DWC-8703 Kenyan Coffee Sector Outlook: A Framework for Policy Analysis Takamasa Akiyama Division Working Paper No. 1987-3 May 1987 Commodity Studies; and Projections Division Economic Analysis and Projections Department Economics and Research Staff The World Bank Division Working Papers report on work in progress and are circulated to stimulate discussion and comment. KENYAN COFFEE SUB-SECTOR OUTLOOK: A FRAMEWORK FOR POLICY ANALYSIS Takamasa Akiyama May 1987 The World Bank does not accept responsibility for the views expressed herein which are those of the authors and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country, territory, city, area, or of its authorities, or concerning the delimitation of its boundaries, or national affiliation. TABLE OF CONTENTS Page No. I. INTRODUCTION AND SUMMARY .. ................................ 1 II. RECENT TRENDS IN KENYA'S COFFEE SUB-SECTOR . . 3 III. SALIENT FEATURES OF KENYA'S COFFEE SUB-SECTOR . . 8 IV. ANALYSIS OF KENYAN COFFEE PRODUCTION . . 22 V. PRODUCTION PROSPECTS .... . 28 VI. PROSPECTS FOR KENYA'S COFFEE SUPPLY/DEMAND BALANCE ........ 44 VIII. POLICY IMPLICATIONS ....... . . ................ . 48 LIST OF TABLES Table 1: KENYA COFFEE: STOCKS, PRODUCTION, AVAILABILITY AND EXPORTS--1969 TO 1984/85 .............................. 4 Table 2: KENYA COFFEE: AREA, PRODUCTION AND YIELD-- 1963/64-1984/85 ....... ... ........ .... 5 Table 3: KENYA COFFEE: PRODUCER PRICES, 1965-84 ........... ..... 7 Table 4: KENYA COFFEE--NUMBER OF SMALLHOLDER FARMS BY CATEGORY AND BY PROVINCE AND THEIR PROPORTIONAL DISTRIBUTION BY SIZE (1979) ........................... 9 Table 5: KENYA COFFEE--PERCENTAGE OF SMALLHOLDERS GROWING SELECTED CROPS BY PROVINCE (1974/75) .......... 10 Table 6: KENYA COFFEE: AVERAGE UTILIZATION OF LAND AMONG SMALLHOLDERS .. 11 Table 7: KENYA COFFEE: PRODUCTION COSTS OF CLEAN COFFEE (PER HECTARE AND PER KILOGRAM) ON SMALLHOLDINGS AND ESTATES UNDER VARIOUS ECOLOGICAL CIRCUMSTANCES (1981/81) 12 Table 8: KENYA COFFEE--COST/BENEFITS OF ADOPTING PROPOSED INPUT PACKAGES RECOMMENDED BY CRF ...................... 13 Table 9: KENYA COFFEE: GROSS MARGIN PER MANHOUR OF THE MAIN ENTERPRISES WITHIN EACH SMALLHOLDER ZONE .......... 14 Table 10: PRODUCTION COSTS PER HECTARE OF GREEN COFFEE IN SELECTED COUNTRIES (1982) ........................ .. 16 - ii - Table 11: KENYA COFFEE: NET PAYMENT BY CBK TO KPCU BY "CLASSES" OF COFFEE--1983/84 .......................... 18 Table 12: KENYA COFFEE: NET PAYMENT BY CBK TO KPCU BY AGGREGATE "CLASSES"--1976/77-1983/84 .................. 19 Table 13: KENYA COFFEE: EXPORT TAX AND COUNTY CESS--1977/78-1984/85 ................................. 21 Table 14: EXPECTED YIELD INCREASE FOR 1985-2000 . . 31 Table 15: KENYA COFFEE: "BASE CASE" COFFEE PRODUCTION PROJECTIONS--1985/86-1999/2000 ........................ 35 Table 16: KENYA COFFEE: SENSITIVITY ANALYSIS ON KENYA COFFEE PRODUCTION, BY PRODUCER PRICES ................. 36 Table 17: PROJECTED PRODUCTION GROWTH RATES WITH DIFFE RENT R EAL PRODUCER PRICES ......................... 38 Table 18: KENYA COFFEE: SENSITIVITY ANALYSIS ON KENYA COFFEE PRODUCTION, BY RUIRU 11 YIELDS ........... 39 Table 19: KENYA COFFEE: PROJECTIONS OF KENYA'S COFFEE SUPPLY/ DEMAND BALANCE--1985/86-1999-2000 ..................... 45 I. INTRODUCTION AND SUMMARY 1. Long-term production prospects, as well as the outlook for domestic and export demand are indispensable in forming sectoral and macroeconomic policies, especially those related to pricing. This paper examines recent trends in the Kenyan coffee sub-sector and evaluates production and export prospects for Kenyan coffee under several scenarios. The analysis was done to provide background for an assessment of policies relating to the coffee sector in Kenya. Statistical and econometric tools were used in carrying out the analysis. 1/ 2. In evaluating the long-term prospects for the Kenyan coffee sub- sector, scenarios with respect to important factors such as the potential yields and adoption rates of a new high-yielding variety and forecasts of international coffee prices have been considered. The analysis shows that under the assumption that future world coffee prices develop more or less as projected by the World Bank and that a price-supporting international export quota system is maintained, Kenya's coffee production would increase at about 2.8% p.a. during the period 1985/86-2000. 3. An important policy issue is whether Kenya should support the International Coffee Agreement (ICA) export quota system. Analysis indicates that Kenya's future coffee export revenues will not differ much whether the quota sytem continues or not, provided that Kenya obtains quota levels 1/ The results of this analysis were incorporated in a model on the Lotus spreadsheet program to facilitate simulation exercises. - 2 - increasing at 2% p.a. However, if Kenya is to expand its robusta production substantially or if adoption of the new high-yielding variety (Ruiru 11) increases arabica production substantially, Kenya's coffee export revenues could be larger in the absence of the quota system. Another problem with the quota system is the possibility of Kenya accumulating large stocks. 4. The paper is structured as follows. Section I describes recent trends in Kenyan coffee production, consumption, exports, stocks, yields and prices. The performances of the estate and smallholder sectors are compared. Section II describes the characteristics of coffee production in the estate and smallholder sectors, covering topics such as costs of production, quality differentials and government policies. In Section III, coffee production in the estate and smallholder sectors is modelled econometrically using the vintage capital approach. Section IV describes projections of Kenyan coffee production using these models. The assumptions underlying the projections are outlined. Sensitivity of the projections to changes in real producer prices and to differing assumptions about the potential yields from the new coffee variety (Ruiru 11) is examined. The prospects for Kenya's coffee exports are described in Section V. The focus here is the International Coffee Agreement and, in particular, the outlook for the Agreement's export quota system and Kenya's likely quota allotment. The results of Section IV and V are brought together in Section VI in an examination of Kenya's future coffee demand/supply balance. Major points of interest are Kenya's export revenues from coffee and likely stocks under the various scenarios. Section VII draws together the various policy implications resulting from this analysis. -3 - II. RECENT TRENDS IN KENYA'S COFFEE SUB-SECTOR 5. Table 1 presents a balance sheet for Kenya's coffee production, exports, domestic consumption, and stocks since the late 1960s. Production has been increasing on average at a rate of 4.7% p.a. since 1968 and exports have increased at 3.9% p.a. during the corresponding period. As production has increased faster than exports, and given the low domestic consumption, stocks have accumulated. The stock level reached 1 million bags at the end of the 1980/81 season, following the imposition of ICA export quotas. As production continued to increase after the quotas were adopted stocks reached over 70% of Kenya's production in the last three years. Since the quotas were suspended in February 1986, Kenya has sold a large portion of its stocks. Had the ICA quota system continuied during the 1985/86 season, stocks could have gone as high as 1.7 million bags (instead of the actual level of 1.2 million bags in October 1986). 6. Table 2 gives area, 1/ production and yield of coffee estates and smallholders combined. Several important features of the sector can be observed from the table: (i) There are no clear long-term trends in the estates' area and production. Estates' mature areas declined slowly from the early 1960s until the mid-1970s--just before the "coffee boom"--and increased thereafter; 1/ "Mature area" in Table 2 signifies area under coffee trees of more than two years old. Table 1: KENYA COFFEE: STOCKS, PRODUCTION, AVAILABILITY AND EXPORTS--1969 TO 1984/85 ('000 bags) -----------------------------------------------------------------__----------__---------------------------------------------- Opening Stocks Net ZL Exports Crop Total Domestic Exportable Gross Avallability To ICA To Year Gross Net ZA Production Consumption Production Availability For Export Total Members Non-Members (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) STARTING I OCTOBER 1968 127 -73 821 19 802 929 729 777 731 46 1969 152 -48 954 21 933 1,085 885 799 784 15 1970 286 86 1,036 23 1,013 1,299 1,099 953 783 170 1971 346 146 990 22 968 1,314 1,114 1,079 941 138 1972 235 35 1,261 20 1,241 1,476 1,276 1,200 1,153 47 1973 276 76 1,244 20 1,224 1,500 1,300 1,224 1,161 63 1974 276 76 1,171 20 1,151 1,427 1,227 1,088 1,014 74 1975 339 139 1,225 21 1,204 1,543 1,343 1,343 1,317 26 1976 200 0 1,699 20 1,679 1,879 1,679 1,428 1,364 64 1977 451 251 1,356 43 1,313 1,764 1,564 1,391 1,300 91 1978 373 173 1,232 40 1,192 1,565 1,365 1,231 i,195 36 1979 334 134 1,651 50 1,601 1,935 1,735 1,366 1,317 49 1980 569 369 1,715 68 1,647 2,216 2,016 1,205 1,110 95 1981 1,011 811 1,474 36 1,438 2,449 2,249 1,702 1,301 401 1982 747 547 1,551 50 1,501 2,248 2,048 1,427 1,291 136 1983 821 621 1,992 51 1,941 2,762 2,562 1,516 1,391 125 1984 1,246 1,006 1,558 50 1,508 2,754 2,514 1,604 1,330 274 1985 1,130 800 2,167 50 2,117 3,247 2,917 2,023 1,796 227 1986 1,224 L& Working stocks were 200,000 bags for the period 1968 to 1983; 240,000 bags thereafter. Source: ICO Table 2: KENYA COFFEE: AREA, PRODUCTION AND YIELD--1963/64-1984/85 ------------------------------------------------------__---------------------__-----------------------------------------------------__---- Estates SmalIholders Total Coffee Mature Production Yield Mature Production Yield Mature Production Yield Year Area Area Area (ha) (tonnes) ('000 bags) (kg/ha) (ha) (tonnes) ('000 bags) (kg/ha) (ha) (tonnes) (i000 bags) (kg/ha) -------------------------------------------------------------------------__--__----------------------------------------------------------- 1963/64 32,538 28,405 473 873 13,000 15,373 256 1,183 45,538 43,778 730 961 1964/65 32,423 22,393 373 691 16,900 14,774 246 874 49,323 37,167 619 754 1965/66 32,267 25,683 428 796 27,400 25,523 425 931 59,667 51,206 853 858 1966/67 31,964 25,231 421 789 42,900 27,558 459 642 74,864 52,789 880 705 1967/68 31,188 13,246 221 425 49,900 20,515 342 411 81,088 33,761 563 416 1968/69 30,690 22,342 372 728 52,300 23,264 388 445 82,990 45,606 760 550 1969/70 29,903 26,521 442 887 54,057 26,275 438 486 83,960 52,796 880 629 1970/71 29,900 28,600 477 957 54,057 26,302 438 487 83,957 54,902 915 654 1971/72 29,535 29,984 500 1,015 55,555 28,362 473 511 85,090 58,346 972 686 1972/73 29,535 38,956 649 1,319 55,555 33,783 563 608 85,090 72,739 1,212 855 1973/74 29,129 31,152 519 1,069 55,600 36,767 613 661 84,729 67,919 1,132 802 1974/75 28,603 29,985 500 1,048 57,786 35,464 591 614 86,389 65,449 1,091 758 1975/76 28,603 37,675 628 1,317 57,786 36,135 602 625 86,389 73,810 1,230 854 1976/77 27,821 49,685 828 1,786 57,786 47,660 794 825 85,607 97,345 1,622 1,137 1977/78 29,402 33,685 561 1,146 59,206 47,744 796 806 88,608 81,429 1,357 919. 