.75 and 18.25 cents per pound, respectively, reflecting higher raw material costs. The October price of 21 cents was the year's high. The price of polymer grade propylene, a precursor for acrylonitrile (a raw material for acrylic fibers) and olefin fibers, was unchanged for January-June 1999 at 12.25-12.50 cents per pound. There was ample supply and slow demand. Rising demand in the summer and fall pushed the price to 18.5 cents in October. The price of acrylonitrile, listed at 53 cents per pound all year, was reported to be discounted as much as 50 percent. The price of ethylene glycol (a raw material used to make polyester fibers) dropped from 20 cents per pound in January to 17 cents in April, reflecting more fiber imports. Later, with higher raw material costs, the price rose to 23-25 cents where it remained through October. The price of caprolactam (a raw material used to make nylon) was listed at 90-93 cents per pound all year, although heavy discounting was reported. Special Article Analyzing U.S. Cotton Acreage Response Under the 1996 Farm Act William Lin, Scott Sanford, and Robert Skinner1/ Abstract: This article presents results for cotton from a recent study on supply response. The own-price supply elasticity is estimated at 0.466 for U.S. upland cotton (a 0.466-percent increase in cotton plantings is associated with a 1-percent increase in the expected cotton farm price)--up 7.9 percent from 1986-90 when little planting flexibility existed. The cross- price elasticities, in all cases, show larger increases under the 1996 Act. The cross-price elasticity with respect to corn prices, for example, is estimated at -0.147, compared with -0.019 during 1986-90. The increase in 1999 cotton planting intentions over 1998 partly reflects the fact that the expected decline in competing crop prices more than offset the expected decline in the effective cotton price. An increase in the expected effective cotton price, along with 4- to 8-percent declines in most competing crop price expectations, points to a slight increase in 2000 cotton planting intentions. Keywords: Acreage response, acreage price elasticities, cotton, normal flex acreage (NFA) U.S. cotton planted acreage has shown an increasing trend since 1983 when it hit a record low of 7.9 million acres (figure A-1). Until the enactment of the 1996 Farm Act, year-to-year variations in acreage were affected not only by costs and returns of growing cotton relative to competing crops, but also by provisions of cotton programs, such as deficiency payments, annual set-aside and acreage reduction programs, paid land diversions, and the Conservation Reserve Program (CRP). However, since 1996, farmers have based their planting decisions primarily on market forces-- expected farm prices and net returns of cotton relative to competing crops.2/ The 1996 Farm Act decoupled program payments from planting decisions and removed concerns over base acreage protection, making fundamental changes towards full planting flexibility. A centralquestion resulting from this policy change is how responsive crop plantings are to movements in market prices under the 1996 Act compared with previous legislation. 1/ William Lin and Robert Skinner are agricultural economists with the Economic Research Service, and Scott Sanford is an economist with the Farm Service Agency, U.S. Department of Agriculture. The authors are listed in alphabetic order. 2/ Admittedly, the marketing loan program became an important factor affecting producers' cotton plantings in 1999, and is likely to be important, again, in influencing 2000 cotton plantings. This impact has been incorporated in the analysis and estimation presented in this article. Since the enactment of the 1996 Farm Act, cotton planted acreage declined 3 years in a row-- from 16.9 million acres in 1995 to 14.6 million in 1996, 13.9 million in 1997, and to 13.4 million in 1998 (USDA, 1998). Cotton acreage planted in 1999 is estimated at 14.6 million acres, an upturn from the previous 3 years. Increased planting flexibility and higher profitability of competing crops were two major factors that contributed to the decline in cotton acreage during 1996-98. Given nearly full planting flexibility, cotton farmers were offered incentives to switch away from cotton to other more profitable crops, such as corn and soybeans. For example, based on the ERS Agricultural Resource Management Study (ARMS) survey, cotton producers in the Southern Plains region averaged a net return of -$29.3 per acre in 1996, while corn producers in the Plains States (including South Dakota, Nebraska, Kansas, Colorado, and Texas) had an average net return of $158.3 per acre. In 1996, U.S. cotton planted acreage declined more than 2 million acres from 1995, primarily due to attractive net returns for corn plantings. Estimating Supply Response The limited time since the passage of the 1996 Act complicates the study of producers' acreage response. To increase the number of observations, State-level information on producers' planting decisions during 1991-95, when producers were granted limited planting flexibility under the 1990 Farm Act, was used. The time series (1991-95) and cross section (State within a production region) data were pooled to enhance the degrees of freedom. This study focuses on producers participating in 1991-95 cotton programs. The approach adopted in this study assumes that cotton producers maximize their expected net returns. Cotton producers' acreage response is estimated using normal flex acreage (NFA), which represents the majority of producers who made their planting decisions, at the margin, within the range of NFA.3/ The results are then extended to the whole farm that includes acreage impacts of a 1-percent change in the expected farm price on both NFA and the rest of the base acres. 