Policy, Resrch, and Extemal Affairs WORKING PAPERS Agricultural Podlkies Agricultural and Rural Development Departmen' The World Bank June 1990 WPS 429 Ghana's Cocoa Pricing Policy Merrill J. Bateman, Alexander Meeraus, David M. Newbery, William Asenso Okyere, and Gerald T. O'Mara Ghana's cocoa production declined because of policies that overvalued the domestic currency and heavily taxed cocoa exports. A variable rate tax on cocoa (above the critical level) that increases and decreases with the world price would distrib- ute the price risk between cocoa farmers and the rest of society, stabilize cocoa farmers' real incomes, and let consumers share in windfall profits when world cocoa prices are high. The Policy, Research, and Extemal Affairs Complex distributes PRE Wor'xing Papers to disserninate the findings of work in progess and to encourage the exchange of ideas among Bank suff and all others interted in development issues. These papers carry the names of the authors, reflect only their views, and should be used and cited accordingly. The findings, interpretations, and conclusions are thc authors' own. They should not be tttnbuted to the World Banrk, its Board of Directors, its management, or any of its m rmhcr countnes Policy, Research, and External Affairs Agricultural Policies WPS 429 This paper a product of' thc Agricultural Plolicics Division, Agriculturc and Rural Dcvelopment Dcpartment- is part of a largercl'fort in PRE to analyze thc cconomics of perennial crops and thus provide policy guidelines for e'flicient agricultural development. Copies are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Ciccly Spooncr, room N8-039, cxtension 30(464 (214 pages with flgures and tables plus 103 pages of annexcs). The long decline in Ghana's cocoa production Taxing cocoa farmers directly lowers their -from half the market 25 years ago to onc- income, indirectly lowers the income of food tenth the market now - was associated with produccrs, and transfers incomes from those policies that overvalued the domestic currency, farmers to food consumers -worsening income implicitly taxed cocoa cxports, and ignored the distribution. realities of world markets. It would probablv help to stabilize the rcal This study addresses the dilemma Ghana's producer price of cocoa. An attractive alterna- govemrnment I'aces: how to provide cnouglh tive is a variable rate tax on cocoa (above the producer incentives to stitnulate Lhe cocoa critical level) that increases and decreases with exports Ghana needs for foreign exchange while the world price and would distribute the world maintaining ilie government revenucs nee(ded to market price risk between cocoa fanners and thc avoid unmalnageable fiscal del'icits. The k-y rest of' society (assuming revciuc requircnts maN N: idt6ing acceptable revenue alterTa- are h'ixed). 'Tihis would stabilize thc real incomes tives to ocoa export taxes, Amonig coniclusioris oflcocoa f'anTlers and let consumers share in the auihlors reach: windlfall prot its f'rom high cocoa prices. 'I'o mainitain the right price incentives in tic 'I'Thc yearly cocoa buying price should cocoa subsector, the exchangc rate regime always be set above !40 cedis (in 1987 prices) if' should be libcrali/cd so that the prevailing rate the objective is to stimulate growth in long-term always roughly equals thc cquilibrium level. supplies. Thc government should explore shifting COCOBOD should terminate its marketing taxes l'rom cocoa producers to all consumilers by activities for col'fec, and all laxes on col'f'ce increasing taxes on appropriate consumcr goods exports shiould cease. Even with heavy taxation, -CspecialIly Lhose that arc income clastic and net revenue f'rom cof'ec is negative to thC price inclastic (such as gasoline, cars, andl government under the current arrangement. consumcr duraibles). Stopping cof'fec taxes would give cofl'ec produc- crs a stronger incentive to expand production, A policy of low cocoa taxcs anl high cocoa particularly where swollen-shoot virus infesta- production lavorably affects thc rural poor. tion has rcduced profits from cocoa production. 't'hc I'RE~ Workingt Pa;cr Sc'ri.s diis .minak's thc. tindiing.s of ork; undc.r w ay in thc R3ank s l'olicy, Rcscarch, and External Affairs Compilex An objectis Lof thc scrics is to gCt thsc tincirigs out quict\, vcen ifl presentations are less thlai fLIu) polished. 1'he Fmdings interpretations, aid conclusions in iltise papers do nol necessarilh represent ofFicial Bank policy. Pmdu( a the t'RE Ii)tss FD niinsiintwn terr EXECUTIVE SUMMMRY i) This report presents the findings of a study of cocoa pricing policy for the Ghana Cocoa Board. The origin of the study was the realization by policymakers in the government of Ghana that past cocoa policy had been misguided. Twenty-five years ago, Ghana was the world's leading producer, wlth a market share approaching one-half. Today, Ghana's production levol Is about one half of what It was 25 years ago, and its market share is only one-tenth. The world leader now Is Cote d'lvoire, whose production of cocoa has tripled over the Intervening years. The key to the divergent paths of the two neighboring countries is In the pollcies their governments have adoptod with respect to cocoa producer Investment incentives Sd other policy instruments such as foreign exchange rates, taxes and interest rates. ii) The long decilne In Ghana's cocoa production from the mid-1960s to the early 1980s was associated with policles that overvalued the domestic currency, and the implicit taxation of cocoa exports vla the high price placed on domestic currency was the major source of excessive taxation of cocoa. In truth, the formulation of cocoa policy appears to have been divorced from the realities of the world market. Yet, cocoa is still the major source of foreign exchange earnings, supplying 50-60 percent of the total. Since a viable subst'tute Is nowhere In sight, cocoa exports will have to remain the major source of foreign exchange earnings for the foreseeable future. This study attempts to sort out the issues and polnt the way to a resolutlon of the dilemma facing the government - - how to provide producer incentives sufficlent to Induce cocoa exports adequate to meet the country's needs for foreign exchange earnings while maintaining govainiment revenues at a level that does not create unmanageable fiscal deficits. The key to resolving this dilemma may be the Identification of acceptable revenue e'ternatives to cocoa export taxatlon. ii 111) The main conclusions of the paper are as follows: a. World cocoa prices have been falling; and given the large stocks of cocoa In the world and the expected Increase In world output, the price will continue to fall Into 1990, declIning 25 percent below the 1987 levol. rollowing the 25-year cocoa cycle, the price will recover to the 1987 level by 1995 and reach a cyclical high somewhat after the year 2000 expected to be two to three times the 1987 level. b. There Is evidence from the Agricultural Economics Survey for Cocoa Producer Pricing that Improved Incentives have Induced new plantings of cocoa. Thirty-three percent of the land under cocoa by farmers interviewed was planted In the past four years. C. It Is recommended that the government sustain a policy of providing adeq4uate price Incentives to cocoa producers. To be realistic and because of fluctuations In the world price, cocoa buying prices should be based on real costs and not Just on an artificially-set proportion of the f.o.b. price. In the context of growing population and Income, a price of 120 cedis (in 1987 prices) Is not adequate to sustain cocoa production over the long term; If a food grain self- sufficiency policy Is adopted, neither Is a price of 140 cedis. Except for this case, a price of 140 cedis will sustain but not Increase cocoa production over the longer term. Over the medium term (say ten years), somewhat larger volumes of cocoa production can be produced at these prices. Note that a price of 120 (1987) cedis Is equivalent to a price of 165 cedis In 1988 prices. d. In order that the government can continue to provide the required price Incentives In the cocoa subsector, the exchange rate regime should be liberalized so that at any polnt In time the prevailing rate Is approximately equal to Its equilibrium level. This Is only one of the ma,iy benefits from a llberalized exchange rate regime. e. In addition to foreign exchange earnings, a low cocoa tax, high cocoa production policy impacts very positively on the Incomes of the rural poor, and this policy Is preferred on Income distributional grounds. In contrast, taxing cocoa farmers directly lowers the Income of cocoa farmers, Indirectly lowers the Income of food producers and transfers Incomes from these farmers to food consumers. Since cocoa farmers are poorer than consumers of taxed goods, and nontraded foods producers are much poorer than food consumers In jeneral, such taxation worsens the Income distributlon. iiI f. Apart from the essential rlrst step of llberalizing the exchange rate ragkne, the Important stop In offsetting a fall In revenue from cocoa taxation Is to explore the extent to which taxes can be shifted from cocoa producers to all consumers by 1 reasing taxes on appropriate consumer goods. The ideai subjects for such taxation are those which are income elastic and price Inelastic -- gasoline, cars, consumers durables are all good examples. g. There are good ground3 for stabilizing the real producer price of cocoa. That is, above the level at which farmers are just willing to plant and maintain cocoa, the tax rate -an Increase with the world price. A variable rate tax on cocoa (above the critical level) results In a sharing of the world market price risk between cocoa farmers and the rest of society (assuming that revenue requirements are fixed). This both stabilizes the real incomes of cocoa farmers, and lets consumers also share in the windfall profits from high cocoa prices. h. Owing to long gestation lags In production, even setting cocoa buying prices for the short to medium term requires a long-term perspective. In brief, the questin of the long- run competitive trade advantage for Ghana must be answered In setting an optknal cocoa export tax even In the short run. Past policy In Ghana can be Interpreted as having assumed that long-run advantage did not Include cocoa, but that assumption has not turned out to be true. At the very least, pollcy ought to consider mixed strategies which admit the possibility that present expectations of the future can be mistaken. If we admit that cocoa may likely be part of Ghana's long-term advantage, then the yearly cocoa buying price should never be set below 140 cedis (in 1987 prices). It should be set above this level If growth In long-term supply Is preferred. L An annual decomposition of selected pricing policles over the 1988-97 decade showed that under a likely scenario - - growth In populatin, income and wages with a cocoa buying price of 140 cedis (in 1987 prices) -- cocoa production would remain at or below 250,000 metric tons per year through 1990 and then rise rapidly to reach a level of 500,000 metrl_ tons by 1997. The reason for the lag In production growth Is the long gestatlon lag after planting before cocoa trees produce cocoa at something approaching their potential. The same scenario shows total cocoa acreage remaining relatively constant as obsolete traditional varleties are replaced by higher ylelding hybrid varleties. Thus, the same acreage with relatively mature hybrid trees Is capable of much greater production. The lag In cocoa productlon also npliles a corresponding lag in farm Income. In the intorkn period before farm Income rlses In real terms, policies which heavily tax cocoa are capable of destroying the expectations of long-term profitability that induced the planting Investments of the past four years and Iv thus abort the resurgence of cocoa production over the onger term. J. It Is recommended that the COCOBOD terminate its marketing activitles for coffee, and all export taxatlon of this commodity be ceased. Even given the present heavy taxation of coffee, net revenue from coffee Is negative to the Government of Ghana under tho current arrangement. Stopping coffee taxation would give coffee producers a much stronger Incentive to expand production, particularly In those areas for which swollen shoot virus infestation has reduced profits from cocoa productlon. TABLE OF CONTENTS EXECUTIVE MWARY .............. I INTDROOCTION . 6 rART ONE: ANALYSIS OF COCOA PRODUCER INVESTMENT RESPONSE TO PRICING POLICY ... 1.1 A. SOME ECONOMIC CONCEPTS CENTRAL TO THE ANALYSIS OF PRODUCER RESPONSE * Opportunity Cost .1.1 * Consumption Discount Rate. 1.1 * Marginal Efficiency of Capital. 1.2 * Risk Premium. 1.2 * Economic Rent .1.3 B. A REVEW OF THE CURENT STATUS OF GHANAIAN AGRICLTURE ......... . 1.4 Land .1.8 * 'Labor ...................................... 1.13 * Purchased Inputs. 1.14 * Farm Operatlons. 1.15 C. qNALYSIS OF PRODUCER INVESTMENT. . .. . 1.23 * Methodology for Dynamic Analysis of Response to Pricing Policy. 1.23 * Tree Crop Investments. 1.25 * Land Allocation. 1.26 * A Brief Technical Note .1.30 * Model Validation Using the Annual Decompositon Model .1.33 * Policy Simulations .1.38 * Scenario 1--Supply Response under Flxed Resource Supplies .1.39 * Scen:rlo 2--Effects of Population Growth on Supply Responses .1.47 * Scenario 3---Effects of Growth In Per Capita Income on Cocoa Supply .1.54 * Scenario 4--Effects of a Policy of Food Self-Sufficiency on Cocoa Supply .1.68 * Annual Decomposition for the 1988-97 Decade of Selected Experiments. 1.72 * Policy Implications of the Simulation Experiments .1.80 IL MANRAC SPEICFICATON OF IANA iwLnT_O AV MTLAPL SECTOR W ...................... 1.3 * Set Dofhnitions ................................ 1.83 * Variable List .................................. 1.84 * Equation List ..1.85 * Paramoter List ..1.87 PAT TWO. COCOA TAX AND REVEME ALTERNATVES .2.1 I. hntroducton .2.1 II. The Analysis of Taxes and Tax Reform .2.5 III. Salient Features of Ghanaian Cocoa Prichg .2.8 IV. Brie History of Cocoa Production Prichg and Taxation hI Ghana .2.11 V. The Consequences of the Dclino In Cocoa Production 2.28 VI. Other C0nsequences of Exchange Rate Liberalization . 2.35 VIl. Issues in Setthg the Producer Price of Cocoa .... 2.37 Vill. Alternatives to the Cocoa Export Tax .2.38 IX. COCOBOD'S Costs .2.40 X. The Optknal Export Tax .2.41 Xl. Standard Arguments for Output Taxes .2.43 XiI. Quallfications .2.54 XiII. The Risks hnoived In Setting the Producer Price Too Low .2.56 XIV. kIpilcatlons of Taxing Cocoa for Input for Input Subsidies .2.58 XV. Spatial Variations In Cocoa Pricing ............ 2.59 XVI. Price Stabilization .2.60 XVII. knplicatlons for Taxes on Consumer Goods .2.62 XVIII. onclusions .2.84 References .................... ...... 2.65 APPENDIX A: A Note on Data . 1-6 PART THEE: GLOBAL PERSPECTIVES ON COCOA SUPPLY AND DBAAND ... ......... 3.1 I. Introduction .3.1 II. World Cocoa Market Environment, 1988-2000 .3.3 A. Cocoa Hectarage, Major Producnrs, 1970-87 ... 3.3 B. World Cocoa Production & Prospects, 1975-2000 3.6 C. World Cocoa Production Summary, 1988-2000 3.21 D. World Cocoa Consumptlon, 1975-85 .3.23 E. World Supply/Demand and Price Forecasts, 1988-2000 .3.24 - 2- IIL Ghana's Cocoa and Foreign Exchange Polcis ...... 3.42 IV. Conchlslons and Sumuary ................ 3.44 FO NMTEATION OF POLICY WPLICATINS FROM IDIVmUAL S1UDES AND RE DAT ............................ 4.1 ANNEXES: 1. Multiporlod Agricultural Sector Model and Numerkcal Shnulatlon Esthuates II. Optimal Trade Taxes on Agriculture In Developing Countries Ill. Ghana's Cocoa Environemnt TMALES: PART ONE 1.I.1: Distribution of Survey Households by Reglon . . . 1.7 I.B.2: Age Distribution of Household Members by Region.. 1.8 1.8.3: Average Landholding per 4ousehold by Region .. 1.9 1.8.4: Average Size of Plot by Region 1.10 I.B.5 Number of Farmers Growhg Cocoa and Oil Pam.. 1.12 I.B.6: Age Distribution of Tradional Cocoa Trees by Region 1.12 1.I.7: Mean Age Distribution of Hybrid Cocoa Trees by Region 1.12 1.8.8: Mean Age Distribution of Oil Pakn Trees by Region.. 1.13 1.8.9: Labor Use on Farms by Region 1.14 I.B. 1 0 Mean Labor Requirement for Farm Operatlons. . 1.16 1.I.1 1: Mean Labor Requirement for Harvesting One Acre of Various Annual Crops (Mandays). . 1.19 I.8.12: Mean Labor Requirement for Threshing Produce from One Acre of Various Annual Crops. . 1.20 l.B.13: Llvestock Numbers and Holders .............. .. 1.21 I.C.1: Ghana Macroeconomic Projection 1.31 I.C.2: Dlsaggregation of Macroeconomic Projections, 1984-85 1.32 I.C.3: Annual Decomposition of Multiperiod Model Validatlon Period 1.34 I.C.4: Comparison of Slnulated and Actual Cocoa Production 1.36 I.C.S: Calculatlon of Farm Income (Concept 3) from the National Accounts 1.37 I.C.6 Arc Elasticlties of Cocoa Supply 1.41 I.C.7 Sbnulated Annual Pattern of Labor Use In 2015-19 1.49 I.C.8 Food Grain knports 1.55 I.C.9 Annual Decomposition of Multiperiod Model Solutions 1.76 I.C.1 0 Cocoa Supply Esthiates 1.82 - 3 - TABLE:- PARr TWO 2.1 Distributional Characterlst',s for Ghana ..... . . . . . . 2.48 TABLES: PART THREE 3.1 Cocoa Hectarage, Brazil Cote d'lvolre & Malaysia (1970-88) ..... . . . . . 3.4 3.2 Major Producer Yleld Patterns by Year ..... . . . . 3.6 3.3 Coto d'lvolre Hectarage, Yields & Prodt..'lon .... . 3.8 3.4 Brazilian Hectarage, Yields & Production ..... . . . . 3.11 3.5 Peninsula Hectarage, Yields & Production ..... . . . 3.14 3.6 Sabbah Hectarage, Yields & Production ..... . . . . 3.16 3.7 Sarawak Hectarago, Yield & Production ..... . . . . 3.18 3.8 Cocoa Productlon Prospects by Major Producing Area 3.22 3.9 World Cocoa Grindings ........ .. . . . .. . . . . . 3.25 3.10 Productlon, Consumptlon Stocks & Prices ..... . . . 3.27 3.11 Cocoa Consumptlon Regression Model (1960-2000) . 3.29 3.12 Cocoa Price Regression Model . . 3.31 3.13 Ghana Cocoa Premium Regression Model (1960-20) . 3.32 FNlGUES: PART ONE I.C.1 Cocoa Supply Response ....... . . . . . . . . . . . . 1.40 I.C.2 Cocoa Acreage Adjustment. . 1.42 I.C.3 Farm Income ......................... . 1.43 I.C.4 Arc Elasticities of Cocoa Supply ...... . . . . . . . 1.41 I.C.5 Labor Uses Under Fixed Resources ...... . . . . . . 1.46 I.C.6 Cocoa Supply Response Under Populatlon Growth . . . 1.48 I.C.7 Cocoa Acreage Adjustments to Price Under Population Growth ...... . . . . . . . . . 1.50 I.C.8 Farm Income Under Population Growth ..... . . . . . 1.51 I.C.9 Land Use Under Population Growth, 2015-19 .... . . 1.52 I.C.10 Labor Uses Under Population Growth, 2015-19 . . . . 1.53 I.C.1 1 Cocoa Supply Response Under Population & Income Growth . 1.57 I.C. 12 Cocoa Acreage Adjustmer Populatlon & Income Growth 1.58 I.C.1 3 Farm Income Under PopL.P don & Income Growth . . . . 1.59 I.C.14 Land Use Under Population & Income Growth, 2015-19 1.61 I.C.1 5 Labor Uses Under Population & Income Growth, 2015-19 . 1.62 I.C.16 Cocoa Supply Response Jnder Population & Income Growth . 1.63 I.C.1 7 Cocoa Acreage Adjustment, Population & Income Growth 1.64 I.C. 18 Farm Income Under Population, Income & Wage Growth 1.65 I.C.19 Land Use Under Population, Income & Wage Growth, 2015-19 .1.66 I.C.20 Labor Uses Under Population, Income & Wage Growth, 2015-19 .1.67 I.C.21 Cocoa Supply Response Under Population & Income Growth Wlth Zero Food knports ....... . . . . 1.69 I.C.22 Cocoa Acreage Adjustment, Population & Income Growth -4 - With Zero Food wtrortc . .e...................... 1.70 I.C.23 Farm hcome Uder Food Self-Sufficlency . . 1.71 I.C.24 Land Use Under Self-Sufficiency, 2015-19.. 1.73 I.C.25 Labor Uses Under Food Self-SuffIcIency. . 1.74 FIGURES: PART TWO 1. Production .2.12 2. Real Cocoa Price (Producer prce' deflated by rural CPI) .......... 2.13 3. Cocoa Production, Prices, Income .. 2.14 4. Production and Lagged Prices .. 2.15 5. Real Cocoa Pricos (at constant $1980). 2.17 8. Ratio of Producer to World Price (Ivory Coast and Ghana). 2.18 7. Ghana Real Producer Prices (Deflate Rural CPI). 2.20 8. Real Crop Prices (Deflated by Rural Food Price. 221 9. Relative Prico of Cocoa (Defla'.. by Maize Prices) . .23 9A. Cocoa-Maize Price Ratio (Ratio of Domestic to World Price Ratios). 2.24 10. Area Under Crops.. ................ 2.25 11. Area Under Cereals. 2.26 12. Revenue from Cocoa Sales . 2.29 13. Government Revenue (Deflated by CPI). 2.31 14. Revenue from Cocoa Sales .2.32 FIGES: PART THREE 1. Cote d'lvoire Cocoa Production .3.9 2. Brazil Cocoa Productlon . 3.12 3. Malaysia--Peninsula Cocoa Production .3.15 4. Malaysla--Sabbah Cocoa Producton .3.17 5. Malaysla--Sarawak Cocoa Productlon. 3.19 6. Cocoa Consumptlon. 3.26 7. Cocoa Spot Price Ghana. 3.28 8. Ghana Cocoa Premium. 3.30 - 5 - INToDUC'ON This report presents the findings of a study for the Ghana Cocoa Board of cocoa pricing pollcy. T;e origin of the demand for this study was the realization of policymakers In the Government of Ghana that past cocoa polcy had been misguided. Twenty-five years ago, Ghana was the world's leading producer, with a market share approaching one half. Today, Ghanas production level Is about one half of what It was 25 years ago, and Its market share Is only one tonth. The world market leader now Is Cote d'lvoire, whose production of cocoa has tripled over the Intervening years. The key to the dIvergnt paths of the two neighboring countries is In the policies their governments have adopted with respect to cocoa producer investment incentives and other policy Instruments such as forelgn exchange rates, taxes and Interest rates. In spite of Its relative decine, cocoa is still the major source of foreign exchange arnig for Ghana, supplyin over 60 percont of total export eamings. Since a vlable ujbstitute Is nowhere In sight, cocoa exWorts will have to remain the major sour:e of foreign exchang earaings for decades to come. Thus, the fate of' the country's development aspirations depends on the vitality and vigor of the cocoa sector. The prosperity of the cocoa sector depends, In turn, on '-hd their farms and most of them sprayed only once a year. When the crops are matured, they are harvested. Some of the produce are consumed by the farm household and the surplus Is sold either In the local market or to a wholesaler who will take them to a large urban market. Groundnuts and the cereals - 1.19 - can be harvested continuously untit the whole crop has been Completely harvested. Due to problems of storage the root crops are harvested as they are needed for home consumption or for the market. The mean number of man days required in harvesting the annual crops Is summarized In Table 1.8.11. The harvested Droduce are carried in baskets or sacks to the house before they are sent to the market. Farms In the North are closor to the house of the households than farms in the South. Northern farmers practice what is termed 'compound farming' because the farm Is close to the house or dwelling place. Farmers In the South have to walk some distance before they get to their farms. Thus, distance from farm to house tend to be shorter In the North than in the South. On the average, It takes the farmer In the South 24.5 percent more time to convey produce from the farm to the house. When the produce Is sent home, some minknal amount of processing Is done In case of cereals before the marketable surplus Is sold. Normally, millet, sorghum and rIco are threshed and maize Is shelled. Groundnuts may be sold shelled or unshelled. The labor requirements Involved In these operations are summarized In Table 1.8.12. Table Mean Labor Requkrwmnt for Harvesthg One Acre of Varbus Aral Crops an Days) Region Casuva Yam Miut Sorghum Rioe Maize Groundnut North 9.7 8.1 5.3 5.2 10.7 9.6 5 5 South 14.7 14.1 . 21.7 24.2 6.3 Ghana 14.4 10.6 5.3 5.2 14.0 11.7 6.1 - 1.20 - Tabt L LS12: Mew LbO e~ibinen fRW 1Iud*q Prou fraU Ore Awe of Vwu Mwu CoP Region Mi Sorghum RiAe Mai North 6.7 4.6 6.2 9.1 South 12.7 13.6 Ghana 6.7 4.7 5.4 12.3 Mixed farming involves the raisig of both crops and anhals. Apart from generating an extra Income for the farm household and providng draft power and as a source of protein, ankual droppings can be used as manure for crop lands. In a country where fertilizer use Is ileted, the maintenance of sol nutrient with anbal waste material could be very essential. The major Ivestock kept by Qhwalan farnrs are cattle, sheep, goats, pgs, chickens and guinea fowls (Table 1i.13). Cattle re largely kept in the northern savannah. Over 90 percent of the cattle belongi to the households In the survey were In the North. Both the North and the South have sheep and goats. There are about 28 percent more sheep and goats In the South than In the North. For religious and cultural reasons, many Ghanaians do not eat pork. The trend Is changing slowly especially In urban centers where port khebab Is becomig very popular. Consumption of pork can be hereased If more processed pork products could be made available on the market. Both the North and the South had about the same number of pigs, although the average holdng per household was higr in the North than In the South. Guinea fowls are reared mainly In the North whereas chickens can be found both in the North and South. Alfost every farm household keeps a few poultry birds either for the eggs, or the meat or occaslonally to sel when the cash flow is low. The average holding i the North Is 16 bids per household and seven birds per household In the South. Commrcial poultry (chickens) keepig Is undertaken aroUnd tho urban centers. None of them was Included In the survey samPle. - 1.21 - Table 1L.1 3: UveeWtok mbers and HodersNo a/ Cob Uwp & Go" PIs Po0I41 Region Number Holderl Number Holders Number Holders Number Holders Ashani 0 0 234 67 9 2 575 58 Brong Ahafo 0 0 43 18 9 3 227 33 Cental 0 0 324 61 10 2 645 70 Efamn 0 0 197 57 72 1 507 64 Greer Accra 12 7 70 20 107 16 84 13 Nortwn 451 102 455 157 145 31 1368 165 UpperEast 182 56 292 81 36 13 1044 89 Upper WeN 342 61 211 69 15 4 447 53 Vora 34 5 295 81 35 5 1056 102 Weeosm 35 2 165 42 14 3 600 67 Nokth 975 221 968 207 187 48 2659 307 South 81 14 1328 348 187 34 3694 408 survey 1066 235 2286 653 363 82 6553 715 a/ Holders may be greater than sample se due to aggregabon of animal t"yp. * PourVy kiudes chckwns, skeys and guWn fowla . Livestock can be reared intensively by putting them In kraals, pens or coups and feeding them with concentrates, gralns, and fodder. Alternatively, they could be left to roam In an open space and look for food. Commercial rearing of cattle may use both techniques so that at certain times of the day the ankials are allowed to graze on a pasture and at other tknes, especially durlng the dry season, the feed Is supplemented with concentrates and grains. Commerclal raising of poultry requires keeping the birds in coups or houses and feeding them with concentrates. On the other hand, farmers who keep a few birds allow them te roam and look for their food although occasionally they are given small quantities of grains. Pigs are normally kept In pens but in the villages they are allowed to roam about. One major cost component In livestock maintenance Is voterinary care and modIches. The anknals have to be vaccinated against some common Illnesses and when they are sick they have to be given drugs. Since the withdrawal of subsidies on veterinary drugs, veterinary care could be expensive. On the average, the Ghanaian - 1.22 - farmer spends $59.70 per cattle, $11.55 per sheep or goat, $13.00 per pig and $0.91 per poultry bird in a year on veterinary care. These expenditures are very low If one considers that the goverrnent has removod the subsidies on veterinary drugs. Apart from veterinary expenditure, other costs are mlnkial. - 1.23 - C. ANALYSIS OF PRODUCER INVESTMENT RESPONSE TO COCOA PRICING POUCY 1. Methodology for Dynamic Analysis of Response to Pricing Policy The analytical approach to cocoa Investment must start with the reality that SUCh Investments are long-lived (30 years on average) and Involve a significant gestation period between initial Investment and start of productlon (about 5 years) with an even longer perlod required to achieve full production. Thus, perennial crop Investments such as cocoa commit present resource outputs, as well as future ones, In order to generate future outputs, which will be sold at unknown future pricos. In calculating the opportunity cost of conmitting land and labor to a perennial crop, the potential Investor must also form some notions of the future prices (and costs) of alternative crops, both annual and perennial. Since no one has the ability to predict the future, the Investor must adopt some strategy In formulating expectations of the future. One such strategy Is to assume that the present situation remains the same In the future--e.g., existing prices, incomes, technology, or In a more sopilsticated form, their trend rates of change remain constant. This myopic strategy Is belled by history, and has little to recommend It. Another approach Is to use the past as a gulde to formulating expectations. This backward- looking strategy Is better, and It often works well. However, It assumes that the pattern of the past will prevail In the future. This can be a dubious assumption In situations where people have no desire to follow the pattern of the past. Since the history in Ghana of economic policy has been of this kind, the past may be a poor guide to the future. Another approach Is to use all currently available informatlon to analyze the effects of present policies in formulathg expectations. This forward-looking strategy seems to be a better way of formulathg expectations for countries in Ghana's situation, - 1.24 - and It Is presumably the one used by farmers and other Investors In forming investment plans. Of course, If the economic policy of the government Is Inconsistent or chaotic, the plans made by private agents will minimize future commitments and seek rather to Invest outside of the country to the extent that this is feasible. For the purposes of this study, a continuation of the economic adjustment program that improves economic efficiency and leads to sustained growth is assumed. It should De noted that this scenario Is quite different from the experience of the past 20 years In Ghana, which have not seen sustalned growth In real per capita Income. In this sense, the study is conditional on the continued success of the adjustment program. Thus, a malntained hypothesis of the study Is that private agents such as farmers have formed similar expectatlons concerning the ongoing adjustment and that the Government of Ghana acts in such a way as to confirm these expectations. Formulation of a simulation model requires careful specification of the domain of analysis--Le., what aspects of the economy (and the political and social environment) are to be taken as given, though subject to parametric variation to assess the sensitivity of model results to basic assumptions. In our context, this specification includes macroeconomic and International prices--e.g., real wage, real Interest rate, foreign exchange rate, world prices of cocoa and other tradeables--and the economic policies of the Government of Ghana. The model must be multiperiod In order to capture the essentlal way that time enters into decisions on long gestation, long-lived agricultural Investments. Model solutions simulate period-by-period multi-market equilibria for the major alternative commodity outputs from the agricultural sector. It Is necessary to model perennlal crops in the context of the agricultural sector since a comparison of expected returns from present and future alternative Investments In annual crops is critical to the efficient choice of perennial investments. - 1.25 - Tree Crop Investments The essence of a perennial crop Investment such as cocoa Is to plant tree seedllngs on cleared land set aside for the crop and to apply labor for cultivation, pruning, spraying of pesticides, etc., before and after the maturing trees begii to produce an output, as well as to harvest, ferment, dry and transport to market the output. In order to reallocate the land to another crop, It Is necessary to disirivast by clearing the land. If the prospective returns from alternative crops do not Justify the Investment In clearing, and the return from the crop Is so low that harvestlng and marketing the output Is unprofitable, then the land and the trees may be abandoned. This outcome may occur under some combinations of misguided policies and exogenous shocks. Therefore, we treat clearing and planting as separate Investment decisions and permit the (dis)invectment decision to abandon as well. Of course, the short-run marginal costs of producing from a mature tree are quite low In comparison with the long-run marginal costs of Investment and operations so that Is more likely that misguided policies will deter investment In planting while some level of harvesting remalns profitable. In modeling tree crops, accounting relationships distinguish each perennial variety by vintage (age) and tkie period. The stock of each perennial vintage Is a process variable that has a specific yield and specific Input requirements. Thus, the age yield curve of the perennial variety Is embedded In the model parameters. Data on Initial tree stocks by variety, production, crop ylelds, technology, input use and prices are mostly taken from the 700 farm household Survey for Cocoa Producer Pricing (1987), although data from Ghana Cocoa Board, Ghana Ministry of Agriculture, Ghana Statistical Services and persons knowledgeable concerning crop agronomy were also used In estimating parameters related to crops. Typically, sample means and proportions were used as point estknates of population parametars. Mapping of task requirements Into Input requirements by months was done using crop calendars and agronomic consultation. While perennial stock proportions were taken from the survey, estimation - 1.26 - of total stocks required a sequence of Iterative model calibration solutions. ThiS expedient route was required diue to the total absence of the usual cansus type Information on perennlal crops. Land Allocatlon Although Ghana has surplus land, such land Is not kimedlately available for agricultural use. A change In land use to agriculture requires migration of farm families and Investment In Infrastructure, e.g., roads, villagos. utilities to permit local Import of manufactured conswner goods and agricultural inputs and local export of agricultural outputs. The extension of the stock of land under cultivation or In bush fallow tones to proceed In parallel with populatlon and Income increases that tend to Increase population density on the land and create demand for greater agricultural outputs. So long as additional land of comparable quality can be cleared for agricultural use by means of migration of farm families to now lands and investment In Infrastructure, more Intensive use of existing agricultural land will not occur unless a now technology becomes available that Is at least cost competitive with tradItional methods using bush fallow. To date, this has not occurred on a signifhjant scale In Ghana or other West African countries. Thus, at any point In tkne, the stock of land available for agricultural use can be regarded as fixed, though ox or the medium term Increases In population and Incomes will create additinal demands for agricultural output that will result in the conversion of now !and to agricultural use under the traditional bush fallow cycle of Ghanaian agriculture. For each five-year period, the stock of agrlcultural land Is fixed, but between periods the stock of such land Is Increased In proportion to the Increase In population. In the multl-agricultural market equilibrium within a t"me period, annual and perennial crops compete for the existing stock of agricultural land, with allowance being made for the required bush fallow under traditional practico. Note that any land In perennia! crops Is not avallable for other uses until It Is cleared. Skimlarly, land In bush fallow Is not - 1.27 - available for cropping use until It Is cleared. In order to keep track ot use in the five- year periods, the following accounting scheme Is used: posbl and uses: xp cropped wlth perennials f fallow xe croPPed wltn annuals poese land tranactons (at begig of pwod) ap abandoning perennlals xp --3 f aa abandonig annual land xa --> t cl clearng f --2 Xa pi plantng peronnials f -->. xp fg addition to land stock --> f and balnce relations: xp(t+1) - xp(t) - ap(t+l) + pKt+1) ;(t+1) a f(t) + aP(t+l) - cKt+1) + fg(t+1) + aa(t+1) xa(t+1) - xa(t) + cK(t+1) - pl(t+1) - aa(t+1) where the letter t Indicates a use or transactlon In period t. Labor Allocation Crops In Ghana compete for two major inputs, land and labor. Unlike land, labor Is regionally and Internationally mobile. Since Ghana has a rural economy, the labor force Is predominantly rural and agricultural. Given that recent census type statistics on the labor force are not avaiable, the esthiates of labor force size, composition and growth are based on demographic statistbcs and projectlons. In our estleates, the labor force was identified as the class of people 15-44 years old. Given the predominantly rural - 1.28 - population of Ghana, this approxlmation Is sufficient for present purposes. Data from the Survey for Cocoa Producer Pricing Indicate that most of the labor empioyed in agriculture Is family labor. For such labor, there Is no market transaction from which wage Information can be obtalned. However, It Is also clear that family labor Is not willing to work in the fields for Infinitesinal returns. At some point, the disutillity of labor exceeds the utility valuation of the product from an additlonal hour of labor. Therefore, there exists a reservation wage (In the sense of a return In real output) below which the individual will not work. While the reservation wage Is a subjective magnitude and thus specific to Individuals, It will have a mean value across a farm population. Since the mean reservation wage Is not observable, it is a parameter that Is estimated In the process of model callbratlon. The estimate used In our modeling Is 50 percent of the daily wage for casual labor. Since labor Is mobile, Interregional migration Is feasible; and, in fact, It Is observed In Ghana. In our modeling exercise, seasonal migrants between the two regions, north and south, are paid a premium for the cost and disutility of migration. Thus, the representative farm household faces a rising supply curve for labor through the Increasing cost over family labor of regional casual labor ard the premium over the casual wage paid to Interregional migrants. Given that unskilled labor Is an economywide resource, Its wage Is determined in the general equlilbrlum of the economy. Thus, sectorwide modeling should treat the wage of labor parametrically; and this Is the way we handle It. This approach does present a difficulty In a multiperiod model: What will GDP and the %age of labor be In the future? In practice, we have estimated future GDP by assuming a steady but slow annual growth In output per worker (about one percent) combined with labor force estimates based on demographic projections, In this framework, two alternatives for projecting the time path. of the wage of unskilled labor are feasible. One assumes nio growth in real wages, with all of the Increment In real product golng to capital, including service of external debt, - 1.29 - and the other is to assume that some share of the Increment goes to labor, say 50 percent. Both alternatives were simulated. The macroeconomic projections used In calibrating the multiperiod model are given In Tables l.C.1 and analogous projections for an associated annual model are given In Table I.C.2. As already noted, the macroeconomic projections have assumed an elastic supply of arable land and a constant ratio of land to agricultural worker. in addition, the yields for annual and perennial crops are assumed to remain constant. In effect, these assumptions kIply the projection over 50 years of the static technology of traditional agricultural practice using bush fallow. As section 1.8 has shown, this Is the current situatlon In Ghanaian agriculture. With the exc0ptlon of hybrid cocoa, high yielding varietles are not In use on a significant scale. Nor are the fertilizer and pesticide packages used In Intensifying agriculture elsewhere In the world In common use In Ghana or other countries In West Africa. In consequence, even the hybrld cocoa yields are 50 percent lower In Ghana and Cote d'lvoire than In Brazil or Malaysia. Nonetheless, projecting the absence of significant technical change In agriculture over a half century Is a very strong assumption, and most agronomists and plant geneticists would be appalled at such a suggestion. The problem Is not that we expect technical change to remain static, but rather that meaningful crop specific estimates of the course of technical change In Ghanaian agriculture are simply not available. At least by projecting the absence of significant technical change against a backdrop of rapid population Increase, we highilght the resource imbalances that will occur unless agricultural productivity does start to Increase. - 1.30 - A Brif Technical Note The core concepts of the Ghana Multiperiod Agricultural Sector Model (GMASM) are multinarket equilibrium and forward-looking investmont planning. Markets are linked over thne by the investmont activities of farmers and labor supply and demand changes due to population and Income growth. Demand Is shuply specified owing to a lack of recent demand studies for Ghana. Demand functions are Inoar in own price and per capita consumption expenditure, and parameter estkiatlon utilized base case price and quantity observations alng with elasticity esthiates obtained from demand studies of skuilar countries. Supply functions are defined hDlicitIy using Ihear activity analysis. Multhiarket equlilbrium Is sbuAlated by maxknizing the sum of consumers and producers surplus In each thke poriod, using a nonlinear programmuing algorithm. To avoid arbitrary selection of termhal conditions. the sibulation experients were run for a 50-year horizon without terminal stock levels. Inspection of the temporal trajectories of key variables revealed evidence of some boundary effects In the later tkne periods. However, trajectories through 30-year horizons were free from any observed boundary effects. A complete algebraic specification of the multiperiod model Is given In Section l.D. For readers with some knowledge of the Generalized Algebraic Modeling System (GAMS), a complete specification of the model In GAMS notation, inclAding parameter derivation and calibration, Is given In Annex A of this report. - 1.31 - Tabl* I.C.1: Chono Wacroeconomic Projectiono Growth in JPrivate Popu- Labor DP/ Labor Private _ Conoumpton lation Force Worker Productivity GOP |Conjuuption of Per Ccpita Year (000.) (000 ) (°°°O 1975) U/yr (1Os 1975) (1O 3975) GOP (19751) 1965 12,737 6,433 .8425 - 5,420 4,181 .77 328.3 1990 14,974 7,56" .9186 1.74 6.*49 5,212 .7C 346.1 1996 17,506 6,954 .9944 1.60 6,904 6,676 .76 381.5 2000 20,291 10,637 1.0734 1.50 11,416 6,449 74 416.4 2001 23,266 12,633 1.1421 1.23 14,453 10,561 .73 463.S 2010 26,290 14,914 1.2091 1.11 16,033 12,964 .72 493.9 2015 29,309 17,442 1.27" 1.06 22,249 15,797 .71 539.0 2020 32,320 20,101 1.3440 1.01 27,016 16,911 .70 SS5.0 2025 35,215 22,724 1.4125 1.00 32,096 22,449 .70 636.1 2030 37,637 25C.200 1.4 1.00 37,412 26,199 .70 692.1 2les 40,501 27,136 1.560 1.00 42,340 29,626 70 731.3 2040 43,203 29,944 1.6 93 1.00 47,449 33,228 70 769.1 Perceetage Itural Labor Urban _ ; M. turel S. Rural Agrieulture N. Agriculture S Agrieulture Labor Force Force Labor Foreo"/ Labor Forces/ Labor Forcec/ Labor Force/ Labor Forceo/ Year Foreo Rural (000's) (000'o) (000 ') (000'.) (000 o) (000g*) (000*) 1986 70.0 4,03 1,930 1,247 3,266 3,260 903 2,367 1990 67.5 5,107 2,459 1,415 6,692 3,697 1,024 2,673 1995 15.0 5,620 3,134 1,612 4,206 4,214 1,167 3,047 2000 62.5 6,648 3, Ng 1,641 4,607 4,613 1,333 3,480 2005 60.0 7,560 5,053 2,100 5,480 1,466 1,620 3,968 2010 11.0 6,650 6,264 2,396 6,254 6,263 1,735 4,528 2015 56.5 9,655 75607 2,730 7,125 7,135 1,976 5,159 2020 55.0 11,05 9,045 3,063 7,993 6,005 2,217 5,708 2025 53.5 12,157 10,5U7 3,367 8,790 0,602 2,436 6,364 2010 61.0 12,852 12,$40 8,S66 9,292 9,305 2,677 6,720 2035 49.5 13,432 13,704 3,720 9,712 9,725 2,694 7,031 2040 46.0 13,694 15,052 3,649 10,045 10,059 2,786 7,273 o/ Computed an 27.7% of rural/agriculture labor force. b/ Computed as 72.3Z of rural/ogriculture labor force. c/ Computed a. 72.4X of rural labor force. - i.a2 - S oL..1L AC DLsaaregation of Macroeconomic Projections. 1964-95 (000'r) (000's 675) Labor (106¢75) (675) (000'r) Labor GDr/ Prod. (10¢75) ri. 2 Pri. Cons. Year Populatioa Force Worker rarowLb/Yr GDP Cons. GOP Per Capita 1964 12,323 6.228 .$2$2 5,158 3,917 .759 317.9 1985 12,737 6,433 .8425 1.73 5,420 4,161 .771 328.3 1986 13,095 6,645 .8581 1.85 5.702 4,321 .756 330.0 1967 13,506 6,664 .8706 1.46 5.976 4.269 .718 317.5 1988 13,900 7,091 8884 2.05 6.300 4,725 .750 338.0 1989 14,468 7,325 .9036 1.70 6,619 4,964 .750 343.1 1990 14,974 7,566 .9185 1.65 6,949 5,212 .750 346.1 l991 15,449 7.825 .9332 1.60 7,302 5,476 .750 354.5 1992 15,940 8,093 .9461 1.60 7,673 5.755 .750 361.0 1993 16,446 5,371 .9633 1.60 8.064 6,048 .750 367.7 1994 16947 6,657 .9767 1.60 6,473 6,355 .750 374.6 1995 17,506 6,954 .9944 1.60 6,904 6,678 .750 381.5 1996 18,031 9.268 1.0101 1.56 9.362 7,003 .748 388.4 1997 18,571 9,593 1.0259 1.56 9,841 7,341 .746 395.3 (000o c) Rurcl LI (000 s) (000'.) (000'° ) (000"' ) (000") (000's) Rural as I AMr Urbn N. Rural S. Rural X.Agr SA&r Year L.F. Total L. F. / L.F. L.F. !' L b/ .P.L V L.F b/ 1984 4,397 70.6 3,183 1,831 1,218 3,179 882 2.301 1985 4,503 70.0 3,260 1.930 1,247 3,256 903 2,357 1986 4,618 69.5 3,343 2.027 1,279 3,339 926 2,417 1967 4.736 69.0 3,429 2,128 1,312 3,424 950 2,479 1906 4,857 68.5 3,516 2,234 1,345 3.512 974 2,542 1989 4,981 68.0 3,806 2,344 1,380 3,601 999 2,607 1990 5,107 67.5 3,697 2.459 1.415 3,692 1,024 2,673 1991 5,243 67.0 3,796 2,582 1,452 3,791 1,051 2,745 1992 5,382 66.5 3.897 2,711 1,491 3,891 1,079 2,818 1993 5.525 66.0 4,000 2,646 1,530 3,995 1,108 2,692 1994 5,670 65.5 4*105 2,967 1,571 4,099 1,137 2.966 1995 5,620 65.0 4,214 3,134 1,612 4,206 1.167 3,047 1996 5,976 64.5 4,326 3,290 1,656 4,322 1.196 3 129 1997 6,140 64.0 4,445 3,453 1.701 4,439 1,231 3.214 a' Computed as 27.72 of rural/agr labor force g/ Com puted as 72.32 of rural/&gr labor force - 1.33 - Since investment In cocoa production is long-term investment, the appropriate model for analyzing farmer Investment response to cocoa pricing and other government interventions Is a multiperiod, long horizon framework. However, many pollcymakers have difficulty thinking In terms of a sequence of five-year perbods; and quite often the focus of policy concern Is the near future. For these reasons, an associatea annual model for short- to medium-term projections has been developed to assist In linking the multiperlod model to current pollcymaker concerns. This annual model takes the perennial investment projections of the multiperlod model as given. Thus, the annual model decomposes five-year Investment projections Into annual projectlons by means of a smoothing procedure and then solves the annual model for the factor allocations, commodity outputs and prices for annual crop Investments In a given year. Model Validation Using the Annual Decomposition Model The first application of the annual decomposition model Is to the recent past In order to compare model solution values with some observed annual quantities. However, In making such comparisons It should be kept In mind that the anniual model solutions are based on Investment results from the multiperiod model, which outputs solution values for five-year periods. Hence, both models necessarily abstract from short-run fluctuations due to the weather, strikes, natural disasters, etc. Rather, these models assess producer response over time to pollcy Intervention. Since farmer supply response In Ghana Is mainly characterized by allocation of land and labor to various crops, this allocation Is given In Table 1.C.3 for the validatlon period of 1984-87. Deferring discussion of cocoa production estkiates for the moment, we turn to allocation of land ar,d labor. Labor use Is estimated to Increase from 477 million mandays In 1984 to 563 million mandays In 1987, a gain of 18 percent over three years, or 5.7 percent per year. While this increase also reflects growth In the agricultural labor force, up 7.7 percent over the period, most of the growth Is - 1.34 - Taft LC.3& Ajui D0oef Otf lUrtOd Mod Valdstton Perod C cooa Acgs abor Uufon Labor Sources Prod'n (000GAcius (MWonManda) (MiNorsMarndays) You (000C MT) Trad. Hyblid Annuul Pown. Total Famnily Casal 198445 226.9 994.8 71.6 266.1 2107 476.6 470.2 6.6 1906861 227.9 670.6 637.6 271.4 239.3 510.7 499.0 11.7 1966.7 228.9 660.5 1066.0 279.4 262.3 541.7 526.0 1s.7 1967.66 230.0 846.4 1264.2 251.5 262.0 563.5 544.9 18.6 Crtp Acrae Farm kcome (00W AncMe (ehllndof 1067 Cedls) Y A nnAua Peren. C Iet 1 CAXOCWe 2 196485 4000 2665 164.3 209.3 19854.6 4700 3201 174.2 223.4 1966-67 4640 3534 197.0 230.2 19874.6 4662 3639 171.8 227.9 due to hproved icentives for cocoa productlon and investment. Of the 86 million manday increase in labor utilization over 1984-87, 71 million mandays wore for work on perennial crops, mostly cocoa. Only relative mnor amounts of casual labor were used, with family labor accounting for 97 percent or better of labor hput each year. The resurgence of peronnial crops, ospecially cocoa, Is also shown In the estkinates of acreages in annual and perennial crops, with annual acreage Increasing by 280,000 and perennial by 930,000 acres, of which 630,000 are In cocoa. The Increase in future cocoa production potential Is even greater since all of the Increased acreage (and more) Is In hybrid cocoa, which enjoys a 50 percont yield differential over traditional varlties. As we shall see, the reolacement of acreage In traditional varieties by planting of hybrld varieties Is desthnd to be an heprtant feature of cocoa Investments over the next decade. This transition has already occurred In the other major cocoa producing countries. - 1.35 - Estimates for two concepts of farm Income are also presented. Concept 1 is utility based in that the cost of labor Is netted from Income, most of wnich Is the estimated disutility of labor by family members. This Is the Income concept that enters Into the multiperlod decislons fo farmers in model solutions, and It Is the concept used for the Income estimates from the solutions of that model. Concept 2 Is more conventional In that labor costs are not netted out. This concept Is practically Identical to the value added concept used In measuring agricultural gross domestic product (GDP). At this point, it Is useful to note that of the osthuates presented In Table 1.C.3, only the cocoa production and family Income (Concept 2) esthuates can be chocked against reliable, Independent estimates, although cocoa acreage can be checked against the Informal estimates of the COCOBOD. Clearly, the absence of census and survey data that documert the labor force and Its utilization, as well as an agricultural census that Integrates annual and perennial crops has created an information void concerning labor and land utilization that should be corrected. Turning to the cocoa production estbmates for which validation Is possible, Table l.C.4 presents the relevant comparison. Clearly, simulated production Is acceptably close to actual productlon In 1985 and 1986. Note, however, that both 1984 and 1987 were years marked by exceptional weather that depressed cocoa production. Moreover, the cocoa buying price was still quite iow In real terms In 1984, provIding little Incentive for a complete picking of the cocoa on the trees. As already noted, the model cannot be expected to pick up the effect of short-run weather fluctuatlons. In fact, the simulated level of production shown, with annual cocoa production approxkmately constant over the perlod, Is precisely what should be expected given an IncreasIng price Incentive but a history such that cocoa Investments five to ten years previously were quite low due to weak Investment Incentives then. - 1.38 - Table LC.4: Cot rlson of Sbuated and Actual Cocoa Production Coooa Producion Cocoa Buying Price (000. Metric Tons (Cedis/kg)* Simulated Currnt 1987 Yer Simulated Actua /Actual Cedis Cedis 1964-85 227 175 52 30.0 54.2 1968.6 228 219 9 56.6 93.9 1968467 229 228 1 85.5 116.8 1967488 230 183 47 140.0 140.0 Source: The WoMd Bank, Ghan Country Economic Meorandum, Report No. 7515- 64 (Green Cover), November3, 1968, Statistical Appendix: Tabes 7.01, 9.01. To compare farm Income estimates (Concept 2) from model solutions with estimates from the natlonal accounts, some adjustments to national accounts data are necessary. These are shown In l.C.5. The national accounts sectoral GDP estimates separate cocoa production and marketing from other agriculture and livestock, as shown (in current cedis) In rows 1 and 2 of Table l.C.8. To calculate the amount of cocoa GDP that accrues as Income to farmers, cocoa GDP (row 2) Is multiplied by the farmer's share of cocoa export revenue (row 3). yielding the desired amount (row 4). Estimated farm Income (row 5) Is then the sum of rows 1 and 4. The estimates In current cedis are then deflated to constant 1987 cedis, using deflators based on the average rural CPI for each year (row 6) to obtain farm Income (in 1987 cedis) In row 7. The GDP based estimates are then computed as percentages of the model estimates (row 9). Note that the exclusIon of Ilvestock and some high vaiued minor crops from the model Implies that the GDP based estimates should bo greater than the model estimates by perhaps 10 to 20 percent. Thus, the model estinates for 1984 are clearly too high by at least 10 percent; and the model estimate for 1987 Is low by perhaps 10 percent. However, given the weakness of the data base, and hence the strong assumptions that had to be made to - 1.37 - Table LC.5: Calculation of Farm hkcoo (Concept 2) from the National Accounts rTEM 1984 1985 1986 1987 1. GDP, Agriculture & Uvestock (billions current cedis) 110.4 118.7 175.6 266.1 2. GDP, Cocoa Production & Marketng (billions current cedis) 11.1 18.8 41.0 66.0 3. Cocoa Farmers' Share, Cocoa Export Revenue .311 .320 .408 .45 (est) 4. Cocoa GOP Paid to Farmnrs (Row 2 x Row 3) 3.5 6.0 16.7 29.7 5. Estmated Farm Income (Row I + Row 4) 113.9 124.7 192.3 295.8 6. Rural CPI, Annual Average (1987=-100) 55.4 60.3 73.2 100.0 7. Estmated Real Farrn Income (billions 1987 cedis) 205.6 206.8 262.7 295.8 8. Model Estimate of Farm Income 209.3 223.4 230.2 227.9 9. (GOP-Based Estmate/Model Estimate) x 100 98.2 92.6 114.1 129.8 Source: The World Bank, 'Ghana Country Economic Memorandum, Report No.7515-GH-, (Green Cover), November 23, 1988, Statistical Appendix: Tabes 2.02, 7.01, 9.01. generate both estimates of farm Income, the relative closeness of the two sets of farm Income estimates Is reassuring. Finally, we note that the estimates of cocoa acreage In 1987 given In Table l.C.3 Imply total cocoa acreage of about 2,100,000 acres, or approximately 850,000 hectares. These estimates were derived first from the sample totals (by age class) of cocoa acreage obtained from the 1988 Agricultural Economics Survey for Cocoa Producer Pricing (by COCOBOD) and then projected to the natlonal level In the calibration of the Multiperiod Agricultural Sector Model. The exact calibratlon procedure Is given In Annex A, which lists the GAMS specificatlon for the model. The official COCOBOD estimate of cocoa acreage Is one millon hectares. There Is no precise way of resolving the difference between these two estimates of about 150,000 hectares short of a detailed census. However, we note that if the model estimate Is low, then model yield estimates (by tree ages) must also be too low. Again, only detailed measurements of cocoa yields (by age class) can answer this questlon. - 1.38 - 2. Policy Simulatkns In order to assess cocoa producer response to price Incentives In a dynamic context, several different sconarios are required since producer response is also a function of the available supplies of the inputs needed for cocoa investment and productlon, as well as the state of the Ghanaian economv and other policies of the Goverrvnent of Ghana. in general, the continued bmpiementation of the policy measures integral to the structural adjustment program Is taken as given. However, the progress of the Ghanalan economy Is stll subject to external economic and climatic influences as the history of the recent decade attests. Therefore, the following sequence of sconarios was simulated: 1. The effects of price incentives, given zero growth in population and per capita icome. 