Ghana's cocoa pricing policy

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 '<he quallty of the government's policy toward cocoa. This study is ded&;ated to tS e proposition that better understanding of the principles involved can contribute to the formulation of better cocoa policy. Now that real world cocoa prlcor are approaching the low point of their 25-year cycle, the need for better policy is greater than ever. This report Is divided into revlews of three major substudies, which follow In sequence: 1. Analysis of Cocoa Producer Investment Response to Policy Intervention; I1 Fiscal Alternatives to Cocoa Taxation; and Ill. Giobal Perspective on Cocoa Supply and Demand. The report concludes with a section that Integrates the policy hmpilcations deveioped In the substudles and presents some recomnendations. Detailed analytica! support for tho analytlcal reulits presented In the main report Is presented in Annexes 1, 11, and Ill.  ]. ANALYSIS OF COCOA PRODUCER INVESTMENT RESPONSE TO vRICING POUCY by Alexander Meeraus, W. Asenso Okyere and G.T. O'Mara  A SOME ECONOMIC CONCEPTS CENTRAL TO THE ANALYSIS OF PRODUCER RESPONSE Prior to the discussion of the current status of Ghanalan agriculture and the analysis of cocoa producer response to price Incentives, It will be helpful to Introduce some economic cor,cepts that are kIportant to the analysis of perennial supply response. Opportunity Cot. The productivity of a resource In Its most produc- tivo alternative use. Thus, the opportunity cost of land and labor employed In cocoa prcduction Is their productivity In their socially most valuable alternative use, which typi- cally will be In the production of some other crop. Therefore, the analysis of cocoa producer response to price and other policy Interventions must be In the context of the entire agricultural sector, since other crops and/or regions may have less or more privately financially rewarding uses and either supply resources to or pull them out of cocoa production. Cosrition Dbsoiunt Rate. The rate at which consumers trade off present and future consumptlon. Such tradeoffs are generally accomplished by saving and dissaving, though the savings may be relative to potential income as when the Individual forgoes present income by Investing In training that knproves future produc- tivity. In a fully monetized economy, the saving and dissaving may be accomplished through Investing and disinvesting In financial assets; In a developing economy, it Is of- ten accomplished by real Investment or disinvestment In property, educatlon, productive agricultural assets, etc. The disinvestment may be accomplished through fallure to replace real assets as they wear out. The basis of the trade off Is a preference for present consumption, with a one-period discount rate being the rate at which the discounted value of future consumption Is equivalent to present consumptlon--i.e., the consumer Is indlfferent between them. Thus, the consumption discount rate can be regarded as the required premkAm for deferred consumption. To the extent that Invest- ment hI cocoa production involves present savng for future consumption out of the proceeds from cocoa production, the increased future consumption after allowing for  - 1.2 - future production costs and when discounted to the present at the farmers consump- tion rate of discount must be at least as large as the amount of present savings. In other words, the cocoa Investment must have a positive return when discounted at the consumption rate. Margial Efficiency of CapitaL The return from an additional unit of in- vestment when discounted at the cost of capital confronting the Investor. Also called the private marginal productivity of capital or the private accounting rate of Interest. If the Investor Is a small cocoa farmer without access to at least medium- term credit, then the cost of capital to the farmer Is his consumptlon rate of discount since the Investment must be financed from his private savings. In such cases, the rate of Investment and accumulation Is necessarily slow since It Is constrained by the small farmer's abillty to generate savings. Removing surplus from the return to small farmer cocoa Investments via explicit or knpllclt taxatkin kuplies that expenditure of an additional cedl of revenue by government for either