Burundi - Arabica Coffee Improvement Project

Document of The World bank FOR OFFICIAL USE ONLY Report No. 1 531 Project Performance Audit Report BURUNDI ARABICA COFFEE IMPROVEMENT PROJECT (Credit 147-BU) March 21, 1977 OnAratinna Vvaliutt-4nn -nav- thirof d ies. I contents maynotothr be dosed wBrecipients a t In io their o0ffi_cialduties. itscontentsm;ay_not; oth;erwise -be -dsclosed without World Bank authorization. Currency: Franc Burundi Exchange Rate: until December 1972 : US$1 = FBu 87.50 Ta..... 1070- U a 1 18074 *T - Tc1 =D.. 70 Iq .jLAL L Lt rJ LA A IUv * - 1- I May 1, 1976 - present : US$1 = FBu 90.00 (NB: The exchange rate US$1 = FBu 78.35 has been used in this report). ABDREVIAIUD .01nDU Iustitut nes Scieuce agLULUUm1qUCb UU DULUILUL (Agricultural Research Institute) OCIBU = Office des Cultures Industrielles du Burundi (Agency for industrial Crops). FOR OFFICIAL USE ONLY TABLE OF CONTENTS Page Preface Basic Data Sheet Highlights PROJECT PERFORMANCE AUDIT MEMORANDUM I. Summary of the Project Completion Report i TT. ORD Commentn General iv 7hp Prniet'A Ficnerimental value v Socio-Economic Environment vii Effort4vanpsnof tha Vwtanainn Qprv4ra 74 4 4 Importance of Price Policies ix An nVArall FAVOrshla Tmpr-asion The Income Issue PROJECT COMPLETION REPORT Table of Contents I. Background 1 II. Identification, Preparation and Appraisal 1 TTT Tma n4,n A IV. Organizational Aspects 22 Ill D.--4,4 ojt Jt I pctjL .- VI. Government and Bank Performance 28 113"r1r. Change sLn the .. Secon in V L.e HO L 1 L%C OCUUa U rLVUJ=L ;L Annexes 1-3 m.1- 1 /. Chart Maps innumunwmnanar=nncwuounnouuonanomymsedenyrecipantsoxyinthepetnene of their official duties. Its contents may not otherw be disclosed without WorMd Bank authoriastion.  Project Performance Audit Report BURUNDI ARABICA COFFEE IMPROVEMENT PROJECT (Credit 147-BU) PREFACE Credit 147-BU for US$1.8 million was signed in April 1969 in support of the Government's coffee improvement program, and closed, fully disbursed, in June 1976. It was followed by a pAcon e gmgy Meme pL ojec t- pr'iv 17 a-(e-dCF ,~L~p.I$ by QI ion, signed in DecembR T . e audit report deals with the first project, though some comments are offered on the role of the second. The bulk of the report is composed of the Project Completion Report, issued by the Eastern Africa Regional Office in September 1976. The PCR's description of the project, and analysis of effects at the levels of the farm and washing stations, is outstanding. The summary prepared for the PCR appears first in the covering memorandum, and is followed by an OED discussion of some interesting and important issues that are illuminated by the PCR. This OED short audit was based on a review of the two appraisal reports and twelve supervision reports, in addition to the PCR. Judging by that limited material, OED finds the PCR comprehensive and accurate.  ARABICA COFFEE TMPROVPNT PRTvT fw CnTrT 1A7.nmr A. Amounts (in US$ million) Outtunulg Original Disbursed (as of 6/30/76) Credit 147-BU 1,800 1,800 1,900 - Foreign Exchange Adjustment 100 ) B. Project Data Original Plan Actual First Mention in Bank Files 1966 Government's Application 1966 Board Approval 0W/08/69 Credit Agreement U4/11/69 Credit Effectiveness 07/15/69 Last Disbursement 03/18/76 Credit Closing 12/31/73 06/30/76 Total Costs (US$ million) 2.1 2.5 Cj coffee growing: 15% coffee processing; less than 10% C. Mission Data Origin Mi -in M-+n+h/Ye. No n P.-n- No. f Weeks Man Weeks Date of Report Preparation FAO/IBRD July 1967 3 3 9 Appraisal HQ March 1968 4 3 12 10/22/68 Technical Assistance HQ March 1969 1 1 1 04/01/69 Total 3-5 Supervision 1 HQ July 1969 2 1 2 10/22/69 Supervision 2 -Q jun 197 2 8/87 Supervision 3 HQ Feb. 1971 1 2 2 03/11/71 Supervision 4 KQ Dec. 1971 2 1 2 01/18/72 Supervision 5 HQ April 1972 2 1 2 06/02/72 Supervision 6 RMEA Sept. 1972 1 1 1 09/22/72 Super- s-; 7 MMtA V-~ Inl7O 0 1.- -0 Supervision 8 RMEA Aug. 1973 la/ 1 1 08/22/73 Supervision 9 RMEA Feb. 1974 lb/ 2 2 03/12/74 Supervision 10 RMEA Aug. 1974-eI 1 1 1 09/27/74 Supervision 11d/ HQ May 1975 1 2 2 07/01/75 Supervision 12 RMEA Dec. 1075 1 1 1 12/2hi/7 Total 22 Completion HQ April 1976 1 1 1 D. Follow-up Pro.iect Credit 593-BU of US$5.2 million, signed December 5, 1975 for Second Coffee Improvement Project. a/ Accompanied by another RMEA staff member to assist Government with the preparation of the Phase II Project. b/ Together with one RMEA staff member and one consultant to give further assistance for Phase II preparation work. c/ The second project was appraised in October 1974 by a four-man mission which staved about three vppk. in Rn-I d/ Combined with further discussions on the second project.  nTTnTKMT ADAtTrA rOVWT TMPZUITMNT PROlTF.rT tr,.Af 1A7_-RTT) HIGHLIGHTS n- e.A audi reviews. progresso uinder t-he A-.1-4c o ffee improvenment ILL~ UU.L 9L I. _ - _ V k'~j..'J~. '~f'4. S.* Project, which is Government's main vehicle to increase coffee productivity. ±IL prisio o.VL UlL L1e-n-dLca.L Lerlies as well~J as imrvmn ofJ ote culti vation Practices,, were the main elements of the project; the improvement of cuLLee quality through Lue wet pruebing uLeLtLU W LLIC ULLM1 LinJU1 pLlJuJo.L component. The agricultural part of the project was implemented almost as planned; the washing stations component, which had been added at appraisal, was implemented slowly. Lack of reliable data makes it impossible to assess the project's impact on the volume of coffee production, but it is unlikely that more than one-third of the appraisal's overly optimistic target was achieved. Improvement of coffee husbandry practices should be regarded as a profitable undertaking, but serious doubts remain with regard to the profitability of the washing stations. An effective coffee extension organization was eventually estaD- lished and, under the second, follow-on project, has been integrated with the general extension service. The following points may be of special interest (single and double digit numbers refer to paragraphs in the Memorandum; triple-digit decimal numbers refer to paragraphs in the PCR): - the Bank support to exploratory projects (paras. 7 and 8) - the project's experimental value; role of the Government and role of the Bank (paras. 9-11) - socio-economic factors ignored by the Bank (paras. 12-14) - advantages and drawbacks of project organization; effective- ness of the extension service (paras. 15-18, and 4.01-4.04) - importance of price policies (para. 19) - incentive value of the low incremental incomes provided to participating farmers (paras. 21-24, and 5.09) -- Uank 4nvolvment in ~rMjeCt prpaato (annr"0 9 l -9 R rolean ,n@.. of4epatr 4te inprojae mfanhaement (pmaras. 3 .23.5) - factors affetin the efcientu and profiblyof the washing stations (paras. 3.32-3.39) Pvn4ert Ppvfnympnc Audi-t ----A 'RTTO~TTK1TT APAR~TrA rnrrvV TVfPnl7rVMT P'PnTT?rT RUNDI 1-0-0 CO M E I- R M N - 0-E M I. SUMMARY OF THE PROJECT COMPLETION REPORT 'T , A,-aI64C. r,-f-F-p T-rvmn Project - ,.,.,,4. .. foermn in 1964 or 1965, identified by an FAO/IBRD Cooperative Program mission in Apri1966C, prpae by another 1 ~- - - Prga sin --------I-In July_ _ 1f-C and_. appraised in the field by a Bank mission in March 1968. The project under- went substanotl changes at all stages between conception and appraisal. Government at first was thinking in terms of securing financing for the pur- chase of 10,000 ton fertilizeL aHnually for a country-wide coffee fertiliza- tion program. The final project, as it emerged after the appraisal, was restricted to th e better half of ----I± rroinc copisn abou -- scotsUu &_ LuLI uLe HLl L Oi6uinl Province comprlsing aDOUL 44,UUU smallholder coffee growers with about seven million trees, or about one sixh ofC Burund 's total area under coffee. The provision of chemical fertilizers was still an important element of the project but improvement o vtner cultivation practices, such as pruning..mulching and insect con- trol, were other features. The improvement of coffee quality through the wet DrocessInz method was included in the project as another major component by the appraisal-team. An IDA credit (147-BU) of US$1.8 million was approved in April 1969 and became effective in July. The project started on schedule in mid-1969 but was completed two years behind schedule in late 1975, and this despite the substantial reduction in the project's size during prepara- tion and appraisal. It appears that the late start of the processing compo- nent, consisting of the construction of four coffee washing stations. was a major reason for the delayed completion of the project. The civil distur- bances of 1972, changes in project leadership and continuous transport problems between Burundi and the Indian Ocean ports slowed progress as well. 2. The agricultural part of the project, in effect the improvement of coffee husbandry practices, was implemented along the lines indicated in the appraisal report. In particular, mulching and pruning of the coffee, trees has much improved and should.heregardedas thp main accomDlishment of the project. The application of chemical fertilizers in coffee was also implemented in accordance with the original plan, but because of the lack of a reliable data basis the effect of fertilizer applications on coffee production can still not be satisfactorily quantified. Common sense, based on experience elsewhere, and the results of trials conducted in the 1960s, suggest that nitrogen must have a positive effect on Burundi's coffee yields. But it is important that the use of fertilizers be com- bined with high standards of coffee husbandry. The profitabilitv of fertiz lizer depends very much on the prices of both coffee and fertilizer- Both prices were subject to large tiuctuations in the last few years and under - ii - present conditions (normal fertilizer prices and extremely high coffee prices) there should be little doubt that the application of fertilizer in coffee is justified. The application of insecticides on coffee somewhat improved under the project but serious questions have been raised (and are not yet answered) with regard to the types of insecticides to be used. 3. The coffee processing component has been the subject of very inten- sive discussions between the Government and the Bank. The Government attaches great importance to the construction of coffee washing stations, but the Bank was not convinced that the undertaking is a profitable one. Firstly, the construction of the stations turned out to be four times more expensivg than anticipated during the appraisal. Secondly, the stations were operated far below their designed capacity during the initifyeas; in 7,'and, par- ticularly in 1976, the performance has improved markedly, however, and throughput no longer is a major concern. Thirdly, and most serious of all, the fully washed coffee has so far failed to command a price premium suffi- ciently large enough to justify its production. The possibility of realizing higher premiums in the future, when Burundi would produce larger volumes of fully washed coffee and when the presently distorted world coffee market would return to more normal conditions, should not be ruled out, but the experience of the last three years does not hold out much hope. The Burundi Government has repeatedly pointed out that the impact of the coffee washing sttion pogrm is greater than can be measured in strict economic r finan- cial terms. The stations can serve as focal pMWfff! I'-e RuerE rural development activities, for example demonstration of better growing practices for both coffee and food crops and sales of farm inputs and con- sumer goods. 4. Lack of reliable data makes it impossible to assess with a reason- able degree of accuracy the project's impact on the volume of coffee produc- 1ion, but is cerain that araisal estimgates gte too optimistic. It is unlikely that more than one third of the appraisal target was achieved. Export Rrices for coffee always exceeded the appraisal estimates, at Present a tleast b afcoof three. In contrast with the wet processing component of the project, which seems a marginal economic proposition, improvement of coffee husbandry practices should be regarded as a profitable undertaking. Average farmers' incomes are estimated to have increased by a modest UT.l per household only, but the number of beneficiaries is very large. 5. Despite serious uderstaf fin of the project with expatriate per- sonnel during part of the project period. and the adverse effPcfR nf the 1972 civil disturbances, the project has in the course of its final years succeeded in establishing an effective coffee extpnAion nannisn4nn Government has consistently displayed a keen interest and staffed the proiect with Personnel in a manner whirb chuni1d hi ronrandf no a taking the severe limitations of the country's human resources into account. Since the Droject was nnnePrned with coffee vmen only the Mi y o Agriculture's general extension service remained in poor shape, because of - iii - lack of resources and proper guidance, and could accomplish little. Under of Ngozi, both extension services have been integrated. The second project, S07=-w-hich. an TTNA CreAl /CO I_DTT1 4 TTroQC . I 1J a 1s 1 - 1- Y L &Us .n-- & wan .Ur % .JJ"DU} V.L UayJ.f million was apprUVeCU u LLe DULu in November 1975, aims at a more balanced development of coffee and other agL ililtua.L M-C-161Vi.-L-Ces LLALIUU6LL Ole incLLlusion oL a .LUUU ULUP ULU1PULLCLLL anIU experiments with other rural development activit iimprove- ment of avyZe rouL.Un. A larger provision ror corree researcn particularly in the field of insect control, has also been madeA wile a small team of internationally recruited advisers would visit Burundi at regular intervals to assist in formulating work programs and advise on specific project problems. - iv - II. OED COMMENTS 6. The audit finds the PCR to be a balanced description of the proiect. The analysis of the attributable economic effects at the farm level, an analysis carried out with little field information, is outstanding. These comments focus on some issues, explicit or implicit in the PCR, to which OED wants to attract special attention. General 7, The smallholder project was presented in 1969 as a profitable rnffpp imnroverent- program, bAsed p-rimarily on nit-rogen feriliza_tion, with an economic rate of return forecast at 32%. The PCR now forecasts the ratp of rpfivirn to t-hp fiPlr pnTnr%nnunt, of the ojc atrO5, nd, to the factory component 9%. The PCR's calculations are based on conserva- may have been offset in part by an assumption about annual benefits continu- C,~m-nIh -_i,v. a -_-,wi Cot O -LJ"' believes t- bebt uVSS t an overall project rate of return should be about 10-13%. That is not an unim- pressive range for a qmallhol1er nrnipnt hiEV 4t h4Aa aho t sfne- *h-4n 4m-- fect farm data presently available suggests that the coffee yields and incre-me-ntal prneiit-t-cinn Tnny 'haie reached only one-third their appraisal estimates, now thought to have been too optimistic. Whatever semblance of in_-___a roftblt that-- accue to the project would be due then to an unexpected increase in export prices. In short, the PCR is not dsrbn a peroven, profitable coffee -- p-m a L LLHPLUVf=L11U1AL JJEOgLalllIUatUall, ana makes no such claim. 8. What is described is a modest, though essential set of activities - some tt, sme new!:- focud un deveLoping cost-effective methods for modernizing Burundi's premier cash crop (80% of exports). Given the Burundi omanuler a traditonally low yields (one-third those on some of the few large coffee estates in the country), and the poor competitive value of his parhmMelt,.t, quality improvement of the cottee crop was and should have been 1/ Similar combinations of such treatments had been tried and proven under S ular ecological conditions elsewhere in Atrica, in Kenya for example. ti ine rollowing terminology is used in this report: Cherry, the complete fruit of the coffee tree, should be picked when it turns red, and nor- mally contains two beans. These beans are individually covered by mem- branes, the silver-skins, and parchment-like shells, all covered by a mucilaginous substance. Immediately surrounding this is the skin of the cherry. Skin is removed in the pulping process, and mucilage during washing and fermentation. Pulped coffee, when dried, is called parch- ment coffee. The parchment and silver-skins are removed in hulling factories, which produce clean or green coffee. All coffee trading is in green coffee, which is eventually roasted before consumption." Source: Appraisal Report for the Second Coffee Improvement Project, October 30, 1975, p. 4. - v - an objective of highest priority for the Government. An exploratory nro4A^11- elf f-14. -4 ---A-A U.,,14- -I .n-- -4: return., forec.ast to be 4-ti fied. The Bank has to learn how to be comfortable with giving more weigl t, to t*, e l-kar "ng.l -'ue of. th-1, 4ro et and- less 1 o t-he economic .orecasts, and to help design and supervise the projects with those syiTnh en -4nA experience. In important respects its experimental nature was not reflected 4" U. . -tuA - e vis o a.nA .LtJZ- Aate -- reacC Uit eAL AO w, 4-t.UcS -L +-L knowing which components of the original package worked and which did not andA wh.is T- particu.lar theret s no convinc--- that --4- --- J~AA -.