2. The effects of prico incentives, given population growth, but no growth in per capita income. 3. The effects of price incentives, given population increase and relatively slow growth in r capita income with some agricultural protection. 4. The effects of price incentive, given growth In populatlon and per capital income and a policy of food self sufficiency. Scenarlo 1 Is included skmply to assess the effects of price incentives without the complications of growth In population and per capita Income. Scenar5o 2 introduces the affects of population growth alone, although with a proportionate Increase in land available for cultivation (as discussed above). This scenario essentially projects a continuation of the lack of growth In per capita Income of the past two decades, but without the distorted price incentives. It also Inpiles Investment in Infrastructure sufficient to extent the margin of cultivation proportionately with population growth. Scenario 3 Introduces the effects of slow growth In per capita income. As already noted, this scenario has two variants: one with no growth In the wage of unskilled labor, and another In which unskilled labor receives a 50 percent share of the Increment in per - l.39 - worker product. Sconario 4 consilers the effects of a policy of food self sufficiency on cocoa supply. Scenario 1 - Supply Response Under Fixed Resource Supplies This Is a reference case, not a realistic sconario; however, It strips the situation to Its essentials, the use of land and labor In the production of cocoa, other perennials crops and annual crops. Except for the export crop, cocoa, all other crops are produced to supply domestic demand; and domestic supply and demand are equilibrated by an endogenous domestic price. The crops maize and rice, which compoet with Imports, are protected by a 100 percent tariff; and thol pricos are ondogenous only when less than twice the world price. The cocoa supply response for the fixed endowment case Is presented In Figure l.C.1 for a set of prico experW nts over a horizon of 50 years. Note that the Initial conditions are the same for each experiment--the stock of cocoa acreage by vintage and variety, cocoa production and the Ghana Cocoa Board buying price at the midpoint of the years 1985-89. Measuring the surply response as the average of the five-year average supplies over horizons of 20, 30 and 50 years and computing arc elasticities of supply yields tho estimates for this scenario presented in Table l.C.6. Clearly, the resporse Is Inelastic Irrespective of horizon except for the 120-140 cedls price Interval; and the estimates are consistent In the sense that they do not exhibit a large coefficient of variation. The clear implication Is that reducing the buying prico below 140 1987 cedis will significantly reduce supply, at least In a context of fixed resources. Note the 30-year cycle In cocoa supply response which corresponds with the productive life of a cocoa tree. Since the tial conditons for cocoa tree stocks are relatively iow for historical reasons, heavy plantig of now trees Is profitable in the early periods when a buying prico that at least covers cost prevails, In additlon, It Is profitabie to scrap the stocks of traditional varieties in the hintlal periods since these are dominated by the hybrid varieties, In this respect, Ghana Is still going through the - 1.40 - Flgure I.C.1 COCOA SUPPLY RESPONSE UNDER FIXED RESOURCES 700- I -s0o I .1 I I' -4 Suoo / 400 / 200 - 1987 1997 2007 2017 2027 2037 PERIOD MIDPOINTS 0 10 * Mo o 30 £ 0 X - 1.41 - Table LC.S: Arc Elasticiltes of Cocoa Supply Scenario & Cocoa BuyIng Price Intervals Horizon 120-140 140-160 160-180 180-200 hbed Paourm 20 years 1.573 0.584 0.472 0.448 30 years 1.213 0.441 0 461 0 468 50 yeaus 1.149 0.537 0.501 0.481 Averag 1.312 0.521 0.478 0.456 Paon 20 years 0.973 0.439 0.264 0.274 30 yer 0.775 0.414 0.285 0.328 50 years 0.916 0.355 0.286 0.315 Aveage 0.888 0.403 0.279 0.306 Pcpusaion and Ioom Growlh Cuas A: Conudant Wage 2sr yeus 1.746 0.714 0.392 0.373 30 years 1.986 0.913 0.579 0.324 50 years 2.634 1.616 0.786 0.476 Average 2.122 1.061 0.586 0.391 Case B: ocreasing Wage 20 yura 2.043 0.815 0.429 0.436 30 yeoas 2.567 1.042 0.690 0.366 50 years 3.493 2.038 1.056 0.658 Averwa 2.701 1.298 0.725 0.487 qiA In Populon & hoame Vf Zo Food h 20 years 0.043 2.019 1.606 0.562 30 yas 0.042 3.059 1.931 0.710 50 years 0.041 4.286 3.106 1.087 Aveagoe 0.042 3.121 2.215 0.786 transition to hybrid varieties that other producing countrles have essentially completed. Thus, supply roughly reproduces the shape of the age yield curve for hybrid cocoa trees In this scenario with fixed supplies of laind and labor. The long gestatlon period of cocoa Investments converts the supply response to price Into a stock adjustment problem skiilar to the stock adjustment of any long-lived capital asset when relative profitability changes. Flgure l.C.2 presents the sequence of stock levels of cocoa acreage defining the adjustment path for each price experklent. Note that acreage stocks converge fairly rapidly (I.e., one to two periods) to the nelghborhood of the desired level, but further adjustments are necessary due to the unbalanced nature of the age composition of Initial stocks. Not surprisingly, acreage stocks show much less variation than cocoa production, which is constrained by the age compositlon of stocks. - 1.42 - Figure I.C.2 COCOA ACREAGE ADJUSTMENT TO PRICE UNDER FIXED RESOURCES 3.8 3.2 - / \ _ _ _ _ _ _ _ _ _ _ _ AK1//N~~~~~~~~~~~~ \\ =\ Ij 2.6- 4167 1907 2007 2017 2027 20:37 PERIOD MIDPOINTS 0 120 * 1 o g _0 q ~ - L43 - FlVU l.C3 FARM INCOME UNDER FIXED RESOURCES (oftI twof SW CeiIe} 240 no am 300 o 30 300 no go - 3002~a 2032 o 30 + No0 o 30MO~ 0 - 1.44 - The anomalous acreage adjustment path under the 120 cedis per kilogram buying price reflects the marginal profitabillty of cocoa at this price. Only limlted amounts of resources can be bid away from other crops, requiring slower adjustment. The sharp rise In acreage at this price (in the 2015-19 perlod) reflects holding trees In production longer than Is profitable at hliher prices. Sknulated varlatlon In total farm Income under a fixed resources scenario Is given In Figure i.C.3. Farm income Is defined as gross revenue (including farm level consumption valued at the endogenous market price) less purchased Input costs (including family labor valued at the reservatlon wage). In general, variation In farm income reproduces the pattern of cocoa supply response (at a constant price). Note the relative constancy of farm Income at the marginally profitable buying price of 120 cedis per kilogram in contrast to the divergent patterns of variatlon in cocoa supply and cocoa acreage at this price. These adjustments were necessary to roughly stabilize Income. Patterns of land and labor use for a representative period (the years 2015-19) are shown In Figures 1 .C.4 and 1 .C.5, respectively. In contrast to cocoa output and farm Income, the allocation of Inputs varles smoothly across prices and periods. Since total land Is fixed under a given scenario, and a proportion of annual land must be held In bush fallow, only the allocatlon between annual and perennial crops contains any Information. The acreage In perennlals Increases almost linearly with Increasing cocoa buying prices (at roughly 100,000 acrea per 20-cedi increase in the price per kilogram). In contrast to land, at least seasonally unemployed labor exists for all solutions. Since total employment Is overwhelningly In family labor (never less than 97 percent), only total employment of ik - *6 % 9 Y family labor and the allocation of total employment between annuals and perennials are shown In Figure l.C.5. More complete data on sknulated labor employment Is given In Table 1 .C.4. As with land, labor al;ocatlon changes smoothly with both total family labor and allocatlon to perennials Increasing with Increases In the cocoa buying price. 1 .45 - Figure 1.C.4 LAND USE UNDER FIXED RESOURCES, 2015-19 (MiLLior of Acres) 4.5 4.0 3.5 3.0 2.5 2.0 1.5 0.5 120 140 160 180 200 Cocoa Price (1987 Ccdib/Kgi a Perennial + Annuol 0 Follow 0~~~~~~~~ 4c~~~~~~~ - 1.47 - Total empoyment Increases by about three pwrcont as the cocoa price Increases from 120 to 200 cedis per klbgram, reflecting the greater labor Intensity of cocoa production. Scenario 2 - Effects of Population Growth on Supply Response This sconaro assesses the effect of parl Passu growth In population and cultivable land. The population projection, which was prepared by Rodolfo Bulatao of the Population and Human Resources Department of the World Bank, inplies declining fertility from the present rather high levobe. As noted before, the assoclated labor force estimates are defind as the class aged 15 to through 64 years. Both sets of estinates are givon in Table l.C.1. The cocoa supply responses generated by this modification of the reference case are presented in Figure l.C.6 for the same set of price experkments. Note the significant werease over the reforence case valiu for cocoa production In all of the exprints. This change Is due solely to a more elastic labor supply over time. Measurin supply response as before and computing arc elasticities of supply, we obtain the estimates presented In Table l.C. for this sconario. These estimates are qualitatively similar to those for the reference case. They Indicate an Inelastic supply response except possibly for the 120-140 codis Interval. Thus, the supply elasticities for this scenario also imply that reducing the buying price below 140 1987 cedis will significantly reduce supply. With growth In population (and associated Increases in land in use sufficient to keep the land-man ratio constant), the pattern of cocoa production over time changes from the aimost pure cycles of the reference case to cycles with upward steps. In this t C sconaro, as with the reforence case, It Is the Initial condition of relatively low cocoa production and acreage that creates the cycle by nduchg heavy planting (at profitable pricos) In the early piod. This pattern of hvistment combinod with the productive life cycle of the cocoa tree and an incoasig endowment of labor (and effective land) results In the pattern of stepped cycles, in contrast, the adjustments of cocoa from - 1.48 - Flgure 1.C.6 COCOA SUPPLY RESPONSE UNDER POPULATION GROWTH 1.2- '.5 0.4 0.37 0.2 1967 1997 2007 2017 2027 2037 PERI;O MIDPOINTS 0 m *+ 140 o 20 0 00 - 1.49 - Table LC.7: Sujlated Aruul Pattern of Labor Uso hI 2015-19 Period (mulons of man-days) USES SOURCES Cocoa Producer Peren. Annual Family Casual Under Scenario Price Crops Crops Total Labor Labor Employ. Fixed 120 354.1 256.1 610.2 597.7 12.5 808.8 Resources 140 358.2 251.2 609.4 599.1 10.3 809.7 160 364.3 247.9 612.2 601.2 11.0 806.8 180 378.1 242.9 621.0 607.4 13.6 798.0 200 392.8 237.8 630.6 614.2 16.4 788.4 Population Growth 120 687.5 603.5 1291.0 1277.8 13.2 1812.4 Only 140 789.4 563.4 1352.8 1322.7 30 1 1750.5 160 810.1 552.3 1362.4 1330.3 32.1 1740.9 180 839.6 540.6 1380.2 1343.6 36.6 1723.2 200 868.0 529.6 1397.6 1356.7 40.9 1705.7 Population & Income 120 330.4 738.4 1068.8 1068.8 00 2034.5 Growth 140 488.7 679.5 1168.2 1168.2 00 1935.2 160 628.8 630.1 1258.9 1254.9 4.0 1844.5 180 720.1 581.8 1301.9 1288.6 13.3 1801.4 200 769.4 563.8 1333.2 1312.2 21.0 1770.1 Population, come & 120 243.4 779.5 1022.9 1022.9 0.0 2080.4 Wags Growthi 140 432.4 699.0 1131.5 1131.5 0.0 1971.9 160 577.9 647.4 1225.3 1225.3 0.0 1878.1 180 662.8 606.4 1269.2 1263.4 5.8 1834.2 200 749.1 571.8 1320.9 1303.4 17.5 1782.5 Food 120 215.7 814.6 1030.3 1030.3 0.0 2073.0 SelffSufficiency 140 215.8 814.5 1030.3 1030.3 0.0 2073.0 160 380.7 743.5 1124.2 1124.2 0.0 1979.2 180 621.4 640.8 1262.2 1262.2 0.0 1841.2 200 717 1 599.4 1316.5 1303 4 13.1 1786.9 acreage (I.e., investment and disinvestment In cocoa trees) at each buying price shown In Figure l.C.7 reduce to almost a pure step functlon since these do not have the age- yield curve of cocoa trees Influencing their form. The variation In farm income at various buying prices under this scenario, shown In Figure l.C.8, also has a rough step function form due to the Induced pattern of cocoa Investment and production and the growing avallabiiity of labor (and associated land) resources. The pattern of land use in response * varlation In cocoa bsylng price, nown In Figure I.C.9 for a representative period (the years 2015-19), now exhibits a sharp nonlinearity reflecting sharp the Increase In profitabiiity of cocoa production between prices of 120 and 140 cedis per kliogram. The pattern of labor use, shown In Figure I.C.10, exhibits this nonlinearity In attenuated form. The Increase In total labor employment between prices of 120 and 200 cedis per kilogram - 1.50 - Figure I.C.7 COCOA ACREAGE ADJUSTMENT TO PRICE UNDER POPULATION GROWTH a. 3 - s 2~~~~~~~~~~~~# / / / "_ _ 2 1- .4 - .. 1987 1997 2007 2017 2027 2037 PERIOD MIDPOINTS o o + f140 ISO I 3 0 *~~~~~~ , . - 1.51 - Figure I.C.8 FARM INCOME UNDER POPULATION GROWTH (Billions of 1987 Csdis) 700 500 so~~~~~~~~~~~~ / 300 e 200- 1992 2002 2012 2022 2032 PERIOD WLDPOINTS 0 120 + 140 o 160 180 = 200 - 1.52 - Figure 1.C.9 LAND USE UNDER POPULATION GROWTH, 2015-19 (Miltiom of Acres) 10.0 9.0 8.0 7.0 4.0 3.0 2.0 1.0- 120 140 160 160 200 Cocoa Price 11987 Cede/Kg) 0 Perennai + Annuol 0 Foilow - 1.53 - Figure I.C.I0 LABOR USES UNDER POPULATION GROWTH, 2015-19 (littions of Nmn-Do") 1.4- to 0.9 - 0.8 - 0.7 0.6 0.5_ 120 140 160 180 200 Cocoa Price (1987 C.&/Kg) 3 POuUAL + ANUAL FAMIY LABOR - 1.54 - is now about 8 percent. The significant increase in cocoa production under this scenario also reflects Increasing reliance on food grain Imports that permit more specialization in the comparatively advantageous cocoa crop. The simulated levels of imports are given In Table 1'.C.8. They rise to levels of 1.0 to 1.3 million tons per year under all of the simulated buying prices, In contrast to simulated Imports of only 0.4 million tons under all buying prices for the reference case. Scenario 3 - Effects of Growth In Per Capita Income on Supply This scenarlo adds growth In per capita Income to the set of conditions that affect cocoa supply response. The income growth stems from an assumed growth In GDP per worker that averages 1.3 percent over the 50 year horizon. Estimates of this productivity growth, Its use in calculating estimated GDP and the assoclated growth In per capita GDP are shown In Tables l.C.1 and l.C.2. Note the full employment assumption implied by the calculation of GDP. A growth rate In output per worker of 1.3 percent Is not very large by International standards, but It represents a significant break with Ghanaian experienco over recent decades. The derived growth In per capita income averages 1.6 percent over the 50-year horizon owing to the Increasing labor force participation rate that follows from the declining fertility assumption. This phenomenon Is generally associated with declining fertility In the process of development. This scenario also distingulshes two cases--a constant wage case In which all of the Increment in domestic product accrues as Income to capital (including external creditors), and a real wage growth case in which labor receives half of the Increment In product. We examine first the constant real wage case. Cocoa supply responses for this case are given In Figure I.C.11 for the standard set of price experiments. It Is clear that the addition of income growth has drastically reduced the cocoa supply response at a given price, and this Is to be expected since the additional Income causes growth In food demand that raises food prices and Induces farmers to shift land from cocoa to food crops. The arc elasticities - 1.5S - Table LC.8: Food Grahi ports ('000 metric torwm) L Mai Upor ndar FRed P_owcu COCOA PRODUCER PRICE (1987 CEDIS/KG) Period 120 140 160 180 200 1990.94 206.1 444.1 444.1 444.1 444.1 199S99 44.1 44.1 444.1 444.1 444.1 2000-04 444.1 444.1 444.1 444.1 444.1 2005-09 4441 444.1 444.1 444.1 444.1 2010-14 444.1 444.1 444.1 444.1 444.1 2015-19 444.1 444.1 444.1 444 1 U4.1 2020-24 444.1 444.1 444.1 444.1 444.1 2025-29 444.1 444.1 444.1 444.1 444.1 2030-34 444.1 444.1 444.1 444.1 444.1 2035-39 319.5 444.1 444.1 444.1 444.1 L Maie Undwr P_xAUon rwh COCOA PRODUCER PRICE (1987 CEDIS/KG) Period 120 140 160 180 200 1990-94 318.0 522.2 522.2 522.2 522.2 1995-99 610.1 610.1 610.1 610.1 610.1 2000.04 707.4 707.4 707.4 707.4 70- 2005409 811.3 811.3 811.3 811.3 3 2010-14 916.5 916.5 916.5 916.5 s 6 2015-19 1021.8 1021.8 1021.8 1021.8 10 1 2020-24 1127.0 1127.0 1127.0 1127.0 112t.0 2025,'9 1091.4 1227.8 1227.8 1227.8 1227.8 2030-4 1080.7 1319.3 1319.3 1319.3 1319.3 2035-39 934.5 1412.1 1412.1 1412.1 1412.1 1 Maie and Rio h.pog Undw PopuIafon wnd cme Gro*,M COCOA PRODUCER PRICE (1987 CEDIS/KG) Peliod Grain 120 140 160 180 200 1990-94 Maize 312.3 534.7 534.7 534.7 534.7 1995-99 Maize 649.7 649.7 649.7 649.7 649.7 200004 Maize 788.9 788.9 788.9 788.9 788.9 2005-09 Maize 942.4 942.4 942 4 942.4 942.4 2010-14 Maize 1110.5 1110.5 1110.5 1110.5 1110.5 2015-19 Maize 1295.2 1295.2 1295.2 1295.2 1295.2 Rice 3.3 2020-24 Maize 1491.7 1491.7 1491.7 1491.7 1491.7 Rice 81.9 81.9 81.9 2025-29 Maize 1706.2 170G.< 1706.2 1706.2 1706.2 Rice 183.7 215.6 215.6 i 2030434 Maize 600.2 988.4 1393.7 1393.7 1393.7 2035-39 Maize 2111.6 2125.5 2125.5 2125.5 2125.5 Rice 169.6 426.4 - 1.56 - IV. Main aid Nos MOM Lkdw Papuluiwi. v & Wag O wt CO;OA PRODUCER PRICE (1987 CEDIS/KG) Period Grain 120 140 160 180 200 1990-94 Maize 284.8 534.7 534.7 534.7 534.7 1995-99 Maize 649.7 649.7 649.7 649 7 649.7 2000-04 Maize 788.9 788.9 788.9 788.9 788.9 2005-09 Maize 942.4 942.4 942.4 942.4 942.4 2010-14 Maize 1110.5 1110.5 1110.5 1110.5 1110.5 2015-19 Maize 1158.0 1295.2 1295.2 1295.2 1295.2 Rios 45. 2020-24 Maize 1491.7 1491.7 1491 7 1491.7 1491.7 Rice 83.5 83.5 2025-29 Maize 1706.2 1706.2 1706.2 1706.2 1706.2 Rice 217.6 217.6 2030-34 Maize 450.3 771.3 1316.8 1393.7 1393.7 2035-39 Maize 1938.4 2125.5 2125.5 2125.5 2125.5 Rice 354.5 implied by this supply response are giver, In Table I.C.6 and these exhibit the same pattern as previous scenarios, elastic at low prices and inelastic at higher prices. The simulated patterns of cocoa production over time for various buying prices under this scenario are In general quite similar to the stepped cycles of the population growth only scenario. However, In this sconarlo, the lowest buying price of 120. cedis per kilogram has cocoa production collapsing to qulte low levels as other crops are more attractive. That is, the growth In Income stimulates demand for foods, raising the prices of nontradable foods, and shift'ng resources away from cocoa production. In addition, the stimulus to food demand from Income growth boosts imports of food grains to levels that reach 1.5 to 2.5 million tons per year In the later time periods. The simulated patterns of cocoa acreage adjustment under this scenarlo, shown In Flgure I.C.12, show the rough step function of the previous scenario at the hlgher prices. At the 120 cedis per klogram price, acreage collapses to under one mililon acres. At a price of 140 cedis per kilogram, acreage Just malntains its Initial level of silghtly over two millon acres. Thus, the price beiow which demand becomes elastic becomes 160 cedis per kilog-am under this scenario. The simulated levels of farm Income in the Income growth, constant wage scenario, shown In Fiqure I.C.1 3, follow the rough step function of - 1.57 - Flgure 1.C.11 COCOA SUPPLY RESPONSE UNDER POPULATION AND INCOME GROWTH (Constant Wag) 1.5 14 - 12 0. o 14 0.5 0.7~~~7/ 0.3 0.2 0.1-19207- 198769 07 2017 2027 2037 PERIOD MIDPOINTS 12~0 * 140 o IS0 IGO AK 20 - 1.58 - Figure I.C.12 COCOA ACREAGE ADJUSTMENT POPULATION AND INCOME GROWTH (Comtant Ws) 6- 3 *=37 1997 2007 2017 2027 2037 PERIOD MIDOINTS o W2 + 140 o 0 IG a IG a _ 1.59 - Flgure 1.C.13 FARM INCOME UNDER POPULATION AND INCOME GROWTH (lilittfa of 1987 Codis) 900 no 6 7W -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~7 400- 300 -~\ ,/~~~~" ', *'; 200- Wti/0=0ofZ l~~~~~~~~~~~~~~~ _ iSi92 2002 2012 2022 2032 PERIOD MIDPOINTS a 120 + 140 o 360 I O A 2X - 1.60 - the population growth only scenario up to the last two periods. These last two show a dip and then a rise that Is due to a boundary effect from the fixed horizon--adjustment of perennial acreages produces this effect. Patterns of land and labor use for this scenario, shown in Figures 1 .C.14 and 1.C.15, respectively, both show Increased allocations of resources to annual crops at all prices, as would be expected. In consequences, the growth In agricultural employment going from a cocoa buying price of 1 20 cedis to one of 200 cedis per kilogram Is 25 percent. This Is a dramatic illustration of the Income redistributlonal effects of Increasing cocoa prices. The Increased pressure on agricultural resources raises prices of nontradable foods, labor wages and land rents. These changes raise the incomes of the poorer households, small farmers producing nontradable foods and landless laborers at the expense of more affluent urban consumers. The patterns of cocoa supply response and acreage adjustment under the Income growth, Increasing wage scenario, given In Flgures 1.C.16 and 1.C.17, follow those of the constant wage scenario, except that levels are everywhere reduced due to higher labor costs. The same can be said for the patterns of farm Income, land use and labor presented in Figures 1.C.18, 1.C.19 and 1.C.20. - 1.61 - Figure I.C.14 LAND USE UNDER POPULATION AND INCOME GROWTH, 2015-19 (MiLlions of Acres) 13.0 7.0 Ito 10.0 9.0 3.0 6.0( 4.00 2.0 120 140 160 180 200 Cocoa Price (1987 Cedis/Kgi O Perennici + Annual ° Follow - 1.62 - Fisure 1.C.15 LABOR USES UNDER POPULATION AND INCOME GROWTH, 2015-19 (liLliam of aN-Dwa) 1.4- 1.3 -_ L2 - Li ' LO l 0.8 0.7 0.5 0.4 120 140 160 10 200 Cocoa Price (1987 Cod&/Kgi o PERIAL + ANUAL FAMLY LAS E . - 1.63 - Figure 1.C.16 COCOA SUPPLY RESPONSE UNDER POPULATION AND INCOME GROWTH (Increasing Wage) 1.4 - 13 - 1.2- 1I" 0.9 ' 0.4- 0.3- 0.2 0.1 1987 1997 2007 2017 2027 2037 PERIOD MNDFOINTS 0 1 + 140 o I0 a 2O !.64 - Figure 1.C.17 COCOA ACREAGE ADJUSTMENT. POPULATION AND INCOME GROWTH (Incrosing WVga) 4 2 -J I I187 1997 2007 2017 2027 2037 PERIOD MIDPOINTS 0 120 + 140 o I0 I O 2 . - I.B5 - Figure 1.C.18 FARM INCOME UNDER POPULATION. INCOME AND WAGE GROWTH (ItLtlfao of 196? CedIs) 900 700 - 600- 300 \\ 200 1992 2002 2012 2022 2032 PERIOD MDPOINTS 0 0 + 140 o mI I SO - Xe - Uies I.C.19 LAND US UNDER POPULATION. INCOME AND WAGE GROWTH. 2015-19 Vitt, of Aem sdo - ILO - ~ ~ ~ ~ ~ MII Lgs n.0 VW - 10.0- 7.0 &o - 4.0- 3.0 2.0- 120a0 s0o IO 200 C Mm GM? C.4F/Kgl 0 ~~.ronnhoI * ~~Annual 0 Follow - l.67 - Figure 1.C.20 LABOR USES UNDER POPULATION. INCOME AND WAGE GROWTH, 2015-19 (Ultlm of mm-oe") 1.4 1.32 11 1. 0.9 0.8 0.7 0.6 0.5 0.4 0.2 120 140 160 IS0 200 Cocoa Pice (1967 C.4/KSI a PERENAL + AN JAL r A.? LuO - 1.68 - Scenario 4: Effects of a Policy of Food Self Sufficiency on the Cocoa Supply -- Response of Scenario 3 The thrust of this scenario Is to remove the possibillty of food Imports from the adjustments of the previous sconario. Solutions to experiments in the previous scenario, particularly at higher price levels, Included Imports of up to 2.5 million tons of maize and rice In later time periods of the 50-year horizon. Elkmination of these removes the gains from trade achieved by importing maize and rico at constant prices and shifting land and labor Into the more valuable cocoa crop. In consequence, cocoa production virtually collapses at the iower price levels once the horizon extends beyond 20 years. As Fgure I.C.21 shows, only a price of 160 1987 cedis or higher assures cocoa production throughout the 50-year horizon. Moreover, the quantities produced, when cocoa is produced, are much less than under previous scenarios. Patterns of cocoa acreage under zero food Imports, shown in Figure I.C.22, confirm the collapse of cocoa Investment at the lower buying prices. Even a price of 160 cedis per kliogram results In a decilne In cocoa acreage of 50 percent. Only at the higher prices does cocoa production and acreage increase. Thus, as Table 1.C.6 indlcates, at the iower prices supply elasticities are essentially zero. In the range from 140 to 180 cedis per kligram, supply Is elastic; and only In the Interval from 180 to 200 cedis is supply Inelastic. However, under this scenario, farm Incomes Increase consistently (ignoring ths dip In the last two periods due to horizon effects) reaching the highest levels In any of the simulatlon experiments, as Figure l.C.23 shows. By ellmInating the galns from trade and thus shifting more land Into producing food grains via higher food grain prices, the prices of other foods and peremial crops tend to Increase In the competitlon for the supply of land and labor. Thus, agricultural prices, land rents and agricultural wages tend to Increase. These changes Increase farm Incomes. As cocoa prices exogenously Increase In this context, resources are drawn away from food production causing food prices to _ I.e. _ Figure 1.C.21 COCOA SUPPLY RESPONSE UNDER POPULATION AND INCOME GROWTH WITH ZERO FOOD IMPORTS 1.2- Li -/\ 0. - 0.6- 0.7- 0.6- 0.5- 0.4- 0.3 0.2 0.1 s97 1997 2007 2017 2027 2037 PERIOD MDPOINTS * o 90 U * 140 o 30 I O X - 1.70 - Figure I.C.22 COCOA ACREAGE ADJUSTMENT, POPULATION AND INCOME GROWTH WITH ZERO FOOD IMPORTS 6 - ,/' '\\ 4 - 2 J =7 l 200t 2017 2027 2037 PERIOD MDP>TS O 1X ~~+ 140 oI8^SO 20 - L1- 3g I.C23 FARM INCOME UNDER FOOD SLF-SUFFICIENCY I. _ 0.9 OA~~~~~~~~~~~~~~~~~~ O I 0. -4 0.3- 02 1 1392 2002 2032 0 3 * "140 A 3L a 200 - 1.72 - Increase and tending to increase land rents and agricultural wages. In short, the food self-sufficiency scenario turns the internal terms of trade toward agriculture, which as already noted has beneficlal Income distributlonal effects in that it Increases the welfare of poorer families. On the other hand, it raises the cost of Wlving in urban areas, increasing the real wage cost to urban employers. The patterns of land and labor use under zero food kmports, given in Figures l.C.24 and l.C.25, confirm the Increased resource aliocatlon to annual food crops. Under this scenario, agricultural employment increases by 28 percent as cocoa buying prices are Increased from 120 to 200 cedis per kliogram. Annual Decomposition for the 1988-97 Decade of selected Experbkents by the Multiperlod Model In order to assist the reader In assessing the medium-term knplicatlons of several of the more likely scenarios, the annual decompositlon model was used to provide annual projections over the next decade. The particular experknents selected were: I. Growth In population, Income and wages at a constant real cocoa buying price of 140 (1987) cedis. II. Growth In populatlon and Income at a constant real cocoa buying price of 160 (1987) cedis. Ill. Growth In population and Income wlth no grain knports at a constant real cocoa buying price of 160 (1987) cedis. Glven the apparent fIscal needs of the Goverrwnent of Ghana and Its feasible revenue alternatives, a real cocoa buying price of 140-160 cedis was judged to be the most likely range of policy choice. In particular, the 140 cedls price has been Identifled In many scenarios as the threshold below which producer supply response bocomes elastic, resulting in large foreign exchange earnings losses. Slmilarly, the 160 cedis price was Identifled In the food self-sufficiency sconarios as the threshold beiow which cocoa S,Apply collapses In the longer term. Results from simulations of the three sconaris are given - 1.73 - Figure 1.C.24 LAND USE UNDER SELF-SUFFICIENCY, 2015-19 (MiI.ttos of Mr") 14.0- 13.0-I Ito- 10.0 1 9.0 3O / 5.0 4.0 - 2.0 T 120 :40 160 180 200 Cocoa Puce (17 C.d/KqJ a Perennsal * Annual O Follow - L74 - Ftpre l.C.25 LABOR USES UNDER FOOD SELF-SUFFICIENCY (Sititon of NMi*ai) t4 0.4 - , Li! LO- 0.4 0.2 - W2 140 360 N60 200 0oo-. P ( C/K a P+NAL v MSWMA ' AMT AOMO - 175 - In Table l.C.9. Starting with Exporkment I, which Is perhaps the scenario most likely to be realized, a striking result Is the slow Increase In cocoa production over the next few years despite significant planting investments in hybrid cocoa acreage. Production stays at 250,000 tons or below through 1990, and then rises steadily to over 500,000 tons by 1997. The main reason for this time of cocoa production Is the gestatlon lags of five years before productlon commences and ten years before full yield Is reached. Since significant new cocoa planting Investments were not occurrlng before 1985, full production on significant now plantings will not be reached until the mid-1990is The other reason for the lag In production Is a consistent decline in acreage In traditional varieties as overage tree stocks reach economic obsolescence. By 1997, the estimated 885,000 acres In traditional varieties In 1984 Is projected to decline to only 36,000 acres. For this reason, estknated total acreage In cocoa does not Increase over the decade. The gain In production comes from Increasing maturity of higher yielding hybrid variety acreage. The lag of cocoa production behind cocoa planting Investments Is also reflected In the farm Income estinates (Concept 2), which remain below 250 billon cedis through 1991, and then steadily rise to over 400 billon cedis by 1997. In contrast, land and labor utilization Increase steadily over the decaae, with an additional 1.7 millior, acres going under cultivatlon and an additlonal 110 millon million mandays of labor employment projected, of which 64 millon are In perennial crops and 45 millon In annual crops. This behavior knpiies significant farm savings and Investment (mainly in the form of labor Inputs) In the early years of the decade, with the payoffs coming only after five to ten years. These Investments will not occur unless adequate Incentives to cocoa farmers are in - 1.76 - Table I.C.9: Annual Decompositon of Multiperlod Model Solutons Estknates for Several ScenarIos, 1988-97 1. Growth in Population, Income & Wages Cocoa Buyug Price - 140 (1987 cedis) Cocoa Cocoa Acreag Labor Utilization Labor Sources Prod'n (ooos Acres) (Million&(Mandays) (Millions/Mandays) Yar (OSOc MT) Trad. Hybrid Annual Pwn. Total Family Casual 1988 229.0 771.0 1540.8 284.2 306.6 590.8 569.4 21.4 1969 234.0 656.5 1818.6 283.4 328.2 611.6 584.6 27.0 1990 250.9 522.0 2077.8 286.8 348.0 634.8 605.7 29.1 1991 276.2 384.9 2283.5 290.4 362.2 652.6 622.5 30.1 1992 309.7 262.4 2397.2 294.2 367.8 662.0 632.6 29.4 1993 352.1 171.8 2381.4 301.6 361.3 662.9 636.6 26.3 i994 399.6 108.4 2339.7 309.1 361.8 67-.9 647.7 23.2 1995 448.2 67.8 2298.7 315.9 364.7 680.6 660.0 20.6 1996 488.3 45.3 2259.7 322.7 368.0 690.7 672.7 18.0 1997 513.6 36.4 2224.2 329.5 370.5 700.0 684.7 15.3 Crop Acrege Grain Farm Income (OOOe Acres) Impoits (Billion of 1987 Ced5I) Year Aual Peren. (OSOs MT) Concept 1 Concept 2 1988 4935 4123 100.0 175.8 235.9 1969 4968 4405 502.9 143.1 207.4 1990 5040 4653 534.6 160.0 227.5 1991 5093 4844 564.1 179.1 248.8 1992 5146 4956 595 7 202.5 273.0 1993 5255 4966 629.3 229.4 299.4 1994 5364 4974 665 2 260.8 331.8 1995 5474 5002 703.7 290.7 362.9 1996 5583 5060 t42.8 315.8 389.4 1997 5694 5113 818.8 332.7 407.6 - 1.77 - 11. Growth in Population & Income Cocoa Buying Price - 160 (1987 codis) Cocoa Coco Acreage Labor Utilization Labor Sources Prod'n (0009 Acres) (Millions/Mandays) (MillionlMandays) Yer (000. MT) Trad. Hybrid Annual Pern. Total Family Casual 1968 229.0 771.0 1576.3 263.3 310.5 593.8 571.8 22.0 1969 234.0 656.5 1921.7 281.7 337.9 619.6 591.0 28.6 1990 250.9 522.0 2280.0 282.7 365.4 648.1 614.7 33.4 1991 276.9 384.9 2610.7 263.8 389.3 673.1 630.5 42.6 1992 314.0 262.4 2873.2 281.7 406.5 666.2 646.0 42.2 1993 365.3 171.8 3027.0 292.7 414.3 707.0 663.1 43.9 1994 428.0 106.4 2339.7 309.1 361.8 67.9 647.7 23.2 1996 499.1 67.6 3106.7 312.7 427.1 739.8 705.8 34.0 1996 568.2 45.3 3081 7 320.3 432.4 752.7 719.5 33.2 1997 626.7 36.4 3046.2 327.5 437.9 765.4 733.2 32.2 Crop Acrege Grain Farm Incone (000 Acre) Imports (Billns of 1987 Crdis) Yew Annual Prn. (0006 MT) Concept 1 Concept 2 1968 4921 4139 100.0 176.6 236.8 1989 4960 4455 502.6 144.4 209.0 1990 5000 4760 474.6 189.5 257.4 1991 5039 5034 460.4 224.0 295.6 1992 5078 5257 542.2 244.1 317.2 1993 5196 5411 629.3 250.3 325.3 1994 5312 5513 665.2 270.2 346.1 1995 5429 5584 703.7 291.1 387.7 1996 5546 5644 742.8 324.5 401.9 1997 5663 5712 784.3 351.3 429.6 - 1.78 - M. Growth In Popuation & Income wih No Grin ImporX Cocoa Ikyn Prce - 160 (1987 oedis) Cocoa Cocoa Acrea Labor Utllzaon Labor Sources Prod'n (000s Acrs (Mililors/Manrdays) M You (000. MT) Trad. Hybrid Annual Pow. Total Family Casual 1968 228.0 757.9 1365.2 296.3 284.0 580.3 562.8 17.5 1960 231.3 623.0 1421.2 310.9 280.0 590.9 576.0 14.9 19o 246.5 466.0 1424.1 325.6 276.8 602.4 589.2 13.2 1991 264.0 312.6 1426.1 340.4 276.2 616.6 604.2 12.4 1992 279.9 183.4 1459.2 355.1 281.1 636.2 622.8 13.4 1903 291.7 102.3 1555.5 360.5 290.5 651.0 637.2 13.8 1994 297.2 58.2 1685.5 366.6 302.0 667.6 653.3 14 5 1995 299.1 30.9 1819.6 371.3 312.5 683.8 669.0 14.8 1906 303.5 36.4 1928.4 376.6 320.6 697.2 683.0 14.2 =ao7 315.5 36.4 1962.2 362.1 324.4 706.5 694.2 12.3 Crop Acreage Famr Income (0009 AMs) (B13o11 of 1967 Codis) Y u A l Peron. Concept I Concept 2 198 5125 3970 174.0 231.6 1980 5366 4011 192.5 250.8 1990 5670 3996 210.6 260.8 1991 5853 3975 231.4 291.9 1992 6o09 3965 245.6 380.1 1993 6162 4073 263.9 3276 1994 6269 4209 281.0 346.6 1905 6355 4366 297.1 364.3 19-6 6442 4514 313.1 381 5 1997 6528 4626 331.9 401.0 to cocoa farmers are In place early In the decade. Note also, the growth In inputs to annual crops Is restrahed by the Increase In grain kmports to 0.8 mIllIon tons by 1997. Experkuents II and III are really a matched pair, differing only In whether or not grain huports are allowed. They will be discussed together sInce all of the differences are due to the reubrement that domestic food demand be met from domestic production In the no-food grains hkorts case. Thib requirement forces additional land and labor Into food production. Since cocoa production at a fixed real price Is an alternative, resources can be bid away from cocoa use only by IncreasIng factor returns and therefore food commodity prices (which are now totally nontraded). The results In 1997 - 1.79 - are cocoa production that Is 100 percent greater than under food self-sufficiency at the cost of 0.8 million tons of grain bports when these are allowed. In both cases, the gestation lag results In remarkably shlllar cocoa productlon estimates through 1991. Though the food self-suffcIency scenario turns the Internal terms of trade toward agriculture, the static effickincy loss means farm income Is lower and the greater labor Intensity of cocoa means that employment Is also lower under food self-sufficiency. However, self-sufficlency Is a standoff hcome and employment wise when compared with Sconario I. It can also be argued that the dynamic learnig by doing gains from higher food prices and greater food production would speed the modernization of Ghanalan agriculture. To put the short-run cocoa prici3 hmplications of these experkments In context, the reader must keep In mind that cocoa nvostments are made when compared with alternativo Investments and In full knowledge of the gestation lag. There Is very llttle effect that a cocoa buying price announced this year can have on cocoa production over the next four years. It will have an increashg effect on productlon beyond five years. From the risk bearing perspective, the government Is assuming all of the external price risk If It sustains a constant real price--this Is equivalent to a variable export tax. If the government levies a constant unit tax on cocoa, then the farmer assumes all of the external price risk. posint a constant ad valorem tax on cocoa results In risk sharlng between the government and the farmer, with the government risk share determinod by the level of the tax. In contrast, the farmer perceives a government prcig risk In that he doos not know what price the government will offor In the future when he makes a cocoa Investment. Shulathg a dynamically varying rIsk hmpiles a stochastic formulation, and this Is currently beyond the state of the art. Moreover, If the risk Is truly random, then In the longer run Its value should appraoch the mean (If the distrbution function Is weN behaved), and a deterstic formulation Is adequate for - 1.80 - assesshg long run hnvostment prospects. As a practical matter, the worst case is vacillating government pollcy that arbitrarily varles the real buying price (or real tax) since Investors will shorten Investment horizons (which Is the kiss of death for long-lived investments like cocoa) until they have formed stable prior distrlbutlons. This might take long tine if government Is sufficiently arbitrary. Pollcy Implications of Simulation Experlments It Is now appropriate to sort out the lmpications of the analytical work on cocoa producer supply response. First and foremost, It Is clear that specification of a producer buying price (even In real terms) Is not sufficient Information to permit unqualifled estimates of cocoa supply. Analytical methods which attempt to do this are really saying that the future will be like the past. In many contexts, this Is an acceptable approach; but givon Ghana's economic history of recent decades, such an approach can be misleading. Who could have predicted the exuberant cocoa Investment resDonse of Ghanaian farmers to Improved price Incentives, given the way that previous policy had worked against the Interests of cocoa farmers? Yet the evidence from the Survey for Cocoa Producer Pricing shows that 33 percent of existing cocoa acreage In trees have been planted In the past four years. The results from the simulatlon experiments are less obvious since they are concerned with medium- and long-term horizons. The Inertia of the past weighs heavily on the adjustment path of long gestation, long-lived Investments. Thus, the differences among the several scenarios with respect to cocoa production and acreage are not great for the first ten years of the multieriod experhnents. By the tkme 40 years have elapsed, the differences can be huge. Perhaps the real Issue of judgment is which scenario Is the relevant one for policy planning purposes? Since no one can predict the future, this Is not a question that responds to economic expertise. For that reason, several varlant scenarios have been shmuiated. However, discounting sconarios with zero population growth or with Income - 1.81 - but not wage growth as very unlikely, the supply for the remaining sconarlos Is given for ton and forty year horizons In Table l.C.10. The sconario wlth parl passu growth in population and cultivable land is In effect a simulation of Ghanalan experience over the past two decades. This Is the set of experiments which show the greatest growth In cocoa production and for which supply response Is inolastic over most price Intervals. Yet the bottom line Is static real Income for the Ghanalan people, and what use Is all the cocoa production? Clearly, the sconarlos with positivo welfare kIplications for the Ghanalan people are those with growth In population and Income, with or wlthout a food self-sufficlency policy; and this kind of scenario Is the objective of the Economic Recovery Program. One knplicatlon of Table l.C.10 Is that a price of 120 cedis per kilogram is not sufficient to sustain cocoa production over the lngor term when Incomes are growing; If a food grain self-sufficiency pollcy Is adopted, neither Is a price of 140 cedis. Although sizeable volmes of cocoa can be produced at these prices over the medium term, say, ten years. At higher prices, cocoa supply response Is sustained over both medium- and long-term horizons. Thus, the questlon of the optknal level of cocoa taxation has a dynamic dknenslon that asks where will the competitive trade advantage for Ghana be In the long run? Past policy In Ghana can be Interpreted as having assumed that long-run advantage did not Include cocoa, but that assumption has not turned out to be true. At the very least, policy ought to consider mixed strategies which admit that present expectatins of the future can be mistaken. Both of the scenarios positing Income growth, with or without food grain self- sufficiency, show substantial growth In employment as cocoa prices are Increased, reflecting the greater demand for resources as incomes Increase and more resources are shifted to labor-intensive cocoa production. Thus, a low cocoa tax, high cocoa production pollcy kmpacts quite positively on the incomes of the rural poor. Of course, the food grain self-sufficiency policy Impacts drastically on cocoa production at lower - 1.82 - prices. The reason Is quite shuple. This policy shifts resources away from more productive employment In cocoa production toward less effiient productlon In maize. It may be argued that this outcome Is really a consequence of the assumption of static agricultural technogy. However, until convincing evidence Is at hand that Ghanaian agriculture Is movlng to cost efficlent cost efficient higher productivity practices, a food self-suffiieWncy policy entalls heavy offilency lsss. On the other hand, combining food self-sufficiency with a high cocoa prie policy does yild a strong kIpact on employment of rural labor and smalholder farmer Incomes. The greater Impact on the welfare of the poor might be an acceptable tradeoff for lower static efficiency. Moreover, If It Is accepted that productivity enhanchg technh;al change In agriculture Is more likely in the context of rising agricultural employment and Incomes, then dynamic efficiency gans are more lkely under the food self-sufficincy scenario. Table LC.1 10 Cocoa Supply Estbates ('000 Metric Tons) CrOO BjoIN Pvlc (1987 cOiM Horizon and Scenario 120 140 160 180 200 Ten-Year Horizon Population Growti Only 338 527 643 687 728 Population, Income & Wago Growth 248 468 548 an6 6/6 Population & income Growth (wliF no food imporb) 216 216 296 440 530 Fort-vYew Hortzon Population Growth Or 1062 1449 1578 1697 1781 Populton. bnom & Wa o 314 636 1191 1346 Population & ncom Growth (no food mo-) 282 689 1179 D. ALGEBRAIC SPECIFICATION OF GHANA MULTI-PERIOD AGRICULTURAL SECTOR MODEL Set Definltions Sot Name D" (r) Agricultural regions (rp)C(r) Agricultural regions with perennial crops tc) Crop commodities (ca)C(c) Annual crop commodities (cp)C(c) Perennial crop commodities (cn)C(c) Natlonally consumed crop commodities (cm)C(c) knported crop commodities (ce)C(c) Exported crop commodities (rc)C(r,c) Crop possbilitles (a) Annual crop prooessos (Ap) Perennial crop processes (ra)C(r,a) Process posslbilltles (t) Thne In months (v) Vintage (in 5-year ordinal periods) (vf)C(v) First vintage (th) Tlme periods (5-year chronolgical periods) (tl)C(th) Initial period (tp)C(th) Tlme periods (vtov)C(v,v) Agelng sequence sets (in) Inputs tk) Tasks (v)-(vv) Alias sets (c)-(cc) Alias sets (cn)-(cnn) Alias sets (v)-(vv) Alias sets _ 1.84 - Variable List Varabb Name Descr"tlon XA(TH,R,A) Land under annual crops TXA(TH,R) Total annual land XP(TH.R,AP,V) Land under perennial crops F(TH,R) Fallow land XPF(TH,R,AP,V) Abandoned perennlal crop land XAF(TH,R) Abandoned annual crop land CL(TH,R) Cleared fallow land LABFAM(TH,T,R) Family labor LABCAS(TH,T,R,RR) Casual labor OUTPUT(TH,R,C) Crop connodity production NATCON(TH,C) Domestic consumptIn of crop commodity EXPORTS(TH,C) Exports of crop commodlty IMPORTS(TH,C) kIports of crop commodity INPCOST(TH,R) Cost of purchased nonlabor Inputs LABCOST Labor cost ICOST Total Input cost CPS Consumers' and producers' surplus - 1.8S - Equation List 1. Perennial Land Baance XP(TH.RP,AP,V) * E XP(TH-1, RP,AP,VV) - XPF(TH,RP,AP,V) VVeVTOV(VV,V) 2. Falow Land Balance F(TH,R) - F(TH-I,R) + XAF(TH,R) + E XPF(TH,R,AP,V) AP,Ve/VF(V) + GF(TH,R) - CL(TH,R) 3. Total Annual Land Det tion TXA(TH,R) - £ XA(TP,RA) AeRA(R,A) 4. ItAhs Falow Requremmnt F(TP,R) 2 FREO(R) TXA(TP,R) 5. Arwuja Land Balance TXA(TH,R) - TXA(TH-1,R) - XAF(TH,R) + CL(TH,R) - E XP(TH,RAP,VF) (AP,VF) 6. Labor Balnce L CLABREQA(R,A,T) + FREQ(R) TLABCL(T)) XA(TP,R,A) + AeRA(R,A) C(LABREQP(AP,T,V)XP(TP,RAPSV) + TLABPL(TAP) XP(TP,R,AP,VF) + (AP,V) LENGTH E LABCAS(TP,T,R,RR) + TLABCL(T) CL(TP,R) - RR LENGTH LABFAM(TP,T,R) + LABCAS(TP,T,R,RR) + t LABCAS(TP,T,RR,R) RR 7. Labor SuyPpy Constraint LABFAM + E LABCAS(TP,T,R,RR) < LABSUIIT(TP,T,R) RR 8. Prt.