Colombia - Upper Magdalena Pilot Watershed Management Project

of Document The World Bank FOR OMCIAL USE ONLY Rqpt No. 11390 PERFORMANCEAUDIT REPORT COLOMBIA UPPER MAGDALENA PILOT WATERSHEDMANAGEMENT PROJECT (LOAN 2069-CO) DECEMBER 1, 1992 Operations Evaluation Department This document hasa restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCYEQUIVALENTS Name of Currency (Abbreviation) = Peso (Col$) CurrencyExchangeRates: Appraisal Year 1981: US$1.00 = Col$ 57.27 Completion Year 1987: US$1.00 = Col$ 194.3 WEIGHTS AND MEASURES Metric System ACRONYMS AGR Agriculture and Rural Development Department AGREP Agriculture Department - Division of Economics and Policy BTO Back-to-Office Report CPS Central Project Staff CVC (Corporation of the Cauca Valley) DNP (National Planning Department) DRI (Integrated Rural Development Program) FEDECAFE (National Federation of Coffee Producers) GDP Gross Domestic Product GOC Government of Colombia HIMAT (Colombian Institute of Hydrology,Meteorology and Land Management) ICA (Colombian Institute of Agronomy) IDB Inter-American Development Bank INCORA (Agrarian Reform Corporation) INDERENA (National Institute of Renewable Natural Resources and Environment) ISA (Electric Services) OED Operations Evaluation Department PCR Project Completion Report PIF Project Information Form PIN (Plan for National Integration) PPAR Project Performance Audit Report PROCAM (Upper Magdalena Watershed Management Project) R&D Research and Development (Upper Magdalena Basin Project Unit) RENORDE (National Network for Technical Cooperation in the Management of Watersheds) SCF Soil Conservation Fund SENA (National Training Institute) TOR Terms of Reference  FOROFFICIL US ONLY AUDffREOR PERFORMANCE COLOMBLA UPPER M1ANAG1NT PROJECf PILOT WAITERSHEID MAGDALErNA (LOAN2069-CO) CONTENTS PREFACE .................................................. i BASICDATASHEET .............. ................................ ia EVALUATIONSUMMAl4RY ............................. v AUDIT REPORT .................................... PERFORMANCE 1 L PROJECTBACKGROUND ANDDESIGN ...................... 1 Background ............. ................................ 1 Identificationand Design .................................. 2 Project Objectivesand Components .......................... 3 InstitutionalFramework ................................... 5 Project Costs,Budgets,and Conditions ........................ 5 IL IMPLEMENTATION EXPERIENCE .......................... 6 Postnegotiationand MobilizationPeriod ....................... 6 Mobilizationto EffectivenessPhase .......................... 8 ExecutionPhase ......................................... 9 IlL PROJECT IMPACT ....................................... 12 Achievementof Project Objectives ......................... 13 Farm and Tree-Crop Systems .............................. 13 InstitutionalImprovement ................................ 14 Planningfor Phase II .................................... 16 Replicability--ThePath to Sustainability ...................... 16 IV. FINDINGSAND MAINISSUES .............................. 18 Project Design and Preparation Shortcomings ................... 20 The ConceptualBasisfor Watershed Managementin Colombia .............. 25 The InstitutionalFrameworkfor WatershedManagement..... ...... 30 The LegislativeFrameworkin Colombia ....................... 31 UnderstandingWatershed ManagementEconomics ..... .......... 32 ANNEXES Annex 1 . ............................................... 35 Annex2 . ................................................ 39 BIBUOGRAPHY. .................................................. 41 MAP- IBRD 15847 This document has a restricteddistributionand may be used by recipientsonly in the performance of their oMcialduties. Its contents may not otherwisebe disclosedwithout World Bankauthorization.  -1- PERFORMANCEAUl)lT REPOR COLOMBIA UPPER MAGDALENAPILOT WATERSHEDMANAGEMENTPROJECT (LOAN 2069-CO) PREFACE This Performance Audit Report (PAR) was carried out in January/February 1992 for the Upper Magdalena Pilot Watershed Management Project in Colombia, which was sup- ported by Loan 2069-CO. Initial project identification and design activities were begun in 197° by the principal implementation agency,Instituto Nacional de Recursos Naturales Renovables y del Medio Ambiente (INDERENA), and were followed up by various Bank preparation missions in 1980. Preappraisal and appraisal missionswere undertaken in the first semester of 1981,while project negotiation with representatives of the Colombian authorities took place in Washington in November 1981, and the project was signed on December 15, 1981. It was not until January 1983,however, that the preliminaryconditions for effectiveness, mostly related to staffing and inter-institutional agreements, were achieved and that the project became effective. The original project loan in the equivalent amount of US$9 million 1/ was intended to support the National Government's activities over a five-year implementation period and with a total cost of US$27.3 million. After numerous implementational difficulties and delays in the years thereafter, the project was terminated on June 30, 1987,its original Closing Date. Actual disbursements under Loan 2069-CO amounted to US$2.37 million (26 percent), and the balance of US$6.63 millionwas cancelled in January 1988. This report has been prepared on the basis of a careful review of project docu- mentation including the Project Completion Report (Colombia - Upper Magdalena Pilot Watershed Management Project Report No. 8481, March 1990),the President's Report (Report No. P-3106-CO),2/the Loan Agreement, and project files and reports available in both Washington and Colombia. An Operations Evaluation Department mission visited Colombia in January 1992, during which time intensive discussionswere held with the concerned governmental authorities. A field visit to the project area was also undertaken. The Project Completion Report provides a broad overview of the project and its implementation experience and problems. The PAR has endeavored to elaborate key issuesof particular relevance for both the Bank's and national government's activities in watershed management. Following Operations Evaluation Department procedures, copies of the draft PAR were sent to the Borrower on June 4, 1992,for comments by July 22, 1992. No comments were received. 1/ Al mferencea to the Bank'sloan amount in this report are in USS equialent Acoordingto a footnote In the openingpaga of the President' Report "Inview of the pilot nature of this project, an appranul reporths not been prepared.0  PERFORMANCEAUDIT REPORT COLOMBIA UPPER MAGDALENAPILOT WATERSHEDMANAGEMENTPROJECT (LOAN 2069-CO) BASIC DATASHEET KEYPROJECT DATA Apprai Actual or Actual s % of lstimrte cw t a Aoisi 13Atlmat Project comt(USS milin) 273 &.1 30 Loan amount (USS milin) 9.0 2.4 27 Economicrate of return (%) n.a La. Insttutionl pefrmance Satidactory Unsti Number of direct benefidiarkeso .A. KEYPROJECT DATES Appil Actualor Actual a % of Estimate Curet Estil ADsor Eatimac Date of board approval 12A1SI1 Date of effectivenes 07/15R82 01131/3 Ckoing Date 06130M NOW307 Project impkmentation peod t 60 53 88 CUMUIATIVEDISBURSEMENTS EM EE El EM EM E8 EM Apprais estimte (USS mInion) 0.9 2.9 48 6.5 8.0 9.0 Actual (USS million) - 0.75 122 1.85 2.16 2.16 237 Actualas*of rappaladmats 0 26 25 28 28 24 26 Date of fin dbbursemnt 09 Amount canelled US$6.6 mllion &I Number of montihm effatvc to aoWUgDate.  Iv I - MISSION DATA Stane Date No. of Staff/Dayz Specialization Performance Types of (Mo/Yrj Pesons in Field Represented I Ratin, h Trend S/ Problemsy Identification 03R0 1 7 c - - - Preparation 10R0 3 63 c,d,e Preappraisal 02R1 2 14 c,unL Appraisal 04R1 4 28 c,d,unk,unk Poetnegotiations 1241 3 42 a,f,g - Supervision 1 09R2 1 14 g 2 1 MF 2 03R3 2 28 unk. 3 2 FMT 3 10/83 1 7 g 2 1 FMT 4 02R4 3 42 g,a 2 2 TM 5 05/O4 1 7 b 3 3 MT 6 02485 1 14 a 3 2 MTF 7 0486 S/ 3 42 unk. 3 2 MTF 8 09R6 1 7 f 3 2 MTF 9 1247 1 3 a - - STAFF INPUTS f/ (in staff-weeks) FY81 FY82 FY83 FY84 FY8S FY86 FY87 FY88 TOTAL Preparation 28.6 28.6 Appraisal 10.7 4.6 15.2 Negotiations 153 15.3 Supervision 8.9 13.6 21.0 10.7 15.6 20.6 3.7 94.1 Total (staff-weeks) 39.3 28.8 13.6 210 10.7 15.6 20.6 3.7 153.2 OTHER PROJECT DATA Borrower INSTITUTONACIONALDE LOS RECURSOS RENOVABLES Y DEL MEDIO AMBIENTE (INDERENA) Executingagency INDERENA Follow-onproject: ua. VI or agricultural economist;c - mision leader, d - environment engineer, a - agriculturist;b - ecoonomist e - watershed management;f - lvestock specialist;S - irripton engineer, unk. - unknown W 1 - problem free or minor problems;2 - moderate problems; 3 - major problems St 1 - Improving 2 - statiooay; 3 - worsening IV M - managerial;F - financial;T - technical V/ Combinedsupervion nd mid-term evaluation g Sou TRS data information from LACCA  v - PERFORMANCEAUDITREPORT COLOMBIA UPPER MAGDALENAPILOT WATERSHEDMANAGEMENTPROJECT (LOAN 2069-CO) EVALUATIONSUMMARY Introduction extension; the Servicio Nacional de Aprendizaje (SENA), for extension and 1. Some of Colombia's most modern training in environmental awareness and natu- development achievements are linked directly ral resource management; and the Instituto to the environmental stability of some of the Nacional de Hidrologia, Meteorologia y most backward areas. The national depen- Adecuacion de Tierras (HIMAT), for dence on hydroelectric power, potable water hydrology and river basin studies. supply for the emerging towns and cities, and the importance of irrigated agriculture for 4. Proyecto Cuencas de Alto Magdalena export earmings are all directly tied to the (PROCAM) was to develop the techniques condition of the watersheds in the Andean and watershed management plans in three Region. In the late 1970s, cognizant of the representative pilot watersheds in the increasing ecological deterioration, the Provinces of Tolima and Huila. These pilot National Institute for Renewable Natural programs were intended to cover a total area Resources and the Environment of 200,000 hectares and address the needs of (INDERENA) requested the assistanceof the approximately 5,000 target farmers, with the World Bank for a Pilot Watershed Manage- followingcomponents: ment Project in the upper reaches of its most important river--the Rio Magdalena. * On-farm investments Objectives * Public works * Extension and training 2. The main objectives of the project * Staff training were: (a) to develop viable farming and tree- * Research and studies crop systems as a basis for sustainable land- * Environmental protection use; (b) to improve the institutional capability * Preparation of phase II. for promoting and supporting watershed management and development; and, (c) to 5. Total project cost was estimated at prepare the informational baseline for an US$27.3 million,for which the Bank agreed to investment project for the entire Upper make available a loan of US$9 millionto cover Magdalena Watershed. the foreign-exchangeneeds of the project. 3. In addition to INDERENA, which had Implementation Experience the overall role of administration and coordination of the project, several other 6. PROCAM, first identified in late Government of Colombia (GOC) agencies 1979,was subsequentlyapproved by the Board were provided specialist services: the Caja on December 15, 1981. It was only shortly Agraria was responsible for a credit program after the completion of negotiations and the aimed at financing small farmer investments in signature of the agreement that those involved improved agriculture and soil conservation in the project, both at the Bank and at measures; the Instituto Colombiano de INDERENA, realized that it was going to be Agricultura (ICA), for farm-levelplanning and extremelydifficultto implement it as designed.  - vi - 7. During the postnegotiation and mobili- needed to come to grips with the overall zation period, a Bank mission delegated to reorientation and to put this flawed project assist with the preparations for loan effec- back on track. Despite continuing Bank tiveness detected a series of shortcomings supervision missions, which outlined the needing attention. Of greatest concem, as was problems and constraints, only marginal repeatedly emphasized during project prepa- successwas achievedin rectifyingthe situation. ration but apparently unresolved, was the Among the most significant implementation complex institutional setting, particularly the problems that ultimately led to the Bank's inter-institutional agreements between refusal to extend the project, were the INDERENA and the other GOC agencies. following: Even after the project had been signed, little had been done to coalesce these agreements. * The failure to design and implement a Only in January 1983, after several post- monitorlngand evaluationsystem essential ponements of the deadline for effectiveness, to guide the pilot activitythrough analysis was INDERENA finally able to formalize of problem areas and justification of cor- these agreements. In retrospect, the nature of rective actions and to provide, in part, the these interagency agreements inhibited the baseline for the second phase. innovative and complex approach that was originallyforeseen as the vehicle for problem- * The failure of the creditsystemfor on-farm solving and for development of integrated interventions: only 270 (5 percent) of the watershed management plans and practices. planned 5,000 loans to small farmers, which were designed to assist them with 8. Another serious shortcoming was the the improved agriculture and soil conser- inadequate elaborationand implementationof vation practices, were approved. The the various components of this complex, inte- credit line and its requirements proved grated approach to watershed management. beyond the capacity of the majorityof the The seventeen subcomponents had appeared smallholders. to have been designed independently, without appropriate sequencing from R&D to field implementation. Bank staff recommended a * A low level of operationalcapabilitywithin team of short-term consultants to work with the INDERENA project unit, which was INDERENA in the final preparation of the charged with the direction and leadership project for effectiveness;it was not, however, of the project, resulted from staff until after effectiveness that this additional shortages, overwhelming administrative technical expertise was contracted. Indeed, responsibilitiesinherent in the multiplicity difficulty in recruiting and hiring suitable of subsidiary agreements, and limitations experts which the Bank deemed essential for to its financial resources because of this project (characterized early on as a successive,austerity govemment budgets. technical assistance project) was to remain a major constraint of implementation. * A number of technical difficultiesmade it 9. Project effectivenesswas finallyrecog- impossiblefor the project to live up to its nized by the Bank, with some caveats regard- implementationschedule. Trialsestablish- ing the need for additionaldevelopment of the ed to test new production with conserva- farm models,strengthening the staffing,and an tion technologies were never completed overall realignment to take account of the and thus undermined the scientific basis delays and lessons leamed during inception. to substantiate many of the pilot inter- ventions. 10. Political uncertainties and a changing cast of INDERENA directors and project * Disbursement shortfalls. Both imple- directors characterized the project implemen- mentation difficulties and constraints on tation period and stifled the determination financial resources, as a result of austerity  - vii - budgets, combined to severely limit the 13. Impacts of other kinds, i.e., conceptual utilization of loan funds. This was and institutional, however, bear both mention exacerbated during the same period by and scrutiny. The local understanding of the the continuing rapid devaluation of the important relationship between the down- Colombian peso. stream user and the upstream resident is still vivid and real, the inhabitants are presently living the dilemma of watershed degradation * The Bank/INDERENA relationship was and appear to have resolved to solve it. hampered by the overall design of the Watershed management is now a part of the project, involving numerous instances local development plans of the municipal requiring Bank approval of specific corporations charged with rural development. agreements, subcontracts, and Terms of Their interests are now in their own hands, Reference (TORs). This situation proved and with them, the authority, responsibility, an overwhelming administrative burden and accountabilityto their constituents for the for INDERENA and its project unit as stewardship of the resource base. Some farm- well as for the Bank staff charged with ers are making changes in their land-use prac- supervision of the project. As might be tices, shifting off the slopes and seeking new expected, Bank supervision was no sub- ways to earn a living. stitute for the longer-term technical assis- tance that was never put in place. 