consumption or Investment purposes Is socially more valuable than expenditure of the same cedl by the cocoa farmer. The evidence from the past three decades casts doubt on the validity of this proposition at the levels of taxation of cocoa output that have prevailed. Alternatively, the calculation of net revenue from cocoa taxation should make allowance for the fiscal cost of reduced cocoa export earnings due to foregone cocoa productlon Investments. Risk PrombAm. The additlonal return that rlsk averse Investors require In addition to the expected return from an Investment If they are to undertake a rlsky Investment. In other words, the risk premium Is the reward for bearing risk. The empirical evidence on the behavior of small farmers In developing countries Is quite clear; they are overwhekningly risk averse. Just as clearly, uncertainty with respect to future states of both the domestic and world economies Implles that all market medlated Investments are Inherently rlsky, though some are more risky than others. The existence of cyclical behavior In world cocoa prices makes cocoa production Investments more ris-  - 1.3 - ky than average. Since risk averslon and the Impiled need for a risk premium Imply that risk averse Investors will undertake less investment than Is optimal, a case can be m.ide for subsidy of cocoa Investments In order to achieve a more efficient allocation of resources. This might take the form of a pollcy that buffered cocoa producer prices from the extremes of world price fluctuations. Economic Rent. The difference between the opportunity cost of a resource and the price recelved/pald by the producer or consumer. Thus, the difference between the price of cocoa beans FOB Tema valued at equilibrium exchange rate, net of efficient marketing costs, and the price the farmer recelves Is an economic rent. In practice, It Is absorbed by implicit taxation and marketing costs greater than the efficient level. The absorption of economic rent by the government (Including COCO- BOD) reduces the return to the producer and hence weakens the Incentive to expand cocoa production via more Intensive cultivation/harvesting or through expansion of cocoa acreage. If the expected return to Investment In cocoa production Is reduced below the expected return to other crops on a sustained basis, farmers will disinvest In cocoa and invest in other crops. Since shifting to other crops from cocoa requires extensive cutting, burning and clearing of the land, which Is costly, disinvestment may not be rapia unless the prospective return from the crop(s) Is sufficiently large to make quick land clearance profitable. Land can be cleared less rapidly by plecemeal cutting of trees whose maturity (over 30 years) has significantly reduced yield, by failing to replace dead trees, by pruning back cocoa trees and Interplanting annual crops, etc. Thus, unless the expected long-term profitability of the most attractive alternative crop Is much greater, the process of disinvestment may be siow and difficult to detect in the early stages.  L A REVIEW OF THE CURRENT STATUS OF GIHANAIAN AGRICULTURE Agriculture plays a major role In Ghana's economic development. The sector employs about 70 percent of the labor force and contributes nearly 60 percent to the Gross Domestic Product. With a high rate of growth of the active labor force (15-64 years) and a relatively low growth of gross capital formatlon, agriculture will continue to play an kIportant role In Ghana's development In the foreseeable future. Although the goverrnnent's food security objectivos call for the productlon of substantial amounts of local food staples, the high debt burden, the need for essential Imports and the absence of viable export alternatives continue to make cocoa which contributes about 60 percent of the country's foreign exchange earnings, the most hkportant export commodity in Ghana. There are three ecological zones In Ghana, namely, the coastal savannah which is minor In terms of area and covers part of the coastal strip of Ghana, the forest zone which occupies about one third of the country, and the northern savannah which occupies most of the remaining two thirds of the country. The forest zone produces the major food and cash crops, namely, maize, tubers, cocoa and oil palm. The northern savannah produces the basic crops: sorghum, millet, rice, beans, and peas, ground nuts and cotton. Cattle Is kIportant In the North and provides power for farm operations as well as a source of meat. It also serves as the place for quarantining and fattening cattle knported from Burkina Faso and Mall before they are moved down South for slaughtering. The major activity In the coastal savannah Is fishing. Lknited numbers of cattle are raised In the Accra plains, In most parts of Ghana, small numbers of poultry, sheep and goats are raised by households as a source of protein and for sale to supplement the Income when the cash flow Is low. A few large scale poultry and pig ventures can be found close to the major urban centers. Vegetables are raised around the major cities to moet the demand of expatriate and urban dwellers.  - 1.5 - There are tradtional regional patterns In food consumption In Ghana. The major staple foods In the North are miliet, sorghum and yams. 'n the coastal regions, maize products are more predominant. Plantains and tubers are consumed more In the forest zone. With increasing migratlon these broad reglonal patterns now follow thin lines. Malze and rice are gradually becoming national food Items, particularly In the South, and they are consumed by a large cross section of the population. Agricultural production in Ghana Is characterized by smailholders who use traditional methods with little or no purchased Inputs. The major inputs are land and labor. There Is not much pressure on land as revealed by the 1984 population census. On the average there are 51 persons per square kilometer. It is estinated that of the total land In Ghana about 10 million hectares of them are suitable for agricultural productlon. However, due to accessibility and other constraints, less than one third of this land is currently being cropped. Farm sizes tend to be small because of the production technology used and also the Inheritance and land tenure systems which lead to land fragmentatlon. Farm extensions therefore occur in the vertical direction with one farmer owning a few plots at different locatlons in the village or in the country. There are variations in the land tenure system depending upon the area. Generally, in the North, all the lands are vested In the government and so outright sale can be obtained only from the government. However, temporary occupation of land can be arranged through the 'tindana' or caretaker of the land In the area concerned. In the South, the chiefs are the custodians of the lands for the people. In most places, In the South, outright sale of land Is forbidden but security of tenure can be obtained through a long-term lease or sharecropping arrangement. People who come from an area (citizens) can obtain land from the chief for farming purposes and the farm can be passed on to one's heirs. The relative abundance of land has given rise to a bush fallow-cropping sequence, especially In the South where the vegetation and rainfall patterns can support  - 1.8 - It. In the bush fallow-cropping system, a plece of land is cultivated for a number of years and then It is left uncultivated for a while for the soil nutrients to be replenished befcre cropping Is done again. The length of the fallow normally depends on the pressure on land In the area. Slnce land Is available the only critical factor in agricultural production is labor. Most of the agricultural labor Is supplied by the .armer and members of the household. High demands for labor during labor-intensive operatlons like harvesting and weeding used to be met by soliciting the help of a neighbor In an arrangement known as nnoboa. Nnoboa labor is not paid for but the farmer also provides the same service to the neighbor at another tkne. Wlth the dwindling In size of the farm household through schooling of children and rural-urban migratlon and the avallability of a pool of farm laborers, nnoboa Is ceasing to be an Important source of farm labor. A farm can hire labor on an annual contract or hire casual labor on a daily basis. To be able to characterize the agricultural sector of Ghana, a nationwide agricultural economics survey (AES) was conducted from October 1987 to February 1988 to gather Information on land and animal holdings, cropping patterns and Input use in Ghanaian agriculture. The survey formed part of the cocoa sector