-ucton Acount g OUTPUT (TP,R,C) - E NYELD(R,CA) XA(TP,R,A)/1000 AeRA(R,A) + E XP(TP,R,AP,V) YIELDP(CAP,V) T E YELDP1 (C,AP)XP(TPSR,AP,V) (AP,V) (AP,VeVF) 9. D_wnd Balance NATCON(TP,CN) - E OUTPUT(TP,R,C) + WPORTS (TP,CU) ReRC(R,C) - EXPORTS(TP.CE) - 1.86 - 10. Wkut Cost ACOunthg INPCOST(tP,R) . E CICOSTA(R,C) + FREO(R)INCOSTCL) XA(TP,R,A) AERA(R.A) + E ICOST(AP,V)XP(TP,R,AP,V) + INCOSTPL(AP) XP(TP,R,AP,V) (AP,V) (AP,VfVF) LENGTH + INCOSTCL CL(TP.R) LENGTH 11. L~LOW 0ost ACWXMV LASCOST(TP,R) a £ (WAGEFAM(R)LABFAN(TP,T,R) T + E WAGECAS(RR,R)LASCAS(TP,T,RR,R)) (1 + WFACTOR(PINDX(TP) -1)) 12. Total Cot DetIfon TCOST(TP) - E (CNPCOST(TP,R) + LASCOST(TP,R)) R 13. Mwket EiUibrkn Cndftlon CPS -a £ EDELTA(TPXE (ALPHA(CN) - GAMMA(CN)(1 + YFACTOR CN (CINOX(TP) -1)))NATCONWTP,CN) + .54ETA (CN) (NATCON(TP,CN]2) 1 + PFACTOR(POPINDX(TP)-1) - E PE(CE)EXPORTS(TP,CE) - £ PM(CM)OAPORTS(TP,CM) - TCOST(TP))) CE Cm 14. Wm Condtw an Lwp Bowids XP(TI,TPAP,V) - CALLOCP(AP,RP) YPROP(RPAP,V) TXA(TI,R) - E CALLOCA(A,R) A FMTI) - FRECKR) TXA (TIR) LAOFAWTH4T,R) O FRAT LR) LABSUPT(TH,T,R) _ 1.87- Parameter List FREO(R) Fallow land required per annual rotation (prop) GF(TH,R) Growth additlon to land cultivated (1 0OOA) LABREQA(R,A,T) Total labor requirements, annuals (MVPA) LABREOP(AP,T,V) Total labor requirements, perennials (.ADPA) LENGTH Length of the periods (YEARS) TLA8CL(T) Labor requirements, clearng land (MDPA) TLA8PL(T,AP) Labor requirements, planting perennials (MDPA) LABSUPT(TP,T,R) Total labor availabiNty (1000 MD) NYIELD(R,C,A) Net yild, annuals (KGPA) YIELDP(C,A,P,V) Perennial yields (KGPA) YELDPKC,AD) Oth peronnlal (first peiod) yields (KGPA) ICOSTA(R,C) Total hput cost, annuals (1000 CPA) ICOSTP(AP,V) Total input cost, Perennials (1000 CPA) INCOSTCL hput cost, clearkig land (1000 CPA) INCOSTPL(AP) Input cost, perennial plantig (1000 CPA) WAGEFAM(R) Family reservation wage rate (1000 cedis) WAGECAS(RR,R) Casual labor wage rate (1000 cedis) WFACTOR Wage growth switch (zero or one) YFACTOR Income growth switch (zero or one) PFACTOR Population growth switch (zero or one) PINDX(TP) Productivity ndex (1985-89-1) (PROP.) CINDX(TP) Consumption Index (1985-89-1) (PROP.) POPINDX(TP) Population Index (1985-89-1) (PROP.) DELTA(TP) Discount factor (PROP.) ALPHA(CN) intercept, commodity demand function BETA(CN) Price gradient, commodity demand function GAMMA(CN) Income shift, commodity demand function PE(CE) Export commocity prices (CPKG) PM(CM) h4ort commodity pricss (CPKG) CALLOCPAP,RP) Gross perennial crop allocation (1987) (1000A) CALLOCA(A,R) Gross annual crop allocation (1987) (1000A) YPROP(RP,AP,V) Perennial yield proportions (1987) (PROP.) FRAT(R) Fraction of famiy labor (PROP.) - 1.88 - Key to Unit AbNrevlatku A Acre(s) MD Man-day(s) MOPA Man-days per acre KGPA Klograms per acre CPA Cedis per acre CPKG Coes per kclbgram PROP. Proportion I. COCOA TAX AND REVENUE ALTERNATIVES by DWdE M. Ns.bw 1. INTRODUCTION Ghana's past cocoa t and pricing policy has caused cocoa exports to decilne steadlly over the past two decades. As part of the Economic Recovery Program the price paid to cocoa farmers has recently been significantly Increased, although In the 1987- 88 season It was stlll less than 40 percent of the export price. The fIrst part of this Interku study examines the way In which Ghana's tax system and agrlcultural pricing policy has Influenced the composition and level of output of her agricultural sector, particularly In the cocoa growing regions. It also examines the consequences of this policy for Goverrnent revenue. The Government must decide on Its future price policy towards cocoa and other crops. These policies will affect future levels of agricultural output, forelgn exchange earnings, government tax revenue, and the standard of living of cocoa farmers. The final part sets out the case for ralshg the producer prico of cocoa as quickly as possible to at least 55 percent of the export price. In 1986, the latest year for which esthiates are available, foreign exchange earnings In real terms were less than half their 1970 level, despite a considerable kIprovement over the recent past. Cocoa accounts for about two-thirds of Ghana's export earnings. World Bank projections show other sources of foreign exchange rlsirg from 32 to 37 percent of total export earnings by 1989, whilst debt service and Imports of necessities will place heavy demands on available foreign exchange. In the medium term it may be possible to diversify Ghana's exports, though most of the alternative agricultural exports are, like cocoa, tree crops with a lengthy gestation period. Despite the heavy discrh,inatlon against cocoa, there has been little evIdence In the past of any significant switch to alterative export crops, and It would therefore seem unwise to rely on any spontaneous Increase in earnings from these sources in the short to medium term. Export earnings from cocoa will remain central to Ghana's neods In thG foreseeable future. Cocoa production will continue to declhe as trees age and yields fall, unless the rate of planting of new trees rises significantly compared with the past two decades or the productivity of existing trees Is -alsed, or both. Future world cocoa prices are - 2.2 - projected to decline as supplies from other countries expand, so wlthout an active policy to stimulate new cocoa planting, Ghana's export earnings will be severely squeezed. The most dire-t way to encourage more planting, and the more Intensive cultivation of existing cocoa trees, Is to raise the price to the farmer. In tne 3hort run this Is likely to reduce Goverrnment revenue, though the amount of cocoa harvested and sold should increase and offset some cf the revenue reductn. In the medium run (4-6 years) if the price Increase Is enough to stimulate new planting, cocoa supplies should Increase further as these trees come Into bearing, and the decline in revenue will be further offset. The Government therefore faces a difficult choice -- In order to reduce the pressure on foreign exchange earnings It Is desirable to Increase cocoa prices and face a reductkn In its cocoa tax rece4ts. This will be acceptable, provided there are alternative sources of revenue which have less damaging consequences on the economy than the present cocoa tax. if COCOBOD Is to be In a position to argL, for ralsing the price of cocoa to farmers, then It needs to know the revenue Implications of this policy, and be able to compare the economic costs of the cocoa tax with those of other taxes that the Government Is currently examining. on addition, alternative taxes may have impacts or cocoa farmers which will affect their supply decision, and these need to be taken Into account when deciding on the appropriate cocoa pricing policy. From a wider perspective, the Government should Oe concerned to choose a tax system and pricing policy which ral*es the required level of revenue efficiently and equitably. It Is therefore Important to compare the present taxes on cocoa producers with possible altemative sources of revenue. If It can be shown that cocoa taxes are less efficient and/or less equitable than altemative taxes, then the case for lowering cocoa taxes and replacing the lost revenue by these alternatives Is further strengthened. Of course, a study such as this whose prime focus is the cocoa industry cannot undertake a full review of the current tax system, which Is in any case the subject of - 2.3 - a World Bank stLAdy t be undertaken in early 1988. Instead this study will consider a limited range of alternative sources of revenue which have been suggested, and which are of direct relevance to the cocoa Industry. One suggested alternative Is to replace, elther fully or partially, the cocoa tax by extending the present Income tax to cocoa farmers. At present, in prlnciple, all non- cocoa agricultural income Is liable for direct taxation, though Income from cocoa Is exempt. It Is therefore Interesting to examine the case for such a chtngo, to see whether It Is warranted on efficiency and equity grounds, and whether It would be administratively sensible. Taxes on reported agricultural Income are not only hard to monitor, hut also suffer from the usual disincentive effects, and varlous economists, together with the NF, have therefare argued for a land tax, effectively a tax on presumptive or po' ntlai Income, rather than a tax on agricultural Income. How far does this argument apply In Ghana? Would a land tax, assuming It to be feasible, be preferable to the present cocoa tax? If so, would It be administratively feasible In Ghana, elther now or in the Rikely future? Both agricultural Income taxes and land taxes are natural alternatives to the current cocoa tax, though both present formidable administrative problems. If, as seems likely, these difficulties rule them out, then alternatNVe sources of revenue will be iequired. One obvious alternative Is a broad based Indirect tax, such as sales taxes and/or hmport tariffs. Another Is an excise tax on transport fuel and/or purchase taxes or Increased licence fees on transport vehicles. Both would affect cocoa producers, though they would also have wider inpacts. Indirect taxes would reduce the purchasing power of cocoa farmers, whiist transport taxes would raise the cost of transporting cocoa from the farmer to the point of export. Nevertheless, because both taxes fall on other sectors of the economy, the net burden of taxes on cocoa farmers would fall, and one would like to know by how much. - :.4 - At present, inputs such as pesticides, fungicides, fertilizer, sprayers and cocoa seeds are subsidized, In large part to offset the Ilaincentlve effects of the low producer price of cocoa. If the producer price Is Incteased, then the case for subsidies Is weakened, though not necessarily ehiminated. A coherent pricing policy for cocoa should therefore also consider th choice of the aproprlate level of subsidies on purchased Inputs. Finally, cocoa Is an Investment with a gestatIon period of 4-5 years and a life of 30-40 years. Any change In tax or price policy will have consequences stretching Into the future, which need to be taken Into account. One potentially attractive way of summarizing theso dynamic apcts Is to discount the future hipacts on revenue back to the present. It may be that raising the price of cocoa Induces a sufficiently large supply response that although there Is an himdbate fall In cocoa tax revenue, future revenues actually Ierease. In an extremo case the present value of tax revenues might rise with an Increase In the producer price, makhg both the goverrwnent and the cocoa farmer better off. Even !f the present value of revenue falls, the not benefit to producers might be large compared wlth the fall In revenue, making It potentially attractive to switch to some altematIve source of revenue which has a lower Impact on welfa )er unit of revenue raised. - 2.5 - 11. THE ANALYSIS OF TAXIS AND TAX REFORM When consIdring a reform of tax*s or agricultural pricing polices, there are two natural ways In which to proceed. The first Is to compare the current system with some concept of a 'desirable' or opthnal alternative, and then to move towards this ideal goal. The second, rather les ambitio aWroach, Is to attert to detbrmine the costs and benefits of possibe changes, and choose changes which have a high benfit-cost ratlo.1 If the benfit of a tax Is the revenue Is rates, then the cost Is the reductIon In welfare of the taxpayer. The natural way In which to rank tax or price changes is to measure the fall In welfare pr urt of revens raIsed - then the most desirable taxes are those which give the lowest faN In elfare per unit of revenue raised. The desirablgty of a tax wtfl depend on two factors. First, the extent to which Jt Is dstrlbutlonally desrable - that Is, the extent to which the tax falt more heavily on groups who are consikered dstributkonally ls deservhg, such as the rich. Second, the inef lfincy of the tax wUI depend on the proportion of the fal In the money value of welf are which Is captured In the form of reesnu. If a tax reduces the real Income of tax payers by SY, and yields revenue SR, then the dead-weight lss of the tax, a measure of Its Inefficiency, Is Y-R, or, In proportional terms, Y/R-1. Taxes on Inelastically supplied or demanded goods cause price changes which result In relatively small changes In demand, and for such taxes the inefficiency will be low. If consumers can readily substituto altemative, untaxed goods for those on which taxes are levied, then little revenue will be collected, though the consumers will have been made worse off since 1 The distinction between these two approaches has some skmilarities to the distinction betwen comprehenlve plannig and project appraisal. With the former the akl Is to choose the entire hnvostmnt plan, whgst the latter Is a piecemeal or decentralized way of selectig Individual projects. Of course, a good method of cost benefit analysis systematically applied will generate a complete Investment plan, ust as a systematic approach to measurhg the costs and benefits of tax changes should result In the selection of an opthial set of taxes. In both cases the merit of the plecer ial approach Is that It Is informatkonally less demanding. - 2.6 - they have now shlf ted to their second-best choices. If they have few alternatives to the taxed goods, then they are effectively compelled to pay the tax, and the efficiency of the tax will be high. The case for uniform and broad based taxes, that Is, taxes which fall on a large fraction of consumer expenditure, Is that they provkid Uttle reason or opportunity for changing consumption patterns to avoid the tax, and hence are relatively efficient. Income taxes, or uniform taxes on consumption (such as a universal value added tax) are good examples where the mair source of inefficiency (assuming that evasion Is not a problem) Is the reduction In work effort since ielsura b untaxed). Land taxes are ahuiarly attractive as the SUppiy of lnd Is fixed, and again the tax camot be avoide by shIfting expenditure to other goods. Heavy taxes on a narrow range of goods are potentially distortionary unless these goods are helastkally demnded (le are ssentlals) or are helastbcaity supplied (ike land). Goods that are hieastlcafy demanded and essential are lkely to account for a larger fractlon of expenditure by the poor than the rlch, and heavy taxes on such goods are therefore lkely to be inequitable. Although taxing such goods may appear to score well on effciency, they score poorly on equity, strengthening the case for broader based taxes which are less inequtable. The suiplest taxes to analyze are those whose incidence falls entirely on the tax payer. Value addod taxes which are only paid by final consumers are a good example - - they resut In an increase In the consumer prlce of the taxed good, but, under reasonable assumptions, no change In the pattern of r.oducer prices. When analyzing changes In taxes on fhal consumpt!oii goods one only needs to study the hnpact on consumers - supply consirations (and supply eiastthies) can be gnored. Taxes on agricutural Inputs and outputS have far more widespread repercussions, and It Is potentially misleading to concentrate only on the hedate tax payers. Further, the efficlency of such taxes wifi depend not only on demand elasticities but also on supply responses and supply elasttities. - 2.7 - The tIrst step hi studyhi Ghana's tax and prico policy Is to sketch the way In which the current system of taxes, prices, and other oovernment polices affects the Ghanaian economy. The sketch Is necessarily rather h4pressoonbstick as data are scarue and often unroeable, and It Is hard to be wue how the economy hi fact Q"pllbrates In rosponse to the varlety of distortion and hIterventlone to which ; Is subject. - 2.8 - 111. SAUENT FEATURES OF GHANAIAN COCOA PRICING Ghanalan agriculture might be roughly characterized as follows. The main export crop Is cocoa, which Is heavily taxed. It competes on the production side with root crops, vegetables and fruit, all of which are, to a first approxhuaton, non-traded goods, and with cereals, which are currently imported. Cocoa competes with maize for land and labor. Ric Is grown under different agro-cikuatic cIrcumstances, and so cocoa does not comto so directly with rice, though there may be an indirect and possibly rather weak link through the labor market. Indeed, In the past a significant fraction of the labor needed for the expansion of cocoa production was migrant labor from the North and from present rice producing areas. The main defence for taxing a non-food export crop like cocoa Is that Its supply Is relatively Inelastic, the tax incidence Is likely to be m*Hiy ,Xrogressivo (especially If cocoa farmers are wealthier than non-cocoa farmers). and the tax Is administratively easy to collect. It we accept the equity and administrative arguments for the moment, then the key Isstne Is the Inelasticity of supply. A cocoa tax will approxlnate a broad-based tax If all of the following conditions hold. I) The supply of total labor in the economy as a whole is Inelastic, that Is, tUe totat number of hours worked In Ghana changes little In response to changes In the cocoa price. ii) There Is little substitutablity between cocoA and traded agricultural (,oods like maize and rice, or, alternatively, these other traded goods are taxed as heavily as cocoa, so that the tax base Is effectively broadened to a tax on traded agricultural goods. As we shall argue below, It may not be necessary that non-traded agricultural goods are taxed. 111) Elther the demand or supply of non-traded agricultural goods Is Inelastic -- that is, either It Is unattractive to switch between producing traded and non-traded agricultural goods, or consumers have relatively fixed demands patterns for non-traded goods (like roots and vegetables). - 2.9 - Iv) The supply of labor to the non-agricultural sector Is elastic, that Is, It responds to differences between urban and rural living stArdards, and the non-agricultural wage responds to chanr,.os In labor supply. To see the role of each of these conditions, suppose that the farm-gate price of cocoa Is reduced. Farmers will now find It relatively more attractive to produce other crops. If there Is llttle possibility of substituting Into traded crops, or If their price has also decilned as much as that of cocoa (condition 2) then the only attractive alternative Is to shift Into the production of non-traded goods, whose price Is set by local supply and demand. If It Is hard to switch Into these crops (condition 3), then their supply will be Inelastic, and the farmer effectively cannot replace cocoa by other agricultural activities. If consumers have an Inelastic demand for these non-traded crops (condition 3) then a small Increase In their supply will cause thelr price to fall, Inhibiting further switches from cocoa to non-traded crops. If farmers cannot divert their resources to other crops, then they will be compelled to elther continue growing cocoa, or to migrate to the urban area, or to wlthdraw their labor from the economy (either by emigrating or by Increasing their leisure). If the urban wage responds rapidly to a fall In the rural standard of living (condition 4), then the relative attractiveness of migrating will be roduced, and cocoa farmers will again be Induced to continue production. An elastic labor supply toge.her with a competitive urban labor market widens the incidence of any agricultural tax, and by broadening the effective base of the tax, reduces Its Inefficiency, and reduces the supply response to the tax. The next section examines the historical record to throw some light on these Issues, and to examine the kmpact of past policies on the economy and on Government revenue. Given the sparsity of reliable data and the problems of Interpreting events In a highly distorted and Inflatlonary economy (issues which are taken up In Appendix 1), the - 2.10 - approach adopted Is to look for persistent long tarm trends or deviations from earlier pollcles, and to study them graphically, rather than econometrically. A visual presentatlon such as that adopted here gives a much better Idea of the extent to which events ir. any year or short period are unusual, significant, or merely reflect lags In adjusting to sharp changes In Internal or external circumstances. For example, measures of effective protection or domestic resource cost ratios can vary dramatically from year to year wlth changes In the degree of overvaluation of the exchange rate, changes In world market prices (which can be dramatic, as the graphical record shows), or changes In domestic supply and local price levels. Any study of agricultural pricing policy using these measures which Is based on one or even a few years' data gives little reassurance of the correctness or robustness of Its concluslons. Graphical prosentations give a quick visual kIpression of the degree of variability and a rough Idea of the significance of devlations or trends (though of course it is always useful to supplement them by more systematic econometric tests). - 2.11 - IV. BRIEF HISTORY OF COCOA PRODUCTION PRICING AND TAXATION IN GHANA Figure 1 gives a graph of cocoa production since the 1947/48 crop year, and the five ysar moving average smoothed trend. It shows reported2 production roughly doubling between 1957 and 1962, then holding roughly constant until about 1972, and then doclinng fairly steadily to Its current levil somewhat below the start of the period. Flgwo 2 gives the real prIce of cocoa (deflated by the rural cost of living Index)3 since 1940/41. It showS the dramatlc rise In prkcos after World War 11, their steady fall up to the mid 1980s, and their second main fall In the mld 1970s. Prices In the recent past have been less than one quarter of their value in the early 19509. The latest figures, based on an assumed producer prico of 150,000 cedis per MT, and using the August 1987 prico inex as deflator (pwrhaps optnisticaliy) shows that real prices appear to have returned to their 1970 level. Figure 3 suporlmposes the prico serles and the hIpiled total producer Income (real prico thues production), showing that Income remained roughly constant until the early 1960s (as productlon Increases offset price falls), but since then the docilne In producer Income has be(en even more dramatic than that of producer prIces.4 Naturally one asks whether the fall In output can be attributed to the fall In prices, and Flgure 4 graphs the movement In output and the real pries of eleven years earlier. Cocoa takes about 4-6 years to come Into frult from planting, and ylelds Increase steadily thereafter. The 2 The fgwuros refer to deliveries to buying agents, and thus Ignore productlon which Is Illegally smuggled IntO nelghboring countries, such as the Coto d'lvoire. Konings (1986), citing World Bank esthuates, suggests that a minhum of 40,000 tons were smuggled out of the country In 1977, whilst a Brong-Ahafo Regional Admuinistrator has asserted that as much as one-third or the region's output Is sold llelly In the Coto d'lvoirs. (West Africa, June 12, 1978.) 3 See the discussion In the Appendix on defiating cocoa prices, and the relative movements In the different price indlces. 4 See also footnote 2. To the extent that cocoa was smuggled, farmers Incomes will be understated by this calculation. GHANA Cocoa Production Tonnes (Thousands) 800 r 400 - 300- 200 100 _ 1947 1952 1957 1962 1967 1972 1977 1982 crop years Figure 1 Total output + 5 yr MA GIHANA REAL COCOA PRICE Producer price deflated by rural CPI 1977 C/MT (Thousands) 6 5 4 3 2 1 O ,,, I' ,,1 f I 1111 11 I_ Ij I I 'ji 'ii ' '11 I' 1940/41 1945/46 1950/51 1955/56 1980/61 1965/66 1970/71 1975/78 1980/81 1985/86 Crop Year Figure 2 - Real producer price Figure 3 GHANA COCOA Production, Prices, Income 00 tons, C/MT, x .2 mill C (Thousands) 8. 6 4f 2 1947 1952 1957 1962 1967 1972 1977 1982 crop years - Total output + real price i real income Income is real price times production Figure 4 GHANA COCOA Production and lagged prices '00,000 tons,'000 C/MT 6 5 4 3 S ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~, 2 1eI 1951 1956 1961 1966 1971 1976 1981 crop years - Total output + real price lag 1 lyr Prices lagged 11 years - 2.16 - figure certainly makes plausible the clakm that farmers plant when prices increase and stop planting, or neglect existing trees as the price falls, with the effect on output showing up after a substantial lag, perhaps as long as eleven years. The obvious questlon to ask is why the real producer price fell so dramatically. Figure 5 graphs the real world price and the real producer prices In Ghana and Cote d'lvolre In constant S1980. The world price would be the same as the fob export price (apart from a relatively small transport component) at a Purchasing Power Parity exchange rate, which the Appendix argues Is the appropriate measure to use for a distorted economy such as Ghana's. The Cote d'lvoire Is a natural country against which to compare Ghana, as it shares a skillar cklkate and ecology, and yet has displayed a dramatically dlfferent cocoa production performance since the start of the period, more or less changing places with Ghana In the world share of cocoa production. 5 The vertical scale in Figure 5 Is necessarily compressed and disgulses the fall In the producer prices In the two countries, so Flgure 6 gives the ratio of the producer prices to the five year moving average world price for the two countrles. Together the two graphs show that the decline in producer prices In Goiana (and perhaps In the Cote d'tvoire, though the data are lacking) In the late 1950s was in response to the fall In world prices - If anything the ratio of the prorducer to the world price Increases in Gharn until 1962/63. Thereafter the world price recovers, but the Ghanaian producer price continues to decline as a fraction of the world price, In contrast, producer prices In the Cote dilvoire are a relatively more stable fraction of the world price and does not exhlbit any noticeable decline, at least In the 19709. Furthermore, the producer price in the Cote d'lvoire remains at roughly the same fraction of the world price as that In Ghana In the l9SOs. 5 Again, though, see footnote 2. To the extent that Ghanalan cocoa was smuggled into Cote d'ivoire, her performance has been exaggerated, though of course the causes of the dlfferential marketed supplies remalns the difference in prices paid In the two countrles. Figure 5 Real Cocoa Prices At constant $1980 1980$/MT (Thousands) 6 5 4 3 21 1 1950/51 1955/56 1960/61 1965/66 1970/71 1975/76 1980/81l 1985/86 year World price Ivory prod pr Ghana prod pr 5 yr MA wld price Deflated by CPI or MUV Ratio of producer to world price Ivory Coast and Ghana 100 Percentage 80 - 60 L 40 - 20 1952/53 1957/5a 1962/63 1967/68 1972/73 1977/78 1982/83 1987/88* year Figure 6 Ivory Coast Ghana 'atios to 5 yr MA (final years forecasL) - 2.19 - Cocoa output mht respond to a faN In the price of cocoa for two different reasons. If the price of cocoa fell relatively to the prices of other crops, then farmers might devote their effort to growhg these alternativo crops instead. If so, then it might be sufficient to restore the previous relative prices of alternative crops, perhaps all at a lower absolute level, in order to encourage farmers to redirect their efforts towards cocoa production. On the other hand, the farmers may negbect cocoa not because of the relative attractivenoss of alternative crops, but because farming has beconme less attractive than migrating either to the urban areas or to neighboring countries. Figure 7 shows the real prices of competing crops (though ric is relatively unknportant In the cocoa growhg area) The real prices of an crops decline over the period, though none so dramatically as cocoa. Figure 8 gives more detaih for the relatively recent past, and corrects for the unstable torms of trade between food and non-food during this period by deflating by the rural food price. (More details on the relative movements in the various price indlcos are to be found In Appondix 1). This shows the lack of any obvious trend in the prices of the ma!n food crops (with the exception of rice whose relative price tends to appreciate), but the continued decline on the cocoa price. Ghana Real Producer Prices Deflated by rural CPI 1977 cedis/MT (Thousands) 6 5 4 3 2 0 1953/54 1958/59 1963/64 1968/69 1973/74 197B/79 1983/84 crop year Figure 7 cocoa i maize X rice ° yam GHANA REAL CROP PRICES Deflated by rural food price 1977 cedis/MT or 100kg 350 300 250 200 150- 100 50 .0 1970 1972 1974 1976 1978 1980 1982 1984 crop years beginning Yams | Cocoyams Cassava a Rice Maize Cocoa Figure 8 - 2.22 - Flgure 9 highlights the relativo price movemonts in cocoa and maio at domestic and world pricos. (The prices of the two other main alternatioeS to cocoa - yams and cocoyams -- follow that of maize farly closely and the graph of their domestic price ratis would be shUar to that of cocoa to maize). The praph brhg out two points clearly -- whilst the domostic prico ratio of cocoa to malze has declhed fairly C eadily from its late 1960* peak, the world price ratio has If anythin drifted upwards, reflecting the steady fall In the real world price of maize, to the point where It now stands at Its nw,st value for over fifty yers. Thus whilt cocoa has steadlly become relatively more valuable than maize in the rest of the world, Ghana has boen attompthg to maintain the domestic real porlce of maize hilst alowi the real price of cocoa to fal. How does one nterpret the evidenco? Although the relative price of cocoa to other crops has deteriorated, the pries of non-traded products such as yams and cocoyams has not faln much In ra tr_m over the past fifteen yars. Had their been a substantial switch Into these non-traded alternatives one might have expected their supply to Ins ease relative to dmnd and to have dopressod their price. Of course, It may be farmers chose to grow maize istead of non.-traded crops, and that the price was prevented from falling by protective trade policies (and, at the Increasingly overvalued exchange rate, the domestic price of iaported crops such as maize and rice might be expected to have fallen rapidly In real term, even faster than the fall In world prices). Even here, though, there Is not much evidenco of a large increase In domestic food productlon. Figuro 10 gives (completo) ovdenco on areas under food crops. The recent rise appears to mahW be a recovery from the poor performance of the mid 19709, and there In no perceptble trond over th period from 1970. Flgure 11 conf Irms that this also hold for maize, which , urnko the other coreals, Is grown In the cocoa producing area. Given the substantlal irease In population during this period It Is hard to accept that the rural labor force redirected Its efforts from cocoa to alterativo crops. The Figure 9 Ghana Relative Price of Cocoa Deflated by maize prices ratio at domestic and world prices 20 15 10 5 0 1953 1958 1963 18 1973 1978 19b83 year + cocoa/maize cocoa/maize $ 3YMA$ prices Ghana Cocoa-Maize price ratio Ratio of domestic to world price ratios percentage 100 80 60 40 20 - 1953 1958 1963 1968 1973 1978 i983 year Figure 9A dom/S price ratios Ratio to 3yr MA world price ratio Figure 10 GHANA Area under crops Thousand hectares 3000 2500 2000 1500L 1000 500 0 1970 1972 1974 1976 1978 1980 1982 1984 1986 * _ Cereals Cassava L Yam Other starch Pulses LI Vegetables Figure 11 GHANA Area under cereals Thousand hewtares 1400 1200 1000 800 600 .I. .. .. 400 S 200 19.0 1972 1974 1976 1978 1980 1982 1984 1986 Maize Rice LI Millet Guinea Corn - 2.27 - more plausble explanation Is that cocoa famers (or potential cocoa *armors and workers) withdrew their labor from the agricultural sector In response to more attractive alternatives elsewhere, either in the urban areas or In countrlos such as Nlgerla and the Cote d'lvolre, which enjoyed a decade of prosperity following the oil price rlse of 1974. This Interpretation Is supported by Thabatabal's careful analysis5. He argues persuasively that the declne In agricutural production In the 1970s was caused by the migration of between one and two million mainly rural workers, or roughly one-fifth of Ghana's population. An esthmatod 1.2 miUlon of these were expolied from Nigeria In 1983 (followed by a further 150,000 In 1985), causing a large Increase In the rural populatlon coincident with the severe drought of 1983. The large Increase In food production in 1984/5 (and the fall In real food prices) was a natural response to both the drought and the increase In labor supply. Konings (1986, p.117-122) argues, without quantitative evidence, that cocoa tarmers were increasingly shifting their efforts from cocoa production to food crops and other agricultural raw materials. Of the farmers he Interviewed In the 1979/80 season 79 percent Intended to expand food production, whilst the remalning 21 percent clahied that the switch was difficult because of transport and/or marketing difficulties. He also found that 59 percent of the cocoa farmers had abandoned weeding of cocoa on (at least a part of) their farms. 6 Hamd Thabatabal, "Economlc Declin, Access to Food and Structural Adjustment In Ghana", World Economic Programme Research Workhg Papr 10-6/WPSO, nternational Labor Organisation, 1986. - 2.28 - V. THE CONSEQUENCES OF THE DECUNE IN COCOA PRODUCTION The offects of the price policy on the icomes of cocoa farmers as a whole has already been Illustrated In Flgure 3. Thabatabal (1988) argues that this must have been an kmportant factor behind the out-migration of rural labor. One could argue that the pull of the oil-rich nelghbors was too strong to resist at any sonsible domestic cocoa price, but this Is beled by the Cote d'lvolre, whose higher cocoa prices durlng this period were associated with a continued growth In cocoa output (or at least, marketings), not wlth a redirection of thelr rural labor into non-agricultural employment. Konings (1986) reported that some cocoa tenants and workers had migrated to the Coto d'lvoire intendig to farm cocoa there. Figure 12 shows the division of the real purchasing power of cocoa revenues over two sub-periods -- the 1950s, the the of prosperity and the lead-up to the rapid expansion of output, and the last decade. The figure shows the real Income of the farmers (deflated by the Natlonal CPI), the reported real costs of the Cocoa Board, the Governmont recelpts from the cocoa export tax and the profits of the Cocoa Board (which, as a public agency may be conskiered as part of the revenue sources of the Government), and the profits on the forelgn exchange recelpts ('Forox Profit'). This last Item Is the dl'forence between the value of the forolgn exchange earnings at the PPP exchange rate and at the official rate (the earnings reported by the Cocoa Board). The are a measure of the potential surplus value available to the Government via the Bank of Ghana. The figur shows the relatively small size of the Board's costs In the earlier period, and the extent to which the prico policy stabilized the producer price, absorbing fluctuations In the world market price In fluctuathg Government revenues. During the second poriod the producer's share was squeezed by the fallig value of exports at the officlal exchange rate, the gradual increase In Board costs, and the attempt by the goverrnent to preserve Its own direct clabl on the nominal revenue (i.e., excluding Its Figure 12 GHANA Revenue from Cocoa sales 1977 Cedis/MT (Thousands) 20- 15 10 a 5 0. -5 50/51 52/53 54/55 56/57 58/59 75/76 77/78179/80 81/82 83/84 85/8B crop years Producers Government U Board Costs Forex Profit Forex profits relative to PPP ex rate - 2.30 - forex profits) -- an attempt which clearly became hmpossible at the end of the 1970s. What Is, however, striking, Is the extent to which the country (or some of Its members) was nevertheless enjoying unprecedented profits per tonne of cocoa exported during this period, most dramatically during the period of high world prices In the mid 1970s, but remaining higher in real terms throughout this perbod as compared wlth the earlier period. Figure 13 highlights the implications of this unrecorded source of profits, Itself reflecting the fallure of the Governmient to makitain the real value of the exchange rate. The figure gives the real value of recorded Government revenue from taxes on cocoa and from other sources, as well as the potential revenue which might have accrued to the Government had the Bank of Ghana auctioned off the cocoa recelpts at a constant real PPP exchange rate. The figure demonstrates the Importance of cocoa tax to the budget, and Its fall (along with other sources of revenue) during this period. What Is striking Is that the potentlal revenue dwarfs the nominal tax revenue on cocoa, and for the past decade has been larger than total Government revenue from all sources. Much, perhaps most of this hidden surplus was dissipated In either consumption or unproductive Investment -- certalnly there Is little evidence of much productive Investment. Part of It presumably created opportunities In Accra which must have strengthened the pull on rural out-migration, thus further worsening rural labor scarcity (at wages cocoa farmers could afford to pay, or those bIplled by the abusa tenancy system). Figure 14 examines the extent to which excessive Cocoa Board costs eroded the sums available to the producers and the Government. It takes as a benchmark the real board costs per tonne averaged ovor the earlier period In Flgure 12 -- these are the 'Standard costs'. The excess costs are the difference between the reported Board costs per tonne and this standard cost -- and in 1980/81 to 1982/83 this excess was negative, suggesting that the standard may well be on the high side. The figure shows GHANA Government Revenue Figure 13 Deflated by CPI Billion 1977 cedis 7 6 5 4 3 I 2 -1 -1 H ' I' I ' I ' I ' g i _ 69/70 71/72 73/74 75/76 77/78 79/80 81/82 83,84 85/86 Cocoa tax Non-cocoa revenue L Forex potential Forex = cocoa exports at PPP - official Figure 14 GHANA Revenue from Cocoa sales 1977 Cedis/MT (Thousands) 6 4 c) t -2 -2 I I I I I ' -1-- " | 1974/75 1976/77 1978/79 1980/81 1982/83 1984/85 crop years Producers Standard Cost LI Excess Costs Profits, taxes Standard Board costs are 1950-60 average - 2.33 - that this excess cost was, In several years, large compared both to the standard costs and to producer revenue, and again represents another avoldable loss to the economy. Several conclusions emerge from these last three figures. First, the problem of raling the price to producers without re4lucing the tax rate per tonne is best achleved by moving the exchange rate into line with Its equilibrium or PPP rate. Second, the most obvious next sourre of revenue either for the Government or the producers Is the excessive costs of the Cocoa Board. Though these were much smaller than the foregone foreign exchange profIts during the second period graphed in Figure 12, they will become Increasingly knPortant as the exchan.i rate regime Is liberalized and the discrepancy between the official and equiilbrium rate Is reduced. Third, It Is potentlally dangerous to specify a target producer price as a percentage of the fob price, as this leaves unspecified how the fob price Is to be translated into domestic currency. The temptation would be to use the exchange rate as a mechanism for reconciling the clalms on the surplus by the Cocoa Board and the Government. The Board might propose a cedi producer price which, givon prevalihg world cocoa prices and the current exchange rate, met the specified target percentage of fob, and the Government, anxious to raise revenue In the following calendar year, might devalue, thus raising the cedi export price of cocoa. Whilst this would be better than maintaining the exchange rate at an overvaled level, It would have the effect of raising domestic prices and reducing the real value of the producer cocoa price. This Incentive could be reduced by contracting to pay cocoa producers a given percentage of the finally realized export price, which In effect would require the payment of a bonus If the cedi were devalued (and possibly also If world cocoa prices huproved). The main objection to this system, apart from Its administratlve complexity, Is that a promise of future payment to cocoa growers Is less attractive than current payment. (The former 'chit' system was apparently very unpopular). On the other hand, It might serve the purpose of keeping the management - 2.34 - of the xcha rate la t', and of convkicb the coca rowrs of the good faith of the Cooa ard. - 2.35 - VI. OThER CONSEQUENCES OF EXCHANGE RATE UBERAILZATION A major component of the reform strategy Is the liberalization of the forolgn exchange market, with the consequent movement of the market rate towards Its equilibrlum level. As argued above, provided the Cocoa Board's costs are held down to historically feasible real levels per tonne (which, given the fall in output, will require a substantial reduction In total costs), and provided the cocoa exrport tax per tonre is not greatly increased in real terms, this should allow the real cocoa price to be substantially raised. It will also have effects on the prices of competing crops. Maize and rice are currently knported (as Is wheat, which Is not domestically grown). Salinger (1988) has demonstrated how sensitive the profitability of cereals production Is to the exchange rate and domestic pricing policies, and If we concentrate on maize (as being the more obvious mmedbate competitor to cocoa) the situation would appear to be as follows. World maize prices have fallen so much In real terms, and transport and handlng costs from rural areas to Accra are so high, that kuported maize enjoys a substantial comparative advantage In Accra at equilibrium exchange rates. At the moment knported cereals are subject to an knport duty and sales tax at a combined rate of 40-45 percent of the cif prico, whilst domestic cereals In practice do not appear to pay the sales tax. Thus cereals are quite heavily protectec (or would be at the equilibrium exchange rate). In addition knports have In the past been controlled, and the free market price appears to reflect the scarcity price of foreign exchange and the Inport tariff, so that the undervaluation of the exchange rate does not appear to have undermined the protective Intent of the cereals kport policy, If this protection were maintained as the exchange rate were llberalized, domestic maize prices might not change markedly, though the government would now receive the scarcity rents and hkport tariff revenue from cereal Imports. Domestic malze prices would remain high relative to world price levels and the maize/cocoa price ratlo at domestic prices would amplify the disincentives to cocoa production relative to maize production. - 2.36 - One solution to tnis would be to reduce the hnport tariff (and do facto tariff In the form of the sales tax) on cereals (or at least on maize). This would load to a fall In the market prico and farm gate price of malze, and would make cocoa relatively more attractive. It would probably load to a fall In the oquIlibrium prices of non-traded food crops like plantalns, yams, cocoyams, and cassava, for as farmOrs switched out of maize into the alterratives, thoe supply would rise. As consumers might be expected to substitute Into the now cheaper maize, demand for these non-traded crops would fall, and thus the equilibrium price would also fall, probably by about as much as the fall In the Price of maize. These tariff and price changos would have further repercussions. Government revenue would fall, as would the cost of livig (as the cost of food would have fallen). If the revenue shortfall were made up by Increases in taxes on other consumer goods, then urban consumers would be no worse off (and should be better off to the extent that non-cereal food prices will have fallen). Tho taxes on other consumer goods would fall partially on rural consumers, who would now be taxed twice over -- their food crops would command lower prices whilst their purchases of consumer goods will be more expensive. Whether on balance cocoa farmers were better or worse off would depend on how much the cocoa price were Increased, and the extent to which the other price changes affected them, but one would expect the effect to favor cocoa farmers relative to non-cocoa farmers, and certainly to encourage the production of cocoa relative to food crops. The next section attemptS to reach more precise policy conclusions and recommendations on setting the producer price of cocoa. - 2.37 - VIl. ISSUES IN SETTING THE PRODUCER PRICE OF COOA Increases In the real producer price of cocoa will have kIpacts on Government revenue, foreign exchange earnings, on the Incomes of cocoa farmers, and, Indirectly, on rural and urban wage rates and the prices of some food crops. At this stage, In the absence of a quantified model of the cocoa sector, one cannot be very precise about the magnitude of these effects, and the followig analysis Is therefore more a framework for dist ission than a set of well-defined and defondod recomnendatlons. - 2.38 - Vill. ALTERNATIVES TO THE COCOA EXPORT TAX The first questlon to ask is whether a cocoa export tax Is desirable, or whether cocoa farmors should be taxed either through an agricultural Income tax, or a land tax (ie a tax on presumptive Income). The arguments In favor of a land tax are sknple -- It Is a tax on an Inelastically supplied factor, which therefore causes no distortions provided It is levied on the unkiproved site vaiue (le on the value of the land before It Is cleared and prepared for cultivation). To the extent that land ownership Is correlated with wealth, It Is also an equitable tax, and so desirable on both grounds. The difficulties with the land tax are equally sinple. It will be unpopular amongst landowners, who frequently have the polltical power to block the tax, and It requires a cadastre (le a land register giving the unimproved value or potential rental of the land, Its locatlon and ownership). Such cadastres are costly to construct, and none yet exists for Ghana. Moreover, It Is not clear how high the unimproved site value of uncleared agricultural land Is, sinre the major cost of bringing land Into production Is the clearing and land preparation. If the land has a relatively low value then the tax base will also be small, whilst If the value of the land Is taken as Its uIproved value, then the tax will essentially be elther a labor tax (le a tax on the labor needed to prepare the land) or a tax on agricultural output, and In particular cocoa output (or potential output, le on trees). It Is therefore difficult to see a land tax replacing other forms of agricultural tax In Ghana, at least In the foreseeable future. If land taxes are not feasible, would an agricultural Income tax be a better a'ternative? In practice this would amount to a tax on gross output at crop-specific rates (the rates being adjusted to reflect notional Input costs). (Usually farmers are allowed to Itemize their costs If they can prove them to be higher than the notional allowance, but In practice this Is rare In countries which have a large farming population which they attemPt to tax by this method). It Is hard to hmagine food crops being taxed on gross output as opposed to marketed sales, and even here the taxation would - 2.39 - typicafly be done by nfluncwhg the prico of hIported compoting crops as It would be hard to monitor (and tax) even wholesale sales In rural areas. In practice, then, the tax would apply malnly to export crops or crops whose whole output Is handled by markoting boards, In which case we are offectivoly back to the type of taxation already employed for cocoa. The next argument to consder Is that, given the difficulty of taxing many agricultural crops, would It not be better to tax the goods that farmers buy (is consumer goods), rather than the goods they sell? In both cases the effect Is to turn the terms of trade agaist agriculture, and It might therefore be sensible to choose the least cost or administratively sriplest alternative. This argument Is compelling to the extent that It would have been deshrable to tax all crops at the same rate. If It were desirable to tax cocoa more heavily, and other crops at a uniform rate, then the same effect can be achieved by taxhg consumer goods and cocoa alone, exempting the other crops. The question then resolves Into whether there Is a case for a differential tax on cocoa. There are two basic arguments for a differential tax on cocoa productlon -- an optimal export tax argument, according to which Ghana has some Influence on the world price of cocoa, and a supply elasticity argument, according to which cocoa Is an attractive crop to tax because It has a relatively low elasticity of supply. These two arguments are examined below, and relate to the extent to which the producer price should be driven below the appropriate farm-gate price. The latter Is related to the fob or export price by the costs Incurred by COCOBOD In handling the cocoa from the buling station to the world market. If the producer prico Is to be specified as a percentage of the fob price, then It Is hportant to detormhe the appropriate margin for handling costs, which Is done In the next section. - 2.40 - DL COCOODS COSTS The World Bank has argued that COCOBOD should reduce Its costs to 15 percent of the fob price. Is this a reasonable target? The average ratio of costs In real terms to the 1950/51 fob price was 6.5 percent for the decade 1950/51 to 1959/60. The projected fob price In 1989 Is, however, only just over 50 percent of Its real value In 1950/51, and on that basis the real costs might be 13 percent of the fob, If COCOBOD was to achieve Its former level of officIncy. If some costs were now higher (because of the need for more disease control, or more research, or more extension), then the 15 percent would seem plausible. If one argues that there Is no reason for COCOBOD's real cost per tonne to change over tkne, then It might be botter to specify the target not In terms of the fob prIce, but as a real cost per tonne. This Is discussed further below. - 2.41 - X THE OPTIMAL EXPORT TAX The first argument for taxing cocoa Is that the world demand is inelastic, so that increases in Ghanaian exports of cocoa will tend to lower the world price. The elasticity of demand facing Ghana wiil depend on the supply response of the rest of the world, which will depend on the extent to which changes In the world price are passed through to other producers, the speed with which the producer price In other countries is adjusted to the predicted now lng run world equilibrium pric, and the speed with which this affects planting and ssWpply. In the short run It may be that other countries will not vary their supply, In which case the (short-run) elasticity of demand facing Ghana will be E/a, where E Is the world short run price elasticity of demand (given as between about 0.2-0.3, as a positive number, by inran and Duncan, 1988, Table 3), and a Is the share of Ghana in world production (roughly one-tenth). in the long run the rest of the world will presumably respond to the lower prico and reduce Its supply. If the long run elasticity of supply of the rest of the world Is n, and the long run demand elasticity Is e, then the long run net demand elasticity facing Ghana Is e/a + n(1-a)/a. lmran and Duncan give the long run demand elasticity as 0.4 (in Table 3) and the supply elasticities of most countries as about 0.3 In the short run and variously esthnated as 0.8-1.8 In the long run for different countirles on the basis of econometric estkiates, but perhaps as high as 1.5-2 (on the basis of rather little hard econometric evidence, see knran and Duncan p14-15). Imran and Duncan then estknate that the Ghanalan short run demand elastilcty might be 3.2, and the long run elasticity would be 12. The opthnal short run export tax (Ignoring other arguments for cocoa taxation) would then be 1/3.2 or 30 percent of the fob price, whilst the long run opthnal tax would be 1/12 or 8 percent. knran and Duncan, using a rather different approach which Involves balancing the advantages of the sWort run inelasticities agaist the lng run costs of excessive export taxes, actually esthuate an opthnal export tax of 21 percent, but this appears to rest of a shaky estknate of the social discount rate In Ghana. The figuro we shall use Is - 2.42 - somwhat arbitrargy "t at 12 prcont, c reflects some sacrifice of long run revenue for short run gain Anothwr way of looking at this Is that tho shadow value or accounting value of cocoa fob s 88 percent of the fob prico, and the shadow value of cocoa at the farm gate Is 88 less 15 percent (COCOBOD's costs), or 73 percent. A recommendatin to pay 55 percont of the fob price s thus to pay 55/73 or 75 percent of Its vahoe -- oquivalent to a tax of one third of the post tax prico (25/75). - 2.43 - XI. STANDARD ARGUMENTS FOR OUTPUT TAXES The remaining arguments for cocoa taxation Involve balancing the need for Government revenue against the cost of raising that revenue In terms of deadweight losses and adverse impacts on the distributlon of Income. The main arguments for taxing cocoa production are elther that cocoa farmers are richer than non-cocoa farmers, and should therefore be subjected to a higher rate of tax than can be achieved by the system of Indirect taxes, or that the elasticity of cocoa supply Is sufficiently low that the inefficlency losses make It an attractive source ot revenue. Appendix 2, entitled 'OpthTal trade taxes on agriculture In developing countries', gives a skiplified model of the economy (even If the mathematics appears somewhat complex!) which aliows one to characterize the optinal rate of taxation of cocoa, on the assumption that other agricultural products are not taxed, (and In particular that cereals are kmported duty free) and that non-food consumption goods are subject to Indirect tax. The technical analysis sets out the mathematical formulas and In the final section quantifies them for Ghana using the results of the Ghana Living Standards Survey and the Agricultural Economic Survey (AES). The discussion here will be confined to a non- technical summary of the main findings of that Appendix paper. The essence of the argument Is that cocoa taxes have to be compared with alternative methods of raising government revenue, and the assumption Is that these will have to be Indirect taxes on non-food manufactured (or knported) goods. The argument here Is simple -- Income taxes have a relatively lknited coverage and may already be set at unsatisfactory levels, while Indirect taxes fall on all consumers and thus have a reasonably broad base. It Is hard to tax domestically produced food and non- manufactured goods, which leaves the main cholce as specified. It Is possible to knposo tarlffs on cereal Imports and that Issue Is addressed beiow. The exact levels at which various tax rate will need to be set will depend