Madagascar - Mangoro Forestry Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 4198 PROJECT PERFORMANCE AUDIT REPORT MADAGASCAR - FIRST MANGORO FORESTRY PROJECT (LOAN 1065-MAG/CREDIT 525-MAG) November 24, 1982 Operations Evaluation Deoartment This document has a restricted distribution and may be used by recipients only in the performance of their officiai duties. Its contents may not otherwise be disclosed without World Bank authorization.  ABBREVIATIONS ADTPA - Air Dried Ton per Annum CTFT - Centre Technique Forestier Tropical FMG - Franc Malgache FOFIFA/CENDRADERU - Center for Agricultural Research and Rural Development FANALAMANGA - Mangoro Forest Company (mixed-capital company, currently 100% state owned) MAI - Mean Annual Increment OPS - Operations Policy Staff of the World Bank  FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT MADAGASCAR - FIRST MANGORO FORESTRY PROJECT (LOAN 1065-MAG, CREDIT 525-MAG) TABLE OF CONTENTS Page No. Preface ............................................................ Basic Data Sheet ................................................... i Highlights ............ ............................................. iii PROJECT PERFORMANCE AUDIT' MEMORANDUM I. PROJECT SUMMARY .......................................... 1 Il. MAIN ISSUES ................., 4 A. Tree Growth .......................................... 4 B. Bank Involvement in the Mangoro Forestry Development.. 6 C. Additional Observations .............................. 9 Annex 1 - Comments from the Department of Forestry and Fishing Research ............................................... il Annex 2 - Comments from the Ministry of Agricultural Production and Agrarian Reform ........................................ 17 PROJECT COMPLETION REPOR'r I. Introduction ............................................. 25 II. Pre-project Developments ................................. 29 III. Implementation ........................................... 38 IV. Operating Performance .................................... 52 V. Institutional Performance and Development .... ............ 53 VI. Bank Performance ......................................... 55 VII. Environmental Impact ..................................... 57 VIII. Financial and Economic Reevaluation ...................... 58 IX. Conclusions .............................................. 59 Annexes 1-6 Maps IBRD 15266 (PCR) Mangoro Forestry Project IBRD 15290 (PCR) Project Location This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without WorId Bank authorization.   -i- PROJECT PERFORMANCE AUDIT REPORT MADAGASCAR - FIRST MANGORO FORESTRY PROJECT (LOAN 1065-MAG, CREDIT 525-MAG) PREFACE This is a project performance audit report of the (first) Mangoro forestry project in Madagascar for which Loan 1065-MAG in the amount of US$6.75 million and Credit 525-MAG in the amount of US$6.75 million were approved in December 1974. The final credit disbursement was made in Decem- ber 1978 and the loan account was closed in April 1982 after cancellation of the undisbursed balance of US$8,838 and after the by one year extended Closing Date of December 31, 1981. The audit report consists of an audit memorandum prepared by the Operations Evaluation Department (OED) and a Project Completion Report (PCR) dated April 7, 1982. The PCR was prepared by the Eastern Africa Regional Office. The audit memorandum is based on a review of the Appraisal Report (590-MAG) dated December 6, 1974, the President's Report (P-1539-MAG) of December 5, 1974, and the Loan and Credit Agreements dated December 23, 1974. Correspondence with the Borrower and internal Bank memoranda on project issues as contained in relevant Bank files as well as documentation related to the follow-on project (Second Mangoro Forestry Project, Credit 1161-MAG, US$20 million) have also been consulted and Bank staff associated with the project have been interviewed. A copy of the draft report was sent to the Borrower on September 2, 1982. Comments received from the Goverament of Madagascar are attached as Annexes 1 and 2. The audit finds that the PCR adequately covers the project's salient features, and the PPAM generally agrees with the conclusions. Some controver- sial points need to be raised, however, concerning the timely estimation of tree growth and the Bank's involvement in the Mangoro forestry development. These issues, as well as a few other points, selected for their relevance for the follow-on and similar projects, are therefore elaborated on in the PPAM to underscore ootential lessons for the Bank Group.   - il - PROJECT PERFORMANCE AUDIT BASIC DATA SHEET MADAGASCAR - FIRST MANGORO FORESTRY PROJECT (LOAN 1065-MAC, CREDIT 525-MAG) KEY PROJECT DATA Appraisal Actual Estimate Item (US$ m) (FMG m) (US$ =) (FMG m) Total Project Cost 17.2 4,000 20.2 4,601 I Cost Overrun 17% % Physical Overrun at Completion 35% Loan/Credit Amount 13.