Delivered by: Dr. Lawrence H. Shaw, Executive Director.

Delivered to: 1995 Annual Conference of the International Textile Manufacturers Federation

Location: Prague, Czech Republic

Date: 17 October, 1995

Forecasting the cotton market is something that we do at ICAC nearly every day. There is hardly a day that goes by that there isn't some development causing a change in our assessment of cotton production, cotton consumption, imports or exports in some country of the world. We make the change. Fortunately we have computers to add up the totals. And we have a model, an econometric equation, to assess what the changes mean for cotton prices. Because it is prices that we are all interested in.

The cotton market situation is constantly changing as we learn more about crop outcomes, about world economic conditions and actual production of yarn. When we make a forecast of prices, it is always based on what we know at a point in time. Most of our errors in forecasting prices are not due to our model, our technique; they are due to corrections in our assessments of the basic fundamentals of demand and supply.

Last year in Brazil I forecast the average value of the Cotlook A Index, the recognized measure of world cotton prices computed by Cotlook Limited a firm in the UK, at 77 US cents per pound of cotton lint, up from 71 cents the previous season. By January 1995, the forecast had risen to 88 cents, due to a half million ton drop in estimates of world cotton production and a doubling of our estimate (it should have been a trebling) of China (Mainland)'s imports of cotton. By March the estimate was 94 cents. The average Cotlook A Index for the 1994/95 season, adjusted by ICAC to reflect the absence of quotes in the final 2 « months of the season, was 94 cents. Our model's estimate based on the latest figures estimates 95 cents, an error of one cent, once we knew all the facts.

As you may know, the ICAC Secretariat has been making regular forecasts of world cotton prices since 1987. Few organizations forecast cotton prices. Cotlook Limited wouldn't dream of forecasting its own index. The US Department of Agriculture is prevented by law from making cotton price forecasts. They make forecasts for every other commodity, but someone in the US Congress a long time ago thought that price forecasts were disruptive to the market. Our view is that the price forecast is already imbedded in assessments of supply and demand. All we are doing is making the market situation more transparent, an agreed upon notion of what an international, intergovernmental commodity organization ought to be doing.

Today we expect the Cotlook A Index for the 1995/96 season to average 86 cents per pound.

There are still uncertainties regarding production in many key countries; and there are still uncertainties about China's imports. The limit up and limit down moves in cotton futures are testament. We will continue to forecast cotton prices during the season, making these forecasts publicly available at least once a month, and the forecasts will change. But today what we know about cotton supply and demand and trading conditions says 86 cents for 1995/96.

The key variable in our model of cotton prices is cotton stocks. We express stocks as a ratio of the stocks at the end of the season to consumption during the season.

Because of the size of Chinese stocks and the fact that regardless of their size they are not generally available for purchase by other countries, we look at the stock situation outside of China (Mainland) and add another variable to account for the impact of Chinese stocks on the rest of the world, China (Mainland)'s net exports as a ratio to the rest of the world's consumption.

These two variables are the major determinants of cotton price variation. Before 1992, they were enough to explain approximately 90% of the movement in prices. With the breakup of the USSR, barter trade and price expectations began to influence prices and it was necessary to add some additional variables.

Changes in cotton stocks on a world level are the result of the difference between production and consumption. Let us look at what the ICAC Secretariat currently anticipates for world cotton consumption and production in this season.

First consumption.

Cotton consumption is determined by the demand for textile products. According to economic theory, the demand for textile products is determined by the size of the world economy and the prices of textile products relative to other goods and services.

At the outset, I would like to assure you that, despite the confusion of the constant barrage of economic ups and downs beamed to us daily by satellite, the world economy is strong. Growth is expected to be above average both this year and next.

In fact, our forecast for world GDP growth this year is now 3.8%, up slightly from the forecast of 3.7% made a year ago. Our forecast for 1996 is growth of 4.2%. Generally, normal long term growth is assumed to be 3%; thus expected growth this year and next is anticipated to be a third faster than the long term average.

Growth in 1994 was also greater than anticipated last year in Brazil, 3.7% as opposed to a forecast of 3%. The world grew at a faster rate, even though we continued to underestimate the decline in Eastern Europe and the former USSR by 50%, which was 9% instead of an expected 6%. Both industrial and developing countries grew more rapidly than expected.

