World Bank Commodity Models Volume Ib. Papers Presented at the Wcrkshop on Commodity hlodels and Policies, Aarhus, Denmark, December 14-1 7, 1979 WORLD BAKK STAFF CO!tlhlOL~I'I3'IYOKKIKG P.4PERS Number o World Bank Commodity Models Volume Ib. Papen Presented at the Workshop on Commodity Models and Policies, Aarhus, Denmark, December 14-1 7, 1979 - - Sponsored by: Economic &dysis and Projections Department - The World Bank Institute &Economics. University of Aarhus - The Worli Rank Washington. D.C., 1JS.A Copyright C -1381 The l n t m u t d &nL la Rrcmrtructm ad Drvrbpnrnt I na wcuu, a<= 1818 H kf&. N W Wal)un~or\.D C ZOIJJ. U S A fint pnnlmg Idy 1961 SccaJ prrntmg Iw 1964 (m two \.duncrl Al n&tr r d m t h UnrlrJ9dcrd A m r mUOlll)TKA UXJW)(lr An E c a m m u C m a OF ITS s m m , PEwomuula Am) PIOSObCrS V - i l i . TABLE OF COSTE!!TS . ................................................. I LNTRODUCTLON ................................ ............................... A.. Producrlon and Exports . .......................................... 8 Conruprlon and Imports C Prlce Trends 111. mDEL STRUCTURE AND ESTIMATION . . . . . . . . . . * . . . . . . . . . . o o o a o e a ~ a o ....................................... ................. A.. Overvlaw of the Model and Past M e l h .. ....................................... 8 Supply Analysls ..................................... C D e r o d k u l y s l s D Prlcer and Stockm . ............................. IV MIDEL SIWLATLON AND PROJsCTIONS .................................... ........... A ..Ex-Port S l u l n r l o n 8 Model Projecrlons and Altarnrrlve Scenarios . .................................................. V CONCLUSIONS APPENDIX 11: TIHE TRENDS IN AREA. YIELD AND PRODUCTION. 1957-1990 .. THE VOW ECONOW: AN ECOXOHBtRIC WDEL OF ITS STWWRE, PERPORiUNCE AND PmSPECTS 1/ Thir paper a t t e q t r to: (a) model the rtructure and , \ a r f o ~ n c eof the vorld tea econoary, taking into account the c'ruractarirticr peculiar to tea; (b) arrerr the performance of the model hirtorieally and for projection purpose#; and (c) analyze the effect of alternative policy r c e ~ r i o ron the future developmnt of the world tea economy vith empbrir on projecting future pricea. . 2 Section XI provider an overview of the vorld tea economy r u a ~ r i z i a g key characterirticr of the product rnd tracing the p e r f o ~ n c eof key indicators (production, co~umption,trade and prices). Section I11 containa the model structure and the econometric e r t i u t e s . In Sectlon I V the hirtoric r h l a t i o n r of the model are prerented, together vith the extrapolation of the w d e l for projection to 1990. Soma policy a l t e r ~ t i v e rin p r o j e c t i e ~key endogenous variables to 1990 are also explored. Finally, Section V rrnr~arizer the conclusionr that can be dram from the tea w d e l simulations (vith cophasis an price projections). - 1/ We vould like to thank Peter Pollak, Jos de Vries and Maw-Cheng Yang for their invaluable assistance in overcoming problems related to the estimation and simulation of this tea model. Thanks are also due to the many reviewers a t the Aarhus Workshop (especially T. Akiyama) for their valuable co-nts vhich have been 1ncorpo;ated i n this paper. 11. AN OVERVIEW OF THE WORLD TEA ECONOHY 3. Ten may bc broadly c I a s s i f l e d i n t o t v o types--depending on whether t h e t e a leaves a r e f e r a e n t e d (black t e a ) o r unfermented (green tea). Green t e a production o r i g i n a t e d i n Chinn and I t s cons\mptIon a s a d r i n k was f i r s t introduced I n t o Europe by t h e Dutch r e t u r n i n g from t r a d e with China i n t h e 1 7 t h cer.tury. Commercial black t e n production s t a r t e d i n Assam, India, i n t h e mid-19th century. 11 S t a r t l n g with t h e B r i t i s h c o l o n i e s , black t e a producti'on h a s s i n c e spread a l l o v e r t h e t r o p i c s , and I t presently r e p r e s e n t s approximately 7s percent of a l l :ea protiucsd. V i r t u a l l y a 1 1 t h e t e a grown i n developing c o u n t r i e s (except f o r Taiwan nnd Lndonesia) is black tea. A. ---. Productlon and Exports 4. Uorld t e a production grew by 3.3 percent annually from 1961 t o 1977 (Table 1). Production I n c e t \ t r a l l y planned economies grew twice a s f a s t a s i n t h e developing count ries and, as a r e s u l t ,t h e c a n t r a l ly planned economies' s h a r e i n u u r l d producr !on i n c r e , ~ s e dfrom 19 p e r c e n t i n 1961 t o 24 percent i n 1977, v h l l a t h a t of t h e d c v e l o p l n ~c o u n t r i e s f e l l from 72 percent t o 67 percent. k a n u h i l e , t h e L t ~ d u r t r l a l i z e dc o u n t r i e s ' s h a r e r e m i n e d a t a h t 8- 10 percent. The respect l v e shares of developing c o u n t r i e s i n world output have changed s u b s t a n t i a l l y : that of t h c biggest t r a d i t l o n a l producers, India and S r i Lanlra dropped f r o n 55 pcrccnt i n 1961 t o 43 p e r c e n t i n 1977, Mile t h e A f r i c a n produccrs increased t h e i r s h ~ r cfrom 5 percent t o 11 percent l a r g e l y because of e x t e n s i v e a r e a cxpanslon In t h e 1960s (Table 2). world production g r o v t h (2.6 percent atri~trallyduring 1961-77) has, however, been f a s t e r t h a a a r e a expansion because of y i e l d improvements n w n g recent plantings. 5. The s t r u c t u r e of production of tca hns a l s o been evolving through t h e years. Productlon i n t h e e n r l y yenrs of t h e t e a i n d u s t r y was through " s t e r l i n g " company estates I n the Asian colonies. Hence i n c o u n t r i e s l i k e I n d i a , S r i Lanta and Indonesia ( t h e e n r l y e n t r a n t s i n t h e production of t e a ) , t h e e s t a t e production s t r u c t u r e s t i l l predominates, but n w primarily u ~ d e r l o c a l p r i v a t e s e c t o r w n e r s h i p . I t is only i n a f e v A f r i c a n c o u n t r i e s l i k e Kenya and Tanzania (whose t e a production u a s introduced l a t e r ) that s ~ l l h o l d e r sc o n s t i t u t e a l a r g e proportion of t o t a l tea acreage. S t a t e ownership of e s t a t e s h a s 3190 increased i n many of t h e s e c o u n t r i e s i n recent y e a r s and o v e r a l l a r e a expansion has, i n any c a s e , been increasingly ' determfned by n a t i o n a l hea development plans. 6. Although production In developing c o u n t r i e s grev h i s t o r i c a l l y a t 2.6 p e r c e n t a n n l ~ a l l y ,t h e i r e x p o r t s expanded a t only about 1.7 percent (Table 3) l a r g e l y because domestic constnption r o s e f a s t e r than production i n key Asian c o u n t r i e s . For i n s t a n c e , Indi.2 and S r i Lanka's share of w r l d n e t exports has d e c l i n e d even more drastic.: l l y ttlnn t h e i r s h a r e of prarluct ion (from 6 9 percent to 51 percent i n 1977). WLth production Increases o u w e i g h i n g population and income growth, African t e a exports grew a t 7.7 p e r c e n r per year and increased t h e i r s h a r e from 9 percent t o 19 percent i n t h e same period. -11 G. S. Ssrksr, nlr World Te;c Fco~~auy, - --- - . - - -- - Oxford Univ. P r e s s , 1972. Table 1: TEA PRODUCTION, BY WIN COUNTRIES AND ECONOHIC REGIONS Grouch Rate /. 196' 1965 1970 1974/76 1377 1961-77 I n ~ u , t r l a l i r a d Count t i e r --- -- - 87.0 90.7 125.5 156.8 167.5 4.5 Japan Turkey Other Asia and Oceania 692.2 738.3 776.0 858.3 940.2 1.6 India 354.4 366.4 418.5 497.3 560.8 2.9 S r l b n k a 206.5 228.2 212.2 204.8 208.6 -0.3 Inaoner l* 73.9 79.8 63.5 69.2 73.5 -0.9 Bangladenh 26.5 27.2 31.3 31.7 37.5 0.9 Other 30.9 36.7 50.5 55.3 59.8 4.2 Kcnya 12.6 19.8 41.1 57-4 86.3 11.4 Halaui 14.3 13.0 18.7 26 .O 31.7 5-9 Uganda 5-1 8.4 18.2 18.5 15.0 8.6 Mozambique 10.1 10.7 17.0 16.5 13-8 4 -9 Tanzania 4.5 5.7 8.5 13.6 16.7 9.1 Other 9.8 12.8 16.0 20 -4 25.4 7.4 Latin Awrica 8.8 23.0 34.3 46.3 44.7 8.7 . Argentina 5.9 15.5 25.8 35.3 34.O 11.3.- - - Other 2.9 7 5 8.5 11.0 11.6 8.5- ' Centrally Planned - Economies 203.5 216.6 294.8 387.1 418.2 5.0 * - --- -. * China, Mainland 156.5 158.8 220.0 291.7 309.5 581 m USSR 39.6 48.3 66.8 86.3 99.2 5z4 Other 7.4 9.5 8 .O 9.1 9 -5 1.6 ~ WORLD T3TU 1,047.9 1,139.0 1,350.1 1,600.9 1,759.5 -.. 3 3 - ..-- - /a Trend regressions against time. Source: PAO, Production Yearbooke, and International Tea Committee. Table 2: TEA AIW PUHTED, BY HAIN COUNTRIES AND ECONOMIC REGIONS -( '000 hectarer)- -(a* qua1 g r w t h i n 2) /a- - Indurtrialized Counttier --- - - - 101.0 123.3 150.6 1.6 4.8 2.4 Japan 48.5 55.5 -7.6 0.9 1.8 1.4 Turkey 13.5 28.8 ...O 6.5 14.O 7.9 Other 0.3 0.1 0.1 -6 -1 0.0 -4.5 Developing Countrier 910.7 1,006.8 1,043.6 - - - 0.8 0.9 0.9 . k i a and Oceania 7 817.2 814.8 0.1 C.3 0 1 India 331.3 360.1 365.0 0.8 0. 3 0.6 S r i Lak. 235.5 241.9 240.6 0.2 -0.1 0.1 Indoneria 135.8 104.1 101.1 -2.5 -0.6 -2.3 Bangladesh 3 7 42.7 43.0 2.8 0.2 2.O Other 70.4 68.5 65.1 0.1 -1.0 -0.1 Africa 68.9 1 4 3 0 171.6 6e l 4 05 6 .O Kcnya 15.9 49.8 66.0 9 2 6.8 9.1 Hal a d 11.8 1 5 8 17.2 2-4 2.1 2 -3 Uganda 6.9 19.1 20.9 8.9 2.0 7.4 Hot-bique 15.0 15.6 15.8 0.1 0.3 0 -3 Tanzania 6.8 14.0 18.4 5.4 7.1 6.0 Other 1 28.7 33.3 7.5 3.7 6.9 Latin America 3 2 42.3 53.3 1.3 5.7 2.1 . Argent inr 3Q.7 U . 3 44.4 , 1.0 6.7 1.9 0 Other 6 5 8.5 8.9 2.8 1.0 2.7 Centrally Planned Econodea --- 280.4 443.2 447.0 - - - 2.7 2.2 2.9 '9 China, Mainland 200.0 317.5 350.0 3.2 2.7 3.6 USSR 64.4 ?4.7 76.0 1.4 0.6 I 1.1 .. Other 16.0 0 21.0 0.7 n-3 0.3 WORLD TOTAL 1,292.2 1,543.3 1,641.2 - - - 1.4 1.6 1.5 - - - - /a Tread regreeeiona agaiart t h e . Source: FA0 and International Tea C o d t t e e . Table 3: TEA EXPORTS, /a - BY WIN COWTRIES AND ECONOMIC REGIONS Indust r la11zed C o u n t r i o --- - - - 25.1 29.1 53.7 61.1 67.9 6.4 Nother lands 0.7 0.9 20.2 21.2 18.O 30.3 Uni tad K i -do. 4 . 4 17.8 19.1 25.3 34-9 4.8 - Clruda 0.5 1.3 2.2 2.7 2.7 10.5 Turkey' 2.6 7.8 3.9 0.1 10.0 Japan4 8.O 4.6 1.5 2.9 3.6 -4.7 Other' 1.5 1.9 2.9 5.1 8.6 11 5 Dovoloping Countrierr 526.4 587.6 644.0 -- - 686.a 706.6 1.7 Asia and Oceania 470.7 508.1 511.7 521.8 523.7 0.7 India 205.0 201.1 200.2 220.9 224.0 0.3 S r i Lank8 193.1 224.3 208.4 195.9 185.5 -0.4 Indones la 32.2 32.3 36.9 47.9 51.2 3.6 Bangladesh 16.3 22.9 29.4 22.6 22.0 0.6 Other 2 4 1 27.5 36.8 34-5 41.0 3.4 Africa 49.1 65.0 108.8 136.5 154.4 7.7 Kenya 12.0 23.3 41.6 56.9 70.1 10.3 k l a v i 13.0 13.2 17.7 26.7 29.9 6.3 Uganda 4.2 6.9 15.1 15.l 15.O 8.8 Mozambique 9.9 10.1 16.6 14.1 12.3 3.2 Tanzania 3.2 4.4 7-1 10.7 12.1 8.- Other 6.8 7.1 10.7 13.O 15.O 5. h t i n b e r ' i c a 6.6 14.5 23.5 28.5 28.5 9 .6 b Argentiaa 5.4 12.3 19.1 ?2.2 27.8 9.1 Other 1.2 2.2 4.4 6 .3 5.7 10.2 Centrally Planned China; Hainland' 34.3 32.2 30.0 58.3 70.0: 5.3 - USSR E 5.7 11.1 9.9 15.1 21.4 1) 5.7 Other' 2.3 3.5 2.3 3.7 5.4 5.5 WORLD TOTAL --- 593.8 663.5 739.9 825.0 871.3 --- - 2.4 - * Indicates producing countries. /a Gro39 exports includ tng reexports. - From trend regressions against tine. Source: FAO, Production and Trade Yearbooks, and International Tea Committee. 7. Nearly 8 percent of world t e a exports a r e reexports from industrial- ized countries. T r a d i t i o n a l l y , a l a r g e proportion of t h e t e a s o l d i n t h e world e c k e t has been purchased by a handful of t e a wholesaling firms through London a u c t i o n s , and these f l r a a have blended and reexported ' e s e t e a s t o o t h e r c o u n t r i e s (including developing countries). Other major rce' q r t e r s include t h c Netherlands and t h e USSR. In recent years, t e a a u c t i c .and 4 d a s t i c blending i n key producing c o u n t r i e s have assumed g r e a t e r ; tance \ and t h e reexports from industrialized c o u n t r i e s have consequently d e c ~ l n e d . 11 Presently, about 25 t~ 30 percent of exports a r e s o l d through crop contracts- and t h e r e s t through auction a n r k e t s located p r i a a r i l y i n London, India, S r i Lanka and Ket~ya. Host of the a a j o r exporting c o u n t r i e s rely on s a l e s through auctions; only some c o u n t r i e s l i k e Argentina, Indonasla, Hozambique, Tanzania and Malawi export the majority of t h e i r production through d i r e c t sales. Uhile t h e q u e s t i o n of p r i c e ~ s n i p u l a t i o nby multinational tradinglblending firma is stlll a matter of contention, FA0 s t u d i e s have bhwn t h a t t h e process of p r i c e a r b i t r a g e appcrrs t o work well among t h e various auction markets. 2 B. Consumption and Imports 8. Given t h c h t s t o r i c n l context i n t h e development of t e a drinking (para. 3). t t i n not s u r p r l s l n g t o find t h e highest per c a p i t a consumption of tea among UK and the former British "colonial" and now industrialized c o u n t r i e s l i k e Ire..ind. Canada and Australia. While these c o u n ~ r i e sa r e key consumers i n t h e w r l d t e a cconoay, s e v e r a l trends i n t h e past one and a half decades a r e becoming apparent (sec Table 4): ( a ) Their share In w r l d consuaption h a s d e c l i t e d from 33 percent i n 1960 t o 20 percent i n 1974176. (b) Per c a p i t a consuaption is d e c l i n i n g i n a number of key indus- t r i a l i z e d countries. I n some c a s e s (notably UK and Australia), even t o t a l consumption is declining. (c) Among the major economic country groupings, the centrally planned e c o n o d e s has been t h e most d y u a d c consumers with t h e s h a r e of USSR and People's Republic of China increasing from 1 5 percent i n 1956 t o 26 percent i n 1974176. (d) ,India has emerged a s the single most important tea consumer i n t h e v o r l d (17% of v o r l d co-swnption i n 1974176). Furthermore, w r l d consumption of t e a has seen a n increase i n t h e proportion- a t e use of t e a bags and soluble ( i n s t a n t ) t e a i n comparison t o "loose" tea. This trend i n f i n a l use has r e s u l t e d i n t h e reduction of t h e q u a n t i t y of t e a -1 The volume of t e a sold through producer a u c t i o n s n w f a r outweighs t h a t sold through London (see Appendix I f o r d e t a i l s ) . -21 Simple c o r r e l a t i o n matrlces of t h e main auction markets gave c o e f f i c i e c t s of 0.85 o r more on both An annual b a s l s (1964-75) and on a monthly basis (1972-76). See Appendix I. Table 4: TEA CONSIRBTION, BY WIN COUNTRIES LXD ECONOHIC ReGIONS Indua t r i a l ired Count.:ier --- - - 476.7 490.3 535.3 566.6 609.5 1.4 United Kingdom 237.6 222.5 220.3 197.7 208.3 -1.2 Other Wartarn Eutopa 37.4 42.6 44.0 52.9 63.1 2.7 United Statas 51.0 60.0 67.1 76 .O 91.3 3.O Canada 20.1 19.9 18.5 21.6 22.1 0.6 Australia 28.0 28.9 27.7 25 .8 25.1 -0 -7 Japan 69.9 78.6 106.0 113.5 112.6 3.6 Other 30.7 37.8 51.7 79.8 87.O 6.8 India 139.6 166.0 213.0 274.0 310.0 5.2 S r i tanka 13.0 15.7 18.9 20.9 21.0 3.3 Indonesia 41.9 45.8 27.0 22.8 27.0 -5.5 Iran 15.6 27.3 37.7 41.3 50.O 7 .2 Other 66.0 77.4 106.0 143.8 167.8 5.0 M r i c a 69.8 75.6 92.9 100.6 109.2 2.6 QSY p t Kenya Sudan Other Latin America 9.9 12.6 20.1 29.5 29.9 6.C Argentina 1.2 2.7 5.O 8-0 8.8 10.4 Other 8.7 9.9 - 15.1 21.O 21.1 5.8 --- Centrally Planned '! Economies 197.8 2 2 0 9 295.5 -- - 389.1 406.7 5.4 China. Hainland 122.2 1 3 1 5 190.0 233.3 239.5 5.6 USSR 60.2 73.5 86.1 130.0 137.6 5.7 Other 12.4 16.2 '"-4 25.8 29.o 5.4 WORLD TOTAL 1,025.3 1,128.6 1,346.4 1,588.1 1,731.1 -- -3.3 - Source: PAO, Production and Trade Yearbooks, a a i International Tea Comulttee. needed per cup drunk and, thereiore, tends to o f f s e t any other f a c t o r s responsible f o r consuaption increase. . 9 Import grovth has essentially reflected the consumption patterns and r e e x ~ j r t i n grequlremcnts i n non-tea producing countries. The absolute decline i n the UK, f o r example, r e f l e c t s a declining trend i n domestic consumpiion and the declining importance of the London auctioa. The most dynamic importers i n the past decade-and-a-half have been Japan LI, the ,tnt r a l l y r lanned economies, Pakistan and Saudi Arabia (Table 5). C. Price Trt ads -- . 10 The ava-age price of the different teas sold i n the London Auctic.ns is a good indicator of changes in global supply and demand balance not only because of the c l e a r evidence of arbitrage and simllar w v a w n t s among major auction markets (sea para. 7 and Appendix I), but b a d o n is still the most important place where the widest variety of teas a r e sold. Using t h e average London auction price of t e a a s a n indicator of t h e general price tread, it may be reen (Tabla 6) t h a t tea prices have k e n coasistaatly declining before 1974. Between 1955 and 1972, t h e t e a prica declined, a t a tread r a t e of 2.3 percent annually i n n d ~ (current) terms and by a higher r a t e (48 percent 1 annually) i n r e a l (coaotant) t e r n , 2/ highlighting the substantial erosion of the purchasing p w e r of tea a s a c o m d i t y . Tea prices during t h i s period have a l s o k e n relatively stable. 11 Since 1973, however, prices have k e n rather e r r a t i c , with a particularly sharp jump of 75 percent taking place In 1977. According to the t e a trade, price fluctuations of such a n amplitude a r e unlikely t o p e r s i s t since they were caused by t h e coiacldental occurrence of short-term f a c t o r s !:k-.s "pull" e f f e c t of unexpected coffee price increase due t o Brazilian f r o s t , market speculation and various goverameat f i s c e l policy intervention). 11. While t h e average London auction price a c t s as a &,od baro-ter of the o v e r a l l world t e a m r k e t s and the t e a p r i c e i n individual countries geaeraly move i n l i n e v i t h each other, there has been changfs i n t h e price d i f f e r e n t i a l among t h e t e a s from d i f f e r e n t countries. Kenyan tea, which had - been s e l l i n g a t pricea substantially below that of S r i Lankan and North Indian tep has been receiving equivalent o r highsr prices i n Lcadon auctions since 1967. This development r e f l e c t s the improvememt i n Kenya's tea q u a l i t y that - - 11 4- Although it a l s o r a o r t s tea, Japan's imp-rts a r e essentially black tea which it does not ::educe i n significant quantities. -21 Using t h e ~ 6 r l dBank's -ndex of international i n f l a t i o n (indust countries' 'ICLP" index of US$ prices of manufactured exports, S developing countries) a s a deflator. - 31 The index of f l u c ~ u a t i o n(based on the root mean square of deviation from 1955 t o 1976, f i v e year moving average) indicates that tea was the lowest i n a range of 4.3 percent t o 30.3 percent f o r 41 commodities. (Source: Commodities aud E x p ~ r tProjections Division, Price Prospects f o r Major Primary Cotmnodities, Report No. 814178.) Table 5: TEA WORTS,/. BY nb;W COUNTRIES AND ECONOHIC REGIONS 1:routh Rats Ib - 1961 1965 1970 1914176 1977 1961-77 Induatrlallted Count r l e r Unltad Kingdom Othar Uartem Europe Clruda United S t a t e r Jrpaa Aurtralia Other Developing Countria. 