ArticlePDF Available

Cotton : Market setting, trade policies, and issues

Authors:

Abstract and Figures

The value of world cotton production in 2000-01 has been estimated at about 20billion,downfrom20 billion, down from 35 billion in 1996-97 when cotton prices were 50 percent higher. Although cotton's share in world merchandise trade is insignificant (about 0.12 percent), it is very important to a number of developing countries. Cotton accounts for approximately 40 percent of total merchandise export earnings in Benin and Burkina Faso, and 30 percent in Chad, Mali, and Uzbekistan. Its contribution to GDP in these and other developing countries is substantial, ranging between 5 and 10 percent. Cotton supports the livelihoods of millions in developing countries (at least 10 million in West and Central Africa) where it is a typical, and often dominant, smallholder cash crop. The cotton market also has been subject to considerable market intervention-subsidization in the European Union and the United States, and taxation in Africa and Central Asia. During the past three seasons, annual direct support averaged $4.5 billion. The author reviews the market setting and policy issues and gives recommendations on how industrial and developing cotton-producing countries can improve the policy environment.
Content may be subject to copyright.
Cotton
Market Setting, Trade Policies, and Issues
JOHN BAFFES
DEVELOPMENT PROSPECTS GROUP
THE WORLD BANK
1818 H Street, NW
Washington, D.C. 20433
tel: (202) 458-1880
fax: (202) 522-1151
email: jbaffes@worldbank.org
World Bank Policy Research Working Paper 3218, February 2004
The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of
ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are
less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, inter-
pretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent
the view of the World Bank, its Executive Directors, or the countries they represent. Policy Research Working Papers
are available online at http://econ.worldbank.org.
This paper is part of a larger effort of the World Bank’s Development Prospects Group and Trade
Department to gain an in depth understanding of the structure of commodity markets, including
policy distortions. I would like to thank Ataman Aksoy, John Beghin, Harry de Gorter, Uri
Dadush, Betty Dow, Phil English, Gèrald Estur, Gaston Gohou, Tassos Haniotis, Ioannis Kaltsas,
Stephen MacDonald, Will Martin, Donald Mitchell, John Nash, Richard Newfarmer, and Carlos
Valderrama for comments and suggestions on earlier drafts of this paper. Access to data published
by the International Cotton Advisory Committee is greatly appreciated. Comments from partici-
pants at two World Bank seminars are also appreciated.
— ii —
CONTENTS
PREFACE
SUMMARY
ACRONYMS AND ABBREVIATIONS
PART I: THE MARKET SETTING
THE FIBER MARKET
THE GLOBAL COTTON BALANCE
Production, Consumption, and Stocks
Trade
Costs of Production
PRICE TRENDS AND VARIABILITY
COTTON AND THE DEVELOPING COUNTRIES
NON-CONVENTIONAL COTTON PRODUCTION
Genetically Modified Cotton
Organic Cotton
THE SECONDHAND CLOTHING MARKET
PART II: THE POLICY SETTING
DISTORTIONS IN THE COTTON MARKET
The United States
The European Union
China
Uzbekistan
PREFERENTIAL ARRANGEMENTS AFFECTING THE COTTON MARKET
The Agreement on Textiles and Clothing
The African Growth and Opportunity Act
IMPACT OF DISTORTIONS AND PROSPECTS FOR REFORMS
Impact of Distortions
Prospects of Reforms by Major Producers
REFORM INITIATIVES IN AFRICA
Uganda
Tanzania
Zimbabwe
Francophone Africa
PART III: THE ISSUES
SYNTHESIS OF ISSUES AND STYLIZED FACTS
RECOMMENDATIONS
Developing Countries
— iii —
Developed Countries
NOTES
LIST OF TABLES
Table 1: Cotton’s Importance to Developing and Transition Economies
Table 2: Government Assistance to US Cotton Producers
Table 3: Effect of Removal of Distortions
LIST OF FIGURES
Figure 1: The Classification of Fibers
Figure 2: Polyester to Cotton Price Ratio
Figure 3: Cotton production in Europe
Figure 4: World Fiber Consumption
Figure 5: World per Capita Fiber Consumption
Figure 6: Annual A Index
Figure 7: Monthly A Index
Figure 8: Cotton Share in Total Fiber Consumption
Figure 9: World Cotton Yields
PART IV: THE FACTS
Appendix A: Concentration Indices
Appendix B: Growth Rate Estimation
Appendix C: Price Variability
Appendix D: The Cotlook A and B Indices
Appendix E: Cotton Price Risk Management
Appendix F: Statistical Tables
REFERENCES
— iv —
PREFACE
The objective of this paper is three-fold. First, it highlights the main aspects and character-
istics of the global cotton market by analyzing developments of the past 40 years. Second,
it identifies the policy distortions in the cotton market including their impact on prices and
trade as well as the prospects for reform. Third, it discusses a number of policy alterna-
tives for both developing and developed cotton producing countries.
The paper is divided into four parts. The first part (5 sections) discusses nature of
the fiber market, the global balance of the cotton market, price trends and variability, the
nature and degree of dependence of developing countries on cotton, nonconventional cot-
ton production practices, and the secondhand clothing market.
The second part (4 sections) deals with distortions and reforms in the cotton market.
Specifically the first section discusses the distortions in the cotton market by major pro-
ducers including the United States, the European Union, China, and Uzbekistan. The next
section summarizes the preferential arrangements that indirectly affect the cotton market,
namely the Agreement on Textiles and Clothing and the Africa Growth and Opportunity
Act. The third section looks at the impact of distortions and prospects for reforms. The last
section discusses the reforms initiatives in East and West Africa.
The third part of the paper synthesizes the issues and stylized facts. It also discusses
some policy directions for both developing and developed cotton producing countries in-
cluding cotton promotion, deepening of policy reforms in developing countries and reduc-
tion (and eventual elimination) of support by major subsidizers.
The fourth part consists of five technical appendices and one statistical appendix.
The technical appendices describe the methodology of calculating the concentration indi-
ces, the model which estimates the growth rates, the measures of price variability, the
measure of world price, and issues related to cotton risk management. The statistical ap-
pendix consists of 13 tables reporting figures on the global balance of production, con-
sumption, and trade, monthly prices since 1950, global production and consumption of
chemical fibers, value of exports of secondhand clothing, a chronology of the US commod-
ity programs with cotton provisions, and data on subsidies.
— v —
SUMMARY
The value of world cotton production in 2000/01 has been estimated at about $20 billion,
down from $35 billion in 1996/97 when cotton prices were 50 percent higher. Although cot-
ton’s share in world merchandize trade is insignificant (about 0.12 percent), it is very im-
portant to a number of developing countries. Cotton accounts for approximately 40 per-
cent of total merchandise export earnings in Benin and Burkina Faso, and 30 percent in
Chad, Mali, and Uzbekistan. Its contribution to GDP in these and other developing coun-
tries is substantial ranging between 5 and 10 percent. Cotton supports the livelihoods of
millions in developing countries (at least 10 million in West and Central Africa) where it is
a typical, and often dominant, smallholder cash crop. The cotton market also has been sub-
ject to considerable market intervention—subsidization in the US and EU and taxation in
Africa and Central Asia. During the last three seasons, annual direct support averaged $4.5
billion. This paper reviews the market setting and the policy issues, and also gives recom-
mendations on how developed and developing cotton producing countries can improve
the policy environment.
THE MARKET SETTING
Global cotton production doubled from 10 million tons in 1960 to 20 million tons in 2001.
More than three-quarters of cotton output is accounted for by developing countries. China
and the United States each account for approximately 20 percent of world output, fol-
lowed by India (12 percent), Pakistan (8 percent), and Uzbekistan (5 percent). Other sig-
nificant producers are seven Francophone African countries, Turkey, Brazil, Australia, and
Greece accounting for a combined 20 percent. Most of the growth in cotton production
came from China and India which tripled and doubled their production during this 40-
year period, respectively. Other countries which significantly increased their share of cot-
ton production were Turkey, Greece, and Pakistan. Some new entrants also contributed to
this growth. Australia, for example, produced only 2,000 tons of cotton in 1960 while it av-
eraged 650,000 tons during the late 1990s. Francophone Africa started with less than
100,000 tons in the 1960s and it now produces almost one million tons. The United States
and the Central Asia Republics (then part the Soviet Union), which were the two dominant
cotton producers during the 1960s, have retained their output levels at about 3.5 and 1.5
million tons respectively, thereby halving their shares. A number of Central America
countries, on the other hand, lost substantial share of world output.
Approximately, one-third of cotton production is traded internationally. The four
dominant exporters—United States, Uzbekistan, Francophone Africa, and Australia—
account for more than two-thirds of global exports. Four major producers, China, India,
Pakistan, and Turkey are generally net importers of cotton to supply their textile indus-
tries. Imports of cotton are more uniformly distributed than exports. During the 2000/01
season, the eight largest importers (Indonesia, India, Mexico, Thailand, Turkey, Russia, It-
aly, Korea), accounted for over half of world cotton imports. Apart from Russia, which
— vi —
prior to 1990 was considered a major producer but not an importer since the Central Asian
cotton production was considered domestic trade, most of the remaining cotton importers
are new in the sense that have been importing cotton to supply their newly-developed tex-
tile industries. Four East Asian textile producers (Indonesia, Thailand, Taiwan, and Korea)
accounted for less than 3 percent of world cotton imports in 1960. Their share in 2002 was
22 percent.
A recent survey (based on a questionnaire of 28 cotton producing countries) on
costs of cotton production suggests that West Africa (especially Benin, Mali, and Burkina
Faso), Uganda, Tanzania, are among the lowest cost producers. High cost producing coun-
tries are the United States, Israel, and Syria. The two European cotton producers, Greece
and Spain, are probably the world’s highest cost cotton producers although they did not
participate in the survey.
In line with most primary commodities, real cotton prices have declined considera-
bly during the last half century; they are currently one fifth of their 1950 levels. This de-
cline has been characterized by considerable year-to-year variability, especially during the
last quarter of the century. In particular, cotton prices followed a smooth declining pattern
throughout the 1950s and 1960s. They increased sharply after the 1973 and begun declin-
ing again albeit with much higher volatility than the pre-1973 decline. A structural break
in cotton prices appears to have taken place in 1985 when the United States changed the
nature of its support policies—from stockholding to price support. Real prices have been
declining less rapidly since 1985, while volatility has been reduced compared to 1973-1984.
However, the post-1985 price volatility is about twice as high compared to the pre-1973
price volatility.
The cotton market has been significantly affected by the rapid expansion of chemi-
cal fibers, mainly polyester. Chemical fibers account currently for almost 60 percent of
global fiber consumption, up from 33 percent in 1960. Global production of chemical fibers
reached 30 million tons in 2002. Polyester prices were 4 times higher than cotton prices in
the early 1960s. Following technological improvements they declined to the level of cotton
prices in the early 1970s and since then polyester and cotton products have been trading at
similar price levels. Most chemical fibers are produced in Asia. For example, Asia’s output
of chemical fibers in 1960 was 2.4 million tons while global production in that year stood at
10.3 million tons; it reached 20 million tons in 2000 compared to 28.3 million tons of global
output.
The long term decline in cotton prices has been aided by technological improve-
ments such as application of improved varieties, fertilizers, chemicals, irrigation, and (in
the case of some developed countries) mechanical harvesting. Between 1960 and 2000
world cotton yields doubled, from 300 to 600 kilograms per hectare, implying an annual
growth rate of 1.8 percent. More recent developments in technology such as genetically
modified seeds and precision farming are likely to further reduce the costs of producing
cotton. In 2002, genetically modified cotton accounted for almost 30 percent of global cot-
— vii —
ton output. The United States is the heaviest user with more than 70 percent of its cotton
area allocated to genetically modified cotton, followed by Australia (40 percent), China (20
percent) and more recently by India. If current trends continue, half of the world’s cotton
will be of genetically modified origin within five years. Organic cotton, another “non-
conventional” way of producing cotton, has been tried on a limited scale; the outlook,
however, looks less promising compared to food crops mainly because of weak demand; it
appears that there is too great a distance between the primary commodity—cotton—and
the final product—cloth—in the eyes of the consumer.
Cotton consumption between 1960 and 2000 grew by an annual average of 1.8 per-
cent, i.e. approximately at the same rate as the population growth, implying zero per cap-
ita growth. Consumption of chemical fibers, on the other hand, has grown by 4.7 percent
per annum (or about 3 percent in per capita terms). Therefore, all per capita growth in to-
tal fiber consumption during the last 40 years has been accounted for by growth in the
consumption of chemical fibers.
THE POLICY SETTING
Although there are no significant border policies such as quotas or high tariffs, a number
of cotton producing countries use domestic measures to support their cotton sector. Ac-
cording to the International Cotton Advisory Committee, support during the 2001/2002 to-
taled $5.8 billion. These numbers, however, must be treated with caution for two reasons.
First, China has reportedly supported its cotton sector by an estimated $1.3 billion annu-
ally during the last three seasons, but it is difficult to substantiate such support since Chi-
nese cotton policies are too complex to be assessed quantitatively and the official figures
are sometimes unreliable. Second, the ICAC figures for the United States do not include all
types of transfers. A more comprehensive look indicates that US support is higher than
what ICAC reports.
Support in the United States is given in various forms. The six most important
components of support are the loan deficiency payments, marketing loan program, pro-
duction flexibility contracts (i.e. decoupled support introduced with the 1996 Farm Bill
which replaced deficiency payments), counter-cyclical payments (i.e. emergency payments
introduced in 1998 in response to low prices), insurance, and Step-2 payments (often re-
ferred to as export subsidies.) During the 2001/02 season producer prices in the United
States were 91 percent higher than world prices. Support in the European Union (i.e.
Greece, and Spain) is given in the form of guarantee prices (i.e. the difference between a
pre-announced target price and the market price). During the 2001/02 season producer
prices in Greece and Spain were 144, and 184 per cent higher than world price, respec-
tively.
The current low cotton prices, which have been influenced by the support provided
by major players, have taken a toll on the rural sector of cotton-dependent countries. Re-
search findings indicate that in Benin, where cotton accounts for 40 percent of total mer-
— viii —
chandise exports and contributes more than 7 percent to GDP, a 40 percent reduction in
farmgate cotton prices—equivalent to the price decline that took place from December
2000 to May 2002—implies a 7 percent reduction in rural per capita income in the short
run and 5-6 percent reduction in the long run. Furthermore, the incidence of poverty
among cotton growers in the short run rises from 37 percent to 59 percent while the aver-
age incidence of rural poverty (i.e. including cotton growers and other farmers) rises from
40 percent to 48 percent.
In addition to low prices and loss of export shares by non-subsidizing producers,
support by major players has triggered a number of noteworthy reactions.
Many cotton producing countries have reacted by introducing offsetting support.
Turkey, Brazil, Mexico, Egypt, and India, totaled $0.6 billion of support during
2001/02.
Brazil has initiated a WTO consultation process claiming losses to its cotton exports
due to subsidies by the United States.
Four West African cotton producing countries (Benin, Burkina Faso, Chad, and Mali)
are pressing for removal of support to cotton sector through the WTO. In an unusual
move, the President of Burkina Faso addressed the WTO on June 10, 2003, asking for
financial compensation for cotton producing low-income countries to offset the in-
jury caused by support. The requested compensation was to be in place for as long as
subsidies are in place.
The cotton sector has found an unlikely ally in the Director General of the Interna-
tional Rayon and Synthetic Fibres Committee. In a letter to the Financial Times on
June 12, 2003 he complained that “recent increases in cotton subsidies have rigged
the market even more dramatically in favor of cotton, depressing demand for every
substitute product. The result is industrial plants being kept idle… that were built in
legitimate expectation that the competitive advantages of manufactured fibers
would create demand to fill the capacity…”
Removal of support is expected to reduce production and consequently boost
prices. Simulations show that if full liberalization in the cotton sector takes place including
removal of both trade barriers and production support (along with liberalization in all
other commodity sectors), cotton prices would increase in the next 10 years by an average
of 12.7 percent over the price that would have prevailed in the absence of reforms. World
cotton trade would increase by 5.8 percent while Africa’s cotton exports would increase by
12.6 percent. Uzbekistan would increase its exports by 5.8 percent, Australia by 2.7 per-
cent, while exports from the United States would decline by 3.5 percent. Cotton produc-
tion in the Unites States and the European Union would decline by 6.7 and 70.5 percent,
respectively—in effect, cotton production in the European Union would fall to levels even
below those prior to the Common Agricultural Policy taking effect. Production in Uzbeki-
stan and Africa would increase by 4 and 6 percent, respectively.
However, complete elimination of support is unlikely. The European Union re-
formed its cotton policy regime in 1999 and is unlikely to eliminate support because: (i)
— ix —
none of the candidates to join the Union in the current expansion are cotton producers and
hence there will be no pressure to increase the budgetary allocation to the cotton sector
and (ii) the current cotton program is viewed as a poverty reduction mechanism since the
support supposedly goes to low-income regions of Southern Europe. The United States
approved the 2002 Farm Bill which, in effect, legitimized the emergency payments that had
been given to its cotton (and other commodity) growers following the 1997/98 price de-
cline; it also established a minimum price of $0.71 per pound (or $1.56 per kilogram, much
higher than the 2001 and 2002 world averages of $1.06 and $1.00 per kilogram). The 2002
Farm Bill will be in place for the next 6 years, consequently guaranteeing US cotton grow-
ers generous support until the year 2007, if the current low prices persist (some support
will be in place even if prices increase considerably).
The cotton market has also been affected indirectly by the Agreement on Textiles
and Clothing (the successor of the Multifibre Arrangement), which through quotas and
tariffs on textiles and apparel has influenced the location of the textile industry, conse-
quently imposing an implicit tax on cotton goods. The Agreement is expected to be phased
out by the end of 2004. However, it is back-loaded with most of the reforms expected to
take place at the end of 2004, thus introducing the risk of non-compliance. The Africa
Growth and Opportunity Act may also help many African cotton producing countries to
expand their textile exports to the United States and thereby increase domestic cotton con-
sumption. For example, currently, the average duty to garment imports into the US is 17.5
percent. Under the Act, apparel imports into the United States from the 14 eligible African
countries will be duty free subject to an upper limit of 3 percent of total US apparel im-
ports. Since total trade in clothing from Africa to the US is very small, the 3 percent cap (to
increase to 7 percent by year 7) is unlikely to become a binding constraint. Another benefi-
cial provision is that for countries with Less Developed status, there is a 4-year exception
to the rule of origin (it expires in 2004). For example under this provision, Tanzania, can
import yarn from China and export cloth to the US. However, it should be noted that
while local cotton consumption may increase due to the Act, global consumption is
unlikely to be affected in any significant way.
A number of cotton producing countries (especially in Sub-Saharan Africa) which
traditionally had been taxing their cotton sectors, undertook substantial policy reforms
during the 1990s in order to increase the efficiency of the cotton sectors. These reforms
have been supported, in part, by multilateral institutions, including the World Bank. In
most occasions reforms were the only feasible alternative because the parastatals handling
the marketing and trade of cotton were crippled by huge debts and either went bankrupt
or they managed to stay alive with state infusions of capital. This, in turn, was caused by
falling world cotton prices, inefficiencies and poor management of the parastatals and of-
ten outright corruption.
Substantial reforms were undertaken by countries in Eastern and Southern Africa,
namely Tanzania, Uganda, Zambia, and Zimbabwe. With the exception of Tanzania
— x —
(where the reform process was never completed), considerable supply response took place
along with higher share of export prices and timely payments to growers. There have been
numerous reports that the quality of cotton declined after reforms, but these reports have
been unsubstantiated. West African cotton producers are also contemplating reforms. Cen-
tral Asian cotton producers, especially Uzbekistan, still tax heavily their cotton sectors and
reforms are unlikely to take place soon.
MARKET OUTLOOK AND POLICY OPTIONS
The prospects for growth in cotton consumption are similar to the patterns followed in
earlier decades. Pressure from chemical fibers is likely to remain strong as technological
improvements are likely to enhance their properties and reduce the costs of production. A
growing second-hand clothing market, especially in developing countries, has displaced
(and is likely to further displace) potential demand growth for new garments in these
markets. Therefore, consumption growth is unlikely to exceed the projected population
growth of 1.2 percent. Prices, on the other hand, while they are expected to recover from
the record lows experienced during the 2001 and 2002, are unlikely to reach the highs of
the 1970s (or even the mid-1990s). Given modest consumption growth and poor price
prospects, reducing the costs of production is an imperative for the cotton-dependent de-
veloping countries in order for them to increase (or at least sustain) their share in global
cotton demand. On many occasions, that would entail further policy reforms.
Undoubtedly, the price prospects (and consequently the export shares of low cost
producers, including many African countries) can be improved considerably if support by
developed countries is reduced substantially or eliminated altogether. However, given the
low probability of outright elimination of support, a second best alternative would be for
support to be given in a non-distortionary manner. A type of support with minimal distor-
tionary effects—the so-called decoupled support mechanisms—has re-gained popularity
recently. Income transfers under decoupled support are based on past production levels
and prices and thus have minimal impact on current production decisions—at least in the-
ory. What makes decoupled support in the cotton sector an interesting (and potentially
applicable) alternative is that almost all support is in the form of domestic measures.
Therefore, changing the nature of support does not require changing the sources of fund-
ing such support as it would in the case of border measures.