on the amount of revenue required by the Government, and this Is not yet known. I therefore - 2.44 - proceed as follows. Cocoa taxes have been an knportant source of revenue In the past, and for this to continue to be true several things must aiso be true -- the Government has heavy requirements for revenue, and the costs or availability of alternative sources of revenue must be tight. I shall therefore make both these assumptions in order to see how strong the case for high cocoa taxes Is under conditions most favourable to a hlgh rate of cocoa tax. The assumption is that the Government has reformed the exchange rate system so as to capture the large foreign exchange rents generated under the previous system of overvalued exchange rates and licensing. In addition, taxes on knported and domestically produced goods that can be taxed are assumed to be taxed at an effective rate of 40 percent of their pretax price,7 and at this level the total taxes collected from Indirect taxes and the cocoa tax when the producer price of cocoa Is set at 55 percent of the fob level will be enough to cover planned government expenditures. These rates would be rather high If the exchange rate alignment took place, for the past situation has been to tax knports quite heavily but allow them In at an artificially low domestic price through the overvaluation of the exchange rate. If lower tax revenue Is needed then the case for heavy cocoa taxation will be further weakened. These tax rates are benchmarks to establish the level of revenue to be raised - - as the cocoa tax rate Is varied In response to alternative views about the elasticities and the distributional goals of the government, so the Indirect tax rate wll also need to be adjusted to ensure that the same total amount of revenue Is collec'ad. Thus lower cocoa taxation will raise the Indirect tax rate and ralse the marginal cost of collecting taxes by this method, and this will Inhlbit further reductions In the cocoa tax rate. Again, we are blasing the results towards making the strongest case for heavy cocoa taxatlon - 7 This roughly corresponds to the present tax system, though of course with an overvalued exchange rate the effective tax rate on knported goods will be lower than that. The actual tax rates are quite varled, belng for example very low on gasollne and knported cars. the effective tax rate can be thought of as the tax rate which has the same marginal cost of raising revenue as an equiproportbonal Increase of the current set of tax rates. - 2.45 - - If alternative sources of revenue can be found or If less revenue Is needed (or expenditure Is cut) then lower cocoa taxes will be Indicated. The other two assumptions are that 20 percent of extra Income Is spent on taxed goods, and that the (compensated) price elasticity of demand is 1.25. These assumptions are also discussed In the section on the taxatlon of consumer goods below. Again this assumption biases the case In favour of high cocoa taxes -- If the price elasticity of their demand Is less than 1.25, then the optinal cocoa tax rate will be correspondingly reduced. Perhaps the most useful way of structuring the discussion is to examine whether the World Bank's recommendation of setting the producer price of cocoa at 55 percent of the export price Is soundly based as a formula, and whether the Implied level of cocoa prices Is roughly right. Here the work of Meeraus, Asenso Okyere and O'Mara is particularly relevant. Their finding was that below a critical level of about 140 (1987)cedis/kg there would be little Incentive for the farmers to plant cocoa. Above this level there appears to be a strong response to price Incentives, summarised by a medium to long run price elasticity of supply of one or higher. The AES survey data shows that there was considerable replanting over the past four years once prices began to Increase, though we do not know what price expectations the farmers then held. If they based their planting decisions of current prices, In the full knowledge of the effect of Inflation on their eventual return, then the critical price below which planting will not occur may be somewhat lower than 140 cedis/kg, but If they planted In the expectation that the Government would continue to attempt to raise the real price of cocoa, so that they were looking ahead to more promising prices, then the figure might be an underesthnate. For the moment let us take It at face value. The formula for the optimal tax on cocoa In the Appendix knpiies a ratio for the producer to the export price which appears to be Independent of the world price level, lending superficial support to the World Bank formula. In fact the tax rate Is particularly sensitive to the supply elasticity, which rises rapidly as the real domestic price falls towards the critical level (of 140 - 2.46 - cedis/kg). If this Is taken Into account then the domestic price should be bounded below by tnis critical level. If, as seems to be likely over the near future, the world price of cocoa Is such that 140 codis/kg represents more than 55 percent of the export price, then this lower bound will Imply that the Bank's recommendatlon Is too low, and the price should be accordingly adjusted. But what If 55 percent of the world price Is above this critical level? And Is there a case for raising the prico above the critical level even if the export price Is low. Finally, how should the producer price respond to varlatlons In the expected world price, for example, as forecast by the World Bank's Commodity Divislon or by other consultants such as Bateman? Here the Ghanaian survey data throws useful light on the standard problem facing the choice of any tax. The level of taxes on any particular commodity or Income source Involves a balance between the need for revenue, the distortlonary costs Introduced by the tax, and the distributlonal consequences of the tax. A tax on cocoa must be compared with alternative ways of ralsing the same revenue -- In particular, we need to ask whether it would be better, on balance, to Increase cocoa taxes and reduce other taxes, or, conversely, to reduce cocoa taxes and Increase other taxes. For present purposes, the alternative source of tax revenue Is considered to be a general increase In taxes on non-food consumer goods produced In the modern sector or kIported. It may well be the case that particular goods are even better choices as bases for Increased taxes, and If so, that will further strengthen the case for switching away from cocoa taxes to these alternatives. If the Goverrunont Is concerned with the distributlon of income, and believes that taxes should fall more heavily on richer members of society, then much will depend on how egalitarian the Government Is. In the technical paper this Is captured by a 'coefficient of Inequality averslon', which may take the values 0.5, 1.0, or 2.0. To take the central value of 1.0, this kupiles that the social cost of a tax of S1 on a person whose annual Income Is $5,000 Is considered to be twice that of a person with twice his Income -- - 2.47 - $10,000. At an inWuallty aversion of 0.5 the social cost Is 2 raised to the power 0.5, or 1.414 thies as costly as taxing $1 on tho $10,000 Income, whl, at an Inequality aversion of 2.0 the cost Is 2 raised to the power 2 or 4 tInos as much. The consensus Is that the range 0.5-2.0 covers most plausbie attitudes to hequalilty, with 0.5 perhaps somewhat hegaltarlan and 2.0 behg rather progressive. An Inequality aversion coofficient of 0 would Imply that the Government attached zero weight to redistributing income, or, equivalently, considered the cost of raisig $1 in tax to be the same whether pald by the person with $5,000 or $10,000. Once we have agreed upon a way of weightin the costs and benefits of Income received or paid by people of difforin kicome, then It Is possible to describe the distributional Impact of changig pricos In a particularly simple way, by caiculating the distributional characteristic of the commodity whose prico changes. Table 1, reproduced from the Appendix, givos the values of the distributional characteristics of the Key commodities relative to that of cocoa In production. (At an hequality aversion of 0, the characteristics would all be 1.0). The table Is Interpreted as foliows. Consider the second llne, which gives the distributional characteristics of maize (good 2 In the paper) in production (suporscript p for producers). If the price of matzo to producers were to be reduced so that $1 was transferred from farmers to the government with no other change In the economy (in particular, no changes In production or consumption or Table 1. Distributional Characteristics for Ghana Values for inequality aversion v Variable Commod v = 0.5 v = 1.0 v = 2.0 dP cocoa 1.00 1.00 1.00 1 dP maize 1.13 1.36 2.90 2 d2 maize 0.93 0.91 1.03 dP cereals 1.12 1.36 3.18 d2 cereals 0.90 0.87 0.97 dP NT food 1.11 1.30 2.40 3 d3 c NT food 0.89 0.83 0.87 d4 mfg goods 0.91 0.88 0.99 d s lab sold 1.01 1.09 1.69 de lab bought 0.98 1.01 1.86 - 2.49 - revenue from other sources) then, looking at the second column for an Inequality aversion of 1.0, the number 1.36 means that thJ soclal cost of this transfer is 1.36 tknes aa high as were the $1 to be collected by reducing the price of cocoa and transferring S1 of Income from cocoa farmers. This I: because maize growers are somewhat poorer than cocoa farmers. On the other hand, looking at the row Immediately below this, which refers to maize In consumption (suporscript c for consumers), the cost of transferring S1 from malze consumers to the goverrinent Is only 0.91 thies as high as making the same transfer from cocoa farmers. Shnilarly, and more to the present point, the cost of transferrlng $1 from consumers of manufactured goods Is only 0.88 tines that of the cocoa transfer. If changing cocoa taxes and taxes on manufactured goods only resulted In changes In the prices of cocoa and manufactures respectively, then the relative advantage of the two taxes would be reasonably simple. If the elasticity of supply of COC03 Is lower than the elasticity of demand for manufactures then on efficiency grounds cocoa would be a preferable tax base to manufactures, but conversely on equity grounds. But changing the price of cocoa changes the Income of cocoa farmers which in turn reduces their expenditure on taxed goods, and reduces the revenue ralsed, and this effect weakens the case for cocoa taxation on efficiency grounds. 3econd, changing the price of cocoa lowers the returns to farming cocoa and eticourages farmers to switch into other non-traded crops, whose price will then fall as their supply Increases. This will redistribute Income from poorer farmers to the richer consumers of non-traded food -- the differences In the table between the distributional characteristics of NT (non-traded) food In production and consumption are quite pronounced. Rural wages will tend to fall (in money terms) as returns to farming fall, which will redistribute Income from sellers of labour to buyers of labour. All these effects make taxing cocoa less attractive on distributinal grounds, so the final choice will depond very much on the relative elasticities (which will affect the - 2.50 - officlency aspects) and the weight placed on distrlbutn. If the Government Is completely neutral as to Income distribution (an inequallty aversion of zero) and If the long run cocoa supply elasticity Is one (corrosponding to a producer price quite close to the critical level) then "he optknal ratio of the producer price to the export price would be 55 percent. If the supply elasticity were as low as 0.5, then the optimal producer price would be 47 percent of the export price. The effect of changes In the world prico of cocoa are quite interesting. Suppose that the best long run estimate of the fob price Is 270 cedis/kg, so that 55 percent of this Is 150 cedis, at which polnt the supply elasticity might be quite high, perhaps 1. At this elasticity, ignoring distrlbutional Issues the right producer price Is Indeed 55 percent of the fob price, and the correct amount of revenue Is raised. Now suppose that the (long run) world price rlses by 40 percent to 380 cedls/kg, so that 47 percent of this Is about 180 cedis, at which polnt the supply elasticity might be 0.5, and again we appear to have found a consistent soiution. The effect of the fall In the supply elasticity with rising domestic producer prices has been to Justify higher cocoa taxes which tend to reduce the response of the producer price to changes In the (longrun) world price -- in this case the producer price only rises 20 percent In response to a rise In the fob price of 40 percent. But thero Is another force at work tending to make the domestic producer price track the fob price more ciosely. The tax collected from cocoa will Increase substantially as the world price rises, and this will reduce the need for Indirect taxes, Justifying a lower Indirect tax rate 'knd with It a lower cocoa tax rate -- so the producer price should In fact be somewhat higher than 47 percent of the fob price. Thus the change In the supply elasticity with Increases In the domestic price means that the domestic price should respond loss than proportionately to fluctuatlons In the world price, even ignoring cases for stabilizing domestic producer Income. (This case will be considered separately beiow). Above a world price of 400 cedls/kg provided the supply elasticity remains roughly constant at 0.5, the ratio of the producer to the fob - 2.51 - price would tend to stabilizo at 47 percent, again, Ignoring any distributional consideratlons, except that as the cocoa tax rovenuo rises with the fob price, so all taxes can be scaled down somewhat, and the producer prico will now tend to move more than changes In the long run world price. 1' the Government Is concer !d about the distrlbution of Income, and if we take the central value of Inequality aversion of 1.0, the optimal rati or producer to export price would be 64 percent at the high supply elasticity of 1.0, and 56 percent at the lower elasticity of 0.5. (Both calculations assume as before that at a producer price of 55 percent of fob the Indirect effective tax rate raquired to balance the budget Is 40 percent. Thus If the producer price Is set at 64 percnt of the fob prico the indirect tax also has to be ralsedj. At an Inequality aversion of 2, the producer prico should be between 63 and 74 percent of the export price (deponding on the supply elasticity), at least according to the formula. But It Is Important to recognise that the parameters upon which these estimates are made might change If the balance of taxation changed as dramatically as this between cocoa and consumer goods, and all that should be Inferred Is that If the Government Is very concerned with Income distribution, then In effect It should be shifting the balance of taxation away from Its concentration on rural producers towards the more broadly directed consumer taxes which fall on everyone, both rural and urban. The reason why a concern for Income distribution can make such a difference to the level of cocoa taxation lies In the Indirect effects which changing the cocoa price has on the prices of non-traded food stuffs. Raising the price of cocoa also raises the price of non-traded foods like yams and cassava, and thus Increases the Incomes of farmers who produce such crops. Since farmers tend to produce roughly similar amounts regardiess of their Income level, whilst richer consumers are more Ikely to substitute cereals for roots. The effect Is roughly simlar to a proportional icome tax which Is returned as a limp sum to producers -- a rather egalitarlan kind of income transfer. - 2.52 - Taxhg traded manufactures Is also rather like a proportional income tax and so Is a good way of financing the shortfall of revenue which reducing the cocoa tax would cause. As explained In the Appendix, the calculations of the optimal producer price of cocoa assume that the tovernment has decided to follow a policy of free trade In cereals. Given that cereals are currently subject to a 30 percent import duty, this raises two questions. Should cereals be protected by this tariff, and how does the presence of this tariff affect the level at which the price of cocoa should be set? The first questlon Is addressed In the technical Appendix, and the answer Is fairly unambiguous. Regardless of attitudes to Inequality, there Is a case for ralsing some revenue from cereals Import dutles If this allows taxes on other goods (either cocoa or currently taxed representative consumer goods -- though not necessarily particular consumer taxes on eg gasoline) to be reduced. The reason Is simple -- tariffs raise revenue, and the effect In encouraging cocoa farmers to switch Into maize production and hence lower cocoa tax revenue is relatively slight, at least at low levels of cereals duty and adequate levels of the cocoa price. In additlon tariffs redistrlbute Income from consumers to cereal farmers. Further, by inducing a switch Into cereais and out of non- traded roots, the price of roots rises and further red:strlbutes Income to the poorer farmers. How high a tariff on cereals can be justlfled on these grounds Is hard to say, for the higher the price of cereals relative to cocoa, the greater will be the loss In cocoa revenue for each extra cedl collected In cereal import duties. The second question Is nut treated formally. Given that cereals are currently subject to Import duty, an increase In cocoa tax (le a fall In the cocoa producer price) will Induce farmers to swltch Into other crops. This will tend to drive down the price of non-traded roots, and so encourage at least part of the resources released from cocoa farmhg Into cereals production. The rise In the price of roots cause by the fall In production will Induce consumers to switch from roots to cereals to some extent. On balance It Is not clear whether the Wicrease In cereals consumption will exceed or fall - 2.53 - short of the Increase in cereal production, and It Is notoriously difficult to measure cross-price elastIties either In supply or demand to answer the question. Taking an agnostic stance, suppose that there were no change In cereal knports caused by an Increase In cocoa tax, there would then be no offect on cereal tariff revenue, and so the earlier analysis would stand. - 2.54 - Xl. QUAUFICAT1ONS The model from which this calculation derives Is not only highly simplifled, but also static, both In modelling the demand for Ghanalan cocoa and the domestic supply response. The first point has already been noted, the second Is potentially more serious. The co..oa supply response consists of two distinct components. The first is the output response of the existing stock of cocoa trees to higher prices. Farmers will intenslfy management (weeding, spraylng etc) and produce more cocoa per tree at higher prices. This productlon response will be rapid, and a plausible estimate of the elasticit) of output per tree might be 0.25. It might well be higher If one recognizes that what Is relevant Is the response of sales to prico, and higher prices will discourage smuggling, raising the sales elasticity above the output elasticity. If, as seems to have been the case In Ghana, the elasticity of labor supply to agricunural (and therefore cocoa) productlon Is quite high (alowing for migration, as well as the number of hours worked per manyear), then again this short run elasticity will be further ralsed. Farmers will also Increase the number of trees If the price Is attractive, and this area response will lead to Increased production with a lag of 4-5 years, with a steady build-up thereafter. It Is more difficult to estimate this iong run response, but the model of Meeraus, Okyere and OMara suggests that It might be qulte high -- perhaps more than 1.0 at low producer prices, though lower as the price Increasingly exceeds the cost of planting and productlon. Again, a high labor supply elasticity will raise this figure, and the elasticity will be higher If the prices of competing food crops are kept down by a liberal Import policy. 1, grain Imports are restricted, or tariffs raised on cereals In an attempt to Increase food self-sufficiency, then cereals prlces will rise, leading to sympathetic Increases In non-traded food prices, and the relative attractiveness of cocoa planting (and tending) will be lower, reducing the supply response. The cocoa supply response thus has two components -- a short run response with an elasticity perhaps In the range 0.25-0.5, and a long-run area response, taking - 2.55 - perhaps 8 years or more, but perhaps having a considerably higher elasticity. If the goverrvnent could credibly announce a future price which would prevail In four years tkne, then the optinal producer price would be one that used the short run supply elasticity to calculate the price for the first four years, and thereafter used the full supply response (the sum of the short-run and area elasticities). Farmers would therefore plant more trees know, knowing that the current low prices would be replaced by more attractive prices Just as the newly planted trees first started bearing cocoa. It Is, however, difficult to kuagine that this would be a credible strategy In Ghana, given the experience of cocoa farmers over the past two decades. If farmers are to be persuaded to continue to plant now trees, and to tend to their existing recently planted trees, then they will have to be given evidence that the price of cocoa will be attractive, and that In practice means raising the price now. It Is possible to show that the best constant producer price Is a compromise between the low initlal price (based only on the short run supply elasticity) and the higher long run equlilbrium price (based on the full elasticity). At this stage, given the present state of knowledge about likely planting responses to the cocoa prico, all one can say Is that the producer prices should be above, and possibly substantially above, the price at which the supply response Is large, which Is near the cost of productlon, and about 140 (1987) cedis/kg. - 2.56 - XiII. THE RISKS INVOLVED IN SEMNG THE PRODUCER PRICE TOO LOW The argument in favor of a producer price at least as high as 55 percent of fob price, and at any rate above the critical supply price of about 140 (1987) cedls/kg, may look somewhat fragile, and once can knagine arguments for elther a higher or (less plausibly) a lower price. Granted this, one would like to know the risks Involved In setting the price either too high or too low, to see on which side It Is safer to err. If the price Is set too high, the obvlous loser Is the Government, which will lose tax revenue. This would be serlous If there were no alterative sources avallable. Granted that revenue is tight, there are several mitigating factors which make this a lower risk strategy than might appear. The first Is that provided the foreign exchange market Is liberalized, the potential extra effectivo revenue flowing to the government sector will be significantly Increased, compared with the recent past. In particular, this will tend to Increase the real value of the cocoa tax receipts, since cocoa exports will now earn the market, not the overvalued, rate of foreign exchange. The second Is that the extra Income transferred to cocoa farmers will be spent ulthiately In considerable part on taxed goods (or services) and will In part return to the exchequer. Flnally, the foreign aid situation Is likely to be relatively more favorable In the earlier period of reform and adjustment than later. A shortfall In cocoa tax revenue which has been urged by the Bank and the IMF Is likely to be more sympathetically treated than Is a subsequent shortfall In foreign exchange earnings caused (or thought to be caused) by a failure to stknulate increased output by failing to pay the agreed price. The risks of setting the price too low look more serious. The most obvious problem Is that It will take some time before the ovidonce -- In the shape of a failure to inrease production and exports -- becomes available to demonstrate that the price was too low. By the tkne that It does, It wlH take a further 4-8 years to reverse the damage, and the patbnce of internatinal creditors will beome strained. There are other poiltical economic reasons why a low producer prico strategy may have harmful - 2.57 - consequences. Raising the producer prico puts prosswe on COCOBOD to cut costs, and on the Government to reduce wasteful expenditure. The argument for taxing cocoa (or anything else) is that the revenue can be bettor or more productively spent by the government than by the taxpayer. This argument has been demonstrably false In the past, and without strong Incentives to hprove offlciency and cut waste, it will continue to be false. The lower Is the value of government expenditure, the lower should be the level of taxation, and the higher should be the producer prico of cocoa. At this point it Is worth brnging out another hportant finding from the GLSS and AES survey data and the technical appendix. It appears that cocoa farmers are rather poorer that the typical consumer of knported or taxed goods, and that producers of non-traded foods are very considerably poorer than consumers of these foods. Given this, it Is, on distribuclonal grounds, desirable to transfer revenue from consumers of food to farmers, and also from consumers of taxed goods to cocoa farmers. Taxing cocoa farmers does the reverse on both scores, for It directly taxes the Income of cocoa farmers, and Indirectly lowers the Income of food producers and transfers Incomes from these farmers to food consumers. - 2.58 - XIV. IMPUCATIONS OF TAXING COCOA FOR INPIUT SUBSIDIES If there is a case for taxing cocoa output, then there Is a corresponding argument for subsidizing cocoa Inputs, so that the relative price ratio of Inputs to outputs facing the farmer Is not moved too far from the efficient price ratio. It can be shown that If the elasticity of substitution between hputs which can be subsidized (like Insecticides, fungicikes, sprayers, hybrld seods and tfrtlizers) and those which cannot (labor, land) Is unity (l the production functlon Is Cobb-Douglas), then the optimal Input subsidy should be set at the same rate as the output tax. If It Is argued that the producer price of cocoa Is set at 55 percont of the world prie, then this corresponds to a ratio of producer price to farmgate vakhe of 55/73 or 0.75, where the farmgate value Is 73 percent of the world price (equal to the fob price less the optiinal export tax of 12 percent and a COCOBOD handlhg cost of 15 percent of fob). The corresponding sales price to the farmer for Inputs Is 75 percont of cost of delivery, equal to the ratio of the cocoa prico to Its farm-gate. If there are boneficial externalities In using disease prevention measures (le If other farmers benefit) then there Is a case for further subsidy, otherwise not. - 2.59 - XV. SPATIAL VARIATIONS IN COCOA PRICING The argument that the optimal output tax on cocoa is positive given above refers to a tax rate on the farm-gate value, Itself equal to the fob price less handling costs. These handling costs will vary depending on the location of the farm, and will presumably be higher In more remote areas. If one were to Interpret the choice of tax rate literally, then the producer prico should be set at different levels In different places, though at the moment the producer price Is uniform across the whole country. There are two reasons why the full handling costs should not necessarily be borne by the farmer. The first Is akin to the argument given above that It Is desirable to subsidize Inputs, in this case the transport and handling services necessary to produce the final output -- In this case, cocoa for export to the world market. This would argue for making the cocoa price equal to 65 percent of the fob price less handling costs at each location, and would not make the price uniform (though there would be less variation than If the handling costs were deducted from the net of tax farm-gate price). The second argument is that the elasticity of supply might vary systematically wlth location, belng higher In more distant or costly locations, arguing for a higher producer price relative to farm-gate value. The main reason for this would be that more distant locations are closer to alternative marketing outlets in nelghboring countries, which make smuggling more attractive than In the more central areas. If this Is true, then a uniform producer price might be Justif led as the sknplest compromise designed to reduce smugglng. Whether It Is true or not Is not clear. - 2.80 - XVI. PRICE STABIUZATION The remaining Issue to resolve In setting the producer price of cocoa Is how to respond to medium run variations In the world price level. (Short run fluctuations are already dealt with by announcing and keeping to one price for the forthcoming season, though, as remarked in the main paper, It is Important to adjust the domestic price In line with exchange rate changes even wlthin the season to achieve a constant real price of cocoa). There are two separate Issues. The first Is how quickly to adjust the domestic price In response to a permanent change in the expected level of world prices -- caused, for example, by a permanent switch (or at least, a long perlod switch, In the level of supply of other producing countries), and how to respond to a deviation over the next year or so from the expected long run level. In the case of a permanent change, the sooner the domestic producers are confronted with the correct producer price corresponding to the world price the better. If the adjustment Is downwards, nothing Is gained by postponing the downward adjustment of the domestic price, as the short-run supply elasticity will be lower than the longrun elasticity, arguing for the shortrun exploitation of the revenue potential of lowering prices (assuming that the longrun consequences of this price change are what Is desired). If the adjustment Is upward, although this argument might appear to work against a rapid Increase In the price, the long run supply response Is unlikely to begin unless current prices Increase. In the case of a shortrun deviation of the world price from trend, there are good reasons for buffering the domestic producers from these transient shifts. MIrrlees (1988), In a recent paper for the World Bank, estimates that In such cases It would be appropriate to take a five year moving average of export prices (including futures prices In the formula). In other words, current producer prices should only respond 20 percent to changes In the export price level. Mirrlees assumed no correlatlon between price and supply, and so his case Is somewhat overstated, If there were a strong correlation - 2.61 - between Ghanaian and other West African cocoa supply, then prices would be high when supply were low, and there would be a case for allowng domestic prices to rise In sympathy with world prico movements to offset the effects of low output. Whether this could be accomplished within the present system of payments la not clear, though It might be possibe to pay a bonus In years of below average total supply. U4 - 2.62 - XVII. IMPUCATIONS FOR TAXES ON CONSUMER GOODS The opthnal producer price for cocoa was derived on the assumptlon that consumer goods were taxed 40 percent on their cost (or Import price), and faced a (rather h!gh) compensated demand elasticity of 1.25. ' It also assumed that the Government has iberalized the exchange rate reghme so that the officlal exchange rate Is cose to the equilbrhkm exchange rate. (If this Is not true, then It Is knportant that the cocoa producer prkce be set at a level which takes account of the overvaluation of the exchange rate, but even then the effective rate of tax on consumer goods will be lower than the nominal rate If the exchanger rate remains overvalued. This In Itself strengthens the case for ralshg consumer tax rates and lowering the effective taxation of cocoa exports). If there are consumer goods with a lower demand elasticity, or which are taxed at a lower rate than 40 percent, or which have a higher than average income elasticity of demand, then there Is an equally strong case for switching tax revenue from cocoa farmers to such sources of taxation. The most obvious examples are cars and petroleum products, which, though currently subject to excise taxes, certalnly meet most, If not all of the above conditlons. If we concentrate on transport fuels (gasoline and diesel), then most, perhaps all, of the current tax Is better considered as a user charge. As a rough guide, the road user charge element of taxes on transport (of which fuel taxes are a major component) should yleld revenue comparable to expenditures on the road network. Specifically, taxes on heavy vehicles should cover perhaps two-thirds to three-quarters of the maintenance expenditures required to keep the road network In Its desired long-run state of repair, whilst taxes per passenger car equivalent (a measure of congestive capacity) should recover the balance of malntenance expenditures and a substantial fraction (between two-thirds and the full amount) of the Interest on the capital stock of the highway system. This last Item Is typically well below the annual Investment In building or reconstructhg roads, unless the highway system Is expanding faster than the rate of Interest. If the highway system has deteriorated to the point at - 2.63 - which substantial reco,structlon expenses are reuirod, then higher road aser charges can be justifled In the period before this reconstructlon Is completed. The pure tax element corresponding to the 40 percent Indirect tax rate would be In addition to this road user charge componmnt. Ideally, the tax would be restricted to fall on passenger transport, leaving frelght transport untaxed, as It Is an Input Into production. in practIce this means concentrating the tax on gasoline and passenger vehicles, with dissuasive taxes on diesel-engined private cars designed ta make it unattractive to buy these Instead of conventlonal gasoline powered cars. Such levels of gasoline taxation would be appreciably higher than current levels, though probably below European levels. Given the administrative ease with which such taxes can be collected, and the difficulty the consumer has In Identifying the tax element given that fuel prices change In response to changes In the world market price and the exchange rate, It Is hard to think of cogent arguments against shifting revenue collection away from cocoa farmers towards gasoline taxes. - 2.64 - xIn. CNCLUSIONS The case for announchig and demonstratlng a commitment to a higher real producer price for cocoa appears strong, and the risks of setting the price too low appear to be higher than the rlsks in the other directin. The first priority in setting the producor price of cocoa Is to ensure that It Is above the level at which farmers are willing to plant and maintain cocoa. While this seems self-evident, It has not been the past polIcy of the Government, and It would not necessarily be ensured by llnking the producer to the export price level. The main way In which the fall In tax revenue from the cocoa tax and the profits cf COCOBOD can be offset are first of all to move the exchange rate to Its equilibrium level so that the Goverrmnent will have access to additlonal purchasing power which Is significant compared to total current tax revenues. The next step Is to explore the extent to which taxes can be shifted from cr- ;a producers to all consumers by Increasing taxes on approprlate consumer goods. The Ideal subjects for such taxation are those which are Income elastic and price Inelastic - - gasoline, cars, onsumer durables, etc are all good examples. The greater the scope for consumer good taxation, the hlgher should be the price of cocoa. There also appears to be good reason to try and stabilize the real producer price of cocoa, both because the tax rate legitinately Increases with tne world price, and because on Income smoothing grounds price stabilization Is moderately efficaclous. Finally, even If there does remain a (rather modest) case for an output or export tax on cocoa, there seems little point In moving to the more complex alternatives of an agricultural Income tax or a land tax. - 2.9 - RAEEBCES Imran, M and R C Duncan, (1988?). 'Opthal Export Taxes for Perennial Crop Erporters', mhoeo, Internatinal Comfodltles Market Division, World Bank, Washington DC. Konings, P. (1986). The State and Rural Class Formation hI hana: A Comparative Analysis, KPI (distr. by Routledge and Kean Paul, London). MIrrlses. J. A. (1988), 'OptUal Comodity Price hntrventIon", _beo, World Bank, August. Thabatabal, Hamld, (1908) 'Economc Decline, Acces to Food and Structural Adjustment In Ghana'. World Ecoromc Programeo Rosech WorkIng Paper 1O-G/WP8O, hnternatIonal Labor Organisatlon. APP_DOt A A NOTE ON DATA IN COCOA TAX AND REVENUE ALTERNATIVES 4 by David N. Nwbry APPENDIX A A NOTE ON DATA Pag I of S Tax analysis and the study of agricultural price responses In Ghana is fraught with a number of serious data problems which should be borne in mind when interpreting the evidence presented here. First, the rate of Inflation has been high and variable for the past two decades, often over 50 percent per year. This presents particular problems In deflating the price of cocoa, which has a crop year running from September to August. The approach adopted here is to assume that the bulk of the crop Is sold near the end of the calendar year, and to take the geometric mean of the adjacent years' price index as the deflator. The next Issue Is which, price Index to use, as between the national CPI, the rural CPI, the urban CPI, or some other. Flg 1 shows the evolution of the first three since 1963 (before that only an urban CPI was available). The rural CPI follows the national CPI qulte closely, and since the emphasis Is on national Issues such as tax revenue, or rural issues such as the real returns to cocoa producers, the national CPI seems appropriate. Fig 2 gives the movements in the rural food and non-food Indexes relative to the national CPI, and shows the rise In tlbe relative prico of food to non-food in the period 1977-1983, though otherwise tne two series move In parallel. The next problem to confront Is the failure of the government to adjust the exchange rate In response to rapid domestic inflatlon. The exchange rate consequently became Increasingly overvalued. Intermittent large devaluatons make estknates of the world price of agricultural commoditles In domestic currency particularly unstab:e and untrustworthy. The solution adopted was to convert International prices Into domestic curency ushg a purchasing power parity exchange rate (or Opp rate), computed by dividig the CPI (Consumer price idex) by the Index of manufacturing unit value (MLV), which the World Bank uses to deflate commodity prices. It Is a measure of the purchashg power of dollars over hkports of manufactured goods Into APPMNDIX A A NOTE ON DATA Page 2 of 6 developing countrlos. It might have boen better to use an idex of non-food prices, though sufficiently long tike series were not readily available. The other reason for preforrlng the CPI Is that domestic prices were deflated using the CPI, and the international prices were then deflated using the intrnational MIJN idex, allowing comparons to be made with other countrlos (particularly Coto d'lvore). The excharge rate was assumed to be in equlibrkAm In 1950 for basing the PPP caklulatns. Two other hdicators of the deUee of overvaluatlon are available - the paralle or 'black market' exchange rate, and th 'oqulilbrum' exchange rate. The former measures the cost of obtaing dolars for margial unauthorized transactions, the latter attempts to estuat the market clan excivinge rate, and Is a simble average of the parallel and offical, rate. Fig 3 shows the ratio of the official, parallel and 'equllbrhiu' exchange rates to the PPP rate. It Is clear that the aralel and hence the equibrhAm rate are both very unstable over the past decade, but the thre yer movng average of the equUIbrkAm exchange rate does not suggest any very significant divergence from the PPP rate calculated, londing support to the use of the PPP rate. The main remainig problm In caulating the real fob cocoa price Is the choice between the varous figuros for the reported export revenue, which froquently differ by significant amounts, the choice of the appropriate date for deflating, and reconciling these figures with the world cocoa price (where the World San's commodity price series for 'Accra cocoa was used.) The problem wlth using the reported export revenue figuros Is that the date of export Is unknown, and In som ease the exchange rate changed by a significant amount during the year. With an Inflation rate of 50 - 60 percent per yer, an error of tring of six months might Introduce an rror of 30 percent nto the estimate of the real vahle of the fob price. The sliper solution Is to use to reported real wod pric, and on the assumption that the export sales are centered on March foNg the start of the cocoa year, the vahle used for the crop APPENDIX A A NOTE ON DATA Pae 3 of 6 year 19xO-19xl was taken as 0.25 x world price hI year 19xO + 0.75 x world price hi year 19xl. Athough there are several reported vakoes for the fob prico of cocoa for later years, at least the producer price Is wel-defhned (hi current cedi). This Is far from true for other agricultural crops. The reported 'producer prices are yearly and country wido averages of wholesale prices hI bcsal markets. The variation over the course of the year at any one place Is large, as Is the spatial variation at anyone date, whit the transport costs from the farm to wholesale may be substantiaL Agahn, the roevant price for the farmr chooshig whether to grow cocoa or other crops Is the farugate prko at the date of harvest, not the aunual average price. AN that one can hope Is that the relatlonhi betwn the farugate price and the reported price. remahs roetIvely stabl over tiss. so that gross movements hI relative prices gve a rough Idea of movements hI farumgate relativo prices. Figure Al GHANA Real price indices 1977=100 120 - 100__ _ _ _ _ _ 80 60 - 40 - 20 - 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 Ye ar z 0 Rural CPI/national # Urban CPI/National Rural/urban prices eOX ORIIR a o > Figure A2 GHANA RURAL PRICES Deflated by National CPI 1963=100 250 200- 150 100 50 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 Year z 0 Pural food Rural non-food Food/Non-food #2o^V No~ Figure A3 GlIHANA Ratio of exchiange rate to PIP rate Index 1950=100 250 200- 150 100 50 1950 1955 1960 1965 1970 1975 1980 1985 Year z 0 Official - Parallel 'Equilibrium' 3yrMA equilibrium {OX PPP is CPI/Manufacturing Unit Value %.X Ill. GLOBAL PERSPECTIV ON COCOA SUPPLY AND DEMAND by imsr1 J. ktem 1. INTRODUCON From the early 1900s to the mid-1970s, Ghana was the world's leading cocoa producer with market share ranging from 30 to 40 percent of the world total. In the ten-year period, 1976-1985, however, Ghanaian cocoa output fell from 394,000 to 159,000 tonnes--a 60 percent decline. A major decline In real farmer prices plus El Nino weather In West Africa (1976-77, 1982-84) were responsible for the drop. Currently, Ghana's cocoa production accounts for 10-11 percent of world output. Recently, the Government of Ghana has Increased the price Incentives to cocoa farmer. Discussions with Ghanalan cocoa farmers and other experienced cocoa people In the outside of Ghana suggest that the Incentives have Induced some new cocoa planting, better farm malntenance and that cocoa output will Increase durlng the next few years If the current Incentives are maintained. I The purpose of the following analysis Is to examine the opportunitles that exist for Ghana to Increase foreign exchange earnings from the cocoa sector during the next 15-20 years; to determine the costs of cocoa productlon In Ghana; and to suggest an appropriate pricing policy that the Government of Ghana should pursue to revitalize the Industry and capitalize on world market opportunities. The study begins with an analysis of the world cocoa market Including projections of world production for 1988-2000. Cocoa hectarage and yleld Information for a number of major producers Is used to project output. World cocoa consumption growth patterns are analyzed and a world cocoa consumption model Is developed which forecasts usage based on world cocoa prices, rates of Inflatlon, world Income, etc. World cocoa stocks are then projected with and wIthout major weather problems. Finally, world prices are projected to the year 20(X based on a cocoa price model which has also been developed. An annex to the report (Annex ill) deals with Ghana's cocoa environment. Estimates of Ghanaian hectarage and yields per hectare are given for the 1975-1987 period. Nominal and real cocoa prices received by the farmer are examine for 1963-1987. - 3.2 - Ghana's costs of producn coooa are esthmated for 1979, 1983, 1986 and 1988. These costs are compared with production costs for the other major producers. The fhal section of the report integrates the Information obtained from the world market analysis and the hnternal Ghana market to formulate a Ghanaian cocoa pricing policy which will hwprove Ghanas foreign exchange earnigs during the 1990-2000 period. As the analysis shows, an opportunity exists for the Goverrnent of Ghana to significantly Improve Its forelgn exchange earng durhg the next two or three decades If it pursues the appropriate foreign exchange and farmer pricing pollcies. It Is Ikportant that the pollelas not only be adopted but that they be sustained over the given the long-term nature of the cocoa Investment decision. - 3.3 - 11. WORLD COCOA MARKET ENVIRONMENT, 19882000 The decade of the 1970s was one of cocoa shortage. The low prices of the 19605 discouraged production In Ghana, Nigorla and Brazil. World cocoa production peaked at 1.508 millon tonnes In 1965 and again at 1.683 million In 1972 and did not rlse above those levels until 1980. Consequently, world cocoa prIces (spot Ghana) soared from 200 sterlhng per tonne In the early 1970s to more than 3000 In 1977 In order to keep world consumption In check and maintain minimal Inventories. The high prices of the 1970s and early 1980s encouraged plantings in the Ivory Zoast, Brazil, Malaysia and other producing nations. Consequently, world cocoa output rose from the 1.500 million tonne level In the late 1970s to 1.627 In 1980, to 1.946 In 1985 and to an estimated 2.150 millon In 1988. m The first signs that the high prices of the 1970s had stimulated cocoa piantings in various countries occurred In the late 19709 and, particularly, In 1980 when output exceeded 1.600 millon tonnes for the fIrst tkme. Between 1980 and 1985, output Increased another 300,000 tonnes with Cote dlvoire accounting for 173,000, Brazil for 118,000 and Malaysian output Increased 71,000 tonnes. The three countries combined had Increases totaling 362,000 tonnes-more than the Increase In world production. Minor producer output also rose 58,000 tonnes while production In Ghana and Nigeria fell by 112,000 and 22,000 tonnes, respectively. The section which follows examines the hectarage changes and yleld patterns for the major producing countrles which Is largely responsible for the growth In world output. A. Cocoa Hectarage, Major Producers 1970-1987 Table 1 contalns official estimates of cocoa hectarage for Cote d'lvoire, Brazil and Malaysla--the thrte countries largely responslble for the growth In cocoa production. The planting patterns of the last tow decades are examkied for each country. _ 3.4 - TabO 3.1: Cocoa iectarage, Balz. Coto civohe, Malaysaa, 1970-S Crop Year Co. MALAYSLA Beg Brazil d'ldre Penin Sabbah Sarawak Total 1969 338 557 3 4 0 7 1970 343 607 7 5 0 12 1971 352 655 12 5 0 17 1972 362 700 16 6 1 23 19!3 374 743 19 6 1 28 1974 392 7865 22 10 3 35 1975 423 843 26 12 3 41 1976 453 901 30 15 4 49 1977 485 968 34 22 5 61 1978 524 1017 45 38 a 89 1979 569 1075 57 58 9 124 1980 606 1133 70 83 11 164 1981 640 1180 82 114 13 209 1962 640 1230 84 133 14 231 1963 640 1310 89 159 17 265 1984 655 1390 101 173 24 298 1965 655 1450 102 160 26 306 1966 e65 1460 103 184 28 315 1987 655 105 187 32 324 Sounw: Brazil USDA estma1s based on CEPLAC document. for 1970-62. 1963-87 estimates basd on CEPLAC estams of Brazl's cocoa area in production as reported in Coffee and Cocoa Inabonal, No. 5, 1967. Cote D'voire USLA estimate obbtand from SATMACI documnwt. Malavsa USDA agrlcula attache eimalis, obtalned from te Miniaty of Agnculture, Deparmunt of Statiics, Malaya. 1. Cots d'ivoire Hectarage (Table 1) Cocoa hectarage Increased from 557,000 In 1969/70 to 785,000 In 1974/5--an Increase of 41 percent. The Increase between 1975 and 1980 was another 290,000 or a 37 percent rise. From 1980 to the present, cocoa hectarage Increased from 1,075 to approxlmately 1.500 mililon--aknost treble the 1970 quantity. At the end of 1987, 17 percent of Ivcrian cocoa was under 5 years of age, and ahnost 40 percent was 10 years or less. Consequently yield Increases from newly bearing trees can be expected In the 1988-1992 perlod and trees which are not now bearing will begin producing In the early 1990s. Expected production durlng the years 1988-2000 from newly bearing trees will be examhnod In a later sectlon. 3.5 - 2. Drail Hectarag. (Tab 1) Approximately 50 percent of the cocoa area In Bahia was planted during the last 17 years. Cocoa hectarage increased from 338,000 to 655,000 between 1970 and 1988 . Wlth the decline In world prices durlng the 1980s, the growth In Brazilian hectarage has also slowed. Approxlmately 9 percent of the cocoa Is 5 years or less while 35 percent Is 10 years or younger. On the basis of hectarage, one might expect Brazilian output to continue rising during the next few years. However, low world and Brazillan farmer prices have resulted In a reduction In fertilizer and other Inputs which has cal -ed output to decline and stagnate In the 350,000-400,000 tonno range. The current price knpact on Brazilian output will be reviewed In a later section. 3. Malaysia Hectarage (Table 1) Malaysia's cocoa area totaled only 7,000 hectares In 1970, rose to 35,000 by 1975, to 124,000 b;y 1980 and currently Is ostinated by the Malaysian Ministry of Agriculture at 324,000 hectares in early 1988. Approxhiately 35 percent of Malaysia's cocoa Is 5 years of age or less. More than 80 percent Is 10 years or younger. Consequently, major Increases In Malaysian output will occ.ur during the next 5-6 years. With yields between 900 and 1200 kilos per hectare, Malaysian output should Increase above tne 300,000 tonne level by 1990 or 1991. The output Increases expected from Sabbah, Sarawak and Peninsular Malaysia will be examined In the next section. lAgain, these years refer to the crop year ending in 1970 or 1988. The USDA Agricultural Attache's estimate of Brazilian hectarage is currently 655,000 which has not changed since 1984/5. CEPLAC's 1988 estimate is 687,926 hectares. The USDA series has been used int he analy;sis as their series fits actual output better than the CEPLAC information. 3.6 - B. World Cocoa Production and Prospects, 1975-2000 The hectarage data In Table 1 has been combined with tree yield Information presented in Table 2 to obtain cocoa production esthnates for the three major producers through the year 2000. The yield Informatlob for the major producers (Table 2) Indicates that West African ylelds for hybrlds peak at 60' kilos per hectare in Table 3.2: Major Producer Yied Pattens By Year (Kg/Ha) Ghans d'lvoire Brazii Year Hybrid Hybed Hybrid Malaysia 1 0 0 0 0 2 0 0 0 0 3 0 0 0 0 4 50 50 200 200 5 200 200 400 400 6 400 400 6oo 600 7 500 500 800 800 -s5 600 600 1000 900-12000 26-2 500 500 1000 900-1200 29-30 400 400 1000 900-1200 Soucs: Ghana anJ Cote d'lvoire yields taken front Ghana Cocoa Board atudy, Cost of Prroduction of One Hectare of Cocoa, Octor, 1965. 