Delivered by: Dr. Terry Townsend, Statistician.
Delivered to: West Texas Flow/Marketing Meeting
Location: Lubbock, Texas
Date: 12 October, 1995
Summary
Slow growth in the world average cotton yield, combined with the policy of the Government of China (Mainland) to maintain stocks at more than 50% of domestic use, have boosted world and USA cotton prices since October 1993. World production is estimated at 19 million tons (88 million bales) in 1995/96 and 20.5 million (94 million bales) in 1996/97. While representing substantial increases over 1994/95, even the 1996/97 estimate is not record high, reflecting continuing difficulties being experienced in four of the five largest producing countries. The factors which led to record prices in 1994/95 will probably continue for several years. Consequently, the world cotton industry may be in the midst of a five-to-ten year period of higher-than-average prices caused by relatively slow growth in world yields, a situation reminiscent of the 1970s and early 1980s. A change in the China (Mainland) stock-holding policy could cause an abrupt change in the outlook for prices, just as changes in the US cotton program affected the world cotton market in 1986.
In 1995/96, world cotton production is rising and may exceed consumption by a modest amount, and production may exceed consumption by a larger amount in 1996/97. World consumption is also rising, despite weakness in Europe and some countries of East Asia. Strong world economic growth and new investment in spinning technology are pushing mill use of cotton and chemical fibers higher. Because of tight supplies and high prices, cotton is losing market share. Nevertheless, growth in cotton use continues, especially in the USA, India, Indonesia, Pakistan, Brazil, Mexico and Turkey, causing the world total to rise. With production exceeding consumption, world cotton stocks are expected to increase this season and next. Declines in the volume of world trade in cotton may occur during 1995/96 and 1996/97, primarily because of reduced imports by cotton producing countries and because of reduced consumption in Japan and Hong Kong.
Based on current estimates of world cotton supply and use, a statistical model of world cotton prices indicates that season averages of the Cotlook A Index will be 87 cents per pound in 1995/96 and 78 cents in 1996/97. The average Cotlook A Index during the 22 seasons since 1973/74 has been 73 cents per pound.
Net imports of cotton by China (Mainland) will be a key factor affecting the level of cotton prices in 1995/96 and 1996/97. World cotton prices were boosted in 1994/95 by inefficiency in the Chinese economy. China (Mainland) imports rose to 874,000 tons while exports fell to 45,000 tons in 1994/95, even though production matched consumption. The Chinese cotton economy suffered from a breakdown in the system of state procurement and distribution, combined with political uncertainty and widespread fear of inflation. The near-record level of imports allowed the government to divert domestic cotton into the state reserve, an inventory held for emergency use; the state reserve may now account for half of Chinese (Mainland) cotton stocks, or one-fifth of world stocks.
Imports by China may decline in 1995/96 and 1996/97, assuming no further increases in stocks are necessary, but imports will, nevertheless, probably remain substantial. Cotton production in China (Mainland) has been equal to or less than consumption in all but two seasons since 1984/85, and production is forecast to remain below consumption during 1995/96 and 1996/97 as well. Actual purchases by China (Mainland) will depend on the harvest size and government procurement results; if production is higher or consumption lower than currently estimated, or if a reduction in stocks is allowed, prices will be affected. A decline of 100,000 tons (460,000 bales) in net-imports by China (Mainland) is associated with a reduction in the season average Cotlook A Index of two cents per pound.
As is always the case, production prospects also affect the current outlook for prices. There is concern that the US harvest could be lower than currently estimated. However, estimates of production in Pakistan and India have recently been raised, and South America and Australia received some relief from drought in recent weeks. Each change of 100,000 tons in world production adds or subtracts approximately one cent per pound to estimates of the season average Cotlook A Index.
In the United States, cotton mill use is being boosted by a competitive, growing national economy, consumer preference in favor of heavy cotton products and rising textile exports. Higher mill use is keeping cotton stocks in the USA relatively tight, even though reduced imports by China (Mainland), Pakistan, India, Turkey and Brazil will probably lead to lower levels of US exports than in 1994/95. The rise in production in the Southeast and Mid-South is pushing an increased proportion of Texas cotton into exports.