rehabilitation project. A sample of 705 farm households constituting a probability sample of farm householas was used. The sample formed a subqet of the sample of 3,200 households being used for the Ghana Llving Standards Survey (GLSS). The GLSS sample was selected by a two- stage sampling technique. In the first stage, 200 clusters or enumeration areas (EAS) were systematically selected from a total of 12,969 with probability proportional to the size of the EAs. The sample frame was a total of 2,444,836 households out of which 3,200 were selected. Through an arrangement with the Ghana Statistical Service which required that the AES should Interview the responoents after the GLSS enumerators had been there, the agricultural households of the first 1,070 households of the GLSS sample were interviewed for the AES. A total of 705 farm households were interviewed using a  - 1.7 - pretasted structured questionnaire In the three ecobgical zones, coastal savannah, forest and northern savannah. Survey results were entered on computer floppy disks using a special data entry program which checked for consistencies In responses. Seven hundred questlonnaires were found to be useable for the analysis. The breakdown of respondents by region Is presented in Table l.B.1 There were 174 farm households in the North and 526 In the South. About 45 percent of the respondents In the South Table 113.1: Dlstrbution of Survey Households by Region Number of Cocoa Region Househids Farmers Ashand 143 91 Brong Ahafo 40 27 Central 85 37 Eas_m 81 42 G-eair Accra 17 NofWiem 10 Upper Eat 32 Upper Wet 32 Vol 84 5 Wstasm 78 35 North 174 South 526 237 T O T A L 70 2S7 respondents In the South cultivated cocoa. The age distribution of household members of the survey Is presented In Table l.B.2. When the distributlon Is compared with the national distribution (last two lines of Table i.B.2) as estinated by the Statistical Service, we see a sknilar pattern and so we can therefore Infer that our sample Is a normal one. The active labor force (age 15 to 64) represents 49.2 percent of all household members In the sample. Nationally the proportion of the population In the active lak'or force is about 49.5 percent (mid-year 1986 esthuate).  Tom AM OWth of HWoMsb MOWS by RS0wn 41 Regkon 0. 10-14 15.19 2024 2544 45054 6644 Ove 64 Aehahr 245 114 71 46 1S 47 30 30 (33.02) (15.36) 9.57) (6.20) (21.43) (6.33) (4.04) (4.04) &WV Ahdo as ~ 34 22 21 42 10 3 6 Brong 66"0 (33.33) (16.43) (10.63) (10.14) (20.29) (4.83) (1.45) CAnI" 154 61 44 36 104 26 20 17 (3.06) (33.19) (13.16) (9.46) (6.19) (22.41) (5.60) (4.31) EansWn 162 s 49 43 104 33 16 26 (SS7) (32.21) (13.52) (9.74) 6.56 (20.66) (6.56) (3.18) OrseW Accx 29 10 10 7 15 6 4 2 (34.94) (12.06) (12.05) (.43) (16.07) (7.23) (4.82) Noitin 254 105 as 40 192 0 21 11 (1.47) (34.00) (14.06) (984) (1.16) (25.70) (7.76) (2.81) Upper Eas 63 23 12 14 45 22 6 a (33.16) (12.11) (6,35 (7.37) (23.08) (11.58) (3.16) upper WeN 71 26 30 24 49 16 11 4 (30.21) (11.91) (12.77) (10.21) (20.66) (766) (4.68) (1.70) Vom 201 76 47 40 114 22 21 12 (37.71) (14.26) (5.62 (7.50) (21.39) (4.13) (3.94) Western 136 57 39 37 102 15 I1 15 (3. 2) (33.33) (13.77) (9.42) (6.54) (24.64) (3.62) (2.66) Ghao.a 1366 576 360 310 251 143 130 (Survey) (33.06) (13.99) (9.47) (7.2) (22.49) (6.24) (3.47) (3.16) T iw Ghws 4_ 12 1161 lOMs UN 734 430 472 (3. (11.71) (9.10 (YA OL14) 5 (33 (3.64) S/ Figw is pwrmnUes deams prcen of ON ib region. * EsUnaftd pOPulnIOn (nVd.yer 1966) by BSte11i SIF5ios. Amr. Febuwy 1967. Land The average lnd of a houshod an hoW the WA Is used Is presented hI Table .1.3. On the average, the Ounalan farme has abOut 10 acrs of land at his disposal. Although there Is more land pr capa the North than In the South, the average  - 1.9 - agricultural land holding In the South is greater than that In the North. In general, the farmer In the South has 32.5 percent more land than the farmer In the North. Leasing or share cropping Is completely absent In the North. About 18 percent of the land in the South Is Involved in some leasing or sharecropping arrangement. A lease Is normally contracted for a given period of time for a certain monetary consideration either paid outright or paid in Instaiments. When a plece of land is transacted under a sharecropping arrangement, the land owner and the tenant farmer