3/ NFA refers to 15 percent of base acres where farmers were allowed to grow the base crop, other program crops, soybeans and other oilseeds, any other approved non-program crops, or leave the cropland idle without loss of base acreage, but received no deficiency payments. Cotton supply response is estimated for major production regions, including the Southern Plains and Southeast and Delta regions combined (figure A-2). Acreage price elasticities are approximated from the estimates developed in a Food and Agricultural Policy Research Institute (FAPRI) study for the Far West region, as reported by Adams, based on the relationship between ERS and FAPRI estimates for the neighboring region-- Southern Plains. Supply response on cotton NFA is estimated by pooling time-series (1991-95) with cross-section (State) data in each production region. The acreage response is specified in two models, where the dependent variable is specified differently. In Model 1, a lower-bound estimate, the dependent variable is specified as the percent of cotton NFA planted to cotton or alternative crops. In Model 2, an upper-bound estimate, the percent of the combined NFA and acreage reduction program (ARP) acreage planted to cotton or alternative crops is used as an alternative dependent variable to derive estimates (Westcott). 4/ 4/ Specifying the dependent variable in the acreage response equation as the percent of cotton NFA planted to cotton may lead to some measurement problems relative to the underlying acreage shifts. This results from changes in the acres covered by NFA for different ARP levels across years. These measurement problems introduce a downward bias into the estimated own-price coefficient resulting from interaction of NFA acreage with the year-specific ARP that reflects an inverse relationship between the expected price and the ARP level. That is, the acreage response reflects the effect of a change not only in the expected price, but also in ARP. A possible adjustment to the dependent variable to address this concern is to incorporate the ARP into the dependent variable by defining it as the percent of combined NFA plus ARP land that was planted to the base crop or alternatives. This alternative reduces the measurement's downward bias, but it does not fully eliminate it. However, it also adds a policy-related upward bias to the measurement of acreage shifts. Explanatory variables in both Models 1 and 2 include expected net returns for the program crop itself (cotton in this case) and competing crops, as well as a set of intercept dummies for States in the region. Expected net returns equal the expected price times the trend yield by State minus variable cash costs of production for the region. The expected price is the new-crop futures price at harvesttime in the month when planting decisions are made by producers. The new-crop futures price for cotton and competing crops has the following time dimensions: Cotton: December futures price at the New York Cotton Exchange in mid-March, current year. Corn: September futures price at the Chicago Board of Trade in mid-March, current year. Winter wheat: July futures price at the Kansas City Board of Trade in mid-October, previous year. Sorghum: the expected price follows the same percentage change in the expected farm price for corn. Soybeans: November futures price at the Chicago Board of Trade in mid-March, current year. The new-crop futures price is further adjusted by the expected basis (the previous 5-year average difference between futures prices and cash prices received by farmers in the month preceding the futures delivery month) to allow for price differentials across States, and to arrive at the farm-price equivalent. The trend yield is estimated using State data for 1975-95. The acreage response equations in terms of cotton NFA planted to cotton, other program crops, soybeans, other crops, and NFA idled are estimated by Seemingly Unrelated Regression (SUR) as a system (Zellner).5/ In addition, to the extent that it is appropriate, theoretical constraints--symmetry and linear homogeneity--are imposed in the SUR estimation procedures. 6/ 5/ Due to the small sample size, any gain in efficiency of the estimation by SUR, relative to ordinary least squares (OLS), is limited. Thus, in many cases, there are only very small changes in the regression coefficients, although t-ratios are increased in some cases. 6/ The symmetry restriction requires that cross-net return regression coefficients across the share equations be equal while the linear homogeneity constraint requires that the sum of all own- and cross-return regression coefficients in each of the equations be zero. The symmetry restriction reflects the notion that the cross-price elasticities are linked to the ratio of the acreage shares between two competing crops. The linear homogeneity constraint reflects the fact that the share of cotton NFA is homogenous of degree zero in prices since the same proportional change in net returns for the program crop and competing crops does not alter the share of cotton NFA planted to a specific crop. Estimated Acreage Price Elasticities Acreage price elasticities for cotton in the Southern Plains and Southeast and Delta regions are derived, in part, from regression coefficients in the estimated acreage response equations on NFA (table A-1). Acreage response on NFA associated with a 1-percent change in the expected price is first estimated, and then extended to the rest of the base acreage.7/ Thus, acreage price elasticities reported in this study reflect the whole farm acreage response, not just the response on NFA. Also, the midpoint average of the acreage price elasticities from Model 1 and Model 2 is used as the "best estimate." 