World Average Yields Below 5-Decade Trend
Between 1950/51 and 1994/95, the average world cotton yield rose at an annual rate of 2%, or 8 kilograms per hectare (7 pounds per acre) per year. However, yields were below the 45-year regression line in the late 1970s and early 1980s as high cotton prices encouraged expansion into lower-yielding areas. Between 1983/84 and 1991/92, the world average yield rose from 450 kilograms per hectare to nearly 600 kilograms (402 pounds per acre to nearly 535 pounds) largely because of productivity gains in developing countries, especially China (Mainland) and Pakistan. However, world yields were below the regression line in 1992/93 and 1993/94 and just reached the regression line in 1994/95, despite exceptional weather, indicating that the most recent period of above-average gains in yields has passed; the world average yield is estimated at 550 kilograms (490 pounds per acre) in 1995/96, 25 kilograms per hectare lower than indicated by the regression line.
World cotton area ranged between 30 and 35 million hectares (74 and 86 million acres) in most years between 1950/51 and the present, with no apparent tendency to rise or fall. An extrapolation of the 44-year regression line through world yields indicates an average yield in 2000/01 of 620 kilograms per hectare (554 pounds per acre). If world area is 33 million hectares, production would be about 20.5 million tons (94 million bales) in 2000/01; if world area rises to 36 million hectares (89 million acres), production would climb to 22.3 million tons (102 million bales) five seasons from now.
Per capita world cotton consumption reached 3.6 kilograms (8.0 pounds) in 1987, but fell to 3.2 kilograms (7.1 pounds) in 1994. If per capita world cotton use is to return to the 1987 level and perhaps rise modestly to 3.7 kilograms (8.2 pounds), world use will need to rise to 23 million tons (106 million bales) by 2000/01. However, to achieve 23 million tons of cotton use with yields averaging 620 kilograms, world cotton area will have to rise two million hectares above the record set in 1984/85 to 37 million hectares (91 million acres).
US Yields Not Rising
Production in the USA is estimated at more than 4 million tons this season and 4.8 million tons in 1996/97 (19.5 and 22 million bales). Drawing on experience from the first half of the 1980s, analysts often project a trend increase in US cotton yields of between 2 and 5 kilograms per hectare (2 to 5 pounds per acre) per year. However, the US average yield reached 791 kilograms (706 pounds) in 1987/88 and was only 3 kilograms per hectare (5 pounds per acre) higher in 1994/95, seven seasons later. Because of changes in economic conditions in the USA, the 1990s may be similar to the 1960s and 1970s when USA cotton yields were flat, despite incremental annual advances in production technology.
The US cotton program up to 1974 emphasized controls on supply through the use of acreage allotments which prevented production from shifting to areas of higher potential yield. After 1974, cotton area began to shift westward, but high cotton prices relative to land values encouraged increased production on land marginally suited to cotton.
It was not until the 1980s, after regional shifts in production had occurred and agricultural land prices had risen to reflect the rise in output prices, that cotton yields jumped; between 1980/81 and 1987/88, the average US yield rose by three-fourths. However, since the mid-1980s, US farmers have faced different economic incentives from those that stressed increases in yields per acre. Because of lower land values, government program benefits and improvements in general economic conditions, financial stress on the US farm sector lessened after the mid-1980s, allowing less intensive use of cotton land. During the entire 1980s, yields tended to rise in only two states, Texas and California, but a decline in yields was evident in Arizona.
In 1995/96, a US national average yield of less than 670 kilograms (600 pounds per acre) is forecast, and production may be similar to the 1994/95 output, despite a 21% increase in cotton area. In 1996/97, US production is likely to be record high. The acreage reduction percentage in the government program, assuming there is a government program, is likely to remain at zero in 1996/97. With prices likely to remain above average during 1995/96, US area will probably expand as producers gain experience with cotton and as additional cotton-production equipment is purchased in the Delta and the Southeast.
