India - Second Maharashtra Irrigation Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 7580 PROJECT CONPLETION REPORT INDIA MAHARASHTRA IRRIGATION II PROJECT (CR_DIT 954-IN) DECEMBER 30, 1988 Asia Regional Office This document has a resticted disbodon and may be used by recipients only In the performance of their official dudes. Its contents may not oterwise be disclosed without Wodd Bank authorizaton.  CURRENCY EQUIVALENTS Name of Currency Rupee (Rs) Appraisal Year Average (1979) USS 1.00 = Rs 8.076 Intervening Year Average (1980-84) US$ 1.00 - Rs 9.730 Completion Year Average (1985) US$ 1.00 Rs 12.237 FISCAL YEAR OF BORROWER April 1 - March 31 WEIGHTS AND MEASURES Metric System  FOR OMCIAL USE ONLY 7liE WORLD BANK Washington. D.C. 20433 U.S.A. OMiCe of DetKc"Wal Opeifatfts E[vaiation December 30, 1988 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on India Maharashtra Irrigation II Proiect (Credit 954-IN) Attached, for information, is a copy of a report entitled "Project Completion Report on India Maharashtra Irrigation II Project (Credit 954-IN)" prepared by the FAOIWorld Bank Cooperative Programme, with an overview memorandum prepared by the Asia Regional Office. Full evaluation of this project has not been made by the Operations Evaluation Department. Attachment T This document has a restricted distribution and may be used by recipients only in the perfonnance | of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  FOR OMCLIL USE ONLY ABBREVIATIONS AND ACRONYMS AD - Agricultural Department AEO - Agricultural Extension Officer ARDC - Agricultural Refinance and Development Corporation CAD - Command Area Development CADA - Command Area Development Authority CCA - Cultivable Command Area CCF - Constructive Conversion Factor CDO - Central Design Organization of GOM DCA - Development Credit Agreement ERR - Economic Rate Of Return FCF - Foodgrain Conversicn Factor GOI - Government of India GON - Government of Maharashtra HD - Health Department HYV - High Yielding Varieties ICA - Irrigable Command Area (see Glossary) ICB - International Competitive Bidding ID - Irrigation Department IFAD - International Fund for Agricultural Development LCB - Local Competitive Bidding LDA - Land Development Agency N & E - Monitoring and Evaluation MPAU - Mahatma Phole Agricultural University at Rahuri MWUP - Maharashtra Water Utilization Project OER - Official Exchange Rate OFD - On-Farm Development (see Glossary) O&M - Operation and Maintenance PA - Project Agreement PW & HD - Public Works and Housing Department QCU - Quality Control Unit RWS - Rotational Water Supply SCF - Standard Conversion Factor SSC - State Seeds Corporation T&V - Training and Visit System of Agriculture.l Extens4"n VLW - Village Level Worker WAPCOS - Water and Power Consultants WMU - Water Management Unit This document has a restricted distribution and may be used by recipients only in the perfonnance of their officlal duties. Its contents may not otherwise be disclosed without World Bank authorifation.  GJ.OSSARY Branch and - Irrigation canal with capacity more than 0.7 m3ts Distributary originat'ng at the main canal Canal Black Cotton - Heavy black clay soil, derived from basaltic rocks (BC) or Vertisol Chak - Block or irrigation service area below the Government outlet Field Channel - Irrigation channel from the lovernment (chak) outlet to individual farms (with lefs - 28 llsec - discharge) Irrigable - Equal to net Cultivable Command Area (NCCA) Commavd Area (ICA) Kharif - Wet Season (June to October) Main Canal - Main conveyance systems originating at the water source (dam, river) Minor _ Irrigation canal with capacity less than 0.7 m3/s and serving group of chaks Murrum - Granular material derived from weathered basalt formations On-farm - Small-scale works needed for efficient water development distribution and soil moisture control at the chak (OFD) level (e.g. field channels and dividers, drains and landshaping). OFD can be with or without land-shaping Outlet - Outlet. discharging from minor to 15 to 25 ha block, designed to take full field channel discharge and be either open or shut Rabi - Winter Season (November to March) Rigid Shejpali - Allocation of water according to Shejpali with delivery at predetermined turns Shejpali - Rotational Water Supply used in Maharashtra on sanction of water by crops, and irrigation from the tail-enders upstream Warabandi - Rotational Water Supply (RWS) used in northern India and based on allocation of water in proportion to the size of the landholding and delivery at predetermined turns  PROJECT COMPLETION REPORT INDIA MAHARASHTRA IRRIGATION II PROJECT (CREDIT 954-IN) 1ARLE OF CONTENTS Pate No. Preface ........................ ..... Co.. * .* * * ** * * ** .......... #. .*.... i Basic Data Sheet ......00.. **..*o......ii Evaluation Sumnary ......... *.***.........*.**....O..*.... {iii Overview ........... 0.....- 6* ..... 0 vii PROJECT COMPLETION REPORT I. Introduction ..... 1 ... ... 0... II. Project Identification, Preparation and Appraisal .... 2 III. Implementation ... *.*....... ...... 3 IV. Institutional Performance and Development ............ 13 V. Project Impact .................................... 17 VII. Economic Re-evaluation ..... ..... ... . ....... 20 VII. Bank Performance .......... **.*. ........... $0 ........ 23 VIII. Conclusions and Lessons Learned ... 24 Tables: 1. Implementation Status for Girna and Mula Schemes 2. Comparison of Main Features of Physical Components 3. ICB Procured Machirery 4. Project's Annual Expenditures per Scheme 5. Un'.t Cost Rates of Earth Works, Concrete and Rasonry over Period 1977 -78 to 1985-86 6. Allocation of IDA Credit 7. Schedule of Disbursements 8. Purchase of Machinery and Equipment under ICB 9. Annual 0 & M Expenditures and Water Charges 10. Summary - Total Cropped Area, Cropping and Irrigation Intensities, Farm Income 11. Cropping and Irrigation Intensities by Crops - SAR and PCR Comparison 12. Crop Yields by Scheme 13. Agricultural Production at Full Development by Scheme 14. Annual Inc--mental Agricultural Production at Full Development - SAR, PCR Comparison 15. Economic Input Costs and Farm-Gate Prices 16. ERR - Krishna 17. ERR - Warna 18. ERR - Kukadi 19. ERR - Upper Penganga 20. ERR - Upper Wardha 21. Total Project - ALT. I 22. Total Project - ALT. II  Figures: 1. Implementation Schedule - Krishna 2. Implementation Schedvl.e - Warna 3. Implementation Schedule - Kukadi 4. Implementation Schedule - Upper Penganga 5. Implementation Schedule - Upper Wardha ATTACHMENT I: Borrower's Comments  - i - PROJECT COMPLETION REPORT INDIA MAHARASHTRA IRRIGATION II PROJECT (CREDIT 954-IN) PREFACE This Project Completion Report (PCR) reviews implementation of the Maharashtra Irrigation II Project in India, for which Credit 954-IN in the amount of US$ 210 million was approved in October 1979. The credit was closed on 31 December 1985 as originally scheduled; with the final disbursements in early 1986 for expenses incurred during the project period, the Credit was fully disbursed. The PCR was prepared by the FAO/CP and was revised slightly after consultation with Asia Region staff. The evaluation summary and IDA Overview prepared by Asia Region staff, is based on the PCR, the Staff Appraisal Report (No. 2529a-IN, dated 13 September 1979), President's Report (No. P- 2625-IN of the same date), Development Credit Agreement, dated 14 April 1980, Bank supervision mission reports, on field visits to two sub-projects, and on reviews of Maharashtran irrigation by FAO/CP and Bank staff during the 'reformulation' missions for Maharashtra Irrigation III (MCIP III) Project. MCIP III is a follow-on to MCIP I not to this project. There has been no follow-on project to date. In accordance with the revised procedures for project performance reporting, this PCR was read by the Operations Evaluation Department (OED) but the project has not yet been audited by OED staff. A copy of the draft report was sent to the Government of India for comments on October 26, 1988. Those comments have been appended as Attachment I.  IMOIA - MAHARASNTRA IRRIGAtION It PROJECT (CR.954.1.1) PROJCT COMPLETION REPORT . OASIC DATA SHEET !x.Pro 1ct Oat Aooral%a1 F Actual as tot Project Cost (USS million) 1/ 452.9 *96.19 iI1 Credit Amount (USS million) Z/ 710.0 210.00 100 Oste of Effectiveness 10/1979 6/1980 Oats Board Aoproval 10/1979 Otte Ph s ca' Conponents 9/1984 1990 Completem Closing Octe * 12/198S l2/198a Economic Rate of Return (S) IS c 1.41A/ (Weighted Average 5 Sli Nem Schemes) If Excluding 'iAO Cradit - USS50 million. jI Source: SAR Vol. Ilt: SAR Vol. I states USS451.0 million. According to Credit Agreement. t Figure In reackets gives a second invesUent optien, explained tn Chapter VI. Cumulative itsbursement Fiscal Years: FY80 FY81 FY82 FY83 FY84 FY85 FY86 Appraisal Estimate S.0 36.0 84.0 :5,00 198.0 - (USS million) Revised CUSS million) . _ lSB.O 195.0 210.0 Actual (USS million) 8.8 35.9 75.3 119.5 162.9 191.7 210.0 Actual as S of aporal. 176 100 90 g0 80 93 100 rtvised estimate Mission .ata Mission 5 n ate No. Man- SciC' - frr 4ncRatin - ( (nth) dir- ays sation f c/ d/ - year) i in Status Trend tio Preparation' - Appraisal HQ Feb. 79 10 supev. I MOO Apr. 90 1 12 I 1 I - superv. ti "DO mar. 81 3 29 1.1.1 1 2 T.,HF Superv. 