Decoupled support was attempted in the European Union with the Common Agri-
cultural Policy reform of 1992, in Mexico with the PROCAMPO program of 1994, and in
the United States with the Freedom to Farm Act of 1996. The outcome of these programs
has not been encouraging because these schemes did not include three essential elements
that would make them successful: (i) substituting all existing support mechanisms with
decoupled support; (ii) limiting the duration of the programs which would have made
them true transition mechanisms and (iii) not requiring that land remain in agricultural
use which would reduce the overall supply of the commodities under consideration and
— xi —
hence lift world prices. Unless these conditions are met, any attempts to restore the credi-
bility of decoupled support policies and ultimately remove support to the cotton (and
other) sector(s) are unlikely to have the intended beneficial impact.
— xii —
ACRONYMS AND ABBREVIATIONS
ADF Augmented Dickey-Fuller (unit root test)
AFD Agence Française de Développement
AGOA Africa Growth and Opportunity Act
ATC Agreement on Textiles and Clothing
CAP Common Agricultural Policy
CEMAC Communauté Economique et Monétaire de l’Afrique Centrale
CFA Communauté Financière Africaine
CFDT Compagnie Française de Développement des Fibres Textiles
COMTRADE Commodity Trade Statistics
EU European Union
FAO Food and Agriculture Organization
FAPRI Food and Agriculture Policy Research Institute
GATT General Agreement on Tariffs and Trade
GDP Gross Domestic Product
GM Genetically Modified
HS Harmonized System
ICAC International Cotton Advisory Committee
IFCP International Forum for Cotton promotion
IRSFC International Rayon and Synthetic Fibres Committee
MFA Multifibre Arrangement
MUV Manufactures (export) Unit Value
OLS Ordinary Least Squares
PP Phillips-Perron (unit root test)
PROCAMPO Programa Nacional de Modernizacion del Campo
SITC Standard International Trade Classification
STO State Trading Enterprise
TRQ Tariff Rate Quota
UEMOA Union Economique et Monétaire Ouest-Africaine
UKP UzKhlopkoprom/UzPakhtasanoitish
USDA United States Department of Agriculture
WTO World Trade Organization
— 1 —
PART I: THE MARKET SETTING
THE FIBER MARKET
Fibers include a wide variety of products which can be divided into two broad categories:
natural and man-made (see figure 1).1 Natural fibers can be further divided into fibers of
plant-origin (such as cotton and linen) and fibers of animal-origin (such as wool and silk).
Likewise, man-made fibers can be further divided into inorganic and organic fibers. Inor-
ganic fibers are materials such as ceramic, glass, and carbon (typically not used in gar-
ments.) Organic man-made fibers, on the other hand, are mostly used in garment produc-
tion either as substitutes or as complements to natural fibers. Organic fibers are further
sub-divided into natural and synthetic polymers. Natural polymers (often called cellu-
losic) are made from pulp (i.e. wood). The most common natural polymer is viscose, also
known as rayon. The synthetic polymers are made from crude oil. The most common syn-
thetic polymers are polyester, acrylic, and polyamide (also known as nylon). It is interest-
ing to note the continuum of fibers in terms of their level of chemical transformation of the
raw material. At one end of the spectrum are the natural fibers with minimal chemical
transformation. In the middle are natural polymers with some degree of chemical trans-
formation. At the other end of the spectrum are synthetic polymers with substantial de-
gree of chemical transformation.
Cotton—by far the most common natural fiber of the 19th and 20th centuries—has
been used as a raw material for clothing for at least 5,000 years. Its use expanded signifi-
cantly after the invention of the ginnery in 1793 (which introduced mechanical separation
of seed and lint consequently reducing the costs of producing cotton lint) and the indus-
trial revolution which reduced the cost of producing textiles.
Although commercial production of man-made fibers on a large scale is a post-
WWII phenomenon, experimentation was taking place as early as the late 1800s. Man-
made fibers first appeared in the market earlier in the 20th century. In 1925, for example,
rayon accounted for 1.6 percent of the world’s total fiber consumption. Its share increased
to 11.8 percent 20 years later.2 Global production of cotton and chemical fibers reached 20
and 30 million tons, respectively, in 2000. The dominant chemical fiber producer is China
accounting for 6.7 million tons, followed by the European Union, the US, and Taiwan with
3.4, 3.3, and 3.2 million tons, respectively.
Between 1960 and 2002, man-made fiber consumption increased at an annual rate of
4.7 percent. Cotton consumption during this period increased only 1.8 percent. Per capita
chemical fiber consumption in 1960 and 2000 was 1.75 and 4.52 kilograms, respectively.
The share of man-made fiber consumption is currently 57 percent, up from 22 percent in
1960; cotton’s share fell to 40 percent in 2002 (see figure 2).
Prices of non-cellulosic man-made fibers have typically traded at comparable levels
with cotton since the early 1970s. Between 1960 and 1972, the polyester price indicator de-
— 2 —
clined form $12 per kilogram to $2.50 per kilogram, mainly a reflection of the technological
improvements (and consequently cost reductions) that took place in the chemical fiber in-
dustry. After reaching parity with the A Index in 1972, the ratio of polyester to A Index has
increased at an average rate of 1 percent per year, implying that while cotton and polyes-
ter are priced at similar levels, polyester has made small pricing gains (see figure 3).
THE GLOBAL COTTON BALANCE
Production, Consumption, and Stocks
Cotton is produced in many countries but the northern hemisphere accounts for 90 percent
of global output and more than two thirds of cotton is produced by developing countries.
During the last 4 decades cotton production grew at an annual average rate of 1.8 percent
to reach 20 million tons in 2001 from 10.2 million tons in 1960. Most of this growth came
from China and India which tripled and doubled their production, respectively, during
this 40-year period. Other countries which significantly increased their share of cotton
production were Turkey, Greece, and Pakistan. Some “new entrants” also contributed to
this growth. Australia, for example, produced only 2,000 tons of cotton in 1960 while it av-
eraged 650,000 tons during the late 1990s. Francophone Africa produced less than 100,000
tons in the 1960s and now produces almost one million tons. The United States and the
Central Asia Republics (then the Soviet Union), the two dominant cotton producers during
the 1960s, have maintained their output levels at about 3.5 and 1.5 million tons respec-
tively, thereby halving their shares. A number of central America countries which used to
produce almost 250,000 tons of cotton, now produce virtually none. The share of East Afri-
can cotton producers has declined considerably during this period. The concentration of
production has been declining since its 1984 peak, when China became an important
player in the cotton market, mainly reflecting increased production by the new entrants
(see appendix A for definition and calculation of concentration).
During the 1990s, cotton production fluctuated between 18 and 20 million tons with
no significant trend. China and the United States each accounted for approximately 20
percent of world output, followed by India (12 percent), Pakistan (8 percent), and Uzbeki-
stan (5 percent). Other significant cotton producers are the countries of Francophone Af-
rica, Turkey, Brazil, Australia, and Greece which account for a combined 18 percent of
global output. The remaining share is accounted for by a number of smaller producers.
The consumption pattern of cotton is primarily determined by the size of the textile
industries of the dominant cotton consumers. China, the leading textile producer, ab-
sorbed more than one quarter of global cotton output during the late 1990s. Other major
textile producers (and hence major cotton consumers) are India, the United States, and
Turkey, which together (including China) account for three-quarters of global cotton con-
sumption. A number of East Asian countries have emerged recently as important cotton
consumers. For example, Indonesia, Thailand, Korea, and Taiwan consumed only 130
thousand tons in 1960 (1.2 percent of global consumption) while they consumed 1.5 mil-
— 3 —
lion tons in 2002 (7.2 percent of global consumption). That is also reflected in the concen-
tration pattern of consumption which increased by 2 percentage points during the 1990s.
Growth in the demand for cotton has been slow. Between 1960 and 2000, cotton
demand has grown at the same rate as population (1.8 percent per annum) implying that
per capita cotton consumption has remained stagnant. A contrasting view of total and per
capita cotton consumption is vividly illustrated in figures 4 and 5.
Stocks, which have historically fluctuated between 20 and 50 percent of global out-
put, have affected the cotton market considerably, especially price variability. Major
stockholders are the United States and China. Consequently, the stockholding policies of
these two countries have affected the level and volatility of cotton prices. Two major cot-
ton de-stocking episodes are associated with periods of considerable price variability: the
1985 shift in US policy from stockholding to price support and the 1999 reforms in China.
Trade
One-third of cotton production is traded internationally. The four dominant exporters—
US, Uzbekistan, Francophone Africa, and Australia—account for more than two-thirds of
the world’s exports. Four major producers, China, India, Pakistan, and Turkey do not ex-
port cotton and occasionally import to supply their textile industries. Imports of cotton are
more uniformly distributed than exports.
During the 2000/01 season, the eight largest importers (Indonesia, India, Mexico,
Thailand, Turkey, Russia, Italy, Korea) accounted for over half of world cotton imports.
Apart from Russia, which prior to 1990 was considered a major producer but not an im-
porter because the Central Asian cotton production was considered internal trade, most of
the remaining cotton importers are new in the sense that have been importing cotton to
supply their newly-developed textile industries. For example, four East Asian textile pro-
ducers (Indonesia, Thailand, Taiwan (China), and Korea) accounted for less than 3 percent
of world cotton imports in 1960, compared to 22 percent in 2002.
The import concentration index fluctuated around 5 percent during the late 1990s
versus an export concentration of 10 percent. The corresponding indices during the 1960s
were 7 and 12 percent, respectively, indicating that cotton trade is less concentrated than
before despite the fact that trade as a percent of global output has not changed appreciably
during this period.
In terms of direction of trade flows, 44 percent of cotton exports went from indus-
trial to developing countries during 2000/01. The shares for 1980/81 and 1990/91 were 38
and 31 percent. The shares of cotton exports from developing to developing countries in-
creased from 13 percent in 1980/81 to 31 percent in 2000/01. This change in the pattern of
trade flows reflects the growth of the textile industries in South-East Asia.
Costs of Production
The International Cotton Advisory Committee collects data comparing costs of production
— 4 —
among cotton producers. In its most recent (2001) survey, based on a questionnaire of 28
cotton producing countries, it suggests that West Africa (especially Benin, Mali, and Burk-
ina Faso), Uganda, and Tanzania, are among the lowest cost cotton producers. High cost
producers are the United States, Israel, and Syria. The two European cotton producers,
Greece and Spain, are probably the world’s highest cost cotton producers although they
did not participate in the survey.
Calculating and comparing the costs of producing cotton in various countries is,
admittedly, a difficult task as it would involve a number of assumptions about the cost of
land and capital as well as various hidden subsidies and distortions. As the publication
warns, “The data must be used carefully. Differences in production practices, variations in
the input supply among countries and direct and indirect technical and financial support
to farmers in the form of free seed, technical advise, etc. makes comparisons difficult
among countries.” (p. 5).
The Long Term Outlook
The average population growth for the current decade is projected at 1.2 percent per an-
num. In the absence of any policy reforms by major players ICAC (2003) projects that con-
sumption during the current decade is expected to be about 1.8 percent per annum, imply-
ing that by 2010 world cotton consumption will be 23.6 million tons. However, that may be
viewed as an optimistic scenario considering that during the last 15 years cotton consump-
tion grew by an annual rate of 0.7 percent.
PRICE TRENDS AND VARIABILITY
Real cotton prices have followed a declining pattern albeit subject to temporary spikes and
troughs, a pattern that has been consistent throughout the last two centuries.3 The reasons
for the long-term decline of cotton prices are similar to those characterizing the price de-
clines in most primary commodities, namely, reduction in the costs of production due to
technological improvements, slow demand growth, and strong competition from chemical
fibers. The declining pattern of cotton prices has not been smooth and it appears that a
structural break has taken place in the mid-1980s (see figure 6). Between 1960 and 1984
real cotton prices averaged $2.62 per kilogram. Following a sharp decline in 1984 (from
$2.45 per kilogram in 1984 to $1.83 in 1985 and $1.27 in 1986), they have been fluctuating
around an average of $1.49/kg in the post-1985 period. There has been a declining trend of
about 0.9 percent per annum between 1985 and 2002 (as opposed to only 0.2 percent dur-
ing 1960-84).4
Reductions in the costs of production have been primarily associated with yield in-
creases which during the last 40 years have doubled, from 300 kilograms per hectare in the
early 1960s to 600 kilograms per hectare in the late 1990s. The phenomenal yield growth
was aided primarily by the introduction of improved cotton varieties, expansion of irriga-
tion, use of chemicals and fertilizers, and mechanical harvesting.5 To these improvements
— 5 —
one should add developments in genetically modified seed technology and precision
farming during the late-1990s, which are expected to further reduce the costs of produc-
tion. Other innovations in the transportation and the information technology sectors have
reduced costs of transporting cotton and also reduced the need for large stockholding.
Substantial technological improvements have also taken place in the textile sectors
whereby the same quality of fabric can be achieved with much lower quality of cotton, a
trend that holds for many products whose main input is a primary commodity.
The principal reason behind the 1984/85 decline in cotton prices was the structural
shift of the support policy by the United States and the shift in China’s trade policy
(McDonald 1997). During the 1950s, the US Commodity Credit Corporation bought and
sold most of the US cotton. For example, between 1962 and 1966, the Credit Corporation
accounted for almost two-thirds of cotton carry-over. While its role was reduced after
1970, the US would still account for 35 percent of non-Chinese stocks. Following the 1985
Farm Bill, the loan rates (i.e. the equivalent of a minimum prices) were substantially re-
duced and most of the US stocks were released and found their way into the world mar-
ket. This year also marked the initiation of large exports by China, which for the previous
20 years was a net importer. In fact between 1980 and 1985, China went from the world’s
largest importer to the world’s largest exporter.6
During the 1990s, nominal cotton prices, as measured by the Cotlook A Index, fluc-
tuated between $2.53 per kilogram (May 1995) and $0.97 per kilogram (December 1999).
The post-1996 decline in cotton prices was a result of a number of factors: first, there was
excess production during the 1997/98 season. Second, demand was weak, especially from
the East Asian textile producers affected by the financial crisis of 1997—Indonesia, Repub-
lic of Korea, and Thailand. Together these countries account for more than 15 percent of
cotton import demand. Third, stocks rose to a record 9.8 million tons in 1997/98, which
pushed the stock-to-use ratio to 0.51, the highest ratio since 1985/86. Fourth, low synthetic
fibers prices, which were a result of currency devaluations in several East Asian chemical
fiber producers. Between January 1997 and January 1998 the South Korean polyester fiber
indicator price declined from $1.66 to $0.79 per kilogram. The strength of the US dollar
during that period also contributed to the price declines.
It appeared that cotton prices would have a sustained recovery when they reached
$1.45 per kilogram in December 2000—up 45 percent from a year earlier (see figure 7).
However, the recovery was short-lived since it mainly reflected the 1998/99 weather-
related short-fall in the US crop. In the 1998/99 season, US cotton output was a little over 3
million tons compared to an average of 4 million tons in the preceding 5-year period. With
production and consumption for the 2001/02 season at 21.1 and 19.9 million tons, implying
a surplus of 1.2 million tons, cotton prices have been under intense pressure. The A Index
declined to $0.82 per kilogram in October 2001, which, excluding August 1986, is the low-
est since November 1972.
In addition to their declining pattern, cotton prices have been volatile. The nature of
— 6 —
volatility, however, has changed considerably during the last 40 years. A simple measure
of volatility shows that during 1985-2002 volatility was 2.5 times higher compared with
1960-72, but half compared to the 1973-84 period (see Appendix C for calculation and
comments on volatility). Note that 1973 reflects the commodity price boom while 1985 co-
incides with the US change in cotton policy regime and the subsequent disposal of large
cotton stocks.
COTTON AND THE DEVELOPING COUNTRIES
Although cotton is insignificant on a global scale (it accounts for only 0.12 percent of total
merchandise trade), it is an important cash crop to a number of developing countries at
both farm and national level. For example, cotton accounted for between 30 and 44 percent
for total merchandize exports in 5 West African cotton producing countries (Burkina Faso,
Benin, Chad, Mali, Togo) during 1998-99. The corresponding figures for Uzbekistan, Taji-
kistan, and Turkmenistan was 32, 15, and 12 percent, respectively. Cotton’s contribution to
the GDP of these countries has been substantial, ranging between 3.6 percent (Turkmeni-
stan) and 8.2 percent (Tajikistan). With the exception of Turkmenistan, the per capita an-
nual GDP in these countries is well below $500. In most of these countries (especially in
Africa) cotton is typically a smallholder crop, it is the main cash crop, it is grown in rainfed
land and the use of purchased inputs such as chemicals and fertilizers is minimal.
According to FAO estimates, there were 100 million rural households involved in
cotton production worldwide during 2001. In China, India, and Pakistan about 45, 10, and
7 million rural households, respectively were engaged in cotton production. The total
number of rural households depending on cotton in major African producing countries,
including Nigeria, Benin, Togo, Mali, and Zimbabwe totaled 6 million.
The high dependence on cotton in these countries has important poverty ramifica-
tions, especially when large price changes take place. In a study which focused on Benin,
Minot and Daniels (2002) found that a 40 percent reduction in farmgate cotton prices—
equivalent to the price decline that took place from December 2000 to May 2002—implied
a 7 percent reduction in rural per capita income in the short run and 5-6 percent reduction
in the long run. They also found that the incidence of poverty among cotton growers rose
in the short run from 37 percent to 59 percent while the average incidence of rural poverty
(i.e. including cotton growers and other farmers) rose from 40 percent to 48 percent.
In terms of policy interventions, the cotton sector in developing countries has been
traditionally taxed either explicitly through export taxes or implicitly through oligopsonis-
tic arrangements or exchange rate misalignments. The pattern, however, has changed
somewhat during the 1990s as a number of cotton producers undertook reforms. How-
ever, several African and all Central Asian cotton producers still tax their cotton sectors.
NON-CONVENTIONAL COTTON PRODUCTION
Recent trends in growing cotton focus on cost reductions through less intensive use of in-
— 7 —
puts, especially chemicals. These include the use of genetically modified (GM) seed tech-
nology and organic methods of production. GM cotton has not faced the degree of opposi-
tion faced by GM food crops and this has allowed more rapid adoption. Organic cotton
has been embraced enthusiastically by environmental activists, but not by consumers.
Hence, while there is plenty of room for expanding GM cotton, the scope for expanding
organic cotton appears to be limited.
Genetically Modified Cotton
Genetically modified (GM) cotton, a result of technological developments of the 1990s, has
the potential of reducing the cost of production and hence increasing profitability of the
early adopters of this technology.
There are two types of GM cotton: Bt cotton (first used in the US in 1996) and herbi-
cide-tolerant cotton (it gained approval by the US Environmental Protection Agency in
1998). Bt (Bacillus thuringiensis) is a naturally occurring soil bacterium that has been used
as a biological pesticide for many years. The gene that produces an insect toxin has been
transferred from that bacterium into the cotton plants. In turn, the plants produce their
own toxin and there is no need for the grower to apply certain pesticides. Herbicide-
tolerant cotton is a cotton plant that has been genetically modified to resist a herbicide that
would otherwise kill both weeds and the cotton plant. Consequently, the herbicide can be
applied without exterminating the cotton plant.
In economic terms, GM-type cotton (as well as all other GM products) act as insur-
ance against pests, insects, or weeds. The growers pay a premium for the resistant-seed (as
they would when buying insurance). If the insect attacks the crop, the grower’s benefit
comes through the lower costs from not spraying. If the insect does not attack the crop
growers simply lose their premium (i.e. the cost difference between conventional and GM
cotton).
GM cotton was first grown in the United States in 1996.7 A number of cotton pro-
ducing countries have introduced GM cotton technology since then including China, In-
dia, and Mexico in the northern hemisphere and Argentina, Australia, and South Africa in
the southern hemisphere. Other countries are in the process of approval or at the trial
stage, including Israel, Pakistan, Turkey, Brazil, and Indonesia. Major producers that have
not used or approved GM cotton (as of 2003) are the European Union, Central Asia, and
Francophone Africa (except Burkina Faso, which is conducting trials).
It is estimated that about 22 percent of world cotton area is under GM varieties, up
from 2 percent in 1996/97. The largest user of GM cotton is the United States, which during
the 2003/04 season is estimated to have sown 70 percent of its cotton area with GM varie-
ties. In Australia, about 44 percent of cotton area was sown to GM varieties in 2002/03 up
from 40 percent in 2000/01. In China, which adopted the new technology at an experimen-
tal stage in 1996, more than 20 million hectares with GM varieties were planted in 2002,
corresponding to over 20 percent of cotton area. In addition to the imported GM varieties,
— 8 —
China has developed its own 11 GM varieties. According to Carl et al. (2001), the major
share of the benefits from growing Bt cotton in China went to farmers (most of whom are
small-holders).8 This is because the Chinese cotton varieties were developed using public
funds. In contrast, most of the benefits associated with GM products in the other cotton
producing countries go to biotech and seed companies.
If the conversion to GM cotton varieties continues at rates experienced during the
last few years, in less than 5 years as much as half of world’s cotton will be of GM origin.
That is equivalent to 40 percent of cotton area.