3razil and Malaysia yieds taken fromn CEPLAC data and Brazil. Malaysan maximum ydds vary from 900 kilos per hectars on the penirnsula to 1200 kikl in Sabbah and Sarawak. The Cots d'lre, Brazil and Malaysian ybid ptterWns were combined with hectarage data and tsted against actual output which verified te reasonabbness of the estimates. the eighth year, re-"ain at the peak through the twenty-fifti and then slowly decline thereafter. Brazil oata from CEPLAC indicates that Brazilian hybrids produce 1000 kilos per hectare at maturity which occurs In the eighth ;year after planting. With appropriate Inputs, peak yields can be maintained for many years. However, If the Inputs are withdrawn (fertilizers, insecticides, fungicides, etc.), significant declines yields occur. Malayslan Informatlon suggests that yields peak In the 900-1200 kilos per hectare range depending on the location. Peak yields in Peninsular Malaysia are approximately 900 - 3 .7 - klos while mature trees on Sabbah and Sarawak yield 1200 kilos at the peak. Richer soils combined wlfh more inputs (fertilizers, insecticides, fungicides, etc.) are responsible for the higher yields In Brazil and Malaysia vis-a-vis West Africa. Detailed Information on the assumptIons used to generate productlon forecasts for Cote d'lvo!re, Brazil and Malaysia appear In Tables 3 through 7. The model estimates of production compared with actual output appear In the sams tables plus Charts 1 through 5. 1. Cote d'vobre Output-1988-2. 3 (Table 3, Chart 1) Table 3 presents the results of a hectarage/yleld model applied to the Cote d'lvoire data. Hectarage Information for the 1965-1987 period Is separated into traditlonal and hybrid. The assumption Is made that all new plantings since tne mid-1960s have used hybrid material. A fLrther assumlpt!on Is tlat yields have used tybrld materbal. A further assumption Is that yields of traditional cocoa trees peak at 400 kilos per hectare while hybrid trees peak at 600 kilos.2 The last two columns In Table 3 compare estknated Cote d'lvolre output derived from the hectarage/yield model with actual production for the 1965-1987 period and Chart 1 graphs the estimated against the actual. The chart Illustrates the excellent fit of actial agalnst potential. The shortfalls In Cote d'lvolre production during 1975-1978 and, again, In 1983-1984 resulted from drought conditions due to El Nino weather. Othierwise, the two series are almost Identical. Using the yield and hectarage Informatlon In Table 3, projections of Cote d'lvoire production ar6 made for the pe i 1963-1995. Output Is forecast to reach 800,000 tonnes by 1992 and stagnate near that level during the rest of the decade due to low 2 The hybrid tree yield data in Tables 3 through 7 represent yields by age of tree and not yields by calendar or crop year. For example, the first three values for the hybrid yield are given as 0 -which means that hybrid trees in Cote d'Ivoire do not produce during the first three years after planting. In the fourth year, hybrid trees produce 50 kilos per hectare. In the fifth year, yields rise to 200 kilos, etc. Tree yields peak at 600 kilos in the eighth year and remain at 600 through the 25th year. Starting in the 26th year, tree yields begin to decline due to age. - 3.8 - COTE 'T. , limcTAP.IC:. :CLC. ,^cJ) 'HC-:cN llctArgc ix. 'cidiF:Xil>.Productionl (000 MT) 196S - 1995 Crop Ydar - I ctar gc--------- ----Yic ld------ ---Cstt.Lcd Lducton-- ; :.*L Oeg. Total Traditional liybrid Tcad:Ct. Iybrid Tradit. Hybr:d Twtal it.:c. .rn 196S 445 445. 0 260 0 11 I ' 1 : . 1966 465 445 20 335 0 149 D 14:9 : 1967 485 445 40 350 0 1 5 6 0 1968 507 445 62 375 SO 167 0 1G7 145 1969 557 445 112 400 200 178 1 1970 607 445 162 400 400 178 5 1971 655 445 210 400 500 178 13 .u 1972 700 445 255 400 600 1768 '5 2 .131 1973 743 445 298 400 600 178 43 22: -:9 1974 785 445 340 400 600 178 67 245 242 1975 843 445 398 400 600 178 94 ''2 231 1976 901 445 456 400 600 178 123 301 230 1977 959 445 514 400 600 178 1SO 328 304 1978 1017 445 572 400 600 178 177 255 312 1979 1075 445 630 400 600 178 206 304 379 1980 1133 445 688 400 600 178 238 416 ;03 1981 1180 445 735 400 600 178 Z71 449 457 1982 1230 445 7J5 400 600 178 306 404 355 1983 1310 445 865 400 600 178 340 526 405 1984 1390 445 945 400 600 178 375 553 565 1985 1450 445 1005 400 600 178 407 585 590 1986 1480 445 1035 400 600 178 440 618 609 1987 1500 445 1055 400 600 178 476 t.54 660 1988 1520 445 1075 400 600 178 517 I95 1989 1.520 445 107S 400 600 178 55 .'S 1990 1520 445 1075 400 500 178 590 760 1991 1520 445 1075 400 500 178 612 -'Jr 1992 1520 445 1075 400 SOO 178 626 004 1993 1520 445 1075 400 400 178 633 1l3 1994 1520 445 1075 400 400 17B 630 H. 1995 1520 445 1075 400 350 178 G25 '3G3 Sources: Hectarage -- 1965-198b. USDA agr2cu±tural attache reports. 1987-1995. Hectarage assumed tuncnanged near ' 5 mil*.on Yields by Age -- Ghana Cocobod study. 'Cost of Production of 1 Hectaro of Cocoa.' October, 1985. 4 4 4 J * 4 CHIART I COTE D'IVOIRE Cocoa Production metric tons (000) 1000- 3 0 0 -..... .. ... ..................... ......... 600 400 .. .. . . . 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 200 = _ -. 1)65 1970 1975 1980 1985 1990 1995 Actual ^ - Estimated Commodity Informatlon, Inc. -3 .10 - world cocoa prices. Given the low world prices of 1987-1988 and those projected for 1989-1992 plus the financial problems that have been and will be created for most world cocoa prod1ucers, it is reasonable to assume that Cote d'lvoire hectarage will stagnate near current levels. . Stagnatlon In plantings will result In no growth of output In the second half of the 1990s so that the output forecast for 1995 Is also relevant for the year 2000. World cocoa prices are currently so lbw that major producers such as Cote d'lvoire may be forced to lower farmer prices below cost of production levels which would result in maxkinum yields and output not being achieved. Consequently, Cote d'lvoire production may not reach the 800,000 tonne level durlng the 1990s. Conservatively, output Is forecast to stagnate In the 750,000-800,000 tonne range. 2. Brazillan OutDut, 1988-2000 (Table 4, Chart 2) Table 4 and Chart 2 present Brazilian hectarage and yield informatlon for the 1965-1987 period with projectlons to 1995. The assumptlon is made that al! trees in 1965 were traditlonal, but that all plantings since 1965 have been hybrid, It Is also assumed that traditlonal hectarage started declining In the early 1980s and was replaced by hybrid plantings. The yield assumptlons employed assume that traditlonal tree yields peak at 600 kilos per hectare while hybrid trees peak at 1,000 kilos. Again, the last two columns and Chart 4 compare the output from a hectarage/yield model with actual output for the 1965-1987 perlod. The correlatlon between the estimated and actual output serles Is not as high In Brazil as in Cote d'lvoire although the fit is reasonable. The pattern of world cocoa prices Is one explanation for the underesthiates In 1965-1972, the overestknates In 1973-1931 and tne underestimates A 1986-1987 Worl& cocoa prices were iow during 3 Cote d'Ivoire producer prices will likely stagnate or decline in real terms given the vtry low cocoa prices in the world mark.et. Consequently, hectarage devoted to cocoa is expected to stagnate as well.. - 3.11 - TAOLC 4 DRA71LIN4 IIECT.RACC. YLCLUS A.: Pl.O8UCTK-. 11ectarcs (000): Yield Ihs,) rvodu,;Li;i -J,, Mlr, 1965 - 1995 Crop Year -------lectarage--------- -----Yldd---- - Cstimated ProductLon-- Actual nOc. Total Traditional hybrid Tradit. IlybrUd Tradit. llyUrid TO1L1 ProductLon 1965 330.7 330.7 .0 600 0 198 0 196 173 1966 332.6 3317.0 - 1.6 600 0 199 0 199 175 1967 334.8 331.0 3.8 600 100 199 0 199 144 1968 336.1 331.0 S.1 600 200 199 0 199 165 1969 338.3 331.0 7.3 600 400 199 1 199 201 1970 343.0 331.0 12.0 600 600 199 1 200 102 1971 351.9 331.0 20.9 600 800 199 2 201 167 1972 362.4 332.0 30.4 600 1000 199 4 203 162 1973 374.4 332.0 42.4 600 1000 199 7 206 246 1974 392.4 332.0 60.4 600 1000 199 il 21C 273 1975 423.4 332.0 91.4 600 1000 199 16 216 258 .L976 453.0 332.0 121.0 600 1000 199 24 224 234 1977 485.0 332.0 153.0 600 1000 199 36 236 283 1978 524.0 332.0 192.0 600 1000 199 52 251 314 1979 569.0 332.0 237.0 600 1000 199 72 272 294 1980 606.0 332.0 274.0 600 1000 199 98 297 349 1981 640.0 309.0 331.0 600 1000 1>85 128 313 314 1982 640.0 294.0 346.0 600 1000 176 163 339 336 1983 640.0 279.0 361.0 600 1000 167 201 369 302 1984 655.0 265.0 390.0 600 1000 159 239 398 412 1985 6SS.0 252.0 403.0 600 1000 151 278 429 376 1986 655.0 239.0 416.0 600 1OOC 143 311 456 369 1987 655.0 227.0 428.0 600 1000 136 342 478 399 1988 655.0 216.0 439.0 600 1000 130 368 497 1989 655.0 205.0 450.0 600 1000 123 384 507 1990 655.0 195.0 460.0 600 1000 117 401 518 1991 655.0 18S.0 470.0 600 1000 .11 416 527 1992 655.0 176.0 479.0 600 1000 106 428 534 1993 655.0 167.0 488.0 600 1000 100 439 539 1994 655.0 158.0 497.0 600 1030 95 450 545 1995 655.0 151.0 504.0 600 1000 91 460 550 Sources: Hectarage -- 1965-1986. USDA agricultural attache reports. 1987-1995. Hectarage assumed unchanged near 1.5 million Yields by Age -- CEPLAC data for Brazil 4 'oul' lJOJDujiojul Allpouwoo Palmi}s3 _ lenjoV lOl S86L0 086L 9L6 O6L 996L . . . . ... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .................... . ... . .. . . , _ 0 0 1 > ' I ~~.. ... . . . . . . . . . .. ... . . . . . . . . ............ ................. .. ...... ..... o o C ................. ....... . . . . . . . . .. . . . . .. . . . . .. . . . . ... .. 0 0 ~............ . . . . . ........ .. ... . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . .. . . .. . . 0 0 F _ ~~~009 (000). suol OIJIaW uoilonp°Jd 2°D° ) llZvde z lUdVIID - 3.13 - the late 1960s and mid 1980s but very high during the second half of the 1970s. It is apparent that Brazillan cocoa farmers withhold Inputs during low price periods and add them during periods of high prices. The hectarage/yleld model forecasts Brazilian output to reach 500,000 tonnes by 1989 and 550,000 tonnes by 1995. These production levels will not be achieved because of low prices wh!ch are currentiy forcing farmers to remove fertilizers, fungicides and insecticides.4 Consequently, Brazilian production Is expected to stagnate in the 350,000- 400,000 tonne range during the remainder of the 1980s and the early 1990s. Given the expected Increase In world cocoa prices In the second half of the 1990s which will be shown later, Brazilian farmers are expected to add Inputs between 1995 and the year 2000 which would raise Brazillan output above 400,000 tonnes by 2000. 3. Malaysian Output, 1988-2000 (Tabls 5-7, Charts 3-5) Malaysia produced no cocoa in 1965 although 2,300 hectares had been pianted. By 1975, 35,000 hectares was under cultivatlon and output had risen to 9,400 tonnes. Ten years later (1985), hectarage reached aimost 300,000 and output totaled 130,000 tonnes. A comparison of the estimated and actual output series for the Peninsula, Sabbah and Sarawak in Charts 3-5 suggest that the "estimated" series mey slight!y overestimate the "actual" for the Peninsula but underestimate output for Sabb?1 and Sarawak. The hectarage/yleld models for the three Malaysian areas perform reasonably well, however, and suggest that Malaysian output will exceed 300,000 tonnes by 1990 and reach 350,000 by 1995. Currently, Malaysian p ov,ucVon is ircreasing at a rate of 30,000 tonnes per year. Malaysian costs of production are among the lowest of the major producers as will be noted in a later section. Consequently, Malaysian planting may slow down during 4 Gill and Duffus Group PLC, 'Cocoa Market Reiprt,' Mp/ 331. August. 1988; Barretto Peat, various crop report during 1988. - 3.14 - tABLE S PCNINSULA IIECTAAGCC. YICLOS AND PRODUCTION Ilctaics (000); Yield (kg): Production (000 MT) 1965 - 1995 Crop --------- lctarage ---------- ----- Yield------ --- Estimated Production--- Actual Year Total Traditional Hybrid Tradit. hlybrid Tradit. 11ybrid Total Production Dcg. 1965 1.5 0 1.5 0 0 0 0 0 1966 1.9 a 1.9 0 0 0 0 0 1967 2.0 0 2.0 0 100 0 .2 .2 1968 2.4 0 2.4 0 200 0 .3 .3 1969 3.1 0 3.1 0 400 0 .7 .7 1970 3.4 0 3.4 0 600 0 1.1 1.1 1971 7.5 0 7.5 0 Soo 0 1.6 1.6 1972 12.0 0 12.0 0 900 0 2.1 2.1 1973 15.7 0 15.7 0 900 0 2.8 2.8 1974 18.7 0 16.7 0 900 0 3.9 3.9 1975 22.0 0 22.0 0 900 0 5.8 5.8 1976 26.0 0 26.0 0 900 0 6.3 8.3 1977 Z.9.7 0 29.7 0 900 0 11.5 11.5 197! 34.3 0 34.3 0 900 0 14.6 14.6 1979 45.2 0 45.2 0 900 0 18.1 18.1 1980 57.3 0 57.3 0 900 0 21.3 21.3 1981 70.0 0 70.0 0 900 0 25.3 25.3 1982 62.2 0 62.2 0 900 0 30.4 30.4 1983 83.9 0 63.9 0 900 0 37.2 37.2 24 1964 69.2 0 89.2 0 900 0 4S.S 45.5 28 196S 100.6 0 100.6 0 900 0 54.5 54.S 50 1986 102.0 0 102.0 0 900 0 63.7 63.7 61 1987 103.0 0 103 0 0 900 0 71.9 71.9 70 1988 105.0 0 105.0 0 900 0 78.3 7863 1989 105.0 0 105.0 0 900 0 63.5 83.5 1990 105.0 0 105.0 0 900 0 87.5 87.5 1991 105.0 0 lOS.0- 0 900 0 91.0 91.0 1992 105.0 0 105.0 0 900 0 93.1 93.1 1993 105.0 0 105.0 0 900 0 93.$ 93.8 1994 105.0 0 105.0 0 900 0 94.3 94.3 1995 105.0 0 105.0 0 900 0 94.5 94.5 Sources: Hectarage -- 1965-1986. USDA agricultural attache reports. 1967-1995. Nectarage assumed unchanged near 1.5 million Yiels by Age -- CEPLAC data. except Cor peak yield of 900 which has been assumed by Conwodity Information CHART 3 MALAYSIA - Peninsula Cocoa Production metric tons (000) 100 - 8 0 ............... ........... ... ... - ...................... . ..... . .. . 6 0 .. . ... ... <..... ............ .. ...... ... . . I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 X. .......... .. ........... .. ..... ............................................... ............ ...... .. ..... .. 20 ........................................................................................................ 60 1983 1985 1987 1989 1991 1993 11995 - Actual - - Estimated Commodity Information. Inc. - 3.16 - TAOLC SAOOAIt IlCCTARACEC YSE_ZS A!:D PRODUCTICV4 liectares (000>. yjeid (k,,. i'rzoduction koi90 'ITa 1965 - 19 5 C r o? ---------lIectarage----- - -----Yield- - ---- Estimated Productioi ----- Actual Year Total Traditional Hybrid Tradit. Hybrid Tradit. Itybrid oL a I'roduction 13cg. 1965 .6 0 .6 0 0 0 0 0 1966 .7 0 .7 0 0 0 0 0 1967 .8 0 .8 0 100 0 .1 1 1968 1.0 0 1.0 0 200 0 .1 1 1969 1.2 0 1.2 0 400 0 .3 .3 1970 4.0 0 4.0 0 600 0 .4 4 1971 4.5 0 4.S 0 0oo 0 .6 .6 1972 5.5 0 5.5 0 1000 0 1.1 1 1 1973 6.2 0 6.2 0 1200 0 1.7 1.7 1974 8.1 0 8.1 0 1200 0 2.5 2.5 1975 9.8 0 9.8 0 1200 0 3.5 3.5 1976 11.7 0 11.7 0 1200 0 4.7 4.7 1977 15.0 0 15.0 0 1200 0 6.1 6.1 1978 22.5 0 22.5 0 1200 0 7.8 7.8 8 1979 37.8 0 37.8 0 1200 0 9.5 9.5 10 1980 58.0 0 58.0 0 1200 0 12.0 12.0 12 1981 83.5 0 83.5 0 1200 0 16.2 16.2 19 1982 114.5 0 114.5 0 1200 0 23.0 23.0 30 1983 132.7 0 132.7 0 1200 0 33.5 33.5 30 1984 159.3 0 159.3 0 1200 0 48.8 48.8 41 1985 172.7 0 172.7 0 1200 0 68.1 68.1 68 1986 180.0 0 180.0 0 1200 0 92.4 92.4 97 1987 184.0 0 184.0 0 1200 0 118.5 118.5 125 1988 187.0 0 187.0 0 1200 0 144.9 144.9 1989 187.0 0 187.0 0 1200 0 168.9 168.9 1990 187.0 0 187.0 0 1200 0 188.9 188." 1991 187.0 0 187.0 0 1200 0 203.1 203.1 1992 187.0 0 187.0 0 1200 0 214.0 214.0 1993 187.0 0 187.0 0 1200 0 219.5 219.5 1994 187.0 0 187.0 0 1200 0 222.4 222.4 1995 187.0 0 187.0 0 1000 0 223.7 223.7 Sources: Hectarage -- 1965-1986. USDA agricultural attache reports. 1987-1995. Hectarage assumed unchanged near 1.5 million Yields by Age -- CEPLAC data. except for peak yield of 1200 which has been assumed by Commodity Information CHART 4 MALAYSIA -- Sabbah Cocoa Production metric tons (000) 250 200 ...... . . I - 150-..-. 100- i- .< .. 50 - - ........ .. .. . . ...7 ..................... ..... .. ... ..... 50.. I 0 I_ I I , , , I I I , I 1980 1985 1990 1995 I Actual Estimated Commodity Informnation, Inc. - 3.18 - TAOLC 7 SARAWAK IICCTAIACE. Y1ELD AND ProvUCTION lcctLrcs 000): Y%aLd (kg; 1'roductLoil (VuO Mlu 1965 - 1995 Crop - Ilectarage.. --- -------Yield -E- timated Production---- Actual Year Total Tcad&tional 8ybrid Tradit. ItybriJ Ircdit. Ilybrid Total ProductLon Beg. 1965 .2 0 .2 0 0 0 0 0 1966 .2 0 .2 0 0 0 0 0 1947 .2 0 .2 0 200 0 0 0 1961 .2 0 .2 0 400 0 .1 .1 1969 .3 0 .3 0 600 0 .1 .1 1970 .4 0 ;4 0 600 0 .2 .2 1971 .e 0 .6 0 1000 0 .2 2 1972 1.2 0 1.2 0 1200 0 .3 .3 1973 1.5 0 1.5 0 1200 0 .4 .4 1974 2.0 0 2.0 0 1200 0 .6 .6 1975 2.6 0 2.8 0 1200 0 .9 .9 1976 3.5 0 3.5 0 1200 0 1.2 1.2 1977 4.0 0 4.0 0 1200 0 1.6 1.6 1978 4.7 0 4.7 0 1200 0 2.4 2.4 1979 6.4 0 6.4 0 1200 0 3.0 3.0 1960 6.5 0 6.5 0 1200 0 3.7 3.7 2 1961 10.7 0 10.7 0 1200 0 4.7 4.7 4 1942 12.7 0 12.7 0 1200 0 6.0 6.0 5 1963 14.4 0 14.4 0 1200 0 7.6 7.6 6 1964 17.1 0 17.1 0 1200 0 9.4 9.4 9 1955 24.3 0 24.3 0 1200 0 ll.S 11.S 12 1966 26.0 0 26.0 0 1200 0 14.0 14.0 15 1967 26.0 0 26.0 0 1200 0 17.5 17.S 19 196a 32.0 0 32.0 O 1200 0 21.0 21.0 1969 32.0 0 32.0 0 1200 0 24.5 24.5 1990 32.0 0 32.0 0 1200 0 26.3 26.3 1991 32.0 0 32.0 0 1200 0 31.9 31.9 1992 32.0 0 32.0 0 1200 0 34.9 34.9 1993 32.0 0 32.0 0 1200 0 36.4 36.4 1994 32.0 0 32.0 0 1200 0 37.6 37.6 1995 32.0 0 32.0 0 1200 0 36.4 36.4 Sources: Hectarcag -- 1965-1966. USDA agricultural attache reports. 1967-199S. HKctarag. assumed unchaged near 1.5 million Yield by Age -- CEPLAC data. except for peak yield ot 1200 uhich has been assumed by Commdity Intor-ation OCIART S MALAYSIA -- Sarawak Cocoa Production metric tons (000) 40 30./ 3 0 . ....... .. ................................... .............. Z-............ . . . . . . . . . . 1 _ '0 0 1980 1985 1990 1995 Actual - Estimated Cornmnodity Inforrnatlon, Inc. . 3.20 - the next few years but still experience some growth which means that output Increases are expected through the late 1990s. 4. Ghana's Cocoa Prospects, 1988-2000 For the first tkne In two decades, optimism exists regarding Ghana's cocoa production. Until pricing poilcies were changed In 1985, Ghana's cocoa production was headed for oblivion. As will be shown In a later section, the farmer price increases which began in 1985 have stabilized production Just above 200,000 tonnes and will lead to substantive growth if sustained. Given current Ghana farmer pricing policies, Ghanaian output will Increase to 300,000 tonnes or more by 1995. Information collected by the World Bank during 1987 and 1988 Indicates that significant new plantings have taken place In Ghana. A field tour in October, 1987 Indicated that 20-25 percent of the cocoa hectarage consisted of new plantings under the age of three years. A survey of the cocoa sector by the World Bank places the new plantings at an even higher proportion of total hectarage. The findings of the survey indicate that output will Increase above 300,000 tonnes by the mid-1990s. However, if the appropriate pricing Incentives are not maintained, the potential Increases will not be maintained and Ghanaian output will stagnate in the 250,000-300,000 range. The appropriate policies for stimulating Ghanalan production growth will be outlined in a later section. 5. Nlgerian Cocoa Prosiects, 1988-2000 Nigerlan output Is also expected to Increase at least 40,000 to 50,000 tonnes during the next decade. The creation of a free cocoa market In Nigeria has resulted In higher prices to the farmer. Observers of Nigerian cocoa have noticed significant Improvements In cocoa farm maintenance and harvesting during the past two years. Consequently, Nigerlan cocoa output will Increase to the 175,000 tonne level during the next few years and If the price incentives continue, output could rise above 200,000 tonnes during the 1990s. - 3. 21 - 6. indne Prospects, 1988-2000 Another country which shows consirable production promise during the next decaoo.% is Indonesia. Although reliable data Is diffcult to obtain, Gill and Duffus data indicates that hndonesiar output has grown from ahbost nothing In 1975 to approxklately 45,000 tonnes in 1987-1988. According to hdonesian sources, hectarage planted now totals 90.000 with excollet prospects for further hcreases. Output Is expected to reach the 90.000 tonne level during the mki-1990s. The Indonesian projection has been incorporated in the Asia/Oceanic total which Is forecast to rise from 87.000 tonnes In 1986-1987 to 150,000 by 1995 and then to 170,000 by the year 2000. C. Warld Coo Production wary. 1988-2000 (Table 8) Taking Into account the above Information, world productin Is expected to Increase from 1,980 million tonnes In 198S-1987 to more than 2.150 million In 1987-1988, to 2.340 million by 1990 and then piateau In the 2.550-2.650 range In the 1995-2000 poriod. Output was stagnant during the 1970s as noted earlier, but grew 3.5 percent per year between 1980 and 1985 and Is proJocted to grow 3.9 percent between 1985 and 1990. Output growth Is expoected to siow to just under 2 percent per annum in the 1990-1995 period as a result of low world cocoa prices. Output Is forecasted to stagnate during 1995-2000. In summari, world cocoa production stagnated In the 19709 as a result of the low world prices of the 19603, but has hcreased dramatically In the 1980s because of hectarage expansion In a number of major producing countrles due to the high prices of the 1970s. The low world prices currently prevailing h'i the 1985-1988 period and projected through 1994 are expected to siow output growth during the 1990s. - 3.22 - TABLE 6 COCOA PRODUCTrON PROSPCCTS BY MAJOR PRODUCUC AMEA 1974/5 - 1999/2000 (000 TONNESD Producing - ------Actual :;__-- -------------- Forecasts---------- Area 1974/5 3979/60 1964/4 1986/7 1967/a 1968/9 1989/90 1994/5 1999/2000 Major Producers 1315 1364 1634 1650 1604 1935 1975 2140 2220 Iraail 273 294 412 369 400 365 375 375 425 Camroon 116 124 122 123 129 130 130 13S 140 Coto dWvoire 242 379 552 609 660 725 735 750 750 tcundor 76 90 122 77 75 55 60 80 so Chans 37? 265 173 223 185 250 250 275 300 Malaysia 13 32 103 164 210 240 275 350 350 Nigeria 214 172 150 s0 145 ISO 150 175 17S Mlaor Producors 234 242 300 330 346 360 365 410 430 Africa 60 61 S9 64 54 60 60 60 60 America 131 134 166 177 196 195 190 200 200 Asia/Oceania 43 47 73 69 96 105 l1S 1SO 170 World Total 1549 1626 1934 1960 2152 2295 2340 2550 2650 Annual Rate .97 3.53 1.16 6.69 6.64 1.96 1.73 .77 of Growth Source: Historical data -- Gill & Oultus plus origin sources. Proj ctions -- CGouodity Information* Inc. - 3 .23 - D. World Cocoa ConsUtloPt 1976-19S6 Table 9 presents historical data on cocoa grindIngs for the 1975-1985 period plus forecasts through the year 2000. Cocoa conswuption is determined by cocoa prices, population and 'ncome and manufacturers' viw reardng the availability of cncoa over a 3 to 5 year period. Cocoa prices have ben extrely Important In the determination of world usage because prices reflect the changes in world output, manufacturers' margins, and the price of confoctxnery relative to other foods. During the 19709, world cocoa usage was unable to grow because of stagnant output and high prices. Glven the secular upward trend In consumption supported by population and Income, world cocoa prices had to rise in order to hold usage near the 1.500 million tonne level World grindig totmied 1.536 mlion In 1972 and 1.488 millon In 1980. During that period of thIe, end-of-year inventories fe1 from 616,000 In 1972 to 287,000 by 1977. A major point Is that world consumption can only grow when world productlon grows. It Is the supply of cocor which determs the usage growth path through the price mechanism. Between 1980 and 1985, world grindigs and consumption grew at a 4 percent annual rate--slightly higher than the output growth rate. The 1980 consumption was 140,000 tonnes below the 1980 production level which provkied the room for faster growth. In 1985, consumption totaled 1.838 wUllon tonne-still well below the 1.934 million tonne productlon level. The growth In consumption was spurred by falling world prices which declined from $2,600 per tonne In 1980 to $2,250 by i98S 5 Between 1985 and 1987, grinding hcreased at a 2 to 3 percent pace. World consumption Is projected to hcrease 5 pereont In 1988. World prices averaged less than $2,000 In 1987 and wIN average ls than $1700 in 1968 In Deutschmarks, cocoa prices 5 The ICCO 1980 indicator price fell from 118.06 cents per pound to 102.27 cents between 1980 and 1985. Prices in 1981-1983 were even lower in the 78.78- 96.10 cent range. - 3.24 - are the lowest in decades at 3.4 DM per kilo--down from S-8 DMs In the mid-1980s and 12 DMa In 1977. E. Word Suppl/Demand and Price Forecasts, 1988-2000 (See Tables 9-13 and Charts 6-8) The world supply/demand and price forecasts for 1995 and 2000 contalned In Tablee 9-13 are based on three econometric modals--a wor!d cocoa price model which uses the Ghana Spot price as the dependent variable; a grhidings modei which Is based on real world cocoa prices, real world Income, Interest rates and lagged cocoa stocks; and a model which explains the Ghana price prom.um. An explanation of the models follows. 1. World Coooa QtrVvs Model (Table 11) The grindIngs model data and paramoter esthiates are provided In Table 11. The model Is as follows: WCGt , 745.25 - 1 5.84*RGSPt - I + 294.22*RGDPt -7.92 TBt (16.3) (3.8) (18.3) (1.8) +.3O*WCSt - 3 (4.6) Corr. R-sq. - .95 SER . 51.3 DW - .93 d.f. - 23 where: WCGt - world cocoa grindngs In year 't"/ RGSPt a 1 Ghana spot price deflated by the U.S. wholesale food price Index and lagged one year. RGDPt A weighted average of the U.S., U.K., West Gman, French and Japanese GOP's deflated by the U.S. wholesale food price Index for year t". TBt The 90 day U.S. treasury bllU rate for year Ot". WCSt - 3 - World cocoa stocks lagged 3 years. - 3.25 - TABLE 9 WORLD COCOA GRINDINGS 1975 - 2000 (000 TONNES) ---- History ------- -------Forecasts------ Region £975- 1980 1985 1990 1995 2000 W. Europe 516 548 686 815 950 1025 E. Europe 272 218 243 255 315 330 Africa 159 139 169 180 225 235 America 438 513 624 710 860 935 Asia & Oceania 77 92 116 185 250 275 TOTAL GRINDINGS 1471 1510 1838 2145 2600 2800 Annual Growth (%) .52 4.01 3.14 3.94 1.49 Source: History -- Gill & Duffus Group, PLC Projections -- Commodity Information, Inc. CHART 6 COCOA CONSUMPTION mietric tonis (000) ;3000 2500 2Q000 10-00 %.o0o () a I I I .L J~ I [ .L~ I I.. I t A - I -.1- J1- 1.. -.I I.J.. 1- .1 I I I I J.i L I I I I 19.f0- I96t. 1970 1975 1980 1986 1990 1995 2000 Acttual Es ti aate.d C lil.Eait)(Ity InI-) i.eiaon. Ilia. - 3.27 - TABLE 10 PRODUCTION, CONSUMPTION STOCKS AND PRICES 1960 - 2000 (000 TONNES) Spot Ghana Crop Stock Price Year . Prodn Consn Stocks Ratio Sterling/tonne 1960 1053 916 385 .42 222 1965 1508 1297 845 .65 138 1970 1435 1354 500 .37 306 1975 1549 1452 419 .29 723 1980 1627 1488 534 .36 1270 1985 1944 1799 497 .28 2028 1986 1963 1835 605 .33 1568 1987 1957 1884 658 .35 1320 Assumes 2* Good Weather Inflation 1990 234(c 2145 1140 .53 970 1995 2550 2600 1295 .50 1255 2000 2650 2790 560 .20 4510 El Nino (1989, 1993, 1998) 1990 2100 2145 905 .42 1280 1995 2550 2540 1040 .41 1575 2000 2650 2665 '30 .16 5810 Source: Historical data from Gill & Duffus. Forecasts by Commodity Information, Inc. CHART 7 COCOA SPOT PRICE GHANA Pounds Sterling (Thousands) l ; t~~960 1965 19T0 t975 1980 1985 190 1995 2000 -Actulal EstimatedJ Cammfity Inor-idtoion. Inc. I f~~~ ;1} ,,~ ~ ~~~~. ; - 3.29a - TABLE 11 COCOA CONSUMPTION REGRESSION MODEL (1960 - 2000) Spot Gross Cocoa Cocoa Price Domestic T-Bill Cocoa Consumption Consumption Year Ghana 1/ Product.2/ Rate Stocks 3/ (Actual) (Pred) 1960 2.45 1.10 2.93 215 1000 1088 1961 1.95 1.18 2.38 259 1095 1133 1962 1.82 1.26 2.78 385 1144 1184 1963 2.26 1.37 3.16 562 1184 1243 1964 2.07 1.50 3.55 596 1297 1289 1965 1.45 1.56 3.95 616 1374 1320 1966 1.91 1.59 4.88 654 1387 1324 1967 2.38 1.71 4.32 845 1403 1395 1968 3.11 1.82 5.34 685 1369 1375 1969 3.81 1.90 6.68 635 1354 1367 1970 2.71 2.01 6.46 572 1399 1408 1971 2.02 2.18 4.35 433 1536 1455 1972 2.22 2.29 4.07 500 1583 150D 1973 4.07 2.21 7.04 585 1512 1442 1974 6.01 2.16 7.89 616 1452 1393 1975 3.99 2.21 5.84 416 1523 1415 1976 7.87 2.56_ 4.99 335 1438 1438 1977 15.72 2.72 5.27 418 1394 1363 1978 9.80 2.80 7.22 394 1457 1473 1979 7.72 2.87 10.04 285 1488 1490 1980 5.34 3.01 11.51 288 1592 1564 1981 4.50 3.13 14.03 408 1606 1625 1982 4.01 3.25 10.69 529 1641 1717 1983 5.78 3.46 8.63 574 1726 1773 1984 7.66 3.59 9.58 698 1816 1798 1985 7.51 3.87 7.48 579 1851 1872 1986 5.67 4.00 5.96 342 1895 1902 1987 4.67 4.16 5.82 449 1990 1992 1988 3.82 4.33 6.00 559 2079 1989 3.65 4.50 6.00 613 2145 1993 3.24 4.68 6.00 777 2241 1991 2.99 4.87 6.00 970 2343 1992 2.98 5.06 6.00 1142 2438 1993 3.09 5.26 6.00 1276 2525 1994 3.36 5.47 6.00 1359 2601 1995 3.80 5.69 6.00 1395 2666 1996 4.46 5.92 6.00 1370 2717 1997 5.43 6.16 6.00 1293 2754 1998 6.84 6.40 6.00 1175 27;7 1999 8.87 6.66 6.00 1032 2787 2000 12.34 6.93 6.00 876 2771 - 3.29b - TABLE 11 (Cont.) REGRESSION MODEL: COCOA CONSUMPTION u 775.54 - 17.74 * (SPOT GHANA PRtCE,'FOOD PRICE INDEX". + 298.38 * (GROSS DOMESTIC:PltODUCT/FOOD PRICE I**;7 - 6.64 * (T-BILL RATE) + .215 * (COCOA STOCKS lAI,;ED THIREE YEAS' Corr R-sq a .96 SEE s 44.67 DW * l.99 1/ SPOT GHANA PRICE IN POUNDS STERLING DIVIDED BY THE FOOD PRICE INDEX 2/ DIVIDED BY THE FOOD PRICE INDEX TO GET ' 7 3/ COCOA STOCKS ARE LAGGED THREE TIME PER3OUS -oul 'uolljgu3ogul Allpowwc-D paleuJlsg. . eo 00O0 9661 0661 9861 0961 SL61 OL61 9961 0961 ' -rr-r rr r lll1 irlr 1- 7 -r r- I- a r rr '/FZ~; .1:1:-F t-^ i-i 1 0 1W .. ... .. ooI* O~~ /. I~~~~~~~~~, ' ' Iy oos 009 I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ .. . . . . . ... . ..... . . ....... . ................ ... . . .. 0 0 8 BuiliJaS spunod vqnIVI3Ud VOOOO VNVHD 9 JUVI) - 3.31a - TABLE 12 COCOA PRICE REGRESSION MODEL (1960 - 2000) Spot Spot Food Stock Cocoa Ghana Ghana Price Consumption Currency Cake Price 2/ Price 2 Year Index Ratio 1, Index Dominant (Actual) (Pred) 1960 90.70 100.00 222.00 1961 90.90 .562 99.23 0 177.00 ;-3 1962 91.70 .544 99.17 0 167.00 167 1963 90.70 .538 99.19 0 205.00 163 1964 90.80 .552 99.19 0 188.00 161 1965 94.90 .652 99.26 0 138.00 144 1966 101.00 .499 99.38 0 193.00 223 1967 100.00 .458 103.46 0 238.00 220 1968 103.00 .408 103.65 0 320.00 267 1969 108.90 .316 103.03 0 415.00 409 1970 113.10 .369 102.82 0 306.00 ' 368 1971 115.10 .418 106.17 0 232.00 304 1972 121.70 .401 107.99 0 270.00 342 1973 143.90 .263 116.16 0 585.00 6b0 1974 164.60 .223 115.70 0 990.01 1056 1975 181.30 .289 117.57 0 723.00 903 1976 177.80 .259 123.47 1 1399.00 1560 1977 187.30 .200 120.54 1 2943.99 2502 1978 204.60 .280 122.24 1 2006.00 1924 1979 223.80 .282 119.94 0 1727.00 1338 1;ao 237.80 .359 114.89 0 1270.00 1255 1981 250.60 .368 117.80 0 1127.01 1271 1982 257.70 .435 121.75 0 1033.00 1017 '983 260.00 .345 125.21 0 1502.00 i274 1984 270.30 .196 133.62 0 2069.99 2300 1985 270.00 .247 128.00 0 2028.01 1935 1986 276.70 .302 135.23 0 1568.03 1402 1987 282.40 .323 136.46 0 1320.00 1315 1988 293.70 .394 136.05 0 1150.56 1130 1989 305.44 .467 132.00 0 1073 1990 308.50 .532 130.00 0 971 1991 311.58 .569 130.00 0 914 1992 317.81 .580 130.00 0 930 1993 324.17 .572 130.00 0 982 1994 330.65 .542 130.00 0 1089 1995 337.27 .497 130.00 0 1256 1996 344.01 .441 130.00 0 1507 1997 350.89 .380 130.00 0 1871 1998 357.91 .318 130.00 0 2402 1999 365.07 .260 130.00 0 3178 2000 372.04 .200 130.00 0 4502 - 3.31b - TABLE 12 (Cont.) REGRESSION MODEL: SPOT PRtCE GHANA * 6.604 + 1.966 * LN(FOOD PRICE INDEX) - 1 .192 * LN(STOCK/CONSUMPTION-1) Corr R-sq = .98 - 2.414 * LN(COCOA CURRENCY INDEX) SEE X .146 + .576 * CAKE DOMINANT DW - 1.83 1/ CONSUMPTION IS LAGGED ONE TIME PERIOD 2/ IN POUNDS STERLING - 3. 39a - TABLE 13 GHANA COCOA PREMIUM REGRESSION MODEL (1960 - 2000) Fosd Stock Cocoa Ghana Ghana Price Consump.ion Currency Cake Premium Premium Year Inde;,q Ratio 1! index Dominant (Actual) 2/ (Pred) 2/ 1960 90.70 100.00 7 1961 90.90 .562 99.23 0 6 6 1962 91.70 .544 99.17 0 5 7 1963 90.70 .538 99.19 0 5 7 1964 90.80 .552 99.19 0 7 7 1965 94.90 .652 99.26 0 9 5 1966 101.00 .499 99.38 0 12 12 1967 1o6.oo .45, 103.46 0 24 11 1968 103.00 .408 103.65 0 25 16 1969 108.90 .316 103.03 0 30 37 1970 13.10 .369 102.82 0 19 28 1971 115.10 .418 106.17 0 9 ' 17 1972 121.70 .401 107.99 0 11 20 1973 143.90 .263 116.16 0 66 60 1974 164.60 .223 115-70 0 176 135 1975 181.30 .289 117.67 0 100 81 :976 177.80 .259- 123.47 1 114 104 1977 187.30 .200 120.54 1 408 275 1978 204.60 .280 122.24 1 128 133 1979 223.80 .282 119.94 0 168 134 1980 237.80 .359 114.89 0 143 111 1981 25Q.60 .368 117.80 0 106 102 1982 257.70 .435 121.75 0 40 58 1983 260.00 .345 125.21 0 84 92 1984 270.30 .196 133z62 0 170 293 1985 270.00 .247 128.00 0 196 210 1986 276.70 .302 135.23 0 115 95 1987 282.40 .323 136.46 0 81 79 1988 293.70 .394 136.05 0 54 1989 305.44 .467 132.00 0 46 1990 308.50 .532 130.00 0 37 1991 311.58 .569 130.00 0 32 1992 317.81 .580 130.00 G 32 1993 324.17 .572 130.00 0 35 1994 330.65 .542 130.00 0 43 1995 337.27 .497 130.00 0 57 1996 344.01 .441 130.00 0 82 1997 350.89 .380 130.00 0 127 1998 357.91 .318 130.00 0 212 1999 365.07 .260 130.00 0 378 2000 372.37 .200 130.00 0 785 - 3.32b - TABLE 13 (Cont.) REGRESSION MODEL: GHANA PREMIUM - 16.802 + 2.702 * LN(FOOD PRICE INDEX) - 2.597 * LN(STOCK/CONSUMPTION-1) Corr R-sq = .93 - 6.227 * LN(COCOA CURRENCY INDEX) SEE = .361 + .329 * CAKE DOMINANT DW 1.88 1/ CONSUMPTION IS LAGGED ONE TIME PERIOD 2/ IN POUNDS STERLING - 3 33 - The figures In parentheses under the coefficient estimates are the "t" statistics. All coefficient estimates are statistically slgnificant at the .0005 level except for the treasury bill rate which Is significant at .05. The weights for the GDP series are: U.S. (.30), U.K. (.20), West Germany (.30), France (.08), and Japan (.07). The German weight Is large because the Eastern European block generally buys cocoa with Deutschmarks. The model explains 95 percent of the variation In cocoa grindings over the 1960- 1987 period. The m )del suggests that world cocoa prices and interest rates negatively Impact consumption "s one would expect, and that real Income arl world cocoa stocks have a positive effect on consumptn. The world cocoa stock variable Is an important determinant In that manufacturers must believe that their decisions to Increase or decrease bar welghts or to adjust confectionery prices will not have to be rescinded within a short period of time. Considerable expense Is Involved In changing machhery to adjust confectlonery bar weights and/or to obtain the cooperation of the wholesale and retail trade in changing confectionery prices. Consequently, cocoa grindings do not fully adjust to Increases or decreases in world cocoa prices and availabilities until manufacturers are assured that the new price levels reflect a longer-term change In world cocoa supplies. One problem with the above equation is the positive first order autocorrelatlon revealed by the Durbin-Watson statistic (DW) which decreases the efficiency of the estimates. The DW statistic of .93 Is below the lower limit (dL) for the 5 percent significance level of "d" and just above the lower lkit at the 1 percent level. Consequently, the evidence points toward positive autocorrelation In the errors. In order to correct the autocorrelatlon problem, Beach and MacKInnon's maximum likelihood procedure was used. The parameter estimates In the corrected model are as follows: - 3.34 - WCGt * 775.54 -17.74*RGSPt t -1 + 298.38*RGDPt - 6.64*TBt (4.3) (13.2) (1.2) +.21 *WCSt - 3 (2.7) Corr. R-sq. - .96 SER - 44.7 DW , 1.99 d.f. - 22 The new parameter estimates suggest that world cocoa grindings are slightly more sonsitlve to world cocoa prices and less sensitive to Interest rates and world cocoa stocks when compared wlth the ordinary least squares estimators. The Durbin-Watson statistic at 1.99 indicates that the null hypothesis (of no autocorrelatlon) cannot be reJected at the 1 or 5 percent levels. Table 11 and Chart 6 Illustrate actual and fitted grinds values for 1960-1987 plus projections for the years 1988-2000. As noted above, the model explains 96 percent of the annual variance In grinds for the last 28 years. In particular, the model describes well the behavior of world consumption during the growth periods--1960-1971, 1980- 1987--and also during a period of stagnatlon-1972-1979. The model Indicates that consumption will grow at a 4 percent annual rate during the 1988-1995 period given low world prices and large cocoa stocks. The growth In world grinds will then slow to 1 percent per year between 1995 and 2000 as production ceases to grow, prices rise and world stocks fall. The slowdown in world cocoa production after 1993-1995 Is key to the price and grinds changes. 2. World Cocoa Price Model (Table 12) The other model required to project world cocoa consumption Is a world cocoa price model. Table 12 presents the variables and data used to develop the price equation. The model Is as foliows: - 3.35 - LGSPt - 7.91 + 2.09*LFPlt - 1.1 1ILSCRt - 1 -2.81 LCCIt (2.2) (9.2) (7.9) (2.8) +.6 *8t Corr. R-sq. - .98 SER - .15 DW - 1.41 d.f. - 23 where: LGspt - log 'of Ghana spot price in year mtN. IFPlt log of U.S. wholesale food price index In year "t". LSCRt_1 - log of world cocoa stock-consumption ratio lagged one year. LCCIt - log of cocoa currency nebx using weighted average of U.S dollar, U.K. sterling, West German mark, French franc and Japanese yen. CDt - Cake dominant varlable for 1976-1978. Again, all of the coefficients are statistically significant at the .0005 level except for one-the cocoa currency Index which Is significant at .005. The U.S. wholesale food price Index represents a general Inflatlon index which has affected cocoa prices durl.g the past three decades. According to the model, a one percent Increase In U.S. food prices translates Into 1 2 percent Increase In world cocoa prices. This seems high but inflatlon has been ore rampant In the producing countries than among consuming nations which may partially explaln the high secular uptrend In cocoa prices. It Is well known that world cocoa prices are heavily Influenced by the world stock/consumptlon ratio. 6 The esthuated coefficient for this variable Indicates that a 1 percent hcrease In the ratio of stocks to consumption decreases prices by 1.1 percent. This coefficlent Is somewhat lower than the 'sthmate found In other models. However, most models do not Introduce a speclal variable for the 1976-1978 period 6 F.H. Weymar, The Dynamics of the World Cocoa Market, M.I.T. Press, 1967. - 3 6 - during which the scarcity of supply forcod the market to operate on the demand curve for cake products rather than the demand curve for cocoa butter. Since the demand for cake or powder Is much more inelasti than the demand for butter, cocoa prices soared to high levels. If one omits the cake variable, more burden Is placed on the stocks/consumption ratlon to explain the high prices of 1976-1978 which ralses the coefficient estiuato but gives a much poorer fit to the equation. The currency variable was Introduced to account for the varying Impact of currency changes on world cocoa pricos durhg the 28 year period. Again, the weights on the currencies used to develop the serles are the same weights as used In deveioping the GDP variable descrlbed above. The currency coefficient suggests that a one percent change in currencles produces a 2.8 pwrcont chane hI the Ghana sPot prico (sterling). The high coefficient probably stems from the fact that stering has been a weak curroncy during most of the period under revlew. Chart 7 Illustrates the fit of the esthated prico against the actual. The largest error occurs in 1977 when prices peaked. This Is also the tiem that the cake ratio reached Its zenith. The cake dominant variable would have performed better If the values for 1976-1978 had more accurately reflected the changes In cake prices. Still the model explains 98 percent of the annual variations In price during the 1961-1988 period. Just as the grinds model was plagued by a high degree of positive autocorrelation In the residuals, so the Durbin-Watson statistic suggests the presence of serlal correlation in the residuals as DW equals 1.41. The dL and dJ limits for the 5 percent test Is 1.03 and 1.85. The 1 percent lilts are .83 and 1.62. The estimate at 1.41 falls in the hideterminate range for both and 5 percent and 1 percont tests. Again, the Beach-MacKirnon maximum likelhood proceduro was used to correct for the autocorrelation. The resuWts are: - 3.37 - LGSPt - 6.60 + 1.97*LFPIt 1.19*LSCRt-1 - 2.41*LCCIt (1.6) (6.9) (8.0) (2.1) + RSO8CDt (4.) Corr. R-sq. - .98 SEE - .146 OW - 1.83 d.f. * 23 The Bea%h-MacKinnon procedure results in a DW estimate of 1.83 which Is above the upper linit for the 1 percent test and marginally below the upper limit In the 5 percent cest. The statistic suggests that one cannot reject the null hypothesis of no autocorrelatlon. The coefficlents for the key variables (LSCR, LFPI) In the new model are within 6-7 percent of the initlal parameter estimates and the other estimates (LCCI, CD) are wIthin 10-12 percent of the eariler ones. Table 12 and Chart 7 contain the fItted results for 1961-1988 with projections for the 1989-2000 period. Again, 98 percent of the variation In annual Ghana spot cocoa prices Is explained by the model for the 1961-1988 period. Prices fluctuated over a wide range during the estimation perlod (1 38-2944 sterling) and the model describes the high prices of 1976-1979 and 1984-1985 well. The model projections Indicate that Ghana spot prices will fall below 1,000 sterling during 1990-1993 provided the crop production estimates are realized. It Is quite likely that the potential crops used In the analysis in Table 12 will not occur. Weather problems are likely In one or more years hetween 1989 and 1995. Consequently, the Ghana spot price forecasts probably underestrniate the actual path that prices will take. By the late 19909, world cocoa prices will rise above 2000 sterling. The low prices of the late 19809 and early 19909 are expected to halt plantings of cocoa In all major producers causing production to stagnate In the 2.60-2.70 million tonne range. As a consequence, consumption will catch up with production during the late 1990s, stocks will fail and prices will rise. - 3.38 - 3. Ghana Spot Price Premiom Model (Table 13, Chart 8) One of the requests of the Ghana Cocoa Board Is for a model which explains the Ghana premium versus other world prices. The premium s defined as the difference between the Ghana spot price and London's first terminal price. The same varlables which appeared In the Ghana Spot Price equation were tested and found significant as explanatory variables for the Ghana premium model. The estkiated equation is as follows: LGPt - 20.10 + 2.82*LFPI - 2.65*LSCRt 1 - 7.07*LCCIt (2.0) (4.8) (6.9) (2.6) + .53*CDt (2.0) Corr R-sq - .92 SER - .38 DW - 1.26 where: LGPt - the Ghana PremhAm h year "ti. Again, the coefficients are significant at the .01 levei or better except for the cake dominant estimate which Is statistically significant at the .05 level. The model explains 92 percent of the variation In the Ghana premium during the past 28 years. The OW statistic once again Indicates the presence of positive serlal correlation in the error terms. To correct the problem, the Beach-MacKinnon procedure was used. The maxknum likelihood results are: LGPt - 16.80 + 2.70*LFPit - 2.60*LSCRt-i - 6.23*LCCIt (1.5) (3.8) (6.7) (2.1) + .33*CDt (1.1) Corr. R-sq - .93 SEE - .36 DW - 1.88 d.f. - 22 The coofficlents for the key variables (LFPI, LSCR) are within 2-4 percent of the Initlal ordinary least squares esthnates. The coefficients for the LCCI and constant term - 3 .39 - are 12-16 percent differ ant from the earlier estimates but the cake dominant coeffi%lent deteriorates 38 percont below the eariner esthlato and Is signlficant only at the 20 percent level. Again, the cake dominant variable cculd be improved using the cake ration which peaked In 1978. Had the cake ratio been used Instead of an instrument constructed with zeroes and ones, the coefficient probably would be stronger and show more statistical Ngnificance. The value of the Ghana premhim model can be seen by comparing its estknate for 1988 with the actual premhum that now exists. Currently, the premium Is very high because of Cote d'lvoire's current withholdig scheme. From March, 1988 through September, 1988, Coto d'lvoire sold approxhnately 100,000-150,000 tonnes of cocoa versus a normal sales rate of 300,000-350,000 tonnes because of the low world price. The lack of Cote' dilvolre sales caused premium West African cocoa to become scarce in the world market. Consequently, the premium for Ghana cocoa soared to 250-350 sterling. Given the large quantities of West African cocoa that would be avaliabie If Cote d'lvolre were to change Its selling policy, the model esthuates that the Ghana premium would decline to the 50-80 sterling range. This means that the current Cote d'lvolre selling policy Is benefitting Ghana by 200 to 300 sterling per tonne-a policy highly beneficlal to the Ghanaians and other producers. 4. World Supply/DEmnd and Pric Forecasts Glven the grindgs and price models outlined above, It is possible to combine them with assumptlons about Inflation, world Income and currencies and forecast the paths of world grindings and Ghana spot prices. The forecasts have assumed that world Inflation will Increase at a 4-5 percent rate durhg the 1988-1989 period and then settle down to a 2 percent rate thereafter. World come In real term Is forecast to grow at a 4 percent pace durhg the remander of the decade compared with a 5 percent rate during 1960-1988. - 3.40 - Low world prices are expected to prevail through the early 19909 barring a crop disaster due to weather. Large Increases In world productlon began In the mid-1980s following the El Nlno years of 1982-1983 and 1983-1984 and will continue through 1988- 1989 at least. Prices fell to 13 year lows during the summer of 1988 and world consumption Is beglnning to grwo at a 5-8 percent pace given the low world prices and large world stocks. Low pricos are expected to continue and stinulate consumption. The usage forecast for 1990 Is 2.145 million tonnes--an Increase of 3 percent per annum for the 1985-1990 period (see Table 9-11). The 1995 forecast at 2.600 million or better represents a compounded annual growth rate of 4 percent during the first half of the 1990s. By 1995, consumption wUl catch up with output. Since output is forecast to stagnate by 1995, prices will begin to rise In the mid-1990s In oroer to slow consumption growth and keep It in line wlth output. Cocoa pries (sterling) in 1988 are 3 to 5 tkmes those of the 1960s and early 1970s because of booming world-wide Inflation In the 1970s and the weakness of sterilng. Ghana spot pricos Increased from the 100-300 sterling range In the 1960 early 1970 period to a peak of 3,740 sterling In 1977. Prices then fell to 2,000 by 1985 and to 1,320 by 1987. Currently, Ghana prices are In the 1,000-1,150 range n London. (a) Good Weather Scenario, 19 3-2000 The price forecasts assuming good weathbr throughout the next decade Indicate that Ghana spot prices will fall to 1,130 sterling In 1988-1989 and remain at or below 1,200 sterling through 1994 (see Tables 9-12). The stocks/consumptlon ratio will peak at 58 percent In 1993 and then begin a lonterm downtrend which will cause prices to rise from 1994 through the end of the century. Durhg the early part of the 19909, consumption Is movig rapidly to catch up with productin wnich should occur In the 1993- 1995 period. The stock ratio ther declhes from 58 percent down to 20 percent by the end of the decade. Prices are stagnant in the 900-1,100 sterlhg range between 1989 - 3.41 - and 1994 and then rise from 1,255 In 1995 to 2,400 by 1998 and then to the 4,000- 5,000 range by the year 2,000. (b) El Nino Weather: 1988-2000 From 1970 through 1987, severe drought around the equator occurred five tkies (1972, 1976, 1977, 1982 and 1983). It is possible (the 1960s) for a decade to pass without an El Nino experlence. However, the probabillty of one or more El Nino's Is quite high during a twelve-year period. Consequently, El Nino type weather was introduced Into the forecasting model In 1989, 1993 and 1998 to determine a potential price path. Each time It Is assumed that the El Nlno reduced output by approximately 250,000 tomes. The outcome Is lIsted at the bottom of Table 10. Prices bottom In 1989 near 1,000 sterling (currently prices are marginally below 1100) and then rise In 1989-1090 to the 1,280-1,300 range due to weather problems, fall In 1991 to 1,100 given a good crop and then gradually work higher reaching 1,575 In 1995 and 4,500 by the year 2000. Obviously, long-term price forecasts will be disturbed by weather and changes in farmer pricing policy In one or more major producing countries. Nevertheless, one can reasonably say that world cocoa prices are near the bottom of the current price cycle and will begin to work higher during the early 1990s. The upswing In prices will depend on weather. A drought in the major growing regions could cause the turn at any time. If the market escapes a drought In the next five years, consumption will catch up with production and the upswing will begin In the 1994-95 perlod. The world cocoa market envirorvnent just outilned Is the one In which Ghana cocoa policies must be Integrated. Given the view that world cocoa prices are near their cyclical lows and that prices will begin rlshg somethle before 1995, an opportunity exists for the Ghanalans to hierase production at a reasonable pace. An analysis of the Ghana cocoa market environment takes place Is given In the aenex. - 3. 42 - 111. GHANA'S COCOA AND FOREIGN EXCHANGE POUCIES The above anaysis generates three iortant conclusions for Ghanaian cocoa policy. The first concerns the producer price required to Increase Ghanalan cocoa output. The second concerns the rate at which cocoa prices should be Increabed. The third concerns the necessity of adjusting the cedi rate against other currencles In order to maintain purchasing power parity. The Government of Ghana should adjust the cedi price of cocoa without distorting Internal food-cocoa Price relatlonships. The following summarizes the recommendations and conclusions. A. Effective Cocoa Producer Price, 1988/9 The real producer price analysis and the cost of production estimates are consistent In suggestig that an appropriate producer price for the 1988/9 cocoa year would be 130-165 cedis per kilo. At 150 codis, the real price of cocoa on a 1963 Index basis Is 74. this Is In Ihe with the real pricos of the late 1960s and early 19709 -- prices which allowed full harvesting, reasonable maintenance and some replanting. From a cost of production point of view, the new 1988/9 farmer price of 165 cedis per tonne covers total production costs Including maintenance, harvesting and establishment costs. IL Rate of Farer Price Adjustmnt What should be the price in 1989/90? Two other questions must be answerecd before the prico question. The first Is "What Is Ghana's inflatlon rate?" The second Is "At what pace do authorities want cocoa production In Increase?" If internal Inflation continues at a 36 percent pace (the rate from the 1986 fourth quarter to the 1987 fourth quarter), then the cocoa pric pald the farmer should be Incroased to 225,000 cedis per tonne in order to keep the farmer's real income constant. If the Inflation rate fals below 36 percent, then the upward price move required to maintain farmer Wicome will be smaller. There Is considerable evidence to suggest that replanthg of cocoa Is taking place In food farms at the current real farmer price (70 on a 1963 basis using 165 cedis per 3. 