MICHIGAN STATE UNIVE 1 III ll llllllllllllllllll \/ 2 7 5828 319300 89 LIBRARY Michigan State University This is to certify that the dissertation entitled AN ECONOMETRIC ANALYSIS OF THE WORLD COTTON AND NON—CELLULOSIC FIBERS MARKETS presented by JONATHAN R. COLEMAN has been accepted towards fulfillment of the requirements for Ph.D degree in Agricultural Economics Major professor / w \ Date MI— MSUI': an Affirmative Action/Equal Opportunity Institution 042771 0!:!(6 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution c:\clr:\dstedue.pm3—p.1 AN ECONOMETRIC ANALYSIS OF THE WORLD COTTON AND NON-CELLULOSIC FIBERS MARKETS Volume I By Jonathan R. Coleman A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1991 ABSTRACT AN ECONOMETRIC ANALYSIS OF THE WORLD COTTON AND NON-CELLULOSIC FIBERS MARKETS By Jonathan R. Coleman The main purpose of this study was to specify and estimate an econometric model of the world fiber market, with emphasis on the cotton sector, and to forecast prices, production and consumption for major world fiber market participants. In addition, the model was used to analyze and measure the impacts of recent market developments and policy changes. The nature of the fiber market is described along with recent trends and market developments, providing the basis for the model specification. The model contains components explaining consumption, production, and pricing of cotton and non-cellulosic fibers for the major world fiber market participants. Estimated equations are combined to form a large simultaneous econometric model. A number of validation statistics are presented that cover various aspects of the model’s ability to forecast. Five sets of simulation results are presented. These are for (i) a forecast of price, production and consumption for the period 1990-2005, (ii) a 10% decrease in cotton production in the USSR, (iii) a 10% increase in cotton production in China, (iv) a 10% decline in the domestic price of cotton in the United States, and (v) an evaluation of the impact of the Multi-Fiber Agreement (MFA) on the cotton and non-cellulosic fibers sector S. 'V.“ fall app simulat: effect c in prod for etc about i cotton and to simula The model forecasts that between 1990 and 2005 the real world price of cotton will fall approximately 25%, while a 10% price increase is forecast for polyester. Three model simulations involve shocking key variables in major producing regions. In each case, the effect on the world market is significant. For example, given a permanent 10% decrease in production in the USSR, the world price rises by about 9%. Over an 11-year period, for every 1% increase in China’s production the world price of cotton falls, on average, about 1% and the price of polyester falls 0.35%. The impact of a 10% decline in the US cotton price during the early 19908 is to reduce US production, on average, less than 3.0%, and to increase world prices an average of 3.7%. The conclusion emerging from the MFA simulation is that the effects on the raw fiber market have not been large. l woo oourst were i indud appre the I nume ACKNOWLEDGEMENTS Iwould like to thank Stan Thompson for his guidance and encouragement throughout course-work stage of my program. His comments on an earlier version of this dissertation were useful. I also thank Jake Ferris, who later became my dissertation advisor, for his support and many helpful suggestions. The other members of my dissertation committee included Ken Boyer, Jim Hilker and Bob Myers. Their comments on the dissertation were appreciated. I am also indebted to my colleagues at the International Trade Division of the World Bank. In particular, Taka Akiyama, Ron Duncan and Elton Thigpen provided numerous helpful comments and suggestions throughout the duration of the study. TABLE OF CONTENTS List of Tables List of Figures 1. Introduction 1.1. Background 1.1.1. The Importance of Cotton and Manufactured Fibers Markets ......................... . 1 1.1.2. Market Instability 1.1.3. Recent Market Developments 1.2. Research Needs 1.3. Objectives of the Study 1.4. Organization of the Study 2. Research Methods 2.1. Modeling Commodity Markets 2.1.1. Overview 2.1.2. Steps in Commodity Model Building 2.1.3. Modeling Methods 2.1.3.1. Market Models 2.1.3.2. Spatial Equilibrium Models and Programming Models ............................. 2.1.3.3. Time Series Models 2.1.3.4. Other Classes of Models 2.2. Modeling the World Cotton and Non-Cellulosic Fibers Markets ............................ 2.2.1. The Cotton and Non-Cellulosic Fibers Markets 2.2.1.1. Cotton Production 2.2.1.2. Cotton Consumption 2.2.1.3. Cotton Price Determination 2.2.1.4. Manufactured Fibers Market 13 13 13 15 18 19 20 25 26 26 27 30 35 3.1.F 32. P 33. l 2.2.2. Model Structure and Specification ........................................................................... 39 2.2.3. Choice of Estimator -- - - . ..... - - - ..................................... 41 2.2.4. Model Validation - ......... - ........................................... 41 2.2.5. Model Simulation ......................................................................................................... 42 2.3. Summary .............................................................................................................................. 43 3. Cotton Demand ........................................................................................................................ 45 3.1. Features of Cotton Demand ........................................................... _ _ _, __ 45 3.2. Previous Studies -- - ............................................................................................... 46 3.3. Theoretical Issues - - - - - ...... 49 3.4. Model of World Cotton Demand .................................................................................... 51 3.5. Empirical Results - ..... . ................................. 54 3.5.1. Cotton Share Equations - . ..................................................... 54 3.5 ..2 Per Capita Total Fiber Use Equations - - - - . - . - 66 3.6. Demand Elasticity Estimates ............................................................................................ 75 3.6.1. Price Elasticities -- - — -- - ........... - - __ 75 3.6.2. Income Elasticities - - - - ................................................................ 78 3.7. Summary - - -- ------------- . 81 4. Cotton Production - -- ......................................................................................... 82 4.1. Introduction -- .. -- - - -- - ..... - - ...... 82 4.2. Theoretical Issues - - - - -- ......................................................... 82 4.3. Literature Review ..... - ...... - ............................ 84 4.4. Model of Cotton Supply ------ - - ...................................................... 86 4.5. Empirical Results - - — -- — ...... 89 4.5.1. Yield Equations -- ~ --------- -- 89 4.5.2. Area Equations ................. 99 4.6. Cotton Area Elasticity Estimates - - ....................... 108 4.7. Summary - - -, - .................................................... 109 vi S. COW 6. Nor 6.1. l 62. l 5. Cotton Price Determination 5.1. Introduction 5.2. Theoretical Issues 5.3. Literature Review 5.4. Cotton Price Determination and Estimation Results 5.5. Price Linkage Equations 5.6. Summary 6. Non-Cellulosic Fibers Model 6.1. Intr “ "inn 6.2. Demand for Non-Cellulosic Fibers 6.3. Supply of Non-Cellulosic Fibers 6.4. Price Determination 6.5. Summary 7. Validation of the Model 7.1. Root Mean Squared Percentage Error 7.2. Mean Squared Error 7.3. Theil’s U-Statistic 7.4. Graphical Validation 7.5. Summary 8. Model Simulation 8.1. Simulation One - Forecast to Year 2005 8.1.1. Rationale for Forecasting Prices 8.1.2. Forecasts of Exogenous Variables vii 110 110 111 113 114 117 119 120 120 120 128 129 130 131 132 135 143 145 149 153 153 153 154 /. (\AAJIH.).. -.8888 ......... 8.1.3. Forecast Results ---- - - .............................................................................. 155 8.1.3.1. Cotton Price Forecasts - -- - -- -- _- _ _ _ _____ 155 8.1.3.2. Cotton Production Forecasts ......................................................................... 158 8.1.3.3. Cotton Consumption Forecasts ..................................................................... 161 8.1.3.4. Non-Cellulosic Fibers Sector Forecasts ....................................................... 163 8.2. Simulation Two - A 10 Percent Decline in Cotton Production in USSR ........... 168 8.2.1. Background - ........ - ............................................................................... 168 8.2.2. Simulation Objectives ............................................................................................... 169 8.2.3. Simulation Results - - -- ................................................................... 169 8.2.4. Conclusion and Implications ................................................................................... 172 8.3. Simulation Three - A 10 Percent Increase in Cotton Production in China ....... 177 8.3.1. Background -- - - _-_.-_ ------ - ______ 177 8.3.2. Simulation Objectives ............................................................................................... 179 8.3.3. Simulation Results ..................................................................................................... 180 8.3.4. Conclusion and Implications ................................................................................... 183 8.4. Simulation Four - A 10 Percent Reduction in the United States’ Cotton Price . 186 8.4.1. Background -- - -- - - -_ - ------- ---- - 186 8.4.2. Simulation Objectives ................................................................................................ 188 8.4.3. Simulation Results ..................................................................................................... 189 8.4.4. Conclusion and Implications ................................................................................... 192 8.5. Simulation Five Analysis of the Multi-Fiber Agreement ...................................... 196 8.5 1. Background _ .............. -- -- ..................................... 196 8.5.2. Past Studies of Economic Effects of the MFA .................................................. 198 8 5 3 Simulation Objectives ..... - - -- - ....................................................... 199 8.5.4. Problems of Modeling the MFA ............................................................................ 200 8.5.5. Method of Incorporating the MFA into the Model .......................................... 202 8.5.6. Simulation Results - -- -- -- -- _- ....................................... 204 8.5.7. Conclusions and Implications .................................................................................. 209 8.6, Summary ............... - _ - -- . ....................................................................... 209 9. Sensitivity Analysis ................................................................................................................. 211 9.1. Introduction ........................................................................................................................ 211 9.2. Procedure .......................................................................................................................... 211 viii 93. Results ................... 93.1. Sensitivity of R 933 Sensitivity of E Total Fiber Us 933. Sensitivity of R Production 93.4. Sensitivity or rt Price 935. Sensitivity of It Non-Cellulosic 9.4. Conclusions .......... 10. Summary, Condusio 10.1. Summary ............ 192 Conclusions ........ 163. Areas of Future Aiitndk A - Variable Appendix B - Etogenou list of References ....... 93. Results ..... - - -- ................................... 213 9.31 Sensitivity of Results to the Equation Explaining World Cotton Price ........ 213 9.3.2. Sensitivity of Results to the Equation Explaining U. 8. Per Capita ................. 218 Total Fiber Use 9.3.3. Sensitivity of Results to the Equation Explaining Chinese Cotton ................ 223 Production 9.3.4. Sensitivity of Results to the Equation Explaining World Polyester ............... 229 Price 9.3.5. Sensitivity of Results to the Equation Explaining World ................................. 234 Non-Cellulosic Fiber Production. 9.4, Conclusions ...... - - -- - - ........................................................................................ 238 10. Summary, Conclusion and Areas of Future Research ................................................ 239 10,], Summary _____ _ - ........................................................................................... 239 10.2. Conclusions ......... . ................................................................................................ 242 10.3. Areas of Future Research ............................................................................................ 243 Appendix A - Variable Definitions ........................................................................................ 246 Appendix B — Exogenous Variable Assumptions for the Forecast Period ...................... 251 List of References _ , _ ........................................................................................ 253 1.1 Export Earnings f1 1987. 12 The Value of Out Fiber Production 1 13 Percent of Textile Merchandise Trad Textile Producing 11 Cotton Productior 1100 Tons. 1.1 Cotton Consumpt ’000 Tons. 23 NonCellulosic F11 1971119893000 To 2.1 Non-Cellulosic F11 Regions. 19711191 3‘1 Price Elasticitieu 2 111011 Elasticities] JIncome Elasticitie 411mme Elasticitie 1Elasticities of Col PM Elasticities 1 R001 Mean. -Squar Variables 12 MeanSquare Err Endogenous Vari Theil Ustatistics :12N0mina] and D131 R21311110111011 pr 2000 and 2005 83M11111 Cotton C( s210011nd2005 8431101181 No“ Cell“ PlOducfiOH for 113 “We Chan no"330111101311 33 3 11 6.1 3 LIST of TABLES Export Earnings from Cotton Production for Selected Countries, .......................... 2 1987. The Value of Output and Export Earnings from the Manufactured ...................... 3 Fiber Production of the Major Producing Regions, 1987. Percent of Textile Market in Value of Manufacturing Output, ............................... 4 Merchandise Trade and Manufacturing Employment for Major Textile Producing Countries, 1987. Cotton Production by Major Countries and Regions, 1970-1989, ........................... 29 ’000 Tons. Cotton Consumption by Major Countries and Regions, 1970-1989, ....................... 31 ’000 Tons. Non-Cellulosic Fiber Production by Major Countries and Regions, ....................... 37 1970-1989,’000 Tons. Non-Cellulosic Fiber Consumption by Major Countries and .................................... 38 Regions, 1970-1989, ’000 Tons. Price Elasticities of Cotton Use. 76 Price Elasticities Estimates Obtained From Previous Studies. ................................. 78 Income Elasticities of Demand for All Fibers. 79 Income Elasticities Estimated Obtained From Previous Studies. ............................ 80 Elasticities of Cotton Area with Respect to Price. 108 Price Elasticities of Non-Cellulosic Fibers Use. 126 Root-Mean-Square Percentage Errors for the Model’s Endogenous .................. 134 Variables. Mean-Square Error and its Decompositions for the Model’s ................................ 140 Endogenous Variables. Theil U-Statistics (U,) for the Model’s Endogenous Variables. ........................... 144 Nominal and Deflated Cotton Price Projections, 1990-2005. ................................. 156 Model Cotton Production Forecasts for the Years 1990, 1995, ............................ 158 2000 and 2005. Model Cotton Consumption Forecasts for the Years 1990, 1995, ........................ 162 2000 and 2005. Model Non-Cellulosic Fibers Forecasts of Price, Consumption and ................... 164 Production for the Years 1990, 1995, 2000 and 2005. Percentage Change in Cotton Variables for a 10% Decrease in .......................... 171 Cotton Production in the USSR. 8,6 Pctcenttlge CW9 10% Decrease In ‘ 1.1 Percentage Chang (Itina’s Cotton Pr 1.8 Percentage Chang Increase in China‘ 19 Percentage Chang Cotton Price. 8.10 Percentage Chang Decrease in 0.8. 8.11 Percent Change it Multi-Fther Agree 911 Impact on Foreca Explaining the Wt 116 Impact on Policy . Explaining the W1 921 Impact on Foreca Explaining U.S. P 926 Impact on Policy Explaining U.S. P 931 Impact on Form Enlamg Chine 93b Inpact on Policy Explaining Chine 9.11 Impact on Forecz E“Planting the W 911 Impact on Mar Equation Explain 953 1mPact on Forec; E‘Phlnmg Worlc‘ Impact on Policy EIPlaflling Work 956 f” p... H )0 N m )0 b.) h? .‘D U) 0‘ 9.4 b? 9.4 0' 9.53 9.5b Percentage Change in Non-Cellulosic Fibers Variables for a ............................... 172 10% Decrease in Cotton Production in the USSR. Percentage Change in Cotton Variables for a 10% Increase in ........................... 181 China’s Cotton Production. Percentage Change in Non-Cellulosic Fibers Variables for a 10% ...................... 182 Increase in China’s Cotton Production. Percentage Change in Cotton Variables for a 10% Decrease in US. ................ 190 Cotton Price. Percentage Change in non-Cellulosic Fiber Variables for a 10% ........................ 191 Decrease in US. Cotton Price. Percent Change in Cotton Variables Associated with the ....................................... 206 Multi-Fiber Agreement, 1979-1986. Impact on Forecasts of Parameter Changes in the Equation ............................... 215 Explaining the World Cotton Price. Impact on Policy Analysis of Parameter Changes in the Equation ..................... 217 Explaining the World Cotton Price. Impact on Forecasts of Parameter Changes in the Equation ............................... 221 Explaining U.S. Per Capita Total Fiber Use. Impact on Policy Analysis of Parameter Changes in the Equation ....................... 222 Explaining U.S. Per Capita Total Fiber Use. Impact on Forecasts of Parameter Changes in the Equation ................................. 226 Explaining Chinese Cotton Production. Impact on Policy Analysis of Parameter Changes in the Equation ..................... 228 Explaining Chinese Cotton Production. Impact on Forecasts of Parameter Changes in the Equation ............................... 231 Explaining the World Polyester Price. Impact on Policy Analysis of Parameter Changes in the Equation ..................... 233 Equation Explaining the World Polyester Price. Impact on Forecasts of Parameter Changes in the Equation ............................... 236 Explaining World Polyester Production. Impact on Policy Analysis of Parameter Changes in the Equation ....................... 237 Explaining World Polyester Production. World Non-Calla and 2005 Comp! Impact of a 10% World Cotton P1 1111 Impact of a 10'! World Cotton C 1.11 Impact of a 10‘! World Cotton P 112 Impact of a 101 World Cotton 5‘ 1.13 Impact of a 109 World Polyester 111 Impact of a 109 World Non~Cell 9.; 3.5 3.6 I. 3.8 11 12 LIST OF FIGURES World Cotton Price, 1960-1989. ....................................................................................... 6 World Polyester Price, 1960-1989. ................................................................................... 6 Model Representation of a Commodity Market. ....................................................... 14 World Cotton Price and Stocks, 1970-1988. ............................................................... 34 Diagram of the World Cotton and Non-Cellulosic Fibers Model. ......................... 40 The Regression of Actual Against Simulated Values. ............................................. 139 World Cotton Production, Actual Vs Simulated Values. ........................................ 146 World Cotton Consumption, Actual Vs Simulated Values. ................................... 147 World Cotton Price, Actual Vs Simulated Values. .................................................. 148 Non-Cellulosic Production, Actual Vs Simulated Values. ....................................... 150 Non-Cellulosic Consumption, Actual Vs Simulated Values. .................................. 151 World Polyester Price, Actual Vs Simulated Values. .............................................. 152 Nominal Cotton Price, Projections, 1990-2005. Outlook Index "A". ..................... 157 Deflated (MUV) Cotton Price, Projections, 1990-2005. Outlook ......................... 157 Index ”A”. World Cotton Production Trends, Years 1990 and 2005 Compared. .................. 160 World Cotton Consumption Trends, Years 1990 and 2005 ................................... 163 Compared. Nominal Polyester Price, Projections, 1990-2005. ..................................................... 165 Deflated (MUV) Polyester Price, Projections, 1990-2005 ....................................... 165 Ratio of Cotton and Polyester Prices. ......................................................................... 167 World Non-Cellulosic Fibers Consumption Trends, Years 1990 ............................ 167 and 2005 Compared. Impact of a 10% Decrease in Cotton Production in the USSR on ..................... 174 World Cotton Price. Impact of a 10% Decrease in Cotton Production in the USSR on ..................... 174 World Cotton Consumption. Impact of a 10% Decrease in Cotton Production in the USSR on ..................... 175 World Cotton Production. Impact of a 10% Decrease in Cotton Production in the USSR on ..................... 175 World Cotton Stocks. Impact of a 10% Decrease in Cotton Production in the USSR on ..................... 176 World Polyester Price. Impact of a 10% Decrease in Cotton Production in the USSR on ..................... 176 World Non-Cellulosic Fibers Consumption. “51111111061101; WorlchnCeIluI nomination 1111mm IlllmpattciaIO‘b 1111110111de tltlmpactoialo'ttt Worlchlyorterl 119 Impadoi110% WorldCottonPn 121 Imptdof110% WorldCottonCu 12111111111011.1111 CottonProduotio MlmpactofalO‘b CottonProdtnio Mlmpactofaw'k 00110an 121 Impactot'110% PclyesterPtice. WImpactofaIO‘Xa CottonStoclrs. 126 Impactola 10% Cotton IlllmpactofMuIti-I Consumption, 19 MlmpadofMulti-I 19794986. I291mpactofMulti-T 8.15 Impact of a 10% Decrease in Cotton Production in the USSR on ..................... 177 World Non-Cellulosic Fibers Production. 8.16 Impact of a 10% Increase in Cotton Production in China on .............................. 184 World Cotton Stocks. 8.17 Impact of a 10% Increase in Cotton Production in China on .............................. 184 World Cotton Price. 8.18 Impact of a 10% Increase in Cotton Production in China on .............................. 185 World Polyester Price. 8.19 Impact of a 10% Increase in Cotton Production in China on .............................. 185 World Cotton Production. 8.20 Impact of a 10% Increase in Cotton Production in China on .............................. 186 World Cotton Consumption. 8.21 Impact of a 10% Decrease in Cotton Price in the US on US ............................ 193 Cotton Production. 3.22 Impact of a 10% Decrease in Cotton Price in the US on World ....................... 194 Cotton Production. 3.23 Impact of a 10% Decrease in Cotton Price in the US on World ....................... 194 Cotton Price. 3.24 Impact of a 10% Decrease in Cotton Price in the US on World ....................... 195 Polyester Price. 3.25 Impact of a 10% Decrease in Cotton Price in the US on World ....................... 195 Cotton Stocks. {.26 Impact of a 10% Decrease in Cotton Price in the US on World ....................... 196 Cotton Consumption. .27 Impact of Multi-Fiber Agreement on United States Fiber ................................... 207 Consumption, 1979-1986. .28 Impact of Multi-Fiber Agreement on BBC Fiber Consumption, ......................... 207 1979-1986. 29 Impact of Multi-Fiber Agreement on World Price of Cotton, .............................. 208 1979-1986. 30 Impact of Multi-Fiber Agreement on World Polyester Price, ............................... 208 1979-1986. 11mm infirm—0 Cotton is an inn n'gtifitantlytofarmino manned Afri earned more than $1 tgrinrlturalexport can ctpandforeiguacha chnrh'naFaso in 198 nadndioetadewhi 1111mm expat gmeratedfisbillionit morethan42% ofitsa Washroom“ manyindustmltzed’ ' cor tyiutlturalcommodity tamed3521 billionfr alesamotmtedtoonly 63% and 2.5% for tht 51%ottheagricultura 1. Introduction 1.1 Background 1.1.1 The Importance of the Cotton and Manufactured Fibers Markets Cotton is an important agricultural commodity in many countries, contributing significantly to farm income and export earnings. Table 1.1 shows the export earnings from cotton of selected African, Asian and Industrialized countries. In 1987, cotton exports earned more than $1 billion for African countries, representing about 9.5% of total agricultural export earnings. In some African countries cotton was the principal export crop and foreign exchange earner. For example, 84% of the agricultural export earnings of Burkina Faso in 1987 were derived from cotton sales, amounting to almost 70% of its nerchandise trade; while in Sudan and Egypt, cotton sales contributed more than one-half if all agricultural export revenues. In Asia, cotton is relatively less important but still enerated $3.3 billion in export revenues in 1987. For Pakistan, the sales of cotton earned tore than 42% of its agricultural foreign exchange earnings in 1987. China and India also erived a large proportion of their agricultural export earnings from cotton exports. In any industrialized countries and centrally planned economies cotton is also an important ricultural commodity and /or industrial raw material. For example, in 1987 cotton exports 316d 5 5.21 billion for the United States and $0.35 billion for Australia, although these as amounted to only a small proportion of total agricultural export earnings, representing 70 and 2.5% for the United States and Australia, respectively. In contrast, more than '7 of the agricultural export revenues of the USSR were obtained from exporting cotton, undammbilli MM!!! ”31 BcrtimFaso Egypl Tanzania aereeaa ussn W 1111: value of out man producing more: in the industrialized 1 leading World supplier. valued almost $55 bi lued at over $1.2 billion. Me 1.1 Export Earnings from Cotton Production for Selected Countries, 1987 :gions Cotton Percent of Agric. Percent of Merchandise 1d Export Export Earnings Trade runtries Earnings ($ million) rkina Faso 43.0 84.3 69.6 ad 31.7 24.9 19.8 ypt 388.9 56.2 13.2 11 46.3 30.6 22.3 lan 185.1 59.5 55.5 tzania 43.2 12.6 10.4 go 34.3 33.1 14.5 RICA 1,069.3 9.5 NA na 756.1 8.9 NA a 175.0 7.5 NA istan 445.2 42.3 NA A 3,319.6 1.9 NA tralia 353.0 2.5 NA . 5,213.0 6.3 NA R 1,221.0 51.0 NA :6: FAQ Trade Yearbook, 1987. The value of output and export earnings from the manufactured fiber production of 'producing regions are shown in Table 1.2. As shown there, production is centered : industrialized and centrally planned countries, although China is emerging as a 1g world supplier. In 1987, total manufactured fibers production of the United States I almost $5.5 billion. Most of the production was consumed by domestic textile 1- 5W E11012 11m Korea Taiwan United States USSR World Solace: Fiber Organo1 FAO. 2131191 Cotton and man in a great many count in the United States higher-income devote Industries provide a la umakinganmportm 3 anufacturers, while earnings from manufactured fiber exports amounted to $0.44 billion. the EEC, over one-half of the total manufactured fiber output, valued at $4.96 billion, LS sold externally. Manufactured fiber sales were also a very important source of foreign :hange for many developing countries in Asian, especially China, Korea and Taiwan. ble 1,; The Value of Output and Export Earnings from the Manufactured Fiber oduction of the Maior P ‘ " ' 1987. gion Value of Value of Output Exports --- $ Billion --- 'ma 1.80 0.72 Europe 2.14 0.26 C-12 4.96 2.54 an 2.63 0.46 ea 1.81 0.91 van 2.49 0.99 ted States 5.48 0.44 R 2.35 0.00 '1d 27.40 8.28 'ce: Fiber Organon. June 1989. FAQ. World Apparel Fiber Consumption Survey, 1989. Cotton and manufactured fibers are consumed by textile and apparel manufacturers great many countries throughout the world. Traditionally consumption was centered 3 United States and Western Europe, but recently this has moved to many of the r-income developing countries in Asia. In these countries, textile and clothing tries provide a large proportion of total manufactured output and employment, as well king an important contribution to merchandise trade. As reported in Table 1.3., 57% oi the value of merchan was provided by its I- contributed more than t 01 the employment for States and Western Eu still make an ilnport: employment MIME. 11111 and Man Region Bangladesh China India Japan Korea “Pug Kong Singapore PM ines United States W. EUI'Ope \ Source, 1% w 1.1. 2 WI The world Cottot rate the early 19603. 4 of the value of merchandise trade and 64% of employment in manufacturing in Bangladesh was provided by its textile sector in 1987. In Korea and Hong Kong textile sales contributed more than one-quarter of the value of manufacturing output and over one-third of the employment for its manufacturing work force. The textile industries of the United States and Western Europe are relatively less important in their respective economies, but still make an important contribution to the balance of trade and as a source of employment. Table 1.3 Percent of Textiles Market in Value of Manufacturing Output, Merchandise Trade and Mr “ L in: F ' ‘ for Maior Textile P. ’ ' Countries 1987. Region Value of Merchandise Manufacturing Manufacturing Trade Employment Output ---- Percent ---- Bangladesh NA 57 64 Ihina 18 24 13 ndia 12 18 23 apan 7 0 14 Lorea 25 23 35 fong Kong 35 39 44 ngapore 4 6 4 iilippines 8 6 26 nited States 5 2 10 . Europe 3 4 7 urce: Industrial Statistics Yearbook, United Nations, 1987. World Development Report, 1988. World Bank. 2 Market Instabilig The world cotton and manufactured fibers markets have undergone substantial change 1 the early 19603. These changes have led to substantial price volatility (Figures 1.1 and 12). Cotton urice iu1973, 1976, 1980, 19 associated with the acct inproduution and com 1960 and 1970 the cot utiutthred a floor on from higher rates of g growth in the United throughout the period harvest in that year, w 1980, resulted from ti United States and in II Mina had a major in 1984 sent prices to 99 be associated with countries. The P01yester p] from increased prod manufactured fibers t aresult of the first 0i With the general fates bthnpiememed Withj manufactured fibers 5 and 1.2). Cotton prices tend to be cyclical with peaks every three or four years (such as in 1973, 1976, 1980, 1983 and 1990). In general, the peaks and troughs in prices can be associated with the accumulation and run-down of cotton stocks, which result from changes in production and consumption decisions by market participants. For example, between 1960 and 1970 the cotton price remained stable as a result of US cotton policy which maintained a floor on the world price. The 1970-74 period of increasing prices resulted from higher rates of general inflation, as well as a simultaneous expansion in economic growth in the United States, Japan and the EEC, which led to declining stock levels :hroughout the period. The drop in price in 1974 was the result of an extremely good rarvest in that year, while the rise in price between 1974 and 1976, and between 1977 and .980, resulted from the acreage diversion from cotton to grains which occurred in the Jnited States and in many other major cotton growing areas. During the 19803, production 1 China had a major effect on the world price. For example, the record production level I 1984 sent prices to very low levels in 1985, while the price increases in the late 1980s m be associated with persistent strong demand for cotton, especially by the industrialized tuntries. The polyester price (fob US plants) declined between 1960 and 1973. This resulted )m increased productive capacity and improvements in technology which allowed mufactured fibers to compete with cotton. The price reversal that started in 1973 was 'esult of the first oil shock, while the rise in price throughout the 1980s was consistent h the general rates of inflation. Given that production and consumption decisions can implemented within a relatively short period, there is very little stockholding of nufactured fibers. As a result, there is an absence of the inter-year fluctuations that are to = O 200 ‘- 175—— 150'— 125 '— 100 _ 75— so~'-L 1960 tIMiddling 1~a/32'. t «2" 275 : 225\~ 175\ 125\ 75 «a 1960 ID 7~'\O 6 Fit}. 1.1 World Cotton Price. 1960-1989 Cotton Outlook 'A' index 1/ 225 200 175 150 125 100 75 50 1960 1965 1970 1975 1980 1985 1990 Year __._ Cotton Price Middling 1-3/32', c.i.f North Europe to X\O Fiq. 1.2 World Polyester Price, 1960-89 FOB US Plants 275 225 175 \ 125 k\ \ r 1 I I 751 l L l l i | 1950 1965 1970 1975 1980 1985 1990 Year i l 2}_l##l } l l l l I l | I I l —*— Polyester Price produce: about 15% o artfully and are seen The recent polio The introduction of 1 production between 1‘ cotton During this p be highly mauve It to them. Given the i Bgorerator of foreign 7 n in the cotton price series. However, the strong competition between manufactured er and cotton in the 1980s caused the prices of these to move fairly consistently, with th price series declining in the mid-19805 and rising during the 1986-89 period. 1.3 Recent Market Developments ‘ There are a number of developments which some market commentators (e.g., ICAC, trld Bank, USDA, Fiber Organon) have argued are of great significance to the recent my and future performance of the fiber market. These developments center on the fits of changes in domestic cotton policy in the major producing countries. The agricultural sector in the USSR is currently undergoing major policy-induced ctural changes. In an effort to improve production and productivity, the government introduced price and property-right incentives, encouraged greater accountability by :ntralizing decision-making, improved input access and quality, and strengthened the 1 between agricultural research centers and the farming sector. Since the USSR uces about 15% of the world cotton supplies, these developments are being monitored Tully and are seen as a major determinant of fiber market prospects in the early 1990s. The recent policy changes which occurred in China are also being watched closely. iintroduction of price incentives in the late 19703 led to a three-fold increase in ’ction between 1977 and 1984, with China emerging as the leading world supplier of . During this period and continuing throughout the 1980s, producers were found to y responsive to changes in the profitability of the various crop enterprises available m. Given the importance of the cotton and textile sectors for employment and as rator of foreign exchange, the government is expected to continue its support for the US crops more com determiningaa’eager world ootton prices I Significance now than will have an important international fibers no Another import the outcome of the no of the ongoing disa mutations on the int (nutty the United SI \_____ 'luituttmm molten 'Wmmmumm “Monumental“ 8 :ton sector during the early 19903. An unofficial target growth rate in production has :11 reported at 3% per year between 1990 and 1995. If this growth is achieved it will 'e great significance for cotton prices during this period and beyond. In the United States, the 1990 Food, Agriculture, Conservation, and Trade Act will ermine cotton policy until 1995. The Act continues the market-orientated provisions oduced in the Food Security Act of 1985. The major changes are provisions to make crops more competitive internationally, and to increase flexibility in the rules :rmining acreage reductions, in order to allow farmers to respond to market signals. target price for the entire 1991-1995 period will be fixed at the 1990 level of 72.9c/lb. : means that real target prices will fall significantly over the next five years. This is in effort to reduce overall agricultural expenditures in accordance with the Gramm- man deficit reduction requirements, as well as in response to the expectation of lower :1 cotton prices throughout the early 19903. While the United States has less Eicance now than in the past’, United States policy provisions in the 1990 Farm Act ‘ave an important impact on the United States cotton sector in particular, as well as rational fibers markets in general. Another important development being monitored closely by many commentators is tcome of the negotiations over the Multi-Fiber Agreements (MFA) which are part 1 on—going discussions in the Uruguay Round of the GATT. The MFA places tions on the imports of textile and clothing items into some industrialized countries y the United States and the EEC) from many developing countries. These evolved and 19703 the US support price acted as a floor on the world price. The emergence of China in the 19803 made nsive and led to the policy changes which were contained in the Food Security Act 1985. A brief discussion of US nted in section 8.4.1. stance of the United S of free access to inte mammal mm 12 MM Anmte cotton uh as producers and me about future entries where and provide an impo addition, forecasts of countries and region: Example, the identifie depend on neonate 1 Remix is net and consumption. W Monetary Fund, the I time to time, these = Mon models, C author is aware. the 9 in the 1960s and have been increasingly strengthened since then, as the market share of domestically produced textile and clothing products in the industrialized countries declined. [t is hard to predict the outcome of these negotiations. This will depend mainly on the tance of the United States, which has to balance its overall accordance with the principle f free access to international commodity and manufactured goods markets, with the owerful political interests of the textile and clothing manufacturers at home. 2 Research Needs Accurate cotton and manufactured fiber price forecasts are useful to many groups, rch as producers and consumers of fibers, governments and lending institutions. Informed lesses about future fiber prices are especially important to decision makers in those tuntries where earnings from fiber sales represent a large proportion of export earnings .d provide an important source of employment in manufacturing (Tables 1.1-1.3). In dition, forecasts of world levels of production and consumption trends for different untries and regions are valuable to producers and consumers of fiber products. For mple, the identification and development of potential export markets for many countries end on accurate forecasts of where demand will be strong in the future. Research is needed to provide accurate forecasts of future fiber prices, production consumption. While a number of institutions (e.g., the World Bank, the International netary Fund, the International Cotton Advisory Committee) publish price forecasts from to time, these are based on a fairly simple analyses, such as using single-equation ession models, or are based on observations of recent price trends. As far as the or is aware, there exists no price and quantity forecasts based on a more complete mnmp therenppemtobeoor rophistietedmethodr Inadditiontode toinvertigateotberiss outheefiectsthatsorr haveonfutureprieel For example, a producer and stock-h analyses have been u world fiber market he1990e Sud! in (him cotton counties. There are on the world fiber rm Another impor (such as those in the prime A quantitatir Participants for men Finally, while u tgreement (MFA) r (Goto mi, 1989), i Measuring the impa fit .0 'amework in which price and quantity forecasts are derived simultaneously. Therefore rere appears to be considerable scope for improving current forecasts by applying a more )phisticated method which captures the interaction of price and quantity. In addition to developing a method of forecasting prices and quantities, there is need investigate other issues and areas in the fiber market. For example, research is needed r the effects that some of the recent markets developments (discussed in section 1.3) may ve on future price levels and production and consumption patterns. For example, as mentioned above, China recently became the most important )ducer and stock-holder of cotton in the world market. Currently, no quantitative rlyses have been undertaken to measure the degree to which China has affected the rld fiber market and how the expected expansion in production will impact on prices in 19905. Such analyses would also be useful in assessing the impact of future changes China’s cotton policy on world prices and supply and demand conditions in other ntries. There are also no studies which measure the impact of USSR cotton supplies the world fiber market. Another important issue concerns how changes in United States cotton programs as those in the 1990 Farm Bill) will affect United States producers and world fiber s. A quantitative analysis of US cotton policy would be of interest to many market cipants for monitoring future changes in United States agricultural programs. Finally, while some research has been completed on the impact of the Multi-Fiber ment (MFA) on the developing countries and the cost of the distortions created §t_a_l, 1989), its effects on raw fiber demand and prices have not been evaluated. uring the impact of the MFA on cotton and polyester producer prices will provide valuable input into the nadtett 13 Objective of the § mprovem' entonthe (hemodelwillbe ' dwelnpmentsinindiv u Quanta—doom this paper is r are dimmed in set determination compr tfipecdvely. In each theoretical issues, an fibers component 01 M “mammalian or. 11 rable input into the debate over the how the MFA may have distorted the world fibers rkets. Objective of the Study The objective of the study is to address the research needs mentioned above. That 0 develop a complete framework with which to make price forecasts for cotton and ester prices, in addition to make projections for quantities produced and consumed. framework will involve the specification, estimation and simulation of an econometric CI of the world cotton and manufactured fibers markets. This approach allows for the 'action of price and quantities to be forecast simultaneously and therefore is an 'ovement on the methods applied currently to fiber market forecasting. In addition, model will be simulated to provide quantitative measures of the impacts of recent lopments in individual countries on the world cotton and manufactured fibers markets. trganization of the Paper This paper is organized into ten sections. The research methods used in the study iscussed in section two. The cotton demand, cotton supply and cotton price ination components of the model are presented in sections three, four and five, tively. In each of these sections a review of the relevant literature, a discussion of tical issues, and the regression results are presented. In section six the non-cellulosic component of the model is presentedz. Section seven presents some model -cellulosic fibers are polyester, nylon and acrylic. These make up about 80% of total world manufactured fibers validation results In reported and discusset policy issues outlined i model variables to kt summary and oonclusir model development an 12 :lation results. In section eight, forecasts from the model up to the year 2005 are )rted and discussed. Also presented are model simulation results which address the :y issues outlined in section 1.3. In section nine, sensitivity analysis of some important lel variables to key parameter estimates are presented. Finally, in section ten a mary and conclusion of the study are provided, along with some proposals for further e1 development and analyses. In order to add analytical tool was req provide the best tech modeling approaches addressing the probler commodity models in noncelluiosic fibers n 21 Modeling Cummt 2.1.1 W A commodity n a quantitative relationships i well as other I A very Supple corm demand/supply diagr quantities demanded themarlret Clearing p: the “imitation mo' “)qu models c iiitnnined by Stern and changing Itastes a demand for raw fiber 2. Research Methods In order to address the research problems outlined in the introductory section an ytical tool was required. A quantitative model of the world fiber market appeared to pvide the best technique to meet the study’s objectives. In this section the various l deling approaches are discussed followed by a description of one most suited to lressing the problems at hand. The stages involved in model building are reviewed for rrnodity models in general, and then specifically for the model of the world cotton and -cellulosic fibers markets developed in this study. Modeling Commodigz Markets Overview A commodity model can be defined as: a quantitative representation of a commodity market or industry; the behavioral relationships included reflect demand and supply aspects of price determination as well as other related economic, political and social phenomena (Labys (1988, p.4)). ry simple commodity model can be represented by the standard Marshallian nd/supply diagram. Building such model would require specifying relationships for ities demanded and supplied and combining them to solve for two unknown variables, rket clearing price and equilibrium quantity. However, for most commodity markets o-equation model is inadequate to capture all components and complexities. Most odity models contain the elements shown in Figure 2.1. Commodity demand is ined by external factors, such as income levels, general price levels, population size, anging tastes and preferences; the demand by end users of the commodity (e.g., the d for raw fibers is derived ultimately from the demand by consumers for 13 131a! Commodi‘ Invento Source: W.C. Wm 14 Figure 2.1 Model Representation of a Commodity Market External Influences 0n, End—Use Demand Demand 1 Commodity IDemand Commodity l 3Commodity i Inventories! Prices Commodity Supply External Influences On Productive Supply Capacity Source: W.C. Labys and P..Pollak Commodity Models for Forecasting and Policy AnalysiS. London: Groom-Helm Publishing Co., 1984. manufactured textile substitutes and oomph weather, the state of disparity, such as the 2 prices of the commod: to determine the prio isrelativeiy simple. h supply and demand countries or regions, 2 2.12 $th in Com Constructing a . it must be decided n W from the mods then. and long-term Used to “fldertake P0: [0 sPeciiic economic used to evaluate the : responses of marken The Second St 8mm” given the s Involves finding a In nature of the marke 15 manufactured textile and apparel items); and the prices of the commodity and its substitutes and complements. Commodity supply is determined by external factors, (e.g., eather, the state of technology and government policy and programs); by productive pacity, such as the area of farm land allocated to producing the commodity; and by the ices of the commodity and its inputs. The demand and supply components are combined determine the price and level of inventories. Even the model represented in Figure 2.1 relatively simple. Many commodity models are more complex and contain, for example, pply and demand relationships for a number of distinct producing and consuming untries or regions, and relationships for the trade of commodities between model regions. .2 Steps in Commodity Model Building Constructing a commodity model requires a number of distinct steps or stages. First, must be decided the purposes for which the model is to be used and what information on from the model’s output. For example, some commodity models are used to make rt- and long-term projections of prices, production and consumption, while others are :l to undertake policy experiments, answering "what if" questions about market responses pecific economic, policy or institutional changes. Commodity market models can be to evaluate the impact of such changes on past market responses, or predict the likely nses of markets participants to policy changes in the future. The second stage in model construction involves choosing the appropriate model ure given the set of model requirements. This is perhaps the most crucial stage. It es finding a model structure which has a strong theoretical basis and reflects the e of the market and how it actually operates, while allowing the model to meet the for the variables and determined by the e equation However. hpractiee more §d_i shooting a specified ’ressonable’. The fourth Sta parameter values for Variables appearing a own-price appearing &.kmh’ Mandala “Ml-limingw “kw"MM the Em 16 objectives of the research assignment. In general, the key issue in determining the model structure is how the price is established in the market. Once this has been determined the odel can be structured to capture this price determination. The price determination is cial because not only is the price variable usually of most interest in forecasting and policy experimentation, but also price enters the demand, supply, stock-holding and trade helationships contained in the model. The third step in commodity model building is model specification. This involves apturing relationships between the model variables in the form of equations that are onsistent with the model structure. In addition to choosing the set of independent ariables for each equation, specification involves choosing the appropriate lag structures sr the variables and functional forms for the equations. Initially model specification is :termined by the economic theory of the behavioral relationships captured in each puation. However, most commodity models are loosely based on economic theory and practice more ad hoc approaches to equation specification are used ‘. This involves oosing a specification that yields a good fit and provides elasticity measures that are sonable’. The fourth stage in model building is estimation which involves obtaining the ameter values for the equations. Most commodity models contain current endogenous bles appearing as regressors on the right-hand side of the structural equations (e.g., -price appearing in demand equations is an endogenous variable in most market ple, standard theory of consumer behavior tells us that consumers maximize their utility from consuming a set of goods income constraint. Solving this maximization problem yields demand functions with income and prices as arguments. by commodity model builders in choosing the variables to go into the demand equations (i.e., demand equations should d income as regressors), however, the theory does not give direction as to which prices to include, whether the variables or not, what the functional form the equation should take, and so on. mm For these n simultaneous aquatic The filth step i using a series of stat historical data or to 4 data. A major prohle m'teria orbenchman 17 clearing models). In this case, the model is simultaneous. Using the ordinary least squares estimator (OLS) in the estimation of a system of simultaneous equations provides estimates ‘of parameter values that are both biased and inconsistent (Johnston (1972); Intriligator 1978)). An estimator within the single-equation method is the instrumental variable rechnique of which two stage least squares (ZSLS) is the best known. In 281.8, the Endogenous variables which appear as regressors are replaced by instrumental variables, :reated (in stage one) by regressing the endogenous regressors on all the exogenous 'ariables in the model. In stage two the instrumental variable replaces the right-hand side -ndogenous regressor and the equation is re-estimated using 01.5. The 251.8 parameters re biased but consistent. Monte Carlo studies have shown 281.8 to be superior to most ther techniques based on small sample properties. Moreover, these desirable properties '6 less sensitive to other estimation problems such as multicolinearity and specification ror than other estimators. In addition to these benefits, 281.8 has low computational ists. For these reasons ZSLS is the most popular and widely used estimator of nultaneous equations 2. The fifth step in model building is model validation which requires testing the model ng a series of statistical tests. Typically, these tests are of the model’s ability to track torical data or to determine how well the model captures turning points in the historical .a. A major problem in validating a multi—equation model is that no statistically objective eria or benchmarks exist by which to accept or reject a validation statistic. The criteria l l methods of estimation include 3SLS and full information maximum likelihood techniques which estimate all the structural lset instead of estimating the structural parameters of each equation separately. The major benefit of using these fat they use all the available information in creating their estimates and provide consistent parameter values. Errors in r, are transmitted to estimates in the whole model and are not confined to the equations in which the errors occur. require a large sample size and have high computational costs. For these reasons systems methods are rarely used for tsedareatbitrarily model as satisfactory er-rnte foremsting a evaluating alternativ: if models are used i The final step system of equau'ons variables Models common algorithms the Newton meth Newton method than other techniqr 7.13 rim-mm. Commodity 1 motleling method different classes 0 Niels There an 18 ed are arbitrarily chosen. As in single-equation estimation, the decision to accept a odel as satisfactory depends upon the intended use of the model. Models designed for j-ante forecasting are typically put through more rigorous tests than those developed for 'aluating alternative policy scenarios. However, having an accurate baseline is important models are used in policy analysis. The final step in commodity model building is simulation. This involves solving the item of equations which make up the model to give values for the unknown endogenous 'iables. Models are normally simulated over successive time periods. The three most nmon algorithms used to solve these systems of equations are the Gauss-Seidel method, Newton method and the Jacobi method. These are available in most of the large nometric software packages. The use of any of these algorithms yields identical ulation results. However, the Gauss-Seidel method is the most commonly used because llows normalized equations to be solved within the system of equations (which the Iton method does not) and is more efficient in terms of speed and computer capacity other techniques. Modeling Methods Commodity models are used to address a wide range of problems and no single ling method has become the standard. However, it is possible to categorize the ent classes of models. Labys (1988) identified the major classes of commodity ls. These are discussed briefly below. 