India - Kandi Watershed and Area Development Project (Punjab)

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 8366 PROJECT COMPLETION REPORT INDIA KANDI WATERSHED AND AREA DEVELOPMENT PROJECT (PUNJAB) (LOAN 1897-IN) FEBRUARY 9, 1990 Agriculture Operations Division Country Department IV Asia Regional Office This document has a restricted distribudon and may be used by recipients only in the performance of their offciad duties. Its contents may not otherwise be disclosed without World Bank authorization.  EXCHANGE RATES Currency Unit s Indian Rupee (Rs) Appraisal Year Average (1981) i Rs 9.1 - US$1.00 Completion Year Average (1988): Rs 14.9 - US$1.00 WEIGHTS AND MEASURES Metric System ABBREVIATIONS AD s Agricultural Department ERR Economic Rate of Return FAOICP F Food and Agriculture Organization Cooperative Program FD s Department of Forestry GOP s Government of Punjab ID : Irrigation Department PAU s Punjab Agricultural University PCR : Project Completion Report PSTC s Punjab State Tubewell Corporation SAR S Staff Appraisal Report SCD s SoiI Conservation and Engineering Department FISCAL YEAR April 1 - March 31  FOR OICIL US ONLY THE WORLD BANK Washington, D.C. 20433 US.A. O1k* ni OivetqrGwCisaI Opfatinm Evstatatm February 9, 1990 MEIORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on India Kandi Watershed and Area Development Project (Punlab) (Loan 1897-IN) Attached, for information, is a copy of a report entitled nProject Completion Report on India - Kandi Watershed and Area Development Project (Punjab) (Loan 1897-IN)" prepared by FAO/CP with Part II of the report contributed by the Borrower, and the Evaluation Summary prepared by the Asia Regional Office. No audit of this project has been made by the Operations Evaluation Department at this time. Attachment _~~~~~~ This docunment has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Dank authofization.  FOR OMCUL USE ONLY INDIA KANDI WATERSHED AND AREA DEVELOPMENT PROJECT (PUNJAB) (LOAN 1897-IN) PROJECT COMPLETION REPORT Table of Contents Pane No. Preface ....... ...... ......... .* ........ , * EVALUATION SUMMARY................ .................. -l PART I - PROJECT REVIEW FROM BANK PERSPECTIVE A. Project Background ....... . ....................................... l Genoral ........ ..... ............ ..... ........... 1 Project Foneulatlon ..................... .........*.. ...... I R. Implemhntation .........tc....t .............................. 2 General .......Rse.c............................................. 2 Rehabilitation of Upper Catcbments ............... 0000.*000*0. .................... 3 Animal usbandry ........................................................ 3 trrigation and Flood Control .............................. 4 F a rm Developmet .......................................... 0 ... 8 Agricultural Rsearch .......................................... 6 BTraining Perf.r.anc.................................. 6 . Costsi Fins, ncng and Disbursemen ts ............................. 7 C. Pesults ......... . .........a Agricultural and Enviroremntal Impacto ......... .o .........................o .......... 8 Economic Rate of Return ... oo.o... o... .0.O..0... 0404........................... 9 Borrowers' Performance .....o.......... o..... o.... o........ .o.................. ll Bank Performance ..................................o............ 13 D. Conclusions, Lessons Learned and Issues .... .................. . ....................... lS Lessons Learned ...... ....................... o................. oo..... 15 Issues ....... . ...................................... 19 PART II - PROJECT REVIEW FROM BORROWER'S PERSPECTIVE | This document has a restrcted distibution 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.  (ront) Pate No. PART III - STATISTICAL INFORMATION Basic Data Tables ................... . 25 Table 1 Project Identity 1 Project Timetable 1 Cumulative Annual Disbursement of Loan 1 Disbursement by Category 2 Project Implementation - Physical Targets 3 Targe,s and Achievements of Irrigation Component 4 Project Costs 5 Project Financing 6 Economic Impact 7 Status of Covenants 8 Supervision Detail  INDIA KANDI WATERSHED AND AREA DEVELOPMENT PROJECT (PUNJAB) (LOAN 1897 -IN) PROJECT COMPLETION REPORT Preface This is a Project Completion Report of the Kandi Watershed and Area D-%velopment Project (Punjab) in India for which Lodn 1897-IN in the amount of US$30.0 million was approved on July 20, 1980. The loan was closed on March 31*, 1988 after a delay of 24 months. The final disbursement was made on September 30, 1988. The Project Completion Report (PCR) comprises a report prepared by staff of FAO/World Bank Cooperative Program (FAO/CP), a report prepared by the Government of Punjab's project coordination office and an evaluation summary prepared by regional staff. The FAO/CP report was prepared following consultation with Bank staff and based on Staff Appraisal Report (No. 2174a- IN), Loan Agreement dated September 20, 1980, supervision reports, a final report of a consultancy team engaged under the project and a final evaluation report prepared by Punjab Agricultural University, department of economics and sociology, who were engaged for the purposes of monitoring and evaluating the project. The evaluation summary has been prepared on the basis of the same documents, the FAO/CP report and visits to the area in connection with identification and preparations of a follow-up project. The PCR was read by the Operations Evaluation Department. The draft PCR was sent to the Borrower on August 11, 1989, for comments by September 29, 1989, but none was received.  - ii - INDIA RAND! WATERSHED AND AREA DEVELOPMENT PROJECT (PUNJAB) (LOAN 1897-IN) PROJECT COMPLETION REPORT Evaluation Summary ITtroduction 1. The Kandi Watershed and Area Development Project (Loan 1897-IN) was appraised June 1978, approved July 1980 and became effective November 1980. It represented the first attempt to tackle, in an integrated way, the ecological degradation problems of the Himalayan sub-mountainous zone in the state of Punjab, known as the Kandi tract. Various state government departments had previously undertaken, with disappointing results, programs to address some of the severe problems that fell in their purview. The project provided for investments in upper catchment rehabilitation, flood control and irrigation, farm development, technical assistance, research and training and project administration. Obiectives 2. The primary objective in this and other watershed projects has to do with slowing down or if possible reversing the so-called man made degradation which is mostly manifest in accelerated run-off usually with sedimentation, reduced agricultural productivity on what is essentially marginal land to begin with, and progressive removal of vegetative cover on non arable land. The project intended to address an additional problem having Lo do vith flooding from a large number of seasonal torrents which each year destroyed farm land on the plains. Implementation Experience 3. A substantial proportion (53Z) of the total project cost and activity, as the project was originally conceived, was allocated for flood control and irrigation. One of five proposals which involved construction of a dam, irrigation distribution infrastructure and flood protection work, was prepared ready for implementation at appraisal and four more were to be prepared. When a consultant team required under the project was eventually hired and feasibility reports prepared, it was clear that the potential for this activity was much less than estimated at appraisal. This, together with slow progress in implementing the one plan that was ready, resulted in recasting of the project to reduce this activity and instead include two small scale irrigation initiatives as well as increasing the number of watersheds to be treated. These changes upset the line department responsible for implementing the irrigation related components and this consternation pervades the Government's completion report. There is some doubt about whether building dams of the size envisaged in this geologically unstable area is wise from an engineering standpoint and considerably more doubt about it from the standpoint of integrated watershed development. This in a sense reflects poorly on project identification.  - iii - 4. The notion of integration was important in the concept of the project and remains important for watershed development. Yet is was not achieved in any sense. There are two significant stages in which integration is essentia'. namely, planning and implementation. In both respects, she line agencies responsible for conceiving and implementing the measures r4quired to address the problems must work together with each other and with the beneficiaries. This did not happen. Individual line agencies undertook their separate agenda according to their own priorities. Except for the irrigation subcomponents (para. 3) they attained their targets rather easily yet not one watershed was completely treated so as to affect some degree of control of run-off and its concomitant effects. 