Organic Cotton
The second trend, organic cotton, may be a small market niche to be exploited by develop-
ing countries. Most developing countries can be classified as “organic” cotton producers,
without altering their production practices, because of their low reliance on chemicals and
fertilizer. Organic cotton potential, however, appears to be limited. Organic cotton initia-
tives have taken place in many countries, including in Africa, but the scale is still insignifi-
cant compared to global production of conventional cotton. Myers and Stolton (1999) re-
ported that in 1997, about 8,150 tons of certified organic cotton fiber was produced world-
wide, of which, 2,600 tons was produced in the US, 1,175 in India, 1,800 in Turkey, 1,570 in
Africa, and 845 in Latin America.
Significant expansion of organic cotton faces a number of difficulties on both the
supply and demand side. On the supply side, the certification process (especially in Afri-
can cotton producing countries where the majority of growers are smallholders) is costly
to implement and monitor; thus any benefits to farmers through higher prices of selling
organic cotton and less use of chemicals are likely to be as high as the certification costs.
On the consumption side, demand for organic cotton is not as strong as is in other com-
modities such as coffee and tea. There are three reasons for this. First, there is a “distance”
in the eyes of the consumer between the primary product (cotton) and the final product
(cloth). Second, consumers of clothing (as opposed to consumers of, say, beverages) must
pay attention to a host of factors before they make their purchasing decision. The decision
involves brand, color, style, size, type of cotton (typically identified by origin, such Peru-
vian, Egyptian, Turkish cotton), content (e.g., 80% cotton, 20% polyester), and care instruc-
tions. Adding to that already congested list information on whether cotton is of organic
origin is rather difficult. Note that this decision making process compares unfavorably
with much simpler labeling for, say, coffee or tea where something like “Organically
grown from Costa Rica” or “Organic of Kenya origin” is likely to suffice. Third, organic
products are typically associated with health-related benefits that do not apply to non-
food products such as cotton.
THE SECONDHAND CLOTHING MARKET
One development that reportedly has affected textile production (and hence cotton con-
— 9 —
sumption) patterns in many developing countries during the last two decades is the dra-
matic increase in the trade of secondhand clothing. Estimates of the level of secondhand
clothing trade are not accurate, however, COMTRADE statistics—the only comprehensive
data source—show that secondhand clothing trade grew from $0.4 billion to $1.4 billion
between 1980 and 2000. More than half of this trade takes place from industrial to develop-
ing countries, with the United States and Germany accounting more than $300 million (see
table F7, Appendix F). While the level of trade may seem small in value terms, it is large in
quantity terms. For example, consider that a typical T-shirt which would be sold for more
than $10 in the US or Europe it is sold for only $0.50 in Africa.
The secondhand clothing trade typically originates through donations from the
public to charitable organizations such as the Salvation Army and Goodwill Industries in
the US, or Oxfam and Humana in Europe. Following collection, these organizations mar-
ket the cloths through normal commercial channels. Sometimes they let other companies
use their name, subject to a fee. The secondhand clothing industry has a number of unique
characteristics. Hansen (2000, p. 122), for example, summarized it as follows:
The secondhand clothing trade is an unusual industry. Few other industries obtain their raw
materials free, as do the charitable organizations, or have suppliers, the clothing donating pub-
lic, who do not know the important role they play at the start of a long commodity chain. In-
deed, the clothing-donating public is generally not aware of either the grant scale on which the
charitable organizations commonly dispose of clothing to commercial middlemen or contract
out the right to solicit and sell under their name.
There are numerous (mostly supply-driven) factors behind the growth of the of sec-
ondhand clothing trade:
Wealth. Consumers in high-income countries respond to changes in fashion quickly,
hence increasing the availability of (seemingly unused) clothing.
Tax incentives. When taxpayers donate to charities they often claim such donations on
their tax returns, thus reducing their tax liabilities.
Lack of Information. Because the public is often unaware that donations go through
normal marketing channels, it probably donates more clothing that it would in the
presence of full information disclosure.
Technological improvements. Technological advancements during the last two dec-
ades have improved considerably the quality of garments thus increasing the durabil-
ity of clothing far beyond what most consumers would consider as “normal use pe-
riod.”
Environmental sensitivities. Because of sensitivity with environmental issues, con-
sumers often prefer to recycle rather than dispose, even when the opportunity cost of
recycling is not zero.
Policy reforms. Liberalization of the textile industries in many low-income countries
reduced the competitiveness of the locally produced clothing thus increasing the mar-
— 10 —
ket share of imported secondhand clothing.
The growth of secondhand clothing trade has a number of important welfare impli-
cations. First, consumers in developing countries gain because they have access to cheaper
clothing. Second, the supply chain industry, such as importers, traders, and merchants of
secondhand clothing gain. Third, the domestic textile industries of the countries that im-
port secondhand clothing lose.
Opinions on these welfare implications differ. Some argue that secondhand clothing
is dumping and should be subjected to prohibitive tariffs. Others believe that it is just an-
other well-functioning sector albeit subject to minor trade distortions. Indeed, the tax
credit adds an incentive to donate, thus it may be viewed as an export subsidy. Further-
more, the quantity and quality of information that the public receives when asked to do-
nate clothing may be considered implicit subsidy—had they known that clothing goes
through normal marketing channels, at least some would have chosen not to donate.
Apart from these two factors, however, all other aspects of the secondhand clothing indus-
try appear to be subject to normal demand and supply conditions. At the outset, the ques-
tion is a quantitative one. If, in the absence of the tax incentive and the presence of full in-
formation disclosure, people would still donate the same (or similar) amounts of clothing,
then the subsidy (and, a fortiori, dumping) issue does not deserve merit.
Because cotton is traded internationally with no significant border distortions (apart
from the provisions of the Agreement of Textiles and Clothing), the location of the produc-
tion and consumption of clothing should not matter. Thus, the only way that global cotton
consumption can be affected by secondhand clothing trade is if the removal of the tax in-
centive along with proper dissemination of information reduces the availability of second-
hand clothing considerably (i.e. secondhand clothing are destroyed).
— 11 —
PART II: THE POLICY SETTING
DISTORTIONS IN THE COTTON MARKET
Cotton has been subject to various marketing and trade interventions. Townsend and
Guitchounts (1994) estimated that in the early 1990s, more than two-thirds of cotton was
produced in countries which had some type of government intervention, including taxa-
tion and subsidization policies—cotton producing countries with little or no government
intervention included: Argentina, Australia, El Salvador, Guatemala, Israel, Nicaragua,
Nigeria, Paraguay, Peru, and Venezuela.
Interventions, whether taxes or supports/subsidies, have occurred (and in most
cases still occur) through domestic market activities by state enterprises, price supports,
and import duties or quotas. These activities result in the following broad (though not al-
ways distinct) types of distortions:
Taxation through a state marketing monopsony. To transfer resources from cotton pro-
ducers to the government, the state marketing agency pays fixed, below-world prices
for cotton. This kind of intervention has been common in Central Asia, where the
state handles both domestic marketing and international trade. In most of Franco-
phone Africa domestic enterprises, along with a French state enterprise, control cotton
marketing and trade.
Taxation through border interventions. Typically to protect the domestic textile indus-
tries, the government uses border interventions to tax cotton producers. Egypt, India,
Pakistan, and Turkey have occasionally exercised interventions of this nature.
Support to producers through price interventions. To increase producers’ income, cotton
producers in the European Union receive support under the Common Agricultural
Policy, amounting to twice the world price in some years, while U.S. cotton producers
receive generous support. This type of intervention accounts for the greatest part of
distortions in the global cotton market.
Support through border interventions. To increase producers’ income some countries,
such as China, impose import tariffs on cotton.
Support through input subsidies. In addition to output distortions, several distortions in
input markets have affected the cotton sector, most notably subsidies on credit, fertil-
izer, and irrigation.
The International Cotton Advisory Committee (2002 and 2003), which has been
monitoring the level of assistance to cotton production by major producers since 1997/98,
found that eight countries provided direct support to cotton production—US, China,
Greece, Spain, Turkey, Brazil, Mexico, Egypt. The level of direct production assistance in
the five seasons between 1997/98 and 2001/02 ranged between $3.8 and $5.3 billion. For
2001/02, direct assistance to US cotton producers reached $2.3 billion, China’s support to-
taled $1.2 billion, and support was $0.8 billion in the EU (Greece and Spain). Producers in
— 12 —
Turkey Brazil, Mexico, and Egypt received a combined total of $150 million in support. In-
dia also supported its cotton sector during the 2001/02 season by an estimated $0.5 billion.
In addition to domestic support, there are some border restrictions, mainly in the
form of import tariffs. Most countries that impose import quotas are cotton exporters,
some with large textiles sectors. Import tariffs rates for 2003 were: Argentina (7.5 percent);
Brazil (7.5 to 10 percent); China (3 percent within quota, 90 percent outside quota; TRQ for
2003 was 856,250 tons, likely to be exhausted); Egypt (5 percent); India (10 percent); US (4.4
cents/kg within quota and 31.4 cents/kg outside quota; TRQ for 2002 was 73,207 tons while
cotton imports totaled 6,295 tons); Uzbekistan (10 percent); Zimbabwe (15 percent duty
plus 5 percent import tax).
The remaining of this section analyzes the structure and degree of domestic inter-
ventions in the US, EU, and China. It also looks at Uzbekistan, a country which taxes its
cotton sector.
The United States
Numerous commodity programs (including cotton) exist in the United States with the ul-
timate outcome of transferring resources from consumers and taxpayers to producers. The
objectives of these programs have evolved around two themes: raising and/or stabilizing
farm income and preserving the small farm. A partial list of these programs includes price
and income support, trade restrictions such as import quotas and tariffs, publicly funded
research, publicly funded irrigation, export subsidies, export credit guarantees, subsidized
land set-aside and conservation schemes, and subsidized crop insurance. The budgetary
outlays for most of these programs are authorized by the Congress (and subsequently ap-
proved by the President) every few years through various Acts, commonly known as Farm
Bills. There have been numerous Acts since the enactment of the first one in 1929, including
a 1934 Supreme Court decision which declared unconstitutional the main provisions of the
1933 Agricultural Adjustment Act.
Historically, the cotton sector in the United States has received proportionally more
support than most other commodities. Al least 20 Acts since 1929 included provisions that
affected the cotton market in one way or another. The most important shift in support dur-
ing the last two decades was the 1985 Farm Bill, which replaced public stock-holding man-
agement by a price support mechanism known as deficiency payments. A second impor-
tant change came with the 1996 Farm Bill whereby deficiency payments were replaced by
decoupled payments.
Currently, the main channels of support are de-coupled payments (formerly known
as production flexibility contracts), deficiency payments (through the loan rate mecha-
nism), insurance, subsidies to domestic mills (the so-called Step-2 mechanism also referred
to as export subsidy), and emergency payments (introduced in 1998 to compensate for the
loss of income due to low commodity prices but became “permanent” under the 2002 Farm
Bill).9 Direct payments are predetermined annual payments based on historical enrolled
— 13 —
areas of cotton and were introduced with the 1996 Farm Bill in order to compensate for the
“losses” due to the elimination of deficiency payments. Market price payments, which
consist of loan deficiency payments, marketing loan gains, and forfeitures, are designed to
compensate cotton growers from the difference between the world price and the loan rate
(i.e. target price) when the latter exceeds the former. Export subsidies, or Step-2 market
payments, are made to eligible cotton exporters and domestic end users of cotton when
domestic US prices exceed North Europe c.i.f prices by a certain level and the world price
is within a certain level of the base loan rate. The objective of the Step-2 payment is to
bridge the gap between higher US domestic prices and world prices so that US exporters
and mills maintain their competitiveness.
According to the International Cotton Advisory Committee, between 1997/98 and
2002/2003, the direct support to cotton production averaged $1.7 billion (see table 2). That
figure, however, does not include support through insurance, emergency measures, and
the step-2 mechanism. When all these measures are taken into consideration, support to
the US cotton growers is much higher (table 2). During the 1996/97 season—the first year
of the 1996 Farm Bill—support to US cotton growers amounted $878 million, almost $700
million from production flexibility contract payments and the rest from insurance subsidy.
In 1997/98 the support was $1.2 billion. When prices begun declining, the emergency assis-
tance measures were introduced, increasing the support to $1.9 billion in 1998/99, $3.5 bil-
lion in 1999/2000, $2.2 billion in 2000/01 and $3.6 billion during the last season.
In 2002 the US passed the 2002 Farm Bill, which is expected to be in place for the
next six years. This Bill retained the earlier support through various Loans, Flexibility Con-
tracts, and Insurance, as well as the Step-2 payment while it legitimized the emergency as-
sistance under the term counter-cyclical payments. If cotton prices remain at their 2001/02
levels, then US support to its cotton sector is expected to be on the order of $3.5 to $4.0 bil-
lion for the next six years, implying the US cotton producers will be receiving close to
twice the world market price.
The European Union
The origin of cotton intervention in the European Union goes back to the early 1980s when
Greece and Spain, the two European cotton producers, joined the Union’s Common Agri-
cultural Policy. During the 1960s, there were three cotton producers in Europe: Greece and
Spain which produced an average of 85,000 tons each and Bulgaria which produced 25,000
tons. Throughout the 1970s, Bulgaria’s output declined while Greece and Spain managed
to retain their cotton production at the levels experienced during the 1960s. Cotton pro-
duction by the three countries taken together, experienced an annual decline of 0.4 per an-
num between 1960 and 1982. With the EU’s expansion and the subsequent accession of
Greece and Spain, cotton production grew by an annual average of 7.3 percent. During the
1990s, Greece’s and Spain’s cotton output averaged 325,000 and 78,000 tons, respectively
(see figure 8).
— 14 —
Under the Common Agricultural Policy, support is given to cotton growers based
on the difference between the market price and a guide (i.e. support) price. Advance pay-
ments, which are made to ginners who pass the subsidy to growers in the form of higher
prices, are on estimates of seed cotton production. The policy also influences the quantity
of cotton produced by a maximum guaranteed quantity of seed cotton for which assistance
is provided—782,000 tons of seed cotton for Greece and 249,000 for Spain, approximately
equivalent to 255,000 and 82,000 tons of cotton lint.
The European Union reformed its cotton program in 1999 (European Commission
2000). While the guide price level and the maximum guaranteed quantity of seed cotton
for which assistance is provided have been maintained, “penalties” (i.e. reduction in sub-
sidy) for excess production over the maximum guaranteed quantity increased. Under the
reformed policy, for each 1 percent of excess production, the level of subsidy is lowered by
0.6 percent of the guide price as opposed to 0.5 percent prior to 1999. As production in-
crease, the “penalty” becomes stiffer, effectively, putting an upper limit on the budgetary
outlays to the cotton sector. It is important to note that this quantitative restriction (the so-
called maximum quantity guaranteed) applies at the aggregate (i.e. country) level imply-
ing that when this restriction is converted to individual basis, it creates not only adminis-
trative complexities but also leads to misallocation of resources. Karagiannis and Pantzios
(2002) found that the current system failed as a surplus containment mechanism and also
resulted in farm income losses.
Between 1995/96 and 1999/2000 the budgetary expenditure on cotton aid ranged be-
tween €740 and €903 million, implying that, on average, EU cotton producers received
more than twice the A Index—the world price of cotton. Note that even in periods of high
prices, EU cotton producers receive support since the amount allocated to the cotton sector
must be disbursed. In addition to output subsidies EU cotton producers receive subsidies
on inputs such as credit for machinery purchase, insurance, and publicly financed irriga-
tion.
China
China is currently the largest producers, consumer, and stockholder of cotton. China’s cot-
ton sector became fully government-controlled in 1953 after the introduction of the first
Five-Year Plan (Zhong and Fang 2003). The central planning policies adopted then were
similar to those of the Soviet Union and remained in place for the next 35 years. The cen-
tral government set production targets and procurement quotas. This monopoly was eas-
ily exercised because all ginning facilities were owned by the Cooperatives. A step to boost
cotton production was taken in 1978 by increasing the price of cotton as well as supplying
more fertilizer. A second boost came in 1980 with the partial abolition of the communal
production system under the Household Responsibility System which gave land use rights
to individual farmers. The Chinese cotton sector consists of primarily smallholders with
plots ranging between 0.5 and 1 hectares.
— 15 —
Evidence suggests that the government of China protects its cotton sector through
support prices, import tariffs, export subsidies, and public stockholding. The government
sets a reference price for cotton, typically above world prices. China also maintains tariffs
on imports that bridge the gap between domestic and world prices. Following its WTO ac-
cession arrangements the tariffs will be reduced to 15 percent but at the same time a tariff-
related quota system will be implemented to manage imports.
The International Cotton Advisory Committee found that support to the cotton sec-
tor in the six seasons beginning in 1997/98 ranged from $0.8 to $2.6 billion. Huang, Rozelle,
and Chang (2003) estimated that during 2001 the nominal rate of protection for cotton av-
eraged 17 percent. Fang and Beghin (2003), however, estimated that between 1997 and
2000, the nominal protection coefficient for cotton has averaged 0.80, implying that China
taxes its cotton sector by 20 percent. The different views on the nature and degree of inter-
vention, however, should not be surprising given the complexities of China’s agricultural
policies as well as the unreliability of the data.10 However, one may conclude that China
subsidizes its cotton sector as indicated by Jinglin (2003, p. 99):
In order to assure the interests of the farmers and to promote their stable incomes, China is
actively exploring the direct subsidy system for the bulk of agricultural products such as
grain and cotton, and having the existing indirect subsidies adjusted to become direct subsi-
dies and changing the subsidies meant to meant previously for the distribution enterprises to
the subsidies directly for the farmers. Nevertheless, such subsidy method and standard will
not surpass the framework of the WTO rules.
In September 1999, the government of China announced reform measures which in-
cluded: (i) the creation of a cotton exchange to facilitate domestic spot trading; (ii) the re-
duction of prices paid to producers; and (iii) a reduction in stocks. In some sense the re-
forms have worked: China’s stocks declined from 4.1 million tons in 1998/99 to 2.3 million
tons in 2000/01.11 In September 2001 further reforms were announced and are currently
under way (Zhong and Fang 2003). First, the internal cotton market would be open to
cross-regional trade. Second, various enterprises would be allowed to buy cotton directly
from producers with approval granted by the provincial government. Third, ginning op-
erations would be separated from marketing cooperatives in effect making them commer-
cial enterprises.
Uzbekistan
Uzbekistan, the world’s fifth largest cotton producer and second largest cotton exporter,
produces more than one million tons of lint, most of which is exported. During 1998-99
cotton exports accounted for one third of total merchandize exports while the sector con-
tributed an average of 6.4 percent to the country’s GDP. Prior to 1991 all aspects of Uz-
bekistan’s cotton sector were under state control (of the Soviet Union). Most cotton was ei-
ther consumed by mills in Russia (then considered domestic trade) or shipped to Eastern
European countries under barter arrangements. Following the collapse of the Soviet Un-
— 16 —
ion, Uzbekistan begun exporting its cotton to Western countries in exchange of foreign
currency (until 1996 some cotton still went to Russia in barter trade terms).
Although 12 years have passed since the change in the trade regime, most aspects of
production, marketing, and trade of the sector closely resemble pre-1991 arrangements.
There are numerous entities involved in all post-production activities of cotton. The three
most important ones are: (i) the state company handling ginning; (ii) the State Trading Or-
ganizations handling exports; and (iii) the Ministry of Foreign and Economic Relations
handling financial transactions.
All pre- and post-ginning operations of cotton are handled by UzKhlopkoprom/
UzPakhtasanoitish (UKP), a state company which used to be a ministry. UKP is responsi-
ble for collecting, storing, ginning, and classifying cotton, making payments to growers,
and providing inputs. UKP owns considerable assets, including all ginning and storage fa-
cilities as well as handling machinery and equipment.
The second important entities are the three state trading organizations (STOs) in
charge of handling all aspects of cotton exports. The main responsibilities of these organi-
zations include contracting cotton merchants for the sale of cotton, organizing the avail-
ability and shipment of cotton, receiving payments and converting them into local cur-
rency, and paying UKP. Although these organizations have a number of other responsi-
bilities (e.g. purchasing machinery and equipment on behalf of the government) exporting
cotton is their core activity. Because each organization has been allocated a quota of cotton
to be exported, there is no competition involved in the export process.
The third important entity is the Ministry of Foreign Economic Relations which re-
ports directly to the government. Its main function is to manage cotton export operations,
including setting prices, selecting buyers, and monitoring dollar receipts. There are a
number of other entities involved in the sector such as the state company responsible for
domestic and international transportation of cotton, the organization responsible for qual-
ity monitoring, and the customs.
It appears that cotton growers are heavily taxed both directly through the lower
price received by UKP (which, in turn, receives a fixed price by the STOs, as dictated by
the Ministry) and indirectly through the exchange rate regime. A recent study found that
at an ex-ginnery price of $1.03/kg, the STOs receive the equivalent of $0.63/kg (these calcu-
lations were based on an A Index of $1.24/kg). With respect to the difference between $1.03
and $0.63/kg, the study concluded: “It is not clear exactly where this profitability figure is
allocated. It is alleged that, after a marketing fee is deducted, the balance is paid to the
Ministry of Finance as an export duty.” The declared price to be paid to farmers by UKC is
126,000 Cym/ton of seed cotton, which, at an exchange rate of 960 Cym/$ and 32 percent
ginning out-turn ratio implies a price of $0.41/kg, about one third of the A Index.