7.1.3.1 Met Moth Marketmodel afiwgq ”Q g r! :4 This class of determined compe‘ “mm 001mm) may not be deter-m alilJFDpl’inte. 2.1.3.1 Market Models Market models are the most commonly applied form of commodity models involving pecifying demand, supply, stock-holding and price behavior in terms of econometrically stimated relationships. Price is determined through a market clearing identity. A typical tructure of such models is as follows, 0", = Q"(P,, P,°, L, T‘) Q‘l = Q'(P,_i, W" Gt ) P. = M. ) I, = Q'. + 1..l - Q“. = quantity demanded at time t, quantity supplied at time t, commodity price at time t, = level of stock-holding at time t, price of other commodities at time t, = income level at time t, = consumer tastes and preferences at time t, = weather at time t, and = government policy at time t. 161'6: Q“. 0‘. P. 1. Pt° Y. T. W. G. This class of model is appropriate for those commodity markets in which price is nrmined competitively. Models of this type have been most widely applied to cultural commodity markets. In modeling minerals and energy markets in which price not be determined by competitive adjustment, other modeling approaches are more ‘opriate. between regions ml solution to the pro! and trade flows be commodity market: “mum: of commodity marl “‘1 marketing a Si “Winds The ob Woman: eqml 20 2.1.3.2 Spatial Equilibrium Models and Pram ' Models Spatial price equilibrium models are the most common form of agricultural trade model used for comparative static analysis of policy changes. The spatial equilibrium model problem is described by Takayama and Judge (1971) as follows: We are given in each of two or more regions demand and supply functions for a given product in terms of its market price at that location. In addition, unit transportation costs are also given for carrying the product between the locations. Under this specification we would like to know what will be the (i) competitive equilibrium price in each location, (ii) the amount of supplied and demanded at each location, and (iii) level and pattern of exports and imports. Most models of this type are solved using quadratic programming. In these models the objective function is given by the maximization of the area under all excess demand curves, minus the area under all excess supply curves, minus transportation costs. This function is then constrained by the following requirements: (i) the quantity entering a region must be less than or equal to the quantity demanded, (ii) the quantities leaving a egion must be less than or equal to the quantities supplied, and (iii) price differences tween regions must not be greater than the transportation costs between them. The olution to the problem gives prices and quantities demanded and supplied in each region nd trade flows between them. Such models have been applied to many agricultural mmodity markets. Programming techniques are also used in transportation and plant location studies commodity markets. The objective function is minimization of the costs of producing d marketing a specific amount of commodity in a given time period or over a number periods. The objective function is constrained in that the total supplies of the producing gions are equal to the total demands of the consuming regions, and that for any intendediate node (8 node is equal to the additional constraint: Some programming establishing a new p solved using mixed i 2.133 Time Series I A method cor Time series models applications to polir Order and Faclrler following componer observations on the lid random disn variable are explain However, more Gite 1“ mixed autoregn generating the Vari trailer as Well as E economic data gem (“mm ‘C. . Pit midtefimewmvlt 21 intermediate node (e.g., transhipment region or processing plant) the quantity entering the node is equal to the quantity leaving the node. Optimal plant location models contain additional constraints limiting the amount of product passing through each processing plant. Some programming models contain zero-one variables allowing the fixed costs of establishing a new processing unit to enter the objective function. This class of model is solved using mixed integer programming. 2.1.3.3 Time Series Models A method commonly used in commodity market modeling is time series analysis. Time series models have been used mainly for forecasting, although, more recently, applications to policy analysis have appeared in the literature (e.g., Myers Lal (1991); Orden and Fackler (1989) ). Time series models typically contain one or both of the following components, (i) a moving average model, in which the process generating bservations on the variable are described completely by a weighted sum of current and aged random disturbances; and (ii) an autoregressive model, in which the values of the rariable are explained by the weighted sum of its past values and a random disturbance. iowever, more often, the moving average and autoregressive representations are combined 1 a mixed autoregressive-moving average model (ARMA model). In this case the process enerating the variable is a function of both lagged random disturbances and its past dues, as well as a current disturbance term (Pindyck and Rubinfeld (1981) ). Often :onomic data series are found to be non-stationary}. When this is the case the ARMA mple, Palaslras and Varangis (1989) found that most commodity prices series are 1(1). That is, they are stationary after ne time. representation is no 3 or more times to ob model can be estim: integrated to obtain autoregressive-mom macroeconomic and Another clam This model has boo model is given by th Whereut = a (m) Identity covariance contemporaneous in defining dl’ltamic in From the eqr determined by com The “‘Odel differs f there are to huge; variablen Seconct Proponents of such data, the true Struc 22 'epresentation is no longer valid. However, non-stationary series can be differenced one r more times to obtain a stationary series and a mixed autorregressive-moving average node] can be estimated with the differenced data. Forecasts from this model can be itegrated to obtain forecasts in terms of the levels of the variable. Such integrated utoregressive-moving average models (ARIMA) have been applied widely to forecast acroeconomic and agricultural variables (Pindyck and Rubinfeld (1981) ). Another class of time series model is vector autoregression (VAR) (Sims, 1980). is model has become a popular tool for forecasting. The basic structure of a VAR odel is given by the equation 2.1. Bysi Baden, (21> i=1 lere: u[ = a (nxl) vector of zero mean, serially uncorrelated disturbance terms with an ntity covariance matrix; A and B are (nxn) parameter matrices representing ttemporaneous interrelationships among y, and u,; and Bi are (nxn) parameter matrices l ing dynamic interactions among yt (Myers _e_t_a_l (1991)). From the equation we see that current values of the endogenous variables are mined by current and lagged values of all the endogenous variables in the system. model differs from a standard simultaneous model in two important respects. First, 6 are no exogenous variables in the system explaining variability in the endogenous bles. Second, there are no restrictions placed on the lagged variables. Also, onents of such models argue that by imposing minimal restrictions on the economic the true structure of the economic system is revealed and that the over-identifying ctions normally imposed on structural models are overly-constraining and their validity often untested (MW While VAR r problems in applimt Sims (1986) discusse One is that, such models usually based into a useful of policy cho understand tt Another argument is Sargent (1984) claim Slstem and therefor (1936) argues that tl models to guide p0 greater structure is i m°d°l tYpe should r Studies in whi i“ totem literature entails first setting 1 by B.1 gives the red Where Ci = BIlBl {C \ MMMVARM 23 often untested (Myers et al (1991), Fackler ( 1988) ). While VAR models have been used widely for forecasting, critics have pointed to problems in applications to policy analysis (e.g., Sargent (1979) and Learner (1985) ). Sims (1986) discusses arguments against the use of forecasting models for policy analysis. One is that, such models are nothing more than summary descriptions of the historical data, usually based on sample correlations. While such a description can be extrapolated into a useful forecast, supposing that it can be the basis for projecting the effects of policy choice amounts to taking correlations to indicate causation, which we all understand to be fallacious (Sims, 1986. p.2). Another argument is that policy variables are treated as random variables in VAR models. Sargent (1984) claims that policy decisions are rarely made without regard to the economic system and therefore should not be treated as exogenous to the model. However, Sims (1986) argues that these arguments do not constitute an objection to the use of forecasting models to guide policy choice. VAR models differ from econometric models in which greater structure is imposed merely from differences in identifying interpretations, and one model type should not be considered superior based on theoretical criteria. Studies in which VAR models have been used for policy analysis have been reported recent literature (e.g., Myers et al (1991); Orden and Fackler (1989) ). The procedure ntails first setting up a VAR model as in equation 2.1. Next, premultiplying equation 2.1 B“ gives the reduced form equation 2.2‘. III y 1:; Ciyt-i+vt (2'2) i=1 ere C, = B"Bi for i=1,2,....,m; and v( = B"Au,. This equation is used to estimate the ion of VAR models is based largely on Myers et al. madame matrix 0f (he equation 22 l5 covariance matrix (6 be expressed as equ lhe likelihood fund Aand B. This ofte inorder for the nm of unique paramete gm information beset equal to zert Policy analysi A“Alysia (IRA) or 1 average representa‘ WhereDo = 1yl A' f, And. the matrices l 24 ovariance matrix of the reduced form disturbances (equation 2.3). o=B-1AA ’3“ (23) ['he equation 2.2 is estimated using OLS. Then, assuming normality and employing the :ovariance matrix (equation 2.3), the log-likelihood for a set of T observations on y, can be expressed as equation 2.4. T A = -0.5110g|B“AA ’B"'| —o.5): v,’B’A “IA ‘le,. (2.4) i=1 The likelihood function is maximized to obtain estimates of the VAR parameter matrices A and B. This often requires that identification restrictions be placed on these matrices, in order for the number of estimated parameters in A and B to be equal to the number of unique parameters in the covariance matrix of the VAR. Restrictions can be based on 1.21% information of the structure of the market, allowing some elements of A and B to )e set equal to zero. Policy analysis can be undertaken with VAR models with either Impulse Response Analysis (IRA) or Forecast Error Variance Decomposition (FEVD). For IRA, a moving verage representation of equation 2.2 is derived as equation 2.5. ”:2 DluH-r-flt) (2.5) i-O here D0 = B"A; f(t) is a function of t that is identically zero if yl is covariance stationary; d, the matrices Di (i=1,2,....) can be computed from the recursion (equation 2.6), Equations 2.5 and 2.t variables y,. For FE A and B (from eqr prediction errors for structural shock in u Myers M (1 polio shocks on de industry. Orden at Variables (such as rm and agricultural pric these macroeconomi Hibl'id prog] model often involvi demand is derived fisheries sector are Optimlltttior 19703 as a result ( 25 mill) D,= )3 0,0,1 (26) j=1 Equations 2.5 and 2.6 are used to show the impact of a shock in one element of up on the variables y,. For FEVD, the covariance matrix of v, depends on the parameter matrices A and B (from equation 2.3), which can be used to decompose the variance of the prediction errors for each element of y, into components due to the variance of each structural shock in up Myers fl (1991) provide IRA and FEVD for the effects of demand, supply and policy shocks on demand, supply, price, and producer revenue in the Australian wool industry. Orden and Fackler report IRA for the effects on various macroeconomic variables (such as money supply, oil price, price level, output, interest rates, exchange rates, and agricultural prices), of shocks to the money supply; and for the effects of shocks to these macroeconomic variables on agricultural prices and general price levels. l l 2.1.3.4 Other Classes of Models Hybrid programming models combine many different modeling methods in a single nodel, often involving the modeling techniques of other disciplines. For example, the :upply side of some energy models are developed from engineering criteria, while the lemand is derived econometrically. Similarly the supply side of many models of the sheries sector are developed from biological growth models. Optimization models were developed in response to the oil price shocks in the early '70s as a result of the non-competitive behavior of the institutions involved in oil production and cons behavior based on t forms of market strut models using this tet System Dyna some form of cyclica from a variety of sor applied to livestock, luputOutpm Itarrsionned into or Primary commodity cons“minim patter macrowonomic vari. mineral markets, as r inoome multipliers) 26 production and consumption. These models can be employed to model non-competitive behavior based on the Coumot equilibrium concept and can accommodate alternative forms of market structure ranging from the competitive market to monopoly. Most of the models using this technique have been applied to the oil market. System Dynamics models have been applied to commodity markets which contain some form of cyclical behavior in production and prices. These models utilize information from a variety of sources including econometrics, engineering and biology and have been applied to livestock, oil, coal, and copper markets. Input-Output models are used to analyze how resources in the form of inputs are :ransformed into outputs. However, they can not be used to explain the behavior of )rimary commodity markets. Instead they provide information on how production and onsumption patterns for different commodities relate to industry structure or macroeconomic variables. Input-Output models have been applied mainly to energy and mineral markets, as well as to the measurement of economic impacts (e.g., employment and tcome multipliers) of an industry in a region. 2 Modeling the World Cotton and Non-Cellulosic Fibers Markets 2.1 The Cotton and Non-Cellulosic Fibers Market l In order to choose the correct model structure it is essential to have a full derstanding of the cotton and non-cellulosic fibers markets, and especially of the ‘rerminants of production, consumption and prices. It is also important to decide which l i the most important producing and consuming countries and regions of the world, as se need to be given special attention in the model structure. This section provides a brief overview of the account in the mode 1.2.1.1 Cotton Prodr Cotton is prodr 40 degrees south. Tl widely in terms of si crops, such as soyb relationship between tends to be the ma production has up improvements in 00 luptoved yields hav. and machinery costs 42%, While the area resulted from Cousin Wider use of fertiliz magi Colton is i order to separate th The "with con increased from 12 r intense Whig b. 27 trief overview of the major features of the world fiber market which must be taken into ccount in the model’s structure. ;.2.1.1 Cotton Production Cotton is produced in over 80 countries, spread widely between 40 degrees north and -0 degrees south. The types of farm on which cotton is grown are heterogeneous, ranging Iidely in terms of size, soil type and technology use. Cotton competes with other arable rops, such as soybeans, rice and coarse grains for farm resources and therefore the elationship between the price of cotton and the prices of these competing commodities ends to be the major determinant of the area planted to cotton. Over time, world roduction has expanded rapidly. This has been associated mainly with significant nprovements in cotton yields in most of the cotton producing countries of the world. nproved yields have maintained gross margins in cotton production despite higher labor Id machinery costs. Between 1970 and 1989, world average cotton yields have increased (‘71:, while the area planted has expanded only 7%. The improvement of average yield sulted from considerable successes in the development of improved seed varieties, the ler use of fertilizers and chemical treatments, and the expansion of irrigated cotton 'eage. Cotton is harvested in the form of seedcotton and is then processed (ginning) in ler to separate the seeds from the fiber or lint. The major cotton producing countries and regions are shown in Table 2.1, along with ntities produced at five year intervals since 1970. World cotton production has eased from 12 million tons in 1970 to 17.5 million tons in 1989, with most of the rase occurring between 1975 and 1985 when production grew almost 50%. Most of the expansion occurred i and 1989. In compé Asa result, the prop about 60% in 1970 ' Of the develt increasing slowly in 1 and 1980s, China en more than 50% be incentives to grower aPPW that mode. fluctuated widely be government India Between 1970 and almostthreedold I greater availability 28 expansion occurred in the developing countries where output increased 60% between 1970 tnd 1989. In comparison, production of the industrialized countries increased only 23%. \s a result, the proportion of world output supplied by the developing countries grew from tbout 60% in 1970 to 67% in 1989. Of the developing countries, production has risen most rapidly in Asia, while ncreasing slowly in South and Central America and stagnating in Africa. During the 1970s and 19805, China emerged as the world’s leading producer and increased its production by nore than 50% between 1980 and 1985 following the introduction of favorable price ncentives to growers, as well as improved access to farm inputs. From Table 2.1 it may tppear that production stabilized between 1985 and 1989. In fact, production has luctuated widely between years as a result of unstable producer prices which are set by the overnment. India and Pakistan have also dramatically increased their cotton production. tetween 1970 and 1989, production in India almost doubled, while in Pakistan it grew lrnost three-fold. In both cases the expansion can be attributed to yield improvements and cater availability of inputs. lhle 2.1 Pr / Region Industrialized NAmerita United States Europe EEC East Europe USSR Asia/Oceania Australia Japan Developing Africa Egypt (‘fiilifaj Africa Asia] Oceania China indie Korea, ep PikiStan Turkey Si Cent America Alienfina Brazil Mexido World 29 T bl 2.1 t on Pro uc ion b M ’or o ntrie and Re ions 1970-1989 ’000 Ton . Region 1970 1975 1980 1985 1989 Industrialized 4,772 4,556 5,648 6,067 5,887 NAmerica 2,220 1,808 2,422 2,925 2,632 United States 2,220 1,808 2,422 2,925 2,632 Europe 2,532 2,723 3,127 2,885 2,928 EEC 165 173 174 238 303 East Europe 22 22 13 12 12 USSR 2,345 2,528 2,940 2,634 2,613 Asia / Oceania 20 25 99 257 327 Australia 20 25 99 257 327 Japan 0 0 0 0 0 Developing 7,272 7,247 8,503 11,126 11,656 Africa 1,267 985 1,144 1,256 1,290 Egypt 509 382 529 435 283 Central Africa 263 381 501 697 821 Asia / Oceania 4,677 4,991 5,703 8,239 8,545 China 2,287 2,374 2,701 4,138 4,138 India 1,017 1,184 1,374 1,829 1,960 Korea,Rep 5 3 3 1 0 Pakistan 544 623 719 1,241 1,546 Turkey 400 480 500 518 599 S.& Cent America 1,328 1,271 1,656 1,631 1,821 Argentina 84 140 85 110 218 Brazil 490 395 622 820 751 Mexico 316 196 347 209 174 orld 12,004 11,803 14,151 17,193 17,543 ource: USDA. World Cotton Situation, FAS, Various Issues. The rise in prc lhe major producer 95% of the total ind the United States an the Australian cotto leading industrialize 2.2.12 Cotton Con: Cotton fibers manufactured textile tnnmon, mixed wit} to give the fabric dr “impedes most stro film the relative foul world fiber 00, ““9 u 40% of m for 10% and 5%) n The major t “ll quantities com risen Sleadily, me world consumption developing 00am 30 The rise in production in the industrialized countries has been much less dramatic. The major producers are the United States and the USSR which together supply about 95% of the total industrialized countries’ output. Between 1970 and 1989, production in the United States and the USSR increased 19% and 16%, respectively. During this period, the Australian cotton sector became established and Australia is now third among the leading industrialized cotton growing countries, producing over 325,000 tons in 1989. 2.2.1.2 Cotton Consumption Cotton fibers are consumed directly by the textile industry for processing into manufactured textile and clothing items. Cotton is used as a sole fiber or else, as is more common, mixed with other natural or non-cellulosic fibers. Manufacturers use fiber blends to give the fabric desired characteristics such as strength, durability and comfort. Cotton competes most strongly with polyester and rayon fibers and manufactures’ consumption reflects the relative prices of these different fibers. In 1989, cotton contributed 45% to total world fiber consumption, while the non-cellulosic fibers (acrylic, nylon and polyester) made up 40% of the market. The cellulosic fibers (acetate and rayon) and wool accounted for 10% and 5%, respectively. The major cotton consuming countries and regions are shown in Table 2.2 along with quantities consumed at five year intervals since 1970. World cotton consumption has risen steadily, increasing from 12.5 million tons in 1970 to 18.5 million tons in 1989. The world consumption expansion has been driven by the rapid increase in consumption in the developing countries. W ___________..—- Region 1“ ”.1 NAmeriea United States Europe EEC East Europe USSR Asia/Oceania Australia Japan Developing Africa Egypt Central Africa Asla/Oceania China India KOIQRep Pakistan Turkey S.& Cent America Argentina Brazil Mexico World 31 T le 2 n um tion b M 'or and Re ions 1970-1989 ’ Ton . Region 1970 1975 1980 1985 1989 Industrialized 6,495 6,297 6,039 6,359 6,605 NAmerica 1,862 1,632 1,342 1,448 1,746 United States 1,787 1,579 1,283 1,394 1,697 Europe 3,838 3,948 3,957 4,205 4,137 BBC 1,263 1,204 1,077 970 1,287 East Europe 705 738 788 829 751 USSR 1,786 1,917 1,993 2,091 1,982 Asia/ Oceania 795 717 740 706 722 Australia 31 28 22 21 25 Japan 764 689 718 685 697 Developing 5,983 7,185 8,361 9,898 11,821 Africa 431 528 656 742 758 Egypt 203 232 326 .337 283 Central Africa 68 88 117 154 163 Asia/Oceania 4,778 5,695 6,672 7,938 9,836 China 2,287 2,505 3,289 3,811 4,356 India 1,136 1,364 1,398 1,566 1,851 Korea,Rep 120 199 315 370 457 Pakistan 442 467 445 510 926 Turkey 175 290 296 451 599 S.& Cent America 774 962 1,033 1,218 1,227 Argentina 105 116 83 114 120 Brazil 303 430 550 675 849 Mexico 155 174 160 146 147 orld 12,478 13,482 14,400 16,251 18,426 ource: USDA. World Cotton Situation, FAS, Various Issues. Between 1970 and 1 in the industrialize‘ consumption by the Consumptior example, China now 19705 India replace Asian countries sue have all expanded t world manufacturer United States and l the abundant supp advantage in prodr 0°“fillies promptec Filer Arrangement Wild in higher Almost three-fold si Willi” 111 Conn; Sim the EEC tl 22.1.3 Comm Prir While there there are a great It or region which h: 32 Between 1970 and 1989 the developing countries almost doubled their consumption, while in the industrialized countries it changed little. As a result, the proportion of world consumption by the developing countries increased from 48% in 1970 to 64% in 1989. Consumption growth in the developing countries of Asia has been dramatic. For example, China now consumes almost one—quarter of the world’s cotton, while since the late 19705 India replaced the United States as the third largest consumer of cotton. Other Asian countries such as Hong Kong, Singapore, Bangladesh, Taiwan, Thailand and Korea have all expanded their consumption substantially. The Asian region now dominates the world manufactured textile and clothing sector, replacing the traditional centers of the 'Jnited States and Europe. Because of the labor intensive nature of textile production and he abundant supply of cheap labor, the Asian countries have found a comparative .dvantage in producing textile products. The loss of market share in the industrialized ountries prompted the establishment of a system of tariffs and quotas (e.g., the Multi- iber Arrangement). In other developing countries the availability of cheap labor also :sulted in higher demand for cotton. For example, Brazil increased its consumption most three-fold since 1970, while demand grew in Egypt and many of the central African untries. In contrast, the major consumers of the industrialized countries (i.e., the United ttes, the BBC, the USSR and Japan) changed their consumption little. 1.3 Cotton Price Determination While there are a few major cotton producing and consuming countries and regions, e are a great many participants in the world fiber market. There is no single country egion which has either monopolistic or monopsonistic control over the market and there is iIGEdOm to e homogenolls comma established and rec man is and“ Given a 00“ aocording to the let clearly in the relatiO Figure 2.2. As worl no and 1981; and 1933; and 1985 and inrecent years the The 'world 1 Cotton Outlook '4 Europe’. The priC cotton throughout t and paid. The cottt standards. Price c' marketing system handlers. Traditior lactorslor determi that influence pro< an ‘ ‘Blhtooe WMBQKM 33 l there is freedom to enter and exist the market at fairly low cost. Further, cotton is a fairly homogenous commodity and international grading and standards of cotton quality are well lestablished and recognized. These conditions indicate that the competitive market framework is applicable to the cotton market. Given a competitive market framework, the price of cotton is expected to move according to the levels of production and consumption in the market. This is borne out clearly in the relationship between world price and world ending stocks which is shown in Figure 2.2. As world stocks increase, the world price falls (e.g., between 1976 and 1977; [980 and 1981; and 1983 and 1985) and vice versa (e.g., between 1977 and 1980; 1981 and 983; and 1985 and 1986). In Figure 2.2 world stocks exclude stocks held in China because at recent years these have been isolated from world markets. The ’world price’ of cotton is generally accepted by market participants to be the Zotton Outlook "A" Index c.i.f price for middling 1&3 /32" staple quoted for North turopes. The price is based on market intelligence provided by buyers and sellers of mm throughout the world, who provide information on a daily basis on the prices offered id paid. The cotton market has a well established set of quality specifications and grading andards. Price determination based on well-known quality factors at all stages of the parketing system is necessary for transmitting the correct incentives to producers and lndlers. Traditionally, grade and staple length have been the most commonly used quality ptors for determining cotton’s trading price. However, more recently, fiber characteristics it influence processing performances have earned a premium. :e is the one reported by the World Bank in Price PMS for Primag Commodities, and by the IMF in the International '19 publications. a? x - \F wV—oonflw UCMW mOmLAHh 34 l-ocm «Eco mEoonm 8.02m uto>> \p $200.5 ll moEa l amm> mew, meme emmF «was 0mm? mama ohms enmp «saw one? m 4 _fl 4 w _ + _ w _ fl _ u _ “ fllTJlOm oow O\_‘£ 0) Dow CON 22.1.4 measure The manufat acetate and triacetz cellulosic fibers don the noneellulosic fil fibers production ‘ production. In cor libero production at in the 1930 production centere celulosic fibers inc 1960s the producti lhe Place of cellulo was largely the res countries since the 991p and energy, t laced stiff price cc 1980s, cellulosic 1 However, productj some developing p World con Per capita World ( kg PEI year Tl‘ 35 2.2.1.4 Manufactured Fibers Market The manufactured fibers market includes the markets for cellulosic fibers (rayon, acetate and triacetate) and non-cellulosic fibers (nylon, polyester and acrylic). While cellulosic fibers dominated in the early stages of manufactured fiber development, recently the non-cellulosic fibers have become more important. In 1988, 83% of the manufactured fibers production was non-cellulosic fibers amounting to 45% of total world fiber production. In contrast, cellulosic fibers production contributed 17% to manufactured fibers production and only 8% to total fiber production. In the 19305, manufactured fiber production was mainly of cellulosic fibers with production centered in the United States and Western Europe. World production of cellulosic fibers increased to over 2 million tons in the 19505 and 19605. However, in the 19605, the production of non-cellulosic fibers (mainly polyester and acrylic) began to take the place of cellulosic fibers in textile products. The decline in cellulosic fiber production was largely the result of pervasive decreases in production capacity in the industrialized :ountries since the early 19705. The cause of this decline was the increased cost of wood )ulp and energy, the main inputs into cellulosic fiber production. Also, cellulosic fibers aced stiff price competition from cotton fibers during the 19605 and 19705. During the 9805, cellulosic fiber production continued to fall to about 1 million tons in 1989. fowever, production remained important in Eastern Europe and the USSR, as well as in me developing countries such as India and China. World consumption of cellulosic fibers has fallen along with production. In 1970, .' capita world consumption was 1 kg per year, while in 1988 it had fallen to below 0.5 per year. The decline in consumptiOn occurred since the 19505 in industrialized countries. Aim ant USSR The major no along with quantitit increased from 4.7 n bythe developing Ct 52 million tons in developing countrie production capacity tspedtlly in China, lube dominated by Sharply, declining i USSR and Eastern ulnently supply ab The major r 2‘4 “99% With quar there is no stockin 99°98- As not. signifiml’ Since and Western Ellro declined lullowinp developing Count; Climb willCli by 1 36 )untries, Africa and Latin America, since 1970 in East Europe, and since 1980 in the ’SSR. The major non-cellulosic producing countries and regions are shown in Table 2.3 ong with quantities produced at five year intervals since 1970. World production creased from 4.7 million tons in 1970 to 14.2 million tons in 1988. Production expansion 'the developing countries has been dramatic, increasing from 0.3 million tons in 1970 to 2 million tons in 1988. In 1970, 6% of total world production was produced by veloping countries and by 1988 their share had risen to over 36%. Rapid expansion of oduction capacity caused the production growth to be particularly strong in Asia and )ecially in China, India and Korea. Production in the industrialized countries continues be dominated by the United States although its share of total world supplies has fallen rrply, declining from 86% in 1970 to 63% in 1988. The non-market economies of the SR and Eastern Europe have also increased their production of non-cellulosic fibers and rently supply about 10% of the world market. The major non-cellulosic fibers consuming countries and regions are shown in Table along with quantities available for consumption at five year intervals since 1970. Since L'e is no stock-holding of non-cellulosic fibers, availability is given by production plus net orts. As indicated earlier, the consumption of non-cellulosic fibers has increased ificantly since 1970. The main consuming areas in the 19705 were the United States ;Westem Europe. Throughout the 19705 and 19805 the market share of these regions l ined following the rapid development of textile manufacturing capacity in the p lloping countries. The most important consumer to emerge during the 19805 was la, which by 1988 had exceeded the consumption level of Western Europe. l Tulz N uulaur / Region / Industrialined NAmerita United States West Europe Germany italy Spain United Kingdom Japan Others Non-Market USSR East EurOpe Developing 3dr Cent. Americ Brazil Mexico Asia China lndia Killed. Rep Africa World 37 Table 2,3 Non-cellulosic Fibers Production by Major Countries and Regions, 1970-1988, ’ Ton Region 1970 1975 1980 1985 1988 Industrialized 4,044 5,421 6,861 6,893 9,007 NAmerica 1,573 2,541 3,364 2,997 3,281 United States 1,509 2,445 3,242 2,864 3,147 West Europe 1,475 1,839 2,128 2,478 2,565 Germany 492 625 720 762 759 Italy 214 277 355 563 565 Spain 64 119 202 275 268 United Kingdom 337 361 288 243 197 Japan 970 1,021 1,357 1,403 1,360 Others 25 21 12 16 63 don-Market 355 792 1,139 1,412 1,738 USSR 167 362 550 413 868 East Europe 188 430 589 699 870 Developing 302 1,139 2,476 4,194 5,232 S.& Cent. America 147 395 612 699 783 Brazil 44 126 321 217 240 Mexico 47 155 239 306 345 Asia 136 718 1,797 3,369 4,299 China 52 280 806 1,756 2,398 India 16 33 70 194 321 Korea, Rep 43 263 536 812 1,115 i tfrica 19 26 66 127 150 brld 4,700 7,353 10,476 12,499 14,239 i 1urce: Textile Organon. Various Issues. i Ta 24 N lltl 1913.95 /- Region ___———-————_ NAmerita United States West Europe France Germany Bah United Kingdom East Europe a/ USSR China b / India indonesia Korea Rep, Japan Other Asia 3' & Cent, Ameri Brazil Mexico Africa s\ ounce; Mlle Or a/ Data for only l M MCludes Tam 38 Table 2,4 Non-cellulosic Fibefi Consumption by Major Countries and Regions, 1970-88, ’fl; Tons Production Plus Net Imports Region 1970 1975 1980 1985 1988 NAmerica 1,719 2,476 2,922 3,068 3,001 United States 1,619 2,332 2,746 2,857 2,978 West Europe 1,343 1,639 1,832 2,053 2,193 France 175 205 206 247 242 Germany 324 332 368 379 341 Italy 194 239 3 13 423 434 United Kingdom 269 335 233 287 363 East Europe a/ NA 189 240 807 764 USSR NA NA NA 790 879 China b/ NA NA NA 2,000 2,280 India 22 44 90 232 252 Indonesia NA 95 124 206 212 Korea Rep. 81 272 514 783 916 Japan 772 604 1,046 1,000 1,043 Other Asia 75 513 866 519 309 S. & Cent. America 176 448 680 700 791 Brazil 55 142 241 236 246 Mexico 48 158 251 230 292 ica 35 38 167 164 175 ource: Textile Organon. Various Issues. Data for only Hungry and Poland in 1975 and 1980. Includes Taiwan 22.2 new From this bl various types of war 2.13.1. was the mt structure characteri allows the forecasti An overview cotton is determine at world production defined in the me Regional cotton or consumption and tr isderived as the pr Bier consumption. behaviorally as m, determined by into film IS estimated the 99199810 price World regiOn (1h 6 39 2.2.2 Model Str_ucture and Specification From this brief review of the world fiber market and from the discussion of the various types of commodity model, it appeared that the market model discussed in section 2.1.3.1. was the most appropriate form. The competitive nature and overall market structure characteristics of the fiber markets are well suited to this model class, which also allows the forecasting and policy analysis objectives of the study to be met. An overview of the model is presented in Figure 2.3. As shown there, the price of cotton is determined by the level of world cotton stocks which is derived as the residual of world production plus ending stocks less consumption. Cotton production in each region defined in the model is derived from behavioral equations for yield and area levels. Regional cotton consumption is derived as the product of cotton’s share of total fiber consumption and total fiber consumption. Similarly, the demand for non-cellulosic fibers is derived as the product of non-cellulosic fibers’ share of total fiber consumption and total fiber consumption. The cotton and non-cellulosic fibers’ share equations are estimated behaviorally as functions of cotton and polyester prices; and total fiber consumption is determined by income, population and general price levels. The supply of non-cellulosic fibers is estimated at the world level as a function of polyester and crude oil prices, while the polyester price is estimated using an inverted demand equation for the Rest-of-the- orld region (there are no non-cellulosic fibers stocks). . I ll. I .—. — — Qua-W whoauh 00min: U-MC—S-unvhvr-hcz LOuWON-Cnu luv-wen). Who's-urn! U-hAv-Au-unvb .ln Bub-u Inc-90 hunt-avg ”it has Efihhfu: MttoN Avis-H533 40 Hi .fi cote—.25 53m? .1 - «05¢ U =c=OU 5:30 .8390 Ric—Each macho 368m 8m...— eoNEtom ”5.2.—:90 mcEem » 8E 9.3.20:th he 02.:— 5590 5:50 =Efiam . _ 9:25 _. a .8on out.— :0 :0 t comes—E3250 zeta—=9:— * it 5:25 2:85 cit—6:5 23E gums—3.0 .5. o z 5.3.—.350 34E :25. :EEEa—EO 5.3.—£3ch max—E Tr is”. 22:..— u_m=_==oo.=cz I vie—£30-52 r 253 he «8% .5533— a ltlllllb ._w ’_tllo.:=_m ESE _. Ewe—330-52 82m conga—em 223 W From the dis to be the most app byMonte Carlo exp In the first regressing current r army models (inclu exceeds the numbe of 2315. To overc used as regressors of exogenous varia is to select endogenctt variables 8 B1B approach wa knowledge of the r marlets‘. A nilmbep Eiillndilon “ages MIC modelrs ablfif .2.3 Choice of Estimator From the discussion of estimation techniques given in section 2.1.2, ZSLS appeared be the most appropriate to avoid the inconsistency of OLS. This choice is supported Monte Carlo experiments summarized by Intriligator (1978). In the first stage of the ZSLS procedure, instrumental variables are created by Fgressing current endogenous variables on all exogenous variables within the system. For any models (including the one developed in this study) the number of exogenous variables 'ceeds the number of observations and the degrees of freedom problem prevents the use ZSLS. To overcome this problem a subset of exogenous variables can be selected and ed as regressors in the first stage. No hard and fast rules exist on how to choose the set exogenous variables used in the first stage. Intriligator suggests one criterion: is to select only those exogenous variables that are most closely related to the endogenous variable in the equation, excluding from each equation those exogenous variables believed to be unimportant on the basis of a priori considerations. 's approach was taken in this study with the set of instruments chosen based on rwledge of the relationships between variables within the cotton and non-cellulosic fibers kets“. 4 Model Validation A number of validation statistics were performed after the specification and nation stages of the model building were successfully completed. The tests evaluated model’s ability to reproduce actual data and to respond to external shocks in a way ely, principle components can be created which are themselves instrumental variables and which capture a specified liability in the set of exogenous variables. The principle components are then used as regressors in the first stage. For be consistent the exogenous regressors on the right-hand side of the equations must be included explicitly in the first stage. consistent with em included: (0 the R0 Error (M35), (ml 7 to withstand these I it could be used for the results for the t 225 MM Forecast at objectives set out i forecast of price, increase in cotton China, (iv) a 10% evaluation of the SGCtors. lhe Gau The results or elasticities. N. model is linear) a 3“ CiOgeuous van model Slimllation an emgeltous Var longrun mUItiplj. variable on an en 42 ‘bnsistent with economic theory and empirical observation. The validation statistics derived included: (i) the Root Mean Squared Percentage Error (RMSPE), (ii) the Mean Squared "irror (MSE), (iii) Theil’s U-statistic, and (iv) graphical validation. If the model was able 5 withstand these testing procedures and to predict actual market values accurately, then could be used for policy experiments and forecasting. A description of these tests and (re results for the fiber model are reported in full in section VII. 2.5 Model Simulation Forecast and policy simulations were undertaken in order to meet the study tjectives set out in the introductory section. Five simulations were performed for (i) a recast of price, production and consumption for the period 1990-2005, (ii) a 10% crease in cotton production in the USSR, (iii) a 10% increase in cotton production in tina, (iv) a 10% decline in domestic cotton price in the United States, and (v) an lluation of the Multi-Fiber Agreement (MFA) on the cotton and non-cellulosic fibers :tors. The Gauss-Seidel method was used to produce the simulation results. I The results of shocking the model could be expressed in terms of either multipliers elasticities. Multipliers can be calculated using matrix manipulation (providing the del is linear) and show the change in an endogenous variable for a one unit change in “exogenous variable (Labys, 1973). Three types of multipliers can be presented in the iel simulation results. First, impact multipliers show the effect of a one unit change in exogenous variable on an endogenous variable within the same time period. Second, fi—run multipliers show the effect of a sustained one unit change in an exogenous able on an endogenous variable given a time period long enough for full adjustment to take place Third, variable t-time peri variable. COW]t multipliers. Elasticities: change in an exoge elasticity can be c percentage change In this stut the basic tool of a the units in which 0f policy changes a and insensitive to ttsnMy In order t markets and to a wed in seed markets Was Chen The varit discussed which Validation, and f 43 take place. Third, cumulative multipliers show the effects on values of an endogenous variable, t-time periods (t=2,3,...,n), following a sustained one unit change in an exogenous variable. Consequently they show the adjustment between the impact and long-run multipliers. Elasticities show the percentage change in an endogenous variable for a one percent change in an exogenous variable or in other endogenous variables. Again, three types of elasticity can be calculated to show initial, cumulative and final effects of a sustained percentage change in an exogenous variable. In this study the results are reported chiefly in terms of elasticities and provided the basic tool of analysis. Multipliers are not employed because their values depend on the units in which the variables are measured making it difficult to tell whether the effects of policy changes are large or small. Elasticities, however, are ratios of percentage changes and insensitive to the units in which the data are measured. 2.3 Summag In order to understand and predict better the cotton and non-cellulosic fibers _ markets and to assess the impact of each of the market developments and policy issues It described in section I, an econometric model of the world cotton and non-cellulosic fibers _ markets was chosen as the most suitable research method. The various steps and stages involved in commodity market modeling were . discussed which include model structure, equation specification and estimation, model validation, and finally, model simulation. Then the different types of commodity market models were reviewed briefly. This review suggested that an econometrically estimated market model was Then the stages of t the world fiber mat MMnm market of the deta analyzing the ques market, and shoult world textile and t 44 market model was the most flexible and best suited for addressing the problem at hand. Then the stages of commodity model building were discussed with respect to the model of :he world fiber market developed in this study. As far as the author is aware, there are no econometric models of the world fiber narket of the detail and country coverage presented below. Therefore, in addition to analyzing the questions posed above, this study is unique in modeling the world fiber narket, and should be of interest to other research institutions involved in analyzing the vorld textile and fiber industry. In this sectii econometricalll'esti modeling cotton de differing results-«65 sensitive to model C an (i) the dive!” and apparel maul govemment inie” determination, W 3.1 W There are 2 different from the theoretical ration?! isdernanded by th other manufacture last that manufact of cotton and p01) insensitive to diffs to the prices of it Ingeneral, the fit 3. Cotton Demand In this section the method used to explain the demand for cotton in terms of anometrically-estimated equations is presented. Previous attempts have shown that )deling cotton demand is a formidable task. Empirical analysis has often led to widely fering results--especially in the estimation of income and price elasticities--and seems asitive to model construction and specification. Monke (1981) associates these difficulties :h (i) the development of manufactured fibers, (ii) cotton’s role as an input into textile d apparel manufacturing, (iii) variations in cotton quality, and (iv) widespread Iernment intervention in cotton production and trade (e.g., input subsidies, price :ermination, MFA). Features of Cotton Demand There are a number of features of cotton demand that make the estimation task sierent from modeling the demand for many other agricultural products. There is no bretical rationale for directly estimating consumer demand for raw cotton. Raw cotton emanded by the processors in response to final consumer demand for apparel items and isr manufactured textile products. This feature of cotton demand is complicated by the l ithat manufactured textile and apparel items are often mixtures of fibers (e.g., blends iptton and polyester), and within a fairly wide range of blends, consumers are relatively l psitive to different textile mixtures. Further, it appears that consumers are insensitive pe prices of individual fibers (e.g., cotton prices relative to polyester prices) because, aneral, the fiber value represents only a small proportion of the final purchase price. 45 therefore, 0015““ supported empirici (1971), Itigpen (19 Textile and: sensitive than cons plant technology e demand by manufa Another cor world trade irt app 1970s. As a resul- may differ signific An appropr lSmeasured by d. have previous stu Miles and Slnrmr 00HSinner consum 32 PM Several ear 0f aPliroaches ha 46 fherefore, consumer demand can be expected to be highly inelastic and this has been imported empirically in a number of recent studies [Dudley (1974), Magleby and Missaien E1971), Thigpen (1978)]. Textile and apparel manufacturers who purchase raw cotton tend to be much more isnsitive than consumers to the relative prices of individual fiber types. Most processing ant technology enables manufactures to substitute fibers quickly at some cost‘. Thus, smand by manufacturers tends to be much more price elastic than at the consumer level. Another complicating factor is that there is substantial world trade in raw cotton and )rld trade in apparel and other textile products has grown dramatically since the early 703. As a result, the quantities of cotton produced and consumed in any one country ry differ significantly. An appropriate measure of consumer demand is the quantity for home use, which measured by domestic mill consumption plus imported textile products less exports. ne previous studies have used domestic mill consumption as the demand variable (e.g., es and Simmons (1988) ). However, this variable does not accurately represent isumer consumption. ‘ Previous Studies Several earlier studies attempted to model the demand for cotton and a wide range :‘pproaches have been taken. Because the demand for cotton is derived from the is now easier to vary fiber content than 10 years ago, some difficulties remain. The cost of substituting fibers differs ' upon the age and design of the machinery and skill of plant managers. A whole assembly line has to be shut down necalibrated. The process can take a week if switching from 100% cotton to a blend. In addition, the resulting product '5in attributes and therefore cannot be marketed as the original product. demand for textile J rather than cotton in an early 5 was estimated as a prices. Dudley use income and lagged global demand for lnthese studies pe was tested, includ semi-log form per fibers falls as con: Many resea Monke (1981) arg Color, fab end-produ. placed ant Previous ti Presumabl ass“mptio: exl’fi‘esion Studies the (1987), Monke a; 1953 to 1975, Be and Behithan e4 47 _ demand for textile products and apparel, some studies have estimated total fiber demand . rather than cotton directly. In an early study by Donald et a1. (1963) fiber consumption for the United States was estimated as a function of real income, the change in real income and an index of fiber , prices. Dudley used a similar specification in which total fiber depended on current real income and lagged prices. Magleby and Missaien (1977) and Thigpen (1978) estimated the global demand for all fibers using time series data pooled over a large number of countries. In these studies per capita income was the only regressor and a variety of functional forms was tested, including the double-log, semi-log and log-inverse functions. Overall, the semi-log form performed best, which supports the hypothesis that income elasticity for fibers falls as consumption rises. Many researchers have used adaptive expectations models to specify their equations. Monke (1981) argues that: Color, fabric coarseness and fiber mix are important characteristics of textile end-products, and at each level of textile fabrication and distribution, orders are placed and/or received for the delivery of goods in a future period. The current demand for cotton thus depends on textile production decisions made in some t previous time period. These decisions, particularly with respect to fiber mix, are presumably influenced by expected prices of cotton and other inputs. The assumption of perfect forecast of income and population changes allows an expression of per capita demand for cotton based on expected prices for fibers. Studies that have used this approach include Adams and Behrman (1976), Ecevit l987), Monkeand Taylor (1983), and Mues and Simmons (1988). Using world data for 1958 to 1975, Ecevit estimated world cotton consumption as a function of lagged world )nsumption, lagged real prices of cotton and polyester staple and a time trend. Adams 1d Behrman estimated per capita cotton consumption equations for industrial and developing countrit form was used to r Consumption in int lagged price ratio specification was er was replaced by pr mpita income and regressors. Mues and S and the Restof-th than current price mills accept order. on cotton suppli e(111360115 includer ‘0 adust producri Another cl: the Work 0i Chou and allllarel prodr that individuals h; haction (lithe dil depreciated 01d S depends on prim current prices) lai 48 developing countries and centrally planned economies (CPEs). A double-log functional form was used to constrain elasticities to be constant over all price and income ranges. Consumption in industrial countries was specified to depend on lagged consumption, the lagged price ratio of cotton to polyester, per capita income and time. A similar specification was employed for the equations for the CPEs except that the income variable was replaced by per capita production. The equation for the CPEs employed real per capita income and a four-period distributed lag of the cotton to polyester price ratio as regressors. Mues and Simmons estimated mill consumption for the United States, West Europe and the Rest-of-the-World. The lagged ratio of cotton to polyester prices was used rather than current prices. The authors argued that lagged prices are more appropriate because mills accept orders for their goods up to 12 months in advance and hence need to secure raw cotton supplies by buying forward to ensure that they meet their contracts. The equations included a lagged dependent variable, "since mills are not expected to be able :0 adjust production levels instantaneously." Another class of model that has been used to estimate textile demand is based on he work of Chow (1960). These models explicitly recognize the durable nature of textile ind apparel products which depreciate over a number of time periods. The model assumes hat individuals have a desired level of stocks of these items and that purchases are some ”action of the difference between the desired stocks at the end of the time period and the epreciated old stock from the previous year. Assuming that the desired stock level also ‘iepends on prices and income, demand equations can be derived as functions of lagged and irrent prices, lagged and current income, and lagged consumption. Using this framework, equations for per C 1970) and Thigpen 33 Theoretical Is The two-eta modeling the demr and Muellbauer ( across broad grou follows. In the c Where X is a vecro function can be p; “M i= 1,... for broad 0011111104 halt the quanrir are alooated am suhutih‘ry fume] In other SUb'lllilit 49 :quations for per capita textile demand were estimated by Houthakker and Taylor (1966, 970) and Thigpen and Mitchell (1988). 3 jlheoretical Issues The two-stage budgeting approach has often provided the theoretical structure for odeling the demand for agricultural commodities. The approach is discussed by Deaton rd Muellbauer (1980) and is based on the assumption that preferences are separable ross broad groups of consumer goods. The model can be described algebraically as llows. Let the consumer’s utility function be given by, U = u(X) .ere X is a vector of all consumption goods. If preferences are separable then the utility action can be partitioned into a set of sub-utility functions, such that U = U[ V109). V209)» , V..(X..) 1 re X, i = 1,...,n are partitions of X. Each V,(X,) can be regarded as a utility function broad commodity groupings, such as food, clothing or housing, while the elements of are the quantities of individual goods consumed within the group. The elements of X allocated among the X, partitions such that the preference structure within any utility function can be determined independently of the quantities of goods consumed ther sub-utility functions. This is known as weak separability. Separability and Muellbauer (l SuhGroups Individual T. Commodities These asst in the first stage This is achiever SUhject to, Where: Pi = a p I = tot: This PTOblcm is Map l Inthese 50 Separability is often illustrated with a utility tree. An example is taken from Deaton and Muellbauer (p. 123). Total Utility ub—Groups Entertainment Shelter Food dividual T.V. Sport Housing Fuel Cereals Meat ommodities These assumptions about the utility function lead to the idea of two-stage budgeting. the first stage of this process consumers allocate expenditures to the commodity groups. his is achieved by maximizing the utility function, U = rV.(X.), V209). . V..