5. Furthermore, lack of coordination negated these achievements. A striking example of this, but by no means the only one, has to do with the livestock component which had as one objective the replacement of poor quality cattle and buffalo with a lesser number of higher quality to improve income and reduce pressure of grazing on land to be revegetated. The replacement target was attained but the supportive actions to develop feed sources did not dogelop so that many beneficiaries sold their animals after milk yields began to decline in successive lactations through inadequate nutrition. Resuits 6. The major expectations entertained for the project were not achieved mainly because of undue emphasis on the irrigation component, and lack of coordination between line agencies both in planning and implementation. None- theless, the necessity of continuing action, to regenerate the Kandi tract and like areas is not in question. Valuable experience was gained in the course of the project and many of the lessons have already been taken into account in the preparation of subsequent watershed projects. The estimated economic rate of return for the project overall, calculated at completion, was 81, ranging from 2-222 for separate analysis of five components. The lowest rate which was for the irrigation component, resulted from delayed implementation and certainly depressed the overall ERR. Findings and Lessons 7. The project was accepted as a learning experience and indeed it was. Rehabilitation of fragile hilly areas such as the Kandi tract requires substantial shifts in thinking towards the promotion of rainfed agriculture in marginal areas away from heavy emphasis on irrigated agriculture, towards low cost stabilization of non arable land away from establishment of conventional forest plantation,. Furthermore, it required a new managenent strategy for applying the resources of the line agencies. The following significant lessons were lesrneds (a) Management. If implementation is best carried out by line departments, and in Punjab it is, then a mechanism is required to coordinated their activities and induce them to coordinate with one another. This was arranged by using a project ccordination office whose head is a competent administrator with a technical background but without the civil service seniority to be able to effect the cooperation and without control over the budget which went directly  - iv - to line departments. The shortcomings have already been recognized and proposals to overcome them have incorporated into new project designs. (b) Time frame. A normal five year time table is too short for integrated watershed development especially if its a first time project. Apart from the lag required to set up the administration and organize the field level planning, many of the project activities need to be implemented sequentially and ought not to start before others are complete. At the same time, watershed development is a continuing process and not a discreet rehabilitation operation. Awareness and acceptance by both government officials and the rural population are only built up by participation. Within the overall objectives, short term goals are constantly evolving and management requires regular interactive planning. A project such as this is in fact a time-slice of a long term process. The optimum length of the time-slice is about seven years. (c) Sequential Implementation. The project provided valuable experience in assessing the efficacy of treatments from a technical view point and their appropriateness from an investment standpoint as well as a good lesson in sequencing. Cattle exchange, if it can be a justifiable investment, should only begin when the beneficiaries have taken the necessary steps to provide regular sources of fodder; revegetation of forest areas ought not to start until the areas have been stabilized and the grazing pressure reduced; and check dams built only after vegetative measures for erosion control put in place. Sequencing involves coordination of line agencies, interactive planning with both the line agencies and the beneficiaries and following a critical path plan. Cd) Sustainability. Sustainability of conservation activities which were implemented with considerable success and the positive environmental impact will depend on the availability of funds for maintenance and the incentives the population have to protect them. Neither is absolutely assured. Production activities that were only partially implemented are expected to be sustained. (e) Other Lessons and Issues. The PCR draws out several other lessons of which perhaps the most valuable is that in-situ moisture conservation as the basis of rainfed farming should be the *central tactic' of future interventions rather than undue emphasis on small scale irrigation where likely benefits have been overestimated. Significant issues requiring consideration have to do with the watershed as a planning unit, economic rate of return as an investment criteria, and the policy relating to subsidies and cost recovery in this and like projects.  INDTA KANDI WATZRSBED AND AREA DEViLOPHENM PROJECT (PUNJAB) (LOAN 1897-IN) PROJECT COMPLETION REPORT PART I - PROJECT REVIEW FROM BAAK'S PERSPECTIVE A. Proiect Backaround General 1. The Kandi Watershed and Area Development Project constitutes the first attempt to tackle the problem of the Himalayan sub-mountainous zone in Punjab. The project was initially appraised in May/June 1978 but it was reviewed and cost tables were up-dated in April 1980. An agreement for a lo&n of US$30 million was signed in September 1980. The project, with a total estimated cost of US$60.8 million, to be implemented over a five-year period, became effective in November 1980 and was completed in March 1988 after a three-year extension. I.. An initial draft project completion report (PCR) prepared by the ?roject authorities for the Government of Punjab (GOP) together with evaluation reports by the Punjab Agricultural University, Ludhiana, and consultant reports form the basis for this PCR. In addition this report is based on the review of the Staff Appraisal Report (SAR), the Development Credit Agreement, supervision mission reports and project files, and field visits in February 1989 by an FAO/World Bank Cooperative Program mission. Proiect Formulation 3. The Kandi tract, identified as the most economically backward area of the Punjab, includes all land within the Punjab Shiwalik Hills, - foothill range of the Himalayas, together with a narrow band of largely under-developed farmland immediately below the hills. The tract covers a total aree of about 4,600 km2 which is 92 of the State, and contains 6Z of Punjab's population. Early concern about ecological d-gradation in this area led to the promulgation of the Punjab Land Preservation (Choes) Act in 1900,1/ which allowed areas to be closed to grazing, controlled the type of livestock permitted within grazing areas and could prohibit tree felling, cultivation and quarrying. Unfortunately the Act was not implemented with any vigor until the late 19309 and even then only for a few years and in a small number of upper catchments. 4. Over the years a number of gove2rnment departments have undertaken programs aimed at developing the Kandi Tract and tackling some of the more severe problems. These have been carried out, however, without reference to other dependent or parallel problems, and their impact has been disappointing. One of the engineering proposals was to construct a number of dams across selected choes just within the hills to act as sediment and flood control /1 A choe is a seasonal sandy river, dry for most of the year.  - 2 - structures as well as providing irrigatio.i supplies. Another proposal was to p channelize all the major choes as a means ef minimizing damage caused by spate flows. Both these proposals gave rise to concern as to whether they might have an undesirable effect on the groundwater regime, which was already being placeI under some strain by the rapid increase in the number of private tubewells. 5. The concept of linking the proposed irrigation and flood control development with improvements in the hill watersheds was formed in consultation between the Government of Punjab and the World Bank. The primary objective of the project as appraised was to lay the basis for sound investment on a large scale to: (a) reverse man-made ecological degradation and (b) protect and develop agricultural land that was subject to serious erosion and flooding. The project therefore was to focus on institution- building, in particular strengthening GOP's capability to plan, implement and monitor similar projects in the sub-mountainous zone. It vas designed as a pilot project, comprising a package of complementary investments and was to introduce technical and social improvements that, through their replicability, would have an impact far beyond the project itself. The project was to select five watersheds out of a much larger number covering the entire Kandi trac. Project interventions includeds (a) rehabilitation of upper catchments, (b) flood protection and irrigation of farm lands below the hills, and *c) development of farm land protected from floods. The project was prepared by the concerned line departments in 1977/78. Feasibility-level planning was completed prior to appraisal by the GOP who had limited planning capability for only one of the five watersheds, leaving the remainder to be identified and planned during the first years of project implementation (para. 7). 6. The project was to be implemented jointly by the Department of Forestry (FD), Soil Conservation and Engineering (SCD), Irrigation (ID), Agriculture (AD), Animal Husbandry, Fisheries, and by the Punjab Agricultural University (PAU). Coordination was to be achieved at Secretary level through a Policy Review Committee and inter-departmental coordination was to be exercised by an Implementation Committee under She Financial Covuissioner vho would also appoint the Project Coordinator. B. Proiect Implementation General 7. With only one of five watershed development plans (Dholbaha) ready for implementation the foremost task after project effectiveness was to prepare similar development plans for the four other watersheds. Although GOP had alreody preliminarily identified these watersheds, the SAR rightly provided for feasibility level studies for a total of ten watersheds, in order to broaden the base from which to select four representative and feasible sub- projects. Reaching agreement on the need for assistance from international consultants in this planning exercise was difficult and prolonged. When a consultant team was eventutllv hired and feasibility reports prepared, it became clear that the potenti.