1SO.2 170.2 227.4 -- - 262.4 287.1 4.o h l a and Oceania 7b.5 80.8 116.2 161.1 185.4 5.9 long Kong 6.8 5.7 8.4 7.9 11.4 3.0 Iran 5.8 9.7 6.2 14.1 12.3 5.9 traq 18.5 12.7 19.5 25.8 2500 1.7 Pakietan 1 6 1 20.5 29.5 47 .O 60.7 900 b u d i Arabia 2.8 308 506 7.8 12.8 8.0 Ocher 24.5 28.4 17.0 58.5 63.2 6.1 Mrica 68.0 80.6 98.0 88.6 93.3 1.5 Egypt 22.8 28.7 29.8 21.0 26.2 -2 05 tlorocco 13.9 8.1 14.2 13.6 12.4 0.7 Sudan 9.4 101 19.4 13.6 12.7 3.2 Other 21.9 33.7 34.6 40.4 42.0 6.1 Latin America 7.7 8.8 13.2 12.7 804 1.9 - - Chile -- '! Other - * Centrally Planned .. Economies --- - - - 23.5 59.2 45.4 86.8 86.8 7.5 USSR Other - /a Groae imports (including reexports) . - Ib Prom trend regressions against tine. Source: FAO,Production and Trade Yearbooke, and International Tea Committee. Table 6: TEA PRICeS (IAHDON AUCTION) BY KEY PRODUCING COUNTRIES (9143) Average All Tea. Sri Lanka India Kenya Xalavi Current US$ Conrtant US$ Current US$ 0 - Source: PAO. Production year*, and International Tea C o dttee. has a r i s e n from the large expansion and !mprove~ent programs i n i t i a t e d since the e a r l y 1963s. 12. With the increasing use of tea bags and soluble " i n s t a ~ t " t e a , i n the part decade and the advent of a m c h l e s s q u a l i t y - d i s c r i d n a t i n g generation of tea dri&.ers i n t r a d i t i o n a l t e a consuming countries, higher quality teas rppeat t o be l e s s prized i n the world aarket. A possible indicator of such a treed is the reduction of the predtm between high g r w n and low ,p.ova S r i Lankan t e a s i n the past decade. 1/ 'This predum (expressed a s a percentage of lou q u a l i t y tea) declined a t a fTtted annual r a t e of nearly 12% from 1967-77. - 1/ This comparison presumes that climate is the determinant of tea quality in t h a t the high grown t e a s generally provide better "liquoring" and "flavoring" characteristics. I n practice differences i n che arethod of manufacture ( l i k e the orthodox method versus the cutting, tearing and c u r l i n g [CTC] method) a l s o provides differences i n liquor and flavor (see Appendix I). Since only orthodox processing methods predominate i n S r i Lanka t h i s presumption is reasonable. 111. HODEL STRUCTURE AND ESfIHAfION A. An Overview of t h e Hodel and P a s t Models 13. The t e a model used i n t h i s study, is a world market model w i t h supply and demand disaggregated by t h e main prodr:cing/consuning c o u n t r i e s (or regions) and a a u r k e t c l e a r i n g equation 'or p r i c e determined s i m l t a n e o u s l y . The g e n e r a l scheme of t h e model s t r u c t u r e is given i n Chart I. 14. World supply is obtained by aggr c g a t i n g eleven caunt r i e s / r e g i o n s analyzed s e p a r a t e l y and world demand by a g g r e g a t i n g seventeen c o u n t r i e s / r e g i o n s a l s o analyzed separately. The exceas of supply over demand d e t e r d n e s t h e a d d i t i o n t o s t o c k s which a r e t h e primary determinant of t h e world i n d i c a t o r price. Data f tom 1960-78 were used f o r demand; supply and p r i c e a n a l y s e s cnvered t h e period 1957-78. Underlying t h e s t r u c t u r e of supply a r e corresponding a r e a and y i e l d equations. 15. Few models o f t h e world tea economy have k e n published. Adam and B e h m n analyzed t h e World Tea Economy i n a t h r e e region framework ( i n d u s t r i a l i z e d c o u n t r i e s , developing c o u n t r i e s eod c e n t r a l l y planned economies) i n t h e i r 1975 paper. 11 An econometric u*del of the world tea mmrket was developed by UNCTAD ir1970 21 but t h e d e r a i l s have not been published i n any o f f i c i a l UNCTAD docuaeTt. A model a t t h e same l e v e l of d i s a g g r e g a t i o n a s M a ~ and Behman was a l s o prepared by UNCTADIFAO i n s 1979. 31 - Only two o t h e r more disaggregated models have been developed: (a) an unpublished Ph.D. d i s s e r t a t i o n by h r t i (which was completed i n 1961 and is -11 F. G. Mars and J. R. B e h r u n , Feven Models of International Commodity Xarkets, U n i v e r s i t y of Pennsylvania, Wharton School, Department of Econornlcs, December 1974. - 21 UNCTAD S e c r e t a r i a t , Econometric Hodel o f t h e World Tea Economy, (mimeo) , prepared f o r p a r t i c i p a n t s o f t h e I n t e r - R e ~ i o n a l Seminar on t h e Planning of - . . - t h e Foreign ~ ; a d e s e c t o r , September 21 October 2 , 1970, Geneva. - 31 UNCTAD/FAO, Tea Simulation Hodel: A Technical Note, Annex t o Document TD/B/IPC/TEA/XC/2. rally Econden Kingdam Stater ia PR Irrdtutrialized America Developing Cent tan Lank. dnited Uaited Canada Auntralia Japan Other ladia Sri Indonen ranI Pakin Kenya Latin Other China, USSR Other Plumed 1 - W 0 R L D C 0 N S u H ECONOMY TEA r I \ WORLD TXE STOQCS V V PRICE London OF - (Implied) Auctiom) WORLD WORLD (Av. L QCUT uT FLOW t 1 - Chart W 0 R L D S U P P L > ) OF TEA Asia Africa America HADE trialized Lanka PnuUUCTXOX I dua Countries India Sri Indonesia Qcher Kenya Tanzania Uganda Other Latin In ~t.n;:t;:&::annqe clearly out of date); 11 and (b) an Econo~l?tricHodel of the World Tea Economy by UNCTAD. 21 16. The model prerented here attempt6 t o analyze the world tea econowy at a f a i r l y high level of diraggregation (eleven rupply and reventeen cogruming countrier o r rubregionr) and t o incorporate i n the rupply analyria both the planting (area planted) and hurbandry (yield) decirionr. Thir, t o our knouledge, har not k e n attempted before. The remaining portion of t h i r chapter r e t r out the rationale f o r the formulation and empirical e r t i m t i o n of rupply ,demand and price. B. Supply Analyair 17. Tea ir a perennial crop. It giver itr f i r a t yield about tw or three yearr a f t e r planting. Depending on location and type of planting mterialr, it taker raven t o nine yearr to fully mature. However, more than SOX of u t u r e y i e l d r a r e attained a f t e r f i v e yearr. A t r e e can continue t o produce u n t i l it ir 60 t o 90 yearr old with only a a l i g h t :ecline in yieldr toward the end of the period. 18. In deciding on t h e expansion of tea production tug typer of decirionr need t o be diatinguirhed: ( i ) a long term invert me^-t decirion on neu area expanrion with new planting andlor replanting of t.-eer; and ( i i ) a decirion on the extent t o - z r o r t t o variour crop hurbandry procedurer which could a f f e c t tea yieldr (and thus production) on a s h o r t t e r n barir. 19. With rerpect t o the investment decision a number of factorr should be borne i n mind: - Inverting i n the i n i t i a l planting of tea, unlike annual cropr, implier a long term commitment of rerourcer (especially Lor e s t a t e r where, i n addition t o factory and ~ c h i n e r y ,housing and h e a l t h f a c i l i t i e s f o r workers have t o be provided). - Tea is d i f f e r e n t from most other crops i n its high per hectare labor requirement and its consequently high employment impact. -11 V. N. )turti, An ~ c o n & t r i c Study of ' t h e World Tea Econom, 1948-1961, Unpublished Ph.D. diqaertation, University of Pennsylvania, 1961. I n t h i s study, supply is d i s regated i n t o India, S r i Lanka and r e s t of :he world; denand covers industrialized countries, India and r e s t of the world. -2/ T. Akiyama Economtric Hodel of the World Tea Economy, mimeo, UNmAD, Nov. 1979 (an e a r l i o r version is i n Document No. TD/B/IPC/TEA/AC/~). The model covers supply equations f o r India, S r i Lanka, Kenya, other Africa, Bnngladesh, Indonesia, Argentina and r e s t of the uorld; and demand equationc f o r India, United States, United Kfngdon, other fcdus:rialized countries, o t h e r developing and E. Europe and USSR. .Also labor costs constitute a large proportion of t o t a l cost of production for tea. L/ - Gross returns per hectare is a l s o among the highest for export crops crown i t r area6 with good tea g r w i n g conditionr. L/ Hence moving 7ut of tea production is, i n practice, d i f f i c u l t b c a u s e of high sunk c o s t s ano the f a c t t h a t s h i f t i n g production from t e a t o another crop a l s o incurs s i g n i f i r a n t expenditures (e.g., i n the removal of trees and factory) and brings abo\:t the reduction i n both labor employment and foreign exchange. . 20 With regard t o the husbandry decirion, tea production can k s i g n i f i c a ~ t l yincreased i n the short term by two methods: - "Coarse pluckingM--the plucking of about f i v e leaves and a bud rather than the usual two leaves >nd a bud during each plucking cycle. Although tea y:eld is thereby increased, l w e r quality rerultr. - Application of nitrogenous f e r c i l i z c r (usually amoniuw sulphate) with yield rerpocues varying from 5 to 8 @ per kilogram of nitrogen. This response could carry over more than one year depending on the amount applied. The a v a i l a b i l i t y of these methods means that, theoretically, tea producers can adjust rather quickly to changing demand conditions. 21. These decisions depend on a s e t of factors which varies with the different modes of t e a production. I n countrier vhere private sector e r t a t e r predominate, one would expect that fiaancial incentives affecting the tea enterprise (product price and input costs) a r e l i k e l y t o be the most important determinants. I n smallholder dominated countries, non-financial factors could be a s inportant as financial factors. 22. The extent the above factors and decisions predodnate would depend on h w r e a l i s t i c is the modelling of producers' investment and yield stimulation decisions and the a v a i l a b i l i t y of relevant data. - - .9 -1/ See f o r example, Y. L. Lim, "Impact of the Tea I n d d t r y on the Growth of ' the Ceylonese Economy", Social and Economic S t u d l e g December 1968; R. W. ) Palmer-Jones, Production and Harkecing of Tea i n w w i , University of 'Reading, Departaunt of Agricultural Economics and Farm Hanagement, Developaent Study No. 15 (1976); N. A. Stern, An Appraisal of Tea Production i n Small Holdings i n Kenya, Paris, Development Centre of OECD (1972); and G. S. Sarkar, Tea: Some Policy Issues, IDS Discussion Paper No. 124, February 1968, University of Sussex. - 2/ J. de Vries, Tea: The P o e s i b i l i t i e s f o r International Action, World Bank C o m o d l t y Taper S o . 28, Janaary 1977, Xikriex ZIT, p . 8. 23. Reviews of t h e supply analyses of perennials have already been under- taken by a number of authors. Rather than provide a global perspective, v i r t u a l l y a l l the s t u d i e s p e r t a i n t o case s t u d i e s of a p a r t i c u l a r crop i n a p a r t i c u l ~ rcountry ( o r region i n a c o u n t v ) where d e t a i l d a t a is more r e a d i l y available. Complex models l i k e that of French and Hathews 21 on US asparrgur ( a perennial v l t h a bearing l i f e of 10-15 years) a r e a b l e to capture t h e d e c i s l o n -king f ramework with respect t o new acreage planting and removalr of o l d plants. 24. nodels on perennial8 i n developin8 countries have, hwever, had t o be more m d e s t because of d a t a l i d t a t i o n r . Batelan postulated a planting d e c i s i o n model f o r Ghanian Coca which incorporated t h e stock of trees (number a f bearing t r e e s ) but proceeded t o e s t i m a t e h i s rupply equation on the b a s i r of harvested acreage a s a proxy. 31 Similarly, Wickenr and Greenfield 41 derived a supply f u n c t i o n f o r c o f h i n Brarll bared on a n invercmnt fTnction r e l a t e d t o atock of treer but fimally ured planted acreage as a proxy by m k i n g simplifying assumptions on constant planting d e n s i t y and random uprooting and abandonment of treer. They did, hwever, attempt an a p r i o r 1 d i s t i n c t l o n b e t w e n t h e i n v e r t r a n t d e c i r i o n and t h e planting decislon. lbra r e c e n t l y , Hartley 51 f o r u l a t e d a d y d c programdng model f o r perenn!sl c r o p supply. He aGlyzed the producerr' decision t o plantlreplant and ~ p r o o t , with t h e aim of capturing i n t e r a l i a t h e i r expectatione on f u t u r e y i e l d (revenues), t h e opportunity c o s t of a l t e r n a t i v e land use, and expected d e p l e t i o n of c a p i t a l stock. The n j o r problem i n estimating t h i s type of model is that of o b t a i n i n g t h e basic data. I n s m r a r y , d e s p i t e r a t h e r s o p h i s t i c a t e d f o r m l a t i o n r of t h e d e c i r i o n process a f f e c t i n g t h e supply ,f perennials, a c t u a l a n a l y s i s (even a t t h e country l e v e l ) has required the estimation of proxy variables, l i r i t e d e s r e n t i a l l y , t o acreage, output and prices. - 11 See: H. J. Batemn, "Supply Relations f o r Perennial Crops i n t h e Less Developed Areas," i n C. R. Ubarton, Jr., ed., Subsistence Agriculture and Economic Development (Chicago: Aldine Pub. Co., 19691, pp. 243-253; D. L. S. Lim, Supply Response of Prirary Producers, Penerbit University Halaya, Kuala Lumpur, 1975; and R. War1 and J. T. CrrPPings, Agricultural Supply Response: A Survey of the Ecot~noctricEvidence. New York: Praeger Pub- l i s h e r s (1976). -21 B. C. French and J. L. Hatheus, "A Supply Response Hodel f o r Perennial Crops," American Journal of Agricultural Economics, No. 53 (August 1971), - pp. 478-490. - 3 See: H. J. Bateman, op. cit. and D. L. 5 . Lim, OD. cit. 8 -4/ H. R . Wickens and J. N. Greenfield, "The-Econometrics of Agricultural Supply: An Application t o t h e World Coffee Market", The Review of Economics and S t a t i s t i c s , Vol. LV 1973, pp. 433-440. -51 H. J. Hartley, Research Proposal on: The Construction of Econometric Models f o r t h e Supply of Perennials, (mimeo), Development Economics Department, World Bank, June 1979. 25. The approach adopted i n t h i s study may be characterized a s pragmatic i n t h a t primary a t t e n t i o n is focused on an a p r i o r i understanding of the key measurable variabler likely t o a f f e c t the development of tea ,upply, including the r e l a t i v e importance of the long-term planting (invertment) decirion and the short-term yield (hurbandry) decirion. N o f o r a ~ l a t i o n r tof the structure and performance of supply a r e attempted: (a) analysis of area and yields separately, and (b) an aggregate rupply analysis (which combines the area and yield effects). 26. For the relevant country o r region, t e a production is decompored a s f o l l w r : - where: Tea production ('OM) ton.) i n year t; Q: - h.lt Hature tea area ('000 ha) i n year t, defined a s planted a r e a lagged s i x years; L1 YLDt Yield per u n i t pature area (tonslha) i n year t. 27. Mature area f o r each country o r region is postulated t o be determined by t h e f o l l w i n g general equation: where: - PTEA D =-the average price of a l l t e a s i n London auction ( i n US cents/lb)-deflated by a proxy f o r input costs. 28. Since mature at. .s defined as the planted area lagged s i x years, t h e formulation of Pature -creage response 1-,lies t h a t price expectations governing t h e decieion on new planting o r replanting of t e a is essentially -11 S t a t i s t i c s on mature o r bearing area is d i f f i c u l t t o obtain f o r the majority of producing countries. However, planted area data is available from FA0 and the International Tea Comnittee. Planted area lagged s i x years was f i n a l l y used because that l a g would have captured more than 60% o t yielding capacity by then. Besides, the use of longer lags did not improve t h e e s t f m t e s and would re4uce t h e d e g r e e s of freedom i n estimating the equations. t h a t of t h e naive "extrapolative expectations" model of Nerlove. 3 I n t h e l o g form, t h e p r i c e c o e f f i c i e n t m y be i n t e r p r e t e d a s t h e long-run e l a s t i c i t y o f p r d v c t i o n s i n c e it measures t h e response a f t e r planted acreage has matured and - ~ t i n gadjustments have been ~ d e . 3 29. The f o r m l a t i o n of t h e r e a l p r i c e of t e a (PTEA D) is e s s e n t i a l l y a s a simple oxy f o r n e t income accruing t o the tea e n t e r p r i s e . F e r t i l i z e r , vage and i n t e r e s t c o s t s c o n s t i t u t e t h e l a r g e s t component i n production costs. I n t e r e s t c o s t s , more relevant t o t h e long t e r n decision, is not r e a d i l y a v a i l a b l e os a country basis. 3/ Similarly, t h e relevant wage c o s t a a r e a l s o d i f f i c u l t t o obtain. For t h e c o u n t r i e s which have r e l a t i v e l y l a r g e coffee i n d u s t r i e s i n o r adjacent t o t h e i r tea a r e a s ( e r r e n t i a l l y t h e African and Latin American c o u n t r i e s ) , t h e p r i c a of coffea (PCOF) A/ is ured as the Port relevant def l a t c r o r proxy f o r u n i t input c o s t s ( p a r ~ i c u l a r l ylabor). 5/ However, u n i t f e r t i l i z e r c o s t (mnoniua rulphate) A/ are used, iarteaa, f o r d e f l a t i n g t h e t e a p r i c a i n the b i a n countries, because they do not face r idlar input competition from industries with readily available w r l d price d a t a ( l i k e coffee). - -1/ The expected p r i c e P* Pt,l. (See )(arc Nerlove, The DynrPi.c s of Supply: Estimation of Farmera' Response t o Price. aaltiwre: Johw Hopkins P r e s s , 1958). Thc use of lagged a r e a i n the above equation (2) is not t o c a p t u r e t h e long-run e l a s t i c i i i e s implied i n ~ e r l o v e ' ; formulation of a d a p t i v e expectations s i n c e such a n i n z e r p r e t a t i o n f o r a perennial crop involving investments with long g e s t a t i o n is r a t h e r tenuous. -21 T h i s is i n l i n e v i t h t h e d e f i n i t i o n of short- and long-run p r i c e responre given i n J. R. B e h r ~ a n ,%nopolistic ,Cocoa pricing", Anerican Journal of A g r i c u l t u r a l Economics, hug., 1968. The short-run e l a s t i c i t y w u l d be t h e response t o harvesting and husbandry over a t i m e period t h a t is not long enough f o r new plantings t o come i n t o bearing. -- -31 It may, i n any case, net be very important especially awng the "older" t e a groving c o u n t r i e s f n Asia (where t h e bulk of t e a production capacity is old and usually f u l b depreciated). * -41 The c o f f e e p r i c e c h o s e r r i s (;uatemalan, prime, washed, s p o t Neo York. -5 The r a t i o n a l e is t h a t d i f f e r e n c e s i n t h e r e l a t i v e p r i c e of t e a and c o f f e e e s s e n t i a l l y r e f l e c t t h e r e l a t i v e value of t h e umrginal product per u n i t input. -61 US wholesale p r i c e of ammonium sulphate expressed a s an index (PPI) with 1970-100. 30. Results of the mature acreage functions a r e given in Tab12 7. Estimations werc by ordinary l e a s t squares regression. k'here auto-correlation may be problematical the Cochrane-Orcutt correct ion procedure uas used. -11 31. Except f o r Latin M r i c a and, possibly, Tanzania, the long run supply e l u s t i c i t i e s a r e very low. Although t\e signs a t e generally correct, significant price e f f e c t s e r e found only f o r Latin America and overall developing countries. Lagged M t u r e area was significant f o r a l l producing countries o r subregions except f o India, indicating t h a t the planting ~ decision is related more t o the previous years planted area rather than price. A systematic increase in the planted (or -cure) areas with time ( a reflection of tha r a t e of expaltrion a t t r i b ~ t a b l et o government tea development program) was found o111y i n the African countries of Kenya and Tanzania. 32. The y i e l d f o r each c o u n t q o r region is postulated t o k determined by tha following senaral equation: where: - Ln PTEA fl and T have already k e n defined; and CLF climatic f a c t o r s (e.g., amount of r a i n f a l l ) . In the log form, the price c o c f f i c i ~ n t smay be interpreted a s the short-run e l a s t i c i t i e s of production since t h i s f o r m l a t i o n permits response rhrough "coarse plucking" and f e r t i l i z a t i o n over a time period which is not long enough f o r newly planted o r replanted areas t o come i n t o bearing (para. 20). Ln P'EA Dt-2 is included to a l l w for the popsible carryover effect of f e r t i l i z e r application. The time trend may be interpreted as representing productivity improvements over time due, f o r example, t o the introduction of high-yielding clonal varieties. 