14. The national government too has continued to give priority to watershed 11. In late 1986,the Bank, havingconclud- management as part of its socioeconomic and ed that the project was some two years behind rural development strategy. The new Consti- schedule and that the financial and technical tution will create the "Corporacion Autonoma feasibilityof the watershed management plans Regional del Rio Grande de la Magdalena,' and practiceswas still highlydoubtful, decided and the Forestry Action Plan for Colombia to close the project on its originallyscheduled also clearlyreflects the concern for watershed completion date-- June 30, 1987. This decision deterioration. There has been some residual was taken despite requests from INDERENA impact among the various institutions involved for an extension and redesign phase. Accord- in PROCAM; both ICA and SENA have insti- ingly, an undisbursed equivalent amount of tuted programs aimed at watershed issues. US$6.6 million or 74 percent of the original INDERENA has regrouped its Division of US$9 millionloan, was cancelled;only US$5.7 Watershed Management and is carrying on million had been disbursed by the GOC over with two programs as a direct sequel to the the five years. project. Project Impact Findings and Main Issues 12. There can be little doubt that 15. The overall design, preparation and PROCAM achieved some impact, albeit far implementation record of the Upper less and in ways not entirely nor originally Magdalena Pilot Watershed Management foreseen for it. Overall,theproject experience Project is so fraught with problems and delays has been, however,unsatisfactory.The physical that it is difficultto isolate particular problems achievements in the field are far less signi- and dissect them for lessons leamed. One ficant as indicators of impact, because ulti- thing is manifestlyclear, however. Underno mately, these farm-level interventions proved circumstancesshould the needs and oppomuni- unsustainable. Little or no progresswas made tiesfor watershedmanagement in Colombia be on establishing the technical and financial judged on the basis of this project'scexpeience. feasibilityof improved conservation practices. The basic rationale for this project is that A return to the old pre-project ways of fire, unless the pervasive and progressive degra- overgrazing,and cultivationof steep slopes has dation of the country's upland watersheds Is become visible in the pilot areas. reversed, the socioeconomic development  - viii - potential downstream will be impacted nega- Colombia. In this regard, the followingpoints tively. This insight remains valid and vivid in have been raised in the full audit report: Colombia today. 21. An important concept is that of land- 16. Throughout the project record, this use capability. While it is often possible to audit could point in numerous instances to "push the envelop" of current practices to issues or failingson the part of the personnel meet land-use capabilitylimitationsby increas- and institutions involved, including both the ing conservation measures, these efforts typi- Bank and the agencies in Colombia. A num- cally raise the cost of production substantially. ber of larger concerns, however, are more sig- Additional costs to achieveconservation objec- nificant and may offer guideposts for similar tives must be either paid for through the endeavors, both for the GOC and the Bank. marketplaceor offset by incentivesor subsidies These concerns are as follows: that allow the farmer to continue to sustain a livelihood. 17. Project design and preparation short- comings. Even a precursory reading of the 22. In Colombia, reforestation has tradi- project file would reveal that the design and tionally been a primary component of preparation was inadequate. The early and programs for watershed management action, pivotal decision to forego the preparation of a whether it was reforestation for production or Staff Appraisal Report led to an uncritical and for protection purposes. There is reason to often ad hoc reviewprocess within the Bank-- believe that for the moment neither is econo- to the point that the first supervision mission, micallysustainable. Reforestation for protec- shortly after signature of the agreement, called tion purposes, i.e., revegetating degraded areas for a reorientation and final preparation of the with traditional plantation techniques is rarely project. necessary, technicallyrecommended, nor eco- nomicallyefficient. In many cases of degraded 18. One of the most notable oversightswas areas, one can reestablish the vegetative cover the failure to analyzethe cost/benefit equation and with it the watershed function, just by of either of the project itself or of the farm protecting the area from fire and grazing. models being proposed as-the basis of the cre- dit program. This lapse was justified on the 23. In Colombia, reforestation for produc- basis of the pilot nature of the project. This is tion purposes is facing an interesting dilemma. difficult to understand considering that both At present, it is very difficultto earn a positive the borrower and the target farmer were rate of return on reforestation investments, expected to take loans at commercialrates of because with the exception of pulpwood, the interest. market is not very attractive. 19. Pilot projects are normallyput in place 24. A traditional bias toward reforestation to field test the findings and achievements of and tree-planting for watershed management preliminary R&D. Their objective is to test purposes has overshadowed the potential of replicabilityand contribute to fine-tuning the interventions in the agriculture and livestock approach so as to maximizeeffectiveness(im- sectors. Unless there is a direct and marked pact) and efficiency (cost-effectiveness). The change in strategy and the country begins to essential and basic components of a pilot deal effectively with the problems brought watershed management project design and how about by agriculture and livestockon the steep the PROCAM design measured up are review- hillsides, degradation will continue. Conse- ed in this audit document. quently, the costs to the country will far surpass the potential losses from a weak 20. The approachto watershed management forestry production sector. in Colombia. On more than one occasion, supervision reports mentioned the lack of an 25. The most important lesson relates to overall strategy for watershed management in the need to make choices and to do so in a  - ix - logicallyand sequentiallysound manner. One Bank's PCR nor in the Final Report prepared of the greatest weaknesses of integrated rural in-country is there any mention of the development projects is that they attempt to economicfeasibilityof watershed management do everything--spreadingcapability, expertise, or of the models put forward during the and resources too thinly and thus failing to project, in either quantitative or qualitative generate the momentum needed for real im- terms. pact. PROCAM has suffered to some degree from this type of problem. 28. Finally, the following lessons of the PROCAM experience must be factored into 26. For this project, institutional cons- future watershed development programs: traints were a principal concern of the Bank at the outset, and, as it turned out, were key elements leadingto inefficacy. Buildingon the * The fundamental principle of a genuine PROCAM experience, institution building participatory approach is that farmers needs to be elevatedto a criticalprerequisiteof must perceive and receive tangible bene- watersheddevelopment. The PROCAMproject fits in return for their efforts to conserve suffered from shallow institutional assessment soil and water. and analysis. Institution-building requires goals and benchmarks--that is, a program to * In areas where resource limitations are follow, specifyinganticipated achievements. severe, incentives or subsidies may be justified to offset the production trade- 27. Perhaps the most serious impactof the offs associated with protection and Bank not having carried out the appraisal of conservation. this project is the continuingfailure to come to gripswith the question of the cost of watershed * The dilemmas involved in administering management. There has been little achieve- such subsidies should be resolved through ment in this realm of inquiry; neither in the pilot schemes.  PERFORMANCEAUDfT REPORT COLOMBIA UPPER MAGDALENAPILOT WATERSHEDMANAGEMENTPROJECT (LOAN 2069-CO) L PROJECT BACKGROUNDAND DESIGN Background 1.1 The Upper Magdalena Pilot Watershed Management Project was identified and designed in a period during which recent development achievements were intrinsic to Colombia's socioeconomicstatus. Steady gains in GDP and per capita income had been recorded over the past few decades and with them significantstructural changes to the economy. Rural migration had led to increasing urbanization throughout the country, manifest in the emergence of urban areas and the rapid population growth in the capital. The establishment of an industrial and service sector centered in these emerging areas transformed the previouslyagrarian economy. The agriculture sector reflected an increase in irrigation capability,in mechanization and in the diversificationof the export crop subsector away from its past dependence on coffee. The population growth rate had also declined significantlyfrom an estimated high of over 3.5 percent to a more modest and manageable rate of 2.1 percent. In general, important improvements were achieved in the overall prosperity and well-being of the population. 1.2 These achievements, however, also created new challenges and opportunities for the GOC and its people. A prime examplewas the increasing need for electrical energy to sustain industrial production and growth, as well as to meet the domestic needs of the urban sector. Directly linked to electricity,because of the vast potential for the development of hydropower in the country, was the concem about water supply and demand. Water supply infra- structure was also among the earliest needs of the urbanized sectors of the nation, both for domestic consumption and industrial use. Thus, a good deal of investment was dedicated to this purpose. Plans for increasing the relatively modest capability for irrigation were, by definition, also directly linked to the availabilityof water resources. 1.3 Accompanyingthis modemization was a concem for the emerging gap between the rural and urban economies. Overall national economic advances fueled rising expectations among the rural poor who had been overlooked, given the emphasis placed on the urban and industrial sector. Migration to the cities had created increasing rural labor shortages, and as a result, there was more recourse to more extensive uses of the land, particularly in the livestock subsector. In the absence of sufficient govemmental services,guidance, and development resources, an increasing number of rural people sought sustenance and a livelihood by clearing new lands, many of which were of marginal quality, for peasant agriculture or for traditional extensive livestock husbandry practices.  -2- 1.4 In the late 1970s,cognizant of the disparities in socioeconomic development, the GOC launched the Plan de Integracion Nacional (PIN). This plan (a) carried on many of the macroeconomicstrategies (aimed at industrialization, export markets, and overall growth) of the past, and moreover (b) began a concerted attempt to raise production efficiencies and to strengthen the institutional capacity of governmental services responsible for promoting and monitoring sectoral development. The PIN also embodied efforts to meet social policy objectives with special emphasis on the rural poor. This was to be achieved by creating greater regional autonomy to promote more decentralized economic and institutional development. Governmental agencies were specificallymandated to focus attention, services, and resources on helping the poor farmers participate in the overall goals of the nation. 1.5 In parallel with developmentsworldwide, in Colombia, this period was also charac- terized by an awakening of national concern about the ecology and the environmental stabilityon which the destiny of many development programs so firmlyrested. This concern was obviously due to the widening occurrence of erosion and watershed degradation in the upland areas of the country. Downstream flooding, siltation, and irregular water flowsin the major river systems of central Colombia provoked high-levelpreoccupation for the future of hydropower as fuel for the engines of growth. Assessments of the forest resource base undertaken in the 1970salarmed many because of the estimated loss of more than 10 millionhectares, or 20 percent, of the forest cover known to exist in 1960. Identification and Design 1.6 It is for these reasons that the Bank reacted positivelyto the notion of a water- shed management project, followingthe request made by Instituto Nacional de Recursos Naturales Renovables y del Medio Ambiente (INDERENA) in 1979for assistance to the forestry and natural resources sector. At the time, the Bank was already committed to a program for Colombia that emphasized power generation and irrigation and had planned to continue contributing to the development of hydropower to increase domestic sources of energy. 1.7 These discussionsfocused on the Rio Magdalena, Colombia's most important river. The GOC had long been concerned about the condition of the natural resources in this key watershed; it was precisely the experience from activities in this watershed that convinced the GOC to create INDERENA in 1968. The Magdalena watershed covers an area of approximately 5.5 millionhectares, much of which lies in the hilly and mountainous terrain of the Central Cordillera and the Eastern ranges. Over twenty major tributaries of the Upper Magdalena rise in these intermountain valleys. Like much of the Andean region, these areas are characterized by steep, unstable slopes (typicallyover 60 percent) and by narrow valleyswith deep canyons. Although rainfall varies between 1,200 and 4,000 millimeters(the latter occurring in the higher reaches of the Central Cordillera, which also feed the watershed with runoff from mountain glaciers), the more important factor related to precipitation is the propensity for occasional torrential rains. These events occur annually, overwhelmingthe infiltration capability of the soils and sending cascadesof floods down upon the towns, many of which are located near the mouth of these narrow mountain valleys,sometimes with disastrous results.  -3- 1.8 During the protracted period of project preparation, in spite of high-levelmanage- ment changes, INDERENA continued to assert the importance and priority it assigned to watershed management in its overall scope. This was undoubtedly due to an emerging perception of the pervasive decline of manywatersheds in the country manifest by the recurrent problems of flooding and interrupted water supply to nearby towns. It was very clear both to INDERENA and the Bank, however, that a large-scale investment, such as the watershed management project, which would encompass the entire Upper Magdalena watershed, was impossibleat that time for limitations in experience, technical capability,and financial resources. Accordingly,a decision was taken to restrict the project under discussionto pilot activities in selected subwatersheds. 1.9 As the population increased and land clearing both for crops and livestock spread throughout the subwatershedsof the Upper Magdalena, the departmental capitals of Ibague in Tolima and Neiva in Huila experienced both flooding and water supply problems, resulting from the deterioration of the subwatersheds on which they were dependent. Both these subwatersheds, the Rio La Ceibas and, in particular, the Rio Combiema,are among the steepest, most populated intermountain valleys in the world, surpassing the conditions often seen in the Himalayas because of their lack of arable land along the valley bottoms. Cultivationsoften extend directly up slopes exceeding 60 to 70 percent with extensive livestock grazing on even steeper slopes. Due to the construction of the large (500 megawats)Betania hydroelectric plant, it was also decided that one of the important affluents of the Upper Magdalena River near to the dam would also be chosen as a pilot watershed. The choice of these three subwatershedsj/ was also of interest because of inherent differences both in their biophysical and socioeconomiccharacteristics, which were expected to generate additional experience and models useful for planning watershed manage- ment activities on a larger scale. Project Objectivesand Components 1.10 The main objectives of the project as described in early Bank briefs referred to the need to develop the institutional, managerial, budgetary, and technical groundwork and to launch research programs for improvingthe basis for future watershed management in the Magdalena and elsewhere in Colombia. It is worth noting, however, that these objectives,which alluded to the need to build institutional and technical capability,were amended in the President's Report, givingthe project a more explicit field and operational orientation. An even more succinct version of project objectives, again of a decisivelyoperational nature, is contained in the loan agreement (Schedule 2--Descriptionof the Project). 1.11 The project objectives identified in the President's Report (which served in lieu of a Staff Appraisal Report) were as follows: . . . to develop, within four years, viable farming and tree crop systems as well as institutional mechanisms for carryingout soil conservation and erosion control and / The term subwatemhedis somewhat misleadingbecause each of these areas is in fact a complete watershed,whichcontributes to the larger watershed of the Upper Magdalena River. A watershed may be defined as the area between the upper limits of the divide from whence water begins to accumulateto the point at which its rivermeets the next river.  -4- maximizingthe long-term economic use of the soils and water. The project would test production and conservation measures (reforestation, civil works, modifying land use practices, forest protection measures) to determine feasible and low-cost ways of introducing soil conservation to farming systems and to the watersheds. The project would also provide the necessary information to prepare a follow-up, larger project covering the entire Upper Magdalena River Basin. An integral part of the project would be the determination of incentives required to motivate land users to adopt conservation practices and an appropriate role for the public sector to play in protecting the country's water resources. 1.12 To reach these objectives,the project was expected to undertake investments, research, and the preparation of a follow-up investment project, organized along the lines of the followingactivities: * On-farm investments, essentiallyby providing credit to participating farmers to carry out soil conservation, reforestation, tree crop establishment, and improved livestock and annual crop practices. * Public works carried out through subcontractual arrangements with both private contractors and regional corporations that include reforestation on vulnerable areas of publiclyowned land and civilworks for river control. * Extension and training primarilythe provision of extension services and training for participating farmers to enable them to understand and adopt new technologies aimed at the soil conservation and watershed management objectives of the project. * Staff training for personnel of INDERENA and other participating institutions to better prepare them to execute future watershed management programs. * Researchand studies of particular relevance to both the project and future programs. These would include studies of hydrometeorology,sedimentation, and erosion of microwatersheds,of forestry and agroforestry practices, on the policy and economic framework for watershed management, and on general technological development. * Environmentalprotection by establishing two field units to monitor and protect forest resources, to manage two national parks within the project areas, and to aid in the enforcement of existing environmental legislation. * Preparationof Phase 11 mandating pertinent studies required to prepare the follow-up investment project that encompasses the entire watershed area of the Upper Magdalena River. Here again, a succinct version of these activities appears in the loan agreement. While these two sets of descriptions of project activities are basicallycompatible, neither contains an explanation of how they were quantified or costed.  -5 - Institutional Framework 1.13 For the intents and purposes of this Loan, INDERENA, an agency within the Ministry of Agriculture, was designated as the 'borrower", and the lead institution. The role assigned to INDERENA, in its own description of the project during the design phase, was to include administration, planning, and coordination of the other organizations involved,monitoring and evaluation of their activities, and the execution of certain action subprograms within their ambit of expertise. The other organizations,whose participation was foreseen during the design phase, and their respective areas of responsibilitieswere to include: * Instituto ColombianoAgropecuanio(ICA) to provide technical assistance to farmers in the development and implementation of farm plans. ICA was also expected to take charge of promoting fruit trees, cacao, and improved pasture and livestock practices. * Servicio Nacional de Aprendizaje (SENA) to introduce the project to the farmer, watershed users, and communities;to take charge of the extension and training related to agriculture; and to coordinate with INDERENA the environmental education of the rural communitiesand project staff. * Instituto Colombianode Hidrologia, Meteorologiay Adecuacdonde TYerros(HIMAI) to carry out the subprogram of hydrometeorologicalresearch and to assist INDERENA in the microwatershedresearch. - FederacionNacional de Cafeterosde Colombia under its program of development and diversificationin coffee zones, to execute the subprogram to renew coffee plantations and to promote fruit trees. * Caja de CreditoAgrario, Industrial y Minero (Caja Agraria), the provision of timely and sufficient credit for the project, as programmed by the other technical assistance organizations. 1.14 Inter-institutional coordination was to be achieved by the establishment and opera- tion of both national- and departmental-level committees. In addition to these principal organiza- tions, the other involvementwas also foreseen, including departmental authorities (both as interested participants and as contributors to the Soil Conservation Fund); the departmental forestry corporations (to carry out reforestation on public lands); and the Instituto Colombiano de Energia Electrica (the parent company of the Betania Hydropower Dam and contributor to the Soil Conservation Fund). Project Costs, Budgets, and Conditions 1.15 The total cost of the project was estimated at US$27.3 million of which a US$17.4 millionequivalent was to be lent in local currency and US$9.9 million in foreign exchange (see Annex 1). The Bank agreed to lend to the "Borrower"(in this case designated as INDERENA) US$9 million, or roughly 33 percent, of total appraised project costs. The loan was to have a  -6 - term of seventeen years, including a four-year grace period, with interest at 11.6 percent per annum. 1.16 The submissionof draft terms of reference, which intended to further elaborate several of the component activities early on in the course of project implementation, included a series of requirements related to the training programs on microwatershedand forestry research and the study of the policy and economic framework. Explicit mention is also made of the need to further improve and document in both technical and financial terms the farm models on which both field investments and the credit program were to be based. In the loan agreement, an additional activitymandating a study on INDERENA's administrative,financial, and information management systems,apparently proposed by INDERENA itself during negotiations in response to Bank concerns about its capabilities in that area, was also included. Several of these requirements were presented as conditions of loan effectiveness, and specific dates for their presentation were scheduled. II. IMPLEMENTATIONEXPERIENCE Postnegotiation and Mobilization Period 2.1 It was only shortly after the completion of negotiations and the signature of the loan agreement on December 15, 1981 that those involved in the project, both at the Bank and in Colombia, began to understand that it was going to be a more complex and difficult to implement project than was at first conceived. The BTO of the postnegotiation mission, dated December 28, 1981,identified a series of concerns which, in view of the past performance audit mission, are recognized to have been of fundamental importance. 2.2 Of greatest concern, as emphasized during preparation but apparently unresolved, was the instiutional setting, particularly the inter-institutional agreements. The postnegotiation Mission (November 30-December 11, 1981)detected the fact that "except for HIMAT staff, for staff of all other institutions it was the first time that they had heard of more than the title of the proposed budget (sic)." This is all the more noteworthy given the fact that the Bank preparation mission one year earlier had identified the exact same situation, to wit: "Regarding the insti- tutional set-up, INDERENA envisaged assuming the planning, coordination and evaluation functions and leaving implementation to specialized agencies, such as HIMAT, ICA, SENA, forest corporations, etc. However, no contacts had been made by INDERENA with these institutions in order to prepare the project adequately and get commitments from the implementing institutions." 2.3 An inadequate understanding of and attention to the institutional framework for the project was to continue to be the Achilles tendon of the project during its entire implemen- tation period. The so-called contractual nature of the agreements between INDERENA, as the borrower, and the other implementation institutions must be contrasted with the innovative and  -7 - complex V nature attributed to the pilot project. During the performance audit mission, several interviewees,both from INDERENA and at least two of the participating organizations, indicated that this contractual relationship stifled the flexibilitynecessary to find the right combination of technical interventions and operational responsibilitiesfor successful project implementation. In addition, this relationship appears to have added a high degree of administrativecost, particularly in terms of staff time for reaching agreements. It was thus only in January 1983, after several postponements of the deadline for reaching the conditions of effectiveness of the loan, that these agreements were finallycompleted with the other participating institutions and the loan was declared effective. 2.4 Another of the difficultiesoutlined by the postnegotiation mission was the lack of a comprehensiveconceptualframework for the elaboration and implementation of the various component subprograms (seventeen) prepared by INDERENA. The aide-memoire prepared by the Bank team at the conclusion of its visit indicated that some of the subprograms appeared to have been elaborated independently without taking into account the actual objectives of the project or the other components foreseen for implementation. According to the team, sequencing of subprograms, the choice of areas to be treated, and the justification of quantitative targets all needed further elaboration. Concern was expressed about the ambitious nature of some of the data collection undertakings (study at the microwatershedlevel) and the parallel lack of attention to the collection of socioeconomic data and information,which would serve eventually to help justify the larger, second phase. INDERENA's attention was also drawn to the need to include activitiesaimed at improvingthe sustainabilityof annual crop production practices given their importance in farmer income generation and in the incidence of erosion. The team acknowledged that while this was a pilot project and that much of its effort would be focused on resolving methods and models, nevertheless they would suggest a systematicrevision to facilitate the final preparation of the project. 2.5 These postnegotiation missioncomments are very similar to those contained in an interoffice memo of July 28, 1981 from the Agriculture and Rural Development Department to the Central Project staff, which do not seem to have been taken into account during negotiations. While clearly the postnegotiation mission,which included a Bank staff economist, a forester, an agricultural advisor, and a consultant livestock specialist, raised these issues, their resolution was certainly beyond the scope and breadth of their assignment. The developments described above are not atypical of watershed management projects (and other projects similarlybased on an integrated approach to rural development). It is much easier to detail what could be done in a watershed undergoing degradation than to select what it is advisable to do based on priorities and the resources available for project implementation. Given that the project was also conceived as a pilot project, the choice of activities, knowing that not all the problems could be dealt with, must revolve around the followingkey elements that are vital to ensuring the step to the next phase of investment, i.e., impact, replicability,and cost-effectiveness. _W The repeated statements about the innovativenessand compledty of the pilot project are surprising in that there was a good deal of aperience withwatershed managementin the country. The Corporacdon del Valle de Cauca (CVC) had been carrying out waterhed managementactivities for almost 15 yeam at the time of the designof the Alto Magdalena Project and the operational model proposed by INDERENA wasi very similar to the model developed succesfully and being implemented by the CVC.  -8- 2.6 Faced with these circumstances, the Bank team recommended a proposal, which INDERENA endorsed, that short-termconsultants be engaged,as soon as possible, "for preparing the project for effectiveness." It will certainly be noted by the careful reader that the steps being recommended, both in substance and degree, would appear to go well beyond what had been agreed between the Bank and the Borrower as the conditions for effectiveness (see Annex 2 for a list of the conditions of effectivenessas contained in the loan agreement). Rather the project, as implied in the comments of the aide-memoire and in the BTO of the postnegotiation mission,was still in the stage of final preparation. This recommendation to hire short-term consultants, although quite pertinent, proved more difficult to implement than might otherwise be expected. Terms of reference were prepared for four short-term consultants, including a principal advisor in watershed planning and management (three person-months), and consultants in agricultural economics (three person-months), in agro-silvo-pastoralsystems (three person-months), and in microwatershedresearch (three person-months). In addition, the letter to the director general of INDERENA resulting from this mission stressed the importance of hiring the long-term consultants foreseen in the project design and of the timely completion of the conditions of effectivenessspecified in the loan agreement. Mobilization to Effectiveness Phase 2.7 During all of 1982,efforts continued both by the Bank and INDERENA to meet the conditions of effectiveness and to get the project fully operational. In both March and September, Bank staff visited Colombia to assist INDERENA to move forward with the project. Aide-memoiresdetailing the steps to be taken were prepared and cleared with the INDERENA project director. In September 1982, the director general of INDERENA was replaced as part of a change in government. The new director, however, was quick to write to the Bank (letter dated September 8, 1982) stating his support for the project and confirming the ongoing activities to meet the loan requirements. This period also coincidedwith austerity budget measures imposed by the new central govemment that further impeded the contracting of both INDERENA personnel and of the consultants (both short-term and long-term), as agreed to during the postnegotiation mission. 2.8 During this period, it was necessary to extend three times the date (originally July 15, 1982) proposed as the limit for the fulfillment of the conditions for effectiveness. The records reflect some significantconfusion over the contracting of the short-term consultants. The Bank mission in November-December 1981 had recommended these consultants to assist INDERENA, and, particularly,the Proyecto Cuenca del Alto Magdalena (PROCAM) Project Management Unit that had been established, with completing the final preparation and meeting the conditions of effectiveness. INDERENA appears to have been under the impression that these foreign short-term consultants could only be hired once the loan had been declared effective, presumably because foreign exchange costs incurred before that date would not be reimbursable. 2.9 During this period, INDERENA, in cooperation with the national electricity entity Interconeccion Electrica, sponsored the First National Congress on Watersheds (held in Medellin, Colombia, March 2-6, 1982). The BTO of the Bank staff member, who presented a paper on behalf of the Bank, demonstrates the duality of opinions regarding the capability for watershed  -9 - management in Colombia that underpinned both Bank supervision and support for PROCAM. It reported that although Colombia was seen as a leader in the field, the papers presented at the congress indicated that the efforts were mostly dedicated to the collection of vast quantities of data on the conditions of the various watersheds without the overall framework or understanding required for resource management decisions and interventions. 2.10 As a result of an intensive supervision mission in September 1982 that produced a detailed aide-memoire and the resulting follow-upby the local project director, the project was finallydeclared to be effective on January 13, 1983. The Bank declared the loan effectiveness, waivingthe right to request a report containing the farm models, on which presumably both the macro- and microeconomicfeasibilityof the project depended, that still needed further refinement. There was also a number of other ongoing constraints, understood by both the Bank and INDERENA, including staffing shortfalls for the project unit and the need to accelerate or postpone compliance with the preparation of a number of key planning activities related to long- term implementation. Execution Phase 2.11 Having reached the effectiveness plateau, the project's next six months were very promising. Considered by the Bank and others as the key feature of the project's design, the Soil Conservation Fund was put in place. This fund whereby downstream users of the water flowing from the watershed, in this case the hydropower company, the municipal water supply systemsand a small irrigated perimeter, agreed to provide financial resources to offset the cost of soil and water conservation in the upper reaches. A series of small properties were acquired to serve as experiment and demonstration areas in the target watersheds. The short-term consultants were brought on board and undertook their work in this period. Work on the diagnostic studies of the larger Upper Magdalena watershed was well underway by the subcontracted institution--the Instituto Geografico Augustin Codazzi. The staffingsituation improved overall, and a proposal to provide salary supplements to attract and keep highlyqualified staff for the project unit was put forward. 