(X..)] ject to, E P,.X,. = I :re: Pi = a price index for commodity group i, I total consumer income. problem is solved to determine I, , the proportion of I allocated to each commodity p i. In the second stage consumers maximize the sub—utility function for each group, subject to the armor subject to, vherepi = a vectv this problem gives Note that the ’dir groups are zero. indirectly on the changes affect the constraints in the is that it provides the consumer’s h consurnPtion of t While it is the {empires 0f I deriVed {mm a {V 51 i‘ubject to the amount of expenditure determined in the first stage. That is, Max v,(x,) [object to, “’ >3 M = I. Sphere pi = a vector of individual commodity prices corresponding to X,. The solution to is problem gives a demand function for each element of Xi, ‘ X. = xr(p. ,1. )- fate that, the ’direct’ cross-price effects of individual commodities in different commodity pups are zero. However, price changes in commodities of one group can impact directly on the demand for commodities in another group. This occurs because price anges affect the income allocations in the first stage, which, in turn, alter the budget astraints in the second. The main advantage of this approach in empirical estimation rhat it provides a justification for omitting the prices of all commodities considered in consumer’s budget and focusing only on those prices considered most relevant to sumption of the commodity being modeled. A Model of World Cotton Demand While it is not possible to apply the two-stage budgeting approach in its purest form, Features of cotton demand discussed in section 3.1 suggest that cotton demand is ed from a two-step process. Textile and apparel products seem a plausible commodity t, being comprised of individual commodities such as cotton, wool and manufactured From the discussion of manufacturers’ behavior, the consumption of these dual items depends on their relative prices. This is captured by the second stage of the twostage built expenditures whiCi total consumer inc To capture Demand in each s an identity. The f the consumption c use was estirnatec from these equati income. However 00min, WOOL cellr constant with res] “filled eather, th income may lead value can be asso 50 on, rather tha The price the lack of data comflililtint of th. and EEC COUnt 52 the two-stage budgeting process. Consumption of all fibers is constrained by total fiber expenditures which can be proxied by total fiber demand; which, in turn, is determined by total consumer income’. To capture these characteristics of textile demand, a two-step approach was taken. Demand in each specified world region was estimated using two behavioral equations and an identity. The first behavioral equation explains total fiber use, which is a measure of the consumption of all fiber types in the form of apparel and textile products. Total fiber use was estimated as a function of current real income. Income elasticities are derived from these equations which express changes in total fiber use in response to changes in income. However, this specification forces the income elasticities for individual fibers (i.e., cotton, wool, cellulosic fibers, and non-cellulosic fibers) to be the same since the shares are constant with respect to income changes. This restriction may not be important since, as argued earlier, the fiber content of most textile items is generally small. An increase in income may lead consumers to purchase higher valued items. However, typically the higher halue can be associated with higher quality products, brand named items, better styling and so on, rather than because of changes in the raw fiber composition. The price of textiles was not included in the total fiber use equations because of he lack of data on textile prices. Indexes of prices of textiles and apparel do exist as a omponent of the overall consumer prices index for some countries (e.g., the United States 1End EEC countries) but use of these was econometrically unsuccessful. The lack of trio-stage budgeting approach were to be strictly applied consistent with the theoretical model outlined in section 3.3, there icquation explaining the consumption of individual fibers as a function of the prices of the fiber and its competing fibers, isumcr income allocated to textile products. The second equation would explain the total consumer income allocated to ':ts as a function of a textile price index, indexes for other major consumption categories, and total consumer income. This 1 not be undertaken given the lack of data for the consuming countries in the model. The approach taken was considered 1 data availability, while still maintaining the general framework of the two-stage budgeting model. statistical signifiml of textile manufact was estimated to 2 it). Variations in l in the prices of fir The secor processors as the estimates the pro cellulosic fibers) tf estimated as a it eqvations the rat equations these I W approaches lobsthe same. l e‘lllatiorts in Whit T0 Comp equating the WW caPita total fiber each legion Was 53 statistical significance may result from the fact that the fiber component of the total cost of textile manufacturing is often small (e.g., the cost of cotton delivered to US mill in 1976 was estimated to account for 10.8% of the retail price of a pair of dungarees made from it). Variations in fiber prices therefore account for only a small proportion of the variation in the prices of finished textile products. The second behavioral equation attempts to capture the behavior of textile processors as they respond to changing relative prices of fiber types. This equation estimates the proportion of each fiber type (i.e., cotton, wool, cellulosic fibers, and non- cellulosic fibers) that makes up total fiber use. Thus the cotton share of total fiber use was estimated as a function of the prices of cotton and polyester staple. In most of the equations the ratio of these prices was used as the explanatory variable, while in a few equations these prices appeared as separate regressors. The major difference between these approaches is that in the ratio form, the elasticities of the two prices are constrained to be the same. While this specification leads to a reduction in flexibility, it was found that hquations in which prices appeared in their ratio form tended to perform better. To complete the demand system, cotton use was derived through an identity ;quating the product of the cotton share of total fiber use and total fiber use (given by per iapita total fiber use multiplied by population). To summarize, the demand for cotton in ach region was derived from: CTU =CI‘SH*PCTFU*POP, CTSH = f( CTP, PSP ), PCl‘FU = r( GDP). An identic: discussed in detai up total fiber use fibers. The non-r consumption of o 35 11mm 3.5.1 Cm The equat mum“ and irc IhebaSis of data Obtained from t Reasonably 00m] Pilitiisiled Slime) for industrialize endogelmlls Vari hostage least . 54 tere, U = Quantity of Cotton for home use, SH = Cotton share of total fibers for home use, )P = Gross Domestic Product, P = Price of cotton, P = Population, P = Price of polyester staple, and TFU = Per capita total fibers for home use. An identical structure was used to estimate the use of non-cellulosic fibers and is cussed in detail in section VI. That is, the proportion of non-cellulosic fibers making total fiber use was estimated as a function of the prices of cotton and non-cellulosic ers. The non-cellulosic fiber share was then multiplied by total fiber use to obtain the tsumption of non-cellulosic fibers. Empirical Results 3.5.1 Cotton Share Equations The equations were estimated using annual data from 1964 to 1986 for developing tries and from 1964 to 1987 for industrial countries. These periods were chosen on asis of data availability. The quantity data used in the demand side of the model was ined from the World Apparel Fiber Consumption Survey conducted by the FAO. onably comparable data from this surVey were available from 1964 to the most recently hed survey results (FAO, 1989 issue) up to 1986 for developing countries and 1987 dustrialized countries. Given that the cotton share equations contain current enous variables on the right-hand side, the model is simultaneous. Consequently, the tage least squares estimator (2SI_.S) was used to avoid the inconsistency of OLS. The price V: of non-cellulosic 1 Domestic prices it these world prices and consumer prit Based on equation for each and the deflated p either one or hot amount of the V2 United States and and signed corret other regions ma: depite the deflal \ JWCNMO 55 The price variables used in the cotton share were the prices of cotton3 and the price ton-cellulosic fibers‘. These prices were denominated in US cents per kilogram. nestic prices for each region were obtained by applying the US dollar exchange rate to e world prices and deflating by a domestic consumer price index. The exchange rates consumer price indexes were obtained from the IMF‘ 6. Based on the framework described above, the cotton share of total fiber use ation for each region was estimated initially as a function of the deflated price of cotton the deflated price of polyester. This specification was not successful in most cases with .er one or both of these price variables not significant or else did not explain a large mm of the variability of the left-hand side variable. Only in the equations for the ted States and Brazil were the price variables included as separate regressors significant signed correctly (equations 3.3 and 3.14). The poor results of this specification in :r regions may have been due to multicollinearity between cotton and polyester prices, site the deflating of both price series. In light of this possibility, equations explaining t Source: Cotton Outlook A Index, Middling 1-3/32 Inch, CIF Europe. he price of non-cellulosic fibers was represented by the polyester staple price, fob, US plants. SurceleF International Financial Statistics. [his method of deriving domestic cotton and polyester prices was somewhat crude since it ignored transportation costs, Ices in cotton qualities, the impact of trade restrictions on domestic prices in many countries, problems associated with g cotton prices and issues over how to obtain exchange rates and deflators for aggregate regions (e.g., the EEC and Africa). However, an important requirement of the model was to solve for the Outlook 'A' Index price and therefore .ecessary to link domestic prices in cotton producing and consuming regions to this price in some way. Given limited :s and the lack of reliable and consistent domestic cotton price series for individual countries, it was infeasible to take sophisticated approach. However, for those countries where price data were available (e.g., China, India and the United domestic prices were used in the share equations. This meant that equations linking the domestic price and the 'A' Index price had to be estimated (chapter 5). In the equations with the prices of cotton and polyester appearing o, the exchange rate and deflation conversions became irrelevant, since identical exchange rates and deflators are in both the numerator and denominator of the ratio. the cotton share of of cotton and poly! the equations of I Japan, the Repubi in the mod income elasticities equations. This v and was pursued consumers to fibe regression period While the itttotal fiber use ‘ SPelifittttions wet that estimating tl abetter lit than i We and consun many of the eq] 56 e cotton share of total fiber use in other regions were estimated as a function of the ratio cotton and polyester prices. This specification performed much better and was used in 3 equations of Argentina, Australia, Central Africa7, China, the EEC-12, Egypt, India, )an, the Republic of Korea, Mexico and Pakistan. In the modeling process other variables were tried. For example, to test whether ome elasticities differed across fiber types, an income variable was included in the share rations. This variable was not significant across all regions of the model, as expected, l was pursued no further. A possible explanation for this finding is that attention by sumers to fiber content is a fairly recent phenomenon which was not captured in the ‘ession period. While the price variables were significant generally, the variability of cotton share )tal fiber use was not explained entirely by price relationships. In many equations, the ifications were modified to improve the “goodness-of-fit'“. For example, it was found estimating the cotton share equations with variables converted into logarithms gave tter fit than in the linear form. In these cases the elasticities were unchanged over all a and consumption levels and were given by the coefficients of the price variables. In t of the equations a zero-one dummy variable was added to the specification to .‘° Central Africa region includes the following countries: Angola, Benin, Burkina Faso, Cameroon, Central African pic, Chad, Ethiopia, Ghana, Ivory Coast, Kenya, Malawi, Mali, Mozambique, Niger, Senegal, Somalia, Sudan, Tanzania, plganda, Zaire, Zimbabwe. is important to define what is meant by goodness-of-fit. Typically, this refers to obtaining high adjusted R—squared .ow standard errors on parameter coefficients, and small sum of squared residuals. However, one can often achieve it if variables are non-stationary and/or if strong trends are present in the data series. In the case of the fiber model, [15 were estimated primarily to obtain elasticities which capture producer and consumer responses to changes in 5. In terms of a market model, elasticities of demand and supply are particularly important for solution of a set of Sieous equations. Therefore, concern over obtaining accurate elasticities led to the inclusion of variables capturing Dr dummy variables (which merely discards observations) in order obtain satisfactory elasticities. No tests for ity and specification error were undertaken. Future work on expanding the model will include these tests. capture abnonnall variable. This var noneellulosic fibe: the time variable ' cellulosic fibers th For example, mar strength and unifc Further, it has be 19305 and 19603 i A potenti Identification ref the reduced form “Mary and sut order and rank engennus variab bis Condition is rank CohditiOn OOefliciems ofth of “lithe Other e One. FOF large idiom Tit finders e‘iltttin 1 l 57 :pture abnormal behavior in the data series, while many of the equations contained a time .riable. This variable was added because much of the competition between cotton and un-oellulosic fibers is not directly price-related. In all the equations in which it was used, 3 time variable was negatively signed. This captured the substitution of cotton by non- dulosic fibers that has taken place following advances in manufactured fiber technology. r example, many important fiber characteristics demanded by manufacturers, such as ength and uniformity of length, are much easier to control by using manufactured fibers. rther, it has been argued (e.g., Monke) that much of the substitution for cotton in the 30s and 19603 depended on manufactured fiber use rather than price. A potential problem in estimation is whether the equations are identified. ntification refers to the problem of whether there are enough restrictions imposed on reduced form parameters to obtain unique values of the structural parameters. A :ssary and sufficient condition for an equation to be identified is that it meets both the r and rank conditions. The order condition is met if the number of excluded enous variables is greater than the number of included endogenous variables less one. condition is a necessary, but not sufficient condition. A sufficient condition is that the condition is met. This occurs when the rank of the matrix containing (i) the icients of the endogenous variables omitted from the equation, and (ii) the coefficients the other equations in the model, is less that the number of endogenous variables less For large econometric models, tests of equation identification are not usually med. This is because when there are a great many predetermined variables, the ’s equations are likely to be over-identified, especially when a lot of structure has been inposed thro usually find equatit although future res carehrlly. lhe estimat selected diagnostic Durbin Watson st: in the model sim Appendix The cotton ! INCISHARG = .02 Within) (corn; MWHARG = DFCH’ARG = DH’SPARG = No uv The cotton Shart “l0 of cotton in 0f~/+ 0.07. A high levels of do Very high levels lhe efillttlion, 58 ten imposed through many parameter restrictions. As a result, tests of identification ually find equations to be over-identified. In this study identification was not tested, l hough future research to expand the model will consider the identification problem more refully. The estimated cotton share equations are presented in equations 3.1-3.14, along with ected diagnostics (i.e., t-values, corrected R-squared, standard error of the estimate, rbin Watson statistic or H—statistic). Reported are the final equation specifications used the model simulation. The variable definitions and specifications are given in the pendix. The cotton share of total fiber use in Argentina is given in equation 3.1. YISHARG = - 0.22 - 0.066 LN DFCI'PARG/DFPSPARG + 0.074 D76 + 0.606 LN CI’SHARG(-1) (3.1) (-2.65) (1.85) (5.48) ‘UARED (coax): 0.89 SEE: 0.0208 H-SI‘AT: 0.64 PERIOD or m: 1964-1986 c:CI‘SHARG == Cotton share of total fiber use, Argentina, DFCI‘PARG = Deflated cotton price, Argentina, DFPSPARG = Deflated polyester price, Argentina, D76 = Zero-one variable, equal 1 in 1976, 0 otherwise, LN = Indicates variable transformed into logarithms. cotton share equation was specified with variables converted into logarithms. The of cotton and polyester prices was significant and gave a short-run elasticity estimate + 0.07. A zero-one variable for 1976 was found to be important and captured the levels of domestic fiber use following a sudden run-down of stocks which had been at nigh levels during the early 1970s. A lagged dependent variable was also included in ruation. The cotton : cmnus = 0544 - o. (a vacuum (CORR) thrtCl'SHAUS =t ovcraaus = onsraus = “ME = The cotton share and a time varial significant when elasticities were . Which was small llPluring the no throughout moo. size of the l-val tStimated price The cottor WARE!) (COR Wham“ DPCIPBRA ”Water 083 IN The equaiiim e with variables a and Price of Pl Elasficill estim 59 The cotton share of total fiber use in Australia is given in equation 3.2. HAUS = 0544 — 0.028 DFCI'PAUS/DFPSPAUS - 0.006 TIME (32) (.242) (-10.06) )UARED (CORR): 0.92 SEE 0.0048 DW: 2.32 PERIOD OF FIT: 1964-1987 tre:CI'SHAUS = Cotton share of total fiber use, Australia, DFCI‘PAUS = Deflated cotton price, Australia, DFPSPAUS = Deflated polyester price, Australia, TIME = Time variable. 3 cotton share for Australia was a function of the ratio of cotton and polyester prices I a time variable. The price ratio was included after f'mding the cotton price not to be tificant when entered in the equation as a separate regressor. The own- and cross-price sticities were estimated to be -/ + 0.06 (when measured at the mean prices and quantity) ich was smaller than expected. A time variable was found to be highly significant, tturing the non-price related substitution Of cotton with manufactured fibers in Australia aughout most of the estimation period. The strength Of this trend is indicated by the t of the t-value (-10.06) and corrected R-squared (0.92) which might explain why the mated price elasticities were small. The cotton share Of total fiber use in Brazil is given in equation 3.3. TSHBRA = - 0.389 - 0.181 LN DFCI‘PBRA + 0.170 LN DFPSPBRA + 0.056 D83 (33) (-6-15) (10.84) (1.92) UARED (CORR): 0.91 SEE: 0.0118 DW: 1.79 PERIOD OF FIT: 1964-1986 :ICI'SHBRA = Cotton share of total fiber use, Brazil, DFCI'PBRA = Deflated cotton price, Brazil, DFPSPBRA - Deflated polyester price, Brazil, D83 Zero-one variable, equal 1 in 1983, 0 otherwise, LN Indicates variable transformed into logarithms. equation explaining the cotton share of total fiber use in Brazil fitted the data better variables converted into logarithms. The share was explained by the price of cotton nice of polyester which entered the equation as separate regressors, and resulted in city estimates Of 0.18 and 0.17, respectively. A dummy variable for 1983 accounted lot the effect of ti cotton at the expt lhe cotton momma = on RSQUARH) (CORR thmClSHCAF = Dl-Cll’GtF = DH’SPCAF = TIME = LN = lhe logarithm ol polyester prices Australia (equat demand (./+ 0' strength of the I Will in consur Aida in the 19 The cottot mllilil = 108-l (~E WARE) (COR DFCl‘pCHl l) D74 he cotton shat tad Wheeler \ v 5“ f0Otitote 7 ”lilting; TIM 60 )r the effect of the debt crisis in Brazil. The recession caused manufacturers to switch to mm at the expense of other fibers, although, overall, cotton demand declined sharply. The cotton share of total fiber use in Central Africa9 is given in equation 3.4. lCI‘SI-ICAF = - 0.143 - 0.042 LN DFCI'PCAF/DFPSPCAF - 0.08 TIME (3.4) (-3~91) (-10.8) SQUARED (CORR): 0.87 SEE 0.0058 DW: 1.42 PERIOD OF FIT: 1964-1986 tere:CI‘SI-ICAF = Cotton share of total fiber use, Central Africa, DFCI'PQAF = Deflated cotton price, Central Africa, DFPSPCAF = Deflated polyester price, Central Africa, TIME = Tune variable, LN = Indicates variable transformed into logarithms. e logarithm Of the cotton share is explained by the logarithm of the ratio of cotton and lyester prices and a time variable. The results were similar to those reported for stralia (equation 3.2) with small estimates of the own- and cross-price elasticities of nand (-/+ 0.04) and a highly significant, negatively-signed time trend variable. The :ngth of the time trend variable in explaining the share was the result of the sharp up rig in consumption of cheaper textile and apparel items which occurred throughout ica in the 19703 and 19808 period. The cotton share of total fiber use in China10 is given in equation 3.5. {CHI = 1.08 - 0.001 DFCI‘PCHI/DFPSPCHI + 0.11 D74 (35) (-8-31) (3.08) UARED (CORR): 0.83 SE: 0.0070 DW: 1.89 PERIOD OF FIT: 1964—1986 :zcrsrrcrn = Cotton share of total fiber use, China, DFCI'PCHI = Deflated cotton price, China, DFPSPCHI = Deflated polyester price, China, D74 = Zero-one variable, equal 1 in 1974, 0 otherwise. cotton share Of total fiber use in China was estimated to depend on the ratio of cotton P01yester prices and a zero-one variable for 1974. The price ratio was highly \ :e footnote 7 for region definition. icludes Taiwan Province. significant with a estimate of -/+ 0 coefficient estima demand shift tow for by price mow The cotton maniac = .10 account) (corn: WbcmCISl-IEEC = DFCI'PEEC : DFPSPEEC = 083 IN The cotton shat. of cotton and p data better whe and cross-price The cotto WARE” (C011 Tile Comm shfl found ‘0 fit tl declined c0nd 61 Eicant with a t-value of -8.31 and gave an own- and cross-price elasticity Of demand rate Of -/ + 0.06. The initial specification with the prices as separate regressors gave icient estimates which were not significant. The dummy variable captured the sudden nd shift towards cotton following the Oil price shock which was not fully accounted 7 price movements. The cotton share of total fiber use in the EEC-12 is given in equation 3.6. THEEC = - 1.009 - 0.140 LN DFCI'PEEC/DFPSPEEC + 0.121 D83 (3.6) (.595) (2.44) \RED (CORR): 0.70 SEE: 0.0472 DW: 1.34 PERIOD OF FIT: 1964-1987 .TSHEEC = Cotton share of total fiber use, EEC-12, )FCTPEEC = Deflated cotton price, EEC-12, )FPSPEEC = Deflated polyester price, EEC-12, >83 = Zero-one variable, equal 1 in 1983, 0 otherwise, .N = Indicates variable transformed into logarithms. >tton share Of total fiber use in the EEC-12 was estimated as a function of the ratio :on and polyester prices and a dummy variable for 1983. The equation fitted the otter when the variables were converted into logarithms. The estimate of the own- ass-price elasticities were -/ + 0.14. he cotton share Of total fiber use in Egypt is given in equation 3.7. TEGY = 1.259 - 0.168 LN DFCI'PEGY/DFPSPEGY - 0.065 TIME (37) (-2.08) (435) {ED (CORR): 0.84 SE: 0.0536 DW: 1.33 PERIOD OF FIT: 1964-1986 SHBGY = Cotton share of total fiber use, Egypt, ’CI'PEGY = Deflated cotton price, Egypt, T'I’SPEGY Deflated polyester price, Egypt, ME = Time variable, = Indicates variable transformed into logarithms. ton share equation for Egypt estimated with variables converted into logarithms was 3 fit the data better than when estimated in the linear form. The cotton share l consistently throughout the estimation period which was captured by a time variable and four cotton and polyet The cotton lNCl'SllIND = 0.039 The cotton share polyester price I found to fit the estimated at ./ responsive to pr coTlsumption. "l The cottor WN=0246+ WARP-D (COR “Mascara . DTCDJPN = DTPSPJPN : on The 001101] shay total fiber use e I (111311011 becat hidden iItcreas 62 able and found to be highly significant. Also statistically significant was the ratio of on and polyester prices with price elasticity estimates Of -/ + 0.17. The cotton share of total fiber use in India is given in equation 3.8. ISHIND = -0.039 - 0.016 LN DFCI’PIND/DFPSPIND + 0.829 LN CI‘SHIND(-1) (3.8) (4.98) (13-12) 'UARED (CORR): 0.95 SEE: 0.0042 H—SI‘AT: 0.73 PERIOD OF FIT: 1964-1986 e:CI‘SHIND = Cotton share of total fiber use, India, DFCI'PIND = Deflated cotton price, India, DFPSPIND = Deflated polyester price, India, LN = Indicates variable transformed into logarithms. cotton share Of total fiber use in India was estimated as a function of the cotton and 'ester price ratio and a lagged dependent variable. A double-log functional form was rd to fit the data better than the linear form. The price elasticity Of demand was mated at —/ + 0.016, indicating that the consumption of cotton in India was not onsive to price changes. The lagged cotton consumption captured the strong trend in :umption. This result casts doubt on the accuracy of the price elasticity estimate. The cotton share Of total fiber use in Japan is given in equation 3.9. JPN = 0.246 + 0.466 CI‘SHJPN(-1) + 0.045 D74 - 0.017 DFCI'PJPN/DFPSPJPN (3.9) (221) (231) (4.47) JARED (CORR): 0.61 SEE: 0.0045 DW: 1.88 PERIOD OF FIT: 1964-1987 :CI‘SHJ'PN = Cotton share of total fiber use, Japan, DFCI'PJPN = Deflated cotton price, Japan, DFPSPJPN = Deflated polyester price, Japan, D74 = Zero-one variable, equal 1 in 1974, 0 otherwise. otton share Of total fiber consumption in Japan was estimated to depend on the ratio :ton and polyester prices, a zero-one dummy variable for 1974, and cotton’s share of fiber use lagged one year. The price variable was not significant but kept in the ion because it had the correct sign. The dummy variable for 1974 captured the 0 increase in cotton imports in that year not explained by price movements. The cotton uvcrsrlKOR = .041 rsouartw (CORR Wherefl‘SHKOR = DPCII’KOR = DFPSPKOR = 073 TIME IN The variables be Korea were tht dummy variable logarithmic funt elasticities of do are more respo Variable 08pm“ Period. Howev sudden demam ToCounted fort The cotto CIWET = 0.091 WARE” (000 “More: BMW DFPSPMEi The Cotton the ‘36de With th. exploited by 0 63 The cotton share of total fiber use in Korea is given in equation 3.10. HKOR = - 0.401 - 0.335 LN DFCI'PKOR/DFPSPKOR - 0.192 LN TIME + 0.38 D73 (3.10) (437) (.2154) (3.63) mm) (com): 0.89 SEE: 0.1318 DW: 2.22 PERIOD OF m: 1964-1986 II‘SHKOR = Cotton share of total fiber use, Korea, )FCI'PKOR = Deflated cotton price, Korea, )FPSPKOR = Deflated polyester price, Korea )73 = Zero-one variable, equal 1 in 1973, 0 otherwise, 1" ['IME = Time variable, .N = Indicates variable transformed into logarithms. ' ariables best able to explain the variability in cotton’s share of total fiber use in were the ratio of cotton and polyester prices, a time variable, and a zero-one y variable for 1973. As with many of the equations reported above, a double- :hmic functional form provided the best fit to the data. The estimated price ities of demand were -/ + 0.335 which indicated that Korean cotton manufacturers are responsive to price movements than in most other regions studied. The time .e captured the non-price related trend away from cotton during the regression However, a zero—one dummy variable for 1973 was needed to account for the . demand shift towards cotton following the oil price shock which is not fully ted for by price movements. he cotton share of total fiber use in Mexico is given in equation 3.11. = 0.091 - 0.043 DFCI‘PMEX/DFPSPMEX + 0.857 CI‘SHMEX(—1) (3.11) (4-77) (37.69) LED (CORR): 0.99 SEE: 0.0024 H-SI‘AT: 1.71 PERIOD or FIT: 1964-1986 SHMEX = Cotton share of total fiber use, Mexico, CI'PMEX = Deflated cotton price, Mexico, PSPMEX = Deflated polyester price, Mexico. ton share of total fiber use in Mexico declined steadily throughout the regression Ilth the result that more than 90% of the variation of the cotton share could be :1 by the dependent variable lagged one year. The ratio of cotton and polyester prices W88 fOUDl elasticity estimat The cotton ClSlll’AK = 0.1572- (_ RSQUARED (COR!l thrcCISHPAK = DPCIl’PAK = DFPSPPAK = lhe downward lagged depende addition to this Specification. ll because it has I The cotto WWW DFCIYIUR D73 The com sha “Mom in Tu year‘ The Goo: mean prim am towards cottor was captured ' 64 s was found to be significant however, and gave own- and cross-price short-run :ity estimates of -/+ 0.09 when calculated at the mean price and quantity. The cotton share of total fiber use in Pakistan is given in equation 3.12. AK = 0.1572 - 0.030 DFCI'PPAK/DFPSPPAK + 0.34 CI‘SHPAK(-1) (3.12) (.172) (10.72) ARED (CORR): 0.91 SE 0.0142 HSI‘AT:- 1.86 PERIOD OF FIT: 1964-1986 HSHPAK = Cotton share of total fiber use, Pakistan, DFCI'PPAK = Deflated cotton price, Pakistan, DFPSPPAK = Deflated polyester price, Pakistan. Iownward trend in cotton’s share of total fiber use in Pakistan was captured by a 1 dependent variable which appeared in the equation with a coefficient of 0.84. In on to this variable, the ratio of cotton and polyester prices was included in the final ication. While this variable was not statistically significant, it was kept in the equation se it has the correct sign. ['he cotton share of total fiber use in Turkey is given in equation 3.13. JR = 0.237 - 0.000007 DFchrUR + 0.090 D73 + 0.750 CI‘SH'I‘UR(-l) (3.13) (-2.46) (2.93) (8.07) RED (CORR): 0.92 SEE 0.0111 HsrAT: - 0.68 PERIOD OF m: 1964-1986 I'SHTUR = Cotton share of total fiber use, Turkey, FCI'P'I‘UR = Deflated cotton price, Turkey, 73 = Zero-one variable, equal 1 in 1973 0 otherwise. tton share of total fiber use in Turkey was estimated to depend on the deflated price )n in Turkey, a zero-one dummy variable for 1973, and the cotton share lagged one 'he coefficient on the cotton price gave an elasticity of -0.13 when estimated at the rice and quantity. The zero-one dummy variable captured the sudden demand shift cotton following the oil price shock which was not fully accounted for by price :nts. The trend away from cotton consumption throughout the regression period tured by the lagged dependent variable. l The ootton : GSHUSA = 0.488 ' 0.l (4.1 vsouaam (CORR WthISHUSA = DFCU’USA ’- Dl'YSl’USA = TIME = For the United 1 variability of dot the fit of the eqr and cross-price tabulated at th the United St: counterparts in oottou in the l elipluined by or The fit 0 00fretted R.“ Tests for not autoWtrelativ motoring a 1 Variable). 1y shale 0f con 19303) which ‘° techno, 65 The cotton share of total fiber use in the United States is given in equation 3.14. ISA = 0.488 - 0.0007 DFCI‘PUSA + 0.0004 DFPSPUSA - 0.004 TIME (3.14) (.491) (4.42) (-2.86) ARED (CORR): 0.89 SEE: 0.0135 DW: 1.40 PERIOD OF FIT: 1964-1987 ZI‘SHUSA = Cotton share of total fiber use, United States, DFCI'PUSA = Deflated cotton price, United States, DFPSPUSA = Deflated polyester price, United States TIME = Time variable. he United States the cotton share of total fiber use was specified to depend on the >ility Of deflated cotton and polyester prices and a trend variable. It was found that t of the equation improved with the prices entering as separate regressors. The own- :ross-price elasticities were estimated to be -0.30 and +0.22, respectively, when rated at the mean values of price and quantity. This indicated that manufacturers in lnited States have been more responsive to fiber price movements than their :rparts in other model regions. The time variable captured the trend away from in the United States during the regression period (1964-1987) which could not be 1ed by economic factors. The fit of the share equations was satisfactory overall. Out of 14 regions, six had a ted R-squared value of over 90% and for only two equations was it below 80%. for auto-correlation (by the DW or H-statistic) failed to find the presence of Irrelation in most of the equations. A major concern was the inclusion of variables ng a trend in the consumption data (e.g., time variable or lagged dependent 3). These variables were included because there has been a steady decline in the If cotton in total fiber consumption during the regression period (1960s to mid- which was not price-related. As mentioned above, this trend was due, in most part, nological improvements in the manufactured fibers industry. Also in many equations. dumrt observations from was to estimate : s' ' cant or in lheir inclusion I become projectiv in the forecasts : (e.g., growth rat- estimated price out-weighed by in the consump 3.52 Per C321 The estin to 3.28 below. deflated gross llllel'lns of 10 incOmes and . Pl0duets falls‘ 66 tions, dummy variables were used, perhaps too frequently, to remove outlying rvations from the data set. However, what was important for the simulation model :0 estimate reliable price elasticities. In most cases, these were unreliable (i.e., not icant or incorrectly signed) with the omission of the trend and dummy variables. ' inclusion raises the question of how will the long-run forecasts be. Will they mainly ne projections of past trends? However, many other factors play an important role : forecasts also, such as the assumptions made about the model’s exogenous variables growth rates of income, general prices, exchange rates and pOpulation), as well as the ated price and income elasticities. It is unlikely that the impact of these factors are sighed by the trend variables; however, this is an issue which is considered carefully consumption forecasts which are reported in section VIII. Per Capita Total Fiber Use Equations 7he estimated equations for per capita total fiber use are presented in equations 3.15 l below. Per capita total fiber use was estimated as a function of current per capita d gross domestic product. In most cases, the income variable entered each equation as of logarithms. This was consistent with empirical evidence that showed that as s and consumption rise, the inCome elasticity of demand for textile and apparel ts f n. In many of the equations a trend variable was included to provide more fl; QT = Demand for textiles, = Per capita income, l = Variable transformed into logarithms, , = Elasticity of textile demand with respect to income (607/61).(I/QT . = Intercept and sIOpe coefficients, respectively, in regression equation. mi-log functional form, the demand equation is represented by, QT = a + b. LN 1. Therefore ET, = b / QT. Thus 5, E77 falls. reliable estimates of thisvariable in using caution wl determine the or The per ca retro/110 = 834 + vsouaan) (cont WARG PDGDPARG 081 D82 LN The equation e capita deflated The income ve tdcvtlated at tl fesuh is explai of which was r The per Willis = . & RSQUARED (C0 WMPCIFUAm PDGDPAL 1383 LN lhe variabiliv 67 liable estimates of the elasticity of total fiber use with respect to income. The inclusion this variable improved the statistical fit of the equations considerably, although it implied ing caution when making long-run forecasts to see that the trend variables do not termine the outcome. The per capita total fiber use in Argentina is given in equation 3.15. WARG = 6.84 + 3.96 LN PDGDPARG - 133 D81 - 156 D82 (3.15) (3.26) (.275) (.323) QUARED (CORR): 0.72 SEE 1.88 DW: 2.42 PERIOD OF FIT: 1964-1986 :re:PCI'FUARG = Per capita total fiber use, Argentina, PDGDPARG = Per capita deflated gross domestic product, Argentina, D81 = Zero-one variable, equals 1 in 1981, 0 otherwise, D82 = Zero—one variable, equals 1 in 1982, 0 otherwise, LN = Indicates variable transformed into logarithms. 2 equation explaining per capita total fiber use in Argentina included the log of per ita deflated gross domestic product and zero-one dummy variables for 1981 and 1982. income variable was highly significant with an estimated income elasticity of 0.58 ulated at the mean total fiber use level. The dummy variables were significant. This It is explained in terms of capturing the recession caused by the debt crisis, the effect 'hich was not fully captured by the income variable. The per capita total fiber use in Australia is given in equation 3.16. UAUS = - 88.14 + 11.84 PDGDPAUS - 1.95 D33 (3.16) (10.32) (-2.08) JARED (CORR): 0.88 SEE: 1655 DW: 1.88 PERIOD OF FIT: 1964-1987 :PCI‘FUAUS = Per capita total fiber use, Australia, PDGDPAUS = Per capita deflated gross domestic product, Australia, 1383 = Zero-one variable, equals 1 in 1983, 0 otherwise, IN = Indicates variable transformed into logarithms. Iariability of per capita total fiber use in Australia was estimated to depend on per mpita deflated iI lududed t0 capo for by the moon The per ca PCllUBRA = 4235 vsouaaeo (corn Wherel’CIi-‘UBRA PDGDPBRA 085 086 LN lhe equation 1 However, this 1 1986 and no e The income va income elasticl The per Ll P000011 = “QUAKE (co WWCI‘FUQ]: Poonrci, 1N lhe Chosen 5 lllClllded p8r ‘ found that fey limit [1‘3“st ll 8“ foolhon 68 capita deflated income and a zero-one dummy variable for 1983. The dummy variable was included to capture the sharp decline in textile imports in Australia in 1983 not accounted for by the income variable. The per capita total fiber use in Brazil is given in equation 3.17. PCI'FUBRA = -1235 + 155 LN PDGDPBRA - 2.46 D85 - 2.86 D86 (3.17) (8.81) (-6.96) (.719) R-SQUARED (CORR): 0.95 SEE: 1.88 DW: 1.29 PERIOD OF FIT: 1964-1986 WherezPCI'FUBRA = Per capita total fiber use, Brazil, PDGDPBRA = Per capita deflated gross domestic product, Brazil, D85 = Zero-one variable, equals 1 in 1985, 0 otherwise, D86 = Zero-one variable, equals 1 in 1986, 0 otherwise, LN = Indicates variable transformed into logarithms. The equation for Brazil fitted the data very well with an adjusted R-squared of 0.95. However, this result was heavily affected by the zero-one dummy variables for 1985 and 1986 and no explanation was found why consumption declined suddenly in these years. The income variable was highly significant (t-value = 8.81) providing an estimate of the ncome elasticity Of total fiber use equal to 0.32. The per capita total fiber use in Central Africa12 is given in equation 3.18. N PCI‘FUCAF = - 351 + 054 LN PDGDPCAF + 0.90 LN PCI'FUCAP(-1) (3.18) (2.56) (6.98) rSQUARED (CORR): 0.92 SEE: 0.06 H—SI‘AT: 0.46 PERIOD OF FIT: 1964—1986 'here:PCI'FUCAF = Per capita total fiber use, Central Africa, PDGDPCAF = Per capita deflated gross domestic product, Central Africa, IN = Indicates variable transformed into logarithms. be chosen specification Of the equation explaining total fiber use in Central Africa cluded per capita deflated income and per capita total fiber use lagged one year. It was und that regressing the logarithms of variables provided a better fit to the data than their rear transformation. The income variable was significant with a coefficient (and elasticity "See footnote 7 for region definition. estimate) of 0.51 period meant th: in the equation l with a coefficier The per 0: lCllUGll=dUl+ ( vsouartn) (con W011 - 10000011 D81 nv Total fiber use domestic prodr be treated wit income trendet We elastic appears to be non-income re The per Phonic.“ WW3 (cc where PDGDPEE 69 estimate) of 0.54. The steady increase in fiber use in Africa throughout the regression period meant that the data contained a very strong upward trend. This was accounted for in the equation by the inclusion of a lagged dependent variable which entered the equation with a coefficient of 0.9 and t-value of 6.98. The per capita total fiber use in China is given in equation 3.19. Pcrrucrn = 6.07 + 2.88 LN PDGDPCHI + 058 D81 (3.19) (18.28) (2.36) R-SQUARED (CORR): 0.95 SEE: 0.24 DW: 1.73 PERIOD OF FIT: 1964-1986 WherezPCI'FUCI-II = Per capita total fiber use, China, PDGDPCHI = Per capita deflated gross domestic product, China, D81 = Zero-one variable, equals 1 in 1981, 0 otherwise, LN = Indicates variable transformed into logarithms. Total fiber use in China was explained by the logarithm of per capita deflated gross domestic product and a zero-one dummy variable for 1981. The regression results should )e treated with caution because both per capita consumption and per capita deflated ncome trended upwards consistently during the regression period with the result that the ncome elasticity was estimated at 0.91 with coefficient t-value of 18.28. This elasticity ppears to be unduly large and could have resulted from the income variable capturing a on-income related trends in total fiber use. The per capita total fiber use in the EEC-12 is given in equation 3.20. mm = - 116.9 + 15.0 LN PDGDPEEc - 0.14 UNEMPEEC - 0.76 DMFAI - 058 DMFAII - 1.77 DMFAIII (320) (7.84) (.171) (-2.39) (-1.71) (.490) SQUARED (CORR): 0.93 SEE: 0.44 DW: 1.62 PERIOD OF FIT: 1964-1987 terezPCI'FUEEC = Per capita total fiber use, EEC-12, PDGDPEEC = Per mpita deflated gross domestic product, EEC-12, UNEMPBEC = Unemployment rate (%), EEC-12, DMFAI = Zero-one variable for MFAI period, equal 1 1974 to 1977 else 0, DMFAII = Zero-one variable for MFAII period, equal 1 1978 to 1981 else 0, DMFAIII = Zero-one variable for MFAIII period, equal 1 1982 to 1985 else 0, LN = Indicates variable transformed into logarithms. In the equation unemployment v of this variable r In periods of g especially the t recessions. Altl signed oorrectly one dummy va operated duriny Which resulted 12 from develo hair initial sp (ie, less than The per PCTFUEGv . .14 IiSQUARED (C0 whtml’Cl'Fumt PDGDPEC D81 LN The per chpit hpita gmss Variable Was Mean fiber u increase in y 70 the equation for the EEC-12, in addition to the per capita income variable, an remployment variable was added to the specification. The hypothesis behind the inclusion this variable was that textile products tend to be durable and purchases are infrequent. periods of general recession the sales of textile products fall significantly. This is pecially the case for lower income households which tend to be most affected by sessions. Although not significant, the variable was kept in the equation because it was :ned correctly. Also included in this per capita total fiber use equation were three zero- e dummy variables, accounting for the three Multi-Fiber Agreements (MFA) which erated during the estimation period. The variables captured the decline in consumption ich resulted from restrictions placed on the importation of textile products into the EEC- from developing countries. An index of prices of textile and apparel products was tried an initial specification of this equation. However, the variable had a very low t-value :., less than unity) and was dropped from the specification. The per capita total fiber use in Egypt is given in equation 3.21. F'UEGY = -14.87 + 355 LN PDGDPEGY + 1.22 D81 (3.21) (3.93) tUARED (CORR): 0.91 SEE- 1.66 DW: 1.84 PERIOD or m: 1964-1986 ezPCI'FUEGY = Per capita total fiber use, Egypt, PDGDPEGY = Per capita deflated gross domestic product, Egypt, D81 = Zero-one variable, equals 1 in 1981, 0 otherwise, LN = Indicates variable transformed into logarithms. per capita use of fibers in Egypt was estimated to depend on the logarithm of per :a gross domestic product and a zero-one dummy variable for 1981. The income .ble was highly significant, and gave an elasticity estimate Of 0.68 when calculated at 1 fiber use. The dummy variable for 1981 in the equation accounted for the sudden ease in yarn output and weak demand in its export markets. lhe variability capita gross dor equation perior lhe per 1 PCll‘Ull’N = - 110. RSOUARED (001 “automotive PDGDPJPN on no Per capita tot WM and a Variable gave _ lhe dummy ( increase ill in 71 The per capita total fiber use in India is given in equation 3.22. PCI‘FUIND = - 1.00 + 0.28 LN PDODPIND - 0.104 LN TIME - 0.08 D82 (3.22) (-3.46) (-2.62) iQUARED (CORR): 0.62 SEE: 0.013 DW: 2.08 PERIOD OF FIT: 1964-1986 remPCI'FUIND = Per capita total fiber use, India, PDGDPIND = Per capita deflated gross domestic product, India, TIME = Time variable, D82 = Zero-one variable, equals 1 in 1982, 0 otherwise, LN = Indicates variable transformed into logarithms. re variability of per capita total fiber use in India was explained by variability in per pita gross domestic product, a time trend and a zero-one dummy variable for 1982. The nation performed better with variables converted into logarithms than in the linear form. The per capita total fiber use in Japan iS given in equation 3.23. TUJPN = - 110.0 + 8.68 LN PDGDPJPN + 4.21 D73 (3.23) (3.30) QUARED (CORR): 0.81 SEE- 32.47 DW: 1.23 PERIOD OF FIT: 1964—1987 :rezPCI‘FUJPN = Per capita total fiber use, Japan, PDGDPJPN = Per capita deflated gross domestic product, Japan, D73 = Zemnc variable, equals 1 in 1973, 0 otherwise, LN = Indicates variable transformed into logarithms. capita total fiber use in Japan was explained by per capita deflated gross domestic duct and a zero-one dummy variable for 1973. The estimated coefficient on the income able gave and income elasticity of 0.58 when calculated at the mean of total fiber use. dummy variable for 1973 captured the increase in consumption due to the sudden ease in imports in that year. The per ca pcmjKOR : - 20.35 rsouattFD (00“ llherel’Cll’UKOR . PDGDPKOR on - LN lhe upward tr regression perir equation specil product and a : Variable in the due to political The per i PURW : . 1: RSQUARED (C0 PDGDPME D82 LN lhe variables capita gross c‘ use lagged on i“ an income heeded in the lhe duhlmy resllli of the 72 The per capita total fiber use in Korea is given in equation 3.24. PCI'FUKOR = - 20.85 + 1.75 LN PDGDPKOR + 0.71 PCI'FUKOR(-1) - 257 D72 (3.24) (2.49) (5.83) (-3.48) R-SQUARED (CORR): 0.97 SEE- 9.13 H-STAT: 2.01 PERIOD OF FIT: 1964-1986 WherezPCI'FUKOR = Per capita total fiber use, Korea, PDGDPKOR = Per capita deflated gross domestic product, Korea, D72 = Zero-one variable, equals 1 in 1972, 0 otherwise, LN = Indicates variable transformed into logarithms. The upward trend in per capita use of fibers in Korea was very strong during the regression period. because of this, lagged per capita fiber use was included in the final equation specification along with the logarithm of per capita deflated gross domestic product and a zero-one dummy variable for 1972. The statistical significance of a dummy variable in the equation for Korea in 1972 captured the substantial economic disruption due to political unrest during the early 19708. The per capita total fiber use in Mexico is given in equation 3.25. PCI'FUMEX = - 15.98 + 1.77 LN PDGDPMEX + 0.43 PCI’FUMEX(-l) - 0.82 D82 (3.25) (2.49) (2.07) (-2.17) R-SQUARED (CORR): 0.84 SEE 2.01 H-SI‘AT: 2.86 PERIOD OF FIT: 1964-1986 Where:PCI'FUMEX = Per capita total fiber use, Mexico, PDGDPMEX = Per capita deflated gross domestic product, Mexico, D82 = Zero-one variable, equals 1 in 1982, 0 Otherwise, LN = Indicates variable transformed into logarithms. The variables used to explain per capita total fiber use in Mexico were the logarithm of per capita gross domestic product, a zero-one dummy variable for 1982, and per capita fiber Ise lagged one year. The coefficient on the income variable was significant and resulted n an income elasticity estimate of 0.33. The lagged per capita fiber use variable was reeded in the equation to capture the rising trend in use throughout the regression period. 'he dummy variable for 1982 was included to account for the decline in fiber use as a :sult of the debt crisis in the early 1980s. The per ca pm?“ = 0.44 ‘ rsouaam (00m WAX D72 Of all the em all regions exr dependent vari regression peri 1972 and 1982. capita fiber us supplies gOing The per W=J whemPClFUm PDonm Die LN lhe equation cepita gross hher Ilse lagg blthe stron dummy vari resulllhg fro 73 The per capita total fiber use in Pakistan is given in equation 3.26. :IPUPAK = 0.44 + 0.84 PCI‘FUPAKGI) + 0.94 D72 - 0.82 D82 (3.26) (8.40) (251) (-2.17) SQUARED (CORR): 0.83 SEE: 2.40 H—SI‘AT: 0.78 PERIOD OF FIT: 1964-1986 rerezPCI‘FUPAK = Per capita total fiber use, Pakistan, D72 = Zero-one variable, equals 1 in 1972, 0 otherwise, D82 = Zero-one variable, equals 1 in 1982, 0 otherwise, LN = Indicates variable transformed into logarithms. fall the equations explaining per capita fiber use, the income variable performed well in . regions except for Pakistan, where it had the wrong Sign (negative). A lagged pendent variable performed well, reflecting the strong trend in the data throughout the gression period. Zero-one dummy variables were included in the regression for the years 72 and 1982, when there was an unexplained increase and decrease, respectively, in per )ita fiber use. The increase in fiber use in 1972 may have been caused by disruption Of splies going to Bangladesh soon after its independence. The per capita total fiber use in Turkey is given in equation 3.27. 'FUTUR = - 7.20 + 0.85 LN PDGDPI‘UR + 0.67 D74 + 0.73 PCI‘FUI'UR(-1) (3.27) (220) (250) (5.76) )UARED (CORR): 0.95 SEE: 1.02 H-SI‘AT: 1.48 PERIOD OF FIT: 1964-1986 :PCI'FUTUR = Per capita total fiber use, Turkey, PDGDPTU R = Per capita deflated gross domestic product, Turkey, D74 = Zero—one variable, equals 1 in 1974, 0 otherwise, LN = Indicates variable transformed into logarithms. equation explaining per capita total fiber use in Turkey contained the logarithm of per ta gross domestic product, a zero-one dummy variable for 1974 and per capita total use lagged one year. The large coefficient on the lagged fiber use variable was caused e strong upward trend in per capita fiber use during the regression period. The my variable accounted for an unexplained increase in fiber consumption in 1974, [ting from sharply lower exports of textiles to other countries which were experiencing a recession. The per ca ycn-‘UUSA = -19l.( RSQUARED (COR Whertcl’ClIUUSA PDGDPUSA UNIM’USA DMFAl DMFAl] DMFAIII LN lhe specificati same as used 1 to account for the variable w: per capita tor: for the three “it and am he variable ‘ illicificatim Otterall, seven had a between 0.3 , lumionelat \ ”M'Wncla Will‘mmptit till] 3 Cm 74 a recession. The per capita total fiber use in the United States is given in equation 3.28. PCI'FUUSA = - 191.0 + 23.1 LN PDGDPUSA - 0.28 UNEMPUSA - 1.21 DMFAI - 230 DMFAII - 2.74 DMFAIII (328) (7.79) (.132) (4.65) (.342) (.312) R-SQUARED (CORR): 0.82 SEE: 0.77 DW: 1.51 PERIOD OF FIT: 1964-1987 WherezPCI'FUUSA = Per capita total fiber use, United States, PDGDPUSA = per capita deflated gross domestic product, United States, UNEMPUSA == Unemployment rate (%), United States, DMFAI == Zero-one variable for MFAI period, equal 1 1974 to 1977 else 0, DMFAII == Zero-one variable for MFAII period, equal 1 1978 to 1981 else 0, DMFAIII = Zero-one variable for MFAIII period, equal 1 1982 to 1985 else 0, LN = Indicates variable transformed into logarithms. The specification used to explain the variability of fiber use in the United States was the same as used for the EEC-12. An unemployment variable was added to the specification to account for declining textile sales during periods of recession. Although not significant, the variable was kept in the equation because it was signed correctly. Also included in the per capita total fiber use equations were the three zero-one dummy variables, accounting for the three MFAS which operated during the estimation period. An index of prices. of textile and apparel products was tried in an initial specification of this equation. However, :he variable had a very low t-value (i.e., less than unity) and was dropped from the specification. Overall, the equations fitted the data reasonably well. Of the 14 equations estimated, ven had a corrected R-squared above 0.9, five equations had a correct R-squared of tween 0.8 and 0.9, and for only two equations was it less than 0.8. Some equations were to-correlated as evidenced by the DW and H-statistics”. Initially this problem was ”Autocorrelation may have arisen from model misspecification. The most important source of measurement error in any ile consumption model is measurement error in the dependent variable. There are standard procedures used by statisticians llecting the raw data in estimations of end use cotton consumption (especially conversion factors used to estimate the raw rer equivalents of textile trade). These methods rest on data and assumptions that are periodically revised and known to ntain errors. Since these errors are unlikely to be random from period to period. the M0! that “"9"“ the dependent riable may contains an autocorrelation structure. corrected using required that a In most cases, t simulation perfi inthe cotton sh While improvit isneeded in de‘ elasticity estim and general p forecasts. 3.6 [M 3.6.1 E The pri elasticities as Korea, With 311d newlnim the EEC .01 ttttl, China. the Use of manufacture relative pllC 75 corrected using the Cochrane Orcutt transformation. However, using this technique required that a lagged dependent variable appear on the right-hand side of the equation. In most cases, the coefficient on this variable was very large and resulted in unsatisfactory simulation performance. Therefore, the equations were left uncorrected. As was the case in the cotton share equations, it was necessary to use time and lagged dependent variables. While improving the fit of the equations considerably, their inclusion means that caution is needed in deriving model forecasts. Nonetheless, the equations provided plausible income elasticity estimates, which together with the assumptions about future population, income and general prices, as well as the trend components, were used to make consumption forecasts. 3.6 Demand Elasticig; Estimates 3.6.1 Price Elasticities The price elasticities for cotton use are presented in Table 3.1. The demand lasticities are highly inelastic, ranging from -0.02 for India to -0.33 in the Republic of orea. With the exception of Australia, the elasticities are higher for the industrialized nd newly-industrialized countries (e.g., the United States -0.3, the Republic of Korea -0.33, he EEC -0.14) and lower in the developing countries (e.g., Central Africa -0.02, Argentina .07, China -0.08). The higher elasticities in the industrialized countries most likely reflect e use of more modern processing facilities in these countries, which enables anufacturers to quickly change the mix of fibers processed in response to changes in lative prices. Table .1 Fri Region2 Argentina Australia Brazfl Cent Africa EEC-12 Egypt India Japan Korea Mexioo Pakistan Turkey‘ USA \ lBased on 3 countries. The elasticitie Shaft equatio Ems/r = (3C *bt-Pr). (at Values for C 2N0 elasncrn demand equ: Where poly. usedillllle: thtng sign. 76 Table 3,1 Price Elasticities of Cotton Use Region2 Cotton Price Polyester Staple Price3 Argentina - 0.07 + 0.07 Australia - 0.06 + 0.06 Brazil - 0.18 + 0.17 Cent. Africa - 0.04 + 0.04 China - 0.08 + 0.08 EEC-12 - 0.14 + 0.14 Egypt - 0.17 + 0.17 India - 0.02 + 0.02 Japan - 0.04 + 0.04 Korea - 0.33 + 0.33 Mexico - 0.09 + 0.09 Pakistan - 0.04 + 0.04 Turkey‘ - 0.13 USA - 0.30 + 0.22 1Based on a regression period 1964-1986 for developing 1964-1987 for industrialized countries. The elasticities of cotton use with respect to price (Emu) were derived from the cotton share equations. Ecru/P = (aCTU/aP).(P/CI'U). Recall CI'U = CTSH * TFU and that CI‘SH = (E:',‘b,.Xi + brPk). (aCTU/aP) = brTFU and Em”, = b,.TFU.P/CTU. Values for CPU, P and TFU were taken as historical means. 