43 - kilo and a CPI factor of 108,000 (1963-100)). If 15-20 percent of the currert cocoa area Is new cocoa as observed by the author In October 1987, then maintaining real farmer prices In the 70-75 range (based on a 1963 real price - 100) should result In additInal planting over the next few years and lad to a Ghana crop of 300,000 tonres or more In the early 19900. If the Goverr,ent of Ghana wants productlon to move to the 400,000 tonne range in the 1995-2000 pIod, then real farmer :rices should be raised above the 70-75 area. This would require prices in 1989 to be incrased to the 250,000 level assuming an Inflation rate of 25 percent. The recommendation in this PaPer Is that real producer prices be maintained in the 70-75 range but that a special sample survey be conducted during the next year to determine the amount of planthg that Is occurring and to measure the percentage of new cocoa that exists. This can be done relatively cheaply If the right methodology and techniques are used. C. Fordg Excae Rate Ad)Astent Fortunately, Ghana has already begun a forelgn exchange auctlon In order to keep the cedi In line with Its true purchasing power. It Is kIportant that the process be continued and that It be Improved to ensure that the vahe coming from the auction approxlmates a free-market value. The wider the participation In the auction, the more representative the currency value will be. One of the problems with the current auction is that the use of the foreign exchange Is restricted. This Is fl key factor In the rate not being representative for the cocoa farmer. The major factor causing the decle In Ghana's cocoa kidustry during the 1965- 85 perod was a fIxed rate of exchange. A serxusly overvalued cedl restrictod the actions of the Government In providing the farme wIth a remunerative price for cocoa. It Is kwrative that the rate of exchange maintain Ghana's pruchashg power parity nternally so that those who gonerate foreign exchange can recelve adequate rewards. - 3,.44 - IV. CONCLUSIONS AND SUMMARY Two major conclusons are dorivod by the study. The first Is that world cocoa prices will be low durlng the next five years which will curtail plantigs In most major producing countries. If Cote d'lvoire resumes a normal sales policy durhig 1988/9, world cocoa prices will collapse to the 800 to 1.000 sterilng level. Producers who are already financially stressod by current prices will exporiece chaotic financial condistlons. PlanWng will cease, puts will be further reovod, and the growth In cocoa production will come to a halt. This wil occur in the 1988-1992 peod. When production stagnates, prices will begin to rise and the prico forecasts for the second half of the 1990s nduIate hlgNy remunerative prices. If Ghana wll continue its current farmer prickg policy whik allows planthV to occur, Ghana's cocoa output and foreign exchange revenues durig the 1995-2005 period wiN be much larger than If the country reacts negatively to the low world price and lowers ts fmr price causing Ghanaian output to stagnate. A major threat to Ghana's ablity to maintain remunerative internal cocoa prices Is Its foregn exchange pollcy. If the oCl is allowed to remain overvalued, It will be dIfftl !t over tkne to keep Internal cocoa prices high enough to ensure moderate growth In production. Ghana's cocoa production costs are in lne with other producers. The key to Ghana's cocoa future will be Its farmer price. IV. INTEGRATION OF POUCY IMPUCATIONS FRCrM INDIVIDUAL STUDIES AND RECOMMENDATIONS IV. INTEGRATION OF POLICY IMPLICATIONS FROM INDIVIDUAL STUDIES AND RECOMMENDATIONS The Integration of the studles will proceed In reverse order, starting wlth the International cocoa market, so that the discussion of Ghanaian policy logically follows the assessment of International market conditions upon which natinal policy must be based. Dr. Bateman expects world cocoa prlios to be significantly below the 1987 level (I.e., 25 percent) by 1990, to advance slightly above the 1987 level by 1995, reaching a level two or three times the 1987 level by 2000. The offect of the price decreases will be to decrease the Intensity of cultivation and hence production In the short run and to roduce plantings of new trees and thus future production. In fact, world cocoa prices tend to follow a 25-year cycle that corresponds to the offective economic llfethue of a cocoa tree. World prices are approaching thelr cycilcal low and they will reach their cyclical high somewhat after the year 2000. Thus, cocoa Investments made In the next few years of low prices will be In full production when the next cyclical high In real cocoa prices occurs. In other words, a counter-cyclical strategy of cocoa plantings will produce larger returns (provided other cocoa producing nations do not buffer cocoa farmers from the effects of low international prices). The ev!Jence from the Survey for Cocoa Producer Pricing Indicates a strong producer supply response to adequate price incentives--the survey showed that for the sample of farmers Interviewed, 33 percent of existing cocoa acreage has been planted in the past four years. The results from the shnulatlon exporknents with the multl-period agricultural sector model strongy support the existence of a robust producer supply response to adequate price Incentives. They also Indicate that any reductions of the real buying price of cocoa below 140 cedis per klogram (in 1987 prices) will result In significant reductins In export earnngs from cocoa. Clearly the sbnulatlon exporkments with positive welfare kmplicatlons for the GhanaIkn people are those with growth In population and income, with or without a food self-sufficiency' pollcy; and this kIlnd of - 4.2 - scenario Is the objective of the Economic Recovery Program. One bmpilcatlon of these experiments Is that a price of 120 (1987) cedis per kilogram Is not sufficient to sustain cocoa production over the longer term when Incomes are growing; If a food grain self- sufficlency policy is adopted, neither Is a price of 140 cedis. Although sizeable volumes of cocoa can be produced at these prices over the medium term, say ten years. At higher prices, cocoa supply response Is sustained over both medium- and long-term horizons. Thus, the questlon of the optimal level of cocoa taxation has a dynamic dimension that asks where will the competitive trade advantage for Ghana be In the long run? Past policy In Ghana can be Interpreted as having assumed that long-run advantage did not Include cocoa, but that assumption has not turned out to be true.' At the very least, pollcy ought to consider mixed strategies which admit that present expectations of the future can be mistaken. If we admit that cocoa may likely be part of Ghana's long-term advantage, then the yearly cocoa buying should never be set below 140 cedis (In 1987 prices). It should be set above that level If growth In long- term supply Is preferred. Both of the sconarios positing Income growth, wlth or without food grain self- sufficlency, show substantial growth In employment as cocoa prices are Increased, reflecting the greater demand for resources as Incomes Increase and more resources are shifted to labor Intensive cocoa production. Thus, a low cocoa tax, high cocoa production policy kIpacts quite positively on the Incomes of the rural poor. Of course, the food grain self-sufficiency policy also bnpacts drastlcally on cocoa production at lower prices. The reason Is quite simple. This policy shifts resources away from more productive employment In cocoa production toward less effticent production In maize. It may be argued that this outcome Is really a consequence of the assumptlon of static agricultural technology. However, until convincing evidence Is at hand that Ghanaian agriculture Is moving to cost officient higher productivity practices, a food self- sufficlency policy entails heavy efficiency losses. On the other hand, combining food - 4.3 - seif-sufficlency with a high cocoa price policy does yield a strong Impact on employment of rural labor and smaliholder farmer Incomes. The greater knpact on the welfare of the poor might be deemed an acceptable tradeoff for lower static efficiency. Moreover, if It Is accepted that productivity enhancing technical change In agriculture Is more likely In the context of rlsing agricultural employment and incomes, then dynamic officiency gains are more likely under the food self-sufficiency scenario. An annual decomposition of selected pricing policies over the 1988-97 decade showed that under a likely scenario -- growth In populatlon, Income and wages wlth a cocoa buying price of 140 cedis (In 1987 prices) -- cocoa production would remain at or below 250,000 metric tons per year through 1990 and then rlse rapidly to reach a level of 500,000 metric tons by 1997. The reason for the lag In productlon growth Is the long gestation lag after planting before cocoa trees produce cocoa at something approaching their potential. This same scenario shows total cocoa acreage remaining relatively constant as obsolete traditional varietles are replaced by higher yielding hybrld varietles. Thus, the same acreage with relatively mature hybrid trees Is capable of much greater production. The lag In cocoa production also knplies a corresponding lag In farm Income, which does not Increase significantly In real terms until 1991 and later. In the Interim period before farm Income rises In real terms, policies which heavily tax cocoa are capable of destroying the expectatlons of long-term profitabillty that Induced the planting Investments of the past four yearn and thus abort the resurgence of coco a production over the longer term. The analysis of cocoa tax and revenue alternatives reached the following conclusions. The case for amouncing and demonstrating a comnitment to a higher real producer prie for cocoa appears strong, and the risks of setting the price too low appear to be higher than the risks in the other direction, The first priority In setting the producer price of cocoa is to ensure that It Is above the level at which farmers are Just willing to plant and maintain cocoa. While this seems self-evident, It has not been - 4.4 - the past policy of the Government, and It would not necessarily be ensured by linking the producer price to the export price level. The main way In which the fall In tax revenue from the cocoa tax and the profits of COCOBOD can be offset are first of all to move the exchange rate to Its equlilbrium level so that the Government will have access to additional purchasing power which Is significant compared to total current tax revenues. The next step Is to explore the extent to which taxes can be shifted from cocoa producers to all consumers by Increasing taxes on appropriate consumer goods. The Ideal subjects for such taxation are those which are Income elastic and price inlastic - gasoline, cars, consumer durables, etc., are all good examples. The greater the scope for consumer good taxation, the hgher should be the price of cocoa. There, also appears to be good reason to try and stabilizo the real producer price of cocoa, both because the tax rate legitimately Increases with the world price, and because on Income smoothhg ground prico stabilization lb moderately efficacious. Finally, even if there does remaln a (rather modest) case for an output or export tax on cocoa, there seems little point in moving to the more complex alternatives of an agricultural Income tax or a land tax. Analysis of data from the GLSS and AES surveys shows that cocoa farmers are rather poorer than the typical consumer of liported or taxed goods, and that producers of non-traded foods are very considerably poorer than consumers of these foods. Results from the multi-period agricultural sector model and the tax Impact analysis show that reducing cocoa taxation pulls more land and labor resources Into cocoa production and thus hiducing hcreases In the price of non-traded foods and the emloyment of agricultural labor. Glven his, It Is desirable If pokcy has some aversion to hequalIty to shift real icome from conumers of non-traded foods to farmers producing them and from consurs of taxed goods to farmers. Taxing cocoa farmers does the reverse on both scores, for It dkectly taxes the income of cocoa farmers, and Indirectly lowers the Income of fooo producers and transfers Incomes from these farmers to food consumers. - 4.5 - There are good grounds for stabilizing the real producer price of cocoa. That is above the level at which farmers are just willing to plant and malntain cocoa, the a tax rate can Increase with the world price. A variable rate tax on cocoa (above the critical level) results In a sharing of the world market price risk between cocoa farmers and the rest of society (assuming that revenue requirements are fIxed). This both stabilizes the real incomes of cocoa farmers, and lets consumers also share In the windfall profits trom high cocoa prices. There Is a case for raising some revenue from cereals hport duties if this allows taxes on other goods (either cocoa or currently taxed representative consumer goods - - though not necessarily particular consumer taxes, eg., on gasoline) to be reduced. The reason Is shiple -- tariffs ralse revenue, and the effect In encouraghg cocoa farmers to switch Into maize production and hene lower cocoa tax revenue Is relatively slight, at least at low levels of cereal duty and adequate levels of the cocoa price, In addition, tariffs redistribute Income from consumers to cereal farmers. Further, by inducing a switch Into cereals and out of non-traded roots, the price of roots rises and further redistributes Income to the poorer farmers. Finally, cereal tariffs reduce national dependince on Imported foods. Our mrin recommendation Is to continue the policy of providing an adequate price Incentive tc cocoa producers. This may entail moving to a pricing forrdula based on real costs ancd not on a proportion of the f.o.b. price. While we support the World Bank advice on giving the farmer at least 55 percent of the f.o.b. price as soon as possible, It should be noted that the recent fall In world prices moves the 55 percent of f.ob. critorion clo to a real cost level at which any rduction would be quite damaging to future export revenues. Any further reductions In the world price level could push the 55 percent of f.o.b. Price below the level consistent with an expanding future cocoa production. A corollary of this main recommendatlon Is that an efficient substitute for revenue from the cocoa export tax should be found knmedlately. We fInd the arguments - 4.6 - for Increased taxation of vehicles and fuel to be persuasive In this respect, although other alternative revenues sources are available. With respect to short-run pricing pollcy, we recommend that in real terms the buying price not be dropped below the level of 140 cedis per kilogram that provalled In 1987-88. The simulation experiments Indicated a sharp decrease In future cocoa export earnings If the buying price Is dropped below this level. Moreover, any drop In the real price offered to the cocoa farmer will inevitably bring back memories of the m!sgulded policies of the past and inpact unfavorably on the expectations of cocoa farmers. We note that with the 40 percent Increase In the general price level over the 1987-88 crop cycle, the equivalent of the 140 cedis hI 1987 prices is 195 cedis per kilograms over, the 1988-89 cycle. At a number of points In our discussions with officials of the Government of Ghana, a pollcy of diversificatIon of exports was discussed, and most of them approved of such a policy. It was then qulte disappointing to discover that very little Is being done to promote alternative exports, and there are policies in place that are actually counterproductive. For example, Ghana produces only a little coftee; and export earnings from this crop are negligible. Yet, coffee Is taxed at the same level as cocoa, despite the probability that the net revenue from this tax is negative. It would be a simple matter to eliminate taxation of this commodity and privatize the marketing of It. Government net revenues would galn, and coffee producers would be given a powerful Incentive to expand productIon, particularly In those areas for which swollen shoot virus infestation has badly reduced the profits from cocoa. ANNEX I MULTIPERIOD AGRICULTURAL SECTOR MODEL AND NUMERICAL SIMULATION ESTIMATES USED IN FIGURES I.C.1-l.C.25 by AIexander Meraus William Aseso Okyore Grald T. O'Mara Abbeo: ndp: man days pe awe cpa: cdis per aa upa: unt era GAMS 2.05 PC AT/XT 89/01, 2 15 i,:, i'A_ 2 AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SET DATA 32 sets 33 34 r agricultural regions / north, south / 35 rp(r) regions with perennials / south / 36 37 c crops / maize, rice, cassava, millet, sorghum, g-nut, 38 bean-pea, cocoyam, yam, cocoa, palmfruit, plantain / 39 40 ca(c) annual crops / maize, rice, cassava, millet, sorghuiii, g-nut 41 bean-pea, coc3yam, yam / 42 cp(c) perennial crops / cocoa, palmfruit, plantain / 43 44 cn(c) nationally consumed crocps 45 cm(c) 4 nported crcps 46 ce(c) exported crops 47 48 49 rc(r,c) crop possibilities 50 51 a annual crop processes / maize-1, maize-2, cassava-l, cassava-2, 52 rice, e-millet, 1-millet, sorghum, g-niuc 53 bean-pea, cocoyam, yam / 54 ap perennial production / cocoa-t, cocoa-h, oil-palm, plantain / 55 56 ra(r,a) process possibilities 57 58 t time in months / jan,feb,mar,apr,may,jun,iul,aug,sep,oct,nov,dec / 59 60 v vintage / 0-5, 6-10, 11-15, 16-20, 21-25, 26-30, 31-99 / 61 vf(v) first vintage / 0-5 / 62 63 th time horizon / 1985-89, 1990-94, 1995-99, 2000-04, 2005-09 64 2010-14, 2015-19, 2020-24, 2025-29, 2030-34 65 2035-39 / 66 ti(th) initial period / 1985-89 / 67 TP(TH) TIME PERIODS/ 1990-94, 1995-99, 2000-04, 2005-09 68 2010-14, 2015-19, 2020-24, 2025-29, 2030-34 69 2035-39 / 70 71 vtov(v,v) aging / 0-5.6-10, 6-10.11-15, 11-15.16-20, 16-20.21-25 72 21-25.26-30, (26-30,31-99).31-99 / 73 74 in inputs / seed 75 suckers 76 wiremesh 77 insect insecticides 78 fuel 79 spray-rnt sprayer rental 80 chain-rnt chain saw rental 81 ax 82 cutlass machete 83 hoe 84 sickle 85 baskets 86 drymat drying mats / GAMS 2.05 PC AT/XT 89/01/2' 15:15:u' PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SET DATA 87 88 k task / tree-fell, burn, clear, land-prep 89 nurse, plant, stake, weed, trap, shade 90 ring, harvest, shell 91 crop-care, harv-proc, tr-home, tr-market / 92 93 alias (v, w), (c,cc), (cn,cnn), (r,rr); 94 95 scalars length length of time periods / 5 /; CAKS 2.05 PC AT/XT 89/01/27 15:15:03 PACE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN PERENNIAL CROP DATA 97 98 parameters 99 100 pricein(in) price of inputs (cedis per unic) 101 102 / seed 1000, suckers 1000, wiremesh 9000 103 insect 1070, fuel 67, spray-rnt 100 104 chain-rnt 750, ax 1500, cutlass 500 105 hoe 500, sickle 250, baskets 100 106 drymat 100 / 107 108 landclin(in) land clearing input requirements (upa) 1.09 110 / cutlass 2, ax .33, chain-rnt 1, fuel 4 / 111 112 landclab(k) land clearing labor requirements (mdpa) 113 114 / tree-fell 15, burn 1.1, clear 12.1 115 116 plantpin(ap,in) perennials input requirements for planting (upa) 117 118 / (cocoa-t,cocoa-h) .seed .5 119 (cocoa-t,cocoa-h,oil-palm,plantain).suckers 5 120 oil-palm wiremesh 1.5 / 121 122 freq(r) fallow land required for rotation of annuals (fractions) 123 124 / south .25, north .2 / 125 126 table plantplb(k,ap) perennials labor requirements for planting (indpa) 127 128 (cocoa-t,cocoa-h) oil-palm plantain 129 nurse 8.5 7.6 130 plant 9.3 3.6 8.8 131 ring 5.0 132 weed 9.3 12.3 13.9 133 trap 17.0 134 shade 12.1 135 136 table labprcl(k,t) labor proportions for land clearing (prop) 137 138 jan feb mar apr may jun jul aug sep oct nov dec 139 tree-fell 1 1 .5 140 burn 1 141 clear 1 142 143 table labprpl(ap,k,t) labor proportions for planting perennials (prop) 144 145 146 jan feb mar apr may jun jul aug sep oct nov dec 147 (cocoa-t, cocoa-h).nurse 1 .5 148 (cocoa-t, cocoa-h).plant .5 1 149 (cocoa-t, cocoa-h).weed 1 1 150 (cocoa-t, cocoa-h).trap 1 151 (cocoa-t, cocoa-h).shade 1 UMS 2.05 PC AT/XT 89/01,27' 15:15:03 PACE ;RICULTURAL SECTOR MODEL FOR GHANA- BASE RUN RENNIAL CROP DATA L52 L53 oil-palm nurse 1 .5 .54 oil-palm plant .5 1 .55 oil-palm ring .5 1 .56 oil-palm weed 1 1 .57 .58 plantain .plant 1 .5 .59 plantain weed 1 1 .60 .61 parameters tlabppl(ap,k) total labor proportions planting (prop) .62 tlabpl(t,ap) total labor requiremnts planting (mdpa) .63 tlabpcl(k) total labor proportions clearing (prop) tlabcl(t) tctal labor requiremnts clearing (indpa) incostpl(ap) total input cost planting (1000 cedis per acre) incostcl total input cost clearing (1000 cedis per acre) .67 tlabppl(ap,k) - sum(t, labprpl(ap,k,t)); tlabpcl(k) - sum(t, labprcl(k,t)).; tlabpl(t,ap) - sum(k$tlabppl(ap,k), plantplb(k,ap)*labprpl(ap,k,t)/tlabppl(ap.k)); tlabcl(t) - sum(k$tlabpcl(k), landcllab(k)*labprcl(k,t)/tlabpcl(k)); incostpl(ap) - sum(in, plantpin(ap,in)*pricein(in))/1000; incostcl - sum(in, landclin(in)*pricein(in))/1000; display tlabppl, clabpc., tlabpl, clabcl, incostpl,incostcl; table ireqp(ap,in,v) input requirements for perennials (units per acre) 0-5 (6-10,11-15,16-20,21-25,26-30,31-99) cocoa-t. insect 1.2 2.0 cocoa-t. fuel 1.3 2.0 cocoa-t. spray-rnt 2.0 2.0 cocoa-t. cutlass 3.0 2.0 cocoa-t. sickle .25 cocoa-t. baskets .5 2.5 cocoa-t. drymat 1.0 cocoa-h. insect 1.2 2.0 cocoa-h. fuel 1.3 2.0 cocoa-h. spray-rnt 2.0 2.0 cocoa-h. cutlass 3.0 2.0 cocoa-h. sickle .25 .25 cocoa-h. baskets 1.5 2.5 cocoa-h. drymat 1.0 1.0 oil-palm. insect oil-palm.fuel oil-palm.spray-rnt oil-palm.cutlass 4.0 4.0 oil-palm.sickle oil-palm.baskets 2.0 5.0 oil-palm.drymat plantain. insect plantain.fuel plantain.spray-rnt GAMS 2.05 PC AT/XT 89/01/27 15:15:03 PACE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN PERENNIAL CROP DATA 207 plantain.cutlass 2.0 2.0 208 plantain. sickle 209 plancain.baskets 1.0 1.0 210 plantain.drymat 211 212 * note that cocoa is interplanced in yesr 0-5 with plantain and 13 * cocoyam (first year only). the inputs and labor tasks are for 14 * both, perennial and annual intercrops. 15 16 table ttreqp(ap,k,v) total task requirements: perennials (mdpa) 17 18 0-5 6-10 (11.15,16-20,21-25,26-30,31-99.) 19 cocoa-c. crop-care 31.4 31.4 21.6 20 cocoa-,. harv-proc 66.6 66 6 21 cocoa-t. rtr-market .2 5.4 5.4 22 cocoa-t. harvest 1.4 23 cocoa-t. tr-home 1.4 24 cocoa-h. crop-care 31.4 31.4 25.0 25 cocoa-h. harv-proc 40.0 66.6 66.6 26 cocoa-h. tr-market 4.2 5.4 5.4 27 cocoa-h. harvest 1.4 28 cocoa-h. tr-home 1.4 29 oil-palm.crop-care 30.0 30.0 27.0 30 oil-palm.harv-proc 21.6 24.8 31 oil-palm.tr-market 2.4 4.0 32 plantain.crop-care 21.6 21.6 21.6 33 plantain.harv-proc 12.0 12.0 9.6 34 plantain.tr-market 3.6 4.8 4.8 35 36 37 table labpp(ap,k,t) labor requirements for perennials (prop) 38 39 Jan feb mar apr may jun jul aug sep oct nov dec 240 cocoa-t. crop-care .6 1 .3 .6 .7 241 cocoa-t. harv-proc 1 1 1 1 1 1 242 cocoa-t. tr-marker 1 1 1 1 1 1 243 cocoa-t. harvest 1 1 244 cocoa-c. tr-home 1 1 245 cocoa-h. crop-care .6 1 .3 .6 .7 246 cocoa-h. harv-proc 1 1 1 1 1 247 cocoa-h. tr-market 1 1 1 1 1 1 248 cocoa-h. harvest 1 1 249 cocoa-h. tr-home 1 1 250 oil-palm. crop-care .5 .6 .5 1 .6 251 oil-palm. .harv-proc 1 1 1 1 1 1 1 1 252 oil-palm. tr-market 1 1 1 1 1 1 1 1 253 plantain. crop-care 1 1 1 1 254 plantain. harv-proc 1 1 1 1 1 1 1 1 1 1 1 1 255 plantain. tr-market 1 1 1 1 1 1 1 1 1 1 1 1 256 257 258 parameters icostp(ap,v) total input cost for perennials (1000 cpa) 259 tlabr'ap,v) total labor requirements (mdpa) 260 ptotpfap,k) totals for normalizing (prop) 261 labreqp(ap,t,v) total labor requirements (indpa); GAMS 2.05 PC AT/XT 89/01/27 15:13:03 PACE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN PERENNIAL CROP DATA 262 263 icostp(ap,v) - sum(in, ireqp(ap,in,v)*pricein(in))/1000; 264 tlabp(ap,v) - sum(k, ttreqp(ap,k,v)); 265 ptotp(ap,k) - sum(t, labpp(ap,k,t)); 266 labreqp(ap,t,v) - sum(k$ptotp(ap,k), labpp(ap,k,t)*ttreqp(ap,k,v) 267 /ptotp(ap,k)) 268 display icostp, tlabp, ptotp, labreqp; CAMS 2.05 PC AT/XT 89/01,'27 15:15:03 PAG'E 8 AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN ANNUAL CROP DATA 70 table ireqa(r,a,in) input requirements for annuals (upa) 71 72 seed cutlass hoe 73 south.(maize-l,maize-2,rice,e-millet 74 1-millet,sorghum,g-nut,bean-pea) 2 75 north.(maize-l,maize-2,rice,e-millet 76 1-millet-sorghum,g-nut,bean-pea) 2 77 * note: seed requirements for cropping above 78 * are netted out from yield (see nyielda) 79 north.(cassava-l,cassava-2) .525 2 80 south,(cassava-l,cassava-2) .525 2 8: north.cocovam 1.600 2 82 south.cocoyam 1.600 2 83 north.yam 7.500 2 84 south.yam 7.500 2 85 86 87 table ttreqa(r,k,a) total task requirements: annuals (md per acre) 88 89 maize-I maize-2 (cassava-l,cassava-2) cocoyain vam 290 south.land-prep 5.9 5.9 5.9 5.9 5.9 291 south.plant 5.7 5.7 8.0 8.0 8.0 292 south.weed 14.2 14.2 14.2 14.2 14.2 293 south.stake 7.5 294 south.harvest 6.3 6.3 14.7 14.7 14.1 295 south.tr-home 9.1 9.1 13.7 13.7 13.7 296 scuth.shell 13.6 13.6 297 south.tr-market 1.0 1.0 2.5 1.3 1.3 298 299 north.land-prep 6.8 11.2 300 north.plant 3.3 4 l 301 north.weed 8.1 L6.2 302 north.stake 7.5 303 north.harvest 5.5 8.3 304 north.tr-home 7.8 13.0 305 north.shell 9.1 306 north.tr-market 1.6 307 308 + rice e-miillc l-millet sorghum g-nuc bean-pea 309 north.land-prep 6.8 6.8 6.8 6.8 310 north.plant 0.8 4.1 2.0 3.1 6.8 6.8 311 north.weed 8.1 8.1 8.1 8.1 4.1 4.1 312 north.stake 8.1 8.l 313 north.harvest 10.7 5.3 5.3 5.2 314 north.tr-home 7.1 7.1 7.2 7.1 9.6 9.6 315 north.shell 5.2 6.7 6.7 4.6 5.6 5.6 316 north.tr-market 12.0 317 GAMS 2.05 PC AT/XT 89/0L,/'27 15:1:u3 PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN ANNUAL CROP DATA 319 table labpas(a,k,t) labor requirements for annuals south (prop) 320. 321 jan feb mar apr may jun jul aug sep oce nov dec 322 maize-i. land-prep 1 1 323 maize-i. plant 1 1 324 maize-1. weed 1 1 1 325 maize-1. stake 326 maize-i. harvest 1 1 1 327 maize-1. tr-home 1 1 1 328 maize-1. shell 1 1 1 1 329 maize-1. tr-market 1 1 1 1 330 * jan feb mar apr may jun jul aug sep oct nov dec 331 maize-2. land-prep 1 1 332 maize-2. plant .5 1 .5 333 maize-2. weed 1 1 334 maize-2. stake 335 maize-2. harvest 1 1 L 336 maize-2. tr-home 1 1 337 maize-2. shell 1 1 1 338 maize-2. tr-market 1 1 1. 339 * Jan feb mar apr may jun jul aug sep oct nov dec 340 cassava-l.land-prep 1 1 341 cassava-l.plant .5 1 1 342 cassava-l.weed 1 1 1 1 343 cassava-l.stake 344 cassava-l.harvest 1 1 1 1 1 1 1 345 cassav;.. l.tr-home 1 1 1 1 1 1 1 346 cassava-l.shell 34; cassava-l.tr-market 1 1 1 1 1 1 1 348 * jan feb mar apr may jun jul aug sep oct nov dec 349 cassava-2.land-prep 1 1 350 cassava-2.plant 1 1 1 351 cassava-2.weed 1 1 1 1 352 cassava-2.stake 353 cassava-2.harvest 1 1 1 1 1 1 354 cassava-2.tr-home 1 1 1 1 I L 355 cassava-2.shell 356 cassava-2.tr-market 1 1 1 1 1 1 357 * jan feb mar apr may jun jul aug sep oct nov dec 358 cocoyam. land-prep 1 1 359 cocoyam. plant 1 1 360 cocoyam. weed 1 1 361 cocoyam. stake 362 cc:oyam. harvest 1 1 363 cocoyam. tr-hom. e 1 364 cocoyam. shell 365 cocoyam. tr-market 1 1 366 * jan feb mar apr may jun jul aug sep oct nov dec 367 yam. land-prep 1 1 1 368 yam. plant .5 1 1 369 yam. weed 1 1 1 1 370 yam. stake 1 1 1 371 yam. harvest 1 1 1 1 1 1 1 372 yam. tr-home 1 1 1 1 1 1 373 yam. shell GAMS 2.05 PC AT/XT 89/01/27 15:L5:U3 PAiE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN ANNUAL CROP DATA 374 yam. tr-market 1 1 1 1 1 1 1 1 .-,t GAMS 2.05 PC AT/XT 89/01/27 15:!D5:0' PAO E RICULTURAL SECTOR MODEL FOR GHANA- BASE RUN NUAL CROP DATA 76 table labpan(a,k,t) labor requirements north (prop) 77 78 jan feb mar apr may jun jul aug sep oct nov dec 79 maize-i. land-prep 1 1 80 maize-i. plant 1 1 81 maize-1. weed 1 1 382 maize-l. stake 383 maize-i. harvest 1 1 384 maize-l. tr-home 1 1 385 maize-i. shell 1 .5 1 386 maize-l. tr-market 387 * jan feb mar apr may jun jul aug sep oct nov dec 388 rice. land-prep 1 1 .5 389 rice. plant 1 1 390 rice. weed .5 1 .5 391 rice. stake 392 rice. harvest .5 1 .5 393 rice. tr-home .5 1 .5 394 rice. shell 1 1 1 395 rice. tr-market 396 * jan feb mar apr may jun jul aug sep oct nov dec 397 e-millet.land-prep 1 398 e-millet.plant 1 1 399 e-millet.weed .5 1 .5 400 e-millet.stake 401 e-millet.harvest 1 1 402 e-millet.tr-home 1 1 403 e-millet.shell .5 1 1 .5 404 e-millet.tr-market 405 * jan feb mar apr may jun jul aug sep oct nov dec 406 1-millet.land-prep 1 1 407 l-millet.plant 1 408 1-millet.weed 1 1 409 l-millet.stake 410 1-millet.harvest .1 1 1 411 1-millet.tr-home .1 1 1 412 l-millet.shell 1 1 1 413 1-millet.tr-market 414 * jan feb mar apr may jun jul aug sep oct nov dec 415 sorghum. land-prep 1 .6 416 sorghum. plant 1 .5 417 sorghum. weed .5 1 .5 418 sorghum. stake 419 sorghum. harvest .1 1 1 420 sorghum. tr-home .4 1 1 421 sorghum. shell .3 1 1 422 sorghum. tr-market 423 * jan feb mar apr may Jun jul aug sep oct nov dec 424 g-nut. land-prep 1 .5 425 g-nut. plant 1 .5 426 g-nut. weed 1 ; 1 427 g-nut. stake 428 g-nut. harvest 1 1 1 429 g-nut. tr-home 1 1 1 430 g-nut. shell .5 1 1 .5 GAMS 2.05 PC AT/XT 89/01/27 15:15:03 PAGE 12 ACRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN ANNUAL CROP DATA 431. g-nut. tr-market 432 * Jan feb mar apr may jun Jul aug sep oct nov dec 433 bean-pea.land-prep 1 .5 434 bean-pea.plant .5 1 435 bean-pea.weed 1 i 1 436 bean-pea.stake 437 bean-pea.harvest 1 1 1 438 bean-pea.tr-home 1 1 1 439 bean-pea.shell .3 1 1 440 bean-pea.tr-market 441 * jan feb mar apr may jun jul aug sep oct nov dec 442 yam. land-prep 1 1 .5 1 443 yam. plant 1 1 1 1 444 yam. weed .5 1 1 445 yam. stake .4 1 1 .4 446 yam. harvest .5 1 1 1 1 1 447 yam. tr-home .5 .1 1 1 1 1 448 yam. shell 449 yam. tr-market .5 1 1 1 1 1 450 451 parameters icosta(r,a) total input cost for annuals (1000 cpa) 452 tlaba(a,r) total labor requirements (mdpa) 453 ptot(r,a,k) totals for normalizing labor (prop) 454 labreqa(r,a,t) total labor requirements:annuals (mdpa) 455 labpa(r,a,k,t) labor requirements: annuals (prop): 456 457 icosta(r,a) - sum(in, ireqa(r,a,in)*pricein(in))/1000; 458 tlaba(a,r) - sum(k, ttreqa(r,k,a)); 459 labpa("south",a,k,t) - labpas(a,k,t); 460 labpa("north",a,k,t) - labpan(a,k,t); 461 ptot(r,a,k) - sum(t, labpa(r,a,k,t)); 462 labreqa(r,a,t) - sum(k$ptot(r,a,k), labpa(r,a,k,t)*ttreqa(r,k,a) 463 /ptot(r,a,k)); 464 display icosta, tlaba, ptot, labreqa; CAMS 2.05 PC AT/XT 89/01,',7 15:15:03 PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN YIELD DATA 466 table yieldp(c,ap,v) yield of perennials in 1987 (kg per acre) 467 468 0-5 6-10 11-15 16-20 21-25 26-30 31-99 469 cocoa. cocoa-t 100 160 160 160 120 80 470 COCOA. COCOA-H 20 200 240 240 240 180 120 471 palmfruit.oil-palm 1350 2330 2330 1350 900 450 472 plantain. plantain 960 2230 473 474 475 table yieldpi(c,ap) additional yield from first period (kg per acre) 476 477 cocoa-t cocoa-h 478 plantain 480 480 479 cocoyam 200 200 480 481 *note the yield figures are averaged over years 0 tO 5 482 483 table yield(r,c,a) process yields (kg per acre) 484 485 maize-I maize-2 cassava-l cassava-2 cocoyam yam 486 south.maize 370 370 487 south.cassava 2870 2870 488 south.cocoyam 1920 489 south.yam 1550 490 491 north.maize 315 492 north.yam 1950 493 494 + rice e-millet 1-millet sorghum g-nut bean-pea 495 north.maize 496 north.rice 457 497 north.millet 243 230 498 north.sorghum 261 499 north.g-nut 397 500 north.bean-pea 250 501 502 table seed(c,a) seed requirements (kg per acre) 503 504 maize-l maize-2 505 maize 40 40 506 cassava 507 cocoyam 508 yam 509 510 + rice e-millet 1-millet sorghum g-nut bean-pea 511 rice 45 512 millet 30 30 513 sorghum 30 514 g-nut 25 515 bean-pea 25 516 517 parameter nyielda(r,c,a) net yield for annuals (kg per acre); 518 519 nyielda(r,ca,a) - max(yield(r,ca,a)-seed(ca,a),O); 520 CAMS 2.05 PC AT/XT 89/01/27 15:15:03 PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN YIELD DATA 521 rc(r,ca). - sum(a$nyielda(r,ca,a), yes); 522 rc(rp,cp) - yes; 523 ra(r,a) - sum(ca$nyielda(r,ca,a), yes); 524 525 option nyielda:O; display nyielda, rc, ra; GAMS 2.05 PC AT/XT 89/01/27 15:15:03 PAGE 15 AGRICULTURAL SECTOR MODEL FOR GHANA-' BASE RUN LAND AND LABOR DATA 527 table stock(r,ap,v) stock of perennials from survey 1987 (acre-s 528 529 0-5 6-10 11-15 16-20 21-25 26-30 31-99 530 south.cocoa-t 29.5 16.0 29.5 112.5 76.5 76.5 226.0 531 south.cocoa-h 441.1 92.5 85.0 95.9 40.0 40.0 63.0 532 south.oil-palm 199.8 149.1 101.1 20.4 9.0 9.0 6.5 533 south.plantain 37.7 14.0 534 GAMS 2.05 PC AT/XT 89/01/2' 1SJ5:0' PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN PRICE AND PRODUCTION DATA 536 table pest(c,*) production estimates for 1987 537 538 production harvested 539 * (lOOOtons) (1000 acres) 540 maize 500 1350 541 rice 80 175 542 millet 170 565 543 sorghum 130 650 544 cassava 2900 1010 545 yam 860 440 546 cocoyam 760 395 547 plantain 880 395 548 g-nut 119 300 549 cocoa 230 820 550 palmfruit 980 420 551 bean-pea 17 380 552 553 554 table pricedat(c,*) demand data 555 556 reference export import m-tax 557 * (c) (c) (c) (fract) 558 maize 50 25 1.0 559 rice 97 50 1.0 560 millet 59 inf 561 sorghum 56 inf 562 cassava 25 inf 563 yam 50 inf 564 cocoyam 40 inf 565 plantain 25 inf 566 g-nut 115 inf 567 cocoa 140 140 inf 568 palmfruit 80 inf 569 bean-pea 73 inf 570 571 parameter pe(c) commodity export prices 572 pm(c) commodity import prices 573 mtax(c) import tax rate 574 price(c) refernce price (cedis per kg); 575 576 pest(c,"a-yield") - pest(c,"production")/pest(c,"harvested"): 577 578 cn(c) - yes; cn("cocoa") - no; 579 ce(c) - yesSpricedat(c,"export"); 580 cm(c) - yes$(pricedat(c,"import") It inf ) 581 price(c) - pricedat(c,"reference"); 582 pe(ce) - pricedat(ce,"export"); 583 pm(cm) - pricedat(cm,"import"); 584 mtax(cm) - pricedat(cm,"m-tax"); 585 586 option eject;display "calibration section" 587 ..................... 588 display pest, cn, ce, cm, price, pe, pm, mtax; GAMS 2.05 PC AT/XT 89/01/V27 15:15:03 PACE ! AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN CALIBRATION TEST FOR 1987 590 591 table palloc(c,r) production allocation by region (prop) 592 593 north south 594 maize .2 .8 595 rice 1 596 millet 1 597 sorghum 1 598 cassava 1 599 yam .2 .8 600 cocoyam 1 601 plantain 1 602 g-nut 1 603 cocoa 1 604 palmfruit 1 605 bean-pea 1 606 607 parameters talloca(r,a,c) technology allocation for annuals (prop) 608 tnuma(r,c) number of technologies 609 prodr(c,r) production by region (1000 tons) 610 calloca(a,r) gross cropping allocation (1000 acres) 611 612 tallocp(r,ap,c) technology allocation for perennials (prop) 613 tnump(r,c) total technology (acres) 614 tlandp(r,ap) all vintages fro 1987 (acres) 615 yprop(r,ap,v) yield proportions (prop) 616 wyield(r,c,ap) weighted yield (kg per acre) 617 wvyield(c,v,ap) weighted yield by rotation length (kg per acre) 618 callocp(ap,r) gross cropping allocation (1000 acres) 61 .j labcal(r,*,*) labor use (1000 md); 621 622 prodr(c,r) - pest(c,"production")*palloc(c,r); 623 tnuma(r,c) - sum(a, l$yield(r,c,a)); 624 talloca(r,a,c)$yiell(r,c,a) - l/tnuma(r,c); 625 cAlloca(a,r) - 626 sum(c$yield(r,c,a), prodr(c,r)/yield(r,c,a)*talloca(r,a,c))*1000: 627 628 tlandp(r,ap) sum(v, stock(r,ap,v)); 629 tnump(r,c) - sum(ap$sum(v, yieldp(c,ap,v)), elandp(r,ap)); 630 yprop(r,ap,v)$tlandp(r,ap) - stock(r,ap,v)/tlandp(r,ap); 631 wyield(r,c,ap) - sum(v, yieldp(c,ap,v)*yprop(r,ap,v)); 632 wvyield(c,v,ap) - sum(vv$(ord(vv) le ord(v)), yieldp(c,ap,vv))/ord(v); 633 tallocp(r,ap,c)$wyield(r,c,ap) - tlandp(r,ap)/tnump(r,c); 634 callocp(ap,r) - 635 sum(c$wyield(r,c,ap), prodr'c,r)/wyield(r,c,ap)*tallocp(r,ap,c))*1000; 636 637 * reduce annual land allocation by intercropped output 638 639 display calloca; 640 calloca(a,rp) - calloca(a,rp) - sum((c,ap)$y.eld(rp,c,a), 641 yieldpi(c,ap)*callocp(ap,rp)/yield(rp,c,a)); 642 643 set col / annual, perennial, perenn-pl, clear ,total /; 644 ;AMS 2.05 PC AT/XT 89/01/27 1 .; 03 PACE LGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN ;ALIBRATION TEST FOR 1987 645 labcal(r,t,"annual ") - sum(a, (labreqa(r,a,t) + freq(r):'tlabcl(t)5 c^lloc lar!! 646 labcal(r,t,"perennial") - sum((ap,v), labreqp(ap,t,v)*callocp(ap,r)V:vprop(Lr, '.- ): 647 labcal(r,t,"perenn-pl") - sum((ap,vf), tlabpl(t,ap)/length*callocp(ap,r) .Vprop(r.ap.vr)) 648 labcal(r,t,"clear ") - 649 tlabcl(t)*sum(apScallocp(ap,r), callocp(ap,r)/sum(v, yprop(r,ap,v) ord(v))/1ength) 650 labcal(r,t,"total ") - sum(col, labcal(r,t,col)); 651 labcal(r,"total",col) - sum(t, labcal(r,t,col)); 652 labcal(r,"max",col ) - smax(t, labcal(r,t,col)); 653 654 display prodr, tnuma, talloca, yield, calloca, tnump, tlandp, yprop 655 wyield, wvyield, tallocp, callocp,labcal; 656 657 658 table demdat(c,*) demand data 659 660 pelas yelas 661 * 662 maize -.9 .4 663 rice -.8 .7 664 millet -.3 .2 665 sorghum - .3 .2 666 cassava - .8 .1 667 yam -.9 .5 668 cocoyam - .9 .3 669 plantain - .7 .3 670 g-nut - 9 .6 671 cocoa inf 672 palmfruit - .8 .5 673 bean-pea - .5 .4 674 675 676 677 parameters alpha(c) demand curve intercept 678 beta(c) demand curve gradient 679 gamma(c) demand curve income shift; 680 681 * p - alpha + beta*q - gamma*y 682 * beta - pO/qO/pelas 683 * gamma - yelas*pO/pelas 684 * alpha - pO - beta*qO + gamma 685 686 demdat(c,"ref-p") - pricedat(c,"reference"); 687 * below is production net of seeds 688 demdat(c,"production") - pest (c,"production"); 689 demdat(c,"ref-q") - pest (c,"production") 690 - sum((a,r)$ra(r,a), calloca(a,r)*seed(c,a))/1000; 691 692 beta(cn) - demdat(cn,"ref-p")/demdat(cn,"ref-qm)/demdat(cn,"pelas"); 693 gama(cn) - demdat(cn,"yelas")*demdat(cn,"ref-p')/demdat(cn,"pelas"); 694 alpha(cn) - d.mdat(cn,"ref-p") - beta(cn)*demdat(cn,"ref-q") + gamma(cn); 695 696 demdat(cn,"dem-a") - alpha(cn); demdat(cn,"dem-b") - beta(cn); 697 demdat(cn,"dem-c") - gamma(cn); CAMS 2.05 PC AT/XT 89/01/27 15:15:03 PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN CALIBRATION TEST FOR 1987 698 699 display demdat; 700 701 702 parameters 703 frat(r) ratio of family labor (fract) / south .75 704 north .90 / 705 wagefam(r) reservation wage rate (1000 ) / south .10 706 north .08 / 707 wagecas(r,rr) casual wage rate (1000) / 708 south.south .200, south.north .300 709 north.north .150, north.south .250 / 710 CAMS 2.05 PC AT/XT 89/01, " 1/ 5: I PACE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN DYNAMIC DATA 712 table indexdata(th,*) index numbers 713 714 pop inc lab cons w-prod 715 716 1985-89 1.000 1.000 1.000 1.000 1.000 717 1990-94 1.176 1.091 1.171 1.060 1.0'0 718 1995-99 1.374 1.195 1.292 1.162 1.180 719 2000-04 1.593 1.322 1.475 1.288 1.274 720 2005-09 1.827 1.460 1.683 1.404 1.358 721 2010-14 2.064 1.612 1.920 1.529 1.435 722 2015-19 2.301 1.784 2.187 1.669 1.514 723 2020-24 2.538 1.964 2.454 1.809 1.595 724 2025-29 2.765 2.142 2.698 1.974 1.677 725 2030-34 2.971 2.324 2.853 1.141 1.762 726 2035-39 3.180 2.454 2.981 2.263 1.852 727 728 * note: pop population index 729 * inc income index 730 * lab rural labor force index 731 * -ons consumption index 732 * w-prod worker productivity index 733 734 parameters 735 736 wforce(r) rural workforce in 1987 (mill workers) / north 1.31, south 3.42 / 737 agwforce(r) agric workforce in 1987 (mill workers) / north .95, south 2.48 / 738 yindx(th) income index (1985-89 - 1) 739 cindx(th) consumption index (1985-89 - 1) 740 pindx(th) productivity index (1985-89 - 1) 741 popindx(th) population index (1985-89 . 1) 742 labindx(th) labor force index (1985-89 - 1) 743 labsupt(th,t,r) labor availability (1000 mandays) 744 yfactor fraction of income inrease 745 wfactor fraction of wage increase 746 pfactor fraction of population increase 747 lfactor fraction of labor increase 748 tland(r) total land from calibration (1000 acres) 749 gf(th,r) addition to fallow land (1000 acres) 750 delta(th) discount factor; 751 752 yindx(th) - indexdata(th."inc'); 753 cindx(th) - indexdata(th,"cons"); 754 pindx(th) - indexdata(th,"w-prod"); 755 popindx(th) - indexdata(th,"pop"); 756 labindx(th) - indexdata(th,"lab"); 757 yfactor - 0; 758 wfactor - 0; 759 pfactor - 0; 760 lfactor - 0; 761 762 labsupt(th,t,r) - agwforc.(r)*25*(1+1factor*(labindx(th)-l))*l000; 763 764 tland(r)- sum(ap, callocp(ap,r))$rp(r) + sum(a, calloca(a,r))* 765 (l+freq(r)); 766 * land is growing proprtional to rural labor force GAMS 2.05 PC AT/XT 89/01,, 27Li.. );K . AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN DYNAMIC DATA 767 gf(th+l,r) -1.OO*eland(r)*lfaceor*(labindx(th+l) labindx(ch)): 768 delta(th) - (1/1.03)**(ord(th)*lengch); 769 770 display labsupt, tland, gf, delta; 771 ;AMS 2.05 PC AT/XT 89/01/27 15:l5:;3 PAGE 22 kGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN IODEL DEFINITION possible land uses xp cropped with perennials f fallow xa cropped with annuals possible land transactions ( at the beginning of the period ap abandoning perennials xp > f aa abandoning annual land xa > f cl clearing f > xa pl planting perennials xa > xp fg addition > f xp(t+l) - xp(t) - ap(t+l) + pl(t+l); f(t+l) - f(t) + ap(t+l) - cl(t+l) + fg(t+l) x aa(t+l) xa(t+l) - xa(t) + cl(t*l) - pl(t+l) - aa(t+l) note that pl(0-5) - xp(O-5) and pl has been substitued out. 796 797 vaUiables 798 xa(th,r,a) land under annual crops (1000 acres) 799 txa(th,r) total annual land (1000 acres) 800 xp(th,r,P;,v) land under perennial crops (1000 acres) 801 f(th,r) land fallow (1000 acres) 802 xpf(th,r,ap,v) abandoning perennial crop land (1000 acres) 803 xaf(th,r) abandoning annual crop land (1000 acres) 804 cl(th,r) clear fallow land (1000 acres) 805 labfam(th,t,r) family labor (1000 md) 806 labcas(th,t,r,rr) casual labor (1000 md) 807 output(th,r,c) production of crop commodity (1000 tons) 808 natcon(th,c) domestic consumption (1000 tons) 809 exports(th,c) national exports (1000 tons) 810 imports(th,c) national imports (1000 tons) 811 revenue(th) regional revenue (mill cedis) 812 profit profit from crop process (mill cedis 813 cps consumers and producers surplu.; (mill cedis) 814 inpcost(th,r) cost of purchased variable inputs (mill cedis) 815 labcost(th,r) labor cost (mill cedis) 816 tcost(th) total input cost 817 818 positive variable xa, vo, labfam, f, cl, txa, xpf, xaf, labfam, 819 labcas, nutput, imports, exports; 820 821 eqiations 822 xpbal(th,rp,ap,v) perennial land balance (1000 acres) 823 xabal(th,r) annual land balance (1000 acres) 824 txadef(th,r) total annual land dof (1000 acres) 825 fbal(th,r) fallow land balance (1000 acres) 826 fmin(th,r) minimum fellow requirement 827 laborbal(th,t,r) labor balance (mill md) 828 laborlim(th,t,r) labor limits (mill md) 829 dprod(th,r,c) production accounting (1000 tons) GAMS 2.05 PC AT/XT 89/01,27 15: 15:;' PAC:E AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN MODEL DEFINITION 830 dem(th,c) demand balance (1000 tons) 831 ainp(th,r) input cost accounting (mill cedis 832 alab(th,r) labor cost accounting 833 atcost(th) total cost 834 objn nonlinear objective; 835 836 xpbal(th,rp,ap,v-l)$tp(th).. 837 838 xp(th,rp,ap,v) -e- sum(vv$vtov(vv,v), xp(th-l,rp,ap,vv)) 839 840 - xpf(th,rp,ap,v); 841 842 fbal(th-l,r)$tp(th).. 843 844 f(th,r) -e- .(th-l,r) + xaf(th,r) 845 846 + sum((ap,v)$(not vf(v)), xpf(th,r,ap,v))Srp(r) 847 848 + gf(th,r) - cl(th,r); 849 850 fmin(tp,r).. f(tp,r) -g- freq(r)*txa(tp,r); 851 852 853 exadef(tp,r).. txa(tp,r) -e- sumCa$ra(r,a), xa(tp,r,a)); 854 855 856 xabal(th-l,r)$tp(th).. 857 858 txa(th,r) -e- txa(th-l,r) - xaf(th,r) + cl(th,r) 859 860 - sum((ap,vf), xp(th,r,ap,vf))$rp(r); 861 862 863 laborbal(tp,t,r).. 864 865 sum(a$ra(r,a), (labreqa(r,a,t) + freq(r)*tlabcl(t))*xa(tp,r,a)) 866 867 + sum((ap,v), labreqp(ap,t,v)*xp(tp,r,ap,v) 868 869 + tlabpl(t,ap)/length*xp(tp,r,ap,v)$vf(v) )$rp(r) 870 871 + sum(rr, labcas(tp,t,r,rr)) 872 873 + tlabcl(t)/length*cl(tp,.) 874 875 -e- labfam(tp,t,r) + labcas(tp,t,r,r) + sum(rr, labcas(tp,t,rr,r)); 876 877 878 laborlim(tp,t,r).. 879 880 labfam(tp,t,r) + sum(rr, labcas(tp,t,r,rr)) -1- labsupt(tn,t,r); 881 882 dprod(tp,r,c)$rc(r,c).. 883 884 output(tp,r,c) -e- sum(a$ra(r,a), nyielda(r,c,a)*xa(tp,r,a))/l000 ;AMS 2.05 PC AT/XT 89/01/2- 15:15: 03 PAGE 24 RICULTURAL SECTOR MODEL FOR GHANA- BASE RUN DEL DEFINITION 85 86 + sum((ap,v), xp(tp,r,ap,v)*(yieldp(c,ap,v) 87 + yieldpi(c,ap)$vf(v)))/lOOO; 88 89 890 891 dem(tp,c). 892 893 natcon(tp,c)$cn(c) -e- sum(r$rc(r,c), output(tp,r,c)) 894 895 + imports(tp,c)Scm(c) - exports(tp,c)$ce(c); 896 897 898 899 ainp(tp,r).. 900 901 inpcost(tp,r) -e- sum(a$ra(r,a), ( icosta(r,a) 902 + freq(r)*incostcl)*xa(tp,r,a)) 903 904 + sum((ap,v), icostp(ap,v)*xp(tp,r,ap,v) 905 906 + incostpl(ap)/length*xp(tp,r,ap,v)$vf(v) )$rp(r) 907 908 + incostc1/length*cj(tp,r) 909 910 911 alab(tp,r). 912 913 labcost(tp,r) -e- sum(t, wagefam(r)*labfam(tp,t,r) 914 915 + sum(rr, wagecas(rr,r)*labcas(tp,t,rr,r)) 916 *(l+wfactor*(pindx(tp)-1)); 917 918 atcost(tp). 9 0 tcost(tp) -e- sum(r, inpcost(tp,r) + labcost(tp,r)) 1 2 3 objn.. 4 cps -e- sum(tp, delta(tp)*( sum(cn, ( alpha(cn) 5 gamma(cn)*(1+yfactor*(cindx(tp)-l))) stlanzcon( t:p, cn; 6 + 5*beta(cn)/(l+pfactor*(popindx(-.p).l)) *sqr(natcoii(cp.cn))) 7 + sum(ce, exports(tp,ce)*pe(ce)) 8 sum(cm, imports(tp,cm)*pm(cm)*(l+mtax(cm))) 929 tcost(tp) )); 930 931 ?32 model ghanan / xpbal, xabal, txadef, fbal, fmin, laborbal, laborlim i33 dprod, dem, ainp, alab, atcost, objn /; 934 3-5 * set initial conditions ?36 labfam.up(th,t,r) - frat(r)*labsupt(th,t,r); 337 xp.fx(ti,rp,ap,v) - callocp(ap,rp)*yprop(rp,ap,v); . GAMS 2.05 PC AT/XT 89/01/27 15:15:03 PAC-E AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN MODEL DEFINITION 938 txa.fx(ti,r) - ?um(a, calloca(a,r)); 939 f.fx(ti,r) - freq(r)*txa.l(ti,r); 940 941 * set starting values and bound to avoid numeric problems 942 rntcon.l (th,c) - demdat(c,"ref-q"); 943 natcon.lo(th,c) - demdat(c,"ref-q")*.25; 944 945 OPTION LIMCOL-O,LIMROW-O,ITERLIM-lOOOO,reslim-500; 946 solve ghanan using nlp maximizing cps GAMS 2.05 PC AT/XT 89/01/127 15:15:03 PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN REPORT DEFINITION 948 set coll; coll(v) - yes; coll("total") - yes; 949 set col3 950 SET COL2 / ANNUAL, PERENNUAL, PERENN-PL, CLEAR ,TOTAL / 951 col4 / annual, perennual,fallow,tocal / 952 col5 / family, casual, casual-imp, slack / 953 col6 / revenue, input-c, labor-c, farm-inc / 954 rall / north, south, ghana /; 955 956 coll(v) - yes; coll("total") - yes; 957 col3(a) - yes; col3("total") - yes; 958 col5(col2) - yes; 959 960 parameter areap(*,th,*) perennial crop area (i000 acres) 961 areaa(*,th,*) annual crop area (1000 acres) 962 areas(*,th,*) land use summary (1000 acres) 963 labu(*,*,*,*) labor use and sourcses (1000 md) 964 gprod gross production (1000 tons) 965 prep price summary (cedis per kg) 966 ISUM INCOME SUMMARY (BILL CEDIS) 967 PIREP PRICE SUMMARY 968 GROSSCREV(TH) GROSS COCOA REVENUE (MILL CEDIS); 969 970 areap(ap,th,v) - sum(rp, xp.l(th,rp,ap,v)); 971 areap(ap,th,"total") - sum(v, areap(ap,th,v)); 972 areap("total",th,coll) - sum(ap, areap(ap,th,coll)); 973 974 areaa(r,ti,a) - calloca(a,r); 975 areaa(r,tp,a) - xa.l(tp,r,a); 976 areaa(r,th,"total") - sum(a, areaa(r,th,a)); 977 areaa("ghana",th,col3) - sum(r, areaa(r,th,col3)); 978 979 areas(r,th,"annual") - areaa(r,th,"total"); 980 areas(rp,th,"perennual") - areap("total",th,"total"); 981 areas(r,th,"fallow") - f.l(th,r); 982 areas(r,th,"total") - sum(col4, areas(r,th,col4)); 983 areas("ghana",th,col4) - sum(r, areas(r,th,col4)) 984 985 labu(tp,r,t,"clear") - tlabcl(t)/length*cl.1(tp,r); 986 labu(tp,r,t,"annual") - sum(a$ra(r,a), 987 (labreqa(r,a,t) + freq(r)*tlabcl(t))*xa.l(tp,rL-i)); 988 labu(tp,rp,t,"perenn-pl") - sum((ap,vf), tlabpl(t.ap)/length*xp.l(tp.rp.ap.vf)): 989 labu(tp,rp,t,"perennual") - sum((ap,v), labreqp(ap,t,v)*xp.l(tp,rp,ap.v)); 990 labu(tp,r,t,"total") - sum(col2, labu(tp,r,t,col2)); 991 labu(tp,r,t,"family") - labfam.l(ep,t,r); 992 labu(tp,r,t,"casual") - labcas.l(tp,t,r,r); 993 labu(tp,r,t,"casual-imp") - sum(rr$(ord(r) ne ord(rr)), 994 lab,as.l(tp,t,rr,r)); 995 labu(tp,r,t,"slack") - round(laborlin.up(tp,t,r) - laborlim.1(tp,t,r)); 996 labu(tp,"ghana",t,col5) - sum(r, labu(tp,r,t,col5)); 997 labu(tp,rall,"max",col5) - smax(t,labu(tp,rall,t,col5)); 998 labu(tp,rall,"total",col5) - sum(t,labu(tp,rall,t,col5)); 999 1000 gprod(r,ti,c) - prodr(c,r); 1001 gprod(r,tp,c) - output.l(tp,r,c) 1002 + sum(a, xa.l(tp,r,a)*seed(c,a))/1000; . GAMS 2.05 PC AT/XT 89/01/27, 15:1;:03 PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN REPORT DEFINITION 1003 gprod("ghana",th,c) - sum(r, oueput.1(th,r,c)); 1004 1005 prep(ti,c) - price(c); 1006 prep(tp,c) - dem.m(tp,c)/delta(tp); 1007 1008 1009 isum(r,tp,"revenue") - (sum(cn, prep(tp,cn)*output.l(ep,r,cn)) 1010 + sum(ce, pe(ce)*exports.l(tp,ce))Srp(r))/1000; 1011 isum(r,tp,"input-c") - inpcost.l(tp,r)/1000; 1012 isum(r,tp,"labor-c") - labcost.l(tp,r)/1000; 1013 isum(r,tp,"farm-inc") - sum.(col6, isum(r,tp,col6)); 1014 isum("ghana",ep,col6) - sum(r, isum(r,tp,col6)); 1015 1016 1017 PIREP(RP,TP,'COCOA-H') - XPBAL.M(TP,RP,'COCOA-H','0-5')/DELTA(TP): 1018 PIREP(R,TP.'ANNUAL ') - XABAL.M(TP,R)/DELTA(TP); 1019 PIREP(R,TP,'FALLOW ) - FBAL.M(TP,R)/DELTA(TP); 1020 1021 GROSSCREV(TP) - (.85*300 - PE("COCOA"))*EXPORTS.L(TP,nCOCOA"); 1022 1023 OPTION AREAA:1, AREAA:1, AREAS;1, LABU:1, PREP:1,GPROD:1,PIREP:l; 1024 display areap,areaa, areas,labu, gprod, prep, exports.l,imports.l,isum 1025 PIREP, GROSSCREV; . AMS 2.05 PC AT/XT 89/01/27 1, 1. 15 e.8. iGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING ,BOL TYPE REFERENCES SET DECLARED 51 DEFINED 51 REF 56 . - 319 376 451 452 453 45-4 -. 458 459 460 461 3*462 4 63 83 517 2*519 521 523 607 6iO 3*626 2*640 641 2*645 3*690 '64 98. 3*865 3*884 2*901 902 938 974 975 ; 986 2*987 2*1002 CONTROL 457 458 461 462 519 521 523 623 624 625, 640 645 690 764 853 865 S84 938 957 974 975 976 986 1o0)0 kGWFORCE PARAM DECLARED 737 DEFINED 737 REF 762 kINP EQU DECLARED 831 DEFINED 901 IMPL-ASN 946 REF -3" ULAB EQU DECLARED 832 DEFINED 913 IMPL-ASN 946 REF kLPHA PARAM DECLARED 677 ASSIGNED 694 REF 696 92. SET DECLARED 54 DEFINED 54 REF 116 ' .: 161 162 165 168 4*170 172 17 7 '1 237 258 259 260 261 263 26-. 3*266 267 466 475 527 612 614 615 616 617 618 628 2*629 3*630 2;'631 632 2*633 3*635 2*641 3*646 3*647 3-649 764 ' 802 822 2*838 840 846 860 2 867 . 2*886 887 2*904 2*906 2*937 970 971 c. 2*988 2*989 CONTROL 168 170 172 263 265 266 628 629 630 631 (32 634 640 646 647 649 764 s36 860 867 886 904 937 970 '971 988 989 kREAA PARAM DECLARED 961 ASSIGNED 974 975 976 REF 976 977 979 2*1023 1024 *REAP PARAM DECLARED 960 ASSIGNED 970 971 972 RPEF 972 980 1024 REAS PARAM DECLARED 962 ASSIGNED 979 980 981 .d82 REF 982 983 1023 1024 kTCOST EQU DECLARED 833 DEFINED 920 IMPL-ASN 946 REF BETA PARAM DECLARED 678 ASSIGNED 692 REF 694 SET DECLARED 37 DEFINED 37 REF 40 - 45 46 49 93 466 475 .S 517 536 554 571 572 573 579 580 581 591 607 608 609 613 616 617 2*622 623 2 '~624 -.62, 631 632 2*633 4*635 640 2>4 , I-.-1 678 679 686 688 689 690 8- 809 810 829 830 882 2*' 884 886 4 4*893 4*895 942 943 1000 1.001 10:02 1005 1006 CONTROL 576 578 579 580 81' 622 623 624 626 629 631 632 - 635 640 686 688 689 882 891 942 943 1000 1001 1003 1005 1006 SET DECLARED 40 DEFINED 40 REF 2*519 52i 523 CONTROL 519 521 523 ,ALLOCA PARAM DECLARED 610 ASSIGNED 626 640 REF 639 645 654 690 764 938 974 "ALLOCP PARAM DECLARED 618 ASSIGNED 635 REF 641 ')96 GAMS 2.05 PC AT/XT 89,/O012'27 5::' PACE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING SYMBOL TYPE REFERENCES 2*649 655 764 937 CC SET DECLARED 93 CE SET DECLARED 46 ASSIGNED 579 REF 582 588 895 2*927 2*1010 CONTROL 582 927 1010 CINDX PARAM DECLARED 739 ASSIGNED 753 REF 925 CL VAR DECLARED 804 IMPL-ASN 946 REF 818 848 858 873 908 985 CM SET DECLARED 45 ASSIGNED 580 REF 583 584 538 895 3*928 CONTROL 583 584 928 CN SET DECLARED 44 ASSIGNED 2*578 REF 93 588 >692 3*693 4*694 2*696 697 893 924 2 925 25926 2*1009 CONTROL 692 693 694 2*696 697 92' 1009 CNN SET DECLARED 93 COL SET DECLARED 643 DEFINED 643 REF 650 651 652 CONTROL 650 651 652 COLl SET DECLARED 948 ASSIGNED 2*948 2*956 REF 972 CONTROL 972 COL2 SET DECLARED 950 DEFINED 950 REF 990 CONTROL 9 990 COL3 SET DECLARED 949 ASSIGNED 2*957 REF 977 CONTROL 9 COL4 SET DECLARED 951 DEFINED 951 REF 982 983 CONTROL 982 983 COL5 SET DECLARED 952 DEFINED 952 ASSIGNED 958 REF 9 997 998 CONTROL 996 997 998 COL6 SET DECLARED 953 DEFINED 953 REF 1013 1014 CONTROL 1013 1014 CP SET DECLARED 42 DEFINED 42 CONTROL 522 CPS VAR DECLARED 813 IMPL-ASN 946 REF 924 946 DELTA PARAM DECLARED 750 ASSIGNED 768 REF 770 924 1017 1018 1019 DEM EQU DECLARED 830 DEFINED 893 IMPL-ASN 946 REF 9 1006 DEMDAT PARAM DECLARED 658 DEFINED 658 ASSIGNED 686 688 6 2*696 697 REF 3*692 3*693 2*694 G9 9 943 DPROD EQU DECLARED 829 DEFINED 884 IMPL-ASN 946 REF 9 3XPORTS VAR DECLARED 809 IMPL-ASN 946 REF 819 895 9 1010 1021 1024 F VAR DECLARED 801 IMPL-ASN 946 ASSIGNED 939 REF 8 2*844 850 981 FBAL EQU DECLARED 825 DEFINED 844 IMPL-ASN 946 REF 9 1019 FMIN EQU DECLARED 826 DEFINED 850 IMPL-ASN 946 REF 9 FRAT PARAM DECLARED 703 DEFINED 703 REF 936 FREQ PARAM DECLARED 122 DEFINED 124 REF 645 765 8 865 902 939 987 GAMMA PARAM DECLARED 679 ASSIGNED 693 REF 694 697 9 GF PARAM DECLARED 749 ASSIGNED 767 REF 770 848 GHANAN MODEL DECLARED 932 DEFINED 932 REF 946 GPROD PARAM DECLARED 964 ASSIGNED 1000 1001 1003 REF 10 1024 GROSSCREV PUAAM DECLARED 968 ASSIGNED 1021 REF 1025 GAMS 2.05 PC AT/XT 89/01,'27 15:15:03 PAGE 3 AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING SYMBOL TYPE REFERENCES ICOSTA PARAM DECLARED 451 ASSIGNED 457 REF 464 901 ICOSTP PARAM DECLARED 258 ASSIGNED 263 REF 268 904 IMPORTS VAR DECLARED 810 IMPL-ASN 946 REF 819 895 92 1024 IN SET DECLARED 74 DEFINED 74 REF 100 108 11 2*172 2*173 177 2*263 270 2*457 CONTROL 17 173 263 457 INCOSTCL PARAM DECLARED 166 ASSIGNED 173 REF 175 90? 