Perhaps, it is not unreasonable to conclude that, apart from the fact that cotton ex-
ports from Uzbekistan moved from a barter to a commercially oriented structure, the sec-
tor is still tightly controlled by the government. Moreover, growers are taxed heavily, re-
— 17 —
ceiving only about one third of the export price of cotton.
PREFERENTIAL ARRANGEMENTS AND THE COTTON MARKET
Unlike the markets for commodities such as cocoa, coffee, and rubber, there has been no
United Nations-backed international price stabilization scheme or stockholding mecha-
nism in the cotton market. The cotton market, however, may be indirectly affected by two
arrangements: the Agreement on Textiles and Clothing and the US African Growth and
Opportunity.
The Agreement on Textiles and Clothing
The Multifibre Arrangement (MFA) was an agreement among developed country import-
ers and developing country exporters of textiles and apparel to regulate and restrict the
flow of trade of textiles. Negotiated in 1973 under the auspices of General Agreement on
Tariffs and Trade (GATT), the MFA replaced a number of bilateral textile agreements as a
temporary exception to the rules that otherwise would apply. The MFA (and its predeces-
sor agreements) influence the location of the textile industry thus increasing the costs of
textile products—in addition to other welfare implications.
In 1995, the MFA was replaced by the Agreement on Textiles and Clothing (ATC).
Under the ATC, import quotas in textiles will come to an end by January 1, 2005 and im-
porting countries will no longer be able to discriminate between exporters. Their elimina-
tion is expected to encourage the relocation of textile processing facilities from developed
to low cost (mostly developing) countries, reduce the cost of producing textiles, increase
cotton demand, and consequently raise cotton prices.
The African Growth and Opportunity Act
The African Growth and Opportunity Act (AGOA) provides African countries with the
most liberal access to the US market available to any country with which the United States
does not have a free trade agreement and covers the 8-year period from October 01, 2000
to September 30, 2008.12 One condition is that these countries must have undertaken sub-
stantial policy reforms. The Act, which was part of the US Trade Development Act of 2000,
was signed by President Clinton on May 18, 2000 while President Bush signed amend-
ments—known as AGOA II—into law on August 6, 2002, expanding the preferential ac-
cess for imports from the eligible countries.
Although AGOA sets stringent conditions, it offers potentially vast benefits for
qualifying African countries because it includes a broad list of commodities. However, one
should note that AGOA is a two-part Act: one applying to non-textile/clothing sectors and
one applying to the textile/clothing sectors. The conditions for the latter are much more
stringent than the conditions for the former (AGOA 2002). Specifically:
The Act provides for duty-free and quota-free access to the U.S. market without limits for
— 18 —
apparel made in eligible Sub-Saharan African countries from U.S. fabric, yarn, and thread.
It also provides for substantial growth of duty-free and quota-free apparel imports made from
fabric produced in beneficiary countries in Sub-Saharan Africa. Under AGOA I, apparel
imports made with regional (African) fabric and yarn are subject to a cap of 1.5% of overall
U.S. apparel imports, growing to 3.5% of overall imports over an 8 year period. AGOA II
doubles the applicable percentages of the cap.
Beneficiary countries for apparel exports must establish an effective visa systems to
ensure compliance with rules of origin, specifically to prevent illegal transshipment and
use of counterfeit documentation, and show that they have instituted required enforce-
ment and verification procedures.
Currently, the average duty to garment imports into the US is 17.5 percent. Under
AGOA, apparel imports into the US from the 14 eligible African countries will be duty free
subject to an upper limit of 3 percent of total US apparel imports. Since total trade in cloth-
ing from Africa to the US is very small, the 3 percent cap (to increase to 7 percent by year
7) is unlikely to become a binding constraint. Another beneficial provision is that for coun-
tries with Less Developed status, there is a 4-year exception to the rule of origin (it expires
in 2004). For example under this provision Tanzania can import yarn from China and ex-
port cloth to the US.
IMPACT OF DISTORTIONS AND PROSPECTS OF REFORMS
Impact of Distortions
On the textile side, Martin (1996) estimated that the MFA imposes an implicit tax of about
20 percent on cotton products relative to synthetic fiber products. A more recent study by
Quirke (2002) simulated the impacts of full implementation of ATC and found that the
world price of cotton will increase by 4 percent (the equivalent of 4 cents/kg at 2001 world
cotton prices), if the ATC is phased out as planned.
On the cotton side, the International Cotton Advisory Committee concluded that in
the absence of direct subsidies, average cotton prices during the 2000/01 and 2001/02 sea-
sons would have been 17 and 31 cents/pound higher. If the US alone removed its subsidies
during these two seasons, world cotton prices would have been 6 and 11 cents higher, re-
spectively. These figures imply cotton prices 30 and 71 percent higher than the actual av-
erages of 57.2 and 41.8 cents/pound. The study, which is based on a short run partial equi-
librium analysis, does acknowledge that while removal of subsidies would result in lower
production in the countries which receive them (and hence higher prices in the short
term), such impact would be partially offset by shifting production to non-subsidizing
countries in the medium to longer terms; similarly higher prices are likely reduce the
growth of cotton consumption making the long-run impact less striking.
Quirke (2002) estimated that removal of production and export subsidies by the US
and the EU are likely to induce a 20 percent reduction in US cotton production, 50 percent
— 19 —
reduction in US cotton exports, with much higher figures for the EU. He also estimated
that if support was not in place, world cotton prices would be 10.7 percent higher com-
pared to their 2001/02 levels. Note that both the ICAC analysis and Quirke (2002) used the
production and export support levels reported by ICAC (see Appendix F, table F11 for
data details).
Based on a partial equilibrium model, Tokarick (2003) found that multilateral trade
liberalization in all agricultural markets (including cotton) is expected to induce a 2.8 per-
cent increase in the world prices of cotton, with 0.8 percent coming from the removal of
market price support and the remaining 2 percent coming from the removal of production
subsidies (removal of market price support most likely applies to the US Step-2 payment).
Tokarick also calculated that global reforms will lead to $95 million in total change in wel-
fare per annum.
FAPRI (2002) found that under global liberalization (i.e., removal of trade barriers
and domestic support of all commodity sectors), the world cotton price would increase
over the baseline scenario by an average of 12.7 percent over the 10-year period (see table
3). The largest gains in trade go to Africa which would increase its exports by an average
of 12.6 percent. Exports from Uzbekistan and Australia increase by 6.0 and 2.7 percent, re-
spectively while exports from the United States decline by 3.5 percent. The most dramatic
impact is on the production side where the European Union’s cotton output would decline
by more than 70 percent. The latter outcome should be a complete surprise considering
that the European Union’s cotton output during the late 1990s was, on average, three times
as much as it was before CAP took effect on the cotton sector.
Prospects of Reforms by Major Producers
Prospects for policy reforms by major producers subsidizing the sector are mixed. Support
for cotton in the European Union is likely to remain at current levels. It is unlikely to in-
crease for two reasons. First the countries expected to join the Union are not cotton pro-
ducers and hence there will be no budgetary pressure. Second, the current support scheme
is, effectively, subject to an upper limit which appears to be a binding constraint as both
Greece and Spain, being among the world’s highest-cost cotton producers, are unlikely to
increase production given the reduced support they would receive if they exceed the cur-
rent output levels. On the other hand, support is not expected to be eliminated because it
supposedly goes to low income areas and hence it is regarded as a poverty reduction pro-
gram. The nature of support, however, may change by shifting away from direct price
support towards partially decoupled payments according to a recent EU proposal (see box
1). The conditions of success and implications of the shift to decoupled support are dis-
cussed in the last section of this paper.
The US took a step in the right direction with the replacement of the deficient pay-
ment system by decoupled payments in 1996 but all progress was eliminated with the 2002
Farm Bill which effectively: (i) legitimized emergency payments introduced in 1998/99 fol-
— 20 —
lowing the sharp decline in prices; (ii) renamed them to counter-cyclical payments; (iii) in-
creased target prices; and (iv) made it more convenient for larger farmers to increase the
support they receive. Historically, US farm bills either give what they promise or give
more than what they promise (as the recent experience showed). Hence, if history is any
guide, it is reasonable to expect that US cotton farmers will be receiving generous support
for the next 6 years, unless the support exceeds WTO commitments.
However, a number of factors may induce some early reforms. First, the substantial
increase of the support to the US cotton sector along with 30-year record low prices and
the fact that 10 percent of US cotton growers receive 90 percent of the support (hence falsi-
fying the claim that support preserve the small farm), is likely to put pressure for altering
the nature of policy sooner. Second, Brazil’s request for consultations at the WTO regard-
ing US cotton subsidies may create some pressure to lower subsidies (WTO 2002). Third,
four West African cotton producing countries (Benin, Burkina Faso, Chad, and Mali)
pressed for removal of support to cotton sector through the WTO. In an unusual move, the
President of Burkina Faso addressed the WTO on June 10, 2003, asking for financial com-
pensation for cotton producing low income countries to offset the injury caused by sup-
port. This compensation, according to the request, should be in place for as long as subsi-
dies are in place.
China appears to be the most promising case of reform. The reforms undertake in
1999 and more recently in 2001 indicate that its cotton sector will be soon exposed to inter-
nal and external competition. China is also in the process of establishing a cotton futures
exchange, indicating that market forces within the sector are likely to play a more signifi-
cant role in the future (Shuhua 2003).
On the international side, while the phase-out of the ATC is supposed to end the
distortions imposed on the location of the textile industries, it is uncertain whether the ex-
pected benefits will be fully realized. First, ATC is back-loaded with most of the reforms
expected to take place in the last year, thus increasing the risk of non-compliance. Second,
a number of (mainly EU) countries have repeatedly sought to impose antidumping duties
on textile imports from Asia in recent years. Third, there are a number of provisions under
the ATC which allow for the imposition of temporary duties in the case that the currently
domestic textiles suffer “significant damage” following the phase out.
REFORM INITIATIVES IN AFRICA
During the 1990s, a number of African cotton producing countries undertook substantial
reforms. The reform process and its outcome have been studied extensively. See, for ex-
ample, Kähkönen and Leathers (1997) for Zambia and Tanzania; Sabune (1996) and Lund-
bæk (2002) for Uganda; Larsen (2002) for Zimbabwe; Baffes (2002) for Uganda, Zimbabwe
and Tanzania; Baffes (2000), Badiane et al. (2002), and Goreux and Macrae (2003) for Fran-
cophone Africa; Baffes (2002) and Gibbon (1998) for Tanzania. Poulton et al. (2003) looked
at the cotton sectors of six African countries while Shepherd and Farolfi (1999) reviewed
— 21 —
export commodity sectors for a number of sub-Saharan African countries. The remainder
of this section draws mainly from these studies and summarizes the reform experiences in
the cotton sectors of Uganda, Zimbabwe, Tanzania, as well as the efforts currently under
way in Francophone Africa.
Uganda
Cotton was introduced to Uganda early in the 20th century, and production grew rapidly
until the mid-1930s, when coffee began to compete as an alternative cash crop. Cotton
production continued to increase steadily, peaking during the early 1970 at 75,000 tons,
making it the third largest African cotton producer after Sudan and Egypt. Most activities
in the cotton industry were administered under monopolistic arrangements. The Ministry
of Agriculture had responsibility for cotton research and seed multiplication. Responsibil-
ity for cotton seed for planting and oil milling fell to the Lint Marketing Board (a state en-
terprise), which was also responsible for lint marketing (both domestic and export) and for
regulating the industry. Primary marketing and processing were the responsibility of co-
operatives, which had their own society networks and ginning operations.
The political instability, poor governance, and inappropriate macroeconomic poli-
cies of the 1970s and 1980s had a devastating effect on the Ugandan economy and hit the
cotton sector especially hard. Cotton production collapsed, plunging to a low of 2,000 tons
in 1987. Seed multiplication activities were disrupted, as were research and extension. Co-
operatives failed to pay farmers cash for their cotton, and inefficient ginning marketing
and operations generated high overhead costs.
In 1992, with World Bank assistance, Uganda embarked on a major reform program
that included the liberalization of the cotton industry (World Bank 1994). The government
redefined its role in the cotton industry taking on some new responsibilities (especially
during the transitional phase) and shedding others. Ginning and the marketing of cotton
and cotton inputs were liberalized, and research, seed multiplication, and extension ser-
vices were strengthened.
At the time of the reforms the cooperatives were crippled by bad debt. Potential
ginning capacity was much greater than actual capacity, and run-down ginneries needed
infusions of new capital. Management knew little about how to restructure their facilities,
however. To address the problems, the government established the Business Advisory
Service, a temporary agency that worked with cooperatives to draw up new business
plans. In return for restructuring their businesses, the ginneries would receive limited debt
relief. Temporary lines of credit were established through the Bank of Uganda’s Develop-
ment Finance Department to provide working capital for the restructured firms. Some co-
operatives were able to sell assets and finance smaller, more efficient business. Others en-
tered into joint ventures with foreign partners.
Government participation in the cotton sector after the reforms takes place almost
entirely through the Cotton Development Organization. The Cotton Organization repre-
— 22 —
sents the cotton industry as a whole and to monitor the production and marketing of cot-
ton. A 12-member board of directors which includes public and private sector representa-
tives governs the organization. Among other things the board approves expenditures, sen-
ior staff appointments, procurement procedures, and business plans. An auditor-general
reviews the Organizations’ accounts and by law must report the findings to the legislature.
To carry out its mandate, the Cotton Organization can charge for its services, bor-
row, manage property, and levy a cess (local tax). The initiating statute also placed explicit
limits on the Organization’s authority, however. The agency cannot levy a cess of more
than 2 percent, although the Ministry of Agriculture, which is responsible for the Organi-
zation, can vary or rescind the cess by statutory instrument. The types of penalties the Or-
ganization can levy for noncompliance are limited. Further, it is obliged to give all new en-
trants registration permits, even if they have not previously been engaged in the cotton in-
dustry. Businesses can renew their registrations automatically by paying the fees.
In many respects the cotton reforms in Uganda have been successful. During the
eight-year period staring in 1995/96 cotton output in Uganda has averaged 17,000 tons, an
almost three-fold increase compared to the eight seasons prior to 1995/96. The correspond-
ing A Index average for before and after 1995/96 was $1.56 and $1.40 per kilogram. The
farmers’ share in world prices rose from less than 50 percent to 70 percent after the re-
forms while a number of new traders and exporters entered the sector. This success came
despite the failure of most credit mechanisms that have been launched after the reforms.
Zimbabwe
Commercial production of cotton in Zimbabwe started in the early 1920s while a compre-
hensive cotton research program along with a research station was set up in 1925. Ad-
vanced technology through insect control and the development of improved seed varieties
increased production turning Zimbabwe into an important cotton producer in Africa. Ini-
tially, cotton marketing was the responsibility of a committee under the Grain Marketing
Board. The Cotton Marketing Board was established in 1969 and handled cotton market-
ing.
The Board controlled most aspects of cotton production until 1994, from the sale of
planting seeds to the purchase of cotton from farmers. Ginning and the marketing of cot-
ton and cotton seed fell under its purview. The Board had eight ginneries and was the sole
buyer of cotton (the country’s single private ginnery had to buy cotton from the Board on
contract). The Board also regulated the industry. It announced producer prices well in ad-
vance, so that the state absorbed all of the price risk.
The Cotton Board grew into an inefficient organization with poor governance and
high operating costs. It developed financial difficulties because of weak management and
subsidized cotton sales—often at half the international price—to the domestic textile in-
dustry. Cotton production fell by almost half during the 1980s. Producers were not paid on
time and often did not receive full payments. By the late 1980s it had become clear that the
— 23 —
board would have to be restructured or the cotton industry would collapse. A severe
drought in 1991–92 contributed to the sector’s woes, causing a further 60 percent decline in
production.
Zimbabwe appointed private sector representatives to the Cotton Board in 1992,
leaving just one government representative. The Board’s mandate at the time was to de-
velop a privatization plan for all aspects of cotton trade and marketing. Various regulatory
controls (such as seed quality regulations and cotton grading) were transferred from the
Board to the Ministry of Agriculture. In 1993 the government announced that the cotton
market would be open to new entrants, bringing the Cotton Board’s monopoly to an end.
In 1994 all subsidies to the textile industry were discontinued.
In July 1994 the Cotton Board began having difficulty paying for cotton, and a
number of commercial growers started to buy their seed cotton and have it ginned at the
only private gin. In September 1994 the Board’s monopoly was formally terminated, and it
became the Cotton Company of Zimbabwe, with the government holding all shares. The
government assumed all of the Cotton Company’s debts, allowing the agency to start out
with a clean balance sheet, and discontinued all subsidies to the textile industry.
Private companies have moved into ginning and marketing in the country. As of
1994 the Cotton Company still owned 80 percent of the ginning capacity in Zimbabwe and
operated a network of buying centers and collection points throughout the major cotton-
growing areas. The Commercial Cotton Growers Association, a cooperative owned by
growers who farm 25 hectares or more, joined with an international cotton company to
form a new firm, Cotpro, that provides competition for the Cotton Company
In 1995 the Cotton Company leased two of its gins to Cargill, a U.S.-based agribusi-
ness. Cargill started buying seed cotton, putting itself in direct competition with the Cot-
ton Company and Cotpro. In 1996 it bought the two gins it had been leasing from the Cot-
ton Company. In October 1997 the Cotton Company was privatized. The government
holds 25 percent of the shares, small-scale farmers 20 percent, institutional investors and
the general public 15 percent each, large-scale farmers and the National Investment Trust
10 percent each, and employees 5 percent.
Following reforms the cotton industry improved in several ways. First, cotton pro-
duction is up substantially. During the eight seasons since 1995/96, cotton output has av-
eraged 115,000 tons, 50 percent higher than the eight-period average prior to 1995/96.
Some 30 percent of the 1997/98 cotton harvest was marketed entirely by private entities.
Private companies now transport most cotton. Competition has pushed the price farmers
receive to close to 80 percent of international prices, and producers are being paid faster.
Zimbabwe has also retained its premium it used to receive in the world market.
Despite its success, the cotton sector of Zimbabwe is currently going through major
difficulties which are beyond the sector’s control. Political and macroeconomic instability,
and uncertainty over land issues has reduced substantially investment in the sector. The
sector is also suffering from implicit taxation through inflation and exchange rate mis-
— 24 —
alignment. During 2002, for example, the official exchange rate was fixed at Z$45 per US$.
However, unofficial reports indicated that it was traded as high as Z$400 in the parallel
market.
Tanzania
Cotton was introduced to Tanzania around 1904 by German settlers as a plantation crop,
but the attempt failed. During the 1920s new efforts focused on smallholder production,
first in Eastern and later in Western Tanzania. Local research during the 1930s led to the
development of a local pest-resistant variety. Cotton output, especially in Western Tanza-
nia, rose considerably with the releases of these local varieties, along with better organiza-
tion of the sector following establishment of the Tanganyika Lint and Seed Marketing
Board in 1956. By 1966 Tanzania’s cotton output was 80,000 tons, or 0.75 percent of world
production of 10.7 million tons.
A turning point came in the 1960s following the spread of the cooperative move-
ment and deterioration of relations between ginnery owners (mostly Asians) and cotton
growers. Several hundred primary societies had sprung up, and the groups began han-
dling crop purchasing. Soon, they formed cooperative unions and began building ginner-
ies, training staff, and taking over ginneries and cotton oil mills from foreign owners. In an
attempt to correct the inefficiencies and poor management, the government abolished the
unions in 1976 and turned over cotton marketing to the Tanzanian Cotton Authority, the
successor of the Lint and Seed Marketing Board. The government set the prices paid to
farmers, establishing uniform national prices for an entire season. This marketing structure
also failed, and the cooperative unions were reinstated between 1980 and 1984. The unions
and primary societies acted as agents for the Tanzanian Cotton Marketing Board, the re-
named Tanzania Cotton Authority. The primary societies stored and sold cotton to the co-
operative unions for a fixed price, and the unions processed the seed cotton for a fixed
margin. The Cotton Board managed domestic and international sales. Because the coop-
erative unions were semipublic entities, they simply added another bureaucratic layer
rather than making a substantial contribution to value added. Most of the unions accumu-
lated huge debts and managed to survive only through government subsidies and donor
support.
The first steps toward cotton reform in Tanzania were taken in 1989/90, when the
government launched the Agricultural Adjustment Program. The program transferred
ownership of seed cotton from the Cotton Board to the cooperative unions, and the board
was converted into a fee-based marketing service for final sales and input purchases. Price
controls on cotton were gradually relaxed. In 1991/92 the government announced only in-
dicative prices, not fixed prices. The cooperative unions were free to determine their own
producer prices for the next season, although they chose to offer uniform prices through-
out the country.
The largest reforms came with the Cotton Act of 1994, when the government for-
— 25 —
mally eliminated the monopoly held by the board and the cooperative unions and allowed
competition in cotton marketing and ginning. At the time there were 14 regional coopera-
tive unions licensed to trade cotton. In 1994/95 some 22 private companies started trading
cotton, and 8 new private ginneries were constructed. That opened up another marketing
channel, especially in Western Tanzania. In Eastern Tanzania, where production was low
and some farmers had no buyers, the Tanzanian Cotton Lint and Seed Board (the new
name of the Tanzania Cotton Marketing Board as of 1995) acted as buyer of last resort.