33. The yield functions a r e given i n Table 8. Climatic data for including C u t a s an independent variable is not available. But, a s w i l l be discussed i n pa;ra. 49, t h i s d s s i o n is not expected t o be particularly serious f o r tea. 34. The r e s u l t s id Table 8 shov that, i n contrast t o the mature area' equations (Table 7). significant price e f f e c t s were obtained from one-year lags f o r India and Uganda and two-year lags f o r Indonesia and other Asia. Significant t i m e trends were obtained f o r a l l the equatiops except f o r t o t c l industrialized countries. Thesc results show that, i n ge.Qeral, the Asian countries a r e more receptive t o short-term price incentives, and thus correspond with the expected tendency f o r private sector-dpminated production f r e receptive t o changes i n price incentives. Tha such an inter- preta ion appears t o be c l e a r e r i n the case of short-run i c e response rather -11 See D. Cochrane and G. Orcutt, " ~ p p l i c a t i o nof Least Squdres Regressions t o Relationships Containing Auto-Correiated Error Terms," J. Am. S t a t i s t . 9 s 3 0 r . : V n l . 6 6 , 3 ? . 72-61, 1949. ~0t.1 Coetrally ? l a r d Croaala* ) .4217 - 0 . 0 1 ~ 0 . 7 2 0.0015 0.r))) 0.0279 2 . a (I.-) (-0.42LL) (S.OB51)- (1.6855) ~0t.1 W*IWIU C c r ~ r i s. 4.1248 0 . 0 2 ~ o . n n 0 . m ~ 0 . ~ 1 5 0 . m ~ 2.27 (3.7547)- (2.4221) ( 1 . W ) 0.57?))- Other t T i n perlod m u d LB tho -1)~ia - -- is 1937-78: a11 rp~.sloosw r o e s t l r t d r u h nl IOR-IO~ ~ptcif~..cim. b Tho f l p n In p.rarhasi8 hlor d c a f f i c 1 . u l a rho com8pmdl.l c auclatlc. S U Stmodad error of a c l u t o . & D.2. R R b l m Y . t m statistic. 0 auto-rogroasln coofflclmcs I r chr b c h r r r Q r c o t c procodof.. Total I r h r t r i r l l s d m t r l r r Total C o t r r l l y T l d L m l r r Iota1 D w m l q l m @ Caatrlom Aml. 1J i b # r l U r n l a r l a 0tb.r Afrler 7-1. o w 0th Ltl. M e . -- T i n )rrid 4L. c h m a l y l r 11 l9S7-rr dl q-mtlmam r r e e o t l r t d umlna b loplo: m p ~ l f h t l o s . lhf l m L.m b e r l a klwrcL d f l c l r t 11 cbermrrr)oaamt mutlmtlc. m stdr+d .~XTPC of UCWO. D.Y. - D u b l a mat- m u t i a t k . 0 rco--mi- c d f l c l e ~ ~In cba rachraoOrcPtt procdsre. m than nature area response may be explained by the f a c t that the deflators used f o r tea prices a r e more relevant t o short-run (yield) response. 35. From the time trend variabie, one could i n f e r t h a t productivity improvenects due t o technological change (better yislding varieties) appear t o be highest in Uganda, the Cent r a l l y Planned Economies, Tantanla, India and other Asia (primarily Papua New Guinea). Theee countries experienced more than a ivo percent increase in yield a f t e r the e f f e c t of short term price had k e n taken i n t o account. The trend declines i n yield f o r S r i tanka and Uganda ~ el efc t e the protracted years of ~ ~ t r u c t u r disruptions i n t h e i r tea a l economies. 36. Total world production can then be obtained by summing up the pro- duction o b t s i n e r f o r each country o r region. - n (4) World Production 9. P Q: 1-1 ...n where: 1-1, a r e the number of countries or regions being studied. Aggregate Supply Formlation 37. An a g ~ r e g a t ef o m l a t i o n of svlpply was a l s o t r i e d jn order t o t e s t whether the e s s e n t i a l d e t c r d n a n t s of acreage and yields, formulated separately, could be captured in a single equation. Taking the measurable variables postulated t o a f f e c t both area and yield, production (QS) f o r each country o r regior. is postulated t o be determined a s follows: - ( 5 ) QS fl' (Ln PTEA Dt-l ,PTeA Dt-2, PTEA Dt-7, T) t where: P T U D is defined i n para. 27 and T is tt.o t i m e trend. The estimated supply functions a r e given i n Table 9. 38. The estimated supply equations ~n Table 9 appear t o have captured the key r e s u l t s from separate area and yield analysis, particularly with respect t o shorc term price responw and the time trend. As obtained i n the yield I equations, the s h o r t term responsiveness of India, Uganda, InConesia, and other Asia were a l s o significant a t approxinately the 5 percent level or better. The time trend, essentially reflecting the combined d f e c t of s y s t ~ m a t i carea expansion (essentially government expansion s emes) and ik technological change, was significant f o r a l l countries and s regions except f o r Indonesia. Africqn c ~ u n t r i e s ' overall production grew a t a n overall r a t e of about 8 ?ercent annually a s a r e s u l t of these factors, countries&xperienced a corresponding growth of l e s s than 2 p cent. ,Isian Countries k i t h s p e c i a l aulti-period t e a development plans (particularly Kenya, Tanzani; and Uganda) had the highest trend coefficients. 39. The long term price (PTEA D 7) effects, while a l s o significant f o r Asia and developing countries as a v k l e , were not transmitted from the mature a r e e equations t o the overall supply formulation a t the country/sub-regional l e v e l except for Latin America. Sri Lanka's negative v i ~ l r lt r e n d s dnminzted the trend of overall supply; but i n Uganda, f a s t e r area expansion compensated f o r the declining trends i n yields. C. Demand Analysis 40. A review of the trend i n conslarption and prices (paras. 8-12) reveal a number of features which need t o be captured i n analyzing t e a demand. The s t ~ a d ypric? decline, the increase i n consumer incomes and population and the f a c t that t e a remains the cheapest beverage per cup have provided favorable conditions f o r consumption expansion. A t the same time, a corresponding s h i f t i n preferences from t e a t o coffee and thh advent of t e a bags and instant t e a have tended t o o f f s e t consumption expaneion. The relative etrength of these opposing developments determine the path of tea coneuption i n varioue countries o r regions. 41. Tha poetulation of tea demand which has given us the most promising r e s u l t is the following: where: - - Q%C, Apparent consumption per capita; PTEA/PCOP Average price of a l l teas i n London auction o r PTEA ( i n US cents/lb) deflated by t h e average price of coffee o r PCOP (Guatemalan, przme washed, spot, New York i n US cents/lb). - CDPPC CDP per capita. 42. Data on tea consumption is d i f f i c u l t t o obtain on a worldwide basis. The d a t a used i n t h i s study is t h a t of "apparent consumption". 11 Annual stock adjustments t o apparent consumption have been made f o r only a few n e t importing countries (mainly UK, US and Australia) and some producing countries (mainly India, S r i Laaka and Bangladesh). 43. The choice of prices was another d i f f i c u l t y i n estimating demand functionp on a time s e r i e s basis. p e only readily available price series- obtained from t h e London t e a auctions (PTEA)-is essentially a wholesale price. To extrapolate t h i s price a s having an e f f e c t on t e a consumption i n other countries v i t h changing t a r i f f s and i n t e r n a l taxes and charges can, therefore, be somewhat tenuous. Time s e r i e s of r e t a i l prices, however, were not available. ! . 44. The demand results, obtained by l e a s t squares regression 'and . * - Cochrane-Orcutt procedures, a r e given i n Table 10. 8 - - I/Fornetexportingcountries, apparentconsumption production minus net exportn; f o r net importing countries, apparent consumption = net imports. -- Othar L 1.5593 0.01W 0.9122 0.0571 1.70 (12.57j1)- (12.8961)- r & t ~ i r parlod l a 1960-78; al? eqnaclms were astlprcod usin&a lot-lo* spcclflcatim. b Ibo fl-· in p.rmthesls klw uch coafflclcac 1s tha corrampoodlq t statlsclc. s l ~ l ~ l c r nc cc tho It 1.r.l. C.slgnlflemc ac tho 1Z 1ovaI. 4 5 . The use of coffee price a s a d e f l a t o r ( t o capture competition between t e a and c o f f e e ) provided expected s i g n s f o r the relevant c o e f f i c i e n t s and prodllced, s t a t i s t i c a l l y , the best o v e r a l l r e s u l t s , but other d e f l a t o r s ( l i k e the World Bank's I n t e r n a t i o n a l Price Index) a l s o yielded n i g n i f i c a n t p o s i t i v e c o e f f i c i e n t s f o r the e f f e c t of t e a prices on consumption. For high i n c k c o u n t r i e s with high per c a p i t a consumption ( i n d u s t r i a l i z e d countries and I r a n ) a time t r e n d was included i n t h e demand functions t o capture the p o s s i b i l i t y of per c a p i t a c o n s u q t i o n reaching s a t i a t i o n l e v e i s and the systelaatic change i n t a s t e away from tea. Because of the d i f f i c u l t y of obtaining comp?rable d a t a only a time trend is included (for o t h e r i n d u s t r i a l i z e d countries, o t h e r developing countries and c e n t r a l l y planned economies). 46. The r e s u l t s i n d i c a t e t h a t high income c o u n t r i e s (especially Australia and, t o a l e s s e r extent, UK) appear t o be systematically switching out of t e a because of t a s t e r a t h e r than r e l a t i v e p r i c e reasons (i.e., t h e i r time c o e f f i c i e n t s r a t h e r than t h e i r r e l a t i v e price c o e f f i c i e n t s were significant). I n s i g n i f i c a n t response of p e r c a r i t a tea consumption t o changes i n the r a l a t i v e p r i c e of tea and coffee were found i n a l l cases except f o r India and %enya. Given t h e tendency f o r compensating s h i f t s i n the demand function (see para. 3) over the years such a r e s u l t may not b a l t o g e t h e r unexpected. 47. Only f o r t h e UK, Japan, I r a n and Pakistan do t h e income e l a s t i c i t i e s deitved i n t h i s study conform with those previously estimated by the Food and Agriculture Organization of the United Nations (PA01 on the basis of cross- s e c t i o n a l expenditure studies. 11 The r e s u l t s a r e d i f f e r e n t f o r t h e o t h e r f i v e c o u n t r i e s f o r which FA0 e s r h a t e s a r e available. A p a r t i c u l a r l y wide d i f f e r e n c e i n estimates exists i n t h e case of Indonesia, where FA0 has an income ( e x p r i d t t u r e ) e l a s t i c i t y of 0.8 (expenditure survey of 1969) while our a n a l y s i s g i v e s a n income e l a s t i c i t y of -1.05. This probably r e f l e c t s the d i f f i c u l t y of using apparent c o ~ ~ n m p t i oans a measure o f "actual" consumption. 21 The income e f f e c t on t e a consumption is p a r t i c u l a r l y strong f o r Iran, ~ e % a , L a t i n America, USSR, S r i La* and India. 48. T o t a l world consumption i n t h e t e a model is obtained by s d n g up consumption obtained f o r each country o r region: m , I (7) World consumption QDTEA = I: QDTEAj 1-1 - 1 FAO, Income E l a s t i c i t i e s of Demand f o r Agricultural Products, General Studies Group, Xarch 1976. - 21 An important problem i n Inrlonesia during recent years has been the limited a v a i l a b i l i t y of green t e a (about 50 percent of domestic consumption) as a r e s u l t of a steadv declfne I n dnnestfc sn2llholEer 7roduc:fon and of import r e s t r i c t i o n s . Such a s i t u a t i o n h a s r e s u l t e d i n a per c a p i t a d e c l i n e i n apparent consumption of t e a i n Indonesia. where: - QMeA, t o t a l cona.umption of tea for the Jh country or - region. ... j 1, a are the nuaber of countrice or rcgionb being studied. D. Pricee and Stocke 49. Tuo characteriatica of tea price behavior ( i t # relative a t a b i l i t y and aecular decline through 1973) have already k e n noted (para. 10). The relative a t a b i l i t y pry be attributed t o the rha~z:teriatic of the tea crop and ite rrrrketing etructure. Although aeaaombility of harvesting (plucking) exiate, moat producere can pluck throughout the year. Hence, for a leaf crop l i k e tea, adverae weather during a few ronthe of the year has leas impact on annual yielde than for seed o r f m i t cropa affected during the brief but c r i t i c a l perioda of t h e i r g r w t h (e.g., flowering atage i n coffee). Tea i q o r t i n g , bleoding and packaging haa alao been characterized ae beiag concentrated i n a few transnational f i r m 1,and the oligopaoaiatic nature of the aarket haa also been attributed t o be a factor i n the relative e t a b i l i t y of tea pricee. However, developrrante i n recent yeara point t o the fact that collusion and price fixing i n the interartional tea market may n w be m c h weaker. Partly due t o ouch a tendency tea prices have aleo become more volatile. 50. The secular decline characterietic of tea prices i e related t o the tendency of the tea m r k e t t o be i n oversupply and t o the relative periehability of tea. L/ Hence, faater rtock turnovers a r e required in the case of tea and rtock a c c u d a t i o n is leer flexible than in other beverages. . 51 Carryover etocke i n any one year should, therefore, be an important d e t e r d n a n t of price. I n addition, the coffee price and f e r t i l i z e r price (a~aoniumsulphate) is poetulated t o have an important effect on tea price because of t h e i r importance a e proxies of input costs (para. 29) and, particularly, the e f f e c t on short-term yield reepoase (para. 32). Price Equation L . 52 The f o l l w i n g price equation has been e s t h a t e d for tea (time period from 1956-78): -- -1/ See Appendix I and G. Sarkar: Tea: Some Policy Issues, IDS Discussion Paper, February 1978, University of Sussex, Bri;{hton, p. 20. -21 While tea s t o r a b i l i t y is still a debated issue, it seem that tea cannot be k e p t f o r y o r e t h a n s i x zontyz v l t h o u t l o s s of q u a l i t y . (See Ap?cndlx 1 ) --- where: P T U average price of a l l tear in the Lo~donAuction. STKTU t o t a l implied rtockr 3f tea (in '000 metric tone). Q M U vorld tea conrumption (in 'OW metric tone). PPI -- Index of f e r t i l i z e r price (amonfun rulphate) 1970-100. PCOP price of coffee (Cuataulan, p r i m warhed, rpot, New York i n US centr pat lb). 53. Total -lied rtockr a t the t is an identity defined as follows: Where the variables have been defined i n paras. 48 and 52 except for QSTEA vhich stands for world tea production. The i n i t i a l stock level (1956) ha8 been taken from PAD 11 for five major net importing countrier (UK, US, Australia, IktherlanTs and Japan) and four major net exporting countries (India, S r i Lanka, Indonesia and Bangladesh). - 1/ FhO, Tea Statistics, 1961, CCP: Tah 67/WP.1. IVm WDBL SMJLATION AND PROJECTIONS A. Ex-Po8 t Sbmulation . 54 A r i u l a t i o n of the model for rupply, demand, price8 and rtockr was undertaken for the 1934-1978 period. kcaure of computer capacity probleou, only the a g s r e y t e rupply f o r u l a t i o n war used i n tne riarlationr. The r e r u l t r for key vrriabler are given i n Table 11 belw and i n Chartr 11 and III. 55 The s i u l a t i o n rerultr indicate t h ~ the w d e l captured the t performnce of the world tea economy quite well. In the fourteen year r i m l a t i o n period, the root-eao-rqrure error ( W Z ) for simulated production did not e.:ceed 2 percent of the actual for each of t ~ three main regions q (indurtrialized carntrier, developing c o u n t r i o and cea:rally p l a ~ e d econodee). The r i d a t i o n for coarurption was only rlightly 188s ratirfactory, largely becatme of the centrally planned econoDdes (with a M E of 3.5 percent). Implied rtocks s h w a much larger simulation +rror, 1 yet the movement of r i m l a t e d rtodrs is i n the correct direction and ta? pTice tracking perforvnce of the w d e l is not affected by it. 1. mdel krojectionr and Utetaative Scenarios Pro lection I - The Dame M e 1 56. The w d e l war used to project the endogenous variables (supply, dcpand and price) t o 1990. The erogenous variablee (population, COP, f e r t i l i z e r price and coffee price) were proiected on the basis of the assumptions used i n the World Bank's Yorld Development Report, 1980. The results (hereafter called Projection I) are given i n Table 12. -11 The largest errors i n etock carryover6 were obtained i n the volatile 1977 azd 1978 perlcds. Table 11: SUMMARY OF EX-POST SIMLATION RESULTS FGR SELECTED VARIABLES (1964-78) World World World World Statietice Production Conelraption Stocke Price ('000 metric tone)-- ( ~ / l b ) Haaa (actual data) 1,403 -9 1,345.2 389.5 62.6 Hean (eimulated data) 1,363.0 1,352.6 317.4 63.0 Induetrialired Cantrally Planned Developing Countria8 Bconoliee Countries . Prod. Cone. Prod. 2 . Prod. Cone. -('000 metric tone) )(can (actual data) 135.5 5 2 274.7 275.2 973.9 548.4 Hean (simulated data) 137.5 538.8 171.9 265.3 953.7 541.6 M E /. 2.2 6.3 2.8 9.8 20.1 6-8 M S B (in X) 1.6 1.2 1.0 3.5 2-0 1.2 /a Root-ocan-square s i d a t i o n error. 7 57. The odly data available to test the abort term projectione-yeare 1979 and 1980-18 for price and production ae shown below: -- World Price World Production - Projection I Actual Error - Projection I /a'! Actual X (a) (b) (c) - --- (thoueand tona)- - /a (b) - (a) x 100 - 0) - / b n.a. not available. Price -mj CPb ion Indurtrialized Conrrrpt RESULTS Developia~ /. MIDEL) Yorld (ME I regionr. CPEa PROJECTION component 12: for Table Indurtrialized rounding Production to due Developing up /s add not World may Totals /a- ART][I; EX-POSTESiMULATlON OF PRODUCTION. CONSUMPTION, PRICEAND STOCKS (1964-1978) The e r r o r f o r p r i c e s r e f l e c t e d a continuation of the tendency of the model :o "undertrack" p r i c e s s i n c e 1976, a period during which e r r a t i c p r i c e Psovemeatr became more pronounced. 58. As t h e above corparison suggerts, a given model, even i f it s t i m u l a t e s h i r t o r i c a l development well, need not n e c e r r r r i l y do a good job i n projection with respect to the degree of r e l i a b i l i t y of r e s u l t r . An evaluation has t o be made of t h e a b i l i t y of the ertimated equatione t o capture s t ~ c t u r a changes which a r e l i k e l y to dominate trends i n t h e future. l If s t r u c t u r a l changes a r e l i k e l y to be s u b s t a n t i a l l y d i f f e r e n t from t h e trend captured i n tho time period of a n a l y r i s , it ir better to undertake rome adjustments of relevant c o e f f l c l e n t r t o t a k e i n t o account such char,ge than to e x t r a p o l a t e from a set of e x l u t i n g equationr which t h e a n a l y s t k n o w vould n o t be realistic. . 