111 MOO Oc. 81 2 19 1.1 2 1 T.M.F Superv. IV NOO May. 82 2 21 III 2 1 T,M Superv. V NOO Nov. 82 2 32 1.1 2 2 T,M Superv. VI MOO May 83 2 24 1.1 2 2 TIM Superv. Vtt MOO Sep. 83 5 51 I I.I.A.A 2 2 TIM Superv. Vill 00 Mar. 84 5 27 1I1,1.R,A 2 2 F." Superv. IX MOO Sep. 84 4 303 i,,i.R 2 2 F.M Superv. X MOO Mar. 85 5 27 I1.1,,lRA 2 2 FIN Superv. XI MOO Nov. 85 6 40 I,II,I,R,A 2 2 F.M4 v PCR FAO/CP Fa. 87 Prepared by GO In cooperation with IAPCOS. WO and FAO/CP In 1977/79. / t * Irrigation Engineer% F * Financial Analyst; E a Economist; R * Roads Engineer; A a AgriCulturiSt. I/ 1 * Problem free or minor probims; 2 a *odorate problems; 3 * rjor problems. I/ 1 * Iproving 2 * stationary; 3 * deteriorating. FJ f * financal; N * managerial; T * technical; P * political. / New format Introduced, rating adjusted to previous system. Other Project UA lorrower: Governeent of India Eaecuting Agency: Goveruwant of Pbharasntra Fiscal Veer: April 1 - March 31 Mame of Currency: Rupee (As) Currency ExchaMe Rate: 179/80 USSl.OO * 8.076 1980/81 * * 7.893 1981/82 8 8.929 1982/83 * 9.628 1983/84 - *10.312 1984/85 11.887 1985/86 1 12.237  - {ii - PROJECT CONPLETIO0N REPORT INDIA HAIARASHTRA IRRIGATION II PROJECT (CREDIT 954-IN) EVALUATION SUMMARY i. This project was defined as the 1980-85 "time slice" of construction of five already-started irrigation sub-projects (Krishna, Warna, Kukadi, Upper Penganga & Upper Wardha) (97Z of base cost) and certain modernization works on two existing sub- projects (Girna & Mula)(3Z of base cost.. The "time slice" of an eighth already-started sub-project, Bhima, wa appraised with the other seven but financed by IFAD; it is the subject of a separate PCR. Li. All seven sub-projects are canal irrigation projects in the Deccan Trap portion of Maharashtra. All seven cozmand areas are predominantly vertiuols. All sub-projects are in rather flat yet undulating terrain, necessitating large dams and shallow reservoirs, as well as quite a bit of cut and fill to construct canals and quite a bit of land shaping/levelling to permit good irrigated farming. All are in rainfall-deficient areas, with the western-most sub-projects in more drought-prone areas than more easterly Upper Wardha and Upper Penganga. Except for the modernization sub-projects, none were or were expected to be completed by the end of the five-year time slice. At full development, incremental irrigated command area from the five sub-projects was expected to be about 494,000 ha. benefitting 120,000 farm families directly and generating 140,000 jobs for hired farm labor and 80,000 non-fomu jobs. Costs. Benefits, and Returns iii. The credit became effective in mid-1980, nine months later than anticipated. The project works were not completed by the end 1985 as anticipated, but all credit funds were disbursed by the end of 1986 as originally planned. In that period, costs were 10 more than had been anticipated. However, theae costs did Iot achieve the anticipated objectives. Comparison of physical targets and achievements is complicated by modifications in project design anticipated at appraisal. By comparing physical achievements reported in the PCR with those costed in the SAR and applying weights as per SAR base cost, the weighted  - iv - physical under-achievement was 352. Cost over-run, therefore, was 702. The PCR cites steeper-than-anticipated cost escalation, in part because the appraisal team worked from out-Aated cost estimates. iv. As the PCR argues, it is premature to estimate benefits when, in 1987, less than 10% of the expected ultimate irrigable area is completed and when much of that 10% is in the breaking-in stage. PCR estimates factor in a) delay in benefits because works' completion is delayed, and b) lower irrigation intensity in kharif (rainy season) because the systems cannot deliver crop water needs for more than 50% of the kharif command area. Also, as compared to the SAR, the PCR finds higher without-project yields and lower cropping intensities, lower full-development yields, and a revised cropping pattern that reflects developments since appraisal, notably restrictions on sugarcane areas, development of hot-season groundnut, etc. These adjustments indicate a major reduction in benefits. v. The PCR may be under-estimating full-development yields by basing its projections on results obtained during a period when farmers are learning irrigated farming by doing (since, as the PCR reports, the extension improvements built into the project did not take place). Conversely, the PCR may be over- estimating full-development yields by basing its projections on the experience of farming the head reaches of two sub-projects. Even based on the PCR's optimistic assumptions about system efficiency (5.6 & 7.3), the lower reaches may get their canal water less reliably than the head reaches. A definitive finding would be premature. vi. The most serious of the PCR's findings rexarding potential benefits is that canal caps.city will permit no more than 50X of the command to be irrigated dv'ring kharif, based on optimistic assumptions about system efficiency mentioned above. SAR estimates ranged from 65 to 93% by project. The PCR does not explain the source of the discrepancy between canal sizing and anticipated kharif irrigation intensity. Underestimation of system water losses is probably not the source because PCR estimates were optimistic ones (5.6), though construction quality seems to have been poor (3.21-2). SAR underestimation of crop water requirements could not have caused the discrepancy because the SAR used the same WATREQ model assumptions that the PCR did. Did the SAR over-estimate the extent to which crop water requirements could be met from rainfall during kharif? Or did the PCR, conducted following several years of extraordinary drought, under-estimate them? Did the PCR inadequately take into account the important water-storage capacity of projects' vertisols? For design of any future cana.& irrigation projects in vertisol areas or re-design of old ones, it is important to have answers to these questions.  vii. With all these caveats, the PCR very tentatively estimates the economic rate of return for the five on-ftoinit sub- proiects as wholes at from less than 0-to-7X, compared with the 14-to-18X estimated by the SAR. Neither SAR nor PCR estimate return to the time slice and completion of these sub-projecta, taking whatever was built by 1980 as a sunk cost. That would, of course, have been the operationally-relevant indicator in 1979 -- was it socially worthwhile to finish the five projects? And would it be today? The O-to-6% finding does not answer this question, but it does suggest that these projects are much less attractive investments than supposed in 1979. Lessons Learned and Some Thints That Went Wrong viii. Quality of works, particularly of those contracted out, has been a persistent problem, partially overcome with help from vigorous IDA supervision. 'ID officers have not been resolute enough to reject substandard work, especially in canal lining and do not make contractors strip back and replace unacceptable sections' (3.21). The quality-control organization set up was not effective. ix. The project aimed to replace the traditional Maharashtran system of allocating irrigation water to crops as agreed between users and Irrigation Department (shejpali) with a rotational water-supply system with irrigation water allocated proportional to land holding and with farmers free to crop as they wish to optimize the use of this pre-determined supply (warabandi). This replacement has not yet taken place, except for a few pilot experiments with 'rigid shejpali", a half-way hc.use between the two systems. The PCR agrees with the SAR that warabandi would 'lessen waste and permit more equitable distribution' and "provide a natural incentive toward higher irrigation intensities, particularly during kharif" (3.28). There is no convincing account, however, of why warabandi has been resisted. One possible reason is that, with only a small portion of the commands developed, water is not yet scarce, as it is in all areas where warabandi now works. x. Canal lining and smaller chaks. The PCR finds the SAR azgument convincing that lined distributaries and smaller chaks, though expensive, are cost effective -- that the higher standard is the least cost solution (8.2). Nevertheless, ID, driven by budgetary considerations, has cut back on lining and has reduced chak size less than agreed. The problem of stabilizing unlined canals in vertisols, however, was not solved. Grass sodding in the invert does not work. xi. Training in land and water management at the Aurangabad WALMI has been successful and significant.  - vi - xii. Road construction and maintenance, carried out by ID during the credit disbursement period, reverted to Public Works Department thereafter. xiii. Adaptive aaricultural research at Mahatma Phule Agricultural University did not turn up any effective solutions to farmers' problems of irrigated vertisol farming. xiv. Extension to help farmers adapt to irrigation, using the T&V system and involving both the Agricultural Department and the Command Area Development Authority, was almost totally ineffective (4.6-7). xv. Malaria control worked. xvi. Monitoring & evaluation was ineffective. xvii. Pilot water manacement" programs to determine the optimum level of irrigation services were inconclusive (3.26-7), except that the PCR finds that flanged semi-circular spun RCC pipe of 30 1/s capacity is best for lining field channels in vertisols. xviii. Cost recoverv through water chartes was much lower than either expected or promised. A study showed that volumetric charges were impractical.  - vii - PROJECT COMPLETION REPORT INDIA MAHARASHTRA IRRIGATION II PROJECT (CREDIT 954-IN) OVERVIEW Introduction 1. The Maharashtra Irrigation II Project was approved in October 1979. It includeds - six years (1980 through 1985) of the construction of five gravity irrigation sub-projects already begun earlier (942 of slated costs) (the same 'time slice" of a sixth project was appraised simultaneously but financed separately by International Fund for Agricultural Development); - modernization works on two existing gravity irrigation sub-projects (32 of costs); - a trial pilot scheme to try out higher construction standards on the Jayakwadi gravity irrigation scheme then already being financed by the Bank under Credit 736-IN (1.5Z of costs); and - establishment of the Water and Land Management Institute to supplement the training of engineers and agricultural officers (1.5Z of costs). 2. The five sub-projects had been started between 1968 and 1972. They were at various stages of completion. They were to be substantially completed by 1994, bringing their net cultivable command area of 494,000 ha. fully into production. The appraisal was accompanied by the most thorough state-wide irrigation-sector assessment made up to that time. Obiectives 3. In the broad sense, this 'composite' project, coming on the heels of Credit 736-IN, a 'composite' project focused on the large Jayakwadi gravity command, aimed at improving irrigation planning, project selection, and the standards to which gravity irrigation systems were built and operated throughout Maharashtra State. As part of this effort, prior to and during appraisal, the Bank surveyed the sector state-wide more comprehensively than  - viii - had been done previously in an Indian state. A number of desirable changes were identified: - because farmers would often not construct smaller canals and field channels on their own as expected and, therefore, a lot of canal water was being wasted, have the state system deliver water to 8 ha blocks (chaks) instead of to much larger ones as previously and construct the field channels on behalf of farmers (and recover the expense of doing so through the credit system); - because Bank analysis showed that canal lining, though expensive, saved enough water to make it the least cost solution, line all canals down to the 8-ha chak; - to facilitate operation and maintenance, build canal- side roads and paths and install radio communication links; - because the Bank was convinced of the superior efficiency of the warabandi rotational-water-supply operating system (though most of the Maharashtran irrigation cadre was not), size all the distribution network for operation under a rotational-water-supply system; - because distribution-network works were being carried out and completed in a haphazard way and drainage was generally being neglected, construct the command area works by blocks, including the drainage, and 'commission' thew -- certify their completion before the Irrigation Department handed them over to the Command Area Development Authority; - because of the manifest inadequacies in construction, raise standards of concrete technology, earthwork compaction, cross regulators and other control structures, dewatering structures; - because of the knotty problem of the expansive black soils' popping out canal linings when they get wet, or simply eroding away when canals are unlined so that canals disappear, condur.t experiments on different kinds of canal lininas, their cost and effectiveness, and then adopt standards; and - looking ahead to future demand for more sophisticated irrigation to meet more specialized crop needs, experiment with intermediate storage reservoirs to meet peak demands and volumetric water charges. 4. All of these efforts to improve state-wide irrigation standards, and others to improve agricultural research and extension for black-soil irrigators, to improve markets and access to them, to increase cost recovery from beneficiaries, to  - ix - improve monitoring and evaluation, for instance, were to be tried through the five sub-projects (plus Bhima, the IFAD-financed sub- project), through up-grading the already commissioned Girna and Mula sub-projects, through an experiment with higher standards on a pilot area of the Jayakwadi project, and through supplementary training of engineering and agricultural staff. 5. The specific goals on the five sub-projects, each at different stages of completior when the Credit began, were to complete specific construction targets to agreed standards by 1984 (with disbursement allowed ..ntil end 1985 to accommodate delays). None of the five sub-projects was expected to be completed during the time slice. Major construction was expected to go on until 1987/8 for the Krishna sub-project, to 1990/1 for the Upper Penganga and Warna sub-projects. The area actually served by irrigation canals, 8,000 ha at the time cf appraisal, was expected to rise to 85,500 ha by the end of the disbursement period and to reach 494,000 ha by full development between 1993/4 and 199718, depending on the sub-project. 6. At full development, on the net cultivable command area, cropping intensity was expected to rise from 100-110% to 124-160X, depeuding on the sub-project. Net farm incomes were to rise from three to five times what they would have been without the project thanks to an combination of much higher yields and shifting to more profitable crops. Annual incremental production at full development was estimated at 945,000 T of grains and pulses, 2,337,000 T of sugarcane, and 58,000 T of cotton. The 120,00 farm owning and/or operating families were to benefit both from their increased incomes and from being able to provide an additional 20 million days of work (equivalent to 67,000 full- time jobs or about half a job/family) on their own farms. Moreover, irrigation tnduced incremental demand was expected to create the equivalent of 142,000 full-time jobs for farm laborers and, based on an assumed multiplier effect, 79,000 additional jobs in the non-farm sector. Implementation Experience 7. The PCR finds that 110X of the projected expenditure was made to complete roughly 65% of the projected works, which amounts to a cost over-run of about 70%. (The comparison cannot be precise due to modifications in project content and in Credit allocation.) The Credit was fully disbursed by the original closing date. Since only two thirds of the works were completed by this time, the implementation lag was substantial. Completing the last third of the works, a Government of Maharashtra responsibility, has been made more difficult by budgetary stringencies since 1985 resulting from droughts and other factors, and by the absence of the "carrot" of a follow-on IDA since MCIP IV has been dropped from the lending program.  8. The PCR stresses the difficulty of evaluating a time slice of sub-projects that will be completed only 15 to 20 years after the end of the time slice, "particularly when only few agricultural activities have been undertaken, especially when these have been poorly monitored" (Para 8.10, p. 26). Indeed, very little of the 494,000 irrigable command to be created by the five sub-projects was actually being irrigated, and most of that was in the breaking-in period. It is understandable that the PCR asks that its conclusions "be viewed with great skepticism" and calls upon the Bank to review its practice of commissioning "conventional' PCRs on time slices of long-term programs. Granting and accepting these observations, certain patterns of implementation do emerge from the experience with this project to date, particularly when they are viewed together with the experience of the Jayakwadi Project (Cr. 736 for Maharashtra Composite Irrigation I Project, Cr. 1383/Ln. 2308 for Maharashtra Water Utilization Project, and Cr. 1621 for Maharashtra Composite Irrigation III Project). Construction Performance 9. The PCR concludes that contractual specifications and procedures, as well as construction, improved, and that GOM demonstrated its capacity to implement high-quality major civil engineering works under force account. However, there were major slippages for many reasons, from shortages of cement and fuel, to shortages of funds from GOM, to inability to acquire needed land on time, to disputes with contractors and between contractors and their workers. Block contracting and "commissioning" does not seem to have got very far during implementation. The PCR notes that IDA supervision, which was heavily oriented towards engineering, helped improve the quality of engineering construction, but tnat IDA supervision can only go so far, and 'further improvement of the construction supervision of ID (Irrigation Department) staff is necessary in order to enforce quality standards in contractual work" (Para 8.7) In fact, IDA supervision did concentrate on policing the quality of construction and was criticized within IDA as too heavy, reinforcing the PCR's conclusion on this point. 