2 N o elasticities can be reported from model regions for which prices do not appear in the demand equations (ie. India, Pakistan, Japan and Egypt). Where polyester price elasticities equal the cotton elasticities a ratio of these prices was sed in the share equation. 6 price of polyester was dropped from the share equation for Turkey because it had the ong sign. Hence no elasticity could be reported. The low ten years less ti mills, with 80% price changes : traded goods, i price variable consumed is is A rathe that it would capacity. Hov inthis case [rt cotton industr be that Chin; In Tal estimates are period 1955. 1954-86). C Price elastici “We elastic studies are r sp‘xlficatior 77 The low elasticity for Australia may reflect the fact that, on average, over the past ten years less than 20% of cotton goods sold to consumers has been provided by domestic mills, with 80% coming from imports. Given that the domestic consumption response to price changes in the model is through the milling sector and not through the price of traded goods, it is expected that the elasticity will be small. This may explain also why the price variable was not significant in the case of Japan, where about one-third of cotton consumed is imported. A rather surprising result is the low elasticity reported for China. It was expected that it would be larger given that China has recently invested heavily in modern plant capacity. However, the elasticity reflects average relationships over the estimation period-- in this case from 1964 to 1986--and therefore encompasses periods of stagnation in China’s cotton industry during the mid-19703 and earlier. Another reason for the low elasticity may be that China’s wholesale cotton prices are different from world producer prices. In Table 3.2 the elasticities obtained from previous studies are reported. All estimates are inelastic and range from -0.09 (Thigpen, for developing countries over the period 1955-75) to -0.34 (Mues and Simmons, for the United States over the period 1954-86). Consistent with the results from this study, other researchers have found the price elasticities to be higher for the higher income countries. Elasticities also tend to be more elastic for the more recent time periods. Generally, the elasticities reported in other studies are not dissimilar to the ones found in this study despite major difierences in model specification, period of estimation and regional coverage. Tail .2 Fri Author Donald et a1 Dudley lhigpen Adams & Be Mues & Sim 3.62 m The i are differen The mOSt st for the higl results do consufltptio reSpectral) m0“ pros 78 Table 3,; Price Elasticities Esgm’ ates thained From Previous Studies Author Region Dep.Var. Period Elasticity Donald et al US. ‘ Fiber Dem. 1948-60 - 0.14 Dudley World Fiber Dem. 1953-70 - 0.25 Thigpen DCs Cotton Dem. 1955-75 - 0.20 LDCs Cotton Dem. 1955-75 - 0.09 Adams & Behrman DCs Fiber Dem. 1955-73 - 0.23 Mues & Simmons U.S. Mill Con. 1954—86 - 0.34 W.Europe Mill Con. 1954—86 - 0.26 ROW Mill Con. 1954-86 - 0.10 3.6.2 Income Elasticities The income elasticities of total fiber demand are reported in Table 3.3. The results are different from what was expected, ranging from 0.12 in Turkey to 1.08 in the BBC. The most striking feature of the results is that, in general, income elasticities are greater for the higher-income countries (e.g., the United States, Japan and the EEC). These results do not support the hypothesis that income elasticities decline as income and consumption increase. The elasticities of 1.08 and 1.04 for the BBC and the United States, I espectively, seem high. A possible explanation is that the large increases in the use of otton products in the United States and the EEC are mainly cheap imports, especially from the newlj consumption It For the reason use equations. influences on Table 3.3 Inc Region Argentina Australia Brazil Central Afric China EEC-12 Egypt India Japan Korea Mexico Turkey United Stats ll Based s IndustrializE 2/ ElaSIlC PCDFGDP region. 79 from the newly-industrialized countries of Southeast Asia as well as China. Increased consumption may have resulted from lower import prices as well as from income increases. For the reasons discussed in section 3.4, prices do not appear in the per capita total fiber use equations. Therefore, the income variable may be capturing both price and income influences on consumption. Table 3,3 Income Elasticities of Demand for All Fibers‘. Region Income2 Elasticity Argentina 0.58 Australia 0.65 Brazil 0.32 Central Africa 0.54 China 0.91 EEC-12 1.08 Egypt 0.68 India 0.28 Japan 0.58 Korea 0.24 Mexico 0.33 Turkey 0.12 United States 1.04 1/ Based on a regression period 1964- 1986 for developing countries and 1964-1987 for industrialized countries. 2 / Elasticities derived from the semi-log functional form PCTFU = 2:',‘b,Xi + bkLN PCDFGDP are given by b, / PCTFU. The historical mean of PCTFU is used in each region. Income study by Thll industrial and elasticities dec period Thigpe: data Interest industrial cou industrial cou Author \ Donald et a1 Dudley Magleby & Missaien Tltigpen Adams & Bthl‘mar 80 Income elasticity estimates from other studies are reported in Table 3.4. In the study by Thigpen, income elasticities for cotton and total fiber are reported for the industrial and developing countries and for the CPEs. His findings show that income elasticities decline as incomes rise. However, using cross-sectional data over a three-year period Thigpen avoids the multicollinearity problems associated with using only time-series data. Interestingly, Adams and Behrman found the income elasticity to be larger for the industrial countries than for the developing countries, although the magnitudes for the industrial countries were much lower than in this study. Table 3.4 Income Elasticities Estimates Obtained Fm Previous Studies Author Region Dep. Var. Period Elasticity Donald et al US. Fiber Dem. 1948-60 0.80 Dudley World Fiber Dem. 1953-70 0.27 Magleby & World Fiber Dem. 1964 0.62 Missaien Thigpen DCs Cotton Dem. 1970-72 0.07 LDCs Cotton Dem. 1970-72 0.50 CPEs Cotton Dem. 1970-72 0.20 DCs Fiber Dem. 1970-72 0.30 LDCs Fiber Dem. 1970-72 1.40 CPEs Fiber Dem. 1970-72 0.60 Adams DCs Fiber Dem. 1955-73 0.60 & Behrman LDCs Fiber Dem, 1955-73 0.47 ll mm This section des demand for cotton. '11 there is no direct oomsr cotton is a substitutal composition and that purchase price, equat The cotton share of t of cotton and manuf relative prices. This PFOVided the theoret Performed well with, iI‘lllortant was that t with previous econo 81 Summagy This section described the econometrically estimated equations used to explain the rand for cotton. The task of modeling cotton demand was complicated by the fact that e is no direct consumer demand for raw cotton but rather for textile products, in which on is a substitutable input. Given that individuals are relatively insensitive to fiber position and that the fiber component represents only a small proportion of the final :hase price, equations for total fiber demand were estimated as a function of income. cotton share of total fiber demand was estimated as a function of the relative prices otton and manufactured fibers and captures the price sensitivity of manufactures to tive prices. This framework was consistent with the two-stage budgeting model which 'ided the theoretical basis for the econometric estimation. The regression equations ormed well with, in general, high explanatory power and significant coefficients. More )rtant was that the price and income elasticity measures are reasonable and consistent previous econometric studies of the textile market. ‘Rw m 4.1 Introduction An inrportant output since the early of area. Since 1963, a an increase of over 6 at around 30 millior separately as import: product of these con Cotton prodr annual crop which T cereals and oilseed: become fairly stand sight of the theore theoretical issues is 4.2 Theoretical loss A theoretic [0 malimize their mputs and 0Ulput. 4. Cotton Production 1 Introduction An important feature of world cotton production is that the substantial increase in utput since the early 19605 has resulted from yield increases and not from an expansion f area. Since 1963, average world yields have risen from 338 kg/ha to 545 kg/ha in 1988, 11 increase of over 60%. Over this period the cotton area has remained fairly constant t around 30 million hectares. This suggests that yields and area should be modeled eparately as important information may be lost if production is estimated directly as the Iroduct of these components. Cotton production is more straight forward to model than demand. Cotton is an nnual crop which has similar production requirements to other annual crops such as ereals and oilseeds. While the specification of models of annual crop supply have ecome fairly standard over the years and can be used here, it is important not to lose fight of the theoretical rationale behind their specification. A brief overview of the heoretical issues is presented below. I .2 Theoretical Issues A theoretical model of agricultural supply assumes that individual farmers attempt ' maximize their profits, subject to technological constraints and exogenous prices of puts and output. Assume farmers produce a number of commodities and are cost 82 minimizefs- The“ rec where, Y = vwor Of x = VCCIOI 01 Z = valor 0‘ W : VOCIOI' 0 V = producti Given a vector of en maximization, with p The profit maximila Finaly using Hotelli Production of each opportunity set (ie. complementary cor Most mode these components : prices and fixed imp the price of the m the area equation. com ' modrty and on resellIces such as 83 mizers. Then technology can be expressed in terms of a cost function, C(Y,W,Z) = minm(WX : (Y,X,Z is in V) ) re, Y = vector of i outputs, X = vector of j variable inputs, Z = vector of k fixed inputs, W = vector of input prices, and V = production possibility set. en a vector of exogenous output prices, P, cost minimization is equivalent to profit Limization, with profit given by, 1r = PY - C(Y,W,Z). : profit maximization for the function G(P,W,Z) is given by, G(P,W,Z) = maxw(PY - C(Y,W,Z) ). ally using Hotelling’s lemma, the vector of production Y(P,W,Z) can be obtained from, Y(P,W,Z) = aG(P,W,Z) / 6P. tinction of each output is a function of all prices of products within the farmers’ l brtunity set (i.e., the price of the commodity itself plus the prices of all competing and lementary commodities), all input prices and the levels of fixed inputs. Most models of annual crop production separate yields from area and estimate lo components separately. According to the theory presented above, input and output lps and fixed inputs are determinants of production. Typically, the relationship between hrice of the modeled commodity and the prices of competing products is captured in .rea equation. The area equation is often specified as function of lagged prices of the nodity and one or two commodities believed to compete with the commodity for farm trees such as land and labor. The relationsh the yield equation. Th Exogenous factors affr the yield equation. 4.3 Literature Revie Compared tr production and of the estimated a United 5 areas (i.e., the Delta variables used incluc incorporated prices and sol’beans). Al diversion Payments States into four 58F appropriate mans! reglons and be Calls. resources. Statbird an‘ United States The Plains (non‘ifrigar Planted and Wear} crop Year, In Addi 84 The relationship between the price of inputs and the price of output is captured in Iield equation. The effect of the fixed inputs are felt through the equation coefficients. genous factors affecting supply, such as weather and technology, are often included in yield equation. Literature Review Compared to cotton demand, relatively few studies have focused on cotton duction and of these most have been for the United States. For example, Evans (1977) mated a United States cotton acreage response equation for the four major producing as (i.e., the Delta region, the Southeast, the Southwest, and the West). The explanatory ables used included average variable and opportunity costs of producing cotton, which trporated prices and costs of cotton and competing crops (e.g., corn, barley, sorghum, ‘ soybeans). Also policy variables were used, such as national acreage allotments, lrsion payments and direct payments. Of most interest is the division of the United es into four separate producing regions and the use of region-specific data. This is "opriate because United States cotton production practices differ significantly across )ns and because of differences in the types of crop that compete with cotton for farm iurces. Starbird and Hazera (1981) estimated cotton yield equations for four regions of the ed States. These were the Mississippi Delta, Texas High Plains (irrigated), Texas High 5 (non-irrigated) and California. Yields were explained by the number of acres ed and weather variables for rainfall and temperature during crucial periods of the year. In addition, a policy dummy variable was used to capture the effect of the skip- row policy rules‘. No whiCh explained a higl Monke (1933) countries for the peri prices of a number 0 coefficients on 00mP‘ equation is that a lag coefficient of 0.95 (8 explains the lack of s Mues and Si States, Australia anc world price of cotton used as regressors. including dummy va years, the guarantee reqtlirement for par POIiCy, were signific World except that t Hamid m Sim regions of P 85 or policy rulesl. No economic variables were included in the United States equations ich explained a high proportion of yield variability. Monke (1983) estimated an equation for the production of all price-responsive mtries for the period 1960-80. The world price lagged one period as well as current ces of a number of competing commodities were used as regressors. None of these efficients on competing commodities were significant, however. A concern about the ration is that a lagged dependent variable was added to the specification which had a :fficient of 0.95 (and t-value of 52.84). This variable drives the model and probably tlains the lack of significance of the coefficients on competing crops. Mues and Simmons (1988) estimated the area planted to cotton for the United tes, Australia and the Rest-of-the-World. For the United States equation the lagged ‘ld price of cotton, the price received for soybeans and a lagged dependent variable were d as regressors. Various policy instrument variables were added to the specification, uding dummy variables for the payment-in-kind (PIK) program and for the soil bank rs, the guaranteed price to growers under the farm program, and the acreage-reduction iirement for participation in the program. All the variables, including those capturing 3y, were significant. A similar specification was used for Australia and the Rest-of-the- 'ld except that the policy variables were dropped from the equations. Hamid et al. (1987) estimated cotton area and yield equations for the Punjab and regions of Pakistan. The equations reported did not perform well with most allotment years of 1954-61, all acreage in fields planted to skip-rowed cotton were counted as cotton acreage, including lth. During 1962-65, these rules were relaxed, and only the planted rows were counted as acreage planted. Acreage imposed in 1966-67. Since 1968, only land actually planted to cotton has been counted as cotton land in determining program provisions" (Starbird and Hazera, p. 18). coefficients not signiflc were significant. The number of tubewells in such as improvements Thigpen and It determined by laggec more competing cmp linear trend, lagged ( the competing cr0p 1 4-4 W A3 suggeste iptfiification for anr equation for yield 3 these components. P C C Where, P D =C com =c CFHA = PCC _ PCT : PF :1 W :1 86 :oefficients not significant and some with the wrong sign. None of the economic variables were si ' 1cant. The most significant variable was water availability, measured by the number of tubewells in operation. This variable may have also been capturing other factors uch as improvements in cotton varieties. Thigpen and Mitchell (1988) estimated equations for both area and yield. Area was letermined by lagged cotton area, lagged cotton revenue and lagged revenue of one or more competing crops such as coarse grains and soybeans. Yields were determined by a [near trend, lagged cotton price and the current fertilizer price. The fertilizer price and he competing crop revenue were exogenous in the model. .4 Model of Cotton Supply As suggested above, cotton supplies were estimated using a fairly standard )ecification for annual crops. Production in each region was determined by a behavioral quation for yield and area planted, and an identity giving production as the product of rese components. That is, PD = CTYD * CTHA, CTYD = f(PCI‘, PF, W), CI‘HA = f(PCT(-1), PCC(-1) ). rere, 3 = Cotton production, [YD = Cotton yield per hectare, FHA = Number of hectares planted, 3C = Price of competing crops, T = Price of cotton, Price of fertilizer, and Weather. ll As mentioned in the theoretical review, yields are determined by profitability of the cotton enterprise whit To capture these relat explanatory variables of the model regions. and temperature play little irrigated cotton temperature data for the specification the cases weather data a period. Another fact Oihigh'yldding and variable was added trended upwards th variation in yield. j the logarithm of tir to measure the per ulitter-limit (i.e., Unfortunately, dat available. H Oweve for new technolog The rationale fer forced Onto lan d 87 cotton enterprise which depends on the relationship between product and factor prices. To capture these relationships the cotton price and price of fertilizer were used initially as explanatory variables in the yield equations. This specification performed badly for many of the model regions. On careful inspection of the yield data it was observed that rainfall and temperature play a crucial role in determining yields, especially in those countries with ittle irrigated cotton area. This led to the collection of large amounts of rainfall and :emperature data for the major producing countries. When these variables were added to he specification the performance of the equations improved dramatically and in some ases weather data alone explained up to 80% of the variations in yield over the historical seriod. Another factor affecting cotton yields is the impact of the rapid growth in the use fhigh-yielding and drought-resistent cotton varieties. To capture this development, a time ariable was added to the specifications. In some regions of the model in which yields 'ended upwards throughout the estimation period, a time trend explained almost all the triation in yield. In forecasting production up to 2005 the time variable was replaced by re logarithm of time. Perhaps a better variable to capture this historical effect would be i measure the percentage of total crop under high-yielding varieties. This would give an )per-limit (i.e., 100%) to increases in yield from the uptake of new technology. nfortunately, data on area allocated to high-yielding varieties of cotton are not readily ailable. However, this variable would raise problems in forecasting as it would not allow r new technology. Another variable that was successful in some cases was area planted. re rationale for the inclusion of this variable was that as area increases, production is "ced onto land of lower quality, forcing average yields to decline. The number profitability of the cot iann resource requi determining cotton at as coarse grains and The producti: competing wmmodit “Sing the United St: index for each mos significant or Wrong maybe eXplained h prices. in many c. farmers facing a se 1“ an attempt to competing crops“ These Prices Were , without exception considerably. Nex Only in the equati. 011“le index I A. prim Cm . if“ lmcx)?:r Pnoe index W8 0i “ailable. H 0 88 The number of hectares planted with cotton is specified to depend on the profitability of the cotton enterprise relative to the profitability of crops which have similar farm resource requirements. To capture substitution between crop enterprises in determining cotton area, the lagged prices of cotton and the prices of competing crops such as coarse grains and soybeans were included in the regression equations. The production equations were estimated using international prices for cotton 2and :ompeting commodities. These prices were adjusted by converting into domestic currency using the United States dollar exchange rate and by deflating using the consumer price ndex for each model regions. The results were unsatisfactory with many prices not ignificant or wrongly signed. The poor performance of price variables in the equations may be explained by the fact that producers in many of the regions do not face world trices. In many countries government intervention in production and trade results in armers facing a set of local prices which differ substantially from the world price levels. 1 an attempt to capture farmer response to prices, local price data for cotton and )mpeting crops were collected for some regions (i.e., India, China, and the United States). hese prices were also in local currency and deflated by the consumer price index. Almost ithout exception local prices were significant and the area equations improved msiderably. Next, local prices were related to world prices using price linkage equations. nly in the equations for the United States were statistically significant relationships found :tween the local and world cotton prices. Local prices of other countries (e.g., China and Index 'A' price. :umer price index was used because for many regions a more appropriate index (e.g., a wholesale price index or producer as not available. However, the consumer price index is likely to be highly correlated with other more appropriate deflators. lodia) were treated a: In many count United States cotton West. in each region crops (e.g., cotton cc region, with sorghum barley in the West). substantially across r and area and total production. A regit for area and yield f( and Western regior 4-5 Mm 4.5.1 m ~lhe esfirna 19644988 and are ‘0 e8timate hem, overall the equati. than 30%. In the variables, In gem 89 India) were treated as policy variables and made exogenous in the model. In many countries, cotton production takes place in distinct regions. For example, United States cotton output is produced in the Delta region, Southeast, Southwest and West. In each region farmers face different prices and choices with respect to alternative :rops (e.g., cotton competes with soybeans in the Southeast, with soybeans in the Delta region, with sorghum and winter wheat in the Southwest and with alfalfa, wheat, corn and )arley in the West). Also, yields are affected by rainfall and temperature which differ :ubstantially across regions. United States production was obtained by the product of yield tnd area and total United States production was obtained by summing the regional troduction. A regional disaggregation was attempted for India and Pakistan. Equations or area and yield for the Punjab and Sind regions of Pakistan and the Northern, Southern nd Western regions of India were estimated successfully. .5 Empirical Results 4.5.1 Yield Equations The estimated equations for cotton yields were estimated using 01.8 for the period 964-1988 and are presented in equations 4.1 - 4.17 below. Yield equations were difficult t estimate because there was often a lot of randomness contained in the series. However, rerall the equations perform well with 14 out of 17 having a corrected R-squared greater an 80%. In many equations the fit was improved by the inclusion of zero-one dummy riables. In general, these were capturing abnormally good or bad weather conditions for tton production. The estimated 4.1. (133) nsoUARED (CORR): 03 Where CI'YDARG = Cl DFCI'PARG = D TIME = TI D70 = 2 D81 = 2 LN = 11 The cotton y a time trend variable was included to cap technology, such 21 responses, and othe this was accounted the 1981 cr0p seas. access to special er in Which the price The BStimz 4.2. (5.1 W“ Crlitmus a on . no mo II The “mo“ Yield 90 The estimated equation explaining cotton yields in Argentina is given in equation DARG = 0.26 CI‘YDARG(—l) + 1.09 DFCI‘PARG + 0.90 D81- 0.06 D70 + 0.06 LN TIME (4.1) (1.88) (3.74) (2.49) (4.62) (3.11) )UARED (CORR): 0.82 SEE; 0.02 H-SI‘AT: 0.03 PERIOD or m: 1964-1988 re: CI'YDARG = Cotton yield, Argentina (m.tons / hectare), DFCI‘PARG = Deflated cotton price, Argentina, TIME = Time variable, D70 = Zero-one variable, equals 1 in 1970, 0 otherwise, D81 = Zero-one variable, equals 1 in 1981, 0 otherwise, LN = Indimtes variable transformed into logarithms. The cotton yield series for Argentina was explained by the deflated price of cotton, ne trend variable and zero-one dummy variables for 1970 and 1981. The time variable included to capture the improvement in yields through time as a result of improved ,nology, such as higher yielding seed varieties, improved fertilizer and chemical onses, and other better farming practices. The rate of increase declined over time and was accounted for by converting the time variable into logarithms. The high yield in 1981 crop season was a consequence of the policy change which gave cotton growers 58 to special credit to purchase inputs. The equation for Argentina was one of the few hich the price of cotton was significant. The estimated equation explaining cotton yields in Australia is given in equation AUS = 0.77 + 0.02 TIME - 0.27 D72 - 0.24 D75 - 0.27 D76 (4.2) (5.29) (-203) (4.31) (-203) IARED (CORR): 0.68 SEE: 0.34 DW: 1.82 PERIOD OF FIT : 1964—1988 CI'YDAUS = Cotton yield, Australia (m.tons / hectare), TIME = 'I'rme variable, D72 = Zero-one variable, equals 1 in 1972, 0 otherwise, D75 = Zero-one variable, equals 1 in 1975, 0 otherwise, D76 = Zero-one variable, equals 1 in 1976, 0 otherwise. :otton yields in Australia were explained by a time trend and zero-one dummy variables for 1972. 19‘ to be significant The uptake of improved < zero-one dummy vari low rainfall in those ‘ The estimate ovum = 0.13 + 027 c (1.68) tsounnru (corn): or Where CIYDBRA = C DFCI‘PBRA = I DFPNPBRA = I 'IIME = '1 D70 = 2 D84 = 2 Brazilian cotton yiel zero-one dummy va and fertilizer prices the variable was ke used in the equati: esPatially in the us included mange , letels, The eSlim: mm = 008 + 08 (13.3 91 iables for 1972, 1975 and 1976. Prices of cotton and competing crops were found not be significant. The trend variable was highly significant (t-value of 5.29), reflecting the ake of improved cotton production technology, especially since the early 19803. The ia-one dummy variables captured yields substantially below trend, mainly the result of rainfall in those years. The estimated equation explaining cotton yields in Brazil is given in equation 4.3. DBRA = 0.13 + 0.27 CI‘YDBRA(—1) + 0.03 DFCI'PBRA/DFFNPBRA + 0.003 TIME - 0.06 D70 + 0.10 D84 (4.3) (1.68) (1.72) (2.21) (-2.56) (4.03) 'UARED (CORR): 0.84 SEE: 0.009 H—STAT: - 0.45 PERIOD OF FIT: 1964-1988 e: CIYDBRA = Cotton yield, Brazil (m.tons / hectare), DFCI'PBRA = Deflated cotton price, Brazil, DFFNPBRA = Deflated fertilizer price, Brazil, TIME = Time variable, D70 = Zero-one variable, equals 1 in 1970, 0 otherwise, D84 = Zero-one variable, equals 1 in 1984, 0 otherwise. ;ilian cotton yields were explained by cotton and fertilizer prices, a time trend variable, -one dummy variables for 1970 and 1984, and cotton yield lagged one year. The cotton fertilizer prices were entered in the equation as a ratio and, although not significant, 'ariable was kept in the equation because it had the correct sign. The time trend was in the equation to capture improved technology use in the Brazilian cotton sector, :ially in the use of higher-yielding cotton varieties. The dummy variable for 1984 was fled because the near-perfect growing conditions in that year sent yields to record The estimate equation for cotton yields in Central Africa is given in equation 4.4. ZAP = 0.08 + 0.87 CI'Y DCAF(-1) - 0.05 D75 - 0.00001 CI'HACAF (4.4) (13.31) (-377) (-154) ARED (CORR): 0.92 sen- o.oo44 H-SI‘AT: - 0.26 PERIOD OF m: 1964-1988 CI‘YDC‘AF = Cotton yield, Central Africa (m.tons / hectare), CI'HACAF = Cotton area, Central Africa, or = Zea in the equation for tl one year, the cotton trend in yields was ca weather-related yielc was significant in the area within the re git where yields were 10 because it had the r The estimat moon = ~258 + one (195 mum (CORR): t incentives Were , introduction of It cotton and fertiliz letting, prices w result was nor Sm 92 D75 = Zero-one variable, equals 1 in 1975, 0 otherwise. In the equation for the Central Africa region cotton yields were explained by yield lagged one year, the cotton area, and a zero-one dummy variable for 1975. The strong upward :rend in yields was captured by lagged yield and the dummy variable for 1975 captured the weather-related yield reduction in many African countries in that year. The cotton area yas significant in the equation. The negative sign on this variable showed that as greater urea within the region was planted to cotton, production moved onto land of lower quality vhere yields were lower. While this variable was not significant it was kept in the equation recause it had the correct sign. The estimated equation explaining cotton yields in China is given in equation 4.5. TYDCHI = - 258 + 0.003 DFCI'PCHI - 0.001 DFFNPCHI + 1.15 LN TLME- 0.16 D88 (45) (1.95) (.345) (9.20) (2.67) SQUARED (CORR): 0.95 SEE- 0.0289 DW: 2.31 PERIOD OF FIT: 1977-1988 here: CI'YDCHI = Cotton yield, China (m.tons / hectare), DFCI‘PCHI = Deflated cotton price, China, DFFNPCHI = Deflated fertilizer price, China, TIME = Time variable, D88 = Zero-one variable, equals 1 in 1988, 0 otherwise, LN = Indicates variable transformed into logarithms. he variability of cotton yields in China was explained by a regression estimated from 1977 1988. This shortened time period was used because prior to the late 19708, before price centives were introduced into China, yields were almost unchanged. With the troduction of market incentives Chinese producers became highly responsive to both tton and fertilizer prices. This was demonstrated in the equation in which cotton and tilizer prices were statistically significant. The logarithm Of time was added to capture :hnological improvements. Rainfall was found not to be significant in the equation. This :ult was not surprising given that most of China’s cotton growing area is highly irrigated. The zero-one dummy during the crucial mo The estimate (6335) nsounneo (com 03 Where CI'YDBGY = CC 078 =2: use =7. Cotton yields in Eg variables for the ye: was evidence of a 5' model (Outlook Int ot the high-quality oi the more comrr Provided by the 10 cotton prices‘. The estjm; e(Nation 4.7. mutton . 032 , 0' (3. RSQUARED (CORR): wittrc: CWDINDN: CHIMNDN 11MB . Dil D73 4 “know, “My 0t the in W "‘° “White in . 93 e zero-one dummy variable for 1988 was included to account for poor weather in China :‘ing the crucial months of August and September which resulted in extremely low yields. The estimated equation explaining cotton yields in Egypt is given in equation 4.6. DEGY = 0.99 CI'YDEGY(—1) + 0.21 D78 - 0.12 D88 (4.6) (63.26) (3.38) (-1.96) IUARED (CORK): 0.81 SEE: 0.0851 HSI‘AT: 0.02 PERIOD OF FIT: 1964-1988 he: CI'YDEGY = Cotton yield, Egypt (m.tons / hectare), 3 D78 = Zero-one variable, equals 1 in 1978, 0 otherwise, D88 = Zero-one variable, equals 1 in 1988, 0 otherwise. ton yields in Egypt were explained by yields lagged one year and zero-one dummy ables for the years 1978 and 1988. The coefficient on the lagged yield variable (0.99) evidence of a strong trend in Egyptian cotton yields. The price of cotton used in the 61 (Outlook Index "A") was not significant. However, most of Egyptian production is te high-quality extra-long staple (ELS) cotton, with a market quite separate from that te more common medium staple cotton. Evidence that the markets are distinct is ided by the low price transmission elasticities between the ELS and medium staple 11 prices‘. The estimated equation explaining cotton yields in Northern India is given in tion 4.7. TNDN = 0.32 + 0.06 LN TIME - 0.0002 CI'HAINDN + 0.0001 RINDN + 0.15 D71 + 0.06 D73 (4.7) (3.49) (-2.78) (2.92) (5.70) (251) ARED (CORR): 0.86 SE: 0.0076 DW: 1.89 PERIOD OF FIT: 1965-1984 CI'YDINDN = Cotton yield, Northern India (m.tons / hectare), CI‘HAINDN = Cotton area, Northern India, TIME = Time variable, D71 = Zero-one variable, equals 1 in 1971, 0 otherwise, D73 = Zero-one variable, equals 1 in 1973, 0 otherwise, ric study of the Egyptian ELS cotton market is being undertaken at the World Bank In the model the ELS price is e variability in cotton production. At a later stage the Egyptian ELS model will be incorporated in to this model. fits reported here should be treated as preliminary. Cotton yields in Nortl region rainfall, time, increase declined 0 logarithms. The are: yields as cotton was ; significant No statis variables, even when equation. The estirna equation 4.8. CWINDS = one + on (3.4 R‘SQUARED (CORR): Where CI'YDINDS a Cl‘tlAlNDs = “ME = 084 RlNDS LN II ll " The equation ex] rainfall, and a zer Vallabies was Sim and of farm inpu 94 RINDN = Annual rainfall, Northern India (mm), LN = Indicates variable transformed into logarithms. ton yields in Northern India were explained by the area of land planted to cotton in the on, rainfall, time, and zero-one dummy variables for 1971 and 1973. Since the rate of l ease declined over the regression period the time variable was converted into l lrithms. The area planted variable was significant and captured the decrease in average is as cotton was grown on land of lower quality. The annual rainfall variable was highly ificant No statistically significant relationships were found between yield and economic ables, even when the local prices of cotton and competing crops were included in the ation. The estimated equation explaining cotton yields in Southern India is given in ttion 4.8. IINDS = 0.10 + 0.05 LN TIME - 0.00001 CIHAINDS + 0.00009 KINDS - 0.20 D84 (4.8) (3.49) (-0.76) (2.19) (8.87) JARED (CORR): 0.94 SEE: 0.“)426 DW: 1.64 PERIOD OF FIT: 1965-1984 CI'YDINDS = Cotton yield, Southern India (m.tons / hectare), CI'HAINDS = Cotton area, Southern India, TIME = Time variable, D84 = Zero-one variable, equals 1 in 1984, 0 otherwise, RINDS = Annual rainfall, Southern India (mm), LN = Indicates variable transformed into logarithms. equation explaining cotton yields in Southern India contained time, cotton area, all, and a zero-one dummy variable for 1984. The rationale for the inclusion of these ales was similar to that for the Northern India equation. Again, the prices of cotton if farm input were not found to be significant. The estimate equation 49- CIYDINDW = 0.“ IN m (30. ) nsoUARED (CORR): 03 A vim CIYD‘NDW TlME ms D83 e: __. In 0.1 k1 -‘ E lhe specification 0' logarithmic time tre 1983. The yield in prohibited timely 5] significant relations The estima mom = 034 + 0.1 (95! Homer) (conn): Where ClYDMEX = TIME D78 D88 LN u II II Cotton yields in variable convertev improvements in varieties, improve prices of cotton 2 95 The estimated equation explaining cotton yields in Western India is given in ration 4.9. DINDW = 0.04 LN TIME + 0.00006 RSUINDW - 0.03 D76 - 0.02 D83 (4.9) (30.69) (6.61) (-5.50) (-2.43) PUARED (CORR): 0.88 SEE: 0.0006 DW: 2.21 PERIOD OF FIT: 1965-1984 re: CI'YDINDW = Cotton yield, Western India (m.tons / hectare), TIME = Time variable, D76 = Zero-one variable, equals 1 in 1976, 0 otherwise, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise, RSUINDW = Summer rainfall (mm), Western India, LN = Indicates variable transformed into logarithms. specification of the equation explaining cotton yields in Western India included a .rithmic time trend, rainfall in the summer months, and a zero-one dummy variable for I. The yield in 1983 was abnormally low as a results of heavy unseasonal rainfall. This tibited timely spraying of insecticides and led to severe pest damage. No statistically ificant relationships were detected between cotton yields and cotton and inputs prices. The estimated equation explaining cotton yields in Mexico is given in equation 4.10. tMEX = 0.34 + 0.18 LN TIME + 0.09 D78 + 0.16 D88 (4-10) (9.59) (3.38) (3.46) JARED (CORR): 0.87 SEE: 0.0422 DW: 2.11 PERIOD OF FIT: 1964-1988 CI'YDMEX = Cotton yield, Mexico (m.tons / hectare), TIME = Time variable, D78 = Zero-one variable, equals 1 in 1978, 0 otherwise, D88 = Zero-one variable, equals 1 in 1988, 0 otherwise, LN = Indicates variable transformed into logarithms. an yields in Mexico increased throughout the regression period and only a time ble converted into logarithms was found to be significant. This variable captured the )vements in yields resulting from improved technology, such as higher yielding seed :ies, improved fertilizer and chemical responses, and better production practices. The s of cotton and inputs were not statistically significant when included in the equation. The estimate is given in equation (735) R—SQUARED (CORR): 0. Where CI'YDPAKP = 1 TIME =T D‘ll =Z D83 = 2 The impact of techr No significant relat Azero-one dummy heavy unseasonal r pest damage. The estimz INCH in equation (7.69) Where: CTYDPAKS DFCTPPAK . IRRPAK . mistress TIME 072 D78 083 The Variables f0 Pakistan Were th Sind (emu, a \ dummy Variable, which the price 96 The estimated equation explaining cotton yields in the Punjab region of Pakistan ven in equation 4.11. (PAKP = 0.27 + 0.03 TIME + 0.09 D71 - 0.26 D83 (4.11) (7.36) (2.05) (-5.61) JARED (CORR): 0.80 SEE: 0.0296 DW: 1.00 PERIOD OF FIT: 1965-1985 CI'YDPAKP = Cotton yield, Punjab region, Pakistan (m.tons / hectare), TIME = Time variable, D71 = Zero-one variable, equals 1 in 1971, 0 otherwise, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise. mpact of technology on cotton yields in the Punjab region of Pakistan was very strong. lgnificant relationships were found between yield and economic or climatic variables. fo-one dummy variable for 1983 was added to the equation to capture the effect of 'unseasonal rainfall which prohibited timely spraying of insecticides and led to severe lamage. The estimated equation explaining cotton yields in the Sind region of Pakistan is in equation 4.12. AKS = 0.00002 DFCI‘PPAK - 0.0005 CTHAPAKS + 0.000011RRPAK + 0.03 D72 - 0.12 D78 - 0.11 D83 (4.12) (2-69) (-13.83) (6.08) (3.41) (-5.46) (4.54) \RED (CORR): 0.91 SEE: 0.00613 DW: 2.01 PERIOD OF FIT: 1965-1985 CI'YDPAKS = Cotton yield, Sind region of Pakistan (m.tons / hectare), DFCI'PPAK = Deflated cotton price, Pakistan, IRRPAK = Number of tubewells constructed in the Sind region of Pakistan, CI'HAPAKS = Cotton area, Sind region of Pakistan, TIME = Time variable, D72 = Zero-one variable, equals 1 in 1972, 0 otherwise, D78 = Zero-one variable, equals 1 in 1978, 0 otherwise, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise. ariables found to best explain variability of cotton yields in the Sind region of in were the deflated cotton price in Pakistan, the area planted with cotton in the egion, a variable for the number of tubewells in the Sind region, and zero-one (variables for the years 1972, 1978 and 1983. This equation was one of the few in the price of cotton was found to be statistically significant. The elasticity of yield with respect to vrict irrigation variable, D “this was consistent tubewells has been captured yield incrr lhe estimat mourn = 054 + on (1m RSOUARED (CORR): . Where: CIYDTUR = CIHAIUR = D65 = Cotton yields in T and a zero-one dr strong upward to function of the at land of lower qua The CStimater gm“ in equatior (~11 Where: CI'YDUSI : Clli/tUS1 no , D82 RSPUSI TSMUSI IFLUsr Variation in cm bl Weather Vari 97 respect to price was estimated to be 0.01 at the mean yield and price levels. The ation variable, measured by the number of tubewells constructed, was very significant. was consistent with the findings of Hamid La]. However, since the number of wells has been increasing over the last 20 years this variable may, in addition, have rred yield increases from technological change. The estimated equation explaining cotton yields in Turkey is given in equation 4.13. TUR = 054 + 0.81 CI'YDTUR(-1) - 0.0006 CI'HATUR - 0.06 D65 (4.13) (17.27) (.705) (-1.98) JARED (CORR): 0.95 SEE: 0.0165 DW: 2.23 PERIOD OF m: 1964-1988 Cl‘YDTUR = Cotton yield, Turkey (m.tons / hectare), CI'HATUR = Cotton area, Turkey, D65 = Zero-one variable, equals 1 in 1965, 0 otherwise. tn yields in Turkey were explained by yield lagged one year, area of cotton planted, l zero-one dummy variable for 1965. The lagged dependent variable captured the g upward trend in yields throughout the regression period. Yields were also a .on of the area planted that captured the decline in yield as cotton was planted on )f lower quality. 16 estimated equation explaining cotton yields in the United States, Delta region is in equation 4.14. [81 = 8.73 - 0.00032 CI‘HAUSI - 0.24 RSPUSI - 3.13 'ISMUSI + 1.95 musr - 0.30 D67 + 020 D82 (4.14) (-3.17) (4.27) (-3.18) (4.18) (-272) (2.21) km (CORR): 0.81 SEE: 0.12236 DW: 1.68 PERIOD OF FIT: 1964-1988 CTYDUSl = Cotton yield, Delta region of United States (m.tons / hectare), CI'HAUSI = Cotton area, Delta region of United States, D67 = Zero-one variable, equals 1 in 1967, 0 otherwise, D82 = Zero-one variable, equals 1 in 1982, 0 otherwise, RSPUSI = Spring rainfall (inches), Delta region of United States, TSMUSl = Summer temperature (degrees C), Delta region of United States, TFLUSI = Fall temperature (degrees C), Delta region of United States. ion in cotton yields in the Delta region of the United States was explained mainly tther variables, as well as by planted cotton area and zero-one dummy variables for 1967 and 1982. No to be significant. 1‘ reduced yield, while with knowledge of t Hazera (1983). The estirna region is given in ( ovnusz = 178-11000 (1.68) nonunion (CORR): Where: CIYDUSZ .-. . CI'HAUS2 = D67 = ' 082 = ISMUSl TFLUSZ The selected equ; for the Spring (a cotton area was a sign, The estir region is given i “mm = ~6353 . t WARE) (Cont CIHAU83 D70 D83 TStruts THUsa SMUS3 98 )67 and 1982. No economic factors, such as the price of cotton or fertilizer, were found ) be significant. In the Delta region, excess spring rainfall and summer temperatures :duced yield, while high temperatures in the fall improved yields. This result is consistent lth knowledge of growing conditions for cotton and the results reported by Starbird and azera (1983). The estimated equation explaining cotton yields in the United States Southeast gion is given in equation 4.15. YDUSZ = 1.78 - 0.00016 CI'HAUSZ - 0.03 TSMUSZ + 0.026 musz + 0.26 D82 - 0.18 D67 (4.15) (-1.68) (-255) (3.48) (2.84) (-172) iQUARED (CORR): 0.67 SEE- 0.1319 DW: 2.02 PERIOD OF FIT: 1964-1988 ere: CI'YDUSZ = Cotton yield, Southeast region of United States (m.tons / hectare), CI'HAUSZ = Cotton area, Southeast region of United States, D67 = Zero-one variable, equals 1 in 1967, 0 otherwise, D82 = Zeno-one variable, equals 1 in 1982, 0 otherwise, TSMUSZ = Summer temperature (degrees C), Southeast region of United States, TFLUSZ = Fall temperature (degrees C), Southeast region of United States. 6 selected equation contained the same variables as the Delta region equation, except the spring rainfall variable which was found not to be significant. The variable for ton area was also not significant but was kept in the equation because it had the correct The estimated equation explaining cotton yields in the United States, Southwest ion is given in equation 4.16. Duss = - 63.53 - 0.000006 CI‘I-IAUS3 + 158 'I‘SMUS3 + 0.04 SMUS3 + 0.02 muss - 0.01D70 - 0.12 D83 (4.16) (940) (3.14) (3.19) (453) (.509) (-4.81) “JARED (CORR): 0.85 SEE- 0.01225 DW: 1.40 PERIOD OF FIT: 1964-1988 PC: CI'YDUS3 = Cotton yield, Southwest region of United States (m.tons / hectare), CI'HAUS3 = Cotton area, Southwest region of United States, D70 = Zero-one variable, equals 1 in 1970, 0 otherwise, D83 = Zero-one variable, equals 1 in 1983, 0 Otherwise, TSMUS3 = Summer temperature (degrees C), Southwest region of United States, TFLUS3 = Fall temperature (degrees C), Southwest region of United States, SMUS3 = Soil moisture level, Southwest region United States. Cotton yields in the fall temperature, an 1970 and 1983. TI Consistent with otl explaining the mi: The estirnz region is given in CIYDUS4 = 237 . 0'02 ' (.131, RSQUARED (coax): When: CIYDUSt DH D78 TSMUS4 TIME The variability of by Summer temp. and 1978. Other gave the Wrong , subsequently (3ij 4.52 AM! The equ below. In gene: corrected R'S‘lu variable Was a d( and to estlmate 99 on yields in the Southwest region of the United States were explained by summer and emperature, an index of soil moisture, cotton area, and zero-one dummy variables for and 1983. This specification is similar to that presented by Starbird and Hazera. tistent with other United States regions, prices were found not to be important in .ining the variability of cotton yields. The estimated equation explaining cotton yields in the United States, Western 11 is given in equation 4.17. U84 = 237 - 0.02 TSMUS4 + 0.007 TIME — 0.23 D71 - 0.33 D78 (4.17) (-137) (2.63) (—2.63) (-3.84) 'ARED (CORR): 0.67 SEE: 0.12612 DW: 1.39 PERIOD OF FIT: 1964-1988 CI'YDUS4 = Cotton yield, West region of United States (m.tons / hectare), D71 = Zero-one variable, equals 1 in 1971, 0 otherwise, D78 = Zero-one variable, equals 1 in 1978, 0 otherwise, 'I‘SMUS4 = Summer temperature (degrees C), West region of United States, TIME = Time variable. 'ariability of cotton yields in the Western region of the United States was explained nmer temperatures, a time trend variable, and zero-one dummy variables for 1971 978. Other variables, such as area planted, fall temperatures and rainfall, and prices he wrong signs when entered into initial specifications of this equation and were [uently excluded. Area Equations The equations for cotton area planted are presented in equations 4.18 to 4.34 In general, the equations fit the data well with ten of the 17 equations having a 3d R-squared of 90% or greater. In many of the equations a lagged dependent t was added. This was included in order to capture adaptively-formed expectations astimate long-run elasticities. In some cases (e.g., Australia, Brazil, and Northern india) the captured a the area lagged or zero-one also inch Both pri. for cottc Price at. central. Previou RSQUA‘ where: 100 Idia) the coefficient on the lagged dependent variable was very large and obviously Iptured a trend in the data series. The estimated equation explaining cotton area in Argentina is given in equation 18. ."HAARG = 038 CI'HAARG(-l) + 5603 DFCI‘PARG(-1) - 3148 DFCGPARG(-1) - 117 D80 — 166 D85 + 1.83 TIME (4.18) (3.48) (456) (-256) (-2.13) (-3.88) (1.30) SQUARED (CORR): 0.85 SEE: 35028 H-SI‘AT: - 0.23 PERIOD OF FIT 2 1964-1988 mere: CI'HAARG = Cotton area, Argentina, DFCTPARG = Deflated cotton price, Argentina, DFCGPARG = Deflated coarse grain price, Argentina, TIME = Time variable, D80 = Zero—one variable, equals 1 in 1980, 0 otherwise, D85 = Zero-one variable, equals 1 in 1985, 0 otherwise. 3 area planted to cotton in Argentina was explained by the deflated price of cotton gged one year, the deflated price of coarse grains lagged one year, a time variable, and .‘o-one dummy variables for the years 1980 and 1985. Cotton area lagged one year was 0 included to capture adaptively-formed expectations and to provide long-run elasticities. th price variables were statistically significant. The derived short-run elasticity estimates cotton and coarse grains were, respectively, 0.87 and 0.38, when estimated at the mean :e and area. The dummy variable for 1985 captured the 22% decline in area in the tral-south region of Argentina due to bad weather and difficulties in marketing the vious year’s crop. The estimated equation explaining cotton area in Australia is given in equation 4.19. MUS = 39.11 + 0.75 CI‘HAAUS(-1)- 0.18 DFCGPAUS(-1) + 44.14 D83 + 96.42 D87 (4.19) (11.76) (-139) (3.83) (8.05) UARED (CORR): 0.98 SEE: 2275 HsrAT: 1.28 PERIOD OF FIT: 1964—1988 :: CI'HAAUS = Cotton area, Australia, DFCGPAUS = Deflated coarse grain price, Australia, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise, D37 = Zero-one variable, equals 1 in 1987, 0 otherwise. Variability deflated pr variables it an initials was signifi mean pric which can Tl CI'HABRA R-SQUARE Where C D D The are: and zerc variable period. followip crops n Variablc in the South PiCVer 101 iriability in the area placed under cotton in Australia was explained by variability in the rflated price of coarse grains, the cotton area lagged one period and zero-one dummy riables for 1983 and 1987. The price of cotton was incorrectly signed when included in initial specification of the equation. However, the coefficient on the coarse grain price .s significant and negative and provided an elasticity estimate of 0.35 when calculated at :an price and area. The coefficient on the lagged area variable was very large (0.75) .ich captured a strong trend in the cotton area series. The estimated equation explaining cotton area in Brazil' is given in equation 4.20. {ABRA = 633.4 + 0.71 CI'HABRA(-1) + 3965 D68 + 401.7 D84 (4.20) (6.32) (3.11) (3.07) QUARED (CORR): 0.76 SE: 310198 H-SI‘AT: 1.77 PERIOD OF m: 1964-1988 :re: CI'HABRA = Cotton area, Brazil, D68 = Zero-one variable, equals 1 in 1968, 0 otherwise, D84 = Zero-one variable, equals 1 in 1984, 0 otherwise. 3 area planted to cotton in Brazil was explained by cotton area planted lagged one year lzero-one dummy variables for 1968 and 1984. The large coefficient on the lagged area able was evidence of the strong upward trend in cotton area throughout the regression iod. The 1984 dummy variable captured the sudden substitution of cotton for soybeans )wing changes in relative domestic producer prices. The prices of cotton and competing S were found not to be significant. The reason for the poor performance of the price ables may be that Brazilian producers face domestic prices which are not determined re international market. Also, there are two distinct growing regions in Brazil (i.e., h and Northeast) with different characteristics and growing conditions. Lack of data :nted these regions from being separated in the model. The estimated equation explaining cotton area in the Central Africa region is given in equatior The area deflated r for 1980 gave an . oi fettih requirer variable enterpr CUiACH R\SQUA Where: The a 00001 “One 102 luation 4.21. .CAF = 1261 + 0.70 CI'HACAF(-1) + 315.4 DFCI'PCAF(-1)/DFFNPCAF(—l) - 2655 LN TIME (4.21) (1039) (3.84) (4.37) - 370.9 D80 + 356.1 D88 (.33) (3.04) JARED (CORR): 0.93 SEE' 215536 H-SI‘AT: — 1.48 PERIOD OF FIT: 1964-1988 CI'HACAF = Cotton area, Central Africa, DFCI'PCAF = Deflated cotton price, Central Africa, DFFNPCAF = Deflated fertilizer price, Central Africa, TIME = Time variable, D80 = Zero-one variable, equals 1 in 1980, 0 otherwise, D88 = Zero-one variable, equals 1 in 1988, 0 otherwise, LN = Indicates variable transformed into logarithms. area of cotton grown in Central Africa was explained by area lagged one period, ted cotton and fertilizer prices, a time trend variable, and zero-one dummy variables 980 and 1988. The cotton price variable was found to be statistically significant and an elasticity estimate of 0.12 when calculated at the mean price and area. The price rtilizer was also added to the specification to account for the fact that fertilizer rements are higher for cotton than for most other competing crops. The time trend >le was included to capture the trend away from cotton into other more profitable )rises. The estimated equation explaining cotton area in China is given in equation 4.22. HI = 470553 + 26.21 DFCI‘PCHI(-1) + 1376.7 D84 -1087.0 D86 (4.22) (2.63) (3.28) (-2.67) .RED (CORR): 0.82 SEE 383.0 DW: 2.32 PERIOD OF FIT: 1977-1988 CI'HACHI = Cotton area, China, DFCI‘PCHI = Deflated cotton price, China, D84 = Zero-one variable, equals 1 in 1984, 0 otherwise, D86 = Zero-one variable, equals 1 in 1986, 0 otherwise. ea equation for China was estimated for the period 1977-1988, since prior to 1977 area was largely a policy decision of the government and not dependent on tic factors. A domestic Chinese cotton price lagged one year was found to be statistica in additi and 1981 increase Even ti respons 1986 w: season of the R500; Where: Cottr zero- estin QXpl; sign intc 103 tically significant and provided an elasticity estimate of 0.11 at mean price and area. dition to price, zero-one dummy variables were added to the specification for 1984 986. The dummy variable for 1984 captured the change in policy which substantially ased producer prices, coupled with significantly more generous fertilizer allocations. though the domestic producer price was included in the equation, the actual area nse in 1984 was in excess of that predicted by the model. The dummy variable for was included to account for the fall in area resulting from the experience of the 1985 11 when income was reduced by weather-induced poor cotton quality for roughly 60% growing area. The estimated equation explaining cotton area in Egypt is given in equation 4.23. EGY = 1276 - 255.1 LN TIME - 188.2 D64 - 100.3 D68 (4.23) (47.02) (-5.06) (-2.96) ARED (CORR): 0.93 SEE: 21531 DW: 1.86 PERIOD OF FIT: 1964-1988 CI'HAEGY = Cotton area, Egypt, TIME == Time variable, D64 = Zero-one variable, equals 1 in 1964, 0 otherwise, D68 = Zero-one variable, equals 1 in 1968, 0 otherwise, LN = Indicates variable transformed into logarithms. n area in Egypt was explained by a time variable transformed into logarithms and ne dummy variables for 1964 and 1968. The decline in cotton area over the .tion period was reflected in the fact that over 90% of the variation of area was red by the trend variable. Economic variables were found not to be statistically :ant in initial specifications of the equation. A reason for this lack of success may ,t, as explained in the description of the cotton yield equation for Egypt, the ctional price used in the equation was for medium staple cotton, while most of Egyptiar India at RSQUAI Where RSOUA Where: RSQU. Where: Cottr COiiC Wer i it“ f001m 104 tian cotton production is of extra-long staples. The estimated equations explaining cotton area in Northern, Southern and Western dia are given in equations 4.24, 4.25 and 4.26, respectively. NDN = 0 95 CI‘HAINDN(-l) + 13.86 DFCI‘PINDN(-1) + 1174 D83- 225 D85 (4.24) (23.14) (1. 95) (2. 66) (4 99) UARED (CORR): 0.94 SEE.