59 Thtr hypothesis t r l l l u r t r a t e d f o r t h e w d e l with rerpect to the implication of t h e d e c l i n i n g a v a i l a b i l i t y of n t u land s u i t a b l e f o r l a r g e scale t e a expansfow-a phenomenon t h a t l a becoming more crlticrrl f o r v i r t u a l l y a l l t h e major t e a p r o ~ u c i n gc o u n t r i e r a s can be seen from the average annual ~ r w t hr a t e of planted a r e a given b u l w : -(Average annual g r w t h i n X)- Ind la S r i Lanka Indoaesia Kenya Tanzania Uganda Except f o r Indonesia where a c t i v e new planting programs were taking place i n the 1 9 7 0 ~t h~e r a t e of new plantinge haye been declining. The negative r a t e s of growth i n i972-78 f o r Tanzania and Uganda r e f l e c t the impact of protracted p o l i t i c a l and economlc d i s r u p t i o n s which a r e l i k e l y t o be reversed (i.e., growth w i l l be positive); but t h e i r 1351-72 growth perforeance is unlikely t o be repeated. Hence, t o expect the 1951-72 planted nrea r a t e s t o c o n t i n u e , ~ a s using t h e estimated model implies, vould be u n r e a l i s t i c . Similarly, basic! trends i n long-term y i e l d s could a l s o be a f f e c t e d depending on the e x t e n t of r e p l a n t i n g with high f i e l d i n g v a r i e t i c s . For example, i n S r i Lanka (and = possibly f o z Tanzania e n d Uganda), average y i e l d s a r e l i k e l y t o increase a m a f a s t e r ( o r a t l e a s t decrease at a slower) r a t e than i n the past couple of decades because of s p e c i f i c government programs focused on t e a r e h a b i l i t a t i o n . Hence a n assessment, based on a knowledge of t h e relevant country o r region, has t o be made on Likely changes i n t h e long term trend of pl, . ~ t e d(and thus harvested) a r e a and y i e l d t o achieve a more c r e d i b l e r e s u l t . ProJectlon 11 Adjustment for Long Term Structural Change i n Supply - 60. An attempt was made t o adjust the time trend coefficients f n r supply i n the proJection period (1979-90) t o take i n t o account the likely s t r u c t u r a l changes i n long-tern area and yield trends discussed above. 1/ The price coeffi:ients a r e retained t o a l l w t o r price effects. The p ~ o j c c t i o nr e s u l t s (ProJectlon 11) a r e In Table 13 t o r the major regions. 61. A comparison of selected s t a t l o t i c s with Projec:lon I an4 with actual (or.ly price, world production and world consumption) l a summarized belw: Uorld Production Uorld Consumption Price P r o j I Proj 11 ~ c t z lProj I Proj 11 Actual P r o j I P w j I1 Actual -----.-- -- (thousand tons! -(c/ 1b)--- 62. The overall reduction i n the time trenda f o r the supply equations i n Projectlon I1 resulted i n its short term price projections (1979 and 1980) c l o s e r t o a c t u a l than Proje=tlon I (the base case). Ptojectlon 11 price projections f o r 1985 and 1990 were respectively 16 percent and 11 percent higher than f o r Projection I. The differences i n production and c o a s u ~ p t i o n between Projection I and Projection 11were i n l i n e with the differences i n prices. Proportionately higher d i f f e r e n t i a l s were, hwever, obtained for the s h o r t t e r n projections than for the longer tern. Projection 111 Exogenous Fixed Supply Projections - 63. Numerous projections f o r tea&/ a r e based on simple mrlre; equilibrium models i n which supply and demand a r e i n i t i b l l y projected independently (implicitly assumlng hat prlces remain unchanged during the projection period). 'The price projection under'such an approach is usually obtained by calculating t h e price change between the base period and the projected period (through the use of price e l a s t i c i t i e s of sapply acd demand) t o e q u i l i b r a t e the imbalance betveen the supply and demnd projected - -11 tf a r a s possible, the adjustments wzre made with t i e price effectn netted out. The im?lled area and yield component una.~rlying the adjustments on the t h e trend coefficients f o r supply .\re given i n Appendix 11. - 21 See, f o r e x a ~ p l e ,the projections t o 1985 undertaken by FA0 and the World Bank's report on Prlce Prospects f o r Hajor Primary Commodities (Report No. 9 l 1 1 ! S ? , Jz-uzry 1?QC), .fi.??t~dlxT 2nd ~ 3 55-77. . Prlce Qk loa robucrhllud coraupr ~velopl~ Uorld 9E.a red rlali Lon Inhrrt Product brvelop1r.g Yorid exogenously. These simple models provide period by period projections, but do not allow for the influence of developamts i n the intervening years. To compare the effect of incorporating supply projections exogenously, the impact of two alternatives 11 a r e attempted: (i) Projecrion IIIa - "Host likely" crupply projections (based on industry judgement of likely ?upply by countriar o r subregions) t o 1990; and (ii) Projection I I I b - o f f i c i a l country supply projection (based on o f f i c i a l g o v e r m n t estinates reported t o the FA0 Intergovernmental Croup on Tea and on o f f i c i a l development plans). The resulte are given i n Table 14. 64. A compariron with Projection I for price, world production and world conrumption ir suamarized below: World Production World Consumption Prica Project ion: Project ion: ProJact ion: - I l11a 111b I i11a I I I ~ I IIIa I I I b . 65 Some intsresting differences m y k gleaned from the comparison of methodologits s w r i z e d above. For price projections, there appears t o be greater differences between Projection I and Projection 111 than within Projection 111 (lee., betweon Projection IIIa and IIIb). By making tea production erogenous (i.e., not p e r d t t i n g any shczt term or long term price response), projected prices are m c h lower than the base case results. This is primarily due t o the fact that (i)the erogenous production i n Projection I11 were proportionately higher than in che base case (Projectiou I) and (if) consumption results were relatively close (despite differences i n methodology). I n comparison with the results from the "gap analysis" approach, 2/ projected prices i n the l a t t e r were consistently higher primatil y because its consumption (projected exogenously) were consistently higher than that from the model projections. Plaintaining price response f o r consumption i n the model results i n less variance i n the consumption = projections. '9 - 11 Sources: PA0 Intergovernmental Croup on Tea and World Bank, Price Prospects for Hajor P r h r y C o m d i t i e s , Report No. 814/80. -21 Prices projected by t h i s approach (see para. 63) is given i n World Bank Report No. 814180, p. 77 (1985 = 139911b and 1990 = 185911b). The PxoTenocIs prod~rctionf o r 1980, 1985 and 1990 a r e t h e same as Projection IIIa. Price 6Es on Indurtrialized Consrrprf Developing ) tons World CPEs (thousand ialized Industr Production Developing Alternative Alternative Plan World ----------------------- bet-likely Government - - IIIa IiIb Projection I V - Sensitivity to Coffee and Fertiiizer Price Increaees 66. The sensitivity of the base model to changes i n coffee and fertilizer prices w r e also examined. The results for a 10 percent increase in coffee price i s given under IVa of Table 15 while that for a similar increase in fertilizer price i s under IVb. 67. Comparison with the base model is stlnaarized belw for price, w r l d production and wrld consqtion: World Production Uorld Consumptiorl Price Pro.jaction: Projection: Projection: I IVa IVb I iva IVb I X V ~ IVb 68. In comparison with the base case (I), an increase in the price of the input cost proxies (coffee price and fertilizer price) reeulted in higher prices for tea in the case of the former and lover prices in the case of the latter. The cost push effect in reducirrg production (and thus tb- 3roduction- coneumption gap) through lagged prices, appears to be more importaut for coffee price than for fertilizar price. Tea prices, a t lcast in the earlier years of the projection, also appear to be more sensitive to charrges in the coffee price than the fertilizer price. A/ -1/ For 1979-1990, 10 percent higher price for coffee resulted, on the a;.erage, ::Ia nlxe percerit ?-iigl:s~tea price, ~ ; l - i i l~~i i a tfor F c r t i l i z e r resulted in a 4 percent lower tea price. Tabla 15: PIOnCTION 1V- BASEt D D U UITII TKl4 INCRUSE It4 corm OR rKXTlLUKE PllaS hoductloo baaupcloo World Ioduatrlalixad D8vaLopla8 0 Uorld IoQlatrla1lr.d MvaLoplq ~ r l c a V. CONCLUSIONS 69. The conclusions for price projections (the primary objective of the study) a r e summarized below for 1985 and 1990. I n a l l the projection alternatives, tne level of prices projected t o r 1985 and 1990 would be ?owat than the actual 1979 level of 1 0 2 ~ 1 l bi f adjustaent f o r international inflation were undertaken. -- - 15'90 Price Projectionr ( ~ / l b ) Projection Alternative8 Currant Price Conatant Price /a - 1985 1996 1985 1990 I (Base Case) 123 146 78 69 I f (with structural adjurtl~ant i n supply trends) 147 164 93 78 1111 (exogenous supply-mst likely) I I I b (exogenous supply-government plan) 78 93 49 44 IVa (10% higher coffee price) 135 117 85 55 IVb (10% higher f e r t i l i z e r price) 159 143 100 68 - /a Current price deflated by the World Bank's Index of International Inflation with 1979-100. 70. On the basis of market knowledge and short term validation of the projections, the most likely price projpction f o r tea i n 1985 and 1990 (93O/lb and 78$/1b respectively) is the alteruative which adjusts f o r structural changes i n the trend i n area and yields f o r key countries. -11 71. Oifferences in-projection methodology results i n markedly different ?= resulta. I f production re exogenously determined (i.e., no provisions were ~ e d ef o r price response i n l i n e with the levels regarded most likely by perket analysts, 21 th? real price of tea could be substantially lower than - the most likely scenari&generated by the model. Allowance for price response 11 The above two values in parenthesis are constant (1979) prices. d -21 See, Price Prospects f o r Major Primary Commodities, World Bank Report c:o/sc ( p . 7 7 ) . and its e f f e c t s on production (and coneumption) i n the intervening years i n the model is important. 72. Tea price is more eenaitive t o coffee price changes than t o f e r t i l i z e r price changee. The importance of coffee prices seems, howeve:, t o be more apparent through its input effecte (used a s a proxy f o r production costa) than through its output e f f e c t (via its influence a s a competing beverage i n consunption). Characterietics of Tea and I ~ l i c a t i o n sf o r Harketing 1. The wide variety of growing and processing conditions has given r i e e t o a vide variety of teas produced. I n addition t o the broad categorization of teae i n t o "black", "green" and the intermediate "Oolong" teas 11 differencee among coneumere have meant that d i f f e r e n t grades of leaf and leaf p a r t i c l e e (brokene) have t o be blended i n d i f f e r e n t combinations. . 2 The quality of teae ( l i k e that of wine) is d i f f i c u l t t o define. Hade (proceesed) t e a may be prized not just by it8 appearance (size, color, brightneee); its method of manufacture (CTC 21 vereue orthodox), but by characterietice which can only be evaluated through infusing r?d tasting. The "liquor" and flavor of tea (there a r e more than 100 worde of description) is determined by the a l t i t u d e and e o i l of the tea area, the type of bush (clonal origin), the t h e acd fineneee of plucking and the s k i l l and method of manufacture. The technical explanation of the differences is manifested i n the amount and combination of aflavinee, and thearubigens (for aroma liquoring quality), and caffeine (for brisknese). The resultant variety of c h a r a c t e r i e t i c s make f o r an almoet i n f i n i t e number of "qualities" and confers on t e a its heterogeneity which cannot be standardized by the usual method of deecription. The t e a trade, hwever, groupe the wide range of teas into three broad categorles: (a) high "quality", (b) "medium", and (c) plain "filler". -31 The combination of theee cater f o r d i f f e r e n t consumer tastes. . 3 Tea is a l s o relatively perishable. While there is still some debate over the length of period tea may be stored without "eignificant" 108s of quality, it is generally agreed that t e a could l o s e i t e flavor a f t e r s i x -11 "Green" t e a is unfermented; "black" t e a is f u l l y fermented; and "Oolong" t e a is p a r t i a l l y fermented. 1 b - 21 Tea produced by the prccess of crushing, tearing and curling. In t h e t e a trade, quality descriptions a r e often linked t o t e a subregions ( o r even s p e c i f i c estates) of a particular country. Currently the ranking of t e a s may be s-rized a s f o l l w s : (a) high "quality1'--Darjeeling and Assam second f l u s h (Hay-June) from India, S r i Ladca high grown and Kenya high grown; (b) "nediuma" (identified by its b r i e t n e s s and useful liquor-particularly f o r tea bags) Assam and Doers (India), S r i Lanka and Kenya medium grown, Halawi teas, "peak season" Indonesian and some t e a s from the People's Republic of China (PRC); and (c) p l a i n s (.'fillersn) (identified by its plain color and liquor--particularly useful a s f i l l e r t e a i n blending) lw quality Aesame and Doers, some Malawi and most Indonesian, Uauritius, some PRC, Bangladesh, Argentina and Mozambique. Even lover quality ('3unk1') t e a a r e made--but t h e i r marketability is icLl:cr 1:dteG (Less Turkey, Fassia, I r a n a n d Tafizar~ianoff-grades a r e of t h i s category). Another category which is important f o r the Middle East market is t h e lowland "tippy" dark t e a s from S r i Lanka and India. months depending on storage conditions (tropical climates o r moist conditions vould mean an even shorter period) and the vapor proofing property of the container i n vhich the t e a is kept.L/ 4. The system f o r physically marketing the product is, therefore, geared towards handling these characteristics. Hade tea is conventionally marketed i n tea chests (usually plywood v i t h f o i l liner). Perishability and the need f o r a large variety of consignments has a l s o meant that traditionally the blending of tea takes place a t the consumption end of the marketing chain. Shipment is usually i n 8-11 conslgnaents t o give u n i f o r d t y and ease of sampling and provldes the readily available teas for blending under the auction system. Methods of Tea b r k e t i n a 5. Tea may be marketed t i t h e r through the auction s y s t e m o r through d i r e c t sales. Of the approximately one d l l i o n metric tons of tea produced by developing countries i n 1975-77, about 652 was sold through auction systems (562 i n producer countries and 92 i n London); the rest (352) was sold d i r e c t l y t o purchasers vho may be traders andlor blenders (see Table 1). :'ti tanka and Bangladesh a r e good examples of producers v i t h more than 902 of t h e i r production sold i n local auctions. On the other hand, Argentina and Indonesia sells the bulk of its crop directly t o purchasers. Virtually a l l of the production sold i n producer auctioar is exported. The main exception is India vhere 152 of the tea production sold i n l o c a l auctions (532 of the t o t a l ) is sold d i r e c t l y o r similarly disposed. 6 Direct s a l e s t o local markets i n producing countries usually consist of poorer qualj t y teas, the price of vhich neither justif its ;heir trans- portation t o auctions nor the payment by the producers of the .-ecessary brokerage fees. On the other h ~ r d ,private sale. t o foreign buy:rs (especially those sold on forward contracts) a r e usually of bettc r quality t e a s since the s e l l e r has t o maintain consistency i n its s a l e s o r r i s k the l o s s of a regular c u s t o m e r . 3 - 11 A good summary of the evidence for storability of manufactured tea may be found i n UNCTMIFAO, The Technical Feasibility of Operating an Inter- national Bufferstock for Teas TD/B/IPC/TEA 5, December 21, 1977 and Tee: - International Stocking Arrangements with Supplementary Measures, TD/B/IPC/TEA/AC/Z, July 31, 1978, pp. 27-30. -21 A good example is the tea exported by Malawi. Table 1: PRODUCTION AND DISPOSITION OF TEA BY TYPES OF SALES, AV. 1975-77 ('000 metric tons) Auction Sales Product ion London Producer Total Direct Salee Country India S r i Laaka Indonee la Bangladesh Kenya Malawi Uganda Mozambique Tanzania Argentina Others TOTAL DEVELOPING COUNTRIES Source: UNCTAD, Uapubliahed Study on Marketing and Distribution of Tea, 1978. Xarket Structure and Xarket Constraints - 7. The e e l l e r a of t e a i n the pajor import markets face three major problems with respect t o t h e i r a b i l i t y t o bargain f o r equitable prices: (a) t h e p ~ i s h a b i l i t yof tea; (3) the oligopsoaistic struc,ture i n its trading, blending distribution; and - - (c) t h e large proportion of s a l e s i n the a u c a o n system. 8. 'The roleof dtinationalcorporations in the coats1of exports from major t& exporting countries is not a s high a s the c o n t r o l b f purchasing and blending i n the major importing countries. I n S r i L a h andW1ndia, f o r example, the s i x largest uultinational companies controlled only 46%and 32% of their respective exports (1975-77). -1/ 9. Since major importing country blenders usually purchase the .bulk of tea imports themselves (or through brokers whoa they w n or control);2/ the concentration of blending by large f i r m is the key determinant of marret control. According to a study conniseioned by the Cormaonwealth Secretariat, A/ blending i n thc UK i s dominated by four companies 41 - and i n Australia and the US by eight corpanies each. 10. But t o what extent are there oligopsony/oligopoly forces a t work manipulating the tea in major mnrkets? A t the auctions (especially in London), the situation of a large nuo'er of sellers facing feu buyers, by i t s e l f , is not suf f icient to mke a case for market manipulation. Sellerr can, however, take advantage of a situation of excess supply and offer lover prices. In such a situation, the teas which are not auctioned (or are deliberately withdrawn) face even l o w r pricer i f sold outside the auction rince the opportmity for holdiog out for higher pricer ir limited by perishability and carrying costs. 11. A useful measure of exploitation ir the extent net return8 to the various levels of the tea industr- are out of line with similar industries. The conclusions from a report by the UK Price Comlsslon's inquiry into tea prices and nmrket mnrgins a t the blending, wholesale and r e t a i l levels are worth summrizing. 5/ The report (published in 1978) found that blenders/wholesaler maintained their historic gross nmrketing margin8 during the period of sharp world price increase i n 1976/77. 61 The resulting increase i n blending costs e r e , essentialy, passed on to the retailer and net margins were substantially increased above the historical rate. ~ e d p i t ethe high concentration a t the wholesale and blending level, no definitive evidence on price collusion was found. Hwever, it concluded that, "while characterized a s a gentlemanly trade, tea blending could afford to be more competitive." The impression of significant oligopoly profits from the sale of tea ty blenders appears t o be further supported by the consistently higher profitability of tea blending coppared with that of soluble coffee 31 Unpublished UNCTAD Study on Warketing and Distribution of Tea, 1978". .r 1 21 Commonwealth Secretariat, Studies Relating to the Development of Exports f from Bangladesh, Vol. 9. Tea - Product Summary Report, Economic Int er- ligence Unit, 1973. * -31 Co~onwealthSecretariat, ibid, p. 7. -4/ Brooke Bond 0x0, Lyons Tetley, Typhoo Tea and CWS. -5/ UK Price C o d s s i o n , Tea Prices, Report No. 32, 1978. -6/ Ibid, paras. 4.15 aad 5.1. manufacturing.(based o n t h e UK Price Commission's r e p o r t s on t e a and coffee). 