Pesticide Resistance Reduces Production in China (Mainland)
Reduced harvests caused by pests and disease in what were formerly the most important cotton producing provinces of China (Mainland) account for much of the rise in international cotton prices during the past two seasons. Production has been less than consumption in China (Mainland) during most years since 1984/85. Problems affecting cotton production in China include poor quality planting seeds and ineffective insect control caused by pesticide resistance, an increase in insect numbers because of intercropping wheat and cotton, a failure to till fields during the winter, and incorrect mixing and application of pesticide. Temperature and rainfall in 1994/95 were nearly ideal for cotton production; the Yellow River region, which often experiences drought, received abundant rainfall, and the Yangtze River region did not experience flooding, as is often the case. Weather in China in 1995 has not been as good, and lower production is likely.
Production in the provinces of Hebei, Henan and Shandong totaled 3 million tons in 1991/92 but is estimated at 1.7 million tons in 1994/95 and less than 1.5 million tons this season. The problems in Eastern China appear to be of a long term nature; yields in Shandong have been declining since 1987/88, a symptom of environmental degradation. Problems in Eastern China will necessitate shifts in the location of cotton production and changes in agronomic practices in the affected areas. Production in China (Mainland) is estimated at less than 4 million tons in 1995/96 because of reports of increased pest populations this year and reduced area because of high grain prices. Production may rise somewhat in 1996/97 because of increases in irrigated area and progress in controlling bollworms.
Indian Cotton Yields Remain Below World Average
Cotton area in India is expanding by an estimated 5% in 1995/96 in response to high prices, and a further gain may occur during 1996/97. However, production in India may remain below the 1992/93 record of 2.4 million tons because of disease in the Northern producing states which account for about one-third of national production.
Average yields in India remain lower than in the rest of the world despite efforts to improve seed varieties and to expand irrigated area. Progress was made during the 1980s, and the ratio of Indian yields to yields in the rest of the world rose from .35 in 1980/81 to .51 in 1989/90. However, since the 1980s, Indian yields have fallen back to less than half the average outside India, reflecting difficulties in recent years with pests. The identification of the leaf curl virus in Northern India during 1994 indicates that India may face additional problems in the 1990s in raising production.
Leaf Curl Virus Present in Pakistan
The 30% decline in production in Pakistan between 1991/92 and 1994/95 is mostly the result of a virus in combination with other pests for which no effective chemical defense has been found. In both India and Pakistan, efforts are underway to control the leaf curl virus and boost cotton production. The new efforts are in addition to ongoing programs to raise yields by providing better seeds and expanded irrigation. All the efforts may be partially successful as new varieties are introduced and as farmers gain more experience in dealing with disease and pests. However, even with successful control efforts, the leaf curl virus is still present and will reduce yields from what might otherwise have been.
Central Asian Production Still Declining
Production in Central Asia fell from 2.8 million tons (13 million bales) in 1988/89 to an estimated 1.9 million (9 million bales) in 1994/95 and appears headed still lower. Reductions in cotton area due to environmental concerns, civil war in Tajikistan and Azerbaijan, low producer prices when adjusted for inflation and exchange rates, and a failure to provide inputs and equipment to farms and farmers are the major reasons behind the declines in output. Central Asian production is not expected to rise during 1995/96 and 1996/97, despite above-average world cotton prices. Reflecting a tightening of Central Asian supplies over the last several years as production has fallen, the average quote in Cotton Outlook for Central Asian middling rose from 92% of the Cotlook A Index in 1992/93 to an average of 96% in 1994/95 and 98% during the first part of 1995/96.
Prices Likely To Be Variable and High
Consumer demand for cotton is growing, and with specific problems of disease, resistance to pests and difficult economic conditions in four of the five largest producing countries, prices will need to be above average during the 1990s to elicit the necessary supply.