10. The project was only partly successful in getting higher enaineerina standards adopted in project implementation, either on the sub-projects being financed or in the irrigation sector state-wide. Building systems to deliver to smaller chaks was largely achieved state-wide (but this design parameter has tended to become the standard India-wide too), though the chak siza was reduced less than agreed. Manuals of construction standards were elaborated and agreed by 1985. However, Government has backed away from 'World Bank" standards which it considers "gold plated" in several critical instances, notably canal lining, drainage. and road standards.  - xi - System Operation Performance 11. Firstly, a major objective of Bank policy in this and other recent Indian projects has been to change the predominant operational mode of irrigation systems in most of India to that used in northwest India. In the former (shejpali), which is practiced in Maharashtra, engineers agree with farmers to supply the irrigation water required for a specific "sanctioned" cropping pattern. Because shejpali allows government officials great discretion in water allocation, it is an invitation to collusion between them and better-off farmers. She4pali generally results in concentration of irrigation water on lucrative, high-water-use crops (e.g. sugarcane, rice, bananas & hot-season groundnuts) towards the head end of the canals and/or in hot season, when water is most valuable, with irregular or no water delivery to tail reaches. Shejpali generally results in very complicated, if not impossible, gate operations to deliver the water where and when it is needed for sanctioned crops, and, in fact, in chaotic system operation and extremely inefficient use of the canal water. In the latter (warabandi), the system operates automatically from the distributary down, allowing no discretion to government operators; water delivered is proportional to land commanded, regardless of cropping pattern, and irrigators below the outlet share water by taking turns. Warabandi operation has led to much greater efficiency in canal water use than shejpali, and to much more equitable distribution of water within the commands. 12. Under Maharashtra Composite Irrigation II, the Bank hoped to convince the State's Irrigation Department (ID) to convert to warabandi (see SAR, Para. 3.10, especially footnote 1). All new canals under the five sub-projects were to be sized for warabandi operation. (However, it is a mark of lack of consensus or of confusion that the system was to be gated below the distributary, which would have been pointless under warabandi operation.) Canal sizing is not a matter covered by the PCR, so it is assumed that they were sized for warabandi operation. However, ID officials were not convinced to adopt warabandi-type operation and this has not been done. The last SPN (November 1985) reported that "rigid shejpali' (a sanctioning system where water entitlements are proportional to crops sanctioned, not to area commanded) was being gradually introduced in Mula, Krishna, Kukadi and Upper Pex)ganga sub-projects, but that no form of rotational water supply was being practiced elsewhere. (see discussion below under Findings and Lessons) 13. A second system-operation objective of the Bank was to change the timing of irrigation and, in so doing, the cropping pattern. The appraisal report assumed that the sub-projects would be operated to carry over water through the hot-season in order to be able to give the command area a pre-monsoon soak,  - xii - thereby making it possible to plant the kharif earlier and, consequently, to grow both an irrigated kharif crop and an irrigated rabi crop on the same piece of land. The cropping pattern and irrigated cropping intensity assumed by the SAR are prAdicated on this development, which did not take place. In the instance, there was no water left for a pre-monsoon soak for a variety of reasons, including droughts, and shifts in the cropping pattern, especially to hot-season crops which used up all water left at the end of rabi. 14. Since the system was not operated as the SAR anticipated in these two regards, the agriculture practiced on the irrigated command was quite different from what the SAR anticipated. Agricultural Performance 15. It is too early to evaluate agricultural performance. The PCR has done the best that can be done at this stage basfd on the two sub-projects that are the most advanced: Krishna etd Kukadi. As f that mission in February 1987, there were 33,400 ha of commissioned cultivable command area (73% of the 46,000 ha expected by that time by the SAR) out of 220,000 ha of cca (cf. 206,000 ha foreseen by the SAR) expected to be irrigated under these two sub-projects at full development. The area being examined here is the 15X of the commands closest to the head works and therefore most likely to be well served with irrigation water. (In other words, the other 85% may well not get as much water as reliably as the part from which the extrapolation has been made.) Based on agricultural experience in these areas, the PCR has made courageous assumptions about what is likely to happen over the entire commands from now to full development and beyond. These are the best that can be made now, and it is not likely to be possible to Improve upon them for some time. Here are the conclusions. 1) Because irrigation works are delayed, irrigation benefits are delayed. 2) Even in these areas, which are the head reaches of the systems, the systems cannot deliver crop water needs for more than 502 of the command area in kharif. As a result, the kharif irrigated cropping intensity is lower than the SAR anticipated (though overall kharif cropping intensity is about as expected because more land has continued to be farmed using the rains). Rabi irrigated cropping intensity has been about as anticipated, and hot season (zaid) intensity somewhat higher than anticipated. 3) Ex-post, it appears that, on balance, the SAR over- estimated the yields irrigation would bring and, perhaps more importantly, under-estimated the yields farmers were already getting under rainfed conditions.  - xiii - 4) Actual input prices have been and are expected to continue to be far higher than predicted at appraisal (Table 15) but then so, presumably, are the output prices (PCR assumptions not specified). Taking into consideration the dangers of generalizing from the present situarion, when farmers who have recently received irrigation are very much still learning by doing, it appears that head-reach farmers are concentrating irrigation water on high value crops like hot-season groundouts and sugarcane (to the extent that new restrictions permit them to). The main reasons benefits are falling short of expectations are differences between the predictions on without-project and with-project yields at appraisal and those being made now, as well as the devastating effect of the delay in benefits on their net present worth. It is not clear whether the shift (relative to SAR estimates) to hot-season groundnuts and sugarcane from low-value irrigated crops like sorghum and pulses raises or lowers the economic irr. However, for farmers who pay little for irrigation water, the financial interest in growing these very water-using crops is clear. 16. As regards agricultural services, the PCR finds no difficulty in marketing outputs or in getting inputs (seeds, fertilizers, pesticides) to farmers. (It is known, from studies preparatory to the Narmada Sagar Project, that Girna sub-project cotton farmers truck their cotton to Madhya Pradesh to be exploited by private middlemen instead of selling to the Maharashtra Government monopoly.) Extension to a 'tpt to irrigation, using the T&V system, was not effective (Paras 4.6-7). The same is true of adaptive agricultural research for irrigated vertisol conditions. The PCR does not comment on land- shaping and credit for land-shaping, a problem in the nearby Jayakwadi project. Health Performance 17. The malaria control measures undertaken in connection with the project were effective. This is a striking contrast to simultaneous developments during implementation of Upper Krishna Irrigation Project (Cr. 788), a short distance down-stream from the Krishna and Warna sub-projects in Karnataka, where malaria was introduced and made significant headway during project construction. (This problem is now being addressed under the Upper Krishna II Project, currently being apptaised.) Training Performance 18. The Water and Land Management Institute (WALMI), started at Aurangabad under the project, has flourished and become a model for WALMIs in other Indian states. It is, of course,  - xiv - difficult to measure to impact of training, but the Aurangabad WALMI is highly