1978/79 29,102 26,809 447 921 62,574 46,079 768 736 91,676 72,888 1,215 795 1979/80 31,232 39,109 652 1,252 71,172 51,900 865 729 102,404 91,009 1,517 889 1980/81 32,861 34,744 579 1,057 84,710 6A,007 1,067 756 117,571 98,751 1,646 840 1981/82 33,635 34,392 573 1,023 97,423 52,531 876 539 131,058 86,923 1,449 663 1982/83 33,605 32,981 550 981 100,967 52,469 874 520 134,572 85,450 1,424 635 1983/84 35,711 54,258 904 1,519 114,235 74,683 1,245 654 149,946 128,941 2,149 860 1984/85 35,711 28,922 482 810 116,328 64,717 1,079 556 152,039 93,639 1,561 616 Source: Coffee Board of Kenya. - 6 - (ii) Smallholders' area and production have increased at very rapid rates of 7.6% p.a. and 6.7% p.a., respectively, since the early 1960s; (iii) Yield has fluctuated fairly widely from one year to the next, but on average the estates' average yield has been considerably higher than that of smallholders.' Average yields for the last 15 years for estates and smallholders are 1019 kg/ha and 664 kg/ha, respectively i.e., over 50% difference between the two sectors; (iv) There are no statistically significant trends in the yields of either the estates or the smallholders; and (v) After stagnating during the preceding 10 years, both area and production of the estate and smallholder sectors increased substantially in the last 6-7 years following the 1976-78 "coffee boom". 7. A key economic variable affecting production is the producer price. As noted elsewhere in the report, payments to smallholders vary substantially from one "society" to another and even among factories in a society. Factors that affect payments to farmers include: (i) quality of coffee produced; (ii) management skill of factory operators; and (iii) financial structure of societies. It is estimated that on average a smallholder coffee farmer receives about 86% of the amount paid by the Coffee Board (CBK) to the Kenya Planters Co-operative Union (KPCU). 1/ The smallholder payment share varies between 76% and 94%. 2/ Payments to the estates, on the other hand, depend only on the quality of coffee as they are paid directly by the KPCU. 1/ The payment channel goes from CBK to KPCU and then to societies and estates. Societies pay farmers after deducting expenses. 2/ KPCU unpublished paper "Payment to farms" Nairobi, March 1982. - 7 - 8. Table 3 shows average net payments by the CBK to the KPCU for 1 kg of clean coffee. The payments are shown in terms of nominal Kenya Shillings, in real terms with the consumer price index as deflator and relative to maize prices. Even excluding the "coffee boom" years of 1975-78, long-term real producer prices have shown an increasing trend. Table 3: KENYA COFFEE: PRODUCER PRICES, /a 1965-84 Year Nominal Prices Real Prices /b (K. sh/kg of (Deflated (Deflated by clean coffee) by CPI) maize price /c) 1965 5.960 20.001 14.863 1966 5.300 17.099 15.014 1967 5.850 18.581 18.994 1968 5.580 17.632 20.217 1969 7.440 23.549 27.055 1970 5.680 17.615 17.057 1971 7.020 20.957 .18.046 1972 8.740 24.652 22.468 1973 9.820 25.364 21.164 1974 9.400 20.579 13.429 1975 22.310 41.028 28.974 1976 39.520 65.240 44.404 1977 26.070 37.455 33.639 1978 26.600 32.699 29.888 1979 24.830 28.256 26.027 1980 21.330 21.330 21.330 1981 27.800 24.866 25.812 1982 34.890 25.903 22.656 1983 36.530 24.322 20.874 1984 46.600 28.070 n.a. /a Net CBK payment to KPCU. 7i CPI 1980=]00. T7 Maize proclucer price in Kenya, FAO data. Source: CBK, FAO. - 8 - III. SALIENT FEATURES OF KENYA'S COFFEE SUB-SECTOR 9. The discussion in this section is provided as background to the analysis that follows on the long-term production prospects. (i) Production System 10. In Kenya, coffee is produced both by estates and smallholders. Before Independence in 1963, almost all Kenya's coffee was produced ,by estates owned by expatriate farmers. Before 1952, when the Swynnerton Plan was implemented, it was illegal for smallholders to grow coffee except for small trial areas in the Meru and Kisii districts. After Independence, legislation was enacted which requirecl all estates to be registered for the purposes of the control of returns on capital and the remittance of profits abroad. The expatriates have gradually been selling the estates since that time. Some of the estates sold have been subdivided and distributed to smallholders. As a result, the small- holder sector expanded rapidly in the 1960s and again in the late 1970s. Its share increased from less than 50% of output in the 1960s to about 70% in the early 1980s. 11. Estate farming is different from smallholder farming in a number of aspects. The estates specialize in coffee growing; they are considerably larger (the mijority being between 8 ha and 200 ha); they hire permanent staff and are capital intensive. Many of the estate farms are irrigated (67%) and use large amounts of chemical inputs. They also own their own factories for pulping coffee and often hire management and farming specialists to evaluate and improve performance. 12. Smallholder farms, on the other hand, are mostly very small. According to the Central Bureau of Statistics (see Table 4), 67% of them are less than 2 - 9 - Table 4: KENYA COFFEE--NUMBER OF SMALLHOLDER FARMS BY CATEGORY AND BY PROVINCE AND THEIR PROPORTIONAL DISTRIBUTION BY SIZE (1979) Farm size Province Total area (Ha) Western Nyanza Rift Central Eastern Coast Total (1000 ha) Valley SmalIholdings less than 2 63 74 66 69 60 58 67 2-4 32 23 20 27 33 31 26 more than 4 5 3 14 4 7 il 7 number (x 1000) 342 708 527 501 456 160 2694 (3216) Intermediate farms number (x 1000) 8-50 4 5 29 4 10 1 53 (1040) Large Farms 50-200 36 78 44 75 46 49 54 112 200-2000 45 18 51 22 23 35 40 895 more than 2000 19 4 5 3 31 16 6 1652 number (x 1000) 0.011 0.116 2.327 1.067 0.163 0.051 3.735 .(2569) All farms number (x 1000) 346 714 558 506 466 161 2751 (6915) Source: Central Bureau of Statistics (1981, 1982); J. de Graaff, "The Economics of Coffee," Pudoc Wageningen, 1986. ha and 93% are less than 4 ha. They do not specialize in coffee but normally have other enterprises such as maize/beans, English potatoes and/or cattle. A typical smallholder farm with 1.5 ha of land in the Central region has 0.5 to 0.6 ha for coffee, 0.4 ha for maize and beans intercropped, 0.1 to 0.2 ha each for English and sweet potatoes and 0.2 to 0.3 ha for pasture (see Tables 5 and 6). In some high altitude areas, tea is also grown as a cash crop. Most of the required labor is provided by the family and chemical input use and yield per - 10 - Table 5: KENYA COFFEE--PERCENTAGE OF SMALLHOLDERS GROWING SELECTED CROPS BY PROVINCE (1974/75) ----------------------------------------------------------__-----------------_ Western Nyanza Rift Central Eastern Coast Kenya Valley ----------------------------------------------------------------__-----------_ Coffee 5 21 5 45 44 1 27 Tea 4 - 15 18 11 - 12 Local Maize 74 80 59 95 99 94 86 Hybrid maize 73 36 92 67 30 19 50 Beans 79 39 22 98 86 28 69 English potatoes - 1 8 86 52 2 32 Sorghum 37 75 1 1 16 2 30 Pyrethrum - 18 16 8 7 - 9 Cotton 20 17 5 - 2 5 9 Source: Central Bureau of Statistics (1977); J. de Graaff, op. cit. unit of land are substantially lower than on the estates. Almost none of the smallholders use irrigation. (ii) Cost of Production and Profitability 13. Cost of production per hectare or per kilogram varies substantially from one estate to another and from one smallholder farm to another. A key parameter affecting production cost is yield; the higher the yield the lower the cost of production and thus the higher the profitability (see Table 7). Due to their much higher overhead costs, the production costs of estates are considerably higher than those of smallholders. A recent study of the smallholder sector done by the Coffee Research Foundation (CRF) indicates 1/ that the yield is strongly correlated with input expenditures. The study also 1/ J.D.M. Roe and J.K. Nyoro, "Towards Improving Coffee Productivity in the Smallholder Sector," CRF. Table 6: KENYA COFFEE: AVERAGE UTILIZATION OF LAND AMONG SMALLHOLDERS (Hectares) Area of Land (hectares) Farm Crop Coffee Maize & English Millet Sweet Pasture /a Size Zone Beans Potatoes Potatoes -----------------------------------------------------------__----------------_ 1.5 ha. Central UMI /b 0.5 0.4 0.2 - 0.1 0.3 Central UM2 7r 0.6 0.4 0.2 - 0.1 0.2 Central UM3 7i 0.5 0.6 0.1 - 0.1 0.2 Kisii 0.4 0.4 - 0.15 0.15 0.5 Machakos 0.4 0.5 - - 0.1 0.5 3.0 ha. Central UM1 /b 1.0 0.8 0.4 - 0.2 0.6 Central UM2 /b 1.2 0.8 0.4 - 0.2 0.4 Central UM3 7r 1.0 1.2 0.2 - 0.2 0.4 Kisii 0.8 0.8 - 0.3 0.3 0.8 Machakos 0.8 1.0 - - 0.2 1.0 /a Pasture area includes homestead and fallow land. /b UM1, UM2 and UM3 refer to marginal coffee growing areas, major coffee growing areas and coffee and tea growing areas, respectively. Source: J.K. Nyoro and J.D.M. Roe "Economics of Agricultural Production on Smallholder Coffee Farms in Kenya," CRF. shows that smallholders' profitability would increase with yield, and that the percentage profitability increase would be largest for the farmers with very low yields (see Table 8). 