share either the proceeds of the farm (in case of annual crops) or divide the farm (in the case of perennial crops) In some ratio. There are mainly two sharecropping arrangements: 'abunu' system whore the owner of the land and the tenant have equal shares; 'abusa' system where the owner of the land's share Is one-third and the tenant gets two-thirds. Land Is cropped more Intensively In the North than in the South. Only 2..6 percent of the land in the North Is left to fallow as compared with 30.7 percent in the South. Farmers In the South have never cleared about 8 percent of the land they own whereas In the North Just 2.6 percent of such lands are uncleared. Table LB.3: Average Land Holding Per Household ' Region (acres) Leased Leased Size Owred or Share Or Share Owned Owned of and Cropped Croppeed and and Not Region Land Operated Out In Fallow Cleared AshanU 13.5 5.3 1.0 1.1 5.3 0.8 Brong Ahafo 14.2 25.8 0.3 3.3 4.8 0.0 Centrai 10.3 4.2 0.4 2.6 2.3 0.7 Eastern 1 1.9 6.6 0.7 1.5 2.5 0At# Greter Accra 7.2 38 0.00 06 2.1 08 NorUwn 7.7 7.4 0.0 0.0 0.1 0.3 Upper East e8 6.7 0.0 0.0 0.1 0.3 Upper West 8.6 8.0 00 0.0 0.5 0.0 Voht 8.1 4.2 0.1 06 1.9 1.3 Wetern 1 1.6 4.2 0.1 1.9 3.6 1.9 North 7.7 7.4 0.0 0.0 0.2 0.2 South 11.4 5.0 0.5 1.6 3.5 0.9 Ghana 10.5 5.6 0.4 1.2 2.6 0.7  - 1.10 - As It has already been polnted out farmers In Ghana may have more than one plot of land either In the same village or scattered in the farming areas. The average size of a plot In Ghana Is 4.44 acres (Table 1.8.4). Plot sizes tend to be bigger In the South than In the North. On the average, the plot size of a farm In the North Is 43.7 percent less than that of the farm In the South. Ashanti Farmers normally plant multiple crops on the same plece of land. This is done to mitigate Brongyleld risk and also provide food security whereby the farmer ensures that most of his food Ahafoneeds are provided from his own sources. Annual crops could be mixed together or there Centralcould be a mixture of annuals and perennials. Thus the cropping pattern of an area could comprise a mixture of a number of crops. Apart from providing food and a cash Income, young cocoa trees are Interplanted with plantain, bananas and cocoyams. As the cocoa trees grow up and form a canopy, the food crops are elkuinatod. Table 1.8.4: Average Size of Plot by Region (acres) Leased Leased Size Owned or Share or Share Owned Owned of and Cropped Cropped and and Not Regio. Land Operated Out In Fallow Cleared Ashant 6.65 2.59 0.50 0.54 2.63 0.39 Brong Ahato 6.40 2.6. 0.15 1.44 2.20 0.00 Central 4.42 1.83 0.17 1.10 0.99 0.32 Eastem 5.34 2.97 0.31 0.67 1.13 0 25 Greater Accra 3.05 1.61 0.00 0.24 0,88 0.32 Noren 3.00 2.88 0.00 0.00 0.02 0.10 Upper East 2.11 2.09 0.00 0.00 0.02 010 Upper West 3.25 3.04 0.00 0.00 0.21 0.00 Volta 3.31 1.76 0.05 0.32 0.74 0.52 Western 4.80 1.72 0.04 0.79 1.47 0.78 North 2.85 2.74 0.00 0.00 0.06 0.06 South 5.06 2.21 0.22 0.71 1.53 0.41 Ghana 4.44 2.36 0.16 0.51 1.11 0.31 Tree crops are cultivated In the forest zone. The most kIportant ones are cocoa and oll paln. The conditions that are suitable for the cultivatlon of cocoa are also suitable for oil pabn and so the two crops compete for land, labor and other  - 1.11 - resources of the farmer. Forty-five percent of the respondents In southern Ghana cultivated cocoa and this represented 33.8 percent of aH the respondents (Table 1.9.1). Oil pakm was cultivated by 21.3 percent of the respondents In southern Ghana or 16 percent of all the respondents In the survey (Table l.B5). Current output from a tree crop farm depends on the acreage and the ages of the trees since the age-yield relationship of trees have a great deal of impact on output. Cocoa and oil pakm trees were disaggregatad by vintage and the results are presented In Tables 1.B.6-8. There Is a large stock of old cocoa trees (20 years and over) of the traditional variety In Ghana. More than one-fourth of the acreage In cocoa have trees which are over 30 years old and most of these trees are of the traditlonal varlety. This result has a lot of consequence for cocoa production In Ghana since ylelds are very iow at that old age. The area under cocoa In Ghana started to shrink from the mid-19609 and reached Its lowest point around the late 19709. It Is noteworthy to point out the connection between acreage under