12. The evidence f o r oligopoly "rents" a t t h e blending/uholesale l e v e l i n 1974-76, however, does not necessarily mean t h a t t h e problem i n t h e UK ir a s s e r i o u s today. Since t h e publication of t h e P r i c e Cotmiss?on Report, t e a blenders and d i s t r i b u t o r s have been under extreme pressure by t h e UK government t o "ratiotmlize" t h e i r operations and p r i c i n s p o l i c i e s . M s o , a number of them ( p a r t i c u l a r l y t r a n s n a t i o n a l s l i k e Brooke Bond and Lyons Tetley) have d i v e r s i f i e d t h e i r product l i n e s u b s t a n t i a l l y over t h e years. Their dependence o n tea a s a n important source of corporate income has diminished; so has t h e i r financial f l e x i b i l i t y for moving i n and out of the market i n a b i g way because of t i g h t e r corporate c o n t r o l s on, and t h e higher opportunity cost of, holding liquid cash. Hence, t h e case f o r c o l l u r i o n and p r i c e f i x i n g i n t h e market is n w orch weaker. 13. The s i t u a t i o n a t the r e t a i l l e v e l is, i n any case, d i f f e r e n t than f o r blending i n t h a t a number of v a r i e t i e s of tea are o f t e n i n t e n t i o n a l l y priced at lou levels as a "106s-leader" i n the s t o r e s t o draw customers. Given 9uch p r a c t i c e s , t h e a s s e s s e n t of r e t a i l margins is not p a r t i c u l a r l y relevant. 14. Furthermore, the evidence f o r oligopoly r e n t s should be separated from t h e competitive n a t u r e of t h e price-oaking mechanism. Despite t h e market s t r u c t u r e discussed, t h e process of p r i c e a r b i t r a g e f o r t e a s e e m t o work very v e l l a c r o s s major a u c t i o n markets. (See the c o r r e l a t i o n matrices 'fable 2.) 2/ Hence, p a r t l y i n recognition of t h i s phenomenon, t h e p r i c i n g of d i r e c t sales is u s u a l l y geared tovards the p r i c e s a t London auctions. -1 C. Uarnigatunga, Desk Study on Tea: Blending Packaging and Production of I n s t a n t Tea, G m m e a l t h S e c r e t a r i a t , London, Nov. 1979 (Draft) compared "net p r o f i t *ginsm and "return on c a p i t a l employed" from t h e Tea P r i c e s r e p o r t with a 1977 P r i c e Commission Report on Coffee P r i c e s and came o u t with t h i s r e s d t (p. 38). * -2/ Given t h e d i f f ~ r e n c e si n market l e v e l s (and :he need t o a d j u s t f o r t r a n s p o r t c o s t changes), c o r r e l a t i o n s among t h e p r i c e s i n t h e f i v e major . markets a r e s u r p r i s i n g l y very high. As expected, c o r r e l a t i o n s a t the annual l e v e l tend t o be higher than a t t h e monthly level. While p r i c e s a & Cochin and C a l c u t t a c o r r e l a t e very w e l l annually, t h e performance on a monthly b a s i s is r a t h e r poor u n t i l t h e d a t a is deseasonalized ( r e f l e c t i n g t h e marked v a r i a t i o n i n s e a s o n a l i t y and q u a l i t y of North Indian t e a s sold i n Calcutta). Table 2: CORRELATIONS BETWEEN TIE AVERAGE PRICES OF TEA, AT VARIOUS AUCTION CENTERS (USc/ke) London Hombasa Colombo Cochfn Calcutta Lnnual Date 1964-75 (12 observations) London Hombaea Colombo /a Cochfn Calcutta /a 7 b n t h l y Data 1972-76 (59 observrtfonr) London Hombdra Colombo /a Cochin Calcutta Jonthly Data Dt-Seaeonallzed 1972-76 (59 observations) London bmbaea Colombo - /a Cochin Calcutta Auction prices adjusted f o r export duties and cesses. ource: PA0 Report CCP: TE 77/4, January 1977. L araa &be plantad &ad treda. lor variable. 1979-00 for productlor Idspaadeat corlflclenta ad aru ot&r rred trerla cl. pkrd 1957-1990 from aru a8 country. b- 6arld eatlmcd earto4 19M. -tow, la w FTUDt-7) hi trc? Iw, (In plancllya tread bod area ktuoan 1CI.D ~wu Y prlce tin warm of MU, term planted 1978-90. )uct&. APPENDTX~ typo. In 1970-84 loma for aod coulrt~acy af for TRENDS ulth lrhatry plan. 1951-18 rav1.u 1984-30. TltU from a rate. for ad and erpaarloo adjuatwntr cwfflclant rater ylqld DO 4 0.7 0.3 .01 9.3 6.5 9.0 4.3 5.6 obtalnd 1910-84 -2.1 ulth tha la plum trend 8 judgawnt for of aovarnwnt rat. tlu Tabla of 1984-90 9. averale averaye indumtry area Grum from llty Planned on ly &la Afrlca Anrlca Table lrnb Dsrlvad Uelghied (1972-78); feaalbl Uelglrtcd Derlved baed Yrom H.rveacrd Kconorlaa Indla Srl Indonaaia Ot1,ar Kenya Tanunla Uganda 0th.; Ltln Central /a- - - - - - /b /C /d /a If Lg STRUCTURE AND PROSPECTS OP TER WORLD PATS AND OILS ECONOHIt Peter PoZtczk I. OVERVIEW OF THE WORLD PATS AND OILS ECONOm ...... ...... ................................... .....,.... A. Structure of the Vorld Oilreed Economy VI- 1 8. Recent Trendr VI- 3 11. ST~UCXUREOF THE mDEL FOR TtlE PATS AND CILS ECONOMY . . ................................ ................................ A. S~lpplyEquationr .. B. Demnd Quatioar C. Prica Equatioar for Fatr and Xi&-Protein 13aalr I HODEL SUKUTIONS AND PROJECTIONS ....................... A. Vali&tion of the M e 1 ......................... B. Prorpactr for the World Oilseed Economy ......... I. OVERVIEW OF THE WRLD FATS AND OILS ECONOMY 11 - 1. Oilseeds and t h e i r products--fats, o i l s and high protein meals-- account f o r about 12 percent of M r l d trade. Although the bulk of these products come from industrialized countries, many developing countries r e l y heavily on export earnings from oilseeds (particularly coconut, o t l palm and groundnuts). Host of these countries ore located i n the tropical belt. 2. An important characteristic of the ~ r l fda t s and o i l s economy is the large number of oil-bearing raw materials. On the baeir of t h e i r source, f a t s and o i l s con be grouped i n t o vegetable o i l s o r animal f a t s (including marine o i l s ) . Vegetable o i l s can be further broken down i n t o o i l 8 from annual oilseed crops and those f roa perennial t r e e crops. Soybeans, cunf lover seeds, cottonseeds, groundnuts and rapeseeds a r e the major annual crops; coconuts, palm f r u i t and olives a r e the main t r e e crops. Cattle, hogs, f i s h and other marine products a r e the main sources of ar~imalf a t s . 11 Structure of the World Oilseed Economy Linkages Between O i l s and Heals . 3 Many f a t s and o i l s a r e byproducts from thc processing of oilseeds and animal products. Their supply, then, depends largely on the demand f o r the primary product. The a v a i l a b i l i t y of tallov and l a r d , for example, is d i r e c t l y related t o the slaughter of c a t t l e and hogs. Since most ojlseeds contain both o i l and high-protein w a l , the supplies of vegetable o i l s a r e frequently determined by the demand for meals. The joint recovery of two products-oil and meal--facing markat demands that a r e largely independent of one another is another characteristic feature of the world oilseed economy. Since the economic forces that determine demand f o r high-protein meals a r e d i f f e r e n t than those f o r f a t s and o i l s , the markets f o r both these two commodities have t o be considered. 31 - - 11 While the authors are responsible for any errors, they wish to thank - - Martin Paldas, i n particular, f o r h i s extensive comments and suggestions. -21 The f a t s and o i l s extracted from these raw materials a r e generally class- - 6 i f i e d a s "edible/soap £fits and oilu," reflecting t h e i r two major end-uses: - edible products and soa s o r detergents. These end-uses account for about ' 97 percent of the prod - t i o n of a l l major f a t s and o i l s . t i I n t h i s paper,* t h e term " f a t s and oils" refers t o t h i s edible/soap category. - 3 1 The close l i n k between the markets f o r f a t s and o i l s and high-protein meals i n j e c t s an additional element of i n s t a b i l i t y i n t o the supply of vegetable o i l s . A strong demand f o r high-protein meals would stimulate the processing of oilseeds, which i n turn wouid increase the supply of v ~ q - t a h lof!?, rp_~arr!lr'sc nf thr. TAT?^! ~ f a t 9 2nd o f l s . Linkages t o Livestock and Grains 4. Feedgrains a?d high-protein w a l s are the tvo main ingredients i n ar.iaal feeds. The exact composition of the feed depends on the nutrltional needs of the animals and the relative prices of the ingredients. Thur, the demand f o r high-protein meals is t i e d t o the production of livestock and dairy products, a s v e l l a s t o the market f o r feedgrains. The g r w i n g shortage of pasture has boosted the demand for animal feeds by producers raising livestock i n feedlots and by other large-scale livestock operations. Linkages Between Individual Pats and O i l s 5. In addition t o the linkage8 between the markets f o r fate, o i l s , high- protein meals, feedgrains and livestock products, there a r e a l s o close t i e s betveen the markets for individual f a t s and o i l s . Host f a t s and o i l s a r e interchangeable, and manufactures of amrgarine, shortening, soaps and other f a t products can substitute one o i l f o r another, within the specific requirements of t h e i r products. Similarly, the animal feed industry can substitute ona high-protein meal for another i n the manufacture of feeds. Considering the scope of t h i s interchangeability, t h e i r markets have t o be looked a t simu1taneously. Annual Versus Perennial Crops 6. &I noted a uniqut feature of the fats and o i l s amrket is that i t s supplies come both f r o a annual crops and perennial t r e e crops. The source of a product a f f e c t s h w supplies can be adjusted i n response t o prices and the s t a b i l i t y of supplies. Producers of annual oilseed crops can adjust t h e i r production plans quickly (usually v i t h i n a year). 11 Producers of tree crops lack t h i s option. Their investraent i n o i l palm, coconut o r other l i v e t r e e s is based on price expectations over a longer period a s compared t o that for oilseed crops. Because of the uncertainty, investment i n tree crops is usually more risky tl-an i n annual crops. 7. The supply of o i l s from tree crops is more s t a b l e than that of annual crops. Once the t r e e s have been planted, producers usually have no choice but t o harvest a s unharvested crops a r e a major source of the diseases vhich endanger the l i f e of the trees and thus future harvests. L , 8. The t r e e s of most perennial oilseed crops remain productive for more than 30 years; some such a s olive trees have a l i f e span of up t o 60 years. - 1/ Many annual oilseed crops are produced i n both the Northern and the Southern Hemispheres. Their production cycles a r e roughly opposite-- the harvest time i n t h e Soutnern Hemisphere corresponds to the planting time i n the Northern Hemisphere. Thus plantjng dectsions In one hemisphere a r e affected by t h e nutcc-c o f t h e har.~esti n :hc o t h e r . Although t h e y i e l d of tree crabs f o l l w s a y a t t e r n y which v a r i e s w i t h t h e c r o p and sometimes even with t h e v a r i e t y of t h e same c r o p , weather, c u l t i v a t i n g p r a c t i c e s and t h e age of t h e tree remain t h e most i n f l u e n t i a l f a c t o r s determining y i e l d s . The n a t u r e of t h e c r o p combined w i t h varying investment horizons g i v e s each o l l s e e d c r o p a s p e c i f i c production cycle. These c y c l e s , o f t e n accentuated by abnormal weather, account f o r a s i g n i f i c a n t p o r t i o n of t h e i n s t a b i l i t y of s u p p l i e s of f a t s and o i l s . B. Recent Trends 9. S i n c e t h e e a r l y 1960s. v o r l d production of f a t s and o i l s has g r w n a t a n average annual r a t c of s l i g h t l y more than 3 p e r c e n t (Table 1). During t h i s p e r i o d , s u p p l i e s of vegetable o i l s grew f a s t e r t h a n t h o s e of animal f a t s and marine o i l s . S e v e r a l f a c t o r s account f o r t h e i n c r e a s e i n t h e s u p p l i e s o f v e g e t a b l e o i l s . F i r s t , t h e g r w i n g demand f o r roybean meal l e d t o a s h a r p i n c r e a s e i n soybean o i l , which is recovered j o i n t l y w i t h soybean meal. Soybean o i l n w p l a y s a dominant r o l e i n t h e f a t s and o i l s market, w i t h market share of about 20 percent. 10. A second f a c t o r behind t h e i n c r e a s e i n v e g e t a b l e o i l s u p p l i e s is t h e sudden g r w t h o f p a l m o i l supplies. The d r o p i n n a t u r a l rubber p r i c e e i n t h e y e a r s f o l l w n g t h e Korean War t r i g g e r e d a massive s h i f t away from rubber t o o i l palm. Crowing c o n d i t i o n s were i d e a l i n Malaysia, Indonesia and some West A f r i c a n c o u n t r i e s and were r e f l e c t e d i n t h e lw c o s t s of production of palm o i l r e l a t i v e t o its p r i c e s , which were l a r g e l y determined by t h e market f o r a l l f a t s and o i l s . These favorable econcmic c o n d i t i o n s s u s t a i n e d t h e expansion of o i l palm production, o f t e n at t h e expense o f rubber, throughout t h e 1960s and e a r l y 1970s. The market s h a r e of palm oil-10.3 percent--is q u i c k l y approaching t h a t of soybean o i l . 11. While t h e market shares of soybean and palm o i l s expanded, tallow and l a r d l o s t some ground as l i v e s t o c k producers began t o respond t o t h e demend f o r l e a n e r meat (beef and pork). T h i s demand grew o u t of a concern about t h e a d v e r s e e f f e c t s on h e a l t h o f polysaturated f a t s . Consumer demand s h i f t e d from b u t t e r and animal f a t s t o vegetable o i l s . T h i s change i n t h e p a t t e r n of consumption o c c u r r e d a t about t h e same time t h e i n c r e a s e d s u p p l i e s o f v e g e t a b l e o i l s a r r i v e d on t h e market. These s u p p l i e s were t h e r e f o r e absorbed without a n a p p r e c i a b l e d e c l i n e i n t h e i r p r i c e s r e l a t i v e t o those o f o t h e r f a t s and o i l s . , Host tree rldtdde. Whfle sow parts of the model, euch ao the regional breakloa~of the deaand for fats and oils, need further refinaent, the siru tion results indicate that the structure of the model i e TP 6 OOUO~. -1/ Since data for some of the key variables -re not available for the years before 1961 and some variables were wed with 12-year lags, it uae not pos8ible t o s i u l a t e the model before 1970. Prospect8 f o r t h e World O i l s e t d E c o n o q 172. The model war used t o determine 1ikc.y developaentr i n t h e world o i l s t e d economy. Table8 25 and 26 r h w t h a projected values f o r production, conrumption and pricer. 173. The projectione confiread t h e vieu tht t h e d a d f o r high-protein walr will g r w farter than the d-nd f o r f a t 8 tnd o i l s . Thir, i n turc, w i l l lead t o g r w i n g i n v s a t o r i e r of f a t 8 and oil. and a d e c l i n e i n t h e i r o v e r a l l p r i c e l e v e l , which eventually (within a n appropriate lag) v i L l deprerr t h e rupply of o i l r e e d r with a hi* o i l coateat. The d e c l i n e of r u p p l i a r from t h e r e cropr, togather with t h e g r w i n g d e v a d f o r 011 f r o o developing c a r n t r i e r , w i l l than cause a decline i n t h e i a v e n t o r i e r of f a t o and o i l 6 and a r i r e i n t h e o v e r a l l p r i c e l e v e l f o r f a t o and 0118 touard the and of t h e period baing projected. ~ s ' b l e258 WORLD PRODUtXION OF PATS AND OILS 1971-75 (ACTUAL) AND 197690 (PROJECTED)BY MOB GROVPS - FIVE-YEAR AVBMBS 1971-75 1976-80 1981-85 1986-90 '000 Tona % Share X Share 8000 Tom '006 Tona % Share VEGETABLE OILS Prom: Oiloecdo Soybean ~unflovar 3 ;886 Cottonoeed 2,783 Groundnut 3,275 Rapereed 2,487 Prom: Tree Cropr 6,982 Olive 1# bt&1'1,532 Coconut Palm Palm Kern01 ANIMAL FATS AND MARINE OILS 14,564 Fish Butter Lard Tallow TOTAL FATS AND OILS .Department Source: U.S of Agriculture (actual) ; IBRD, Economic Aarlyaia and Projectioaa Depar u a n t (projected). VI-Y b ...................................... SLtO(ARY AND alICLtKIONS ................................ ......... A htemlruntr of C m r c i . 1 n . e r m a d ........................ I... Data AvrLlrbility C tcoaowtric Specifiutlon ......................................... ............... A The D8ta ................... I Tbe Sirultanoour Bl~hmtioakthod ............ C....The Two-step B8thmtioa netbod D A Brief Cog.rL.oa of tba ILo ktbodr i. During the 1960s, p r i m ? coxmercial ccrergy demand for a l l developing countries increased by about 72 annually, a s compared with slightly o&r 52 i n the industrialized couatries (exclusive of centrally planned economies). During the 19708, the world energy c r i s i s affected the growth of energy demand i n developing countries only marginally. Whereas i n industrialized countries the increase in energy demand decelerated to an annual rate of about 22 betveen 1970 and 1978, the rate of increase remained a t a high 6 t o 72 i n tha developing countrio. ii. If real enargy prices remain at their 1975 level, energy demwd i n tha damloping c o m t r i e s w i l l increme by about 62 annually betveen 1978 and 1985 and a t a s l i g h t l y higher r a t e of about 72 betvean 1985 and 1990, assuring that ecoaoric g r w t h and structural change f o l l w the pattern pro- jected i n tha World Bmk's World Damlopmaat Report 1979. A conetrnt r e a l anargy price is, haramr, quite unrealistic. I f it is assumed, therefore, t h a t real p r i c u Fa 1990 w i l l be about SOX higher than in 1975, energy demand i n the developing countries is projected t o grow a t only about 5% annually betveur 1978 and 1990, with the growth rate t o 1985 slightly lover than that batuean 1985 and 1990. ill. Energy d e n ~ di n the indretrialized countries is expected t o grow a t considerably lover rates. Therefore, by 1990, energy demand in developing c o m t r i e s should exceed that of Western and Northern Europe, and those corntries sbould become the second largest energy conaunrers behind North America and Oceania. iv. The w d e l on which these projections are based has been spc=ifically designed t o accomt f o r the i q a c t of struct*lral change, a s well as economic g r d and energy prices, on energy dem:..l and t o all- the use of macro data projected 3y the Bank's mdel as exogenous variables. The basic e s t h a t i o n equation is a l o g linear function of per capita energy demand dependent on real per capita income, the real energy price and structural variables, defined as the shares of agriculture, mining and conatruction, ranufacturirg and, e l e c t r i c i t y , and the tzansport sector i n Cross,Domestic Product. The e s t h a t i o p a have been nade f o r t o t a l primary commercial eneray expressed in thousand metric tons of coal equivalent. The data a r e time series, ranging from 1960 t o 1975, f o r 80 developing countries am6 16 industrialized count~ies. V. lbo alternative estimation =tho& were used. In the simultaneous estimation (SE) procedure, all parametera were ~ s t i m a t e d13 one atep. Structural and price elasticities weze based o q t h e data for each of 14 country groups; t h e income e l a s t i c i t i e s were based on the data of the respective couttry. In the two-step (TS) method, the parameters of the structural variables and the energy price were estimated i n the f i r s t step. The impact of the change i n econondc structure and energy prices was then deducted from the original energy consumption data. The reduced energy correqtion variable obtained by t h i s calculation war, regressed against real per capita Income for all cormtrier. In t h i s way country-bpedfic incol~se l m t i d t i e a were obtained that reflect variations in energy damand with respect t o chmges in income, a f t e r the -act of 8tructurd chmgr and mergy prices on energy demand har been eliminated. vf. The projection8 for the cormtry groups were obtained by adding