- 29026 H-srAT: 0.46 PERIOD OF FIT: 1966-1986 ere: CI'HAINDN = Cotton area, Northern India, DFCI'PINDN = Deflated cotton price, Northern India, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise, ‘ D85 = Zero-one variable, equals 1 in 1985, 0 otherwise. HAINDS = 1504 + 44.65 DFCI'PINDS(-1) - 236.9 D73 + 233.2 D76 (4.25) (1.92) (-3.08) (8.76) iQUARED (CORR): 0.87 SEE: 87647 DW: 1.65 PERIOD OF FIT: 1966-1986 ere: CI'HAINDS = Cotton area, Southern India, DFCI'PINDS = Deflated cotton price, Southern India, D73 = Zero-one variable, equals 1 in 1973, 0 otherwise, D76 = Zero-one variable, equals 1 in 1976, 0 otherwise. IAINDW= 6231 + 134.1 DFCI'PINDW(-1)- 5103 LN TIME- 497.1 D74 (4.26) 97) (40 27) (4 26) JUARED (CORR): 0.91 SEE: 95789 DW: 2.03 PERIOD OF FIT: 1966-1986 re: CI'HAINDW = Cotton area, Western India, DFCI‘PINDW = Deflated cotton price, Western India, TIME = Time variable, D74 = Zero-one variable, equals 1 in 1974, 0 otherwise, LN = Indicates variable transformed into logarithms. ton area in the three regions of India was Specified to depend on the local price of on, zero-one dummy variables and variables capturing a strong trend in the area data , lagged area in the Northern India equation and the logarithm of time in the Western 3 equation). Producer prices in regional markets were used in the equations and these ormed well, giving elasticity estimates ranging between 0.07 and 0.17. The prices of peting crops (i.e., corn in the North, sorghum in the South, and millet in the West) : included but no statistically significant relationships were found. The dummy variable: CDIIIIIIOI RSQUAI Where: The er logarit 1975. of the elastic Paki asn When RSI Whr 105 tbles included were most likely capturing the effect of the prices of competing nodities, as well as changes in policy in certain years. The estimated equation explaining cotton area in Mexico is given in equation 4.27. MEx = 1146 - 255.1 LN TIME + 0.05 DFchMEX(-1) + 131.1 D68 - 214.1 D75 (4.27) (—16.53) (4.02) (2.45) (4.03) JARED (CORR): 0.94 SEE- 52065 DW: 151 PERIOD OF FIT: 1964-1988 CI'HAMEX = Cotton area, Mexico, DFCI‘PMEX = Deflated cotton price, Mexico, TIME = Time variable, D68 = Zero-one variable, equals 1 in 1968, 0 otherwise, I D75 = Zero-one variable, equals 1 in 1975, 0 otherwise, LN = Indicates variable transformed into logarithms. equation explaining cotton area in Mexico contained a time variable (converted into ithms), the deflated price of cotton, and zero-one variables for the years 1968 and The trend variable captured the decline in cotton area in Mexico throughout much : regression period. The cotton price was statistically significant and provided an :ity estimate of 0.56 at the mean area and price. The estimated equation explaining cotton area in the Punjab and Sind regions of an are reported in equations 4.28 and 4.29, respectively. m = 0. 957 CI'HAPAKP(- 1) + 0. 070 DFCI‘PAKP( 1) + 163.6 D71 270. 1 D75 276. 8 D76 (4.28) (31. 41) (2. 94) (2.68) (4. 33) (4. 64) .RED (CORR): 0.90 SEE: 50293 H-STAT: 0.27 PERIOD OF FIT: 1965-1985 CI'HAPAKP = Cotton area, Punjab region Pakistan, DFCI‘PAKP = Deflated cotton price, Punjab region Pakistan, D71 = Zero-one variable, equals 1 in 1971, 0 otherwise, D75 = Zero-one variable, equals 1 in 1975, 0 otherwise, D76 = Zero-one variable, equals 1 in 1976, 0 otherwise. LKS = 41.48 + 0.94 CI‘HAPAKS(-1) - 210.8 D76 - 159.7 D77 (4.29) (11.84) (-6.53) (450) RED (CORR): 0.91 SEE: 15664 H-SI‘AT: - 1.45 PERIOD OF FIT: 1965-1985 CI'HAPAKS = Cotton area, Sind region Pakistan, D76 = Zero-one variable, equals 1 in 1976, 0 otherwise, 137’ = Zero-one variable, equals 1 in 197], 0 otherwise. 106 fotton area in the Punjab and Sind regions trended strongly throughout the regression eriod. This trend was captured by cotton area lagged one year. The international price :‘cotton was statistically significant in the equation for the Punjab region. Prices for Mon and competing crops were not available on a regional basis for Pakistan. Use of gional price data may have improved the results considerably and their omission could count for the significance of a number of zero-one variables included in the equations. The estimated equation explaining cotton area in is given in equation 4.30. HATUR = 686 + 0.01 DFCI‘PI‘UR(-1) - 80.74 LN TIME - 102 D70 (4.30) (353) (-339) (-1.84) DUARED (CORR): 056 SEE: 54831 DW: 1.83 PERIOD OF FIT: 19641988 ere: CI'HATUR = Cotton area, Turkey, DFCI‘PTUR = Deflated cotton price, Turkey, TIME = Time variable, D70 = Zero-one variable, equals 1 in 1970, 0 otherwise, LN = Indicates variable transformed into logarithms. riability in cotton yields in Turkey was explained by the deflated price of cotton lagged 3 year, a logarithmic time variable, and a zero-one dummy variable for 1970. The time iable captured the trend away from cotton over the regression period. The deflated ton price was significant and provided a supply elasticity of 0.33 when calculated at the in of the price and area series. The estimated equations explaining cotton area in the Delta, Southeast, Southwest West regions of the United States are given in the equations 4.31, 4.32, 4.33 and 4.34, rectively. \U81 = 1426 + 220 DFCI'PUSl(-1) - 0.87 DFFNPUSA(-1) - 18.8 TIME - 396 SKRWUS1 + 4395 D72 - 437.9 D83 (4.31) (3.02) (-2.48) (-394) (-394) (2.96) (-3.00) UARED (CORR): 0.81 SEE: 350928 DW: 1.02 PERIOD OF FIT: 19641988 CI'HAUSI = Cotton area, Delta region, United States, DFCI'PUSI = Deflated cotton price, Delta region, United States, DFFNPUSA = Deflated fertilizer price, United States, TIME = Time variable, ClllAUSE RSQUM Where: RSQUI Where: Moe Where: The each row the the sea at 107 SKRWUSl =Zero-one variable for Skip-Row policy, equals 1 1966-67, otherwise 0, D72 = Zeno-one variable, equals 1 in 1972, 0 otherwise, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise. .USZ = 4795 + 105.3 DFCI'PUSZ(-1) - 051 DFFNPUSA(-1)- 11.1 TIMB- 244 SKRWUSZ + 0. 621CI'HAUSZ(-1) (4.32) 2 60 5) . (-331) (-3. 86) ( 5. 09) (6.1 - 0.67 DFSBPUSA(—1) ) (-2.88 JARED (CORR): 0.94 SEE: 58595 H-SI‘AT: 0.38 PERIOD OF FIT: 1964-1988 CIT-IAUSZ = Cotton area, Southeast region, United States, DFCI'PUSZ = Deflated cotton price, Southeast region, United States, DFSBPUSA = Deflated soybean price, Southeast region, United States, DFFNPUSA = Deflated fertilizer price, United States, TIME = Time variable, SKRWUSZ = Zero-one variable for Skip-Row policy, equals 1 1966-67, otherwise 0, JS3= 596 + 417 DFCI‘PUS3(-1)- 617 SKRWUS3 + 055 88CI'HAU-S3( -1-) 89 DFSGPUSA(-1)- 650 (D82+D83) (4 33) (215) ( 3 09) (488 (4)59) (3 3) ARED (CORR): 0.76 SEE 1280438 H-SI‘AT: 0.70 PERIOD OF FIT: 1964-1988 CI'HAUS3 = Cotton area, Southwest region, United States, DFCI'PUS3 = Deflated cotton price, Southwest region, United States, DFSGPUSA = Deflated sorghum price, United States, SKRWUS3 = Zero-one variable for Skipr policy, equals 1 1966-67, otherwise 0. D82 = Zero-one variable, equals 1 in 1982, 0 otherwise, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise. r84 = 175.4 DFCI'PUS4(-1) + 10. 01 TIME + 043 CI'HAUS4( -.1) 16.88 DFRIPUSA(-1)- 223.4 D83 (4.34) (4.0 2) (3.32) (358) (-1.39) (-2- 68) IRED (CORR): 0.83 SEE: 124078 H-STAT: 2.12 PERIOD OF FIT: 1964-1988 CI'HAUS4 = Cotton area, West region, United States, DFCI'PUS4 = Deflated cotton price, West region, United States, DFRIPUSA = Deflated rice price, United States, TIME = Time variable, D83 = Zero-one variable, equals 1 in 1983, 0 otherwise. {nations for the United States were estimated as functions of the price of cotton in igion, the price of a competitive crop, a policy variable to account for different skip- les, and a lagged dependent variable. A dummy variable for 1983 was included in rations for the Delta, Southwest and West regions of the United States to capture Inge in farm policy as a result of large stock levels in 1982. For the 1983 crop the acreage reduction and PIK programs resulted in more than 7 million acres being lut of cotton production. Since the fertilizer requirements were higher for cotton 108 r for most other competing crops, the price of fertilizer was used in the equation for Delta region. Cotton Area Elasticigg Estimates The elasticities of planted cotton area with respect to price are presented in Table All elasticities are inelastic in the short-run, ranging from 0.07 for Northern India to for Argentina. In general, the elasticities are larger for the higher-income countries ‘3 may reflect greater flexibility and choice in producing alternative mom to cotton. : 4.1 Elasticities of Cotton Area Planted with respect to Price. '— m Short-run Long-run Elasticity Elasticity Itina 0.87 1.40 a] Africa 0.12 0.40 t 0.11 North 0.07 South 0.17 West 0.09 0 0.56 an Punjab 0.08 y 0.33 ioutheast 0.27 )uthwest 0.36 0.95 elta 0.27 0.60 'est 0.41 0.72 ttes of Mues & Simmons (1988) I States 0.48 0.64 lia 0.59 2.46 f World 0.06 0.25 109 Of the studies reviewed in section 4.3 only Mues and Simmons reported elasticities. ese are shown in Table 4.1 also, and were similar in magnitude to the ones obtained in 8 study. For the United States the short-run elasticities were slightly smaller than those Mues and Simmons but the long-run ones were larger. The long-run elasticity estimate Australia (2.46) seems very high and resulted from the large coefficient on the lagged )endent variable in the area equation. Summary In this section the method used to explain the production of cotton in terms of nometrically-estimated equations was presented. Based on theory of individual firm avior and drawing on earlier models of cotton supply, production was derived from the duct of yield and area planted. Both of these supply components were estimated in a Irate equation for each region of the model. The econometric equation results were :ented along with supply elasticity estimates. Overall, the econometric results were :factory although in many cases world price variables for cotton, competing crops and ts were not significant. However, when local prices were used in some equation the lts improved significantly. 5. Cotton Price Determination .1 Introduction In an earlier version of the model, cotton demand, production and stock equations cere estimated for each region and combined in an identity equating production and eginning stocks with consumption and ending stocks. When simulating the model, this lentity solved for consumption in one region and price solved as a right-hand side ndogenous variable in the demand function for this region. This formulation did not erform satisfactorily in that forecasted values of the price did not track the actual value ell, even though consumption and production forecasts performed adequately. The reason for this result was twofold. First, the estimates of stock equations were Isatisfactory; only in a few equations were the price variables significant. This may be cause in many regions (e.g., China) stock levels were more the residual between oduction and consumption, than the result of profit-maximizing behavior of stockholders. cond, as seen from Table 3.1 the elasticities of demand for cotton were highly inelastic. rerefore, small errors in quantity variables tended to lead to relatively larger errors in the ice variables. These problems were discussed by Ghosh QLEL (1987) who noted that most studies rploy inverted stock demand functions where price is specified as a function of stocks. ey noted that otherwise, in forecasting or model simulation the price is forced to move too much in order to clear the market. This is a consequence of the incompleteness and inaccuracy of stock data which result in poorly fitting demand equations. 110 111 a result of the poor results obtained using the market-clearing approach, a pricing ation was estimated as an inverted world stock demand function. Theoretical Issues Demand for stocks comes from both producing and consuming countries and ves from two separate motives. The first is from a transactions demand for which ts are held to ensure that unanticipated changes in demand can be met. The second :rived from a speculative demand. Stocks are held if prices are expected to increase e future in excess of the cost of storage. Suppose stock levels are determined by current prices. That is, St = s(P,). :h et al. discuss two simple models in which price dynamics can be introduced. First, an assume that stocks follow a partial adjustment process in which stock levels adjust year towards a certain desired stock level 8'. The S' is given by, S' = a0 - alPt (5.1) St - 8,1 = (1 - A )(S't - SH), and (O< A <1). (5.2) tuting the partial adjustment equation 5.2 into the stocks equation and inverting gives, P, = b, - 13,5, + b,s,_,. (5.3) 112 )sh et a1. noted that almost all researchers find a lagged price term necessary in ation 5.3 and rarely is the coefficient on the SP1 term significantly different from zero. [6 S,I term in equation 5.3 is substituted j times using equation 5.3, the price equation Iven by, Pl = C0 - clS, - czziA 3PH. (5.4) equation is not useful for estimation because it implies an infinite series of past prices a negative relationship between current and past prices. Therefore investigators have [fled equations in which prices adjust to stock levels (instead of stocks adjusting to :s). This gives an estimable equation of the form, P[ = bo + blP,_1 - b,s,, 1 has been applied widely. Researchers often find the coefficient on lagged price to )se to one, and that the coefficient on stocks is not significantly different from zero. is case, the equation provides forecasts that are too smooth and unresponsive to :es in quantity variables. The second method of incorporating dynamics into the pricing equation relates levels to an expected price. This approach was developed by Hwa (1981, 1985) who :d a stocks equation given by, St = a0 + alCl + a,(ln P2,”, - In P,) - a31n rl (5.5) 113 are C‘ is consumption, pi“), is the expected price in time t for the period HI and rt is rate of interest. Assuming that prices adjust to the market clearing value, equation 5.5 be inverted to give, aln P, = b, + b,C, + b,(ln 131,, ,, - ln 1),) - b3r, - b,S,. (5.6) difficulty with empirically estimating equation 5.6 is to formulate an expectations hanism for the (ln P1,”, - In P!) variable. Gilbert and Palaskas (1989) argued that instead of reacting to expected future :8, stockholders respond to expected future excess supplies, which are conditioned on ually-formed price expectations. In their pricing equation, expected future supplies used as one of the regressors. This variable is obtained by estimating the entire 31 (excluding the price equation) with price set at its mean value. That is, a variable, ESHl/t = (Qt+1/I(P) " Ct+l/t(P))/ Q 'e P and Q are mean values of price and quantity) is added to the price equation 161', with lagged price and interest rates. Leview of Literature Most of the econometric studies of the cotton sector discussed in previous sections lsed single equation models where prices have been exogenous. Only in the studies res and Simmons (1988) and Agbadi (1988) were the models closed with price 114 :ndogenous. In the study by Mues and Simmons, a market clearing identity was used to olve for price and- no price equation was estimated. Agbadi estimated the mill cotton rice for the United States as an inverted mill demand function using domestic mill ansumption of cotton and manufactured fibers as regressors. The CPI for textile products as estimated as an inverted consumer demand function, with price explained by consumer )nsumption and income. The ICAC (1988) estimated a single-equation regression model of the price of >tton in order to forecast near-term price movements. Price was regressed on net exports ' China (expressed as a percentage of non-Chinese world consumption) and a ratio of )cks held outside China to use outside China (i.e., world ending stocks net of China’s mics and trade, divided by world consumption net of China’s consumption). The rationale hind this specification was that, cotton prices are clearly related to the actual or perceived tightness of supply. Perhaps the best single indicator of this tightness is a ratio of available stocks to use. In recent years the size of Chinese stocks and the fact that a large pr0portion of those stocks were isolated from world markets have made it necessary to look at world stocks and consumption net of China. e equation fitted that data very well for the period 1974 to 1986 and provided accurate ecasts of the 1987 and 1988 prices. A Model of Cotton Price Determination and Estimation Results The pricing equation used in this model was based on the ICAC model. The price ation and identities are presented in equation 5.7. .TPWOR = 7.24 - 0.78 LN CI'ESWORXCHI + 0.92 LN MUV + 0.31 LN CTCONWORXCHI - 0.86 LN TIME (5.7) (-6.32) (5.12) (1.75) (.333) 'UARED (CORR): 0.94 SEE: 0.03 DW: 2.09 PERIOD OF FIT: 1964-88 1 15 re: CI'PWOR = World cotton price (Outlook Index 'A'), CI‘ESWORXCHI == World cotton stocks excluding stock held in China (also see equation 5.8), MUV = Manufacturing unit value], CI‘CONWORXCHI = World cotton consumption excluding consumption of China (also see equation 5.9), TIME = Time variable, LN = Indicates variable transformed into logarithms. rld price was Specified as a function of ending world stocks net of China’s stocks, a eral deflator (MUV)’, world cotton consumption net of China’s cotton consumption, and me trend variable. All variables were converted into logarithms which fit data the er than in the linear formz. The coefficient on the stocks variable was highly significant correctly signed. The flexibility of price with respect to stocks, given by the coefficient the stocks variable, was estimated to be -0.78. No other study has reported a esponding elasticity for comparison. In order to capture the transactions demand for m stocks, world consumption, net of China’s consumption, and time were added to the ification. Consumption was used based on the assumption that a certain fixed ortion of the quantity consumed is held each period to cover unanticipated changes emand. Time was added to account for these stocks having declined over the ration period. This is because improved milling technology and marketing and portation channels have meant that less stocks are required for unforeseen demand 1ations. Since prices were estimated in levels, the MUV deflatorl was added to the fication to capture the effect of inflation. lb— US dollar terms of manufactures exported from the G-5 countries (France, Germany, Japan, United Kingdom and weighted proportionally to the countries’ exports to the developing countries. : approach is to estimate the equation using a semi-logarithm functional form. This would have the benefit of allowing to change over different price and quantity levels (e.g., the price flexibility may be higher at low stocks-to-use levels). approach was not taken in this study. 116 The stocks and consumption variables used in the pricing equation 5.7 were .btained from the identities shown in equations 5.8 - 5.13. TBSWORXCHI = CI‘ESWOR - CI'ESCHI (5.8) TCONWORXCHI = Cl'CONWOR - CI‘CONCHI (5.9) I'ESWOR = CI‘PDWOR + CI‘ESWOR(-1) - CI‘CONWOR (5.10) I‘ESCHI = CI'PDCHI + CI'ESCHI(-1) - CI‘CONCHI - CINECHI (5.11) I‘PDWOR = 2,.(CIHAfCI'YDi) + CI'PDROW (5.12) I‘CONWOR = z,(CI’SH,‘TFUi) + CI‘CONROW (5.13) here: CI'ESWORXCHI = World cotton stocks excluding stock held in China, CI‘ESWOR = World cotton stocks, CI'ESCHI = China’s cotton stocks, CI‘CONWORXCHI = World cotton consumption excluding consumption of China, CI‘CONWOR = World cotton consumption, CI‘CONCHI = China’s cotton consumption, CI'PDWOR = World cotton production, CI'PDCHI = China’s cotton production, CINECHI = China’s net exports of cotton, CIT-IA, = Cotton area in country i, D, = Cotton yield in country i, CI'PDROW = Rest-of—the-World cotton production (production by countries not explicitly modeled, i.e., exogenous), CTSHI. = Cotton share of total fiber use in country i, TFU, = Total fiber use in country i, CT CONROW = Rest-of-the-World cotton consumption (consumption by countries not explicitly modeled, i.e., exogenous). World stocks were derived from a world market clearing identity (equation 5.10) :l the stocks of China were derived from China’s market clearing identity (equation 5.11). r China’s market clearing identity, net exports were estimated (equation 5.14) as a dim of lagged net exports and a variable capturing excess supplies. Price variables re found not to be significant in determining net exports. ECHI = - 326 + 0.23 ’ (CI'PDCHI + CI‘ESCHI(-1)- CI‘CONCHI) + 0.34 CI‘NECHI(-1)- 528.4 D84 (5.14) (5.78 (2.70) (-2.80) )UARED (CORR): 0.92 SEE: 142.9 DW: 1.80 PERIOD OF FIT: 1974~88 re: CINECHI = China’s net exports of cotton, CI'PDCHI = China’s cotton production, CI'ESCHI = China’s cotton stocks, CI‘ CONCHI = China’s cotton consumption, D84 = Zero-one variable, equals 1 in 1984, 0 otherwise. 117 Price Linkage Equations In the supply equations for the four regions of the United States, for the three ms of India and for China, local prices (in terms of local currency) were used instead vorld prices adjusted for exchange rates. These were the prices most relevant to rers’ decision making and this was borne out by the estimation results. These local es were linked to the world price of cotton in order to make these regions responsive 'orld market conditions. In the case of the United States, price linkage equations were estimated linking '. and world cotton prices. These are presented in equations 5.15 - 5.18. [‘PUSI = 0.25 + 0.92 LN CI‘PWOR - 0.31 D74 (5.15) (15.39) (-2.38) JARED (CORR): 0.92 SEE: 0.34 DW: 1.81 PERIOD OF FIT: 1964-1987 CI'PUSI = Cotton price in the Delta region of the United States, CI'PWOR = World cotton price, D74 = Zero-one variable, equals 1 in 1974, otherwise 0, LN == Indicates variable transformed into logarithms. PUSZ = 0.34 + 0.91 LN CI'PWOR - 0.30 D74 (5.16) (15.10) (-2.32) IARED (CORR): 0.92 SEE: 0.34 DW: 1.85 PERIOD OF FIT: 1964-1987 CI'PUSZ = Cotton price in the Southeast region of the United States, CI'PWOR = World cotton price, D74 = Zero-one variable, equals 1 in 1974, otherwise 0, LN = Indicates variable transformed into logarithms. ’US3 = 0.32 + 0.90 LN CI'PWOR - 0.31 D74 (5.17) (15.05) (-2.38) t\RED (CORR): 0.92 SEE: 0.34 DW: 1.83 PERIOD OF FIT: 1964-1987 CI'PUS3 = Cotton price in the Southwest region of the United States, CTPWOR = World cotton price, D74 = Zero-one variable, equals 1 in 1974, otherwise 0, LN = Indicates variable transformed into logarithms. 118 .‘I‘PUS4 = 0.09 + 0.96 LN CI'PWOR - 0.32 D74 (513) (13.74) (—2.11) lUARED (CORR): 0.90 SEE: 0.46 DW: 1.79 PERIOD OF FIT: 1964-1987 re: CI'PUS4 = Cotton price in the West region of the United States, CI'PWOR = World cotton price, D74 = Zero-one variable, equals 1 in 1974, otherwise 0, LN = Indicates variable transformed into logarithms. Iouble-log functional form was used in all cases and a dummy variable for 1974 was ed to account for very low prices in that year. In 1974 target price policy was oduced, with direct payments made to growers if market prices fell below the target :e. The transmission elasticities ranged from 0.90 in the Southwest region to 0.96 in the st region. Price transmission equations were also estimated linking local prices in India and na with world prices. These equations performed badly, with world price not significant. ; finding was consistent with price policy in these countries which is formed in isolation vorld markets. However, it is unlikely that over the long-run world prices can be red by policy-makers in these countries. Rather than have administered prices genous in the forecast period these prices were linked to the world price assuming an icity of 0.25 for China and 0.50 for India’. These elasticities entered the model using dentity shown in equation 5.19. *P, = LN LPPo + x ' LN(WPP, / wrro) (5.19) LPP, = local producer price in period t, LPPo = local producer price in period 0, WPP, = world producer price in period t, WPPo = world producer price in period 0, and x = 0.5 for India and 0.25 for China. ‘— f elasticities was based on a study by Mundlak and Larson (1990) of the relationships between local and international Summagg In this section, the equations for the determination of cotton prices were discussed. 3 equation explaining the world cotton price was specified as an inverted world stocks ration. The world price was explained by world stocks (net of stocks held in China). nsactions demand for cotton stocks was captured in the equation by a time variable and 1d consumption (net of consumption in China). The regression results were satisfactory gave a flexibility estimate of the cotton price with respect to stocks of -0.78. In the el simulations this method of price determination gave results far superior to those ined when price was derived through a world market clearing identity. This finding nsistent with Ghosh et al. AN ECONOMETRIC ANALYSIS OF THE WORLD COTTON AND NON-CELLULOSIC FIBERS MARKETS V0 lume I I By Jonathan R. Coleman A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1991 6. Non-Cellulosic Fibers Model (.1 Introduction l Elasticities of demand, supply and price equations for the non-cellulosic fibers (i.e., plyester, nylon and acrylics) are described in this section. The market for cellulosic fibers j,.e., acetate, rayon and triacetate), the other major group of manufactured fibers has not teen modeled. The reason for this was that cellulosic fibers contributed less than 10% of ital world fiber use in 1985 and less than 23% of manufactured fiber use. Further, these larcentages have decreased significantly over time and are expected to continue to fall fespite a recent resurgence of rayon consumption). A cellulosic fiber model can be added a later stage, if necessary. In previous sections the theoretical issues and relevant literature were presented d reviewed. The theoretical assumptions for non-cellulosic demand were the same as for :ton demand (section 3.3). The theoretical basis for the pricing equation is similar to t of cotton, except that prices were determined from an inverted consumer demand [ation rather than from a stocks demand equation. There are no stocks data for non- ulosic fibers. The supply equation for non-cellulosic fibers was developed from the tmption of profit maximization by manufacturers. As far as the author is aware there no published econometric studies of the non-cellulosic fiber sector upon which to draw. Demand for Non-Cellulosic Fibers The demand equations for non-cellulosic fibers was specified using a structure :ical to that for cotton use. The proportion of non-cellulosics fibers of total fiber use 120 121 as estimated as a function of the prices of cotton and polyester. The non-cellulosic share bers was then multiplied by total fiber use to obtain the consumer demand for on-cellulosic fibers. That is, NCU = NCSH * PCTFU * POP, NCSH = f( PCT, PSP ), PCTFU = f( GDP ). ere: NCU = Quantity of non-cellulosic fibers for home use; NCSH = Non-cellulosic fibers share of total fibers for home use. e non-cellulosic fibers share was not specified as one minus the cotton share because ere are other fiber types that make up total fiber demand. The per capita total fiber use quations were discussed in section 3.5.2. The non-cellulosic share equations are presented equations 6.1 - 6.14 below. The equation specifications are very similar to the cotton share equations reported section 3.5, which in most cases fitted the data well. The non-cellulosic share is a action of polyester and cotton prices1 which appears as a ratio in most of the equations. many of these equations, price parameters are significant and with the correct sign. ice variables that were not significant but correctly signed were maintained in the model ., for Argentina and Central Africa). In initial equations for Egypt, India and Pakistan price variables were not significant and were incorrectly signed and were dropped from specification. To capture the effect of prices in these equations the cotton share of 1 fiber use was added to the specifications which, as shown in section 3.5, were ctions of relative prices. lyester price used was the polyester staple price fob U.S. plants taken from Cotton and Wool Situation and Outlook, gton D.C. otton price was the cotton outlook 'A' Index cif N. Europe, taken from Cotton Outlook. 122 As reported in section 3.5.1, the non-cellulosic fibers share in these regions followed ry strong time trend throughout the estimation period. A time variable was significant any of the equations, capturing the non-price determined substitution of cotton with -cellulosic fibers. This point was discussed in section 3.4. In some cases a logarithmic tional form gave a better fit to the data. Dummy variables were used to capture ers in the data. For example, dummy variables for 1972 and 1973 in the equation for epublic of Korea captured the impact of political unrest on that economy during the I 1970s. The non-cellulosic fibers share of total fiber use in Argentina is given in equation ARG = 0.016 - 0.034 DFPSPARG/DFCI‘PARG + 0.81 NCSHARG(-1) (6.1) (-1.82) (10.84) IARED (CORR): 0.92 SEE: 0.0106 H-SI‘AT: 0.47 PERIOD OF FIT: 1964-1986 NCSHARG = Non-cellulosic fibers share of total fiber use in Argentina, DFPSPARG = Deflated polyester price in Argentina, DFCI'PARG = Deflated cotton price in Argentina. The non-cellulosic fibers share of total fiber use in Australia is given in equation .Us = 0.018 - 0.049 DFPSPAUS/DFCI'PAUS + 0.85 NCSHAUS(-1) (6.2) (.301) (15.79) RED (CORR): 0.96 SEE: 0.0087 H-STAT: — 0.86 PERIOD OF FIT: 1964-1987 NCSHAUS = Non-cellulosic fibers share of total fiber use in Australia, DFPSPAUS = Deflated polyester price in Australia, DFCTPAUS = Deflated cotton price in Australia. The non-cellulosic fibers share of total fiber use is given in equation 6.3. HERA = - 1.55 - 1.239 LN DFPSPBRA/LN DFCI'PBRA (6.3) (-1185) RED (CORR): 0.87 SEE: 0.8879 DW: 1.19 PERIOD OF FIT: 1964-1986 NCSHBRA = Non-cellulosic fibers share of total fiber use in Brazil, DFPSPBRA = Deflated polyester price in Brazil, 123 DFCIPBRA = Deflated cotton price in Brazil. LN = Indicates variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in Central Africa is given in ation 6.4. NCSHCAF = - 4.98 - 0.10 LN DFPSPCAF/LN DFCTPC‘AF — 1.03 LN TIME (6.4) (4.11) (.1166) UARED (CORR): 0.95 SEE- 0.2824 DW: 1.31 PERIOD OF FIT: 1964-1986 NCSHCAF = Non-cellulosic fibers share of total fiber use in Central Africa, DFPSPCAF = Deflated polyester price in Central Africa, DFCI'PCAF = Deflated cotton price in Central Africa, TIME = Time variable, LN = Indicates variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in China is given in equation 6.5. VCSHCHI = - 3W - 1.02 LN DFPSPCI-II/LN DFCTPCIII - 6.40 LN CISHCHI (6.5) (-9.41) (-16.42) )UARED (CORR): 0.97 SSE: 0.6521 DW: 1.85 PERIOD OF FIT: 1964-1986 re: NCSHCHI = Non-cellulosic fibers share of total fiber use in China, CI’SHCHI = Cotton share of total fiber use in China, DFPSPCHI = Deflated polyester price in China, DFCI‘PCHI = Deflated cotton price in China. LN = Indicates variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in the EEC-12 is given in equation szEEC = - 1.00 - 0.58 LN DFPSPEEC/LN DPCIPEEC + 0.38 D86 (6.6) (-1043) (3.05) [JARED (CORR): 0.85 SEE' 02.525 DW: 1.45 PERIOD OF FIT: 1964-1986 NCSHEEC = Non-cellulosic fibers share of total fiber use in the EEC-12, DFPSPEEC == Deflated polyester price in the EEC-12, DPCI’PEEC = Deflated cotton price in the EEC-12, D86 = Zero-one variable, equals 1 in 1986 0 otherwise. LN = Indicates variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in Egypt is given in equation 6.7. BGY = - 0.497 + 029 LN TIME - 0.38 CTSHEGY (6.7) (7.65) (.751) ARED (CORR): 0.99 SEE: 0.0062 DW: 2.65 PERIOD OF FIT: 1964-1986 NCSHEGY = Non-cellulosic fibers share of total fiber use in Egypt, CI'SHEGY == Cotton of total fiber use in Egypt, 124 TIME = 'I'Ime variable, LN = Indicates variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in India is given in equation 6.8. ~1CSHIND = - 3.01 + 0.73 LN TIME + 0.67 LN NCSHIND(-1) (6.8) (2.93) (5.33) IUARED (CORK): 0.98 SEE: 0.3198 H-SI‘AT: 1.12 PERIOD OF FIT: 1964-1986 re: NCSHIND = Non-cellulosic fibers share of total fiber use in India, TIME = 'I'Ime variable, LN = Indicates variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in Japan is given in equation 6.9. lCSHJPN = - 034 - 0.19 LN DFPSPJPN/LN DFCI'PJPN + 0.65 LN NCSHJPN(-1) - 0.28 D72 - 0.32 D74 (6.9) (4.19) (7.69) (-2.68) (-399) UARED (CORR): 0.92 SEE: 0.0965 H-STAT: - 0.28 PERIOD OF FIT: 1964-1987 e: NCSI-IJPN = Non—cellulosic fibers share Of total fiber use in Japan, DFPSPJPN = Deflated polyester price in Japan, DFCI'PJPN = Deflated cotton price in Japan, D72 = Zero-one variable, equals 1 in 1972, 0 otherwise, D74 = Zero-one variable, equals 1 in 1974, 0 otherwise, LN = Indimtes variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in Korea is given in equation 6.10. KOR = - 0.70 - 0.82 DFPSPKOR/DFCI'PKOR - 0.47 D72 - 0.61 D73 (6-10) (.299) (.117) (-3.76) JARED (CORR): 0.83 SEE- 05866 DW: 1.02 PERIOD OF FIT: 1964-1986 NCSHKOR = Non-cellulosic fibers share of total fiber use in Korea, DFPSPKOR = Deflated polyester price in Korea, DFCI'PKOR = Deflated cotton price in Korea, D72 = Zero-one variable, equals 1 in 1972, 0 otherwise, D73 = Zero-one variable, equals 1 in 1973, 0 otherwise. The non-cellulosic fibers share of total fiber use in Mexico is given in equation 6.11. vIEX = - 0.058 DFPSPMEX/DPCTPMEX + 0.92 NCSHMEX(-l) (6.11) (-3-60) (2734) . ARED (CORR): 0.99 SSE: 0.0093 H-SI‘AT: 1.04 PERIOD OF FIT: 1964-1986 NCSHMEX = Non-cellulosic fibers share of total fiber use in Mexico, DFPSPMEX = Deflated polyester price in Mexico, DFCTPMEX = Deflated cotton price in Mexico. 125 The non-cellulosic fibers share of total fiber use in Pakistan is given in equation l2. SHPAK = 0.64 — 0.70 CTSHPAK (6.12) (-21.81) QUARED (CORR): 0.97 SSE: 0.0001 DW: 1.91 PERIOD OF FIT: 1964-1986 ere: NCSHPAK = Non-cellulosic fibers share of total fiber use in Pakistan, CI'SHPAK = Cotton share of total fiber use in Pakistan. The non-cellulosic fibers share of total fiber use in Turkey is given in equation 6.13. NCSHTUR = - 4.75 - 0.27 LN DFPSPTUR/LN DFCI'PTUR + 1.12 LN TIME (6.13) (.278) (11.92) QUARED (CORR): 0.96 SSE: 0.1992 DW: 1.64 PERIOD OF FIT: 1964-1986 re: NCSHTUR = Non-cellulosic fibers share of total fiber use in Turkey, DFPSPTUR = Deflated polyester price in Turkey, DFCI'PTUR = Deflated cotton price in Turkey, TIME = Time variable, LN = Indicates variable transformed into logarithms. The non-cellulosic fibers share of total fiber use in the United States is given in ation 6.14. {USA = 0.17 + 0.0012 DFCI‘PUSA - 0.0005 DFPSPUSA + 0.012 TIME (6-14) (6.13) (-372) (6.27) UARED (CORR): 0.94 SSE: 0.0209 DW: 1.78 PERIOD OF FIT: 1964-1986 e: NCSHUSA = Non-cellulosic fibers share of total fiber use in the United States, DFPSPUSA = Deflated polyester price in the Unites States, DFCI‘PUSA = Deflated cotton price in the United States, TIME = Time variable. The own- and cross-price elasticities are reported in Table 6.1. They ranged from in Central Africa to 1.24 in Brazil. In most cases the elasticities were inelastic but all are higher than the own-price elasticities estimated for cotton. This may have cted the high degree of substitutability Of non-cellulosic fibers with cellulosic fibers. untrast to the cotton own-price elasticities, the responsiveness to prices does not tend » greater or less in the industrial countries than in the developing countries. Cross-price elasticities can be compared using Tables 3.1 and 6.1. The comparisons 126 able 6.1 Price Elasticities of Non-Cellulosic Fibers Use‘. egion2 Polyester Staple Price Cotton Price3 rgentina - 0.15 + 0.15 ustralia - 0.13 + 0.13 razil - 1.24 + 1.24 entral Africa - 0.10 + 0.10 hina - 1.02 + 1.02 EC-12 - 0.58 + 0.58 gypt4 - 0.42 + 0.42 Ipan - 0.19 + 0.19 orea, Repub. of - 0.83 + 0.83 [exico - 0.12 + 0.12 akistan‘ - 0.32 + 0.32 urkey - 0.27 + 0.27 'nited States - 0.22 + 0.42 lased on a regression period 1964—1986 for developing countries and 1964-1987 for tdustrialized countries. he elasticities of non-cellulosic use with respect to price (Emu/P) were derived from the )n-cellulosic share equations. (cu/P = (aNCU/aP).(P/NCU). Recall NCU = NCSH * TFU and that NCSH = (bi + .P,). (aNCU/aP) = bk.T'FU and ENG”, = b,.T'FU.P/NCU. rlues for NCU, P and TFU were taken as historical means. 10 elasticities can be reported from model regions for which prices did not appear in the mand equations (i.e., India) here polyester price elasticities equal the cotton elasticities a ratio of these prices was ad in the share equation. asticities calculated using cotton share elasticity estimates. W” " WM. - p.1- _ - .flm. 127 w that non-cellulosic demand is more responsive to cotton price changes than is cotton nand to changes in the price of polyester. This may reflect the fact that historically the Is of raw cotton in textile manufacturing have exceeded those of non-cellulosic fibersz. :refore, the income effect of a change in the cotton price is larger than for a change in polyester price. This finding was consistent with the Slutsky condition from consumer nand theory. This states that, e,,. = (w,/w,).ej, + wj.(ejy - ey) ,ere: eij = the demand elasticity of good i with respect to the price Of good j, eji = the demand elasticity of good j with respect to the price of good i, wi = the income expenditure share allocated to good i, = the income expenditure share allocated to good j, = the income elasticity of good i, = the income elasticity of good j. 1:9 <9 3 good i be cotton and j non-cellulosic fibers. Since historically manufacturers have spent :e on cotton than non-cellulosic fibers, (i.e., wi exceeded wj), then the elasticity of cotton Iand With respect to the price of polyester was expected to be less than the elasticity on-cellulosic fiber with respect to the cotton price (assuming the w,.(ejy - ey) term were 11). so for the 19705 than 19805. Another factor is that polyester prices change less frequently than cotton prices. At times '1 the United States remain constant for several months while cotton prices change daily. 128 Supplv of Non-Cellulosic Fibers The supply of non-cellulosic fibers has been described by Thigpen and Mitchell 8) as follows: Non-cellulosic fibers are produced by industrial processes from long-chain non- cellulosic polymers and are usually of petroleum origin. The levels of production can be adjusted quickly to market conditions within the limits of plant capacity. These fibers are produced by industrial processes which give producers a considerable degree of control of output over a relatively short period of time and within the limits of total capacity. The non-cellulosic fiber market is comprised of a relatively small number of firms with individual firms large enough to influence the level of production and price. The supply of non-cellulosic fibers is expected to depend on the profitability of lucing non-cellulosic fibers, which is determined by the price of non-cellulosic fibers ive to input costs, such as oil costs and interest rates. Equations were estimated for major producing regions. However, difficulties were encountered in estimating uction at a regional level with the prices of oil and polyester not significant in many ;. Poor results may have resulted from the lack of accurate regional level data. uction of non-cellulosic fibers was therefore estimated at the world level using a single tion which is reported in equation 6.15 )D = - 27116- 4397. DFOILPR- 127.6 RIRUSA + 1077. DFPSPWOR(-1) + 12211. LN TIME (6.15) (-2-77) (-253) .49) (13.18) ARED (CORR): 0.99 SEE: 488.7 DW: 1.65 PERIOD OF FIT: 1964-88 NCPROD = Non-cellulosic fibers production, world, DFOILPR = Deflated price of oil, (OPEC petroleum average prices), = Real rate of interest (long term U.S. bond yield), DFPSPWOR = Deflated polyester price, = Time variable, LN = Indicates variable transformed into logarithms. 129 m-CCIIUIOSIC production is estimated as a function of the deflated price of oil, the deflated cc of polyester and the real rate of interest. The coefficients have the right signs and : all significant. The elasticities of supply with respect to the price of oil and own-price re - 0.08 and 0.36, respectively. The logarithm of time was added to the specification capture the dramatic increase in production of non-cellulosic fibers associated with l Fnological innovations in the non-cellulosic fiber sector. i Price Determination The polyester price was determined by an inverted demand equation (6.16). nand in the Rest-of-the-World was included, and, in turn, was determined endogenously 1e difference between supply and the demand from regions modeled explicitly (equation ). The lagged price of oil captured the effects of oil prices on non-cellulosic fibers, e the MUV deflator accounted for general inflation throughout the estimation period. equations fitted the data well with all coefficients significant and correctly signed. The )ility of the polyester price with respect to the demand in the Rest-of-the-World is 0.90 h is consistent with the elasticity estimates reported in Table 6.1. The polyester price icity with respect to the MUV was 1.04. R = 79.6 + 1.74 MUV — 0.05 NCUROW + 1.81 OILPR(-1) (6.16) (6.83) (—5.66) (6.16) ARED (CORR): 0.96 SEE: 556 DW: 2.61 PERIOD OF FIT: 1969-88 ’SPWOR = Polyester price, World, VIUV = Manufactures unit value (deflator), JCUROW = Non-cellulosic fibers use, Rest-of—the-World, )ILPR = Price of oil. 130 NCUROW = NCPROD - 2,(NCSH,‘TFU,) (6.17) WherezNCUROW = Non-cellulosic fibers use, Rest-of-the—World, NCPROD = Non-cellulosic fibers production, world (from equation 6.15), NCSH, = Non-cellulosic fibers share of total fiber use in country i (from equations 6.1-6.14), TFU. = Total fiber use in country i (from equations 315-3.28). I 6.5 Summary ‘ In this section, an econometric model for the non-cellulosic fiber sector was presented. Equations for the non-cellulosic fibers share of total fiber use were estimated :"or each of the model regions, which were then combined with total fiber use to determine 3 on-cellulosic fibers consumption. The supply of non-cellulosic fibers was estimated for the yorld. The polyester price was determined from an inverted demand equation for non- ellulosics in the Rest-of—the-World which was derived from a market clearing identity. The iodel was linked to the cotton model through the polyester price which entered the cotton emand equations. Overall, the econometric equations were satisfactory with good tplanatory power and coefficients and elasticity estimates at reasonable levels. 7. Validation of the Model The single equation estimates presented in earlier chapters were accepted or jected on the basis of a set of standard diagnostics such as the corrected R-squared, the .Irbin-Watson statistic and standard error of residuals. The decision to accept or reject equation often ultimately depends on the purpose for which the equation is being :irnated. For example, models estimated for forecasting should have small standard on, while those used for evaluating alternative policy scenarios or calculating structural sticities should be Specified to be consistent with economic theory. Once such single rations have been put together to form a multi-equation model, a similar evaluation icedure is necessary to test the properties of the entire model. This section presents 'erent statistics which cover various aspects of model evaluation. A major problem in validating a multi-equation model is that no statisically active criteria or benchmarks exist by which to accept or reject a validation statistic. criteria used are arbitarily chosen by the modeler. As in single equation estimation, decision to accept a model as satisfactory depends on the intended use of the model. lels designed for forecasting are typically put through more rigorous tests than those loped for evaluating alternative policy scenarios. In this section, four sets of validation statistics are presented. These cover various :ts of the model’s ability to plot historical data and to respond to economic stimuli in net consistent with both economic theory and empirical Observation. 131 132 1e validation statistics include: ( 1) Root Mean Square Percentage Error (RMSPE); (2) Mean Squared Error (MSE); (3) Theil’s U-statistic; (4) Graphical validation. validation statistics presented below were based on a simulation period from 1966 to 8 for production and price variables and from 1966 to 1986 for consumption variables. Validation using the Root Mean Squared Percentage Error The root mean square percentage error (RMSPE) statistic shows how well ulated values of the endogenous variables match with their actual historical values. The SPE is defined as, RMSPE = (1#12.((441-130/4402)1’2 ’e: A,= the actual value Of an endogenous variable, P,= the simulated value of an endogenous variable, and n = the number of periods in the simulation. :tatistic is very useful in that it provides a single value measuring the variation Of the :ted values around actual values of the endogenous variables. The statistic does have fawbacks. First, the RMSPE is an average which as a measure of central tendency ask the true nature of the series which it represents. For example, a few very large 133 rrors can raise the RMSPE of a series that otherwise tracks very well. Second, in cases here the actual values are small (e.g., net exports), small errors in absolute terms give rise substantial errors in percentage terms. The RMSPES for the endogenous cotton variables of the model are reported in able 7.1. At the world level the RMSPES for cotton use and cotton production are 1.56% d 2.2%, respectively. The RMSPE for the world price of cotton is 8.42%. Thus, the acking Of the quantity variables tends to be better than for prices. This can be explained the inelasticin of the supply and demand curves in which inaccuracies had a greater ect on prices than quantities. For the individual regions the results for cotton use tends to perform better than use for production. Of the 14 consumption regions in the model all recorded a RMSPE less than 15% and only for two regions (i.e., Pakistan and the Republic of Korea) have lues of less than 10% been reported. On the cotton production side of the model, the :ults are less good. However, only for three producing regions were RMSPES of more :n 20% obtained (i.e., Australia, Southeast region, United States and Sind region, ' tan), while for seven regions the simulation gave RMSPES Of less than 10%. As mentioned above, the RMSPES are sensitive to the levels of the variables. This lains the very high values recorded for stock and net trade in China (CTESCHI = 73.88 CPNECHI = 191.34). Both these variables are very close tO zero during some periods he Simulation, resulting in very large percentage errors between actual and simulated es. This is also the reason for the very high value reported for per capita total fiber in Central Africa (PCI'FUCAF = 83.3). In these circumstances, other validation stics must be used to evaluate the performance of the variable in the model. 134 ble 7.1. Root Mean S uare Percenta e Errors for the Model’s Endo enous Variables]. egion RMSPE Region RMSPE Region RMSPE Region RMSPE ['YDARG 1050 CI'YDAUS 11.63 CI'YDBRA 750 CI'YDCAF 6.78 ['YDCHI 3.35 CTYDEGY 13.90 CI‘YDINDN 5.92 CI'YDINDS 11.43 YDINDW 4.15 CI'YDMEX 5.08 CIYDPAKP 12.33 CTYDPAKS 9.21 'YDTUR 6.35 CI‘YDUSI 9.19 CFYDUSZ 15.18 CI'YDUS3 7.90 YD US4 6.63 ‘HAARG 11.82 CI'HAAUS 29.78 CI’HABRA 536 CI'HAC'AF 251 EACH! 5.94 CI'HAEGY 5.38 CIT-IAINDN 10.49 CI'HAINDS 4.02 HAINDW 1.24 CI'HAMEX 17.41 CIHAPAKP 5.48 CI'HAPAKS 11.76 HATUR 7.47 CIT-IAUSI 11.88 CI'HAUSZ 2351 CI‘HAUS3 11.68 HAUS4 14.44 PDARG 13.72 CI‘PDAUS 3851 CI'PDBRA 9.68 CI‘PDCAF 6.79 DCHI 4.75 CI'PDEGY 14.48 CTPDINDN 8.28 CI'PDINDS 11.93 DINDW 3.87 CI'PDIND 4.30 CI‘PDMEX 17.44 CI‘PDPAKP 11.71 DPAKS 21.72 CTPDPAK 10.19 CI'PDTUR 7.96 CI'PDUSI 12.18 DUSZ 2750 CI‘PDUS3 15.88 CI'PDUS4 16.14 CI‘PDUSA 8.31 DWOR 2.20 HARG 2.71 CI‘SHAUS 3.17 CI‘SHBRA 2.60 CI‘SHCAF 1.33 H CHI 2.48 CI‘SHEEC 3.33 CI‘SHEGY 5.78 CI‘SHIND 3.38 iI-UPN 3.33 CI‘SHKOR 7.50 CI‘SHMEX 3.07 CI‘SHPAK 2.86 ll-lTUR 5.51 CI‘SHUSA 5.28 FUARG 4.70 PCTFUAUS 454 PCI'FUBRA 6.11 PCI'FUCAF 83.30 FUCHI 6.11 PCFUAEEC 2.82 PCI'FUEGY 5.07 PCI'FUIND 1.70 FUJPN 6.95 PCFUAKOR 11.05 PCI'FUMEX 4.69 PCI'FUPAK 12.81 UATUR 3.59 PCTFU USA 2.63 ARG 4.36 CI‘UAUS 5.45 CI'UBRA 6.11 CI'UCAF 652 CHI 651 CI‘UEEC 3.79 CI‘UEGY 7.95 CTU IND 3.02 JPN 754 CI'UKOR 13.80 CTUMEX 752 CI'UPAK 14.19 TUR 7.13 CI'UUSA 5.97 CI'UW OR 156 CTESWOR 9.69 CHI 73.88 CINECHI 191.34 CTPAWOR 8.42 WOR 1158 NCUW OR 8.62 PSPWOR 14.40 RG 11.24 NCSHAUS 9.62 NCSHBRA 25.43 NCSHCAF 22.32 CHI 26.62 NCSI-IEEC 12.53 NCSHEGY 12.05 NCSHIND 17.88 N 6.28 NCSHKOR 24.21 NCSHIND 17.88 N CSHJPN 6.28 KOR 24.21 N CSHM EX 14.94 NCSHPAK 112.92 N CSI-ITU R 10.49 USA 11.00 RG 1252 N CUAUS 1251 NCUBRA 25.17 NCUCAF 24.27 HI 29.36 NCUEEC 13.03 NCUEGY 13.78 N CUIND 18.25 N 1259 NCUKOR 22.15 NCUMEX 20.80 NCUPAK 104.87 R 10.78 NCUUSA 11.70 NCU ROW 14.97 'able definitions are provided in Appendix. 135 The RMSPES for the non-cellulosic fibers variables are also reported in Table 7.1. e RMSPES for world supply, demand and price are 11.58, 8.62 and 14.4, respectively. ese are higher than their cotton counterparts but still indicate good model predictions the historical series. Again, the price variable performs less well than the quantity riables, reflecting the inelasticity of demand for non-cellulosic fibers. The non-cellulosic ers share equations do not perform as well as the cotton share equations. This may fleet the fact that non-cellulosic fibers share has been historically below the cotton Share that percentage differences are computed from a lower base level-~accounting for the ge RMSPES reported for China, the Republic of Korea and Pakistan. The RMSPES for non-cellulosic fibers use reflect errors in the share and total fiber use equations which ve been discussed already. The value for non-cellulosic fibers use in the Rest-of-the- )rld (NCUROW) is relatively low at 14.97. Accuracy in tracking this variable is portant as it was the demand variable in the polyester price equation. Validation Using Mean Squared Error (MSE) The mean squared error (MSE) is similar to the RMSPE in that it measures the n of the squared difference between actual and Simulated variables. It can be defined rms of the differences in the levels of the variables (MSEL) with, MSEL = 1/n 2:,(P,-A,)2 e: A: the actual value of an endogenous variable, P,= the simulated value of an endogenous variable, and n = the number of periods in the simulation, 136 ' in terms of percentage changes (MSEP) with, MSEP = l/n 2,(p,-a,)2 Jere: P. = (P. - AID/A... 3. = (A. - 41)/A... ace the MSEL will depend on the units in which the variable is measured, the MSEP is ore useful in providing comparisons of forecasting accuracy for variables measured in fferent units. The major usefulness of this statistic is that it can be broken down into separate mponents to reveal the sources of discrepancy between actual and simulated values. Two ethods of decomposition can be derived. Theil (1966) suggested that the MSE should broken down into its bias, variance and covariance components and these are derived follows, MSE = (P - A)2 + Szp-a, MSE = (P - A)2 + SZP+S2, - 2rSpS., MSE = (P - A)2 + (Sp-S.)2 + 2(1-r)SpS,, 1 = (P - A)2 + (Sp-S.)2 + 2(1-r)SpS,, MSE MSE MSE :re: P = the mean of the simulated values, A = the mean of the actual data, S, = the standard deviation of the simulated data, 5, = the standard deviation of the actual data, r = the correlation coefficient between the simulated and actual data. 137 [hell defined, (P - A)2 as the bias component (Ub), MSE (Sp-S.)2 as the variance component (U’), MSE 2(1-r)S,,S, as the covariance component (Uc). MSE iote that U'3 + U + Uc = 1. The bias component shows whether the simulated values tend to be higher or lower ran the actual values, while the variance component indicates to what extent the MSE is fluenced by the variance of the actual and simulated values. The covariance component easures the unsystematic error (i.e., that which remains after errors in average values and erage variabilities have been accounted for). Maddala (1977) argues that there is no reason to insist that the variances of actual d simulated data should be equal and suggests that a decomposition into bias, regression d disturbance terms is more illuminating. These are derived as follows, MSE = (P - .4.)2 +32”, MSE = (P - A)2 +szp+si - 2rsps, MSE = (P - A)2 +(S,,-rs,)2 + (14652,, 1 = (P - A)2 +(sp-rs,)2 + (1-r2)s2,, MSE MSE MSE 138 Maddala defines (p - A)2 as the bias component (Uh), MSE (Sp-rs.)2 as the regression component (U'), MSE (1-r2)S2, as the disturbance component (Ud). MSE Iote again that U" + Ur + Ud = 1. Maddala describe the benefits of this approach using the regression of actual on 'mulated values as follows, A,=a+b*P, perfect forecast yields a = 0 (U’ = 0) and b = 1 (Ur = 0). Figure 7.1 shows a 'gression line between actual and simulated values in which the 45° line represents a :rfect forecast (P,=A,). The error in the intercept (a 4: 0) is accounted for by Ub while 8 error in the slope is accounted for by U’. The Ud represents unsystematic errors, rived from random disturbances that are contained within the actual data series. Since 3y are random and cannot be explained by the model, the forecast cannot be expected capture these disturbances. Given that a perfect forecast yields U" = O and Ur = 0 the idation statistics improve as, *06Uv90,U°-91,U'40, and Ud-ol. 139 Figgre 2,1 The Regr_ession of Actual Against Simulated Values, Actual i'O. b.1 o l~0. b‘l 45‘ . Statuatcd The MSE and its decompositions are presented in Table 7.2. Overall the model performs well in terms of this criterion. Most of the U" and U' values are close to zero, ‘ indicating that the simulated values do not tend to be higher or lower than their actual values. The U‘ for most of the variables are close to one. This indicates that most of the errors in the simulated values are associated with randomness in the actual data series. Using the Theil decomposition of MSE into bias, variance and covariance, the model results further suggest that the errors in predicted values can be associated with unsystematic errors. The RMSPES showed poor performance for some model variables-- especially cotton stocks and net exports of China, per capita total fiber use in Central Africa, and the non-cellulosic fibers share in Pakistan. Using the MSE criterion it appears that net trade of China and fiber use in Central Africa validate better. However, the Chinese stocks variable do not perform well, with a Ub as high as 27.7%. 140 Table 7.2. Mean- are Error and its Decom sitions for the Model’s Endo enous Variables' MSE Ub U’ Ud Uv Uc CTYDARG 0.001 0.013 0.055 0.932 0.011 0.976 CTYDAUS 0.011 0.016 0.033 0.952 0.028 0.956 CI‘YDBRA 0.000 0.001 0.002 0.996 0.036 0.993 CI'YDCAF 0.003 0.155 0.002 0.843 0.025 0.820 CI‘YDEGY 0.016 0.008 0.022 0.971 0.110 0.883 C'I'YDINDN 0.001 0.037 0.035 0.929 0.138 0.625 CI'YDINDS 0.001 0.007 0.021 0.972 0.069 0.923 1 CI'YDINDW 0.000 0.001 0.034 0.965 0.000 0.999 ' CI'YDMEX 0.002 0.043 0.041 0.916 0.177 0.780 CTYDPAKP 0.001 0.005 0.116 0.879 0.004 0.990 CI'YDPAKS 0.000 0.104 0.008 0.888 0.000 0.396 CI'YDPRR 0.251 0.065 0.080 0.855 0.159 0.775 CI'YDTUR 0.002 0.368 0.089 0542 0.017 0.615 CI'YDUS] 0.003 0.003 0.008 0.989 0.140 0.858 crvnusz 0.006 0.010 0.009 0.980 0.062 0.928 CI'YDUSB 0.005 0.008 0.022 0.971 0.110 0.883 CTYDUS4 0.004 0.100 0.000 0.899 0.055 0.845 CI'HAARG 3286.0 0.003 0.000 0.997 0.102 0.895 CI‘HAAUS 273.69 0.003 0.245 0.752 0.417 0580 CTHABRA 15898. 0.009 0.069 0.922 0.252 0.739 CTHACAF 5345.7 0.001 0.071 0.928 0.126 0.874 CTHACHI 45086. 0.073 0.037 0.889 0.072 0.354 CI‘HAEGY 10943 0.000 0.016 0.984 0.085 0.915 CTHAINDN 5081.3 0.095 0.169 0.736 0.265 0.340 CIHAINDS 4306.7 0.000 0.000 1.000 0.030 0.969 CI'HAINDW 4186.2 0.000 0.002 0.998 0.013 0.986 CTHAMEX 2344.4 0.012 0.012 0.976 0.070 0.918 CI'HAPAKP 6013.9 0.040 0.199 0.761 0.111 0.450 CI'HAPAKS3 474.9 0.071 0.013 0516 0.000 0.428 CI'HATUR 2365.3 0.012 0.024 0.954 0.270 0.718 CI'HAUSI 18037. 