regarded. It is to be presumed that it has been useful, not just for the project but for the Maharashtra irrigation sector as a whole, in supplementing the normal training of civil engineers and agriculturalists called upon to work in the irrigation sub-sector. Resettlement & Rehabilitation Performance 19. The SAL estimated that the five sub-projects would force resettlement of about 16,000 families (20,000 counting Bhima, the IFAD-financed sub-project) from 282 affected villages. This is a rather large number for the size of the commands. The submergence area equals 11% of the net cca (cf. 25% for Bhima sub-project). As the SAR noted "The relatively large area of submergence in relation to the areas put under irrigation is a result of the flat topography at the storage sites, and also includes land lost to canals." (SAR, Para 3.17) 20. The PCR does not report on resettlement experience. The f:nal IDA SPN report found that resettlement was encountering moderate problems, except for "major problems but being adequately addressed" for the 24 villages being affected by Upper Wardha. Of 132 villages reportedly affected (cf. 282 in SAR), 74 were reported to have been moved and 21 in process. The aide memoire stated that resettlement had been completed where dams were storing water and was in progress elsewhere, but that "there is no monitoring of post settlement status of oustees, and this should be followed up by detailed inspections." Resettlement was an on-going concern of SPN missions, though more often in terms of scheduling to avoid delays in civil works than in terms of adequacy vis-a-vis Bank guidelines. For instance, the November 1984 SPN recommended, "organizing resettlement before end of November 1984 of persons affected by reservoir submergence in Varna Project so that gorge filling car be completed this year and further postponement of construction of dtam can be avoided." The penultimate SPN mission of February-March 1985 reported, "The norms adopted by GOM for resettlement are satisfactory according to the guidelines of the Bank." Rate of Return 21. The SAR calculated rates of return for each of the five sub-projects as a whole, that is from their inceptioas (1968 to 1976) to their expected completions (1987 to 1991) with benefits and recurrent costs extending beyond. The PCR did the same. The SAR reckoned that the economic irr to the resources engaged in the sub-projects ranged from 14Z for Warna to 18% for Upper Wardha and Krishna. The PCR re-estimate, based on projections of the agricultural results from the two most-advanced sub-projects (see caveat under "Agricultural Results" above) and based on the  - xv - most realistic projections of future funding levels by GOM, give economic irrs ranging from less than 0% for Krishna and Warna to 61 for Upper Wardha. ,he SAR average for the five was 15%; the PCR average, 1%. To the extent that PCR estimates can be trusted, they show that the over-all investment in the five sub- projects was not well spent, giving less return than the presumed marginal investment in India -- about 12%. 22. What neither the SAR nor the PCR tell us is whether the investments from 1979 foiward have been worth making. That, after all, was the relevant question in 1979, when, according to the SAR, Rs 1,359 million in 1979 Rupees (1979 $ 158 million) had already been irretrievably sunk into the five sub-projects. Were they worth finishing? Disregarding all costs and benefits before 198011 (and still accepting all other SAR and PCR assumptions), the economic returns t. resources engaged and to be engaged in the five sub-projects from 1980 onwards are as follows: Returns to post-1979 Investments Sub-proiect To Finish Projects SAR Estimates PCR Estimates Krishna 67% 0% Kukadi 26% 5% Upper Penganga 22% 5% Upper Wardha 20% 7% Warna 19% 7% In short, if all costs prior to 1979 are considered as sunk costs, there is a higher return to completing the sub-projects than the rates shown, which are the returns to the sub-projects as wholes. Based on the PCR assumptions, completing the sub- projects does have a positive net social value, but not as high as that presumed to be available on other investments -- 12%. Sustainability 23. Irrigation in India can be eminently sustainable, as the orderly functioning of some systems more than fifteen centuries old demonstrates. The principal menaces to sustainability in this instance are four: 1) filling of reservoirs due to siltation; 2) deterioration of systems due to inadequate maintenance; 3) deterioration of systems due to destruction of canal linings by expansive vertisols and their disappeerance through erosion; and 4) loss of command area to farming due to build-up of salinity and water-logging.  - xvi - 24. Regarding the first, the slopes up-catchment from the reservoirs of the five sub-projects (and of the two modernization sub-projects too) are modorate. Therefore, siltation is moderate. Even though reservoirs are shallow, their useful life should be long. 25. Regarding the second, Bank missions have repeatedly found maintenance to be inadequate in the MCIP II sub-projects, as in most others in Maharashtra, in India, and in the world. However, improvements have been realized. It is reasonable to expect that, once farmers see benefits from the projects, they will see to it that Government somehow maintains the systems or will somehow do it themselves. 26. Regarding the third, there is understandable concern that there may not be a way of lining small distributaries and field channels that is both sustainable and affordable. The Bank has recognized the need to keep looking for the cheapest sustainable way of stabilizing small canals by supporting continuing experimentation in irrigation creditsiloans made since MCIP II, notably MWUP (Cr. 1383/Ln.2308). MCI? III (Cr. 1621), and in Upper Krishna I (Cr. 788). A number of Indian irrigation departments and academic institutions are also working on this problem, with and without Bank support. 27. Regarding the fourth, trends suggested by the PCR are ominous. In black-soil projects from Bargi and Tawa in Madhya Pradesh to Tungabhadra in Andhra Pradesh and Karnataka, failure to allocate irrigation water equitably (coupled with failure to pass on a reasonable portion of the public cost to beneficiaries) has allowed head-reach farmers to concentrate a lot of water on a small part of the command area, generally to grow sugarcane, rice and bananas. This has frequently led to waterlogging ad sometimes to salt build-up. These conditions can be overcome, but it is better all around to avoid them in the first place by encouraging economical irrigation-water use through rationing to produce water scarcity (as in warabandi), or much higher water charges for high-water-use crops (a simple proxy for volumetric charges), or by encouraging ground water development, preferably by some combination of these. The authorities responsible would be well advised to monitor such trends in the MCIP II sub- projects. Findings and Lessons 28. I. Economic worth of vertisol gravity irrigation projects. The MCIP II PCR confirms an impression from other gravity-irrigation projects in the Deccan on vertisols. Given the way these projects are now being constructed and operated, the benefits do not justify the costs. This may not be so for completion of a project where there are substantial sunk costs,  - xvii - but new starts should certainly be avoided until some of the obvioua problems are overcome. 29. The findings of the PCR for MCIP II confirm findings of the appraisal missions for MCIP III (Jayakwadi) in 1984 and of Upper Krishna II in 1988. In both of those cases, as in the MCIP II sub-projects, construction was so substantially delayed and unit costs rose so much above projections that projects expected to give decent rates of return, above but not radically above the presumed opportunity cost of resources, in fact, taken as wholes, do not. In the case of MCIP III and UKP II, given that sunk costs are sunk, appraisal missions found that it paid to build the next section however. (But the MCIP III appraisal mission found that it would not pay to extend the Majlegaon canal in a third phase, even with all earlier phases sunkl) 30. There are a number of reasons why these projects are not spectacular investments from an economic point of view. Terrain is flat, yet undulating. River flow is highly seasonal, so that substantial storage is required. But the best darm sites are not good ones, requiring long, low dams that create shallow reservoirs from which a high proportion of the water left after monsoon evaporates. Since reservoirs are shallow, to store the needed water, they must be large, flooding a lot of the best riverine farmland relative to the size of eventual commands, and forcing the resettlement of a lot of people relative to those benefitted. And since tErrain is undulating, construction of canals is costly, as is land levelling and/or shaping so that canal water can be used. Where the command is narrow besides, as it is in Narmada Sagar and in the third phase of Majlegaon, the ratio of main-canal length to area commanded is low so that, in some listances, the irrigation system is not worth building. (Irrigation tmder Narmada Sagar is justified, up to a point, because the immense power benefits justify dam construction.) Moreover, because vertisols are both expansive and easily transportable, canal lining is essential to maintain and to operate the system, and also very expensive because of the steps that need to be taken to protect lining against soil expansion. Furthermore, the very water-retentive (that ls, hard to drain) vertisols, while very rich, are also very hard to irrigate or even to farm under rainfed conditions during monsoon. 