2.12 Soon after the third Bank supervisorymission (early October 1983), another change in the leadership at INDERENA took place--the fourth such event in the short life of the project. There was also a decision by the new director general to replace the PROCAM project director, who had served in that capacity practicallyfrom the project's inception. 2.13 Soon thereafter the project entered into a period of considerable uncertainty in implementation, from which it appeared never to recover. Despite some promising achievements, the inadequate conceptual framework,which undermined the sense of mission and purpose for the staff and the institutions, combined with other design misconceptions,began to make themselves felt. Some of the most pertinent considerations 3/ of this period were: _N Many of the facts and details of the implementationrecord are discussedin both the PCR and the FinalReport of the project prepared by INDERENA, albeit without the degree of analysisthis report is attempting to apply.  - 10- * Project rorientation and realignment. An overall reorientation and realignment of the project to take account of the delays and findings during its inception, expected to be guided by the short-term consultant in watershed planning, was not achieved to the degree necessary or anticipated. * Monitoringand evaluationsystem. The design and implementation of a monitoring and evaluation system was never realized, although it was also expected of the short- term consultant in watershed planning and subsequently of another longer-term consultant hired late in the project. These systemswere considered essential for two purposes: (a) to guide the pilot activitythrough analysisof problem areas and justification of corrective changes, and (b) to provide part of the baseline for the subsequent design of the larger investment project encompassing the entire Upper Magdalena Watershed. * Creditsystemfor on-farm interventions. Perhaps the most salient example of both the institutional and design issues hampering the project was the almost total failure of the credit component. During 1983 and part of 1984, of the almost three hundred beneficiaries preselected for credit, only one loan was approved. A considerable effort to examine and reconcile this issue with the Caja Agraria was made and some slight improvement achieved. As it turned out, however, the guarantees and requirements to obtain a loan under this credit line proved beyond the capacity of the majority of the smaller farmers residing along in the pilot watersheds that the project was trying to reach. The Caja Agraria stood its ground, some would say adamantly,on the contention that it had agreed to provide credit only within the existing systems and norms currently practiced in the country. In the final analysis,only 270 loans (only 5 percent of the estimated 5,000 originallyplanned) were made, amounting to roughly US$400,000(again about 5 percent of the projected US$7.1 million planned for the credit component) over the life of the project. v Operational capabilityof the project unit. The project unit that was created as a separate entity within INDERENA continued to have difficulties responding to the demanding role of overall manager and coordinator of PROCAM. This was due to continuing staff shortages; disapproval by central government of the salary supple- ment; continually growingadministrativeand financial accountability responsibilities, resulting from the proliferation of subsidiary agreements and subcontracts with other government and private organizations for different facets of project activity; further changes in leadership (both the director general and the project director changed again between 1984and the end of the project in 1987), and limitations to its financial resources as a result of successivegovernment austerity budgets. * Isolaion and resentment of the promject unit. Added to its own inherent difficultieswas the emerging resentment felt among the other sections of INDERENA, which was perhaps exacerbated by the period of high-levelindifference and even outright opposition to the goals of the project. This situation was also part of the reason that the project unit experienced difficultiesin hiring consultants for technical assistance and in coalescingits training program,which were subject to the reluctant clearance and approval by other units within the organization. A certain plausibilityexists to the  - 11 - claim made by project participants that the planned study on the administrative and financial management systems never took place because of direct obstruction by other units within INDERENA At the height of the most problematic period in 1985, and despite Bank expressionsof concern, by executive fiat, both financial and human resources destined for the project were deviated for other purposes within INDERENA. * Contractualnature of the relationshipsamong the participatinginstitutions. In general, the principal implementation organizations chosen to participate in the Project (ICA, SENA, HIMAT) were selected by virtue of their mandates and functions established by law. The contracts, though complyingwith the conditions set in the loan agreement, proved to be ineffective during implementation (e.g., unclear interagency arrangements, including budget allocations). The overall orientation of their activities was that of completing tasks ("el cumplimientode metas fisicas")prescribed in their contracts rather than participatingwith INDERENA in the joint conceptualization and testing of an inter-institutional and appropriate model(s) for an integrated approach to watershed management. This situation was even more accentuated in the numerous cases of smaller direct subcontracts awarded to both private and public entities for specific studies and tasks. For example, INDERENA commissioneda series of engineering studies for civil works, as foreseen under the project, without, in retrospect, the appropriate in-house capabilityto judge their adequacy. Few of the planned civil works were actuallybuilt, some of which were built washed out. The director general later decided that this kind of work was beyond the mandate of INDERENA and, thus, cancelled further work in the subprogram. * Technical difficulties. A number of technical difficultiesand the phasing of activities to overcome them made it almost impossiblefor the project to live up to its implementation schedule. Trials established to test new production cum conservation technologies were found to be statisticallyinvalid and had to be redesigned and re- established. The ICA cooperated by providing specialized research design skills, outside the terms of its contractual relationship, with the net effect that the research program is still ongoing. This left the project without the scientific basis to substantiate many of the pilot interventions that were being tried on-farm in the watersheds. * Disbursementshortfalls. Both implementation difficultiesand constraints on financial resources as a result of successiveausterity budgets combined to severely limit the utilization of the loan funds. While after the first spate of mobilization expenses cumulative disbursementsin early 1984 reached approximately US$1 million, outlays were considerably off the pace intended thereafter--to the point where by mid-1986 the total had only reached US$2 million. It must be borne in mind, however, that part of this slow disbursement was due to the continuing and rapid devaluation of the Colombian peso over the same period. * Bank/lNDERENA relationship. As part of the implementation record, a number of facets of the relationship between the Bank and INDERENA bear scrutiny.  - 12 - The review of the project's correspondence files shows the preponderance of exchanges (with the exception of selected aide-memoiresleft as a result of supervision missions)mainly of administrativenature, i.e., exchanges regarding of loan requirements completion of terms. Even more notable is the numerous instances in which the Bank was reviewingthe TORs, draft agreements, draft subcontracts, or some other implementation planning tool. Several things seem apparent from this situation: the simple exchange of correspondence, views, and approvals related to the multiplicityof these arrangements of a subcontractual nature (for participating institu- tions, subcontractors, consultants, and project procurement) was an overwhelming administrativeburden, both for INDERENA and the project unit, as well as for the Bank. During the audit mission, INDERENA personnel mentioned that the administrativeminutiae of the project obliterated time (and energy) for anything else. Thus, it seems that the file contains very little professional or technical exchanges that might have influenced some of the more basic terms of project implementation. In retrospect, the overall design of the project seems to have imposed a burden on INDERENA, in an area in which the Bank judged it to be the weakest--administra- tion and management. It is also clear from the record that Bank supervisioncannot and should not be expectedto substitutefor needed technical assistance. Another point is that there were several instances and many valid reasons throughout the course of implementation that would have allowed the Bank to exercise its rights (contained in the loan agreement--Article VI--Remedies of the Bank) to suspend the loan and move for reappraisal. This did not take place. What is perhaps more surprising is that in reviewingthe entire set of project files, no mention of such a possibilitywas ever raised. This may be due in part to the fact that, after a certain point, supervision missionswere undertaken by a changing cast of Bank staff members. This may have required a high learning curve on their part; one that did not allow staff adequate time to reach the ultimate conclusion that the project was so severely flawed that it would never be successful. Then again, it may have been a factor of the high administrativeoverhead associated with backstopping this project, which made it difficult to see or reflect on the larger picture. III. PROJECT IMPACT 3.1 Undoubtedly, PROCAM achieved some impact, albeit far less and in waysnot entirely or originallyforeseen. Overall, the project experience,however,has been unsatisfactory. Indeed, only the absolutist language used in the Bank's project information form for defining very unsatisfactoryprevents the audit from evaluating it as such. This rather negative assessment of the Upper Magdalena Pilot Watershed Management Project is based on two basic considerations: (a) the project failed, almost completely, to meet the objectivesset for it, and in doing so, (b) was unable to demonstrate the replicabilityof its approach to watershed management. The section that followsis a discussionof this evaluation of performance and the criteria on which the judgements put forward immediatelyabove are based.  - 13 - of Project Objectives Achievement 3.2 Project documentation mentions three principal objectives: (a) development of viable farming and tree crop systems, (b) improvement of institutional capability, and (c) preparation of a follow-up investment project. Farm and Tree-CropSystems 3.3 In general, the physicalachievements in the implementation of farm- and field- level interventions, some of which in the agricultural area exceeded the set targets, are far less significantin terms of impact than might be hoped for. Although it was not possible in the time available for the performance audit to view more than a limited sample of the work sites, field observationsin the two (out of three) watershed areas visited suggested that few of the farm-level interventions have continued to be applied by the residents. This is hardly surprising in that the need to further develop and perfect these farm-levelinterventions was the subject of considerable Bank intemal discussionregarding the conditions for effectiveness of the Bank loan. 3.4 PROCAM set up a fair number of experimental runoff plots to test the effective- ness of suggested improved agricultural-cum-soilconservation technologies for the sloping lands typical of the watersheds. This research program on the PROCAM experiment stations in several locations throughout the pilot watersheds is still ongoing. The plots have as yet to provide the data and information necessaryto confirm either the anticipated improved yields or their effec- tiveness in terms of soil conservation. Even more regrettable is the fact that they have not been assessed from the point of view of cost/benefit analysisto ensure their viability as recommenda- tions for the farmer or as the basis for a credit program. 3.5 Trials of this nature, with several important variables, require fairly exacting statistical and economic analysisto assess and extrapolate the results for application under the full spectrum of diverse conditions in the watersheds. It remains to be seen if the INDERENA staff currently manning the stations will be able to perform these analyseswithout recourse to the technical assistance now unavailable to them because of the closure of the project. 3.6 Reforestation was also an important part of the farm and public works investment components of the project. Although the implementation targets were almost reached, the whole question of the role of reforestation must be reviewed in the context of watershed management in Colombia and, in particular, as applied in the Upper Magdalena watersheds. For one thing, reforestation has been successfullycarried out in these areas for some time. And indeed, almost all of the trees one sees in the area have been planted, e.g., the eucalyptus and Cupressus. This is especiallytrue in the middle elevations where for years, according to INDERENA, refores- tation campaigns have been going on for years for both public and private land. Regrettably, many of the areas successfullyplanted and growing have either not been managed (thinned) or should not be because of the steepness of the sites. Hence, there is reason to question the financial viabilityof tree planting under such conditions, whether for production or for protection (see next section for more detailed discussionof this topic).  - 14 - 3.7 Furthermore, evidence of fires, the renewed cutting of trees and clearing of vegetation on steep slopes for farming, and the continued overgrazing all indicate that many of the farmers have gone back to their older ways. Even the promisingpractice of establishing fodder banks (pastosde corte with King grass and other species) has not apparently been expanded despite its high potential still evident in the remanent areas. It would seem that the efforts by SENA to carry out a campaignto sensibilize the peasants to the dangers of natural resources and watershed degradation fell on deaf ears! 3.8 The reality, however, is otherwise. Too many projects of this nature, i.e., inherently top-down, start with the premise that they must sensibilizethe farmers. Farmers in such conditions observe the degradation from day-to-day and are very aware of the impact of their practices. The watersheds of the Upper Magdalena are no exception; floods, landslides, declining yields, and the challenge of survivalin these areas' are paramount considerations of every small farm family. In most cases, the lack of alternatives (that is, technologies applicable under the biophysicalconstraints of steep lands, farm inputs, markets or off-farm employment opportunities) compels farmers to behave in seeminglyirrational ways. Within the Andean watersheds of Colombia, many farm familieshave already taken the only real route open to them: immigration to the towns and cities or to the new lands of the lowland tropics of the country. Perhaps it is the project personnel who should be the first target of the sensibilization campaigns--shorthand for saying that, to be effective, watershed management projects must start at the grass roots level. 3.9 In short, the pilot farm- and field-levelinterventions, some of which undoubtedly hold promise for achievingboth production and protection targets in the context of watershed management, were not successfullydeveloped, promoted, nor implemented during the project. Moreover, the actual definition of this project as a pilot field exercise can be questioned. In addition to these needs to further develop the field-level technological packages, it would appear that the project design grosslyoversimplifiedand underestimated the complexities associated with watershed management in the conditions typicalof the Upper Magdalena. Had there been even the slightest possibilityof the overall feasibilityof the proposed packages of interventions, one would have expected to see them being adopted by progressive farmers in the area, even in the absence of incentivesor credit. And indeed, progressive farmers have moved off the slopes and are now raisingpoultry for sale to nearby towns. A more fulsome discussionof the whole concept of watershed management in Colombia is part of the followingsections of this report, which treat the findings and main issues. Institutional Improvement 3.10 In general, the demise of the Upper Magdalena Pilot Watershed Management Project is moot testimony to the fact that it failed to achieve its institutional strengthening objective. It should also be noted, however, that this objective was never particularly well described in project documentation nor was it well articulated with the operational aspects of the project. Too much emphasis was placed on improvingwhat was seen as INDERENA's adminis- trative and financial management weaknesses. In retrospect, it could be argued that these areas of capabilitywere actually put under more stress and hindered in their development by the decidedlycomplex nature of the institutional framework. The contractual posture of the relationship between the lead agency and its other institutional partners also impeded the learning  - 15 - and adaptation processes implicit in the pilot stage, especially as this concerns the integrated approach to watershed management. The reluctance to utilize the technical assistancefequired and the failure to truly launch the training program, both foreseen as essential parts of the project strategy, undermined institution-building. 