~L 1- -AL -LL inUg LUL L I t.ir respLose to nitrogen boosters, given the kind of prices Burundi expects to pay and WULLL & LL LC 1HU.LV UAL uIs llhoULde or to tUe economry. Whatever rise in average yields was reported in the project area, and there is evidence that yields auvanceU relative to otner corree regions, is due to some mixture of improved husbandry practices - including fertilization, punng, muluhig, anU insecticUes - whic cannoT De separately evaluated from available evidence. Moreover, as shown in the appraisal report of the secunu project, whatever has been earned by those farmers of the first proj- ect who continue the new practices is the sum of the effects of increased cherry qualities and yields. The quality improvement affects incomes only when the cherries are sold to the stations; it is worth nothing when the coffee is sold, as most of it is, withintraditional marketing arrangements. If the yield estimates presented in the PCR are reasonably accurate (some comaentators feel the estimates both of present yield increases and of future levels are too pessimistic) then the yield improvement is small enough in aDSOlute value to be almost matched byan alternative strategy of expanding h small plantation onto marginal land (for farmers with a choice). Even if the yield improvement is somewhat better than the existing data indicates, it is difficult to see how the project extension officer can convince the participating farmer that he ought to persevere. The present situation makes one wonder whether the PCR does not demonstrate that the smart farmer has been the one who planted additional accessible acreage to coffee, once the new supplies of free seedlings became available in 1974,, but otherwise did nothing new during the project period unless it was free, except perhaps to continue the pruning practices he had learned under a smaller, predecessor project in 1967-69. He did not pay for fertilizer, did not use more pesticide, and did not sell to the washing stations unless forced by Government to do so. And he does not now intend to start doing these things, even when he runs out of marginal land. That model is not an encouraging one for the future, and we do not pretend that it is correct. But the Bank cannot show the model is not correct. If the research components of the project as appraised had been pressed vigorously, - vi - and if the warning of the second supervision mission (June 1970) about the 4prance of corecing th rel Aar.-n a of -1a.- 4 14 --F4nl. I-Iian .naros had been followed up in subsequent supervision, some of the researchable questions raisedA by appraisal may by now ha-"-V- U-- --A 1/ T-+-4aA they are to be investigated in the second project./ 10. This comment on the low experimental value of the first project dosnot apply to all ilts aspects. r_TIA th rspect to tI1e techLILUU 3 economics of the four coffee washing stations established by the project unit an enormous amount us iterest and useful analysis nas veen gener- ated, as discussed in the PCR. Also, some of the technical issues of tue farm improvement program - ror example tne errects ana siae errects of chemical pesticides - have been investigated to the extent project finance and complementary Government services permitted. Much more work of this sort could have been financed (the Government's agricultural research agency, iSABU, had only one plant pathologist to serve the country and all its crops), but the issues were not neglected. What seems to have Deen neglected in particular was the economics or the farm package,,ana the other influences bearing upon the smallholder's decision to adopt. 11. The previous paragraphs carry an implicit criticism of project supervision. In fairness to the supervisory staff itself three qualifying comments are in order. First, staff constraints at the Bank's regional mission in Nairobi did not permit exhaustive field3 nalysis. Second, the project financed research equipment and materials,- but not research 1/ This is not to belittle the difficulty of coffee research, given the long gestation period and problems of enforcing farmer trials throughout the period. 2/ The Bank can be credited with having greatly reduced the fertilizer com- ponent at appraisal of the first project. The original proposal called for 10,000 tons distributed per year, or 67 kg each for 150,000 farmers. The appraised project called for 500 tons, or 11 kg each for 44,000 farmers. It is not known whether this reflected doubts in the Bank about the economics of fertilization. The PCR suggests it was done partly to make the project more manageable (paragraph 2.07). In any case the Government always referred to the proposal as a fertilizer campaign (project d'application d'engrais). 31 In fact, the project did not even finance research equipment and materials (although envisaged) since Government was not aware that provision for this had been made in the Credit Agreement (see 3.42 of PCR). - vii - staff. At the same time, ISABU's limited budget and staffing were not commensurate with the needs of either this proiect or the country as a whole (which points to a possible flaw in the project design). Third, supervision must respect the appropriate balance between technical assis- tance and direct involvement in project management: it would have been inappropriate to intrude far into the latter area. Socio-Economic Environment 12. Within the socio-economic enuironment tht AhnnApm the nntPvn nf the farmer's response, a number of factors that are absolutely essential to understanding him hAhauinr wmild annnr tn han been 4gnovaA by annpr- vision and other commentators on project developments, at least as recorded bI the -oi n1~ parntpiular th connection between. the farmers and the traders, including the roles of the extension officers in the delivery of~~~ esicde a1zp ndr t~ he saeo ofe snowhere well explained. The only significant reference to the give and take at this "int-Prfnt-P1 is In the i&i+ 1Q7/iIo pr. M- .-~"- is made to expain why the farmers were not bringing the cherries-to-thenew, _.Q_a____42_t*_A__4_e%_n_ rioi,.'= -+16 L - L ..J Em D ... L CL.LLLUt_LLL p 9U£J- By hand on the farms or at the small, communal and pulping stations as was tf 4 1- 0 - 1 _^~ 1i 4-A,.., ff~ *~L L L~~ L 5C i lpoQnts on designated Coffee-tr-ading days. The report said that the.tralders, r;:77:6 nmW sor -- -o. ote offeredU -UVM-nC=-S tO LLLC 'LULIUt-L: S ssouUI"4 Bank staff feel'this description may have been incorrect) and had shared 5m L ILeMUu witu respct to picLig L[e crop. iL aiso-mentions that the extension ge__t themselves were involved-in the trade - "one incentive of an wouhewise puutty PAtu 0I U Heweb of traditionLrela- EtRps In E =T6cTf 'economy Tn be expected to r tnew patterns of -r-0r--- Iusuues eatend from the towns too. Several supervision Y siake brief mention of the political power of the traders' organi- vA., tLlu =mumaLun ue Commercanrs Barunaais, set up after independence to promote Burundians vis-a-vis non-indigenous traders. However, what the upJ.JaLLULLun LLoJ5 power nave Deen for project success has also never been carefully examined, as far as OED can tell. (The Association itself an ionerex4 a4±Cs . .the problem is not just that the traders can fight the washing stations. The problem is also that the financial incentives one would 11wI pLov±ueU for quality improvements in coffee husbandry are not peen n price paid for parchment by traders. Government sets the local marKeL price ior traders.tUDscounts in the tield reflect only grees faults such as excessive stone or water. Cheating is reported to be common. 14. How the increased wealth flowing into the project area of Ngozi province, wealth attributable to improvements in quality and quantity of - viii - coffee deliveries, is shared among the farmers, the traders, and the exten- sion agents isalso no exnrn1enA ThORA rosanrrhal mnatters that should be included in the project definition. They are not mentioned in t-he first4 apris4al repr, so~ supevision in, a sensea did not fail to follow through. But another function at the "interface", the fertilizer of a fertilizer sales program") and there has been no progress to date by the project unit t1oward creating a- self-sustaining "ivte griultra supplies system and no comment on that fact by supervision, even though ment. It would be interesting to know for example whether the traders or indeed have basis to hope to harness them as salesmen. To date OCIBU has ruAnte Suppflies.L~ syCsem, g.L 4LL L*Le fertlize fo free. LthroughNJS5L 1 an,.. charging subsidized prices thereafter. The project cries out for careful socLLU.iLcaIlwe aD wel1. as ConUomi-L.. LIHVCbLgLakLLUL-- U. LLLC ULLLLLU LoL.LUeImaL UL CH coffee economy: investigations of all factors that seem to influence small7 LueUe CLV.LUL igirL.Li.LnLLLLy, .LfuctuULig, iL LL LN LIm1pUrLeLIL, LLIC LrLVUIL question. Effectiveness of the Extension Service 15. Another part of the micro climate that is not well enoughIreated LI supervi5ion reports is the efficiency of the extension service. The fact that the position of Training Officer was not refilled either by another expatriate or by a Burundian, arter the first expatriate incumbent was promoted to Project Manager in early 1971 has drawn the criticism (apparently without effect) of several supervision missions (and the POR: see paragrap! 3.03). They did not accept the Government argument that the Extension Officer, a position filled Dy a Burundian after mid-1972, could handle Doth of these appraisal assient. Supervision reports make occasional comment that suggests the quantity and quality of the extension work at the farmer level was not up to appraisal standards. But there is practically no explanation or the training program given to this staff or of its qualifi- cations to discuss with f armers theproject practices and answer difficult farmer questions. OED is left with no sense as to whether the shortfalls in farmer performance - in his average yield, in the extreme variability of performance among farmers, and in the high rate (50%) at which the first project farmers are thought to be abandoning the fertilizer component of the husbandry package, once a price was charged for fertilizer- are attribu- table mostly to weakness in extension or in the technical package the staff! has to extend. Comparable problems with extension components have emerged in other smallholder projects in Eastern Africa, and the Bank is aware that the development of an effective extension service is a long, dispiriting process that invariably exceeds in time the years allowed by appraisal. Performance in this project may be relatively good, but the documentation is not enlightening. - ix - 16. What is of most concern about extension, however, is that the second nrdiet intends to expand the assignment of the project extension service to cover other crops and livestock: to absorb, in effect, the para- 111 work of the reIgular ertnsion servie of the Ministry of Agriculture. Experience in other parts of the world shows that the development of an efectiveanrnan4n .mITe nftAn etArts with single cron campaigns. which allow the fledgling and under-tralF YIer3T- E-f-Te "tsusiness'. Single crop evens.li .arw4ca ne usally more effective than multi-ourDOse exten- sion services, though of course a developing country cannot afford a prolif- eration of single crop sev4ces There are A few instanePA where a single crop service, once proficient in the art of extension, has expanded success- .AJ.U Iy o cver othe crops. Some pesn identif th4-- -- ------------Rratpv for building a multi-purpose extension system. What must be avoided is expaing.IL before LLe prjectJ extenion servic is red to. fv f-IkA Premature expansion would exhaust the capacity of the existing service, even bef.ore .Jt La learne -toLLV had le~± LLe cofee ssinmet wll fI. WhLthe tHE pjLUJcvL s raLVC Yr 4O . D's e Wos t c asr in the files - and the absence of material - is not encouraging, notwith- staUG1Hg Ze 1Upre81VC appeUZCLU UL LLH: LLLLL jJLuje a CcuoL v in the field. That appearance can be traced partly to new buildings and vehicles, an esprit-de-corps that rates wel1 cumpared to regular Ministry staff (to some extent explained by better emoluments), and the busy-ness of Uelivering free fertiizers Unu uter supplies. It s nueacesarily ind ative of the underlying technical quality of the extension contact with the farmer or of the respect and demand that that contact engenders. Le second appraisal report does not forecast any rapid expansion of the extension assignment, in fact only 6,UU of 110,UU target coffee farmers will be advised about food crop improvements as well. T ex2ansion asarrectly seen as a step by step process. But the base nevertheless must be strong. The need for expansion is obvious: the great maorit of the coffee farmers and of their children, cannot raise their incomes much aDove present levels rFom"oeeao. 18. The reports do not discuss another issue, related to extension, that is mentioned in the first appraisal document: "the expected spread of improved coffee husbandry to other coffee areas". Whether the first project has had any demonstration effect, either on farmers outside the project area or on other extension and promotional activities, is not known. The second project will attempt to spread the new technolo and adminis system to the rest of Ngozi province. Importance of Price Policies 19. The PCR shows how relative changes in fertilizer and coffee prices have up-ended much of the economic analysis conducted up to and including appraisal. When the ratio of the two prices has risen, the justification - x - for fertilization has been undermined. When, as after July 1975, coffee prices soar A_ coffee of all quantities have benefited and the differential between grades was almost lost. The incentives for quality mprvment by washing at the stations were diminished. Clearly program success is depen- ent upon astute manageent of and response to prices, a point which the Bank fullyappreciates. An Overall Favorable Impression 20. The first project has sensible qualities and we suspect compares favorably in some ways with many other smallholder crop projects. It has helped strengthen the basis for further development of the coffee sector, even though it did not offer as many lessons as it should have. Government is described in the PCR as having done a reasonably good job. The assess- ment contrasts with many complaints - about inadequate administrative services, poor accounts, initial shortages in Government counterpart con- tributions - reported by the supervision missions. But the PCR assessment is probably fair on balance, particularly when considering Government's limited resources and how these were turned to support the project - les often than required by the project agreement but more often than not The fac'that the project recovered from the 1972 violence - one senior under- study, four of the seven agronomists, nine of the 72 lower level staff and many participating farmers were "lost" in the action - is in some senses a statement on the priority Government gave it. The violence caused the project to pass into a period of stagnation, nevertheless most of the missing staff were replaced within a few months and the largest volume of fertilizer distributed during the project was distributed just after the troubles. The fact that the Minister of Agriculture swung his influence behind the washing stations in 1975, causing local government staff and youth brigades to a1 pressure on the smallholders to deliver cherry to the stations, is another welcome indication. The PCR presents the story in detail. The Income Issue 21. The final point to be added to those in the PCR concerns t l9w incremental incomes Drovided to narticinatina farmprn- One must wonder wheTer theaounts offered by either the first or second projects are enourh to attract sustained farmr intrpat_ The PrP til1m11Atac thAs I-h farmers in the area of the first project will have earned for their efforts an extra TTqA6 in 197; snel TTRAI in 107. Thna syarsaa figures, hiding a wide variation. The increase from 1975 to 1976 reflects the extraordinarv riQP In onffee inrit- _int- tix 107% Tn m44107r +,- I__- ------ --- -- __ In. mid-1975 the appraisal report of the second project estimated that farmers entering the naream in i h teant-a ofeth first prj 47 con t-in to use pthene areas $2.30. Veterans of the first project who continue to use the new inpuits weare then expected-n ton earn~ 0 3.80. Th copst farm budget.., - xi - presented in the second appraisal report represents nine farmers, three of whom mnprove and fert414z.other crops as well eascffee. he bPRt of them could expect an increment of $35.00. But only 2,000 of the 110,000 target Ninety-five percent of the target group are expected to take only the level, within the range $2.30 to $7.50. Seventy-nine percent will average $3.20. r Both apprals-I reo- say that the pecuinary Incent-Tve are strong enough to attract the target farmers and convince half of them. over pre-project coffee earnings - not insignifican - and maybe 5% to su. of total cash earnings. Since other income opportunities are generally unavailalDe, famliles lookng for more cash will not ignore these potential benefits. However, shifting up to the newer husbandry regime requires levels of entrepreneurial and labor inputs - and loss of leisure - which only under the strangest of assumptions would be valued lower than $3 or $4 per year. The incentives which the two projects offer participating smallholders would seem therefore to provide only a short term encourage- ment. The coffee components of the projects may lead to yield improvements higher than those now reported, and may promote a self-sustaiing techno- logical transformation of the coffee economy. But the orders of magnitude of available data indicate that these coffee components cannot offer to most farmers a means to continuing income advances. The limits are tightly drawn by the absence of extra land and the tremendous population pressure. The coffee smallholder schemes must be considered therefore as only small steps in the right direction. A larger solution must still be sought, as, in a way, the second project does by extending out to cover food crops.et 1/ This point is not mentioned in the summary section of the appraisal report, which describes only the $35 farmer. 