9v INCOSTPL PARAM DECLARED 165 ASSIGNED 172 REF 175 906 INDEXDATA PARAM DECLARED 712 DEFINED 712 REF 752 753 75 755 756 INPCOST VAR DECLARED 814 IMPL-ASN 946 REF 901 920 101 IREQA PARAM DECLARED 270 DEFINED 270 REF 457 IREQP PARAM DECLARED 177 DEFINED 177 REF 263 ISUM PARAM DECLARED 966 ASSIGNED 1009 1011 1012 1013 101 REF 1013 1014 1024 K SET DECLARED 88 DEFINED 88 REF 112 126 13 143 161 163 168 169 4*170 4*171 216 237 260 264 265 3*266 267 287 319 376 453 455 458 459 460 461 3*462 463 CONTROL 168 169 170 171 264 26 266 458 459 460 461 462 LABCAL PARAM DECLARED 620 ASSIGNED 645 646 647 649 65 b1 6-br2 REF 650 651 652 655 LABCAS VAR DECLARED 806 IMPL-ASN 946 REF 819 871 2*87 880 915 992 994 LABCOST VAR DECLARED 815 IMPL-ASN 946 REF 913 920 101 LABFAM VAR DECLARED 80Q IMPL-ASN 946 ASSIGNED 936 REF 2'-31 875 880 913 991 LABINDX PARAM DECLARED -42 ASSIGNED 756 REF 762 2' -6- LABORBAL EQU DECLARED 827 DEFINED 865 IMPL-ASN 946 REF 93 LABORLIM EQU DECLARED 828 DEFINED 880 IMPL-ASN 946 REF 93 2*995 LABPA PARAM DECLARED 455 ASSIGNED 459 460 REF '61 46 LABPAN PARAM DECLARED 376 DEFINED 376 REF 460 LABPAS PARAM DECLARED 319 DEFINED 319 REF 459 LABPP PARAM DECLARED 237 DEFINED 237 REF 265 266 LABPRCL PARAM DECLARED 136 DEFINED 136 REF 169 171 LABPRPL PARAM DECLARED 143 DEFINED 143 REF 168 1VD LABREQA PARAM DECLARED 454 ASSIGNED 462 REF 464 685 36 987 LABREQP PARAM DECLARED 261 ASSIGNED 266 REF 268 ',46 86 989 LABSUPT PARAM DECLARED 743 ASSIGNED 762 REF 770 880 93 LABU PARAM DECLARED 963 ASSIGNED 985 986 988 989 99 991 992 993 995 996 997 998 REF 990 996 997 998 1023 1024 LANDCLIN PARAM DECLARED 108 DEFI.4ED 110 REF 173 LANDCLLAB PARAM DECLARED 112 DEFINED 114 REF 171 LENGTH PARAM DECLARED 95 DEFINED 95 REF 647 649 76 869 873 906 908 985 988 LFACTOR PARAM DECLARED 747 ASSIGNED 760 REF 762 '67 MAX FUNCT REF 519 AMS 2.05 PC AT/XT 89/01P27 15:15:03 PACE ': AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING SYMBOL TYPE REFERENCES MTAX PARAM DECLARED 573 ASSIGNED 584 REF 588 *-8 NATCON VAR DECLARED 808 IMPL-ASN 946 ASSIGNED 9-.2 REF 893 925 926 NYIELDA PARAM DECLARED 517 ASSIGNED 519 REF 521 ,2" 2;'. 884 OBJN EQU DECLARED 834 DEFINED 924 IMPL-ASN 946 REF OUTPUT VAR DECLARED 807 IMPL-ASN 946 REF ' 819 88. 8i93 1001 1003 1009 PALLOC PARAM DECLARED 591 DEFINED 591 REF 622 PE PARAM DECLARED 571 ASSIGNED 582 REF 588 927 'v! 1021 PEST PARAM DECLARED 536 DEFINED 536 ASSIGNED 576 REF F- 588 622 688 689 PFACTOR PARAM DECLARED 746 ASSIGNED 759 REF 926 PINDX PARAM DECLARED 740 ASSIGNED 754 REF 916 PIREP PARAM DECLARED 967 ASSIGNED 1017 1018 1019 REF >22 1025 PLANTPIN PARAM DECLARED 116 DEFINED 118 REF 172 PLANTPLB PARAM DECLARED 126 DEFINED 126 REF 170 PM PARAM DECLARED 572 ASSIGNED 583 REF 588 928 POPINDX PARAM DECLARED 741 ASSIGNED 755 REF 926 PREP PARAM DECLARED 965 ASSIGNED 1005 1006 REF 1009 1023 1024 PRICE PARAM DECLARED 574 ASSIGNED 581 REF 588 10O5 PRICEDAT PARAM DECLARED 554 DEFINED 554 REF 579 580 8S, 582 583 584 686 PRICEIN PARAM DECLARED 100 DEFINED 102 REF 172 173 26,3 457 PRODR PARAM DECLARED 609 ASSIGNED 622 REF 626 625 1000 PROFIT VAR DECLARED 812 PTOT PARAM DECLARED 453 ASSIGNED 461 REF 462 463 PTOTP PARAMY DECLARED 260 ASSIGNED 265 REF 266 26' '68 SET DECLARED 34 DEFINED 34 REF 35 49 56 93 122 270 287 451 452 _,'3 455 457 458 461 3*462 463 -.83 519 521 523 527 591 607 608 610 612 613 614 615 616 618 620 622 623 2*624 4*626 628 629 3<:630 3*633 4*635 3*645 2*646 2*647 3>649 65' 652 2*690 703 705 707 736 '37 748 749 762 3*764 765 767 '98 800 801 802 803 804 805 806 814 815 823 824 825 826 827 828 829 831 832 3*844 2*846 27848 '-850 3-853 4*858 2*860 4*865 867 2*869 871 873 4-' 875 3*880 882 4*884 886 2*893 3*901 27902 904 2*906 908 3*91, 2*915 2*920 2*936 938 2*939 974 975 976 977 979 981 982 983 985 986 3*987 990 991 2*992 993 994 2*995 996 1000 1001 1002 1003 1009 1010 1011 1012 1013 1014 1018 1019 CONTROL ,57 458 461 462 519 521 523 622 623 GAMS 2.05 PC AT/XT 89,0 'I 2 15:;, ¼ PACE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING SYMBOL TYPE REFERENCES 624 625 628 629 630 631 '33 645 646 647 648 650 651 652, 762 764 767 842 850 853 356 863 878 882 893 899 911 920 q36 ?38 939 974 975 976 977 979 S 81 983 985 986 990 99! 992 993 s95 996 1000 1001 1003 1009 1011 10O12 1,013 1014 1018 1019 RA SET DECLARED 56 ASSIGNED 523 REF 525 690 853 865 884 901 986 RALL SET DECLARED 954 DEFINED 954 REF 997 998 CONTROL 997 998 RC SET DECLARED 49 ASSIGNED 521 522 REF 525 382 893 REVENUE VAR DECLARED 811. ROUND FUNCT REF 995 RP SET DECLARED 35 DEFINED 35 REF 2*640 2 .6'41 7,4 822 2*83S 840 846 860 869 906 2-937 970 988 989 1010 1017 CONTROL 522 640 836 937 970 980 988 989 1017 RR SET DECLARED 93 REF 707 806 871 875 880 2*915 993 994 CONTROL 871 875 880 915 993 SEED PARAM DECLARED 502 DEFINED 502 REF 519 690 1002 SQR FUNCT REF 926 STOCK PARAM DECLARED 527 DEFINED 527 REF 628 630 T SET DECLARED 58 DEFINED 58 REF 136 143 162 164 168 169 170 171 237 261 2,65 266 319 376 454 455 459 L60 461 462 2*645 646 647 649 650 651o 652 743 805 806 827 828 2:-865 J67 i69 871 873 3*875 3*880 913 915 '26 'd5 2*987 988 989 990 991 992 996 997 998 CONTROL 168 169 o 265 266 459 460 461 462 6-5 .,5 647 648 650 651 652 762 86 373 913 936 985 986 988 98? 990 "91 992 993 995 996 997 3;98 TALLOCA PARAM DECLARED 607 ASSIGNED 624 REF 626 654 TALLOCP PARAM DECLARED 612 ASSIGNED 633 REF 635 655 TCOST VAR DECLARED 816 IMPL-ASN 946 REF 920 929 TH SET DECLARED 63 DEFINED 63 REF 66 6 12 738 739 740 741 742 743 749 750 752 753 754 755 756 762 2"6 17 -63 798 799 800 801 802 803 804 805 806 807 808 809 810 811 814 815 816 822 823 824 825 826 827 828 829 830 831 832 833 836 2-838 840 842 3*844 846 2*848 856 4*858 860 936 960 961 962 968 970 971 972 976 977 979 980 981- 982 983 1003 CONTROL 752 753 754 755 756 762 767 768 836 842 856 936 942 943 970 GAMS 2.05 PC AT/XT 89,'01 ' 1 '5 )' PACE 33 RICULTUPAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING SYMBOL TYPF REFERENCES 971 972 976 977 979 980 981 982 983 1003 SET DECLARED 66 DEFINED 66 REF 939 C'O':TROL 937 938 939 974 1000 1005 TLABA PARAM DECLARED 452 ASSIGNED 458 REF 464 TLABCL PARAM DECLARED 164 ASSIGNED 171 REF 175 645 649 865 873 985 987 TLABP PARAM DECLARED 259 ASSIGNED 264 REF 268 TLABPCL PARAM DECLARED 163 ASSIGNED 169 REF 2''171 175 TLABPL PARAM DECLARED 162 ASSIGNED 170 REF 175 647 869 988 TLABPPL PARAM DECLARED 161 ASSIGNED 168 REF 29 170 175 TIAND. PARAM DECLARED 748 ASSIGNED 764 REF 767 '70 TLANDP PARAM DECLARED 614 ASSIGNED 628 REF 629 2''630 633 654 TNUMA PARAM DECLARED 608 ASSIGNED 623 REF 624 654 TNUMP PARAM DECLARED 613 ASSIGNED 629 REF 633 654 SET DECLARED 67 DEFINED 67 REF 836 842 2'850 2*853 856 865 867 869 871 873 3*875 3*880 2*884 886 2*893 2*895 901 902 904 906 908 2*913 915 916 3'920 924 2'925 2*926 927 928 929 975 985 987 988 989 990 991 992 994 2>,995 996 997 998 1001 1002 2*1006 2*1009 1010 1011 1012 1013 1014 2*1017 2*1018 2*1019 1021 CONTROL 850 853 863 878 882 891 899 911 9i8 924 975 985 986 988 989 990 991 992 993 995 996 997 998 1C)01 1006 1009 1011 1012 1013 1014 1017 1018 1019 1021 REQA PARAM DECLARED 287 DEFINED 287 REF 458 462 REQP PARAM DECLARED 216 DEFINED 216 REF 264 266 TXA VAR DECLARED 799 IMPL-ASN 946 ASSIGNED 938 REF 818 850 853 2*858 939 TXADEF EQU DECLARED 824 DEFINED 853 IMPL-ASN 946 REF 932 SET DECLARED 60 DEFINED 60 REF 61 2''71 93 177 216 258 259 261 263 264 266 466 527 615 617 628 629 630 6 631 2*632 2*646 2*649 800 802 822 2;'-838 840 2*846 2*867 *2*869 2*886 887 2'904 2-906 937 970 971 2*989 CONTROL 263 264 266 628 629 630 631 632 646 649 836 846 867 886 904 937 948 956 970 971 989 VF SET DECLARED 61 DEFINED 61 REF 647 846 860 869 887 906 988 CONTROL 647 860 988 VTOV SET DECLARED 71 DEFINED 71 REF 838 vv SET DECLARED 93 REF 2*632 2*838 CONTRO' 632 838 WAGECAS PARAM DECLARED 707 DEFINED 707 REF 915 WAGEFAM PARAM DECLARED 705 DEFINED 705 REF 913 WFACTQR PARAM DECLARED 745 ASSIGNED 758 REF 916 WFORCE PARAM DECLARED 736 DEFINED 736 WVYIELD PARAM DECLARED 617 ASSIGNED 632 REF 655 ;AMS 2.05 PC AT/XT 89/01/27 15:15:03 PAGE AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING SYMBOL TYPE REFERENCES IELD PARAM DECLARED 616 ASSIGNED 631 REF 633 2';635 35 VAR DECLARED 798 IMPL-ASN 946 REF 818 853 865 884 902 975 987 1002 BAL EQU DECLARED 823 DEFINED 858 IMPL-ASN 946 REF 932 1018 XAF VAR DECLARED 803 IMPL-ASN 946 REF 818 844 858 VAR DECLARED 800 IMPL-ASN 946 ASSIGNED 937 REF 818 2*838 860 867 869 886 904 906 970 988 989 BAL EQU DECLARED 822 DEFINED 838 IMPL-ASN 946 REF 932 1017 XPF VAR DECLARED 802 IMPL-ASN 946 REF 818 840 846 YFACTOR PARAM DECLARED 744 ASSIGNED 757 REF 925 ELD PARAM DECLARED 483 DEFINED 483 REF 519 623 624 2*626 640 641 654 YIELDP PARAM DECLARED 466 DEFINED 466 REF 629 631 632 886 YIELDPI PARAM DECLARED 475 DEFINED 475 REF 641 887 YINDX PARAM DECLARED 738 ASSIGNED 752 YPROP PARAM DECLARED 615 ASSIGNED 630 REF 631 646 ' 647 649 654 937 SETS ANNUAL CROP PROCESSES AP PERENNIAL PRODUCTION CROPS CA ANNUAL CROPS CC ALIASED WITH C CE EXPORTED CROPS CM IMPORTED CROPS CN NATIONALLY CONSUMED CROPS CNN ALIASED WITH CN COL COLI COL2 C0L3 COL4 COL5 COL6 CP PERENNIAL CROPS IN INPUTS TASK AGRICULTURAL REGIONS RA PROCESS POSSIBILITIES RALL RC CROP POSSIBILITIES RP REGIONS WITH PERENNIALS RR ALIASED WITH R TIME IN MONTHS TH TIME HORIZON I INITIAL PERIOD CAMS 2.05 PC AT,'XT 89/01/2 15i 3>i)3 PAGE 2 AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYM1BOL LISTING SETS TP TIME PERIODS V, VINTAGE VF FIRST VINTAGE VTOV AGING 72Tt' ALIASED WITH V PARAMETERS AGWFORCE AGRIC WORKFORCE IN 1987 (MILL WORKERS) ALPHA DEMAND CURVE INTERCEPT AREAA ANNUAL CROP AREA (1000 ACRES) AREAP PERENNIAL CROP AREA (1000 ACRES) AREAS LAND USE SUMMARY (1000 ACRES) BETA DEMAND CURVE GRADIENT CALLOCA GROSS CROPPING ALLOCATION (1000 ACRES) CALLOCP GROSS CROPPING ALLOCATION (1000 ACRES) CINDX CONSUMPTION INDEX (1985-89 - 1) DELTA DISCOUNT FACTOR DEMDAT DEMAND DATA FRAT RATIO OF FAMILY LABOR (FRACT) FREQ FALLOW LAND REQUIRED FOR ROTATION OF ANNUALS (FRACTIONS) GAMMA DEMAND CURVE INCOME SHIFT GF ADDITION TO FALLOW LAND (1000 ACRES) GPROD GROSS Pi'ODUCTION (1000 TONS) GROSSCREV GROSS COCOA REVENUE (MILL CEDIS) COSTA TOTAL INPUT COST FOR ANNUALS (1000 CPA) CCSTP TOTAL INPUT COST FOR PERENNIALS (1000 CPA) NCOSTCL TOTAL INPUT COST CLEARING (1000 CEDIS PER ACRE) NCOSTPL TOTAL INPUT COST PLANTING (1000 CEDIS PER ACRE) N,DEXDk.'A INDEX NUMBERS P.F-'-% INPUT REQUIREMENTS FOR ANNUALS (UPA) REQP INPUT REQUIREMENTS FOR PERENNIALS (UNITS PER ACRE) SUM INCOME SUMMARY (BILL CEDIS) AiBOAL LABOR USE (1000 MD) LABINDX LABOR FORCE INDEX (1985-89 - 1) LABPA LABOR REQUIkEMENTS: ANNUALS (PROP) LABPAN LABOR REQUIREMENTS NORTH (PROP) LABPAS LABOR REQUIREMENTS FOR ANNUALS SOUTH (PROP) LABPP LABOR REQUIREMENTS FOR PERENNIALS (PROP)- LABPRCL LABOR PROPORTIONS FOR LAND CLEARING (PROP) TABPRPL LABOR PROPORTIONS FOR PLANTING PERENNIALS (PROP) LABREQA TOTAL LABOR REQUIREMENTS:ANNUALS (MDPA) LABREQP. TOTAL LABOR REQUIREMENTS (MDPA) LABSUPT LABOR AVAILABILITY (1000 MANDAYS) LABU LABOR USE AND SOURCSES (1000 MD) LANDCLIN LANu CLEARING INPUT REQUIREMENTS (UPA) LANDCLLAB LAND CLEARING LABOR REQUIREMENTS (MDPA) ENGTH LENGTH OF TIME PERIODS FACTOR FRACTION OF LABOR INCREASE TAY IMPORT.TAX RATE NYIELDA NET YIELD FOR ANNUALS (KG PER ACRE) PALLOC PRODUCTION ALLOCATION BY REGION (PROP) GAMS 2.05 PC AT/XT 89/01/27 15:15,03 PAGE 3 AGRICULTURAL SECTOR MODEL FOR GHANA- BASE RUN SYMBOL LISTING PARAMETERS PE COMMODITY EXPORT PRICES PEST PRODUCTION ESTIMATES FOR 1987 PFACTOR FRACTION OF POPULATION INCREASE PINDX PRODUCTIVITY INDEX (1985-89 - 1) PIREP PRICE SUMMARY PLANTPIN PERENNIALS INPUT REQUIREMENTS FOR PLANTING (UPA) PLANTPLB PERENNIALS LABOR REQUIREMENTS FOR PLANTINC (MDPA) PM COMMODITY IMPORT PRICES POPINDX POPULATION INDEX (1985-89 - 1) PREP PRICE SUMMARY (CEDIS PER KG) PRICE REFERNCE PRICE (CEDIS PER KG) PRICEDAT DEMAND DATA PRICEIN PRICE OF INPUTS (CEDIS PER UNIT) PRODR PRODUCTION BY REGION (1000 TONE.) PTOT TOTALS FOR NORMALIZING LABOR (PROP) PTOTP TOTALS FOR NORMALIZING (PROP) SEED SEED REQUIREMENTS (KG PER ACRE) STOCK STOCK OF PERENNIALS FROM SURVEY 1987 (ACRES) TALLOCA TECHNOLOGY ALLOCATION FOR ANNUALS (PROP) TALLOCP TECHNOLOGY ALLOCATION FOR PERENNIALS (PROP) TLABA TOTAL LABOR REQUIREMENTS (MDPA) TLABCL TOTAL LABOR REQUIREMNTS CLEAR.NG (MDPA) TLABP TOTAL LABOR REQUIP-MENTS (MDPA) TLABPCL TOTAL LABOR PROPOR1.ONS CLEARING (PROP) TLABPL TOTAL LABOR REQUIREMNTS PLANTING (MDPA) TLABPPL TOTAL LABOR PROPORTIONS PLANTING (PROP) TLAND TOTAL LAND FROM CALIBRATION (1000 ACRES) TLANDP ALL VINTAGES FRO 1987 (ACRES) TNUMA NUMBER OF TECHNOLOGIES TNUMP TOTAL TECHNOLOGY (ACRES) TTREQA TOTAL TASK REQUIREMENTS: ANNUALS (MD PER ACRE) TTREQP TOTAL TASK REQUIREMENTS: PERENNIALS (MDPA) WAGECAS CASUAL WAGE RATE (1000) WAGEFw^M RESERVATION WAGE RATE (1000 WFACTOR FRACTION OF WAGE INCREASE WFORCE RURAL WORKFORCE IN 1987 (MILL WORKERS) WVYIELD WEIGHTED YIELD BY ROTATION LENGTH (KG PER ACRE) WYIELD WEIGHTED YIELD (KG PER ACRE) YFACTOR FRACTION OF INCOME INREASE YIELD PROCESS YIELDS (KG FER ACRE) YIELDP YIELD OF PERENNIALS IN 1987 (KG PER ACRE) YIELDPI ADDITIONAL YIELD FROM FIRST PERIOD (KG PER ACRE) YINDX INCOME INDEX (1985-89 - 1) YPRO' YIELD PROPORTIONS (PROP) VARIABLES CL CLEAR FALLOW LAND (1000 ACRES) CPS CONSUMERS AND PRODUCERS SURPLUS (MILL CEDIS) EXPORTS NATIONAL EXPORTS . (1000 TONS) F LAND FALLOW (1)0O ACRES) IMPORTS NATIONAL IMPORTS (i TONS) ;AMS 2.05 PC AT/XT 89/01,2' 15: 1,:'! PAGE 37 kGRICULTURAL :iECTOR MODEL FOR GHANA- BASE RUN iYMBOL LISTING ?'ARIABLES INPCOS COST oF PURCHASED VARIABLE INFUTS (MILL CEDIS) LABCAS CASUAL LABOR (1000 MD) LABCOST LABOR COST (MILL CEDIS) LABFAM FAMILY LABOR (1000 MD) qATCON DOMESTIC CONSUMPTION (1000 TONS) )UTPUT PRODUCTION OF CROP COMMODITY (1000 TONS) PROFIT PROFIT FROM CROP PROCESS (MILL CEDIS ) kEVENUE REGIONAL REVENUE (MILL CEDIS) rCOST TOTAL INPUT COST rXA TOTAL ANNUAL LAND (1000 ACRES) LAND UNDER ANNUAL CROPS (1000 ACRES) KAF ABANDONING ANNUAL CROP LAND (1000 ACRES) LAND UNDER PERENNIAL CROPS (1000 ACRES) (PF ABANDONING PERENNIAL CROP LAND (1000 ACRES) JATIONS INP INPUT COST ACCOUNTING (MILL CEDIS) kLAB LABOR COST ACCOUNTING kTCOST TOTAL COST )EM DEMAND BALANCE (1000 TONS) )PROD PRODUCTION ACCOUNTING (lCjO TONS) FBAL FALLOW LAND BALANCE (1000 ACRES) 'MIN MINIMUM FELLOW REQUIREMENT LABORBAL LABOR BALANCE (MILL MD) LABORLIM LABOR LIMITS (MILL MD) )BJN NONLINEAR OBJECTIVE rxADEF TOTAL ANNUAL LAND DEF (1000 ACRES) KABAL ANNUAL LAND BALANCE (1000 ACRES) KPBAL PERENNIAL LAND BALANCE (1000 ACRES) 4ODELS 3HANAN k*** FILE SUMMARY INPUT D:\GHANA\NEWP.REP )UTPUT D:\GHANA\NEWP.LST -PILATION TIME - 0.329 MINUTES ANNEX 1 Numerical Shnulatlon Esthkatos Page 38 Scenario 1: Cocoa Supply Experiments Under Fixed Resources (Thousand of Metric Tons) Cocoa Buying Price 1987 Cedis/KG) Period 120 140 160 180 200 1990-94 243.6 298.7 307.1 311.5 313.8 1995-99 325.8 502.8 580.5 623.2 647.7 2000-04 489.2 595.4 635.4 674.9 713.5 2005-09 489.7 577.9 611.8 647.2 683.3 2010-14 456.6 450.0 480.5 578.1 558.9 2018-19 318.4 376.5 356.1 362.1 378.3 2020-24 438.5 561.7 613.0 651.2 683.0 2025-29 513.2 610.8 666.0 710.3 747.9 2030-34 513.2 610.8 666.0 710.3 747.9 2035-39 496.9 531.8 579.1 621.4 658.3 Average 1990-2009 387.1 493.7 533.7 564.2 589.6 1990-2019 387.2 466.9 495.2 522.8 549.2 1990-2039 428.5 511.6 549.6 583.0 613.3 Arc Elasticit.es of Cocoa Supply Fixed Resources Scenario Cocoa Buying Price Intervals Horizon 120-140 140-160 160-180 180-200 20-years 1.573 0.584 0.472 0.448 30-years 1.213 0.441 0.461 0.468 50-years 1.149 0.537 0.501 0.481 Average 1.312 0.521 0.478 0.456 ANNEX 1 Numerical Simulation Estimates Page 39 Scenario 1: Cocoa Acreage Adjustment to Price Under Fixed Resources (Thousand of Acres) Cocoa Buying Price (1987 Cedis/KG) Period 120 140 160 180 200 1985-89 2061.1 2061.1 2061.1 2061.1 2061.1 1990-94 1552.3 2494.7 3139.8 3360.4 3477.4 1995-99 2298.9 2666.3 2782.3 2936.0 3105.4 2000-04 2202.5 2542.6 2658.5 2812.2 2981.6 2005-09 2040.4 2407.8 2557.5 2702.9 2846.9 2010-14 2312.4 3081.7 3330.8 3517.7 3697.7 2015-19 3082.8 308i.7 3330.8 3519.8 3361.1 2020-24 2138.2 2545.1 2775.1 2959.7 3116.3 2025-29 2138.2 2545.1 2775.1 2959.7 3116.3 2030-34 2138.2 2545.1 2775.1 2959.7 3116.3 2035-39 2138.2 2545.1 2775.1 2959.7 3116.3 Scenario 2: Cocoa Supply Experiments Under Population Growth (Thousand of Metric Tons) Cocoa Buying Price (1987 Cedis/KG) Period 120 140 160 180 200 1990-94 244.5 300.4 312.7 316.9 320.9 1995-99 338.2 527.5 643.8 687.1 728.2 2000-04 529.3 714.2 77E.5 836.5 892.4 2005-09 541.6 707.2 796.1 851.2 909.8 2010-14 421.8 608.8 712.6 744.9 794.3 2018-19 665.9 638.0 662.7 738.9 828.2 2020-24 953.9 1282.8 1402.5 1477.6 1551.9 . 2025-29 1062.0 1449.4 1578.0 1697.2 1780.9 2030-34 1062.0 t.449.4 1583.1 1709.8 1851.4 2035-39 958.9 1336.9 1444.4 1547.5 1678.2 Average 1990-2009 415.9 562.3 632.3 672.9 712.8 1990-2019 458.6 582.7 650.7 695.9 745.6 1990-2039 678.8 901.5 991.2 1060.8 1133.6 ; ANNEX 1 Numerical Sknulatlon Estkiates Page 40 Arc Elasticities of Cocoa Supply Population Growth Scenario Cocoa Buying Price Intervals Horizon 120-140 140-160 160-180 180-200 20-years 0.973 0.439 0.264 0.274 30-years 0.775 0.414 0.285 0.328 50-years 0.916 0.355 0.288 0.315 Average 0.888 0.403 0.279 0.306 Scenario 2: Cocoa Acreage Adjustment to Price Under Population Growth (Thousands of Acres) Cocoa Buying Price (1987 Cedis/KG) Period 120 140 160 180 200 1985-89 2061.1 2061.1 2061.1 2061.1 2061.1 1990-94 1596.9 2806.6 3420.0 3631.7 3833.7 1995-99 2515.2 3131.5 3426.2 3687.9 3926.2 2000-04 2391.4 3131.5 3360.5 3582.5 3829.1 2005-09 2256.7 2971.5 3360.5 3593.9 3829.1 2010-14 3333.9 4305.0 5152.0 5575.8 6142.5 2015-19 5541.4 6788.5 6956.7 7295.6 7613.9 2020-24 4425.2 6039.2 6654.4 7145.2 7486.4 2025-29 4425.2 6039.2 6596.3 7126.8 7765.3 2030-34 4425.2 6039.2 6596.3 7126.8 7765.3 2035-39 4425.2 6039.2 6596.3 7115.4 7765.3 ANNEX I Numerical Sinulation Esthates Page 41 Scenario 3: Cocoa Supply Under Growth in Population and Income (Thousands of Metric Tons) Cocoa Buying Price (1987 Cedis/KG) Period 120 140 160 180 200 .Case A: Constant Wage 1990-94 237.0 296.2 305.6 312.4 316.8 1995-99 253.7 475.0 571.5 636.8 683.3 2000-04 291.1 540.9 678.1 734.6 810.9 2005-09 273.8 518.6 660.6 746.1 817.3 2010-14 119.6 368.0 528.8 685.3 675.0 2018-19 125.3 246.7 378.7 464.8 530.1 2020-24 130.4 459.2 926.2 1152.7 1248.8 2025-29 156.5 542.4 1057.7 1272.2 1414.1 2030-34 156.5 542.4 1057.7 1250.0 1433.8 2035-39 156.5 499.2 788.9 1121.4 1329.7 Average 1990-2009 263.9 457.7 554.0 607.0 657.1 1990-2019 216.8 407.6 520.6 596.7 638.9 1990-2039 190.0 448.9 695.4 837.6 926.0 Case Be Increasing WAif 1990-94 234.9 296.C 303.2 312.8 316.0 1995-99 248.0 469.4 548.1 638.4 675.8 2000-04 221.2 494.8 654.3 714.6 800.8 2005-09 194.9 463.5 637.8 705.7 807.0 2010-14 40.7 309.3 498.1 728.0 665.4 2018-19 39.6 224.8 344.7 414.1 529.7 2020-24 0.0 262.0 725.5 1048.0 1195.6 2025-29 0.0 314.4 836.0 1191.3 1346.2 2030-34 0.0 314.4 836.0 1135.5 1346.2 2035-39 0.0 314.4 662.9 874.7 1235.6 Average 1990-2009 224.8 430.9 535.9 592.9 649.9 1990-2019 163.2 376.3 497.7 585.6 632.5 1990-2039 97.9 346.3 604.7 776.3 891.8 ANNEX I Numerical Sknulatbon Estinates Page 42 Arc Elasticities of Cocoa Supply Population and Income Growth Scenario Cocoa Buying Price Intervals Horizon 120-140 140-160 160-180 180-200 Case A: Constant Wass 20-years 1.746 0.714 0.392 0.373 30-years 1.986 0.913 0.579 0.324 50-years 2.634 1.616 0.788 0.476 Average 2.122 1.081 0.586 0.391 Case B: Increasing WAr 20syears 2.043 0.815 0.429 0.436 30-years 2.567 1.042 0.690 0.366 50-years 3.493 2.038 1.056 0.658 Average 2.7C1 1.298 0.725 0.487 ANNEX 1 Numerical Skmulatlon Estimates Page 43 Scenario 3: Cocoa Acreage Adjustment Growth in Population and Income (Thousand of Acres) Cocoa Buying Price (i987 Cedis/KG) Period 120 140 160 180 200 Case A: Constant Wage 1985-89 2061.1 2061.1 2061.1 2061.1 2061.1 1990-94 1221.5 2594.2 3064.7 3405.4 3624.4 2095-99 1489.2 2419.2 3011.1 3209.9 3562.9 2000-04 1275.4 2295.4 2887.3 3238.1 3439.1 2005-09 1140.7 2160.7 2752.6 3473.4 3439.1 2010-14 498.3 1698.2 3230.2 3864.3 4469.3 2015-19 1150.3 3116.3 4778.1 5681.6 6245.7 2020-24 652.0 2259.8 4407.2 5452.6 5882.8 2025-29 652.0 2259.8 4407.2 5300.7 5992.4 2030-34 652.0 2259.8 4407.2 5300.7 5992.4 2035-39 652.0 2049.9 3287.7 4930.7 5992.4 Case B: Increasin Wage 1985-89 2061.1 2061.1 2061.1 2061.1 2061.1 1990-94 1116.9 2588.7 2945.1 3424.3 3587.4 1995-99 1210.2 2189.7 2916.1 2121.4 3520.0 2000-04 946.8 2065.9 2792.3 2997.6 3396.2 2005-09 812.0 1931.2 2657.6 3793.2 3396.2 2010-14 169.6 1288.8 2736.2 3600.6 4462.7 2015-19 169.6 2404.8 4158.7 5085.4 5966.4 2020-24 0.0 1310.0 3483.3 4964.0 5609.3 2025-29 0.0 1310.0 3483., 4964.0 5609.3 2030-34 0.0 1310.0 3483.3 4964.0 5609.3 2035-39 0.0 1310.0 2762.3 3664.9 5609.3 ANNEX 1 Numerical Simulation Estknates Page 44 Scenario 4: Cocoa Supply Under Growth in Population and Income With Zero Food Imports (Thousand cf Metric Tons) Cocoa Buying Price (1987 Cedis/KG) Period 120 140 160 180 200 Increasing Wage Cast 1990-94 224.0 234.5 264.6 291.0 300.3 1995-99 216.2 216.2 295.7 439.5 529.5 2000-04 186.5 186.5 408.7 637.7 717.8 2005-09 154.2 154.2 405.1 647.0 720.8 2010-14 0.0 0.0 257.0 508.5 599.1 2018-19 0.0 0.0 250.5 464.4 602.6 2020-24 0.0 0.0 246.6 772.0 1047.5 2025-29 0.0 0.0 281.5 889.0 1179.0 2030-34 0.0 0.0 281.5 889.0 1179.0 2035-39 0.0 0.0 209.2 701.7 .976.3 Average 1990-2009 195.2 197.8 343.5 503.8 567.1 1990-2019 130.2 131.9 313.6 498.0 578.4 1990-2039 78.1 79.1 290.0 624.0 785.2 Arc Elasticities of Cocoa Supply Population and Income Growth With Zero Imports Scenario Cocoa Buying Price Intervals Horizon 120-140 140-160 160-180 180-200 20-years 0.043 2.019 1.608 0.562 30-years 0.042 3.059 1.931 0.710 50-years 0.041 4.286 3.106 1.087 Average 0.042 3.121 2.215 0.786 ANNEX 1 Numerical ShAtlon Esthat. Page 45 Scenario 4: Cocoa Acreage Adjustment Under Growth in Pouplation and Incomo With Zoro Food Imports (Thousand of Acres) Cocoa Buying Price (1987 Cedis/KG) Period 120 140 160 180 200 1985-89 (Base) 2061.1 2061.1 2061.1 2061.1 2061.1 1990-94 1040.6 1098.9 1571.3 2337.6 2803.3 1995-99 900.9 900.9 1946.5 2954.5 3261.8 2020-04 777.1 777.1 1822.7 2830.7 3138.0 2005-09 642.4 642.4 1688.0 2696.0 3003.3 2010-04 0.0 0.0 1344.5 2833.8 3985.6 2015-19 0.0 0.0 1892.9 4745.3 5795.5 2020-24 0.0 0.0 1172.8 3704.0 4912.5 2025-29 0.0 0.0 1172.8 3704.0 4912.5 2030-34 0.0 0.0 1172.8 3704.0 4912.5 2035-39 0.0 0.0 871.9 2923.8 4327.8 ANNEX 1 Numerical Shiulation Esthnates Page 46 StiMULATED LAND USE IN 2015-19 PERIOD (MILLIONS OF ACRES) Cocoa Producer Perennial Annual Scenario Price Crops Crops Fallow Total Fixed Resources 120 4.390 4.323 0.973 9.686 140 4.498 4.237 0.951 9.686 160 4.583 4.169 0.934 9.686 180 4.692 4.082 0.912 9.686 200 4.800 3.995 0.891 9.686 Population Grovth Only 120 8.743 10.142 2.299 21.184 140 9.789 9.305 2.090 21.184 160 10.040 9.105 2.039 21.184 180 10.252 8.935 1.997 21.184' 200 10.447 8.779 1.958 21.184 Population & Income Grovth 120 5.742 12.543 2.899 21.184 140 7.172 11.399 2.613 21.184 160 8.402 10.415 2.367 21.184 180 9.554 9.493 2.i37 21.184 200 9.926 9.196 2.062 21.184 Population, Inceme & Wage Growth 120 4.872 13.139 3.073 21.184 140 6.671 11.800 2.713 21.184 160 7.970 10.761 2.453 21.184 180 8.990 9.945 2.249 21.184 200 9.759 9.330 2.095 21.184 Food Self Sufficiency 120 4.455 13.573 3.156 21.184 140 4.456 13.572 3.156 21.184 160 6.015 12.324 2.844 21.184 180 8.277 10.515 2.392 21.184 200 9.120 9.841 2.223 21.184 ANNEX 1 Numerical Simulation Estinates SIMULATED NET FARM INCOME, 1990-2039 Page 47 (BILLIONS OF 1987 CEDIS) COCOA PRODUCER PRICES (1987 CEDIS) Time Period 120 140 160 180 200 Fix Resources 1990-94 87.1 81.7 96.7 109.7 119.4 1995-99 87.9 132.8 168.9 199.6 226.0 2000-04 106.6 150.3 185.4 219.0 251.0 2005-09 116.8 150.4 179.0 210.0 242.0 2010-14 104.0 122.9 149.6 178.0 207.8 2015-19 86.5 113.6 128.4 147.9 168.6 2020-24 119.7 159.9 192.7 224.7 255.2 2025-29 108.8 151.7 190.4 224.3 256.1 2030-34 108.6 141.0 175.9 211.3 243.4 2035-39 119.1 133.4 164.8 194.0 227.1L Population Grovth Only 1990-94 125.5 123.2 140.5 155.4 169.3 1995-99 123.0 176.4 227.5 262.3 297.5 2000-04 162.9 224.8 273.2 321.1 365.9 2005-09 192.2 249.1 294.4 338.4 389.3 2010-14 175.0 236.0 288.7 329.2 375.8 2015-19 212.3 248.4 293.4 346.7 402.8 2020-24 302.4 407.2 484.4 557.1 629.4 2025-29 276.2 394.2 487.8 577.8 659.4 2030-34 286.7 396.0 465.0 556.6 655.2 2035-39 342.8 401.9 474.4 550.0 636.8 SIMULATED NET FARM INCOME, 1990-2039 ANNEX 1 (BILLIONS OF 1987 CEDIS) Numerical Simulation Estimates Page 48 COCOA PRODUCER PRICES (1987 CEDIS) Time Period 120 140 160 180 200 Population and Income Growth 1990-94 135.0 134.6 147.6 162.0 176.3 1995-99 124.4 183.0 227.4 269.6 305.4 2000-04 158.1 232.2 282.8 329.0 381.8 2005-09 187.7 268.6 322.4 371.1 419.2 2010-14 221.0 291.7 352.9 420.3 447.8 2015-19 210.5 280.9 335.9 386.2 433.4 2020-24 290.7 399.7 529.8 627.2 706.2 2025-29 356.0 475.2 595.2 706.4 803.5 2030-34 210.7 247.5 342.2 427.3 573.4 2035-39 449.9 624.2 726.0 838.8 889.3 Population, Income & Wage Grovth 1990-94 134.4 132.2 143.0 159.2 172.0 1995-99 113.8 173.7 212.3 260.2 294.4 2000-04 133.4 210.6 266.5 311.7 365.1 2005-09 155.2 238.6 298.8 342.5 400.0 2010-14 181.4 257.1 322.2 399.9 422.4 2015-19 174.6 237.6 296.0 338.3 388.3 2020-24 219.8 309.6 449.6 552.2 636.9 2025-29 273.6 370.0 527.4 616.8 721.4 2030-34 151.7 175.7 231.0 343.6 423.8 2035-39 386.3 494.0 630.4 724.2 806.9 Food Self-Sufficiency 1990-94 171.4 177.7 187.1 195.4 201.5 1995-99 191.5 196.3 213.7 245.5 292.3 2000-04 231.6 235.3 280.6 332.2 374.6 2005-09 280.5 283.6 334.4 388.5 '26.4 2010-14 322.8 322.8 368.1 420.6 453.2 2015-19 362.3 362.3 413.0 456.4 486.3 2020-24 444.3 444.3 490.9 589.2 692.3 2025-29 521.5 521.5 574.4 693.5 768.1 2030-34 267.9 267.9 327.2 468.6 564.5 2035-39 752.7 752.7 799.2 910.3 979.8 f ~~~~~~a ~~~~~ I ANNEXU OPTIMAL TRADE TAXES ON AGRICULTURE IN DEVELOPING COUNTRIES by David IL No"y Optimal trade taxes on agriculture in developing countries by David M Newbery' University of California at Berkeley March 3, 1988 (Revised November 6, 1988) t. Introduction Food crops produced on family farms and marketed by the private sector are often hard to tax at any point betweer. local production and consumption in developing countries (though it may be possible to subsidize urban consumers by subsidizing the costs of transporting, storage and handling). It is easy, and common, to tax agricultural imports and exports. If domestic transactions cannot be taxed, how should agricultural trade taxes be set? A number of developing countries (Ghana is a good example) export non-food agriculturai products (such as cocoa, coffee, rubber, sisal, cotton) and import cereals, at the same time as meeting a consiramble fraction of cereal demand from domestic production. The non--food exports are often taxed, either by an export tax, or by means of a state monopoly on the domestic purchase and export of the crop. The domestic cereal price may be held below the import price (equivalent to import subsidies), or, more usually in the recent period of low ';;orld cereal prices, imports may be taxed and the domestic price kept above the import price. This policy is sometimes defended as improving food self-sufficiency and hence the food security of the country. The second objective of this paper is to .xamine the case for either a tariff on cereal imports, or a subsidy which lowers domestic consumer and producer prices below the import parity price. I Written whilst on leave from Cambridge University, and revised at the Departmnent of Applied Economics, Cambridge, England. I am indebted to Go, ion Hughes tor his extensive computer analysis of the Ghanaian data reported later in this paper, and to Gerry O'Mara for his help in obtaining the data and providing some of the results used in the -mpirical section. . . . .~~~~~~~~~~ Finallv, if agricultural exports are taxed, then there is a case for subsidising agricultural inputs (such as fertilizers, pesticides, and equipment). How should the level of subsidies be set, given that different agricultural products may be subject to different rates of tax? Opdmal tax theory (as set out by Diamond and Mirrlees, 1971; Atkinson and Stiglitz, 1980, and Newbery and Stem, 1987, for example) has been developed to answer such questions. Given the government's objectives (summar, d by a social welfare function), and a specification of the tax instruments available, and a descripotion of the economy (production possibilities. endowments, preferences), the theory characterises the optimal rates at which to set the specified taxes. Optimal tax theory has been most successful in characterising tax systems in developed market economies satisfying strong, but not unreasonable assumptions- for example, nat indirect taxes can be restricted to fnal consumption, that all goods can be taxed, and that there are constant returns to scale in production. Under stronger assumptions the optimal tax system takes a very simple form in which th rate of indirect taxation is the same for all goods (and could equally be zero) with all redistribubon confined to the income tax2. Two important lessons can be drawn from the opdmal tax literature. The first is that the optimal rate at which to set any tax depends sensitively on the set of tax instruments available for reform. If income cannot be redistributed through changes in the income tax system, or by making uniform lump-sum transfers, then cotunodity tax rates will reflect the tension between equity and efficiency pulling in opposite directions. If altemative redistributive options are available, com iity tax rates wiU primarily reflect efficiency conisiderations. The second lesson is that the choice of functional form for estimating consumer demands may lrgdy prejudge the design of optimal taxes, by imposing such 2 The conditions for this are discussed in Atkinson and Stiglitz (1980, ch 14), and Newbery and Stern (1987, pp86-7). If the income tax may 1'. non-linear, a sufficient condition is that consumers differ only in their wage raes and that the utility function is weakly separable between labor and all other consumption goods. [f the income tax must be linear (which is equivalent to uniform commodity taxes plus an optimal uniform lump-sum grant) then in addition Engels curves must be linear and identical for all consumers. critical features as separability, linearity, and homotheticitv (Deaton, 1987). Despite the apparent fragility or sensitivity of the conclusions of optimal tax theory, there appears to be a consensus that in developed countries in which the whole tax system is open to reform, it is hard to justify differential commodity taxes (Stiglitz, 1986, ch 19; Davis and Kay, 0985). Useful though these conclusions are, they do not readily apply to developing countries, at least, those with a substantial peasant agriculture. The agricultural sector violates the assumptions on which the optimal tax results rest in a number of important ways - agricultural (and other) incomes) may be hard to tax, it may not be possible to tax final consumption without also taxing production, and it may be un-easonable to assume constant returns to scale (see for example NewbeTy, 1987a). Diamond and Mirlces (1971) recognised these limitations, and pointed out that the agricultural sector could be formally included in their model as part of the household sector, rather than part of the productive sector. This is satisfactory if the sales prices of all agricultural goods can be adjusted by trade taxes or taxes on marketed surplus, but this may not be possible for non-traded agricultural goods. The marketed surplus may be hard to tax by sales taxes, and, being non-traded, their domestic price cannot be directly set by trade taxes. Restrictions on the range of consumer goods which can be taxed have been explored by Heady and Mitra (1982, 1987), and shown to have important consequences for the rates at which other goods should optimally be taxed. Rclatively little work has been done on the optimal taxation of different agricultural goods. Most theoretical studies assume that there is a single agricultural good (Heady and Mitra, 1987; Atkinson, 1987). Sah and Stiglitz (1987) have characterised the optimal set of agricultural prices for different products, and have derived appealing rules for special cases - notably for the taxation of different non-food agricultural outputs - but admit that the general problem of setting prices for food and non-food crops optimally requires a daunting amount of infonnation. Faced with this problem, a number of researchers have tackled the more tractable problem of identifying desirable directions in which agricultural prices or taxes might be changed. Starting from the present set of taxes, is it desirable to lower cereal prices whilst raising taxes on other consumer goods? Braverman and his collea-ues (Braverman, Hammer and Ahn, 1987; Braverman and Hammer, 1986; Braverman, Hammer and Gron, 1987) and Newbery (1987b) have applied this technique of reform analysis to agricultural price reform in a number of different countries. The strategy adopted in this paper represents a compromise between the methods of optimal tax theory, or tax design, and those of tax reform analysis. The main problem in characterising an optimal set of taxes is that the tax rates depend on own and cross-price elasticities of demard and supply, whose number increases rapidly with the number of separate goods distinguished. Tractability requires that the number of goods be severely restricted by judicious aggregation. The model analysed below examines the optimal taxes on just two goods - an aggregate modern sector or imported non-food consumption good, and the non-food export crop. The non-traJed informal or agricultural sector goods - various foods and non-foods - are assumed (realistically) to be non-taxed (modern non-traded services like transport and electricity are aggregated with the taxed consumer goods). This leaves cereals and agricultural inputs as potentially taxable. The tax on inputs can be discussed after the other tax rates have been set, and its uptimal rate can be determined without much greater effort. Cereals are treated rather differently. The question posed is whether, starting from free trade in cereals with no tariffs or subsidies, it is desirable to raise or lower the domestic price. In particular, what appear to be the main factors which might justify imposing an import tariff or providing a subsidy? Thus, the pricing of the export crop is considered to be one of tax design, or of setting the producer price, whilst the pricing of cereals is considered one of tax reforn, though starting not from the current price, but from free-trade. For Ghana, this is a realistic set of policy options, as the government must each year announce the cocoa buying price, whilst its food policy can be gradually adjusted but is unlikely to be radically revised, given the political sensitivity of the price of food, itself closely affected by the price of imported cereals. The main features of the problem which distinguish it from standard optimal tax 5 calculations are - the potential importance of non-taxed, non-raded consumption moods. and their response to tax changes, and the sensitivity of wages (and/or migration) to agricultural taxes. A minor, but practically important feature is that the economy is assumed to have some market power in the world market for its export crop. The main analytical result is that although the price changes of the non-traded goods induced by changes in the price of the export crop and cereals can be ignored as far as their impact on revenue goes, they have potentially important distributional effects which will modify the choice of agricultural prices. This confirms the earlier, more limited analysis in Newbery (1987a) and suggests that it is a reasonably robust rule which considerably simplifies the analysis. The informational requirements for setting the tax rates or identifying the direction in which to change the cereal price appear to be manageable. 2. The model An optimal set of taxes (or consumer prices, q) maximizes social welfare W(Vh(q,mh)) subject to raising the required level of government revenue, R, and subject to production feasibility. (Vh(.) is the indirect utility function of household h, with lump sum mh, facing consumer prices q). A convenient way of characterising the optimum is to define the; marginal social cost of raising an extra dollar of revenue by changing the ith price, qi, (1) 1 (I) Ai 'dRid where the numerator is the fall in social welfare caused by raising qi, and *he denominator is the extra revenue raised. (See Stem, 1987a). An optimum is then a set of tax rates (or consumer prices, q) such that Xi X., all i. If Xi > XA for some pair of goods i, j, then social welfare could be increased by reducing the tax collected from the ith good and replacing it by additional taxes collected on the jth good. The Xi's can therefore be u,sed to identify directions of desirable reform. 6 It is algebraicallv more convenient to work in terms of the inverse of these marginal social costs, and to that end define the benefit-cost ratio, or BCR, of the ith tax as t2) ~~~~~~~~~~aR/aq. _,1 (2) *= xi I Thus vi is the revenue raised per unit of social cost, and is a measure of the productivity in revenue terms of the tax. Farmers grow the export crop (cocoa), cereals, and non-traded food crops (roots, such as yams and cassava, vegetables, etc). They buy taxed consumer goods ('manufactures') and untaxed, non-aded consumer goods. For most purposes, there is no need to distinguish between non-traded, untaxed agricultunal and non-agricultural goods, and the goods will therefore be numbered 1-4, with cocoa as good 1, and manufactures as good 4. The world prices of cocoa and cereals are Pi, P2, and the domestic producer prices of goods are pi, i - 1-4. The relevant domestic consumer prices are P2, p3. and q4, equal to the domestic producer price except for taxed manufactures. Initially cereals are untaxed so that P2 = P2. Government revenue is (3) R = (P1-p1)Yl + (P2-P2)(Y2-X2) + (q4-p4)X4 + tc'4(P4, w), where Yi is aggregate output and Xi is aggregate consumption of good i, w is the wage rate, tc is the tax rate on profits, and it4 is the profits from producing manufactures. Incomes outside the modern sector are assumed untaxed, and taxed iiscomes are subsumed into profits taxes for simplicity (though they could be sepamtel, distinguished if necessary). The second term will be zero, but its derivative with respect to P2 will noL The producer price of manufacturing, p4. is set on world markets, and q4 - p4 + t4, so the consumer price is altered by changing the tax rate. The producer prices of cocoa and cereals are set directly, or by setting trade taxes. The BCR of taxing manufactures is found as follows. The evenue derivative is 7 (4) =RM +,C 4 q- 4 4 4 g = E4(l ' rz"44)' where ,4 = (q4 - p4)/q4 is the tax rae on the tax-inclusive price, and £4- alogX4/alogq4 is the price elasticity of aggregate demand for manufactures. The impact on social welfare is (S) G4 haVhaq 4-- dhXA, where oh =aw/avh.aVh/amh is the socia value of tsfemng income to household h, Xh is the consumption of manufactures by h, (and aVh/aq 4 - -x4I.Vh/dnh from the properties, h of the indirect udlity function), X4 - ;x4 is agrepte consumption of manufactures, and d4 is the distributional charcteristic of manufacturs, defined as h - a measure of the extent to which the good is consumned by the socially desering. Putting (4) and (5) together, the BCR of taxes on good 4 is (6) 1 d Now consider the problem of setting the price of cocoa, which will affect the level of cocoa exports, and hence the world price, P1. The impact on revenue will be (7) r~~~ ~ ~~~p1 ay1 dh dx Y1+Y1 + { I P1 r p 1 + t4{hb4 pI+ q4 .ipI c as 84 8w h h h where br q^x4/axm is the marginal expenditure share of household h on manufactures. The first term in (7) is the direct impact on revenue, holding output constant. The 8 second term may be written (8) PI - Pll y where p1 * Pl(1- _/en) is the marginal revenue from extra cocoa exports and en is the elasticity of net demand for cocoa by the rest of the world, (as a positive number, allowing for supply responses elsewhere), and Ti is the elasticity of cocoa supply, alogY1/alogp,. Thus pi is the shadow price of exports, rather thar P?, and (p1 - pl)/p, = tl is the shadow tax rate on cocoa. (See the discussion of shadow taxes in Stem, 1987b). The third term in (7) is the tax revenae collected on manufactures by changing incomes, mh, and other prices. This, together with the final term, will be argued to be h negligible. The marginal expenditure shares, bN, will be assumed identical for all hcuseholds (equivalent to identical linear Engels curves), in which case the first term in the bracket is (9) ,4b4la-pi+ (t4+ (1-t )3- ) 51}I where rl((p1, P2, p3) is total agricultural (strictly informnal sector) profits, and t4 is employment in the formal sector. All wage changes within the agricultural sector cancel out, as they merely redistribute income from labor hirers to laborers, and, with linear Engels curves, this will have no impact on aggregate demand. Profits after taX in the formal sector, it4, are assumed to be distributed and spent, and the last term in (9) measures the induced income change on these profit recipients. The first term gives the total impact on profits of the change in cocoa price (hence the full, rather than the p&Ldal derivative). It may be evaluated by noting that ar/api - Yi from the properties of the profit function: drl arl an3 3 Y 9 where *3 alogp3/alogp is the response of the price of non-traded gcods to the price or cocoa. The second part of the third term in (7) is similarly dX4 q4X4e43Y31 [q4X4643 - q3X3b4l Y31 Caip I Pp1 P1 where ei= - alogx./alogpj is the uncompensated cross-price elasticity of demand for good i, and eij is the compensated elasticity. Finally, a74/aw a £4, again from the properties of the profit function. Putting all these expressions together gives (10) - -Y I'-ll - 14(b4 - j-4e3Y3)} +bt4 ( 4b4 The fourth and fifth terms are likely to be small, and to be of opposite sign, and the error in ignoring them will therefore also be small. As an approximation, then, (11) aR w _yl{l -tlTll-t (This confirms the earlier analysis in Newbery (1987a), that the revenue impact of price and income changes mediated through price changes of untaxed, non-traded goods can typically be ignored, as they effectively amount to the tax consequences of redistributing a fixed total incnme from one side of each market to the other - as the demands equal supplies for all non-taded goods). The welfare impact is aw , h (d7ch +h dw n dP3) a5P, h lap-, s Jpl X3 P 10 , h J vh h (V-h ap3 + Ih aw where fs is the net sales of labor by h (negative for farms hiring labor). The welfare impact can be written as (12) ; a Y dP + 3(4 - )y wL (d' d al where dP= Z3hyi/Yi is the distributional characteristic of good i in production, d= h/X is its distibutional characterisdc in consumption, d, - _ ht/L is the distributional characteristic of labor that is sold, and corfespondingly db is that for labor that is bough, and L is the total amount of labour that is sold. Whercas it was not necessary to calculate the price responses, yij, for the revenue impact, it is necessary for the distributional impact, and if direct estimates are not available, the appendix shows how they may be calculated in terms of elasticities of demand and supply. Putting these expressions together (and substituting for the price response tern from the fornula of the appendix gives the BCR of raising cocoa taxes as (13) I - - where l.1aW, P+d c (113) wL- s alloo -5 1 ( 9 3)s3+e P p(d 4S!L is the effective distribudonal chaxcteristic of cocoa, allowing for the various market mediated redistribudonal effects. Whereas the revenue impact of these redistributional effect is small, their effect on a redistributionaily sensitive social welfare function may be significant. 11 The optimal shadow tax on exports, rt1 can now be found from (6) and (1S): (14) 1- IT1 - T4b4 = (P{1 + ?4(644 - b4)} where p - dld is the ratio of the distributional impacts of the two taxes. Thus if p = 1, 4~~~~~~~~~~~ (14') T4( E44) and the ratio of the tax rates is equal to the ratio of the compensated elasticities (as a positive number). The value of (p can be calculated from a household and farm budget survey, but in general one would expect d3 > d3 2 dL k d4, on the grounds that non-raded agricultural goods are likely to be produced by typical farmers who are usually poorer than typical consumers of such goods, whilst cocoa farmers are likely to be relatively wealthier than other farmers, though possibly poorer than the representative consumer of taxed manfuactured goods. It is therefore not clear whether p will be greater or less than one. Rearranging (14) gives (14") 1= (1- 1. (Later in the paper the calculation is done using Ghanaian data, for which (p > 1. The reason is that although dI > d4,' the difference is slight, and not enough to offset the consideable excess of dS over d3). Thus the explanation for why, with (p > 1, cocoa taxes should not be quite so high is that reducing the tax on cocoa farmen cause a rise in the price of non-,raded foods, as cocoa farmers switch resources out of their production into cocoa and reduce supply relative to demand at the old price. This in turn has advantageous redistributive effects, given the assumptions about the relative incomes of 1 consumers and producers. 2.1 The Government Budget Constraint If no other taxes are to be varied, then the optimal cocoa tax can be solved from equations (14") and (3), once the level of government revenue R is specified. The main problem is that most of the the variables on the right hand side of (14") are functions of prices, and hence of the tax rates, while the level of both taxes T1and T4 will depend on the amount to revenue to be raised. Nor is it sufficient to hold the amount of revenue in money terms constant - instead, the government revenue constraint might be specified as a set of activities which must be undertaken, whose value will depend on the prices prevailing in the new tax equilibrium. The problem will be geatly simplified if the government undertakes activities whose prices do not change in response to tax changes (perhaps because they involve traded goods only), and if everything is then valued at accounting prices. The practical problems involved are discussed in section 3 below. 22 Cereal tariffs The next question to address after setting the optimum tax on cocoa is whether, starting from a zero tariff on cereals, it is desirable to raise or lcwer the domestic price, p,. It will be desirable to impose a tariff and raise the price if the marginal social product of a cereal tariff, W2, is higher than the altematives, xV, = W4. The revenue impact of a cereal tariff is (15) ~~~aR y (15) F (-pl) + i X2 - Y2 + T4b4(Y2-X2) -T 021y 2+ (X2-Y2)(l - s4b4), where T21- alogY2/alogp, is the cross price elasticity of supply (which will be negative), and aY 2 Y2/ap1 from standard duality theory. Again, the revenue impacts arising from the price changes of non-tmded goods and labor have again been ignored as 13 negligible. The welfare effects will be aw hJyh h (h h aP3+ hawI Up2 =h5 {2 ~ 2 +(3h PI ShW + Ph 2 or 52 d2qY2 diX2 + X2{I (dp -2d3)Y32 + (dsdb)a } If s2= Y2/X2 is the degree of self-sufficiency in cereals, this can be written as (16) 3 ^-X2{(L-s2dq) + 52. 3( 3Y(dc p + di-d (using the formula of the appendix). If we deltme the effective distributional characteristic for cereals in terms of net sales of cereals (Y2 - X2) (which will be negative for an importing country), so that 2 - X2)d2 = -(-s2)X2d2, then the BCR for cereals taxation will be (17) W2= 2tl2l + (l-s,2)(i- 4b4) d2( I-s2) where d2(1-s2) is minus the expression in brr,es in (16). If the cocoa tax is being optimally set, then, using the formula for the optimal value of t from (14), (1-c4b4)(1 - X(l'-)) - (PXt4(E44) ~~~18) ~ ~ 'Y 14 where x= - >0 measures the relative importance and responsiveness of cereals to price changes. The direction in which to change the cereal price (or the tariff) starting from free trade will depend on the sign of W2 and its magnitude relative to ,W4. If w, and d2 are negative, then it is socially desirable to raise cereal prices because it redistributes incomes from consumers to more deserving producers whilst at the same time increasing revenue. In that case tariffs are unambiguously desirable regardless of the cost of raising revenue by other taxes. If 12 and d2 are both positive, so that there is a cos: in raising revenue by cereal tariffs, then the direction in which to change the cereal tariff is given by the sign of W2 2-4'4 (or w,2 - ,). If f= d2/d4, then (19) (''2 1414)d2 (-t4b4) l +X(l 4cl. the smaller number of producer (s2 < 1) offseu this. The second term is negative, as producers are poorer than consumers and for non-aided goods production is equal to demand. The last term is also neptive but qantadl y malL For hiher values of inequality averion, v, the first term decreases towds o ad the soeond term becomes increasingly negative, until the net effcct is neative. 4. Conclusions for Ghanaian Pricing Policy 9- - / The high supply response of cocoa at a price near the cost of replantina, which talls otf as the price rises above this, means that the producer price should not be allowed to fall below this level, even when the world price falls to the point at which this supply price exceeds 55 percent of the export price. Given the assumptions made, there appears to be a good case for rising the cocoa price to at least 55 percent of the export price, and possibly higher if the govemment is reasonably egalitarian (v 2 1). Only if the worli price rises subst.1ntially so that 55 percent is significantly above the very elastic part o' the supply schedule is there a case for a lower percentage, and then only if the govemment is rather inegalitarian. Setting the tariff rate on imported cereals is a more cor,,'.cated issue, and the paper only addresses the rather simpler question of whether there should be tariffs at all. It would appear that there is a case for protecting cereals on revenue and distributional grounds, though whether the currently high levels of protection are wamnted has not been addressed. It would in principle be possible to extend the analysis to ask whether cereal tariffs should be reduced starting from their current values, but that will have to await another occasion. 