7/ For details on steps to estimate the own- and cross-price elasticities, the reader is referred to an upcoming report (Lin, et al). The own-price supply elasticity of U.S. cotton is estimated at 0.466 under the 1996 Act, 16 percent higher than during 1991-95, and 8 percent higher than during 1986-90 (table A-2). The increase in cotton's own-price elasticity becomes much more pronounced at the regional level. In the Southern Plains, where over 40 percent of U.S. cotton is grown, the own-price elasticity is estimated at 0.480, 83 percent higher than during 1991-95, and 39 percent higher than during 1986-90 (table A-3). All cross-price elasticities estimated under the 1996 Act show a greater increase from previous legislation than cotton's own- price elasticity. At the national level, corn, wheat, sorghum, and soybeans are the primary competing crops for cotton. For example, a cross-price elasticity of -0.147 with respect to the corn price means that a 0.147-percent decline in U.S. cotton planted acreage is associated with a 1- percent increase in the expected corn price. Similarly, the - 0.09 cross-price elasticity with respect to the soybean price means that a 0.09-percent decline in U.S. cotton plantings is associated with a 1-percent increase in the expected soybean price. While the own-price elasticity shows an increase of 16.2 percent from the 1991-95 period, the increase in cross-price elasticities are much more pronounced. For example, the increase in cross-price elasticity with respect to the corn price is nearly tenfold. Much lower capital requirements for growing competing crops, such as corn and soybeans, entice cotton producers--as opposed to many other crop producers--to more readily switch their planting decisions. Also, producers continued to grow cotton to protect their base under previous legislation because cotton offered one of the largest deficiency payments on a per acre basis. Supply elasticities vary among major production regions. The own-price elasticity in the Southeast and Delta (0.435) is the lowest (table A-3), while that for the Far West region (0.570) is the highest. Relative to 1986-90, when little planting flexibility existed, the own-price elasticity shows the largest increase in the Southern Plains--39.1 percent. In contrast, the elasticity shows only a 6.1-percent increase in the Southeast and Delta region and an 8.9-percent increase in the Far West region. This suggests that cotton plantings under the 1996 Act associated with a decline in cotton prices would result in a gain in U.S. share of cotton acreage for the Southeast, Delta, and Far West regions at the expense of the Southern Plains. Cross-price elasticities at the regional level, in many cases, are greater than at the national level. For example, the cross- price elasticity with respect to corn prices is estimated at - 0.171 under the 1996 Act in the Southern Plains, compared with - 0.147 at the national level. Similarly, the cross-price elasticity with respect to soybean prices is estimated at -0.183 in the Southeast and Delta region, compared with the -0.090 at the national level. All cross-price elasticities have an expected negative sign. Factors Contributing to the Increase in 1999 Cotton Acreage The cotton acreage response model (based on acreage price elasticities estimated in this study) lends itself to use in acreage-forecasting applications. In particular, the model was implemented into USDA's short-term market analysis and long-term baseline projection activities about a year ago. In general, the model performed well in forecasting cotton acreage intentions for 1999. Planting intentions for 1999 cotton were estimated by USDA last March to total 13.94 million acres, up 0.73 million from 1998's planting intentions (USDA, 1999). Taken by itself, the lower cotton price expected by producers in 1999 (based on new-crop futures prices) would have suggested a reduction in cotton acreage. As of March 15, 1999, the December 1999 new-crop futures settled at 59.15 cents per pound on the New York Cotton Exchange, for a 56.65-cent farm price equivalent. The expected cotton farm price represented a 19.0-percent decline from the expected farm price based on December 1998 futures quotes a year earlier. However, assuming that producers expected a 7-cent-per- pound loan deficiency payment (LDP) or marketing loan gain (which lies between the 2-cent average LDP for 1997/98 and the 10-cent LDP for 1998/99) on their 1999 crop, cotton producers would have anticipated an effective producer price of 63.65 cents (the sum of the expected LDP or marketing loan gain of 7 cents and the expected farm price of 56.65 cents based on new-crop futures. Hence, the effective cotton producer price in 1999 would have only declined by 8.94 percent from the previous year's expected producer price. Given the estimated own-price elasticity of 0.466, this decline in the expected cotton price implies a 4.2- percent (or 0.55 million acres) decrease in cotton plantings. However, the decline in the expected prices of competing crops (again, based on new-crop futures prices) more than offset the effect on cotton plantings due to the decline in expected cotton prices. For cotton producers, the decline in the expected prices of competing crops (corn, wheat, sorghum, and soybeans) generated an increase of 0.88 million acres in cotton plantings. Combining own- and cross-price effects suggests a model forecast of 13.55 million acres for 1999 cotton planting intentions, compared with the 13.94 million estimated by USDA in the March 1999 intentions report. A large portion of the 0.39-million-acre deviation between the model forecast and March intentions can be attributed to the effect of non-price factors on 1999 cotton plantings. One of these non-price factors was concern over aflatoxin (a crop fungus) for corn crops in the South, which appeared to have caused farmers to switch a part of their corn acreage into cotton plantings. In addition, wet weather might have prevented some winter wheat acres from being planted as intended in the South, ranging from Arkansas to South Carolina (USDA, 1999). A part of this acreage not planted to winter wheat could have been switched to cotton plantings. Cotton Planting Intentions in 2000 Could Be Slightly Larger Than 1999 Intentions The above cotton acreage response model can be used to forecast cotton planting intentions for the year 2000, just as it was used last year for forecasting 1999 intentions. Despite a lower expected farm price for 2000 cotton (based on current new-crop futures prices), the effective producer price is still somewhat larger than last year's expectations based on the current marketing loan program differential. Also, expected prices for competing