2/ One must ask whether the two appraisal reports erred in the same direc- tion, by including so many farmers in an experimental program. Concen- trated extension, on the progressive few, may have given higher "ve " earnings and a better base for expansion. That argument cannot be pushed too far, because the project extension staff probably did con- centrate whatever time they had for classical extension work (technical guidance) on the progressive few. Also, it was essential to include all farmers in the pesticide program. - xii - These postage stamp coffee plantations in northern Burundi are in no sense comparable to the smallholder coffee farms which have attracted attention elsewhere in Eastern Africa: in Kenya, for example, where they have helped provide the liquidity that has fueled a minor technical revolution in sec- tions of that country's rural sector. A look at map 1 shows that the first proiect was located in Burundi's most favorable smallholder zone. The $4 contribution to the solution of rural poverty which the project seems to offer - by attackinz the best of the country's cash crops in its heart- land - sets a ceiling to what this type of project offers for the rest of the country. 23. What the permanent solution to this problem is will not be easily determined. It is clear, however, that the Bank is not at all close to disbursine excessive amounts per target farmer in Burundi to find it. To provide a background to a discussion of rural strategies, of those that ffe-nit spent in thNe rirst profject oaroximate - -on each o abot 55,000 Ruru ian es mlTnlderc to enahl them to inrrPa. 1hpir annual in- comes by $4, frm sam8 ~7~J~gT~ ost exactly the sae pr nitT;~~r n the or_nlnan nf Pnnvin MoJw Ciiinon n-rnria Ann-rr%--- mately $5,400 each to about 1,600 oil palm settlers at Hoskins to enable them to incrpanp their annual inenmn hv ahni- A 1,00, startino f-ron einea to zero (see PPA Report No. 1400, dated 12/18/76). Even at Hoskins there ic~ CnMP miioctt4nn W11thA Oint-F 4nr.,em -.411 1k 1, 4-1, .- -1, 1,- to inspire optimal husbandry performance. Nevertheless the Hoskins incomes na 4n +- vnge, whkichive plnners cofr h-t thbe poverty problemn (LDC context) has been solved for the participants. Those persons in the Pnn~l- Tdhn na praesngy tnoa eep per cait cost as4 low asepossibl ment schemes, or to focus efforts on traditional village areas rather than over more farmers - have never suggested that there is a $45 per family __ _- r-- LL~ i dul patici~pani C 24.ni i-at th ncmparison of Uoskins -A WgOZi offers then~ is theCon- trast between an immediate solution for a tiny number of farmers, and a - * Jl J .4.. .y MJ . Q lLLL C * ±I1 L a LL I-. LUL Li LiLe i L L L THe-y Th- 7 wide coverge. Te"tow iTncremental income is partly offset Lvyn p u.L.uI.y InLL tLat further tecunica MIovemence may De self- generating and that the coffee crop activities can thus help accelerate the purce uf growth. what the $4 figure does is remind the Bank that the improvement is, after all, terribly small compared with the target range mentioned in the previous paragraph, and that, if this is the chosen strategy, or the only feasible one, whatever progress has been achieved to udCe must not oe arrowed to distract planners from seeking the next incremental steps, or more radical measures, toward the longer term solu- tion. There is also the danger of exhausting some of the country's best extension talents on assignments where opportunity costs may be very high. - xiii - Given the extreme population pressures in Burundi, however, no project pro- posals are likely to look much more attractive than these cottee projects, provided their lessons are put to work.  E4TRTTNDT AR A'RTr,A C)WWR TMPROVEMNT PRO.TCT - CREDIT 147-BU rnMPT.ETTON REPORT1/ Table of Contents Pa a T l A rTFi flfeTTNTT TT Tr^TmT AmThT DDVnAnAMATT ATM ADDDATCAT 1 'r 'I 1'f1 M TMnffMAWTMI A Mr^N 14 Stuart-up Staffing Position during Implementation 5 Agricultural Aspects Coffee Quality and Processing 14 Washing Stations 16 Other Project Activities 20 Project Costs 20 Financing and Disbursements 21 Procurement 21 IV. ORUuANZATIUNAL ASPECTS 22 V. PRECUET IvACT Coffee Production 2 Coffee Prices 26 Value of Incremental Coffee Production 2o Farmers' Incomes 27 Economic Performance VI. GUYVERN1ME;NT AND BALNA PEFORUMANCE VII. CHANGES IN THE SECOND PRUjECT ANNEX 1 Project Description ANNEX 2 Memorandum on Allocation of Proceeds for the Project ANNEX 3 Economic Evaluation Agricultural Component TABLE 1 Distribution of Fertilizers TABLE 2 Project Costs TABLE 3 Coffee Exports, 1960/61 - 1975/76 TABLE 4 Coffee Exports, 5-Year Moving Averages CHART 1 Expatriate Staffing Schedule 1/ This report was prepared by Mr. H. von Voorthuizen, who visited Burundi in April 1976,and was issued on September 27, 1976.  BURUNDI ARABICA COFFEE IPROVEMENT PROJECT - CREDIT 147-BU COMPLETION REPORT I. BACKGROUND 1.01 The introduction of arabica coffee as a cash crop in the traditional subsistence farms in Burundi (and also Rwanda) started only in the 1930's and most of the coffee was planted between 1947 and 1961. By now more than 400,000 or well over half of Burundi's 800,000 farmers have small plots of on average 0.1 ha with about 150 coffee trees each and coffee is by far-the most ipexort cro and source of cash income. Of Burundi's foreign exchange earnings, which averaged about US 30 million per annum over the last few years, no less than 85% comes from coffee and probably few other countries d so much on one single export commodity. During the last decade Government has started to diversify its agricultural production, mainly through the implementa- tion of tea projects with"iancialassitance from Belgium and the European Development Fund. But simultaneous efforts have been undertaken to increase the production of coffee as well. Dartly through a further expansion of the areas but, since not much unused land is left, mainly through activities to increase the productivity of existing plantations. 1.02 The "Arabica Coffee ImDrovement ProJect". for which a first IDA Credit (147-BU) of US$1.8 million was approved by the Board in April 1969 and a second Credit (598-Bu) of US$5.2 million in November 1975, is Government's main vehicle to increase.coffee productivity. The first coffee project-! , for which the last disbursement was made in March 1976- was the nlv ariiultural nrniect in Burundi supported by the Bank Group until the Credit for the second coffee project was anroved and has only renentlv been fnllnwe by An TnA Crerit. fnr n iqbhrip Project. II. IDENTIFICATION. PREPARATION AND APPRAISAL 2.01 The first coffee improvement project was conceived by Government, nrohahly in 1Q6 eh 10i- and 'Mrnt.3ffa v an wA/Trn rnn*+v PatA/Pm mission in April 1966. The project underwent substantial changes at all stages from idAnf.i-PatAnn t.hrmnah prearation 'Mlevr PAO/TAT) +n n"a4i*al Government's principal objective seems to have been to spread the use of fArf.i17wr +.r all ^ffA, p-r4"&nvmareSinund and it was therefor- ail interested in securing financing for fertilizer purchases. A small pilot project of fertilizer distribution ad already been started in a few ocations in ngoz / Project description is at Annex 1. province with Belgian assistance, following what were regarded positive results4 of a seri.es of on-farm- tri-l wbUi ~c.N. 1ad- comme,neA during the e~-I1arly, 1960s, also in Ngozi province. The project, as identified in 1966, was estim--ed +o cos+ .bk-+ TTOk C -4-1 14m - nP u aa aAwel or involved the procurement d7i-s'tribution of 10,000 tons of fertilizer (up from 100 ton in the pilot project). Burundi's total production of V .LA -.L-VC6S%=0IX-0111 CLL1 E =16 5,U 03. 1 , %A IJ LUIL0.11 1,0ton per ear ais a RIVcftEsa-tin 2.02 The Bank was apparently not very eager at the time to finance the prJ.L . UL.±L1LLLy L %ACLUL10.L.t%.UL1 .L .LOUII. %0 L±U1b er aie whether Burundi's production goals were compatible with the quotas allocated by thie InternatiLonal1 Coffee Organi zati'on 'I,LV WLLe' UL I L '.LLLJCZ was fully justifiable on economic grounds, whether there would be enough inceunives uo tne faermers to responu to tne proJec nu Wnu:ner a11 Ulle needed new staff could be found. In the fall of 1966 the Bank came erroneously under the impression that Burundi was no longer interested in DaniDA f11anCig and no further actions were taken for a while. It took an early 1967 letter by Burundis Prime Minister to the President of Tne Bank, complaining about the lack of progress on the matter, to get things moving again. In May 1967 - more than a year aftUer the miLsio had visLteu b use country - traUslateu copies of the identification report were sent to the Burundi Government. 2.03 A 3-man FAO/IBRD preparation mission was subsequently mounted in uly 1967 and its report was circulated the following November. The project which emerged was somewhat wider in scope but smaller in size than the one iuentified a year earlier. The project would be confined to Ngozi Province (where about one third of the country's _coffee is grown and which has the best potentiali instead of being extended over all corree growing areas. One half of the Province with seven million trees would be e osed to intensive action- , Including-the application or 2,000 ton or fertilizers per 6num fromyear 3 onwards (down from the 10,000 ton mentioned_by The 'entifi a- ton mission), wnile the other naif the Province, with another seven million trees, would be treated in a more "extensive" manner with empHasis on better 4msmmov? .lip - crop nusbandry methods without fertilizers. At the same time the project would carry out pilot studies in order to determine the possibilities of diversifying agriculture away from coffee in the marginal areas, and to introduce some other crops into the existing cropping pattern. The idea was " to make full use of the existing administrative structure", but the division of responsibilities remained pretty much unclear in the proposals. The estimated project cost was down to US$2.5 million over four years (from US$5.6 million during identification), and the incremental production of green coffee resulting from the project was estimated at 2,800 ton (down from 6,000 ton during identification). 2.04 It is interesting to note that much of the preparation mission's thinking was in terms of "rehabilitating" Burundi's coffee industry. This was reflected in the project's name at that time: "Coffee Rehabilitation and Agricultural Development Project". The perception was that the productivity of the coffee trees had declined rapidl yduring the years following the country's Independence in 1262, There have indeed been a few lean years - 1963 production was particularly low, but this was probably more due to normal cyclical Datterns CmairLty caused by weather conditions) than to a breakvdown in Government services and reduced farmers' efforts. The five year movinR average of Burundits coffee production staRnated somewhat in the early 1960s but continued to increase regularly from 1964 onwards, There was. however. indeed a breakdoma of coffee hand nutina facilities. which was later correeted by the project, and a deterioration of coffee Quality which continues to the present day CTee further nara 3.30). 2.05 During apraisal in March 1968 the nrolect area was further reduced to the better half of Ngoz's coffee roing areas Lwith seven million trees arown by some # .000 small holders for whinh tht nrenara- tion mission had envisaged the 'rintensi-ve' treatment, nBooster" applications of fertilizers would be 2iven to all seve n mMilnn tret free of charge to the farmers, during three subsequent years and some 50 percent of the farm1Rs Were exneeted to Lnnthnue fee.rP sg ther trf#C theRyfter at their own expense; the total volume of fertilizers to be sold by that time was estimated at In tn no- v-A" Cae%n tPrm P Ann +^n ai,"ho rennncti n ine with the proposals of the preparation mission the pro ectwould also nnav'id f'armi nT~os to partctiengn nmweeeaw a inraf r.e+ s'in 1o a aign free of charge to the farmersY, rehabilitate 'e2sting o 'ee hand rnnn PAn+tsvc ^ntn+m, t,ha, -P"41v +-iale "n O"fa A m" -triAd handt- i e o- e a d . . d for expaia and urundian management and extension staff. The appraisal misSion did not re-tain h nrapsals for pilot wt+k o+h+ and rejected also the proposed improvement of coffee marketing facilities. The mission included were-rl a ne-wr compon am n=ely -se+ procesigo coffee in four pulping cum washing stations to be constructed under the mn 9t..o+ r +deastn prov tne qua e.U14.4w uJ J mUAJ. a.@J.L . . Ann tion and the stations- were to be modelled after similar centers which had ben built ad suu operated since the early 19uc in KLenya _se- further para 3,321. The mission also included a small amount for the repair the p r eT_earer ta ath preparation stage: a special depart- --t was to ue crea in Glovrnment i mpleentu-u agency, OnTu, TTo would be responsible for the extension of improved coffee production and procssng methods MU A~ -- TTUIZ. L %UZL ws UUW1 U UOr..L HuJiL0 UVe Iyc years compared with US.7miion oe four ears es ifaed during preparation PLoVJ ue UurLemenLaC green coftee production was estimateu at 1,600 ton,down from e 27 t07on figure mentioned in the preparation report. 2.06 The leader of FAO/IBRD's preparation team raised objections against some of the changes made by the appraisal mission. The composition of the envisaged expatriate team, which included experts for training and coffee quaLity control, was considered too specialized and, according to the preparation team, more emphasis on the establishment of an efficient extension service was necessary through the assignment of two experienced expatriate extension staff. The preparation team also disagreed with the inclusion of the four washing stations, since these wouTd7probably raise staffing poems. and recommended to 1imit the processing component to une eabiltatxon of the existing small pulping centers,. Other objections concerned the types, of fertilizers recomiended by the appraisal mission tpure nitrogen instead of compoundsl and the,wv the pest control campaign was to be set up &ndividual dusting by, the farmers themselves instead of organized spraying team.l. There are no indications in the files that the appraisal report raised questions, within the Bank. The main concern was again whether the International Coffee Organization would object to the project, The ICO Secretariat was informally contacted by the Bank and the Executive Director of ICO gave his personal, but obviously authoritative, assurance that the other exporting members of 100 would not object to the project since notnial production increases would not raise Bhrundi's exports be d 2.0T Looking back over the pre,-implementation stage of the project it can be concluded that the Government started out demanding ubig" and that FAOIZBRD and Bwnk staff gradually redueddt7heyroLect toa sizwich a.6 considered more compatible with ernment's implementing capacity.12 Tese reduetions were jusTed and it turned out that even tbe-smaler project took two additional years to complete. Because of an increased sense of urgency to transfer financial resources to the poorest countries, like Burundi (and Rwanda for that matterl, there may at present exist a greater willingness to accept proposals for what are essentially over-sized projects, The Bank should, however, resist the temptation to go along with such proposals and continue to carry, out one of its principal tasks, namely to ensure that projects do not overstress Governmentts implementing capacities. 