3.11 Nevertheless, some residual impact has been discerned among the various institutions who played a principal role in the implementation of the Upper Magdalena Pilot Watershed Management Project. Both SENA and ICA, drawing on their experience in PROCAM, have established programs aimed at watershed issues. SENA has included watershed degradation as an important component of its National Prevention of Disasters Program. Concerns about natural resource management and the environment are core elements of their standard rural development promotion and training campaigns. They are also collaborating with INDERENA on a series of training bulletins related to watershed management (see below). ICA, for its part, has created a Division of Natural Resources in its new structure, and although fundamentally agriculture-production oriented, ICA has become ever more conscious of the watershed issue, particularly as it applies to irrigation, on which, in many cases, programs for agricultural commodityexport depend. 3.12 Despite what must have been an arduous and troubling experience with PROCAM, INDERENA has regrouped its Division of Watershed Management and, with the limited government funding at its disposal, is executing two programs as a direct sequel to the project. The Program for Research in Renewable Natural Resources for Watershed Management is undertaking the investigativework begun during PROCAM, maintaining the experiment stations and continuing the testing of a limited number of other promising technologies. The Development Program aims at applying the techniques to field conditions in support of other entities that have defined watershed management as a need. Through this latter program, INDERENA is presently contributing to the GOC's decentralization efforts by assisting municipal corporations to plan and implement watershed management in the microwatersheds that supply municipalwater systems, under the overall national campaign of "MunicipiosVerdes." INDERENA, in collaboration with SENA, has produced a series of training brochures 41 to assist municipal leaders, teachers, farmer leaders, and others to become familiar with watershed management issues and activities. 3.13 The GOC too has given priority to watershed management as part of its overall socioeconomicand rural development strategies. In the case of the Magdalena River, for example, specific mention is made in the new 1991 Constitution regarding the creation of the Corporacion Autonoma Regional del Rio Grande de la Magdalena, charging it with, among other things, the utilization and preservation of the environment and renewable natural resources. Similarly,in the Forestry Action Plan for Colombia, the first of the four major problem areas is described as follows: "The deforestation rate reflects the overexploitation of natural forests in Colombia, resulting in the erosion of 20 percent of the national territory and varyingdegrees of Among the titles in this series, which acknowledges the contribution of the Bank through PROCAM, are: 'Todos Necesitamos del Suelo"; "Especies Vegetales para Proteccion del Recurso Hidrico"; "Manejo de Cuencas Hidrograficas Abastecedoras de Acueductos Municipales"; "El Componente Amnbiental en el Plan de Desarrollo Municipal"; "Autogestion Comunitaria;" and, "Communicacion Participativa."  - 16 - imbalance brought to the different watersheds, the sedimentation of the main bodies of water, the fluctuations in flow volume, the reduction of the useful life of reservoirs, the decreased navigabilityof rivers, and the destruction of estuary zones." In response to watershed degradation, the plan proposes the establishment of the National Network for Technical Cooperation in the Management of Watersheds (RENORDE). This project involving the regional corporations, INDERENA, and other public entities under the initial direction of the Corporacion del Valle de Cauca (CVC) is expected to take approximatelysixty-four months and to cost almost US$5 million. RENORDE will provide only the basic technical assistance intended to leverage much larger amounts of funding to meet the investment cost of watershed management activities associated with municipal and local water supply systems it helps to get underway. Planning for Phase II 3.14 Given the many difficultiesencountered in the implementation of this pilot project, it is not surprising that efforts to meet this objective were, at best, half-hearted. The major study of the Upper Magdalena watersheds, undertaken by the Instituto Geografico Augustin Codazzi, while fairly comprehensive,proved to lack the level of analysisrequired for the design of a follow-upproject. As discussed at some length in the section that follows,designing integrated rural development efforts, such as watershed management, requires making prioritized choices of activities based on a sound understanding of the constraints and opportunities, both biophysicaland socioeconomic. PROCAM never reached this stage. 3.15 The other major shortcoming related to the preparation of the investment project, as pointed out on numerous occasionsby Bank supervisors,was the total failure to design and implement the monitoring and evaluation system for PROCAM. Without monitoring and evalua- tion, project personnel were unable to systematicallydocument progress and problems, to make adjustments to the implementation of the pilot stage, nor to lay the groundwork for a follow-up investment project. Unfortunately, although a concerted watershed management program is still much needed in the Upper Magdalena area, PROCAM will be much harder to analyze, and its experience will be necessarilymore anecdotal than scientific. Replicability--thePath to Sustainability 3.16 Successfulwatershed management projects must reach a high degree of auto- sustainabilitybased on the production and returns from the application of technologies appropri- ate to the inherent land-use capabilities. In Colombia and particularly in the steep, unstable soils of the watersheds of the Upper Magdalena, the recurrent costs of a watershed management model based or dependent on a social welfare approach (incentives to farmers to survive without cultivating steep areas) would be, at least for the near- to medium-term, prohibitivelyexpensive. Any analysisof PROCAM that attributes impact to the physical achievements is therefore simply inadequate. 3.17 One of the implicit aims of PROCAM as a pilot project was to test the replica- bility of and to improve its approach to watershed management--and thus to lay the foundation for future efforts. Unfortunately, the project's record is so replete with delays and problems that it was unable to achieve any real testing of the various components of its approach. On the  - 17 - technical side, as mentioned previously,the application of new technologies to the problems of sustainable production under the hillside conditions was at best incipient. While the experiments are still ongoing, on-farm experiments and introductions are no longer being monitored, thereby limiting the amount of field-informedfeedback, which is the normal test of replicability. On the institutional side, the best that can be said is that the PROCAM institutional framework-- involvingmany agencies all with operational responsibilitiesunder contract-like arrangements--is not the route to follow. The quantitative assessmentsof the project activitieswere so scarce that even overall economic feasibilitycould not be confirmed, nor could cost efficiencies be determined to ensure replicability. 3.18 The project has, however, served to confirm replicabilityin a few areas, and this should have some bearing on the long-term sustainabilityof watershed management in Colombia. The achievements are mainly related to the learning process among the participants. They are the principal actors on.the stage of watershed degradation and will have to continue to play that role to achieve watershed management. 3.19 The local understanding of the important relationships between the downstream users of the water resources and the upstream residents of the watershed is still apparent in the Upper Magdalena area of Colombia. During the field visit, municipal authorities voiced their strong convictionsregarding the need to support conservation and management in the upstream portions of the watersheds that supply their urban water systems. In many urban areas in the valley, the water supply is drawn from intermountain rivers and streams. These areas are presently experiencing the reality of watershed degradation. The high cost associated with the maintenance of upstream intakes and the occasionaldrought periods serve as testimony to these convictions. 3.20 Although it may still be difficult to tap financial resoui'cesdirectly from concerned municipalitiesto support watershed management in their supply areas, many are now including this issue in their local development plans and are seeking government resources to address it. The Soil Conservation Fund attempted under PROCAM was an excellent idea--that is, downstream beneficiaries help to finance upstream costs. Sustainingthe willingnessto contribute either directly through SCF-like arrangements or through the tax structure, creates a demand for tangible results. Fortunately, it would appear that the foibles of PROCAM, which never reached the stage of producing results that could be detected downstream, have not undermined the overall concept. Perhaps even more illustrative of the situation, given the new shift within government towards decentralization, local entities have been given the right to issue use permits, for such activities as tree-cutting, gravel extraction, water use, and effluent discharge, all directly affecting the watersheds. The vested interests of local entities are now in their own hands and with it the authority, responsibility,and accountabilityto their constituents for the stewardship of this fundamental resource base. 3.21 There is also an evident and growing awareness among farmers of the watersheds that their situation is a precarious-one. Several individualswho spoke up during the audit field visit to the project areas demonstrated a sound grasp of watershed degradation issues. They evinced a willingness,indeed an urgency,in finding other viable means of earning a livelihood. A  - 18 - number of the farmers in the Rio Combeimawatershed have turned to poultry raising on a semi- commercialscale rather than continue the numbing existence of exploiting the poor hillsides for subsistence. These more progressive farmers expressed dismaythat PROCAM had ended before it had reached its goals and pleaded for its re-establishment. When the unlikelihood of commercialfarmingwas apprised, several farmers vowed to continue their own efforts and to pressure local authorities to reinstate a similar program.:/ Still other farmers, however, look to the government to solve their problems and seem little aware that their destiny is inextricably linked to that of the watersheds. IV. FINDINGS AND MAIN ISSUES 4.1 The overall design, preparation, and implementation record of the Upper Magdalena Pilot Watershed Management Project is fraught with so many problems and delays that it is difficultto isolate particular problems for lessons learned. Several things are, nevertheless, very clear. Under no circumstancesshould the needs and opportunities for water- shed management in Colombia be judged on the basis of this experience. The basic rationale for this project remains valid and vivid today, i.e., that unless the pervasiveand progressivedegradation of the country'supland watershedareas is reversed,there will be significantnegativeimpact on the socioeconomicdevelopmentpotentials downstream. 4.2 The Project Completion Report prepared by the Bank, in the opinion of this performance audit, seems to be followingthe same path as the design and preparation period - oversimplifying the situation, making vague allusionsto institutional concerns, and providing an essentiallydescriptive rather than analyticaltreatment of the project.&/ While the persistent concern with the institutional dimensionsof this project, during preparation and implementation phases and in the final synthesis, is well-founded, it overlooks the fact that the Bank was experiencingsimilar problems with its other projects in Colombia during this period of changing governments and of austerity budgets.2/ 4.3 Numerous instances exist throughout the project record where, with the luxury of hindsight afforded this performance audit, one could point to issues or problems on the part of the personnel and institutions involved. Many of these can best be characterized as instances of human failingswhere either the institutions involved,both local and Bank, or the personnel who represented them, were unable to meet their responsibilities. Examples of individualsituations Watershed management plans have been prepared now for two watersheds included in PROCAM, both the Rio Combeima (whichsupplies the municipalwater systemof Ibague,the capital of the Department of Tolima) and the Rio La Ceibas (which supplies the municipalwater system of Neiva, the capital of the Department of Huila). It should be noted that accordingto the written record of the project files,the PCR appears to have been prepared by a Bank stafferwho wasnot previously involvedin the project and who wasallocated one dayas part of an unrelated TOR to meet with INDERENA and to review the status of that institution's final report, before initiating the preparation of the Bank's PCR. 2/ See the Bank's letter of September 11, 1986to Dr. Jose Fernando Botero, Minister of Agriculture, regarding the five Bank supported projects under execution at that time.  - 19 - where the staff involveddid not follow up on some of the important dimensions of the project preparation or implementation specificationsinclude the following: * The decision at the Bank not to carry out a full-scaleappraisal of the project and in particular the lack of a satisfactory examination of the cost/benefit implications of watershed management activities proposed as the basis for the loans, both to the borrower and to the farmers. * Inadequate preparation of the institutional project framework by INDERENA, especiallyas concerns the working relationships between the principal GOC agencies and lack of follow-up on this important consideration by Bank staff preparing the project. * Lack of clarity in budget estimates, reflected in the confusing budgets attached to the President's Report, and an overall lack of breakdown of the operational budget. * Lost opportunities at the Bank to improve the design and preparation of the project. For instance, apparently no action resulted from a July 1981 Agriculture and Rural Development Department memo raising several concerns about the project; the Bank accepted farm models despite their known shortcomings;little systematicfollow-up occurred on many serious issues raised by the postnegotiation supervision mission (November/December 1981) or on issues raised by the staff member in a BTO after visiting Colombia in March 1982. * INDERENA's reluctance and overall failure to hire the technical assistance consultants foreseen in the loan agreement. Those hired were not contracted in a timelyway and the overall level of effort foreseen for consultancies fell far short of the estimate made during project preparation. 4.4 Although the cumulative effect was quite serious in its impact on the project, particularly the missed opportunities during design and appraisal, an analysisat this level would do little to bolster future efforts. It would appear to this audit that once the project was approved, both INDERENA and Bank staff were unable to muster the time, energy, or will to undertake a systematicanalysisof the operational problems and to proceed with the redesign, which was undoubtedly necessary. Had the consultants foreseen in the agreement been hired as planned (and one can rationalize that both Bank and INDERENA project personnel felt that this technical assistance would solve the problems), many of these problems might have been overcome. 4.5 A number of more significantconcerns, however, may offer guideposts for similar endeavors both in Colombia and at the Bank. Four major issues,considered the most salient and far-reaching, are discussed here below in the hope of further contributing to the outcome of this pilot project, which ultimate goal was to advance the state-of-the-art of watershed management in Colombia.  - 20 - Project Design and Preparation Shortcomings 4.6 It is clear from a careful reading of project files that a significantdegree of ambivalenceand confusion existed right from the start regarding the original intent of the project. Although it was nominallycalled a pilot project from the beginning, frequent statements about the need for technical assistance suggest that its target should have been more basic. 4.7 The interest and enthusiasm of the Bank appears to have been centered on two innovativefeatures of the project. The most important of these was the precedent of a well- recognized and functional relationship between the upstream and downstream users, now legally established in Colombia, which requires, as an example, that 2 percent of the net proceeds of the operation of hydroelectric facilities must be reinvested in the upstream areas. The proposal contained in the project documents to establish the Soil Conservation Fund with voluntary contri- butions by downstream users (hydroelectric companies, municipal water corporations, and the irrigation associations)was indeed a significant achievement and thus worthy of support. The second characteristic was the integrative approach to watershed management, involvinga wide segment of govemment services to achieve a common goal of enhanced and sustainable produc- tion in critical watershed areas. These two characteristics were seen as exemplary features of the project and explain the high profile of the project, transcending both the Bank and Colombia. It is somewhat ironic that PROCAM was much discussed in development circles around the world even before it became operational. 4.8 If the record is a legitimate reflection of the thinking of the times, the GOC had perhaps more practical aspirations. In the words of the then director general of INDERENA (in an April 1980letter to the Bank personnel), their intention was that of learning by doing ("aprender haciendo"). INDERENA staff, which were involved in the early design activities and interviewedduring the audit exercise, asserted that it was their intention to investigate the various watershed management models and approaches existingin Colombia.J Their vision of the project, at least as expressed in recent meetings in Bogota, was a project geared to research, experimentation, and data collection on different approaches and technologies applicable to watershed management in the Upper Magdalena and its tributary watersheds. 4.9 These inherent differences of viewpoint, now difficult to substantiate, are certainly plausible,especially given the fact that the director general of INDERENA, who was the original proponent and promoter of PROCAM, was replaced during the preparatory period (October 1980). Whatever the circumstances,the fact remains that the project as described in the authorizingdocuments of the Bank was definitely considered to be a pilot project. y Here again, it is necessary to set the record straight. In 1979, there were, according to sources at the National Planning Department (DNP), at least ten ongoing watershed managementprojects in the country, not the least of whichwas the large undertakingof this nature beingcarried out by the Corporacion del Valle de Cauca (CVC). It would also appear in retrospect that this was not the firt time either that the Bank was involvedin watershed management,and depending on one's outlook about the DRI Project activities that addresed land rehabilitation,pouibly not even the first time the Bank was involvedin watershed managementin Colombia. This "whetheror not it was the firt" issue is relativelymeaninglessin practice although it does point to a certain superficialityof the preparation process, i.e., no knowledgeof past activitiesin the field in Colombia.  - 21 - Pilot projects are normallyput in place to field test thefindings and results of preliminary R&D efforts. Their objective is to test replicabilityand contribute to fine-tuning the approach, so as to maximizeeffectiveness (impact) and reach highest efficiency(cost- effectiveness). 4.10 The essential and minimumcomponents of a pilot watershed management project and how the PROCAM design measured up are: * Sound understanding of the problem situation. The documentation, which reflects the experience of the project, is clear in this regard. In absolutely none of the official project documentation is there any in-depth analysisof the situation of the watersheds in question, neither from the biophysicalnor socioeconomicstandpoints. This is in sharp contrast with the statements in the early project brief (May 1980) that justified the expanded President's Report in lieu of an appraisal: "The preparation report would include detailed technical knowledgeof climate, soils, crops and water flows as well as social and economic information about the regions." Although some of this information may have been available during the early preparation phases,2/ the basic watershed management plans for the Rio Combeima and, subsequently, for other watersheds in the pilot area, are only now being finalized. At the very least, and to evaluate the planned interventions from an appraisal point of view, one might have expected to find information on the following: - topography and altitudes - soils (althoughsome information on the soils is included in the President's Report, the data is insufficientfor subsequent analysis); - climate, includingrainfall amounts and periodicity,life zones or basic ecologicalinformation; - an assessment of the impacts of past practices and abuse and erosionrisk assessments; - socioeconomicdata includingpopulations, density, land ownershippatterns, occupations,labor availability,off-fann employment, and social, technical, creditand developmentservicesin the area; - an overviewof land use and existingfarm production systems; 2/ The President's Report mentions the eistence of "A project file .. prepared as a basic guide for the project." Neither this document nor any further reference to it in the preparation or implementationperiod was found in the record nor recalled by the personnel interviewedin Colombia.  - 22 - - an understandingof consumdtion and commercializationpatterns of farm products; - the extent and condition of infrastructurein the project area; an4 - activities in the project area. a synopsisof past and present devewlopment Without this basic information, it is hard to see how either the Bank or INDERENA, as the borrower, were able to judge the overall feasibility,from either the technical, social, or economic point of view of the project. Part of the problem lies in the general approach to watershed management in Colombia that is still rather directive and patemalistic, despite high-level government policy decisions regarding decentralization and the admitted importance of the communitydimensions. At present, many of the best-intentioned, ongoing efforts of INDERENA are stili aimed at the development of technologies,with little regard to a thorough analysisof the problem, particularly from the point of view of the principal actors--the peasant farmers. * Technological package to addressperceived constraints. The project was premised on the existence of suitable on-farm technologicalmodels applicable to the conditions of the watershed, which would enhance or at least maintain productivitywhile achieving soil and water conservation. The improvement of the existing models was part of the conditions for effectiveness. After substantial delay, the models were finally accepted by the Bank in 1983 with the full knowledge that they needed to be further elaborated and that the cost/benefit analysisneed to be carried out. Some of these models, aimed at the introduction and expansion of perennial crops, such as coffee, cacao, and fruit trees, are by definition more suitable to the conservation goals of the watershed management. their productivity,however, in practice on the soils of the area (with the possible exception of coffee), has as yet to be thoroughly analyzed. As concerns annual crop production--the main cause of erosion in the project area-- much remains to be done. Much of the groundwork for understanding the production, conservation, and financial dimensions of these models is still being studied. There is simplyno shortcut for documenting the soil and water retention efficiencyand production capabilitiesof new approaches to peasant agriculture. The extensive network of runoff plots and trials put in place on experimental areas is still being monitored, albeit with the limited local budget now available to INDERENA and its field personnel. These types of experiments are long-term in nature. Newer techniques of monitoring soil movement (with pins) and soil condition can provide indicators of improvingsoil conditions and stability. They are still, nevertheless, hostage to the longer-term farming systemsstudies required to assess the productivity of new technologies, particularlyfor the intercropped systems typicalof the project area.  - 23 - Owing to the institutional orientation and capabilitiesof the lead agency, INDERENA, the project has emphasized conservation, without adequate attention to the existing agricultural, livestock, and forestry production systemsexisting in the area. Sedimentation studies, river control structures, and water quality improvementswill do little to ameliorate the basic problem of watershed degradation. The key is to resolve the pervasive and degenerative spiral of improper land use, soil degradation and ero- sion, and decreased agricultural production leading to opening up new lands. Through the efforts of ICA under the Integrated Rural Development Program (DRI), some achievements have been made in assisting small farmers (less than 20 hectares) to more rationally and fullyexploit their holdings. Too little of ICA's resources (according to informed sources, less than 5 percent of its total budget) is devoted (a) to research topics relevant to the small, hillside farmers of the Andean zone, and (b) for effective technological interventions with annual crop schemes so keyed to watershed management. * Size and scale suitablefor pilot operadons. One cannot help but remark that for a pilot project, this was quite a large undertaking. With a total cost estimated at US$27.3 milliondollars, it was by far the largest project that INDERENA had ever undertaken. Perhaps more to the point is the contrast with the early project brief (May 1980) that described a project where "the total costs are expected to be in the US$8.14 million range. An estimated Bank loan would amount to US$3-6 million...' How this escalation in size took place is now unclear. As is stated elsewhere in this report, however, it is virtuallyimpossible to understand the reasoning behind the various budget synopses provided with the loan documents. The earliest indicationsof size of the project mentioned an area of approximately 15,000hectares; by the time the preparation was complete, the area had expanded to cover 200,000 hectares, slightlyless than one-third of the entire Upper Magdalena watershed area. Here again, there is a contrast with the stated target areas for the various treatments that in total did not amount to 6,000 hectares (5,200 hectares of on-farm investment and 660 hectares of reforestation on public lands). This small area was justified as "an adequate test of the suitabilityof the particular farming systemsand their acceptabilityto farmers. Herein lies the dilemma of PROCAM and, for that matter, of many other watershed management projects worldwide. Five thousand hectares is too large an area on which to test the suitabilityand acceptabilityof these new interventions, particularly with the recognized institutional weaknesses described in the project documentation. It is also too small an area on which to develop the integrated watershed management model. The most important question related to size and scale of watershed management projects is that of impact threshold, i.e., what percentage of the farmers would have to accept the conservation mandate or what percentage of the land would have to be treated to actuallybring the area into sustainable productivity and relative  - 24 - environmental stability to restore the watershed function. This concept is often overlooked in watershed management planning, particularly in those programs that take the technologicalapproach, i.e., the conviction that the challenge lies in introduc- ing new technologies to the farmers. Directly related to and resulting from this point, is the question of what it will take to achieve this impact threshold. This question is pertinent whether the approach be one of govemmental enforcement of land-use regulations or the more sociallysensitive and community-orientedapproach, such as that adopted by PROCAM, involvingtechnical assistance, services,credit, and incentives for the farmer. * Representativepilot test areas. The three watersheds included in PROCAM appear to have been well chosen, both for their economic importance and for their representativeness of the conditions in the larger watershed of the Upper Magdalena. The choice, however, in this case was somewhat fortuitous in that the overall assessment of the conditions of the full Upper Magdalena watershed was only begun in year three of the project implementation period, through the sub-contract awarded to the Instituto Geografico Augustin Codazzi. In effect, or at least as far as the project documentation is concerned, neither INDERENA nor the Bank had any way of knowing beforehand that the three pilot watersheds were generally representative of the larger area of the Upper Magdalena. * Preliminaryunderstandingof watershed management economics. Engaging in watershed management activities involvinga credit program for the on-farm investments component demands at a minimumpreliminary assurances of the microeconomicfeasibilityof the technological interventions to avoid putting the farmer participant at risk. The President's Report stated that, 'The agricultural technologies in themselves, are financiallyviable--preliminaryestimates carried out on one hectare farm models suggest that all investments have financial rates of return greater than 15 percent." It is unclear whether these analyses of financial rate of return took into account the "effectiveinterest rates on average between 24 percent and 26 percent" and the impact of the inflation. What is clear, however, is that both the Bank and INDERENA were advisedof difficultiesassociated with the on-farm models on several occasions,1JQ/ both during preparation and in the postnegotiation period prior to loan effectiveness. It is difficult to comprehend in retrospect why this fundamental issue was not treated with more rigor by the Bank. In January 1983, in a memo by operational staff to the project file, the followingdecision was recorded: "While the (farm) models are considered satisfactoryfor effectiveness . . ., it is understood that economic analysistechniques will be refined as new models are developed during project implementation." jo/ See for example,the July 28, 1981 memo from AGR to CPS with commentson the green coverdraft of the President's Report; the BTO by the livestockconsultant on the PostnegotiationMision dated January 1982; and the June 24, 1982 mcmo from AGREP to the Latin America region commentingon the methodologyof the farm investmentmodel.  - 25 - It is ironic that despite the rhetoric about the unresponsivenessof the Caja Agraria in facilitating the credit component, they may in fact have saved many farmers from incurring unnecessary debts. Little or no appreciation of that reality currently exists in Colombia, primarilybecause only limited progress has been achieved in assessingthe cost/benefit dimensionsof the farm model. * An appropriateinstitutionalenvironment. A good deal of the Bank's early concerns and efforts with project preparation were in establishingthe appropriate institutional environment for project implementation. It was a most pertinent issue at the time, and one with which the country is still struggling. The present government is reinforcing the earlier decisions to allocate the role and responsibilitiesfor development projects to departmental, municipal, and local corporations as the major thrust of its decentralization strategy. Both of the prescriptions chosen to overcome the institutional concerns, i.e., the establishment of an independent project unit within INDERENA and the emphasis placed on signed subcontractual agreements with the other participating institutions compounded dramaticallythe administrative and managerial problems of the project. * A practical implementation plan. Although detailed implementation planning is usually undertaken early on during the mobilizationand planning phase of a project, a significant and indicative implementation plan should be part of any design effort. No such plan was included in the various project documents. It is therefore small wonder that according to several individualsinvolved in the project, they felt compelled to move quickly to achieve the physical targets after the long delay in reaching effectiveness. Overwhelmed by the continuing and intensive negotiations between INDERENA and the Bank and with the other participating institutions, there was little time for planning and reflection. This only served to compound the day-to-day problems of implementation. The ConceptualBasis for WatershedManagementin Colombia 4.11 On more than one occasion, supervision reports have mentioned the lack of "an overall conceptual framework" as the basis for the subprograms being implemented by the project. This observation is particularly pertinent, for the reasons inherent in the supervision remarks and because, in the opinion of the audit, some of the fundamental premises of watershed management in Colombia also need to be re-examined and possiblyamplified. 