31. In spite of all the above, the evidence suggests that the completion of already-started vertisol gravity-irrigation projects usually give good returns on investment provided they are constructed and commissioned expeditiously and operated more or less according to plan, so that anticipated farming results can be realized. That is what the MCIP II SAR and PCR show. So does Bank analysis of other vertisol gravity-irrigation projects in the Deccan.  - xviii - 32. However, in almost all cases, these results have not been obtained in fact, due to lengthy construction delays with consequent delays in benefits, due to large cost over-runs, and due to system operation and farming on the command radically different from what was anticipated at appraisal. 33. II. Using composite proiects to achieve sectoral aoals. MCIP II was, in a sense, only incidentally about time slices of five sub-projects, etc. It had the broader and, in economic terms, more important objective of changing irrigation sector policy in Maharashtra in many important ways. As the discussion above shows, these ranged through mode of system operation, design and service standards, mode of contracting and of controlling quality of construction, cost recovery, and training the professional engineers and agriculturalists who deal with irrigation. In these matters, the Bank was sure that it knew better ways of doing things. In some instances, as of appraisal, the Bank was sure th:.t its sector work and its earlier dealings with Maharashtra had convinced the relevant decision makers about new standards; canal lining is an example. In other instances, it was assumed that studies to be carried out under MCIP II would furnish the needed proof; volumetric water charges and warabandi operation are examples. 34. During implementation, at management levels, the Bank was sure that its strategy was aucceeding. When the next Maharashtran 'rrigation project was presented to the Board on June 9, 1983, the speech took for granted the success of MCIP II; "...new design and operational standards were introduced for the whole State" the most important of which were "upgrading and lining of minor canals." As the PCR shows, and as a careful reading of SPR reports shows that technicians knew all too well, institutinnal and policy change was quite modest; Maharashtran irrigation policy makers were much less convinced about the system changes the Bank believed in than was assumed. 35. To some extent, the Bank itself is less sure about some of the 'irrigation for the 21st Century" innovations now than it was in 1979. There is less emphasis on volumetric water charges and on small balancing reservoirs to make possible flexible water delivery on demand. The MCIP III appraisal mission of 1985 backed away from earlier insistence on full canal lining and full conversion to warabandi operations. A look at the Jayakwadi project, different parts and aspects of which wre covered under four Bank projects (MCIP I, MCIP II, MWUP, and MCIP III) shows that the Bank itself has been learning from experience and adapting its standards. 36. The Bank's aspirations for 'irrigation for the 21st Century" conflicted with its desire to convert the system to warabandi operation. The desire to be able to deliver water to  - xix - meet the individualized demands of farmers let to experimentation with or acceptable of structures that are inconsistent with warabandi, or which, if not technically inconsistent, would be extremely difficult to operate in order to run the system to warabandi specifications. The many gates are an example. During Implementation, consultants under the Bank-implemented UNDP project had the difficult task of resolving these conflicting aspirations, which was made more difficult by the size of the systems. A certain amount of confusion ensued. This undoubtedly compounded what was probably a strong reluctance on the part of many in ID to convert to warabandi in the first place. 37. Some attempts to chan.e policv have succeeded. The WALMI, to give engineers and agriculturalists specific, practical training on irrigation, is a fixed part of the landscape; the ex- Director of the Aurangabad WALMI is much in demand to help in duplicating the Maharashtrian experience in other states. More canals are being lined. Chaks are smaller. Although the PCR doesn't say so, progress has been made in resettlement and rehabilitation, and Maharashtra's policies are being imitated by other states. 38. Nevertheless, the MCIP II experience, and the Bank's subsequent experience in the Maharashtran irrigation sector, show how hard it is to bring about policy change, even where the policy dialogue is sustained and intense, and where it is backed up with substantial loans, with very specific covenants and pilot trials to bring out the facts where the borrowers was not convinced. Most attempts to get policy and standards changed under MCIP II failed. Though the PCR is circumspect in saying so, construction quality is still poor. State-wide, Maharashtra is still operating its gravity systems under a sanctioning system, and the shejpali is not even more noticeably 'rigid' than before. As part of discussions on "restructuring" MCIP III, Maharashtra is proposing cutting back radically on canal lining, drainage, land-shaping and road investments, all design and policy matters which the Bank thought were settled in 1979. So the policy dialogue continues. 39. III. Unanswered questions and a research agenda. Review of the MCIP II PCR highlights some of our uncertainties regarding gravity irrigation and irrigated farming in Deccan vertisols. 40. Farmers persistently use less irrigation water and do less irrigated farming in kharif *"han anyone anticipated. There are several plausible theories, but no conclusive answers. The PCR suggests that system design is responsible. FAO Cooperative Programme work "reformulating" MCIP III has generated a model that gives the same conclusions. Canal sizing will permit kharif irrigation of less than half the command. There is not enough  - xx - storage to irrigate more kharif and still have enough left for rabi, hot season and perennials. But the results of this quick and rough model are somewhat counterintuitive. As one Bank staff member has commented on the FAO/CP model and its rivals, 'None of the methods adequately represents the moisture carried over to the rabi in these very highly retentive black cotton soils. None were adequately calibrated against known scheme performance." (Albinson-to-Lafourcade of 10 February 1988) All of the models are deficient in taking into account two basic facts of farming in the rabi-dominant, vertisol areas: soil moisture storage and carry-over into dry periods, and deliberate water-stressing of plants by farmers that is economically optimal, not agronomically optimal (as are the modelled water requirements). Missing these critical facts about rabi-dominant, vertisol farming means not knowing whether farmers' reluctance to use canal water in kharif stems from inadequate canal size, inadequate storage, poor system operation resulting in unreliable or no canal water, the difficulty of controlling weeds on deep vertisols because they are so mucky, inadequate land-shaping, farmers' not having learned how to irrigate, adequacy of rainfall, inflexibility of system operation and inability to respond to droughts, the sanctioning and water-charge system, or something else. 41. A properly specified simulation calibrated against known performance would shed light on this and other questions, possibly even answer them definitively. At least three simulations exist or are being created that might helps the SIMYIELD model being constructed by S.G.Narayanamurthy as part of the Rajasthan Canal Project; the Danish Hydraulics Institute's SHE ; and a new simulation being constructed ab initio for MCIP III by Hydraulics Research under a trust-fund arrangement. By applying them creativelv, it should be possible to test some of the hypotheses mentioned above. It s'ould also be possible to improve the quality of our answers to other questions, such as: - what would be the economically optimal allocation of canal water between kharif, rabi, hot season, and perennials, and what incentives would bring farmers' interests into line with that national interest?; - given climate and reservoir configuration, is it better to hold water over for a pre-monsoon soak, thereby losing a lot to evaporation but advancing the kharif and making more- kharif-rabi double cropping possible, or to use the same water in rabi or hot season?; - given canal-water scarcity and historical, uncertain patterns of rainfall, how much irrigation in kharif is economically optimal?; what is the best degree of water- stressing?; what is the minimum irrigation-water slug that farmers can handle and the system can deliver?;  - xxi - - what degree of 'water-spreading' or "water- concentrating is optimal?; should sugarcane, rice and bananas be banned or restricted?; - what are the comparable results of operating these gravity-fed, vertisol systems under shejpali or under warabandi or under different variants of these?; and - what is the optimal sizing of canals and storage dams? Throughout our protracted policy dialogue with Maharashtra and with other Indian states on these issues, the Bank has evolved answers to these questions heuristically. The Bank has, to some extent, allowed those answers, sometimes explicit, sometimes implicit, to evolve in response to increasing experience. Now, there is a convergence of issues arising in the MCIP II PCR with those constantly axising in other vertisol gravity-irrigation projects. The accumulation of experience has improved our ability to specify what has to be modelled. Therefore, there is an opportunity to make simulation useful and to improve answers to the questions that have been posed in the design and implementation of MCIP II, as well as of both its predecessors and successors.  