-/a Fully disbursed Date Physical Components Complete4b December 1979 June 30, 1979 Proportion of Time Underrun 10% Economie Rate of Return 13% 7% Financial Rate of Return 1O0 6% CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS (US$ million) Credit 525-MAG/Loan 1065-MAG FY75 FY76 FY77 FY78 FY79 FY80 FY81 FY82 Appraisal estimate 1.10 5.70 7.55 9.25 11.15 13.50 - - Actual - 1.50 2.60 6.20 8.60 11.85 12.26 13.50 Actual as Z of estimate - 26 34 67 77 88 91 100 OTHER PROJECT DATA Item Original Plan Actual First Mention in Files 1972 Government's Application for Loan 10/31/72 Negotiations 08/21-27/74 Board Approval 12/17/74 Loan/Credit Agreement Date 12/23/74 Effectiveness Date 03/24/75 07/08/75 Closing Date 12/31/80 12/31/81 Borrower Malagasy Republic Executing Agency FANALAMANGA Fiscal Year of Borrower January 1 - December 1 Follow-on Project Name Second Mangoro Forestry Project Credit Number 1161-MAC Amount US$20 million Credit Agreement Date October 29, 1981 MISSION DATA Date: No. of Manweeks Specializations Performance Types of Mission (Month/'?ear) Persons in Field Represented-/c Rt Trend/e Problems/f Identification 03//2 3 6 Preparation 01/13 4 + 1 17 Appraisal 4 12 Total 35 Supervision I 06/15 1 1 FO 2 2 F and T Supervision II 12/15 2 2 FO and AE 2 2 F and T Supervision I1l 07//6 1 1 FU 2 2 F and T Supervision IV 02/17 1 1 FU 2 1 F aoid T Supervision V 10/17 3 6 FO, AE and PPS I 1 Supervision VI 09/178 2 3 FO and AE I 1 Supervision VII 04/19 2 2 FO and E 1 1 Supervision VIII 11/79 1 4 E 1 1 Supervision IX 06/80 2 6 FO and E 1 1 (and Appraisal Phase Il) Supervision X 03/31 2 1.5 FO and E 1 1 Total 27.5 CURRENCY EXCHANGE RATE Name of Currency (Abbreviation) Franc Malgache (FMG) Average Exchange Rate: Appraisal Year Average 1974 US$ 1 = FMG 240.5 1975 US$ 1 - FMG 214.3 1976 US$ 1 = FMG 239.0 1977 US$ 1 - FMG 245.7 1978 US$ 1 = FMG 225.6 1979 US$ 1 = FMG 212.7 1980 US$ 1 = FMG 211.3 Completion Year 1981 US$ 1 = kMG 235.7 /a Credit of US$6.75 million Flus loan of US$6.75 million. /b See para. 3.01 and 3.16. Physical targets achieved ahead of time but project unit managed to extend project period and finance two additional planting seasons (June 1981). /c FO = Forester; AE - Agricultural Economist; E - Economist; PPS = Pulp and Paper Specialist. 7d 1 = Problem free or minor problems; 2 - Moderate problems; 3 = Major problems. /e 1 Improving; 2 - Stationery; 3 = Deteriorating. 7f F = Financial; T - Technical.   - iii - PROJECT PERFORMANCE AUDIT REPORT MADAGASCAR - FIRST MANGORO FORESTRY PROJECT (LOAN 1065-MAG, CREDIT 525-MAG) HIGHLIGHTS The first Bank/IDA supported forestry project in Madagascar financed a time-slice of a long-term silvo-industrial development program in the Mangoro Valle-x. Following pre-investment activities and trial plantations supported in part by other external donors the project aimed at establishing new forestry plantations and providing related infrastructure, research and training. Over 47,000 ha of forestry plantations were established, exceeding the appraisal target by 35%. The targets for road construction and fire- breaks also were significantly exceeded. However, only a fraction of the small pasture development component could be realized. Modifications were introduced during implementation regarding tree planting operations, construc- tion standards and methods for roads and buildings, and research and expatri- ate consultancy. The project's production objectives and returns on investment are, however, unlikely to be achieved. Tentative tree growth estimates for all plantations established up to 1981, of which about one-third are pre-project plantings, averaged orly 5.3m3/ha/year. The yields for project plantings averaged 8.0m3/ha/year as compared with 13m3/ha/year anticipated at appraisal. These results, if confirmed under the on-going verification, are jeopardizing the originally envisaged pulp mill development despite the fact that some improvements may be possible in current and future plantings. The project's reestimated economic rate of return is 7% compared with 13% esti- mated at appraisal. A decision concerning investments in silvo-industrial facilities could not be reached by the end of the first project phase, mainly due to the lack of a reliable forestry inventory and the questionable viability of such facilities, particularly for pulp production, and the lack of interest among potential foreign partners. IDA approved a second-phase project in 1981 for implementation over a three-year period. The following points may be of special interest: - tree growth estimates becoming available recently indicate very disappointing results. These crucial estimates, which to a certain extent contradict earlier findings and are still subject to verifi- cation, have appeared at a rather late stage; a lesson to be learned from this experience is that strict monitoring of yields is neces- sary in forestry projects where the type of species planted is not indigenous or clearly proven to adapt adequately (PPAM, paras. 