2.08 A second observation concerns the lack of any scheduled food crop developments in the project as it emerged after appraiss. It is a pity that the preparation teams proposals for pilot food crop activities were dropped for the sake of what was believed to be a clearer project organiza- tion. An opportunity was missed to-aim at a more balanced development of both coffee and food crops which would have been Particularly relevant because of th-e-igh population density in the_pro ect area, 111. IMPEuaMTATION Startup 3.01 Prior to 1969 OCIBU had already started coffee extension work in Naozi Province. includin2 the a oqf' for t{wzear on a piln cale. The new project assumed responsibility for the ongoing activities and there were no delays in the antual start of the AVYe A expatriate p= ec+ manager and a training officer both arrived in April 1969, before the IDA Cretdit b%am Tn"er .m 15, 10o An expatriate extens-on officer joined the project in September 1969 and at about the same time about 30 project. - 5 - Buru=di extension staff vere made. available to the -roject by OCIBU and the Ministry of Agriculture, tenders were written out for the supply of next season's fertilizers. vehicles were nurnhased- and extension work was effectively started., Staffin Position during Iplementation 3.02 Exatriate Staff. In terms of numbers, staffing of the project with exnatriate essential in view mf the Taek of nualified nationals) was better during the first tve project years than thereafter (see schedule chart 1I Of fthp Pour exnatriate in t+ nv--t- by the aPnnrai1 mission that of the coffee processing officer remained vacant until the end nf IQ71 And anniumntTv no re c we fforts iin1Prt%kPn during the intervening two years. This can be explained by the fact that r1rTEJL sand nnttnT w? he -+"+nw4a + moa m " mnrw ^"i a ly inke_ -m cirr port to the idea of constructing four coffee washing stations because of the nsatisfat.ry results oa an Oe%stnIM mub larger, central uOping and washing center operated by OCIBU in the project area. This attitude changed ranidTy af'ow +'Ne A.n. --r w_.. +"-A -P4. + -S -4-+ ma. JK,+ +I n- time lost in recruiting the processing officer is certainly one of the ma%4m w r eans o r tift -hn+ T + A-T1r - n 4 h- n-4.l-a -4 Ah' Lm -+-- and hence of the project. 3.03 From 1972 onwards there were in effect only two ex-patriates n officer was never replaced after the first officer was promoted to project manger in April r711a "LM_ AUAIHU HC ACMU liC E 8& LP L- ..L7 I L .LJ. 1 P siLul3l U1. l".LeJ .& Al L UI UJ.LLu WM.b ±1 J.L U 1972 taken over by a capable Burundi agronomist who still heads the extension a . vLU IIUAo .vsy LU wras a m 11e nWCVt:r, UnMU ne.-unertU M Wit t:AP L1A_MUt training officer not the extension officer were replaced since the pgJect WCL u epr-ve O any rea teclical s ort in whaU shoUJ ve regarueu as Ts main area o acivity. At various occasions Bank supervision missions Uave pointed this out but to no avail. No real efforts were undertaken by Government to fill the gap. 3.,o4 In terms of ui.itythe exerience.with the ex-patriate s ecial's g wa U Ac, ie irs' project imanager anu the 1arEt extension oliicer were TpparenMy both n representatives nor with eacn other. This led first to the dismissal of the extension officer in September 1970 and six months later to the resignation of the project manager. After a short unsuccessful search for a new project manager the training officer was elevated to that position, He stayed in ofiice ror nearly four years.DUt his tenure cannot be characterized as a e1Z -ood one, either. He was youngand rather inexperienced and, moreover) not y active. He tended to stay around in his office in Ngozi is predecessor lived and spent all his working time in Bujumbura), was not well in touch with the field work - not to speak of directing it - and developped increasingly poor relations with OCIBUls head office. During that period the project benefitted most from the second extension officer and especially from the processing officer. With the arrival of a third project manager towards the end of 1974 the working atmosphere, the level of activities and the relations between the project and OCIBU, as well as other - 6 - Government authorities improved markedlyv -4 n; rrph= P4"a+ Tnainaffa"m~oa wa maAo- airnlnhl 'h-r f.hp P-r-nnh firm BDPA and the last manager came from ADS. The other expatriates were recruited direc+1 r 'yVTWIT Pn44?ma,+ of theseo +h avnare thrinah .,nn.on1 t4na f^+1qmM - J.~.. U- --~~ Of0~ U-- --h- 1 was repeatedly considered by OCIBU but rejected because of the much higher cos+s involve T+ A4i is P4 +ifcl o renAd r 4a ugema- e a+h'er tnis was riht or not because recruitment through a firm does not necessarily guarantee good re + snO . w have, however, a defini+e impression +hnat in some cas.. publicity should have been given to job openings and that too frequently n1T'sTTs choice -s restricted to +Jst a fe, orn-4 e+ R4 especially in a country like Burundi, the uality of ex-patriate staff should be reardeD ascriticl f the orderly imupleumentation of a pro+ect, selection of those staff should have received more attention than it actually did. 3.06 Burundi Staff. The number of agricultural officers ("Agronomes") VaAc-u uCUweI CLu. u L6cU ulLLUMb U1. Us LIC PrVJUkL 1=.L. aJ L%. WC6D 4n LJ_1ue with the appraisal estimate Te-In-uer of field extension workers L vulgarisateurs anu moultteurs J varieudvwcc UeUWt:l C u wl"L;iil WaD n.L61,ul than the h4 estimated during appraisal. On the olher hand the project did appraisal. The appraisal mission's idea was to recruit these assistant- moniteurs amoug the vetter farmers ou ass-are active particip'-ion vy tne majority Hutu tribe in the extension activities. The idea apparently never appealeu to te Burunu auuorities but the project UU make use Uo_the.- paid services of some 1,250 leaders of farmers groups Chifo 50 members per for fertilizers from the farmers. The project suffered a serious set-back during a period of civil violence in 1972 wich, amongst ouers, causeu the disappearance of four agricultural officers and nine extension workers. 3.07 Thanks to the good leadership of the Burundi chief of extension a .1 1 I t .. I ift .1 -_ s * services there is now a good esprit de corps among the extension staff and most of them are clearly devoted to their work. Extension activities. are well scheduled and supervised and a satisfactory reporting system from -the -feld level to project headquarter fs in existence. The qualy of the extension advice coud, however, have been better if the chier or the service had continuously been supported by,a well qualified and experienced coffee agronomist (such a specialist has recently been recruited under the second project). 3.08 The position of deputy project-mAnager (not envisaged during appraisal) was created in 1973 and has since then been filled by a young university graduate. His attitude is good; he maintains excellent relations with both the project manager and OCIBU's headquarters and has the potential to become an effective, mature, manager. Agricultural Aspects 3.09 In line with the appraised project, agricultural extension work was restricted to coffee only and work on other crops was left entirely to the -7- Miniatrys Provincial extension services Csee para 4,03). Coffee extension ' m rv;7-rCr AMvOl P--rAnAPA f1rnm T7" Qe%r-n"-m mnAr in t.he- t-ilr-, rf' IQ71 the whole project area was covered, By 1975 the number of coffee growers in the proJetf are^a 'NA--5-vNmc- +n c; (inn anAr +'ha, niirm*hAy nf onffpp i-~~t about 8.5 million Ccompared with an estimated 44,000 and 7 million at the i-Tnp n-f annvAicl) Tlio nn4nfl n---- and, aciveet in Ehe ver -.us r J go or. -- V--J --- - ---- - - - ..- - -Cr areas of coffee husbandry are reviewed below. 3.10 Pest Control, The Antestipsis lineaticollis is the most serious campaigns for many years free of charge to the farmers. At the time of the appraso- 1.ne effectiveness U the UI.,I was re-rucu ue Iuw 4au o1 main project objectives in this field were to improve the application method and to assure full coerg uf all Coffee trees in the Pro-Jet ara.egla insect counts during the last four years suggest that the insect control campai-u as IneUU been better supervised inside the project area than out- side (on average 3 insects per tree in the project area and 6 to 7 in other areas of Ngozi Province). But apart from that there has not been a real improvement in the application method. The farmers still use small bags of tinlLy woven Cotton lauric il.Led with the dust mixture D.DT malathion) which they shake against the trees. This rather primitive technique allows no control over the quantities applied to each tree and it is moreover generally believed that substantil amounts of theinsctici.es.. Aed lor domestic p.uposes. The project did not purchase the hand operated dusters envisaged during appraisal (sprayers were ordered instead, apparent- ly because of errors in the French translation of project documents prepared by the Bank) nor did the project organize special spraying teams as recom- mended by the FAO/IBlD preparation mission. 3.11 In recent years serious questions have been raised with regard to_ Eneotypes or insecticides to be used. In the country as a whole the presence of antestia insects has increased from 1970 onwards, which coincides with a further deterioration of the ualit of Burundirs coffee e Usee also para . Is feare that and continuous applications of DDT malathion may also destroy useful predators and create a DDT resistant antestia population. According to OCIBU's 1970 annual report, ISABU had in that year already been requested to recommend other insecticides to replace DDT malathion but up till now ISABU has no gel its recommendation. Mean- while the price of DDT malathion has increased sharply from FBu 22,000 (US$280) per ton in 1970 to FBu 60,000 (US$765) per ton in 1975; in 1975 UCIUU paid FBu 65.5 million LUS$840,000) for 1,100 ton DDT for its national program. For both efficiency and cost reasons it is necessary to identify and introduce alternatives WYthout de3yE 3U-aDs-that the matter s l rceive the i'es riori u7er the second project. 3.12 Mulching. Mulching of coffee trees is important for enriching the organic matter contents of the soil, to control both weed gr5EiEEanhili17 erosion, and to conserve mistue in the soil during dry weather. The project has clearly been successPul in encouraging better mulching standards and it is estimated that this is now satisfactorily done in about 70% of the planta- tions, against some h0% befor e project started. TaMana leaves and other food crop residues are normally used for mulching and the project has established nurseries of Trypsacum laxum grass plants in areas where natural trash was short in supply. The combined size of the nurseries varied between 5 and 7 ha during the last four years and after initial distribution of the slips, farmers have continued the multiplication in their own fields. The imnact - 8 - of this part of the project was visible in all areas visited in the course of the completion mission. 3.13 Pruning. The project's extension services also paid much attention to improvements of pruning techniques. Some 20,000 pairs of pruning saws and secateurs were sold to the farmers at a subsidized price of FBu 50 per pair and an estimated 5.8 million trees, or about 8g_ of the trees which existed in 12618, were pruned during the last four proje_ct yers. During the 1960s pruning had largely been neglected or done too late. It is now done on a selective basis and the noral runin cycle varies between four and six years. depending on the stand of the trees. 3.14 Coffee Fertilizer Trials. A-first series of six fertilizer trials on coffee in farmers' fields in Ngozi province was conducted if6m I 6_-199 The 1968 appraisal mission for the first project was of the opinion that the trials had firmly established the overall importance of good husbandry practices and of the application of nitrogenous fertilizers for bringing about substantial increases in the yield of parchment coffee. Only a small effect of potash was observed in the six trials and there were no indptions thatate applications contributed to the yields. The re-estatlishment of the research program was regardd necessary b the mission to provide a basis for continuing advise on fertilizer use in Ngozi Province. The 1974 appraisal mission for the second project was less convinced by the results of the trials conducted in the 1960s. The 1974 mission concluded from the available evidence that three out of six trials hadshown a resnonse ratio of 2.5 to 3.5 gram incremental" parchment coffee for 1 am of uenitFogenetil he other three trials so ratios of ess than 2PL The main conclusion to be drawn from the six trials was, however, that yields vary considerably, both from year to year and in different locations, whether or not fertilizers are-annlied. Dif- ferences in location, altitude, type of soils, the standard of maintenance of the nlantations, the stage in the pruning cycle and the weather conditions were all of overriding importance and made it impossible to attach statistical significance to the trial results- Th 1Q74 mission concluded that the results of the six trials did not suort the 169 mission' ss on of an incremental ~ld/rt~r to o' I and that a response ratio ofthe ord &eri ofr EN5Vl seemed to be a more reslistic estimate (appraisal report second project, Annex 1, ara 8-io) 3.15 ~ ~ ~ ~ . sh idng f't ecoand series of trials, conducted under this project from 1970-1975 in five other iocations, were summarized and commented upon in a short paper prepareA by TSABT Apr41 1076 a+ +ha ocasion of the completion mission. But unfortunately the outcome does not contribute to the stt uofu kno-W-11edge o the sub~O~Lject. Theuu main cocuso of th paper isagi that the small effects of fertilizers, which were observed are statistically not sInfcn.As in the FirT series"this is caused by much lwrgw, GuxplTi erences between the yields of individual plots which received the same treatment. it should be noted that there were serious shortcomings in the design of the trials. Originally, none of the trials had control plots nu no plus were veing teated with tne recomueuuu projecu formula. Accoruiug to an August 18, 1970 supervision report, ISABU's director said that these - 9 - plots had been omitted because of the limited area for the trials. After the visit of the 1970 supervision mission, control plots were added to three trials but the other two were carried through without control plots. The application formulas were not changed. The lowest application rate of about 123 gram pure nitrogen per tree ,-oble the rate of 64gram per year ali:edinfYhe project. Besides, otash and phosphate were on applied in ;8mL7-tfiffitt"VVd1 other - !a.:h - ad with varying quantities of pure nitrogen but not separately. Hence, the design of the trials makes it impossible to draw any conclusion on the effects of either potash, or phosphate. 3.16 Apart from design errors implementation problems have also played an adverse role. The results are too fluctuating (and sometimes conflicting), to even support the conclusion that fertilizers have only a slight, non- significant effect on production levels. They rather suggs a ack of consistency and, probably of proper care and supervision, during the execution of the trials. The officer-in-charge discussed some of his problems with the October 1974 appraisal mission for the second project (he was on leave in April 1976). Since he had a number of other responsibili- ties and because of his occasional absence for leave or other reasons, he could not always personally supervise the application of the fertilizers and the harvest of the coffee. Mainly because of frequent staff changes (especially after the 1972 events) he felt that he could not fully rely on his local staff and he was not sure that the correct types and amounts of fertilizers were always applied at thg_ rtgo meto the correct trees in the correct manner. Similar problems may have jeopardized the FarE- of cof eich_ normally is done at regular intervals over a period of several months. Even on a research station a trial of this kind, which takes at least five years to complete, causes problems of consistent, and high quality supervision. If implemented on farmers' fields, spread over an entire Province, there is a further problem that growers sometimes change the standards of maintenance and care of their plantations for one reason or another which can affect the yield levels more than any other single factor, including fertilizers. There is good reason to doubt whether farm level trials of this kind, even when properly designed and consistently supervised by an officer who has no other trials to worry about, will produce the right answers as long as there is no guarantee that the level of maintenance (mulching, pruning, insect control) is consistently the same throughout the trial period. It may well be that this can only be achieved if the research organization assumes full Map~~b i y o h an ance o: attbs 6a yon ht atation tvme of situation. 