By 1996/97 private businesses were purchasing almost half of all cotton. Private
traders and ginneries were able to capture a considerable share of the market because they
offered higher prices than cooperative unions and paid promptly. Some private ginneries
also engaged in contract farming, providing inputs (seeds and occasionally fertilizer) to
producers who agreed to supply cotton in return. The ginneries and producers usually es-
tablished a minimum price at planting time, but the price could be adjusted if the market
price was higher during the harvest.
The outcome of cotton reforms in Tanzania has been mixed. On the positive side,
the share of producer prices increased to 51 percent (from 41 percent prior to the reforms).
Furthermore, growers farmers receive payments quickly, a major achievement compared
to 6-month delays encountered prior to the reforms. And, contrary to what many reports
show, quality of cotton appears not to have suffered considerably. On the other hand, cot-
ton production after 1995/96 has averaged less than before reforms (55,000 and 61,000 tons,
respectively). On the policy side, the Cotton Board along with the two line Ministries (Ag-
riculture and Food Security as well as Cooperatives) still play a major role in the sector
which goes far beyond the regulatory role they are supposed to play. Collection and dis-
semination of data (as well as accuracy of statistics) are poor even by the government’s
own admission.
Francophone Africa
The modern cotton industry in West African countries was pioneered by the French state-
owned Compagnie Française de Développement des Fibres Textiles (CFDT). As countries
gained independence they established their own national cotton companies, but the CFDT
retained a minority shareholding position (usually holding around one-third of shares) in
these companies and entered into technical agreements with them. It also retained an
ownership interest in companies engaged in processing cotton by-products. The national
cotton companies have a legal monopsony in seed cotton, and most have a monopoly in
ginning, marketing, and providing inputs to farmers. They announce a base buying price
for seed cotton before planting starts, sometimes supplementing this price with a second
payment. Any supplement is based on the company’s financial results for the season and
is paid on production the following season. Village producer associations handle interme-
diate input credit and seed payments, and input credits are deducted from payments for
cotton.
— 26 —
Aided by research in Africa and France, cotton growing expanded rapidly in West
Africa, increasing more than fourfold in the past 25 years. During the 1999/ 2000 season,
West and Central African countries produced 930,000 tons, or 5 percent of world produc-
tion. The region is the third-largest cotton exporter after the United States and Uzbekistan,
accounting for almost 15 percent of world exports. Farmers use chemical inputs and seed
varieties well adapted to local conditions to produce high yields and consistent quality
cotton. The industry is bolstered by high repayment rates for input credits and well-
organized producer associations.
The system has a number of weaknesses, however. The prices West African pro-
ducers receive tend to be very low. For example, farmers in Zimbabwe and India received,
respectively, 37 and 60 percent more for similar types of cotton from 1983 until 1995. Since
1994 (when the CFA franc devaluation took place), prices in Zimbabwe and India have
been 80 percent to 100 percent higher than prices in West Africa. The services the cotton
companies provide (extension services, rural road maintenance, and transportation to
move seed cotton to gins) are only partially responsible for the price differences. A large
part of the gap between domestic and export prices of cotton has been absorbed by gov-
ernments in the form of various taxes (export taxes, taxes on parastatal profits, etc.) This
money has, in part, been used to subsidize other groups, especially domestic textile firms
through low prices charged for cotton, and low prices for cotton seed supplied to the do-
mestic oil and meal companies. The absence of any competition in domestic markets has
often allowed costly operating inefficiencies on the part of the parastatal companies, espe-
cially in the export marketing of lint. More generally, the system creates opportunities for
rent seeking and corruption, generally at the cost of farmers and the economy.
The determination of annual cotton prices reflects the relative bargaining power of
a number of groups; namely, producers, governments, managers of the state-owned cotton
companies, the CFDT, and in some cases private ginning firms. Pan-territorial pricing of
cotton and farm inputs means that transport costs are not properly taken into account in
the determination of where cotton is grown. Furthermore, the uniformity of cotton input
and output prices across entire countries, effectively transfers resources from producers
who are close to ginning or distribution centers to those who are further away. Farmers
have limited autonomy as regards decisions on the types and quantities of the seeds and
other inputs they use. Pan-seasonal pricing of seed cotton and planned delivery schedules
to ginning plants severely limit farmers’ choices on holding seed cotton inventories. The
system also restricts the use of seed cotton as collateral to borrowing for cotton inputs
only, which underutilizes the potential of cotton production to support rural credit flows.
Finally, the system does not respond flexibly to changes in world market conditions. For
example, in the late 1980s and early 1990s low world prices and an overvalued currency
led to the de facto bankruptcy of a number of state cotton companies. The companies had to
be drastically restructured, supported by injections of money from national governments
and international aid organizations.
— 27 —
During the past several years, in conjunction with the AFD (Agence Française de
Développement), the World Bank has held intensive discussions with the governments
and other stakeholders in West and Central Africa, including the cotton parastatals, CFDT,
and input suppliers. These discussions resulted in two broad proposals: retaining the cot-
ton companies but reforming and regulating them, and introducing free entry and compe-
tition.
The first proposal involved a number of steps including:
setting prices at levels appropriate to a competitive environment;
giving producers equity in the national cotton companies and more influence over
key decisions, especially price setting;
subcontracting activities such as providing inputs and transportation to private
firms; and
eliminating subsidies on sales of cotton lint and cotton seed to domestic textile firms
and oil mills.
The advantage of this proposal is that it reduces the risk of damaging the current
system—which has many desirable aspects—with more far-reaching reforms. Maintaining
the current system’s ability to recover research and extension costs and its high repayment
rates on input loans is especially important. The proposal has two weaknesses, however.
First, domestic prices are unlikely to move in line with world prices (a supposed goal of
the reform process), because large shares of national income are at stake. The price-setting
mechanism has been (and is likely to remain) political because a number of interest groups
are involved in the negotiations. Second, the proposal is incompatible with initiatives to
establish free trade among countries in the region under the two regional arrangements,
(Union Economique et Monétaire Ouest-Africaine, UEMOA and Communauté Economi-
que et Monétaire de l’Afrique Centrale, CEMAC). New trade arrangements would require
reforming the cotton industry again.
The second proposal, which involves free entry calls for:
opening the sector to competitive entry at all levels and hence linking domestic
prices to international prices which would vary according to transportation costs
and the season;
maintaining and strengthening research, extension, and phytosanitary regulations
where the government has an essential role;
strengthening farmer groups and facilitating their participation in voluntary contract
farming arrangements;
freeing the cotton industry from sector-specific taxation and subjecting it only to
economy wide taxes; and
increasing the efficiency of regional ginneries by harmonizing reforms of cotton
trade across West African cotton zones.
— 28 —
In some countries free entry may be all that is needed to generate a competitive system. In
other countries privatization of the national cotton companies and subsequent restructur-
ing into a number of successor companies may be necessary. Such a move would signal
the government’s commitment to open markets and ensure that producers in every region
have access to competitive markets for their seed cotton.
The World Bank has argued that the discipline and responsibility that a free-entry
competitive system imposes on market participants would make for a more resilient, flexi-
ble, self-reliant, and more innovative national cotton sector. Improved competition
through market reforms offers important opportunities for regional trade and cooperation,
the latter in areas such as research, phytosanitary regulations, and seed development and
certification. Most importantly, improved sector performance would contribute to alleviat-
ing poverty by raising cotton prices to levels enjoyed by farmers elsewhere in the world.
Significant developments have taken place during the last few years, which indicate
the future direction of institutional changes in the region’s cotton sector. Three producing
countries, Benin, Côte d’Ivoire, and Togo, have now opened their sector to private ginners.
Benin and Côte d’Ivoire, with a combined production of about 290,000 tons in the
1999/2000 season, have eliminated the monopoly power of their national companies and
transferred key responsibilities to the private sector.
— 29 —
PART III: THE ISSUES
SYNTHESIS OF ISSUES AND STYLIZED FACTS
A number of market- and policy-related issues and stylized facts emerge from the preced-
ing analysis.
Cotton is very important to a number of low income African and Central Asian
countries, in some cases contributing as much as 40 percent to merchandize exports
and between 5 and 10 percent to GDP. Considering that in most countries cotton is a
smallholder crop, the implications of price changes (either induced by market forces
or policy interventions) as well as changes in market share are enormous. For exam-
ple, a 40 percent reduction in price (i.e. equivalent to the price decline that took place
from December 2000 to May 2002) implies a 7 percent reduction in rural income in
Benin—a typical cotton producing country in West Africa.
Cotton faces intense competition from chemical fibers, especially following techno-
logical improvements of the early 1970s which brought their prices down to cotton
price levels. Since 1975, polyester and cotton have been traded at roughly the same
price levels. Currently, the share of cotton in total fiber consumption is 40 percent
(down from 68 percent in 1960).
Consumption growth of cotton has been the same as population growth during the
last 40 years, implying that if past trends continue, cotton consumption at the end
than the current decade will be about 23 million tons (up from the current level of 21
million tons).
Cotton prices, as is the case with most primary commodities, have been declining in
real terms. Price variability during the post-1985 period has been 2.5 times higher of
what it was during the pre-1973 period and it is about half of what it was during
1973-84.
Minimal border restrictions but considerable domestic support. Major subsidizers
are the United States, $3.7 billion in 2001/02 and the European Union—Greece and
Spain—$0.7 billion (compare this to $20 billion, the value of world's cotton produc-
tion, evaluated at 2001 prices and quantities). This level of support implies that
prices received by US and EU cotton producers are 87 and 160 percent above world
prices. China has been reportedly supporting its cotton sector during the last few
seasons by an estimated $1.5 billion annually, but this figure cannot be substantiated
due to the complexity of Chinese policies and the quality of data (this does not nec-
essarily imply that China does not support its cotton sector).
Many cotton producing countries have reacted by introducing offsetting support.
Support in Turkey, Brazil, Mexico, Egypt, and India, totaled $0.6 billion during
2001/02.
Domestic support by major producers has triggered some noteworthy reactions.
Brazil initiated a WTO consultation process claiming losses to its cotton exports due
— 30 —
to subsidies by the United States. More recently, four West African cotton producing
countries (Benin, Burkina Faso, Chad, and Mali) pressed for removal of support to
cotton sector through the WTO. In an unusual move, the President of Burkina Faso
addressed the WTO on June 10, 2003, asking for financial compensation for cotton
producing low income countries to offset the injury caused by support. This com-
pensation, according to the request, should be in place for as long as subsidies are in
place.
The cotton sector has found an unlikely ally. The Director General of the Interna-
tional Rayon and Synthetic Fibres Committee in a letter to the Financial Times on June
12, 2003 complained that “recent increases in cotton subsidies have rigged the mar-
ket even more dramatically in favor of cotton, depressing demand for every substi-
tute product. The result is industrial plants being kept idle… that were built in le-
gitimate expectation that the competitive advantages of manufactured fibers would
create demand to fill the capacity…”
If support is completely removed (including full implementation of the Agreement
on Textiles and Clothing), cotton prices would be at least 15 percent higher than they
would have been in the absence of reforms.
The prospects for eliminating support by major producers are slim at best. The
United States introduced the 2002 Farm Bill, which is expected to be in place for the
next 6 years—historically Farm Bills have given more than what they promise, not
less. The European Union reformed its cotton policy in 1999. Its forthcoming expan-
sion does not affect the cotton sector since none of the new entrants is cotton pro-
ducer. A shift from price support to decoupled payments, however, if designed and
implemented properly, is likely to reduce production by the main supporters and
consequently boost world prices.
A number of East African cotton producers undertook reforms during the 1990s.
Research has shown that whenever the reform process was completed or there was
no backtracking supply response took place and growers received higher share of
f.o.b. prices, both considerable achievements in an environment of declining world
prices.
The World Bank and other institutions have advocated reforms in the cotton sec-
tors of West Africa during the last 6 years and extensive consultations have taken
place among the relevant stakeholders. The World Bank's position (which has been
clearly articulated on several occasions) is in favor of competition in the cotton sector
while retaining the positive elements of the current system, not outright privatiza-
tion.
RECOMMENDATIONS
Given the highly distorted nature of the cotton market and the fact that millions of rural
poor households in developing countries depend on this commodity, what are the alterna-
tives? On the demand side, cotton promotion is something that must be pursued by all cot-
— 31 —
ton producers. There are two encouraging sings regarding cotton promotion. First, the US
has an active cotton promotion program with an annual budget of about $60 million. The
main feature of this program is raising consumers awareness of cotton through the “Seal of
Cotton” campaign. According to Skelly (2003) there is a strong correlation between the
program’s advertising campaign and cotton’s market share in the US.
Second, an initiative was undertaken recently through the International Cotton Ad-
visory Committee to establish an international cotton promotion program. To that end, the
International Forum for Cotton Promotion (IFCP) was established in 2003. The principal
objective of the Forum is to encourage and facilitate national market development pro-
grams, organized by associations and commercial organizations in individual countries,
and funded from domestic resources. The Forum is to achieve this objective by serving as
a clearinghouse for the exchange of proven ideas and strategies to be implemented by na-
tional organizations, and by facilitating the establishment and expansion of national de-
mand enhancement efforts. While it is too early to assess the performance of the Forum, it
is certainly a step in the right direction.
Developing Countries
As discussed earlier, a number of developing countries, especially in sub-Saharan Africa
have undertaken policy reforms during the 1990s. Setting aside the lively debate on the
motives of the reforms, in many respects the reforms have been successful. For example, in
the few cases reviewed here, cotton growers received higher share of fob prices, they also
received payments more promptly, and there was considerable supply response. In an en-
vironment of declining commodity prices, these are not trivial achievements. However, in
a number of cases, the reform process has not been completed (Tanzania), it has been re-
versed (Zimbabwe), it has been slow (West Africa), or it has not even started at all (Uz-
bekistan). In these cases further reforms are the only feasible alternative.
A second issue that should receive attention is the enabling policy environment re-
garding the use of genetically modified cotton. In China for example, where GM cotton is
used extensively by small-holders, the costs of producing cotton declined by 20-25 percent.
This cost reduction meant doubling the net income for cotton growers. One should also
note that GM cotton has not been subject to negative consumer reaction as has been the
case with GM food products.
A third issue (and closely related to GM cotton) is organic cotton. Producers of or-
ganic products typically command significant premia. However, organic cotton produc-
tion has not been as profitable as other organic crops (such as coffee and tea). The main
reason is weak demand which appears to be a reflection of the “distance” between the
farm product—cotton—and the final product—cloth. It is because of this distance that GM
cotton has not faced resistance by the consumers, which further reinforces the conclusion
that GM cotton is something that developing countries should consider seriously.
— 32 —
Developed Countries
The price prospects (and consequently the export shares of low cost producers, including
many African countries) can be improved considerably if support by developed countries
is reduced substantially or eliminated altogether. However, given the low probability of
eliminating support, a second best alternative would be for support to be given in a non-
distortionary manner. A type of support with minimal distortionary effects—the so-called
decoupled support mechanisms—has re-gained popularity recently. Income transfers un-
der decoupled mechanisms are based on past production and prices and thus have no ef-
fect on current production decisions.
Decoupled support was attempted in the European Union with the Common Agri-
cultural Policy reform of 1992, in Mexico with the PROCAMPO program of 1994, and in
the United States with the Freedom to Farm Act of 1996. The outcome of these programs
has not been encouraging. The reason was that these schemes did not include three essen-
tial conditions that would make them successful: (i) substituting all support mechanisms
with decoupled support; (ii) limiting the duration of the programs which would have
made them true transition mechanisms and (iii) not requiring that resources remain in ag-
ricultural use which would reduce the overall supply of the commodities under considera-
tion and hence lift world prices (Baffes and Meerman 1998). Unless these conditions are
met, any attempts to restore the credibility of decoupled support policies and ultimately
remove support to the cotton (and other) sector(s) are unlikely to have the beneficial im-
pact intended.
What makes decoupled support in the cotton sector an interesting (and potentially
applicable) alternative is that almost all support comes in the form of domestic measures.
Therefore, changing the nature of support does not require changing the sources of fund-
ing as it would in the case of border measures.
Recently the EU proposed changes in its cotton policy (see Box 1). According to the
proposal, about €800 million (approximately equal to the current budgetary outlays of the
cotton program) will be allocated as follows: (i) €420 million on decoupled payments; the
only requirement for these payments is that producers must have planted cotton during
1999, 2000, and 2001; (ii) €280 million on area payments (if plantings exceed 425,360 hec-
tares, support per hectare will be reduced proportionately); and (iii) €100 million to be
spent on rural development measures. Although the proposal is a step in the right direc-
tion, two shortcomings must be highlighted. First, it unties 60 percent of support, not all
support, which is one of the main shortcomings of current decoupling schemes. Second, 40
percent of the support is linked to total area, thus not fully removing the incentive to
overproduce.
— 33 —
NOTES
1 The terms man-made and chemical fibers are used interchangeably in this paper. Often, the lit-
erature refers to man-made or chemical fibers as synthetic fibers, probably a reflection of the fact
that non-cellulosic fibers are also called synthetic polymers as opposed to cellulosic fibers whish
are called natural polymers.
2 The shares in world’s total fiber consumption in 1925 were: cotton, 84.2 percent; wool and flax,
13.5 percent; silk, 0.7 percent; and rayon, 1.6 percent. In 1946 they were: cotton, 72.6 percent; wool
and flax, 15.4 percent; silk, 0.2 percent; and rayon, 11.8 percent.
3 The only period when prices increased in a sustained manner was between 1730 (the earliest
price records) and 1790, 3 years prior to the invention of the ginnery (Baffes 2003).
4 The validity of these percentage changes depends on the stationarity properties of the variables
under consideration. In order to highlight the importance of this issue, all growth rates reported in
this paper have been estimated with OLS and all relevant statistics along with their interpretation
are reported in Appendix C.
5 Not all these technological improvements have been shared equally by all cotton producers. For
example, mechanical harvesting is being practiced in the United States, EU, Australia, and to some
extent in Brazil, representing approximately 30 percent of cotton production. The remaining cot-
ton is hand-picked.
6 The 1984/85 price decline is considered as a structural break. One may argue, however, that if the
policy shift in the US, which caused the massive de-stocking, was the main reason behind the price
decline, a new stock equilibrium level would cause a price increase. What is argued here is that the
policy shift accelerated the price decline which would have taken place in the long run. Real cotton
prices did rise later but never reached the pre-1984 levels.
7 Although GM cotton was first grown in the US, China was the first country to commercially
produce GM tobacco and tomatoes.
8 Carl et al. (2001) also found that the increased use of GM cotton in China was associated with
considerable positive health effects, i.e. fewer hospitalizations from pesticide poisoning. That was
expected as according to the survey farmers who did not use Bt cotton had to spray, on average 12
times, while farmers who used Bt cotton had to spray only 3 to 4 times.
9 The Step-2 mechanism is often referred to as export subsidy. Its objective is to eliminate the dif-
ference between the higher US domestic prices and world prices so the US cotton exporters main-
tain their competitiveness. Because it is a subsidy to millers, whose end product may either be con-
sumed domestically or be exported, the Step-2 payment is equivalent to a production subsidy.
Ironically, US cotton producers are subsidized in all three conceivable scenarios: (i) when prices
remain constant (through the production flexibility contracts); (ii) when prices fall (through the
loan rate and counter-cyclical payments); and (ii) when prices increase (through the Step-2 mecha-
nism.)
10 Following the September 1999 reforms, the Chinese Ministry of Agriculture in collaboration
with FAO holds conferences every two years in order for local officials to familiarize themselves
— 34 —
and exchange ideas concerning developments in the cotton market. These conferences have be-
come useful venues for information and data dissemination.
11 Chinese stocks have been subject to lively debate within the cotton industry, because (i) the ac-
curacy of the data is questionable and (ii) the quality of cotton stocks is unknown.
12 As of May 2002, there were 36 eligible AGOA countries. Cotton producing countries which have
been declared eligible for apparel provision under AGOA are: Benin, Cameroon, South Africa,
Uganda, Tanzania, Zambia. Chad, Côte d’Ivoire and Nigeria are eligible under AGOA but non-
eligible under the apparel provision. Notable exceptions from AGOA are Angola, the Democratic
Republic of Congo, and Zimbabwe.
— 35 —
Box 1:The 2003 EU Reform Proposal for Cotton
On September 23, 2003, the EU Commission proposed to reform its cotton, sugar, and tobacco
sectors. Under the cotton reform proposal, the EU support to the cotton sector will consist of the
following parts: (i) a single farm payment scheme, (ii) production aid scheme, granted as an
area payment, and (iii) development measures. The proposed reform is schedules to be effective
during the 2005 season. The projected annual budgetary expenditure for the new scheme will be
about €800 million, approximately equal to the current level of budgetary outlays. The first and
second components of the new program will absorb €700 million, 60 percent of which (€420 mil-
lion) is earmarked for the single payment and 40 percent (€280 million) for the area payment.
The only requirement for the single farm payment is that recipients must have planted cotton
during the 1999, 2000, and 2001 seasons, making it a fully decoupled payment. The area-
coupled payment is given in order to prevent production disruption in areas with a high eco-
nomic dependency on cotton.