Three consecutive years of disease and pest damage have reduced production in Pakistan and China (Mainland) by one-third and indicate that basic shifts in varieties and agronomic practices which may require several years to implement will be needed before yields in each country can return to 1991/92 levels. Further, there is potential for the leaf curl virus to spread from Pakistan and Northern India to Central Asia during the next several years. The introduction of the leaf curl virus into Central Asia would compound already difficult conditions.
The last twenty-one years have seen two distinct periods of cotton prices. Between 1973/74 and 1983/84, the Cotlook A Index averaged 76 cents per pound and was above the 22-season average of 73 cents eight times out of eleven years. However, in the ten seasons between 1984/85 and 1993/94, the Cotlook A Index averaged 67 cents a pound and was below 73 cents seven times. Higher cotton prices during the first period coincided with slow growth in world yields, rising prices for most commodities and increased imports by China (Mainland); cotton prices fell after 1984/85 because of record production in China (Mainland), changes in the US cotton program to expand exports, and reduced cotton consumption in Eastern Europe and the former USSR after 1989/90.
The development of the leaf curl virus in Pakistan, and the possibility of its spread to neighboring countries, combined with economic difficulties in Central Asia and bollworm resistance to pesticides in China (Mainland) indicate that increases in world consumption during the 1990s may need to come from expanded area and production in countries currently accounting for less than half of world output. In the past, above-average prices have been required to sustain increases in area and production in the Western Hemisphere and Africa. Consequently, it is likely that average prices in the 1990s will be higher than in the 1980s; depending on competing crop prices and rates of growth in world cotton consumption, prices in the 1990s could be above the 76-cent per pound average of the 1970s.
Tight Supplies Limiting US Disappearance in 1995/96
The US cotton situation presents a classic economic picture of high prices resulting when strong demand is constrained by reduced supply. Textile producers in the USA have become among the most competitive in the world and investment in open end and air jet spinning is continuing. US cotton use has risen during nine of the last ten seasons, and growth of 7% occurred in 1994/95. Growth in mill use in the USA is estimated at 3% in 1995/96 and 4% in 1996/97, when cotton supplies are expected to be greater. Mill use in the USA slowed during April, May, June and July 1995 as supplies of cotton tightened and because of slower economic growth. But expanded supplies with the 1995 harvest and faster economic growth since the second quarter of 1995 could lead to increased US mill use in 1995/96.
Textile mills in the USA are adding capacity; denim production capacity alone may be higher in 1995 by as much as 40,000 tons (200,000 bales). In a survey conducted by the International Textile Manufacturers Federation (ITMF), machinery manufacturers reported 1994 shipments to USA spinning mills of 118,000 open-end rotors and 30,000 spindles. At average rates of use, the rotors and spindles installed in 1994 could account for 125,000 tons of increased US mill use. Machinery shipment statistics for 1995 are not available, but the preliminary estimate of 1995 capital spending by USA companies in the textile mill, apparel and finished textiles category is 14% greater than 1994 capital expenditures by the same group.
The technique of forecasting US mill use by extrapolating from recent seasonally adjusted annual rates of monthly use based on surveys of conditions in the spinning industry has yielded inaccurate results. Between 1985/86 and 1994/95, early-season estimates of US mill use were too low every season; the average error of the August forecasts over the past ten seasons was 140,000 tons (625,000 bales).
Because of reduced imports by producing countries, world trade in cotton is likely to fall during 1995/96, and US exports are estimated at 1.6 million tons (7.3 million bales), down 450,000 tons (2.1 million bales) from 1994/95. The US share of world trade is estimated at 28% in 1995/96, compared with 33% in 1994/95 and an average share since the breakup of the USSR after 1990/91 of 26%. As is the case with mill use, exports of US cotton would be greater if supplies were ample; supply is constraining demand.
Total imports by China (Mainland) are expected to drop by about 300,000 tons (1.2 million bales), even if production in China is lower than in 1994/95. China (Mainland) rebuilt the state reserve in 1994/95 and will not need to do so again this season, and US exports to China (Mainland) may fall from 500,000 tons to 350,000.