crops (also based on new-crop futures prices), as of November 11, 1999, show 4- to 8-percent declines from 1999 levels. Thus, an increase in the expected effective cotton price, along with declines in most competing crop price expectations, points to a slight increase in 2000 cotton planting intentions. The marketing loan program is the major factor contributing to larger effective producer prices for cotton in the year 2000. LDPs or marketing loan gains anticipated by producers for the 2000 crop could be higher than the 7 cents per pound expected for the 1999 crop. As of mid-November 1999, LDPs had risen to nearly 20 cents for 1999/2000. However, producers may not view this high LDP rate as sustainable, given historical patterns. In this analysis, we assume a 13- cent-per-pound LDP or marketing loan gain (which lies between the current 20-cent rate for 1999/2000 and the 10 cents estimated for 1998/99). At this expected LDP rate, producers could anticipate an effective cotton price of 67.13 cents per pound, which is the sum of the 13-cent estimated LDP and an expected cotton farm price of 54.13 cents (based on the December 2000 futures price of 56.63 cents per pound as of November 11, 1999). Thus, the expected cotton producer price of 67.13 cents per pound would be about 5.5 percent larger than the 63.65 cents in 1999. Given the own-price supply elasticity of 0.466, this means that cotton plantings would be up 2.5 percent (or 0.36 million acres) due to a rise in cotton's expected producer price. Declines in the expected price for competing crops would also contribute to an increase in 2000 cotton planting intentions. Based on current new-crop futures, price declines appear for wheat, corn, and sorghum--8.1 percent for wheat and 4.3 percent for both corn and sorghum. The combined effect of the decline in the expected price for these competing crops could add another 0.29 million acres to 2000 cotton plantings. In contrast, the soybean marketing loan would have an effect of lowering 2000 cotton acreage. Per unit revenue for soybeans--another competing crop--could exceed the $5.26 per bushel loan rate for the 1999 crop if producers choose to lock in their LDP rates or marketing loan gains at higher levels and sell their soybeans for a higher price at a later date after harvest within the "60-day" window (according to the CCC new procedures announced on October 28, 1999). This per unit revenue for soybeans could be about 4.6 percent higher than last year, reducing cotton acreage by 57,000 acres. As a result, the combined cross-price effect is to increase cotton acreage by 0.24 million acres. Overall, given market conditions as of November 11, 1999, the acreage response model presented in this study suggests that U.S. cotton planting intentions in the year 2000 could reach 14.54 (13.94 + 0.36 + 0.24) million acres. This would be about 0.60 million acres larger than the 13.94 million acres reported in the 1999 Prospective Plantings report. Actual cotton plantings in 2000, whether greater or smaller than the 14.6 million acres planted in 1999, will continue to be affected by changing market conditions between now and next spring, as well as by non-price factors. References Adams, Gary. "Acreage Response Under the 1996 FAIR Act?" Speech presented at ERS-USDA seminar series on Supply Response Under the 1996 Farm Act. June 24, 1996. Lin, William, Paul C. Westcott, Robert Skinner, Scott Sanford, and Daniel G. De La Torre Ugarte. "Supply Response Under the 1996 Farm Act and Implications for the U.S. Field Crops Sector." ERS-USDA technical bulletin (forthcoming). Westcott, Paul C. "Policy and Modeling Issues Affecting the Estimation of Supply Elasticities." Speech presented at ERS-USDA seminar on Supply Response Under the 1996 Farm Act. April 8, 1997. U.S. Department of Agriculture. Prospective Plantings, NASS- USDA, March 31, 1999. U.S. Department of Agriculture. Cotton and Wool Situation and Outlook Yearbook, CWS-1998, ERS, Nov. 1998. Zellner, A. "An Efficient Method of Estimating Seemingly Unrelated Regressions and Tests for Aggregation Bias," Journal of the American Statistical Association, June 1962, pp. 348-368. Special Article An Economic Analysis of U.S. Total Fiber Demand and Cotton Mill Demand Leslie A. Meyer 1/ Abstract: This article presents an economic analysis of factors that influence total U.S. fiber demand (cotton, wool, and manmade) and cotton mill demand. The study examines changes that have occurred in U.S. cotton consumption since the early 1960s and provides insights for future demand, particularly in the context of expanding textile trade. A system of simultaneous equations is developed to explain the structural relationships of U.S. per capita demand using annual data for 1962-97. The equations are used to project future levels of total fiber demand, cotton mill demand, and subsequently cotton fiber demand. Keywords: Cotton, fiber demand, mill demand, textile trade, per capita consumption. Cotton, wool, and manmade fibers are among the most important textile fibers in the world. Over the past decade, these three fibers collectively accounted for 98 percent of the world's annual textile fiber production, reaching about 100 billion pounds of fiber in 1997. Although demand for fibers by the textile and apparel industry has generally risen over time with population, changes in demand for specific fibers, such as cotton, are normally dictated by changes in fashion trends, product acceptance, and consumers' lifestyles. 