4.12 One of the most striking points of the performance audit mission was the dif- ference in impressions between the written description of the project area (admittedly, as stated above, not well documented) and the field visit to this area. In many cases, the topography of these watersheds, with slopes sometimes averaging above 60 percent, and the unstable soils suggest to the trained observer that the only option is protection. Some of PROCAM's interventions were seen to be taking place in areas where continued production under any formula simplycould not be sustained. This was true for both reforestation and improved fodder  - 26 - resources planting. This has a number of important implications for the watershed planner, to wit: * Projects must avoid influencingthe farmer to think that there is really a future as a farmer under circumstancesof "minifundio"in some of these steep and highlyerodable watersheds. Theseprojectsshould not be palliativesfor more drastic changes that may be needed. Projects must avoid reinforcing the unfortunate status quo or further weakening the farm household's tenuous hold on financial and subsistence stabilityby burdening them with credit that will not or cannot pay off, given the inherent resource limitations in which they find themselves. Doing so will at best postpone the final reckoning, but it is also likely to increase the negative effects downstream and the eventual costs of social and land rehabilitation that might otherwise have been avoided. * An important corollary to the first point is that prevention, that is, management and conservation in watershed areas, no matter how costly,will almost always be cheaper than rehabilitation--bothof the watershed areas and of the social systemsundermined in the process of watershed degradation. * Another important concept is that of land-use capability. While it is often possible to 'push the envelop" of current practices to meet land-use capability limitations by increasing conservation measures, these efforts typicallyraise substantiallythe cost of production. There is a point of diminishingreturns in attempting to do so. Additional costs to achieve conservation objectives must either be paid for through the marketplace or must be offset by incentives or subsidies that allow the farmer to continue to sustain a livelihood. In other cases, where small farmers must absorb production tradeoffs to achieve conservation objectives of importance to larger segments of society, there must be some means for compensating them for their losses. No nation has ever been able to sustain a conservation program on the backs of those least able to afford it. * In Colombia,reforestationhas traditionallybeen a primary component of watershed management action programs, whether for production or for protection purposes. There is reason to believe that for the moment neither is economicallysustainable. Reforestationfor protectionpurposes, i.e., revegetating degraded areas with traditional plantation techniques, is rarelynecessary,technicallyrecommended, nor economically efficient. In many cases of degraded areas, one can re-establish the vegetative cover, and with it the watershed function, just by protecting the area from fire and grazing for a period (both of which are necessary if tree planting has taken place on the site) and allowingnatural succession and regeneration to do the work. If for some reason, one feels compelled to replant a degraded area for protection purposes (protection meaning for the soil and slope conditions, the area should not be cut in the future again), then one needs to find more effectve and efficient ways to do so than the traditional plastic-baggednursery-raised seedlings employed in Colombia.  - 27 - These techniques are simplytoo expensive to invest in an area that will only yield indirect returns through the re-establishment of the watershed function. For example, to lower costs, one could try direct seeding of some of the fast-growingleguminous species (e.g., Leucaena spp., Calliandraspp. or Gliricidiasepium or in the higher elevations,Acacia mearnsii or Robiniapseudoacacia). There are also other ways to lower reforestation costs such as planting bare root stock, using the stump technology, pseudoestacas, or revegetating in strips along the contour to help arrest erosion and to allow nature to do the rest in between. Reforestationfor productionpurposes in Colombia is facing an interesting dilemma. At present, it is very difficult to earn a positive rate of retum on reforestation investments, because with the exception of pulpwood, the market is not very attractive. Plantations have to be seen for what they are, i.e., capital investments. The amount invested for a plantation should be capitalizedover the length of the rotation, which means that the wood produced has a certain value. The present high-interest rates and low-stumpage rates or forest royalties for cutting in the natural forest mean that it is very difficult for the plantation wood to compete economically. This is especially true for the eucalyptus and pine plantations, which are destined for use as service woods, i.e., not for fine fumiture. In essence, either Colombia needs to raise forest royalties to reflect the real value of the natural forests (which might then also have the ancillaryeffect of inducing people and companies to manage rather than just exploit the natural forests), so as to even out the competition, or it needs to destroy more natural forests so the supply/demandcurve starts to drive up prices. The latter is undesirable but likely to happen unless the country takes several serious, practical steps. 3 This bias in the past toward reforestationand tree-planting for watershed management purposes has also overshadowedthe importanceof interventionsin the agricultureand livestocksectors. Unless there is a direct and sharp change in strategy and the country begins to deal effectivelywith the problems brought about by agriculture and livestock on the steep hillsides,degradation willcontinue and the cost to the country will far surpass the potential losses from a weak forestry production sector. Colombia has some of the most potentially severely erodable watersheds in the world, coupled with an economy that is highlylinked to the availabilityof water, both in quantity and quality. The fact that tree-cutting is sanctioned, but you can do whatever you want on the land after the trees are gone, reflects best this narrow viewpoint that forestry is the critical concern for watershed management. The focus on interventions to bring soil and water conservation to peasant farming systemsduring PROCAM was weak and muddled. Efforts to encourage improved pasture management and the establishment of the cut-and-carry system of fodder production were pursued with more vigor and with greater acceptance and success. No discussionwas undertaken, however, of the need to take on the more difficult but possibly longer-term solution to the livestock problem by reducing herd numbers. Future watershed management action projects need a wider variety of tools to choose from in addressing,where feasible, agriculture- and livestock-basedconservation needs. The followingis a brief list of some of the  - 28 - promising technologiesbeing applied around the world to achieve conservation objectives in hillside farming systems: - Alley cropping or contour hedgerows producing both fodder and green manure while helping to arrest erosion. - Soil conservation techniques for the enhancement of annual cropping systems,including: improved crop varieties, seed selection, improved planting distances and plant distribution, green manuring and mulching, zero tillage, vegetative and similar barriers to slow water runoff, improved fallow,strip-cropping of vetiver and leguminous creepers for site and soil enrichment, and the selection and application of appropriate fertilizers. - Perennial tree crops, includingcoffee and cacao (as was tried in PROCAM) as well as grafted fruit trees on individualterraces. - Farm-level conservation engineering works, such as side hill ditches, gully plugging,and grassed water runoffs. - Coppice-based fuelwood production woodlots with fast-growing leguminous tree species on steeper slopes and in degraded areas, established through direct seeding or by vegetative means. - Livestock improvement programs (in addition to pasture improvement practices) to increase the productivityof individual farmer-owned animals through improved nutrition, simple veterinary practices, and selected animal breeding. * Even the cursory visit to the PROCAM project area was sufficient to point out that both roads and paths through these steep areas are probably responsible for a good deal of the erosion problem. Municipal authorities must ensure that the engineering design of the road system avoid channeling water to the point where it gathers such power and momentum that it cuts deep gullies and causes severe erosion. In a study carried out in El Salvador under comparable conditions, the road and path system was determined to be responsible for as much as 25 percent of the total erosion in the watershed. 4.13 All of these points serve to highlight another principal element of the concept of watershed management--that is, the need to make choices and to do so in a logicallyand sequentiallysound manner. One of the greatest weaknesses of integrated rural development projects is that they attempt to do everything, diffusingcapability, expertise, and resources and thus failing to generate the momentum needed for real impact. PROCAM has suffered to some degree from this type of problem. 4.14 Making choices in watershed management is no simple matter. Professional resource managers assessingthe project area--the topography, climate, soils, and land-use  - 29 - conditions--cannotfail to note a high degree of variable conditions all needing attention. These biophysicalconditions are further compounded by the socioeconomicvariables, such as land tenure, household objectives, and farming practices, thereby providing,if the project is indeed an integrated approach, a potentially staggering array of intervention possibilities. 4.15 Knowing what could be done in watershedmanagement is far easier than knowing what should be done. In general, the choice based on a simple assessment of the biophysical needs generates a list much too broad for an action program. A thorough assessment of the socioeconomicparameters of the watershed can help enormously to filter the basic list. Finally, watershed management planners must applya set of widely acceptedselectioncriteria,based on the larger development goals of the project. While each project team may wish to develop its own set of selection criteria, the followingexamples,which might have been applied to PROCAM, are illustrative. The interventions chosen would: * best meet the needs of the majorityof the target population of hillside farmers; * be most efficient in reversing the trends of watershed degradation in the project area, given the area's socioeconomicconstraints; * have the highest probability of engendering sustainable, productive farming systems. 4.16 Several other techniques may also be required to further narrow the scope of project activities to a manageable and focused, overall effort. A full explanation of all these techniques is beyond the scope of this report. The following,however, seem pertinent in light of the PROCAM experience: * Land use assessment and planning techniques carried out at a scale most directly applicable and practical for local conditions and avoidingcompilation of endless data and information sets that are difficult to synthesize and analyze. * The application of the newer techniques of rapid ruralappraisal (the "sondeo" approach), rather than detailed socioeconomicsurveys. * Using both of the above techniques to establish the biophysicaland socioeconomic baseline essential for a monitoring and evaluation system. * A monitoring and evaluationsystem set up to institutionalize the communication, feedback, and flexibilitynecessary for an effective participatory development process. PROCAM was criticized for failing to set up a monitoring and evaluation system, but the rationale behind such a system was never succinctlyexplained as part of the concept of participatory watershed management. Because of the high degree of diversityinherent in participatory watershed management programs, flexibilityis very important. This flexibility,however, should not be ad hoc. Monitoring and evaluation can help assure purposeful change as a result of the analysisof cause and effect regarding constraints (and for that matter successes) in project implementation. It can  - 30 - do so as a result of a number of situations that may arise during implementation: an improved or changed understanding of the problem by either technicians or lay persons; the need for adjustment resulting from unforeseen impacts of an activityor intervention; the need to hasten reaction time in the face of failure; and the need to adjust implementation schedules resulting from mobilization and delivery delaysthat often occur. * Th delimitation of prority work areas within the watersheds that also takes account of the relationships between the upper and lower reaches of the watershed area. The Institutional Framework for Watershed Management 4.17 As is now well known, institutional constraints were a principal concern of the Bank at the outset of this project and, in actuality, were key elements leading to its lack of success. This can be understood from the statement in the Plan de Accion Forestal para Colombia as follows: "Low Managerial Capabilityof the State--The institutional structure of the sector exhibits problems mainly originating from the multiplicityof agencies created for the same purpose without national and regional task differentiation, due also to the haphazard delegation of responsibilitiesand to the lack of an adequate institutional hierarchical structure to effectively direct the activities undertaken by the State and the private sector." 4.18 The question, of course, is what to do about it. Building on the PROCAM expe- rience, institution-buildingor strengtheningneeds to be further enfranchisedas an immediateproject objective. PROCAM suffered from shallow institutional assessment and analysis. Institution- building requires goals and benchmarks--a path to follow, indicating anticipated achievements. Too often, institution-building,while a principal concern of the project, is relegated to vague notions of reorganization and staff development. In the case of PROCAM, constant references were made to the Bank-imposedstudy of the administrativeand financial management procedures of INDERENA, which eventually was never accomplished. There is good reason to suspect that the institutional problems of PROCAM were considerablymore fundamental than that. 4.19 A project needs to bring in the specialists in institutional analysis,determine in understandable terms what the problems are, and provide indicative designs and models with clearly stated end results. Moreover, local personnel at all levels must be fully apprised of these efforts. The Bank needs to bear in mind, and remind Colombian and other colleagues,that these complex inter-institutional arrangements are likely to be subject to the old adage that the whole will only move as fast and as effectively as its weakest component. 4.20 Now that the project has ended, the Watershed Management Division of INDERENA is valiantly trying to carry on with both the developmental and investigative programs begun under PROCAM. The resources are extremely limited, and as a result INDERENA is seeking cooperative relationships with both other line agencies and the departmental and municipal corporate entities now responsible for land management under the decentralization mandate. They need to be wary of not repeating the institutional problems of PROCAM. These relationships between institutions, based on collegial and professional understanding between like-mindedfunctionaries of the participating organizations, tend to be …-~ _ _ ~~~~~ _…_  - 31 - rather ad hoc. As these relationships gather momentum, there is the inevitable urge to try and officializethem; this brings with it a high surcharge of administrativeadjustments and bureaucracy, which may ultimately be self-defeating, as it was to some degree during PROCAM. 4.21 A simplified,basic model for participatory watershed management is currently emerging in Colombia. Regional or municipal corporations are now responsible for both planning and implementation of watershed management efforts.Jj They will draw on the knowledge and capabilitiesof the more specialized line agencies for subject matter advice both in the planning and implementation stages. For example,INDERENA will provide advice and infor- mation on natural resource management; ICA, on agricultural and livestock improvement and development approaches and techniques; SENA, on promotion, extension and training techniques, packages and materials; HIMAT, advice and assistancewith hydrometeorologicalconcerns; and the agricultural banks, credit assistance and guidance. Lacking in this equation as yet is an institution to guide the communityparticipation process and it would have an in-depth knowledge of local constituencies. These corporations willbe represented in the field by a cadre of multipurpose general extensionists,who will deliver project interventions and support elements to the rural people and local communityorganizations. 