0.044 0.006 0.949 0.103 0.852 CI'HAUSZ 50025 0.077 0.051 0.872 0.000 0.923 CTHAUS3 68040. 0.036 0.118 0.846 0.341 0.623 CTHAUS4 8397.2 0.041 0.635 0324 0.766 0.192 CTPDARG 452.45 0.000 0.001 0.999 0.065 0.935 CI'PDAUS 435.45 0.009 0.272 0.720 0.430 0561 CI‘PDBRA 31535 0.000 0.037 0.963 0.002 0.998 Cl'PDCAF 2144.8 0.156 0.000 0.844 0.048 0.795 CIPDCHI 31460. 0.054 0.244 0.702 0.307 0.637 CIPDEGY 5445.0 0573 0.087 0.339 0.005 0.421 CTPDINDN 622.6 0.154 0.006 0.839 0.014 0.832 CIPDINDS 713.4 0.017 0.074 0.909 0.017 0.966 CI'PDINDW 697.5 0.002 0.025 0.973 0.001 0.997 CTPDIND 26723 0.055 0.006 0.939 0.003 0.942 CI'PDMEX 1808.7 0.058 0.002 0.939 0.029 0.913 CI'PDPAKP 2529.9 0.037 0.251 0.712 0.112 0.851 CI'PDPAKS 13755 0.026 0.025 0.948 0.000 0.973 CI'PDPAK 3026.2 0.128 0.153 0.719 0.064 0.808 141 Table 7.2. Mean-Square Error and its Decommsitions for the Model’s Endqggnous VariablesI MSE U” 0' U: u" u‘ CI'PDTUR 1454.4 0.011 0.077 0.911 0.002 0.987 CTPDUSl 27883 0.139 0.002 0.859 0.033 0.828 CIPDUS2 1813. 0.158 0.197 0.645 0.056 0.787 CI'Pnusa 11783. 0.039 0.070 0.891 0.218 0.743 CI'PDU84 11661. 0.010 0.442 0548 0.624 0.366 CI'PDUSA 47389. 0.000 0.104 0.896 0.283 0.716 CI‘PDWOR 79929. 0.036 0.294 0.671 0.354 0.610 CTSHARG 0.000 0.210 0.073 0.716 0.146 0.644 CI‘SI-IAUS 0.000 0.004 0.002 0.994 0.031 0.965 CI'SHBRA 0.000 0.003 0.013 0.984 0.082 0.915 CI‘SHCAF 0.000 0.030 0.129 0.841 0311 0.659 CI'SHCHI 0.000 0.001 0.000 0.999 0.017 0.982 CI‘SHEEC 0.000 0.003 0.002 0.995 0.084 0.913 CI‘SHEGY 0.002 0.556 0.054 0389 0.154 0.289 CISHIND 0.000 0.002 0.050 0.948 0.165 0.833 CI‘SHJPN 0.000 0.032 0.004 0.964 0.151 0.081 CISHKOR 0.001 0.005 0.003 0.993 0.011 0.984 CI‘SHMEX 0.000 0.097 0.026 0.877 0.043 0.860 CI‘SHPAK 0.001 0.007 0.019 0.974 0.080 0.913 CI’SHTUR 0.001 0.061 0.022 0.917 0.140 0.799 CI‘SHUSA 0.000 0.000 0.018 0.981 0.074 0.926 PCTFUARG 0.105 0.001 0.015 0.984 0.119 0.880 PCTFUAUS 0.686 0.008 0.030 0.962 0.132 0.860 PCTFUBRA 0.104 0536 0.227 0238 0.284 0.181 PCTFUCAF 0.002 0.005 0.021 0.974 0.044 0.951 PCTFUCHI 0.039 0.004 0.004 0.991 0.004 0.992 PCI‘FUEEC 0.146 0.022 0.011 0.967 0.001 0.976 PCTFUEGY 0.079 0.004 0.002 0.994 0.011 0.985 PCI’FUIND 0.002 0.005 0.021 0.974 0.044 0.951 PCTFUIPN 0.005 0.005 0.015 0.980 0.014 0.982 PCTFUKOR 0349 0.005 0.010 0.986 0.039 0.956 PM 0.173 0.022 0.409 0569 0.631 0.348 PCTFUPAK 0.169 0.000 0.051 0.949 0.283 0.717 PCTFUTUR 0.069 0.031 0.000 0.969 0.033 0.936 PCTFUUSA 0309 0.002 0.010 0.987 0.080 0.918 CI‘UARG 21.18 0.090 0.013 0.897 0.105 0.805 CI'UAUS 32.71 0.006 0.080 0.914 0.000 0.994 CI'UBRA 535.81 0551 0.104 0.344 0.147 0.301 CTUCAF 344.20 0.118 0.060 0.822 0.001 0.882 CTUCHI 18677. 0.004 0.008 0.988 0.001 0.995 CIUEEC 3152.2 0.003 0.019 0.978 0.003 0.994 CTUEGY 198.96 0.329 0.018 0.653 0.057 0.614 CI'UIND 1101.7 0.008 0.008 0.984 0.022 0.970 CTUJPN 3055.4 0.023 0.000 0.977 0.053 0.924 CTUKOR 128.97 0.006 0.014 0.979 0.073 0.920 CTUMEX 110.81 0.018 0.206 0.776 0.006 0.975 CIUPAK 691.70 0.000 0.033 0.966 0.119 0.880 CTUTUR 18837 0.016 0.131 0.853 0.008 0.976 142 Table 7.2. “A" “mm" Error and its " for the Model’s " ’ VariablesI MSE U” u’ ()4 Uv If crUUSA 10276 0.001 0.000 0.999 0.052 0.948 CI'UWOR 149.8 0.001 0.000 0.999 0.011 0.988 CI'ESWOR 271926 0.231 0.000 0.768 0.047 0.722 CI'ESCHI 160018 0.277 0.012 0.711 0.074 0.649 crNECHI 23664 0.004 0.05 0.937 0.295 0.701 chw0R 149.89 0.001 0.000 0.999 0.011 0.988 NCPDWOR 1883880 0.083 0.123 0.794 0.082 0.835 NCUWOR 1307980 0.004 0.086 0.910 0.124 0.872 PSPRWOR 396.9 0.119 0.014 0.867 0.029 0.853 NCSHARG 0.0006 0.027 0.000 0.973 0.026 0.947 NCSHAUS 0.0007 0.255 0.082 0.662 0.141 0.603 NCSHBRA 0.0010 0.007 0.002 0.992 0.059 0.935 NCSHCAF 0.0004 0.001 0.148 0.850 0.325 0.673 NCSHCHI 0.0004 0.000 0.048 0.952 0.005 0.995 NCSHEEC 0.0010 0.007 0.013 0.979 0.132 0.861 NCSHEGY 0.0012 0.008 0.002 0.980 0.105 0.893 NCSHIND 0.0000 0.040 0.113 0.848 0.174 0.786 NCSHJPN 0.0003 0.013 0.000 0.987 0.021 0.967 NCSHKOR 0.0042 0.004 0.042 0.954 0.004 0.991 NCSHMEX 0.0008 0.044 0.122 0.833 0.174 0.781 NCSHPAK 0.0003 0.012 0.044 0.944 0.134 0.854 NCSHTUR 0.0005 0.000 0.018 0.982 0.000 1.000 NCSHUSA 0.0010 0.028 0.256 0.716 0.361 0.611 NCUARG 27.76 0.018 0.007 0.975 0.008 0.974 NCUAUS 91.86 0.078 0.095 0.828 0.169 0.753 NCUBRA 387.79 0.171 0.079 0.750 0.160 0.668 NCUCAF 79.01 0.004 0.042 0.954 0.186 0.810 NCUCHI 6629.4 0.010 0.049 0.941 0.008 0.982 NCUEEC 19270.37 0.001 0.097 0.902 0.208 0.791 NCUEGY 90.91 0.005 0.041 0.954 0.183 0.813 NCUIND 98.85 0.056 0.162 0.781 0.224 0.720 NCUJPN 3400.4 0.000 0.089 0.911 0.179 0.821 NCUKOR 288.5 0.008 0.015 0.976 0.047 0.945 NCUMEX 214.23 0.003 0.234 0.763 0.301 0.696 NCUPAK 22.06 0.115 0.009 0.876 0.004 0.881 NCUTUR 51.86 0.003 0.022 0.975 0.000 0.994 NCUUSA 27161.7 0.011 0.210 0.779 0303 0-686 NCUROW 64444.1 0.009 0.243 0.747 0.148 0.843 1/ Variable definitions are provided in Appendix. 143 7.3 Validation Using Theil’s U-Statistic A useful statistic related to both the RMSPE and the MSE is Theil’s inequality coefficient. Theil’s inequality statistic has been defined as, U1 = [1/ “21(PI'A1)2]1/2 II/nx(P.)’]”’ [1/n2.(A.)’l”’ This statistic is scaled so that U1 will lie between 0 and 1 (U1 = 0 represents a perfect fit while Ul =31 indicates a predictive performance as bad as it can possibly be). A major shortcoming is described by Leuthold (1975), for both actual data and changes the error depends on the predictions themselves, that is, the purpose is to assess P, but the assessment is made relative to P, itself since P, is in the denominator. For this reason, a second Theil statistic was used given by, U. = [MSE / 1/n ate-41.01"” U, takes on a value of 0 for a perfect forecast (as in the case of U,) but has no upper limit. It can be shown also that U,= 1 indicates a prediction performance the same as a naive, no-change extrapolation. The U, statistics for the endogenous variables of the model are presented in Table 7.3. All variables have a U, very close to zero, indicating that the model performed well. The variables that perform least well are stocks and trade in China and the non-cellulosic fibers share in Pakistan. This result is consistent with the findings suggested by the other validation statistics discussed earlier. . .4“ 144 Table 7.3. Theil U-Statistics ,) for the Model’s Endogenous Variables’. Region U2 Region U, Region U, Region U, CI'YDARG 0.0525 CTYDAUS 0.0517 CI'YDBRA 0.0369 CI'YDCAF 0.0380 CI'YDCHI 0.0502 CI‘YDEGY 0.0829 CI’YDINDN 0.0316 CI‘YDINDS 0.0424 CIYDINDW 0.0201 CTYDMEX 0.0248 CI'YDPAKP 0.0579 CI'YDPAKS 0.0402 CI'YDTUR 0.0332 CTYDUSI 0.0446 CI'YDUSZ 0.0756 CI'YDUS3 0.0300 CI'YDUS4 0.0307 CI‘HAARG 0.0637 CI‘HAAUS 0.1196 CI‘HABRA 0.0285 CI'HACAF 0.0124 CI‘I-IACHI 0.0210 CI'HAEGY 0.0281 CIHAINDN 0.0428 CI'HAINDS 0.0202 CTHAINDW 0.0061 CI'HAMEX 0.0531 CTHAPAKP 0.0277 CI'HAPAKS 0.0525 CI'HATUR 0.0360 CI'HAUSI 0.0524 CI'HAUSZ 0.0796 CTHAUS3 0.0585 CI'I-IAUS4 0.0699 CI‘PDARG 0.0755 CI‘PDAUS 0.1283 CI'PDBRA 0.0462 CI'PDCAF 0.0369 CI'PDCHI 0.0309 CI'PDEGY 0.0849 CTPDINDN 0.0397 chDINDS 0.0503 CIPDIND 0.0183 CIPDIND 0.0200 CIPDMEX 0.0593 CTPDPAKP 0.0575 CI'PDPAKS 0.0884 CTPDPAK 0.0428 CTPDTUR 0.0382 CTPDUSI 0.0347 chnusz 0.0960 CIPDUS3 0.0632 CI'PDUS4 0.0756 CTPDUSA 0.0435 CTPDWOR 0.0104 CTSHARG 0.0138 CI’SHAUS 0.0159 CTSHBRA 0.0129 CI‘SHCAF 0.0068 CISHCHI 0.0121 CISI—IEEC 0.0167 CI‘SHEGY 0.0306 CTSHIND 0.0167 CI'SHJPN 0.0306 CISHKOR 0.0334 CISHMEX 0.0107 CI‘SHPAK 0.0145 CI‘SHTUR 0.0248 CI‘SHUSA 0.0256 PCI‘FUARG 0.0238 PCI‘FUAUS 0.0226 PCFUABRA 0.0341 PCTFUCAF 0.0330 PCI'FUCHI 0.0313 PCI'FUEEC 0.0142 PCI'FUEGY 0.0266 PCFUAIND 0.0096 PCI‘FUJPN 0.0333 PCTFUKOR 0.0396 PCI'FUMEX 0.0390 PCFUAPAK 0.0585 PCI‘FUTUR 0.0181 PCTPUUSA 0.0126 CTUARG 0.0222 CI‘UAUS 0.0273 CI'UBRA 0.0335 CTUCAF 0.0321 CTUCPII 0.0285 CUAEEC 0.0182 CI'UEGY 0.0418 CTUIND 0.0150 CTUJPN 0.0372 CUAKOR 0.0550 CTUMEx 0.0370 CI’UPAK 0.0649 CTUTUR 0.0374 CI'UUSA 0.0285 * CTUWOR 0.0078 CI'ESWOR 0.0452 CI'NCHI 0.1528 CINEPR 0.5029 CI‘PWOR 0.0448 NCPDWOR 0.0280 NCUWOR 0.0226 PSPRWOR 0.0784 NCSHARG 0.0520 NCSHAUS 0.0343 NCSHBRA 0.0673 NCSHCAF 0.0742 NCSHCHI 0.0851 NCSHEEC 0.0430 NCSHEGY 0.0481 NCSHIND 0.0585 NCSHJPN 0.0267 NCSHKOR 0.0674 NCSHMEX 0.0322 NCSHPAK 0.1046 NCSHTUR 0.0477 NCSHUSA 0.0316 NCUARG 0.0603 NCUAUS 0.0464 NCUBRA 0.070 NCUCAF 0.0889 NCUCHI 0.0916 NCUEEC 0.0431 NCUEGY 0.0402 NCUIND 0.0584 NCUJPN 0.0481 NCUAKOR 0.0501 NCUMEX 0.0433 NCUPAK 0.1113 NCUTUR 0.0479 NCUUSA 0.0334 NCUROW 0.0588 1/ Variable definitions are provided in Appendix 1. 145 7.4. Graphical Validation A common method Of validation involves examining plots Of actual and simulated values against time. Graphs for world production, consumption and price for cotton and non-cellulosic fibers are presented in Figures 7.2-7.7, providing visual evidence of how well the model tracks. They also may indicate that for some periods within the simulation period the model tracks better than in others. While providing an instantaneous perceptual measure, graphical validation may be misleading because the size of the differences between actual and predicted values, when portrayed graphically, depends on the scale Of the graph. The smaller the scale, the better the validations appear. Simulated and actual values for world cotton production are presented in Figure 7.2. The model tracks well, although tending to underestimate supplies in the late 19705. However, all of the major turning points in the actual data are captured by the model, including the wide fluctuations in production which occurred in the early 1980s. The graph for cotton use is presented in Figure 7.3. Again the model tracks the actual data fairly well, especially before 1975 and after 1982. From 1976 to 1981 the model predicts consumption to be much more variable than actually occurred. During this period consumption did not seem to respond to the widely fluctuating prices and tended to trend slowly upwards. In the model, consumption is muoh more responsive to price fluctuations. Simulated and actual price data are plotted in Figure 7.4. Overall, the model appears to track cotton prices reasonably well, especially in the late 1960s and early 1970s. Most of the major turning points of the data are captured such as in 1976, 1980, 1983 and 1986. 626355 1T .6363. ll mm? mm? vwmw Nmmp ommw mumCFamMan 452. «how one momw. comp 3 L _ .4 Tarim a w _ u _ L4 _ m _ 1+ _ n a or - \47... _ i + 3 m \>// S. R) Q 3 ON mm:_e> eBmSEE m> .m:.o< ciozozngn cro..:oo uto>> «K 9:9... 2_——_oc Foam 147 86326 It .622 ll or NF 39 mmmp #4me Nwmw ommv wnmP cum? #4an mum? Dump wwmw mm? _ n _ u _ __ _ fl — A. — n _ Jr _ “ _ _fi _ -.h i. L: mu:_m> “622355 m> .m:_o< cozoEzmcoo cozoo Uto>> Wu 9:20. we or mp l—ocw §_——_O: 148 xouE .<. x0230 \F cmeflngw IT .m30< ll 89 mmmw on? 33.. Nmmw ommw mhmp @509 char N59 052 mmmp mom? ____ _ __ __4_i_aal_3______qdna“.om . ll- 00F /\ om— K < . com mo:.m>. “#52365 m> .m30< \F meta P521010 Uto>> VS 93min. omcww\xm 149 The underprediction of price during the late 19705 explains the underestimation Of production during the same period. This underestimation should not persist if the demand side of the model is responsive to price levels. The production and consumption Of non-cellulosic fibers at the world level trend up throughout the simulation period as shown in Figures 7.5 and 7.6. This is captured well by the model except for the 1974/75 and 1981/82 periods. The polyester price shown in Figure 7.7 is tracked well by the model, with the simulated series catching the decline in the late 19608, the increase in the 1970s, and the slight decline in the early 1980s. 7.5 Summary The simulation results for cotton stocks and net exports in China and non-cellulosic fibers use in Pakistan were poor for all validation tests undertaken, which suggested that future modeling efforts should focus on these variables. However, in general, the model performed well and was able to predict actual market values reasonably accurately. 150 B59385 IT 89 23°F 3.9.5 2 mmmw mmm— vmmw Nmmw ome mum? $59 #4an N52 one am? mam? _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \r/ F \fm :. §.__—.—OC l—OCU) mw mo:_o> eBEBEw m> 333. \F cozoznoi O_mo_:=moicoz mK 9:2“. mw 151 :30... U225 \P 626.35% IT .622 i Lmo> mmmp vmow Nmmw ommw mnmp ohm? vnmw N59 one comp mmmr T _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ N .1 3V1 NF 3. moEm> “322285 m> .m30< ) :00 0622601202 m: oSEM li‘ l—OCCD 2._——-— O C 626356 IT .632. ll Cmo> mmmw owmw #4me N09. Omar aha? mum? whmp thr 059 moor 009 a? 8 3 . // 7 - mm, _\ x} /. . P: 152 OCDCHCO\_‘£U) mow ii mo:_m> $92355 m> 333‘ m tn. $6628 u_._o>> KN 831E ill-II 8. Model Simulations A number of model simulations were undertaken in order to meet the research Objectives outlined in the introductory chapter. In all, five model simulations were performed. These were: (i) forecast of price, production and consumption variables for the period 1989-2005, given a basic set of macro-economic assumptions, (ii) a 10 percent decrease in cotton production in the Soviet Union, (iii) a 10 percent increase in cotton production in China, (iv) a 10 percent decrease in domestic cotton prices in the U.S.,’and (v) an evaluation Of the effects Of the Multi-Fiber Agreement on the cotton and non- cellulosic fibers sectors. 8.1 Simulation One - Forecast for the Fiber Market to Year 2005 8.1.1 Rationale for Forecasting Prices Accurate price forecasts are useful to many groups, such as governments, lending institutions, as well as producers and consumers of cotton. Informed guesses about future fiber prices are especially important to decision makers in those countries where earnings from cotton represent a very large percentage of export earnings. Fiber price forecasts are essential for planners and policy-makers who make decisions about fiscal and investment policy, as well as set the levels of cotton prices and other forms of cotton market interventions. Other groups interested in price forecasts are producers and consumers of 153 154 cotton and synthetic fiber. Accurate forecasts Of future fiber prices improve resource allocation decisions in all stages from production at farm and industry levels to textile and apparel purchases by final consumers. In addition to price forecasts, production and consumption forecasts are also important to decision makers. While forecasts of total world levels of production and consumption determine expected futures prices, projections of production and consumption trends for different countries and regions are valuable to producers and consumers of fiber products. For example, the identification and development Of potential export markets for many countries depend on accurate forecasts of where demand will be strong in the future. This is especially important for developing countries, many of which have recently emerged as major participants in world fibers markets. 8.1.2 Forecasts of Exogenous Variables The model was Simulated through to 2005 in order to provide forecasts of price, production and consumption in the cotton and non-cellulosic fibers sectors. This required that values for all the exogenous variables be evaluated for each year to the end of the simulation period. Annual percent growth rates in income and population for each region were used to project the levels of these variables for each year between 1990 and 2005. Exchange rates and CPIs were given their 1989 values, thus assuming that purchasing power parity was maintained over the long-run. Weather variables which appeared in the yield equations were set at their mean values and production and consumption in the Rest- of-the-World region were forecasted based on a regression against time. More details of the assumptions made for the exogenous variables in the forecast period are given in 155 Appendix B. 8.1.3 Forecast Results 8.1.3.1 Cotton Price Forechs The nominal and deflated price forecasts are presented in Table 8.1 and Figures 8.1 and 8.2. The nominal cotton price is forecast to be 189.8 c/kg in 1990 and to increase at an average of 1.8% annually over the next five years to reach 207.2 c/kg by 1995. Between 1995 and 2000 price is projected to grow at an average of 2.9% p.a., reaching 239.4 c/kg by the end of the century. The annual rate of increase in price is forecast to remain at 2.9% between 2000 and 2005, and for the nominal price to be 276.4 c/kg by the year 2005. Nominal prices were deflated by the MUV and US GDP deflator (1985= 100) and are reported in Table 8.1. Based on the MUV deflator, the real price is predicted to decline throughout the forecast period, falling at an average annual rate of 1.76%, 2.25% and 2.84% for the periods 1990-1995, 1995-2000 and 2000-2005, respectively. The differing rates of change in prices can be associated with cyclical movements in cotton prices that were captured by the model. The projection of a decline of 12%-13% in real cotton prices between 1990 and 2005 is a troublesome result for cotton producers. This suggests that improvements in yields, through, for example, the development Of better seed varieties, the use of lower cost fertilizers, more effective pest and disease controls and improved management, are crucial for cotton to be a profitable enterprise. Alternatively, domestic producer price incentives Will have to be maintained and even extended at high budget cost. This would involve a move away from the current trend, especially in industrialized countries, of reducing levels 156 of government agricultural supports. The forecast of declining real cotton prices is significant for the main user of cotton, that is, the textile manufacturing sector which accounts for approximately 50 percent of consumption. A lower price of its major input should encourage the sector to further invest in textile manufacturing. Table 8.1 Nominal and Deflated Cotton Price Projections, 1990-2005. Year Nominal MUV Deflated1 US GDP Deflated2 (c/kg) (c/lb) (1985= 100) (1985= 100) 1990 189.8 93.7 133.0 159.5 1991 192.5 95.1 131.6 154.9 1992 195.4 96.5 128.3 151.2 1993 202.0 100. 1 126.3 151.9 1994 204.4 101.0 121.7 148.1 1995 207.2 102.3 119.1 145.3 1996 211.4 104.4 117.6 143.7 1997 216.5 106.9 116.9 142.8 1998 223.5 110.4 117.3 142.5 1999 232.0 114.6 118.0 143.0 2000 239.4 118.2 117.7 142.9 2001 246.1 121.6 117.2 141.9 2002 252.8 124.9 116.6 141.0 2003 259.9 128.4 116.0 140.4 2004 267.2 132.0 116.2 139.9 2005 276.4 136.5 116. 1 140.4 ’ Unit value index in US dollar terms of manufactures exported from the G-5 countries (i.e., France, Germany, Japan, United Kingdom and United States) weighted proportionally to the countries’ exports to the developing countries. Source: IECAP, World Bank. 2 Source: International Monetary Fund. 157 Fi re 8.1. Nominal Cotton Price Pro'ections 1990-2005. Outlook Index "A". 280 ‘ /l 260 // c 240 / k g 220 / 200/ 180Lrld'i’+Ln;_Ji*i 1990 1992 1994 1996 1998 2000 2002 2004 Year Figure 8.2. Deflated (MUV) Cotton Price Projectiorgd990-2OOS. Outlook Index "A". 135 130\ ‘i k 125 9 115 L 1 1— 1 _L m 1 1 1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year 158 8.1.3.2 Cotton Production Forecasts The model production forecasts are presented in Table 8.2. In 1990 production is projected to reach 18.3 million tons and to increase at an annual growth rate of 1.1% per year, reaching 21.8 million by 2005. The largest contributor to world production is China whose production is projected to grow on average at 1.5% per year, increasing from 4.4 million tons in 1990 to 5.5 million tons in 2005. The increase is mainly associated with yield increases, since the potential for area expansion is limited given that China’s total Table 8.2 Model Cotton Production Forecasts for the Years 1990. 1995. 2000 and 2005. Region 1990 1995 2000 2005 Growth Rate1 (’000 tons) ---(%)--- Argentina 168 245 304 391 5.8 Australia 232 259 277 290 1.5 Brazil 711 761 793 838 1.1 Central 754 799 822 833 0.7 Africa China 4,449 4,799 5,176 5,529 1.5 Egypt 341 335 328 322 -0.4 India 1,901 2,167 2,347 2,674 2.3 Mexico 144 148 151 154 0.4 Pakistan 1,367 1,427 1,689 1,830 2.0 Turkey 578 607 645 661 1.1 United States 2,995 3,331 3,702 4,090 2.1 Rest-of- 4,702 4,639 4,392 4,214 -0.7 the-World2 WORLD 18,342 19,517 20,626 21,826 1.1 1 Average percentage growth rates per annum 1990-2005. 2 Exogenous in the model. 159 cultivated area is being used at a maximum and that the costs of bringing new land into production are very high. Since the introduction of market incentives in China in the late 19703, production has been determined largely by Government administered prices of cotton. In the model, China’s cotton area and yields were explained by the administered price, which was exogenous. However, over the forecast period the Chinese price was linked to the world price with an assumed transmission elasticity of 0.25 (equation 5.19). As it was not possible to predict accurately the future direction of China’s cotton price movements, the forecasts must be treated with caution. However, in order to provide a sensitivity analysis of the cotton price assumption on the model forecasts, a simulation was performed with Chinese production set 10% above its historical level. The results of this simulation are presented in section 8.3. Production in the United States is projected to increase at an average of 2.1% per year, from 3.0 million tons in 1990 to 4.1 million tons in 2005. This expansion results in large part from greater cotton area in the West and Southwest regions. On average, production is expected to grow at 2.3% per year, in India. This growth is associated mainly with improved yields as irrigated acreage continues to increase in the North and West regions of India. Pakistan’s output is forecast to increase from 1.4 million tons in 1990 to 1.8 million tons in 2005. This expansion is the result of improved yields in both the Punjab and Sind regions. World cotton production trends are illustrated in Figure 8.3 which compares the shares of world production produced by the major suppliers for the years 1990 and 2005. The most striking feature is that shares do not change significantly over the time period. For example, in 1990 the three largest producers - China, the United States and India - will 160 account for 530 of world production. In year 2005 these three countries are expected to supply 56%, with a small increase in the United States’ share and shares in Brazil, Central Africa, China, India and Pakistan all expected to be maintained throughout the next 15 years. To some extend the stability of market shares over the period 1990 through 2005 1 results from the fact that the exogenous variables used for forecasting are based on constant growth rates through time. Due to uncertainty over future production levels in China and the USSR where cotton growing decisions are made administratively, simulations of the model were performed to determine the effect on the world market of a 10% change in the production in these regions. The results from these simulations are reported later in this section. Figure 8.3. World Cotton Production Trends. Years 1990 and 2005 Com ared. CHINA 21$ CHINA 25$ .:~..:':::“ %// BRAZIL (I BRAZIL Ii PAKISTAN 1S UNITED 87ATE5 15$ OTHER 345 Year 1990 Year 2005 161 8.1.3.3 Cotton Consumption Forecasts A comparison of the model’s consumption forecasts are reported in Table 8.3. World consumption is forecast by the model to rise from 18.0 million tons in 1990 to 21.6 million tons in 2005. This is an average annual growth rate of 1.2%. The largest growth rates are forecast for some developing countries. For example, consumption in Argentina, Central Africa, Egypt, Mexico and Pakistan are expected to increase at growth rates of 2.8% 5.2%, 2.1%, 3.2% and 3.0%, per year, respectively. These growth rates can be explained by the high population growth rates assumed in these regions. With low base consumption levels in these regions large percentage growth rates are be recorded for a relatively small expansion in consumption in absolute terms (e.g., Argentina and Mexico). World cotton consumption trends are shown in Figure 8.4 which compares the preportions of total world consumption by the major purchasers of cotton for the years 1990 and 2005. Overall the pattern of consumption does not change dramatically over the period, with the t0p three consumers (China, the EEC and the United States) decreasing their share of consumption from 51% in 1990 to 49% in 2005. Again, the fact that world consumption patterns do not change significantly is partially explained by the fact that the growth rates of the exogenous variables used in the projection period are held constant. 162 Table 8.3 Model Cotton Consumption Forecasts for the Years 1990, 1995, 2000 and 2005. Region 1990 1995 2000 2005 Growth Rate1 (’000 tons) ---(%)---- Argentina 114 129 148 173 2.8 Australia 101 106 110 114 0.8 Brazil 543 588 640 679 1.5 Central Africa 184 187 249 393 5.2 China 4,597 4,803 5,001 5,121 0.7 EEC-12 2,405 2,553 2,697 2,876 1.2 Egypt 196 218 242 269 2.1 India 1,383 1,523 1,673 1,781 1.7 Japan 983 1,074 1,168 1,263 1.7 Korea, Rep. of 176 189 207 231 1.8 Mexico 129 140 167 207 3.2 Pakistan 240 292 338 377 3.0 Turkey 222 233 260 291 1.8 United States 2,314 2,420 2,569 2,686 1.0 Rest-of- 4,456 4,696 4,936 5,173 1.0 the-World2 WORLD 18,043 19,151 20,328 21,634 1.2 ’ Average percentage growth rates per annum 1990-2005. 2 Exogenous in the model. 163 Fi re 8.4. World Cotton Consum tion Trends. Years 1990 and 2005 Com ared. CHINA 29$ BfiAZIL 7‘ Bquu. 3‘ EEC“ 12$ JAPAN IS umrso sures 11s ' umrao sures as Year 1990 Year 2005 INDIA 75 OTHER 35$ JA'AN 05 OTHER 305 8.1.3-4 MW The model forecasts for consumption, production and price for non-cellulosic fibers are presented in Table 8.4. The nominal polyester staple price is expected to increase from 167.8 c/kg in 1990 to 357.0 c/kg in 2005. This is an annual average growth rate of 5.2%. In real terms, prices are expected to increase on average 1.6% per year over the same period (Figure 8.5 and 8.6). While a 12%-13% decline has been predicted for cotton price, the model forecasts a 10% increase in the price of polyester. This result should provide non-cellulosic fibers manufacturers an incentive to increase productive capacity through new investments. However, the forecast is closely associated with the forecast of future oil prices which is not easy to make accurately (Appendix B). 164 La_ble 8.4 Model Non-cellulogic Fibers Forecasts of Price, Consumption and Production. Variable Region 1990 1995 2000 2005 Gr.Rate1 Polyester Price (c/kg) ----(%)---- Nominal World 167.8 196.3 258.3 357.0 5.2 Deflated2 World 117 .6 112.8 127.0 149.9 1.6 Consumption (’000 tons) --------------- Argentina 70 83 93 101 2.5 Australia 154 180 188 187 1.1 Brazil 204 274 293 280 2.1 CAfrica 33 32 36 46 2.2 China 645 695 749 807 1.5 EEC-12 2,084 2,142 2,112 2,048 -0.1 Egypt 56 92 135 187 8.4 India 308 452 632 853 7.0 Japan 923 1,028 1,044 1,026 0.7 Korea 269 286 284 282 0.3 Mexico 243 285 309 318 1.8 Pakistan 46 57 69 82 3.9 Turkey 188 232 290 357 4.4 U.S 4,651 5,646 6,743 7,946 3.7 World 13,770 15,706 17,219 18,615 2.0 Production World 14,870 16,806 18,319 19,715 2.0 1 Average annual growth rates 1990-2005. 2 Deflated by MUV (1985= 100). Wasto O X\O 165 Fi re 8.5 Nominal Pol ester Price Pro'ections 1990-2005. 350 .’/// I 300 4' 250 4 200 / . ' l 150 1 p 11 4% 1“, g, L 1 1 n_ n_ l _1 p _J l I 1990 1992 1994 1996 1998 2000 2002 2004 Year Fi re 8.6. Deflated Pol ester Price Pro'ections 1990-2005. 130 a l 125 ‘4//// 120 l 10 I _1 1 l l l I I 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 ‘kar 166 Also, higher polyester prices will result in lower demand by the textile manufacturers who will demand less non-cellulosic fibers in response to the ratio of cotton to polyester prices. This ratio is depicted in Figure 8.7 which indicates a small increase in the price of cotton compared to polyester in 1991 and 1992. After 1992 the ratio falls at a fairly constant rate through to year 2005, declining more than 25 percent over the period. Non-cellulosic fibers production is expected to increase at 2% per year, increasing from 14.9 million tons in 1990 to 18.6 million tons in 2005. This expansion is driven by the increasing price of polyester (Figures 8.5 and 8.6) relative to the price of oil which is the largest input in non-cellulosic fibers production. Non-cellulosic fibers consumption is also expected to increase at 2% annually between 1990 and 2005. The United States is the largest contributor to this expansion with an annual average grth rate of 3.7% over the forecast period. Large rates of growth are also expected in India, Egypt and Turkey, but these are all increases from small base levels. The important trends in non-cellulosic fibers consumption between 1990 and 2005 are shown in Figure 8.8. As well as increasing production substantially over the period, the United States is expected to increase its share of world consumption--from 34% in 1990 to 43% in 2005. India is also expected to increase its share, reaching 5% of world consumption by 2005. All the other regions show small declines in market share (e.g., EEC from 15% to 11%; China from 5% to 4% and, Japan from 7% to 6%). Overall these trends suggest that over the next 15 years the United States will strengthen its position in the non-cellulosic fibers market at the expense of the rest-of-the-world. 167 Figure 8.7. Ratio of Cotton and Polyester Prices. Cotton Price / Polyester Price 1.15/— 1.05 0.95 0.85 \ 0.75 L l ’ b 1 ‘r 1 41 ’ 7‘ ' i i i N 1990 1992 1994 1996 1998 2000 2002 2004 Year Figpre 8.8. World Non-Cellulosic Fibers Consumption Trends. Years 1990 and 2005 Compared. UNITED srArEs 431. UNITED srAtEs an. cmNA u CHINA 41. JAPAN 711 ’ JAPAN n 0111511 on EEC 157. EEC mt OTHER 321. INDIA :1. Year 1990 INDIA 5% Year 2005 1/ Non—Cellulosic Fibers %- Percent of Total World Consumption Also Given Quantities Consumed (M.Tons) 168 8.2 Simulation Two - A 10 Percent Decline in Cotton Production in the USSR 8.2.1 MEG—11.119 The performance of the USSR cotton sector has been stagnant during the 19805 with production ranging from 8 million and 9 million metric tons annually. A primary cause has been poor management practices involving the over-irrigation of cotton land, resulting in salinization and leaching of nutrients from the soil. Another problem has been the failure of producers to establish effective crop rotations with cotton. It had been demonstrated that a cotton-alfalfa rotation increases yield, improves efficiency, reduces fertilizer and irrigation costs, and combats wilting. In addition to management practices, cotton production growth has been impeded by poor quality inputs, inadequate infrastructure and lack of price incentives. To combat these problems the Government is taking steps to improve cotton production performance. The main thrust of its program is to improve rotations and to move away from the monoculture of cotton by restricting cotton area. However, this policy has not met with much success.1 Other initiatives affecting all USSR agriculture include price reforms; property right expansion (e.g., Lease Terms); greater accountability and decision-making at local level (e.g., Self-Financing provisions introduced in 1988); efforts to improve the quality of agricultural inputs, especially chemicals and fertilizers; and, improving the linkages between agricultural research and science and the farming community. These will impact on the cotton sector substantially if they are implemented 1 ’Despite plans to restrict cotton area and to improve crop rotation cycles, Urbek cotton area increased 118,000 hectares from 1985 to 1987 and the cotton-alfalfa rotation areas fell from 55% of the arable land to 42%.’ USSR Agriculture and Trade Remrt, May 1989, HRS, USDA (p.37). 169 effectively. Despite these measures it is unlikely that USSR cotton production will increase in the 19908. More likely is a decline, as cotton land is taken out of production in order to establish cotton-alfalfa rotations and to replenish the soils previously damaged through extensive over-irrigation. 8.2.2 Simulation Objectives The USSR is the world’s third largest cotton producer, contributing about 15% to world production in 1988. USSR cotton production is not determined by world market prices. Instead, output is controlled administratively by Government planning agencies which set area limits and determine seed and fertilizer use. Since data were not available to model the decision making of these Government agencies, USSR production was exogenous in the model. However, in order to assess how USSR production affects the world market, the model was simulated over an 11-year period with production set 10% below its historical level. The percentage changes from the base simulations for some of the endogenous variables are reported in Tables 8.5 and 8.6. 8.2.3 Simulation Results The initial effect of decreased production in the USSR was to decrease world production by 2.07%. Lower production led to the cotton price rising by 5.07% above the base level and to consumption decreasing by 0.19%. The net effect of lower production and consumption was for world stocks to decrease 5.76% below the base simulation. In most production regions output was unchanged initially. This was because lagged prices 170 were used as regressors in most of the supply equations. Decreases in cotton consumption led to an increase in consumption of non-cellulosic fibers and caused the price of polyester to rise 2.09% above the base level. This rise partially offset the impact of high cotton prices on demand and accounted for the small changes in consumption in the initial period. On average a 10% decrease in USSR production caused the world cotton price to rise by 9.16%. Production in the price reSponsive-regions increased (e.g., United States up 2.77%, Mexico Up 5.36%, Pakistan up 1.42%) and this partially offset the effect of lower USSR production. The net effect was for world production to decrease an average of only 0.94% below the base simulation level. The rise in price caused consumption to decrease on average 0.48%, and for stocks to fall 10.13% below the base level. The average effect on the non-cellulosic fibers market was for non-cellulosic fibers consumption to rise by 0.35%. This increased the price by 3.72% and caused non-cellulosic fibers production to rise by 0.34%. The long-run (or final) impact of the USSR production shock was to increase price by 12.60%. By the end of the 11-year simulation period the increase in production in the rest-of-the-world offset almost all of the USSR production decrease, with production only 0.69% below the base simulation. Consumption continued to decline, but at a slower rate, in response to the falling rate of price increases in the later periods. The net effect of production and consumption changes was to lower ending stock levels 16.15% below their base values. The final impact on the non-cellulosic fibers sector was to raise the price of polyester by 5.39% above the base level and for production and consumption to increase by 0.40% and 0.43%, respectively. 171 Table 8.5 Percentage Change in Cotton Variables for a 10% Decrease in Production in 1.13253. Variable Region Impact Average Final Production Soviet Union -10.00 —10.00 -10.00 Price World 5.07 9.16 12.60 Production Argentina 0.54 10.22 12.57 Brazil 0.34 1.12 1.42 Cent. Africa 0.43 1.00 1.53 Mexico 0.00 5.36 11.13 Pakistan 0.15 1.42 1.53 Turkey 0.00 0.56 0.76 USA 0.00 2.77 3.76 World -2.07 -O.94 -O.69 Consumption Argentina -0.06 -0.63 -1.10 Australia -0.13 -0.34 -0.51 Brazil -0.70 -1.80 -2.59 Cent. Africa 031 -0.32 -0.39 China -O.23 -0.57 -1.06 EEC -0.25 -0.74 -1.05 Korea -0.68 -2.04 -2.85 Mexico -0.31 -2.61 -4.41 Turkey -0.60 -3.75 -5.81 USA -0.74 -1.13 -1.67 World -0.19 -0.48 -0.80 Ending World -5.76 -10.13 ~16.15 Stocks 172 Table 8.6 Percentage Change in Non-Cellulosic Fibers Variables for a 10% Decrease in Cotton Production in the USSR. Variable Region Impact Average Final (% changes) Price World 2.09 3.72 5.39 Production World 0.00 0.34 0.40 Consumption Argentina 0.15 1.31 2.10 Australia 0.28 2.21 3.77 Brazil 2.35 6.30 9.54 Cent. Africa 2.07 5.34 7.23 China 2.81 8.63 13.40 EEC 0.97 2.54 3.55 Japan 0.35 2.37 3.69 Korea 1.28 3.51 4.91 Mexico 0.24 2.10 3.71 Pakistan 0.00 0.00 0.00 Turkey 0.20 0.49 0.83 USA 0.88 1.23 1.55 Rest-of-the-World -2.67 -4.75 -7.30 World 0.00 0.35 0.43 8.2.4 Conclusions and Implications With the current emphasis within the USSR on environmental protection and soil conservation by improving crOpping rotations, USSR cotton production is expected to decline in the 19903. This is in spite of new innovations (e.g., the introduction of the Self- Financing Program and changes in land leasing arrangements) which are hoped to improve agricultural performance. The effects of the 10% fall in USSR cotton production on the model forecasts are presented in Figures 8.9-8.15. Figure 8.9 shows that while real cotton prices are forecast to fall between 1990 and 2000, the effect of declining production 173 in USSR is to slow substantially the rate of decline and to hold real prices relatively constant over the period 1995 through 2000. This result illustrated the responsiveness of world cotton prices to exogenous production shocks. Figure 8.10 shows how consumption forecasts are affected by the USSR production decline. Higher cotton prices reduce the level of forecast production by a fairly constant margin between 1990 and 2000, ranging between 35,000 tons and 160,000 tons below. The impact on the world cotton production forecasts are presented in Figure 8.11. While production forecasts fall initially as a result of the exogenous decline in the USSR production, higher forecast prices create a supply response from other regions. The effect is to close the gap between the two forecasts especially in the 1990 to 1995 period. The forecasts for cotton stocks are shown in Figure 8.12. Stocks are forecast to be lower between 1990 and 2000. Initially the major contributing factor is the decline in USSR production. However, the gap between the two forecasts tends to increase throughout the forecast period. This is because the decline in production exceeds the decline in consumption and the forecast of stock levels falls by over 1 million tons by 2000. The impact of the USSR production shock on the non-cellulosic fibers forecasts is presented in Figures 8.13, 8.14 and 8.15. While only small changes are shown for non- cellulosic fibers production and consumption, polyester prices are significantly higher than in the base forecast. The message from this simulation is that production in the USSR is a very important factor in price determination in the world market. For example, given a permanent 10% decrease in production, the world price rises by about 9.0%. This indicates that forecasts of price, both near- and long-term, must embody some prediction of USSR 174 cotton policy and producer incentive structures and some analysis of how these might impact on its production performance. Fi re 8.9 The Im act of a 10 Decrease in Cotton Production in the USSR on World Cotton Price. MUV Deflated (1985-100) \ 130 1 15 l l 1 l l; 4* _L 1 l 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —°— Base Simulation "‘t- Shock Fi re 8.10 The Im act of a 10 Decrease in Cotton Production in the USSR on World Cotton Consumption. 20.5 i ,i, 20 g l / ! I 0 19.5 n / s M , 19 T / o n s 18.5 18 L 1 1 m _L 1 i j 1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —°'— Base Simulation + Shock 175 Fi re 8.11 The Im act of a 10% Decrease in Cotton Production in the USSR on World Cotton Production 20.75 //I M 20.25 2 / : fl : 19.75 0 / 11 19.25 M f T 18.75 0 / fl ° 18.25 / 17.75 I— j 7* j— T I I I I 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —*— Base Simulation —+-' Shock Figure 8.12 The Impact of a 10% Decrease in Cotton Production in the USSR on World Cotton Stocks. ' ,\ i NA\\\ 3 \ 1— «304-z :o~———z 25 L 1 1 I _m l l ' 1 . 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —‘— Base Simulation —+— Shock oxxo 176 Fi re 8.13 The Im act of a 10 o Decrease in Cotton Production in the USSR on World 1391M; MUV Deflated (1985-100) 135 130 / 125 2//(///1//// / .1/ 1,. / ‘ ::»/// '— 115 110 1' L 1 1 L 1 1 1 1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year -—‘- Base Simulation —+"' Shock Fi re 8.14 The Im act of a 10 a Decrease in Cotton Production in the USSR on World Non-Cellulosic Fibers Consumption. 17.5 /l ”i” 15.5 I l I O " 15.5 M .i. O n 14.5 8 / J 135 l l I l l l l l 4 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —‘— Base Simulation —+— Shock 177 Fi re 8. 5 The Im act of a 10 o Decrease in Cotton Production in the USSR on World Non-Cellulosic Fibers Production. 18.5 .- 17.5 16.5 15.5 / 14.5 1 l I 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year 5:04.: ao—-*Z l l I 11 l l * Base Simulation *1— Shock 8.3 Simulation Three - A 10 Percent Increase in Cotton Production in China. 8.3.1 Backggound The Chinese cotton and textile sectors have undergone significant changes since the 19505. Prior to the mid-19605, the State’s agricultural development strategy emphasized the need to expand the use of improved industrially produced agricultural inputs, such as chemical fertilizers, equipment and machinery. State investment in agricultural production and inputs increased dramatically during the late 19503 and mid-19603. Input prices, which were determined administratively by the State, were cut by 30-50% during this period. Also, an Agricultural Bank was established in 1963 to allocate credit in rural areas. Production was further stimulated by higher producer prices. Between 1966 and 1976, the input orientation of agricultural development was 178 further intensified as physical directives again became the dominant mode of agricultural planning. However, agricultural policy was dominated more by political slogans than economic considerations and technical criteria. During this time, the major emphasis was on self-sufficiency in grains and as a result, the production of cotton stagnated along with the production of most other crops. Under new leadership, wide ranging policy changes began in 1978. Economic incentives were introduced and growers were given much greater freedom over planning and production choices. Between 1977 and 1983, average procurement prices for cotton increased progressively, rising 64.5% over the period. In response to the improved price incentives, production expanded rapidly, increasing from 2 million tons in 1977 to 6.25 million in 1984. Record stock levels resulted and led the State to reduce procurement prices for the following crop year, as well as to introduce a special consumer subsidy to increase low-grade cotton use in padded furnishings and clothing. Following two years of low producer prices and production in 1985 and 1986, stocks had reached a desirable level in 1987 and the State once again raised producer prices. Production in 1987 and 1988 remained fairly constant at about 4.2 million tons. Meanwhile, the demand for cotton had been strong following the eXpansion of investment and production capacity in the textile manufacturing sector. In 1989, the lack of available cotton supplies created a black market for cotton where the price paid by textile mills was as much as twice the State’s procurement price. Exports of cotton fell almost 40% between 1987 and 1988, while imports increased six-fold over the same period. The State responded to the excess demand by increasing procurement prices 34% for the 1990 crop. Also the State is providing cotton growers with heavily subsidized 179 fertilizer and diesel oil and improving the availability of other crucial inputs such as pesticides and plastic sheeting. A production target for 1990 had been set at about 4.3 million tons. Given the importance of the textile sector for employment and as a generator of foreign exchange, the State is expected to continue to support the cotton sector further into the 19903. Unofficial forecasts are for production to increase at 3% per year during the early part of the 19903. As well there is possibility of State support for a policy of increasing cotton production at the expense of wheat which can be purchased relatively more cheaply in the world market. 8.3.2 Simulation Objectives While production is expected to increase significantly in the future, it is hazardous to make guesses about the future direction of the Chinese cotton sector. Therefore, it is useful to measure of the effects on the world fiber market of a given change in China’s cotton policy, to which, as the 19803 have shown, farmers are highly responsive. The objective of this simulation is to quantify the effects a production change in China has on the rest of the world fibers markets. A simulation of the model was performed with Chinese production set at 10% above its historical level. This simulation provides useful information on how China may affected the world market in the recent past and how developments in China might impact on the cotton and non-cellulosic fibers sectors in the future. 180 8.3.3 Simulation Results The results of the simulation are reported in Tables 8.7 and 8.8. The initial effect of a 10% increase in Chinese production is to raise world production 1.82% above the base level. This increase in cotton production causes the price of cotton to fall by 1.02% which increases demand for cotton by 0.04%. The increase in cotton demand causes the demand for non-cellulosic fibers to decline and for the polyester price to fall 0.41% below the base levels. This polyester price increase partly offsets the impact of lower cotton prices on cotton demand. On average, the 10% production increase in China raises world production by 1.84%. This is the net effect of higher Chinese production and lower production in the rest of the world in response to lower cotton prices. World consumption rises by 0.42% above the base run. The changes in world production and consumption cause stocks to rise 17.94% above the base level and world price to decrease by 10.10%. In the non-cellulosic fibers sector, consumption decreases 0.30% below the base simulation, causing the price of polyester to fall by 3.63%. Lower polyester prices reduces production by 0.29% and slows the rate of cotton consumption increases. By the end of the 11-year simulation period, world cotton production is only 1.53% above the base level. This expansion is the net effect of the increase in China’s production and the decrease in production in the price-responsive regions. Despite lower cotton prices, consumption is only 0.91% higher. The consumption increase is dampened by lower polyester prices in the non—cellulosic fibers sector. The fall in non-cellulosic fibers demand continues to the end of the simulation period, which reduces production and price by 0.49% and 8.25%, respectively. 181 Table 8.7 Percentage Change in Cotton Variables for a 10% Increase in China’s Cotton Production Variable Region Impact Average Final (% changes) Production China 10.00 10.00 10.00 Price World 1.02 —10.10 2217 Production Argentina -0.11 -9.27 -13.30 Brazil -0.07 -1.10 -2.79 Cent. Africa -0.08 -1.63 -3.65 Mexico 0.00 -2.23 -5.50 Pakistan 003 -1.03 -1.95 Turkey 0.00 -0.23 -1.20 USA 0.00 -1.97 -3.80 World 1.82 1.84 1.53 Consumption Argentina 0.12 0.40 1.00 Australia 0.03 0.32 0.72 Brazil 0.14 1.65 3.53 Cent. Africa 0.01 0.18 0.42 China 0.04 0.49 1.01 EEC 0.05 0.68 1.43 Korea 0.13 1.87 3.82 Mexico 0.06 1.73 4.09 Turkey 0.12 2.90 6.19 USA 0.14 1.04 2.23 World 0.04 0.42 0.91 Ending World 5.47 17.94 31.23 Stocks China 36.19 39.91 58.34 Net Trade China -8.71 43.87 468.91 182 Table 8.8 Percentage Change in Non-Cellulosic Fibers Variables for a 10% Increase in China’s Cotton Production Variable Region Impact Average Final (% changes) Price World 0.41 3.63 8.25 Production World 0.00 0.29 0.49 Consumption Argentina 0.03 0.83 2.04 Australia 0.05 1.48 3.61 Brazil 0.46 7.21 14.87 Cent. Africa 0.40 5.22 10.94 China 0.54 8.38 18.60 EEC 0.19 2.47 5.33 Japan 0.07 1.83 4.31 Korea 0.25 3.43 7.39 Mexico 0.05 1.33 3.29 Turkey 0.04 0.53 1.16 USA 0.17 1.18 2.35 Rest of World -0.52 -4.33 -10.39 World 0.00 0.30 0.54 183 8.3.4 Conclusions and Implications The results show that China’s cotton production has a major impact on the world fiber market. Over the 11-year simulation of the model for every 1% increase in China’s production the world price of cotton falls, on average, about 1% and the price of polyester falls 0.36%. Therefore, if production increases at the unofficial forecast growth rate of 3% per year, the cotton price can be expected to fall substantially. Figure 8.16 shows the forecast of world cotton ending stocks following the 10% production increase in China. By 2000 world stocks are forecast to be more than 1 million tons above the base forecast, causing the world cotton price to be more than 20% below the base forecast level (Figure 8.17). The possibility of higher future cotton production in China has important implication for the price of polyester as depicted in Figure 8.18. This shows considerably lower polyester price forecasts than in the base scenario, with the difference between the two forecasts increasing over time. Figures 8.19 and 8.20 show the change in cotton production and consumption forecasts, respectively. Although most of the pricing and production decisions are made internally, China is now the largest player in the world fiber market. The importance of China in determining world price and thereby production and consumption levels in the price- responsive regions of the world is demonstrated by the model. While future cotton policies of China are unknown, the results show that developments in China during the early 19903 must be monitored closely and included in future fiber sector forecasting and policy analyses. 184 Figpre 8.16 The Impact of a 10% Increase in'Cotton Production in China on World Cotton Stocks. 5 M 4 5 I ° ‘/ I l i 3 4 \ M .i. O . n 3.5 s \l 13990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 ' Year * Base Simulation —*— Shock Fi re 8.17 The Im act of a 10% Increase in Cotton Production in China on World Cotton Price. MUV Deflated (1985-100) 135 130\ 125 \K\\ 120 \M 115 \ 110 105 1 ' 1 1 1 4 ' ’ ‘ 1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —°‘— Base Simulation —*—' Shock 185 Fi re 8.18 The Im act of a 10 Increase in Cotton Production in China on World Polyester Price. MUV Deflated (1985-100) 130 125 / 120 // i\ 115 ID ”\0 110 L 1 _L l 1 1 '1 1 I 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year “‘* Base Simulation "i— Shock Fi re 8.19 The Im act of a 10 Increase in Cotton Production in China on World Cotton Production. 20.75 / ' l 20.25 I ’ fl 3 19.75 . M V 1' 19.25 /// 0 n 3 18.75_/ 1825‘ 1 1 1 1 1 1 1 1 1 l 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year -°— Base Simulation —+— Shock 186 Fi e . O The Im act of a 10 Increase in Cotton Production in China on World Cotton Consumption. ENLZS I 19.75 19.25 18.75 / 18.25 / ./ 01:10-1- I ao—-—-g 17.75 ' 1 1 1 11 m 1 1 11 1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 ‘mar ‘—°‘ Base Simulation —+— Shock 8.4 Simulation Four - A 10 Percent Reduction in the United States Cotton Price 8.4.1 Background Since 1933 US cotton growers have been supported by several government programs. While the level of support given to growers has changed over the years, the policy orientation has remained the same. This has been to provide cotton producers with price incentives consistent with the global cotton market conditions and to adjust cotton acreage and production to meet market needs. Cotton growers are provided with a floor price through the loan rate. This a nonrecourse loan meaning that repayments can be made by delivering stored cotton to the Commodity Credit Corporation (CCC). Typically the loan rate is set at some proportion 187 of the average of recent past world cotton prices. In addition to the loan rate, growers who participate in the program by setting aside a certain amount of cotton acreage from production are entitled to a deficiency payment. This payment is equal to the difference between the target price (first established in 1973) and the loan rate. The target price is based on a cost-of-production formula as well as current market price trends. During the 19703 and 19803 a number of agricultural acts were brought into legislation’. Each of these Acts established new specifications for the loan rate, deficiency payments, target price, acreage diversion requirements and so on. However, while the specifications were changed through time to accommodate the prevailing market conditions, the actual instruments of support have been preserved 3. Other SUpport measures provided in recent agricultural legislation have included limits on the total amount of payments received by any one grower, disaster relief provisions and the payment-in-kind program. The most recent legislation came in the Food, Agriculture, Conservation, and Trade Act of 1990. This Act continues the market-orientated legislation established in the 1985 Food Security Act“. The Act includes provisions designed to make US producers competitive in the international cotton market, by fixing the target price for the 1991-95 period at the 1990 level at 72.5c/lb. This is in an effort to meet Gramm-Rudman deficit reduction targets, and to support domestic producers at levels consistent with expected 2 The major pieces of legislation were the Agricultural Act 1970, Agricultural and Consumer Protection Act 1973, Food and Agricultural Act 1977, Agriculture and Food Act 1981, Agricultural Programs Adjustment Act 1984, and Food Security Act 1985. The provisions of each of these Acts are discussed by Starbird et al (1987). 3 Perhaps the most significant policy change was the introduction of the payment-in-kind (PIK) program which was introduced in 1983. This was in re3p0nse to an unprecedented build up of stocks in 1982 and government deficiency payment expenditures of $520 million. 4 . . . ' I ' . ' . . . . . These conditions were mainly the result of the production boom 1n Chma wrth production increasmg from 3.3 million tons in 1982 to 5.9 million tons in 1984. 188 falling world cotton prices during the 1991-95 period. 8.4.2 Simulation Objectives In the past US cotton programs have had a major impact on world prices. For example, the establishment of target prices in the mid-19703 led to a reduction in planted acreage in 1975 and 1976 and was largely responsible for the 46% decline in the world price between 1975 and 1976. Then, following the introduction of the Payment-In-Kind (PIK) program in 1983, the dramatic reduction in United States stocks led to strong world prices in 1983 and 1984 despite the large expansion of production and stocks in China. The 1985 Food Security Act was aimed at reducing world prices and reducing the cotton stocks held in the United States. This policy was a major contributory factor in the 42% drop in the world cotton price between the 1984 and 1985 crop years. The 1990 Food, Agriculture, Conservation, and Trade Act came into law in November 1990. The most significant impact of the Act on the US cotton market is the fixing of the target price for the 1991-95 period at the 1990 level. Assuming a 4%-5% annual inflation rate for 1991-95, the average real target price for the 1991-95 period is eXpected to be about 25% lower than for the 1986-90 period. This represents a significant fall in the support level given to domestic producers of cotton. Given that real US target prices are expected to fall over the 1990-1995 period and the importance of the United States cotton sector on the world fiber market, the model was simulated with the prices in the United States production equations set 10% below the base simulation levels. The results on this simulation will provide useful estimates of the impact that the 1990 legislation will have on the world fiber market. 