28 References Atkinson, A.B. (1987), ch 14 of Newbery and Stern (1987). Atkinson, A.B. and J.E. Stiglitz, (1980) Lectures on Public Economics, McGraw-Hill. Braverman, A., J. Hammer and C.Y. Ahn, (1987) ch 17 of Newbery and Stern (1987). Braverman, A., J. Hamuaer and A. Gron, (1987), "Multimarket Analysis of Agricultural Price Policies in an Operational Context: The Case of Cyprus", The World Bank Economic Review, 1(2), 337-56. Braverman, A. and J. Hammer, (1986), "Multimarket analysis of agricultural pricing in Senegal", in I. Singh, L. Squire, and J. Strauss, eds. Agricultural Household Models: Extensions, Applications and Policy, Washington DC. World Bank Davis, E. and J. Kay. (1985), "Extending the VAT base: problems and possibilities", Fiscal Vtudies, 1-16. Deaton, A.S., (1987), ch 4 of Newbery and Stern (1987). Deaton, A.S., (1988), 'New Approaches to Household Survey Data from Developing Counties", Discussion Paper 139, Woodrow Wilson School, Princeton. Diamond, P.A. and J.A. MirlIces, (1971),"Optimal Taxation and Public Production, Part I: Production Efficiency", Amer. Econ. Rev. 61(1), 8-27. Heady, C. and P. Mitra, (1982), "Restricted Redistributive Taxation, Shadow Prices and Trade Po'icy", J. P - Econ. 17(1), 1-22. Heady, C. and r. Mitra, (1987) ch 15 of Newbery and Stern (1987). Newberv, D.M. (1987a) ch 13 of Newbery and Stern (1987). Newbery, D.M. (1987b) ch 18 cf Newbery and Stern 1987). Newbery, D.M. and N.H. Stern, (1987), The Theory of Taxation for Developing Countries, Oxford University Press for the World Bank. Sah, R. and J.E. Stiglitz, (1987) ch 16 of Newbery and Stem (1987). Stiglitz, J.E.. (1986), The Economics of the Public Sector, Norton. Stern, N.H., (1987a), "The Theory of Optimal Commodity and Income Taxation", ch 2 of Newbery and Stern (1987). Stem, N.H., (1987b), "Aspects of the General Theory of Tax Reform", ch 3 of Newbery and Stern (1987). '9 Appendix The price response terms Yi; can be found by differentiating the equations for supply and demand of the non-raded god i totally with respect to pj: Yi -xi -0, whence Yij - t lij Some of the formulas in the text can be simplified further by noting that PiyiTlij - p.Yj.iT. (This comes from the standard duality result that aY;apj - Similar relations hold for the compensated elasticies, eij. GHANA'S COCOA ENVIRONMENT by MuTE J. katm ANNEX IlIl GHANA'S COCOA ENVRONMENT PAGE 1 OF 26 A. Real Farmr Prices and Ghana's Cocoa Production, 1963-87 A 1972 World Bank report on Ghanalan cocoa written by the author suggested that the cocoa policies pursued by the Goverrnent of Ghana durhg the 1965-72 period were In the process of destroying the industry. Planting had stopped because of a fall In real producer prices (see Table A-111-1) and capsid spraying and other forms of maintenance were on the decline. Although the process was slow, the deterioration was inevitable In the absence of a policy change. Producer prices In real terms fell from 100 In 1963 to the 44-52 range In 1965-6. They rebounded cose to 70 In the late 1960s and early 19709 but then tumbled below 50 again In the late 1970s and 1980s. The decline In production was clearly exacerbated by the droughts that occurred in the 1970s and 1980s. The droughts not only affected cocoa production but food output as well. As food prices soared, Ghanalan farmers had to protect themselves from starvation. Moreover, the price Incentives shIfted In favor of food. Food farmers replaced coco. farms as food prices rose and cocoa prices received by the farmer remained at relatively low levels due to government policy. Cocoa production, which had fallen slightly below the 400,000 tonne level prior to the droughts, fell to 300,000 in 1977 and then to 225,000 by 1981/2. Further deterloratlon occurred during the droughts of 1982/3 and 1983/4 as production fell below 200,000 tonnes. In 1984/5, a good weather year, output totalled only 175,000. As Table A-111.1 Indicates, the goverrvnunt'b cocoa price policy began to shift In 1985 when real cocoa prices ncreased from 30-51 and then to 60 In 1986 and to 69 In 1987. At 140,000 codis per tonne, farmer prices are back to the levels of the late 1960s and early 1970s relative to other prices In th economy. The Increase In prices has brought a short-run producer response In the form of reclaiming moribund cocoa, better farm maintenance and more extensive harvesting. The result has been a production -- --- ----- ----- ----- -- * I cr~ Cj T alima 6tuua F| , W.. ANNEX Ill GHANA'S COCOA ENVIRONMENT PAGE 3 OF 26 data available by region. The pod count Information provides tree yields for each year and, given the number of trees per hectare, it is possible to estimate yield per hectare. Data from Cocoa Board Indlcate that approximately .04542 kilos of cocoa Is produced per pod. The pod count Information Indicates that an average tree produced 20-25 pods In the mid-1970s but that the number had fallen to the 12-17 range by the Table A-111-2: Mature Cocoa Hectarage and Yieds by Region, 1967-87 (00 Hectares, KIM) 1976 190 1965 1967 Pegion Ho oT Kg/Ha Hot Kg/Ha HotC Kg/Ha HoCT Kg/Ha Ashanti 225 500 210 450 100 500 100 550 Brong-Ahafo 125 650 115 6oo 75 400 75 400 Easernm 140 500 90 500 70 400 70 450 C'ntral 120 500 so 400 S0 500 47 580 Westem 110 400 95 500 108 500 120 550 Volta 33 400 15 350 5 350 5 350 Souvo: Conyriodfty Ie*rrnaUon, Inc., based on COCOC300's regio* pod court data. 20-25 pods range by the mid-1980s (see Tables A-111-3-8). The crltlcal piece of information missing is the number of trees per hectare. The Ideal spacing cal:s for 1200 trees. It Is generally known that farmers plant clser than the Ideal knowing that some seedlings will die and that there will be a natural thinnig process take place over tie. Since most cocoa farms In Ghana were planted prior to 1962, considerable thinning has taken place - espeialy In light of the low malntenance levels In recent years due to low real prines. Consequently, the assumption Is made that 800-1,000 trees per hectare existed In the early 1970s, but that only 600-700 trees remained by the mid-1980s except In the Western Region where planting did take place in the late 19603 and 19709. ANNEX IIl GHANA'S COCOA ENVIRONMENT PAGE 4 OF 26 :n this region, the assumptlon Is made that the number of trees per hectare fell from 950 to 750 between 1970 and 1987. Given the assumptions outlWied above and the analysis performed In Tables A-1ll- 3-8, the data In Table A-111-2 has been aerived using the pod count data and output figures for each region. ANNEX III GHANA'S COCOA ENVIRONMENT PAGE S of 26 Table A-III-3: Ashanti Hectarage and Yield Estimates, 1969170-1987188 Crop Total Total Podsl Pods/Tree Prod'n/Tree c/.d/ Trees Yield.:a (kg) Prod'n Year Pods Sites Site Actual a4ov Ave bIActual Mov Ave Ha Actual Hov Ave (H.T.) Hectares 1969/70 38844 71 547 21.9 -- .99 . 800 795 -- 121487 152780 1970/71 35104 71 494 19.8 20.7 .90 .94 800 719 753 126262 175702 1971/72 36347 71 512 20.5 19.9 .93 .90 750 744 723 141009 189513 1972/73 34513 71 486 19.4 197.0 .88 .77 750 662 579 121722 183771 1973/74 9697 35 277 11.1 14.8 .50 .67 750 378 504 103634 274514 1974/75 12104 35 346 13.P 15.3 .63 .69 750 471 520 106371 225733 1975/76 18265 35 522 20.9 16.3 .95 .74 750 711 555 120449 169884 1976/77 12360 35 353 14.1 18.0 .64 .82 700 449 573 100959 224794 1977/78 16712 35 477 19.1 14.9 .87 .68 700 607 473 89619 14758Z 1978/79 9940 35 284 11.4 16.0 .52 .73 700 361 508 89613 240636 1979/80 15273 35 436 17.5 14.5 .79 .66 700 555 4eo 100353 180847 1980/81 12788 35 365 14.6 15.3 .66 .69 650 431 450 91537 212149 1981/82 11980 35 342 13.7 13.8 .62 .63 650 404 407 70790 175131 1982/83 11409 35 326 13.0 15.8 .59 .72 650 385 467 55310 143682 1983/84 18100 35 517 20.7 16.8 .94 .76 650 611 497 47095 77116 1984/85 14691 35 420 16.8 17.7 .76 .80 600 458 483 44929 98194 1985/86 11751 30 392 15.7 15.7 .71 .71 600 427 429 52087 121988 1986/87 11035 30 ,68 14.7 . .67 . 600 401 . 54665 136333 a/ 25 trees per site. bl 3-year moving average centered on middle year. c/ Kilos. assumes the number of pods on the tree in August/September equals the number of nature pods produced by the tree during the crop year. d/ According to COCOBOD data, the bean weight in each mature pod averaves .04542 kilos. ANNEX III GHANA'S COCOA ENVIRODIENT PAGE 6 of 26 Table A-III-4: Brong-Ahafo Hectarage and Yield Estimates, 1969/70-1987/88 Crop Total Total Pods/ Pods/Tree Prod'n/Tree clod/ Trees Yield/Ha (kg) Prod'n Year Pods Sites Site ActualalMov Avebl Actual Mov Ave Ha Actual Mov Ave (M.T.) Hectares 1969170 27576 46 599 24.0 -- 1.09 . 850 926 -- 111787 120752 1970/71 27467 46 597 23.9 24.1 1.08 1.09 850 922 930 108574 117746 1971172 28026 46 609 24.4 23.4 11.1 1.06 850 941 902 115433 122688 1972/73 25139 46 547 21.9 20.3 .99 .92 800 794 738 109230 137517 1973/74 8453 23 368 14.7 17.5 .67 .79 800 534 636 76049 142369 1974175 9156 23 398 15.9 17.4 .72 .79 800 579 633 78985 136512 1975176 12457 23 542 21.7 18.5 .98 .84 800 787 674 85716 108887 1976/77 10363 23 451 18.0 19.1 .82 .87 800 655 693 75879 115869 1977/78 10058 23 437 17.5 15.1 .79 .69 800 636 550 69541 109411 1978179 5688 23 247 9.9 '4.4 .45 .66 800 359 525 50408 140240 1979/80 9158 23 398 15.9 12.5 .72 ^57 800 579 453 75894 129413 1980/81 6641 23 289 11.5 14.1 .52 .64 800 420 512 47598 113419 1981/82 8498 23 369 14.8 13.7 .67 .62 700 470 435 49747 105870 1982/83 8459 23 368 14.7 15.8 .67 .72 700 468 502 35173 75199 1983/84 10294 23 448 17.9 16.6 .81 .75 700 569 528 29756 52103 1984/85 9908 23 431 17.2 17.1 .78 .78 650 509 505 28756 56526 1985/86 8067 20 403 16.1 15.4 .73 .70 650 476 456 35391 74300 1986/87 6473 20 324 12.9 . .59 . 650 382 . 31659 82833 a/ 25 trees per site. bI 3-year moving average centered on middle year. c/ Kilos. assumes the number of pods on the tree in August/September equals the number of nature pods produced by the tree during the crop year. d/ Acc.ording to Cocobod data, the bean weight in each mature pod averages .04542 kilos. * F""sE~~~~~~A III GIIAM'8 COCOA 1=1IOC1 PACE 7 of 26 Table A-III-5: Western Hectarage and Yield Betimtes. 1969170-1987188 Crop Total Total Pods/ Pod ITree yrod'nlTree c/.d/ Trees Yield/Ha (kg) Prod'n Year Pods Sites Site Actuala Mow Aveb Actual Nov Ave Ha Actual Mov Ave (M.T.) Hectares 1969/70 5540 12 462 18.5 -- .84 . 950 797 -- 30141 37779. 1970/71 8282 12 690 27.6 20.8 1.25 .95 950 1191 899 35319 29650 1971/72 4928 12 411 16.4 20.9 .75 .92 950 709 871 49302 69501 1972/73 4960 12 413 16.5 15.4 .75 .70 900 676 628 41781 61821 1973/74 1967 6 328 13.1 13.4 .60 .61 900 536 548 40046 74707 1974/75 1586 6 264 10.6 10.8 .48 .49 900 432 442 50478 116789 1975/76 1309 6 218 8.7 9.4 .40 .43 900 357 385 43272 121301 1976/77 1342 6 224 8.9 9.0 .41 .41 850 345 348 39083 113151 1977178 1401 6 234 9.3 10.7 .42 .49 850 361 414 41968 116387 1978/79 2078 6 346 13.9 12.0 .63 .54 850 535 463 45873 85770 1979/80 1917 6 320 12.8 12.5 .58 .57 800 493 484 52301 106002 1980/81 1646 6 274 11.0 11.4 .50 .52 8oo 399 413 45148 113230 1981/82 1552 6 259 10.3 10.9 .47 .50 800 376 396 43703 116245 1982183 1707 6 285 11.4 11.6 .52 .53 800 414 422 35509 85873 1983184 1968 6 328 13.1 12.5 .60 .57 800 477 455 40243 84415 1984/85 1954 6 326 13.0 12.3 .59 .56 750 444 418 52489 118284 1985166 5332 20 267 10.7 14.3 .48 .65 750 363 487 59564 163967 1986187 9600 20 480 19.2 . .87 . 750 654 . 72731 96975 a/ 25 trees per site. b/ 3-year moving average centered on middle year. c/ Kilos. assumes the number of pods on the tree in August/September equals the number of nature pods produced by thle tree during the crop year. di According to Cocobod data, the bean weight in each mature pod averages .04542 kilos. ANNEX III GHANA'S COCOA ENVIRONHEENT PAGE 8 of 26 Table A-III-6: Eastern Hectarage and Yield Estimates, 1969170-1987188 Crop Total Total Pods/ Pods/Tree Prod'n/Tree c/,d/ Trees Yield/Ha (kg) Prod'n Year Pods Sites Site Actuala/Mov Avebl Actual Mov Ave Ha Actual Nov Ave (M.T.) Hectares 1969/70 12109 36 336 13.5 . .61 . 800 489 . 67262 137583 1970/71 13966 36 388 15.5 15.4 .70 .70 800 564 559 71557 - 126907 1971/72 15491 36 430 17.2 17.6 .78 .80 800 625 638 83313 133211 1972173 17986 36 500 20.0 16.4 .91 .74 750 681 559 72248 106127. 1973/74 5395 18 300 12.2 15.4 .54 .70 750 408 523 63566 155646 1974175 6345 18 353 14.1 13.3 .64 .60 750 480 454 71100 148028 1975176 6239 18 347 13.9 13.1 .63 .61 750 472 455 67038 141943 1976177 5436 18 302 12.1 j .55 .59 700 384 413 51781 134823 1977178 5844 18 325 13.0 12.7 .59 .58 700 413 403 41289 99998 1978179 5848 18 325 13.0 13.7 .59 .62 700 413 435 50200 121496 1979180 6786 18 377 15.1 18.5 .68 .84 700 479 587 45051 93963 1980181 12298 18 683 27.3 18.3 1.24 .83 650 807 540 46632 57797 1981/82 5617 18 312 12.5 18.4 .57 .84 650 369 544 36890 100105 1982/83 6972 18 387 15.5 13.3 .70 .61 650 457 393 31254 68328 1983/84 5404 18 300 12.0 13.5 .55 .61 650 355 397 25525 171925 1984/85 5789 18 322 12.9 13.5 .58 .61 600 351 368 28540 81408 1985/86 5848 1S 390 15.6 13.1 .71 .59 600 425 357 31764 74741 1986187 4053 15 270 10.8 . .49 . 600 295 . 31713 107670 a/ 25 trees per site. b/ 3-year moving average centered on middle year. cl Kilos, assumes the number of pods on the tree in August/September equals the number of nature pods produced by the tree during the crop year. d/ According to Cocobod data, the bean weight in each mature pod averages .04542 kilos. AlMI III GHAA'S COCOA EUIVIOUT PAGE 9 of 26 Table A-III-7t Central Hectarage and Yield Estietea, 1969170-1987188 Crop Total Total Pods/ Pods/Tree Prod'n/Tree cI,di Trees Yield/Ha (kg) Prod'n Year Pods Sites Site Actuala/Hov Avebl Actual Hov Ave Ha Actual Mow Ave (M.T.) Hectares 1969/70 10554 25 422 16.9 . .77 . 600 614 . 53510 87209 1970/71 11513 25 461 18.4 17.4 .84 .79 800 669 632 57944 86569 1971/72 10547 25 422 16.9 19.9 .77 .90 800 613 722 56156 91583 1972/73 15213 25 609 24.3 18.7 1.11 85 750 829 636 42110 50786 1973/74 4802 13 369 14.8 18.1 .67 .82 750 503 617 46217 91823 1974/75 4951 13 381 15.2 14.6 .69 .66 750 519 497 49180 94770 1975/76 4465 13 343 1-1.7 13.2 .62 .60 750 468 451 58202 124375 1976/77 3494 13 269 10.8 10.6 .49 .48 700 342 337 37343 109249 1977/78 2379 13 183 7.3 10.1 .33 .46 700 233 321 21553 92609 1978/79 3984 13 306 12.3 9.2 .56 .42 700 390 292 25700 65940 1979/80 2590 13 199 8.0 11.9 .36 .54 700 253 378 19034 75122 1980/81 5013 13 386 15.4 12.9 .70 .59 650 455 380 25563 56135 1981/82 4963 13 382 15.3 14.7 .69 .67 650 451 435 22069 48951 1982/83 4391 13 338 13.5 12.4 .61 .56 650 399 367 17604 44134 1983/84 2763 13 213 8.55 12.6 .39 .57 650 251 373 13702 54592 1984/85 5178 13 398 15.9 15.2 .72 .69 600 434 415 19071 43923 1985/86 5325 10 533 21.3 18.2 .97 .83 600 580 496 25666 44216 1986/87 4345 10 435 17.4 . .79 . 600 474 . 75681 54221 a/ 25 trees per site. b/ 3-year moving average centered on middle year. c/ Kilos, assumes the number of pods on the tree in August/September equals the number of nature pods produced by the tree during the crop year. d/ According to Cocobod data, the bean weight in each mature pod averages .04542 kilos. ANNEX III GHANA'S COCOA IWlIROHNT PAGE 10 of 26 Table A-III-8t Volta Hectarage and Yield Estites. 1969/70-1987/88 Crop Total Total Pods/ Pods/Tree Prod'n/Tree cOMd/ Trees Yield/Ha (kg) Prod'n Year Pods Sites SiteActuala/Hov Ave b/Actual Hov Ave Ha Actual Nov Ave (M.T.) Hectares 1969/70 2859 10 286 11 . .52 . 900 467 20878 44660 1970/71 3553 10 355 14 10.7 .65 .46 900 581 437 15348 26418 1971/72 1599 10 160 6 10.2 .29 .34 900 261 408 10289 39353 1972/73 2476 10 248 10 7.6 .45 .38 850 382 297 22188 58030 1973174 800 5 160 6 8.4 .29 .37 850 247 324 14469 58640 1974175 1106 5 221 9 8.0 .40 .38 850 342 311 14010 41012 1975176 1112 5 222 9 8.3 .40 .53 850 343 316 13622 39661 1976/77 908 5 182 7 11.6 .33 .48 800 264 429 9228 34963 1977/78 2336 5 467 19 10.6 .85 .49 800 679 385 7369 10852 1978179 734 5 147 6 10.8 .27 .35 800 213 392 5980 28027 1979/80 978 5 196 8 7.7 .36 .36 800 284 272 4776 16800 1980/81 1165 5 233 9 7.9 .42 .40 750 317 275 1496 5712 1981/82 814 5 163 7 8.7 .30 .46 750 222 297 1683 7587 1982/83 1291 5 258 10 10.1 .47 .51 750 352 344 3776 10733 1983/84 1681 5 336 13 11.3 .61 .42 750 458 378 2656 5798 1984/85 1273 5 255 10 9.4 .46 .35 700 324 308 1028 3175 1985/86 554 5 111 4 7.7 .20 700 141 245 855 6068 1986/87 1059 5 212 8 . .38 . 700 269 . 1806 . 6705 a/ 25 trees per site. b/ 3-year moving average centered on middle year. c/ Kilos. assumes the number of pods on the tree in August/September equals the number of nature pods produced by the tree during the crop year. di According to Cocobod data, the bean weight in each mature pod averages .04542 kilos. ANNEX IlIl GHANA'S COCOA ENVRON1ENT PAGE 11 OF 26 1. Total Hectarage, 1976-87 The hectarage estimates derived Indicate that hectarage in 1976 had fallen to 753,000. output was 394,000 and yields were In the 500-525 kilo range. By 1980, hectarage had fallen to 585,000 and output per hectare was still approximately 500. By 1985, however, hectarage yields were down to 435 kilos and output had fallen to 175,000. With improved price Incentives during 1986 and 1987, pod count data indicates that a small amount of hectarage prices Is an increase In yields -- from 435 kilos back to the 500-550 range. 2. RegbInal Hectarage and Yield Changes. 1976-87 The underlying changes In regional hectarage and ylelds are consistent with what' one knows about the cocoa area In Ghana and supports the total estimates as well. For example, large declines In hectarage have occurred In Ashantl, Brong-Ahafo, Eastern and Central regions. According to the pod count Information, A major decline in yields has also occurred In Brong-Ahafo where yields have fallen from 650 kilos per hectare to 400 -- a drop of akmost 40 percent. Yields on Eastern farms have also declined, but yields in Ashanti and Central appear to have recovered to the 5O kilo norm. The only reglon where the data suggests that hectarage hAs increased is Western. The data suggests that the old Western farms produced 400 kilos per hectare on 110,000 hectares in 1976. Hectarage In this region fell to 95,000 by 1980 but has increased to 1 20,000 In 1987. On the other hand, yields steadily Increased in Western during the last decade which Is consistent with new trees coming on stream. Yields improved from 400 kilos In 1976 to 500 by 1980 and to 500 by 1987. The fIgures In Table 3.7 are not meant to be exAct but are rounded using the pod count and production data. Nevertheless, the author believes the figure are approximately correct and can serve as a guide to Ghana's cocoa area and yields. . ANNEX Ill OHANA'S COCOA ENVIRONMENT PAGE 12 OF 26 . implications of Hectarage and Yield EstiAtes Glven the above rrea and yield estkiates, It Is apparent that the short-run price effects of knproved real farmer priceu have already run their course. It Is highly unlikely that Ghana has one millon hectares under cocoa cultivation with ylelds near 200 kilos which Is the assumptlon that many analysts have made. If the lower yields prevailed, higher farmer prices could lead to a doublhg of cocoa output In a relatively short time (three years). Better farm maintenance would result In yields rising from 2000 to 300 to 400 kilos per hectare and output would double. Given Ghana's hectarage In the 400,000-500,000 range and yields between 500 and 550 kilos per hectare, the short-run upside potential given good maintenance and weather' Is probably 250,000-275,000 tonnes. Production above 275,000 will require additional hectarage producing cocoa. The now hectarage planted In the 1985-88 period should begin to bear fruit from 1989 onwards. t is the evidence of new plantings which gives hope that Ghana's recent farmer pricing policies will lead to Ghanalan production above 300,000 tonnes in the early 1990s. In order to better grasp the knpact of Ghana's current farmer price policies, the next section examines the cost of production by type of activity In Ghana. COCOA PROOUCTION COSTS, 1979-88 In order to understand the Inpact of cocoa prices received by the farmer on planting, maintenance and harvesling decisions, It Is necessary to know the costs of production for each activity. This section examines Ghana's cocoa production costs and compares the results with other maJor producers. Ke Economc Data Table A-11l-9 contains the key economic data required to estkiate cocoa production costs. Labor rates per manday and Inflation rates provide the basics. In additlon, exchange rates allow the costs to be estimates In dollars as well as In local currencies. ANNEX IlIl GH.NA'S COCOA ENVIRONMENT PAGE 13 OF 26 Mandays Per Task Both labor and nonlabor costs per type of activity by year were obtained from Ghana's Cocoa Board for 1988. Three members of the Ghana Cocoa Board worked with the author during June, 1988. Thelr knowledge of current wage rates In cocoa, prices received for alternative food crops, yields for those alternatives plus planting, maintenance and harvesting practices In cocoa were extremely valuable In calculating the Ghanaian costs of production. They also helped determine the appropriate inflation and discount rates from the cocoa farmers' polnt of view. Wages were set at 300 cedis per day plus food (worth another 100 cedis) for the establishment and maintenance functions and 350 cedis per day plus food for harvesting. The number of mandays was also modlifed by personnel from the Ghana Cocoa Board In order to reflect actual practice of the cocoa farmer In establishing a cocoa farm. In past analyses, the author and others have used an Oldeai" set of mandays for estimating cocoa production costs. For example, the "Ideal" versus the "actual" mandays are as follows: No. of Mandays Year kJeal Actual 1 159 64 2 85 37 3 51 32 4 54 30 5 54 32 6 59 32 7 64 35 8-25 64 40 26-30 54 30 31-35 49 25-27 U . ANNEX III GHANA'S COCOA ENVIRONENT PAGE 14 OF 26 Table A-II-9- Economic Data for Major Cocoa Produchg Countries January, 1979 to January, 1988 cote Ghana Grvoire Brazil Malaysia Exchange Rate (Cur S) Je nuary 1979 10 225 024 2.21 January 1983 25 336 .275 2.32 Jaruary 1966 100 380 10.4 2.45 January 1966 180-275 280 77 0 2.55 Labor Rates (Per Manday) January 1979 10 475 .05 7.48 January 1933 40 750 7 9555 January 1986 140 817 30 10.03 January 1968 350 950 150 10.25 Inflation Rates (%) 1979 (Ave. of 1978480) 87 16 58 4 1983 (Ave. of 190-85) 56 6 249 5 1986 25 7 145 0 1987 45 5 365 1 Inflation Rate (USS) 1979.83 73 -8 -6 5 1963-8 .290 4 .14 .1 198648 0 23 24 -1 Fertilizer No No Yes yes Spraying Yes Yes Yes Yes InWerop Pntain Plantain Baaa & Cocoyam & Bananas Soue Exchange Rates; IMF, International Financial Statistics, August 198, except for 1979. 83 Ghanaian rates wh ch ar unofficial rnarket rates, The 1988 Ghanaian rate is officially 180 but conversations with Ghanaian offials indicat a black market rate of 275 would be more appropriate. Labor Rates: Brazil wage data obtained form sources in Bahia Cote D'tvoire and Ghanaian data obtWned from World Bank and ILO reports. Malaysian data obtained from ILO data for 1979.86 and estmated for 1988. Ferblizer & S W;; Aft: Brazilian data trom CEPLAC protect dated March 1979, January 1963 and Noverter 1965. 1988 Brazilian data estimated. Cote dvoimr data from February 1978 survey plus various SATMACI repors. Ganaian data trom May 1979 survy, Pre-fasbity Study for the Rehablitation of the Cocoa Industry in Ghana', Ghana Cocoa Maketing Poard, October 1981 pkus adjustTent for infation through 1985. The 1987 data basd on inormation provided by tw Ghana COCOBOD October 1987, and June 1988. Infation Rats IMF, Intrational Financial StatisIcs, November 1987. According to the Ghanalan officials, the NactualN number of mandays spent in establishing, maintaining and harvesting are considerably less than the ldealN. Glven the iow farmer prices of the past, the farmer could not afford the Nldeal. If he were to remain In cocoa farming, he could not afford to perform the tasks in the manner recommended by the extnslon service. It Is obvious that yields could be improved If the ANNEX Ill GHANA'S COCOA ENVIRONMENT PAGE 15 OF 26 farmer were to pay more attention to his cocoa farm. However, cocor revenues at the farm level did not cover the marginal costs of various functlons recommended by the government. Nonlabor Costs Nonlabor costs vary with the year and the activity. Land rent, seedlings, cocoyam and plantain suckers, matchets, axes, buckets, forks, hoes, steel files bags, etc., are the major items Included. The costs and quantitles of these Items were estkiated by the Ghanaian officials for 1988. Revenue from Food Crops Food crops are planted with newly plantod cocoa for two reasons. One reason is to provide shade for the young cocoa and the other Is to generate revenue for the farmer. Cocoyam and plantain prices for early 1988 were used to estknate the revenues derived from the food crops. The cassava price was 1,500 cedis per bag and plantain prices were set at 250 cedis per bunch. One a new cocoa farm, cassava yields were set at 25 bags per hectare n year 2 and zero in all other years. Plantain output was set at 250 bunchss per year for years 1 through 3. Inflatlon and Discount Rates The current agricultural rate of Interest In Ghana Is In the 20-30 percent range. Inflatlon has varied between 20 and 45 percont durhig the past three years. The Ghanalans suggested an inflation rate of 20 percont and a discount rate of 26 percent as representative rates for the kmng run. Given the long gestation proid required to bring cocoa tress Into fuU production (eight years) the costs of production are significantly affected by the particular rates choson. The higher the inflation rate, the higher the costs. The higher the discount rate, the iower the costs. . ANNX Ill GHANA'S COCOA ENVRONMENT PAGE 16 OF 26 However, rates chosen should be related. One would expect the discount rate to exceed the Inflation rate In order for the real cost of money to be positive. In most developed countries, real rates of Interest are In the 3-5 percent range. For example, the current real cost of money In the U.S. Is 5.5 percent (I.e., the Inflaton rate Is approxknately five percent and the prkle lendIng rate Is 10.5 percent). Consequently, a real rate of Interest for the Ghanaian farmer of six pereont Is not out of line with current rates in other countrls. Ghana's Productlon Costs In Cedis, 1988 (Tables 10, 12, 13) Table 10, summarIzes Ghana's costs of producing cocoa In cedhs per kilo given the assumptions outlined above and assuming peak production yields In the 500-600 kilo range. Tables 12 and 13 present the cost details for a new cocoa farm and A cocoa farm established on an old food farm. The data In the first cokumn of Table 10 gives the maximum yieid fgures. The next four columns present the estktated costs of production for establishment, maintenance, harvest and the total. If one assumes 500 klos per hectare on a new farm, total costs that the farmer must recover are 161 cedis per klo. Establlshment costs total 34 cedls, maintenance costs are another 58 coels per kilo and harvesting costs 69 cedis per kilo. In order to determine how the figwes are wrrived at, an examination of the analysis In Tables 12 and 13 follow: (a) New Farm Costs (Table 12) At 500 kilos per hectare, the lng-run production costs to establish a now farm In 1988 totalled 161 cedbs per klo or 161,000 eodt per tonne. This assumes that 41 . . ANNEX III GNANA'S COCOA ENVIRONMENT PAGE 17 OF 26 Table A-III-10. Ghana Cocoa Production Costs by Typ of Activity, 1988 (Cs per Klb) (inflation - 20X; DUsount Rate - 26X) Maximum Yield Per Heca. (101S) Establsh Mahinkin Harves Tota New Frm 500 34 So as 161 550 30 53 63 146 600 28 48 58 134 Pladnt in Food Farm 500 24 60 73 156 550 20 50 61 130 600 17 43 52 112 Table A-1I-1l: Cocoa Fam Production Costs By Type of Actlvfty, 1979-88 Wb) fypa Of CoC GHANA BRAZIL d oire Now Farm ReplaFt Nw Farm Replant Malaysia Estabshment .10 .06 N/A .08 .04 Mainea .09 .13 N/A .24 .24 Hwvesn 1s .21 N/A .27 .27 Total Farm Cost .37 .42 N/A .59 .55 .39 -New Farm- Janary 1963 Jan. '83 May '83 Establishment .14 .09 .06 .11 .07 Makntnanc .09 .13 .13 .21 .17 Harsn .18 .22 .24 .19 Totl Farm Cost .41 .44 .41 .se .43 44 Jany 196 EstablishMe .13 .06 N/A .08 N/A M .09 12 N/A .20 N/A Har u .16 .22 N/A .22 N/A Totl Farm Cost .40 .42 N/A .50 NUA .43 anwy 1966 Eatabblwent .17 .11 .A7* .07 N/A MaIntnance .12 .19 .18 .23 N/A Haim " .2O 23 .22 .24 N/A Tol Fam Cot .57 .53 .47 .53 NJA .42 *apiwady mm ~i a _m Pi *. . _ u-e mm in a *M low MM Coa of pmweoC E d an - amiad f_g Ghltm AM in ftt Cas dl,ra CEPLAC in bua ad Mr=- in "" "Mu am nm k*1" a" WE" OU Rn*6 ftpg5g &W' dagug Afwiy on lil bm o i moAr fh to =wqbtf Urn 11 Pigo- m ba ytrsa u hM oo baum md. Tha Wm". of mmo Wltl prnt h b* _ _ d M a Wnn=. =j go oMa d 5d Cab &W & An wA~ lof 7aa a d=wd id 'a as n * _ e wd M a,Wig V. Gwi eons G An Au a1 h emSo* Wu pw dolw km ba um md in afmau Vn SfwiimW aoe 0 ANNEX IlIl GHANA'S COCOA ENVIRONMENT PAGE 18 OF 26 mandays are required in the first year to clear the land, develop the nursery and plant the seedlings. Approxbnately 18 mandays are required to weed and maintaln the cocoa and food crops and five days are required for harvesting during the first year. In the second year It takes five days to coonplete the estabilshment, 22 days for maintenance and 10 days to harvest. The revenue generated the food crops Is allocated to the various functions (establishment, maintenance, harvesting and nonlabor) on the basis of the number of days devoted to each food crop function. During the first year, the value of the food crop exceeds the costs Incurred In planting cocoa and grown food by 6,275 cedis. In the second year, not revenues total 178,475 cedis. Food revenues exceed total costs during the first three years after which food planting on the land ceases as the cocoa canopy begins to close In and food crops no longer be planted. Given an inflation rate of 20 percent, the cost figures are inflated to obtain a nominal estimate of costs for future years. The Inflated figures are presented In the last four columns of Table 12. In order to estknate the total cost per kilo over the lifetIme of a cocoa hectare, the net present value, of the Inflated cocoa cost stream Is determined using a 26 percent discount factor. The net present value of total costs Is then divided by the net present value of the yield stream (at SOO kIlos or 550 kilos or 600 kilos). This procedure is followed not only for the total cost stream but also for the establishment, maintenance and harvesting costs. The results are presented at the bottom of Table 12. At 500 kilos per hectare, establsent costs per kIlo total 34 cedis (approximately 21 percent of the total costs), maintenance costs total 58 cedis (36 percent) and the remainhg 69 cedis per kilo Is spent harvesthg. If maxhnum yields, are 550 kilos per hectare, total costs fall to 134 cedis. These costs figures are In line with the current farmer price of 165 cedis per kilo. In real terms, the farmer price has been near this 0 ;5 liPMlIX III UHMA'S COCOA 63W1UWW Pag 19 of 26 Tabl- A-I11-13: Obee Cost of Productlon N.o Cocoa Form (20 1 Irf lotlon; 26v Discount Rate) Labor (Nandays); Cost (Cedi) N of Max Yield Inter Estab Maint Hrust Estab Maint Hrust Non- Total Infitn Infltd Infltd Infitd Inf ltd Far. Max 050 0550 0600 Crop Labor Labor Labor Cost Coat Cost Labor Cost 020X Estab Maint Hrust Total Age Yield kg kg kg Incoma Cost Coat Cost Cost cost 1 .00 0 0 0 62500 41 16 6 32734 -12862 -26027 30876 -6275 1.00 32734 -12362 -26627 -4475 2 .30 0 0 0 200000 6 22 10 -7159 -77499 -93613 6226 -178476 1.20 -8591 -9299 -112581 -214170 8 .00 0 0 0 62600 22 10 0 -16877 -24896 6926 -43276 1.44 0 -26402 -36864 -42316 4 .08 42 46 60 0 20 12 0 14206 7602 9307 21807 1.73 0 24546 13137 87662 6 .U3 164 182 196 0 20 12 0 18409 664 9665 22066 2.07 0 27806 17926 46733 6 .67 888 867 440 0 20 12 0 18309 S5W 8496 21696 2.49 0 33118 21364 54462 7 .98 417 46 500 0 20 15 0 13866 10999 9915 246S6 2.99 0 40806 32844 78649 a 1.00 500 660 600 0 15 26 0 9009 16266 6025 26275 3.68 0 32282 68283 9s65 9 1.00 500 550 600 0 15 25 0 9009 16266 6025 25275 4.80 0 38739 69939 106678 10 1.00 500 550 600 0 1s 25 0 9009 16266 6026 25275 5.16 0 46466 68927 180413 11 1.00 500 50 600 0 15 26 0 9009 16266 6026 25276 6.19 0 66784 100712 156496 12 1.00 500 560 600 0 15 26 0 9009 16266 6026 26276 7.48 0 66940 120855 167795 18 1.00 500 550 6oo 0 i5 25 0 9009 16266 8025 25275 6.92 0 80326 145026 225254 14 1.00 500 650 600 a 15 26 0 9009 16266 3025 25276 10.70 0 96394 174031 270425 15 1.00 600 560 600 0 15 26 0 9009 16266 6025 25276 12.64 0 116873 206887 825410 16 1.00 50 550 600 0 15 26 0 9009 16266 6025 25275 15.41 0 138806 250606 839412 17 1.00 500 550 600 0 15 25 0 9009 1626 6025 26275 18.49 0 166569 800726 467295 13 1.00 0OO 550 6oo 0 15 25 0 9009 16266 6026 25276 22.19 0 199663 860671 560754 19 1.00 50 550 600 0 15 25 0 9009 16266 6025 25275 28.62 0 239860 438045 672905 20 1.00 500 650 600 0 16 25 0 9009 16266 3026 26276 31.96 0 267632 619654 607496 21 1.00 500 550 60 0 15 26 0 9009 16266 6025 25276 38.34 0 345396 w 23686 956968 22 1.00 500 560 600 0 15 25 0 9009 16266 6025 25275 46.01 0 414477 748302 1162779 23 1.00 50 55O 6000 0 15 25 0 9009 16266 6025 25276 66.21 0 497373 897962 1895855 24 1.00 50 560 600 0 15 25 0 9009 16266 6025 26276 66.25 0 596647 1077666 1674402 25 1.00 600 550 600 0 15 26 0 9009 16266 3025 25275 79.50 0 716217 1298066 2009268 26 .3U 417 453 500 0 15 15 0 10018 10763 9025 20775 95.40 0 965166 1026702 1961656 27 .3U 417 456 600 0 16 16 0 10018 10768 9025 20775 114.48 0 1146166 1232042 237322 26 .83 417 46 5W00 0 16 16 0 10018 10768 6026 20775 137.37 0 1376423 1473461 2658678 29 .67 83U 867 400 0 16 12 0 10465 3967 6025 19425 164.34 a 724000 1473107 8202106 U0 .67 U8 867 400 0 15 12 0 10453 3967 s026 19425 197.81 0 206880 173729 3842629 31 .60 250 276 800 0 16 10 0 10816 7710 9026 18626 237.88 0 2667226 1630171 4387395 82 .60 250 276 300 0 16 10 0 10916 7710 8026 18626 234.86 0 3809670 2196206 5276767 88 .50 250 275 800 0 15 10 ( 10315 7710 6025 16525 841.62 0 8696304 2685447 6382251 34 .60 250 275 800 0 16 10 0 10916 7710 9026 16625 410.19 0 448616 38126326 759701 86 .60 250 275 800 0 15 10 0 10161 7710 9025 13525 492.22 0 5323397 8795043 9116441 Estab Maint Hrust Total Cost Cost Cost Cost Yield nNVP) (WPV) (NPV) fNPVi 550 84 58 69 161 550 80 58 68 146 600 28 46 53 184 AFPPIX III SIU'S COCOA BW1U111T Pag 20t f 29 Table A-III-14: Ghaa Cost of Produactio Mm Cocoa F retmN Feed Fare (20 1 lnf lotimo; 260 hloe*t Rnae) Labor (sWd ye); Coot (Cedla) s of Max Yield Inter Inter MLint Hrust Estab Maint Hrust Non- Total Infitn Infiltd Infltd Infltd Infltd Fare Max 0600 0660 o060 Crop Labor ILabor Labor Cost Cost Coat Labor Cost 0201 Estab alint Hrust Total Age Yield kg kg kg Income Cost Cost Cost Cost Cost 1 .00 0 0 0 162500 35 16 5 19349 -54763 -72147 21500 -107650 1.00 19439 -54753 -72147 -107550 2 .00 0 0 0 120000 5 22 10 -3176 -41673 -53851 6100 -99660 1.20 -3611 -49888 -s4422 -110820 8 .00 0 0 0 50000 22 10 0 -12213 -1 866 5600 -30900 1.44 0 -17686 -26910 -44496 4 .06 42 W0 5u 0 20 10 0 14121 7561 9162 21682 1.78 0 24402 13065 87486 S .25 125 150 175 0 20 12 0 18321 6599 0630 21930 2.07 0 27644 17830 45474 6 .60 250 S00 85o 0 20 12 0 18231 0659 6370 21770 2.49 0 32924 21247 54171 7 .07 388 400 467 0 20 15 0 13894 10946 9790 24540 2.99 0 40592 82664 78276 6 .8 417 500 568 0 15 25 0 6963 16168 7900 25150 3.56 0 32114 58003 90117 9 1.00 oo0 600 700 0 15 25 0 6963 16166 7900 25150 4.80 0 38637 89603 106140 10 1.00 500 600 700 0 15 25 0 6963 16166 7900 251W0 5.16 0 46245 36324 129766 11 1.00 5o0 600 700 0 15 25 0 9713 18166 7900 25150 6.19 0 60137 100229 155722 12 1.00 500 600 700 0 15 25 0 6963 16166 7900 25150 7.43 0 66592 120274 196687 18 1.00 500 600 700 0 15 25 0 6963 16166 7900 15150 8.92 0 79911 1443a 224240 14 1.00 500 600 700 0 15 25 0 3963 16168 7900 25150 1C.70 0 95893 113195 269066 15 1.00 500 600 700 0 15 25 0 n963 16168 7900 26150 12.64 0 115071 207834 322906 16 1.00 0w0 600 700 0 1s 25 0 6963 16186 7900 25150 15.41 0 138065 249401 867467 17 1.00 500 600 700 0 15 25 0 6N93 161e 7900 25150 18.49 0 165703 299281 464964 1 1.00 W00 600 700 0 15 26 0 8963 16168 7900 25150 22.19 0 198643 369138 657961 19 1.00 500 600 700 0 15 26 0 893 16166 7900 26150 26.62 0 238612 430966 669577 20 1.00 500 600 700 0 15 25 0 8963 16166 7900 25150 31.96 0 286334 617168 603492 21 1.00 600 6o0 700 0 15 26 0 6963 16168 7900 25150 36.34 0 343601 620690 964191 22 1.00 oo0 600 700 0 15 25 0 8968 16188 7900 25150 46.01 0 412321 744706 1157029 23 1.00 500 600 700 0 1s 25 0 8f96 16186 7900 25150 65.21 0 494786 893649 1886435 24 1.00 oo0 600 700 0 15 25 0 896sa 1618 7900 25160 66.25 0 693742 1072379 1666121 25 1.00 50o 60 700 0 15 25 0 8983 16188 7900 25160 79.60 0 712490 1286655 1999346 26 .68 417 600 68a 0 15 15 0 9950 10700 7900 206e0 95.40 0 949192 1020740 1969932 27 .63 417 500 568 0 15 15 0 9950 107000 7900 20860 114.48 0 1139031 1224887 2863916 26 .83 417 500 668 0 15 15 0 99s0 10700 7900 20860 187.87 0 1366637 1469066 2036702 29 .67 838 400 467 0 15 12 0 10389 8911 7900 19300 164.64 0 1712563 1468949 3181502 8o .67 838 400 467 0 15 12 0 10369 6911 7900 19300 197.61 0 2065063 1762739 3617602 31 .o0 250 8W 850 0 15 10 0 10740 766o 7900 16400 287.86 0 2549422 1618303 4867724 82 .50 250 80 850 0 15 10 0 10740 7660 7900 16400 264.85 0 3069306 2161953 5241269 38 .60 250 300 360 0 15 10 0 10740 7660 7900 16400 341.62 0 3871167 2613636 6269623 4 .59 2w0 800 860 0 16 10 0 10740 7660 7900 16400 410.19 0 4405401 3142027 7547427 35 .60 250 800 850 0 1s 10 0 10740 7660 7900 16400 492.22 0 5286461 3770432 9056913 Estab Maint Hrust Total Cost Coat Cost Cost Yi-ld (NVP) (WV) ()l (WV) 500 24 60 78 156 600 20 c0 61 1SO 700 17 43 52 112 APP9IX III awu4Ms COCOA EWUIIIT Pegs n f tS Table A-I11-1l: Broil I Cost of Producties Na Cocoa Fore (7% Imflatiso; 1lC Disecont Rate) Labor (Needay.); Cost (Dollors) Non- Infltd InfItd InfItd InftItd IntfItd Inter Intor Total Labor Total Estab Maint Haruat Non- Fare Bax Crop Crop Estab Maint Harust Cost Cost InfItn Cost Cost Cost Cost Labor Age Yield 1000 1200 Yield Incom Labor Labor Labor (USS) (USP) 07% US$) (USS) (US$) (USC) (USS) 1 .00 0 0 0 .00 264 726.71 560 1.00 1286.71 1146.71 140.00 .00 6e0.00 2 .00 0 0 aso 104.S6 30 60 267.14 90 1.07 268.58 163.94 207.50 .00 06.80 a .13 126 150 700 224.40 18 60 14 262.86 180 1.14 225.36 170.61 233.48 46.30 143.34 4 .26 250 800 700 240.11 3 60 26 264.29 130 1.23 230.66 129.94 249.82 91.00 159.20 5 .50 500 600 700 268.92 40 48 261.43 130 1.31 243.06 .00 235.01 284.97 170.00 6 .68 625 700 660 219.92 40 66 271.43 130 1.40 343.10 .00 261.48 311.67 132.a3 7 .76 760 900 420 176.49 40 60 286.71 13O 1.60 447.39 .00 269.06 354.62 195.09 a 1.00 1000 1200 290 126.89 40 72 320.00 130 1.61 596.71 .00 287.89 434.71 200.76 9 1.00 1000 1200 140 67.36 40 72 320.00 130 1.72 706.33 .00 308.06 466.14 223.76 10 1.00 1000 1200 0 .00 40 72 320.00 130 1.84 827.31 .00 329.61 497.70 289.00 11 1.00 1000 1200 0 .00 40 72 320.00 130 1.97 886.22 .00 362.8s 632.54 255.78 i2 1.00 1000 1200 0 .00 40 72 320.00 130 2.10 947.18 .00 377.37 569.61 278.6 18 1.00 1000 1200 0 .00 40 72 320.00 130 2.26 1013.49 .00 403.79 609.70 292.76 14 1.00 1000 1200 0 .00 40 72 320.00 130 2.41 1004.43 .00 462.06 662.36 813.23 15 1.00 1000 1200 0 .00 40 72 320.00 130 2.60 1160.34 .00 462.29 90.056 385.21 10 1.00 1000 1200 0 .00 40 72 320.00 130 2.76 1241.56 .00 494.65 746.91 356.67 17 1.00 1000 1200 0 .00 40 72 320.00 130 2.95 1328.47 .00 629.28 799.19 Su8.73 18 1.00 1000 1200 0 .00 40 72 320.00 13O 3.16 1421.47 .00 668.33 866.14 410.66 19 1.00 1000 1200 0 .00 40 72 320.00 130 8.38 1620.97 .00 605.97 915.00 4u9.89 20 1.oo 1ooo 1200 0 .00 40 72 820.00 13t0 3.62 1627.44 .00 646.39 979.06 470.15 21 .93 980 1116 0 .00 40 72 320.00 1SO 8.87 1741.38 .00 698.78 1047.6 506.06 22 .87 670 1044 0 .00 40 72 820.00 130 4.14 1863.25 .00 742.34 1120.91 86.2t7 23 .0 O60 960 0 .00 40 72 820.00 13O 4.48 1998.66 .00 794.84 1120.91 530.27 24 .75 750 900 0 .00 40 72 320.00 1SO 4.74 2133.24 .00 349.91 1263.33 616.27 2n .70 700 40 0 .00 40 72 320.00 13O 5.07 2262.57 .00 909.40 1873.16 659.41 26 eso 730 0 .00 40 72 320.00 130 5.48 2442.a4 .00 973.06 1489.23 706.57 27 .60 600 720 0 .00 40 72 320.00 130 5.U1 2613.31 .00 1041.18 1672.13 754.96 26 .55 550 660 0 .00 40 72 320.00 130 6.21 2790.24 .00 1141.06 le2.13 807.60 29 .50 500 C30 0 .00 40 72 320.00 18O 6.66 2991.9 .00 1192.04 1799.94 664. 6 so .45 450 540 0 .00 40 72 320.00 18O 7.11 3201.42 .00 1275.46 1926.98 024.3 Estab Maint Hrust Total Cost Cost Cost Coot Yield on LP (hV) ! C). 1000 .06 .27 .29 .64 1200 .07 .23 .24 .53 IPDIN III GNA'S coCa 8wUUmI Tabl A-l1-10: Cote D'I,eir Cost of Production NOm Cocoa Form (73 1.4 -Ton; 153 Biscoont Rate) Labor (enadeys); Coot (Dollars) Non- Infltd InfItd Infltd Inf ltd Infltd Inter Inter Total Labor Totel Estab Maint Harust Non- Form Max Crop Crop Estab Maint Harust Cost Cost lnfItn Cost Cost Cost Cost Labor Age Yield 800 1000 Yield Income Labor Labor Labor (USS) (USi) 07% US2) (USS) (USS) (USS) (USN) 1 .00 0 0 0 .00 a66 1241.79 569.65 1.00 1851.84 1685.95 147.39 .00 569.55 2 .00 0 0 0 58.98 46 24 244.29 157.21 1.07 876.08 300.42 12" .1 .00 108.22 a .15 120 150 10 96.17 45 24 13 278.21 124.24 1.14 864.61 231.49 126.79 50.50 142.25 4 .80 240 800 o8 102.90 15 24 24 213.75 258.75 1.23 280.86 295.82 177.41 99.75 810.64 6 .60 400 500 8o 110.11 24 41 220.54 89.80 1.81 248.06 .00 132.50 206.10 51.52 6 .0 640 6oo 240 94.25 19 65 801.96 89.80 1.40 843.10 .00 141.77 886.66 66.18 7 1.00 00 1000 13O 75.64 19 62 842.66 89.80 1.60 447.89 .00 126.24 447.02 u6.96 6 1.00 0O 1000 120 58.96 19 82 842.66 89.80 1.61 s59.71 .00 185.07 478.81 08.11 9 1.00 00 1000 60 26.67 19 62 842.66 89.80 1.72 705.68 .00 144.58 511.79 67. u 10 1.00 o0 1000 0 .00 19 62 U42.66 89.80 1.64 627.81 .00 154.64 647.61 77.26 1 1 1.00 S0 1000 0 .00 19 62 842.66 89.80 1.97 *86.22 .00 165.47 656.96 77.82 12 1.00 00 1000 0 .00 19 82 842.66 89.3c 2.10 947.19 .00 177.05 626.96 62.78 1s 1.00 600 1000 0 .00 19 82 342.66 89.80 2.26 1018.49 .00 169.46 670.65 6652 14 1.00 100 1000 0 .00 19 82 842.08 89.80 2.41 1064.43 .00 202.71 717.61 94.72 15 1.00 m0o 1000 0 .00 19 62 842.0 89.80 2.56 1160.84 .00 216.90 708.06 101.85 16 1.00 600 1000 0 .00 19 62 842.0 89.80 2.76 1241.56 .00 282.08 621.21 106.44 17 1.00 600 1000 0 .00 19 82 842.0 89.80 2.95 1826.47 .00 246.32 679.86 116.08 1i 1.00 60o 1000 0 .00 19 62 842.66 89.80 8.16 1421.47 .00 265.71 940.90 124.15 19 1.00 S0o 1000 0 .00 19 62 842.66 89.30 8.86 1620.97 .00 264.31 1006.77 182.64 20 1.00 *00 1000 0 .00 19 62 342.66 89.80 8.62 1627.44 .00 54.21 1077.24 142.14 21 .98 744 980 0 .00 19 62 842.66 89.80 8.87 1741.86 .00 326.60 1152.656 12.09 22 .67 696 670 0 .00 19 62 842.66 89.80 4.14 1166.26 .00 38.29 1288.88 162.74 28 .0 640 6oo 0 .00 19 62 342.66 89.80 4.48 1998.66 .00 872.67 1819.67 174.18 24 .76 600 750 0 .00 19 82 842.66 89.80 4.74 2188.24 .00 896.76 1412.04 196.82 25 .70 560 700 0 .00 19 62 842.08 89.80 5.07 2292.57 .00 426.67 1510.69 199.36 26 .66 620 650 0 .00 19 62 842.0 89.80 5.48 2442.84 .00 456.65 1616.65 21 3.2 27 .60 430 6o0 0 .00 19 82 842.0 89.80 5.61 2618.81 .00 488.49 1729.61 226.26 28 .66 440 550 0 .00 19 62 842.0 89.8o0 6.21 2796.24 .00 522.69 1650.00 244.23 29 .60 400 500 0 .00 19 62 842.0e 89.80 6.6 2991.96 .00 559.27 1960.46 281.82 a0 .45 8a0 450 0 .00 19 62 842.0 89.80 7.11 8201.42 .00 596.42 2119.10 279.62 Etab. Maint HMrest Total Cost Cost cost cost Yield gm gm (WV) (W!V) 600 .26 .17 .87 .62 1000 .22 .14 .80 .66 APPENDIX M SANAS COCOA E NVPONMET Page 23 ot 26 level for the past three years. This explains the considerable Inprovement In farm maintenance and husbandry during the past three years and also explains the new plantings which are taking place. At 165 cedis per kilo, the price covers the costs of establishing a new cocoa farm as well as malntaining and harvesting exlstig farms. At higher price levels, the farmer would spray more than he Is currently sprayhg and also maintain the farms at a higher level (e.g., better weeding). Nevertheless, given the current cultural prActices, the farmer Is able to cover the costs of establishhin new farms. Conversatlons with cocoa farmers and observers who trek the cocoa area year after year suggest that the farmer lost faith in the Goverrnent's willigness to maintain his purchasing power given the very low pricos of the late 1970s and early 19809 which In turn caused the farmer to shift from cocoa to food. However, those same farmers and trekkers Indicate that the farmer began planting cocoa again In 1985 because o' the price adjustments that were made which have changed his view of cocoa's profitability. (b) Cocoa Production Costs In an Estabished Food Farm (Table 13) The bottom half of Table 10 contains cost osthnates for planting cocoa In an already established food farm. Establishment costs are lower because clearlng and fellng activities are reduced. Maintenance and harvesting costs are essentially the same as In the "New Farm" case. Table 13 presents the detailed cost analysis for the "Now Cocoa Farm -- Cid Food Farm" case. Only 35 days are requied to establish a cocoa farm In the first year versus 41 days If the land must be cleared. Ghanaian officlals indicated that gross revenues from food crops on an old food farm wIN be highoest In the fIrst year (152,500), drop marginafy in the second (120,000 codb) and then suffer a major drop In the third year (50,000). Consequently, not revenues total 107,550 codbs In year one, fall to 98,600 In year two, total 30,900 in year three and then costs of maintainhg and harvesting APPEX U WHAS COCOA ENVRONENT Paoe 24 of 26 cocoa total 20,000 to 25,000 per year thereafter. Agaln, the last four colmns Indlcate the nominal or Inflated costs of producing cocoa on one hectare assuming an inflation rate of 20 percent. Total cocoa production costs on an old food farm are 158 codls per klos for the SO klbo farm, 130 cedis for the 550 kilo and 112 for the 800 kilo output. Agaln, these costs are below the prIco level currently received by the farmer and support the observations and survey data that cocoa farms are being estabilshed In Ghana at current prico levels. The evidence obtahod during the October, 1987 Intervews with cocoa farmers supports the above statement. In Ashanti, Brong-Ahafo and Eastern regions, farmers were planting cocoa In a corner of their food farm. On the farms vislted, as much as 20 percent of the cocoa appeared to be now cocoa planted since 1985 In the food crop area. PRODUCTION COSTS FOR MAO PRODUCS (US Cents per Pound) Tables A-111-11 and A-111-14/15 contain cost of production esthuates fcr Coto d'lvolre, Brazil and Malaysla plus Ghana measured in US cents per pound. In dollar terms, the costs of productlon among the major producers varles between 42 (Malaysia) and 57 cents (Cote d'lvolre) In 1988 and has been In that range since 1983. (a) Cote dchovro Production Costs Costs of production In Coto d'lvolre were quire stable until 1988 when the CFA/dollar rate dropped to 280 from 380 In 1988 (see Table 9). The declkin In the value of the dollar raised the costs of lvolran production In dollar terms. The Increase reflects the challenge faced by the Ivolrans In trying to pay the farmer In CFA when the forolgn exchange eamgs have dropped preciptously because of the low dollar cocoa price and even lower French franc price. APPENDIX III GHANAS COCOA ENVIRONENT Page 25 of 26 Historically, Cote d'lvoire costs have generally been 10-14 cents for establishment, rine cents for maintenance and 18 cents tor harvesting. Until 1988, total costs at the farmgate were near 40 cents. The dollar decline against CFA francs ra!sed Cote d'lvoire's total costs to 57 cents in 1988 with establishment at 17, maintenance at 12 and harvesting costs soaring to 28 cents. (b) Ghanaian Costs h Dollars Ghana's cost of productlon have been shown for both new farms and cocoa established on an old food farm. The latter Is less costly because there are no old trees to cut down and there are food crops to generate revenue during the first year. The foreign exchange rate used to convert the cedl costs of production Is 275 cedis. Although the official rate versus the dollar has been declining at a rate of two to three percent per quarter during the last three years, Inflation has been running at a two- three percent per month rate. Discussions with various Ghana officials plus other Ghanalan sources, the black market rate during early 1988 was in the 250-300 cedi range. Productlon costs on a new farm were stable In the 42-44 cent range during the late 1970s and early 1980s. In 1988, however, costs are estimated at 53 cents at the farm gate. Establishment cost are 11 cents, maintenance costs approximate 19 cents and harvesting costs are near 23 cents. Planting In an old food farm reduces total costs to 47 cents. The establishment costs are seven cents. Harvesting and maintenance costs are 18 and 22 cents, respectively. (c) Brazil's Phoduction Costs Brazil's production costs have fluctuated In the 50-60 cent range during the past decade. The ke, -ariable has been the Brazilian exchange rate relative to the internal APPENiX M GHANS COODA ENVO#ENTf Pago 26 of 26 Inflation rate. Starting In 1983 with the maxl-devaluation, the Brazilian policy has been one of devaluing the currency at a rate equal to the Inflaton rate. The ahi has been to maintain purchasing power parity. The Initial effect of the devaluatin was to reduce producton costs from 56-50 cents. Internal Inflation outstripped exchange rate adjustments between May, 1983 and January, 1986 but have been quite consistent since. Consequently, Brazil's production costs In 1988 look much Ike those estimated for 1986. Establishment costs are 7-8 cents, maintenance costs are 20-23 cents and harvesting costs are In the 20-22 cent range. (d) Malaysian Costs Detalled information Is not avaliable for Malaysia. The Malaysian costs are based on plantation reports for 1979 and 1983. These costs have boon adjusted for Inflation and exchange rates to obtain 1986 and 1988 estimates. it Is quite possble that the Malaysian estimates understate costs In the cocoa sector In that the Internal Inflation rates do not adequately cover the Increases In the costs of production. Regardless, Malaysian estimates are most likely In the 40-50 cent range. PRE Working Paper Series Contact Til Author for paper WPS429 Ghana's Cocoa Pricing Policy Merrill J. Bateman June 1990 C. Spooner Alexander Meeraus 30464 David M. Newbery Wiliam Asenso Okyere Gerald T. O'Mara WPS430 Rural-Urban Growth Linkages in Peter B. Hazell May 1990 C. Spooner India Steven Haggblade 30464 WPS431 Recent Developments in Marketing Panos Varangis May 1990 D. Gustafson and Pricing bystems for Agricultural TakamasaAkiyama 33714 Export Commodities in Sub-Saharan Elton Thigpen Africa WPS432 Policy Choices in the Nowly Bela Balassa May 1990 N. Campbell Industrializing Countries 33769 WPS433 Education and Development: Wadi D. Haddad Evidence for New Priorities Martin Carnoy with Rosemary Rinaldi and Omporn Regel WPS434 Tax Sensitivity of Foreign Direct Anwar Shah Investment: An Empirical Joel Slemrod Assessment WPS435 Rational Expectations and Boum-Jong Choe Commodi'y Price Forecasts WV"S433 Commodity Price Forecasts and Boum-Jong Choe Future Prices WPS437 Institutional Developmen' Work in Cheryl V,'. Gray the Bank: A Review of 84 Bank S. Khadiagalla Projects Richard J. Moore WPS438 Productivity Effects of Milan Vodopivec Redistribution in a Socialist Economy: The Case of Yugoslavia WPS439 Indicative Planning in Developing Bela Balassa May 1990 N. Campbell Countrios 33769 WPS440 Financial Sector Policy in Thailand: William Easterly A Macroeconomic Perspective Patrick Honohan WPS441 Inefficient Private Renegotiation Kenneth Kletzer of Sovereign Debt WPS442 Indian Women, Health, and Meera Chatterjee Productivity PRE Working Paper Series Contact ,ti Author DMe for pager WPS416 Improving Data on Poverty in the Paul Glewwe May 1990 A. Murphy Third World: The World Banks 33750 Living Standards Measurement Study WPS417 Modeling the Macroeconomic William Easterly May 1990 R. Luz Requirements of Policy Reform E. C. Hwa 34303 Piyabha Kongsamut Jan Zizek WPS418 Does DAvaluation Hurt Private Ajay Chhibber May 1990 M. Colinet Investment? The Indonesian Case Nemat Shafik 33490 WPS419 The Design and Sequencing of Brian Levy May 1990 B. Levy Trade and Investment Policy Reform: 37488 An Institutional Analysis WPS420 Making Bank Irrigation Investments Gerald T. OMara May 1990 C. Spooner More Sustainable 30464 WPS421 Taxation of Financial Intermediation: Christophe Chamley May 1990 W. Pitayato- Measurement Principles and Patrick Honohan nakarn Application to Five African Countries 37666 WPS4?2 Civil Service Reform and the World Barbara Nunberg May 1990 R. Malcolm Bank John Nellis 37495 WPS423 Relative Price Changes and the M. Shahbaz Khan May 1990 WDR Office Growth of the Public Sector 31393 WPS424 Mexicos External Debt Sweder van Wijnbergen Restructuring in 1989-1990: An Economic Analysis WPS425 The Earmarking of Government William A. McCleary Revenues in Colombia Eva-naria Uribe Tobon WPS426 Growth-Oriented Adjustment Riccardo Faini Programs: A Statistical Analysis Jaime de Melo Abdel Senhadji-Semlali Julie Stanton WPS427 Exchange Reform, Parallel Markets Ajay Chhibber May 1990 M. Colinet and Inflation in Africa: The Case Nemat Shafik 23490 of Ghana WPS428 Perestroyka and Ks Implications Bela Balassa May 1990 N. Campbell for European Socialist Countries 33769