4.22 The key to making the model work is an understanding by all involved of specific roles and responsibilities,as well as an overall sense of service and accountability. The departmental corporations will provide servicesand be accountable to their local constituents, while the line agencies willprovide services and be accountable to the corporations. In cases of larger corporations, such as the CVC, they will directly employ subject-matter specialists to amplify and accelerate their capabilitiesin the field. It is also conceivable that, if necessary,line agencies could be called on to second staff members as subject-matter specialists. The RENORDE network will allow for inter-institutional cooperation for the purpose of expanding and improving the state-of-the-art of watershed management in Colombia. 4.23 In short, simplifyingthe inter-institutional structures for watershed management in Colombia may prove to be as effective as any single technological innovation or achievement in the near term. The Legislative Framework in Colombia 4.24 One further point for the Bank directly related to the institutional concerns is the legislativeframework. There is absolutely no mention of the legislative dimensionsof watershed management in the PROCAM documentation. In Colombia, as in few other countries, a good deal of the activities undertaken by government organizations are either prescribed or facilitated or agreements of a more localized nature, usually by numerous legislative mandates, decretos-leyes, D Manyof the peopleinterviewedduring the performanceaudit asaerted that one of the leasonsof PROCAM andother watemhed managementprojects in Colombia is the need to address the problems on a subwatershedbasis sub-cuenca.Thia seems to be a rationale approach, but there is alsothe issue about whether the subwatershedboundaries are compatiblewith those of the administmtive boundaries of the regional or municipal corporations. This issue needs further study.  - 32 - promulgated by regional entities. For example, Decreto-Ley 2811/74--CodigoNacional de Recursos Naturales Renovables y Protection del Aznbiente;Decreto-Ley 09/79--CodigoSanitario Nacional; and, Decreto-Ley 56/81--Normassobre Obras Publicas de Generacion Electrica, Acueductos y Sistems de Regadio, all have major implicationsrelated to watershed management. This legislation is in many cases normative and provides the fundamental base for organizational policy, philosophy,and strategy. Incidentally, many specific and facultative references exist to natural resources in the new Colombian constitution; (see for example, Capitulo 3 Articulos 78- 82, Articulo 318, and Articulo 331, the latter of which creates La Corporacion Autonoma Regional del Rio Grande de la Magdalena). Future project endeavors must conform to the legislativeframework. Understanding WatershedManagementEconomics 4.25 Perhaps the most serious impact of the Bank not having carried out the appraisal of this project the continuingfailure to come to gripswith the question of the costs of watershed is management. Little achievement has been obtained in this realm of inquiry--in neither the Bank's PCR nor in the Final Report prepared in-country is there any mention of the economic feasibility of watershed management or the models put forward during the project, in either quantitative or qualitative terms. This audit mission,given the relatively short time at its disposition,was no more successful in assemblingany kind of a quantitative assessment of the outcome of the project. This series of circumstances,however, should not and cannot be allowed to minimize the fundamental importance of the economic dimensions,both micro and macro, of watershed management in Colombia. The discussionthat followsdraws on some of the implicit economic assumptions of the PROCAM experience and raises some points that must be factored into such programs in Colombia in the near future. * As a fundamental principle of a genuine participatory approach to watershed management, farmers must perceive and receive tangible benefits in return for their efforts to conserve soil and water. 4.26 Colombia, unlike few other countries, is presently experiencing the direct impact of watershed conditions on its overall development perspectives. The national dependency on hydroelectric energy, large numbers of urban water systems,irrigation potential for export crops and transport by means of navigable rivers are all directly and inextricablylinked to the destiny of mountainous watersheds throughout the country. The magnitude of deforestatioh and the subse- quent inappropriate land-use practices are spreading wide and far. It is doubtful that government, either central or more local, with or without donor or multilateral support, can muster the resources necessary for projects in which the full measure of direct inputs, even at relatively modest costs per hectare treated, would have to be provided and disbursed. 4.27 Fortunately, broad agreement exists in principle in Colombia that what is needed is a pragmatic integration of agriculture, animal husbandry and natural resource management development efforts that foster popular participation. These programs must build on the traditional systemsand on the inputs already available to the farmer--the land, labor, capital, and technology utilized on a day-to-daybasis. Only the farmers, controlling erosion, improvingsoil conditions, managing pastures and forest areas, because they are interested in the direct economic  - 33 - benefits, can begin to make the substantial impact required to restore ecological stability and to deal with the socioeconomic problem of watershed degradation. 4.28 Where sustainable production systems can be achieved by improvingfarmer access to agricultural services,inputs, and markets, the costs are those associated with providing these servicesto the farming community. Thus, efficiencybecomes the major economic concern. In cases in which credit is necessaryto achieve sustainability,the microeconomicsof local farming systemsneeds careful study, especially in cases such as PROCAM, in which interest rates are high. * In areas in which resource limitations are severe, incentives or subsidies may be justified to offset the production tradeoffs associated with protection and conservation. 4.29 In many areas of the Upper Magdalena watershed, resource limitations are such that either land must be taken out of production to be protected, or the cost of additional measures, typicallylabor-related, to achieve conservation, distort the cost/benefit equation of the farming system, making it difficultfor the farmer to obtain a profit. The economic issue here is that of calculatingthe value of soil and water conservation and of watershed management. This calculation is necessarily more complex, because it involvesnot only the local production and socioeconomicsystems but also an evaluation of the downstream effects. 4.30 For government to be able to justifywatershed management incentives or subsidies,it is necessary to estimate the value of benefits downstream. This was the basic premise behind the Soil Conservation Fund attempted during PROCAM and is the rationale for the 2 percent surcharge on hydroelectric generation. These values are relatively easy to derive although it is fundamental to state the assumptions made in doing so--an aspect that is often overlooked. These values can be quite direct and dramatic. The direct costs to the Municipal Water Supply System of Ibague for continued and extensive maintenance of its water intake facility on the Rio Combeima could be offset, if the watershed were stabilized. When this issue was broached during the visit to the area, the immediate reaction was that the watershed management program would require far greater inputs. Perhaps, but the subsequent discussion revealed that, in fact, no one had any realistic estimate of the cost involved, either for the water system maintenance or for the magnitude of the watershed management program needs. 4.31 Often in sensitive watersheds, people live in poverty and misery,while trying to sustain themselves on steep areas and thereby provoking erosion. In some countries with high overall population density (e.g., Nepal, Haiti, El Salvador, and Bangladesh), this situation must be dealt with in situ. Watershed planners in Colombia, however, have to consider the option of resettlement. It may be far less costly in social, economic, and ecological terms. In the 1970s, under the Caqueta Land Colonization Project, an INCORA project with Bank support estimated that "the average investment per family for farm development," for farms averaging over 100 hectares in size, was less than US$5,000. PROCAM spent much more than that, admittedly with a low level of efficiency,on its 270 farmer beneficiaries. Ultimately and unfortunately, no real record exists of how much was spent on achievingconservation on these farms.  - 35 - Page 1 of 3 ESIMATED PROJECT COSTSAND DISBURSEMENTS Estimated Cost: Local Emiorrn Total (USS million euivalent) On-farm investment 5.3 1.8 7.1 Extension and training 1.2 0.4 1.6 Public works (including reforestation) 1.1 0.3 1.4 Environmental protection and control 1.3 0.5 1.8 Research and studies 2.4 2.3 4.7 Preparation of Phase II project 1.7 1.7 3.4 Project management 1.2 0.5 1.7 Total base costs 14.2 7. 21.7 Physicalcontingencies 0.4 0.5 0.9 Price contingencies 28 1.9 4.7 Total project cost 1/ 17.4 9.9 27.3 Project Financing Plan: Government 9.5 - 9.5 Participating banks 4.8 2.2 7.0 IBRD 1.3 7.7 9.0 Beneficiaries 1.8 - 1.8 Total Project Financing Cost 17.4 9.9 27.3 Estimated Disbursements: EY8 Y FY84 FM E6 E2YZ (USS million equivalent) Annual 0.9 2.0 1.9 1.7 1.5 1.0 Cumulative 0.9 2.9 4.8 6.5 &0 9.0 Rate of return: Not applicable. Project file: A project file containing working papers, which include detailed descriptions of all components, has been prepared as a basic guide for the project. I view of the pilot nature of the project, an appraisal report has not been prepared. SOURCE: President's Report 1/ INDERENA is exempt fmm custom duties.  - 36 - ANNEX 1 Page 2 of 3 C. Preparationof Phase n1 1. A major goal of the project is to assist Colombia in developing appropriate techniques and systemsfor carryingout large-scaleefforts in watershed management programs in the future. In line with this, preparation of a second, larger project would be carried out under the proposed project. The pertinent studies would be initiated in early 1983 and would serve to monitor the results of the proposed project. The preparation of the second project is Lobe carried out jointly by INDERENA and mostlylocal consultants. Specialized foreign consultants would be engaged as needed and would be used to provide training to INDERENA staff. In order to prepare the necessary manpower to carry out large-scalewatershed management activities,short- and long-term training would also be provided under the project (para. 45). sts and Fnning 2. The total cost of the project (net of taxes) is estimated at US$27.3 million equivalent, of which US$9.9 millionequivalent (36%) correspond to the foreign exchange costs. Physicalcontingencies amounting to 15% for civil works, studies and equipment have been included in the estimates. Price contingencies amount to 17% of base costs and are based on forecasts of yearly domestic and international price increases. (Project costs by component are presented in the Loan and Project Summary.) By type of expenditure, project costs are as follows: Loa Foreign Total (in USS million equivalent! Contracts for civil works and reforestation 1.5 0.4 1.9 Vehicles,equipment, and scientific instruments 0.8 2.1 2.9 Consultant services, and training 1.0 1.0 2.0 Studies and extension - Operational Costs A/ 4.4 1.7 6.1 Project administration 1.2 0.5 1.7 On-farm investment 5.3 1.8 7.1 Base cost 14.2 7.5 21.7 Physicalcontingencies 0.4 0.5 0.9 Price contingencies 2.8 1.9 4.7 Total cost 17.4 9.9 27.3 V requiringperiodicreplacement, Includesresarch and studiesto be caried out by publicsectoragencies,equipment mait of vehicles,etc.  - 37 - ANNEX I Page 3 of 3 3. Despite extensive preparation, in greater detail than is usual for a pilot project, the project is too new in conception for every aspect to have been considered in detail at this point. Therefore, in addition to physicaland price contingencies,a small additional contingency amounting to US$500,000has been included so that financing would be available for studies and training heretofore unforeseen or ultimatelyexpanded beyond current expectations. Assurances have been provided that the terms of reference of any study or training program to be carried out under the project would be satisfactory to the Bank (Section 3.01(g) of the draft Loan Agree- ment). 4. The credit component (US$7 millionor 26 percent of project costs) equivalent to about 80 percent of the on-farm investments (US$8.8 million, including contingencies) would be financed entirely by the banking system. Since the resources are non-incremental, they are not included in Bank financing. The proposed Bank loan of US$9 million (corresponding to 91 percent of the foreign exchange cost and about 33 percent of total project costs) would finance the estimated foreign cost portion of all components (except on-farm investments) and US$1.3 millionof local expenditures for vital project components with low foreign exchange components (such as reforestation and extension). The government would finance US$9.5 million equivalent (35 percent) in local costs, and the beneficiaries would contribute US$1.8 millionequivalent (6 percent).  - 39 - ANNEX2 EXCERPTS FROM THE LOAN AGREEMENT UPPER MAGDELENAPILOT WATERSHEDMANAGEMENTPROJECT (LOAN 2069-CO) 1. Special conditions of the loan are listed in Section HI of Annex Im. Conditions of Effectivenesswould be that: (a) INDERENA had signed and duly authorized subsidiary agreements with HIMAT, ICA, and SENA and a Cooperating Agreement with either Banco Ganadero or Caja Agraria for carrying out the hydrometeorology,extension, training and credit components, respectively(Section 3.01(c) and (d) and 7.01(c) and (d) of the draft Loan Agreement); (b) The Project Director and Research Coordinator had been appointed (Section 3.05 and 7.01(a) of the draft Loan Agreement); (c) INDERENA had established the Soil Conservation Fund (Section 3.07(a) and 7.01(b) of the draft Loan Agreement); (d) INDERENA had entered into arrangements, satisfactory to the Bank, for contri- butions to the SCF of no less than Col$60 million (Section 3.07(b) and 7.01(b) of the draft Loan Agreement); and (e) Financial farm models of a scope and detail satisfactory to the Bank had been submitted to the Bank (Section 3.010) and 7.01(e) of the draft Loan Agreement). Conditions of disbursement would be that: (a) For the advance of the working fund, that the fund had been established and that the Government had approved the transfer of its corresponding share (Section 2.03(b) of the draft Loan Agreement); and (b) For subsequent replenishments to the fund, that the Government, the Bank and INDERENA had reached agreement on the size of the fund for the subsequent period and that the Government had approved the transfer of its corresponding share (Part 4(b) of Schedule 1 and Section 2.03(c) and (f) of the draft Loan Agreement). 2. I am satisfied that the proposed loan would complywith the Articles of Agreement of the Bank.  - 41 - BIBLIOGRAPHY Departamento Nacional de Planeacion, Plan de Accion Forestal para Colombia Bogota, n.d. INDERENA, Informe Fmal - Proyecto Cuenca Alto Magdalena (PROCAM). Bogota, June 1988. World Bank. 1966. "Colombia- Livestock Development Project. Report No. TO-523a. Washington, D.C. 1971. 'Colombia - Caqueta Land Colonization Project." President's Report. Report No. P-921. Washington, D.C. -----. 1981. "Colombia- Upper Magdalena Pilot Watershed Management Project." President's Report. Report No. P-3106-CO. Washington,D.C. -----. 1982. "Colombia - Upper Magdalena Pilot Watershed Management Project - Loan Agreement, Loan No. 2069-CO." Washington, D.C. 1982. "Colombia- Agricultural Research and Extension Project L" Staff Appraisal Report. Report No. 4384-CO. Washington, D.C. ----- 1983. "Colombia - Agricultural Research and Extension Project L" President's Paper. Report No. P-3579-CO. Washington, D.C. ------ 1983. "ColombianAgriculture: Selected Issues and Some Directions for Strategy." Report No. 4275-CO. Washington, D.C. -----. 1991. 'Forestry Development: A Review of Bank Experience." Report No. 9524. Operations Evaluation Department. Washington, D.C.  CO LOMBIA UPPER MAGDALENAWATERSHED MANAGEMENTPILOT PROJECT X ~~~~~~~ [uno ~ ~ ~~~To \,, \ j -) X \ n ° s e C A S A N A R E j A N T I O O U A 5 X -</{ /~~~~~~~~~ 'pi/ \'' '' ; r,(~l S ) . A COMBEIMA / CA S bt~ _ ~ C IT9 .I)3 Sedl 1[2g < ) r> llesCi Udr tuy 960 MW -I;~ ~ ' 1' ~ X ° e/ V JC o~ThM t I ~ ~~~~~~~~~~~~~~Selected vii Rosds Unde, Sur-ey. 5,64 3 MW //l t] *2{ x = R,,e,s T.t.1 iled C.pcity S487 MW. Burga \ fA 4 taO ( t ~~~~~~~~~~~~~~ILOMETERS ? 25 50 715 100 k cl > ¢\ :] )t >)LA CE1315 NIILES0 ~~ ~ ~~~~~~~~~25 50L \t X J/tz LC~~~O, P to 9 i2c! Noioo Coio Uee Co v Ec N EZ.U2t7EMW S A1 W~~~~~~~ATEQ'H(ED ) < vJ X~~~~~~~~~~~~~~~~~~~~~Tlol7 Ineiold Cooct 8r8 MW. 4ECOh _X\~/ YAGUARA 'O' -V \ \ Pereirp-~ WATERSHED H 8WAE AD Toluc - Li'I - DeormeI oudoi, KILOMEERS o 25 OL75 MB0 --. _ Santo.deANIAHD / ' w PROJECTJ P WH OOER t. i rit |~og j Reoodor Bo neo o 2 2o 50 Pelnitro~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 /~~~~~~~~~~~~~~~~~~~~~~.I NEIA20 d.1~~~~~~~~~~~~~~~~~~~~~~~~\