189 8.4.3 Simulation Results In order to test how such policy developments may impact on the world market, the model was simulated with US domestic prices set 10% below their historical values. The results are reported in Tables 8.9 and 8.10. The initial values are for the second period of the simulation. In the first period, no changes are experienced in any variables. This is because only lagged values of producer prices enter the supply equations. Initially the 10% fall in the US cotton price led to a decline in United States production by 3.16% below the base simulation. Lower supplies caused ending stocks to fall by 1.63% and for the world price to increase by 1.30%. The higher cotton price led to a reduction of world cotton demand by 0.58% and to a small expansion of production in other regions, partially offsetting the effect of lower United States supplies. The reduction in the demand for cotton created an expansion of non-cellulosic fibers demand, which caused the price of polyester to increase 0.42% above the base simulation. Higher polyester prices dampened the impact of higher cotton prices on cotton consumption. The average impact was. for United States cotton production to fall by 2.93% and world production to fall 0.31% below the base simulation. The reduction in supplies caused the world demand for cotton with world consumption falling 0.19%. As expected, lower cotton consumption increased the demand for non-cellulosic fibers which increased on average 0.13% above the base level. Strong demand increased the polyester price by 1.46% and production by 0.13%. 190 Table 8.9 Percentage Change in Cotton Variables for a 10% Decline in U.S. Cotton Prices Variable Region Impact Average Final (% changes) Price United States -10.00 -10.00 -10.00 World 1.30 3.70 5.30 Production Argentina 0.10 4.29 4.15 Brazil 0.08 0.45 0.73 Cent. Africa 0.10 0.39 1.00 Mexico 0.00 2.73 4.90 Pakistan 0.03 0.48 0.64 Turkey 0.00 0.21 0.36 USA -3.16 -2.93 242 World 058 -0.31 -0.15 Consumption Argentina Ending Stocks Australia Brazil Cent. Africa China EEC Korea Mexico Turkey USA World World -0.02 -0.03 -0.18 -0.01 -0.05 -0.07 -0.16 -0.07 -0.14 -0.14 -0.58 -1.63 -0.23 -0.13 -0.67 -0.09 -0.23 -O.28 -0.76 -0.95 -1.44 -0.43 -0.19 -3 .34 -0.54 -0.19 -0.96 -0.14 -0.26 -0.39 -1.06 ~1.40 -2.00 -0.55 -0.27 -454 191 Table 8.10 Percentage Change in Non-Cellulosic Fibers Variables for a 10% Decline in U.S. Cotton Pri_c<§ Variable Region Impact Average Final (% changes) Price World 0.42 1.46 2.12 Production World 0.00 0.13 0.15 Consumption Argentina 0.03 0.48 0.71 Australia 0.06 0.82 1.22 Brazil 0.77 2.81 3.85 Cent. Africa 0.70 2.06 2.87 China 1.17 3.50 4.68 EEC 0.25 0.99 1.40 Japan 0.10 0.90 1.37 Korea 0.34 1.36 1.93 Mexico 0.05 0.76 1.13 Turkey 0.05 0.22 0.31 USA 0.18 0.46 0.59 Rest of World 068 -1.82 -2.67 World 0.00 0.13 0.16 192 Over the long-run, the world price increased by 5.30% above the base level, with much of the fall in United States production offset by an expansion in other regions. Higher prices continued to constrain world consumption which fell only 0.27% below the base level. In the long-run the polyester price increased by 2.12%, stimulated by higher non-cellulosic fibers consumption following the decline in cotton demand. 8.4.4 Conclusion and Implications Overall the results indicate that effects of a decline in the US price will not have a substantial effect on the world market. The US production falls, on average, less than 3% below the base simulation, while world prices increase an average of 3.7%. Consumption is affected by less than 1% in all regions of the model. While a fall in both loan rate and target price are expected in the early 19903, the declines are not expected to be as large as the 10% in the simulation. Therefore the actual effects are expected to be much smaller than those reported in Tables 8.9 and 8.10. Also the decline in world production will tend to increase the world market prices and this will offset, to some extent, the decline in the US domestic price. Some of the effects of the 10% decline in US cotton prices on the model forecasts are shown in Figures 8.21-8.25. The forecast of production in the United States is significantly below the base forecast, especially in the early 19903. On average, the price shock reduces the United States cotton production forecast by about 100,000 tons. Figure 8.23 shows that the forecast of deflated cotton price falls when the United States price is set 10% lower, but that the rate of decline is less than in the base scenario forecast. The effects on the cotton stocks and consumption forecasts are shown in Figures 8.25 and 8.26, 193 respectively, and illustrates that, in generally, the effects are quite small. Fi re 8.21 The Im act of a 10 Decrease in Cotton Price in the United States on United States Cotton Production. 3.7 M / i 3.5 1 ' l i . 0 n M 3.3 .i. 0 2 3.1 a 2.9 L _I I 11 J1 J1 l_ 1 J1 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year "“ Base Simulation —l— Shock 194 Fi re 8.22 The Im act of a 10 o Decrease in Cotton Price in the United States on World Cotton Production. 20.75 20.25 19.75 1.... / / 0:10-1- 3 :o—-——-z 18.25J _L L I I J1 _I _L I L 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —°— Base Simulation —+- Shock Fi re 8.23 The Im act of a 10% Decrease in Cotton Price in the United States on World Cotton Price. ‘ 135 4m 130 (I: \ k 125 9 .. 120 \/\‘[ 115 1! I I _L 11 l l_ l I 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —'— Base Simulation ‘l— Shock 195 Fi re 8.24 The Im act of a 10% Decrease in Cotton Price in the United States on World waste—rm; 130 125 120 OX\O 115 — I I L I 110 I' J I I I 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —‘— Base Simulation —1— Shock Fi re 8.25 The Im act of a 10% Decrease in Cotton Price in the United States on World Cotton Stocks. 402-1 4 - M l I 3.8 . . i V \\\ 0 . " 3.6 M \\1\ .i. o 3.4 . \ n 8 3.2 \ 3 I I L I 1 1L 1 I g 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year —°— Base Simulation —l— Shock 196 Fi re 8.26 The Im act of a 10 o Decrease in Cotton Price in the United States on World Cotton Consumption. 2: I 2' 19.5 / 19 // I 11990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year 3:10-1- : =o~—--Z “'— Base Simulation —*" Shock 8.5 Simulation Five - Anal sis of the Multi-Fiber A eement 8.5.1 Background The exporting of textile and clothing (T&C) products is very important to the economies many developing countries, especially as a source of foreign exchange and employment. Given that T&C production technology is labor-intensive, developing countries have achieved a comparative advantage in production relative to the industrialized countries where wages are much higher. In 1986 developing countries produced about one-half of total world supplies of manufactured T&C products, while accounting for less than 17% of total manufactured production. More than 50% of the T&C exports of developing countries are purchased by industrialized countrie3--especially Japan, the EEC and the United States. The largest T&C exporters are Hong Kong, Korea 197 and Taiwan, and recently China has emerged as a major supplier to the world market. Despite the importance of the international T&C markets to the developing countries, tariff and non-tariff restrictions have been a constant feature of T&C trade. Trade between the deveIOping country exporters and the industrialized countries has been restricted since January 1974 through the Multi-Fiber Agreement (MFA), and before then by the Short- and Long-Term Arrangements (1961-1973). Negotiated under GATT, the MFA provides for the imposition of bilateral quotas on deve10ping country T&C exports to the industrialized countries. Such agreements are negotiated to avoid "market disruption" in the industrialized countries from (i) sharp increases in the import of a particular T&C item from a particular source, (ii) import prices being below domestic prices, and (iii) imported T&C items causing injury to domestic producers 5. So far four MFA agreements have been negotiated since MFAI was introduced in 1974. The current MFA (MFA IV) is due to expire in 1991. With each new agreement the controls have become more stringent, with more developing countries restricted and wider coverage of T&C items falling into restricted categories. During the MFA I period (January 1974 through December 1977) the United States placed heavy restrictions on many countries, while the EEC and Japan imposed relatively minor quantity restrictions. 5 Some researchers have Questioned how the MFA became to be ne8°tiatcd ““dcr GATT While being in violation of its most- favored-nation clause and requirements that, except under special circumstances, trade restriction should be in the form of tariffs only (Keesing and wolf (1980), Sampson (1986) ). Possible explanations for this are that T&C trade has had a long history of trade restrictions and is dominated by the United States as the major player in the market, “Pam” in the 19505 and 19608 when the foundations of the MFA were being laid. At this time the United States textile sector employed between 15 and 20 P9"cent 0f the 1°13! manufacturing sector labor force and was dominated by very powerful interest groups. Further, Keesing and Wolf argue that many developing countries had good reason to agree to the restrictions because they felt that uncontrolled restrictions would impair the long-run development of their T&C sectors. 198 In the MFA H (January 1978 through December 1981) the EEC substantially tightened its quota levels and in MFA III (January 1982 through July 1986) a greater number of unilateral and bilateral agreements were installed. MFA IV includes further restrictions. 8.5.2 Past Studies of the Economic Effects of the MFA A number of studies have appeared which estimate the effects of the MFA on both importing and exporting countries. In terms of the effects on importing countries (i.e., the United States, Japan, and the EEC), Cline (1985) estimated that MFA cost United States consumers $20.3 billion in 1985, while saving 434,200 jobs ($47,000 per job saved). Hufbauer M (1984) reported that the consumer cost was $27 billion in 1984, saving 640,000 jobs at a cost of $42,000 per job. Both studies concluded that the MFA was an inefficient method of protecting employment from cheap T&C imports. Another set of studies has looked at the costs incurred by the developing countries as a result of the MFA. For example, UNCTAD (1986) estimated that by removing all restrictions on T&C imports into the United States, BBC and Japan, export revenues would increase by $15 billion--and increase of almost 100%. Whalley, using a general equilibrium model, estimated the loss to be $11 billion, while Kirmani (1984) showed that by removing tariff and non-tariff barriers from imports of T&C products to OECD countries, the textile and clothing export revenues of the developing countries would increase 82% and 92%, respectively. Other effects of the MFA observed by other researchers are trade diversion, upgrading of exports and the impact of the MFA on economic development. Some deve10ping countries fill all their quota under the MFA (e.g., Hong Kong and Korea) while others use only a small part of their allocation. This provides incentives for the restricted 199 countries to set up T&C production capacity in countries where quotas are not binding. This has provided a stimulus to economic development in some countries. For example, according to Spinanger (1987), with the help of a Korean company, Bangladesh increased its T&C from nothing in 1979 to $500 million in 1987. However, this rapid growth led to the imposition of quotas under the MFA and as a result 400 plants were forced to close down. 8.5.3 Simulation Objectives Most of the past studies into the effects of the MFA have focused on measuring the costs and benefits of the agreement and to whom they have accrued, such as consumers and producers of T&C products. Also research has been aimed at measuring the cost- effectiveness of the agreement on employment and job-saving. No research, as far as the author is aware, has tried to estimate the effects of the MFA on the supply and demand of raw fibers. Clearly with the restrictions imposed by the MFA, the demand for T&C products has fallen, lowering the demand for raw fibers by the T&C manufacturers. In turn, the lower demand by manufacturers will have reduced the prices of raw product which will impact on suppliers, such as cotton growers and non-cellulosic fibers producers. In view of the recent negotiations of the MFA in the current Uruguay round of GATT, it is timely to measure fully the impact that the T&C trade restrictions impose. One such impact, not yet explored in the literature, is the effect of the MFA on the primary producers of raw fibers. In this simulation the effects of the MFA on primary producers, intermediate consumers and raw fiber prices are measured. The results may provide a better picture of how the MFA affects other sectors of the world fiber market. 200 8.5.4 Problems of Modeling the MFA For a number of reasons modeling the impact of the MFA on the fiber market was found to be extremely difficult and eventually a rather crude method of capturing the MFA in the model was employed. The difficulties encountered were as follows. Restrictions on trade in textiles has evolved slowly since the early 19603 within the scope of the Short- and Long-Term Arrangements followed by the four MFAs. With each new agreement, restrictions have become more stringent, but it has only been with MFA III (January 1982 - July 1986) and MFA IV (August 1986 - July 1991) that restrictions have been significantly prohibitive. The FAO data set used to estimate the demand side of the fiber model ends in 1986 for developing countries and in 1987 for industrialized countries and therefore does not cover the most restrictive time period. Another problem is that the fiber model is based on quantities measured in tons of raw fiber equivalents. The MFA restricts the numbers of specific manufactured textile products and clothing items, such as the number of pairs of gloves or the number of table- cloths; it is almost impossible to measure these items in terms of fiber content and weight. Therefore, the MFA cannot be quantified in a way compatible with the econometric model. The MFA is negotiated bilaterally whereby trade flows between individual countries are specifically restricted (e.g., there are restrictions on the number of shirts imported into the United States from Hong Kong). To model the impact of such restrictions requires that bilateral trade flows be estimated for each importing and exporting country party to the MFA. Time series data are insufficient to allow such trade flows to be modeled. Further, where restrictions have become binding, countries have managed to maintain trade 201 levels by either exporting through a third country whose exports are not restricted, or by establishing new processing plants in such countries. These ’leakages’ are widespread. A final problem is that most of the textile trade has grown rapidly since the 19603. The strong trend in these data prevented estimation of useful response parameters. Several approaches were attempted in an effort to capture the effects of the MFA in the fiber model. Given that the MFA cannot be quantified in terms of raw fiber, dummy variables were constructed for each for the three MFA regimes operating during the estimation period (i.e., MFA I, 1974-1977; MFA H, 1978-1981; MFA III 1982-1986). It was hypothesized that the MFA, by restricting imports, would result in higher textile prices which would, in turn, reduce fiber availability. The textile and clothing component of the Consumer Price Index for the United States and the EEC (the two major importers under the MFA) were obtained and regressed on the following variables: prices of cotton and polyester, a measure of efficiency in the textile manufacturing sector, wage rates, and MFA dummy variables. While the equation fitted the data well for both regions, the MFA dummy variables were incorrectly signed. This is because, despite what effect the MFA has had, deflated textile prices have declined consistently throughout the estimation period. Even the regression of textile price growth rates against the MFA dummy variables revealed no statistically significant relationships. Further, the price of textiles was not significant in the total fiber use equations in both regions. Another approach was based on the assumption that the decline in textile prices in the United States and the EEC is due to the penetration of cheap imports into these markets and that this penetration might have been slowed over time with the imposition of tighter MFA restrictions. Again, this hypothesis was not borne out by the data, with 202 market penetration increasing at an increasing rate, despite increasingly stringent MFA controls. Thus no statistical evidence could be found that supported the hypothesis that the MFA has led to increases in the prices of textile and clothing products, or that the agreements have slowed the penetration of imported products into the United States and the EEC markets. 8.5.5 Method of Incorporating the MFA into the Model While analysis of the MFA cannot be easily handled in this model some crude estimates of the impact of the MFA on the fiber markets were obtained from a model simulation. The approach taken was to measure the percentage consumption reduction in the importing countries which resulted from the MFA (using elasticities derived in other studies) and then to use the model to measure how this decline affected fiber prices, production and consumption. This procedure involved a two-step approach. First, tariff equivalents of the MFA quotas were taken from Pelzman (1988), showing the percentage change in the price of textiles resulting from the trade restrictions imposed by the agreements. Second, an elasticity of textile demand with respect to the price of textiles was used, based on the study by Houthakker (1965). Combining these elasticities gave the percentage change in textile consumption resulting from the MFAs. The model was then simulated with demand set below the historical level according to these percentage declines and the results compared with the base simulation. The effect of the MFA was incorporated into the model using the results from a study by Pelzman (1988). The approach taken by Pelzman was to estimate a model of 203 textile and clothing trade between developing countries and the United States which takes into account market disequilibria that is introduced by the MFA quotas“. Supply, demand and price relationships are estimated separately for trade-restricted and trade~unrestricted markets. The consumption of textiles and clothing in both markets is estimated as a function of prices in both restricted and unrestricted markets as well as of prices in the importing country and variables capturing the levels of economic activity. The supply function for the restricted market is estimated for two scenarios. First, when the quantity supplied is below the quota level, equilibrium prices and quantities are derived from the intersection of the import supply and import demand relationships. Second, when the restriction is binding, Pelzman estimated a predicted value of excess demand or supply using the tobit two-stage method (Maddala, 1983) as a function of prices in the restricted, unrestricted and import markets and of the level of the quota. The estimate of excess supplies or demand was then used in a pricing equation for the restricted market. This price equation was used to generate equilibrium prices that would have existed in the absence of the trade restriction. Supplies in the unrestricted market were estimated as a function of the prices in the restricted and unrestricted markets. The equilibrium price in the unrestricted market is determined by the market clearing identity. Pelzman reports estimates of tariff equivalents of the MFA for numerous textile and apparel items based on the three digit textile category system for the period 1979 through 1986. These estimate the percent change in price as a result of the MFA. These 6 Pelzman criticized previous studies, such as those by Halbauer g1] and Tarr and Morke, which estimate the dead-weight loss of the MFA based on market clearing partial equilibrium models. The constraints on trade through quotas creates excess demand at the price prevailing in the absence of any restrictions. Thus the market equilibrium, crucial to the standard methods of estimating the welfare costs of trade distortions, is violated. 204 differ substantially over time and between the individual textile and apparel items. The tariff equivalents of the MFA quotas were estimated for the United States for the period 1979 through to 1986. This period covered most of the MFA H (January 1978 - December 1981) and the entire MFA IH (January 1982 - August 1986). Pelzman provided tariff equivalent estimates for over 80 separate textile and apparel items. These were aggregated using a weighted average, with weights assigned to individual items according to their proportion of the total value of imports in each year. The aggregated tariff equivalents are reported in Table 8.11. Since no estimates were provided by Pelzman for the BBC, the reduction in fiber consumption for this region was based on the percentage declines in the United States. The elasticity of textile demand with respect to the price of textiles was set at -0.282. This estimate has been used widely in other studies (e.g., T arr and Monke, 1984; Erzan, Goto and Holmes, 1989). 8.5.6 Simulation Results The effects of the MFA H and MFA III on the world fiber market are presented in Table 8.11. The impact of the MFA was to raise textile prices, ranging from an increase of 12.11% in 1986 to 19.10% in 1981. This increase reduced total fiber consumption in the United States and the EEC as shown in Figures 8.27 and 8.28. In both regions during the 1979-81 period (MFA H), consumption was on average 4.98% lower than it would have been without the agreement, while consumption averaged 3.84% lower for the MFA IH period (1982-86). The reduction in fiber consumption in the EEC and United States led to a small decline in world cotton consumption, ranging from 0.58% in 1983 to 1.23% in 1979. The 205 reduction in consumption caused cotton stocks to accumulate which led to a decline in price (Figure 8.29). Over the MFA II period the price of cotton fell an average of 4.87% as a result of the agreement, while during the MFA IH period price averaged 5.51% below the base level. The decline in price caused a reduction in cotton production. However, the production response was less than 1% below the base level in each year of the simulation. In the non-cellulosic fibers sector the results were similar to those in the cotton sector (Table 8.11). The reduction in fiber consumption in the United States and the EEC caused world consumption of non-cellulosic fibers to decline by up to 0.7% below the base level. The decline in consumption caused the price of polyester to fall. The price decline for the MFA II period was 5.01, and there was a decline of 4.44% during MFA HI (Figure 8.30). The lower prices resulted in a small reduction in production, but the decline never exceeded 1% below the base level. With cotton and polyester prices declining by different percentages, consumption of individual fibers in each region either increased or decreased. For example, in Argentina, Australia and Mexico during the MFA H period, the effect of lower cotton prices out-weighed the effect of lower polyester prices, so that cotton consumption increased. Conversely, during the MFA III period, the response to relative prices was reversed, with non-cellulosic fibers consumption increasing at the expense of cotton. Overall the consumption effects of the MFA on the countries which have not imposed restrictions were quite small, with most changes less than 1% above or below the base simulation levels. However, the effect on consumption in the United States was much higher. 206 Table 8.11 Percent Chan e in Cotton Variables Associated with the Multi-Fiber Agzeement, 1979-1986. Variable Region 1979 1980 1981 1982 1983 1984 1985 1986 MFA TariffI US & EEC 16.93 16.90 19.10 162.5 13.39 13.29 12.98 12.11 Equivalent(%) Total Fiber2 US & EEC -4.77 -4.’/'7 -S.39 458 -3.78 -3.75 -3.66 -3.41 Use Cotton Price World -3.38 -5.62 -5.61 -6.90 -7.47 -4.74 -3.63 -4.83 Production World —0.03 -0.26 -0.56 -0.65 -0.89 -0.82 -0.71 -057 Consumption Australia -O.20 0.02 0.02 0.12 0.17 0.08 0.01 0.01 Brazil 0.47 0.92 1.11 131 150 1.09 0.83 0.93 China 033 0.03 0.04 0.19 0.27 0.15 0.02 0.02 EEC -5.12 -4.74 -5.34 -4.36 -3.45 -3.57 -3.63 338 Korea -0.95 0.09 0.11 0.62 1.06 052 0.08 0.09 Merdco -0.46 -0.37 -0.27 0.04 0.44 0.56 0.51 0.45 Turkey 026 0.94 154 2.33 3.26 3.64 3.25 2.86 U.S. -5.07 -4.30 -5.12 -3.94 -2.95 -3.36 -3.58 -3.34 World —1.23 -1.01 -1.08 -0.77 -0.58 -0.73 -0.69 —0.65 Non-Cellulosic Fibers Price World -S.07 -4.93 -5.04 -4.92 -4.74 -4.20 -3.93 -4.40 Production World 0.00 -0.60 -0.63 -0.74 -0.64 -0.63 -056 -0.37 Consumption Argentina 0.30 0.22 0.16 -0.01 -0.20 -0.27 -0.25 -0.24 Australia 0.44 0.32 0.22 —0.05 -0.36 -0.47 -0.42 -0.39 Brazil 3.41 -0.30 -052 —2.16 -2.84 -2.45 -0.30 -0.35 China 4.34 -0.35 -0.77 -2.83 -3.18 —2.18 -0.36 -0.39 Cent Afr. 2.50 -0.26 -0.47 -1.63 -2.53 4.34 —0.22 -0.26 EEC -3.38 4.89 -554 -5.41 4.94 4.39 -3.76 -353 Japan 0.50 0.30 0.13 -0.23 -0.60 0.65 -0.48 -0.37 Korea 1.91 -0.16 -0.23 -l.l6 -1.61 -0.92 -O.15 -0.16 Mexico 0.37 0.28 0.21 -0.01 -0.28 -0.38 -0.36 -0.35 Turkey 0.34 -0.03 —0.04 -0.18 43.24 -0.13 -0.02 -0.03 U.S. 4.84 -5.59 -6.05 -5.53 -4.83 -4.40 -3.93 -3.64 World 0.00 -0.57 -0.61 -0.70 -0.63 -0.61 -058 -0.41 1 Tariff equivalent of MFA is the percentage increase in the price of textiles resulting from the MFA trade restriction. 2 Note: the decline is equal to the tariff equivalent ’ the elasticity of fiber demand with respect to textile price (= 0282). Fi re 8.27 Im act of Multi-Fiber A cement on United States Total Fiber Consum tion 1979-1986. 207 5.5 // / a364 z :o——-§ 01 T 4.5 “If \I/ T 4 l I L I I L 1979 1980 1981 1982 1983 1984 1985 1986 ‘%m —*"' Base Simulation ‘t- Shock Fi re 8.28 Im act of Multi-Fiber A eement on EEC Total Fiber Consum tion 1979- 1986. 5.2 5\ ,,,‘\ \ '8‘ I I l ' I o \ \ / n 4.6 i. 4.4 \/'//———/ 0 n e \ 4'2 V/lr 4 1 l I l I l 1979 1980 1981 1982 1983 1984 1985 1986 Year ‘—°— Base Simulation + Shock 208 Fi re 8.29 Im act of Multi-Fiber A cement on World Price of Cotton 1979-1986. 200 /\ /\ J/ \ //X\ 180 V‘ \ k 0 ,. 14° \\ 7 120 \v/ 100 1L J1 I I I I 1979 1980 1981 1982 1983 1984 1985 1986 Year —°— Base Simulation —l— Shock Fi re 8.30 Im act of Multi-Fiber A cement on World Pol ester Price 1979-1986. 180 170 //\\ //\\ 160 // \/ \\ c 150 / k //l/ \\ 130 ,F N- 120 110 ' ‘ ' ’ ‘ I 1979 1980 1981 1982 1983 1984 1985 1986 Year —'- Base Simulation ‘1— Shock 209 8.5.7 Conclusions and Implications Given the crude method used to derive these estimates, it is important to down-play the results somewhat. However, some general conclusions can be made. For example, the MFA has reduced world prices of cotton and polyester by, on average, about 5% between 1979 and 1986 (within a range between 3% and 7%). The analysis does not cover the most restrictive period since 1986 and therefore the current impact of the agreements on world fiber prices is likely to be at the upper end of this range. While the impact of the MFA was to reduce raw fiber prices by around 5%, the impacts on production and consumption were small. As a result of the agreements, cotton and non—cellulosic fibers production and consumption were changed less than 1% in most years. These results lead to the conclusion that while the effect of the MFAs have been substantial in terms of welfare losses and economic inefficiency in the T&C manufacturing and trade sectors (Pelzman, Tarr and Monke, Keesing and Wolf), the effects on the raw fiber market have not been large. This reflects the small size of the price transmission elasticity between raw fiber and textile prices and of the elasticities of raw fiber supply and demand with respect to price. However, the results are a useful contribution to the overall debate on the economic impact of the MFA that has been ignored previously by researchers. 8.6. Summagy The simulation results discussed in the section provide many important insights into the how the world fiber market operates and where it is headed in the future. Especially interesting are the forecasts for the period 1990 through 2005, while the simulations 210 involving policy shocks in individual countries (i.e., the USSR, China and the United States) provide information on how the forecasts must be modified in light of new developments in these markets. The MFA simulation sheds light on an important aspect of the agreement that has not been the subject of earlier research. 9. Sensitivity Analysis 9.1 Introduction The simulation results are conditional on the model’s assumptions and equation specifications. These were described fully in earlier sections. Parameter estimates were also presented and compared with those from other studies, which gave rise to a fairly good set of validation statistics. While the results from previous stages in model building lead to confidence in the model’s results, it is important to assess how robust the model forecasts and policy analyses are to different parameter estimates. A model’s robustness can be tested with sensitivity analysis, which is defined by Anderson (1974) as ’testing of the robustness of the model through recognition of its imperfections’. Sensitivity analysis shows how the simulation results change under different settings of the parameter estimates. If the model’s output is shown to change little with adjusted parameter estimates, then the results can be considered robust and more confidence can be placed in the forecasts and policy analyses. Conversely, if the simulation results are shown to differ substantially with changes to unsure parameter estimates, then the forecasts and policy impacts should be treated with caution. In this section, the results from a number of sensitivity analyses are reported, based on the discussion and a procedure outlined by Anderson. 9.2 Procedure Complete testing of the model’s sensitivity to changed parameters requires performing an almost infinite number of analyses. The dimensionality of reporting the 211 212 results can easily get out of hand in terms of, (i) the number of unsure parameters, (ii) the number of performance variables, (iii) the number of accounting intervals, and (iv) the number of measures of sensitivity. Therefore, it is necessary to limit the analysis by selecting a subset of all the possible analyses. The selection for the fiber model relate to each of (i)-(iv) above and are discussed briefly below. Changes in unsure parameters were confined to equations explaining five key variables in the model--the world cotton price, per capita total fiber use in the United States, cotton production in China, the world polyester price, and world non-cellulosic fibers production. These were considered the most important equations in determining the simulation results. The choice of performance variables was limited to cotton and non- cellulosic fibers prices, world production, and world consumption. These were selected as variables of most interest in the policy analysis and forecasting exercises. Following Anderson, the accounting interval by which parameters were changed is one standard deviation. In each equation, a single parameter--that of an important explanatory variable-- was changed by this amount in each direction’. Sensitivity to changing parameters on the performance variables is measured by merely reporting the results from the models with changed parameters along side the original model results. This allows the absolute changes and percentage changes in the performance variables to be assessed with ease. I Initially, for each equation, all parameters were changed one standard deviation in the direction giving the greatest change in the performance (independent) variable (i.e., to increase the value of a performance variable, parameters with a positive coefficient were raised by one standard deviation, while those with a negative coefficient were lowered by one standard deviation). However, for some equations it was found that the models failed to converge when all parameters were adjusted. For example, the cotton price equation is estimated in terms of logarithms. Adjusting all the coefficients by one standard deviation resulted in huge changes in CI'PWOR when the exponential was taken. This suggests that the model is highly sensitive to the parameter estimates of this equation. 213 9.3 _R_egm§ 9.3.1 Sensitivity of Results to the Equation Explaining World Cotton Price The equation explaining the world cotton price (equation 5.7) is, LN CI'PWOR = 7.24 - 0.78 LN CI'ESWORXCHI + 0.92 LN MUV + 0.31 LN CI‘CONWORXCHI - 0.86 LN TIME (5.7) (-6.32) (5.12) (1.75) (-3.33) R—SQUARED (CORR): 0.94 SEE: 0.03 DW: 2.09 PERIOD OF FIT: 1964-88 Where: CI'PWOR = World cotton price (Outlook Index 'A'), CI'ESWORXCHI = World cotton stocks excluding stock held in China (also see equation 5.8), MUV Manufacturing unit value , CI‘CONWORXCHI = World cotton consumption excluding consumption of China (also see equation 5.9), TIME = Time variable, LN = Indicates variable transformed into logarithms. A base model was simulated with parameters set at these levels. Sensitivity of the model’s results to the parameter estimate on the stocks variable (CI'ESWORXCHI) was tested by increasing and decreasing it by one standard deviation, giving high and low simulated values of CI‘PWOR. To raise CTPWOR, the cotton price equation was adjusted as shown in equation 9.1., and substituted for equation 5.7. This created a new model (model A) which was then simulated for policy analyses and forecasts. LN CI'PWOR = 7.24 - 0.66 LN CI'ESWORXCHI + 0.92 LN MUV + 0.31 LN CI'CONWORXCHI - 0.86 LN TIME (9.1) A second model (model B) was created by replacing equation 5.7 with equation 9.2 below. This contains a parameter adjustment on the stocks variable, giving low simulated values of the cotton price. LN CI'PWOR = 7.24 - 0.90 LN CI‘ESWORXCHI + 0.92 LN MUV + 0.31 LN CTCONWORXCHI - 0.86 LN TIME (9.2) ZUnit value in US dollar terms of manufactures exported from the G-5 countries (France, Germany, Japan, United Kingdom and the United States) weighted proportionally to the countries‘ exports to the developing countries. -r.._ «2: 214 The results from the original and adjusted forecast simulations are reported in Table 9.1a. For 1990, the cotton price forecasted by model A is about 14% higher than by the base model. Similarly, the cotton price is about 14% lower in model B. Between 1995 and 2005, in all three models, the cotton price declines, and by the end of the sixteen -year simulation period, the cotton price forecasts from the adjusted models are about 12%-14% different from the base model forecasts. The narrowing of the gap between base and adjusted models is due to differences in the levels of the stocks variable in each case. The results for cotton production are consistent with results for the cotton price. A higher forecast of the real cotton price by model A, leads to a higher cotton production forecast for 1990, increasing to 19,369 thousand tons, compared to 18,342 thousand tons forecasted by the base model. This is a difference of about 5.3% and this gap remains fairly stable over the sixteen-year simulation period. For model B, production forecasts are lower, with the difference never exceeding more than 6%. Cotton consumption forecasts are consistent with the price forecast adjustments reported above. From the base model, for 1990, consumption is forecasted to be 18,043 thousand tons, increasing to 21,634 thousand tons in 2005. Model A gives a forecast of 19,906 thousand tons in 1990, increasing to 20,365 by the end of the sixteen-year forecast period, while using model B, the forecast are 19,179 thousand tons in 1990, and 20,903 thousand tons in 2005. Overall cotton consumption forecasts from the adjusted models are always within 5%-6% different from the base model forecasts. 215 Table 9.1a. Impact on Forecasts of Parameter Chances in the Enuation Explaining the World Cotton Price. Variable 1990 1995 2000 2005 World Cotton Pricel/ Base Model” 133.0 119.1 117.7 116.1 Model A3/ 151.6 135.8 134.2 132.4 Model B” 114.4 102.4 101.2 9.8 World Cotton Consumption5/ Base Model 18,043 19,151 20,328 21,634 Model A 16,906 17,971 19,127 20,365 Model B 19,179 20,330 21,528 22,903 World Cotton Production” Base Model 18,342 19,517 20,626 21,826 Model A 19,369 20,429 21,746 22,999 Model B 17,315 18,429 19,505 20,652 World Polyester Pricey Base Model 117.6 112.8 127.0 149.9 Model A 122.9 1185 133.8 1583 Model B 112.3 107.1 120.7 141.5 World Non-Cellulosic Consumption” Base Model 13,770 15,706 17,219 18,615 Model A 13,091 14,819 16,166 17,427 Model B 14,449 16,593 18,272 19,803 World Non-Cellulosic Productions/ Base Model 14,870 16,806 18,319 19,715 Model A 15,003 16,991 18,533 19,961 Model B 14,737 16,621 18,105 19,469 V Deflated by MUV, 1985=100. 2/ Base Model - Model with parameters of the cotton price equation set at base model level (equation 5.7). 3/ Model A - Model with parameters of the cotton price equation set according to equation 9.2. 4/ Model B - Model with parameters of the cotton price equation set according to equation 9.3. 5/ ’000 metric tons. The adjusted forecasts for the non-cellulosic fibers market are also reported in Table 9.1a. From model A, for 1990, the polyester price is about 4.5% higher than in the base model, while it is about 4.5 % lower in model B. A similar trend in the polyester price forecasts was found using the adjusted cotton price equations. That is, the real price of 216 polyester falls in the first part of the simulation period, and then increases after the mid- 19905. As a consequence of dynamic relationships captured in the model, the gap between the base forecasts and the adjusted forecasts narrows, so that by the end of the simulation period, there is about a 4.5% difference between the adjusted models and the base model. The non-cellulosic fibers production and consumption levels differ from the base model in a consistent manner with differences between the base and adjusted forecasts of the polyester price. Over the entire simulation period, the gap between the base and adjusted forecasts of non-cellulosic fibers production and consumption never exceeds 2%. The results show that the model forecasts are fairly robust to changes in the parameter on the stocks variable in the cotton pricing equation. However, the price forecasts are more sensitive than the quantity forecasts, and the adjusted model cotton price forecast does differ by more than 15% from the base model in some years. The effects of the parameter change in the cotton price equation on policy analysis are shown in Table 9.1b. In general, the results are fairly similar between the original and adjusted models. However, an interesting observation is that the policy effects are larger with model B than the base model, and smaller with model A. This is because, by adjusting the parameter on the stocks variable, creates a model that is more inelastic with respect to price (model B). Therefore, this model is relatively more sensitive to policy shocks, since price has too move more to equate consumption and production for any given quantity change. This results in larger policy effects. For example, taking the cotton price results for simulation 2, we observe that the impact, average and final percentage changes are 5.07, 9.16 and 12.6, respectively. For the more inelastic model (model B), the equivalent results are 5.89, 11.37, and 15.79, respectively, and for the more elastic model 217 (model A) the results are 4.45, 7.50, and, 10.06, respectively. Table 9.1b. Impact on Policv Analvsis of Parameter Chances in the Enuation Explaining the World Cotton Price 1[. Variable Simulation 2” Simulation 3’/ Simulation 4‘/ Impact Average Final Impact Average Final Impact Average Final World Cotton Price Base Modelj/ 5.07 9.16 12.60 -1.02 -10.10 -22.17 1.30 3.70 5.30 Model A6/ 4.45 750 10.06 -0.89 -7.66 -16.66 1.14 4.23 4.12 Model 87/ 5.89 11.37 15.79 -1.19 -13.33 -29.47 1.51 4.72 6.86 World Cotton Consumption Base Model -0.19 -0.48 -0.80 0.04 0.42 0.91 -058 -0.19 -0.27 Model A -0.20 -056 -0.95 0.04 051 1.11 -0.68 -0.14 -0.24 Model B -0.18 -0.49 084 0.04 0.45 0.98 -055 -0.12 -0.21 World Cotton Production Base Model -2.07 -0.94 -0.69 1.82 1.84 153 -058 -0.31 -0.15 Model A -1.96 0.74 -0.46 1.72 1.76 1.41 -055 -0.26 -0.08 Model B -2.19 -0.82 -0.51 1.93 1.95 157 -0.61 -0.29 -0.07 World Polyester Price Base Model 2.09 3.72 5.39 0.41 3.63 8.25 0.42 1.46 2.12 Model A 2.00 350 5.10 0.39 3.39 7.79 0.40 1.31 1.98 Model B 2.19 3.95 5.79 0.43 3.97 8.81 0.44 1.61 2.31 World Non-Cellulosic Consumption Base Model 0 00 0.35 0.43 0.00 0.30 054 0.00 0.13 016 Model A 0.00 0.36 0.47 0.00 0.31 0.59 0.00 0.13 0.16 Model B 0.01 0.34 0.42 0.00 0.27 052 0.00 0.14 0.16 World Non-Cellulosic Production Base Model 0.00 0.34 0.40 0.00 0.29 0.49 0.00 0.13 0.15 Model A 0.00 0.34 0.42 0.00 0.29 0.48 0.00 0.13 0.15 Model B 0.01 0.34 0.45 0.00 0.29 0.50 0.00 0.13 0.15 I/ Results reported in terms of percentage change between base and shock. 2/ Simulation 2 - 10% decrease in USSR cotton production. 3/ Simulation 3 - 10% increase in Chinese cotton production. I/ simulation 4 - 10% decrease in U.S. cotton price. Base Model - Model with parameters of the cotton price equation set at the base model level (equation 5.7). 6/ Model A - Model with parameters of the cotton price equation set according to equation 9.2. 7/ Model B - Model with parameters of the cotton price equation set according to equation 9.3. It is important to keep in mind that the results are reported in terms of percentage 218 changes between base and shock simulations, and that different bases are used in each of the models. Therefore, the percent changes reflect changes in the levels of model forecast, as well as the change in the responsiveness of the model to different policy shocks. In general, however, the differences are fairly small, and would not lead to different policy conclusions or recommendations. 9.3.2 Sensitivitv of Results to the quation Explaining U.S. Per Capita Total Fiber Use The equation explaining per capita total fiber use in the United States (equation 3.28) is, PCTFUUSA = — 191.0 + 23.1 LN PDGDPUSA - 0.23 UNEMPUSA — 1.21 DMFAI - 2.30 DMFAII - 2.74 DMFAIII (3.23) (7.79) (-132) (4.65) (-3.42) (.312) R—SQUARED (CORR): 0.82 SEE.- 0.77 DW: 151 PERIOD OF FIT: 1964-1987 Where: PCI'FUUSA = Per capita total fiber use, United States, PDGDPUSA = per capita deflated gross domestic product, United States, UNEMPUSA = Unemployment rate (%), United States, DMFAI = Zero-one variable for MFAI period, equal 1 1974 to 1977 else 0, DMFAII = Zero-one variable for MFAII period, equal 1 1978 to 1981 else 0, DMZFAIII = Zero-one variable for MFAIII period, equal 1 1982 to 1985 else 0, LN = Indicates variable transformed into logarithms. In the base model, parameters used in the equation explaining per capita total fiber use in the United States were set at levels reported in equation 3.28 above. In order to test the sensitivity of the model to uncertain parameter estimates, the coefficient on the per capita income variable (LN PDGDPUSA) was changed in both directions by one standard deviation and the adjusted models were simulated. The PCTFUUSA variable was adjusted in model A by replacing equation 3.28 with the US per capita total fiber use equation given by equation 9.3 below. This was equivalent to changing the income elasticity of total fiber use in the United States from 1.04 (see Table 3.3) to 1.1. 219 PCTFUUSA = - 191.0 + 26.06 LN PDGDPUSA - 0.28 UNEMPUSA — 1.21 DMFAI - 2.30 DMFAII - 2.74 DMFAII] (9.3) To obtain low simulated values of PCTFUUSA, the US per capita income parameter was adjusted according to equation 9.4 below, giving rise to a new income elasticity for the United States of 0.98. This model (model B) was also simulated over the sixteen-year period. PCI'FUUSA = - 191.0 + 20.13 LN PDGDPUSA - 0.28 UNEMPUSA - 1.21 DMFAI - 230 DMFAII - 2.74 DMFAIII (9.4) The forecasts from the three models are reported in Table 9.2a. The results from model A are for higher levels of per capita total fiber use in the United States. This change is equivalent to a right-ward shift of the cotton and non-cellulosic fibers demand curves of the United States and the world. (The slopes of the demand relationships are unchanged because the U.S. per capita income variable is exogenous in the model.) The forecast of cotton consumption in 1990 is adjusted upwards to 18,674 thousand tons, an increase of about 3.4% from the base level forecast. This is consistent with the fact that the United States consumes about 14% of world cotton and that the change in the coefficient causes the U.S. fiber consumption to increase about 24%. The higher consumption forecast leads to higher price forecasts. For example, the cotton price forecast from model A in 1990 is 143.3, 7.7% higher than the forecast of 133.0 given by the base model. Throughout the sixteen-year simulation period, the gap between model A and the base model remains fairly constant at this percentage level. Forecasts using model B, give a similar proportional difference between the base model, but with 220 lower consumption forecasts leading to lower price forecasts. As a consequence of higher price forecasts, the production forecasts from model A are higher also. Production is forecasted to be 18,342 thousand tons in the base model for 1990. Model A gives a forecast 3% higher at 18,912 thousand tons. Forecasts for 2005 from the base model and model A are 21,826 thousand tons and 22,466 thousand tons, respectively, an increase of 2.9%. Using model B, cotton production is expected to be lower than in the base model, by 2.5%-3.5% throughout the simulation period. The major conclusion emerging from this sensitivity analysis is that changing the income parameter in the per capita total fiber use equation does not result in large differences in the model cotton forecasts and therefore can be considered quite robust with respect to the income variable in the US per capita total fiber use equation. The non-cellulosic fibers forecasts are reported in Table 9.23. It is shown there that, for model A, higher total fiber consumption leads to greater consumption of non- cellulosic fibers, resulting in higher polyester price forecasts, which, on average, are also about 5% higher than in the base model forecasts. Differences in the non-cellulosic fibers production can be explained by differences in the polyester price forecasts. The results for model B are more or less symmetrical to those from model A. That is, higher fiber consumption forecasts, lead to higher price and production forecasts. The most striking feature of the differences between the results for the base and adjusted models is that they are almost identical, showing that the non-cellulosic fibers results are robust with respect to this parameter change. 221 Table 9.2a. Impact on Forecasts of Parameter Changes in the Equation Explaining U.S. Per Capita Total Fiber Use. Variable 1990 1995 2000 2005 World Cotton Pricel/ Base Model” 133.0 119.1 117.7 116.1 Model A” 143.3 128.3 127.0 125.0 Model B‘/ 122.6 109.9 108.4 107.2 World Cotton Consumption” odel 18,043 19,151 20,323 21,634 Model A 18,674 19,802 21,004 22,326 Model B 17,411 18,500 19,651 20,941 World Cotton Production‘y Base Model 18,342 19,517 20,626 21,826 Model A 18,912 20,117 21,257 22,466 Model B 17,771 18,916 19,994 21,185 World Polyester Pricel/ Base Model 117.6 112.8 127.0 149.9 Model A 118.4 113.7 128.1 1513 Model B 116.8 111.9 125.9 1485 World Non-Cellulosic Consumption” Base Model 13,770 15,706 17,219 18,615 Model A 13,662 15,562 17,041 18,412 Model B 13,878 15,850 17,397 18,818 World Non-Cellulosic Production” Base Model 14,870 16,806 18,319 19,715 Model A 14,891 16,836 18,355 19,757 Model B 14,849 16,776 18,283 19,673 V Deflated by MUV, 1985:100. 2 Base Model - Model with coefficient on the income variable in the equation explaining U.S. per capita total fiber 3/ use set at original level (equation 3.28). Model A - Model with parameters of the U.S. per capita total fiber use equation set according to equation 9.3. :/ Model B - Model with parameters of the U.S. per capita total fiber use equation set according to equation 9.4. / ’000 metric tons. The impacts on policy analysis of parameter changes in the equation explaining US per capita total fiber use are reported in Table 9.2b. 222 Table 9.2b. Impact on Policy Analysis of Parameter Changes in the Equation Explaining U.S. Per Capita Total Fiber Use 1/. Variable Simulation 2” Simulation 3” Simulation 4” Impact Average Final Impact Average Final Impact Average Final World Cotton Price Base Model” 5.07 9.16 12.60 4.02 -10.10 -22.17 1.30 3.70 5.30 Model A” 4.71 850 11.68 -0.95 -9.38 4.055 1.21 3.43 4.91 Model 87/ 550 9.93 13.68 -1.11 -10.95 -24.02 1.41 4.01 5.75 World Cotton Consumption Base Model -0.19 -0.48 -0.80 0.04 0.42 0.91 —0.58 -0.19 -0.27 Model A -0.18 -0.46 -0.77 0.04 0.41 0.88 -0.56 -0.18 -0.26 Model B -0.20 -050 -0.83 0.04 0.43 0.94 -0.58 -0.20 -0.28 World Cotton Production Base Model -2.07 -0.94 -0.69 1.82 1.84 153 —058 -0.31 -0.15 Model A -2.01 -0.91 -0.67 1.77 1.79 1.49 -056 -0.30 -0.15 Model B —2.14 -0.97 -0.71 1.88 1.90 1.58 -0.60 —0.32 -0.15 World Polyester Price Base Model 2.09 3.72 5.39 0.41 3.63 8.25 0.42 1.46 2.12 Model A 2.08 3.69 5.34 0.41 3.66 8.18 0.42 1.45 2.10 Model B 2.10 3.75 5.44 0.41 3.66 8.32 0.42 1.47 2.14 World Non-Cellulosic Consumptio Base Model 0.00 0.35 0.43 0.00 0.30 054 0.00 0.13 0.16 Model A 0.00 0.34 0.42 0.00 0.29 055 0.00 0.13 0.16 Model B 0.00 036 0.44 0.00 0.31 053 0.00 0.13 0.16 World Non-Cellulosic Production Base Model 0.00 0.34 0.40 0.00 0.29 0.49 0.00 0.13 0.15 Model A 0.00 0.34 0.48 0.00 0.29 0.49 0.00 0.13 0.15 Model B 0.00 0.34 0.40 0.00 0.29 0.49 0.00 0.13 0.15 I/ Results reported in terms of percentage change between base and shock. 2/ Simulation 2 - 10% decrease in USSR cotton production. 3/ Simulation 3 - 10% increase in Chinese cotton production. V Simulation 4 - 10% decrease in U.S. cotton price. 5 Base Model — Model with coefficient on the income variable in the equation explaining U.S. per capita total fiber use set at original level (equation 3.28). 6 Model A — Model with parameters of the U.S. per capita total fiber use equation set according to equation 9.3. 7/ Model B - Model with parameters of the U.S. per capita total fiber use equation set according to equation 9.4. As mentioned above, in model A, the parameter change causes an upward shift in the demand curves for cotton and non-cellulosic fibers, but involves no change in their slopes. Therefore, the impacts of each of the policies on the model’s endogenous variables in 223 absolute values do not changed. However, the percentage changes are different, because the changes are measured from different base levels. For example, the absolute difference between the shock and base run for the base model is the same as the absolute difference between the shock and base run for model A and for model B. However, since the base run forecasts from the base model differ from the base run forecasts from the models A and B, the percentage changes are different. For example, using the base model, the impact elasticity of a 10% reduction in the USSR cotton production is to raise the world price of cotton 5.07% over the base run cotton price. In model A, the same absolute increase in the cotton price gives only a 4.71% increase in percentage terms, because the price forecasts from this model are higher. In this respect, the results from this sensitivity analysis are less useful than others presented. However, it can be seen that there do not appear to be large differences between the model results, except with the possible exception of those for the cotton price. 9.3.3 Sensitivity of Results to the Equation Explaining Chinese Cotton Production Another key variable in the model is cotton production in China. This was demonstrated in section 8.3, where the effects of a 10% increase in China’s cotton production on the world fiber market were reported. To test the robustness of these results, forecast and policy simulations were undertaken to assess to what extent the model is sensitive to changes in the supply elasticity of Chinese cotton with respect to the cotton price. Recall that production in each region model is derived from the product of yield per 224 hectare and the total number of hectares. Therefore, it was necessary to change parameters in equations explaining both these variables. The equations explaining China’s cotton production are given by 4.5 and 4.22 below. CI'YDCHI = - 258 + 0.003 DFCI'PCHI - 0.001 DFFNPCHI + 1.15 LN TIME - 0.16 D88 (45) (1.95) (.345) (9.20) (2.67) R-SQUARED (CORR): 0.95 SEIEL~ 0.0289 DW: 231 PERIOD OF FIT: 1977-1988 Where: CI‘YDCI-II = Cotton yield, China (m.tons / hectare), DFCI‘PCHI = Deflated cotton price, China, FFNPCHI = Deflated fertilizer price, China, TIME = Time variable, D88 = Zero-one variable, equals 1 in 1988, 0 otherwise, LN = Indicates variable transformed into logarithms. l, CI'HACHI = 4705 + 26.21 DFCTPCHI(-1) + 1377 D84 - 1087 D86 (4.22) (2.63) (3.28) (-2.67) R-SQUARED (CORR): 0.82 SEE; 383.0 DW: 2.32 PERIOD OF FIT: 1977-1988 Where: CI‘HACHI = Cotton area, China, DFCI‘PCHI = Deflated cotton price, China, D84 = Zero-one variable, equals 1 in 1984, 0 otherwise, D86 = Zero-one variable, equals 1 in 1986, 0 otherwise. As in previous sensitivity analyses, 8 base model containing the equations 4.5 and 4.22 was simulated to provide base model forecasts and policy analysis results. Next, the base model was adjusted by replacing equations 4.5 and 4.22 with equations 9.5 and 9.6. These are shown below. CI‘YDCHI = - 258 + 0.00453 DFCI‘PCHI - 0.001 DFFNPCHI + 1.15 LN TIME - 0.16 D88 (95) CI'HACI-II = 4705 + 36.21 DFCI'PCHI(—1) + 1377 D84 -1087 D86 (9.6) These differ from 4.5 and 4.22 in that the coefficient on the price variable (DFCTPCHI) in each equation is raised by one standard deviation. Using this model (model A), 225 simulations for forecasts and policy analysis were undertaken and the results reported in Table 9.3. In addition, a model B was created when equations 4.5 and 4.22 were replaced by 9.7 and 9.8 below. These contain parameters on the price variables lowered by one standard deviation. CI'YDCHI = — 258 + 0.00146 DFCTPCI-II - 0.001 DFFNPCHI + 1.15 LN TIME - 0.16 D88 (9.7) CI'I-IACHI = 4705 + 16.21 DFCI'PCHI(—1) + 1377 D84 - 1087 D86 (9.8) The forecast of cotton production in China in 1990 from model A is about 6% higher than the base model forecast. Given that China produces about 22% of world production, total output increases by about 1.4% from 18,342 thousand tons to 18,599 thousand tons. Between 1990 and 2005, model A production forecasts are about 1.2% higher than in the base model. Higher production forecasts lead to lower price forecasts. For example, for 1990, the base model price forecast is 133.0, while the model A forecast is 128.3, a decrease of about 3.5%. The lower price forecast makes the cotton consumption forecast to be 0.5% higher at 18,327 thousand tons. Comparing the base model and model A, over the entire sixteen-year forecast period, the cotton price and world consumption are 2.1% lower and 0.4% higher, respectively. This indicates that the model forecasts are robust to changes in the supply elasticity of Chinese cotton. 