1/ Agricultural economist, Economic Research Service, United States Department of Agriculture. This article examines the changes that have occurred in U.S. fiber demand--particularly for cotton--since the early 1960s. The analysis measures the effects of principal factors that help determine U.S. demand on a per capita basis. This study also provides insights for future cotton demand--by mills and end- users--particularly in the context of expanding trade associated with the liberalization of textile and apparel products. Background In this analysis, total fiber demand is defined as the sum of annual mill demand for cotton, wool, and manmade fibers plus the net textile trade balance (raw-fiber equivalent basis) in manufactured products for these fibers. On a per capita basis, U.S. total fiber demand doubled during 1962-97, ranging from a low of about 39 pounds in 1962 to more than 80 pounds in 1997, or an average increase of slightly over one pound per year (figure B-1). While the increase shows a general upward trend, per capita demand also tends to follow economic cycles. For example, contractions of the U.S. economy during 1974-75 and 1981-82 are clearly reflected in falling demand, while subsequent expansion has moved total per capita fiber demand to its highest level to date. Also illustrated in figure B-1 is cotton's contribution to total fiber demand. As a major raw material of the U.S. textile and apparel industry, cotton has seen its popularity decline and rebound since the early 1960s. In 1962, cotton accounted for 60 percent of total fiber demand. However, over the next 20 years, cotton's share was cut in half because of manmade fibers' popularity. Demand for cotton reversed its downward trend in the early 1980s when consumer preferences shifted back to natural fibers. Over the past 15 years, U.S. consumer demand for cotton products has risen a dramatic 20 pounds per person, to over 32 pounds in 1997. To illustrate this level of cotton demand in terms of apparel products, the 32 pounds is equivalent to each person in the United States purchasing one pair of jeans and one sport shirt every month or two pair of dress trousers and one dress shirt or blouse every month. While cotton has many uses other than clothing, these examples give some perspective to the magnitude of consumer end-use demand for cotton in the United States today. Cotton now accounts for 40 percent of U.S. total fiber demand, up from 35 percent just 10 years ago. Contributing to this increase are imported textile products that became more readily available as various bilateral agreements provided relatively inexpensive foreign-produced products to U.S. consumers. While bilateral textile and apparel agreements have been in existence to some extent for several decades, more recent liberalization in the 1990s under the North American Free Trade Agreement (NAFTA) and the World Trade Organization (WTO) has expanded the trend. And although U.S. cotton textile exports have also increased, imports have grown faster, as shown by the gap between cotton fiber demand and cotton mill demand (figure B-1). This divergence also represents cotton's textile trade deficit, which has nearly doubled over the last 10 years. Whereas per capita cotton fiber demand has surpassed the level reached during the early 1960s, cotton mill demand remains a few pounds below this earlier period. Model Specification The models developed for this analysis are based on the assumption outlined in Studenmund (1997) that most econometric applications are by nature inherently interdependent. Applying the concept in this analysis means that total fiber demand and cotton demand are interrelated and each influences the quantity demanded of the other. Following this econometric argument, a system of simultaneous equations was developed to capture this interdependence. To this end, structural relationships among the three demand equations were first determined and are listed in their functional form below: 1) Total Fiber Demand = f1 (Economic Activity, z1, 1) 2) Cotton Mill Demand = f2 (Total Fiber Demand, Textile Deficit, z2, 2) 3) Cotton Fiber Demand = Cotton Mill Demand + Textile Deficit Similar past analyses by Evans (1977) and Sanford (1988) found that total fiber demand is influenced by economic activity as well as other exogenous variables (z1), including fiber prices. These studies used ordinary least squares (OLS) to estimate total fiber demand, as is achieved in this analysis. Evans and Sanford continued by estimating cotton mill demand separately using OLS, with total fiber demand given. However, expanding on these previous works and taking into account the interdependence discussed earlier, a system of equations was established in this study to determine cotton mill demand, which is influenced not only by total fiber demand, but also the cotton textile deficit. Other predetermined variables (z2), like substitutes, also affect demand. As a result, the first two equations listed above are solved simultaneously as a system. The use of OLS in simultaneous systems produces coefficient estimates that are biased. Therefore, a two-stage least squares approach is employed--assuming interaction via the error terms-- with total fiber demand and cotton mill demand jointly determined. The third equation--cotton fiber demand--is solved outside the system as an identity once the cotton mill demand equation is estimated. Although the cotton demand equations will be the main focus in this article, a brief discussion of the following total fiber demand equation is in order: Ln (TFD) = a + b Ln (DPI) - c Ln (Fiber Price) + d NAFTA + 1 where Ln represents natural logs and the parameters a, b, c, and d are to be estimated. The dependent variable (TFD) is U.S. total fiber demand for cotton, wool, and manmade fibers and is reported in pounds per capita. Total fiber demand is hypothesized to be influenced by income, prices, and other exogenous variables. As previously mentioned, total fiber demand tends to emulate the general economy and is positively related to demand. This is captured by the variable DPI, which is the per capita real disposable personal income. If the economy is growing, consumers are expected to have a larger disposable income. And as a result, U.S. consumers tend to use this purchasing power, raising demand for various products like clothing. In addition, economic theory suggests that fiber prices are inversely related to demand. In this study, nominal prices were used as a result of past research that found analysis of "real" prices inadequate. Consequently, nominal cotton mill prices lagged one year are used as a proxy for fiber prices, as cotton accounts for an increasing share