8 and 12-15; PCR, para. 3.20);   - iv - it has not been possible to date to identify the type and organiza- tional arrangement of processing industries for the utilization of wood produced on project plantations. For this reason, and because forestry development is a long-term process, a final evalua- tion of the whole program of which the present project is a part must await the mnaturing of the plantations in another 5-10 years (PPAM, paras. 16-20; PCR, paras. 3.19 and 6.06); a follow-on project was approved by IDA in June 1981. The project was designed as an interim project to provide an opportunity for an assessment of future investment options but also contains a signif- icant plantation component, rendering it controversial in view of the reported poor yields of past plantings (PPAM, paras. 21-23; PCR, paras. 3.19 and 9.01); and construction standards for roads and houses were lowered during implementation, raising the question whether the case points to a wider tendency to over-design similar infrastructure components at appraisal (PPAM, paras. 24-25; PCR, paras. 3.13 and 3.26).   PROJECT PERFORMANCE AUDIT MEMORANDUM MADAGASCAR - FIRST MANGORO FORESTRY PROJECT (LOAN 1065-MAG, CREDIT 525-MAG) I. PROJECT SUMMARY-/ 1. Beginning in the mid-1960s, the Malagasy Government undertook surveys and studies, assisted by UNDP, FAO and the Centre Technique Forestiers Tropical (CTFT) aimed ai: increasing the country's forestry production. Trial plantations were started in 1968 and a gradually increasing area was planted yearly thereafter. This project, approved for Bank/IDA financing in 1974, was designed to support a five-year time-slice of a forestry plantation program which constituted the basis for developing a silvo-industrial complex in the Mangoro Valley. It was anticipated that investment decisions concerning industrial facilities would be made later, in time for such facilities to become operational in the mid-1980s. 2. Total project cost amounted to FMG 4,601 million compared with the appraisal estimate of FMG 4,400 million. However, the US$ equivalent of actual cost was US$20.2 million, or 17% above the appraisal estimate of US$17.2 million. The higher cost overrun in dollar terms resulted from the actual exchange rate during the implementation period having averaged about FMG 228/US$ rather than the FMG 255/US$ estimated at appraisal. Costs were considerably higher than anticipated at appraisal for afforestation and the purchase of vehicles and equipment, about the same for project management, and considerably lower for roads, research, training, studies and pasture improve- ment. The IDA Credit and Bank Loan of US$6.75 million each (as planned) together financed about 67% of the project cost rather than 78% as appraised. The balance was financed by the Malagasy Government. The credit and loan became effective in July 1975, four months behind schedule due to political developments in the country and a delay in setting up the required institu- tions. The Credit was fully disbursed by December 1978. The loan account was closed in April 1982, about 15 months behind the original schedule,A/ after cancellation of the undisbursed balance of about US$8,838. 3. The project, in addition to supporting the establishment of plan- tations, also included the construction of firebreaks and implementation of a fire prevention system, the construction and maintenance of roads, the construction of headquarter buildings, research and staff training, the preparation of a second-phase forestry and other projects, and pasture improvement. 1/ Adapted from the I>CR. 2/ The original Clos[ng Date of December 31, 1980 was, however, extended to December 31, 1981.  - 2 - 4. An area of over 47,000 ha was planted with trees, exceeding the appraisal target for the project by some 35%, but overall roughly in line with the original long-term planting program. Significant modifications were introduced in planting operations, in cooperation with the Bank, to alleviate trace element deficiencies, improve site selection, bring about better soil preparation for planting, and diversify into more promising plantation tree species. A total of 2,671 km of roads and track were built. This was 48% above the target in terms of road length, albeit construction standards were lowered during irmplementation to achieve this greater expansion of the road network. The standard of houses and buildings also was lowered so that over 2,000 units, or 78% more than planned, could be constructed. Over 1,000 km of firebreaks were established as compared to 250 km envisaged at appraisal. Pasture development, although very successful with the local villagers, remained nearly 60% below the target due to the high cost. Project activities continued about two years beyond the original completion date, largely because agreement could not be reached sooner on a second-phase project. 