14. Among inputs, fertilizer (C.A.N. and 20:20:0 are most commonly used) and fungicides (mainly copper) are most important; the latter often costing more than the former. In Kenya, these inputs constitute an important part of the production cost of coffee. Their share in total smallholders' production costs varies between 20 and 30% and the corresponding range for the estates is 30 to 40% (see Table 7). Fungicide is the largest cost component in the material inputs; it is required for the control of Coffee Berry Disease (CBD). - 12 - Table! 7: KENYA COFFEE: PRODUCTION COSTS OF CLEAN COFFEE (PER HECTARE AND PER KILOGRAM) ON SMALLHOLDINGS DAND ESTATES UNDER VARIOUS ECOLOGICAL CIRCUMSTANCES (1981/82) -----------------------------------------------------__----------------------__------------------ SmaIlholdings (coffee zone) Estates UM2 UMI UM3 Non- Irrigated irrigated ---------------------------------------------------------__------------------__------------------ Yield (kg/ha) 700 600 400 950 1250 Plant density (plants/ha) 1200 1300 1300 1300 - 1700 Area in coffee (ha) 0.3 0.2 0.15 90 110 Wage rate (Ksh/man-day) 11.70 11.70 11.70 11.70 11.70 Total costs per ha (Ksh '000) 8.1 7.5 6.0 19.3 23.6 Costs per kilogram (Ksh) 11.5 12.5 15.0 20.3 18.9 (USS) 1.21 1.32 1.58 2.14 1.99 -------------------------------------------------------------------__--------__------------------ Sources: Updated information from Ruthenberg (1980) (UM2 & non-irrigated estates), Jaetzold & Schmidt (1983) (UMI1), Schall & Schmidt (1976) (UM3), Magogoni estate records (pers, communication 1974) (irrigated estates), J. de Graaff, op. cit. 15. According to a CRF study, coffee is the most profitable enterprise in the smallholdter sector in most of the coffee growing areas (see Table 9). In most areas, coffee is closely followed in profitability by maize and beans. 16. Some mention should be made, however, of the interpretation of the data comparing the profitability of coffee and other crops. Coffee is a cash- crop; meaning that a well-established marketing system exists for farmers to sell it and obtain cash in return. Maize/beans and other crops in the coffee- growing areas are mainly grown for on-farm consumption and marketing systems for these crops are either poorly organized or do not exist. In fact, it is generally illegal to transport maize from one district to another unless special permits are obtained. Farmers thus do not decide whether to grow coffee or maize solely in terms of the relative profitability of the two crops - 13 - Table 8: KENYA COFFEE--COST/BENEFITS OF ADOPTING PROPOSED a!' INPUT PACKAGES RECOMMENDED BY CRF ----------------------------------------------------------------------__-----__------------------ Yield Group (from low to high) 1 2 3 4 5 6 Current Input Inputs: Fertilizers 271 517 479 634 749 1,065 Expenditure Fungicide 186 272 480 795 943 1,178 (Ksh) Insecticide 74 111 111 148 148 259 Sub-total 531 9000 1,070 1,577 1,838 2,502 Knapsack Sprayer (hired e 55/round) 55 110 165 275 275 330 Total 586 1,010 1,235 1,852 2,113 2,832 Proposed Input Fertilizer 271 377 494 681 864 1,153 Expenditure Fungicide 529 847 1,147 1,297 1,448 2,048 (Ksh) Insecticide 37 37 37 37 37 37 Sub-total 837 1,261 1,678 2,015 2,349 3,238 Knapsack Sprayer 250 350 350 350 350 350 Total 1,087 1,611 2,028 2.355 2,699 3,588 Additional Fertilizer - (140) 15 47 117 88 Expenditure Fungicide 343 575 667 502 505 870 (Ksh) Insecticide (37) (74) (74) (111) (111) (222) (including hired labor) Sub-total 306 361 608 438 511 736 Knapsack Sprayer 195 240 135 75 75 20 Hired labor - 1,065 1,901 1,666 1,288 1,214 Total 501 1,666 2,694 2j179 1,874 1,970 Current Revenue (Ksh) Q4.30/kg cherry 1,806 4,440 7,600 11,965 18,135 25,585 Expected Revenue (Ksh) 6,020 9,783 13,921 18,060 22,575 30,100 Additional Revenue (Ksh) 4,214 5,343 6,321 6,095 4,440 4,515 Net Benefit (Ksh) of adopting input package 3,713 3,677 3,627 3,916 2,566 2,545 a/ The table shows how typical farmers in six yield level groups can increase their net revenues by switching to "proposed input expenditure" from "current input expenditure", farming practice. Source: J.D.M. Roe and J.K. Nyoro, ibid. - 14 - Table 9: KENYA COFFEE: GROSS MARGIN PER MANHOUR OF THE MAIN ENTERPRISES WITHIN EACH SMALLHOLDER ZONE (KSH) Zone Central Central Central Kisii Machakos Enterprise UMi UM2 UM3 Coffee 7.6 7.0 6.4 5.9 5.4 Maize & Beans 7.4 6.2 5.6 7.4 4.4 English Potatoes 4.4 3.8 3.2 - - Dairy 5.4 5.3 5.0 16.5 4.0 Source: J.K. Nyoro and J.D.M. Roe "Economics of Agricultural Production on Smallholder Coffee Farms in Kenya--Supplementary Analysis," CRF. calculated from their market prices but also on the basis of the cash-food requirements of the family. A smallholder farmer would be hesitant to grow only coffee, even if coffee prices were high, because with the present payment system he would not be sure when and how much he would be paid for his coffee and thus planning for basic living would be very difficult. On the other hand, as a farmer ne!eds cash to pay for school fees, clothes, etc., he would also be hesitant to grow only maize/beans. 1/ Other factors to be taken into account in analyzing competition for land between coffee and maize/beans are: (i) coffee is a perennial crop with a lengthy gestation period; and (ii) in Kenya it is illegal to uproot coffee trees. 2/ 1/ The very high real maize prices in the mid-1970s encouraged farmers to allot at least some portion of their land for growing maize. 2/ Although illegal, it is difficult to enforce this law. Uprootings of live coffee trees are known to take place when coffee prices are low. - 15 - (iii) Cross-Country Cost Comparisons 17. A comparison of production costs among major coffee producing countries was undertaken recently. The data are shown in Table 10. The figures under "cost per kg coffee" are slightly misleading as they do not take quality price premiums into account. To take quality price premiums into account, these cost figures were adjusted by adding 17% to Brazil, 10% to other arabica producers, 25% to robusta producers except Indonesia and 30% to Indonesia. These adjustment percentages correspond to the price differentials in international markets between Kenya's arabica coffee and other coffees. Kenya's relative production cost compares favorably in the column under "Comparable cost per kg coffee." 18. Cost comparisons with Brazil and robusta-producing countries are not directly relevant because coffee grown in these countries does not directly compete with Kenya's and because their production costs are lower due to the sun-drying process used (instead of washing). Among the washed arabicas, Kenya's production costs compare favorably with those of Colombia but not with Costa Rica. Part of the reason for the adverse comparison with Costa Rica is the higher labor requirement needed to produce high-quality coffee; but the figures suggest scope for increasing efficiency of labor in Kenya, especially in harvesting. (iv) Quality 19. Kenya produces the finest coffee in the world in large quantities. On average, Kenya's coffee fetches a 10% premium over standard arabica coffees of Central America and Colombia. Creat care is taken in picking, processing, - 16 - Table 10: PRODUCTION COSTS PER HECTARE OF GREEN COFFEE IN SELECTED COUNTRIES (1982) ------------------------------------------------------------__---------------__----------------------- Costs of Labor Costs of Overhead & Total Yield Cost Comparable (man- Wage per Costs Material Establish- Costs (kg/ha) per kg Cost per kg days) Worker (US$) Inputs Costs (USS/ Coffee Coffee /a (USS/day) (USS) ha) (USS/kg) (US$/kg) Arabicas Brazil 75 3.0 220 220 280 720 600 1.20 1.40 Colombia 150 4.0 620 200 520 1,340 800 1.70 1.87 Costa Rica 150 2.4 360 380 580 1,320 1,200 1.10 1.21 Kenya Estates 400 1.5 600 750 800 2,150 1,100 1.95 1.95 Smalgholders 220 1.2 270 220 280 770 600 1.30 1.30 Rwanda 275 1.2 340 190 290 820 700 1.20 1.32 Cameroon Arabica 90 2.0 180 70 90 340 200 1.70 1.87 Robustas Robusta 110 2.0 220 50 100 370 400 0.90 1.13 Cote d'lvoire 70 2.5 180 20 60 260 300 0.90 1.13 Indonesia 120 1.7 210 60 120 390 500 0.80 1.04 /a Adjusted to take quality price premiums into account. Source: J. de Graaff, op. cit. sorting, grading and classifying. 1/ Strict guidelines are observed at each stage. Coffee cherries are picked one-by-one and sorted depending on ripeness (classified to Cherry I, Cherry II and Mbunis) before pulping at co-operative factories. Co-operative factory washing and drying processes are carefully 1/ "Grading" in Kenya refers to sorting green coffee beans by size. This operation is carried out by KPCU. In general the larger the size, the higher the price fetched at the Nairobi auction. "Classifying" refers to labelling coffee on the basis of liquoring tests done by the Coffee Board of Kenya. - 17 - monitored by factory managers. Estates have their own factories, so they wash and dry coffee themselves and deliver it to KPCU. When delivered to KPCU, parchment coffee is hulled and the resulting coffee beans are milled and graded. Before auctioning the coffee at the Nairobi Auction, the Coffee Board liquors and classifies them into ten "classes." This "classification" forms the base to determine how much a particular lot and thus factory should be paid for coffee delivered. However, at auctions only "grades" are disclosed to traders. Traders undertake liquor tests to decide how much they are willing to pay for lots offered at the auction. The Coffee Board keeps track of the prices each "class" of coffee fetches at the auction and calculates payment schedules to factories and estates. Table 11 shows the CBK's payments for all "classes" of coffee for 1983/84. Table 12 shows CBK's payments for some aggregate "classes" of coffee in recent years. As these tables show, there are large price differentials among "classes." Thus payments to factories and estates also can vary widely. 