The area payment would be given for a maximum area of 425,360 hectares (340,000 ha in
Greece, 85,000 ha in Spain and 360 ha in Portugal) and would be proportionately reduced if
payment claims exceed the maximum area of a Member State. It could be differentiated on the
basis of specific criteria, relating to the participation of producers in an inter-branch organiza-
tion which would be approved by Member States and subject to controls. A maximum of half of
the area payment to members of an inter-branch organization could be determined according to
inter-branch scales, rewarding production deliveries in quality and quantity terms. The activi-
ties of each inter-branch organization would be financed by its members and by a Community
grant of €10 per hectare. That support is expected to be about €4.5 million. The balance with the
total market expenditure for cotton will be included in a restructuring envelope for cotton areas.
This last component of the proposal (€100 million) would be shared between Member
States according to the average area eligible for aid over the reference period and would serve
an additional financial instrument within the second pillar of the CAP in order to fund rural
development measures. It may be used either for more beneficiaries, more measures, or even an
increased aid intensity of existing rural development measures.
Source: European Commission (2003).
— 36 —
TABLE 1
COTTON’S IMPORTANCE TO DEVELOPING AND TRANSITION
ECONOMIES: 1998-99 AVERAGES
COTTON EXPORTS
Million
US dollars
percent of mer-
chandize exports
percent
of GDP
MERCHANDIZE
EXPORTS
(million US dollars)
PER
CAPITA
GDPa
Burkina Faso 127 43.9 5.1 289 249
Benin 164 39.1 7.1 419 398
Uzbekistan 1,038 32.2 6.5 3,227 467
Chad 76 32.2 4.7 236 224
Mali 180 29.5 6.7 611 285
Togo 67 21.3 4.7 315 341
Tajikistan 97 15.1 8.2 643 352
Turkmenistan 110 12.3 3.6 891 1,126
Tanzania 44 7.6 0.5 576 185
Syria 214 6.7 1.4 3,177 858
Sudan 41 6.0 0.4 688 290
a. Constant 1995 US dollars.
Source. FAO (FAOSTAT) and World Bank (World Development Indicators).
— 37 —
TABLE 2
GOVERNMENT ASSISTANCE TO US COTTON PRODUCERS, 1996/97-2001/02
1996/97 1997/98 1998/99 1999/2000 2000/01 2001/02
Assistance (million US $)
Loan Deficiency Payments 0.0 6.0 320.7 687.3 151.4 732.1
Marketing Loan Gains 0.0 26.2 239.8 859.8 390.3 1,512.8
Forfeitures 1.6 0.3 3.3 1.1 17.2 0.1
Production Flexibility Contract 699.3 597.5 637.0 614.0 575.2 473.9
Market Loss Assistance 0.0 0.0 316.2 613.5 612.8 523.6
Insurance 157.2 147.7 154.9 223.3 215.8 266.4
Step-2 19.8 466.7 214.4 486.1 252.7 125.1
Total 877.9 1,244.4 1,886.3 3,485.1 2,215.4 3,634.0
Production (thousand tons) 4,124 4,303 3,251 3,832 3,742 4,420
A Index (US$ per kilogram)a 1.73 1.60 1.30 1.16 1.26 0.92
Assistance (US$ per kilogram) 0.21 0.29 0.58 0.91 0.59 0.82
Assistance (% of the A Index) 12% 18% 45% 78% 47% 89%
a. August to July average.
Sources: United States Department of Agriculture (assistance), International Cotton Advisory Committee
(production), and author’s calculations the rest.
— 38 —
TABLE 3
EFFECT OF REMOVAL OF DISTORTIONS (PERCENTAGE CHANGES OVER BASELINE)
2003/04 2005/06 2007/08 2009/10 2011/12 Averagea
World Price 15.6 13.7 13.0 12.2 11.7 12.7
Exports
Africa 12.1 15.1 14.0 13.1 12.3 12.6
Australia 3.9 3.0 2.7 2.3 2.1 2.7
United States -8.4 -6.6 -4.0 -1.5 0.9 -3.5
Uzbekistan 5.4 6.9 6.7 6.4 6.2 6.0
World 3.9 5.6 6.2 6.7 7.3 5.8
Production
United States -18.3 -7.9 -5.9 -4.1 -2.3 -6.7
European Union -77.4 -77.7 -78.3 -78.8 -79.0 -70.5
Uzbekistan 3.1 4.7 4.6 4.4 4.2 4.0
Africa 4.5 7.5 7.1 6.7 6.3 6.0
a. Average is taken over the 10-year period 2001/02 to 2010/11.
Source: FAPRI (2002).
— 39 —
Figure 1: The Classification of Fibers
ORGANIC INORGANIC
TEXTILE FIBERS
ANIMAL ORIGIN PLANT ORIGIN
Most common are wool,
cashmere, and silk.
CELLULOSIC
NATURAL MAN-MADE
Most common are cotton,
linen, sisal, and jute.
Most common are carbon,
ceramic, and glass.
NON-CELLULOSIC
The raw material used to produce cel-
lulosic fibers is wood. Also called
natural polymers. Viscose (known as
r
ayo
n
)
is
t
h
e
m
ost
co
mm
o
n
.
Also called synthetic polymers, they
come from crude oil. Polyester,
acrylic, and polyamide (known as
n
y
l
o
n
)
a
r
e
t
h
e
m
ost
co
mm
o
n
.
— 40 —
Figure 2: Cotton's Share in Total Fiber Consumption (percent)
30
40
50
60
70
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
Figur e 3: Polyester to Cotton Price Ratio
0
1
2
3
4
5
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
— 41 —
Figure 5: World Per Capita Fiber Consumption (kilograms)
1
2
3
4
5
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
Cotto n
Non-Co tto n
Figure 4: World Fiber Consumption (thousand tons)
0
6,000
12,000
18,000
24,000
30,000
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
Cotto n
Chemical Fibe rs
4
2—
Figure 6: Annual A Index (US$ per kilogram)
0.0
1.0
2.0
3.0
4.0
5.0
6.0
1950
1953
1956
1959
1962
1965
1968
1971
1974
1977
1980
1983
1986
1989
1992
1995
1998
2001
Current Constant
Figure 7: Monthly A Index (US$ per kilogram)
0.5
1.0
1.5
2.0
2.5
3.0
Jan-90
Jan-91
Jan-92
Jan-93
Jan-94
Jan-95
Jan-96
Jan-97
Jan-98
Jan-99
Jan-00
Jan-01
Jan-02
Jan-03
4
3—
Figure 8: Cotton Production in Europe (thousand tons)
0
100
200
300
400
500
600
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
Figure 9: World Cotton Yields (kilograms per hectare)
200
300
400
500
600
700
1960
1963
1966
1969
1972
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
4
4—
PART IV: THE FACTS
APPENDIX A: CONCENTRATION INDICES
One way to measure the degree of concentration is to use the Herfindahl concentration
index. This index is defined as the squared sum of, say, export shares of all countries;
values close to unity indicate that a single country accounts for most exports; values
close to zero, indicate that a large number of countries have equal export shares. In
technical terms, let X denote global cotton exports and Xi cotton exports of country i.
Then concentration of exports, IX, is defined as IX = Σi(Xi/X)2. If IX = 1, a single country
accounts for all exports. If IX 0, a large number of countries have equal shares.
The figures that follow depict the evolution of the concentration of exports and
imports (figure A1), concentration of production and consumption (figure A2), and
concentration of stocks (figure A3). During the last 40 years, concentration of cotton im-
ports declined from a high of 8 percent in 1960 to a low of 4 percent during the late
1990s. Cotton exports have been far more concentrated than imports. Moreover, the
concentration of cotton exports is more variable than the concentration of imports. The
spike in the index of cotton exports in the early 1990s is due to Central Asian cotton
production, which is being recorded as trade after the disintegration of the Soviet Un-
ion.
Concentration of consumption fluctuated around an average of 9 percent during
the first 2 decades of the sample and increased thereafter. The peak in the concentration
of production during the mid-1980s reflects China’s production boom. The first peak in
the concentration of stocks is due to the US (stockholding policies of the Commodity
Credit Corporation). The latter two peaks are due to large stocks held by China.
4
5—
TABLE A1: Concentration Index--Exports, Imports
0%
4%
8%
12%
16%
20%
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000
Im por t s
Exports
TABLE A2: Concentration Index-- Production, Consumption
5%
8%
11%
14%
17%
20%
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000
Pr odu ct ion
Consumption
4
6—
TABLE A3: Concentration Index--Stocks
5%
10%
15%
20%
25%
30%
35%
1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000
4
7—
APPENDIX B: GROWTH RATE ESTIMATION AND UNIT ROOTS
Typically, the growth rate between two-periods, say ρ, is calculated as:
ρ = (QtQt-1)/Qt-1, (1)
where Qt is the variable of interest. Equation (1) can also be written as Qt = (1+ρ)Qt-1.
Now let Qt grow at rate ρ in all periods, while in each period it is subjected to stochastic
shocks ηt. That is,
Qt = (1+ρ)Qt-1ηt. (2)
It is assumed that the error term, ηt, is log-normally distributed with mean to equal 1,
reflecting the proportionality nature of shocks.
From (2) it follows that Qt-1 = (1+ρ)Qt-2ηt-1, which upon substitution back to (2)
gives Qt = (1+ρ)2Qt-2ηtηt-1. Recursive substitution yields:
Qt = (1+ρ)tQ0ηtηt-1... η1. (3)
Taking logarithms in (3) and setting µ = ln(Q0), β = ln(1+ρ), and εt = Σtln(ηt), gives:
ln(Qt) = µ + βt + εt. (4)
β is typically estimated with OLS and the growth rate is calculated as ρ = exp(β)-1. Note
that for small values of β, ρ β.
In order to assess the performance of the model both conventional and stationar-
ity statistics are used. Conventional statistics include t-ratios and the R2 while the sta-
tionarity statistics include the augmented Dickey-Fuller (ADF) test (Dickey and Fuller,
1981). ADF is based on the regression (xt - xt-1) = µ + γxt-1 + lags(xt - xt-1) +
υ
t, where xt de-
notes the series under consideration, the error term of (4) in this case; the lag length is
selected so that
υ
t is rendered white noise. A negative value of γ significantly different
from zero indicates that xt is stationary, often denoted as I(0). To identify the presence of
one unit root we test H0: xt is not I(0) against H1: xt is I(0). Another second test is the Phil-
lips-Perron test (PP). If the unit root hypothesis is rejected (i.e. a high ADF or PP statistic
in absolute value), which implies that the variable under consideration is stationary, we
conclude that the growth rate can be summarily expressed by a single point estimate
(see Baffes and Le Vallèe 2002 for details on estimation).
All growth rate estimates discussed in this paper have bee estimated as men-
tioned above for the 1960-2002 period and results are reported in Table B1. In four cases
(real price of cotton, polyester to A Index ratio, world cotton consumption, and cotton
production in Europe) growth rates have been estimated subject to structural breaks.
With the exception of world cotton consumption, the respective dates for structural
breaks reflect events that are believed to have altered the stochastic behavior of the
variable under consideration. For the real price of cotton it was the change in the US
support policy in 1985. For the Polyester to A Index ratio the break is in 1972, when
polyester and cotton prices reached parity. For cotton production in Europe it was the
4
8—
first EU expansion. Because no even is behind the world cotton consumption’s struc-
tural break in 1985 (apart from visual inspection), the growth rate estimate based on the
entire period is also reported.
TABLE B1
GROWTH RATE ESTIMATION RESULTS, ANNUAL DATA, 1960-2002
Period
µ
β
R2 DW ADF PP
ρ
Real price of cotton (A Index deflated by the MUV)
1960-1984 1.01
(20.2)
-0.004
(-1.31)
0.03 1.64 -3.21** -3.88*** -0.44
1985-2002 1.06
(4.02)
-0.020
(-2.61)
0.26 1.52 -2.51 -3.01* -1.96
Price of polyester to A Index ratio
1960-1971 1.72
(23.3)
-0.123
(-12.2)
0.93 0.99 -1.88 -1.96 -11.53
1972-2002 -0.32
(-3.42)
0.010
(3.18)
0.23 1.27 -3.94*** -4.20*** 1.03
World chemical fiber consumption
1960-2002 8.40
(180.5)
0.046
(24.9)
0.94 0.11 -2.43 -2.33 4.69
World total fiber consumption
1960-2002 9.56
(559.7)
0.030
(64.0)
0.98 0.25 -2.70* -1.99 3.02
World cotton consumption
1960-2002 9.20
(658.2)
0.017
(31.9)
0.96 0.30 -2.96* -2.21 1.78
1960-1985 9.21
(804.6)
0.016
(22.1)
0.95 0.68 -3.20** -2.41 1.65
1986-2002 9.61
(308.4)
0.007
(7.62)
0.78 0.85 -1.12 -2.09 0.67
World per capita cotton consumption
1960-2002 1.19
(87.5)
-0.000
(-0.69)
0.01 0.31 -3.04** -2.34 -0.04
World per capita non-cotton fiber consumption
1960-2002 0.66
(20.4)
0.022
(17.1)
0.87 0.18 -2.34 -2.19 2.22
Cotton’s share in total fiber consumption
1960-2002 4.15
(269.2)
-0.010
(-12.2)
0.87 0.20 -1.95 -2.03 -1.04
continued
4
9—
TABLE B1 (continued)
GROWTH RATE ESTIMATION RESULTS, ANNUAL DATA, 1960-2002
Period
µ
β
R2 DW ADF PP
ρ
Cotton Production in Europe
1960-1982 5.31
(119.5)
-0.004
(-1.34)
0.04 2.24 -5.56*** -6.46*** -0.44
1983-2002 4.13
(24.3)
0.053
(10.4)
0.86 1.40 -3.07** -3.51** 5.39
World cotton yields
1960-2002 5.73
(323.2)
0.016
(24.9)
0.94 0.88 -2.44 -3.39** 1.77
Real world GDP
1960-2002 3.81
(261.2)
0.035
(60.9)
0.99 0.08 -2.45 -2.35 3.58
World population
1960-2001 8.02
(1853.3)
0.017
(99.5)
0.99 0.03 -2.63* -0.67 1.76
Notes: Numbers in parentheses denote t-ratios. DW denotes the Durbin-Watson measure of serial corre-
lation. ADF and PP refer to the Augmented Dickey-Fuller and Phillips-Perron stationarity statistics.
ρ
de-
notes the annual growth rate. Asterisks denote rejection of non-stationarity at 10% (*), 5% (**), and 1%
(***) level of significance. Rejection of non-stationarity implies that the variable under consideration is
trend stationary and hence the estimated growth rates are valid representations of the true growth rate.
—50—
APPENDIX C: PRICE VARIABILITY
Calculating price variability is a complicated issue for at least two reasons. First, vari-
ability must be defined around some statistic, typically the first moment of the distribu-
tion. As expected, different statistics around which price variability is defined will yield
different outcomes. Second, the statistical properties of prices may also pose difficulties
in calculating variability. Prices are typically non-stationary, implying that as we move
further away from a certain point, the probability that the price will return to that initial
point becomes smaller. In technical terms, non-stationarity implies that the first two
moments of the distribution do not exist. The remaining of this appendix measures cot-
ton price variability and whether it has changed during the past 40 years by taking into
consideration these two difficulties.
The most widely used indicator of variability is the standard deviation. It meas-
ures how widely values of a sample are dispersed around its average. It is defined as
the sum of the square of each deviation from the mean,
µ
, divided by the number of ob-
servations, n. Let Pt denote the price at time t. Then, the standard deviation, Sn, is
Sn = [Σt(Pt -
µ
)2/(n – 1)]1/2 (1)
If Pt is subject to a linear trend, then (Pt - µ) is replaced by (Pt - t) in (1) and becomes:
Sn = [Σt(Pt - t)2/(n – 1)]1/2 (2)
These measures however, require that the variable under consideration is either station-
ary (when
µ
is used) or trend-stationary (when t is used).
In the case of non-stationarity (the most likely outcome for price series), one has
to induce stationarity first and then calculate variability. Because non-stationary vari-
ables require taking first differences to induce stationarity, the measure of variability
used is the Z-statistic. It is defined as:
Sn = [Σt(PtPt-1)2/(n – 1)]1/2 (3)
Notice that when Pt is nonstationary (or alternatively I(1)), its first difference, (Pt - Pt-1),
is stationary, I(0), thus making the Z-statistic a proper measure of price variability.
Therefore, the first step in calculating price variability is to examine the stationar-
ity properties of Pt. This is done by utilizing the testing procedures outlined in appendix
B. The sample consists of monthly data from January 1960 to December 2002 (a total of
504 observations). The sample is divided into 3 periods consistent with the two struc-
tural changes that have taken place in the cotton market. The first period goes up to
1972 (prior to the commodity crisis of 1973). The second period goes up to 1985 (when
the US changed its policy stance and released huge stocks of cotton). The third period
covers 1985 to 2002. Because the MUV—the preferred deflator for dollar-denominated
commodity prices—is not available on monthly basis, the US consumer and producer
price indices were used instead.
—51—
Table C1 reports the stationarity test statistics for levels without trend (upper
panel), levels with trend (middle panel) and first differences (lower panel), deflated by
the producer price index (first 3 columns) and the consumer price index (right 3 col-
umns). Results from the upper panel indicate that non-stationarity against the alterna-
tive of stationarity is not rejected (i.e. cotton prices are not stationary at levels). Results
from the second panel indicate that, with the exception of PP test during the second pe-
riod which indicates trend stationarity at the 10% level, in all other cases the null hy-
pothesis of non-stationarity cannot be rejected. Thus, the prices have to be differenced
once to become stationary.
Therefore, in order to calculate price variability, one must rely on the Z-statistic.
However, for the sake of comparison, the upper panel of table C2 reports statistics for
both variability around trend (first column, denoted “Trend”) and the Z-statistic (sec-
ond column denoted “First differences”). The third column (denoted “Annual average”)
reports the period average of the standard deviation for each crop year (i.e. first the
standard deviation for each consecutive 12-month period is calculated, starting in Au-
gust an then the average of these standard deviations over the three periods is calcu-
lated). This statistic, in a sense, is as a measure of short term variability.
The lower panel of table C2 gives the percentage change of price variability from
one period to another. All three measures give the same qualitative outcome while there
is no discernable difference when the producer or the consumer price index is consid-
ered. The main message from table C2 is that price variability during the post-1985 pe-
riod has been 2.5 times higher of what it was during the pre-1973 period while it is
about half of what it was during the 1973-84 period. Figures C1 and C2 depict the
within year variability. In particular, figure C1 shows the variability for each year aver-
aged over the three periods defined above—the averages are the ones reported in the
last column of table C2, upper panel, 1.5, 7.5, 3.5, respectively. Figure C2 depicts the
same numbers averaged differently. Specifically, the second period consists of only two
years, the highly inflationary 1973 and 1974. Variability during the second and third pe-
riod increases to 16.8 (from 7.5) and 4.2 (from 3.5), respectively.
—52—
TABLE C1: STATIONARITY STATISTICS
Deflated by Producer Price Index Deflated by Consumer Price Index
ADF PP ADF PP
Levels
1960-72 -2.38 -2.21 -1.90 -2.00
1973-84 -3.40* -2.63 -2.48 -3.37*
1985-2002 -2.53 -2.91 -2.67 -2.46
Trend
1960-72 -2.76 -2.63 -2.07 -2.02
1973-84 -4.25** -3.18* -4.57*** -3.44**
1985-2002 -2.69 -2.84 -2.06 -2.46
First Differences
1960-72 -4.40*** -10.20*** -5.32*** -8.31***
1973-84 -3.83** -11.03*** -3.87** -6.39***
1985-2002 -6.31*** -9.01*** -6.12*** -7.78***
Notes: ADF and PP denotes the Augmented Dickey-Fuller and Phillips-Perron statistics. The McKinnon
critical values (properly adjusted for the number of observations) are: -4.00 (1%), -3.44 (5%), and –3.13
(10%). Asterisks denote rejection of the null hypothesis of non-stationarity at the 1% (***), 5% (**) and 10%
(*) levels of significance.
TABLE C2: PRICE VARIABILITY MEASURES
Deflated by Producer Price Index Deflated by Consumer Price Index
Trenda
First
differencesb
Annual
averagec
Trenda
First
differencesb
Annual
averagec
Measure of Variability
1960-72 11.7 3.7 4.8 3.8 1.2 1.5
1973-84 39.6 12.8 19.7 13.8 4.6 7.0
1985-2002 25.7 7.0 12.1 7.7 2.0 3.5
Change (percent)
1960-72 to 1973-84 238% 250% 309% 268% 279% 368%
1960-72 to 1985-2002 120% 90% 151% 105% 66% 133%
1973-84 to 1985-2002 -35% -46% -39% -44% -56% -50%
a. Variation around the trend for each period—equation (2).
b. Variability around the average of the first differences of each period—equation (3)
c. Variability around the average of each crop year (August-July) for each period; each period consists of
12 observations—period average of equation (1).
Source: Authors’ calculations.
—53—
Figure C1: Monthly Cotton Price Variability
0
4
8
12
16
20
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
Figure C2: Monthly Cotton Price Variability
0
4
8
12
16
20
1960
1962
1964
1966
1968
1970
1972
1974
1976
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
—54—
APPENDIX D: THE COTLOOK A AND B INDICES
The Cotlook A Index is the average of the 5 lowest quotations of 16 styles of cotton
(Middling 1-3/32’’) traded in North European ports from the following origins: Austra-
lia, Brazil, China, Francophone Africa, Greece, India, Mexico, Pakistan, Paraguay, Spain,
Syria, Tanzania, Turkey, the United States, and Uzbekistan. The Cotlook B Index is the
average of the three lowest quotations of eight styles of coarser grades of cotton from
Argentina, Brazil, China, India, Pakistan, Turkey, the United States, and Uzbekistan.