cocoa and its real pt 'ce. Producer prices In real terms fell from 100 In 1963 to the 44-52 range In 1965-66. However, recent events ;ook promising for the cocoa subsector. Farmers have responded to new Incentives and carried out now plantings. Survey results Indicate that new plantings or cocoa under five years old cover 33 percent of all land under cocoa. Of these new cocoa trees, 90.9 percent of them are of the hybrid variety. Nearly 60 percent of these new farms planted to hybrid cocoa occurred In Ashantl and Western reglons. When all these trees start bearing they can boost Ghana's cocoa production by a large amount. In response to reduced Incentives In the growing of cocoa and therefore reductions In the average size of cocoa farms beginning In the mid 1960s, Interest started growing In oil palm cultivation with the average size of farms beginning to rise about two decades ago.  - 1.12 - Tabt LL.: NI& r of Farmers ok Cooa aWd 01 Palo COCOA OL PALM % Of % Of % Of % ot Souwm S ySouem Survey Region Number Total Tot Number Totbl Total Ash"nX 91 17.3 13.0 12 2.3 1.7 Brong Ahao 27 5.1 3.9 6 1.1 09 CenVl 37 7.0 5.3 33 6.3 4.7 Ea*rn 42 8.0 6.0 14 2.7 2.0 VoIl 5 1.0 0.7 13 2.5 1.8 Wee Ae 35 6.7 5.0 34 6.5 4.8 Ohatr' 237 45.1 33.6 112 21.3 16.0 Tabr 6: Ae AV trbuluon of Tracltuil Cocoa Trm" by Region YEARS (S) Region 0.4 5,9 10.14 15.19 20-29 30 plus Al Ages (Aores) Ashanl 1.39 0.00 4.18 11.68 36.55 46.20 214.5 Brono Ahab 3.21 4.59 6.42 23.85 23.85 36.08 109.0 Cenmi 5.78 5.29 1.65 20.66 30.58 36.04 1210 Eastmn 2.18 0.35 0.65 2.46 0.77 1.50 89.5 Volta 0.00 0.00 100.00 0.00 0.00 0.00 1.0 Welern 3.30 0.00 6.35 14.21 4.82 71.32 31 5 Ghana 5.21 2.82 5.21 19.I6 27.0 35.89 566.5 Table LB.7: Mean Age Dstrbuton of Hybrid Cocoa Tres by Region Y E A AR8 I() Region 0-4 5.9 10-14 15-19 20.29 30 pFu All Ages (Acres) AehanU 58.06 12.62 12.28 7.77 3.11 1.16 257.5 Brong Ahab 28.91 1.99 12.63 1.33 31.23 25.91 150.5 CenVa 37.50 25.86 2.59 26.02 6.03 0.00 118.0 E_sn 49.55 13.43 8.71 11.69 7.96 8.46 201.0 Volt 72.73 0.00 18.18 9.09 0.00 0.00 5.5 WoeeSn 81.89 0.00 0.00 13.39 1.57 3.15 127.0 Ghana 51.44 10.79 9.91 11.18 9.33 7.35 857.5  - 1.13 - Table 1X.8: Mean Agt Distrbution of ON Pal Tree by Rgon Y E A R S (%) Region 0-4 5-9 10-14 15-19 20-29 30 plus All Ages (Acres) Ashanti 37.20 29.95 13.b3 4.83 14.49 0.00 103.5 erong Ahab 2.04 40.82 51.02 6.12 0.00 0.0C 49.0 Centa 40.57 46.38 4.95 5.24 2.8e 0.00 105.0 E"JeYn 65.40 26.34 6.03 2.23 0.00 0.00 44.8 vont 1177 12.71 67.72 7.80 0.00 000 117.0 Wesern 67.96 23.93 2.56 0.00 0.00 5.55 117.0 Ghtana 40.38 30.13 20.43 4.12 3.64 1.31 494.9 Labor Slnce generally land can be obtained for agricultural purposes If the need be, the most constraining factor may be labor. Labor for farm operations can be obtained from the farm household or hired either on permanent or casual basis. Most of the labor Is supplied by the farmer and his spouse and other members of the household. In the Survey, 91.4 percent of the heads of households In the North and 74.2 percent of those in the South were married. About 85 percent of heads of households were under 65 years of age and so they formed part of the active farm labor force. The average size of a farm household is presented in column 2 of Table 1.8.9. There are more members in a norttiern household than a southern one wlth the survey average being about S. Since households have multiple farms, the small average household size puts a lot of stress on farm labor requirements. The use of family and permanent labor per tarm Is 2.2 adult persons (child labor Is valued at rialf adult labor) and they work 238 days on the average In a year. There are more family and permanent laborers per farm household in the North than in the South and on the average they work 44.6 percent more days than those In the South. More workers are Involved In casual farm work  _ 1.14 - although they work loss number of days in a year than family and permanent workers. On the average there. are five casual laborers involved In :he farm operations of a household and they work about 62 days in a year. There are fewer people per household Involved In catual farm work In the South but they work more days than in the North. Table L!3*: Labor Use on Farms by Region Fai arid Pwnmnt Cawil Mean Days Mean Days Mean Size Worked Worked of Farm Per Person Per Person Region