5. Training, executed internally by the project unit, was satisfac- tory. Research was carried out through the agricultural research institute, CENRADERU, now called FOFIFA. A four-year research contract becoming effec- tive in 1976 covered forestry research together with plantation trials and pasture-related agronomic research. A second-phase forestry project was prepared as required, but pre-investment work for other projects was not undertaken; the latter is not a serious shortcoming as funds and staff ear- marked for that purpose were put to good use elsewhere under the project. An expatriate forest adviser assisted project management for only one year instead of the originally proposed two, although his performance was very satisfactory. 6. To implement the project, a special unit, FANALAMANGA, was created. The unit has been competently staffed and has operated quite satisfactorily in most respects. Its reporting activities and relationship with other agencies and villagers in the project area have been given high marks. Regular project activities and special problems, as perceived by management, were handled competently. Initial difficulties with procurement, due to unfamilarity with procedures, were resolved in due time. FANALAMANGA's performance was less than satisfactory, however, in the monitoring of tree development on its plantations. Its trial inventory was greatly delayed and not fully processed, with consequences which are difficult to assess at this time. 7. Compliance with Loan/Credit Agreement covenants was generally good. One significant exception was a prefeasibility study undertaken in 1978 with UNIDO assistance without consulting the Bank.  - 3 - 8. Project benefits will derive primarily from wood production as raw material for industrial use. Tree growth was estimated at appraisal at 13m3/ha/year mean annual increment under bark (MAI). The estimate was based on initial growth trials and experience in other countries. With such yields, annual available cutting volume for age class 15 trees would have been 1.17 million m3, sufficient to eventually supply a 200,000 ADTPA (air dried tons per annum) bleached long fibered sulphate pulp mill.L/ However, mainly due to locational factors (slope, soil quality, rainfall, etc.) an MAI of only 5.3 m3/ha/year is estimated to have been achieved for the 12-year pre-project and project period combined; the estimate for the seven-year project period was 8.0 m3/ha/year.2/ The estimated cutting volume after 18 years (in- stead of 15) ranges only from under 400,000 m3 to 700,000 m3. Some en- couragement can be derived from the fact that yields from pre-project plant- ings are much worse than from plantings established under the project, and that some further improvements can be expected for current and future plant- ings as experience gained to date continues to be incorporated in cultural practices applied in the area. 9. Project evaluation at appraisal assumed that a second plantation phase and establishment of a logging industry and a pulp mill would follow completion of this project. Taking the cost and benefits of the project and related follow-on investments into account, the financial rate of return at appraisal was estimated at 10% and the economic rate of return at 13%. 10. Given the continuing, and even heightened, uncertainty concerning tree growth and, consequently, wood supply for industrial processing, an evaluation of the whole silvo-industrial complex using the appraisal method- ology is problematic at best at this stage. The evaluation in the PCR is thus based on stumpage rates to avoid speculation about future developments. Using this method, the project's financial rate of return has been estimated at 6% and the economic rate of return at 7%. 11. When it became apparent, in the late 1970s, that the viability of establishing a pulp mill in Mangoro as originally envisaged was doubtful, the processing of a second-phase project was delayed. In June 1981, IDA finally approved a second credit to finance continued forestry development expendi- tures over a 3-year period. The Credit (1161-MAG; US$20 million) became effective in July 1982. The second phase aims at expanding the planted forest area by 18,500 ha (17.5% of project costs) and supports studies and project preparation to help resolve problems of silvo-industrial development (9.0% of project costs). Nearly three-fourths of project costs are related to planta- tion maintenance, administration, and provision of physical infrastructure. 1/ Meanwhile, studies have shown that the economic size of such a plant is larger, requiring about 1.4 million m3 per year. 2/ Tree growth estimates are still subject to verification due to the underlying low sampling rate (PCR, para. 3.20).  - 4 - Since conclusive tree growth estimates were not available at the time of appraisal of the second-phase project, the assumptions used in the evaluation are similar to those used for the first phase. Il. MAIN ISSUES A. Tree Growthl/ 12. The central issue of this project is related to the disappointing growth of the trees planted in the Mangoro Valley. The first full inventory, undertaken in 1981 but still controversial because of the low sampling rate (0.007% vs. 1.0% recommended), produced an estimate of 5.3 m3/ha/year MAI for all 72,000 ha planted up to 1981 at Mangoro. This estimate includes 25,000 ha of pre-project plantings, most of which have been judged to be non-yielding, and about 47,200 ha of project plantings with an average MAI of 8.0 m3/ha/year. The MAI appraisal estimate was 13.0 m3/ha/year. A combination of factors contributed to the shortfall in yields, including: (a) planting on steep slopes where, because of erosion, soils have become unsuitable for afforestation; (b) trace mineral deficiencies, particularly of zinc, in most locations, often leading to stunted growth or death of trees; (c) occasional droughts; and (d) general adaptation problems of the main species planted, i.e, Pinus kesiya, which continue to manifest themselves in various symptoms as the trees grow older. 13. An alarming aspect of the problem is that forestry experts are not able to clearly attribute the degree of retarded growth to individual factors. Furthermore, adverse effects which were not observed earlier, occur as the trees grow older, reducing the prospects of enhanced growth towards the end of the growing cycle, even of those trees originally well established. While progress has been made in reducing new planting in non-yielding areas, i.e., areas where harvesting would not be worthwhile because of the poor stands, such areas to date have not been eliminated completely. The latest survey shows most of the pre-project plantings as being in non-yielding areas compared with 19% for the project period. 1/ This point is also extensively covered in the comments received from Government (Annexes 1 and 2).  - 5 - 14. Timing of Tree (:rowth Estimate. It is disconcerting that firm indications of likely poor yields did not become available earlier. One explanation, namely that yields in young trees are difficult to determine, carries considerable weight considering the fact that many international experts visited the young plantations without being overly alarmed by signs of stunted growth. However, since the difficulties in judging tree growth are well known to foresters, special efforts to carefully monitor yields through extensive sampling would have been required. FANALAMANGA staff was not successful in carrying out these responsibilities. In fact, the 1979 inventory results, delayed as they were, blurred rather than clarified the picture and, to some extent, misled authorities as they failed to direct the necessary attention to this key problem area. Bank supervision failed to provide the necessary pressure and guidance and thus may have contributed indirectly to creating the information gap.L/ Together with project manage- ment, supervision staff concentrated mainly on areas planted as success indicators and not enough on monitoring yields. One clear lesson to be learned from this experience is that the Bank should insist on strict moni- toring of yields in forestry projects where the type of species planted is not indigenous or clearly proven to adapt adequately. 15. Actual vs. Ant:;cipated Tree Growth. Plenty of expert advice was obtained during the late 1960s and early 1970s and, generally, the findings were positive and encouraging. Surveys and experiments were conducted and trial plantations were established, leading up to the formulation of the Bank/IDA project. The question can be asked how was it possible that a majority of the pre-project plantations must now be considered to have been established in unsuitable areas and how it happened that yield prospects have been so disastrously misjudged.2/ The yield problem is a fundamental one and not just a result of initial silvicultural mistakes which can be easily corrected. The trees simply do not grow sufficiently in many parts of 1/ The Region comments as follows: "Although the 1978 trial inventory was not a success, this was due in part to the fact that FANALAMANGA did not follow Bank guidance (PCR, para. 4.04). In addition, the trial inventory was followed shortly thereafter by the decision to carry out a full- fledge inventory by experts with professional skills. Thus, we do not believe there was much scope for improving Bank supervision. However, if a clear or comprehensive monitoring system had been designed at the outset to face the risk of a delayed start-up of the follow-on project, many significant delays would have been avoided and the supervision mission's pressure and guidance would have been more effective." 