With increased production in 1995/96, imports from all sources by Turkey, Pakistan, India, Brazil and Mexico are expected to drop by a combined 500,000 tons (2.2 million bales). Ending stocks were drawn tighter in a number of importing countries during 1994/95, and imports by Russia, Thailand, Indonesia, Italy and Poland are expected to increase by 225,000 tons (1 million bales) in 1995/96.
Since 1973/74, two variables have explained most of the change in the US share of world cotton exports. Increases in the season average of the Cotlook A Index signify a tightness in world stocks and a need for increased supplies from the USA; US exports tend to rise as a share of the world total when the cotton prices are high. The availability of cotton from the USA also affects exports; when the US exportable surplus meaning beginning stocks plus production minus mill use is high, exports tend to increase. In 1994/95, the high level of the Cotlook A Index and an increase in the US exportable surplus to 2.6 million tons (12 million bales) suggested that US exports might increase to 34% of world trade, and the actual US share was just 1.5 percentage points lower. In 1995/96, the season average Cotlook A Index is expected to be 7 cents per pound lower than last season, indicating an increase in supplies available outside China (Mainland), and the US exportable surplus is estimated at 2.2 million tons (10 million bales). Consequently, the US share of trade is likely to fall about 5 percentage points.
The ratio of ending stocks to use in the USA has been at .23 or lower during five of the last six seasons. Only in 1992/93, when barter sales from Central Asia, slow world economic growth and increased exports from China (Mainland) pushed US exports lower, did the US stocks-to-use ratio rise to .30. In 1994/95, because of increased mill use and imports of US cotton by China (Mainland), the US stock-to-use ratio fell to 13%. With barter sales from Central Asia declining, world economic growth remaining near 4% and China (Mainland) importing cotton in 1995/96 and 1996/97, US ending stocks are likely to remain near to, or lower than, 23% of use.
Stocks Tight in Texas/Oklahoma
The outlook for cotton from the Texas and Oklahoma region is for continued tightness of stocks, with exports constrained by a lack of supply. In 1994/95, mill use of Texas/Oklahoma cotton was about 570,000 tons (2.6 million bales), and exports reached 700,000 tons (3.2 million bales). Total disappearance of cotton from Texas and Oklahoma reached 1.26 million tons (5.8 million bales) in 1994/95 and exceeded production, causing ending stocks to fall. Mill use of US coarse cotton fell about 6% in 1994/95, even though total US mill use rose 7%, because increased production in the Southeast and the Delta is displacing cotton grown farther from textile mills in the East.
In 1995/96, reduced beginning stocks and production will limit the supply of Texas/Oklahoma cotton, and exports are expected to drop to about 500,000 tons (2.4 million bales) while mill use remains near the 1994/95 level of less than 600,000 tons (2.7 million bales). The ratio of 1994/95 ending stocks to use of cotton in Texas and Oklahoma was just .05, compared with a national stocks-to-use ratio of .13. In 1995/96, Texas/Oklahoma stocks may rise by a modest amount, reflecting reduced US exports, but the Texas/Oklahoma stocks-to-use ratio will probably remain tighter than the national average.
Reflecting an increase in world demand for coarse cotton over the past decade, the spread between quotes in Cotton Outlook for Memphis Territory and Orleans/Texas SLM is narrowing. The Orleans/Tx discount averaged 11% during 1990/91 and 9% in 1992/93, but narrowed to zero in 1993/94 and has averaged just 4% during the early part of 1995/96. The narrowing of the spread between Memphis and Orleans/Tx is part of a broader tendency for coarse cotton prices to rise relative to prices of medium cotton, as reflected in the narrowing of the spread between the Cotlook A and B Indexes since the mid-1980s. Increased use of open-end spinning equipment and stronger demand for denim and other coarse-cotton products is causing a shift in the structure of world cotton demand in favor of coarse cotton. Combined with a rise in average cotton prices during the rest of the 1990s, as compared to the 1980s and early 1990s, the reduced discount for Orleans/Tx cotton should provide ample incentive for increased Texas cotton production over the next several years.