of total fiber demand in the United States. The final variable in this equation takes into account the positive effect that NAFTA has had on per capita fiber demand. This variable is a dummy variable equal to one beginning in 1994 (NAFTA's inception) and zero in prior years. The cotton demand equation is represented by the following functional form: Ln (CMD) = a + b Ln (TFD) - c Ln (Ratio) - d Ln (Deficit) + e Ln (1 + TradeLib) + 2 where once again Ln represents natural logs and the parameters a, b, c, d, and e are to be estimated. The dependent variable (CMD) is U.S. cotton mill demand and is reported in pounds per capita. Cotton mill demand is hypothesized to be influenced mainly by three variables: total fiber demand, the ratio of the cotton mill price lagged one year divided by the polyester staple mill price lagged one year, and the per capita cotton textile deficit. The Ratio variable represents the substitutability of fibers, while the Deficit variable accounts for the net trade balance of cotton textiles entering the United States. These will be discussed in more detail in the next section. The final variable included in the model specification for cotton mill demand attempts to capture the effects of trade liberalization. TradeLib is a trend variable (1984=1) multiplied by a dummy that is equal to one beginning in 1984 and zero in previous years. This variable is constructed so that the change in value becomes smaller in subsequent years. It is theorized that cotton mill demand has been positively affected by the opening of foreign markets to U.S. products. The rise in textile exports since the early 1980s can be attributed to trade agreements like the Caribbean Basin Initiative. However, the positive effects are expected to diminish over time as more of these products are further processed in the foreign country and returned to the United States as imports, contributing to the cotton textile deficit. Liberalization of textiles under the WTO will also reduce benefits further as additional foreign products are allowed to compete in the U.S. market in the future. BEGIN BOX The Estimated Regression Equations Ln (TFD) = 2.121 + 0.980 Ln (DPI) - 0.178 Ln (Fiber Price) + 0.128 (NAFTA) (0.080) (0.038) (0.033) t = 12.21 t = -4.68 t = 3.85 Adjusted R-squared = 0.9096 Standard error of the estimate = 0.0559 Durbin-Watson statistic = 1.253 Degrees of freedom = 32 and Ln (CMD) = -2.448 + 1.315 Ln (TFD) - 0.494 Ln (Ratio) (0.186) (0.035) t = 7.05 t = -14.31 - 0.252 Ln (Deficit) + 0.136 Ln (1 + TradeLib) (0.024) (0.025) t = -10.68 t = 5.38 Adjusted R-squared = 0.9312 Standard error of the estimate = 0.0592 Durbin-Watson statistic = 1.672 Degrees of freedom = 31 END BOX Model Results The U.S. per capita cotton mill demand model was estimated in a system of equations using annual data for calendar years 1962 through 1997 As the box above shows, over 93 percent of the variation in (the log of) annual per capita cotton mill demand is explained by the equation. All variables are significant at the 1-percent level, with the standard errors and t-statistics shown below each coefficient. Also, the Durbin-Watson statistic suggests no autocorrelation problem. In addition, because the equation is in log form, elasticity measures are captured by the estimated coefficients. Cotton mill demand (CMD) is positively related to total fiber demand and the TradeLib variable, and, as expected, negatively related to the price ratio and the textile deficit. The price ratio is specified to capture the competitiveness of cotton with respect to polyester. An increase (decrease) in the ratio indicates that cotton is becoming more (less) expensive relative to polyester. The negative sign of the coefficient indicates that if cotton becomes relatively higher priced, fiber substitution may occur and less cotton is likely to be used by mills. Based upon the price ratio coefficient, a 10-percent increase in the ratio would be expected to lower per capita cotton mill demand by about 5 percent. Similarly, a 10-percent rise in the cotton textile deficit would be expected to reduce per capita mill demand by about 2.5 percent, all other things being equal. Furthermore, current levels suggest that a 10-percent increase in the deficit equates to 1.5 pounds per capita and a 2.5-percent decline in cotton mill demand is approximately half a pound. Therefore, given the specified structural form of the estimated system of equations, the 1-pound increase in per capita total fiber demand is expected to raise cotton mill demand by approximately one-third of a pound. In sum, the net effect of a 10-percent rise in the cotton textile deficit is a decrease in cotton mill demand of approximately 0.2 pounds per capita or about 1 percent. Model Performance Figure B-2 illustrates the estimated U.S. per capita cotton mill demand along with the actual values for 1962 through 1997. The derived values from the system estimation track in-sample demand fairly well, especially the rebound that occurred in the early 1980s. Most differences between the actual per capita mill demand and the model's estimates are 1 pound or less. The largest difference occurred in 1966 when the model underestimated actual cotton mill demand by 2.2 pounds per capita. However, the 1966 underestimation can be partially attributed to the dramatic jump in the cotton textile deficit, a 70-percent increase over the previous year. While a change of this magnitude would be expected to reduce mill demand, an overabundance of U.S. cotton at relatively inexpensive prices kept per capita mill demand from falling that year. Mean absolute errors and mean absolute percentage errors were calculated for the estimation period. The mean absolute error was determined to be approximately three-quarters of a pound, while the mean absolute percentage error was 4.5 percent. In addition, these measurement errors were also determined for the rebound period (1982-97) with similar results. These statistical measures indicate the good performance and fit of the cotton mill demand model. Sensitivity Analysis A sensitivity analysis was then