Households Workers Per Year Workers Per Year Ashanti 5.2 1.7 214.8 2.9 49.4 Brong Ahao 5.2 1.8 218.6 3.5 55.0 Cental 5.5 1.7 233.0 7.0 74.7 Easrn 6.2 1.8 230.6 4.1 99.6 GOmter Accra 4.9 1.6 166.5 3.4 20.5 Norwrn 6.8 3.0 310.4 6.0 58.5 Upper East 6.0 5.0 264.2 7.0 38.2 Upper West 7.3 3.4 338.5 10.1 43.8 VoNt& 6.3 1.8 180.3 5.5 68.4 Western, 5.5 1.7 211.5 4.4 56.4 North 6.7 3.4 307.4 7.0 52.2 South 5.6 17 2125 45 65.1 Ghana 5.9 2.2 238 2 5.0 62.3 Purchased Inputs Apart from farm knplements like axe, hoe, matchets, sickle, very few of the farm inputs are purchased. Lnilted amounts of fertilizer are used on food crops like maize, rice and pineapple. The major fertilizers used are NPK, urea and ammonium sulphate. Some cocoa farmers spray their trees wlth unden and gammalin Insecticide. Farmers may purchase the seeds of knproved varieties from the Seed Company or other organizations Involved In seed multiplication. Hybrld cocoa seedlings may be obtained from the Cocoa Services Division of COCOBOD and oil pakl seedlings are obtained from the Oil Pakm Research Station at Kusl.  1.1 - Farm Operations The first farming operaton Is land preparatlon. It can be done by manual labor with axe, cutlass and hoe, or with animal or machinery drawn implements. It Involves bush clearing, tree chopping, buo-ning and clearing of debris (apam). Due to the topography aimost all land proparaflons in the forest zone Is done by manual labor. To a large extent land preparation In the North is also manually performed althougn there is soime use of animal powor. There are a few tractofs avallable which may be used for paowinSg and harrowing. The amount of labor Involved in preparing a plce of land for cropping depends upon the previous use of the land or type of vegetatlon on It. Crops can be grown on virgin land, fallow land or a land already under a crop. Some of the big trees on a virgin land would have to be cut down with an axe during bush clearng. Therefore, land preparatin on virgin land may take several man days If the forest Is thick. Since land Is not left in fallow for too long, It may not take as long to cut down the trees during bush clearing as the land may not have big trees on It. Replanting on land on which an annual crop has been grown requires light bush clearing and so the demand on labor Is low. However, If the land had a tree crop on It, there may be the need to uproot the trees and so labor requirement could be substantial. In general, since the vegetation Is thicker In the forest zone then the savannah zone, bush clearing takes more man days In the forest areas than In the savannah areas. The mean labor requirements in man days for land preparation and weeding are summarized In Table l.B.10.  - 1.16 - Tabt LS.1 0: Mean Lbr Requiremnt for Farm OperatioB (man days) Bush Tres Cleaning W"ding/ Region Ceaing Chopping Buming (' am) Hoeing North 6.9 3.2 1 4.8 7.1 South 13.2 8.1 1 9.1 14.2 Ghana 'V.S 7 6 1 8.7 14 3 Land preparation begins before the onset of the rains. In the South, the months of January, February and the fIrst half of March are devoted to land preparation. The rains do not start In the North until late April or early May. Land preparation therefore starts around February and continues until April. Plants can be planted In different ways, somethnes depending upon the crop and other tkmes upon the practice. For some crops, seeds can be sowed directly and for others, the seeds would have to be nursed Into seedlings before transplanting. The time required for planting depends upon the type of crop. It can take anything fr m one day for rice to 17 days for cocoyam. For instance, whereas rice can be broadcast, seed drills would have to be made for maize, mounds are made for yams, and holes are dug for oil pakm seedlings, cocoyam corms and plantain suckers. Fresh cocoa beans can be sowed directly. This Is the normal practice during rehabilitation of a farm when some dead trees are replaced. As recommended, replanting requires that the cocoa seedlings are raised In a nursery before transplanting them In the field. When the seeds or seedlings are planted, certain operatlons are carrled out to take care of the plants. Cocoa trees need shade when they are young and so apart from leaving a few trees on the land durng land preparation, plantains and cocoyams are planted. These