- for project activities for catcbment rehabilitation, irrigation, flood control and farm development components was considerably less than estimated at appraisal. A major disappointment was the revelation that storage and flood attenuation dams with associated irrigation  - 3 - development similar to the Dholbaha model would be economically feasibly only in two of the ter. watersheds (Janauri and Haili) in addition to Dholbaha. a, nae reduced irrigation-cum-flood control development potential together with slow progress in the implementation of Dholbaha dam and irrigation system, provoked a reorientation of the project, as expressed in the mid-term review of December 1982. There the strategy was changed from the Dholbaha model, to groundwater development, water harvesting tanks and coverage of more waters'neds with corresponding upward revision of project components other than dam building. The Bank thus agreed to extend project activities from five to eleven watersheds, thereby accelerating the reafforestation and protection of the upper catchments and previding on-farm development and livestock development to other beneficiaries within the Kandi tact. 9. While at appraisal no groundwater development was included under the project, the findings of the consultant studies indicated some potential and, probably as a trade-off for the much reduced surface irrigation development, the Bank agreed to the installation of 15 tubewells of which 13 were actually implemented. Rehabilitation of Uoyer Catchments 10. Rehabilitation was composed of revegetation using indigenous as well as imported tree and shrub species, and grasses (Bhabbar and Vetiver). Since 1985 shrubs and grasses have been planted to form vegstative hedges (screens) for erosion control and moisture conservation. Cor=ercially remunerative species such as eucalyptus and teak and fodder trees were also planted on privately owned land. These vegetarive methods were reinforced through check dams, silt detention basins, anl closure to grazing. 11. In response to expansion of activities over an additional six watersheds the Forestry Department established new targets for rehabilitation substantially above SAR provisions. Yet, in the final balance, these targets were again exceeded by a comfortable margin (Table 2). It is noted, however, that areas quoted as rehabilitatcd are consistently larger than the corresponding watersheds, suggesting repeated rehabilitation efforts in certain areas, probably in response to high plant mortality due to continued grazing, drought, fire and other causes. It also reflects the u satisfactory maintenance provision of only two years, and indicates that closAre to grazing has been ineffective. Animal Husbandry 12. Under the livestock program nearly 3,000 head of improved buffalo were exchanged for low-yielding cows and buffaloes. Artificial insemination of over 15,000 cattle units is also reported achieved. However, the Animal Husbandry Department did not succeed in its ultimate objective of substantially reducing grazing in the upper watersheds. PAU estimates that the cattle population grazing in the upper watershed was reduced by only 15z. This failure can be ascribed to:  -4- (a) lack of fodder and feed provision which should have come from irrigated land as well as from fodder harvesting in the upper watersheds. This led ;o declining milk yields and consequently the sale of the improved cattle;2/ (b) lack of marketing facilities for surplus dairy production, and hence inadequate incentive for farmers to exchange their traditional animal husbandry for stall-feeding of improved animals; (c) lack of understanding by project authorities of the social and cultural factors inhibiting adoption of the proposed technology, and weakness of livestock extension staff in the Kandi area. Irrigation and Flood Control Works 13. Prior to the project the ID was represented in the Kandi tract by only a small tubewell section and a drainage division. For the implementation of the dams and related works the ID established a new construction division and a division for carrying out the hydrological and groundwater investigations. Both are being maintained in operation for completion of spill-over works as well as for planning and implementation of further dam projects. At the outset little experience with the construction of earth dams was available, which contributed .o a slow start. Also, ID had little experience with contractual work as irrigation and flood control works were generally carried out departmentally on force account. Meanwhile construction of a fourth dam. although entirely GOP-financed, is also being undertaken through a contractor. 14. Construction of Dholbaha dam and irrigation system was started in mid-1981 and was intended to be completed in two to three years. Instead. completion took over eight years, and most of the tertiary irrigation system is still not operational. The main cause of delay was the inadequate performance of the initial contractor, whose contract was eventually cancelled, land acquisition problems, delays in budgetary provision, inadequate project preparation and poor timing between dam raising and irrigation infrastructure were other causes as well as lack of staff continuity. An important factor in the non-completion of the tertiary system has been the omission of the AD services under the project, which were required in order to undertake, with farmer participation, the implementation of the farm distribution systems in the chaks 31 below the outlets from the ID implemented system. The uncertain law and order situation in the State over the last several years has certainly also contributed to slower progress than anticipated. 2/ It is estimated that at least 50? of cattle supplied under the exchange scheme were sold out of the project area within four years. 3/ Lowest order or sub-division of irrigation command, here about 8 ha.  15. Implementation experience with the Janauri and Maili sub-projects has by and large been encouragingly positive, although even here the implementation of the tertiary irrigation systems is lagging behind for the abovementioned reason. 16. The need for flood control works and corresponding provisions under the project had been greatly over-estimated at appraisal. The proposed 3diversion embankments" t'f 80 km length were entirely dropped from the implementation plans while *channelization' was also greatly reduced. Only some reinforcement of the left bank flood protection levee on Dholbaha Choe, with a total length of 5.2 km, was included. Flood protection works under the project on other choes have not been reported. 17. Implementation of the tubewell sub-component through the Punjab State Tubewell Corporation (PSTC), civering four watersheds, proceeded without major delays. However, at Maili only one of two wells was drilled and later this was abandoned because of insufficient yield. 18. The SAR rightly had made provision for a five-year program of extensive groundwater investigations, which was outlined to include the establishment of an observation network (wells plus stream runoff stations), recharge experiments, pumping tests, groundwater mapping, and related monitoring activities. The urgent need for such a program was reaffirmed by the consultant team in connection with their advocation of a hydrogeological master plan for the Kandi tract in particular, but also for the entire *8ist Doabl, between the Beas and the Sutlej river systems. The importance of such studies was subsequently reiterated by at least one supervision mission. However, implementation of the program was slow, due to lack of budget support, shortage of staff and general lack of commitment. It is quite obvious that the latter was in inverse proportion to the attention given to the promotion and support of dam building. The surface hydrology monitoring appears to have been geared to providing necessary data for further surface storage feasibility studies. To date no report on groundwater-related studies and interpretations has been prepared. Farm Development 19. In the land development sub-component implementation proceeded much in line with SAR expectations through conventional high-cost bench terracing and block levelling, with emphasis on Dholbaha watershed where nearly half of the total coverage of 10,800 ha (Table 2) was concentrated. The remainder was shared approximately evenly between the ten other watersheds. Low-cost vegetative methods and contour hedging for in-situ moisture conservation were only applied in the last two years of the project on a trial scale. 20. Fruit tree planting, predominantly citrus, was achieved on about 2,300 ha of which 90Z was irrigated land and over half was located outside the project area on more appropriate soils that were irrigated. Some plant spacing and fertilizer problems indicate lack of adequate extension coverage. 21. Two fish seed farms were established by the project and have substantially boosted fingerling production for statewide use, although stocking quantities in the project area have not reached expected levels.  - 6 - This was mainly due to reduced and delayed reservoir construction and lesser than expected numbers of village and f&rm ponds. Because of marketing problems 4/ most of the fish produced was marketed outside the Kandi tract, as far away as Calcutta. Fish harvesting was auctioned away and benefits seem to have largely by-passed the targetted beneficiaries. Agricultural Research 22. Under the project a new agricultural research station designed to generate region-specific technology was established in 1981 at Ballowal Saunkhri. With most of the planned infrastructure installed the 120 ha farm, which is operated by the PAU, is performing very useful work on all important aspects of rainfed cultivation. Activities initiated under the project include: (a) research on suitable field crops and green fodder crops; (b) horticultural experiments; (c) testing of tree and shrub species under conditions of the Kandi tract; (d) development of more remunerative farming and land management systems, including development of trees and tree products; (e) soil and moisture conservation and land development practices; and (f) water harvesting techniques. Experimental results indicate substantial scope for technological improvements but have not yet matured into replicable models and further work is thus required and deserves support under any future project. Training 23. Various training activities were undertaken under the provisions of the project. These includet (a) nine senior officers (1 from the Animal Husbandry Department, 2 from FD, 5 from ID and the Joint Development Commissioner) attended Colorado State University for 3-5 month courses; (b) during the project approximately 20 middle-level officers attended short courses at the Central Arid Zone Research Institute, Jodphur, and the Indian Fodder and Grassland Research Institute, Jhansi; (c) in-service training was given to junior FD staff at the Forest Training School, Hoshiarpur; and 4/ Local people dislike the bony carp variety which was predominantly stocked in the reservoirs.  (d) in the latter years of the project bi-annual farmer fairs were held at the Ballowal Research Station. Costs. Financing and Disbursement 24. Total project costs for the originally-envisaged implementation period of five years were estimated at Rs 510.5 million (US$60.8 million at the exchange rate prevailing at the time of appraisal). Of these costs over 50? were to be allocated to irrigation and flood control work. 25. Revised cost estimates (resulting from the extension of the project period to seven years (later extended to eight) and the extension of project scope from the rehabilitation of five to eleven watersheds) totalled Rs 598.8 million. The proportion of total costs allocated to irrigation/flood control components remained constant at over 502, even though it was already clear at the time of revieion that this component was failing to meet designated targets. 26. Actual project costs totalled Rs 595 million, very slightly below the revised estimated of Rs 599 million. Most components' expenditure slightly exceeded revised estimates, with the exception of animal husbandry, horticulture and irrigation. The irrigation component absorbed over 50? of total costs. It will be noted that had five dams been constructed, as originally envisaged, rather than thre-. there would have been a very considerable cost over-run. Cost details are given in Table 4. 27. Due to the low proportion of foreign exchange costs in relation to total costs (estimated at 9? at appraisal), it was considered appropriate for the Bank to assist in financing part of the project's local currency expenditure. Hence it was agreed that the Bank would finance some 45? of local costs in addition to all foreign exchange costs. Remaining financing requirements were to be met, according to the standard formula, 70? by the Government of India and 30? by the Government of Punjab. 28. The IBRD loan finally accounted for 54? of total project financing as opposed to the 49? envisaged at appraisal, in spite of the non-di3bursement of nearly US$4 million. This was because of the falling value of the rupee as against the US dollar. It should be noted that the figure of US$48.4 million for the total dollar cost of the project is only approximate, again because of the shifting exchange rate. Financing details are given in Table 5. 29. Disbursements were to be made for: (a) 100? of the cost of overseas training and consultancy services, including project evaluation undertaken by PAU; (b) soil conservation, land levelling and revegetation works in the upper catchments at the rate of Rs 2.300/ha; (c) 70? of other civil works expenditures; (d) 1002 of the cost of locally manufactured goods procured ex-factory or 702 of the cost of other locally procured goods,  - 8 - (e) 702 of the cost of cattle and sheep; and (f) 502 of the cost of salaries for incremental project staff. 30. These disbursement provisions were not altered when the project timetable and targets were revised. Details of planned and actual disbursement are given in Table 1. 31. Final disbursement from the IBRD loan was US$26.1 million, considerably less than the agreed US$30 million. This shortfall can be attributed to two factorst (a) Irrigation expenditure was Rs 24 million less than the revised estimates. At the exchange rates prevailing in the latter years of the project and with 702 IBRD financing it can be assumed therefore that less than estimated irrigation expenditure accounted for over US$1 million of the undisbursed amount; (b) In a context where foreign exchange costs were relatively low, the US dollar appreciated by nearly 90X in relation to the rupee over project lifetime. Largely because of tardy implementation over 502 of loan disbursement took place in the last three years of the project. C. Proiect Results Agricultural and Environmental Imnact 32. The primary objective of the project was to reverse, on a pilot scale, man-made ecological degradation in the Shiwalik hills. This, when replicated, would generate beneficial effects for the population fringing the hills, as well as for a much larger section of the farmer population in the plains south-west of the Kandi tract. The former were supposed to be the actual target group of beneficiaries under the project. Their temporary sacrifices through foregone grazing and other uses of the watersheds were to be compensated by profit-generating farm development programs as well as reforestation and afforestation in privately-owned land (902 of Shiwaliks). However, soliciting farmers' cooperation, with the limited set of incentives available was a formidable task, requiring strong department coordination/ cooperation. This was not achieved (para. 47). A major shortcoming was that proposed interventions were too dependently linked to irrigated rather than rainfed farming, and that irrigation development suffered both delays and reduction in scope. 33. It is worth noting that the prospect of dam construction not only strongly influenced watershed selection but also provided the rational for much intardepartmental coordination. This was particularly true in the case of the Forestry and Irrigation Departments because of the obvious priority need to stabilize those watersheds where dams were to be built. In fact the watersheds selected for dam constructions possessed a relatively dense vegetative cover already at project inception. There is some indication that FD woul. have given priority to more denuded watersheds, had it not been for the sake of lending support to the dam projects.  - 9 - 34. Nevertheless FD was able to undertake substantial rehabilitation work in all 11 watersheds. The long-term effect of these efforts will however greatly depend on the extent to which maintenance operations can be sustained. Severe financial resources constraints clearly point to the need for wider adoption of contour aligned vegetative screens (vetiver grass, shrub species) and related low-cost in-situ moisture conservation methods. 35. The dependence of dams on relatively stable w-tersheds does not appear to have been sufficiently analyzed. Since watershed stabilization has to be achieved to such an extent that siltation rates of the reservoirs are reduced to economically acceptable levels, it follows that the torrential, sediment-prone nature of floods would already have been moderated so substantially through rehabilitation that little room is left for justifying additional flood attenuation through dams. Apart from initial and running costs, the possible benefits of such dam storage ought to be weighed against the likely negative effects (e.g. evaporation) on the natura'. regime of flood water infiltration along the sand rivers (choes) as an effective means of groundwater recharge for irrigating the fertile plains. It therefore appears that watershed rehabilitation without intervening storage may constitute the ecologically most appropriate, technically least risky and economically most viable solution in the context of most watersheds of the Kandi tract. As far as gravity irrigation from dams is concerned the project has provided little evidence to suggest that it is economically viable on its own merits (paras. 38 and 39). The main thrust of future agricultural development on the marginal soils adjacent to the Shilwaliks therefore should be on rainfed farming and adapted animal husbandry, even at the price of heavy subsidy. The prospects for a moderate further groundwater development were considered quite good by the project's consultant team. However, the relevant studies provided for under the project have not yet yielded conclusive results (para. 18). 36. While the production-related objectives of the project have been only partially ach.eved, the conservation-oriented activities (para. 10) have been implemented with considerable success. On balance, therefore, the environmental, impact of the project has been highly positive, although scope for further improvement remains and continuous vigilance will be required. The Economic Rate of Return 37. The economic rate of return (ERR) estimated at appraisal was based on Dholbaha alone. On this basis the ERR was estimated at 16Z. A less optimistic model assuming few project benefits to the south of the Chandigarh- Hoshiarpur road yielded an ERR of 13Z. A separate analysis was undertaken, estimating the ERR on reforestation and soil conservation measures in the upper catchments--taken as a self-contained activity--as 12Z. 38. A more refined economic analysis was produced by the consultancy team engaged under the project to make feasibility studies of other watersheds. Apart from Dholbaha, which might be suitable for early rehabilitation, ERRs were calculated by major component and by watershed. For the forestry component the ERR averaged 182 over 10 watersheds. The ERRs for the cattle exchange program were exceedingly high. Irtigation and flood control with  - 10- dams were found to have ERRs of less than 71 in all watersheds except Maili and Janauri (201 and 132 respectively), while on the other hand deep tubewell irrigation was found to generate satisfactory ERRs in the majority of watersheds. On the basis of these findings dam construction under the project was thereafter confined to Dholbaha, Maili and Janauri. 39. The PCR estimate S/ of an overall project ERR of 82 therefore represents a substantial shortfall from earlier expectations. Separate ERRs have been estimated for each component, as follows: Forestry 12S Animal husbandry 9S Soil conservation 10l Horticulture 221 Irrigation 2S 40. The reasons for the poor performance of the irrigation component, notably delayed implementation and restricted area actually brought under irrigation, have already been discussed. The less than satisfactory performance of the soil conservation component is largely due to the failure of the irrigation component, since the full benefits of many of the soil conservation activities undertaken were dependent on the availability of timely irrigation. 