The indices are compiled daily by Cotton Outlook, a private company located in
Liverpool, UK. Staff collect quotations by interviewing cotton traders and merchants in
North Europe, and they also look at other market developments likely to affect cotton
prices. These offering prices and the day’s indices are published at about 2:30 p.m. UK
time.
The prices are expressed in U.S. cents per pound, c.i.f. North Europe, cash
against documents on arrival of vessel, including profit and agent’s commission. When
a particular cotton growth is not offered in large volume, the quotation is still reported,
but it is not eligible to participate in the index. The index is based on the three or five
least expensive quotations because quotations reflect offering prices, not the level at
which business has been arranged, so a buyer would normally expect to succeed with
bids that are slightly lower than quoted.
The quotations represent nearby delivery, normally between two and four
months. A dual quotation is reported when information on supply conditions of the
next cotton season is readily available: one for nearby delivery and one for forward de-
livery. For example, quotations on July 12, 2001 referred to July/August 2001 delivery
(nearby) and October 2001 to May 2002 delivery (forward). Table D1 depicts the compo-
sition of the Cotlook A and B Indices for three randomly selected dates: July 12 (nearby
and forward delivery), August 30 ( nearby delivery), and October 18, 2001 (nearby de-
livery).
Because the US quotations are among the highest, they are included in the A In-
dex very infrequently. On the contrary, the Francophone Africa quotation is included
very often in the Index because the quotation is relatively low (and hence the chances of
its eligibility high) and it also traded throughout the entire season.
—55—
TABLE D1
COMPOSITION OF THE COTLOOK A AND B INDICES (US CENTS PER POUND)
ORIGIN
JULY 12, 2001
(NEARBY)
JULY 12, 2001
(FORWARD)
AUGUST 30, 2001
(NEARBY)
OCTOBER 18, 2001
(NEARBY)
Australia 50.00 52.25 51.50 44.25
Brazil 46.50 46.00* 43.00* NQ
China NQ NQ NQ NQ
Franc Zone 47.00* 45.50* 43.25 38.00*
Greece 44.00* 44.75* 42.00* 35.25*
India NQ NQ NQ NQ
Mexico NQ NQ NQ NQ
Pakistan NQ NQ NQ NQ
Paraguay 45.00* NQ NQ NQ
Spain 49.50 46.00* 43.00* 38.25*
Syria 47.00* 46.50 42.00* 36.00*
Tanzania NQ 49.00 47.00 NQ
Turkey NQ NQ NQ NQ
US (California/Arizona) 50.50 53.75 50.50 42.25
US (Memphis/Eastern) 52.00 53.00 49.75 40.50
Uzbekistan 47.00* 45.50* 42.50* 36.00*
A Index 46.00 45.55 42.50 36.70
Argentina 45.50@ NQ NQ NQ
Brazil NQ 43.50@ 40.50@ 37.00
China NQ NQ NQ NQ
India NQ NQ NQ NQ
Pakistan NQ 44.50@ 41.50 33.50@
Turkey NQ NQ NQ NQ
US (Orleans/Texas) 43.00@ 45.00 41.25@ 34.00@
Uzbekistan 45.00@ 44.00@ 41.00@ 34.50@
B Index 44.50 44.00 40.90 34.00
NQ no quotation from this origin available that week.
* quotation is one of the five less expensive origins.
@ quotation is one of the three less expensive origins.
Source: Cotton Outlook, July 13, August 31, and October 19, 2001 issues.
—56—
APPENDIX E: COTTON PRICE RISK MANAGEMENT
Cotton was one of the first commodities to be traded in futures markets. Earlier in the
20th century there were at least 10 active cotton futures exchanges (Baffes and Kaltsas
2002). Currently, there is only one major cotton futures and options contract, which is
traded at the New York Board of Trade. The New York contract, whose size is 50,000
pounds, uses Memphis No. 2 cotton as the cash price equivalent for quality specification
and delivery purposes. There are five delivery months (March, May, July, October, and
December), and the nearest 10 delivery months are available for trade, extending the
time span of the contract to almost two years—a July 2001 contract could be traded as
early as August 1999. Table E1 reports the closing futures prices for the March, May,
and July 2002 contracts on January 3, 11, and 13, 2002. It also reports the strike prices
and costs for the corresponding put options.
The New York contract is appropriate only for U.S. cotton, however. Non-U.S.
cotton traders and merchants have no access to a hedging instrument. Cotton Outlook
(December 12, 1997, p. 3) observed that the lack of an international trading instru-
ment—one that consistently reflects broad world cotton market developments but is
capable of being used as hedge—continues to be a shortcoming of the current pricing
system. The 59th ICAC Plenary Meeting in Cairns reached a similar conclusion (ICAC
2001):
Futures contracts traded in New York are limited to the delivery of U.S. cotton to U.S.
locations. Accordingly, prices in New York reflect primarily U.S. conditions. As a conse-
quence, prices for cotton in non-U.S. locations can diverge from New York futures prices,
limiting the utility of the New York market for many in the world industry.
Comovement between the New York contract and the Cotlook A Index is low, confirm-
ing the inadequacy of the New York contract as a hedging tool for traders and mer-
chants of non-U.S. types of cotton. As an example, in December 31, 1990, the May 1991
New York contract closed at 76.19 cents a pound, 8.21 cents below the Cotlook A Index,
and it expired on May 8, 1991, at 92.22 cents a pound, 8.92 cents above the Cotlook A
Index. In a study using an error-correction specification and weekly cotton prices (com-
ponents of the A Index) from August 1985 to December 1987 and August 1995 to Janu-
ary 1997, Baffes and Ajwad (2001) found that, unlike Central Asian, West African, and
(to some extent) Greek prices, U.S. prices moved relatively independently of other
prices.
Several efforts have been made since the late 1990s to establish an international
trading instrument for cotton (Baffes and Kaltsas 2002). Brazil reintroduced its cotton
contract in 1996 and India in 1998. China, Euronext (the European Trading Alliance),
Turkey, and the United States are also contemplating new initiatives. The Cotton Ex-
changes in Brazil and India are not used by foreign traders, but the one in Europe, if
launched, is expected to trade Central Asian and West African cotton, making it a use-
—57—
ful hedging tool for traders and merchants of non-U.S. cotton. The Common Fund for
Commodities has recently launched a project investigating ways to manage cotton price
risk in Tanzania, Uganda, and Zimbabwe, with the Cotton Company of Zimbabwe as
the project executing agency.
TABLE E1: INDICATIVE COTTON FUTURES AND PUT OPTIONS STRIKE
PRICES AND COSTS (CENTS PER POUND)
MARCH 2002 MAY 2002 JULY 2002
JANUARY 3, 2002 CLOSING FUTURES PRICE
36.46 37.90 39.30
STRIKE PRICE COST OF PUT OPTION
38.00 2.44 2.58 2.50
36.00 1.29 1.64 1.70
34.00 0.56 0.95 1.10
JANUARY 11, 2002 CLOSING FUTURES PRICE
37.40 38.88 40.40
STRIKE PRICE COST OF PUT OPTION
40.00 3.15 3.15 2.69
38.00 1.75 2.08 2.04
36.00 0.83 1.26 1.33
35.00 0.53 0.95 1.04
JANUARY 13, 2002 CLOSING FUTURES PRICE
36.80 38.30 39.80
STRIKE PRICE COST OF PUT OPTION
40.00 3.43 3.32 3.20
38.00 1.84 2.18 2.22
36.00 0.75 1.30 1.45
Source: The New York Board of Trade
—58—
APPENDIX F: STATISTICAL TABLES
TABLE F1: PROFILE OF COTTON: 1998/99
COUNTRY
PRODUCTION
(thousand tons)
EXPORTS
(thousand tons)
YIELD
(kilograms per hectare)
PER CAPITA GDP
(1995 US dollars)
AMERICAS
United States 3,362 1,208 699 30,620
Brazil 611 4 800 4,495
Mexico 177 42 691 3,583
Argentina 167 162 364 8,270
Paraguay 74 60 442 1,768
Peru 46 5 631 2,344
Total 4,489 1,490
AFRICA
Egypt 232 103 776 1,166
Mali 207 209 419 285
Côte d'Ivoire 165 145 585 782
Benin 138 135 360 398
Zimbabwe 117 96 343 674
Burkina Faso 115 112 392 249
Cameroon 79 71 459 655
Chad 69 67 230 224
Togo 68 69 435 341
Nigeria 58 18 199 252
Sudan 54 51 398 290
South Africa 42 9 369 3,946
Tanzania 39 30 186 185
Total 1,556 1,228
EUROPE
Greece 396 275 934 12,400
Spain 118 58 1,143 16,802
Total 514 384
ASIA
China 4,165 258 1,044 747
India 2,729 27 306 438
Pakistan 1,703 46 576 503
Uzbekistan 1,064 897 700 467
Turkey 831 65 1,127 3,072
Australia 734 678 1,332 23,293
Syria 321 226 1,409 858
Turkmenistan 208 195 378 1,126
Iran 141 0 641 1,580
Tajikistan 104 87 413 352
Kazakhstan 67 64 572 1,350
Myanmar 54 27 216
Israel 38 38 1,660 16,453
Total 12,349 2,703
WORLD 18,907 5,805 585
Source. International Cotton Advisory Committee and World Bank (World Development Indicators).
—59—
TABLE F2
GLOBAL BALANCE OF THE COTTON MARKET (THOUSAND TONS), 1960-2002
1960 1970 1980 1990 1998 1999 2000 2001 2002
PRODUCTION
China 1,372 1,995 2,707 4,508 4,501 3,830 4,350 5,100 4,700
US 3,147 2,219 2,422 3,376 3,030 3,835 3,818 4,393 3,879
India 1,012 909 1,322 1,989 2,710 2,650 2,350 2,459 2,450
Pakistan 306 543 714 1,638 1,480 1,800 1,750 1,743 1,717
Uzbekistana 1,491 2,342 2,661 2,593 1,000 1,150 960 1,055 1,035
Franc Zone 63 140 224 562 897 928 700 991 902
Turkey 192 400 500 655 871 826 740 900 900
Brazil 425 549 623 717 420 648 848 725 785
Australia 2 19 99 433 726 733 704 658 386
Greece 63 110 115 213 405 428 420 410 355
Egypt 480 509 529 296 230 229 206 279 293
Syria 112 150 118 145 335 325 362 350 233
World 10,201 11,740 13,831 18,970 18,551 18,887 18,901 20,856 19,076
ENDING STOCKS
US 1,574 915 653 510 849 860 1,174 1,826 2,222
China 0 412 299 1,550 4,124 2,814 2,263 2,347 1,949
India 635 376 59 539 1,011 910 848 812 583
Brazil 144 321 391 231 317 370 505 536 401
Pakistan 52 55 204 313 353 463 353 616 373
Australia 5 13 61 150 424 431 371 378 348
World 4,643 4,605 5,152 6,653 9,699 8,710 7,917 9,896 9,092
CONSUMPTION
China 1,481 2,016 3,300 4,225 4,400 4,800 5,050 5,500 5,700
India 1,006 1,076 1,371 1,958 2,781 2,939 2,924 2,899 2,942
Pakistan 245 429 461 1,343 1,625 1,700 1,760 1,900 2,000
US 1,803 1,786 1,083 1,885 2,265 2,230 1,929 1,681 1,578
Turkey 109 184 293 557 1,000 1,200 1,150 1,300 1,365
Brazil 272 296 566 723 797 852 871 860 875
Indonesia 10 43 104 336 438 470 480 530 500
Mexico 109 146 165 170 484 525 435 430 418
Thailand 15 65 127 328 290 340 360 385 400
Russiaa 1,350 1,821 1,796 1,190 190 280 320 345 362
Korea 59 117 322 436 320 325 320 325 330
Italy 226 201 209 333 284 307 300 290 284
Taiwan 46 137 229 346 290 295 250 260 260
Bangladeshb na na 45 98 153 169 196 215 240
Uzbekistana na na na 205 125 185 220 250 225
Japan 739 766 715 650 275 280 251 221 212
World 10,231 12,173 14,215 18,585 18,674 19,756 19,753 20,152 20,535
a. Uzbekistan and Russia refer to USSR prior to and including 1990.
b. Included in Pakistan prior to 1970 and including 1970.
Source: ICAC, Cotton: Review of the World Situation, various issues.
—60—
TABLE F3
GLOBAL TRADE OF THE COTTON MARKET (THOUSAND TONS), 1960-2002
1960 1970 1980 1990 1998 1999 2000 2001 2002
EXPORTS
US 1,444 848 1,290 1,697 915 1,481 1,470 2,134 2,056
Uzbekistana 381 553 616 397 900 900 820 718 717
Australia 0 4 53 329 650 710 720 650 609
Greece 33 0 13 86 230 294 293 257 249
Mali 2 19 35 114 216 201 125 126 221
Syria 97 134 71 91 210 180 245 220 171
Benin 1 14 8 58 119 151 131 132 164
Burkina Faso 0 9 22 73 117 106 107 127 154
Tajikistana na na na 200 90 83 110 120 147
Côte d’Ivoire 0 7 42 81 130 160 150 115 137
Zimbabwe 0 32 55 38 89 121 128 67 105
World 3,667 3,875 4,414 5,081 5,274 6,054 5,875 6,167 6,377
IMPORTS
Indonesia 7 36 106 324 500 455 520 559 537
India 204 155 0 0 136 200 340 425 509
China 65 108 773 480 78 30 52 102 400
Turkey 0 1 0 46 250 459 285 385 358
Thailand 4 46 86 354 271 302 360 387 356
Mexico 0 1 0 43 302 436 473 396 352
Russiaa 0 238 28 37 179 284 325 341 338
Italy 218 178 193 336 330 365 310 323 315
Korea, Rep. 51 121 332 447 330 350 315 318 298
Japan 800 796 697 634 270 276 242 247 240
Pakistan 1 1 1 0 192 103 101 280 224
Taiwan 47 160 214 358 293 322 269 225 214
Brazil 0 4 2 108 296 340 131 57 200
World 3,804 4,086 4,555 5,222 5,429 5,811 5,875 6,167 6,377
a. Uzbekistan, Tajikistan, and Russia refer to USSR prior to and including 1990.
Source: ICAC, Cotton: Review of the World Situation, various issues.
—61—
TABLE F4
DIRECTION OF COTTON TRADE: 1980/81-2000/01
1980/81 1990/91 2000/01
Million of $ US (Current)
Industrial to Industrial Countries 1,287 1,631 726
Industrial to Developing Countries 1,728 2,114 2,800
Developing to Industrial Countries 951 1,232 898
Developing to Developing Countries 599 1,896 1,951
TOTAL 4,466 6,872 6,375
Share (percent)
Industrial to Industrial Countries 28 23 11
Industrial to Developing Countries 38 31 44
Developing to Industrial Countries 21 18 14
Developing to Developing Countries 13 28 31
TOTAL 100 100 100
a. Industrial countries are: US, EU (Greece and Spain), Australia, and Israel, currently accounting for 40
percent of world exports (2.3 million tons out of 5.8 million tons).
Source: COMTRADE
—62—
TABLE F5: COTTON PRICESa (US DOLLARS PER KILOGRAM), 1950-2002
MONTHLY ANNUAL
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC NOMINAL REALb
1950 0.92 5.05
1951 0.86 0.88 0.94 1.04 1.08 0.96 4.56
1952 1.05 1.02 0.96 0.96 0.90 0.92 0.94 0.99 1.00 0.93 0.90 0.86 0.95 4.31
1953 0.82 0.83 0.83 0.83 0.38 0.82 0.84 0.83 0.82 0.83 0.83 0.84 0.83 3.87
1954 0.86 0.82 0.85 0.85 0.85 0.84 0.83 0.85 0.89 0.88 0.87 0.88 0.86 4.10
1955 0.88 0.88 0.85 0.84 0.85 0.85 0.83 0.82 0.78 0.76 0.77 0.75 0.82 3.84
1956 0.77 0.82 0.84 0.82 0.75 0.71 0.68 0.68 0.68 0.70 0.72 0.73 0.74 3.34
1957 0.74 0.74 0.75 0.74 0.72 0.72 0.72 0.74 0.73 0.74 0.77 0.78 0.74 3.28
1958 0.79 0.76 0.74 0.75 0.75 0.72 0.68 0.68 0.66 0.68 0.66 0.63 0.71 3.09
1959 0.63 0.62 0.62 0.63 0.62 0.62 0.61 0.62 0.63 0.63 0.64 0.66 0.63 2.78
1960 0.65 0.65 0.65 0.64 0.65 0.66 0.66 0.65 0.66 0.67 0.67 0.66 0.65 2.81
1961 0.66 0.68 0.67 0.67 0.68 0.68 0.68 0.67 0.67 0.68 0.66 0.66 0.67 2.85
1962 0.66 0.66 0.66 0.66 0.66 0.66 0.65 0.64 0.63 0.63 0.63 0.65 0.65 2.73
1963 0.66 0.65 0.65 0.64 0.64 0.64 0.64 0.64 0.65 0.65 0.64 0.64 0.65 2.71
1964 0.66 0.66 0.65 0.65 0.66 0.66 0.65 0.64 0.65 0.65 0.64 0.64 0.65 2.68
1965 0.64 0.67 0.65 0.65 0.65 0.63 0.63 0.63 0.63 0.62 0.62 0.62 0.64 2.59
1966 0.62 0.62 0.64 0.62 0.62 0.62 0.62 0.62 0.61 0.63 0.63 0.63 0.62 2.42
1967 0.64 0.66 0.66 0.66 0.66 0.67 0.67 0.68 0.69 0.70 0.71 0.72 0.68 2.57
1968 0.72 0.71 0.71 0.71 0.71 0.70 0.69 0.68 0.66 0.66 0.63 0.62 0.68 2.68
1969 0.63 0.63 0.63 0.63 0.63 0.62 0.61 0.60 0.61 0.63 0.65 0.66 0.63 2.31
1970 0.64 0.65 0.66 0.66 0.66 0.67 0.67 0.68 0.69 0.70 0.71 0.72 0.63 2.25
1971 0.73 0.74 0.73 0.73 0.75 0.79 0.80 0.83 0.83 0.82 0.82 0.85 0.74 2.51
1972 0.89 0.90 0.87 0.87 0.86 0.83 0.78 0.74 0.73 0.78 0.81 0.86 0.79 2.46
1973 0.89 0.91 0.96 1.03 1.15 1.22 1.46 1.67 1.91 1.93 1.75 1.82 1.36 3.63
1974 1.99 1.84 1.69 1.61 1.47 1.40 1.32 1.33 1.32 1.26 1.17 1.09 1.42 3.11
1975 1.05 1.06 1.09 1.16 1.22 1.23 1.25 1.31 1.33 1.32 1.30 1.35 1.16 2.30
1976 1.47 1.52 1.54 1.56 1.61 1.80 2.00 1.92 1.88 1.92 1.91 1.84 1.69 3.31
1977 1.75 1.86 1.92 1.89 1.78 1.61 1.57 1.51 1.43 1.39 1.41 1.42 1.55 2.81
1978 1.46 1.53 1.58 1.60 1.63 1.60 1.55 1.59 1.63 1.67 1.73 1.71 1.57 2.45
1979 1.68 1.67 1.69 1.64 1.68 1.70 1.70 1.71 1.72 1.72 1.77 1.82 1.69 2.36
—63—
TABLE F5: COTTON PRICESa (US DOLLARS PER KILOGRAM), 1950-2002 (continued)
MONTHLY ANNUAL
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC NOMINAL REALb
1980 1.94 2.13 2.06 1.99 1.95 1.85 1.94 2.11 2.22 2.18 2.16 2.18 2.05 2.60
1981 2.18 2.12 2.04 1.98 1.95 1.93 1.89 1.77 1.70 1.67 1.65 1.58 1.85 2.34
1982 1.55 1.54 1.55 1.58 1.69 1.67 1.73 1.66 1.60 1.55 1.52 1.54 1.60 2.09
1983 1.58 1.64 1.74 1.77 1.81 1.90 1.95 2.00 1.98 1.94 1.97 1.97 1.85 2.49
1984 1.93 1.93 1.95 1.96 1.97 1.85 1.74 1.66 1.61 1.62 1.60 1.59 1.79 2.45
1985 1.57 1.51 1.48 1.46 1.43 1.39 1.34 1.26 1.18 1.08 1.06 1.06 1.32 1.83
1986 1.14 1.20 1.15 1.07 1.00 0.90 0.83 0.82 0.96 1.13 1.16 1.30 1.06 1.27
1987 1.45 1.45 1.39 1.46 1.69 1.75 1.84 1.91 1.84 1.68 1.67 1.66 1.65 1.81
1988 1.59 1.49 1.46 1.45 1.45 1.52 1.40 1.27 1.25 1.27 1.29 1.35 1.40 1.45
1989 1.39 1.39 1.46 1.63 1.71 1.74 1.83 1.83 1.80 1.81 1.81 1.71 1.67 1.74
1990 1.66 1.68 1.74 1.83 1.89 1.99 2.01 1.79 1.79 1.79 1.82 1.85 1.82 1.82
1991 1.85 1.87 1.86 1.83 1.82 1.78 1.70 1.62 1.55 1.50 1.40 1.36 1.68 1.64
1992 1.31 1.24 1.22 1.28 1.34 1.41 1.44 1.32 1.25 1.17 1.16 1.20 1.28 1.21
1993 1.26 1.33 1.36 1.35 1.33 1.30 1.27 1.22 1.22 1.21 1.21 1.31 1.28 1.20
1994 1.53 1.78 1.80 1.85 1.90 1.89 1.80 1.69 1.66 1.63 1.71 1.92 1.76 1.60
1995 2.11 2.23 2.44 2.51 2.53 2.00 1.93 1.88 2.01 2.01 1.97 1.94 2.13 1.82
1996 1.90 1.87 1.83 1.83 1.83 1.83 1.76 1.68 1.66 1.66 1.68 1.75 1.77 1.59
1997 1.76 1.77 1.78 1.74 1.75 1.78 1.79 1.79 1.76 1.72 1.70 1.64 1.75 1.69
1998 1.59 1.52 1.51 1.45 1.42 1.52 1.54 1.50 1.46 1.36 1.24 1.23 1.44 1.45
1999 1.23 1.24 1.25 1.27 1.32 1.29 1.20 1.12 1.09 1.05 1.02 0.97 1.17 1.18
2000 1.05 1.18 1.26 1.29 1.33 1.31 1.29 1.34 1.36 1.34 1.41 1.45 1.30 1.34
2001 1.41 1.33 1.20 1.13 1.10 1.05 1.00 0.96 0.91 0.82 0.84 0.95 1.06 1.12
2002 0.96 0.94 0.93 0.91 0.88 0.96 1.03 1.09 1.08 1.09 1.15 1.22 1.02 1.09
2003 1.25 1.30 1.35 1.34 1.28 1.29 1.33 1.34 1.41 1.60 1.70 1.62 1.40 1.40
a. Mexican c.i.f. North Europe up to July 1973; A Index since August 1973.
b. Real prices have been deflated by the manufacture import unit value (1990 = 1.0).