food crops provide the shade and also give the farmer some food and cash ncome during the non-yIelding years of the cocoa trees. Traps are made to  - 1.17 - catch rodents and other small anknals which may prey on the young plants. Apart from preventing crop storage, the animals that the traps catch provide meat protein fcr the farm households. In addition to the traps, collars or rings made from wire gauze or raffla are put around young oil palm seedlings to protect them from rodent attacks. Artlficlal scarecrows are placed In cereal farms to scare away birds who may feed on the grains. The most knportant farm majntenance or croP catr' operation Is weeding and/or hoelng. Weeds abound during the rainy season. The weeds compete with the crop for the available soll nutrlents and moisture. The more the weeds and the longer they co- habitate the soil with the plants, the lower the crop yield Is. It is therefore Important to weed around the farm a few times before harvesting. Weeds reduce as the plants outgrow them and form a canopy. On the average farmers weed around their plants twice In a season. The labor requirement for weeding depends on the spacing of the plants and the number of crops on the land. Thus, an acre of a monocrop land would take less man days to weed than It would a multiple-crop land. In the South, weeding Is largely done with a cutlass but In the North the hoe Is mostly used. The hoe Is also used to open up the soil for better aeration and seepage of water. Survey results Indicate that on the average It takes 14.2 man days to carry out weeding on an acre of cropped land In the South and half as long In the North. The large difference In the two figures may be due to the higher vegetative growth and the greater degree of multiple cropping In the South than In the North. To maintain soll fertility, land that Is continuously cropped should be fertilized. Also to realize the full potentlal of an Inproved variety the soil should be fertilized to supply the optkial level of all soll nutrients In addition to adopting some approved cultural practices. Fertilizer needs water to be effective. In a ralnfed agriculture like Ghana's, there Is a risk In applying fertilizer when an uncertainty exists about the cilhate. The financial loss can be substantlal when there Is a drought. Rainfall patterns  - 1.18 - Indicate that the chances for a drought In a particular year In Ghana Is one out of five. At this low risk level, the risk averse farmer can use fertilizer on his farm. Despite Its advantage, fertilizer is used by a small proportlon of the farmers. Only 14.3 percent of the farmers surveyed used fertilizer, and of this number, 66 percent were either maize, sorghum or milet farmers. None of the cocoa and oil palm farmers used fertilizer. As the population grows and food demand increases, Ghana would have to increase the food production o, resort to bIports. With a precarious export earn!ng potential, the viable alternative Is expansion In domestic production. Such an expanslon would put pressure on land and so the bush-fallow system would have to give way to more Intensive cultivation which would require the application of fertilizer to sustain the system. Even without an acute population pressure currently Ghana Is a high cost producer in many food conmmoditles like maize and rice. The cost per unit of output can be reduced substantially If ylelds would Increase. Yields can Increase through the use of fertilizer. If the widespread use of fertilizer can be postponed because of the abundance of land, It can hardiy be neglected In the cultivation of long cycle perennial crops. it is therefore disturbing that none of either the cocoa or oil palm farmers In the sample used fertilizer on their crops. There Is very little use of insectIcides or fungicides on annual crops. Only 2.4 percent of the annual crop farmers used some sort of insecticide/fungicide. The situation was slightly better for perennials with 14.1 percent of the farmers spraying thear crops with insecticide or fungicide. Although the Cocoa Board recommends that cocoa farmers should spray their farms with Insecticide four tkmes In a year in order to control capsids and thereby Increase thelr yields, only 38.4 percent of the cocoa farmers In the survey spra>-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 ) (