2/ These statements are based on the findings of the latest available forestry inventory. Regional staff caution that, since the inventory findings are still somewhat controversial and subject to verification, the statements may be premature.  the project area no matter how well they are cared for.lI Even today, with more than 12 years of experience, there is doubt that large-scale new planta- tions with acceptable yields can be established in the vicinity of the project area; smaller pockets of suitable locations have been located, however. With hindsight, it can be said that the pro,ject might have benefitted from a larger research and consultancy inputS- Research activities carried out under the project produced some good results with respect to disease control but were less successful in clarifying fertilizer response and in promoting other measures aimed at enhanced tree growth. The amount spent for research and training was below th.e levels anticipated at appraisal. Moreover, if the expatriate consultaney had been prolonged rather than cut short, the tree growth problems might have come into sharper fOcus at a more appropriate time. The last, alheit overoptimistic, yield estimate produced by the expatri- ate forestry expert dates back to mid-1976 and it is conceivable that the 1979 inventory would have been more successful had he been available to assist. Extreme caution must thus be exercised in the future when expert opinion regarding forestry yields, not backed up by long-term practical results under field conditions, is being used to justify a massive forestry investment program. B. Bank Involvement in the Mangoro Forestry Development 16. Forestry-based industries are a new development in the M4angoro Valley, hence, investment planning has been an exceedingly difficult task. In addition ta unproven yields, the long growing period of trees, as is the case with forestry investments, has aggravated problems related to determining appropriate end uses of wood and forecasting associated production costs, markets and prices. Consequently, when a decision was made in the early 1970s to establish large-scale plantations, alternative development models were discussed but, understandably, no decisions and investments related to wood processing were made. It was expected that information becoming avail- able during the 1970s would facilitate the planning of transport and proces- sing infrastructure so that by the time raw material would become available in the mid-1980s, the infrastructure would be in place. 1/ See footnote 2/ page 5. 2/ The audit considers the research and consultancy inputs made available under the project as highly beneficial to the project. It wishes to emphasize, therefore, that an expansion of these inputs beyond appraisal levels could reasonably have been expected to be fruitful. Instead, there was a regrettable shortfall in spending for these purposes below appraisal targets.  - 7 - 17. In retrospect, events did not occur as expected. A viable proces- sing industry has not been identified, partly because of the uncertainty about tree growth and partly for reasons related to adverse world market conditions. Furthermore, it has not been possible to bring prospective foreign partners into the discussions on future development, nor are the prospects for future involvement of essential foreign partnerships promising in Madagascar. The original target date for logging and processing operations to begin in 1985 has thus become illusory, and even the revised date of 1988 will not be met unless essential investment decisions, e.g., on the type of end product to be produced, are made soon. Although the follow-on project provides support for investment planning, it has suffered delays and is encountering ojerational difficulties (para. 21). 18. Initial Bank Involvement. It is appropriate to distinguish between the circumstances facing the Bank at the time the decision to participate in the project was made, and the experience gained to date on the basis of which an interim assessment of that decision might be attempted. The Bank faced a situation in the early ]970s when forestry development ranked high in Govern- ment priorities and, given the paucity of development projects in the Madagas- car agriculture and forestry sectors, it was not foregoing more attractive proposals by considering this project. The appraisal, though having to work from a scanty factual basis, did take into account the results of experimental plantings and the investigations undertaken with the assistance of UNDP/FAO and CTFT, a French bilateral research organization. Available market studies indicated international market potential. With the assumptions underlying the appraisal analysis, pacticularly a potential yield of 13 m3/ha/year, the project was judged to be viable. 