u? the cotmtry-spedfic projections rod then inflating the aggregate by the rate of undercoverage ( w u l l y only 832 o r -re of the grows' energy damand w a s covered by the raqdu). di. Bued on the rerultr from the two e r t i m t i o n metho&, the f o l l w i n g coaclwioaa can be dram: - The r t a t i r t i c r l f i t hardly differr between the two metho&. If anything, the SE mthod producer rlightly better rerultr. Iiwewr, the parawterm of the rtructural v r r i a b l u are r t a t i r t i c a l l y rignificmt portly with the IS mthod, and r a r d y with the SE mthod. In the SE mthod, a courtlot term VU wed for every corntry; apparently it captured otrh of the rtructural differencer between corntries, vhile rtructural chmge vithin countrier ir too rmll for pro- ducing rignificmt paramterm. - The TS mthod ten& t o -hasize the illpact of structural change, v h e m u the SE method probably producea rtructural p a r a w t e n that are too lw, the remult of the collstmt terps. The true rtructural parameters U e somewhere ln betueen thore estimated by the two methods. - Specification eaperimentr v i t h the TS method show that the projecttons are rtnrarkably stable with respect t o alternative specifications (e.g. using country d d e e , bmitting variables, etc.). - Stabflit3 teats of the inc- elasticities indicate that specifiurtioar v i t h variable elabtic,itiea are not superior t o those 'with constant elasticities. - The aggregated estimates of a cormtry group f i t the observed -- values vell even i f the f i t for the individual countries waa not too good. Apparently, errors tend t o cancel each '9 . other out in the aggregate. viii. -- The projections for the individual countries were done rather s d e m a ~ c a l l yby applying the output of :he World Bank model and without taking into account the special circmnetancee that prevail i n almost every cotmtry. They therefore should be used for discuseion only. Comments by country specialists that would enable us t o improve future projections are velcoue. RIERCY DEMND IH ME DEVELOPIN L'ORU); rSTIHATIONS AND PROJEmIONS TO 1990 3Y REGION AND COUNTRY INTRO3UCTION 1. The futur-c energy ace& of the developing vorld have becoue a mtter of unJor concern of policymakerr i n developing comtrier r q v e l lM in indmtrialized oaer. Vherear e n e r n d e w d i n industrialized corntrier ir now beiug prcjected oa a regular basir by national a d international inrtLtutionu, i n particular the International Ecergy Agency, ouch pro- jtctionr are still lacking for the developing world. Thir paper report. on a w j o r f i r r t r t e p tauard bridging thirr gap. I/ 2. A t present, the profectiom a r e confined t o t o t a l primary ccmmercial enel-gy-the rrw of coal and lignite, petroleum, natural and maaufactured gar, and electricity, a11 converted into thourmd metric ton8 of coal equivalent. The demand for am-corpparcfal energy, although quite important i n many developing c o m t r i e r , ir not covered here, primarily because of a formidable lack o t data, but a l r o bec;.ue the concept of nm-comercia1 a r g y rtill h~ not been clarified. 'Ibe breakdown of c o m r c i a l energy demand by individual fuels w i l l hopefully he achieved a t a l a t e r date. 3. Uben t h i s project WM started, it w u roon realized that avail- able e r t i m a t i m and projection approaches were inadequate conceptually, a r w e l l M with respect t o available data. It was therefore necereary to develop a suitable projection methodology. I n the course of the wo*, additional p r o b l e ~v i t h data and ertiuatioua made reveral revieion8 t o the osthod necerrary. Sections 11 t o I V of t h i s paper dercribe the theory mdarlying the method m d the mcthod i t s e l f . 4 . Projections for the developing world a s a vhole are neither wry useful f o r the deweloping countriee themselves, nor could they reasonably b e asseseed by others. From the beginning, therefore, the authors decided t o disaggregate t&e projections by region. For technical reasone outlined b e l w , the regions coincide v i t h o r a r e sub-regioar of thoee wed by the Morld Bank f o r its ucrarconomic projection framework. 'Ibe projection8 were not, however, mdertaken for the regime directly, but were built up from projections f o r individual countries. This approach, dictated by ~cthodologicalconsiderations, had the double advantage of providing information for individual countriei, as vell ae leaving the regional aggregation completely flexible. Me d l f ficrilty that should be noted, however, i e that the projections f o r some countriee are much lees reliable than those f o r the regional aggrsgaFes. s - -11 Research f o r t h i s paper was supported by the World Bank. Walter Oberhofer of Regensburg University provided considerable help i n solving a number of i n t r i c a t e econometric problems. He also permitted the authors sub- s t a n t i a l l y more time by taking over a couple of the more time-consuming computer runs. 5. 'Ihia p a p t ir the f i t r t c q l e t e p n r m t a t i o u of the nrultr aehiemd wing thL. ruv rthodolow. k ouch, u ruth information u porrible h u beea proddad i n both n u u r i u l aad graphical form on the e ~ ~ g e n o uur v e i l u the u t b a t e d vui.bler. Ibe method of prueagatioa o f the rerulw i m outliord bt Sectiolr V. The rerultr by couatry group apprar in Sectioo ?I, which dm contlinr the nnrltr for individual c o ~ t r i uirr the f o m of r m x y tables. Iludrza primarily i n t r r u t e d in th projectloxu t h m e l v u can go Wdluelr to Section VX, coarultiag Sectioo V for axplrrwtiou~of the tabla u nuded. 6. Producers require energy t o produce c m d l t i e r and re-cer; private houaehol& require it for h e a t l q , cooking, lighting, traamportation, etc. The demand of p m d u u n &pen& on the oolrry of their yroduction, u well u on the ucrat to which energy L wed u m input, u coclpand t o other inputr ruth u labor, capital goods m d r a r mterialr. The n l a t i o ~ b i pbe- inputr m d output cm be e q r e r r e d by a production frrrrction which, in its rort w e r d fom, m y be w r i t t e a ua where : X real output of a rector EI energy input L ----labor input C c 8 p i t d input To obtain a prultacer'r energy requirarmt, function (1) m y be r o l n d f o r X I , i f poarible: 7. lhie gmcrd functional relationship docs not account for production techaologien, but it is u r r n r d that a (&am) met of technologies exirtr that can be traced capiricdly by t m t i n g alternativa rpedficationr of furction (2). Input pricar a l r o do not appear in (2) b e u u r e it l a m r u ~ r d I in equilibrirn that a certain ( d r o a priori mkaovn) ralatbnahip e h t r bent- Lnputr and input pricer and that changes in m u t o m l y changer in input prices and vice Persa. 8. Because of the latter asrumption, inputr and input price8 are interchangeable in a production function, and function (2) m y therefore a l r o be written a: - PLC = r e a l price of labor i n t e r n of c l p i t a l - PEC r e a l price of energy i n term of capital r e a l price of rau materials i n ternre of capital PRC = For NO inputs. any energy and e a p i t d , ftmction (3) can be i l l w t r r t e d by the follovirrg f a d l i a r graph: Figure 1 Capital 'Ih. figure doum thu certain output (3,repruarted by an iwqumt, aad cartrio real price GEC)axactly , daterriru tha dmmd for mergy m. 9. harm demand by private howeholdr follow the a m principle, A howehold c a u r r v r mrrgl, f i r r t , on the b u i . of gmerd ability to buy poda an -reared by real incme, ord, second, on ths b m i r of the price of energy as coqmred to other c o m e r goods. Hence, t o c-ute h o ~ c h o l d energy drund, it Fs d y neccsaary t o s&stitute in fmction (3) red incora for output and tbe real pAce of energy in t a m of other cauuaer goods f o r the real prices of inputa: .; EB energy coaa-aption by ho-mehold. - Y --- real incorz ?P= r-1 m e r g price in term of consumer good.. Figure 1 applies to function (4) also. Inwme is expressed by the budget l i n e which is tangent t o an indifference curve; the gradient of the l i n e represents the real energy price. B. Data Avallabilitr 10. Given the availability of data for developFag countrier, it ir generally not possible t o trawlate fmctionr (2) t o (4) into ertimatioa equations directly. Data problem also exirt vith regard to: - the quantities and prices of energy conrumed by the varlour categorier of ladvatrial and non-industrial users (industrial, households, etc.), - the quurtitier md pricer of noa-energy factor iaputr; and - the m a l wtputr of variova indurtrial activltier. 11. Even i f moor of the data could be obtained for a particular country, they usually are mavailable for a tioa period rufticiently loag for regresrion e r t i m t a r . k loag ar there dara probltam eldot, it ir aacesrary to coqrooLse batveen h a t i.ideally required for a r t h a t i n g specified functions of the theoretical nlatioashlps and what c m be estimated vith the data a t hand. 12. The lack of data on energy coarunptim by different users Puke8 it necersary to luq energy c o n a ~ t i o nfor productive purposer, as raprcscnted, for inrtance, by fmctioa ( 3 , vith that of houeeholQ [fraction (4) I. 13. The estimation of a epecificatioa for (3) requires data on output in real te-. Uafortvaately, for u n y developing comtries, only total output, defined by Czorr U r t i c Product (Y), and not aectoral outputr (X), are available in real t a m . X ir, therefore, defined by: (5) X = yX . Y - u . Y 14. I f it is assmed that inflation i n total nominal output does not differ very much frm that in sectoral nondnal outputs, the eectoral rharu, (u) calculated from aomlnal v a l w should be reasonable approxilaationa of the re~pectiveshares in coastlmt prices, because inflation would cancel out la the ratio. 15. t_ further sfntplification has t o be made with respect to the price variables. It is assrrnred that the two energy prices in functions (3) and (4) move parallel over time, thus permltt$ng use of just one price index - (P . The prices for other inputs are mavailable and, therefore, have t o q e omitted. s 16. Given n sectors In the econorny,ahe function for total energy demand (E) can n w be written as: 17. Traditionally, energy demand han been estimated wing eirupler functions, with the energy variable regresoad only against incoma o r income and prices. I f income were the only explanatory variable, the utirnated parameter of the income variable ought t o be the qame an i n a coopparable specification of (6), a r long a s tha acoaomic structure aad the real energy price remain conetant. I f , on the other hand, real entrap price8 f a l l , u they did i n the 19608, auargy consuqtion can ba expected t o riee f u t e r thsn otheruire, and the parameter of Y i n the riuqler function w i l l therefore be b l u e d rrpvarde. If there 1s a shift t w a r d highly energg-intensive iaduetriea, ae hae been tha c u e i n moot developing couatrier ,energy caarumption w i l l rime f ~ t a than r otharuire, and the parameter of Y w i l l again ba b i u e d w a r d s . 18. For projection purpose., r I q l a r furctioar a n acceptable u long u it can reuoarbly be ur-d that rtructural churgu v i l l f o l l w tha 8.p. trends u i n the p u t and that the r e d enemy price is coartmt o r changar i n keeping v i t h p u t treads. Given t h u a US-~IODI, the e r t i m t e d p a r m t e r e would not produce i n c o m c t projectione, though they vould be blued. 19. Both asr\tlptionm are, hwever, questionable, particularly in the longer rm. Tha price of energy h u been aoving rp a f t e r a decline i n the 1950s and 196011, and rtnwtural change h u been tapering off ae countries reach certain letnlr of development. 20. In the prtment exarctre, the mtructural variableo i n function (6) were &fined u the sham of m j o r mectorm much u agriculture, manufacturing, etc., in national incooc, raising the quertioa of the level of aggregation a t which the rtructural variables ought to be daf ined. Ideally, they rhoutd be dimaggregated a s nu& u possible i n order to capture dl rtructural effecte. In practice, the availability of data necemeitater a compromise. 21. Data are available for m j o r rectors for h s c all countries from National Accomts statistics. Any further disaggregation is difficult if the estimates are t o be performed over a longer time period and for a larger n d e r of countrice. Stmut (1976). for imtcmce, has propoeed dinaggregating manufacturing into energy-intensive and non-energy-intensive industries. This certainly is desirableDcbut,as his calculations show, it is only feasible for a limited nuder of couutries and only for a feu, =re recent years. - - C. Econoetric Specification ~! 22. The econonetric specification is primarily a matter of conveaienceE For the estimation functions for production and demand, it has become common* to use logarithmic specifications because they can account for non-linearitie% and therefore tend to produce better s t a t i s t i c a l results t h m do unlogged forms. I f the specification is log-linear, the functions estiamte elaeticities directly, facilitating interpretation. 23. The &vantages of logarithmic specificatioas, however, have to be weighed ayi-&sttheir shortcoPings. h e problem .that for the larger values, the de~iationof obscrved from estimated values tends to be ucderltated, and the coefficieat of determiaation therefore is overstated. Ia tir- series analysis, which usually iavolves increaslag fmctions, the deviations for the most recent years are generally unders:ated. This retult is undesirable ao it la believed that the Pore recent developmentn are a t least ss im~ortaat,i f aot more important, in the estimations than are the earlier values, for iartance, vhea maklag projections. While :heoretically the problem could be solved by weighiag the observations, i a practice it is difficult to detrrPLae the correct mgnitude of the weights. 24. Anothar problem a r i r u with respect to rggregation, if it is prtulated that tha demand fmctioa for total eaergy can be disaggregated jato the fmctioar of the user categories In such a way that the p a r m t e r s of the latter f m c t i o n ~cur be di~tfaguirhedfrom the former. Such a po~tulatererioualy conrtraiar the range of permissfile specifications. Ia particular, it exelder all logarithmic ~pecificatioar,because such functioar generally cannot be &aggregated. 25. Ia .pita of thcse rhorccorLnp, log-linear specificatioa wan chosen i a this rtudy, prlrufly Leuuse the structural variables do c h g e i a a aou-llaear faohion. F r a the work of Kuzaecs (1971), Felr/Schatz/Uolter (1971) urd Churergy/Syrqufa (1975), it is known that the sectoral s b r e s fn Cross Doacstic Product (CDP) approach certain ceilfnga with rising per capita fncoae aad at those points follw a reverne trcnd. 26. The ftmctioa actually used l a the eatimatiorm is: popula,tion share of agriculture i n GDP share of rdning in QP share of conatr*rctioa in GDP share of nanufacturing in GDP share of utilities in GDP share of transport Fn CDP error tern. 27. In ftmctim (7), the rharu of -lag .ad wnrtruction, u uell u thore of mmufacturing .ad utilitier, a n lrrqed together nrpectively, u they cmoot be obtained repantaly for 8evera.l couatrfeo. To avoid that the cDP'~ of .PI11 md large cotmtriu -ate the reoulto, uurgy coaamrpticm md ='a of all corntrim were divided by their populatia~ (B) 111. DATA ASD ESTIMATION STHODS 28. Functio? (7) was e s t i ~ t e dusing pooled time series-cross-rection samples, in uhich the ti- series for individual c o m t r i e s , uhich ranged from 1960 t o 1975, were l u ~ p e dtogether. There vere tuo possible different estimation methods. The f i r s t , which appears t o be -st appropriste for pure projections, is the sinultaneous estimation method. The second, vhich is probably w r e suitable for rimulationr, has two steps. 29. The application of aquation (7) to a country pool iartead of individual comcrier v u necessary for r e w r a l rauonr. P i n t , the wxiwu n d e r of years available for any country vas 16, too feu for the estimation of s t a t i s t i c a l l y significant parameters, given a multiple regrerrion v i t h mix independent variabler. Second, the rtructural changu vithin corntrier over a period of 16 year8 a n generally too 8-11 and the apraad of the rerpective variables therefore too narrou, so that, .gain, r t a t i r t i c r l l y s i ~ l f i c m parametera cannot be expected. t Third, st-uctural chmge is a long-nm phenomcnoa, beat explained by croso-rectron eat{u t e r that can indicate long-r\ra adjustments mxh b e t t e r than can time series. A. The Data 30. The estimation of a ftmctioa such as (7) require0 a rubstuntid amount of data vhich, f o r the =st p a r t , are not readlly available for d-wloping countries. The data used here for the exogenous variables other than prices vere obtalned fron the data banks of the World Bank. The data for energy c a a s ~ t i c ma r e from a tape of the United Nations. L/ 31. Data on prices veta obtained fn part from B.J. m o t of the Eccmoaic Analysis and P r o j e c t i o ~Department of the World Bank and i n part vere collected by the authors. The price s e r i e s f o r individual fuels vere used t o construct energy price indices, taking a s veighte the fuel c o n s ~ t i o nshares of the respective countries i n 1970. That year was also chosen as the base year. 32. lls income +aria; e s are expressed in real terms, the nomlnal' price indices also had t o oe deflated. For t h i s deflation, e i t h e r the consumer price index, r e t a i l price index, wholesale price index, cr CDP - - deflator vere used, depending on d ~ t aavailability. 33. The price s e r i e s were converted into indiccs'%ecause it appeared irmpossible t o derive an average energy price for or-. c w t r y that vould be '*P arableestiamtedto that of another. The consequence of the t h t the price e l a s t i c i t i e s reflect on-y the mpact of price 4rocedure ' adopted is changes over tire and not that of intercountr:. differences. -1/ Published as United Nations, S t a t i s t i ~ a lPapers, Seriee J, World Energy Supplies, 1977 cmc. 34. Ibe r i d r a ~ c o u nestimation (SE) and the ruo-atop (1s) mthoda d i f f e r b u i c a l l y in their treatburt of the reaction of deerad to & q o r in income. h w u argued &ow, the impact of rtructural change ir a long-rm pheno~moothat c m only be ertiarted uith very long ti- rrriee, whi& are not available, o r with a t h mcrie8-crorr-mection ram?Le. However, c5rt problem &anot apply to estimating the I-act cf changss ln incoma, as there c t m g a are certainly much -re pronormcad over ti- than a m rtrucfurrl chmgu. rurtherwm, a t e r g damand c m bn expected to nrpoad to chaager in i n c a r m r e i m d i a t e l y tbua t o & w e 8 Fa the ecoaory'r rtmcture md the price of energy. Imce, tin rerier u t i m c e r oS the m~my-ineoar relationship for individual cowtrier cm 1 ~ 1 to produce rtatirtically ; ratlalactory ruultr. In addition, ertinrting tba eae-*-Lncom relatlorubip for each country indiPidurrlly u l l l yiald craada over t b for the variow couatriu. 35 . The S t mthod b l l r with thir problem by Lacroduci~m income variable and a c w r t u i t term for rvrry country. Each h u a poritivr value for the napectivr country mad * zero for .dl otker comtriar. For a r u p l e of two corntrier, the dc.4 a n arruased u follow8 (error term omitted) : 36. 'fhis ertimatioa producer parameterr of the structural variables md the price vttiable for tbc group as a vbole, as uell a8 paraacterr of the inctnae variab',e and a caartm: tera fot each country iadividually. 