41. The poor performance of the animal husbandry component, particularly in relation to estimates by the consultant team, is ascribed to the exodus of high-quality cattle provided under the project (only 601 of these animals were retained in the project area from the 4th year onwards) and to a sharp drop in predicted milk yields. Both these negative factors relate, in turn, to defective provision of fodder and feed, inadequate extension services, and lack of marketing facilities for surplus milk production. 42. The forestry component, based on a sample area evaluation, achieved the same ERR as estimated at appraisal. It is now acknowledged that results from this component could have been considerably better had adequate provision been made for maintenance, so that high tree mortality rates could have been significantly reduced. The occurrence of a severe drought in 1987 must also be considered. 43. The horticultural component yielded by far the highest ERR. Satisfaction at this result should, however, be moderated by the consideration that a very high proportion of horticultural activity took place on already irrigated land on the fringes of and outside the project area, and therefore contributed little towards the achievement of project objectives. 5/ The PCR estimate is largely based on a cost-benefit analysis undertaken by the evaluation unit at PAU and a supplementary evaluation undertaken at the Delhi Office of the Bank.  - 11 - 44. No ERR was calculated for the fisheries component since data on this component were scanty and fishery development played a minor role in the project as a whole. 45. It is relevant to make the following two observations in connection with procedures for calculating the ERR which appear to have considerable bearing on the design, implementation and outcome of the project: (a) Although the project was primarily addressed to the arrest of soil erosion and environmental deterioration, no coefficients of declining yields in a without-project situation were calculated. This was justified on the grounds that the assumption of no decline in yields would, although unrealistic, make the economic analysis more conservative and hence more robust. Furthermore, little reliable information was available on which to base estimates of declining yields. Finally, declining yields without project intervention would, in the long term, be most damaging outside the project area, that is in the plains below the Shivaliks, as a result of deterioration of water resources, sedimentation, etc. Nevertheless, as a partial consequence of this procedure, hcavy emphasis was laid on the achievement of economic benefits through production-oriented activities and the primary objective--to reverse man-made ecological degradation in the area--was to a large extent lost from view. Given the nature of the project and the importance of the externalities involved, the Bank's insistence at the time of project formulation that an ERR of not l8ss than 12Z should be achieved was perhaps inappropriate. If a more robust estimate of the opportunity cost of capital were accepted in this context (e.g. 8Z), the project's failure to achieve the predicted ERR would not be an indicator of overall failure, as long as the problr-' of ecological degradation had been tackled in any substantive way. (b) In the SAR ERRs were calculated separately for the irrigation and forestry components. Subsequently the consultants' report calculated ERRs separately for each component in each watershed, a procedure which was maintained at the component level throughout monitoring and evaluation. While this is a useful practice and can constitute a valuable check against excessive investment in only marginally beneficial activities, there is a danger that in a project where components are mutually interdependent it can promote a fragmented approach to project implementation. Indeed, this appears to have become the case in the Kandi project; the calculation of separate ERRs for each component reinforced a tendency for line departments to purs:te their separate agendas with little reference to other departments' plans or activities. Borrower's Performance 46. Overall, borrower performance was characterized by a high degree of technical competence. With the exception of irrigation, most line departments ultimately achieved the targets assigned to them, in spite of a slow start. As shown in Table 2 most departments came very close to their financial  - 12 - targets and the majority exceeded their physical targets, as expressed in crudely quantitative terms.61 47. The most disappointing feature of borrower performance was the failure to achieve significant coordination or even communication between departments, and thus the lack of an integrated or coherent approach towards project implementation. Individual departments undertook their separate agendas with varying degrees of energy and according to their own priorities. This resulted, in some cases, in inadequate participation in project activities; it is clear, for example, that the installation of deep tubewells in the Kandi area was not a high priority for the Punjab State Tubewell Corporation, although it is acknowledged that these could have been highly beneficial in many watersheds. 48. In other cases poor coordination between departments seriously reducea the effectiveness of project activities. As an example, the provision of fodder harvested from the upper catchments was essential to the success of the cattle exchange program's objective to substitute stall-feeding for uncontrolled grazing. Yet the FD merely auctioned fodder harvesting rights to the highest bidder, with the result that fodder from the upper catchments rarely found its way to the intended beneficiaries. Similarly, additional fodder was to be grown in the lower watersheds, yet AD contribution to this activity (as in other aspects of the project, with the exception of horticulture) appears to have been minimal. 49. In yet other cases, an individual department energetically fulfilled targets assigned to it, without significant reference to overall project objectives. So, for example, horticultural targets expressed in crude terms of hectares to be planted were over-achieved, but planting under rainfed conditions in the upper regions watershed were almost non-existent. Lack of interdepartmental coordination led, then, to a situation where project objectives were translated into department targets, which then came to obscure those very objectives. Furthermore, activities which were not clearly the responsibility of one single department were, in general, not initiated. So, for example, the consultants' report recommended that establishment of small- scale agro-industries in the upper watersheds as supplementary forms of income-generation, but no steps along these lines were taken. 50. in short, while the technical competence of the executing agencies was of high order, coordinating and planning capacity and the ability to respond flexibly to a developing situation were limited. In cannot be claimed that such capacity was significantly enhanced during project implementation. Although general monitoring and evaluation was undertaken satisfactorily by PAU, there appears to have been poor dialogue between PAU and the technical agencies involved, indicating a lack of monitoring and evaluation capability within the technical agencies themselves, as well as shortage of staff within the evaluation cell of PAU. Discontinuity of project personnel hampered the 6/ The physical target and achievement of the forestry component, for example, specifies trees to be planted and actually planted, but does not take mortality rates into account.  - 13 - development of planning and coordination capability. To take one example, there were reported to have been four different superintendent engineers at the Dholbaha dam over a 3-year period. 51. In gene.al, individual departments administered their budget allocation efficiently, although it was reported that within departments funds were not always channelled promptly down to field implementation level. Financing did not--with the exception of provision for forestry maintenance- -set a significant constraint on successful project implementation. The provision of separate budgets did, however, in addition to reinforcing department autonomies, inhibit the transfer of funds from underperforming components to components which could have been developed on a more ambitious scale. 52. In one sense, however, the financial climate prevailing in the Pubjab did contain successful project implementation. As part of a GOP austerity program, from 1985186 all government and civil service posts were frozen. This freeze included incremental project staff, even though 502 of incremental salaries were to be financed by the World Bank loan. Precise figures were not available on incremental staff to be assigned and actually assigned to the project by departments. Nevertheless, the acknowledged inadequacy of the extension effort in agriculture and animal husbandry may well have been due in a large measure to the freeze on addit_'onal staffing. Bank Performance 53. PrAect Desizn and Anpraisal. An integrated approach to the problems of environmental degradation in the Kandi tract was long overdue, and the Bank played a valuable role as a catalyst in setting in motion through this largely pilot project a large-scale remedial program which will take many years to complete. Many of the problems encountered in the course of project implementation can be attributed directly or indirectly to faulty design and appraisal. Design and appraisal features vhich appear, with hindsight, less than satisfactory are listed below: (a) a central weakness of project design, particularly in view of the predictable difficulties in interdepartmental coordination described in the preceding sections, was the lack of clear definition of different departments' roles and responsibilities. This lack of definition is eloquently illustrated in the confusing and elaborate organizational chart (SAR p. 93); (b) closely related to the above was