19. There undoubtedly were substantial risks associated with the pro- ject. Government's strategy, in supporting such a relatively large, long-term and capital intensive venture with sizeable maintenance and follow-on invest- ments, in itself constituted a high risk factor. The technical risks associa- ted with converting existing grassland on extremely poor soils into commercial forests, using species without extensively documented local growing history, were especially high. The risky nature of Malagasy forestry development was also evident from the poor performance of other plantations in the country and from the absence of an experienced wood products export industry.!/ Proposed means to alleviate those risks included research and continued experimentation to promote tree growth, and the possibility to adopt alternative end uses of the harvest from the plantations in the event of changed market conditions. 1/ The Region does not agree "that managerial and economic risks were evident from the poor performance of other plantations in the country. There are many excellent plantations in other parts of Madagascar".  20. Eight years after the appraisal the results look disappointing. The project's risks turned out to be formidable and the means to cope with them not quite adequate. In retrospect, the yields of non-project plantings were completely misjudged and yields of project plantings overestimated by 40%, according to currently available, but tentative, inventory data. The foreign investment climate is not as favorable in Madagascar as originally anticipated, and world market prospects are more uncertain. Decisions on silvo-industrial investments have proven intractable so far. While the audit finds these adverse circumstances to be of great concern, it recognizes that forestry development is a long-term process and that final evaluation of the whole program has to await the maturing of the plantations in another 5-10 years. 21. Continued Bank Involvement. The project was supposed to be comple- ted in late 1979, with a second-phase planting (and possibly mill investment) project to follow. In fact, the essential physical targets (afforestation, roads, fire breaks and buildings) were achieved on schedule. The processing of the second phase ran into difficulty, however, as a result of disagreements between the Bank and Government about plans for silvo-industrial development. A UNIDO study (1978) indicated that, under changed market conditions, a pulp mill, to be viable in Mangoro, would have to be somewhat larger than origi- nally proposed and Government wanted to proceed on that basis. The Bank, however, refused to endorse plans for the construction of a larger pulp mill, for good reasons, as can be seen now. When an IDA credit was finally approved in 1981, it was a compromise stopgap measure designed to gain further time (i.e., 3 years) for assessing tree growth and conducting studies related to the processing industry to be established.L/ However, additional plantings (18,500 ha) were also included. 22. In retrospect, the second-phase project appears correct in principle because it has served to keep the dialogue going with all parties concerned and to preserve, for the time being, the existing forestry institutions. The 1/ The Region considers this ".....an interim project designed to give the Government an opportunity to undertake a systematic and thorough assess- ment of the future investment options." Specifically, it is a condition under the project that a "Feasibility Study" be carried out under an existing agreement with UNDP (Document No. MAG/80/002/C/01/12, dated October 3, 1980), with FAO acting as the executing agency. While no IDA Credit funds are involved, IDA obtained assurances that the consul- tants employed for the study will be satisfactory to IDA, reports and information will be provided, views exchanged, and meetings held with IDA participation. The study is to be completed by June 30, 1984.  - 9 - planting of completely new areas at this stage, however, must be questioned.l/ There is no guarantee that tree growth in the new areas will be better, nor will this additional area generate sufficient wood supplies for what experts believe to be the minimum viable size of a pulp plant; for major alternative uses, existing stands are already adequate. Had the results of the tree growth survey been available at appraisal, the follow-on project either would have been designed differently or not approved at all. 23. The audit concludes that, in the follow-on project, even greater emphasis should have been given to the design and expeditious execution of end use studies for wood produced at Mangoro and the consolidation of existing plantations.2/ The