20. Two points should be clarified here on the quality-payment relation- ship. First, no detailed analysis on "classes"-"grades"-"prices at auction" relationships have been made in Kenya. Such analysis would be helpful to capture the impact of trends in demand and prices on quality, which, in turn, would help Kenyan authorities in making future policy on payment of quality premiums in a dynamic world coffee market. Another point is that although Kenya's coffee marketing system is very quality conscious, there are very limited incentives for a smallholder to improve quality. "Average" prices are paid to factories and not to individual smallholders. Thus, a quality improvement made by a smallholder is rewarded only to the extent that his improvement raises the average quality of coffee his factory delivers to KPCU. - 18 - Table 11: KENYA COFFEE: NET PAYMENT BY CBK TO KPCU BY "CLASSES" OF COFFEE--1983/84 -------------_.-------------------------------------------__------------- Ksh/kg Ksh/50 kg -------------_.----------------------------------------------------__---- Clean Coffee 1 42.45 2,122.50 2 42.15 2,107.50 3 41.95 2,097.50 4 41.50 2,075.00 5 40.80 2,040.00 1-5 41.40 2,069.95 6 37.80 1,890.00 7 16.40 820.00 8 15.35 767.50 9 14.30 715.00 10 13.25 662.50 ----------_.----------------------------------------------------_ 6-10 29.00 1,450.04 ----------_.--------------------------------------------------__- 1-10 38.06 1,903.09 ----------_.--------------------------------------------------__- Hulled MBUNI I 32.55 1,627.50 II 13.95 609.50 III 13.15 657.50 ----------_.----------------------------------------------------_ 1-III 25.03 1,251.41 ----------_.--------------------------------------------------__- AVERAGE 36.65 1,832.68 -------------_.-----------------------------------------------__--------- SOURCE: CBK. 21. The "acid" taste of Kenyan coffee is much preferred by West Germans and Scandinavians. Due to its altitude and climate, Kenyan coffee is a unique kind with limited direct competition from other coffees, such as from Ethiopia, Tanzania and Papua New Guinea. It should be noted, however, that although some countries such as West Germany, Sweden, Finland and Austria - 19 - Table 12: KENYA COFFEE: NET PAYMENT BY CBK TO KPCU BY AGGREGATE "CLASSES"---1976/77-1983/84 Classes 1976/77 1977/78 1978/79 1979/80 1980/81 1981/82 1982/83 1983/84 1-5 Quantity (mt) 63,995 62,669 60,248 66,854 63,653 53,832 55,608 84,068 Price (Ksh/50 kg) 2,205 1,447 1,369 1,322 1,285 1,666 1,903 2,070 6-10 Quantity (mt) 28,921 13,268 7,506 17,192 23,911 22,690 19,340 30,945 Price (Ksh/50 kg) 1,684 987 1,208 1,035 692 1,023 1,500 1,450 MBUNIS Quantity (mt) 4,430 5,492 5,134 6,963 11,186 10,400 10,504 13,931 Price (Ksh/50 kg) 1,266 707 1,056 973 622 764 1,354 1,251 -----------------------------------------------------------------__----------__----------------- Source: CBK. prefer Kenyan coffee and pay high premiums for it, traders/roasters of other countries such as the United States or Japan do not value "acidity" much and are not willing to pay premiums for Kenyan coffee. 22. The premium Kenyan coffee enjoys over other coffees-depends on supply and demand. A statistical analysis shows that the premium has been strongly related to Kenya's export quantities. The larger the exports, the lower the premium. In elasticity terms, a 10% increase in Kenya's exports reduces its premium over the "ICO Other Milds Indicator" by 1.7%. 1/ 1/ The following regression equation was estimated to evaluate the impact of Kenya's exports on the premiums paid: log XUV/ICO = 1.20 - 0.17 log XKEN - 0.13 DMNQ (3.63) (2.77) R2 = 0.61 D.W. = 1.29 Period of Estimation: 1968-1984 where XUV/ICO = Kenya's coffee export unit value over ICO Other Milds Indicator Price. XKEN = Kenya's export quantity ('000 bags) DMNQ = Kenya's export quantity to non-quota markets ('000 bags). - 20 - (v) Government Policies 23. The government intervenes in the coffee sector in the following major areas: (a) gazetting of land, (b) control of planting materials, and (c) imposition of export taxes. Some details of each intervention are given below: 24. Gazetting of Land. Coffee is allowed to be grown in Kenya only in areas specified or gazetted by the government. The gazetting is a means to control quality and, to some extent, quantity. Although the gazetting makes it illegal to grow coffee in other areas, this law is not strictly enforced. Following the "coffee boom" in the late 1970s, coffee was planted in a number of "ungazetted" areas. 25. Control of Planting Materials. The government attempts to control production through its control over the availability of seedlings. A policy of restricting area expansion, and thus production, by this means was adopted during the period 1966 to 1972 and 1980 to the present when Kenya had to restrain production due to the imposition of the ICA quotas. 26. This measure, like the gazetting regulation, has not been very effective--especially in recent years. Even during years of tight control, farmers were able to obtain seedlings for infilling purposes which many farmers used to expand coffee acreage. 27. Export Tax. As shown in Table 13, the coffee export tax in Kenya is low in absolute terms; it is especially low compared with other coffee- exporting courntries. As the export tax schedule is progressive, the effective tax rate increases with coffee prices. Thus tax rates were 6-7% during the years 1978/79-1979/80 and over 9% in 1984/85. The high rate in 1984/85 was due to the high coffee prices in terms of Kenyan shillings which in turn was due to the substantial depreciation of the Kenyan currency against the US dollar. - 21 - Table 13: KENYA COFFEE: EXPORT TAX AND COUNTY CESS--1977/78-1984/85 1977/78 1978/79 1979/80 1980/81 1981/82 1982/83 1983/84 1984/85 --- -- --- -- --- -- --- -- --- -- --- -- --- -- -- I-------T- - - - - -----------------------__ _ _ _ _ _ _ _ _ _ …---------------------…(K. sh/kg of clean coffee)-------------------- Export Duty 1.32 1.79 1.57 0.75 1.24 1.91 2.32 4.30 County Council "cess" 0.74 0.78 0.87 0.63 0.83 1.05 1.10 1.39 Net CBK payment to KPCU 26.07 26.60 24.83 21.33 27.80 34.89 36.53 46.60 ------------------------------------------------------------_-_--__----------__----------------- Source: CBK. 28. Other government influences on the coffee sector include extension services provided through the Ministry of Agriculture and research work done at the CRF and control of imported quantity and pricing of fertilizer. - 22 - IV. ANALYSIS OF KENYAN COFFEE PRODUCTION 29. This section analyses coffee production in Kenya using the vintage capital approach. 1/ Since the estate and smallholder sectors are quite different, as discussed above, and since separate data on area and production are available, the analysis has been done separately for the two sectors. 30. Consistent long-term data relevant to coffee production in the two sectors are available only on production and mature growing areas. No data are available on new plantings and infillings. As information on new plantings is indispensable in the vintage approach, new planting areas were derived from the data on mature areas in the following manner. When mature area (defined as area under coffee trees of more than two years of age) increases from one year to another, it is assumed that new plantings equivalent to the difference between the two mature area figures have taken place three years previously. For example, if smallholders' mature area expanded by 15,167 ha in 1980/81 this increase in the mature area is assumed to be the result of new plantings of 15,167 ha in 1977/78. When mature area declines it is assumed that the trees were abandoned or uprooted in that year and no new plantings have occurred. 31. To calculate production capacity through the vintage approach, yield curves, i.e., changes in the average yield with age of tree, have to be provided. Yield curve data were obtained from agronomists at the CRF and from other coffee specialists in Kenya. Areas classified by age were multiplied by 1/ For details on the vintage capital approach, see, for example, T. Akiyama and P. Trivedi, "A New Global Tea Model," World Bank Staff Commodity Working Pa,per Number 17, March 1987. - 23 - the yield curves to obtain the basic production capacities. The basic production capacity variables were adjusted by the trend in yield over the last 20 years to take into account possible changes in yield over time. This adjusted production capacity, together with real coffee prices, was then regressed against actual production to give a relationship between actual production and production capacities and short-term price effects. Details of the main results obtained are discussed below. (i) Smallholder Sector 32. The estimated new planting acreage for the smallholder sector was derived from the mature area data in the way described above. As no statistically significant trend in yield could be found from this sector (see Table 2), all the increase in smallholder production was attributed to increases in acreage. Since the early 1960s, most of the area of mature coffee trees held by smallholders has occurred in two periods--the 1960s and from 1979 to the present. The acreage expansion in the 1960s was mainly due to the changes in the laws which allowed smallholders to grow coffee and to the breakup of some estates which were sold to smallholders. The second large area expansion occurred in the late 1970s as a consequence of the "coffee boom". The virtual stagnation in acreage in the 1970s before the "coffee boom" was not only due to the low real coffee prices but also to the restrictions which government put on seedling availability to restrict production in line with the ICA export quotas. - 24 - 33. Statistical analysis shows that smallholder new plantings are very sensitive to real coffee prices. 1/ The medium-term price elasticity (3 years) is estimated co be 3.6; i.e., a 10% increase in real coffee prices increases new plantings by 36% over a period of 3 years. This elasticity, however, should be interpreted with caution. The large new plantings which took place during the period 1975-1980 were caused not only by the high real prices but also by the change of law which allowed smallholders to expand acreage during the period. So it is likely that the estimate is based upwards. 