3.17 Distribution of Fertilizers. As a foliow-un on the pilot distribu- tion program involving about 3,000 farmers (1966) the project set out to a arrange for large scale distribution of fertilizer to coffee growers. The Burundi authorities regarded this activity as the main thrust of the project whic-h -.aq rf1PetPd in it. loc(Al1 titleo "Prnipt. show until the third year and the farmers had only small cash incomes (mainly from co%ffee) it was oq Ai 4 +*rr ++fg.4 ~ ~ distributed free of charge to all 44,000 farmers during three years within the first -P +4--f + . dy 6 e n project yearsfou fert4ixi we( to be sold0 ' s=bsdy of 66 percent in project years four to six to ~ 4.~ ±~5UJO4 ~ U~Ji * ii ~ ~ V ~.L,~J~1U dIU.11±i~ ~1I~ ~ LU~i4, - 10 - would be reduced to 33 percent and thereafter be removed. At the recom- mended application rate of 140 gram urea per tree per year it was expected that by year ten about 50 percent of the farmers, or 22,000, would purchase annually about 500 ton of fertilizer. The appraisal report suggests (para 4.05 and 4.06 and Annex 6, Table 1) that the fertilizer sales program would start in prolect year three and be implemented simultaneously with the initial, free of charge, distribution program for late starters. In a sideletter to the Credit Agreement the Government confirmed.amongst others, the following understandings, reached during the negotiations: (i) discontinuation of distribution free of charge in the project area of fertilizers - after 1974t and (ii) subsidv on fertilizer sales, i.e. 66 percent from 1972 through 1976, 33 percent from 1977 through 1979, and no subsidy thereafter- 3.19 Details on the actual distribution of fertilizers are at Table 1. About 2,665 ton urea were distributed free of charge over a four year period h,tv,n Ne)vPTmhFr 1QQ an Nnvpmhpr 1071- Tht ciqtribhti.-n -rPahfd n. npak during the 1972/73 season when a total of about 870 ton urea were distributed to about 80% of the coffee growers: about 75% of all trees in the project area were fertilized during that season. The projected 100% coverage during the initial "booster ann1iation" vans wa thus not achipvpd hnt. t.'hp totl volume of urea distributed free of charge corresponded closely to the appraisal estimate bennq an imnrtant numihr n fP?mPrQ rt-nivpA Pf-p fPr.i1i-r for more than three years. The fertilizer sales program through private coffee tradrt9'n a penv,qpopv aniivv-nu annrnia1 -s ra'~ 'ndderPt unun.kshla anr' nvt executed and it was also felt that the free distribution program could not be npr-ipa mi qimultan.nely with a rogram of salling frtilizes Tnstead the Government decided in 19 tosop the ree,,distribution entirely and to to chice c r resente fr bou fe rtil ea S.. nfJ. - -. . 571- ll'. J-- - -fl -~ . .LL. . OJ ~ La. - - _ . JJU 27,1400 grow-rs 5,,-Mt-%1e r"t'-'i 'area wh Tih _Xpojec ar, were pr ce c age represented abu n hl fte actual cost of the fertil- and the agricultural officers. It was the first time that inputs were sold tU coffee gruwers ou cre it - ll LUZUn plt:V.LUULAy rCt.q ub L ufworanv.U4 under BurunE--co o The results were encouraging- about FBu 1.9 mllion. (UWr$,000) or 82% of thle receivable amount was indeed collected. Bedause the prices of fertilizer tripled fu p e e rchase ULCL I. L .L-.L L.Z.L L 0 'tL- 'Au K--aA_-4'y .12 I) FAQ made 500_ton urea available to Burundi at a subsidized price of US$12Q per ton c.1.f., Bujumbura and the entire consignment was made available to the project; the cost price was about FBu 11-12 delivered at Ngozi. Some 27a ton were distributed in Ap ri L-;5 to about 200u farmers and the arrangements for repayment were similar to those made in 1974 except that the price had been - 11 - increased from FBu 6 to FBu 9 per kg. However, the repayment rate turned out to be worse than in the first year: 64% against 82% in 1974. 3.21 A number of reasons, in addition to the fertilizer Price increase have been put forward to explain the deteriorating repayment rate: (i the total coffee production was much 2Qam J.915 than in 1974 f16,000 m ton and 28,000 m ton respectively); 'hP bauPnsn of the slumn in thp onffF-P Pynnrt marketsq ov%rnment. was forced to decrease the official parchment coffee price from Fi hc nr- ka ;_ _ Onn hA wq- n- n nv -rn I wan r n r, ', r 0r production was below normal and many of the farmers are said to have spen+ par+ of +heir aAo mc"l inco-m-e from c-offeoe on food stuffs. These are all valid reasons, and the final uRement on the merits and disadvantages of the SyStem .+.-+-A - - T 07)1 -aA +b- tleast ++,-l+;1 +k- resl+ of the current 1976 season are known. The project has made an effort to prevent Agriculture (who was not aware of the 1975 repayment problem) told the nmio4n misslnde tvhate m ro, rl surge lh loal a;inA ur+onm UO h1ei-p. supervise the repayments. The 3rospects for e current season are better. 4-~d J4 4.~ O4.parch ent .PAlLJIi .L%LC I.LCUO VJccI± d.O ULy 11iJ -Lq=.. ULIuI 1)/41 %,U V1.UUL U,7 per kg, following the sharp increase in coffee export prices, and the 1976 harves -11' - -&uliCLL- 1~L -- 'T J4_ 4.Lt-- le Uj± .eL _.L has meanwhile been maintained at FBu 9 per kg. 3.22 In hindsig it may have been a mistake to distribute fertilizers free o hUarge for such a 'uu period because it may have encouraged the use thereof under marginal crop husbandry or soil conditions. It would probably have been better if a nominal charge had been made from the beginning. Tne project should be commended for its experimentations with credit sales since -1.7yI, uCo.Ue -ne m-Acu ruC o, -uu Z-4 -be epArie ton- buouu coUn--A - U. under the second project until a satisfactory system has been developed. The generally helu, unu proUVly curreu view is thau Ute larmers lac. tne cash to pay for fertilizers at the time it is applied (November and March-April) "n' ]Un hould M 16be asked to pay at The time the corree crp is sold May- July). Either delaying - as presently out - or advancing of payments for fertilizers seems therefore to be necessary. 3.23 The most important question is of course whether the application of fertilizer pays, but as discussed in paragraphs 3.14 - 3.16 the research efforts undertaken sofar have failed to provide a clear basis to answer that question. Common sense, based on experience elsewhere, and the results of the trials conducted in the 1960s suggest that nitrogen must have a positive effect on the yields in Burundi but the appraisal mission went too far when it stated - 12 - that inputs were of "proven value". ISABU als rpcommends.annual appligafRg§g of potash a phosphate but one application once evey e he A well be sufficient if coffee is properly mulched. The question has been raised whether the fertilizer strategy was well conceived in view of the shaky research basis and whether the project should have concentrated more on other elements of coffee husbandry. There is no basis to answer these questions unequivocally either way. But there is good reason to believe that it would be more difficult to encourage farmers to do a better mulching and pruning job (which costs them etemel if not combined with the supply of fnputs like insecticides and fertilizer. Again, it is difficult to believe that nitrogenous fertilizer wou uThave no effect under Burundi conditions. Very much depends also on the price relation bet]oen coffee and fertilizer. The economics look quite ETferent when coffee sells at US$0.50 per pound and urea costs anywhere from US$350 to US$450 per ton, as in early 1975, or when the coffee price approaches US$1.50 per pound and urea costs less than US$200 per ton, as at present (for some detailed calculations see the appraisal report for the second project). These are god.reasms-to-be,prudent and for not pressing too hard on the farmers to apply fertilingesTh.der is also important that the farmers' prices for fertilizers reflect the economic costs and the project should rely on the judgement of the individual farmers to establish whether they consider fertilizers worthwhile to spend their money on. Quite a number of farmers, whose opinions were asked in the course of the completion mission, seemed convinced that the use of fertilizers is beneficial to them but so far they have never paid more than half the true economic cost. It seems clear that in any event the use of fertilizers should be combined with high standards of coffee husbandry. 3.24 New Coffee Planting. The original prolect did not include a Drovision for lanting of new coffee trees to replace oxising_old or diseased trees or to expand the cofee areas. For many years small coffee nurseies were maintained in all adMinistrative communes of Burundi's coffee growing areas. Seeds were purchased by the Ministry of Agriculture from ISABU and maintenance of the nurseries involved little cash expenditures because all work was done by communal labour under the manaiement of the Youth Revolutionary Movement. During preparation and appraisal of the project this arrangement was apparently considered satisfactory and the sublect is not even discussed in the annraisAl report. 3.25 A November 1971 supervision mission observed, however, that Burundi's arnhinn nnr~f'fPA ndiintinn Pyppp t.he anmnl TrnO Pyn+ min+nc fonr them caqnn. year in a row. This was attributed to a massive pruning campaign conducted in effects of t IDroj ct, and especially, increased lantings. The number of cofe in the coutr iv'as a whole ~b . illion trees p ar, which corresponded to an anulL rthof 5- . The International Coffee Agreement and that the Bank had agreed with ICO to take success of ICO's production limitation policy. The mission suggested that Uwo.0 practical wa.ys~ I.o. reuceL neVw p.P-a.IIL1e5Z WUU.LU Uq= U ±U -Lwt.ULUU..L- PLutIZ5 - - 13 - wuI.n was Veu . vuuc vy I-n1e Government at tnat 4-e- u u im WILLI O. LC..Liles %;V1O.L~UC1 L 3 U .Y li UV hatit~ t. ,L~U imeI. - aiL to. 1iJ.it plant distribution to replacements only. Taking also into account that many ___ -1t.----a. ...& VfI tL 1 e A ~W~.,LU1 werLe es U .L;D11CU L11 PJUU1 V a,kiU WJLk11 ULI much supervision the mission concluded that the project should control nurseries and pleaut distribution in the projeCt area while in other areas the Government should undertake to reduce new plantings. 3.26 The Government at first did not react to the mission's opinion to uring the coffe- nurseries under the project-s control. Accoruing to the September 1972 supervision report, OCIBU's director maintained that this was unnecessary since OCIBU received all its requirements. Later during that year, however, all communal nurseries in the project area were closed down, except for one (near tne new Musnonge coffee washing station) whicn was taken over by the project. The capacity of the Mushonge nursery was 35,000 trees at that time and according to the March 973 supervision report the project proposed to establish a further six nurseries of equal capacity. With these nurseries it would take 8-10 years to replace an estimated 15; of the total trees in the project area which were assumed to be either neglected or dead. The March 1973 supervision report flashed a warning signal in reaction to the authorities apparent determination to issue young plants prior to-urooting of the trees they were supposed to replace. The mission commented that this would almost certainly lead to unwanted extensions of the coffee areas, since the farmers would be disinclined to uproot any of their old trees, no matter how bad they were. In the customary letter following supervision, Government was urged that seed.ing should be sold at FBu 3 each in order that farmers could appreciate their value, and that seedlings must not be handed over until after old or dead trees had been removed. 3.27 Both points were ignored in the ensuing years. There is still an old law on the books in Burundi prohibiting farmers to uproot their coffee treesvwithout specific intu T Me n F ieiTe o o hr is i hvid.en etasuch-instrt tosh-ave been issued in the last few years; to qualify for new plants farmers were only required to prepare the plant holes before the distribution of the young trees. No prices were charged for the plants either, mainly because it had never been done before and because it was considered difficult as the farmers continued to provide free labour for the establishment of the nurseries (though not for their maintenance). This point was overlooked during the appraisal of the second project and in hight it would have been better if the Bank had insisted on a small payment. In the course of the completion mission the subject was discussed again and the authorities seemed inclined to agree with the idea in order to help reduce wastage of young plants and to discourage plantings on marginal soils. 3.28 From 1973 onwards the project's nursery program expande ag sk1YA After attempts to renew the ICO agreement failed in 1973, Burundi followed a policy of expansion of coffee areas since other coffee growing countries were believed to be doing the same. The number of plants distributed from the project's nurseries increased from Eabout 25,000:n 1973 o , - 14 - in .Lg IUnu nearly 1.c mLoilo in e. uLYI I icUS Qurretpoullucu oU nearly 157 of the then ex ing tree population in the project area. The new LIO agreement, wnicn wIll become ellect.iyein pepember 1q, Uoes not proEibit-Burundi trom expaning its plantations at such a rate but the major problem in this connectdon is one of space. Ngozi province is very densely populated and s lms inevitable that young trees are planted either on marginal soils or on land previously cultivated x-itnjo P, of whicn there are frequently shortages. During the completion mission it was observed that many of tne young plantations_are planted on very steep slopes, on marginal soil in the strips between the old hpantaEfons anEe road[s nhich ereprioslyle" uncultivated to allow the passage of cattle. In tElatter case many of the young coffee trees are planted in the shade of eucalyptus or pine trees along the road sides, which does not promise a healthy development for the coffee trees. Government's position is that many young famtiles want, to establish their own corree gardens and that there is a backlog in the supply of young plants because only small numbers were issued in recent years. For the second project it was agreed that the number of trees to be isuedac year would not exceg. ifexisting LaDout e0A ror replacement and 1.575 to establish new gardens for young families&. Enforcement of this condition seems difficult unless Government agrees to a nominal payment for young plants or the Bank refuses to reh,Aburse for claims for excessIve nursery costs, 3,29 In retrospect it seems that the Bank's insistence on bringing the nurseries under project control has encouraged rather than discouraged the expansion of the coffee areas. It also appears that our insistence was based on the oviously wrong assumption that through the project the clear and consistent policy of Government to encourage coffee expansion wherever and whnever possible could be changed. The dialogue with Government on this subject was not fruitful and the only effective message would have been, as just suggested, to stop disbursing against nursery costs. Coffee Quality and Processing 3.30 Coffee Quality. The juality of Burundi's arabica coffee xports has steadily deteriorated over the last 15 years as shown in the summary below (in percentages of total arabica exports)7 Grade 1962-65 1966-71 1972-75 Fully wished 1 1 1 OCIRU 2._/ 30 2 1 OCIRU 3A 58 77 61 OCIRU 3B 6 16 31 Broken 6 5 6 Source: OCIBU Annual Reports 1/ Quality standards are those established since before Independence by the Office des Cafes Indigenes du Ruanda.- Urundi (OCIRU) and are the inter- nationally recoRnized standards on the basis of which exnort sae are meap - 15 - The proportion of OCIRU 2 arabica coffee dropped from nearly one third in the early 1960s to about 2% from 1966 onwards, and no OCIRU 2 coffee was ex-orted in the last two years. The export premium for OCIRU 2 coffee is normally about er pound over the standard OCIRU 3A coffee which was sold in early 1975 at about US cts 45 per pound fob Dar-es-Salam and at present (after the Brazilean frost and some other adverse developments) at about US cts 125-130 per pound. The proportion of OCIRU 3B, which normally sells at bout U et und less than OCIRU 3A. increased steadily from just over 5% in the early sixties to nearly one third in recent years. This development causes great concern in Burundin ET t-TR"MAse of the obvious loss fuayitinvolved but also because it is almost0MTTS-1_1"' couDled with increased losses in nroduction volumes. Poor eoffee proess- singFacTM an ower rop usiah standards are seen as the main reasons for the decline. The decreasing effectiveness of the insect control program is probably the single most important factor because it both reduces the volume of nroduction and increase the pronortion of insect damaged beans. TheRprin_ofJd%aged beans increased from about 15% in the early lQ70s to more than 'Rn -tr- Go No rois1 This show cle---arly-that improve- ments in the insect control programs are urgently needed (see also para 3.11). 3.31 Hand Pulpers. In general terms the firstapoject has thus not su-y in revers'ing the tre-nd o~f deterf-Ang4+1- the exception oA e _Tywashed coffee produced in the new coffee washing washin and drying has been done by the farmers themselve At first the than 300 simple hand-operated disc pulpers were established in the coffee groin aras of -~1 whichabou 70 inK- the-project are. The.