Growth in the USA in 1994 was much as expected, around 4%; growth in the European Union was more than twice as fast as expected a year ago, 2.8% as opposed to a forecast of 1.3% even as late as the end of September 1994.

The Japanese economy grew little. The major change in the forecast is that the Japanese economy is also now not likely to grow significantly in 1995 and way less than average in 1996.

Growth in the world economy in 1994 led to substantial gains in textile fiber consumption at the end-consumer level, 3.4%, the highest percent increase since 1987.

The gains in 1995 will be somewhat more modest, around 2.7% as textile prices have risen relative to the price of other goods and we won't be getting the impact of a post-recession boost in textile consumption which characterized several economies in 1994. In 1996, textile fiber demand is expected to rise 3%.

As world income growth approaches the 3% average level in the period to 2005, annual gains of 2% are expected.

Out of this textile demand comes cotton demand. As my colleague Carlos Valderrama reported at the Fiber Session on Sunday, supply shortages, higher prices and reduced competitiveness led to a significant drop in cotton's share of market in 1994, from 48% to 46%. As a result, total world cotton demand is now likely to fall short of the 20 million tons foreseen last year for 1996. Cotton demand is however expected to rise from 18.5 million tons to 19 million tons this year, 2.5%; and to 19.4 million tons in 1996, a 2.1% gain.

Gains of 2% are expected in the period to 2005, with cotton consumption reaching 20.9 million tons in 2000 and 23.1 million tons in 2005. These longer term forecasts are somewhat lower than those made two years ago. World textile fiber demand is estimated in 2000 to be about 5% lower than foreseen earlier, while cotton demand is foreseen to be about 9% lower, due to a smaller total market and a reduced share.

China (Mainland) remains the world's largest user of cotton.

Cotton consumption has not grown over the last ten years and was in fact lower in 1994/95 than in 1984/85 at 4.3 million tons. Efforts are being made to reduce the number of spindles in China and to increase the use of chemical fibers as there has been a growing realization that Chinese cotton production will fall short of current usage for the foreseeable future. Chinese mill consumption of cotton may rise to 4.7 million tons by 2000.

Cotton consumption continues to grow rapidly in the USA.

Cotton use doubled from 1.2 million tons in 1984/85 to 2.4 million tons in 1994/95.There has been an enormous growth in the consumption of cotton products in this country in the last ten years, enough for imports and domestically produced goods to all prosper. In addition, investment and exchange rates have made it possible for the US textile industry to become a significant exporter as well. Currently, there is some question whether cotton use in the US textile industry will remain at high levels. Our assessment at ICAC is for continued growth through 2000, based on the high levels of investment in US textile mill capacity and continued growth in the US economy and US consumer demand for cotton products.

Cotton consumption has grown significantly in a number of other major cotton producers as well. Indian consumption has grown to 2.2 million tons.

Pakistan's growth in cotton use and as a supplier of cotton yarn was sharp until the last two years, when production problems curtailed supplies and reduced the economic advantage for Pakistani spinners who suddenly found themselves paying world prices for cotton. As production problems are gradually solved, Pakistan's cotton consumption is likely to begin rising again, though not at the rate of the beginning of the 1990s. Pakistan's use is expected to reach 1.6 million tons this season and rise to 1.9 million tons by 2000.

Turkey's textile industry has grown rapidly in recent years. While government policies in Turkey have tended to foster free trade, high cotton prices at the end of 1993 led to the imposition of export taxes on raw cotton to provide protected supplies for the domestic industry. Much of Turkey's textile production is exported to Europe. This trade will be facilitated by the expected accession of Turkey to a full customs union with the EU at the beginning of next year.

Brazil's consumption has risen from 700,000 tons at the beginning of the decade to 800,000 in 1994.

While consumption is expected to increase less rapidly this year, the result of austerity measures of the Brazilian government, consumption of one million tons is expected by 2000.

With the growth of cotton consumption in these countries, a large percentage of total cotton use now takes place in producing countries. 70% of world consumption is now used in the ten largest producing countries, up from 55% ten years ago.

EU cotton use has held up at 1.2 million tons and is expected to remain at this level in the next two seasons. Consumption in the EU declined somewhat in 1991 but has recovered the lost ground and is expected to remain stable in the remainder of the decade.