34. The production capacity and real producer price were regressed against the actual smallholder production to evaluate the validity of the production capacity variable 2/ and to estimate the short-term price elasticity of supply. The short-term price elasticity of supply refers to the extent, say, of output increase when real coffee prices increase and farmers 1/ "Real" coffee prices here refer to the CBK unit payment price to KPCU deflated by the Consumer Price Index. Statistical analysis using maize prices as a deflator gave inferior results, although maize prices are strongly c:orrelated with the CPI (correlation coefficient of 0.97 for the period 1966-1984). The estimated regression: KYNPSH = 10142 + 132.9 CBKR + 172.6 CBKR-1 + 163.7 CBKR_2 + 106.1 CBKR_3 (6.31) (3.12) (9.12) (6.33) (4.77) R = 0.769 S.E.E. = 2606.0 D.W. = 2.31 RHO = -0.487 Period of Estimation: 1965/66-1981/82 where KYNPSH = Smallholders' new planting acreage (ha) CBKRt = Real coffee price (1980 K. Shillings/kg of clean coffee). 2/ The coefficient for the production capacity variable should be significant and be fairly close to unity, as indeed it is. - 25 - apply more inputs and improve husbandry to increase production. The short-term price elasticity is estimated to be 0.33 1/ which suggests that smallholders change yields substantially when prices change. (ii) Estates 35. The estates' coffee acreage declined fairly steadily up to the mid- 1970s (at the rate of 1.2% p.a.) which was the result of low and stagnant real coffee prices. The "coffee boom" in the second half of the 1970s reversed the trend and by 1984/85 the estates' mature area was 28% larger than it was in the mid-1970s. Statistical analysis shows that new plantings and acreage 1/ Period 1963/64-1984/85 KYPDSH = 0.92 KYPDCSH + 8.87 CBKR_1 + 134.6 DM6768 (14.3) (6.37) (2.03) -2 R = 0.891 S.E.E = 91.7 D.W. = 1.88 where KYPDSH = Actual smallholder production ('000 bags) KYPDCSH = Smallholder production capacity ('000 bags) CBKR = Real coffee price (1980 K Shillings/Kg of clean coffee) DM6768 = Dummy variable for 1967 and 1968 to account for the incidence of Coffee Berry Disease. - 26 - reductions by estates are very sensitive to real coffee prices. 1/ The results of the analysis suggest that the estates' acreage will remain stable if clean coffee prices are Ksh. 24-25/kg in 1980 prices. If prices are higher, area will be expanded and vice versa. An increase of one 1980 K.Sh. would increase area by about 70 ha. 36. Based on statistical analysis, the average yield of the estates has been increasing at about 2% p.a. if price, weather and CBD effects are taken out. (These effects are taken into account later.) This secular trend in yield was incorporated in the production capacity variable. 37. Results of statistical analysis also show that the production capacity variable captures the main production trend and that the short-term 1/ Two regression equations were estimated to evaluate the effects of real coffee prices on estates' acreage; one on new planting acreage and the other on an algebraic sum of new planting acreage and area reduction. KYNPE'S = -1410.0 + 34.7 CBKR + 34.0 CBKR.1 + 2009.3 DM80 (-8.61) (5.32) (5.23) (8.62) -2 R = 0.913, S.E.E. = 223.6 D.W. = 2.37 where KYNF'ES = Estates' new planting acreage (ha) DM80 = Dummy variable for 1980 Period of estimation: 1964/65 - 1981/82 KYUE'NPES = -1713.1 + 24.3 CBKR + 44.8 CBKR_1 + 2242.6DM80 (-10.23) (6.75) (9.41) -2 R = 0.920 S.E.E. = 228.7 D.W. = 2.01 Period of estimation: 1964/65 - 1981/82. where KYUE'NPES = Algebraic sum of new planting acreage and area reduction (ha). - 27 - price elasticity of supply is significant and is about 0.3 to 0.4. 1/ As in the smallholder sector, the short-term price elasticity is high. The similarity of the short-term price elasticities of supply for the smallholder and estates' sectors suggests that these elasticities are fairly robust. 1/ The estimated regression for the estates' production is: log KYPDES = 0.818 log KYPDCES + 0.371 log CBKR - 0.671 DM67 - 0.418 DM84 (13.5) (3.9) (4.14) (2.58) -2 R=0.723 S.E.E. = 0.157 D.W. = 1.76 Period of estimation: 1964/65-1984/85 where KYPDES = Actual estates' production ('000 bags) KYPDCES = Estates' production capacity ('000 bags). - 28 - V. PRODUCTION PROSPECTS 38. Basedl on the analysis described in Section III, coffee production projections are made in this section under several scenarios. In order to make projections, assumptions have to be made for exogenous variables. These variables include international coffee prices, the exchange rate of the Kenyan Shilling against the US dollar, the inflation rate in Kenya, yield trends and population growth rates. The following assumptions were made. (i) International coffee prices. Future international coffee prices, which affect Kenya's F.O.B. export unit values, are assumed to take the values as projected by the World Bank. (ii) Kenya's exchange rate and inflation rate: A statistical test was made to see if the purchasing power parity (PPP) assumption holds between Kenya's exchange rate and the world inflation rate proxied by the WJorld Bank's Manufacturing Unit Value (MUV). 1/ The results show that the Kenyan shilling depreciated during the period 1970-85 to adjust for the differences between the domestic and international 1/ The following regression equation was used in the projection exercises: log ( KYCPI * 7.4204 ) = 0.9705 log MUV KYER (189.9) -2 R 0.987 S.E.E. = 0.046 D.W. = 2.03 Rho = 0.503 Period of estimation: 1970-1985 where KYCPI = Kenya's consumer price index (1980=100) KYERt = Kenyan Shilling exchange rate against U.S. dollar (Number of Shillings per U.S. dollar) MUV = Bank's international inflation index. - 29 - inflation rates. As projected values for MUV are available, it was assumed that the PPP will hold in future. For the purpose of the current exercise, it was assumed that the Kenyan hilling exchange rate against the US dollar will be constant and Kenya's inflation index will increase at the same rate as the MUV. (iii) Kenya's producer prices. The "Base Case" projection assumes the present payment system to continue. A statistically derived relation- ship between the CBK's payment and the ICO "Other Milds" Indicator Price was assumed to hold in future. 1/ (iv) Yield. As noted in Section III, for the last two decades yield has been increasing at 2% p.a. in the estate sector while no increase has been observed in the smallholder sector. Projections of yields are difficult as the new high-yielding variety, Ruiru 11, is expected to be introduced in large quantities in a few years' time. Although the yield of Ruiru 11 can attain twice that of the present variety in experimental stations, little information exists on yields at the farm level. Discussions with agronomists suggest 50% above the 1/ The following equation was used in the projection exercises: log ( CBK ) -3.894 + 0.9926 log POMICO KYER (44.2) (49.7) -2 R = 0.991 S.E.E. = 0.058 D.W. = 1.80 Period of estimation: 1962-1984 where: CBK = CBK payment (K. Shillings/Kg of clean coffee) POMICO = "Other Milds" Indicator Price (USC/lb) - 30 - traditional variety to be a good estimate. 1/ The figure of 50% is assumed here. The future shares of Ruiru 11 in total trees in the smallholder and estate sectors are those assumed by Roe and Nyoro. 2/ Yieldls of existing areas are likely to increase with improvement in the standard of coffee management and with infilling by Ruiru 11. Table- 14 shows the expected increase in yield of existing coffee areas. The Table also shows the expected increase in yield for cases where the Ruiru 11 yield is 25% and 75% higher than the traditional varieties. (v) Effects of population growth. Population in Kenya is expected to grow at well over 3% p.a. in the medium term and by year 2000 to be about 66% more than in 1986, according to the estimates by the Central Bureau of Statistics. 3/ Population in rural areas probably will not grow as much as the total population but even a population increase of 44% in rural areas would have an important impact on coffee production. Possible effects include: A larger population implies more land will be required for housing and for other infrastructure such as schools and shops. This would reduce the available farm land including that for growing coffee. 1/ This figure corresponds to the "conservative" assumption in "Economic Implications of Introducing the New Hybrid Variety of Arabica Coffee" by J.D.M. Roe and J.K. Nyoro, Manuscript, CRF, 1985. 2/ Ibid. 3/ See Population Projections for Kenya 1980-2000, Central Bureau of Statistics, 1983. - 31 - Table 14: EXPECTED YIELD INCREASE FOR 1985-2000 /a ----------------------------------------------------------------__----------- Ruiru 11 Yield Smallholder Estate Improvement Base case (50%) 25% 75% Base Case (50%) 25% 75% ---------------------------------------------------------------------__------ …___________-----------…(% p.a)------------------------- Period 1985-1988 0.0 0.0 0.0 2.0 2.0 2.0 1989-1995 0.8 0.4 1.3 2.8 2.3 3.4 1996-2000 1.8 1.0 2.7 3.6 2.5 4.5 /a These estimates were made in the following way for the "Base Case": 1985-1988. No significant quantity of Ruiru 11 will be available. Expected yield increases are based on the historical trends. 1989-1995. Roe and Nyoro in their study 1/ expect adoption of Ruiru 11 by 1995 to be 10% and 15% for smallholder and estates, respectively. Ruiru 11 is assumed to yield 50% higher than the traditional variety. Thus for the smallholder sector, the yield index in. 1995 will increase to 1.05 (=0.9 + 0.1 * 1.5) from 1.00 in 1989 or 0.82% p.a. The gorresponding figures for the estates sector are 1.18 (=0.85 * 1.02 + 0.15 * 1.5) and 2.8% p.a. 1996-2000. Roe and Nyoro expect adoption of Ruiru 11 to be 30% and 40% for smallholders and estates, respectively, by year 2000. The yield indices for 2000 are 1.15 (=057 + 0.3 * 1.5) and 1.408 (= 0.6 * 1.02 + 0.4 * 1.5 * 1.02 ). For smallholders and estates, respectively, it is assumed that after 1996, yield of estates' Ruiru 11 will increase by 2% p.a. This gives yield growth rates of 1.8% p.a. and 3.55% p.a. for smallholders and estates, respectively. Source: EPDCS, World Bank. 1/ Ibid. - 32 - - Smallholder farm plots will be subdivided as they are passed from one generation to another. This implies that the average size of smallholder farm plots will decline even further from the current average of about 1.5 ha. - With less land per farm family, yields may increase as farmers s'trive to obtain enough production to support their food and cash needs. - A larger population may imply larger numbers of casual laborers which many estates and smallholders employ, especially during harvesting time. This may lead to lower wages for these workers, which could reduce the production cost of coffee. - Another possibly important effect is the share of land farmers will allocate to coffee as population increases. As the average land plot becomes smaller, farmers might allocate a larger share of land to maize to maintain their subsistence level. Or, they could allocate more land to cash crops to maximize their cash income. 