------------- always built near perennial streams and included some 6-10 small concrete ha"n +-5 soak" andLi.1 th parchme t offee. afte puI5lpJing. in thie lat.e 1950s and early 1960s most farmers made use of these facilitiesbut later in u o mo.JO Ay W uhn puAprs vecame uperauional yU vecausq 0 Mr andmMS.theand puloejrs en ties sof te pro ecT was to and by the end of 1975 some 60 hand pulpers were in use (nine pulpers wer close after the consotruction of four coffeec washn statiUns). The hand"Upulpers certanyfclte i of the coffee bu donot seemJ~ Tu e washn uun,Mies at tne pulping centers are are s d most coffee is not washed before drying. Short and irregular drying at the f=5ers nomesoften necessitates further drying before hulling and stimulates the spread of micro-organisms that produce off-flavou T problems a armers are not paid in accoruance with the quality or their coffee and are thus aged to do a better jo Introduction of prmrs lur udierent qualities is considered impracticable in the framework of the present marketing system, involving hundreds of small private coffee trauers. Yet, it would be the only effective way of ensuring better care during picking and processing. .L/ No separate figures are available for the project area. - 16 - Washing Stations 3.32 With the objective of improving coffee qualiy the project also provided for the construction and operation of four mechanized coffee ulpin and washing stations. As stated earlier (para 2.05), the four stations were not included in Government's application but added during the appraisal of the project. The idea was to model the stations on the successful Kenya Coffee Authority (KCA) pattern and to achieve a greater increase in quality through the purchase from the farmers of coffee-- cT-re,rte h2_p.soesr a more closely 9 gr s a n pa rchment coffee, and thusoenueamrcley conrolled processing. At the timeof appraisal it was expected that for fully washed coffee an export preMium of about US cts 6 per ound would be received which corresponded to about 18% of the price of standard OCIRU 3A coffee at that time. 3.33 Implementation of this part of the project was slow. The first project manager expressed serious reservations on the economic prospects of the washing stations and the Bank stressed the point that the_price of coffee cherries, to be paid to the farmers, needed first to be raised to make delivery at the stations attractive. The construction of the first station (Mushonge) started finally in December 1971 (two years after the start of the project) but, because of late arrival of equipment, the station became fully operational during the 1973 harvest season only. The results of the 1973 and 1974 seasons were disappointing since no more than 40-45 ton parchment coffee (35% of the designed capacity) were produced during each season. The main reasons were a notorious lack of cooperation from local administrators, who associated themselves more closeIy witthe iTrests -t 'te pr'ia e-coffTee-tradTers hE-nh!jt,Fe ,an ratvecof?e e prThe green lightf or commencing the construction of the three other stations was delayed for some time by the Bank on the assumption that the early results of the Mushonge station should first be evaluated. But after Mushonge failed to become operational in 1972 the Bank authorized the construction of the three stations in February 1973. The actual works on the new stations did not start, however, until October 1973 because of slow processing of bidding documents. The new stations thus did not come into operation until April 1975 and the drying areas were not even fully completed at that time because of late delivery of plastic sheeting. During the 1975 season (a full six years after the project had started) for the first time an important volume of fully washed Darchment coffee was _roduced:'94 ton- or about 80% of the desiwnPd enanitv of the four stationsl7. The conditions for cherry deliveries by the farmers ha much imnroved. The Government maintained the rnf-rPP cherry nice at the 1974 levels while the official price for semi-washed parchment was reduced fmm PR7 FTh 1r5 TPer ka t.n~ TRn '4Q 'he-.ui f - theii vrY low exort p rce levels i early 1975. And the local authorities, after personal intervention by the Minister of a"rictue m- ia aae ser-4ius and sccessfu--lefrto ovic farmers that they had to deliver the cherries at the stations rather than 1/ Production for the just completed 1976 season is now estimated at C:C:( -4--L - -u - 4. --U~LL --J I op LJJ.. -- ..3 LOL *L-L1 - 17 - 3,34 Although it is now finally proven that the project is capable of producing a higher quality fully washed coffee, serious doubts remain_ with regard to the profitability of the washing operation. First of all the constructi costs of te stations trned out to be much higher than estimated at appraisal: about FBu 6.5 million (US$83,000) actual cost per station compared with an appraisal estimate of FBu 1.6 million (US$18,000 at the 1969 exchange rate). Inflation can be blamed for only part of the cost escalation. In the Burundi situation. and taking the late construc- tion of the stations into account, one would normally have expected a cost overrun of at.the most 50-75%. For one thing it seems that the base cost estimate was too low; the appraisal report specified that prices were based on actual cost of washing stations built in Kenva plus 10 percent to cover transportation to and within Burundi. A supplement of 50% over the Kenva cost figures would have been more appropriate for Burundi condi- tions. It further seems that the appraisal mission had a far simpler and less costly station in mind than those which were actually built. It should indeed be possible to construct satisfactory stations by using less concrete and less steel and by installing less engine power. There was also some, probably unavoidable, wastage resulting from experiments with diffprpnt tv-npq f drvinz fantitT-mnnliv nt the Muhone sation. The moving "Warnier" and vertical "Henderson" dryers both turned out to be unastsf'actory under Burndi cnnditions a now only fixed, hriAntal, drying tables (which are cheaper) are being built. Finally the project did not alwv- , t tho hpqt nonsible dAl on eninment nrnon1urPd Vnr example, the equipment for the last three stations was ordered too hastily in an unsccessful effo+r to get +tem into npea+nn dur +h 1O7k season. Besides, some of the materials (amongst others plastic sheeting) were nrnocal-d loaliv a+ 'h4oh yv4^0a - afarn^= +Ina +ha ff4^neamainl because of disruptions in the supplies from abroad. 3.35 The second, even more important factor for the profitability of the this premium has been lower than projected during the appraisal: about US cts per pound I n7l: TTC -.4--. ), -~ ,v7), --..A TTO -4--.. -Q [_ O - - .-' ~ ~ ~ ~ ~ W .kVQ -'M.J4 LLI J_I.., 'JL .U . 1 C611"0.IU UW. %-UD _) %VL -IU U~VIU1 UI L k±~j'JJ4 ing price for OCIRU 3A coffee) in 1975. The premium obtained in 1975 was nov4 -nin r. , A4...nr.4vm+4... o,,A no.. ,'...hkh1w, h-e a,rrmle4omA hw.. +ha r.a,ra,e shortage of coffee following the July 1975 Brazilean frost. Prices for a.verage. grade.L coffee.~, lik nrJ'IDT'.JA .I, were~ A-J4.* -4, --J. P-4-0. U --.L A J4 0k - apparently attached less value to quality than in normal times or in times .J. .J'W 4. so,%U.lV4 UA . u"rthher be noted~U lchadc. L.L ULAU 1L11= .4.7 I )o Y)J.Luw U.&LUII 0±L Ulle1 four stations about 75% was classified as FWA for which the export premium woo A .GVL, %L C Uk1 blaCv WaD Cblafe as eiher VCIRU DA U UCIRU 3D and sold together with traditionally processed coffee classified in those graudes. On te outer nduu te hulling elliciency of the ully wasneu parchment coffee (80%) was markedly better than for traditional parchment coffee (about 71). iThe implications are not fully clear. Tne appraisal mission for the second project found it difficult to believe that equal amounts of coffee cherries produce different amounts of green coffee when processed differently. Traditionally processed parchment coffee certainly contains more moisture and remaining mucilage and therefore shows a better cherry-parchment ratio but a lower parchment-green ratio than coffee processed in the coffee washing stations. Project management believes, - 18 - however that strict quality control on cherry dliveries at the washing stations (and for which two different prices are paid, according to grade), encourages better coffee picking practices by the farmers which ultimately leads to a better overall cherry-green coffee ratio for fully washed coffee. The nroject is currently trving to establish the facts throngh tvi.ls* meanwhile it may be assumed that the quality control measures at the washing wtations will indeed lead to some (but difficult to quantify) improvement of harvesting practices. 3.37 A full evaluation of the merits' of the washing prCess i Lsn tye ~sible but the prospects, at least in" economic rms,A ~ ___ good. The opeRa cost of the kushonge station, comprising, amogs others, salaries and wages, maintenance and repairs, fuel, 4insanae transport costs, but excluding depreciation and the overhead costs (mainly the cost of an avtri+v4t+ proce ssn oP4fficr) +t+oti1A for 1075 abo+ FBu 700,000. The station produced 141 ton parchment coffee which, after hall4ng gaa 110 +on ran coffa* +he nra+i, ng e+e 14+ A ohv,n +hu worked out at about FBu 6.3 per kg green coffee or about US cts 3.6 per pound of green cSffee. TicldAi ne AJ. 4 0nk + dpeai ,a FBu 55,0 TT or+U jSct 3pe rAound green coffee) and overhead costs (about FBu 375,000 per station when ten s taQ,tons wil.l. beJ in operat.Lin, or1 US) ct 2C per pound theLJLL total cos.LtJ of produc tion would be approximately 8.5 US cts per pound of green coffee. These costs woulA be +1 Pe+ r epartly ofse bynsaving+ foaAnn4ng +he +trA4+4inal semV washed coffee (elimination of traders commission and additional drying of the coffe 4. j-4-"ura, +oge+her e.A stimated at FBu 3 per kg or TT As 147 per pound of green coffee). The balance, or about US cts 7 per pound of green coffee wouU neeU to be coverCe vy tne capurt premium fur fully wasueu cUofee and does not include a return on the capital invested. The experience of tS ao three yearsoues not hold out much hupe for achieyiug a premium of that size; 3.38 The Government, however, attaches great importance to a continua- tion of the program and hna at various occasions expressed a number of reasons for taking that position,including the following: Ci in a situation of oversuply on the world coffee markets, which was very much the case in 1914and early 1975 and which may well occur again after another three or four years, buyers will be more selective in their corree purchases and thus be more inclined to pay a premium for good quality coffee. Therefore, it is in Burundi's interest to invest in quality improvement schemes; li1) so far Burundi has produced only small amounts of fully washed coffee and a production of say at least 1,000 ton FWA coffee per year is needed to attract sufficient buyer interest and to establish a name on the market as a quality producer; (iii) the impact of the coffee washing station progry is greater than can be expresse-Tin-triCt con mi cEFfl ft .10 errp. The washing stations, where farmers of the surrounding areas deliver their coffee, can serve as focal points for a large number of rural development activities, for example demonstration of better growing practices for both coffee and food crops and sales of farm inputs and consumer goods; - 19 - U1v) not expressed in so many words, but probably very much on the minds of Burundi Government officials:the washing stations are visible signs of development which is of great importance in the Burundi country situation. 3.39 For the second project Government reguaste iancial_assistance for the construction ofLng less than ;i new stations. Since there was clearly not sufficient basis at the time of appriasal to have confidence in the prospects of such a scheme, it was, after protracted discussions and great reluctance on the Bank's part, agreed to provide for the construction of four additional stations which would have to be of a simpler and less costly design (see appraisal report for the second project). The number was later increased to ten in an agreement between Burundi and the Kuwait Fund for Arab Economic Development, which co-finances the project. To enhance the prospects for successful implementation, Burundi will have to focus very much on the following aspects: (i) strengthening of coffee research, especially insect control. Insects not only destroy or damage the coffee beans but there are also indications that they cause off-flavours such as potato-taste (the fruit-fly is suspected to be the culprit) which results in down-grading of coffee qual (and damages Burundi's reputation on tI ional coffee markets; (ii).gict control of_the quality of cherries delivered at the stations, gQupled with_Drip± j4ffeentials for different grades. The project's record in this areais satfsfactory so far; present standards should be maintained and if possible improved; (Mi) strict gualt control duln and_drying. again the project's record is not bad so far Vu= uch attention should be given to avoid deterioration of standards when the number of stations increases; a well qualified expatriate processing officer is almost certainly needed for quite some time to come; (iv) the production per station should be increased as much as possible to reduce theexcessiveburden of the fixed coRst, which aeffar T greate HE'T$ variab e costsR.The bottlenecks at this time are the drying tables which at most stations can be easily expanded at relatively little cost. A production of 150-200 ton parchment coffee per station, rather than 100-150 ton, would make a big difference in the costs ner kl/: (v) the Dossibility of alternative marketing arranaempnts for fully-washed coffee should be exlored. Organization of a lclauction saleis rein_ rf'faeih r)nr- A qiiffinipent. volume of fully-washed coffee becomes available (800-1000 ton). 1/ See foot note page 16. - 20 - Other Project Activities 3.40 Outside the project's main areas of activity, namely extension on improved coffee husbandry practices and the supply of inputs, coffee research and coffee processing, the project has undertaken xew ouner actono. = prec administration office with clerical staff was established in Ngozi-town during the third project year when it turned out that, because of both the volume Of work, and the long distance to Bujumbura, OCIBU's headquarters was unable to do a satisfactory job. On the other hand the envisaged improvement o culverts and small bridges on existing coffee tracks was not done by the project but left to the local communal administrations, which are responsibue for mainnanou e nf the tracks. Not much seems to have been accomplished: the March 1973 super- vision report says that one small bridge had been improved. Project Costs 3.41 It is not possible to fully reconcile the appraisal estimates with ly weak, underwent several changes during the implementation of the project O.IA UL J. J L D J O.~.JiLLOJ ~.L.. A I i.L-L L' LA-L =0 UVA.)CD ILV L, CL.LW .y ZD LU.L i 1CDVJYJIIU to the categories used in the appraisal report. The closest possible comparison estimate more than four times (see para 3.3) and total capital LUb ~ ui-z u-me hi.L~ hIIC IUI~i±er thanI thte Or-!-!-i V_L_L) actual expenditures for pesticides an d fertii' r Wtfre both1 only slightly higher than the appraisal estimates. On the other hand expenditures for pruning tools and particularly for drying trays remained far below the appraisal estimates. For pruning tools the appraisal estimate was too high since the project farmers were in fact already well served with tools. But drying trays, which were originally envisaged as a major expedient to improve coffee quality, were dropped after a small initial distribution program in the first year and the project continued thereafter to focus entirely on the washing stations for coffee improvement purposes; (iii) staff and operational costs exceeded the appraisal estimate by some 257; the costs of the coffee nurseries are included in this category but the costs of the coffee fertilizer trials are not; (iv) the project records indicate that the total cost of the project was nearly 20% higher than originally estimated. There are indications that the project's records include items not included in the appraisal estimates, namely expenditures for land compensation on building sites and tax components of a number of purchases. To determine the amounts involved would require a major research effort which is beyond the scope of this report. - 21 - .J.LLI JL . JL-.Le$.L"mL J." .L %;.L1Le5 PJ.Lu CLLu wl, 0A . U=L ± L.L6L..LJ V.± JJ±.VJ1=%.