Source: World Bank Commodity Price Data.
—64—
TABLE F6: PRODUCTION OF CHEMICAL FIBERS (THOUSAND TONS)
1975-2000
COUNTRY 1975 1980 1985 1990 1995 2000
China (Mainland) 156 418 912 1,557 2,719 6,711
European Union 2,450 2,694 2,816 2,781 2,713 3,429
United States 2,785 3,609 3,117 3,115 3,465 3,308
China (Taiwan) 284 636 1,145 1,769 2,550 3,264
Korea, Rep. 273 564 825 1,286 1,865 2,665
India 155 203 335 653 1,000 1,866
Japan 1,380 1,755 1,747 1,701 1,613 1,434
Indonesia 8 96 187 323 820 1,261
Thailand 39 103 111 203 539 822
Turkey 51 104 194 305 464 747
Mexico 185 267 326 370 530 602
Pakistan 5 4 42 75 258 504
Malaysia 3 36 39 42 61 414
Brazil 175 283 269 262 281 347
Americas 3,415 4,488 4,074 4,452 4,702 4,660
Africa 36 77 143 166 235 245
Europe 4,443 5,143 5,590 5,579 3,797 3,429
Asia 2,421 4,010 5,668 7,293 12,064 19,999
Worlda 10,314 13,718 15,475 17,672 20,797 28,335
a. World denotes the world total which includes countries not listed here.
Source. International Cotton Advisory Committee.
—65—
TABLE F7: GLOBAL FIBER CONSUMPTION, 1960-2002
CONSUMPTION (THOUSAND TONS) SHARE (PERCENT)
YEAR TOTAL COTTON WOOL CHEMICAL COTTON WOOL CHEMICAL
1960 15,153 10,356 1,495 3,302 68.3 9.9 21.8
1961 15,102 10,085 1,505 3,512 66.8 10.0 23.3
1962 15,339 9,902 1,501 3,936 64.6 9.8 25.7
1963 16,003 10,147 1,475 4,381 63.4 9.2 27.4
1964 17,256 10,830 1,460 4,966 62.8 8.5 28.8
1965 18,182 11,318 1,473 5,391 62.2 8.1 29.7
1966 18,796 11,539 1,545 5,712 61.4 8.2 30.4
1967 19,212 11,695 1,473 6,044 60.9 7.7 31.5
1968 20,434 11,763 1,565 7,106 57.6 7.7 34.8
1969 21,248 11,911 1,604 7,733 56.1 7.5 36.4
1970 22,041 12,405 1,500 8,136 56.3 6.8 36.9
1971 23,037 12,493 1,480 9,064 54.2 6.4 39.3
1972 24,417 12,903 1,578 9,936 52.8 6.5 40.7
1973 26,031 13,288 1,443 11,300 51.0 5.5 43.4
1974 25,267 12,986 1,262 11,019 51.4 5.0 43.6
1975 24,717 13,047 1,358 10,312 52.8 5.5 41.7
1976 26,537 13,211 1,515 11,811 49.8 5.7 44.5
1977 27,025 13,117 1,478 12,430 48.5 5.5 46.0
1978 28,246 13,415 1,481 13,350 47.5 5.2 47.3
1979 29,440 13,897 1,558 13,985 47.2 5.3 47.5
1980 29,580 14,295 1,567 13,718 48.3 5.3 46.4
1981 29,731 14,124 1,576 14,031 47.5 5.3 47.2
1982 28,895 14,248 1,556 13,091 49.3 5.4 45.3
1983 30,166 14,548 1,612 14,006 48.2 5.3 46.4
1984 31,251 14,830 1,621 14,800 47.5 5.2 47.4
1985 32,813 15,768 1,625 15,420 48.1 5.0 47.0
1986 34,956 17,462 1,708 15,786 50.0 4.9 45.2
1987 36,546 18,226 1,754 16,566 49.9 4.8 45.3
1988 37,427 18,210 1,904 17,313 48.7 5.1 46.3
1989 38,228 18,677 1,861 17,690 48.9 4.9 46.3
1990 37,882 18,602 1,628 17,652 49.1 4.3 46.6
1991 38,070 18,563 1,801 17,706 48.8 4.7 46.5
1992 38,872 18,628 1,757 18,488 47.9 4.5 47.6
1993 39,109 18,544 1,649 18,916 47.4 4.2 48.4
1994 40,392 18,427 1,723 20,242 45.6 4.3 50.1
1995 40,792 18,425 1,554 20,813 45.2 3.8 51.0
1996 42,296 18,821 1,440 22,035 44.5 3.4 52.1
1997 45,086 19,059 1,361 24,666 42.3 3.0 54.7
1998 45,481 18,707 1,293 25,481 41.1 2.8 56.0
1999 47,071 19,162 1,350 26,559 40.7 2.9 56.4
2000 49,416 19,791 1,281 28,344 40.0 2.6 57.4
2001 49,603 20,080 1,401 28,122 40.5 2.8 56.7
2002 50,021 20,656 1,357 30,008 39.7 2.6 57.7
Source. International Cotton Advisory Committee.
—66—
TABLE F8: FIBER PRICES (US $ PER KILOGRAM), 1960-2002
NOMINAL REAL
Year A Index Wool Rayon Polyester MUVa A Index Wool Rayon Polyester
1960 0.65 1.34 0.62 2.78 0.23 2.81 5.81 2.67 12.01
1961 0.67 1.39 0.57 2.60 0.24 2.85 5.91 2.43 11.03
1962 0.65 1.22 0.57 2.51 0.24 2.73 5.09 2.39 10.46
1963 0.65 1.43 0.60 2.51 0.24 2.71 6.05 2.52 10.66
1964 0.65 1.51 0.62 2.16 0.24 2.68 6.30 2.58 9.04
1965 0.64 1.26 0.60 1.85 0.24 2.59 5.22 2.47 7.67
1966 0.62 1.36 0.57 1.79 0.25 2.42 5.45 2.29 7.14
1967 0.68 1.22 0.53 1.28 0.25 2.57 4.83 2.09 5.06
1968 0.68 1.17 0.55 1.21 0.25 2.68 4.66 2.20 4.85
1969 0.63 1.32 0.57 0.99 0.26 2.31 5.00 2.17 3.76
1970 0.63 0.98 0.55 0.90 0.28 2.25 3.50 1.96 3.22
1971 0.74 0.80 0.60 0.84 0.29 2.51 2.70 2.02 2.84
1972 0.79 1.18 0.68 0.79 0.32 2.46 3.66 2.12 2.47
1973 1.36 3.05 0.73 0.84 0.37 3.63 8.18 1.95 2.25
1974 1.42 2.52 1.12 1.01 0.45 3.11 5.54 2.47 2.23
1975 1.16 1.82 1.12 1.10 0.51 2.30 3.61 2.23 2.18
1976 1.69 1.98 1.19 1.19 0.51 3.31 3.87 2.33 2.33
1977 1.55 2.27 1.32 1.23 0.55 2.81 4.11 2.39 2.23
1978 1.57 2.35 1.28 1.19 0.64 2.45 3.66 1.99 1.85
1979 1.69 2.55 1.43 1.30 0.72 2.36 3.55 2.00 1.82
1980 2.05 3.02 1.63 1.61 0.79 2.60 3.84 2.07 2.04
1981 1.85 3.28 1.90 1.74 0.79 2.34 4.16 2.40 2.21
1982 1.60 3.07 1.85 1.69 0.77 2.09 4.01 2.42 2.21
1983 1.85 2.70 1.76 1.61 0.74 2.49 3.62 2.37 2.16
1984 1.79 2.81 1.85 1.74 0.73 2.45 3.86 2.54 2.39
1985 1.32 2.59 1.74 1.46 0.72 1.83 3.59 2.41 2.03
1986 1.06 2.38 1.67 1.37 0.83 1.27 2.87 2.01 1.66
1987 1.65 3.44 1.79 1.47 0.91 1.81 3.78 1.97 1.61
1988 1.40 5.67 2.00 1.63 0.97 1.45 5.86 2.07 1.68
1989 1.67 5.21 2.42 1.91 0.96 1.74 5.41 2.52 1.98
1990 1.82 4.47 2.64 1.83 1.00 1.82 4.47 2.64 1.83
1991 1.68 3.08 2.69 1.62 1.02 1.64 3.02 2.64 1.59
1992 1.28 3.03 2.52 1.62 1.06 1.21 2.86 2.37 1.53
1993 1.28 2.40 2.46 1.60 1.07 1.20 2.25 2.30 1.50
1994 1.76 3.23 2.27 1.65 1.11 1.60 2.92 2.06 1.50
1995 2.13 3.96 2.62 1.96 1.17 1.82 3.38 2.24 1.67
1996 1.77 3.26 2.59 1.75 1.11 1.59 2.94 2.33 1.58
1997 1.75 3.59 2.54 1.51 1.03 1.69 3.49 2.45 1.46
1998 1.44 2.75 2.42 1.34 1.00 1.45 2.75 2.43 1.34
1999 1.17 2.38 2.17 1.14 0.99 1.18 2.40 2.19 1.15
2000 1.30 3.53 2.15 1.26 0.97 1.34 3.64 2.21 1.29
2001 1.06 2.60 2.17 1.33 0.96 1.14 2.71 2.26 1.39
2002 1.02 3.72 2.16 1.32 0.96 1.04 3.88 2.24 1.37
a. Denotes the deflator, Manufactures Import Unit Value (1990 = 1.0).
Source. World Bank (A Index and MUV) and International Cotton Advisory Committee.
—67—
TABLE F9
VALUE OF EXPORTS OF OLD CLOTHING (MILLION US DOLLARS): 1975-2000
COUNTRY 1980/81 1985/86 1990/91 1995/96 2000/01
Industrial to Developing Countries
United States 81.0 81.1 116.6 227.3 188.5
Germany 8.8 14.9 36.0 103.5 123.0
United Kingdom 3.5 8.1 20.8 71.4 92.4
Belgium 28.7 35.9 64.9 97.9 70.8
Netherlands 21.3 22.6 42.0 69.8 69.6
Italy 1.5 3.4 28.3 57.1 52.7
Japan 20.7 27.6 40.8 49.9 46.5
Canada 2.5 4.8 11.0 43.0 39.0
France 7.2 8.0 19.4 29.9 31.9
Australia 5.2 5.1 7.6 19.1 13.4
Switzerland 0.1 0.2 0.7 6.0 9.8
Sweden 2.2 1.1 4.2 5.9 7.8
ALL 185.9 217.2 399.2 795.4 765.1
Industrial to Industrial Countries
ALL 223.0 218.4 262.1 393.5 247.9
Developing to Developing Countries
ALL 8.9 11.0 49.3 159.4 220.1
Developing to Industrial Countries
ALL 5.0 10.5 24.3 127.7 163.6
WORLD 422.7 457.2 734.9 1,476.0 1,396.8
Source. COMTRADE.
Notes: The 4-digit SITC code is 2690 and is defined as “Old clothing and other old textile articles; rags.”
The two 5-digit SITC codes (26901 and 26902) are defined as: “Clothing, cloth accessories, travel rugs &
blankets” and “used or new rags, scrap twine, cordage, rope & cables.” The 26901 code share of the 2690
code for the five two-year periods reported in the table are: 22.6%, 45.2%, 55.9%, 66.5%, 75.2%, and 80.7%.
—68—
TABLE F10
A CHRONOLOGY OF THE US COMMODITY PROGRAMS WITH COTTON PROVISIONS
PROGRAM YEAR MAIN PROVISIONS
Agricultural Marketing Act 1929 This Act was the first comprehensive program with the objective to stabilize commodity prices and farm
income. It created the Federal Farm Board, which made loans to marketing cooperatives for the purchase
and storage of surplus commodities, including cotton.
Agricultural Adjustment Act 1933 Aimed to control production and increase prices of designated “basic” commodities, including cotton to be
achieved by restoring farm purchasing power to its 1910-14 average level, a concept which became known
as “parity.” In response to low prices during 1933, the “non-recourse” loans (a form of floor price) were in-
troduced. Marketing quotas were also legislated in 1934 to prevent non-participants in the acreage control
program from sharing its financial benefits.
Supreme Court Decision 1936 The production control and financial features of the 1933 Act were declared unconstitutional.
Soil Conservation and Domes-
tic Allotment Act
1936 Provided for payments to farmers who agreed to adopt soil-building practices and shift land from “soil-
depleting” surplus crops such as cotton to “soil-conserving” crops such as legumes.
Agricultural Adjustment Act 1938 Provided for mandatory prices support loans and marketing loans keyed to acreage allotments. While the
cotton acreage declined considerably, output did not because of increasing yields.
Agricultural Act 1948 Provided for mandatory price support for cotton at 90 percent of parity if producers approved marketing
loans. Subsequent legislation extended this level of support through 1954.
Agricultural Act 1954 Renewed acreage allotments and marketing quotas due to increased production and stocks. Marketing quo-
tas continued until 1970.
Agricultural Act 1956 Established the Soil Bank, the objective of which was to: (i) reduce the amount of land planted to allotment
crops and (ii) provide long-term retirement of cropland to conservation uses.
Cotton-Wheat Act 1964 Authorized the Secretary of Agriculture to make payments to domestic textile mills in order to bring the
price of cotton used in the United States down to the export price. The allotment was also reduced. This Act
was the beginning of voluntary program for reducing cotton production.
Food and Agriculture Act 1965 Established a cropland adjustment program and introduced price support, set at 90 percent of estimated
world price level. For the first time, trade of allotments with a state was allowed.
continued N
—69—
TABLE F10 (continued)
A CHRONOLOGY OF THE US COMMODITY PROGRAMS WITH COTTON PROVISIONS
PROGRAM YEAR MAIN PROVISIONS
Agricultural Act 1970 Provided for a cropland set-aside program while it suspended the marketing quotas. It also imposed an
upper limit of $55,000 on program payments. This limit, however, had no impact as large producers di-
vided ownership of their farms.
Agriculture and Consumer
Protection Act
1973 Introduced target prices and disaster payments in recognition that agriculture faces weather and market
extremes which can result in low incomes. Payments would be made only if target prices fell below a cer-
tain level. The set-aside program continued.
Food and Agriculture Act 1977 Set target prices on the basis of the costs of production and a formula using cost estimates was used for
subsequent adjustments. The Act facilitated a shift of cotton production to lower cost regions of the West
and Southwest.
Agriculture and Food Act 1981 Focused on price and income support and provisions affecting their adjustment. Support was based on his-
torical moving average of per acre costs and actual yields. The Act also had provisions for acreage reduc-
tion. High target prices and weak demand led to large stock accumulation which in turn led to the pay-
ment-in-kind program.
Agricultural Program Adjust-
ment Act
1984 Reduced the target price from 86 cents per pound set by the 1981 Act to 81 cents per pound. It also re-
quired more land to be set-aside for conservation.
Food Security Act 1985 Retained major features of past programs, namely acreage limitations, nonrecourse loans, and target prices
but it gave more power to the Secretary of Agriculture. It specified declining target prices and introduced
the concept of deficiency payments.
Agricultural Reconciliation Act 1987 Reduced minimum price to 75.9 cents per pound from 81 cents per pound.
Food, Agriculture, Conserva-
tion, and Trade Act
1990 Target prices and deficiency payments continued. It introduced a new 3-step procedure in order to keep
US cotton competitive. The second step, the so-called Step-2 payment, is an export subsidy.
Federal Agricultural Im-
provement and Reform Act
1996 Introduced de-coupled support, i.e. payments based on historical area and output but it retained loan
rates. The Act was supplemented by a number of emergency payments during 1999-2001.
Farm Security and Rural In-
vestment Act
2002 Legitimized the emergency payments introduced in 1999. It raised the target prices and loosened the eligi-
bility criteria for support.
Source: Stults et al. (1989) and Glade et al. (1995).
—70—
TABLE F11
DIRECT GOVERNMENT ASSISTANCE TO COTTON PRODUCERS, 1997/98-2001/02
1997/98 1998/99 1999/20000 2000/01 2001/01 2002/03
Total production assistance (million US dollars)
US 597 1,480 2,056 1,020 3,001 1,996
China 2,013 2,648 1,534 1,900 1,196 750
Greece 659 660 596 537 735 718
Spain 211 204 199 179 245 239
Turkey na 220 199 106 59 57
Brazil 29 52 44 44 10 na
Mexico 13 15 28 23 18 7
Egypt 290 na 20 14 23 33
Total 3,812 5,279 4,764 3,822 5,287 3,814
Assistance per kilogram produced (US dollars)
US 0.15 0.49 0.56 0.27 0.61 0.54
China 0.44 0.59 0.40 0.43 0.22 0.16
Greece 1.94 1.85 1.37 1.27 1.72 1.97
Spain 1.82 1.96 1.51 1.90 2.33 2.42
Turkey na 0.25 0.36 0.12 0.07 0.07
Brazil 0.07 0.10 0.06 0.05 0.02 na
Mexico 0.06 0.07 0.21 0.19 0.20 0.18
Egypt 0.85 na 0.09 0.11 0.07 0.11
Cotlook A Index (July/August Average, US dollars per kilogram)
1.60 1.30 1.16 1.26 0.92 1.23
Assistance as a percent of the Cotlook A Index
US 9% 38% 48% 22% 75% 44%
China 27% 45% 35% 34% 24% 13%
Greece 121% 142% 118% 101% 187% 160%
Spain 114% 151% 130% 151% 253% 197%
Turkey na 19% 31% 10% 7% 5%
Brazil 4% 8% 6% 4% 2% na
Mexico 4% 5% 18% 15% 22% 15%
Egypt 53% Na 7% 9% 7% 9%
a. Data for 2001/02 are preliminary.
Source: International Cotton Advisory Committee (2002 and 2003).
—71—
TABLE F12
COTTON PRODUCTION IN EASTERN AND SOUTHERN AFRICA, 1970/71-2002/03
NIGERIA
SOUTH
AFRICA
TANZANIA
UGANDA
ZAMBIA
ZIMBABWE
1970/71 39 19 76 75 4 49
1971/72 38 20 66 75 3 60
1972/73 48 18 77 78 2 51
1973/74 30 40 65 50 1 57
1974/75 52 40 72 31 1 58
1975/76 61 19 42 25 1 47
1976/77 82 35 67 14 3 51
1977/78 40 51 50 20 3 60
1978/79 38 55 56 7 5 57
1979/80 29 65 61 5 9 65
1980/81 27 58 43 4 6 62
1981/82 21 38 40 5 5 56
1982/83 20 27 44 10 12 60
1983/84 13 36 48 12 16 91
1984/85 16 46 31 16 11 103
1985/86 10 47 67 5 12 89
1986/87 28 61 78 3 7 87
1987/88 30 78 54 2 24 116
1988/89 48 78 35 3 12 92
1989/90 41 60 48 4 9 67
1990/91 36 49 85 8 20 72
1991/92 38 20 96 7 9 88
1992/93 63 15 45 9 9 75
1993/94 52 27 40 5 13 60
1994/95 45 24 82 6 9 38
1995/96 60 45 87 10 30 104
1996/97 50 31 62 21 35 101
1997/98 70 42 36 7 42 105
1998/99 65 53 35 15 36 130
1999/2000 50 30 42 22 30 128
2000/01 55 36 45 19 24 152
2001/02 60 18 63 20 35 75
2002/03 55 21 67 22 43 122
Source: International Cotton Advisory Committee, Cotton: Rev