39. It is beyond the scope of the current study to analyze all the effects set out above in detail. However, a cross-section analysis was made to evaluate the relationship between population density and area expansion. Population densities of eight major coffee growing districts were regressed against the percentage increases ifn- total smallholder coffee acreage of the corresponding districts in the late 1970s. The results of the analysis indicated a st-rong3 relationship between population density and acreage expansion; larger acreage expansion occurred in the late 1970s in the districts where population densities were lower such as Meru, Embu and - 33 - Machakos. 1/ (An exception was Kiambu which had high population density and large acreage expansion.) The estimated elasticity was -0.48. In other words, area expansion was reduced by 4.8% when population density increased by 10%, ceteris paribus. 40. This relationship was incorporated into the projections as follows. It was assumed that rural population density will be 44% higher in 2000 than at present and that 10% higher population density lowers area expansion by 5% in the smallholder sector, ceteris paribus. In other words, if the model projects, mainly because of increases in prices, new plantings of 3,000 ha in 1999/2000, the simulation run with the population effect will reduce new plantings to 2,340 ha (.=3000 * (1-.44*.5)). As it turned out, the simulation results showed no significant impact of the assumed increase in population pressure on coffee production through the year 2000. The production-depressing effects of increasing population density are likely to be significant, however, beyond the year 2000. 41. With the above assumptions, (i) to (v) above, the "Base Case" projections were made by first estimating new plantings and production 1/ The estimated regression equation is: ARX = 2.66 - 0.0044POPD + 0.704DM6 (11.14) (4.00) -2 R = 0.683 S.E.E. = 0.291 D.W. = 2.20 where ARX = Districtwise coffee acreage expansion between 1977/78 and 1984/85 POPD = Population density of districts DM6 = Dummy variable for Kiambu. - 34 - capacities. These were then used to calculate production for both the smallholder and estate sectors. 42. Table 15 shows the projected new plantings and production in both the smallholder and estate sectors together with some key assumptions. 1/ The "Base Case" run shows that Kenya's production by the year 2000 will be about 3 million bags or 50% higher than the current level. 2/ This increase translates to 2.8% p.a. growth for the period 1985/86-1999/2000. Mainly due to the assumptions on the adoption rate of the high-yielding variety, Ruiru 11, and thus on yield, production growth rates are not uniform throughout the period. Production is projected to increase at a low rate of 1.2% p.a. for the period 1985/86-1990/91 because the production in the base year 1985/86 was temporarily inicreased due to high prices and because low real producer prices in the following years are expected to keep production low. Production is expected to increase at the rate of 3.3% p.a. for the period 1990/91-1995/96 and at 4.1% p.a. for the period 1995/96-1999/2000 as the Ruiru 11 adoption rate accelerates (see Table 16). The *production share of smallholders is projected to remain about the same during the period (about 60%). 43. By the year 2000 smallholders' acreage is projected to be about 170,000 ha or 40% (48,200 ha) larger than currently, while that of estates is projected to be about 39,000 ha or 10% larger than the current level. The 1/ The projections were made using a spreadsheet program (Lotus 1-2-3) incorporating all the elasticities and coefficients estimated by regression analysis. It was possible, therefore, to perform sensitivity analysis instantaneously. 2/ The model has not taken uprooting or abandoning of coffee trees into account. The rate of these activities in the last 12 years appears to have been small. To the extent that uprooting becomes more important in the future the production projections will be biased upward. Table I1: KENYA COFFEE: "BASE CASE" COFFEE PRODUCTION PROJECTIONS--1985/86-1999/2000 Crop ICO MUV Exchange Kenya CBK Real CBK SmalIholders Estates Total Year "Other Milds" 1983=100 Rate CPI Payment Payment New Produc- New Produc- Production Indicator Price (1980A/. Planting tion Planting tion (US'/lb) (K.Sh./$) (K.Sh/kg) Sh.kg) (ha) ('000 bags) (ha) ('000 bags) ('000 bags) 1981/82 La 139.90 101.20 10.90 134.70 27.80 20.64 2,093 876 0 573 1,449 1982/83 La 131.60 100.00 13.30 150.20 34.90 23.24 2,758 874 98 550 1,424 1983/84 La 144.30 98.20 14.30 166.00 36.50 21.99 1,907 1,245 140 904 2,149 1984/85 La 145.60 99.20 16.40 187.40 46.60 24.87 2,978 1,079 194 482 1,561 1985/86 195.10 118.00 16.10 194.70 61.65 31.66 4,194 1,190 509 829 2,019 1986/87 142.90 119.20 16.10 196.68 45.25 23.01 4,595 1,271 437 754 2,025 1987/88 145.20 127.40 16.10 210.21 45.97 21.87 4,307 1,220 122 745 1,965 1988/89 150.00 131.80 16.10 217.47 47.48 21.83 3,457 1,247 84 760 2,007 1989/90 156.50 128.40 16.10 211.86 49.53 23.38 2,574 1,297 130 809 2,105 1990/91 160.60 131.40 16.10 216.81 50.81 23.44 2,660 1,359 177 842 2,201 1991/92 164.70 134.40 16.10 221.76 52.10 23.50 2,884 1,402 178 868 2,270 1992/93 168.80 137.50 16.10 226.88 53.39 23.53 3,024 1.439 178 876 2,316 1993/94 172.90 140.60 16.10 231.99 54.68 23.57 3,024 1,485 178 893 2,379 1994-95 177.00 143.90 16.10 237.44 55.97 23.57 3,008 1,537 176 927 2,464 1995/96 182.30 147.20 16.10 242.88 57.63 23.73 3,004 1.591 178 972 2,564 1996/97 187.50 150.60 16.10 248.49 59.26 23.85 3,012 1,662 183 1,012 2,674 1997/98 192.80 154.00 16.10 254.10 60.92 23.98 3,033 1,735 187 1,030 2,764 1998/99 198.00 157.60 16.10 260.04 62.56 24.06 3,057 1,809 190 1,056 2,866 1999/2000 203.30 161.20 16.10 265.98 64.22 24.14 3,073 1,886 191 1,103 2,989 La Actual figures except new plantings. Source: EPDCS, World Bank, "Kenya Coffee Production Model." Table 16: KENYA COFFEE: SENSITIVITY ANALYSIS ON KENYA COFFEE PRODUCTION, BY PRODUCER PRICES Real Producer Prices Crop Base Case 101 Higher 201 Higher 101 Lower 201 Lower Year Small- Estates Total Small- Estates Total Small- Estates Total Small- Estates Total Small- Estates Total holders holders holders holders holders -------------------------------------------------------------------__--------__--------------------------------------------------------- 1985/86 1,190 829 2,019 1,190 829 2,019 1,190 829 2,019 1,190 829 2,019 1,190 829 2,019 1986/87 1,271 754 2,025 1,271 781 2,052 1,27i 807 2,078 1,271 725 1,996 1,271 694 1,965 1987/88 1,220 745 1,965 1,241 772 2,012 1,261 797 2,058 1,200 716 1,916 1,180 686 1,865 1988/89 1,247 760 2,007 1,261 787 2,054 1,286 813 2,100 1,227 731 1,958 1,208 699 1,907 1989/90 1,297 809 2,105 1,319 840 2,158 1,341 869 2,210 1,274 776 2,051 1,252 742 1,994 1990/91 1,359 842 2,201 1,388 877 2,265 1,417 911 2,328 1,330 806 2,136 1,301 770 2,071 1991/92 1,402 868 2,270 1,440 907 2,347 1,478 945 2,423 1,364 828 2,192 1,326 791 2,117 1992/93 1,439 876 2,316 1,489 921 2,410 1,539 964 2,503 1,390 833 2,222 1,340 795 2,135 1993/94 1,485 893 2,379 1,551 943 2,494 1,617 992 2,610 1,419 845 2,264 1,353 807 2,160 1994/95 1,537 927 2,464 1,621 982 2,603 1,705 1,036 2,741 1,453 873 2,326 1,369 832 2,202 1995/96 1,591 972 2,564 1,693 1,034 2,727 1,795 1,095 2,890 1,489 912 2,401 1,387 868 2,255 1996/97 1,662 1,012 2,674 1,784 1,080 2,864 1,905 1,149 3,054 1,541 944 2,485 1,419 897 2,316 1997/98 1,735 1,030 2,764 1,876 1,104 2,981 2,018 1,179 3,197 1,593 957 2,550 1,452 908 2,360 1998/99 1,809 1,056 2,866 1,971 1,138 3,109 2,133 1,220 3,353 1,647 977 2,625 1,485 927 2,412 1999/2000 1,886 1,103 2,989 2,069 1,191 3,260 >,252 1,280 3,532 1,703 1,017 2,720 1,520 963 2,483 Source: EPDCS, World Bank, "Kenya Coffee Production Model." - 37 - share of coffee area held by smallholders is projected to increase to 81% by the year 2000 compared with 77% at present. 44. The large difference in the projected area expansion rates by small- holders and estates seems reasonable as the smallholders have the possibility of switching land currently planted to food crops and/or used for pasture to coffee; but there is extremely limited land available for the establishment of new estates or for adding more acreage to the existing estates. 45. Simulation runs were made to evaluate the effects of changes in real producer prices on production. Table 17 shows the results of the simulation runs with different price levels--runs with real producer prices 10% and 20% higher and lower than the "Base Case". These simulation results show that Kenya's coffee production is fairly sensitive to changes in real producer prices, especially in the long run. Long-run price elasticities, defined as the percentage change in production 10 years after the prices are changed by 10%, are 7.3% for the smallholder, 6.7% for the estates and 7.1% for the total. 1/ The long-term price elasticities of the smallholders are higher than those of the estates because the smallholders' new plantings respond much more to prices. 46. Table 17 shows the growth rates to be fairly high in the 1990s even in cases where producer prices are reduced by 10-20%. This is because of the assumption on yield which is greatly affected by the adoption of Ruiru 11. It is assumed in the model that the adoption rate will not vary with the real producer prices, obtaining an estimate of the price effect on the adoption rate has not been possible. If the adoption rates vary with real producer prices, which remains more realistic, the effects of changes in real producer 1/ As discussed in Akiyama and Bowers, price elasticities change with time. - 38 - Table 17: PROJECTED PRODUCTION GROWTH RATES WITH DIFFERENT REAL PRODUCER PRICES Real Producer Prices 10% 20% 10% 20% Period -Base Case Higher Higher Lower Lower -------------_.