the lack of any effective control mechanisms vested in the Joint Development Commissioner's (Project Coordination) office; since budgetary allocations were made directly to concerned departments no institutional or budgetary check was created against their inherent autonomist or even separatist tendencies; (c) preparation and appraisal were constrained by data availability, largely confined to the Dholbaha watershed. It was, perhaps, therefore inevitable that surface-water irrigation should have come to appear the dominant component of the project, although with  - 14 - hindsight it is clear that reforestation and stabilization of the upper catchments supported by farm development should have been given the highest prioritys (d) as already discussed, the techniques of economic appraisal employed tended to emphasize directly production-oriented activities at the expense of an integrated remedial program; (e) given the long-term nature of the activities to be undertaken, the slow-maturing nature of many of the benefits, the lack of institutional experience in this type of project and the paucity of data available at the planning stage, it would have been more realistic to have phased project activities over a longer period, planning for a seven-year project from the outset; (f) insufficient attention was given to potential marketing problems, In particular in respect to input supply (fodder for livestock) but also in respect to output (milk, and possibly also citrus and eucalyptus); (g) the assumption that existing levels of extension staffing were adequate to meet project needs proved unjustified; (h) only perfunctory attention was paid to socio-economic and income- distribution effects of project intervention. Likely winners and losers under the project were not clearly identified. Thus, as already mentioned, livestock owners who received fodder harvested from the upper catchments were not necessarily those who had previously grazed their cattle there. The livestock exchange scheme enabled cattled-owners to exchange two inferior animals for one improved animal. Owners of only one animal were thus excluded from the scheme. Planting material and credit at subsidized rates were provided under the horticultural component to farmer whose land was already irrigated. The value of lands to be irrigated increased sharply, without necessarily benefitting the local population. It is reported that some 10-15S of the most fertile land to be irrigated following construction of the Dholbaha dam was bought up by outsiders before land values began to rise. Many of these eocio-economic problems are admittedly, more associated with project implementation than with design. Nevertheless, a more formal socio-economic input at the appraisal stage could have predicted them and recommended appropriate policies;7/ and (i) finally, as already mentioned, provision for forestry upkeep and maintenance costs were inadequate. 54. Supervision. The mission was unable to locate formal supervision reports either in the Bank Delhi office or at project headquarters for the period up to June 1983. It would appear that up to that date supervisions were conducted on a relatively informal one-man basis, with observations and recommendations presented in the form of memoranda. Thenceforth, however, 7/ It should be noted that the post of sociologist at the Joint Development Commissioner's office provided for under the project was never filled.  - 1S - supervisions were undertaken more rigorously and formal supervision reports were presented. Insufficient attention was paid however to verifying the suitability of technology adopted for the livestock sub-component. 55. The Bank's decision to limit dam construction under the project to three watersheds rather than five and to encourage greater emphasis on vegetative methods of soil conservation and in situ moisture conservation was, in the view of the PCR mission, correct and in keeping with the consultancy team's recommendations. It was also consistent with the SAR's proviso (p. 23) that dam construction would only take place where ERas of not less than 122 were foreseen. Nevertheless, it should be noted that this decision was perceived by many involved in project implementation as a capricious change of policy. It would seem that on this point communication between the Bank and project authorities, in other respects both cordial and effective, was not completely satisfactory. D. Conclusions. Lessons Learned avi Issues 56. Although shortcomings in project design and implementation are amply documented in the preceding pages, as well as implicitly summarized in the following section, "Lessons Learned', it should be recalled that the project was conceived as being pilot and therefore largely experimental in nature. Given the complerity, scale and necessarily long-term character of any systematic program to halt or reverse man-made degradation of the Shilwalik range and the lack of experience in planning and implementing multi- disciplinary interventions of the type required, lessons learned or learnable from the Kandi project can be seen as positive project benefits. The candor with which those involved are prepared to acknowledge planning and implementation errors augurs well for the success of future interventions in the Shiwaliks. 57. Rehabilitation of the Shiwaliks requires a very considerable shift in the conceptual framework that has dominated agricultural development in the Punjab over the last two decades--a shift from the belief that technical solutions devised at departmental level are necessarily the most appropriate, a shift from the emphasis on capital-intensive irrigation and earth-moving works towards more emphasis on improvements to rainfed agriculture and, finally, a shift from the belief that effective interventions of an environmentally conservationist or remedial nature must be justifiable by short-term production targets alone. It would have been unrealistic to expect more than the beginnings of such a shift during the course of this single project. Nevertheless, it achieved a significant first step in reversing the long-estrblished mismanagement of the Shiwaliks Lessons Learned 58. As already mentioned, one of the benefits of the project was that it generated a great number of lessons highly relevant for any further interventions of this type. The principal ones are summarized below. (a) Ortanization, Management and Financina. The core problem throughout project implementation was the failure of the different agencies involved to coordinate their plans or their programs or even,  - 16 - frequently, to communicat%. effectively. The Joint Development Commissioner's office, charged with the role of coordinating project activities, did not posses the institutional or financial power necessary to do so; it was not, for example, automatically given access to departmental files. It is widely acknowledged that in future projects of this type, "one-window" financing (in which the project coordinating unit is responsible for disbursement to departments) would enhance the status of coordination and reduce the tendency of departments to pursue separate agendas. However, this may have its price in increased bureaucracy and some loss of initiativelenthusiasm in the department. (b) General Proiect Design. The long-term nature of the activities embarked upon by the project, the necessity of feasibility studies as preliminary to many project activities and the need to phase out interdependent activities mean that a project such as this cannot be effectively compressed into a five-year span. A project period of 7-8 years, with a mid -term review, would be more appropriate. Departmental targets frequently come to obscure overall project objectives. These objectives should be more clearly defined at project formulation and kept in view throughout implementation. A strengthened coordination unit would assist this. Insufficient attention was focussed during both formulation and implementation on project beneficiaries and losers. A project of this nature, requiring radical departures from traditional rural practice (e.g. livestock grazing in upper watersheds), demands a more formal and sustained socio-economic input than was made. Little thought appears to have been given to small-scale agro-industry which might offer an alternative source oi ircome for farmers who traditionall) grazed their cattle in the upper watersheds. The viability of such activities should be assessed before implementation of further projects of this kind. (c) Irriaation and Flood Control. This project ccmponent has clearly. been the least successful, absorbing over 50Z of project costs in exchange for less than 102 of benefits. In addition to problems arising from delays in dam construction, it appears that the benefits to be derived from surface water irrigation were overestimated, and that the damage caused by floods was overestimated too, so that the benefits to be derived from flood attenuation were also correspondingly overestimated. Accordingly, in future projects a key role should not automatically be assigned to surface water irrigation and flood control. Instead, a more pragmatic case by case approach should be adopted, working from the assumption that, unless firmly indicated to the contrary, the soundest procedure would be to: (i) attenuate floods through watershed rehabilitation; (ii) recharge as much as possible of the small and medium floods to the groundwater along the sand beds of the choes, and;  - 17 - (iii) exploit the water resource thus harnessed through groundwater development and rainfed cultivation in and around the Shilvaliks. (d) Forestry. It follows from the above that in-situ moisture conservation as the key to successful rainfed farming should be the central tactic in future interventions in the area, and therefore that reforestation and afforestation and revegetation of the upper watersheds should be given the highest priority. This could in favorable cases include water harvesting tanks. Although significant activity took place under the forestry component, tree planting was largely concentrated in watersheds were dams were to be built. These were also the watersheds where vegetation