INDIA MAHARASHTRA IRRIGATION II PROJECT (CR.954-IN) PROJECT COMPLETION REPORT I. INTRODUCTION 1.1 Maharashtra is India's third largest State, extending to 308,000 kn,2 and with a population of 64.2 million (1985). Total cultivable area is abDut 18.2 million ha (1985) of which about 2.5 million ha are under ir.-igation. The present irrigation potential is estimated at 6.3 million ha and with improved irrigation efficiencies would increase to 8.9 million ha, including 2.7 million from groundwater sources. The extent of surface irrigation is constrained by the seasonality of the rivers, by topography and by the fact that most of the soils are black basaltic vertisols (black cotton soils) which are difficult to handle under irrigation. Groundwater resources are limited to shallow, low-yielding albeit widespread aquifers. 1.2 Commencing in the early sixties the Bank has supported a substantial share of Maharashtra's irrigation development programme. The first project was the Purna Irrigation Project (CR.23-IN, US$13 million, 1962). This was followed by the Maharashtra Irrigation Project I (CR.736-IN, US$70 million) approved in 1977, which was the first in a series of composite irrigation projects in the State. The project provided support for development in the Jayakwadi area using water from the Godavari river impounded by a dam at Paithan. It further included the upgrading of irrigation and drainage systems in the Purna Project. The project achieved most of its physical targets but, like several other schemes, remained widely under-utilized. This led to the financing of the Maharashtra Water Utilization Project (CR.1383 and LN.2308-IN) in 1983, which was designed to bridge the gap between the speedy "creation of surface irrigation potential" <1> during the 1970s and the slow utilization of this potential. 1.3 The MCIP II <2> project (CR.954-IN) was approved in October 1979 and was to finance a five-year time slice of a composite programme consisting of six new irrigation schemes (Shima, Krishna, Kukadi, Upper Penganga, Upper Wardha and L.arna), and partial modernization work on two existing schemes, Mula and Girna. The Bhima scheme was subsequently taken out of the package and financed by IFAD. c3> The total area commanded by the six new schemes would ultimately comprise 620,000 ha (ICA), all located in the drought-prone Deccan Plateau. The MCIP II was to cost USS452.9 million of which US$210 million was to be firanced by the IDA credit. -------------------------------------- 4l> Construction of irrigation infrastructure with emphasis on major works. '2> t MCIP II is the common abbreviation used for the Maharashtra Irrigation Project II. 43' Separate PCR is available with IFAD.  1.4 This report is based on a review of the SAR, the Project Agreement (PA), the Development Credit Agreement (OCA), IDA Supervision reports, project files, a working paper prepared by GOM, and data compiled by an FAO/CP mission that visited GOM and the project areas in February 1987. II. PROJECT IOENTIFICATION. PREPARATION AND APPRAISAL 2.1 In 1977 the GOI and GOM presented six major on-going surface water irrigation schemes - Bhima (126,000 ha), Krishna (74,000 ha), Kukadi (132,000 ha), Upper Penganga (125,000 ha), Upper Wardha (75,000 ha), and Warna (87,0O0 ha) for IDA financing. Reports for these schemes had been formulated in the 1960s and construction work had been initiated by GOI between 1968 and 1972. Project preparation, involving staff from the Bank, the FAO CP and GOM's project preparation agency (WAPCOS) started in November 1977. The traditional project-by-project evaluation was supplemented in this case by an assessment of the role of the irrigation sector as a whole in Maharashtra's agricultural and economic development context. This sector review took place in late 1978 and was the first in India for major irrigation projects. Specific recommendations for the planning, design, construction and operation of the proposed schemes were developed and set out in a separate Sectoral Background Paper which was incorporated in the SAR as Volume II. The principal findings and recommendations were finalized with GOM during appraisal in February 1979. The recommendations were formalized during negotiation with an assurance that GOM would implement the project works in accordance with agreed planning criteria, and design and construction standards and that it would operate the schemes in accordance with agreed operational criteria. ProJect Components 2.2 The five-year time slice project, as planned at appraisal, was to assist the continued construction of the five major irrigation schemes, Krishna, Kukadi, Upper Wardha, Upper Penganga, and Warna, starting from October 1979, together with the partial modernization of the existing Girna and Hula schemes. Full development of the new schemes was estimated to be reached between 1993 and 1998. -c> The project comprised the following seven components as given in the SAR Vol. I. a) Construction work on 11 dams, construction of about 900 km of main and branch canals together with construction of distribution canals (down to the farmgate); drains to serve an area of 95,000 ha; and about 620 km of new and improved link roads to serve an area. of 85,500 ha in the five new irrigation schemes. b) Modernization of canals and drains in the Mula and Girna Irrigation schemes. This is not specifically indicated in the SAR, but can be gathered from the benefit streams given for the calculation of the economic rate of return (ERR)  c) Strengthening of the CAD organization, including reorganization of agricultural extension and applied research and water managenent services and operation and maintenance on all seven schemes. d) Procurement of vehicles and equipment for project construction and for operation and maintenance. e) Expansion of a GOM pilot water management programme. f) Project monitoring and evaluation. g) Implementation of a training programme for and development and water management specialists. 2.3 Moreover the subsequent on-farm development (OFD) works, as described in the SAR Vol. III <1> are as follows: a) completion of sub-minors and field channels on 76,300 ha; b) lining of sub-minors and field channels down to the eight ha outlet on 55,800 ha of ICA. III. IMPLEMENTATION A. Credit Negotiations and Effectiveness 3.1 A major issue raised by the World Bank during negotiation of the IDA credit, held in August 1979, concerned the proposal of lining all canals down to the eight ha block. It seems that the analysis to Justify this proposition did not take into account that at the initial stage of the canal construction (creation of irrigation potential) water would be relatively abundant. It could therefore be argued that only those canals which would be permanently used needed initial lining while lining of distribution networks which are rotated 7 days on/7 days off could be deferred (at the expense of higher costs) until a later stage when the project would approach full water utilization (para 3.17). Faced with financial constraints the Government, towaras the end of the project, followed this course of action and thereby ignored one of the covenants (para 4.15). 3.2 In spite of the fact that the OCA became effective on July 16, 1980 and GOM made strong efforts to put into motion the necessary institutional arrangements for project implementation, the establishment of the engineering and CADA divisions within the five command areas took at least six months to one year. This was mainly due to a shortage of funds and unavoidable bureaucratic procedures. In the meantime, construction of the dams continued and credit withdrawals for expenditure which occurred after 16 October 1979 were made immediately, which was in line with thr DCA. 1> Not cleatly specified in Vol. I.  - 4 - B. Implementation Schedule 3.3 At appraisal the project's completion was scheduled for 30 September 1984. In the OCA the loan closing date was however extended to 31 December 1985. Although the project closed in time, its physical achievements remained well below SAR targets. The project works under the IDA time slice are now expected to be completed by 1990, i.e. with a delay of five years. However, completion of the overall programme, including CADA works, of the five new schemes is expected to be reached around 2000. Comparison of the SAR estimated completion schedule with p-esent estimates is made in Figures 1 to 5 for the five schemes. 3.4 In order to optimize returns on investments for these large-scale irrigation projects particular care was to be taken in synchronizing implementation of dams, canals, roads and distribution networks. This effort was only partially successful as it was distorted by events (discussed below) that were largely beyond the control of the individual project construction management. Some of the schemes thus show a somewhat skewed development, with considerable progress in dams and main canals but less in distributaries and in command area development. In Uppe, Penganga an encouraging effort was made to bring canal development into line with creation of storage potential by only partial lining <1> of main canals that would be completed when funds permit and by other careful scheduling of works. 3.5 Activities scheduled under the Girna and Mula Irrigation Modernization Schemes are shown in Table 1. At Mula, they have been more or less completed according to plan while at Girna progress has been mixed. C. Implementation of Individual Components <2> Construction of Dams, Canals, Drains and Roads 3.6 The numerous sub-components of the project and the intricate and inconsistent SAR targeting, <3> together with subsequent revisions, preclude an accurate quantification of achievements attributable to the time slice. However, an approximate account of progress on major items is given below and is also summarised in Table 2. 3.7 Oams.The SAR and OCA indicate the construction of 11 dams. Actually according to the project funding only seven dams were to be financed, of which three (Manikdoh, Dimbhe and Simbhora) turned out to be only partially eligible for reimbursement because of already awarded major contracts that were unacceptable to IDA. The degrees of completion of the seven dams at closing date were as follows: two were in service, four were completed for 70% to 95% and one had barely reached 30% completion (Table 2). <1> Not to be confused with selected or deferred lining, this is lining of the bed and toe portion only. <2> Strengthening of CADA, and Monitoring and Evaluation are covered under Chapter IV. '3> Inconsistency between Volume I and III.  -5- 3.8 Major conveyance system. The description of the project in Schedule 2 of the OCA stipulates the construction of about 900 km of main and branch canals, while the SAR Vol. III provides for only 601 km. The actual cumulative length of canal target in the five schemes over the time slice works out at about 660 km, taking note of slight revisions made during implementation in Kukadi and Upper Penganga systems. The overall progress at closing date was about 500 km which is 83% of the SAR target. 3.9 Distribution systems. Against an appraisal target of 88,600 ha of distribution network, about 58,300 ha or 66% of target was achieved. Lining and completion of canal structures is lagging considerably. <1> 3.10 OFO works. These comprise construction of field channels/drains with relevant structures, about 20% of lining, and land shaping, the latter at farmers' expense. An irrigation potential of 76,300 ha (ICA) was to be created by closing date according to SAR Vol. III. The overall progress up to 31 December 1985 has been 67,500 ha or about 88% of target. Lining was completed on only 68% of the targeted 55,800 ha. 3.11 Drains. Soils in the areas where irrigation has started are reasonably well drained with no waterlogging observed so far. GOM therefore has decided to defer investments into drainage systems and to take them up on a selective basis at a later stage. Quantitative targets were not given in the SAR but costs were estimated at 0.0 to 0.4% of total base cost for the schemes. 3.12 Roads. According to appraisal projection, a total of about 600 km of CAD Foaids <2> were to be constructed/improved along with the requisite cross drainage works. During implementation GOM had reduced the length to 552 km in order to restrict the investment on this activity to the original cost as provided in the SAR, although no such restriction was placed on the cost of irrigation components which have escalated by over 70%. The importance of the road sub-component was not fully recognized by ID. It is therefore not surprising that ID's financing of the road programme was discontinued after closing date. Any future improvement or extension of the road network in new project command areas would be planned and implemented under PWD's regular programmes. The road test reaches of the project however, are continuing to be monitored. 3.13 ImPlementation experience. The simultaneous implementation of eight major schemes (including Bhima) at widely apart locations was in itself an important achievement. Regular Bank supervision missions with considerable input of engineering expertise gave much appreciated professional support and led to close cooperation in overcoming numerous technical and other problen.s. Support of the agronomic aspects has, however, been extremely weak. Major factors affecting and delaying implementation are discussed below. -l> The estimate is very rough as all canals are in various stages of construction. <2> It is noted that canal service roads are not included in this component except when serving both substantial public uses and canal service staff. No mention was made on the length of such roads in the SAR.  -6- 3.14 Severe shortages of cement, steel and fuel persisted until 1984 and were a major setback in all construction works. The traditionally critical land acquisition and resettlement problems also turned out to be major retarding factors in several schemes. They were particularly damaging in Kukadi, Warna and Upper Wardha but could have been softened by better planning (para 3.33). Protracted procedures for acquisition of forest lands were delaying at least two schemes. Unexpectedly high price escalations also slowed down the pace of implementation (para 3.37). 3.15 The project was also not spared contractor-induced construction delays (especially Upper Wardha and Warna) and suspension of works due to labor disputes. In the case of poor contractor performance at the masonry dam of Upper Wardha, ID reduced slightly the volume of the contract and awarded that volume of works to a competing contractor which had a salutary effect on the former. Planning and Design Standards 3.16 One major objective of IDA support for the project was to demonstrate the effectiveness of new planning and design standards and to strengthen quality control. This objective was to be reached through application of new technologies, the upgrading of design manuals, the continuous experimentation towards improved technical solutions, especially with regard to black cotton soils and through manpower development. 3.17 Key elements in the upgrading of standards were the lining of canals from the source of supply to the 8 ha outlet and the reduction of chak size from 40 ha and more down to 8 ha. Another criteria was that canal systems should be capable of operating effectively at one half of design capacity. While the latter was adopted in all systems, the concept of lining all canals turned out to be a proposal that GOM was not prepared to follow indiscriminately. First lining of field channels was restricted to 20% of total length, then, towards the end of the project ID introduced selective lining of the distribution network (disnet) below capacities of 100 cusecs (about 3 m3/sec). Lining would be limited to reaches adjacent to structures and to specially exposed or permeable reaches. ID is now considering reducing lining of field channels to as low as 10%. As all disnets would be operated on a rotational basis, the omission of lining in black cotton soils is bound to have negative effects on canal performance (reduced efficiency) and O&M. Furthermore if lining is deferred until the systems have become operational, the cost could increase significantly, as was experienced in the existing Hula and Girna schemes. The stability problem of unlined canals cannot be overcome by resorting to grass sodding of the invert as some ID staff have proposed. In fact the present trend marks a return to the low- standard approach to irrigation that prevailed before MCIP II and which IDA analysts proved to be wasteful in both a physical and economic sense. 3.18 The 8 ha chak concept evolved in other parts of India was also greatly mod fied as this size turned out to be too small for the adopted delivery system <1> and under existing land tenure and other socio-economic conditions. Therefore the new chaks developed ranged from 15 to 25 ha. -