37. Vhtn rhr IS wrbod 18 used, egwtion (7) la f frat artimtad using a l l &ta for a corntry group Incorporatit.g a11 data oa Income into only oar varL5la. Coarcqwntly, t k r e ir also oaly me coartaat tam. ACtrr thir rzep ir c m l e t e d , r reduced r e a y variable l a calculatrd by d.du$ting fros the arlginal vrrlable t t a rtmctural aad thr price rffactr: Tbe ~ d u c e denergy v u l a b l r l a then ragrerqed for cvcry ccuatry rgainst ttcr 172...avariable: - (1C; h :*- GDP ln b,. A tPl ln -4 'Ibs txplhhd vrtiurct of t.44 two-step rstlmtlm can be calculated M follow: - h e r e s2 the variance of the error tcr. in equation (10). and ? - Y ALsa the v8zi.a:. of tbr origdanl energy variable as it ap?ccrrs on a tbe l e f t side of (7). 38. For a fev country groups, d ~ l e for hdivjduel countries were s dded to equation (3) i f it vnz foued that the observed e n e r n cms.uqtioa of the respective cwmr- 1.9 tatireiy erbova or belw ,ne regression l h e . There con be atveral r k ma for such deviations. :a sa;w cases rich and t h r a f o r e cheap energy rol.r:ea u y t e available. The nric- variab* used h e r t - a c r t a ~ r t e dindex-c..pnot capture suck cases. In cthere, the compo- sition of ootputo vithda Itduatries may be -m~m:alor c;imatic conditions u c e p t ~ a n a l . Hcwever. da tLilitea number of couat:ies, in ?articular Africa, i t appear8 that the dfator~eaexchsagc rater used for converting CDP to real d o l l a n were r a 8 p o ~ i b l af o r many of the deviation#. D. A Brief Coapariaon of the Tuo Hrthodr 39. Ut ther o r ?ot the two m t t o & produce the mama o r different reaulta depeodr on tna axirtcnce of ?attrcorrtlation and iarufficient specification. k in alwwt ray regreraioa buad on ti- aerier, aomr intercorrelation did e d a t barreta the independent vartableo i n the armplea, and the l i a t of erglamtory variable8 v u certainly not c o ~ l e t e . The r a u l t a could not, therefore, be eapected to be identical. 40. Ths d iffanncea cur be mdentaad u folloua. In the TS wthod thrce ir, i n tbe f i r r t rtep, only one conatmt term md m a variable each f a r i n c o r m d the CDP aharer. The e a t i u t e d p a r m t e r r therefore mflect cbaager aver ti= u well as diZfarmcer .raoog countries. In effect, i n the rccoed rtep the ti- eerie. effect ir meparated frcm the croar-aectlon effect for the 3 -corn variable. Iba f i r a t effect i a reflected la the para- m t c r of the In- vaciable, the second i n the coaatlnt tern. 41. 'Lh error that dght aria. with chb approach is that all intef country d i f f ~ r e n e ua n forced to ahou up i n the parametera of the f i r a t step of the u t i u t t o n , crwn i f thay are the reault of factor, vhich haw been o d t t e d from the model, for inrtmca, exchmge rat8 d l t o r t i o n a , -tic d i f f e r m u a , large margy roourcaa, etc. 'Lhir ir a problem of fsuufficient rpacifiution. Rwevar, for the inc- variable, thir error im "correctedn for i n tte aecmd rtap. The corutmt term then capturer thoae inter-cornrtry differtacaa vhich w e n allocetzd te the income variable ln the f i r r t rtep, vherau the parameter of the intopa variable reflect, the Impact of i a c o v chaagea over ti- only. The parnmaterr of the rtructural variables are, houever, not "corrected" and might therefore be overstated. 42. In rba SE method, a separate conmtant term and aa income variable are ir:roduced for every corntry from the beginning. Consequently, differarcee rrong c o m t r i u are reflected i n the respective paramecare. However, because of intercorrehtioa, the parameters, in particular the comtant 'term, might also capture robc of the structural differences which, without intercorrelation, would be allocated t o the structurtrl variables. That this occurs with the SE approach is indicated by the tendency of the parameters of the rtructural variables t o become s t a t i s t i c a l l y insignificant d smaller i n value ae coqarrd vith the TS ~ethod. I f the SE r t h o d $applied with only one c o n s t a t term, the parameters again become more aignificaat and larger. It can, therefore, be concluded that the SE meth tends to mdereetimate 8 the parasctem of the structural variables. The r-.x p a r e t e r s may l i e sonevhere between those estimated 3y the tvo methob. 63. As che paraecers of the s t r ~ c c u r a lvariables are lese significant with the SE nethod than with the TS method, the f i r s t method is not very suitable for siaulacions. It is, hwever, more appropriate for pure projactiom, since it achieves an overall raLalmm of the deviationm, vhereas the IS method eventually reaches only a parcial miahnu . 44. The l a t t e r is 11 ely t o occur i f the income variable and tha structural variables are intarcorrelated. Thim m y ba illustrated am follous: 45. I f the parameter of the income vari&le is al and that of a struc- tural variable a2-asa&g for a m l ~ e n tthat there is only one structural variablt--it i o poaeible to drav isoquants of the e m of equated errors which la to be minimized. In the f i r s t step of the TS rrrcthod, a minlmum i e obtained a t a* f o r m estimated ;verage a1 (-at). In the second step, the 2 cour.try specific r, (-a!*) is eetimted for at. Point (at*, a;) apparently is not the overall minimum. This would only be the case i f the two variables were orthogonal t o each other or, s t a t i s t i c a l l y speaking, not intercorrelated. Then (a?*, a21 would coincide with the center point C. IV. THE PROJECTION -OD 46. The b u i c objective i n derigoing the projection method w u t o c o n s t r x t aa energy model that could a u i l y be linked t o ths World B i a k macro model. By taking a0 input8 into the energy m d e l the emganow variables an projected by the World Bank uacro r~odel,it van poasible t o avoid the vol\rrpioow~tank of developing mother, coarietent vorld macro model. Ihrf~rtunately,the linklng procerr war fairly difficult becaws the output from the World 9.nk mdel var a t a higher level of aggregation than the input needed f o r the energy mdel. It baa therefore neceerary t o dirageregrte the emgenoum variabler u rupplied by the World Bank. 47. The lkalr r ~ o d . 1rwplied projected valurr for the folloving varinbler for 14 world regiotrr : - t o t a l Grorr Domertic Product a t 1975 pricer, - Grorr hmeetic Product i n agriculture a t 1975 price8 , - Crorr t o u r t i c Product i n industry, which includer mlnlng, cooettuction, manufacturing and u t i l i t i e s a t 1975 prices, - Groer Domestic Product in rervicer, which alro includes trarmport at 1975 prices. 48. The projections derived i n t h i s study are based on equation (7); therefore, the data had t o be dieaggregated at tvo levels. First, it was necessary to separate industry fnto nining and construction and into manu- facturing and u t i l i t i e r , and services into tranrport and other rervicer. Second, the group totala for population, CDP and aectoral outputs had to be dfsaggregated into country-specific values. The procedures are demonstrated here for Portugal, vhich belongs t o country group 1. 49. The f i r s t step vas straightforward and did not pose any major problen. A preliminary projection of the share of manufacturing and u t i l i t i e s i n Lndustry for group 1vae obtained from the follcving trend function, esti- rsa:ed over the period 1960-1975: - w . (12) RS7 + RS8 0.7541 - 0.0011 tine 8 Z RSi (109.9) ( ~ . 6 ) i-5 Prelidnary projection: 1980: 0.7306 1990: 0.7194 S i d i a r l y , for the #hare of trrnrport in u r p i c e r of group 1we ham '. Prrlidnary prolectioa: Udalonr of t h u e projectioor vrrr u d a in 8 fru c u u vhar the dir.ggregatioa into country-rpedfic vlluu led to taruaonabla r u u l t r . U). The reoood #tap--tb b r i t n t i o a of cormtry-rpecific valuer-18 8 b i t rot. u>glicatrd. kr-, for irutmu, that tb objectiw ir to brim i the rhat. of agriculture in CDP for Portugal. ihL. can be dorw fn the follarLog vw: - Rb Groom -tic Product in .griculture - i any C0tnrtX-y in the group. h p m r r i o n (14.3) v u b u e d on output f igurem from tba World B8nk lpodel for the projection years. In the c u e of the other three aectorr, (14.3) ru mprasented by the values derived in step 1of the diraggregation proced*zes. 51. E x p r e e i o m (14.1) and (14.2) vem f i r a t astimated by llnear t h treudm. 'Ibe projected t r d v d w r vem then crorr-checked by calculating the sectoral aha- in the'- for each cormtry. I f it turned out that a particular aham rbaued a very rmreuoruble movemt, it v u adjuated on the b u i r of p l a w i b i l i q . 'Ih/rurplw or deficit output of the group resulting from such ar adjtmatnt v.Cthen dirtributed proportionately over the r d i n g cotnrtriem. In ro&i cue. revcral iterative adjuatmeats were necessary in order t o obtain remonable share. for a11 sectors i n a11 couatries. 52. The disaggregation procedure for per capita income war elightly different. F i r r t , tread valuta for the couatry-specific rharer i n t o t a l income and t o t a l population sere e r t h a t e d . I f the ertimated rharer appeared unreasonable, a d j u s t ~ r n t rwvre made. After the final round of adjwtmentr, the valuer of per u p i t a income were calculated. B. Projections 53. After all the erogenous variable8 had been derived as described above, the projections could be made by applying the rtructural and price eluticititr, a2 to i6, by equation (7). u well u the conatant aa estimated texm and the income e l u t i c l t y estimated by eqlution (10). 54. Four venioeu were produced for every couutry rrrd couutry group. For each of the two parameter set8 obtained with the SE r t h o d and the I S method, there i8 one projection period; a second one war provided when it w u asr-d that tb r e d price increme between 1975 and 1990 would .IDDM~o SOX, o r ur awragm of 2.n uzaually. t Th.price inctaue for the three fiva-year periob 1975 t o 1980, 1980 to 1985, and 1985 to 1990 vere -8-d t o be 10. 18 and 15%r e ~ e c t i v a l y . 55. Ibe ptojectioaa for the country grow8 were obtalned by riaply adding the projectloll. for the individual couutrier. In the c u e of per capita emrgy coarrrqtim, the arergy end population projectioru of the variow countries were f i r s t added and then divided by each other. Tbe agsreg~teaof the rtrpctural variabl- vere obtained by calculating weighted me- of the country-specific r t r u c t u r d variabler, v l t h the inwme rharea rervLng as weigtrts. In the case of a group'r per capita Lacowe, t o t a l income an rupplied by the World Bank v u divided by t o t a l population. V. PRESENTATION OF THE RESULTS 56. There ir a s e t of eight table8 for every comtry group. Tablea 1to 5 provide the eetinrater and projection8 from the TS method, wheredo Table8 6 t o 8 pre8-t rerultr obtained with the SE mcthod. *.a content of the eight tables ir runmarFred b e l w . 57. t able 1 lirtr the parameterr ertimted by equation (7). The coefficient of d e t e d n a t i o n , the n d e r of obrervationr and the F-valw for that frmctim a m alro o h m . k tha coefficient of determination ir biaaed i n logarithmic regrerrioar ( m a l l . C above), the rum of the rquared deviationr i n abmluta t e r m in relation t o the run of the rquared obrarvationr i n abmluta terms w u added u a mra neutral measure of the gooduerr of f i t . Tbir r a t i o ir denoted by SUX l/SIJn 2. Itr nmerator and denominator are alro rhova. Plollly, tha table c o n t a m a l i r t i n g of all countries included i n each grow. In the care of Southern Europe, there are eight corntrier. 58. Tha upper half of Tabla 2 rhwr, in tbe f i m t CO~UQ, t o t a l ortimated c m r ~ t i o nof energy. firera data were obtdned by multiplying the artinatad per capita val*ns (colunr 3) by the population data (coltam 5 The secmd calm givus actual energy c o p s q t i o n , the fourth the r e ~ p e c t i v eper cupita values. Ihe lover part of the table ohms the time aerier for the raaining exogenou variabler. The price aerier wan calculated an the weighted averaga of the country-specific price indices, with curremt energy comumption d u a r a r the usightr. (For the projection period ertiuated energy coluumpticm valuer were wed as weights.) 59. Table 3 plots actual, entipated (1960-1975) and projected (1980-1990) per capita conswticm of energy. 60. Tables 4 cmd 5 plot the exogemous variables, except for population, i n arder to f a c i l i t a t e anresement of the projectionr of these variables. 61. ' Tables 6, 7 and 8 provide inforsation for the 8~ projections comparable to that for the TS projections in Tables 1, 2 and 3. It may be noted that Table 6 contains neither an income elasticity nor a constant term because these values were different for every country. 62. The main results for the Individual countries are summarlzed i n Tables A and B. - 8 63. Table A rhous the income elaetlcitles and the constant t e r m obtained with the two methods. Kel~euresof the goodness cf f i t are also given. The coefficient of determination is not shown for the SE method because it was the same for all countries within a group and can be found in Table 1. 64. Tabla B p r u m t r a c t d c a u u q t l o n of anargy by corntry batvean 1975 urd 1978, u -11 u tha p r o j a c t i o ~for 1985 .ad 1990. For each of the tvo mthodr, a w a with c a u t a n t real anerg7 p r l u r aad on. vith a 5OZ prlca i n c n u a are .bcm. 65 Tabla C a h -911 granh r a t u for tarrgy conrrrptloa -for th. c u e of wprtrnt nrl e n a m p r i c u . VII-21 VI. ME RESLZTS 66. Uhen looking a t the nrultr presented below, it should be kept i n mind that the projectionr of energy d e w d depend on the projected valuer of the exogenous variabler. k mntioned earlier, the projectionr of the emgenoue variables are those underlying the World Dtvelopment Report 1979 of the World E d . The authom believe that the projected growth rater for C;DP and i t a cowonentr are generally on the high ride. However, rince all the ertimated paramaterr are prertnted here, it wuld be fairly eary t o recalculate the projectionr of energy demand a m d g lover gmvth ratea . 67. The reader rhould a l r o note that i n the care of the conatrat real energy price, the year 1973 w u used u a base. However, the o i l price h u Pore than doubled rlnce 1973, .nd r e a l energy pricer night already h a w increued by around 302 l n quite a n d e r of developfng corntrier. In general , therefore, the c u e involving Lacmared pricer ir probably clorer t o r e a l i t y t h m the m e v i t h conatant pricer. Thir u r m r , of courre, that no furt5er d r u t i c i n c r e a ~ e ri n price w i l l occur La the 1960s. Bcuever, it ir not m l a u a i b l e that f o r a couple of yearn inflation vill keep pace v i t h the ria. in o i l pricer, so that m a l pricea vill remln rougbly coortmt. 68. In a feu c u e s the ertimatioa wan mde v i t h an a priori given value for a certain p a r m t e r . Thia van done i f Latercorrelation betveen the independent variablea led t o unreuonable reaults. Group 1: Southern Europe 69. The coverage i n the a q l e f o r Southern Europe vaa almost 1002. Gibraltar was omitted. 70. The a i d t a n e o u s entimation f i t s the data significantly b e t t e r than does the two-step estiaation. However, the difference in the projections f o r 1990 ia only about 52, a rather small variation. -.a** a o - I e- m - . Y e . I O I Y C 000 m m m 0 0 - m m m o m 0 m m o o m - m m m . 0 0 0 o o m e a * -0- * * * 8 . * C Y C C I a e - I . Y o n o *La-. *I0 a m - o o o --* ,-.-.- - 4 - a 0 0 3 3 0 0 09 an* ... *.a Y-- **Y -** a 1 3 Y O - I I- . rl 4 urn .. W . u 4 Dm. - D D D a s s m w a 0 . 0 0- * * * *2 4 % n *a 0.- m . a - croups 2 6 3: High and Cpper Hiddle Incoae North Africa and Hfddle E a r 8 71. The coverage for this group wan 77.12. The high incom group 2 (capital surplw o i l exgortere) and the qper middle income group 3 had to be lumped together because it v u not porrible t o obtain rufficient.data and ~ a t i s f a c t o ~r ye t i r t i c a l rarultr for each group reparately. t Lebmoa, Bahrain, United Arab EmIratea, Qatar, Oaran ?ad Kwait were omitted, the l a s tNO because of missing price data. 72. In the t v e a t e p aathod, the ahare of agriculture wm omftted am an explanatory variable because i a 80- couetriar the agricultural rector w u axtrcmly reall. In a logrrithplc epecificatioa, changer i n vary r u l l rharer have a rtroag lmpact on thr explained variable and therefore make the projection rather mrtable. An 8 prior1 p a r a t e r of 0.3 w u wed for the rhare of tranrpor:. 73. The f i t of t h t SE totimation ir a g a k better than that of the TS ertimatioa. The projection8 for 1990 differ by about 10%. 0 4 - h Q 3 O J O n u - - n-.-- P O 3 * * * > w = a 0 3 h a 4 h Ir* z 3 0 3 - - n C b moo D O 0 -00 ... 0 -0 mmo moo - -we 6 a * m - Y me 0 m w o w -**L..-*eo-aa-a**m T v m ~ ~ - 0 m w ~ * a e 8 a m * oa- 6 v a *Aa-am*an*vaOeeD 4 a s *,-a-.**a-o-s..0 O % r aa..a.e*--**.DCa * 8 L a *.U*mm---**.*wCaa a < e . . . . . . . . . . . . . . C C C : I s * o s 0 s * 9 s s 0 o a s 3 0 a > s .......... 26.- & - * * a m - e m e * m * 0 4 0 4 .......... - n m a a * m m * * n r w r * m r . * Y w .......... O O O I O O O O O P 0 I . m o m m m m ....... m o m 0 m m m m o m o m m m m o a * . . i . .0. . e e - l D O I O W O O O D O ~ I O I O D *el)* eeo- n a m e nn*n - l * * a 1 * a * Y + X Z a * r e m a - n a - r e= *r a- a e - * m - * D I Y * a * s o m a * - a r w n r a n r e e w a a r m m * m w m a m * s o - * Croup 4 : Laver Klddle Income North Africa and Kiddle East 74. The coverage of t h i s group waa 97.82. The two Yeraccu were omitted. The share of agricultcre was omitted for thr: same reasons aa i n groups 2 and 3, and the same ~ p r i o r pi a r m t e r w a s introduced f o r transport. 75. The f i t of the SE wtimation is b e t t e r than that of the TS estimation. The p r o j e c t ~ o a rf o r 1990 -.in d i f f e r by about 10X. 'I. 1 - < r:- m' J rn 2 clr pi ur -r I C Z (I w - C 1 . 1 1 a m u L* & Vrr -... w u u a C \ r D 0 - O o a 0 16 *a Pb C Q V # U P . - - - - - - - - - - O O O O D O O O O D -,- e m - --- nry* ... * * a m a . 0 0 s * Y Uh Y O 0 a r,* IryY . . 3 3 83% 3:z.. m o m .I I.) a Y Croup SA: Lou Incot-. crc Africa '- 76. fhe World lank h 3 put the 21 countrirr of lou incoma Africa into onr group. For thr r r t b a t i o l u fn t h i r study, t h r 8rot.p uar rubdfvf&d Loto cuo ro8iolu, thr f i r r t (5A) containin8 7 urd the rrcond (51) 8 corrartirr. fir covrrrgr of both group8 together v u 88t. S o u l l r , Esain, Guiara, Llrotho, Comror urd M a p s c a r u r n d t t r d . I f Boaln a d Cufnra h d bron included i n 5A and tho o t h r r d c t e d c o a t r i r r ia 5b, thr c w r r a l o of 5A would br 792 and tbat of Sb 89.12. 77. Tho r x o w o w data ior indrptry rupplird by tho bar& apprrrod t o br f a r too lou. Hot oaly w u Fodrptrld output in 1975 r u b r t u r t i d l y lover tbm thr corrrrpondint f i p x r ia t h r k t a f i l r of t b r rtudy, but tbr ahare of indwtry la CDP war projrctrd to daclior coaciauowly \rp to 1990, r m thou& it had rprr than doublrd rfaco 1960. Mlo the Bank projection w u uard, t h r r r w u r rapid rirr in thr & a n of revricer. For t h h rcudy the r h u r of indwtry u u d j u a t r d up to 3Ot, 3U 4462 for tho y u n 1960, 1985 and 1990 nrprctfvaly. E m n with thrro rubrturtirl e h u r p r , thr r h a n of rrrvicrr rtill romo. 78. For t h h country group, t& f i t of the TS utimtioa i r rlightly trttrr chm tbat of thr SE ucinrtion. Tbr TS r t h o d producu aignificmtly highrr rrrultr. T%rdiffrrencr for 1990 m m t r to about Mt. 79. Tho rharea of rining +ad eoartructiorr a r wall u manufacturing and d e c c r i c i t y v e n o d t t a d fn the TS ortimation b r u u s r the rrtimatrd paramtterr vere vozy iorignif icant. " .. 