Japan and the so called Newly Industrializing Countries -- Hong Kong, China (Taiwan), the Republic of Korea and Singapore -- are expected to represent the same amount of cotton use in 1995/96 as the EU, 1.2 million tons. This market was 50% larger than Europe in the 1980s. High costs of production, especially in comparison with facilities owned by the same textile companies in Southeast Asia, are likely to lead to a continued decline to 1.1 million tons in 1996/97.

The Southeast Asia market is expected to surpass the Japan and NICs market in 1996/97. This market, taken as Indonesia, Malaysia, Philippines, Thailand and Vietnam, grew 5% last year and is expected to grow 6% both in 1995/96 and 1996/97. Together East Asia is expected to continue to consume about 2.3 million tons a year in the period to 2000.

What about cotton production, the other part of changes in cotton stocks?

World cotton production is currently estimated at 19.2 million tons, slightly above the 19 million tons expected to be consumed in the 1995/96 cotton season.

The balance is fragile, as the cotton futures market tells us. Further reduction in crop prospects will result in consumption exceeding production, an important psychological factor in the market. Next season, based on normal yield prospects and expanded area from above average prices, production should exceed expected consumption by perhaps a million tons.

The USA has become the world's largest producer of cotton once again.

It is expected to remain the world's largest producer in the period to 2000, as we currently see Chinese production remaining at something like 4.1 million tons as a result of Government policies. US production could rise from around 4.4 million tons this season to 4.8 million tons next, based on expanded area.

Indian production is expected to be below last year's production because of pest, disease and weather problems (too much rain in North India).

Pakistan is expected to make further progress in recovering its production, although it is not likely to regain the 1991 level until sometime after 2000.

Central Asian production is holding steady. Government plans to reduce cotton area, combined with difficulties in procuring the necessary inputs to boost yields is likely to hold back cotton production in the region in the rest of the century. Climate changes have also perhaps turned negative for Uzbekistan as the season is now cooler and shorter than in previous years, perhaps due to the changes in the Aral Sea.

Production in other countries has risen by a million tons in the last four years, following the declines due to low prices in 1992. Another million ton increase is likely in the period to 2000.

The changes which I have reviewed in consumption and production of cotton are likely to lead to little change in ending stocks this cotton season but increased stocks next season and in 1997/98.

The Cotlook A Index is forecast based on these changes, and an assumption that China (Mainland) will continue to buy between 400,000 and 600,000 tons of cotton from other countries in the rest of the decade, to be in a range of 75-79 cents per pound in the period to 2000.

The most likely reason for prices higher than these indicative forecasts would be production shortfalls in particular seasons due to insects or weather. The most likely reason for prices lower than the indications would be a reversal of China's policy to maintain a significant cotton stockpile as a tool in controlling the internal Chinese cotton market.

Cotton of course is not just cotton, though price premiums in the market may be narrowing over time as mills are able to produce higher quality yarns and fabrics with lower cost cotton. Although it is a statistical oddity because of the shortage of quotations, the Cotlook B Index of cottons for coarse count yarns has actually been higher than the A Index (there is still a discount for these cottons if the same countries are compared) . Quotations for fine count and higher than medium count cottons have are still more narrow than they have been historically and seem to be less related to supplies of these qualities than in the past.

One kind of cotton, that for producing extra fine cotton, will be in short supply this season, due to reduced production and stocks in major producing countries. The premium for Extra Fine cotton is currently quoted at around 175% of the Cotlook A Index and it is questionable how much will be available for export this season.

Production would be expected to expand next season, at least in the USA.

To sum up, production and consumption of cotton at the world level are likely to remain about equal this year, with little change in stocks. Prices are expected to about 9% lower than last season due to a decline in import demand from China (Mainland). Further production shortfalls could lead to higher prices, but further declines in Chinese buying in the face of high prices could temper the rise. Prices are expected to fall another 9% in 1996/97 and stay at these lower levels in the remainder of the decade.

This prepared text is dated September 22, 1995 and will be updated orally. Detailed statistics and charts referred to in this address are found in WORLD TEXTILE DEMAND and the OUTLOOK FOR COTTON SUPPLY IN 1995/96, reports available for sale from the Secretariat of the International Cotton Advisory Committee, 1629 K Street NW, Washington DC 20006 USA. Telephone 202-463-6660, Fax 202-463-6950, Internet Publications@icac.org. Regular monthly forecasts of the cotton situation are available by fax and on the Internet.