% costs compare as follows: Appraisal Estimate Actual roDu rerceun Million Million IDA 157.5 84 163.1/ 74 ur4 m. . Ca ^C U ov rnMLC-U-t r- ±2. ) C Total Mr. -9 100 221. U 100 During the completion mission Government made the point that it had been financing a larger proportion of the project's cost than envisaged. The above summary supports that view although it should be added that some of nthe expenditures may not actually De regarded as part or the project cost (para 3.41). Government obtained 100% IDA financing for expenditures totalling Fnu 83 million (38% of total cost) which represents part of the foreign exchange costs of the project. Actual foreign exchange costs were proDDly higher, however, since there is ample evidence that occasion- ally imported goods were procured locally and paid for in local currency. The true disbursement percentage for these and other local expenditures was 707 throughout the project period. Government never claimed for expenditures for the coffee fertilizer trials, conducted Dy 1SABu, and during the comple- tion mission it appeared that Government was not aware of the fact that provision for this had been made in the Credit Agreement. During the entire project period research costs were financed from OCIBUts normal operational funds. Government financed its share of project costs almost entirely through a special tax on all coffee exports. The tax rate was determined each year depending on anticipated expenditures and the proceeds of the tax (FBu 55.5 million over six years) were made available to the project. 3.43 Original disbursement estimates and actual disbursements are shown in the attached copy of a memorandum by IDA's Disbursement Division (Annex 2). The credit was fully disbursed on March 18, 1976, more than two years after the original closing date. As indicated earlier (para 3.33) late completion of the four washing stations was the main reason for the delay. Procurement 3.44 Procurement procedures were in conformity with Bank guidelines 1/ The actual amount disbursed differs from the original estimate because of exchange adjustments in both the Burundi and US currencies. - 22 - but on some occasions insufficient time was allowed for prospective bidders (see also para 3.34). Delays in delivery of ordered equipment and materials was, however, many times a very real problem and sometimes essential building materials, like cement, were not available for several months (in 1975 for example). Due to Burundi's isolated location and long transport lines all works need to be planned far ahead to be reasonably sure of a timely implementa- tion. IV. ORGANIZATIONAL ASPECTS 4.01 OCIBU, a financially autonomous parastatal institution, has since 1964 been responsible for a number of major aspects of Burundi's coffee industry, including: Ci) establishment and supervision of quality and packing standards; (ii) organization of country-wide coffee pest control campaigns; (iii) promotion of coffee sales; Civ) maintenance of communal hand coffee pulping centers; and (v) advise to Government on general coffee policy, in particular, producer prices. Moreover OCIBU was charged with the imlementation of a Dilot fertilizer distribution_program in Ngozi Province and for that purpose it employed a small number of aricultural officers and extension workers in the Province. 4.02 In view of its nrior experience it was understandable that OCIBU was called upon to implement the project. But the precise nature of OCIBUIs resnonsibilities had been the sub-ect of extensive discussion durini preparation and appraisal. The preparation report assigned the "main responsi- hilitv" for the nrnoent to nCTRT urder th' zpnerA.1 superviqion of thp Ministry of Agriculture while the project would "make full use of the existing 'tionwiTE The Director of OCIBU" but the preparation report did not indicate to whom he would report. Duvin the appanis1 4+ Ame-;Ad A +ce a+oh4lish a special Project Department in OCIBU for the implementation of the project and 4+roa co-"^aA +'he+ +ho 'hmsA P +Tim P" 4m^+ D)amew+mnn " nulA hk- +hkme4+ it asageed that the- hedofte rJetDeatent V-U- e hepoJc manager. He "would be responsible under the authority and contfo o:the Chairmen wnA Aname -ia+ f nrTTtTT f-w +h- Ao- + A- ma-nam + - ----P +h proe . The appraisal report stipulatXs fiFtheYThat coordination of ISABU, and the Ngozi provincial administration would be through the Minister . Agr J- , ca ru -1SA .J.440J. IItOU %AL..J WII L %II '.±O±Ld11 i L UJ111 /iJ iU _LLJVLUU Department does not have other responsibilities beides the implementation of ULte proetU. .Lue JAUmU1 as an Cutuy uz Lveu up UV UsZ expecuuions and has performed reasonably well. Relations between the Department Director ILIe. the project manager) aru OCiDu's General vjrector were excellenV .urng - 23 - the final year of the project but were strained at times in the preceding years. The Minister of Agriculture. in his capacity of Chairman of OCIBU. has maintained a consistent and keen interest in the proj ect's progress. The organizational set-up entailed however one ma.or draw-back. as in effect it divorced coffee extension from general agricultural extension work, which -remained under the Ministry of Aariculture. Almost all aualified staff of the Ministry's extension service in Ngozi Province were transferred to the new nrolect entity and the result has been that the Ministry's staff was reduced to about 80 extension workers of which less than ten had a minimum of basic agricultural education. Durina the project's implementation the two ex i.n organizations fuioned Rately, but the Ministry's extension service was unable to do a satisfactory lob since its staff was largely untrained and because of an almost total lack of transport means and funds. Thf 1Q74 annaial minn for tht qefnnr nroint observed that the Ministry's chief extension officer in Ngozi Province hitch-hiked regular- lv with nroict vh-Il- vhinh vn for him thp onlv way to at n touch with his field agents. Jealousies between OCIBU's and the Ministry's staff oyerr diff~~r~nrpq pr____o tbe Minitrylpc staff. The roe of the Ministry's extension service, and exeso worki oIther cropsn in genrAl, W1."i Pn+.i"Plir ianrs-a n +.iA appnraisAl rponrt fo-r thp first project although a number of pilot food crop activities had been proposed in the pearation report+ Theoraniaonn of the Proi+t+ cnffee extension service was more straight forward, its size smaller and its tasks simpler than rwould have enn the casse if it had not been taken out of the Ministry's General Extension Service. And it should be added that its nyrqf4ianny ho- I nas cler l wyimrovaA 4n +he cars P 4-ha t'k 4ea+ ts 41--l amemr,+-o+ nn "RE these are no good reasons for leaving the Ministry's extension service modest start in the area of other crops. 4.04 For the second project the situation has been improved, The two aiffsen prUc w18)apraen aranemn hao obfore u n agricu ff iD,. n wozi P ovince a e now uAben project's supervision. vie cnterted extoensiocn sevice now certainly in a beter shape to exyand its activities to arZ. -9 ia area and to comm (e work on hrcrp han-i -was lb, 4--atd n.1--A -P ~ LuJ.44~i L,. 0.L .. L wo efore the± first projct st U u ULJ Ullq= CIIU Ul the sec onWproject C1980) a permanent arrangement has to be worked out and the loia solio woU-ud be= that thle entire eXtenSiOn Service (inicluding6 that for coffee) would transfer back to the Ministry of Agriculture while the: coffee processing services would remain under OCIBU. V. PROJECT IMPACT Coffee Production 5.01 Trying to assess the project's impact on the level of coffee produc- tion in Ngozi province turned out to be one of the most difficult tasks during preparation and appraisal of the second project and, again, in the course of the evaluation of the achievements of the first project. The authors of the preparation report for the second project observed in 1974: "Quantitative information on Burundi in general, and on the coffee sector in particular,is extremely limited and highly unreliable. For coffee, about the only accurate - 24 - figures are those of realized FOB prices, and of exported quantities, which are usually equated with production". This is still true today and despite provisions in the second project to improve this situation it may take many years before reliable additional information will become available. 5.02 The annual coffe_production figures for the country as_a whole are quire accurately known since virtually all of it is being exported. Only insignificant amounts are consumed domestically in a few urban areas and the I gers r5Z;;ZTI;Z-18"E ;k-coff;ee. The export figures may be slightly influenced by unrecorded sales in the border areas with Zaire and. Rwanda but, again, the quantities"Tvo1ved are probably insignificant. The most striking feature in the series of annual production figures is the large fluctutation from one year to another. A record production of 28,0 ton green coffee in 1974. for example. was a 7 year low of 16.000 ton 1T57Table-7Weather conditions re i to^be the major determinant factor for the volume of the harvest and a year of unusual high production is invariably followed by one of lowlylds during which the trees recuperate from the previous year's heavy bearing. It is generally believed that consistent good maintenance ofthe plantations reduces the annual fluctuations to some extent since the trees are then in better shane to withstand the demanding impact of a heavy bearing. To eliminate the effect of the yearly fluctuations. 5 year moving averages of the annual roduction 0 een calcul N e CT e4)''. The figures show that the produc- tion of arabicac.offee (arown by smallholders) has been increasing at an average rate of 3% eEara r i the last n 3ea and that ofr&Mta coffee rownontfew nrivate etates by 2 nr year. Durin the last five yearq_including the excellent 197h and poor 1975 seasons, the average annual growth rate of the arabica praduction was somewhAt higher. namely 4.-% - 5.03 Information o~n the aeoaraphical disti-ibuti-np of ti -n lini very poor and unreliab Except for the production of the coffee washing tion, all coffee is pulped and dried by the farmers themselves and sold to local traders, who in turn arrange for transport to one of the hulling factories in Bujumbura. On arrival in Bujumbura it is impossible to determine the origin of the cof fee. All gricultural of ficers are' charged with the o etia t offe repeciYe end of each season. Both the manner in which coffee trading is organized and the ~ ~ ~ J men a hedspsl fte agricuIltural± oficr p±C~~reventu thlem from doing a satisfactory job and their reports necessarily involve a great deal of guesswork. in the past CIBU published the amounts of coffee traded on the -loc kets, as reported by the agricultural officers, in its annual reports. But it has not done so siue 1973, apparently because of an understandable lack of confidence in the figures. A comparison of the actual quantities exported in the various crop years with the quantities reported by the agricultural officers for the country as a whole indicate that the latter are usually some 10-25%0 too low. According to those estimates the production of ozj_province averaged 35% of the national production dring the period o 19-l.nThe year-to-year figures 01-37 in 199, 1972 and 19Y4, 357, in 1970 and 1971, and 40-42% in 1973 and 1975) suggest that the production of vgozi province has been growing at a somewhat faster rate than the national production. On the basis of the figures put forward by the agricultural - 25 - officers the 4-year-moving average production for Ngozi indicates an average annual growth of about .5% du ring_th-last-fo.ur__AAr*. As said earlier, this basis is rather shaKy. It also proves difficult to assess to what extent the growth in coffee production was caused by a_higx_ productivit of existing trees or by an increase in the number of trees. here are neither reliable statistics on the number of coffee trees in the country, nor is it known what the proportion of new plantings has been in the last decade. In 1974 OCIBU estimated the number of coffee trees in the country as a whole at about 50 million and during the appraisal of Ee second project t e number Ngozi province was estimated t17.5 million, QBrn Lr undi' 1. According to project management there were in 1974 8.5 million trees in the area of the first project, which corresponds to one-half of Ngozi's and one sixth of_the nationtstoa, Another estimate of OCIBU suggests that out of Ngozi's 17.5 million trees, 2.5 million were planted in the period 1967-1973. This implies an average annual growth rate of about 2%, which corresponds with the approximate increase in the number of families (new plantations are said to be estab- lished mainly by newly wed couples). 5.o4 If the above assumptions are correct (and it is impossible to either confirm or reject them) it can be concluded that in Ngozi province the coffee production has been growing by -about 2% per lear because of a corr Donina n --se i n iPlanat nd by- a further 37o because of increased productivity of the trees. 5.05 On the basis of an averaeaua 2pdU _2La2.1 IQ. m cKEn and a atiH! of 10 million the average annual production in the country as a whole was estimated by the appraisal team for the second project at 380_gram green coffee, or 1.33 x 380 = 500 gram parchment coffee, per tree. There is no good reason to change the second appraisal mission's estimate nor to alter the estimated increase of the produiny the project area from 7_ gram per ree t These estimates are inEline VKiEh a _ncrease in roductivity of 3% per year (para 5.0) and with the notion that the productivity in Ngozi is higher than elsewhere in the country. The actual increase may have been somewhat higher but certainly not as high as estimated during' the appraisal of the first project, namely an increase of 200 gram (from 425 to 625 gram parchment coffee per tree) with improved husbandry, excluding fertilizers, and another 200 gram (to 825 gram per tree) with improved husbandry, including the application of fertilizers. On the basis of an incremental production of 75 aram narchment coffe er tree and 8.5' million rees in tepoect arpa tetotal annual neremenut at ton archment. or 500 ton areen coffee. THe last Figure compares with an incremental production of 1,580 ton per year, as given.in the appraisal report. 5.06 There is clearly a need for better information on the nroductivity of the coffee trees at various levels of tree husbandry. In the absence of reliable yield figures for the prolect as a whole- it iq maningless +. tr +n d tangLthe effects various improved husbandry practices, including fertilizer. may have had on the vields- Unw the i pmjst management - 26 - has attempted several approaches at measuring coffee cherry production, but these have all failed or are statistically not sound. A small data collecting and evaluation unit has now been included in the second project and the individual cherry purchase records, which are now being kept at the coffee washing stations, may in the future also provide some better information. But still, in view of the long pruning cycle and the large annual fluctuations in production, it takes at least five years of consi tenA_monitoring before any useful concluiio-ns-cante drawn. Coffee Prices 5.07 The average annual export prices for Burundi's arabica coffee and the ofricial growers' prices during the project period were as follows (in currentrterMs): Average Export Price Official Growers' Price (US cts per lb green coffee (FBu per kg parchment f.o.b. Dar-es-Salam) coffee) 1969 29 32 1970 47 36 1971 35 35 1972 41 34 1973 52 38 19T4 48 45 1975 49 39 1976 n.a. 69 On first sight the average export prices during the last three years have been stable but the figures conceal very large fluctuations within the year. The average monthlX export prices fluctuated between US cts 74 and US cts 43. n betweend US cts 32 in 1975. Since July 1975 the exporEFr-Ts ie-finEreWed dramat ically and they are at pres-entheeeing US$1.25 per lb_fob Dar-es-Salam. In view of these sharp fluctuations the GovrenaE asTgeaerTITeen-successful in maintaining rather stable growerPs g reg, with the exception of .1975. In ri 1T75. Tus erre the start of the harvest season, the export market was extremely weak and the prices f.o.b. D 1~~'f ~~o ~ t 0~er-ISi. 