cover was already fairly satisfactory. In future projects efforts could be directed towards the more severely denuded watersheds, where the marginal benefits from investment are likely to be more significant. The impact of forestry activities under the Kandi project was limited by the high mortality rate of trees planted. It is now appreciated that in future projects maintenance costs should be more realistically estimated both in terms of magnitude and duration. It should be recalled that 902 of forest land in the Kandi tract is under private ownership. The Forestry Department responded with commendable pragmatism to this situation by balancing long-term conservationist plantings with the planting of economic species (e.g. eucalyptus) offering a relatively early financial return to farmers. Future projects in the Shiwaliks should bear in mind the necessity of some such trade-off. (e) Soil Conservation. As in the case of forestry, soil conservation activities were excessively linked, particularly in the lower watersheds, to irrigated farming conditions. Thus, the emphasis was on conventional high-cost methods of bench terracing and block levelling. It should be recognized that agricultural production in most areas of the Shiwaliks will remain rainfed and, therefore, that there is a distinct need for large-scale application of low-cost vegetative methods of soil conservation and contour hedging. "Modern approaches' based on protection of soil surface from rain-splash and improvement of internal drainage should also be applied. (f) Livestock Development. It is clear that a significant reduction in the numbers of livestock grazing on the Shiwaliks is an essential condition for improvement of their vegetative cover. The PCR mission does not, however, share the view that more rigorous enforcement of the grazing ban would be the solution to the problem. Such policing would merely aggravate social tensions, while the ban would continue to be flouted as long as alternative sources of low-cost fodder (and fuel) were not forthcoming. It would therefore appear that any further projects in the Shilwaliks would need to include some livestock component along the lines of that attempted in the Kandi project, but it would need to be reinforced by: (i) improved fodder and feed supply and distribution, based mainly on rainfed production;  - 18 - (ii) inclusion within the scheme of poorer livestock owners and non. landowners: (iii) strengthened facilities for marketing surplus production; and (iv) a vigorous educational program geared to farmers' socio-economic and cultural conditions and accompanied by a more effective extension message. (g) Farm Development and Extension. Project achievements in establishing citrus orchards contributed little to development either of the target area or the welfare of target farmers. Very little was achieved under the project in terms of farm development under rainfed conditions. In this context the total absence of women extension officers in the area is to be noted. This is particularly damaging where off-farm employment is available to men and consequently farms are managed by women. How far the livestock component was affected by this could not be verified. (h) Fisheries. Fishery development was of little relevance to the Kandi project, largely because of the lower than expected level of reservoir and farm pond construction. This should not automatically be taken to suggest that fishery development should automatically be excluded as a useful component of any future project in the Shiwaliks. Such a component should, however, not be dependent on dam construction under the ID, and careful study of marketing constraints would be required. 'i) Research. The establishment of the PAU research station at Ballowal was, in the view of the PCR mission, an outstanding project success. Progress in determining appropriate technologies for rainfed agriculture in the region is very promising, and will ultimately be beneficial to a much larger area than the Kandi tract alone. Future interventions in the Shiwaliks should sustain and strengthen ongoing research activities at Ballowal (para. 22). Suggested improvements for the future include: (i) better infrastructure (in particular access) to ensure the continued commitment of high-calibre personnel; (ii) imoroving the technology of vegetation establishment and maintenance for soil and water conservation; (iii) attachment of a sociologist to identify factors constraining the adoption of new technologies and to devise appropriate strategies for overcoming these constraints; and (iv) strengthening links with both extension staff and the farming population. (j) Monitoring and Evaluation. Project evaluation was very competently carried out by the evaluatien cell at PAU, Ludhiana. Input into the project during implementation was, however, limited as a result of  - 19 - lack of staff and poor communication with and between line departments. It is, again, to be hoped that improved coordination would remedy this situation in any future project. It should also be stressed that the existence of a project evaluation cell should not be taken as a reason for line departments to neglect their own technical monitoring function. Issues 59. Under this heading we refer to lessons derived from the Kandi project which could be but have not entirely been learned, or to areas where the nature of the lesson to be learned is not yet clear. (a) The Watershed as Planning Unit. The consultants' report made the point that the watershed did not constitute an entirely satisfactory unit as a basis for planning, a point taken up by the PAU evaluation report. This is because in the watershtd, -s a purely physical unit--and one that in the Shivaliks, furthermore, is frequently not clearly delineated--cuts across more widely recognized social and administrative boundaries. It is thus suggested that the village or group of villages would constitute a more realistic unit on which to base planning and future interventions. Such a strategy would have the additional advantage of encouraging active participation by affected communities (something notably lacking hitherto). There is little evidence that detailed study of how such a strategy might operate is being undertaken. The consultants' report also noted that watershed planning and interventions were being undertaken in the absence of a master plan for the Kandi tract. Future interventions in the Shiwaliks should be carried out in the context of an overall long-term strategy for resource management and investment in the area as a whole, taking into account physical and human resources, tourist and industrial potential. (b) Continuity of Proiect Personnel. The planning and implementation of the Kandi Project was harmed to some extent within the coordinating unit, but in particular at departmentai level, by frequent changes of staff involved with the project. If project planning and implementation capacity is to be strengthened it is important to devise a method of attracting and retaining high-calibre staff, without jeopardizing their career prospects and position within their 'mother departments. (c) The Economic Rate of Return as an Investment Criterion in the Shiwaliks. As already argued (para. 45), inflexible use of the ERR as measure of investment profitability is not appropriate in projects such as the Kandi one. This is because no systematic methodology has been elaborated for incorporating into the economic analysis declining yields and other costs of non-intervention which are likely to be incurred outside as well as inside the project area. If the ERR is to be used in appraising projects of this type it would be advisable to estimate the opportunity cost of capital at significantly less than 12X.  - 20 - (d) Subsidies and Cost Recovery. Under the Kandi project policies regarding subsidies to and cost recovery from beneficiaries were in line with policies applying in the rest of the Pubjab. These correspond, in general, to conditions of high-yielding irrigated agriculture. The extent to which such policies are appropriate in rainfed and relatively unfavorable conditions needs to be reviewed.  - 21 - PART II - PROJECT REVIEW FROM BORROWER'S PERSPECTIVE Operational Perform ce 1. Physical and financial achievements of the components indicate that the operational performance of the project is quite satisfactory. However, there are short-comings which need to be improved upon in a follow up. The lessons learnt from the implementation of the Project and the shortcomings would be noted for future guidance. More emphasis would be laid on soil moisture conservation in-situ and as such, further research development in rainfed practices for agriculture, horticulture and forests. Obviously, agriculture is to play a key dominant role during a second phase project and as such, has been included in the proposals. Proiect Justification and Obiective 2. In the light of circumstances, the assessment of components and their priorities and problems are relevant to the project and are appropriate. However, a subsequent change in the reinforcement of the irrigation component (construction of dams) has affected this component in respect of validity of its original emphasis. The reasons for it are technical. Proiect Contents and Scheduling 3. The size, scheduling and complexities of the project were appropriate and balanced. It is obvious from the results under all the components involved. Proiect ImDlementation and Ol'eratina Outcomes 4. Project Design i.e. integrated multidisciplinary, interdependent, multifarious development approach has proved quite successful in achieving the ultimate goal of scoio-economic development of the Kandi Tract, the most backward area of the State of Punjab. 5. The implementation capacity of the executing agencies has been strengthened to their entire satisfaction. There is a little required for strengthening the esecuting agencies under a second phase project. 6. Scheduling of Civil Works and procurement do not tally with the time schedule proposed. It needs some structural changes so that works (dams and civil works) must be completed as per schedule and the procurement is also to the desired extent. 