conducted to capture the effects of changing total fiber demand, textile trade deficits, and price ratios. Several features presented in this article illustrate the effects on per capita cotton mill demand at different textile trade deficits as well as at various total fiber demand levels. Specifically, per capita mill demand curves are determined for assorted lagged price ratios, with other exogenous variables held constant. Figure B-3 shows total fiber demand of 80 pounds per capita and is illustrated at two cotton textile deficit levels (10 and 15 pounds). For example, if the price ratio were equal to one-- meaning cotton and polyester prices were the same--cotton mill demand would be expected to total 22.7 pounds per capita with a textile deficit of 10 pounds or 20.5 pounds if the deficit were 15 pounds. On the other hand, if the cotton price were 25 percent higher than polyester, cotton mill demand would be expected to fall to 20.3 and 18.3 pounds, respectively. Similar differences along these curves would be noted at other price ratios. Accordingly, adding the appropriate per capita deficit to the mill demand estimate produces a total cotton fiber demand estimate as highlighted earlier. Another feature of this analysis illustrates the effect of differing total fiber demand levels on cotton mill demand. Figure B-4 shows total fiber demand at 80, 85, and 90 pounds, while holding the cotton textile deficit constant at 15 pounds per capita. If the price ratio were equal to one using these examples, cotton mill demand would be expected to range between 20.5 and 23.9 pounds per capita based on the estimated equation. Similar ranges along these demand curves would be noticed at various price ratio scenarios. Out-of-Sample Forecast Since this analysis uses data through calendar year 1997, the effective performance of the cotton mill demand model can be evaluated by comparing the out-of-sample estimate for 1998 given actual 1998 data. In addition, estimates for U.S. total fiber demand, cotton mill demand, as well as cotton fiber demand are made using the latest available data for calendar years 1999 and 2000. For 1998, total demand for cotton, wool, and manmade fibers was 83 pounds per capita and the cotton textile deficit equaled 15 pounds. Using these numbers along with the cotton/polyester price ratio (1.19) and the trend variable, the cotton mill demand equation provided a per capita estimate of 19.5 pounds for 1998. A 90-percent confidence interval for this estimate would range from 17.6 to 21.6 pounds per capita. Examination of the actual data indicated that 1998 U.S. cotton mill demand exceeded 19.3 pounds per capita, slightly below the 19.5-pound estimate. Subsequently, the addition of the textile deficit to the mill demand estimate yielded a 1998 U.S. cotton fiber demand total of about 34.4 pounds per capita, the highest since 1943. Based on the latest available data, the tentative estimates for calendar 1999 indicate a slight increase in total fiber demand and cotton fiber demand, but a decrease in the cotton mill demand. Based on the estimated equation for total fiber demand, a figure of 84 pounds was calculated. Inserting this 1-pound gain in total fiber demand and an estimated 1-pound increase in the cotton textile trade deficit into the cotton mill demand equation yields an estimate of only 18.8 pounds per capita for 1999 (see figure B-4). Based on population estimates, this projection would cause U.S. cotton mill demand to fall to approximately 5 billion pounds this calendar year, the lowest since 1993. However, total cotton fiber demand would continue to rise to 34.8 pounds per capita. For calendar year 2000, U.S. per capita total fiber demand is projected to climb above 85 pounds to a record high, with cotton mill demand perhaps returning close to the 1998 level and cotton fiber demand moving closer to the 1943 level of 36.6 pounds per capita. Assumptions about the cotton textile deficit will play a vital role in the outcome, however. On the one hand, a decline in the U.S. dollar relative to foreign currencies may reduce the recent double-digit gains in cotton textile imports. On the other hand, the continued liberalization of cotton textiles and apparel--and the associated U.S. industry adjustments in preparation for these changes--may keep the import pace on a similar track as the last 2 years, resulting in a rising per capita cotton textile deficit. Because of this uncertainty, a table was developed to provide insights into various scenarios that could develop for 2000 and beyond. Holding the trend variable (TradeLib) constant at the 2000 level, estimates for future demand can be determined using the model for various levels of total fiber demand, cotton textile deficits, and lagged cotton/polyester price ratios. Table B-1 presents these per capita mill demand estimates at three selected levels of total fiber demand and cotton textile trade deficits. For example, if the lagged cotton/polyester price ratio equaled 1.20 for 2000 and total fiber demand totaled 85 pounds per capita with a cotton textile deficit of 17.5 pounds, cotton mill demand could be expected to reach 19.6 pounds per person based on the model presented in this analysis. On the other hand, if total fiber demand increased to 90 pounds, the model indicates that per capita cotton mill demand would reach 21.2 pounds. While there are a wide range of estimates presented in the table, cotton/polyester price ratios have only varied from 0.9 to 1.3 during the last 15 years and are likely to remain in this range in the near future. Consequently, the likely outcome for U.S. per capita cotton mill demand in 2000 and beyond would result in a much narrower range than what the table presents. Conclusions This analysis examined the changes that have occurred in U.S. fiber consumption since the early 1960s and provides some insight for future demand. The effects of factors that determine both U.S. per capita total fiber demand and U.S. per capita cotton mill demand were measured using ordinary least squares and two- stage least squares equations. Given