--------------------------------------------------__----------__ …(%---------------------(X p.a.)-------------------- 1985/86-1990/9i1 1.7 2.3 2.9 1.1 0.5 1990/91-1995/96 3.1 3.8 4.4 2.4 1.7 1995/96-1999/2000 3.9 4.6 5.1 3.2 2.4 1985/86-1999/2000 2.8 3.5 4.1 2.2 1.5 Source: EPDCS, World Bank. prices on procluction growth rates will be more pronounced than those indicated in Table 17. - 47. A sensitivity analysis was also done for the yield of Ruiru 11. The "Base Case" assumed 50% increase in yield. Simulations under the assumption of 25% and 75% yield increase were carried out. These assumptions translate to the overall yield increase, given in Table 18. The figures in Table 18 could also be interpreted as the expected yield increase with different Ruiru 11 adoption rates while holding the yield increase of Ruiru 11 at 50%. The "25% Case" corresponds to the case where the smallholders' adoption rates are 5% and 15% in 1995 and 2000, respectively, and when the estates' adoption rates are 5% and 13% in 1995 and 2000, respectively. Similarly the "75% Case" corresponds to the case where the smallholders' adoption rates are 15% and 45% in 1995 and 2000, respectively, and when the estates' adoption rates are 25% and 67% in 1995 and 2000, respectively. - 39 - 48. In Table 18 no significant differences in total production can be observed up to 1995. However, with the increase in adoption rates in the latter half of the 1990s, the differences become larger. Expected production differs by 5-6% in 2000 between the 25% and 50% yield increase assumptions and also between the assumptions of 50% and 75% yield increase. Table 18: KENYA COFFEE: SENSITIVITY ANALYSIS ON KENYA COFFEE PRODUCTION, BY RUIRU 11 YIELDS --------------------------------------------------------------------__-------__------------------ Crop Base Case (50%) 25% 75% Year Small- Estates Total Small- Estates Total Small- Estates Total holders holders holders -------------------------------------------------------------__--------------__------------------ 1985/86 1190 829 2019 1190 829 2019 1190 829 2019 1986/87 1271 754 2025 1271 754 2025 1271 754 2025 1987/88 1220 745 1965 1220 745 1965 1220 745 1965 1988/89 1247 760 2007 1247 760 2007 1247 760 2007 1989/90 1297 809 2105 1292 805 2098 1301 812 2114 1990/91 1359 842 2201 1350 835 2185 1369 850 2219 1991/92 1402 868 2270 1388 856 2244 1418 879 2297 1992/93 1439 876 2316 1420 861 2281 1461 891 2353 1993/94 1485 893 2379 1461 874 2335 1514 913 2427 1994/95 1537 927 2464 1506 903 2409 1573 951 2524 1995/96 1591 972 2564 1554 943 2497 1635 1002 2637 1996/97 1662 1012 2674 1612 973 2585 1721 1050 2771 1997/98 1735 1030 2764 1671 983 2653 1809 1077 2886 1998/99 1809 1056 2866 1730 1000 2730 1902 1113 3014 1999/2000 1886 1103 2989 1791 1036 2826 1997 1171 3168 Source: EPDCS, World Bank, "Kenya Coffee Production Model." - 40 - VI. PROSPECTS FOR KENYA'S COFFEE EXPORTS 49. Kenya's coffee sector is very much dependent on developments in the international coffee market as most of the coffee produced is exported. In this section, prospects for the international coffee market and the impact of likely developments in the world market on Kenya's exports and production are examined. 50. The short-run prospects in the international coffee market have become more uncertain following the suspension of the ICA export quota system in February 1986. The suspension took place following the skyrocketting of coffee prices on the news of a prolonged drought in Brazil. With the boom over, the quota system is likely to be re-implemented sometime in 1987. It is expected that a new international coffee agreement, with export quota provisions, will come into force after the present one expires in September 1989. However, the allocation of quotas during the remaining period of the current agreement and under the new agreement could well be different from the allocation between September 1980 and February 1986. Judging from the recent discussions held at the ICO, future quota allocations will likely be based more on recent production capacities; heavy penalties, probably in terms of quota reductions, could be imposed on countries which "undership" 1/ or sell large quantities of coffee to non-quota markets at prices substantially lower than those of the quota markets. 51. The projected world coffee prices used in the next section are based on the continuation of the export quota arrangement through the year 2000. 1/ "Undershipment" occurs when a country exports less than its quota. - 41 - Under this scenario, Kenya's exports can be projected by projecting the global export quotas and Kenya's export quotas as well as Kenya's non-quota exports. 52. Prospects for Kenya's coffee exports could be substantially different if the quota system is not implemented during the period 1987-2000. Thus two scenarios are considered: with the export quota system in operation and without the quotas. In neither case were explicit assumptions made about the occurrence of adverse weather in major producing countries, e.g., frosts or droughts in Brazil, but their implications are discussed for each scenario. (i) Prospects for Kenya's Exports under the ICA Export Quota System 53. As the quality of Kenya's coffee is very high, its demand is strong-- especially in north European countries. Kenya has had no export "shortfalls" and has been producing considerably larger quantities than its quotas in recent years. These factors will work in Kenya's favor in its representations to obtain higher quotas in the future. 54. Kenya's export quotas have been between 1,200,000 and 1,400,000 bags for the period 1980/81-1985/86 with the most recent quota being 1,330,000 bags. As quotas were suspended for most of 1985/86, Kenya exported a substantially larger quantity in that season than in the previous season. Assuming that the quotas will not be implemented soon, Kenya will also export a large quantity during 1986/87. If the quotas are reintroduced for the 1987/88 season, Kenya's allotment should be higher than for 1985/86 due to the factors discussed above and an expected increased global quota for 1987/88. It is assumed, therefore, that Kenya's 1987/88 quota will be 1,436,000 bags, or 8% higher than that for 1985/86. For the remainder of the current agreement,. i.e., until September 1989, it is assumed that Kenya's quota will increase at 2% p.a. In the likely event that Kenya obtains favorable treatment in quota - 42 - allocations beyond September 1989, it is assumed that its quota will be 1,500,000 bags for 1989/90 and increase at 2% p.a. thereafter under a new agreement. 55. The assumed annual increase in Kenya's quotas is considerably higher than Kenya's actual quota increases in recent years, which have been stagnant. This rather high quota increase assumption is based on another assumption, i.e., that quotas in future will be aLlocated more on the basis of production capacities of exporting countries and on the demand for specific types of coffee than in the past. Another consideration is that severe adverse weather in major producing countries in the future will lead to Kenya obtaining larger quotas. 56. Non-quota markets are not expected to expand as rapidly as in the recent past since the ICA restrictions on discounts and export quantities to these markets are expected to become much more severe in the future. Kenya's non-quota exports are assumed to increase at 4% p.a. for the period 1987/88- 1999/2000. 57. The lprojected coffee prices for the quota market under this scenario are the most recent projections of the World Bank. Export prices for the non- quota markets are assumed to be 25% below those of the quota markets. This price differential is due mainly to the difference in the quality of coffee exported to the two mairkets. (ii) Prospects for Kenya s Exports Wi-t-hout the Export Quota System 58. ALth ough` K'yan and world& )xport quantities will be larger in the event that where tQherea is no ICA quota system, world prices would likely be lower. How much--worbdprices would be lower in the absence of the quota system depends greatly on the assumptions made about the quota system. If, for - 43 - example, the price floor is set lower and the global quota set at a higher level, then in recent years the quotas may not have very much of a price- raising effect. On the other hand, if the price floor is increased and the global quota is reduced, the price-increasing effects of the quotas could be significant. For the purpose of the current exercises, two "no quota" cases were considered: one with price 10% lower and the other with prices 20% lower than in the "with quota" case above. The differences for Kenya between the existence or non-existence of the quota system are not only in terms of export prices and quantities but also in terms of production. If world prices are lower, Kenya's production will also be substantially lower as a result of the price effect. (It is assumed that all the exportable production, i.e., production minus domestic consumption, will be exported.) 59. An extreme scenario of very low prices could be considered. In such a case, the production cost data in Table 10 suggest that Kenya's smallholders should be able to survive but estate production could fall sharply. 