7. The implementation of schedules of the irrigation component were lagging behind due to some administrative/policy decisions by the lending agency. 8. Operating forecasts proved to be reasonable to a greater extent. The responsibility of the participating components though specifically laid out in the SAR of the Bank, but the role of Forestry and Animal Husbandry components was not clear. This resulted in some confusion concerning development of  - 22 - fodder for the livestock. Similarly, planting of vegetative hedges along the contour, etc., to check soil erosion, surface run-off, etc., were not clearly spelled out and aS such, these practices were not carried out adequately by the components of Forestry and Soils. However, the planting of vegetative hedges--Vetiver Grass (Vetiveria zizanioides) was resorted to by these components on the recommendations of the World Bank supervision mission visits to the area at a later stage. As such, most of the shortfall were remedied and in this connection visits of the World Bank Review Mission were quite helpful. There has been a major shortfall in case of irrigation component where dam construction activity was not done completely in accordance with earlier proposals. Irrigation component was to construct 10 dams whereas, only three have been constructed. In case of irrigation, it cannot be said that the shortfall occurred due to inadequate provision or specifications. Part B of Schedule 2 of the Loan Agreement (IN-1897) is very clear on this issue and reproduced hereundert Part Bs Flood protection, dams, irrigation works and construction and remodeling of bridges, to control and utilize run-off in Dholbaha and about four other watersheds to be selected; The shortfalls in case of irrigation as such, could not be remedied even at a later stage even in spite of administrative sanctions given by the GOP for construction. 9. The role of external aid agency--in the present case--financed with external assistance of IBRD (World Bank) has been laudable. But for the Bank's assistance and guidance--this sort of a Project and program implementation may not have become possible. From time to time World Bank Supervision Mission has been visiting the Project area for periodic reviews of Project implementation. The visits have been as a matter of fact biannual (six monthly) and were generally conducted every year in the months of March and September. The timings of the reviews were very appropriate keeping in view the climatic conditions of the region and such reviews were very helpful in the achievement of Project goals expeditiously. Reviewing staff has been found to be much experienced and familiar wita the Project. The reporting requirements and the reports were quite reasonable and useful to the executing agency, as already stated above, and also to the Government. Evaluation of Project Implementation and Outcome 10. The Project implementation has been satisfactory and the targets as envisaged at the planning stage, were accomplished. Now there is no question of any degree of uncertainty since the PCR is prepared after the closure of Project on March 31, 1988. The benefit cost ratio and the financial internal rate of return of the Project (overall--li watersheds), as assessed by the PAU, has been 1.70 and 16.73Z respectively (excluding Irrigation and Fisheries). Even though the role of irrigation has been controversial, but the construction of dams in Dholbaha, Janauri and Maili watersheds brought about a visible change in the area, not by way of merely as big structures, but by way of facilitating conservation of water for irrigation purposes, drinking, enriching flora and fauna, flood attenuation, containing seasonal hill torrents. Thus, as was originally envisaged, irrigation has a major role to play to check flood onslaught in the area by not only containing water in  - 23 - the larger watersheds, but also to take flood protection measures along riversibigger choes, etc., and providing improved drainage facilities and thereby improving In the drainage system as well. The conservation of water is all the more vital keeping in view of the fact that the rains in the region are seasonal and occur only during two months in a year's time--average mean rainfall varying 900 to 1,100 mm. To check the havoc of floods construction of a few bigger dams is necessitated, specially where the upper catchments are well stabil. zed. 11. Moreover, as has been discussed in the Feasibility Reports prepared by Messrs. Halcrow--MacDonald Consultancy Team and others, the benefits of dams constructed by the irrigation component, are to accrue at a later stage when the irrigation channel system is fully developed, people trained and water for irrigation properly channelized. During the last two years there were unprecedented drought in the history of the country and not in Punjab alone and as such, water could not be collected te full capacity in the built- up reservoirs of the three dams for irrigation purposes, etc. This obviously resulted in lesser irrigation facilities in the command area--and hence, apparently the causative factor of lesser rates of return by the irrigation component. 12. Due to non-construction of dams, the role of Fisheries also remained restricted. Fish fingerlings were to be released in the built-up dams--which could not take place due to obvious reasons--non-construction of dams. However, it is anticipated that fisheries have still a role to play by releasing fish fingerling in the water harvesting tanks (IWHTs) constructed by the soils component and in other village ponds. But, unfortunately fisheries component has also been excluded in the future proposed Project. 13. At this stage it cannot be said that the Project was not well enough identified or prepared because the formal start of the Project in July, 1980 was marked by the arrival of small team of international consultants, drawn from the two Firms of Sir M. MacDonald and Partners, and Sir. W. Halcrow and Partners, at the behest of the Bank, to assist the GOP during the first two years to prepare integrated development plans and feasibility reports for 10 selected watersheds. Eventually the international consultants did prepare the feasibility reports of the watersheds undertaken for development and construction of dams in 10 out of 11 watersheds was prescribed. Messrs. Halcrow-MacDonald also prepared a Final Report in 1983 for the on-going program of implementation of the Project. As regards shifting of the Project- -too quickly to easier goals, is concerned, it is mentioned that at much belated stage, construction of dams by irrigation component was not allowed and this resulted consequently in non-utilization of loan money (US dollars) whereas a major part of the proceeds of the loan (US$30 million) was attributable to civil works (Dams) as per Schedule I of the Loan Agreement. The unspent amount of the loan money could not have been utilized by the other components because of their lesser role. For instance, forestry component could not have covered more than 30,000 ha through vegetation-revegetation and soil conservation programs and they did achieve this figure of about 30,000 ha during the Project period. The que'tion of such a shift whether it was correct or not, could be assessed in view of the position already explained. For maximizing the benefits of the Project in future, there is no denial of the fact that the additional training (overseas), improving supplementary  - 24 - services, better maintenance standards, progress monitoring followed with physical checks, reviews, time-bound framing of program implementation, are very essential and key to success. For early and time-bound program implementation, it is all the more necessary that procurement of staff and vehicles should be provided at the very commencement of the Project in order to achieve better results. Without staff and vehicles particularly, no physical monitoring is possible in such a sub-mountainous terrain of Shiwaliks. 14. In the last, needless to mention that the Project area is nearly 4,600 sq kms and during implementation of the Project, only 78,000 ha of the area has been tackled. There are nearly 58 major watersheds which have been identified whereas only 11 watersheds have been treated so far during implementation and as such, continuation of the Project is very essential in order to reverse ecological degradatLon, stabilization of upper catcbments, prevention of floods and in elevating soclo-economic status of the people of this backward area of the Punjab State.  - 25 - Table 1 INDIA RANDI WATERSHED AND AREA DEVELOPMENT PROJECT (PUNJAB) (LOAN 1897-73) PART III - STATISTICAL INfORMATION Project Identity Project Names Kandi Watershed and Area Development Project (Punjab) Loan Numbers 1897-IN Country: India Sectors Irrigation Original Loan Amount: US$30 million Cancellationss US$3.9 million Net Loan Amountt US$26.1 million Proiect Timetable Date of Appraisal Missiont May/June 1978 Date of Board Approval: 20 July 1980 Date of Loan Negotiations: October 1978 Date of Signatures 12 September 1980 Date of Effectiveness: 18 November 1980 Date of Mid-Term Review: December 1982 Planned Credit Closing Date: 31 March 1986 Actual Credit Closing Date: 31 March 1988 Planned Date of Completion: September 1986 Actual Date of Completion: September 1988 Planned Credit Completion: September 2000 Cumulative Annual Disbursement of Loan Fiscal Year: 1981 1982 1983 1984 1985 1986 1987 1988 ------------------(US$ million)-------------------- Appraisal Estimate 1.8 6.8 14.4 22.4 30.0 30.0 30.0 30.0 Actual 1.0 2.6 6.3 9.3 12.7 16.6 21.1 26.1 Actual as Z of appraisal Estimate 56 38 44 42 42 55 70 87 Disbursement by Category Planned 1/ Actual Soil Conservation and Revegetation 3.7 7.1 Civil Works 16.4 15.1 Well Drilling. etc. 0.7 Equipment and Vehicles 1.3 0.4 Cattle and Sheep 2.9 0.8 Consultants and Overseas Training 2.8 1.4 Salaries for Incremental Project Staff 2.9 0.6 U3disbursed 3.9 1/ Figures under planned disbursement categories represent very approximate estimates. This is because categories of expenditures used at appraisal were not those used during implementation.  Protect 1w5.mtatton - hw%icai aot au IO vised for Emi f . J AcpulJ