the results of ongoing trade liberalization, U.S. cotton textile trade has expanded substantially over the past decade and will continue to play a major role in the quantity of cotton demanded by U.S. consumers. Estimates of U.S. per capita total fiber demand and per capita cotton mill demand were made for calendar years 1998, 1999, and 2000. Subsequently, U.S. per capita cotton fiber demand was then estimated. For 1998, actual cotton mill demand confirmed the effectiveness of the model presented. The initial analysis displayed here also provides insights into future alternative cotton demand scenarios in light of global liberalization of the textile industry and furnishes a reference point for further study concerning U.S. fiber and textile demand in the new millennium. References Barlowe, R.G. and J.V. Lawler. "The Impact of Cotton Textile Imports on the Domestic Market," Cotton and Wool Situation, CWS- 5, May 1976, pp. 24-31. Evans, S. "Factors Affecting Domestic Mill Demand for Cotton and Apparel Wool," Cotton and Wool Situation, CWS-12, September 1977, pp. 25-28. Sanford, S. "Factors Influencing Total Fiber Consumption and Domestic Mill Demand for Cotton and Wool," Cotton and Wool Situation and Outlook, CWS-51, March 1988, pp. 24-28. Studenmund, A.H. Using Econometrics: A Practical Guide. New York: Addison-Wesley, Third Edition, 1997. U.S. Department of Agriculture. Economic Research Service. Cotton and Wool Situation and Outlook Yearbook, Various issues. U.S. Department of Commerce. Bureau of Economic Analysis and Bureau of the Census. Various reports. List of Tables Text Tables A. U.S. cotton supply and use, 1997/98-1999/00 B. World and U.S. cotton prices, August 1998 to present C. World cotton supply and use, 1997/98-1999/00 D. ELS cotton supply and use in foreign producing countries, 1997-2000 E. Wool supply and disappearance, clean content, 1995-99 F. U.S. imports of raw wool for consumption, clean content, 1995-99 G. U.S. mill consumption of raw wool, clean basis, quarterly, 1995-99 H. World wool supply and disappearance, 1991/92-1999/00 I. Reported prices of raw materials for noncellulosic fibers, 1999 Appendix Tables 1. U.S. cotton supply and use, 1960/61-1999/00 2. U.S. upland cotton supply and use, 1960/61-1999/00 3. U.S. ELS cotton supply and use, 1960/61-1999/00 4. Upland cotton: Planted acreage, by State, 1960/61-1999/00 5. Upland cotton: Harvested acreage, by State, 1960/61- 1999/00 6. Upland cotton: Lint yield, by State, 1960/61-1999/00 7. Upland cotton: Production, by State, 1960/61-1999/00 8. ELS cotton: Planted and harvested acreage, by State, 1960/61-1999/00 9. ELS cotton: Production and yield, by State, 1960/61- 1999/00 10. U.S. cotton supply and disappearance of all kinds, by months, 1995/96-1998/99 11. Upland cotton farm, spot, and mill prices, 1970/71-1998/99 12. Fiber prices: Landed Group B mill points, cotton prices, and manmade staple fiber prices at f.o.b. producing plants, actual and estimated raw-fiber-equivalent, 1960-99 13. Index of prices of selected growths and qualities of U.S. cotton c.i.f. Northern Europe, monthly, 1990/91-1999/00 14. Index of prices of selected growths and qualities of U.S. cotton, c.i.f. Northern Europe, annual, 1975/76-1998/99 15. World cotton supply and use, 1965/66-1999/00 16. Foreign cotton supply and use, 1965/66-1999/00 17. Cotton exports, major foreign exporters, 1965/66-1999/00 18. Cotton imports, major importers, 1965/66-1999/00 19. Former Soviet Union cotton supply and use, 1965/66-1999/00 20. Brazil cotton supply and use, 1965/66-1999/00 21. Turkey cotton supply and use, 1965/66-1999/00 22. China cotton supply and use, 1965/66-1999/00 23. India cotton supply and use, 1965/66-1999/00 24. Pakistan cotton supply and use, 1965/66-1999/00 25. U.S. fiber consumption: Total and per capita, by type of fiber, 1990-99 26. Per capita domestic cotton consumption, 1975-98 27. Cotton and manmade staple fibers: Mill consumption on the cotton spinning system, 1960-98 28. U.S. wool supply and use, 1975-99 29. U.S. imports of raw wool for consumption, clean yield, 1970- 98 30. U.S. raw wool imports by country of origin, clean yield, 1994-98 31. U.S. mill consumption of raw wool, scoured basis, annual, 1970-98 32. U.S. raw wool exports by country of destination, clean yield, 1995-98 33. U.S. trade in wool tops, 1995-98 34. Shorn wool prices: U.S. farm price, Australian offering prices, and graded territory shorn wool prices, 1978-98 35. U.S. consumption on the woolen system and worsted combing, annual, 1988-98 36. World wool supply and disappearance, 1990/91-1999/00 37. Sheep population, wool production, and wool exports, major producing foreign countries, 1992/93-1999/00 38. World wool trade by major importing and exporting countries, 1991/92-1998/99 39. Wool sales and government-owned stocks, major foreign exporters, 1988/89-1998/99 40. International wool prices, 1990/91-1999/00 41. U.S. mohair, clean, exports by country of destination, 1992- 98 42. Manmade fiber production and capacity, 1997-2000 43. Domestic shipments of manmade fibers by major category, 1997-99 44. World textile fiber production, 1980-98 45. Raw-fiber-equivalent of textile manufactures, 1960-99 46. Raw-cotton-equivalent of U.S. imports of cotton-containing textile manufactures, 1995-99 47. Raw-cotton-equivalent of U.S. exports of cotton-containing textile manufactures, 1995-99 48. Raw-linen-equivalent of U.S. imports of linen-containing textile manufactures, 1995-99 49. Raw-linen-equivalent of U.S. exports of linen-containing textile manufactures, 1995-99 50. Raw-wool-equivalent of U.S. imports of wool-containing textile manufactures, 1995-99 51. Raw-wool-equivalent of U.S. exports of wool-containing textile manufactures, 1995-99 52. Raw-silk-equivalent of U.S. imports of silk-containing textile manufactures, 1995-99 53. Raw-silk-equivalent of U.S. exports of silk-containing textile manufactures, 1995-99 54. Raw-manmade-equivalent of U.S. imports of manmade-containing textile manufactures, 1995-99 55. Raw-manmade-equivalent of U.S. exports of manmade-containing textile manufactures, 1995-99 END_OF_FILE