'Gover-=ent had at that time no choice but to reduce the growers, price as well, Recently the growers' price has been increased to FBu per o o' the 1976 season (appraisal estimate FBu 35 per kg). Value of Incremental Coffee Production 5.08 Calculated at the very low early 1975 coffee export prices of about US cts 40 Der lb the incremental troduntion resulting from the p%o-eAt (estimated at 500 ton - see para 5.05) had a value of approximately US$440 000 fob Dar-es-Salam but Pt todav's nrinpq of nt lpt TTq.tl P5p 7%mv- 1+ ,project production would be worth about US$1,325.000. This compares with an i incremanta nronution of 1,580 tn with a i i f aou+ Ut$ 1250,000 estimated in the appraisal report. - 27 - Farmers' incomes 5.09 Since the average os are generalUy smal in se an oly a modest increase in the level of production has been achieved, average farm incomes have only small amn. No detaled imormusan is available but it can be safely assumed that incremental incomes varied substantially depending on the extent farmers have adopted the improved husbandry methods. On average the total increase was probably of the order of US*b per farm household in 1 . and UBz1U in 15q0 karter the increase of the inc This does not look very impressive but two observations are relevant in this connection: Ci) the daily wage rate in Burundi is or the order of US cts 35 per ; a farmer who would be lucky enough to find employment would need to work for at least a month to earn the equivalent of his incremental coffee returns; (ii) the number of participating families is very large, about D and this type of project rates well from the point or view of equitable distribution of project benefits. Farmers who delivered their coffee cherries at one of the four new coffee washing stations received about the same income from their coffee in 1975 as those who did not, since the coffee cherry price was about equivalent to the official coffee parchment price. But those farmers obtained that income with a lesse" gLor since they did not need to pup and dry the coffee t emselves. About 8,000 farmers lived in the vicinityiid mide use of the four washing facilities in 1975 for all or part of their coffee production, 5.10 Therr for several hundreds of R!Ple, mostly of a temporaryr sasonal nature on its construction sites, nurseries and coffee washing stations. Econoic Performance 5.11 The economic rate of return for the first project as a whole was estimated at 32X during appraisal; no separate rates of return were calculated for the project's agricultural and processing components at that time. The following observations are relevant for updating that estimate: (i) actual incremental coffee production cannot be accurately estimated but probably no more than one third of the appraisal target was achieved Cpara 5.05); CiiI actual export prices for coffee fob Dar-es-Salam exceeded appraisal estimates on average by about 25-30% during the period 1970-1975 and are expected to exceed those estimates in the future as well; at present export prices are more than three times as high\as forecasted in 1969; Ciii) the total project cost was about 15% higher and the project period about one third longer than estimated during appraisal; the actual annual expenditures were thus lower than originally estimatedt and (iv) about one auarter of actual project cost was related to the coffee processing component. This corresponds with the ax _ L U .L ~U ULL &I L 4 .OU L U ±L± 4L U c.1L r on drying trays, as estimated during appraisal, almost the the four coffee washing stations; 5.12 The up-dated economic rate of return for the project's agricultural component Vimprovement of coffee Cativa-LULIu z .. ,)p wLuln IuIeq=ru Ckunue shadow priced at 120% of the official rate (see Annex 4). This is an estimate and as indicated in paragraph 3.42 the cost figures uSeu in the analysis may include items which in fact should not be regarded as part of the projectts cost. But in any event the rate of return (ealculated at 1975 constant prices) is much lower than estimated during the appraisal of the project because the project*s impact on coffee production was over-estimated. The shortfall in production was thus far partly compensated by unexpected increases in coffee export prices. Actual economic performance of the project will ultimately very much depend on future coffee export prices and fertilizer import prices. During the last two or three years both prices were subject to large fluctuations. 5.13 With regard to the coffee processing component not much can be added to the arguments presented in the appraisal report for the second project. Unless the stations produce in the future at least 150 ton parchment coffee per year and a price premium of the order of US cts 8 per pound fully washed coffee can be obtained,the rate of return would be marginal. Recent information suggests that the production target can be reached but there is as yet no such proof with regard to the price premium. VI. GOVENMENT AND RANK FERFORMANCE 6.01 Government and Bank performance has already been commented upon in the earlier chapters of this report, sometimes in a negative sense, with regard to problems encountered during project implementation. As a more general statement it is fair to say that neither has performed badly. Government has kept a consistent and keen interest in the project and has staffed the project with personnel in a manner which should be regarded as satisfactory taking into account the severe limitations of the country's human resources. Government's persistent drive to expand the areas under coffee and to increase the number of coffee washing facilities conflicted with what the Bank regarded to be appropriate in the particular Burundi situation. The differences in yiewpoints have, however, never led to a breakdown in communications. The discussions continue under the second project and only the future can tell whose opinions are right. The correct understanding of the Bank s role in these discussions is to follow develop- ments closely and to spell out the risks and possible implications of specific Government policies anact ions. The Bank's record on this matter has been satisfactory during the last years. 6.02 During the completion mission Government officials expressed their satisfaction with the way the Bank has been supervising the prolect. There - 29 - was a complaint, however, that in many cases it took too long before the Bank responded to Government's queries. There exists indeed a genuine problem of communication between the Bank and Government which often is caused by trivial matters such as late arrival, or even loss, of correspondence and cables. Having also regard for the often limited understanding of Bank prat e a d aeureson teprt pQvnet ,y like Burundi, every a Bank reresentative on the scene. DIrecfcontacts between the Bank and the Government are now limited to one or two supervision missions per year and there is ample evidence that actions were delayed or problems remained unresolved until a supervision acquired the first hand knowledge often necessary for the Bank to take a position. Within the framework of the present supervision system the Bankts performance should be regarded as good, however. There was an excellent con- tinuity in the _superyision missions; all of these missions (twelve in total), e5kept for the last one, included at least one staff member who had visited the project on one or more earlier occasions. 6.03 There were no flagrant breaches of the covenants included in the Credit Agreement. VII. CHANGES IN THE SECOND PROJECT 7.01 Reference to some of the changes in the second prolect has already been made in the earlier chapters. They can be summarized as follows: (i) e of the entire (rince of jg2&j whereas the first Drolect was limited to the hiah Dotential coffee arowina areas; (ii) balanced develpme of coffee and other agricultural activi- ties through the inclusion of a food cron comnonent and experiments with other rural development activities, including the imnrovement of livestock nroduction; (iii) a unified extension service, rather than a separate service for -c-ffee C6y? Ee irst project) and other crops (Ministry of Agriculture): Uiv) a larger provision for cn??eP rpereh- narfAinla-rly in t.hp field of insect control; (v) a provision for a small team of internationall, recruited - v ntnrvals t assist in formulating work programs and advise on specific  ANNEX 1 BURUNDI FIRST COFFEE IMPROVEMENT PROJECT Project Description The project is the improvement, over a five year period, of coffee production and quality from about seven million existing coffee trees. These trees are grown on some 4,400 ha and are owned by about 44,000 smallholders. The project consists of: (a) the creation-of a specialized department in OCIBU for the extension of improved coffee production and processing methods; (b) the provision of in-service training for Burundian coffee extension and processing personnel; Cc) the implementation of a program of fertilizer trials on farmers coffee Dlots: (d) the provision of farm inputs and tools to Darticipatini farmers; (e) the establishment of a fertilizer sales program; (f) the construction of four pulping and fermenting factories, and the imorovement of about 80 communal pulnina centerst and (g) the improvement of about 350 small bridges on existing coffee tracks. Source: Appraisal Report , October 1968 WORLD BANK/INTERNATONAIL FINANCE CORPORATION ANNEX 2 A N ~KIDløtMi UM-IUt- IVLIVIUr /-\iNjUI T Nr. John Schwartz, ountry Programs II, E. rica DATE: May 6, 1976 FR Peter Gil, Diabursmenta Division SUBJECT: I7 BL - Ooffee Iprovement Proect rAabursements uder the above eredit were coipleted on March 18, 1976. For your information the folloving is the final allocation of proceeda for Category n Iber Total Total aud Description Allocation Diburøød la Vehicles & Pulping Paot. "quipr-nt $ 38,000.- $ 259,131.38 lb Peaticides I. Fert4iiser Fara 53,000.- 424,790.94 le Tbola and Materiale 84.,000.- 33,999.00 2a Civil Wrks, Const. Material & Services 197,000.-- 506,018.24 ing orm. Ta5,0.- IL,314.52 2c Salaris & Vages for Techni- cal Services 621,000.- 571,745.92 3 Unallocated 163,000.- $1,800,000.- $1,800,000.- Exahange Adjustment 129,019.45 Total $1.929.019 BURUNDI Coffee Improvement Project - Credit 147-BU Economic Evaluation Agricultural Comonent Benefits Costs Incremental Coffee Production,i.2/ Green Coffee- Prices - Value Incremental Actual 2/6/ In 1975 Constant 3/ SLiontre)_ fob Dar- am Local rrket Production at Local Prices Parchment Total Total Actual Constant 1975 equivalen' constant )4rket, constant Coffee parchment Green prices ices A / 1975 prices, 1975 prices per tree coffee Coffee USc/in USc/fib FBr?7S FBukg FBu'000 FBu'000 FBu'O00 Gram Ton Ton 1969 . - - - - - - 5,219 9,289 1970 - - - - - -- 20,383 35,262 1971 15 123 97 35 56 97 74 3,589 30,745 49,806 1972 30 246 194 41 62 107 84 9,312 37,909 57,621 1973 45 369 291 52 73 126 103 19,497 27,435 38,409 1974 60 492 388 48 55 95 72 23,280 29,041 33,251 1975 75 615 485 49 49 85 62 30,070 13,916 13,916 1976 65 533 421 - 100 173 150 63,150 - - 1977 55 451 356 - 89 154 131 46,636 - - 1978 45 369 291 - 86 149 126 36,666 - - 1979 40 328 259 - 81 140 117 30,303 - - 1980 40 328 259 -- 79 136 113 29,267 - - 1981 40 328 259 - No projections; value incremental 28,000 - - 1982 40 328 259 - production assumed to decrease 26,500 - - 1983 40 328 259 - gradually from 1980 to 1985. 25,000 - - 1984 40 328 259 - 24,000 - - 1985-88 40 328 259 - 64 111 88 22,797 - - Internal Economic Rate of Return, without shadowpricing of foreign exchange, 13% Same, with shadowpricing of foreign exchange, 15% NB. Benefits 100% foreign exchange, Costs 50% foreign exchange Foreign exchange valued at 120% of official exchange rate 1 Hulling efficiency (from parchment to green coffee): 79%. 2/ Decreasing incremental production from 1976-1979 and exclusion of ongoing maintenance costs from 1976 onwards consistent with appraisal report second project; Annex 1, paragraph 23; Annex 12,, paragraph 2 and Table 5. 3/ Actual prices have been converted in 1975 constant prices by using IBRD's Index of International Inflation. 4/ Green coffee prices, fob Dar-es-Salam, from 1976 onwards are assuned to be UScts 7/lb lower than projected prices for Guatemala, prime washed, spot New York (see appraisal report second project, Annex 3, Table 1). 5/ Local market value of green coffee FBu 23 per kg below projected fob Dar-es-Salam prices (see appraisal report second project, Annex 12, Table 7). 6/ Actual costs of the agricultural (coffee growing) component represent 75% of total project costs; the balance was spent on the project's coffee processing compoe nt. September 13, 1976 BURUNDI FIRST COFFEE Distribution of Fertilizers Year Month Number of Farmers Number of Coffee Trees Quantity of Fertilizers Farmers Total Under Proportion Total Under Proportion Total Per Tree Price of Project Using Project Fertilized Fertilized Fertilizer Administration Fertilizers Administrat ion (Number) (%) ('000) CNumber) (%) (ton) (gram) (FBu/kg) 1969 November 7,000 4,4oo 62 1,300 470 36 51 l09 Free - 1970 April 14,000 12,700 90 2,300 1,575 68 170 l08 Free November 28,000 23,000 82 4,4o0 3,210 73 224 70 Free 1971 April 43,000 32,600 76 6,900 4,750 69 332 70 Free November 47,000 33,400 71 7,700 4,800 62 336 70 Free 1972 April 50,000 34,600 69 8,000 5,630 70 368 65 Free November 51,000 40,800 80 8,250 6,340 77 439 70 Free 1973 April 51,500 42,600 83 8,300 6,230 75 428 69 Free November 52,000 36,200 70 8,460 4,770 56 317 67 Free 1974 April 52,000 27,4oo 53 8,480 3,710 44 386 104 6 November 53,000 - - 8,600 - - - - 1975 April 53,000 27,900 53 8,840 3,330 38 270 81 9 November 55,000 24,900 45 9,000 2,990 33 207 69 9 July 1, 1976 TABLE 2 BURUNDI FIRST COFFEE IMPROVEMENT PROJECT Prjct Costs Appraisal Actual Estimate Capital Costs Vehicles 2.1 6.0 Washing Stations 5.6 26.T Hand Pulping Center Improvement 2.8 Track ITmprovement 0.4. Moniteur Housing, Equipment Shed 1.5 Constructions (unspecified) 21 12. Sub-total 12.4 45.6 Contingency (10%) 1.2 - Total Capital 13.6 45.6 Inputs Pesticides 12.5 16.6 Fertilizer 34 .7 40.6 Dusters 1.9 Pest Control Materials - 1.1 Pruning Tools 5.6 1.2 Drying Trays 19.2 1.2 Sub-total 73.9 60.7 Contingency (10%) 7.4 - Total Inputs 81.3 60.7 Technical Services Staff Costs Expatriate 32.3 Burundian 45.2 Other, includes Research 6.1 Materials and Services 3/ 29.1 Wages and Social Security 86.2 Sub-total 83.6 115.3 Contingencies 8.4 - Tota,l Techntc4l Services 92.0 115.3 Grand Total l8. 221.6 1/ Source: Project Report, April 1976. 2/ Included in "Constructions" (unspecified), a category not included in the appraisal estimate. This category covers the cost of one house in Ngozi, a number of low cost houses for moniteurs and a total of eight fertilizer sheds. 2/ The break-down of the cost of Technical Services in the project report differs from the categories used in the appraisal report. The actual figures do not include the cost of the fertilizer research trials. July 1, 1976 TABLE 3 BURUNDI FIRST COFFEE IPROVE4ENT PROJECT Coffee Exports, 1958/59 - 1975/76, m ton Coffee Year -- 1/. Coffee rear- Arabica Robusta Total 1958/59 9,640 300 9,940 1959/60 20,727 900 21,627 1960/61 9,991 278 10,269 1961/62 12,937 196 13,133 1962/63 12,972 204 13,176 1963/64 5,856 197 6,053 1964/65 17,689 600 18,289 1965/66 12,826 731 13,557 1966/67 14,094 917 15,011 1967/68 17,929 753 18,682 1968/69 15,517 1,023 16,540 1969/70 13,299 1,251 14,550 1970/71 21,184 900 22,084 1971/72 23,365 1,780 25,145 1972/73 17,960 1,600 19,560 1973/74 19,601 1,794 21,395 1974/75 26,445 1,693 28,138 1975/76 14,528 1,701 16,229 1/ The coffee year in Burundi runs from the 1st of May to the 30th of April. Sources: Banque de la Republique du Burundi and OCIBU. July 1, 1976 rPA MT.V I BURUNDI FIRST COFFEE IMPROVEMNT PROJECT Coffee Eports, 5-Year Movinñ Averages, ton 195r/r9 - 1962/63 1,25 376 13,629 Jý'J.Lfl Y7 J..7 (U M -IuJ- -2- t in959/6n -1^. 193/64 1,97 355 .2,852 160/6V - 196Y/65 V',* 29ý «2,>8l 1962/63 - 1966/67 12,687 530 13,218 1963/6I - 1967/'68 13,679 640 14,319 1964/65 - 1968/69 15,611 805 16,416 -1969/70 14,733 935 15,668 1966167 - 1970/71 16,405 969 17,374 1967/68 - 1971/72 18,259 1,141 19,400 1968/69 - 1972/73 18,265 1,295 19,553 1969/70 - 1973/74 19,o81 1,465 20,546 1970/71 - 1974/75 21,711 1,553 23,264 1971/72 - 1975/76 20,379 1,713 22,093 Growth Rate (cumulative) - last 13 3.4% 12.5% 3.8% - last 5 years 4.4% 12.0% 4.9% July 1, 1976 BURUNDI FIRST COFFEE IMPROVEMENT PROJECT Expatriate Staffing Schedule 1969 1970 1971 1972 1973 1974 1975 1976 2 3 1 2 3 4 1 2 3 4 1 2 3 h 1 2 3 4 1 2 3 4 1 2 3 4 1 2 _ i ± L W - i -| _ | W . - Project Manager Carbon . b Dubois Heidebroek Training Officer Dubols Vacant Extension Officer Vacant Warnier Gely Pasteels (coffee) Processing Officer Vacant Veeger Construction Officer Gasparini (second Project' July 1, 1976 IBRD 3971 C6TRAL MAY 19771 7 BURUNDI REPUBLIC OP ZAIRE %ØF CO F EErrr DIN IC~ AHIUN RWANDA N IURUND f IA ANGOL A LAWE RWERU< RWANDA (éAN7ANI \'* I I *ÆI"UMBURAV///7// /OENSITY OF COFFEE TREES ?Very High High Medium Low Very Low .- Inteirnaonal Bokindane THE Bo UNDARIES SHOWN ON THIS MAP DO NOT 0 10 20 30 40 Km bo IMPlY ENDORSEMENT OR ACCEPTANCE By rHE '-- -a-- WQRLD BANK AND LT$ AFFILIATES. 1300   30* BURUNDI Projec t Area - NGOZI Province Prjet re -NOZ Pavnc'SECON[D COFFEE IMPIROVEMENT PROJECT ToMnz Naiional Roads ' ISABU Coffee Fertilizer Trials S Moin Roads - -limit of High Potential CofUee Area PROJECT AREA Secondory Roads .".." Existing Project Area (Credlr 147-BU) . Rivers 0 Coffee Washing Stations 55 1 c 15 20 Provincial Boundaries (Built Under Credit 147- BU) Ø j J - -- international Boundones Provincial Capital To KIh N Birarnbi 0 MUSHONGE Bwl BlURrEZ sl 1~ihg RU~0 <oso%e.we ..'. KAGOMA eushor MUY NGA G- NChash wo w KAYNA ubuga Rugeva , >a SRukagui * 'Pugazi3° - Ruhororo Buhig -Njo30 30' BUBANZA Motonga Bugeny R- W~c WA 1DA> BC g oa - K A R U Z 1 Bodanga cKARZ c 'Muninga OMwakiro Bonga !£A OF --- -(,,'-t ~ N \ Mor L.m bci,d ,Kibogoye - Gatonde C _______ ac M~co~ 'URAMAVYA 30i~ ~ iø o, !O -<- eo'*- Gtro ø 'j. _ _ _ __ar_m_ _li__a_ _ _ _ _ _ _a