60. Export projections under the scenarios (i) and (ii) above are analyzed in the next section. - 44 - V1. PROSPECTS FOR KENYA'S COFFEE SUPPLY/DEMAND BALANCE 61. The analysis and projections of Kenya's production and exports made in the previous two sections can be combined to project Kenya's coffee supply/demand balance. Table 19 shows Kenya's coffee supply/demand balance for the period 1985/86-1999/2000, together with projected export revenues. World coffee prices are assumed to be 10% and 20% lower when there are no quotas compared with prices when there are quotas. Domestic consumption is assumed to increase at 3% p.a. Estimated export revenues are calculated by multiplying world prices and export quantities. Real export revenues are nominal export revenues deflated by the MUV. 62. The projections of the Kenyan coffee sector shown in Table 19 do not take into ac:count the likelihood of future "coffee booms." From past experience, it could be assumed that some major weather abnormality would occur at least once before the year 2000 and this would imply lower stock accumulations by Kenya. However, "adding in" the likelihood of coffee price booms would riot contribute much to the relevance of the above analysis for coffee sector pricing policy per se. This is not to argue that such windfalls are irrelevant to policy but rather to suggest that they should be discussed in a wider context which includes exchange rate, money supply and trade policy, becau;e-their: impact is economy-wide. 63. The iollowing points can be drawn from Table 19: (i) As world prices are expected to decline in 1987 and 1988, Kenya's coffee export revenues are likely to decline sharply during this period from the 1986 level. In real terms, the coffee export revenue in 1988/89'is projected to be at least - 45 - Table 19: KENYA COFFEE: PROJECTIONS OF KENYA'S COFFEE SUPPLY/DEMAND BALANCE--1985/86-1999/2000 Crop Product- Quota Exports. Total Domestic Year-end Export Revenue Year ion Market Non-quota Consump- Stocks (Current (1986 '000 (1986 '000 /a Markets tion '000 USS) US$) Ksh) -------------------------------------------------------------------__--------__--------------------------------- …('-------------------------------- --------- ('000 bags)…--------------------------------------- A. WITH QUOTAS 1985/86 2,019 2,021 0 2,021 50 1,098 572,519 572,519 9,217,562 1986/87 2,025 2,021 0 2,021 52 1,050 419,339 415,117 6,683,390 1987/88 1,965 1,436 300 1,736 53 1,227 347,027 321,422 5,174,895 1988/89 2,007 1,465 312 1,777 55 1,402 366,584 328,201 5,284,032 1989/90 2,105 1,500 324 1,824 56 1,627 392,471 360,682 5,806,980 1990/91 2,201 1,530 337 1,867 58 1,903 411,867 369,866 5,954,836 1991/92 2,270 1,561 351 1,912 60 2,201 431,959 379,250 6,105,926 1992/93 2,316 1,592 365 1,957 61 2,498 452,771 388,560 6,255,814 1993/94 2,379 1,624 380 2,003 .63 2,810 474,327 398,084 6,409,145 1994/95 2,464 1,656 395 2,051 65 3,158 496,652 407,261 6,556,907 1995/96 2,564 1,689 411 2,100 67 3,554 523,217 419,426 6,752,765 1996/97 2,674 1,723 427 2,150 69 4,009 550,469 431,310 6,944,095 1997/98 2,764 1,757 444 2,202 71 4,501 579,023 443,667 7,143,036 1998/99 2,866 1,793 .462 2,254. 73 5,038 608,320 455,468 7,333,032 1999/2000 2,989 1,828 480 2,309 76 5,643 639,004 467,757 7,530,889 B. WITHOUT QUOTAS (World Prices 10% Lower) 1985/86 2,019 - - 2,021 50 1,098 572,519 572,519 9,217,562 1986/87 2,025 - - 2,021 52 1,050 419,339 415,117 6,683,390 1987/88 1,937 - - 1,884 53 1,050 357,454 331,080 5,330,388 1988/89 1,959 - - 1,904 55 1,050 373,216 334,138 5,379,629 1989/90 2,055 - - 1,999 56 1,050 408,753 375,645 6,047,887 1990/91 2,145 - - 2,087 58 1,050 437,963 393,300 6,332,125 1991/92 2,204 - - 2,145 60 1,050 461,624 405,295 6,525,250 1992/93 2,237 - . - 2,175 61 1,050 479,883 411,827 6,630,418 1993/94 2,279 - - 2,216 63 1,050 500,590 420,125 6,764,016 1994/95 2,340 - - 2,275 65 1,050 526,233 431,518 6,947,439 1995/96 2,416 - - 2,349 67 1,050 559,626 448,613 7,222,675 1996/97 2,501 - - 2,432 69 1,050 595,963 466,956 7,517,996 1997/98 2,567. - - 2,496 71 1,050 628,865 481,857 7,757,903 1998/99 2,641 - - 2,568 73 1,050 664,455 497,498 8,009,715 1999/2000 2,736 - - 2,661 76 1,050 706,902 517,459 8,331,094 contd...next page - 46 - Table 19 contd... ------------------------_----------------------------------------------__----__---------------------------------- Crop Product- Quota Exports Total Domnestic Year-end Export Revenue Year ion Miarket Non-quota Consump- Stocks (Current (1986 '000 (1986 '000 /a Markets tion '000 USS) USS) Ksh) ------------------------_.-----------------------------------------__--------__---------------------------------- ------------------------------------------------('000 bags)---------------------------------------- C. WITHOUT QUOTAS (World Prices 20% Lower) 1985/86 2,019 2,021 0 2,021 50 1,098 572,519 496,608 9,217,562 1986/87 2,025 2,021 0 2,021 52 1,050 419,339 360,076 6,683,390 1987/88 1,906 11,436 300 1,853 53 1,050 312,606 251,150 4,661,606 1988/89 1,908 11,465 312 1,854 55 1,050 322,972 250,816 4,655,406 1989/90 2,002 1,500 324 1,946 56 1,050 353,760 282,000 5,234,219 1990/91 2,086 1,530 337 2,028 58 1,050 378,403 294,758 5,471,011 1991/92 2,139 1,,561 351 2,080 60 1,050 397,875 303,007 5,624,132 1992/93 2,162 1,592 365 2,101 61 1,050 411,905 306,619 5,691,178 1993/94 2,187 11,624 380 2,124 63 1,050 426,527 310,504 5,763,274 1994/95 2,229 1,656 395 2,164 65 1,050 444,831 316,402 5,872,755 1995/96 2,284 1,689 411 2,217 67 1,050 469,412 326,402 6,058,354 1996/97 2,346 1,723 427 2,277 69 1,050 495,965 337,079 6,256,532 1997/98 2,391 1,757 444 2,319 71 1,050 519,444 345,242 6,408,055 1998/99 2,443 1,793 462 2,369 73 1,050 544,901 353,889 6,568,545 1999/2000 2,514 1,828 480 2,438 76 1,050 575,741 365,568 6,785,320 a/ Calculated by multiplying "1986 USS export revenues" by exchange rates given in Table 15. Source: EPDCS, World Bank, Kenya Coffee Production Model. - 47 - 40% less than that of the 1985/86 level, regardless of whether the quotas are reintroduced or not. (ii) As world prices and Kenya's exports increase after 1988/89, Kenya's real export revenues are likely to be on an increasing trend. The rates of increase differ by the assumptions made: 3.3% p.a. for the case with quotas; 3.5% p.a. for the case without quotas and 20% lower prices and 4.0% p.a. for the case without quotas and 10% lower prices. (iii) Kenya's total real coffee export revenue for the period 1987/88- 1989/2000 is projected to be, in terms of 1986 US dollars, $5.17 billion with quotas, $5.52 billion without quotas and 10% lower prices and $4.66 billion without quotas and 20% lower prices. In other words, the results show that Kenya's total real export revenues for the next 14 years would not differ very much whether there are quotas or not--if world prices decline by up to 15% in the absence of the quota system. (iv) If the quota system continues, Kenya's stocks accumulate quickly, especially in the second half of the 1990s when production is expected to increase rapidly. This situation is very likely if one considers that the current production capacity is about 2 million bags and the recent export quantities have been only about 1.5 million bags. - 48 - VII. POLICY IMPLICATIONS 64. Analysis of the Kenyan coffee sub-sector has revealed a number of important features of the sub-sector. The analysis also leads to several policy implicaLtions. 65. Optimal policies critically depend on whether some form of inter- national expoirt quota system continues. For example, as the results in Table 19 show, a continuation of the recent large production increase and a restrictive world export quota system (at least for Kenya) would lead to large stock accumulations, whereas little stock accumulation would occur if the quota system were to be discontinued. So, a first policy question is whether Kenya should siupport the quota system. 66. Simulation results shown in Table 19 suggest that future export revenues are not likely to be very different whether the quota system continues or not. It should be noted, however, that these results apply only if the assumptions made are realized. The simulation results and other analysis revealed that the following factors are critical in evaluating the effects of the quota system on the coffee sub-sector. - Restrictiveness of the quota system - Kenya's future quota dallotment - Kenya's future production increases - Cos;ts. oĢ-.ma--intaining-st.ocks.. 67. The ideal: situation tor Kenya would be one in which the quota system is sufficiently restrictive to support world prices and Kenya obtained increases in export quotas which by and large corresponded to the increases in - 49 - production, so that no large stock accumulation occurred. If Kenya's quota increases are small compared with its production increases and large accumulations of stocks occur, the quota system would be of little benefit to Kenya's coffee sub-sector. It should also be noted that the operation of the quota system could exclude the possibility of Kenya expanding its robusta production substantially as this could push Kenya's production level considerably above its quota. Close monitoring of future developments in the world and Kenya's coffee markets should be observed to evaluate the effects of the quota system on Kenya's coffee sub-sector. The Kenya coffee production model developed here should be a useful tool to carry out the evaluations. 68. Simulations of the Kenyan coffee sector also show that if Kenya's production increases and its export quota growth is low, Kenya could accumulate large quantities of stocks. In such a case, the current producer payment system would cause the Coffee Board to accumulate large debts. 1/ In that case, the CBK may consider changing its payments system. Under an alternate system, the CBK could pay producers only from the money received from coffee sold. This would cause producer prices to be lower but would not lead to borrowing by the CBK. 69. Coffee competes for land use with food crops in the smallholder sector. This competition could become much more severe with the expected increases in population. Rapid increases in food crop yields, especially of maize, are necessary for coffee production to increase. Otherwise, population pressures could force farmers to divert land from coffee food crops. The 1/ Under the current system, all the coffee produced, including that held in stock, is paid for at market prices. This system forces the CBK to borrow to pay the growers for coffee that is stocked. - 50 - government imlposes restrictions on domestic maize trade, i.e., on prices and on inter-district transport. More liberal policies on maize trade would allow farmers greater freedom in deciding which crop to grow and in what proportions on the basis of profitability considerations.