0. rU - a n .w n t u 3 u + n m a - L I O N -a e 3 n n n -0 m e - ee 0-0 e- 0-0 r n e m m e :% :s: ii5::: I... 2 a- ee 0- -0- 6e.a -$st?--# - e m s ;aa*a:ss Cad o m e m e - e - 0 - 0 I . . . . . . . . . . . e * * e e m e e m m m e m - *- w a-n --- - 0 - a.0 -.. an* w m a - 0- e m - 0-4 a-n I... m e * * w w ow me-2 a lu' O L L I - e m - - I .a 0 a = * 0 4 a-m I 4 a -* 04 a-4 0 4 - *-a I 4 3 I4 u -* 0 a a a-v I a : 4 0 a u-s I 0 1 ~ " 1 " ~ ~ 0 0 0 0 0 1 0 ~ 1 1 ~ ~ 0 - . - . 1 " ~ 0 - . ~ 1 1 ~ 0 1 0 1 0 ~ 0 0 ~ 0 a u n --- - .* 2:: - o w * w wn-· 08a- 222r& 2- 8 -. 0.. 0- 0 - - * a * *-9 .**)- a m 0 w w m o , . - o m I . . . I . . . . m w * I . . . . 'sow* w..* 80. Sea ducrQ:loa wader group $A for ooprtaga d other &tailr. ma rS utlvtiorr for thir &rowfit. the 6.t. better than the Sll eetlutioa doem, and it ouco *.in producer higher nrultr thaa the arcsad, Rouevet, the differroce 10 about 172 r d l e r thaa for group 5A. a m - m o * a*. a-rn m m o ??: - 0 0 m a m a m- - 0 0 00. w a n a=---- 0 ll- O r m o o o m - *oooor-na mm n 0-eeee a * . . . . s = a a 0 0 n C m e r a '.a* 0-- I . . . L L o m ;; urn t 7 7 7 l n m 7 -8 a n m n 1 - l -8 t " l -a : l-n 8 L -- a - w - 8 C a I l-0 8 L e m * 8 4 a m * 8 4 Y r-L 8 . * - a 8 4 - 8 n r-n 8 . m * 8 a n r - A 8 . r) *-r* 8 4 * Y m - h Y, * Y a s - ) 0 r. a G n - m - m ---o--% 22t:::g m o o o o m @a**.*.. I . . . . . . . 0--or-. 0.0 0.0 o m * o m 0 *-- 0.0 iii 0-a 0 - w -0 ::?33 0 - w n 0 0 0 0 ~ 00000 I..... * o m 0 0 0- en 3 ----o- I . . . . . . --o--o r r r n 0 C . r a a r r a n n r r o m o r o e f f f Croup 6: LGIer a d d l e Iccov Afrier South of S l h u r U. Croup. 6 m d 7, for uhich u p r r r t r ratfmtfoar r n provibd h r n , vrra d o o diati3pdahrd by t h r World B d c , rltho- it rtill u& ita projretionr for th. NOtcgathrr. Tha wvarre for grow 6 v u 9 U ; Caps Vrrdo, Equatorial Cr?lnrr, Botrvam, CuiarrBbaau, S v u i l m d m d Sao T o m and P t i a d p a v r r r o d t t r d . 82. w i n , cb. f i t of tha TS r r t F v t i o o ir b r t t a t thrn that of thr S1 ratisation. Bcrvevmr, in chlr c u a t h S I mtbod producu higher rruulu. 'Tha d i f f a r m m ir about 17% *Y wmener rn a n - m o . . . . a 1WYm* 8 qa c r i a o Y 3 n.. n n n n n n n c. -- Y t a a a r* .Y n L u l 3 a w n 8 l a a - W Y I a m m m m m m m m m m ,..... o m - - 0 '' * o m 0 0 o o m m o m O * a m * *r- m a a m moo-** I ~ 0 . 0 ~ Ztjili f ems:: r m - D ) - I . . . . . . m o m - - 0 a m * =: 4 - 0 C . L O * ) a 0 . m a* O L . * r * m m a w a m * .........--.......:7 u a- m m a- r * a n o m a r r * a - a * ,;: .*..I r * Y Y * n m Yr O O mo r~ymerr )*mr cn an+m* * - .......... I) o m - 0 0 1 10*- m C I m o o m 1 - 0 m 8 ' . I o o w a W O m ~ r* * - o a m u o r* 0 '**r*mmr*r*r*.*r* -- n r o O W w e * .a e m -* s 0 - D 1 0 0 0 0 0 a m m r 0 - o a r ) W Y r ) 0 w -a * -- I' e n - : *-* 0 w r y :" w o o 0 a * - a :" . l -* 0 . s n m 0 0 -a m-n m o o - d m c.me ... a a m CIro m- m 4 4 0 o w OCI drl IC)m 0- Y O rn n Group 7: Intermediate Middle Income Africa South of Sahara 83. The coverage of t h i r aroup v u 92.52. Seycheller, Ceuta, k l i l l a and h'ardbia were omitted. The rhare of mining and coartructioa v u omltted i n the TS ertimatioa becaure it van i n ~ i @ f i c m t . 84. The f i t of the SE ertimatioa ir better thaa that of the TS utimtioo. The TS ertimater are about 172 higher than thore of the SE ertimatea. 85. The ahare of mining and comtructfon had to be oaitted b e c r u e it v u rtatirticdly iorimificaat. m-n -i::sszw; ~ Z * S Sa m . .:2t=:f* * * *ti:..* * * * * * * * * * * 0 * 0 ~ 0 . * 0 0 *** m.. *** 5;: one * * * --- e-. *-- m e - 3 - I 0 L C Y - - = * .a- - n o I - - c oa--*-**a S Y - a -** . 3 3 n-a- m - 0 - e r m a % W m * * * - S a # a * - * * 8 Y s ..................: - . e m S n - - m a n * I 1 a e a a e a * m - * w w a * n o a 3 - - m * * - % e ~ * a * I + m ae * r - e m - m * - a-e--*.-=n w w o m 3 ................... 2 ( UI a o w m o a - - m a - - m * ) = D r I m - * - - * 0 l ; C 8 . - a - - n m m - m a n ~ b - b * * n o o a ~ e e w e w e e 8 e e w a ~ n n n na a .maYmY - . r a m % t - % r ) a m O I a ¶ 0 0 3 0 0 0 a 1 0 3 ¶ 0 0 0 ~ a a 3 0 ~ ~ ~ c . L * o o I )e * . I m-n -- . 00- **- * -' 0 no-'a*- a -a m a 2 m a - .*- ........ ..... ?-- - 4 % E - S 'aa.mr-m.*e 03* rn* w e - * - m a o m m ~-r- m - 0 *.we 3 0 -m* -* a * . wC*a- = w .........s r w r * rr*or m * o m o m 4 ? Y . * * ' 0 . . ;I';t. n o s f-0s E) 0 3 0 4-a hLI " -- @ C) r. @ a @ -.- a Y * U ---:" I I a :" a -m : r. - r y - 8 a U 8 a 7: r. d* 8 a J - I 8 a a-- I a - - a I. I s - n I . I a-. I. 8 a-n @ a * - u - Y I . I a -3 4 . a *-a I S I LIIIII-LLLIII-.I-L~I-.~IILLLL-.L-.-.LLLL-L~L-.L~mm~----------- I-VALUES SUM ( t - k E S T ) e.28 71&4Y6lbL0bbb699 SUM Ee.2 t 147S4450604b.4t0979 Slrhl I YIJML t 0.004&7(, S Z E E f 3 $ i : X 2 : : 2 f S :m= s n m a L I O m - L I e m - y r r r r r r r r r - r r r r r m ................... * ; ~ s ~ mo -omr rr mmm-toeem- m r 0 m 0 m m m m m m m m m m m m m m m o m 0 0 o m o i i o i i o o Cro\tp 9: L a o r Mddlr sad 1ntrrmdi.tr Income k i a - 89. 'Ihcuvmralp of thi. grow v u 91.OX. Sol- Irl.ea&, S m r , T w , Papue U w W,Urr Eabridu, Cilbart Irlro&, Eepublic of Chiaa (Trivro) sad Pacific I a k & u r n a i t t r d . 90. k vith Crow 8, the projected drclLna in tho rharr of agriculturr in CPS appeared to br too rapid. -cultural output v u thoreion ralrrd by ZiX, 322 rod U1S for tha y u n 1980, 1985 rod 1990 mprttivrly. lo opitr of thrrr at- h c n u r r , rtructural chmgr v u rtill very frst. Tho projrctrd rirr f n .orrm damad v u t t u n f o s quit. high. 91. Tha f i t of tha I S r r t i u t i o a ir c l u r l y b r t t r r than that of the Sll e r t h t l o r r . '2hr pmjrctioar IJO- v a n .loo highor. Hwrvrr, tho dlffrrenca vaa only about 10X, not wry l a r p . -- e- m.. *m aa E ......: X m e = e m - u - - . . a * o m - :5S e o * u n * ................ aum e u u . . . --0 4 - 0 m m * * m o * m m n e o m m * m t o - * - - m m m * m m m m m m m * m - m m e n * 2 = = r . . . 0 - 1 n r y t Y Y - . a 0 1 n 9 2 2 2 * a = n - a a - 0 tabla 5 m-roo m a . a m - - - - a *oo 2:: 2:z C I I * * a COO t r t - 0 0 C.*h m e w 0 - w he. 2 3 : ::I ** ** ** -0.. n-0 h a - w w -0- m o o fff a * a * , . . a Y . ? t m o o m o m * a m m s m Group 10: W a r Hiddlr Income Latin America 92. Thr covrragr of t h i r group v u 96.6%. H d t i ( l w income), Crrarda, St. Vincrnt, Dominica, Gutemda, St. Lucia, St. Kittr-Nrvia, Balitr, Nicaragua and Antigua v r r r omittrd. 93. For maufacturiag rod r l r c t r i c i t y , an 8 priori p a r ~ w t r rof 0.3 v u introducrd. In order t o avoid m irglawibla drclinr i n tha rturr of indurtry i n a a d r r of corratrirr, projrctrd i a d w t r i a l output var ralrrd by 15.8%. 16.12 and 15.3% for the g u t 8 1960, 198S and 1990 rrrprctively. 94. Thr f i t of the S t wthod ir b r t t r r thm that of tho TS mathod. Hwrvrr, thr dif f r r m c r i n thr projrctiona-lrrr than 3%-ir marginal. - *Sf ? : sr o2 - ; w e . . - 0 . 0 t .. Y r G c i I.. 0000 C a r) t u 00900 00 m o o m n - 0 - u :$:;$ '.?f $ 8 . . i no* n n w I C 0-4 r m * I 0 -0- - 0 - I .... e o n % S t : r). r)-n. r m m r a o m - 0 u o n e n 7 2 7 : r m n - a - o w n - t ? ? I 2,'s I I 0 0 0 - m e a m * * I o w . -0 0 - --· - - 0 2 : I I a o n o (.I L z z ze m 0 r x i = = 3 ana - Group .L1: Uppe r -3Iddlc Incow Latin A 95. TT\e coverage of this group was M.42. Surinaa, Owdeloupr, Nutherlanb Antil:er, French Cuiacr, Trtr '.dad and T o b t s , P w r t o Rico, Xartinique, B a h m u , Virgin Itlradu, be mud^, Barbadou and Uruguay werr omitted. Rtwoaable data could not be obtained f o r aay of the countries. For transporc. an a prior1 parameter of 0.2 was 'ntroduced fn *he TS eutimrtion. 96. The f i t of the SE e a t l u t i a a i a trttrr thra that of thr TS wtimcrtion, but th. rtavlta o b t a b e d with t 4 r tvcl -tho& a r e almost idea tical. - * CIA -a O C c t 1 1 C \ C I * * n * + * C A J l A +-*&-a& cn-3t r r a a 4 V Y Y I Y I - C Z ; L Z L L V A A A A A A * * m a r l r m * a * r m m * - a . m r a a o r - - - h a m '*)NN.vSry I . . . , . . m m m m m m .. m e * 4 6 - - n s .... sac 0-1 = m a T I . rrc: zm. sz: zs: --. C C . w = . nn. = #?-. r I £S! n rn . t n l . tb * n l n = n t E t C. l Iz = rn rn rn rn rn rn rn rn rn. rn rn ...I m m r n ff: m * * m * Yen*- *- ..... was:: 0 * 0 * 0 m m - e = 3 0 0 0 0 000oa,o*o a 0 0 0 0 0 m w o ....... - n o nnaa-n ... Y Y " aa.0-a a-- '-*a o m * -1- .*a*-w I . . . . . . -.a -LI-.n-4 .*o r m n r - a r*oor--- ------- r-n fia-r*n- .%o-.-m - e n a * - o m 9 * m e w --- s * 0 o - r c n n n a a r a i r * ---r*r**r* 7 I. e :a - I. I. e 0 e L. : r e I.. e 0 e - 0 e . Group 12: .Northern America and Oceania 97. The coverage of thi. group w w 99.72. Only New Ztalmd w u omitted. Strangely enough, it w u ilposrible t o emtimate a reuonable price e l a s t i c i t y f o r t h i r group. An a p r i o r i p a r m t e r of -0.4 war therefore introduced. 98. The f i t of the SE e r t i m t i o n ir rignificantly b e t t e r than that of the TS ertimtim. Hwever, in both projection8 the f i r u l valuer c o w out very high. Thi8 ir particularly ttua for t h a SB mthod, which produced a rwult 242 above that obtained with the TS archod. The rather high rerultr are probably the rerult of the high grwth urmptioar. r r r r r c o = ; z z z z c - n - n - c 2 F X X r X C ' Z z*-v*c- - - u . w V . Z T D w e + - \ 2 - X X 0 4 t u VI w. G w e * - ..-*S$ r e m a a . : 23::: - 0 Y e 1 I I . . . . . . oo*e** I I I I I a . w I a t - I 2 % : : - 8 - I Y2 " . : z ; a - t - * a :*** % a = - - 0 I - 0 - I I rr 'I) * F f I . . . . O O O O . r a w ' O r ) * w m w a o B E E I **r)* .. a a I n r .. - s w 0 'I) Croup 13: Uutern and tlorthern Europe 99. Tha covarage of *his group v u 97.62. Ireland, Iceland and Svitrerland vere omitted. 100. The f i t of the tvo mthodr ir a l m r t identical, but the SE method produces r e a t higher resultr. The difference la about 42. 101. & with Croup 12, it v u felt that the demand projectionr for thir grow were on the high ride. The CDP aruuth r a t e projected by the Bank v u r i ~ i f i c . n t l ya b m the rate projected by OICD. I f the OtCD grouth rater were wed, &maid for energy i n a11 iadurtrial corntrier in 1990 vould probably be r o n 10X l a t e r than rham h e n . - r(r n r a 3 ~ r y n a a- .J a \ -u: I L L .*- - 3 f 2 n m r ; t r u n t mm currr tuurrrtoa r r r r n r currr trrrrurroa a l t t ).#)4?.. ~--1--~--1--1--1--1--1--I--1--t--1--I--1--~--~--1--~--1--1-~-~--~--1-I-1-1--J-!--~--l--~ hu h l 64 n.3 6- b) hb hl hd hY I P ?I)d /A I* 8% *. as e ve ...... .. .. .. w w.. * a w * . a a -* C 2 . L 3. 0 ------- - a am- O D - w e a =--.* e-Ala a *--an m y - - A'. . A * & > L & * a - a s - m 0 *- - . . a . w w rn lan- w E w l . a a n a 1: ---- a- a. *w -*e)rl A' Q, Croups i6, 17 6 18: Other Africa and k l a 102. fie coverage of ?hie grow woe 70. U for Asia and 68.9Z for Africa, excluding South M r i u , f ~ South Africa, r a riaple utrapolatiea was made on pcut trands, as tha required data uare aot rvr:lable m tiar ',ape. I n Africa, Reuaion and nllbouti ware o d t t t d ; in hi.,)(.cao, Fmcb Yolyneeia, New Caledonia, Cw-4, American S m a and Br\esi veto ositted. 103. The f i t is better for the SE e e t h c f o n that for tha IS eatLmatIoa. The projectionr are aligbtly hiahet for the St &',hod. ;be olffennce 16 almost 9%. 0 . . a m . tnf fj 2 2:t*:! ;..;rtZ'i 2s; 23 4-4.. a*-*)* I..... . . m a . - o % z - = * * * . . a w o o , nt-ur 104. I h e r u u l t o o b t a i m d l o r rho iodivid-1 cotatria8 are r w t ! .ad i a Tableo A and 8. Y l t h r e w r d t o rho u t l u t e d lnc- e l u t i c i t i e r , the major coaclwioa l o taat tba e l u t l c i t l ) e r r h a t e d by the SE r r h o d la laraer than that e r t l m t a d by the tS rthJ and &at thn overall i a c m e l u t i c i t y (the averaaa grarth r a t e o f per capita o a e r t j d e n a d i a r e l a t i o a t o th. averaao greurh r a t e o l COP) ir l a r m r than both. 105 Table C 8baur che tnMb r a t e r l o r o n 8 r w c o o r q r i o o by c o w t r y t-rg* am #.a &I am an 0.I 4 I am & I - La (L. - an 1.h 0.u ~ I.. LY 1-1) M 1.. a*&# 1.a (La I.. 4.m 0.. 1.m a-m 1-w (La aI* 1.m a u an -- an e.m - 0.1 a.u a.m e.n a.w a.n -- -- - a-a a.m a.* - - W L .YO.&-. LII b.O -0.11 L I B e m 8.U *L)( @.M L I I1.r) 4 Y -- kr. d LI. 1.m LIS 0.u 4r* w 0.6s n.6) r.8, e.n r.w 6 . h w4mr 1.a LI) a.a I.U n.o r.o e.m e.m r.hs L* W-0- 1.a I.U O C ~ ..a 1.6 1.n 16.- r.n 0.1) e.08 e.H 1.n I.BB m-61 1.s~ 1.6s IJO.%) 1.11 i i r.n - 1-16 1.mm.1) +.os e.16 LY 01.1) r . ~ r.n r.n r.m 1.1s I.- m.n +.SI LPI 8.0 t9.6~ r.n e . ~ @.a1 e.1~ 8-19 I.- -9 +m 0.0 0.1) O.U *a r.u r.n r.w 8.n 1.1rm.1) 4.n a.18 ~ n n . , ) O.BS ~m e.n a. BB I.n 1.n (LO 4u 0.16 4.) 0.1) a.11 r.n r.n r.11 - y..lm - 12a I.m LW 06.1 0. e.m L ~ S (1.0 4.17 @.at r . ~ s e.71 m -I.* 1.1 I.Y DLI) 0.1 LW 1-11 a s ) -1.u e n e.81 e.05 I..J a.w ~ 1 . m 0.1 @.I3 L a r.n r.w r.n LW- I.U a rr *a LI 0-1) e.m e.u r.n r.1, -- W lh m m 1.m LY -11 4 a LI* a-Y 1s.n 6-98 e . ~ r.n r.19 u 0. n 1.0 01.0 4 a o.ss 0.11 a.s) r.16 @.I e.n r.u 3 8.n 8.- a- 4 a a m 0.61 tv.18 4.m e.n r.17 r.r, -. w a.61 1.a -4 4.0 r.81 1.6 t -1.n r.14 LI LI m.n 4.u e-m e.61 f6.n @.IS r.n r.m o. n MLI-d 0.1 L.. ass 4.u LO LY m.1) e. m e n r.n r.11 I.= I-JT 1.n 1-16 m-m 4.u e.n 1.a m.n 4.u r.91 r.m @.IS - I-- I-uOLU a r . ~ 0-u 1.w e m - 8 s r.n LI~.I.~ 4- r.n I-m a1.u 4 - 6 ~ r.u r.n @.I# kl 1.13 I.) aru -0.u ~n e - n m.6) 4-19 @.a r.w r.19 . - - YI 1.n 1.1 a1.n 4-u e.m I-@I 16-71 4.w r.u r . 6 ~ I.U u- Lhs I - ~ C L I ) 4.u r.w e-e r.n e.11 - uninh - rn Lm 8-16 a.m -1.W 8-01 1-U f5.b)8 -1.- @.Y @.W 8.11 -4 I .a 0.6 -LY e-a ~ . S Bu.n -1.n r.m @.so B.~I I.U LII a.11 -1.n r.u e.n cr.n ' -1.n e m O.Y 0.- 2 % n.an to.- a m.m n.mn ~m,mu 1 I a,." *.m 8.b. r,," r u n m.w -.w or- I M . u.w ur.n U.Y r t ~ n urn I 1 u.-o o.m u.m *.- ( a m u %= .II) .II Y C) ua 0 c. I C . ~ nmu m n -.u YL.- I . **AH I r Y ) n * U* *.U m* U m m J l *Y 9.- LI.Ml -.-I U V . - r.r u.r * m* u * . 4 u C W c C .v nmn *- -.n 8m.n d r a w I-- 8 r . u n4-u -.Y- I-.Y- 1am.w un.rU III)LI I I ~ Y -- u. *u CW ma- U*.U fI.*..t. u.an 1.1) :% -.MI m*- am* u r r 1-C m a n Y.U8 I-.- I R a p 41 b r r M U l a W.*rtb Atttc. b Ic(U1. lut 3.9 Sob 12.0 807 5.6 3.S 6.0 S.7 A l l 0...10pi.a -trim b.9 7 1 6.2 7-0 5.9 6.1 6.9 7.6 Prof. P. Gerard M.r Uaiverrity of P 8 ~ l y L r r n i a 3718 tocurt Walk CR Philadelphia, Pa. 19!:& USA Prof. CLokrt bcoo t'd .rsit7 of Word Inotitute of I k o o c d u a d S t a t i s t i u St. Cross buildla8 )(.nor Rord Oxford. OX1 3UL Prof. J. A h a Brwn Unimnity of W o r d Institute of k d u and S t a t i r t i u St* Cross buildiq )(.nor 031 d Orford 0x1 3uL m. Rof E e d - . B o n r e l ILUtimte of Economics Uniterrit7 of hrhs OK-8000 kcbur C h m c k I .C.P. Prof C...ido ILUtitoto d i Scieato Politsche Ooirersita &l Bolopa . Via Zoboni 1 40125 Bologna :ITALY *)Mr. Qaaa* Yorld Bank 1818 E Street 10Y Yuhiugtoa. C.C. 20133 USA Hr. S. Cole Science Policy Uarurch Unit Univerrity of Suerex Brighton UK nr. A.S. Courakia Brareaora Colleaa Oxford UI. Mr. #. C,frta Untwrmity of Wa~ua Caquavaj 55 OK-5230 Odema H kmrk Ur. l b o Crilli Confinduetria Viala dall' ktruraada 30 6m 00144 Ibu Italy Prof. Dr. lmtz Eoffuna Unlntait7 of R.senabur8 L D.8b00 bgeashrrg C~iveraitatamttume31 Ceranuy, FOR. ~ r o f . vend EylIaberg 3 Ia~tltuteof Ecoaodcr : - Univermity of krkrm ~p.8000 krbw c r Dcnmrk Prof. Karsten b u r r e n I n s t i t u t e of Gcaaoricr Uaiverrity of Aarhur DK-8000 k r h u r C Den~ark Ht. m t t h i a r Herr Cniverrity of hgensburg Lb8400 Regensburg Univerrltatrrtrarre 31 G ~ N ~ Y , Mr. b u r l t f o da Ni$rir FA0 C o d l t l e r aad Trade Dlvirlon V l r dell. T a r n d l C a r a u l l a 00100 R o n I t a l y Prof. Jean Pselibdr Netherlands k d c r I n a t i t u t a RDtterdam 16 SO h t ~ ~ s tOudlun e r Netheclan& Prof. a r t i n P.fd.1 Iartitut. o f ~oaoricr Vaiverr iry of krhru DK-8000 k r h u a C Daauck W. Peter Poll& Vorld Bad 1818 E Street 11.V. U!maNngton D.C. 20433 USA nt. Robert *.a der !%ode 1nterrutiott.i Iron and Steel*Inatitute Avenue Wit 14 a-1180 B-~I. & l @ u m - Ur. S h s r h e r Singh World Bank 1818 A Street N.V. Washington? D.C. 20433 USA Ur. htonio la $pad8 Intrtrutional t i n Council 1 Oundon Sttrat London. SYLt 4IQ m lk. Jos dr ttri.8 votld Bank 1818 R Stnet 1Y Y l s h i ~ g tDeC.~ ~2Q133 ~ USA h o t . Alan Yinten 1 Dmparrrnt of Applied Xcoaorlcr Ooinrrity of Cubti- 1 Cambridw a PI NatunlRobbu -I.- E a u n h m t . o n r ~ h r h r @mlNbkrmukthuhnbwba- c r , b a ~ W a l d b u r l t ~ bornMurlNbkr uc- awn & d f drhrworldnrbkracmauythuwu ~ ~ r h ~ m d r b k n o t e d d ~ ~ n b p n m t , t h r c u r ~ r o 3 b c ( r h r dcavndbormdw~ddrub. I bu. S&rorpidLlp.P. 1982.67p.ta. lJ&V M m . $5. montiookforRiluuy Commoditlcr NEW Enoo R Clm. editor cmlmddy W ~ P l p r N o1.. 1982. _ 1 6 2 p p f c l d ~ 3 ~ ) . Ssat No. BK mu.$5. A Slmul- Mobrlof RI# MJtutnunbin WorldPrfray Muiub s & # W ~p.ga.w.4H.1981.u P P f d I*. lab rrfnncrr.w *J. srort No. WPo(99. $3. Tropical HudnoodTnde In Tbe Wotld R u b Ecoaorny. the AsbPuifk ~ ~ C I L I I ) G C I . . ~ ~ KmflTJIuchi k=w- "%r- - trooR<;rSLDubu10tnnttt WdInQ IYU dusmah-.rrrr u d m m r h u r ~ ~ u r d M u u J - ' t - - '-T%r -erhrovChrrprnwm nWQw mdpasdbr* promyopr#onr unlr.l=Nbkr T k ~ ~ ~ P I1974.C I nbrvnqvrrcudl,dbrrdwi*. . 2 fDutm43mra4a- cortrdpodvroondrhsrrnU u.-. drubkr. U744214.fSDNWl&lUtO. JbrcL mlobtwamLk*m+l?m& two. .Ur. IM 1627.U pcI.kot mmm World wak C#ty ccw.rJIN W l b W l 4 . Srrt wr.$10 Mobb .Ur.IH Sh.mrh*r *.& cdJa -w&7*(V+.4 1 s t WCljcr -No. B & W .m, HPuMlca#omOrderF Thb Bank a p ~ SBNDTO: YOUR lRCALDISTRIBUTOR OR TO W O R 'BANK PUBUCATIONS l k r k * u l , d r h h ) P.O. BOX 37525 WASHINGTON. D.C. a 1 3 U.S.A. Date N u n S b S o m ~ ~ i f d d f e m t f m n p u c ~ TItk Nunt P L ~ ntk ----- - - - -- I I - : 1 I I '¶ I I All p h subject to change. Prices m y vary by country. AUow 6-8 weeks for delivery. Subwal Cost I-- WI copies -Air mail surcharge if desired lS2.OO each1 I, - w e MCI hading f a mare Ihntwo complimentary item 1S2.0.eachl S- Total S- Thank you for your order. AIGPrrPU Cll(orMnchSu AteEWr-bk mddB165Cpbo C l b i r G u m a BwllooAbow A U I T I A U A . ? I V A r n G \ ~ -)dUND6 X h N u.uu.- VANUATU Thhamm8Ckr#YIlr*k Woraaorkrvb z s m ~ ~ w- *, kSsv=r maauu ? w a b m 6 . N w l a v ~ matMr. detr#rq .".&a& la0w CANADA taouhmmar AcbrLbrsrrt.mrvarrcr ThSrd C?.# - w s A m d P - QPrbrc r r S O P a r I b d - COdTA auA M uwnDL;IIIK;DOYAND NoumuN- LlrrdnbUd S r r b Acbr Lb. Iq DmMARK ~ ~ , W A b s n - p d I i U ~ r r C AcbrLb.WPdhdIbbll w I&adanr&11 orc-1mCopaaVraV. UMTEDSTATrs ThewoddllntboaLw H;YIT.Ambm@&d 19th%N.W. A1AJmm c-' W- D.C lout AChr:At+.smydZICbrl dQsP.O.ba sm Alc.lu- D.C.30013,USA) Crho CPRANO msouch-Armr %EzzzB LbawnaPbrdc609W roMirdrw8y ~ 1 , S F m f a D ~ a r o . k k . d . . ~ ' He&& 10 z!E%k.N.rknry.mm A- --= 21ZEz- kmrnArabbcr ~ . h d a r l ~ i -v* -Ml.JOdriaKnra l B u b 3 d W - k -13 sAumAIA#A lOOIMaMPIRod Oaa,sonnl 1 ~ ~ 0 0 4 ~ . r k n r y m 2 ='=A HONCItONC.LUCAU KtraaBoo& -- hfaaorr,ua P.O. h3196 1 Sbat N.W. A m r M r ~ W ~ S l l i e h m Wahbvgbn D.C aOOM b R , 146PllmEdwadRad TAIWAN, lUlWA un~sdN.tbnrbootrhop lbrbon hdararabn-wRt*.c..Ltd, Unltrd N.cbnr PLu INOU New Yo& N.Y. 10017 m ~ ~ t d Amr Mr.D.P. Vca Addhd YPltLMLA tikrrl.ddw 5 ~ i R o d P o s t b a r ~ a*musbwlbd AltnMr.JurnPmbs N.rr eahi 1100112 AvdafruwfraDd.Whudcno.52 WAIN -E@ E4Mdo Ap4&. a337 IAmo. S.A.