226 Table 9.3a Impact on Model Forecasts of Parameter Changes in the Equation Explaining Chinese Cotton Production. Variable 1990 1995 2000 2005 World Cotton Pricel/ Base Modelz/ 133.0 119.1 117.7 116.1 Model A” 128.3 115.1 113.8 112.2 Model B” 137.7 123.1 121.6 1200 World Cotton Consumption” Base Model 18,043 19,151 20,328 21,634 Model A 18,327 19,437 20,615 21,939 Model B 17,758 18,865 20,041 21,329 World Cotton Production” Base Model 18,342 19517 20,626 21,826 Model A 18599 19,780 20,894 22,107 Model B 18,085 19,253 20,357 21,544 World Polyester Price” Base Model 117.6 112.8 127.0 149.9 Model A 116.7 111.8 125.8 148.3 Model B 1185 113.8 128.2 1515 World Non—Cellulosic Consumption” Base Model 13,770 15,706 17,219 18,615 Model A 13,649 15,547 17,025 18,394 Model B 13,891 15,865 17,413 18,836 World Non-Cellulosic Production” Base Model 14,870 16,806 18,319 19,715 Model A 14,846 16,773 18,280 19,669 Model B 14,894 16,839 18,358 19,761 ’/ Deflated by MUV, 1985=100. 2/ Base Model - Model with parameters of China’s yield and area equations set at original levels (equations 45 and 4.22). 3/ Model A - Model with parameters of China’s yield and area equations set according to equations 95 and 9.6. 4/ Model B - Model with parameters of China's yield and area equations set according to equations 9.7 and 9.8. 5/ ’000 metric tons. As seen in Table 9.33, the impacts of this parameter change on the forecasts for the non-cellulosic fibers markets are very small. For model A, a lower cotton price forecast leads to lower consumption of non-cellulosic fibers, decreasing, on average over the 1990- 2005 period, less than 1% below the base model forecasts. Less consumption results in a '5 227 lower polyester price forecast, also falling less than 1% below the base model forecast throughout the sixteen-year simulation period. Non-cellulosic production forecasts are lower for model A as a consequence of the lower polyester price forecast. These results indicate that the model is robust to changes in the made to the cotton price elasticity of supply in the China production equations. The impacts of changing the cotton price parameters in the production equation in the China region of the model are shown in Table 9.3b. The effect of changing the price elasticity of supply in China is to change the slope of the supply curve of world cotton. In terms of a Marshallian supply-demand diagram, when the coefficient on the price variables are increased (model A), the supply curve becomes flatter so that the model becomes less responsive to policy shocks. Conversely, in model B, when the parameters are reduces, the world cotton supply curve becomes steeper, and the policy shocks will have larger effects on the model’s endogenous variables. Clearly from Table 9.3b the models’ sensitivity to the parameter changes is small. For example, the results for the world cotton price for simulation two give, for the base model, impact, average and final elasticities of 5.07%, 9.16% and 12.60%, respectively. The equivalent results from model A are 5.26%, 9.27% and 12.64%, respectively; while for model B they are 4.90%, 9.06% and 12.56%. The differences are very small. The effects of the policy Shocks based on model B are smaller in percentage terms, as a result of using a larger base, however, they show greater variability because of the lower cotton supply elasticity. Similarly, using model A, the percentage changes are larger due to the smaller base price levels, and are less variable, due to the higher cotton supply elasticity. 228 Table 9.3b Impact on Polig Analysis of Parameter Changes in the Equation Explaining Chinese Cotton Production ll. Variable Simulation 2” Simulation 3” Simulation 4‘/ Impact Average Final Impact Average Final Impact Average Final World Cotton Price Base Models 5.07 9.16 12.60 -l.02 -10.10 -22.17 1.30 3.70 5.30 Model A” 5.26 9.27 12.64 -1.06 -9.98 -21.84 1.35 3.70 5.27 Model B” 4.90 9.06 1256 -0.99 -10.81 -22.48 1.26 3.70 5.32 World Cotton Consumption Base Model —0.19 -0.48 -0.80 0.04 0.42 0.91 -0.58 -0.19 -0.27 Model A -0.19 -0.49 -0.82 0.04 0.43 0.94 -057 -0.17 -0.25 Model B -0.19 —0.50 -0.84 0.04 0.45 0.97 -0.59 -0.17 -0.26 World Cotton Production Base Model -2.07 -0.94 —0.69 1.82 1.84 153 -0.58 -0.31 -0.15 Model A -2.04 -0.83 -057 1.79 1.82 150 -0.57 -0.29 -0.13 Model B -2.10 -0.85 -058 1.85 1.89 154 -059 -0.30 -0.13 World Polyester Price Base Model 2.09 3.72 5.39 0.41 3.63 8.25 0.42 1.46 2.12 Model A 2.11 3.75 5.44 0.41 3.66 8.33 0.42 1.47 2.14 Model B 2.07 3.69 5.34 0.41 3.60 8.17 0.42 1.45 2.10 World Non-Cellulosic Consumption odel 0.00 0.35 0.43 0.00 0.30 0.54 0.00 0.13 0.16 Model A 0.00 0.35 0.43 0.00 0.30 054 0.00 0.13 0.16 Model B 0.00 0.35 0.43 0.00 0.30 055 0.00 0.13 0.16 World Non-Cellulosic Production Base Model 0.00 0.34 0.40 0.00 0.29 0.49 0.00 0.13 0.15 Model A 0.00 0.34 0.40 0.00 0.29 0.49 0.“) 0.13 0.15 Model B 0.00 0.34 0.40 0.00 0.29 0.49 0.00 0.13 0.15 1/ Results reported in terms of percentage change between base and Shock. 2/ Simulation 2 - 10% decrease in USSR cotton production. 3/ Simulation 3 - 10% increase in Chinese cotton production. 4/ Simulation 4 - 10% decrease in U.S. cotton price. Base Model - Model with parameters of China’s yield and area equations set at original levels (equations 45 and 4.22). 6/ Model A - Model with parameters of China’s yield and area equations set according to equations 9.5 and 9.6. 7/ Model B - Model with parameters of China’s yield and area equations set according to equations 9.7 and 9.8. 229 9.3.4 Sensitivity of Results to the Equation Explaining the World Polyester Price The equation explaining the world polyester price (equation 6.16) is, PSPWOR = 79.6 + 1.74 MUV . 0.05 NCUROW + 1.8101LPR(—1) (6.16) (6.83) (-5.66) (6.16) R-SQUARED (CORR): 0.96 SEEL' 556 DW: 2.61 PERIOD OF m: 1969-88 Where: PSPWOR = Polyester price, World, MUV = Manufactures unit value (deflator), NCUROW = Non-cellulosic fibers use, Rest-of-the-World, OILPR = Price of oil. In the base model, parameters used in the polyester price equation were set at the values shown in equation 6.16 above. In order to test the sensitivity of the model to uncertain parameter estimates, the coefficient on the non-cellulosic fibers use in the Rest-of-the- World (NCUROW) was changed in each direction by one standard deviation. The new models created by these adjustments were simulated as before. The NCUROW variable was adjusted in Model A by replacing equation 6.16 by equation 9.9 shown below. This was equivalent changing the elasticity of demand for non-cellulosic fibers with respect to the price of polyester from -1.1 to -1.35. PSPWOR = 79.6 + 1.74 MUV - 0.04 NCUROW + 1.81 OILPR(—1) (9.9) To simulate the model for low values of PSPWOR, the non-cellulosic fibers use parameter in the equation was adjusted up according to equation 9.10 below, changing the elasticity of demand to -0.85. This model (model B) was also simulated over the sixteen- year period. 230 PSPWOR = 79.6 + 1.74 MUV - 0.06 NCUROW + 1.81 OILPR(-1) (9.10) The results of the forecast simulation are reported in Table 9.4a. For 1990, the polyester price forecasted by model A is about 15% higher than by the base model. Similarly, the polyester price forecast is about 18% lower based on model B. All three models gave the same pattern of price movements. That is, a price decline between 1990 and 1995 ranging between 3.5% and 5% over the three models. Then price is predicted to increase to the year 2005 at similar growth rates. By 2005, the price forecast by model A is 14% above the base model forecast, and from model B the forecast is 16% below. The results for non-cellulosic fibers production are consistent with price forecasts. The higher forecasted prices by model A lead to higher production forecasts--approximately 3% higher in each year reported in Table 9.4a. A decline of about 3% below the base model forecasts of non-cellulosic fibers production is reported for model B. Non-cellulosic fibers consumption forecasts are consistent with the price forecast adjustments reported above. From the base model, for 1990, consumption is forecasted to be 13,770 thousand tons, increasing to 18,615 thousand tons in 2005. Model A gives a forecast of 11,498 thousand tons, increasing to 15,644 thousand tons by the end of the sixteen—year simulation period, while using model B, the forecasts are 16,042 thousand tons in 1990, and 21,586 thousand tons in 2005. Overall, non-cellulosic fibers consumption forecasts from the adjusted models are consistently less than 20% different from the base model forecasts. 231 Table 9.4a Impact on Model Forecasts of Parameter Changes in the Equation Explaining the World Polyester Price. Variable 1990 1995 2000 2005 World Cotton Pricel/ / Base Model2 133.0 119.1 117.7 116.1 Model A” 195.3 167.1 161.8 156.7 Model B“ 70.0 71.1 73.6 755 World Cotton Consumption” Base Model 18,043 19,151 20,328 21,634 Model A 14,237 15,757 17,111 18,461 Model B 21,849 22,545 23,545 24,807 World Cotton Production” Base Model 18,342 19,517 20,626 21,826 Model A 21,781 22,646 23,627 24759 Model B 14,903 16,388 17,625 18,893 World Polyester Price” Base Model 117.6 112.8 127.0 149.9 Model A 135.2 129.1 145.1 170.9 Model B 100.0 96.4 108.9 128.9 World Non-Cellulosic Consumptionj/ Base Model 13,770 15,706 17,219 18,615 Model A 11,498 13,155 14,397 15,644 Model B 16,042 18,757 20,040 21,586 World Non—Cellulosic Production” Base Model 14,870 16,806 18,319 19,715 Model A 15,361 17,340 18,893 20,330 Model B 14,424 16,272 17,744 19,099 V Deflated by MUV, 1985:100. 2/ Base Model - Model with parameters of the polyester price equation set at original level (equation 6.16). 3” Model A - Model with parameters of the polyester price equation set according to equation 9.9. 4/ Model B - Model with parameters of the polyester price equation set according to equation 9.10. 5/ ‘000 metric tons. The adjusted forecasts for the cotton market are also reported in Table 9.4a. From model A, for 1990, the cotton price is about 47% higher than in the base model, while it is about 47% lower in model B. These differences between the base and adjusted models are fairly constant between 1990 and 2005. These differences between base and adjusted model results are the largest reported so far in the sensitivity analysis exercise. The reason 232 for these large differences is due to the inelastic supply of cotton. The higher polyester price forecasts made with model A lead to lower non-cellulosic fibers consumption forecasts, and to higher cotton consumption forecasts. This can be thought of as a shift to the right of the demand curve for cotton, and given an inelastic supply curve (i.e., one that is steeply sloping), price has to increase substantially for the market clearing identity in the model to hold. The cotton production and consumption forecasts differ from the base model in a consistent manner with the differences between the base and adjusted forecasts of cotton prices. Over the entire simulation period, the gap between the base and adjusted forecasts of cotton production and consumption never exceeds 30%. The major conclusion to emerge from this analysis is that the non-cellulosic fiber results are fairly robust to the specification of the polyester price equation, with the difference between base and adjusted model forecasts never exceeding 20%. However, different consumption forecasts in the cotton component of the model give rise to substantial differences between the models. The effects of the parameter changes in the polyester price equation on policy analysis are shown in Table 9.4b. As in section 9.3.1, increasing of the parameter in the non-cellulosic fibers use equation causes the non-cellulosic fibers component of the model to become more elastic, resulting in smaller percentage changes than in the base model for the policy shocks. Conversely, the policy effects are larger because lowering the parameter makes the model more inelastic with respect to price. Also, as pointed out earlier, the percent changes are measured from different base levels, so that the policy impacts appear smaller for variables whose base values have increased, and larger for variables with lower base values. 233 Table 9.4b Impact on Model Poliqy Analysis of Parameter Changes in the Equation Explaining the World Polyester Price 1]. Variable Simulation 2” Simulation 3” Simulation 4V Impact Average Final Impact Average Final Impact Aver-age Final World Cotton Price / 5.07 9.16 12.60 1.02 -10.10 -22.17 1.30 3.70 5.30 Model A” 3.45 653 9.17 0.69 -7.20 -16.13 029 2.64 3.86 Model 87/ 9.63 15.34 20.15 1.94 46.92 -35.45 2.47 6.20 8.48 World Cotton Consumption Base Model -0.19 -0.48 -0.80 0.04 0.42 0.91 -0.58 -0.19 -0.27 Model A -0.24 -058 -0.95 0.05 051 1.08 -0.74 -O.23 -0.32 Model B -0.16 -O.41 -0.69 0.03 0.36 0.79 -0.48 -0.16 -0.23 World Cotton Production Base Model -2.07 -0.94 -0.69 1.82 1.84 153 -058 -0.31 -0.15 Model A -1.74 -0.81 -0.60 1.53 159 1.34 -0.49 -0.27 -0.13 Model B -255 -1.12 -0.81 2.?A 2.19 1.79 -0.71 -0.37 -0.18 World Polyester Price Base Model 2.09 3.72 5.39 0.41 3.63 8.25 0.42 1.46 2.12 Model A 1.82 3.39 5.01 0.36 3.45 7.91 0.37 1.37 2.00 Model B 2.46 4.16 5.90 0.48 3.87 8.71 0.49 159 2.27 World Non-Cellulosic Consumptio Base Model 0.00 0.35 0.43 0.00 0.30 0.54 0.00 0.13 0.16 Model A 0.00 0.38 0.46 0.00 0.32 058 0.00 0.14 0.17 Model B 0.00 0.27 0.33 0.00 0.23 0.42 0.00 0.10 0.12 World Non-Cellulosic Production Base Model 0.00 0.34 0.40 0.00 0.29 0.49 0.00 0.13 0.15 Model A 0.00 0.30 0.35 0.00 0.25 0.43 0.00 0.11 0.13 Model B 0.00 0.32 037 0.00 0.27 0.46 0.00 0.12 0.14 I/ Results reported in terms of percentage change between base and shock. 2/ Simulation 2 - 10% decrease in USSR cotton production. 3/ Simulation 3 - 10% increase in Chinese cotton production. 4/ Simulation 4 - 10% decrease in U.S. cotton price. 5/ Base model - Model with parameters of the polyester price equation set at original level (equation 6.16). Z Model A - Model with parameters of the polyester price equation set according to equation 9.9. Model B - Model with parameters of the polyester price equation set according to equation 9.10. In general, the differences in policy effects between the base and adjusted models is fairly slight. This shows that, overall, the model is robust with respect to changing the demand elasticity of non-cellulosic fibers with respect to the price of polyester. 234 9.3.5 Sensitivity of Results to the Equation Explaining World Non-Cellulosic Fibers Production The final sensitivity analysis is for the production of non-cellulosic fibers. To test the robustness of the results, forecasts and policy simulations were undertaken to assess the extent to which the model is sensitive to changes in the supply elasticity of world non- cellulosic fibers with respect to the price of polyester. The equation explaining world non- cellulosic fibers production (equation 6.15) is, NCPROD = - 27116 - 4397. DFOILPR - 127.6 RIRUSA + 1077. DFPSPWOR(-1) + 12211. LN TIME (6.15) (-277) (253) (5.49) (13.18) R—SQUARED (CORK): 0.99 SE: 488.7 DW: 1.65 PERIOD OF FIT: 1964-88 Where: NCPROD = Non-cellulosic fibers production, world, DFOILPR = Deflated price of oil, (OPEC petroleum average prices), RIRUSA = Real rate of interest (long term U.S. bond yield), DFPSPWOR = Deflated polyester price, TIME = Time variable, LN = Indicates variable transformed into logarithms. As in earlier sensitivity analyses, a base model containing equation 6.15 was simulated to provide base model forecast and policy analysis results. Next, the base model was adjusted by replacing equation 6.15 with equation 9.11 below. NCPROD = - 27116 — 4397 DFOILPR - 127.6 RIRUSA + 1273 DFPSPWOR(-1) + 12211 LN TIME (9.11) This differs from equation 6.15 in that the coefficient on the price variable (DFPSPWOR) is raised by one standard deviation. Using this model (model A), simulations for forecasts and policy analysis were undertaken and the results reported in Table 9.5. In addition, a model B was created when equation 6.15 was replaced by equation 9.12 below. This contains a parameter on the price variable lowered by one standard 235 deviation. NCPROD = - 27116 - 4397 DFOILPR - 127.6 RIRUSA + 881 DFPSPWOR(-1) + 12211 LN TIME (9.11) The results of the forecast simulation are reported in Table 9.5a. Between 1990 and 2005, forecasts of non-cellulosic fibers production by model A are about 2.5% higher than in the base model. Higher production forecasts lead to lower price forecasts. For example, for 1990, the base model polyester price forecast is 117.6, while the model A forecast is 105.8, a decrease of about 11%. The lower price forecast makes the non- cellulosic fibers consumption forecast to be 0.5% lower at 15,285 thousand tons. Comparing the base model and model A, over the entire sixteen-year forecast period, the polyester price and world non-cellulosic fibers consumption are 9.5% lower and 8% higher, respectively. This indicates that the model forecasts are fairly robust to changes in the production elasticity of non-cellulosic fibers. As seen in Table 9.5a, the impact of this parameter change on the forecasts for the cotton markets is quite large. The reasons for this were given in section 9.3.4. For model A, a lower polyester prices leads to lower cotton consumption, decreasing, on average over the 1990-2005 period, about 12% below the base model forecasts. Lower cotton consumption results in a lower cotton price, falling as much as 34% below the base model forecast during the sixteen-year simulation period. Cotton production forecasts are lower for model A as a consequence of the lower cotton price forecasts. These results indicate that the model is quite sensitive to the changes made to the price elasticity of supply of non-cellulosic fibers. 236 Table 9.5a Impact on Forecasts of Parameter Changes in the Equation Explaining the World Non-Cellulosic Fibers Production. Variable 1990 1995 2000 2005 World Cotton Pricel/ 2 Base Model / 133.0 119.1 117.7 116.1 Model A‘V 86.8 87.7 90.1 91.4 Model BV 179.2 2505 - 145.3 140.8 World Cotton Consumption” Base Model 18,043 19,151 20,328 21,634 Model A 20,862 21,375 22,340 19,704 Model B 15,224 16,927 18,316 23,564 World Cotton Production” Base Model 18,342 19,517 20,626 21,826 Model A 20,890 21,567 22,503 23,610 Model B 15,795 17,467 18,749 20,042 World Polyester Price” Base Model 117.6 112.8 127.0 149.9 Model A 105.8 103.0 116.6 138.] Model B 129.4 122.6 137.4 161.7 World Non-Cellulosic Consumption” Base Model 13,770 15,706 17,219 18,615 Model A 15,285 17,232 18,839 20,280 Model B 12,255 14,180 15599 16,950 World Non-Cellulosic Productionj/ Base Model 14,870 16,806 18,319 19,715 Model A 15,242 17,125 18,649 20,060 Model B 14,498 16,487 17,989 19,370 ’/ Deflated by MUV, 1985: 100. Base model - Model with parameters of non-cellulosic production equation set at original level (equation 6.15). “V Model A - Model with parameters of non-cellulosic production equation set according to equation 9.11. 3/ Model B - Model with parameters of non-cellulosic production equation set according to equation 9.12. V ’000 metric tons. The impacts of changing the polyester price parameter in the non-cellulosic fibers production equation model are shown in Table 95b. The results are similar to those reported in section 9.3.4 and in Table 9.4. In general, the results show that the non- cellulosic fibers component of the model fairly is robust to these made, while the cotton component shows more variation for the different models. 237 Table 9.5b Impact on Polig Analysis of Parameter Changes in the Equation Qplaining the World Non-Cellulosic Production 11. Variable Simulation 2” Simulation 3” Simulation 4“ Impact Average Final Impact Average Final Impact Average Final World Cotton Price Base Mode 5.07 9.16 12.60 1.02 -10.10 -22.17 1.30 3.70 5.30 Model A” 7.77 11.61 14.98 -156 -10.10 -22.17 1.99 4.54 6.14 Model 87/ 3.76 6.78 9.29 -0.76 -6.91 -15.39 0.96 2.64 3.81 World Cotton Consumption Base Model -0.19 -0.48 -0.80 0.04 0.42 0.91 -0.58 -0.19 -0.27 Model A -0.16 -0.47 -0.81 0.03 0.43 0.95 -050 -0.12 -0.20 Model B -0.33 -059 —0.99 0.05 0.53 1.15 -0.69 -0.15 -0.25 World Cotton Production Base Model -2.07 -0.94 -0.69 1.82 1.84 153 -058 -0.31 -0.15 Model A -2.04 —0.86 -0.53 2.11 2.06 1.64 —0.30 -0.30 -0.09 Model B -1.82 -0.70 -0.44 1.60 1.69 1.36 -051 -0.24 -0.08 World Polyester Price Base Model 2.09 3.72 5.39 0.41 3.63 8.25 0.42 1.46 2.12 Model A 2.32 4.07 557 0.40 7.98 8.99 0.47 1.60 2.31 Model B 1.90 3.47 4.78 0.37 3.34 7.63 0.38 1.34 1.96 World Non-Cellulosic Consumption Base Model 0.00 0.35 0.43 0.00 0.30 054 0.00 0.13 0.16 Model A 0.00 0.37 0.45 0.00 0.21 057 0.00 0.14 0.17 Model B 0.00 0.45 0.55 0.00 0.38 0.69 0.00 0.17 0.20 World Non-Cellulosic Production Base Model 0.00 034 0.40 0.00 0.29 0.49 0.00 0.13 0.15 Model A 0.00 0.33 0.39 0.00 0.28 0.48 0.00 0.13 0.15 Model B 0.00 0.35 0.41 0.00 0.30 0.50 0.00 0.13 0.15 I/ Results reported in terms of percentage change between base and shock. 2/ Simulation 2 - 10% decrease in USSR cotton production. 3/ Simulation 3 - 10% increase in Chinese cotton production. 4/ Simulation 4 - 10% decrease in U.S. cotton price. 5 Model - Model with parameters of non-cellulosic production equation set at original level (equation 6.15). 6/ Model A - Model with parameters of non-cellulosic production equation set according to equation 9.11. 7/ Model B - Model with parameters of non-cellulosic production equation set according to equation 9.12. 238 9-4 Cam In this section, the results from a number of sensitivity analysis were reported and discussed. The procedure used was one proposed by Anderson, and the analysis was limited to changing a few important parameters in the model and seeing how they impacted on a small number of key performance variables. Overall, the conclusion which emerged from these sensitivity tests is that the model is fairly robust to the unsure parameters tested. However, the analysis showed the importance of having accurate parameter estimates for equations estimated in logarithms, since small errors in logarithms become magnified considerably when the exponential is taken. Also, the inelasticity of demand and supply caused price forecasts to be different, by sometimes large amounts, for relatively small changes in quantity forecasts. 10. Summary, Conclusions and Areas of Future Research 10.1 Summagy The main purpose of this study was to specify and estimate an econometric model of the world fiber market, with emphasis on the cotton sector, and then, after testing and validation of the model, to forecast prices, production and consumption for the major world fiber market participants. In addition to forecasting, the model was used to estimate the impact of some important market and policy developments. Model simulations were undertaken to analyze: (i) the impact of the expected expansion of China’s cotton production; (ii) the impact of continued stagnation in the USSR cotton sector; (iii) the likely effect of the cotton provisions contained in the 1990 Farm Bill on the world fiber market; and (iv) the impact of the MFA on the raw fiber market. Analysis of these developments provide timely and relevant information for many groups and individuals with interests in the fiber market. In section H, the nature of the fiber market was described along with recent trends and market developments. This description provided the basis for the model specifications presented in later sections (e.g., world price determination, choice of model production and consumption regions, and treatment of textile demand). In section 3, the cotton demand component of the model was discussed. For each demand region in the model, two equations were estimated. The first was for per capita total fiber use which was specified to be related to per capita income. In the second, the cotton share of total fiber use was estimated as a function of the cotton price 239 240 relative to the polyester price. This specification captures the price sensitivity of manufacturers to changes in the relative prices of fibers. The econometric results were satisfactory and provided price elasticity measures ranging from -0.02 for India to -0.33 for the Republic of Korea. These conform closely to price elasticity estimates reported in previous studies. The income elasticity estimates ranged from 0.12 for Turkey to 1.08 for the EEC. These also were similar to income estimates found in other studies. The production component of the model was described in section 4. Based on previous econometric studies of annual crop production, cotton production in the model was derived from the production of area planted and average yield. Each of these components was estimated separately. Area planted equations contain as regressors the price of cotton and the price of crops in competition with cotton for farm acreage, as well as lagged area, based on the assumption that producers form price expectation adaptively. In the yield equations, weather variables were used if the data were available and were significant in most cases. The short-run supply elasticity estimates ranged from 0.07 for the north region of India and 0.87 for Argentina. The model was closed by formulating a cotton pricing equation as an inverted world stocks demand equation. This was discussed in section 5. In an earlier formulation of the model, the world price was solved using a world market clearing identity. This did not perform well in simulation and an alternative approach was adopted involving the use of a price equation. The world price of cotton was specified as a function of world stocks, net of stocks held in China. This was because, historically, a large proportion of stocks in this country were isolated from the world market. In addition to world stocks, world cotton consumption (net of Chinese consumption) was included in the equation to capture the 241 transactions demand and gave an estimated flexibility of cotton price with respect to stock levels of -O.78. The non-cellulosic fibers component of the model was presented in section 6. Only equations for the non-cellulosic fibers (polyester, rayon and acrylics) were estimated. However this fiber group makes up almost 80% of the non-cellulosic fibers market. Non- cellulosic share of total fiber use equations were estimated for each consumption region of the model, which were then combined with total fiber use to determine non-cellulosic fibers consumption. The supply of non-cellulosic fibers was estimated for the world. The polyester price was determined in an inverted demand equation for non-cellulosic fibers in the rest-of-the-world region and influences the cotton market through the cotton share of total fiber use equations. A number of validation Statistics were presented in section 7 that cover various aspects of the model’s ability to reproduce actual data. The validation statistics reported were: (i) the Root Mean Squared Percentage Error (RMSPE), (ii) the Mean Squared Error (MSE), (iii) Theil’s U-statistic, and (iv) graphical validation. In general, the model withstood these testing procedures and predicted actual market values accurately enough to be used for policy experiments and forecasting. The forecast and policy simulation results were reported in section 8. Five sets of simulation results were presented. These were for (i) a forecast of price, production and consumption for the period 1990-2005; (ii) a 10% decrease in cotton production in the USSR; (iii) a 10% increase in cotton production in China, (iv) a 10% decline in domestic cotton price in the United States, (v) an evaluation of the impact of MFA on the cotton and non-cellulosic fibers sectors. Finally, in section 9, sensitivity analysis of the model was 242 resented, indicatin the model’s robustness to chan ' arameter estimates. P g gmg P 10.2 Conclusions In general, the study met the research objectives outlined in section I. A model of the world fiber market was developed and provided a number of important insights into how the world fiber market operates and where it might be headed in the future. The model forecasts that between 1990 and 2005 the real world price of cotton will fall approximately 25%, while a 10% price increase is forecast for polyester. This suggests that cotton should maintain, or even expand, its share of the total fiber market in the coming decade. The forecast simulation results also show that the individual countries’ shares of both production and consumption of cotton and non-cellulosic fibers change very little up to 2005. However, to some extent, this results from the fact that the exogenous variables used for forecasting are based on constant growth rates through to 2005. Three model simulations involved shocking key variables in major producing regions (i.e., the USSR, China and the United States). In each case, the effect on the world market was significant. For example, given a permanent 10% decrease in production in the USSR, the world price rises by about 9%. This indicates that forecasts of price, both near- and long-term, should include information on USSR cotton policy and producer incentive structures. Over an 11-year simulation of the model, for every 1% increase in China’s production the world price of cotton falls, on average, about 1% and the price of polyester falls 0.35%. The impact of a 10% decline in the US cotton price during the early 1990s was to reduce US production, on average, less than 3%, and to increase world prices an average of 3.7%. 243 The conclusion emerging from the MFA simulation is that, while the effect of the MFAs have been substantial in terms of welfare losses and economic inefficiency in the T&A manufacturing and trade sectors, the effects on the raw fiber market have not been large. World prices of cotton and polyester were reduced by, on average, about 5% between 1979 and 1986 (within a range between 3% and 7%), although the period of analysis does not cover the most restrictive period since 1986. Therefore, the current impact of the agreements on world fiber prices is likely to be at the upper end of this range. 10.3 Areas of Future Research A number of areas have been identified for improving the model and for further policy simulations using the model. Some of these areas are listed briefly below. While a large proportion of world production and consumption is covered by the regions already included in the model, more countries will be added in the future. In particular, some of the major African countries will be included, such as the Sudan, Nigeria and other West African countries. This will allow the effects of exogenous world market shocks on the cotton sectors in these countries to be measured. While cotton and non-cellulosic fibers make up about 90% of the world fiber market, cellulosic fibers and wool are also important, especially for the major producers of these commodities (e.g., Australia and New Zealand in the production of wool). Given the framework on the demand side of the model, the inclusion of these fibers would be relatively simple, requiring the estimation of share equations which could then be combined with the total fiber use equations to derive demand. )1 244 Another area of future work will be to obtain and incorporate into the model more country-specific data such as local prices, regional production and weather information. As reported in section 4, the use of local price data (e.g., in China) and the breakdown of country production into specific regions (e.g., India and the United States) improved the estimation results dramatically. Also, the inclusion of more weather variables in the yield equations is likely to improve the quality of these equations, as in the case of the US and India yield equations. To meet the objectives outlined in section I, trade and stock demand equations were not needed in the model. However, within the framework of the model these could be added easily. In fact, by using country market clearing identities, either a stock or net export equation need be estimated and the remaining variable derived from the identity. In practice, the estimation of these equations may prove troublesome. In an initial Specification of the model, country-level stock equations were tried but performed unsatisfactorily with the price insignificant in most equations. Also, trade equations are problematic because it is not possible to determine a priori the sign on the price variable if included in the specification. In the current version of the model the production of non-cellulosic fibers is estimated at the world level. At a later stage world production will be disaggregated and equations estimated for each of the major producing areas of the world. This is important as many developing countries are increasing their non-cellulosic fiber production capacity (e.g., China, India and Pakistan) and it will be important to assess the impact of this development on the world fiber market in the future. Given the unpredictability of cotton yields, production often fluctuates widely from 245 year to year. It is possible to include a stochastic element into the yield equations and then to simulate the model. When the simulation is repeated a number of times the variances of the endogenous model variables can be estimated. This provides interesting information such as the likelihood that a certain market outcome (e.g., a given production or price level) will occur. m m m u APPENDIX A VARIABLE DEFINITIONS APPENDIX A Variable Definitions Endogenous CI‘CONCHI = Cotton Consumption (’000 tons), China. CI‘ CONWOR = Cotton Consumption (’000 tons), World. CI‘CONWORXCHI = World Cotton Consumption less China Consumption (’000 tons). CI'ESCHI = Cotton Ending Stocks (’000 tons), China. CI'ESWOR = Cotton Ending Stocks (’000 tons), World. CTCONWORXCHI = World Ending Stocks less China Ending Stocks (’000 tons). CI'HAARG = Cotton area (number of hectares), Argentina. CI'HAAUS = Cotton area (number of hectares), Australia. Cotton area (number of hectares), Brazil. Cotton area (number of hectares), Central Africa. CTHACHI = Cotton area (number of hectares), China. ii CI'I-IAEGY = Cotton area (number of hectares), Egypt. CI'HAINDN = Cotton area (number of hectares), North India. CI‘HAINDS = Cotton area (number of hectares), South India. CI'I-IAINDW = Cotton area (number of hectares), West India. CI'HAMEX = Cotton area (number of hectares), Mexico. CI‘HAPAKP = Cotton area (number of hectares), Punjab Pakistan. CI‘HAPAKS = Cotton area (number of hectares), Sind Pakistan. CTHATUR = Cotton area (number of hectares), Turkey. CTHAUSl = Cotton area (number of hectares), Delta Region, United States. CTHAUSZ = Cotton area (number of hectares), Southeast Region United States. CI‘HAUS3 = Cotton area (number 01' hectares), Southwest Region United States. CI'IIAUS4 = Cotton area (number of hectares), West Region United States. CINECHI = Cotton Net Export (‘000 tons), China. CTPDWOR = Cotton production (’000 tons), World. CI'PRUSl = Cotton Price, Delta Region, United States. CI‘PRUSZ = Cotton Price, Southeast Region, United States. CI‘PRUS3 = Cotton Price, Southwest Region, United States. CI'PRUS4 = Cotton Price, West Region, United States. CTPRWOR = World Cotton Price, Outlook Index 'A’ (c/kg). CI‘SHARG = Cotton share of total fiber for home use, Argentina. CT SHAUS = Cotton share of total fiber for home use, Australia. CI‘SHBRA = Cotton share of total fiber for home use, Brazil. CI‘SHCAF = Cotton share of total fiber for home use, Central Africa. CTSI-ICHI = Cotton share of total fiber for home use, China. CI‘SHEEC = Cotton share of total fiber for home use, EEC. CFSHEGY = Cotton share of total fiber for home use, Egypt. CI‘SHIND = Cotton share of total fiber for home use, India. CTSHJPN = Cotton share of total fiber for home use, Japan. CTSHKOR = Cotton share of total fiber for home use, Korea. CT SHMEX = Cotton share of total fiber for home use, Mexico. CI‘SHPAK = Cotton share of total fiber for home use, Pakistan. CI‘SHTU R = Cotton share of total fiber for home use, Turkey. 246 DFPSPARG DFPSPAUS DFPSPUSA DFPSPWOR NCPROD NCSHARG NCSHAUS II II It ll 11 247 Cotton share of total fiber for home use, United States. Cotton yield (tons per hectare), Argentina. Cotton yield (tons per hectare), Australia. Cotton yield (tons per hectare), Brazil. Cotton yield (tons per hectare), Central Africa. Cotton yield (tons per hectare), China. Cotton yield (tons per hectare), Egypt. Cotton yield (tons per hectare), North India. Cotton yield (tons per hectare), South India. Cotton yield (tons per hectare), West India. Cotton yield (tons per hectare), Mexico. Cotton yield (tons per hectare), Punjab Pakistan. Cotton yield (tons per hectare), Sind Pakistan. Cotton yield (tons per hectare), Turkey. Cotton yield (tons per hectare), Delta Region United States. Cotton yield (tons per hectare), Southeast Region United States. Cotton yield (tons per hectare), Southwest Region United States. Cotton yield (tons per hectare), West Region United States. Deflated cotton price, Argentina. Deflated cotton price, Australia. Deflated cotton price, Brazil. Deflated cotton price, Central Africa. Deflated cotton price, EEC. Deflated cotton price, Egypt. Deflated cotton price, India. Deflated cotton price, Japan. Deflated cotton price, Korea. l Deflated cotton price, Mexico. Deflated cotton price, Pakistan. Deflated cotton price, Turkey. Deflated cotton price, Memphis, United States Deflated cotton price, Montgomery, United States Deflated cotton price, Dallas, United States Deflated cotton price, Fresno, United States Deflated cotton price, United States. Deflated polyester staple price, Argentina. Deflated polyester staple price, Australia. Deflated polyester staple price, Brazil. Deflated polyester staple price, Central Africa. Deflated polyester staple price, China. Deflated polyester staple price, EEC. Deflated polyester staple price, Egypt. Deflated polyester staple price, India. Deflated polyester staple price, Japan. Deflated polyester staple price, Korea. Deflated polyester staple price, Mexico. Deflated polyester staple price, Pakistan. Deflated polyester staple price, Turkey. Deflated polyester staple price, United States. Deflated polyester staple price, World. Non-Cellulosic Production (’000 tons), World. Non-Cellulosic share of total fiber for home use, Argentina. Non-Cellulosic share of total fiber for home use, Australia. 248 NCSHBRA = Non-Cellulosic share of total fiber for home use, Brazil. NCSHCAF = Non-Cellulosic share of total fiber for home use, Central Africa. NCSHCI-II = Non-Cellulosic share of total fiber for home use, China. NCSHEEC = Non-Cellulosic share of total fiber for home use, EEC. NCSHEGY = Non-Cellulosic share of total fiber for home use, Egypt. NCSHIND = Non-Cellulosic share of total fiber for home use, India. NCSHJPN = Non-Cellulosic share of total fiber for home use, Japan. NCSHKOR = Non-Cellulosic share of total fiber for home use, Korea. NCSHMEX = Non-Cellulosic share of total fiber for home use, Mexico. NCSHPAK = Non-Cellulosic share of total fiber for home use, Pakistan. NCSHTU R = Non-Cellulosic share of total fiber for home use, Turkey. NCSHUSA = Non-Cellulosic share of total fiber for home use, United States. PCI'FUARG = Per capita total fiber use (kg), Argentina. PCI'FUAUS = Per capita total fiber use (kg), Australia. PCI‘FUBRA = Per mpita total fiber use (kg), Brazil. PCI'FUCAF = Per capita total fiber use (kg), Central Africa. PCTFUCHI = Per capita total fiber use (kg), China. PCI'FUEEC = Per capita total fiber use (kg), EEC. PCTFUEGY = Per capita total fiber use (kg), Egypt. PCI‘FUIND = Per capita total fiber use (kg), India. PCTFUJPN = Per capita total fiber use (kg), Japan. PCI'FUKOR = Per capita total fiber use (kg), Korea. PCI'FUMBX = Per capita total fiber use (kg), Mexico. PCI'FUPAK = Per capita total fiber use (kg), Pakistan. PCI‘FUTUR = Per capita total fiber use (kg), Turkey. PCI‘FUUSA = Per capita total fiber use (kg), United States. PSPWOR = Price of polyester (US$/ton), World. RCI'FNPCAF = Ratio Cotton to Fertilizer Price, Central Africa. Exogenous Variables C1” CONROW = Cotton Consumption (’000 tons), Rest of the World. CI‘PDROW = Cotton Production (’000 tons), Rest of the World. D66 = Annual dummy variable, equals 1 in 1966, 0 otherwise. D67 = Annual dummy variable, equals 1 in 1967, 0 otherwise. D68 = Annual dummy variable, equals 1 in 1968, 0 otherwise. D69 = Annual dummy variable, equals 1 in 1969, 0 otherwise. D70 = Annual dummy variable, equals 1 in 1970, 0 otherwise. D71 = Annual dummy variable, equals 1 in 1971, 0 otherwise. D72 = Annual dummy variable, equals 1 in 1972, 0 otherwise. D73 = Annual dummy variable, equals 1 in 1973, 0 otherwise. D74 = Annual dummy variable, equals 1 in 1974, 0 otherwise. D75 = Annual dummy variable, equals 1 in 1975, 0 otherwise. D76 = Annual dummy variable, equals 1 in 1976, 0 otherwise. D77 = Annual dummy variable, equals 1 in 1977, 0 otherwise. D78 = Annual dummy variable, equals 1 in 1978, 0 otherwise. D79 = Annual dummy variable, equals 1 in 1979, 0 otherwise. D80 = Annual dummy variable, equals 1 in 1980, 0 otherwise. D81 = Annual dummy variable, equals 1 in 1981, 0 otherwise. D82 = Annual dummy variable, equals 1 in 1982, 0 otherwise. D83 = Annual dummy variable, equals 1 in 1983, 0 otherwise. D84 = Annual dummy variable, equals 1 in 1984, 0 otherwise. D35 = Annual dummy variable, equals 1 in 1985, 0 otherwise. D36 = Annual dummy variable, equals 1 in 1986, 0 otherwise. DFCGPARG DFCG PAUS DFOILPR DFSBPUSA MUV OILPR PDGDPARG PDGDPMEX PDGDPTUR PDGDPUSA 249 = Annual dummy variable, equals 1 in 1987, 0 otherwise. = Deflated coarse grain price, Argentina. = Deflated coarse grain price, Australia. = Deflated cotton price, China. = Deflated cotton price, Punjab, India. = Deflated cotton price, Karnataka, India. = Deflated cotton price, Maharashtra, India. = Deflated fertilizer price, Argentina. = Deflated fertilizer price, China. = Deflated fertilizer price, United States. = Deflated oil price, World = Deflated soybean price, United States. = Deflated sorghum price, United States. = Deflated rice price, United States. = MFA dummy variable, equals 1 in 1974-77, 0 otherwise. = MFA dummy variable, equals 1 in 1978-81, 0 otherwise. = MFA dummy variable, equals 1 in 1982-1986, 0 otherwise. = Irrigated area ('000 hectares), Pakistan. = Irrigated area (’000 hectares), Pakistan. = Manufacturing Unit Value, World Bank. = Price of oil (S/bbl), OPEC. = Per capita deflated GDP, Argentina. = Per capita deflated GDP, Australia. Per capita deflated GDP, Brazil. Per capita deflated GDP, Central Africa = Per capita deflated GDP, China. Per capita deflated GDP, EEC. = Per wpita deflated GDP, Egypt. Per capita deflated GDP, India. = Per capita deflated GDP, Japan. Per capita deflated GDP, Korea. Per capita deflated GDP, Mexico. Per capita deflated GDP, Turkey. Per capita deflated GDP, United States. Annual Rainfall (mm) North India. = Annual Rainfall (mm) South India. Real interest rate (US T.Bill), United States. = Spring Rainfall Delta Region United States (inches). Summer Rainfall (mm) West India. Dummy Variable for Skip Row Policy (equals 1 1966-67, otherwise 0). = Soil Moisture Level, Southwest Region, United States. Fall Temperature Delta Region United States (Degrees C). Fall Temperature Southeast Region United States (Degrees C). = Fall Temperature Southwest Region United States (Degrees C). = Summer Temperature Delta Region United States (Degrees C). UNEMPEEC UNEMPUSA 250 = Summer Temperature Southeast Region United States (Degrees C). = Summer Temperature Southwest Region United States (Degrees C). = Summer Temperature West Region United States (Degrees C). = Time trend. = Unemployment Rate (%), EEC. = Unemployment Rate (%), United States. APPENDIX B EXOGENOUS VARIABLE ASSUMPTIONS FOR THE FORECAST PERIOD APPENDIX B Exogenous Variable Assumptions for the Forecast Period 1. Consumer Price Indexes and Exchange Rates. For the forecast period, the assumption of purchasing power parity (PPP) was made. PPP argues that exchange rates move according to the differential between inflation rates in countries. For example, say US inflation rate is 10% greater than the Japanese inflation rate. PPP says that, over the long-run, the Japanese Yen will appreciate 10% against the US dollar. Therefore, the purchasing power of the two currencies stays the same. In the model, the consumer price indexes and exchange rates were constant at their 1989 level in the forecast period. This means that the model forecasts are generated in terms of real 1989 dollars. 2. Gross Domestic Product (GDP) and Population. GDP forecasts were derived from expected growth rates for the 1990-2005 period. These growth rates are reported in . Population was derived in a similar way, using growth rates published by . The growth rates used in the model are shown in Table A.1. Table A.1. Growth Rates of Gross Domestic Product and Population. Region GDP Population Argentina 0.8 1.5 Australia 3.5 1.4 Brazil 3.3 2.1 Cent. Africa 4.0 3.0 China 2.0 2.5 EEC-12 2.5 1.0 Egypt 3.1 2.5 India 4.0 2.9 Japan 4.3 0.6 Korea 6.5 1.1 Mexico 1.5 2.1 Pakistan 3.5 3.0 Turkey 5.0 2.1 United States 3.0 0.7 251 252 3. Commodigg Prices. Forecasts of commodity prices (e.g., oil, fertilizers, soybeans, rice, and coarse grains prices) were obtained from Price Prospects for Major Primagz Commodities, Report No.814/ 88. International Trade Division, World Bank. This publication contains price forecasts for all the major primary commodities, based on econometric commodity models. 4. Weather Variables. Weather variables were set at their historical mean levels. 5. Production and Consumption in the Rest-of-the-World region of the model. Forecasts of production and consumption in the Rest-of-the-World region of the model were obtained by regressing these variables against time. The equations gave growth rates of between 1% and 1.5% per year. LIST OF REFERENCES List of References Adams, PG. and J .R. Behrman (1976), Econometric Models of World Agricultural Commodity Markets, (Cambridge, Mass: Ballinger Publishing Co.). Agbadi, I. (1988), An Econometric Analysis of the U.S. Cotton Man-Made Fibers and Textile Sectors unpublished Ph.D dissertation, Texas Tech University, May. Anderson, J. (1974), "Simulation: Methodology and Application in Agricultural Economics," Review of Marketing and Agxicultural Economies, Marchz3-55. Chow, G. (1960), ”Statistical Demand Functions for Automobiles and their Use for Forecasting," in AC. Heberger (ed.), The Demand for Durable Goods Chicago, University of Chicago Press. Cline, W. (1987), The Future of World Trade in Textiles and Apparel, Institute for International Economics, Washington, DC. ContiCotton (1990), The Administrations 1990 Farm Bill Proposal, Vol. IV, No. VI. Crook, F.W. (1987), Aggicultural Statistics of the People’s Republic of China, 1949-86, Situation and Outlook Series, Various Issues, Economic Research Service. Deaton, A. and J. Muellbauer (1980), Economics of Consumer Behavior, (New York: Cambridge University Press). Donald, K.R., F. Lowenstein and MS. Simon (1963), "The Demand for Textile Fibers in the United States," Economic Research Service, USDA, Technical Bulletin No. 1301. Dudley, GE. (1974), "U.S. Textile Fiber Demand: Price Elasticities in Major End-Use Markets," Economic Research Service, USDA, Technical Bulletin No. 1500. Ecevit, L.U. (1987), ”Recent Developments in the World Cotton Market and the Future Outlook," International Monetag Fund Staff Papers,380-405. 254 Evans, S. (1977) "Competitive Relationships Between Cotton and other Crops, by Region, 1976 and 1977," Economic Research Service, USDA, Cotton and Wool Situation, February. Fackler, PL. (1988), “Vector Autoregressive Techniques for Structural Analysis," North Carolina State University Department of Economics and Business, Faculty Working Paper No.113, January. FAQ, World Apparel Fiber Consumption Survey, Various Issues. Fiber Economics Bureau, Inc., Fiber Organon, Various Issues. Fiber Economics Bureau, Inc., Textile Organon, Various Issues. Ghosh, 8., CI. Gilbert and AJ. Hughes Hallett (1987), Stabilizing Speculative Commodig Markets Clarendon Press, Oxford. Gilbert, CL. and TB. Palaskas (1989), "Modelling Expectations Formation in Primary Commodity Markets," unpublished. Goto, J. (1988), Effects of the Multifibre Arrangement on Developing Countries. A Survey, World Bank, PPR Working Paper, WPS 102. Hamid, N., T. Pinckney, S. Gnaegy and A. Valdes (1987), "The Wheat Economy of Pakistan: Setting and Prospects," IFPRI, November. Houthakker, HS. and L.D. Taylor (1966), Consumer Demand in the United States 1929-1970, Analyses and Projections, Cambridge, Mass, Harvard University Press. Houthakker, HS. and L.D. Taylor (1970), Consumer Demand in the United States 1929-1970, Analyses and Projections: Second and Enlarged Edition, Cambridge, Mass, Harvard University Press. Hufbauer, G., D. Berliner and K. Elliot (1986), Trade Protection in the United States: 31 Case Studies Institute for International Economics, Washington, DC. Hwa, EC. (1981), "A Simultaneous Equation Model of Price and Quantity Adjustments in World Primary Commodity Markets,” World Bank Staff Paper, No.99. Hwa, EC. (1985), "A Model of Price and Quantity Adjustments in Primary Commodity Markets," Journal of Policy Modeling, 7:305-38. 255 ICAC (1988), Cotton: Review of the World Situation 3:3-9. Intriligator, D.M. (1978), Econometric Models, Techniques and Applications, New Jersey: McGraw-Hill Inc. Kirmani, N., et al., (1981), Effects of Increased Market Access of Exports of Developing Countries IMF Staff Paper, Washington, DC. Labys, W.C., (1988), ”Recent Developments in Commodity Modeling: A World Bank Focus," World Bank, PPR Working Paper, WPS 119. Learner, EB. (1985), Vector Autoregressions for Casual Inference? In Understanding Monetagy Regimes, ed. K Brummer and A. Meltzer. Caregie-Rochester Conference Series on Public Policy, 22:255-304, Amsterdam: North-Holland. Leuthold, RM. (1975), "On the Use of Theil’s Inequality Coefficients," American Journal of Agricultural Economics, 54:84-91. Maddala, GS. (1977), Econometrics, New York: McGraw-Hill Inc. Magleby, RS. and E. Missaien (1977), "World Demand Prospects for Cotton in 1980, with Emphasis on Trade by Less Developed Countries," Economic Research Service, USDA, Foreigp Aggicultural Economic Report No. 1000. Martin, W. and S. Suphachalasai (1988), “Gainers and Losers from the Multi-Fiber Arrangement," presented at 1988 Australian Economic Congress, Australian National University. Monke, EA. (1981), "International Market Demand for Cotton: A Survey of the Literature," Working Paper No. 2, The Interaction of Demand, Supply and Government Polig: A Case Study of Egyptian Cotton. Monke, EA. and L.D. Taylor (1983), “A Trade-Constrained Model for Cotton," unpublished. Mues, C. and P. Simmons (1988), “The Effects of US Farm Policy on Australian Cotton Revenues." presented at 32nd Annual Conference of the Australian Agricultural Economics Society, La Trobe University, Melbourne, 8-12 February. Mundlak, Y. and D.F. Larson, (1990), On the Relevance of World Aggicultural Prices, World Bank, PPR Working Paper, WPS 383. 256 Myers, R.J., R.R. Piggott and W.G. Tomek (1991), "Estimating Sources of Fluctuations in the Australian Wool Market: An Application of VAR Methods," Forthcoming in Australian Journal of Agricultural Economics. Orden, D., and PL Fackler (1989), "Identifying Monetary Impacts on Agricultural Prices in VAR Models," American Journal of Aggicultural Economics, 71:495-502. Palaskas, T. and P. Varangis (1989), Primag Commodig Prices and Macroeconomic Variables World Bank, PPR Working Paper, WPS 314. Pindyck, RS. and BL Rubinfeld (1981), Econometric Models and Economic Forecasts 2nd Edition, McGraw-I-Iill Book Company. Pelzman, J. (1988), "The Tariff Equivalents of the Existing Quotas under the Multi-Fiber Arrangement," (memo). Sampson, G. (1986), "Market Disturbances and the Multi-Fiber Arrangement," in Snape, R., ed., Issues in World Trade Polig, Macmillan Press, London. Sargent, T. (1979), "Estimating Vector Autoregressions Using Methods Not Based on Explicit Economic Theories," Federal Reserve Bank of Minneapolis Quarterly Review, Summer:8-15. Sargent, T. (1984), "Autoregressions, Expectations, and Advise. American Economic Review 74:408-415. Sims, CA. (1980), "Macroeconomics and Reality," Econometrica 4811-48. Sims, CA. (1986), "Are Forecasting Models Usable for Policy Analysis?," Federal Reserve Bank of Minneapolis Quarterly Review, Winterz2-16. Spinanger, D. (1987), "Will the Multi-Fiber Arrangement Keep Bangladesh Humble," The World Economy, March. Starbird, I., E. Glade, W. McArthur, F. Cooke and T. Townsend (1987), The US Cotton Industgg, Economic Research Service, USDA. Starbird, I. and J. Hazera (1983), "Analysis of Factors Affecting Cotton Yields," Economic Research Service, USDA, Cotton Outlook and Situation. Tarr, D. and M. Morkre (1984), Aggregate Cost to the United States of Tariffs and Quotas on Imports, Federal Trade Commission, Washington, DC. 257 Teigen, L. and F. Singer (1988), Weather in US Agriculture, Economic Research Service, USDA. Theil, H. (1966), Applied Economic Forecasting, Chicago: Rand-McNally. Theil, H. (1978), Introduction to Econometrics Englewood Cliffs, New Jersey: Prentice Hall. Thigpen, ME. (1978), "International Cotton Market Prospects," World Bank Staff Commodity Paper No. 2, June. Thigpen, ME. and DO. Mitchell (1988), "Modelling the World Fiber Market," presented at the 25th International Conference on International Commodity Market Modelling, Applied Econometric Association, World Bank, Washington, DC, October. UNCTAD (1986), Protectionism and Structural Adjustment, Geneva. US Department of Agriculture (1987), Maior World Crop Areas and Climatic Profiles. US Department of Agriculture, Cotton Situation and Outlook Series, Economic Research Service, Various Issues. US Department of Agriculture, China Aggiculture and Trade Report Situation and Outlook Series Economic Research Service, Various Issues. US Department of Agriculture, USSR Agriculture and Trade Report Situation and Outlook Series, Economic Research Service, Various Issues. WEFA (1990), World Economic Outlook Volume 1-4. Whalley, J ., et al. (1988), "Do Developing Countries Lose from the MFA," (memo). i W MICHIGQN ST "ll”!ll' 'lllgilfllmgnligfi L