latter includes the disposal of trees in non-yielding areas and the salvaging of threatened stands. While the project still can (and probably will) be modified in the direction indicated, the loss of time since achieving the initial planting target in 1979 and the controversies surrounding the tree lnventory and, of late, the on-going silvo-industrial studies, have hampered forestry development at Mangoro. To avoid recurrence of the experience with the First Mangoro Forestry Project in the future, the Bank ought to require more rigorous monitoring standards than were applied there and set stricter intermediate planning targets for time-slice forestry projects for which (a) crucial information is to be generated during the initial years of the project, and (b) the planning process for all related investments has not been completed when beginning project implementation, i.e., the prevalent case. C. Additional Observations3/ 24. The project exceeded its physical targets by 48% for roads and 78% for houses and buildings. It is explained in the PCR (paras. 3.13 and 3.26) that cost savings were achieved for these components by (a) undertaking construction by force account instead of by contractor, and (b) accepting deviations from the appraisal norms, i.e., switching to lower-grade standards. The audit has reservat:ions about attributing cost savings to force account activities vis-a-vis zontractor work without a detailed explanation. Of 1/ OPS advises: "The forestry related issue should be resolved before proceeding further with the industrial studies and [IDA] should obtain some independent assurance that, if properly maintained from here on, the plantations will grow on to produce logs of merchantable size. Some consideratiokn needs to be given to appropriate thinning schedules. If such assurance is not forthcoming, planting should be stopped alto- gether." 2/ OPS states: "The impact on economic viability of the Mangoro Project needs to be re-assessed on the basis of both lumber and pulp production alternatives." The focus of the on-going studies may thus turn out to be too narrow. 3/ Further remarks on this point are also included in Government comments (see Annex 2).  - 10 - particular interest would be to know whether the method of accounting used by the organization undertaking the construction work accurately attributes all capital and overhead costs to the tasks performed and what the actual savings achieved by not employing contractors were, both financial and economic. 25. The feature of lowering construction standards during implementation would imply that either the project was overdesigned at appraisal or quality or safety were being compromised during implementation. There is no evidence that the latter is the case. This experience suggests, therefore, that there could be a tendency in similar projects to overdesign infrastructure compo- nents and evaluations and audits should focus on this aspect more closely in the future. 26. A final observation concerns the use of residual Bank loan funds. In the last (eighth) year of disbursement, about US$1.6 million was disbursed for the purchase of vehicles and equipment, accounting for 47% of total expenditures for vehicles and equipment and resulting in overdisbursement for this purpose compared with appraisal estimates of 62%. Since these expendi- tures were not needed to achieve the immediate planting targets, which were already exceeded by about 20% at that time, such disbursements at the end of the implementation period must be viewed with skepticism.L/ Theoretically, the equipment could have been used for non-project related purposes or even sold after the closing date of the Loan. The practice is of less concern in projects where progress is satisfactory, follow-on investments are contem- plated and the risk of controversies developing around any of the investment phases is minimal. 1/ In the Region's view: "Although such skepticism would normally be quite justified, in this case the heavy disbursement at the end of the project period was made so as to enable the project entity to continue, in spite of Madagascar's very difficult foreign exchange situation, operating and maintaining its existing assets until Phase II could be effective. When project requirements were assessed during the appraisal of Phase II the existence of the recently purchased equipment was taken into account. Madagascar's foreign exchange situation had deteriorated drastically during the project period - and the apparent [over-disbursement] at the end of the project, when Phase II had already been agreed on, repre- sented a conscious shift in resources, undertaken with the agreement of the supervisory staff. In retrospect, in our view, the decision seems reasonable and appropriate."  - ll - Annex l Page 1 October 27, 1982 E-407/83 French (Madagascar) OED JCB:cc CO1